Handbook on Transport Pricing and Financing (Research Handbooks in Transport Studies series) 1800375549, 9781800375543

Citation preview

Handbook on Transport Pricing and Financing

HANDBOOK ON TRANSPORT PRICING AND FINANCING

RESEARCH HANDBOOKS IN TRANSPORT STUDIES This important and timely series brings together critical and thought-provoking contributions on the most pressing topics and issues in transport studies. Comprising specially commissioned chapters from leading academics these comprehensive Research Handbooks feature cuttingedge research, help to define the field, and are written with a global readership in mind. Equally useful as reference tools or high-level introductions to specific topics, issues, methods, innovations, and debates, these Handbooks will be an essential resource for academic researchers and postgraduate students in transport studies and related disciplines. For a full list of Edward Elgar published titles, including the titles in this series, visit our website at www.e-elgar.com.

Handbook on Transport Pricing and Financing Edited by

Alejandro Tirachini Associate Professor, Civil Engineering Department, Universidad de Chile, and Instituto Sistemas Complejos de Ingeniería, Chile

Daniel Hörcher Research Associate, Department of Civil and Environmental Engineering, Imperial College London, UK, and Research Associate, Department of Transport Technology and Economics, Budapest University of Technology and Economics, Hungary

Erik T. Verhoef Professor of Spatial Economics, Department of Spatial Economics, Vrije Universiteit Amsterdam, the Netherlands

RESEARCH HANDBOOKS IN TRANSPORT STUDIES

Cheltenham, UK · Northampton, MA, USA

© Alejandro Tirachini, Daniel Hörcher and Erik T. Verhoef 2023 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Published by Edward Elgar Publishing Limited The Lypiatts 15 Lansdown Road Cheltenham Gloucester GL50 2JA UK Edward Elgar Publishing, Inc. William Pratt House 9 Dewey Court Northampton Massachusetts 01060 USA

A catalogue record for this book is available from the British Library Library of Congress Control Number: 2023934007 This book is available electronically in the Economics subject collection http://dx​.doi​.org​/10​.4337​/9781800375550

ISBN 978 1 80037 554 3 (cased) ISBN 978 1 80037 555 0 (eBook)

EEP BoX

Contents

vii

List of contributors

Introduction to the Handbook on Transport Pricing and Financing 1 Daniel Hörcher, Alejandro Tirachini and Erik T. Verhoef

PART I   THEORETICAL FOUNDATIONS 1

History of transport pricing Roger Vickerman

9

2

Transport pricing: theory and methodologies Achim I. Czerny and Stefanie Peer

24

3

Transport pricing beyond the social optimum Erik T. Verhoef

39

4

Pricing and other instruments for climate change mitigation in private transport Henrik Andersson, Davide Cerruti and Cristian Huse

59

5

Urban form and the pricing of transport and parking Sofia F. Franco

73

6

Equity and distributional issues in transport pricing Christophe Heyndrickx and Inge Mayeres

107

7

The political economy of transport pricing and investment Bruno De Borger and Antonio Russo

124

PART II   TRANSPORT MODES 8

Road pricing and provision of capacity Se-il Mun and Daisuke Fukuda

146

9

Public transport: design, scale, and pricing Sergio Jara-Díaz, Antonio Gschwender and Daniel Hörcher

171

10

From taxis to ride-hailing: market equilibrium analysis and implications for regulations Xiaolei Wang and Fangfang Yuan

11

The economics of airports’ pricing Tiziana D’Alfonso, Martina Gregori, Hugo E. Silva and Leonardo J. Basso v

190 207

vi  Handbook on transport pricing and financing

12

Pricing in freight transport Edoardo Marcucci, Valerio Gatta, Michele Simoni and Ila Maltese

229

13

Connected and automated vehicles: effects on pricing César Núñez and Alejandro Tirachini

252

PART III   TRANSPORT FUNDING AND FINANCING 14

Transport funding and financing: a conceptual overview of theory and practice José Manuel Vassallo and Laura Garrido

273

15

Investment appraisal: links between finance and economics Georgina Santos, Iven Stead and Tom Worsley

295

16

The regulation of public–private partnerships Eduardo Engel, Ronald Fischer and Alexander Galetovic

311

17

Financing sustainable transport infrastructure in emerging markets and developing economies José C. Carbajo

18

Transport financing and regional development Javier Asensio and Anna Matas

330 348

PART IV   REGIONAL PERSPECTIVES 19

Road transport pricing and financing in Africa Leonard Mwesigwa, Moez Kilani and Matti Siemiatycki

364

20

A review of selected transport pricing, funding and financing issues in Asia Wei Liu, Fangni Zhang, Xiaolei Wang and Yili Tang

380

21

Transport pricing in Europe Chris Nash and Heike Link

394

22

Pricing urban transport in Latin America Andrés Gómez-Lobo and Tomás Serebrisky

417

23

Road pricing applications in North America Mark Burris, John Brady and Sruthi Ashraf

436

24

Transport pricing and financing in Oceania John Stanley and David A. Hensher

452

Index

472

List of contributors

Henrik Andersson is Associate Professor at the Toulouse School of Economics, University of Toulouse Capitole, Toulouse, France Javier Asensio is Associate Professor at the Department of Applied Economics, Universitat Autònoma de Barcelona (UAB), and Institut d’Economia de Barcelona (IEB), Spain Sruthi Ashraf is a Senior Consultant working with Traffic Operations Strategy Business line of WSP USA Leonardo J. Basso is Professor at the Civil Engineering Department, Universidad de Chile, and Director of Instituto Sistemas Complejos de Ingeniería (ISCI), Chile John Brady is Head of Traffic and Analytics at Cintra USA Mark Burris is Herbert D. Kelleher Professor at the Department of Civil and Environmental Engineering, Texas A&M University, USA José C. Carbajo is a Director in the Independent Evaluation Group, The World Bank Group (until October 2021) Davide Cerruti is a Postdoctoral Researcher at the Centre for Energy Policy and Economics, ETH Zürich, Zürich, Switzerland Achim I. Czerny is Associate Professor at the Department of Logistics and Maritime Studies, The Hong Kong Polytechnic University, Hong Kong, China Tiziana D’Alfonso is Associate Professor at the Department of Computer, Control and Management Engineering, Sapienza Università di Roma, Italy Bruno De Borger is Professor at the Department of Economics, University of Antwerp, Belgium Eduardo Engel is Professor at the Department of Economics, Universidad de Chile Ronald Fischer is Professor at the Department of Industrial Engineering, Universidad de Chile and Researcher at Instituto Sistemas Complejos de Ingeniería (ISCI), Chile Sofia F. Franco is Assistant Teaching Professor at the Department of Economics, School of Social Sciences, University of California – Irvine, USA Daisuke Fukuda is Professor at the Graduate School of Engineering, The University of Tokyo, Japan Alexander Galetovic was formerly Professor at Universidad Adolfo Ibáñez, Chile, Research Fellow at the Hoover Institution, Stanford University, USA, and Research Associate at CRIEP, Interuniversity Research Centre on Public Economics, Italy

vii

viii  Handbook on transport pricing and financing

Laura Garrido is Assistant Professor at the Transport Research Center (TRANSyT), Universidad Politécnica de Madrid, Spain Valerio Gatta is Professor at the Department of Political Sciences, Roma Tre University, Rome, Italy and the Department of Logistics, Molde University College, Molde, Norway Andrés Gómez-Lobo is Associate Professor at the Department of Economics, Universidad de Chile Martina Gregori is Postdoctoral Researcher at the Department of Computer, Control and Management Engineering, Sapienza Università di Roma, Italy Antonio Gschwender is Lecturer at the Civil Engineering Department, Universidad de Chile, and Researcher at Instituto Sistemas Complejos de Ingeniería (ISCI), Chile David A. Hensher is Professor and Founding Director of the Institute of Transport and Logistics Studies (ITLS), The University of Sydney Business School, Australia Christophe Heyndrickx is Senior Researcher at Transport and Mobility Leuven and Research fellow at KU Leuven, Belgium Daniel Hörcher is Research Associate at the Department of Civil and Environmental Engineering, Imperial College London, UK, and at the Budapest University of Technology and Economics, Hungary Cristian Huse is Professor at the Department of Economics, University of Oldenburg, Germany Sergio Jara-Diaz is Professor at the Civil Engineering Department, Universidad de Chile, and Researcher at Instituto Sistemas Complejos de Ingeniería (ISCI), Chile Moez Kilani is Professor of Economics at Université du Littoral, Côte d’Opale, France, and at Université de Sousse, Tunisia Heike Link is Head of the Transportation Research Team, DIW, Berlin, Germany Wei Liu is Associate Professor at the Department of Aeronautical and Aviation Engineering, The Hong Kong Polytechnic University, Hong Kong, China Ila Maltese is Research Fellow and Lecturer at the Department of Political Sciences, Roma Tre University, Rome, Italy Edoardo Marcucci is Professor at the Department of Political Sciences, Roma Tre University, Rome, Italy and the Department of Logistics, Molde University College, Molde, Norway Anna Matas is Professor at the Department of Applied Economics, Universitat Autònoma de Barcelona (UAB), and Institut d’Economia de Barcelona (IEB), Spain Inge Mayeres is Senior Researcher at Transport and Mobility Leuven and Research fellow at KU Leuven, Belgium Se-il Mun is Professor at the Graduate School of Economics, Kyoto University, Japan

List of contributors 

ix

Leonard Mwesigwa is a PhD candidate at the Department of Geography and Planning, University of Toronto, Canada Chris Nash is Research Professor at the Institute for Transport Studies, University of Leeds, UK, and Visiting Professor at Masaryk University, Brno, Czech Republic César Núñez is a PhD candidate at the Chair of Transportation Systems Engineering, Technical University of Munich (TUM), Germany Stefanie Peer is Associate Professor at the Department for Socioeconomics, Vienna University of Economics and Business, Austria Antonio Russo is Senior Lecturer at the Department of Economics, University of Sheffield, UK Georgina Santos is Reader at the School of Geography and Planning, Cardiff University, Wales, UK Tomás Serebrisky is Principal Economic Advisor at the Inter-American Development Bank Matti Siemiatycki is a Professor at the Department of Geography and Planning, University of Toronto, Canada Hugo E. Silva is Assistant Professor at Instituto de Economía and Departamento de Ingeniería de Transporte y Logística, Pontificia Universidad Católica de Chile, and Researcher at Instituto Sistemas Complejos de Ingeniería (ISCI), Chile Michele Simoni is Assistant Professor at the Division of Transport and Systems Analysis, KTH Royal Institute of Technology, Stockholm, Sweden John Stanley is Adjunct Professor at the Institute of Transport and Logistics Studies (ITLS), The University of Sydney Business School, Australia Iven Stead is Senior Economist at the Department for Transport, UK Yili Tang is Assistant Professor at the Faculty of Engineering and Applied Science, University of Regina, Canada Alejandro Tirachini is Associate Professor at the Civil Engineering Department, Universidad de Chile, and Researcher at Instituto Sistemas Complejos de Ingeniería (ISCI), Chile José Manuel Vassallo is Professor at the Transport Research Center (TRANSyT), Universidad Politécnica de Madrid, Spain Erik T. Verhoef is Professor at the Department of Spatial Economics, Vrije Universiteit Amsterdam, the Netherlands Roger Vickerman is Emeritus Professor at the School of Economics, University of Kent, and Visiting Professor and Chair of the Transport Strategy Centre, Imperial College London, UK Xiaolei Wang is Professor at the School of Economics and Management, Tongji University, Shanghai, P.R. China

x  Handbook on transport pricing and financing

Tom Worsley is Visiting Fellow at the Institute for Transport Studies, University of Leeds, UK Fangfang Yuan is a PhD student at the School of Economics and Management, Tongji University, Shanghai, P.R. China Fangni Zhang is Assistant Professor at the Department of Industrial and Manufacturing Systems Engineering, University of Hong Kong, Hong Kong, China

Introduction to the Handbook on Transport Pricing and Financing Daniel Hörcher, Alejandro Tirachini and Erik T. Verhoef

I.1 TRANSPORT PRICING AND FINANCING IN THE FACE OF MAJOR CONTEMPORARY SOCIETAL CHALLENGES Pricing and financing constitutes one of the main battlefields in the design and implementation of transport policy, and represents what many may associate with transport economics. Indeed, the price of a transport service is the most evident cost when moving around in geographical space, and funding and financing decisions are necessary for achieving the stable operation of both physical infrastructure and the various organised services provided on it. From this point of view, it may seem like pricing and financing are just unavoidable bitter pills to swallow for users as well as decision makers, where only in a fairy tale world, mobility could be ensured without painful monetary payments by any of the parties involved. To refute this perception, our Handbook summarises decades of research on the usefulness of carefully crafted transport pricing and financing schemes in the pursuit of societal objectives. The 24 chapters to follow reflect that pricing is probably the most sophisticated tool of transport service management, potentially ensuring that social and environmental resources are utilised according to the value they truly represent. With careful oversight, pricing policy becomes the key instrument in the control centre of the transport system, from which the volume of interactions within the spatial economy and the (re)distribution of resources among members of society can be fine-tuned in line with higher-level policy objectives. From the viewpoint of funding and financing, the Handbook delivers theoretical foundations and practical lessons on the wider toolbox that complements pricing in achieving financial stability and sustainability for transport projects and services. We believe the Handbook is timely, as transport pricing and financing policies will have an important role to play in the strategies needed to address some of the main contemporary societal challenges.

I.2 CLIMATE CHANGE AND SUSTAINABLE DEVELOPMENT Pricing will play a leading orchestrating role in the battle of humanity against what many of us foresee as the single biggest societal challenge of the century to come: climate change. The substantial share of the transport sector in global energy consumption, pollution, and greenhouse gas emissions is undeniable. Also, there is little doubt about the need for change in habits and everyday human behaviour, notably also in transport, to keep our energy reliance under control. However, there is no sign of agreement on the exact magnitude of the behavioural adjustment needed, and on what share in the global effort should be acceptable, to ensure the sustainability of our future development. Pricing should be one of the key elements in the debate on transport demand management, and a naive belief in a technological fix will 1

2  Handbook on transport pricing and financing

most likely not resolve the existential challenge the planet and its inhabitants face today. We postulate that the implementation of a universal and global system of incentives through transport pricing is a necessary step to take to make every single household realise and understand the level of contribution needed for stabilising energy use in an efficient way. Besides this relatively new challenge that may not have seemed as obvious to so many a few decades ago as it appears to be today, the more classic policy challenges of accessibility and traffic externalities, such as congestion and accidents, will also remain with us. Economic science recognises, and we intend to disseminate this throughout this book, that road congestion and equivalent phenomena in other modes cannot be fully neutralised as we often imagine: congestion is the inevitable consequence of the spatial and temporal concentration of human interactions. The existence of congestion is therefore pretty inevitable, but its magnitude can be controlled with appropriate tools – notably including pricing. We learn from spatial economics that congestion management is to an important extent an art of balancing the benefits and costs of spatiotemporal concentration. This is a particularly tricky endeavour when many benefits and costs are external to the traveller herself; research on agglomeration economies (external benefits) and congestion and crowding burdens (external costs) go hand-in-hand with the aim of identifying the travel volume that we can consider the optimal balance between offsetting these forces. Technological development and new channels of human interaction may shift this optimum one way or another, but the intellectual challenge of travel demand regulation remains a cornerstone of policymaking, and thus also of this Handbook.

I.3 BALANCING NEGATIVE AND POSITIVE EXTERNALITIES OF SPATIAL DENSITY In the early days of the COVID-19 pandemic, billions of urban residents temporarily witnessed a congestion-free state. The pandemic has also shown that disruption in face-to-face human interactions comes at a huge cost in terms of substantial economic downturn, collapsing supply chains, inflation, rising inequality, and tensions within and between societies. Recent data suggests that, as ever since the early days of human history, cities survive, and accordingly also congestion is now gradually rebuilding in metropolitan areas. But this costly natural experiment demonstrated that the adverse effects of transport are better not neutralised simply by completely eliminating motorised transport flows; accessibility is an undeniable ingredient of modern life, and our tools better be more sophisticated when it comes to regulation in mobility. This Handbook documents the most recent findings in the literature of transport pricing and financing, but the underlying theoretical foundations already have decades of history – in fact, as Roger Vickerman highlights in Chapter 1 of this Handbook, more than a century has passed since Pigou’s seminal writings on traffic congestion. Transport economists often feel that their policy recommendations are not fully understood and acknowledged in the political arena or among the wider public. Urban road congestion pricing is a leading example, but pricing and funding in public transport is also heavily debated, and not always in ways that do full justice to fundamental economic insights on these matters. Two hypotheses could explain this observation. First, researchers may need to pay more attention to the (joint) dissemination of research findings, and explain the intuition behind them more clearly. Second, it is very well possible that the research community overlooks or underrates some of the unwanted

Introduction  3

consequences of the proposed policies, which makes them appear unattractive at least for a fraction of end users, whose political influence is substantial. Both conjectures have been recognised by researchers. This Handbook as a dissemination tool represents a major effort of almost 50 authors from around the world with the aim of making academic outputs more accessible for students, practitioners, and early-career researchers. To address the second concern, there is a clearly visible shift in the recent literature to extend traditional normative analyses towards positive investigations of social acceptance and redistributional consequences.

I.4 EQUITY AND DISTRIBUTIONAL TENSIONS IN TRANSPORT POLICY There is a genuine desire in modern societies, in both developed and especially developing countries, to make public interventions contribute to a more equitable outcome – whichever definition one may use to operationalise this. The underlying motivation is clear and hardly questionable: the unequal distribution of wealth, information, access to amenities, and influence in decisionmaking is considered morally intolerable by many, causing tensions within society. On top of that, if transport policies are to be developed via democratic institutions, a strongly unequal distribution of perceived benefits is more likely to prevent the implementation of pricing reforms through majority voting, even if they deliver a net increase in aggregate welfare. Researchers react to these societal challenges by looking more closely at the distributional impact of new policies, identifying winners and losers in the population affected, and predicting the conditions of political acceptance under the surrounding institutional setting, such as elections or referenda. Distributional analyses are sometimes hampered by the lack of a clear understanding of what quantitative measures are better aligned with the perceptions and preferences of individual users and voters. Even if the traditional utilitarian measures of individual welfare were in line with these perceptions, the way in which utility functions are established in a quantitative model might have profound implications for projected distributional outcomes. For example, low accessibility in locations where low-income households live is often perceived as unfair, suggesting that reducing the price of transport service provision or improving transport infrastructure might be a progressive policy in such areas. However, when one translates this setup into a general equilibrium model in which transport and housing costs are interlinked, the improvement in accessibility can easily make the local area more attractive for richer households as well, thus leading to rising housing costs and reducing the intended distributional effects of this intervention. Spatial heterogeneity in access to mobility is often just a consequence instead of the cause of inequality. Therefore, naive equity-oriented transport policies may not be as effective as direct redistributional interventions targeting the original cause of income inequality, including differentiated income taxes. The aggregation of individuals or households into income segments, as is usually done in distributional analyses of transport policy, can also be problematic, as shown in Chapter 6. When segmenting into income quintiles, within each quintile there are households who are affected positively and negatively by a pricing policy, as the intensity of car use may vary substantially between households with similar incomes. Therefore, aggregating over income to analyse the average impact per income quintile hides these divergent effects. Demand management through pricing incentives is often perceived as an unfair practice simply due to the higher marginal utility of income and the higher price elasticity of

4  Handbook on transport pricing and financing

low-income households, both triggered by the tighter budget constraints they face. More generally, monetary payments seem to be more salient from an equity perspective than other sources of travel inconvenience, such as time loss, schedule delay and crowding. To the extent that people’s perception is biased in this sense, education and dissemination might be an effective way to inform the wider public about the non-pecuniary benefits of pricing, and on the benefits that may arise from the use of revenues. These are key issues for the acceptability of almost any transport pricing reform, because it is often neglected that a pricing policy does not end at the moment when the tolls or fares are paid by the user; recycling revenues from transport pricing enables the implementation of almost any redistributional pattern as part of a wider policy package combined with tax cuts for low-income households, as well as targeted transfers and subsidies. Finally, with the rapid spread of advanced technological solutions for pricing in every mode of transport, price discrimination based on equity considerations is an available but mostly unlocked possibility in practice. This Handbook is meant to showcase that pricing is too effective as an incentive-based regulatory tool to be sacrificed in response to platitudinous claims about its fairness in general.

I.5 THE RISE OF THE SHARING ECONOMY AND NOVEL PRICESETTING MECHANISMS The past decade has brought about substantial technological innovations in (usually) deregulated transport markets, such as ride-hailing (also called ride-sourcing in the academic literature) and car sharing. Pricing is an important component in the wider trend frequently described as the sharing economy. A noticeable demonstration of the power of time-varied pricing and online matching could be observed in the ride-hailing industry, where new entrants became dominant over traditional taxi service providers. The price of hailing a car using smartphone applications reflects the instantaneous cost of operating the service and the occupancy rate of available vehicles, with the application of so-called ‘surge pricing’ when there is a temporary mismatch between demand and availability of drivers. It is very likely that vehicle sharing will become even more popular in the future, especially if automated private cars overcome the barriers that currently impede their proliferation. Understanding competition in this emerging sharing economy market generates new challenges from the regulatory point of view, given that (i) tight competition could make price-setting mechanisms extremely volatile on the one hand, whilst (ii) identifying market power and optimising market power are complex tasks in the presence of economies of network size and fluctuating prices. Interestingly, despite the high efficiency and general popularity of ride-hailing companies (Transportation Network Companies in the United States), simplicity in pricing is still a frequently voiced expectation in the case of publicly provided transport services, including road tolling and fare setting in public transport. Public opinion rarely distinguishes revenue-generating motivations from the aim of efficiency enhancement behind time-varying pricing strategies. The Handbook explains that price differentiation between peak and off-peak periods or network segments of different demand intensities can be more important than identifying the ideal average toll or fare level. In other words, tariff structures are at least as important as tariff levels, and a departure from pre-existing flat (undifferentiated) or subscription-based pricing traditions can be more impactful than a marginal adjustment in the cost recovery ratio of

Introduction  5

transport services. Contactless payment methods and other emerging technological solutions are paving the way for a wide range of feasible ticketing solutions – it is up to the transport economist to make the necessary contributions to the development of new algorithms for fair and efficient resource allocation.

I.6 THE FORESEEN UPTAKE OF AUTOMATION, ELECTRIFICATION, AND SHIFTING MODAL PRIORITIES The electrification and automation of road vehicles will not change the fact that pricing is among the most pressing issues in transport policy, with important questions already emerging for the near future. Traditional fuel duties and vehicle excise duties together constitute a substantial source of tax revenue for local and national governments. The transition to a ‘net zero’ transport sector will create an imminent funding gap in many countries, heavily affecting the financial sustainability of infrastructure development and maintenance, and social programmes outside the transport sector. At the same time, fuel duties are among the most efficient environmental taxes, given their proportionality with the amount of fuel consumed. Due to the financial pressure, it is becoming more likely that novel (electronic) road pricing instruments will replace fuel duties, thus ensuring that electric vehicle owners also pay their fair share of infrastructure provision and other uses that governments have for the current fuel tax revenues. The early signs of such developments are already visible in several European countries with a growing penetration rate of private electric vehicles, including the United Kingdom, Norway, Denmark, and the Netherlands, and proposals for a per-km tax on electric cars in countries such as Australia (Chapter 24). Electronic road pricing had been promoted by transport economists since the infancy of the discipline, and it now seems it may be an external financial pressure that will finally turn this policy proposal into reality. Nevertheless, it is perhaps too early to declare as certain the adoption of road pricing, if only because a poorly designed road pricing mechanism could cause more harm than good compared to a simple tax linked to fuel or electricity consumption. Amid intensifying debates on the ideal electronic road pricing scheme, we have encountered fierce proponents of flat road tolls, which would enable even less differentiation in the price of road use than what fuel duties achieved so far through their proportionality with CO2 emissions. In such debates, a proper understanding of the principles of road externality pricing is more important than ever. We trust the readers of this Handbook will find munition on the forthcoming pages to enter this dialogue with sufficient knowledge on what an efficient road pricing system would look like, and what could make it acceptable for the majority of society as well. Whilst transitional changes are eagerly anticipated in private car use, enthusiasm around public transport remains far less pronounced. This is partly due to the aftermath of the COVID-19 pandemic, which had a huge impact on the public image of public transport. We are much less pessimistic about the future of bus and rail-based modes. First, electrification in bus transport is already a reality and in the situations that allow for automated driverless bus operation, recent findings suggest that automated bus networks should be denser, more frequent, and less reliant on public subsidies than how human-driven buses operate today (see Chapter 13 in the Handbook). Second, in the age of mounting energy prices and geopolitical tensions in the global energy sector, it is no longer just the environmental emission that

6  Handbook on transport pricing and financing

matters in transport, but also the absolute level of energy consumption, no matter how our vehicles are propelled. Railways and other forms of mass public transport feature unparalleled energy efficiency within and between dense urban areas. It is unlikely that technological innovation will ever neutralise scale economies in the capacity of vehicle operations.

I.7 THE HANDBOOK’S OBJECTIVES This Handbook’s primary objective is to explain the economic theory of pricing, including the most recent contributions in the field, and to transform research findings into relevant policy recommendations. Although the scope we cover in 24 chapters is narrower than in some of the well-known more general handbooks of our field (e.g. A Handbook of Transport Economics by de Palma, Lindsey, Quinet, and Vickerman), we provide a wide range of detail and analysis of pricing theory and financing that has not been covered in earlier transport economics handbooks. Most of the forthcoming chapters follow a welfare economic approach to transport pricing. That is, our initial standpoint is that the price of a transport service is a key decision variable in the process of policy optimisation, where a public decision-maker’s objective is to do well according to some predefined social welfare function – and in theory would aim to optimise it. Naturally, this exercise becomes more complicated in a liberalised transport market where the regulator has no direct control over prices set by profit-oriented companies. Also, we do recognise that welfare economics is not a consensual toolbox for policy optimisation in the transport sector. There might be many other driving forces behind pricing-related policies, e.g. strictly political, equity-oriented, or ecological ones. Several clearly expressed and theoretically underpinned policy recommendations of the transport economics community have remained neglected in the policy arena. Therefore, the book presents a critical assessment of the practical applicability and social/political acceptance of the outcomes of theoretical analyses. Due to the narrowly defined welfare economic focus of several pricing studies in the past, the financial aspects of transport pricing have received less attention in the literature on transport economics. This is in contrast to industrial practice, where pricing is often understood as a tool for ensuring financing stability, neglecting its impact on transport demand and resource allocation. The Handbook recognises that pricing and funding/financing are closely related and intertwined subjects in transport policy, where the role of pricing goes way beyond fulfilling the financial requirements of project implementation and service provision. In doing so, the book addresses a critical disharmony in the transport sector between the economic and financial definitions of pricing. The book devotes ample attention to new technologies, thus providing another important contribution in comparison with earlier publications. First, new developments will be discussed in relation to digital payment technologies, information provision, and the pricing techniques they now allow for (e.g. price differentiation and dynamic pricing). Second, we address the challenges and opportunities caused by the appearance of new transport technologies (i.e. new modes), such as ride-hailing and automated vehicles. This way the handbook reflects on the quickly evolving landscape of the urban transport scene and becomes an up-to-date source of information on hot topics in the policy arena.

Introduction  7

I.8 BOOK STRUCTURE The Handbook’s 24 chapters are structured into four parts. Part I provides comprehensive coverage of the theoretical foundations of transport pricing. The journey begins in Chapter 1 with a unique overview of the history of thought behind the discipline. The microeconomic principles of transport pricing are then introduced in two chapters, focusing first on what we call ‘first-best’ in welfare economics and then extending this scope to alternative policy objectives and various constraints on welfare maximisation. In the rest of Part I, we cover cross-cutting aspects of transport pricing that remain applicable across various modes of transport and geographical areas: these include climate change (Chapter 4), general equilibrium implications and urban form (Chapter 5), equity and distributional issues (Chapter 6), and the political economy of transport pricing and investment in Chapter 7. After the general theoretical discussions, we guide the reader’s attention to mode-specific models of transport pricing in Part II of the book. Our aim here is to highlight the unique features of traditional and new transport technologies and the most relevant implications in terms of policy interventions. For example, Chapter 8 covers a very broad literature on road congestion pricing, Chapter 9 highlights the economic rationale behind subsidies in public transport. This part pays special attention to emerging technologies in the transport sector, such as ride-hailing (covered in Chapter 10) and connected and automated vehicles (Chapter 13). We believe that these subjects will draw considerable attention among Handbook readers in the upcoming years. The book covers transport financing in the five dedicated chapters of Part III. The methodological backbone of the related chapters is public and corporate finance. In other words, this part differs from earlier chapters in the sense that economic efficiency and welfare economics in general do not play a leading role here. The foundational concepts of transport financing, including the distinction between funding and narrowly defined financing, are introduced in a unique contribution in Chapter 14. Chapter 15 focuses on investment appraisal, where the authors identify a crucial connection between transport pricing and project financing. Chapter 16 is devoted to public–private partnerships in the context of infrastructure development, whilst Chapters 17 and 18 present global and local approaches to transport financing, respectively. Finally, the six condensed chapters in Part IV of the Handbook review and discuss a variety of regional specificities of Africa, Asia, Europe, Latin America, North America, and Oceania. The role of the regional overview is again to take the theoretical insights of transport pricing and financing closer to practical applications by focusing on local achievements and challenges of the future. We believe that the Handbook is valuable for a wide range of readers, including undergraduate and postgraduate students, early-career researchers, senior researchers, practitioners, and anyone with an interest in transport pricing, financing and policy. We hope you enjoy reading it. Daniel Hörcher, Alejandro Tirachini and Erik T. Verhoef May 2022

PART I THEORETICAL FOUNDATIONS

1. History of transport pricing Roger Vickerman

1.1 INTRODUCTION To understand why transport pricing is in any way different from pricing more generally in economics requires consideration of the specific characteristics of transport as a good or service. Transport as an activity cannot be stored and is directly consumed as it is produced. Hence the price of a transport service has both a spatial and a temporal dimension; it relates to the interaction of demand and supply at a specific location and at a specific point in time. The production of a transport service typically requires the use of infrastructures such as a road or rail network or a port or airport and the provision of vehicles to use this infrastructure. The infrastructure is large and can typically provide a service to many competing users. The size of the infrastructure can lead to a natural monopoly, and in some cases, especially rail, this has led to vertical integration of the infrastructure and service. Unbundling this vertical integration has often been seen as necessary to remove the inefficiencies from hidden transaction costs. The existence of scale economies in the provision of transport services on a given infrastructure is less obvious, but there may be economies of scope and network economies. Direct competition between different operators is possible as long as there is some transparent mechanism for the allocation of rights (slots) to use the relevant network. It is these specific characteristics of transport that has fascinated some of the founders of modern economics as a discipline and have continued to provoke interest in successive generations of economists. It requires the definition of marginal costs for a fixed infrastructure and a variable service; and whether the short- or long-run marginal cost is more relevant. It requires an analysis of the potential for monopoly and an appropriate balance between potential scale and scope economies and the risks of monopolistic exploitation. It requires an assessment of the need for regulation against the possible benefits of allowing free competition. It also increasingly requires a consideration of the external costs of transport as a major contributor to greenhouse gases, atmospheric pollution, and the damage from noise and vibration, as well as the costs of accidents and risk to life. In this chapter, we highlight some of the major themes that have run through the works of successive generations of economists as they have sought to reconcile the theoretical firstbest solution with the practicalities of fixing prices that are both efficient and acceptable, not least in the context of the politics of transport policy. The chapter does not set out to provide a comprehensive bibliography of the subject, nor to give a detailed review of the implementation of pricing by transport operators or public authorities, but rather to provide a framework for understanding both the historical and analytical developments which are covered in detail in subsequent chapters. We start with a brief timeline of historical developments in Section 1.2. Section 1.3 then deals with the development of pricing in practice. Section 1.4 looks specifically at the core question of paying for roads and congestion pricing. Section 1.5 9

10  Handbook on transport pricing and financing

examines the role of the public sector in meeting public service obligations and in regulation. Section 1.6 explores issues in implementing ideal prices before Section 1.7 offers some concluding thoughts.

1.2 A TIMELINE OF HISTORICAL DEVELOPMENTS 1.2.1 Nineteenth-Century Origins Many would now regard the earliest formal statement of the basic issue of pricing in transport to be that of Jules Dupuit (1844, 1849). Dupuit, however, acknowledged the contributions of Say and Ricardo in the development of his thinking as he saw the pricing issue as deriving from an understanding of utility, and thus placed the issue firmly in what has become known as welfare economics (see, for example, the review in Ekelund and Hébert, 2002). Dupuit, however, rejected the basic classical assumption in Say that assumed that through perfect markets the price of a good would reflect exactly its value and hence utility to each consumer. He saw that each individual would have a maximum price that they would be prepared to pay, and this defined an implicit marginal utility curve. Utility was not therefore something that could be uniquely determined by a market price. As Rothengatter (2018) has suggested, this was later to cause confusion with the neoclassical analysis of marginal utility in Marshall (1920). Dupuit’s contribution was also not an abstract piece of theorising; his concern was the eminently practical one of how to define and measure the public utility (utilité publique) of public works as defined in an administrative law of 1841 (see Bonnafous and Crozet, 2018). The main thrust of this was to identify the sum of the utility derived by users. There are two main points to take from Dupuit’s arguments. First is the best-known observation that the optimal tariff to be charged for the use of an uncongested bridge would be zero. This was put forward in his 1844 paper but, following some criticism based on a misunderstanding of the argument, was reiterated in his 1849 paper which had the more focused title of “On the influence of tolls on the utility of means of transport”. The second, and often overlooked, point is that Dupuit focused on the effect of the delivered price of goods from an improvement in transport (in the example, it is a new canal). In this way, he can be distinguished from the later focus of Marshall and Pigou as he is essentially dealing with a general equilibrium solution rather than the partial equilibrium of “the marginalists” (as identified by Rothengatter, 2018). The price of transport per se is of less interest than the effect on the prices of goods. 1.2.2 The Marshallian Approach Marshall was writing some 50 years after Dupuit and was clearly aware of Dupuit’s work, albeit in a couple of footnotes (Marshall, 1920 pp. 85, 394). Marshall focused on the importance of the idea of “increments of utility” and clearly drew parallels between his graphical representation of marginal utility and the diagrams used by Dupuit. It is from this that the common belief that Dupuit and Marshall were talking about exactly the same concept arose, or that Marshall simply took Dupuit’s idea, renamed it as consumers’ surplus and conveniently published it in English. In Marshall’s view, the marginal unit of a good would be sold at a price which exactly reflected the marginal utility to the consumer, but intra-marginal units would generate an implicit surplus as the consumer would have been prepared to pay a higher

History of transport pricing  11

price. Marshall’s influence on the development of economics through the use of the partial equilibrium analysis could be argued to have created a barrier to a wider understanding of the problem initially posed by Dupuit. 1.2.3 The Contribution of Pigou This dichotomy was reinforced by the work of Pigou (1920, 1932), Marshall’s chosen successor as Professor of Political Economy at Cambridge. Pigou is often regarded as the one who set out the basic economics of transport pricing as an element in the foundations of welfare economics. Whilst Marshall focused on the way that variations in competition in the market would affect the price and hence the consumers’ surplus, Pigou’s principal focus was on the deviations between private and social net product. This divergence was another way of looking at the impact of various forms of competition, but it went further in that it also applied to differences in the ownership of resources and hence to situations where one person’s production or consumption of a good or resource had consequences, which could be positive or negative, for another’s consumption or production. Pigou’s concern was how to modify the behaviour of an individual or firm, which by pursuing its interest in maximising private net product, would be unconcerned with the social net product, which would maximise social welfare. This led to the suggestion of imposing a tax, which became known as a Pigouvian tax, on the difference between the implicit marginal private cost and the relevant marginal social cost. This would make those imposing what became known as externalities on others recognise the true costs of their production or consumption and hence modify their behaviour towards achieving a social optimum. The Pigouvian tax has become the mainstay of work on optimal transport pricing, although Pigou was more concerned, as was Marshall, with the implications for controlling monopoly and subsequently understanding the issue of imperfect or monopolistic competition, which had become a major topic in the 1930s and is picked up in the Appendix to a later reprint of the 1932 fourth edition (Pigou, 1952). Interestingly Pigou only makes passing reference to the question of charging for roads (1920, p. 193) in the context of charging fuel duty and vehicle licence tax as an example of payment for the imposition of costs. He remarked in a footnote that such funds were in the United Kingdom at that time only supposed to be devoted to the construction of new roads and not, as he argued they should be, to the maintenance and repair of roads damaged by the use of vehicles. In the first edition of the book (Pigou, 1920, p. 194), he proceeded to follow this discussion with an example of two roads where applying a differential tax could reallocate traffic more efficiently. This is often taken to represent the first statement of a theory of congestion pricing. Pigou’s example was challenged by Knight (1924), and Pigou dropped the model from subsequent editions (e.g. Pigou, 1932). McDonald (2013) shows that Pigou was mistaken in trying to use the two-road model to illustrate the case for a tax on the output of industries subject to decreasing returns, but that the basic principle in justifying congestion pricing was picked up by later writers such as Walters (1954) and Beckmann et al. (1956). Pigou’s larger transport example was that of “the special problem of railway rates”, to which he devoted a whole chapter (Pigou, 1932, Chapter XVIII). In this, Pigou focused on the debate over using a “cost-of-service principle” or a “value-of-service principle” to set railway rates. This hinges on the basic characteristic of transport referred to above, that it is consumed as it is produced with no possibility of storage or resale. Relying on a cost-of-service principle raises the problem of discrimination between those for whom a service can be provided at

12  Handbook on transport pricing and financing

low cost and those, because of either the difficulties of construction or the low density of population, where it can only be provided at a high cost. Defining cost-of-service involves the double problem of allocating fixed (sunk) costs and dealing with joint costs. A value of service principle implied price discrimination on the basis of willingness to pay, in extreme perfect price discrimination as would be charged by a discriminating monopoly (Pigou, 1932, p. 290), thus expropriating all the consumers’ surplus. Both of these issues are found in many common current issues in pricing, such as peak-load pricing and yield-management pricing used by airlines and rail operators to vary seat prices in real time (see Chapter 3 for a deeper coverage of the theory of profit-oriented pricing techniques). In his discussion of railway rates, Pigou was clear that perfect use of either cost-of-service or value-of-service, by setting rates that exactly reflected the costs of provision or the utility derived, would pose problems. However, he also considered that, using parallels with the electronic metering used in electricity distribution, it may be possible to get closer to a solution to this problem. This anticipates both yield management pricing and the use of electronic road pricing. 1.2.4 An Alternative Tradition Whereas much of the subsequent theories of transport pricing are grounded firmly in this marginalist tradition of Marshall and Pigou, it must not be forgotten that there were other approaches to the fundamental questions of economics. The Austrian tradition, as outlined in the works of Böhm-Bawerk (1891) and Wicksell (1901), focused on the process of production through time and the need to recover the cost of capital incurred. Hicks (1973, p. 12) referred to this as “the typical businessman’s viewpoint, nowadays the accountant’s viewpoint, in the old days the merchant’s viewpoint”. This underlies a focus on the pricing of transport to cover the cost of the capital employed and hence a cost recovery approach to pricing.

1.3 PRICING IN PRACTICE 1.3.1 Marginal-Cost Pricing Identifying the marginal cost of providing a specific transport service and its further division into private and social elements became, following Pigou, the dominant theme of contributions to the debate. Initially, this was how to relate price to marginal cost in such a way as to maximise social welfare in the presence of a balanced budget constraint. The problem arises because transport, and particularly transport infrastructure, may display falling long-run average costs due to scale economies such that marginal cost will be below average cost and setting price equal to marginal cost will result in a deficit (Chapter 9 shows that this principle governs much of the literature of pricing in public transport, for instance). The problem of setting taxes on different goods in such a situation was first examined by Ramsey (1927), a problem that Ramsey indicates was suggested by Pigou. However, the general presumption of decreasing costs (increasing returns) and the existence of externalities was challenged by Knight (1924). Although Ramsey does refer again to the question of motor taxes (p. 59), the more direct application to transport was further developed by Boiteux (1956). So-called Ramsey–Boiteux pricing adjusts the price–cost margin by the relevant price elasticity of demand, and this has

History of transport pricing  13

become a key result for price setting in public transport. The problem identified by Boiteux (1956) and implicit in the earlier discussions was formalised by Baumol and Bradford (1970). This raises the question of how to implement such pricing when the public authority faces a budget constraint. The issue of transport pricing in practice was examined further by Hotelling (1938), formalised by Beckmann et al. (1956) and that of road user charges, in particular by Walters (1968) and Newbery (1988a, b). The formal development of the theory of marginal-cost pricing applied to transport, following the seminal work of Beckmann et al. (1956), occurred in the 1960s as the pressure for dealing with the growing problem of congestion and the need to value transport investments in developing economies became stronger. Walters (1961) gave the first presentation of the main concept of pricing according to increasing short-run social marginal costs, reinforced by Mohring and Harwitz (1962), and Vickrey (1963) extended the analysis to a more general analysis of urban transportation. The question is then whether the relevant reference for pricing is the short-run or longrun marginal cost and hence whether the primary goal is short-run allocation or long-run provision of optimal capacity (see, for example, Vickrey, 1948). This leads to a debate about the extent to which any optimal pricing system can recover the costs of providing transport without resorting to the need for subsidy (see Section 1.5 below). A cost recovery approach contrasts with the optimal pricing approach and owes more, as we have seen above, to the Austrian tradition than to the marginalist tradition rooted in utility. Following the contributions of Beckmann et al. (1956) and Walters (1954, 1961, 1968), the theory of congestion pricing has dominated much transport research over subsequent decades. This has looked to build beyond the essentially static model of identifying the first-best social optimum implied in the Pigouvian tax to models of dynamic congestion and the search for second-best solutions. Arnott et al. (1990, 1998) discuss the idea of bottleneck congestion in networks and Verhoef (2001, 2003) explores the concept of hypercongestion in which there is no self-adjustment to a stable equilibrium. Verhoef et al. (1995) and Verhoef (2002) develop second-best solutions. These theoretical foundations are well summarised in Small and Verhoef (2007, pp. 120–148) and Chapters 2 (Peer and Czerny) and 3 (Verhoef) of this Handbook. 1.3.2 External Costs, Pricing, and Optimal Transport Systems Whilst much of the early work had focused on the principle of pricing under conditions of limited capacity and the particular problems of paying for roads and dealing with the natural monopoly problem of railways, attention began to be devoted to how to use pricing to achieve a better balance of mode use. The theory is essentially simple, one of equating (social) marginal costs, but the application of this depended on knowing what those marginal costs were. Considerable effort was therefore expended on identifying the components of such costs. This needed to cover the private costs in situations involving increasing returns to scale and network economies as well as economies of scope where, for example, joint and common costs had to be shared across separable markets such as those for freight and passengers on a rail network. Moves towards privatisation of transport operators often hinged on questions of the existence and extent of potential sources of increasing returns from large operators. At the same time, the implementation of pricing to achieve a socially more efficient transport system depended on identifying the external costs of each mode.

14  Handbook on transport pricing and financing

The external costs of transport have assumed a much larger role in recent years. The issues that dominated in the late nineteenth century and early years of the twentieth have been replaced to a large extent by the contribution that transport makes to environmental degradation through its carbon footprint and emissions. Increasingly this has been about more than just the direct impacts of individual journeys but also the external cost of constructing infrastructure to take increasing volumes of traffic and the sustainability of current and expected levels of mobility. Pricing clearly has a major role to play in this debate, both in terms of reflecting the true cost of mobility in each mode and in attempting to influence behaviour in terms of both levels of trip generation and mode choice. Valuing the external costs of transport is not just a simple matter of identifying the source of these externalities and applying a market price. Whilst there are well-developed methods for obtaining estimates (see, for example, Rizzi and Ortúzar, 2015), getting consensus on the external costs of transport has not proved simple. Political considerations weigh heavily on this according to the impacts that any increases in the price of transport might have on voters and on other sectors of the economy, such as the automobile or aircraft production industries or in those countries with significant oil production. Some of the differences between the United States and Europe can be seen in Delucchi and McCubbin (2011)1 and Friedrich and Quinet (2011), whilst the latter demonstrates the compromises necessary to achieve a consensus view in European discussions. These reflect the differing preferences and perceptions of consumers, plus the differences in the spatial organisation of the economy and the consequent pattern of demand for transport. Andersson, Cerruti and Huse cover the most recent features of this literature in Chapter 4, whilst Nash and Link (Chapter 21) document the evolution of related policies in the European Union. 1.3.3 Price Discrimination The nature of the demand and supply of transport services makes it a potentially important area for the use of price discrimination (for a useful review, see Anderson and Renault, 2011). This was recognised from an early stage and features strongly in Dupuit’s (1849) original analysis where he seeks to build up the demand curve from the maximum amount that users are prepared to pay. Pigou (1932) also addressed the issue from the perspective of the impact on consumers’ welfare. Whilst cases of pure first-degree price discrimination,2 where each individual user is made to pay the price equal to their own willingness to pay, are rare, the use of yield-maximising booking systems on planes and trains starts to get close to this. Here prices vary in real time according to demand and available capacity. Potential users with larger price elasticities of demand who are prepared to book either well in advance, or take a chance on a last-minute booking, can get lower prices, albeit often with zero flexibility. Those with low price elasticities of demand, typically business users or others with limited flexibility in the time of travel, have to pay higher fares. This means that users on any particular flight or train may have paid significantly different fares for what is essentially the same journey, although one for which the implied utility and willingness to pay is very different. This arises because the service paid for cannot be transferred either between different users or at different times by the same user. More common is third-degree price discrimination, where it is not possible to distinguish between individuals except through their response to a specific characteristic. The most common form in use in transport is time-of-day pricing which allows price discrimination

History of transport pricing  15

between peak and off-peak or weekday and weekend. Whilst this reflects the cost of providing capacity in the peak, it also exploits the different price elasticities of demand between different groups of users but averaged across the groups. A variant of this is the use of discounts for specific groups, for example, by age, so that older users or students are given discounts, or in some cases free travel, albeit with restrictions on, for example, time of day when they can be used. More generally, many operators use the idea of discounts for regular users such as season tickets, railcards, frequent flyer clubs, or “carnets”, which provide for cheaper travel on the individual journey but may require an upfront payment. This is not dissimilar to the twopart tariff idea often used in energy pricing in which users pay a fixed fee for connection and then a cheaper price per unit used. These give greater ease of use for the individual user and a greater certainty of revenue for the operator, but also carry implications for equity (covered in more depth by Heyndrickx and Mayeres in Chapter 6) and may work against the general principle of charging higher prices in peak hours leading to crowding issues (see, for example, the discussions in Carbajo, 1988; Jara-Díaz et al., 2016; Hörcher et al., 2018).

1.4 PAYING FOR ROADS The paying for roads question has been a major theme in the political economy of transport pricing. On the one hand, it is the relatively simple recognition that congestion causes marginal social cost to rise faster than private costs and that this situation will continue unless there is some form of intervention. On the other hand, the more radical approach is the idea of effectively privatising the road network such that the problem becomes internalised (Roth, 1996, 2006). A Pigouvian tax is the obvious analytical solution to the first approach, but the practical problem of estimating an optimal rate of tax and charging it and the political problems in implementing it have largely prevented the large-scale introduction of time-varying or traffic volume-dependent tolls. This is further complicated by the recognition of the dynamic nature of congestion and the need to resort to second-best solutions, as noted in Section 1.3.1. The focus of this section remains on practical implementations of road pricing; the underlying theoretical problems are covered by Mun and Fukuda in Chapter 8. Toll roads have existed since the early days of road transport, largely as a means of covering the costs of maintaining the road in the light of the damage caused by vehicles passing along it. Some countries have seen a proliferation of tolled highways more recently, usually roads which provide a better standard of service than existing roads, but which largely run parallel to existing routes, so users have a choice of a faster road at a higher cost or a slower route at a lower cost. This includes the use of high-occupancy toll lanes reserved for vehicles with more than a just a driver in some parts of the United States, as Chapter 23 describes. The development of more sophisticated electronic toll collection systems based either on automatic number plate recognition or by using onboard receivers has made tolling both more feasible in high-density urban areas and more able to differentiate by vehicle type, time of day, and traffic levels. But drivers have been very resistant to such universal charging, even if the net result would be lower costs of a journey due to lower levels of congestion, saving both time and fuel (for a comprehensive discussion of the issues, see Jensen-Butler et al., 2008; Santos and Verhoef, 2011; Lehe, 2019). The largest scale implementations of urban road pricing have been the electronic system in Singapore (Santos, 2005) and the area-based systems in London and Stockholm. These

16  Handbook on transport pricing and financing

required either the courage of a political leader, as in the case of London, or in the case of Stockholm a well-managed trial that demonstrated the potential benefits to drivers before a referendum on implementation (Eliasson, 2008). The London experience has been that traffic and congestion did decrease significantly on implementation, in fact to the extent that predicted revenues were below the target (Peirson and Vickerman, 2008; Santos, 2008). Most estimates of the impact of price changes on the demand for transport are based on revealed preference studies which observe the impact of marginal changes in price (usually increases) on demand. It has been difficult therefore to predict accurately the effect of a step-change in both the level and the relative price to other modes. This suggests that it will be difficult to predict the outcome of any attempts to reset prices on a large scale. The London system earmarked revenues for improved bus services and thus went some way towards an integrated system for urban traffic, although an area-based system allows free movement once inside the zone, and there are exemptions and discounts for residents and certain types of vehicle such as electric cars. This tends to distort the system and it has become complicated by the introduction of an ultra-low emission zone aimed at reducing pollution. A flat fee loses the direct link between congestion and charging. Since the charge is an addition to all other taxes on the use of vehicles, it is often perceived as inequitable with negative impacts on businesses inside the zone. Increases to the charge and attempts to extend the zone have been part of the rescue package for Transport for London following the collapse of public transport revenues post-pandemic (Department for Transport, 2020). This is a long way from having an integrated system for charging for urban transport.

1.5 PUBLIC PROVISION, THE PUBLIC SERVICE OBLIGATION, AND REGULATION 1.5.1 Public Provision and Public Service The question of whether pricing is a means of covering the costs of providing a transport good or service with the elements of a public good or a means of ensuring an optimal allocation between different types of transport with different cost structures has dominated more recent debates. This has coincided with the move to deregulate and privatise the provision of transport. If transport can be seen as a competitive good, then the problem of regulating prices recedes, and normal anti-trust legislation can be used to avoid the abuse of a monopoly position. This does, however, leave the problem of what happens when operators withdraw from a market that is seen as unprofitable, the public service obligation. Franchising or competitive tendering removes competition from within the market and replaces it with competition for the market. In this case, the public sector will again become responsible for setting price levels. The public sector regulator may wish to use the level of prices (or any permitted increases) to squeeze profit levels and secure greater efficiency, but this is typically driven by a need to reduce public sector budgets rather than evidence of any gross inefficiencies. Despite many studies estimating the price elasticity of different modes of transport, real evidence of the effects of using price as a means of changing transport users’ choice is limited as discussed above in the context of congestion charging. Building on an analysis of Mohring (1972) into scale economies in urban bus transport, Turvey and Mohring (1975) extended this

History of transport pricing  17

to derive the basis for optimal, first-best, bus fares, including passengers’ time costs. Glaister and Lewis (1978) explored the potential for a fares policy for London, which explicitly took into account the price elasticities of demand for all modes, including private car transport, assuming that the first-best conditions were not available for the alternative mode. Cervero (1990) used estimates of fare elasticities to examine the case for eliminating fares on public transport. There have been studies of the impact of making public transport free in individual cities (e.g. De Witte et al., 2006; Cats et al., 2014, 2017; Hess, 2017), but relatively few of these have become permanent as political pressure on budgets dominated. An experiment to hold down prices in South Yorkshire in the late 1970s and early 1980s (Goodwin et al., 1983) was successful in maintaining public transport ridership at higher levels than in comparable areas but was ended by pressure from central government over the impact on local finances. As with the imposition of congestion charging, the effect of step-changes in prices proved difficult to predict. A detailed discussion of the issues in the provision of public transport is provided in Chapter 9 (Jara-Díaz et al.), and the political and governance issues in Chapter 7 (Russo and De Borger). 1.5.2 Regulation The twin issues of monopoly on the one hand and destructive competition on the other dominated much of the discussion of transport pricing in the first half of the twentieth century. This led in turn to the regulation of transport operators to try and kerb the potential for destructive competition through low barriers to entry in the growing road transport sector and fear of exploitation from natural monopolies in sectors such as rail, which required large investments in infrastructure. In many cases, this led ultimately to public ownership. Regulation focused largely on entry and unfair competition through licensing of both operators and their services and in ensuring a minimum level of provision through a public service obligation placed on operators (Ponti, 2011). In such a regime, prices were typically based on average costs across all operations rather than marginal costs, allowing for cross-subsidy. The decline in demand for public transport, alongside the rise in private car ownership in the second half of the twentieth century, posed problems for this regime. More services became marginal, threatening the viability of the public service obligation and in turn more expensive to maintain as pressure mounted on public sector budgets. At the same time, the growth of demand for air travel made the traditional highly regulated and often exclusive mode of delivery increasingly outdated. The simultaneous desire for economies and liberation from regulation led to a push for deregulation and privatisation across the whole sector. The belief that competition would reduce the cost of providing expensive public service obligation services whilst allowing innovation and the development of new services fuelled this move. It was recognised that complete deregulation leading to a free for all could be destructive, so some residual regulation remained. This was first of all to maintain safety through entry regulation but second became the introduction in many markets of competition for the market rather than in the market through some form of franchising or competitive concessions. As in the privatisation of other utility markets, the public sector usually retained some overall control of prices. This could be either in the form of retaining the right to set prices with operators competing for services on a gross cost basis or, in effect, a control on profits typically by setting the allowed annual change

18  Handbook on transport pricing and financing

in prices related to an overall price index. In some cases, this ensured that prices should rise by less than the price index to try and force increases in efficiency. Some of the biggest changes occurred in local public transport and the railways. In the former, there was a move from direct public sector provision in many countries towards a regime of competitive tendering (Gagnepain et al., 2011; van de Velde, 2015). Changes in the railways involved a number of dimensions, but in Europe there was a general move towards the vertical separation of infrastructure and rail services and an increase in the use of competitive tendering at least for some regional services (Nash, 2011, 2015). Price setting in these regimes has two basic dimensions. The first is essentially a political and policy decision on to what extent the individual user should cover the cost of any journey. In some jurisdictions, the United Kingdom being an obvious example, the individual user is expected to bear a larger share of the cost, so fares are high in both nominal terms and real terms relative to, for example, wage levels. This leads to high farebox ratios and a greater revenue risk to the operator in the event of a downturn in traffic levels, such as experienced during the COVID-19 pandemic (Vickerman, 2021a). The second is the way prices are set in any franchising or competitive tendering process. Under a gross cost approach, the regulator or public sector commissioning body sets the fare level and the services and receives all the revenue whilst the operator takes only the supply side risks of the costs. Under a net cost approach, the operator takes on both revenue and cost risks, subject to any overall control over price levels. In UK rail franchising, for example, basic fares, including season ticket prices, are regulated with annual fare changes determined by the Department for Transport, but operators are free to set other fares. The bidding for franchises thus depends on operators’ ability to forecast potential growth in the market to secure the revenues that will enable the operator to pay the government for the franchise. On several occasions, operators have failed to meet the contractual terms of the franchise and have had to hand the franchise back (Vickerman, 2021b). It is clear from this discussion that the prices of public transport services are increasingly unrelated to the cost of providing that service or a conventional relationship between supply and demand. They have become largely a matter of political choice and what public sentiment or voters’ opinions will bear. Regulation is discussed in the context of European approaches to pricing in Chapter 21 (Nash and Link) and in how it is applied to questions of public–private partnerships in Chapter 16 (Fischer et al.), as well as the wider discussion of public transport in Chapter 9 (Jara-Diaz et al.).

1.6 IMPLEMENTING IDEAL PRICES Whilst theoretically ideal prices could be defined for most transport services their adoption has been hindered by the lack of an appropriate means of implementing them in practice. Online booking for airfares and many long-distance rail journeys has meant that operators could get closer to individuals’ willingness to pay for a specific journey. This is in part because these are likely to be larger expenditures which occur infrequently. Payments for everyday journeys have remained until recently largely dependent on cash payments except where season tickets or some form of regular user card were available. New ways of charging for journeys can make pricing more sensitive both to the user and to the circumstances of the journey, such as

History of transport pricing  19

time of day. The move in many large cities to integrated fare cards has allowed for a single means of payment to cover journeys by different modes or allowing for multiple interchanges, thus simplifying the journey for the user with a greater degree of transparency. This includes the potential for giving discounts to the regular traveller by imposing a maximum daily fare. However, it also moves us away from the first-best optimal pricing by blurring the transparency of the fare paid for each trip. Transferring this technology to the use of contactless bank cards simplifies the situation for the occasional traveller. Similarly, the ability to charge private motorists directly for their journeys by means of onboard transponders that interact with roadside stations has increased the potential for more sensitive road pricing than that provided by cordon-based systems. In both private and public transport, the ability to transfer this technology to mobile phones provides the opportunity to integrate all transport pricing in a way that makes the relative cost of journeys by different modes transparent and hence aids optimal decision-making. Thus, integrating the journey planning technology used by many with the relevant charging systems may bring us closer to the theoretical solutions of the textbooks once the privacy concerns raised by such a situation have been resolved. The next stage of development is to bring both public and private modes of transport into an integrated system. This “mobility as a service” (MaaS) idea identifies that the basic demand is for mobility rather than particular modes of transport. Different operators, including taxis, ride-hailing services, and active transport (such as cycling and walking) can provide services, but these are sold through “integrators” who can offer a subscription service for mobility (Hensher, 2017; Matyas and Kamargianni, 2019; Mulley et al., 2018). Whether this really represents a further step towards being able to achieve optimal pricing in practice has been questioned in more recent attempts to model the effects of the supposed business models (Hörcher and Graham, 2020; van den Berg et al., 2022).

1.7 CONCLUDING THOUGHTS The main message of this chapter is that a consistent theoretical model of pricing has been developed over the past two centuries. The model recognises the specific characteristics of transport both in the charging for the use of transport infrastructure and for transport services. What has proved less straightforward has been the implementation of theoretically correct first-best prices in practice. This has been both for technical and logistical reasons on the one hand and for political reasons on the other. It has been simpler on both grounds to opt for second-best solutions. Hence there has been a reluctance to move away from solutions which subsidise public transport as means of approximating the right relative prices of public and private transport. It is possible that the technical and logistical constraints are now much less with modern digital technologies that will allow for prices closer to the first-best for each user. However, if this implies increases in the prices for public transport, there will be pressure on equity grounds to find ways of limiting the increases for certain classes of user. The political will to move transport prices such that they reflect the full cost, including all external effects, will differ according to the regimes currently in place. Regulated prices and subsidies can be used to manipulate both levels of mobility and the choice of mode; changing the paradigm can be difficult and can produce unexpected consequences. For example, to encourage cleaner transport, electric cars are often subsidised and exempted from congestion

20  Handbook on transport pricing and financing

charges. The growth of electric car numbers then reduces the tax revenues governments receive from the sale of fossil fuels, and there is no incentive to reduce car movements in congestion charge zones. Unless and until there is a consistent approach to pricing mobility, there will continue to be problems of this type (see discussion in Poon and Vickerman, 2020). The COVID-19 pandemic caused problems for transport operators across the world. Reduced ridership has led to large falls in revenue that have required government bailouts. It seems likely that the recovery, certainly in Europe and the Americas, will take several years. It is not yet clear whether ridership in large cities will ever return to the levels seen before the pandemic as working from home, at least for part of the week, becomes part of the normal pattern of work. In many cities, car usage has returned more strongly than public transport usage. Does this provide a major opportunity for a rethink of how to price public transport and reset the relative prices of the various modes (Vickerman, 2021a)?

NOTES 1. 2.

See also Berechman et al. (2011). The basic definitions and a formal analysis are given in Quinet and Vickerman (2004), pp. 276–283.

REFERENCES Anderson, S.P. and Renault, R. 2011. Price discrimination. In de Palma, A., Lindsey, R., Quinet, E. and Vickerman R. eds. A Handbook of Transport Economics. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 527–560. Arnott, R., de Palma, A. and Lindsey, R. 1990. Economics of a bottleneck. Journal of Urban Economics. 27, 111–130. Arnott, R., de Palma, A. and Lindsey, R. 1998. Recent developments in the bottleneck model. In Button, K.J., Verhoef, E.T. eds. Road Pricing, Traffic Congestion and the Environment: Issues of Efficiency and Social Feasibility. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 78–110. Baumol, W.J. and Bradford, D.F. 1970. Optimal departures from marginal cost pricing. American Economic Review. 60, 265–283. Beckmann, M.J., Maguire, C.B. and Winsten, C.B. 1956. Studies in the Economics of Transportation. New Haven, CT: Yale University Press. Berechman, J., Bartin, B. Yanmaz-Tuzel, O. and Ozbay, K. 2011. The full marginal costs of highway travel: methods and empirical estimation. In de Palma, A., Lindsey, R., Quinet, E. and Vickerman, R. eds. A Handbook of Transport Economics. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 444–475. Böhm-Bawerk, E. von. 1891. The Positive Theory of Capital. Translated by William Smart. London: Macmillan. Boiteux, M. 1956. Sur la gestion des monopoles publics astreinte à l’équilibre budgétaire. Econometrica. 24, 22–40. Bonnafous, A. and Crozet, Y. 2018. Consumer surplus and pricing of transport infrastructures: the legacy of Jules Dupuit. Transport Policy. 70, 8–13. Carbajo, J.C. 1988. The economics of travel passes: Non-uniform pricing in transport. Journal of Transport Economics and Policy. 22, 153–173. Cats, O., Reimal, T. and Susilo, Y. 2014. Public transport pricing policy: empirical evidence from a farefree scheme in Tallinn, Estonia. Transportation Research Record. 2415, 89–96. Cats, O., Susilo, Y.O. and Reimal, T. 2017. The prospects of fare-free public transport: evidence from Tallinn. Transportation 44, 1083–1104. Cervero, R. 1990. Transit pricing research. Transportation. 17, 117–139.

History of transport pricing  21

Delucchi, M. and McCubbin, D. 2011. External costs of transport in the United States. In de Palma, A., Lindsey, R., Quinet, E. and Vickerman, R. eds. A Handbook of Transport Economics. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 341–368. Department for Transport. 2020. Transport for London Settlement Letter. https://assets​.publishing​. service​.gov​.uk ​/government ​/uploads​/system ​/uploads​/attachment ​_ data ​/file​/931802​/transport​-for​london​-settlement​-letter​.pdf. De Witte, A., et al. 2006. The impact of free public transport: the case of Brussels. Transportation Research. Policy and Practice. 40, 671–689. Dupuit, J. 1844. De la mesure de l’utilité des travaux publics. Annales des Ponts et Chaussées. 8, 332–375. Dupuit, J. 1849. De l’influence des péages sur l’utilité des voies de communication. Annales des Ponts et Chaussées. 2, 170–248. Ekelund, R.B. and Hébert, R.F. 2002. Retrospectives: The origins of neoclassical microeconomics. Journal of Economic Perspectives. 16, 197–215. Eliasson, J. 2008. Lessons from the Stockholm congestion charging trial. Transport Policy, 15, 395–404. Friedrich, R. and Quinet, E. 2011. External costs of transport in Europe. In de Palma, A., Lindsey, R., Quinet, E. and Vickerman R. eds. A Handbook of Transport Economics. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 369–395. Gagnepain, P., Ivaldi, M. and Muller-Vibes, C. 2011. The industrial organisation of competition in local bus services. In de Palma, A., Lindsey, R., Quinet, E. and Vickerman R. eds. A Handbook of Transport Economics. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 744–762. Glaister, S. and Lewis, D. 1978. An integrated fares policy for transport in London, Journal of Public Economics. 9, 341–355. Goodwin, P.B., Bailey, J.M., Brisbourne, R.H, Clarke, M.I., Donnison, J.R. Render, T.E. and Whiteley, G.K. 1983. Subsidised Public Transport and the Demand for Travel: The South Yorkshire Example. Aldershot: Gower. Hensher, D. 2017. Future bus transport contracts under a mobility as a service (MaaS) regime in the digital age: Are they likely to change? Transportation Research. Part A Policy and Practice. 98, 86–96. Hess, D.B. 2017. Decrypting fare-free public transport in Tallinn, Estonia. Case Studies in Transport Policy. 5, 690–698. Hicks, J.R. 1973. Capital and Time: A Neo-Austrian Theory. Oxford: Oxford University Press. Hörcher, D., Graham, D.J. and Anderson, R.J. 2018. The economic inefficiency of travel passes under crowding externalities and endogenous capacity. Journal of Transport Economics and Policy. 52, 1–22. Hörcher, D. and Graham, D. J. 2020. MaaS economics: Should we fight car ownership with subscriptions to alternative modes? Economics of Transportation. 22, 100167. Hotelling, H. 1938. The general welfare in relation to problems of taxation and of railway and utility rates. Econometrica. 6, 242–269. Jara-Díaz, S., Cruz, D. and Casanova, C. 2016. Optimal pricing for travelcards under income and car ownership inequities. Transportation Research Part A. Policy and Practice. 94, 470–482. Jensen-Butler, C., Sloth, B., Larsen, M.M., Madsen, B. and Nielsen, O.A. ed. 2008. Road Pricing, the Economy and the Environment. Berlin and Heidelberg: Springer-Verlag. Knight, F. 1924. Some fallacies in the interpretation of social costs. Quarterly Journal of Economics. 38, 582–606. Lehe, L. 2019. A history of downtown road pricing, Transportation Research Part C: Emerging Technologies. 100, 200–223. Marshall, A. 1920. Principles of Economics, 8th ed. London: Macmillan. Matyas, M. and Kamargianni, M. 2019. The potential of mobility as a service bundles as a mobility management tool. Transportation. 46. 1951–1968. McDonald, J.F. 2013. Pigou, knight, diminishing returns, and optimal Pigouvian congestion tolls. Journal of the History of Economic Thought. 35, 353–371. Mohring, H. 1972. Optimisation and scale economies in urban bus transportation. American Economic Review. 62, 591–604.

22  Handbook on transport pricing and financing

Mohring, H. and Harwitz, M. 1962. Highway Benefits: An Analytical Framework. Evanston, IL: Northwestern University Press. Mulley, C., Nelson, J.D. and Wright, S. 2018. Community transport meets mobility as a service: On the road to a new a flexible future. Research in Transportation Economics. 69, 583–591. Nash, C. 2011. Competition and regulation in rail transport. In de Palma, A., Lindsey, R., Quinet, E. and Vickerman, R. eds. A Handbook of Transport Economics. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 763–778. Nash, C. 2015. Rail. In Nash. C. ed. Handbook of Research Methods in Transport Economics and Policy. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 359–370. Newbery, D. 1988a. Road user charges in Britain. Economic Journal. 98, 161–176. Newbery, D. 1988b. Road damage externalities and road user charges. Econometrica. 56, 295–319. Peirson, J. and Vickerman, R. 2008. The London congestion charging scheme: the evidence. In JensenButler, C., Sloth, B., Larsen, M.M., Madsen, B. and Nielsen, O.A. eds. Road Pricing, the Economy and the Environment. Berlin and Heidelberg: Springer-Verlag, pp. 79–91. Pigou, A.C. 1920. The Economics of Welfare. 1st ed. London: Macmillan. Pigou, A.C. 1932. The Economics of Welfare. 4th ed. London: Macmillan. Pigou, A.C. 1952. The Economics of Welfare. 4th ed. Reprinted with 8 New Appendices. London: Macmillan. Ponti, M. 2011. Competition, regulation and public service obligations. In de Palma, A., Lindsey, R., Quinet, E. and Vickerman R. eds. A Handbook of Transport Economics. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 661–683. Poon, J.F. and Vickerman, R.W. 2020. Workshop 8. Beyond the Farebox: Sustainable funding of public transport by better understanding service values. Research in Transportation Economics. 83, 100923. Quinet, E. and Vickerman, R. 2004. Principles of Transport Economics. Cheltenham, UK and Northampton, MA: Edward Elgar. Ramsey, F.P. 1927. A contribution to the theory of taxation, Economic Journal. 37, 47–61. Rizzi, L.I. and Ortúzar, J. de D. 2015. Valuing transport externalities, in Nash. C. ed. Handbook of Research Methods in Transport Economics and Policy. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 93–111. Roth, G.J. 1996. Roads in a Market Economy. Aldershot, Hants: Avebury Technical. Roth, G.J. ed. 2006. Street Smart: Competition, Entrepreneurship and the Future of Roads, New Brunswick, NJ and London, UK: The Independent Institute, Transaction Publishers. Rothengatter, W. 2018. Mr. Dupuit and the marginalists. Transport Policy, 70, 32–39. Santos, G. 2005. Urban congestion charging: a comparison between London and Singapore. Transport Reviews. 25, 511–534. Santos, G. 2008. The London experience, in Verhoef, E., Bliemer, M., Stieg, L. and van Wee, B. eds. Pricing in Road Transport: A Multi-Disciplinary Perspective. Cheltenham, UK: Edward Elgar, pp. 273–292. Santos, G. and Verhoef, E. 2011. Road congestion pricing. In de Palma, A., Lindsey, R., Quinet, E. and Vickerman R. eds. A Handbook of Transport Economics. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 561–585. Small, K.A. and Verhoef, E.T. 2007. The Economics of Urban Transportation. Abingdon: Routledge. Turvey, R. and Mohring, H. 1975. Optimal bus fares. Journal of Transport Economics and Policy. 7, 280–286. van den Berg, V. A., Meurs, H. and Verhoef, E. T. 2022. Business models for Mobility as a Service (MaaS). Transportation Research Part B: Methodological. 157, 203–229. van de Velde, D. 2015. Local and regional public transport. In Nash, C. ed. Handbook of Research Methods in Transport Economics and Policy. Cheltenham, UK and Northampton, MA: Edward Elgar, pp. 345–358. Verhoef, E.T. 2001. An integrated dynamic model of road traffic congestion based on simple carfollowing theory: Exploring hypercongestion. Journal of Urban Economics. 49, 505–542. Verhoef, E.T. 2002. Second-best congestion pricing in general static transportation networks with elastic demands. Regional Science and Urban Economics. 32, 281–310. Verhoef, E.T. 2003. Inside the queue: hypercongestion and road pricing in a continuous time-continuous place model of traffic congestion. Journal of Urban Economics. 54, 531–565.

History of transport pricing  23

Verhoef, E.T., Nijkamp, P. and Rietveld, P. 1995. Second-best regulation of road transport externalities. Journal of Transport Economics and Policy. 29, 147–167. Vickerman, R. 2021a. Will Covid-19 put the public back in public transport? A UK perspective. Transport Policy. 103, 95–102. Vickerman, R. 2021b. Intercity modal competition. In Mulley, C., Nelson, J.D. and Ison, S. eds. The Routledge Handbook of Public Transport. London: Routledge, pp. 61–71. Vickrey, W.S. 1948. Some objections to marginal-cost pricing, Journal of Political Economy. 56, 218–238. Vickrey, W.S. 1963. Pricing in urban and suburban transport. American Economic Review, Papers and Proceedings. 53, 452–465. Walters, A. A. 1954. Track costs and motor taxation. The Journal of Industrial Economics. 2, 135–146. Walters, A.A. 1961. The theory and measurement of private and social cost of highway congestion. Econometrica. 29, 676–699. Walters, A.A. 1968. The Economics of Road User Charges. World Bank Staff Occasional Papers, No.5. Baltimore, MD: Johns Hopkins University Press. Wicksell, K. 1901. Lectures on Political Economy. Vol.1. Translated by E. Classen. London: Routledge and Kegan Paul.

2. Transport pricing: theory and methodologies Achim I. Czerny and Stefanie Peer

2.1 INTRODUCTION Pricing objectives in transport markets include profit maximisation, the recovery of infrastructure investment costs, and tax revenue generation. Following the tradition of welfare economics, this chapter focuses on welfare-related pricing, which is the central pricing objective in the field of transport economics (see Chapter 1 of this Handbook for an overview of the history of transport pricing). Prices are considered “optimal” from a welfare perspective if they lead to (Pareto) efficiency, which is the necessary condition for the maximisation of social welfare functions that respect the Pareto principle: welfare increases as long as at least one person is better off and no-one is worse off. This principle is less restrictive than sometimes assumed, as it is consistent with a wide variety of welfare functions, and hence also allows for distributional preferences to be taken into account. Section 2.2 of this chapter introduces the economic principle of marginal social cost pricing in the context of transport markets. Section 2.3 provides an overview of the role of this principle for different (static and dynamic) models of capacity scarcity, mostly (but not exclusively) focused on road congestion. Section 2.4 discusses optimal pricing in the presence of other externalities such as pollution, noise, and accidents. Section 2.5 highlights several other aspects relevant to pricing in transport markets: atomic users, tradable permits versus pricing, networks, and heterogeneity in time valuations. What these topics have in common is an overarching importance for transport pricing, reflected by them being mentioned in several other chapters of the Handbook. Section 2.6 provides an overview of solution and optimisation methods for deriving optimal prices in the context of transport, and Section 2.7 concludes the chapter. Many topics are only briefly mentioned in this chapter and discussed more extensively in other chapters or sources. To guide the reader through the different book chapters, crossreferences are used wherever relevant.

2.2 CORRECTIVE TAXES AND SUBSIDIES The standard economic theory states that the allocation of resources will not be (Pareto) efficient if prices do not reflect the marginal social costs. In the transport sector, prices are often imperfect, as many travel activities are sources of externalities. Externalities can be negative or positive depending on whether they represent costs or benefits, respectively, that accrue for society but are not included in the price and hence ignored by travellers in their decision-making. For road transport, negative externalities prevail and include time losses and schedule delays, local pollution, greenhouse gas emissions, noise, accidents, and travel time variability; for public transport, positive externalities tend to prevail with the most relevant one being scale economies (when capacity is responsive), but also negative ones such as 24

Transport pricing 

25

in-vehicle crowding can be present. From a social welfare perspective, negative externalities lead to an over-consumption and positive externalities to an under-consumption of the corresponding transport services. Corrective taxes or subsidies can be used to internalise externalities by adjusting the price, which in turn provides incentives for travellers to adjust their behaviour. In the short run, the behavioural responses may involve a change in the number of trips, a shift in departure time, a switch to another transport mode or route, or a change in driving style; medium-run adjustments might concern the purchase of a different car, and in the long run also locational decisions may be adapted in response to price adjustments. Such corrective pricing principles have been pioneered by Arthur Pigou (1920) and are usually referred to as “marginal social cost pricing”. In the absence of other distortions in the economy (so-called “first-best” scenarios), marginal social cost pricing will lead to the social optimum by providing an incentive for individuals to only realise trips for which the marginal social benefits exceed the marginal social costs. By doing so, marginal social cost pricing maximises the difference between total social benefits and total social costs (i.e. the social surplus). This outcome is (Pareto) efficient: even though the payments may make some users worse off, the revenues from the tolls exceed the loss of the users’ surplus and hence can potentially be used to compensate them for their loss. In the presence of other distortions and pricing constraints (so-called “second-best” scenarios), deviations from the marginal social cost pricing principle can be optimal. Second-best scenarios exist if the externalities under consideration are not the only distortion present in the economy, but other distortionary elements such as income taxes, sub-optimally priced alternative transport modes, transaction costs, and information asymmetries are present. Another instance of a second-best scenario is if the use of the pricing instrument is constrained, for instance, due to technological reasons, cost recovery requirements, social equity motives, or lack of acceptance among users. In second-best scenarios, which tend to correspond more closely to reality than first-best scenarios, the pricing instruments are optimised subject to the distortions and/or constraints, and the marginal social costs/benefits need to be adjusted by corrective terms. These often correspond to mark-ups based on relative price elasticities, a principle referred to as Ramsey pricing (Ramsey, 1927). The resulting pricing schemes are sometimes referred to as the “marginal cost based price” (Rouwendal & Verhoef, 2006), with the corresponding welfare gains relative to naively applying first-best pricing (“quasi firstbest pricing”) varying substantially. For public transport, second-best considerations usually provide arguments for the subsidisation of fares. They include reducing distortions of labour supply caused by the distortionary effect of taxes, social equity concerns, and public transport competing against under-priced private car travel. Clearly, pricing and capacity decisions are not independent of each other. In the short run, capacity is usually considered fixed and prices should be set optimally, conditional on a given capacity, whereas in the long run capacities can also be optimised. The self-financing theorem by Mohring and Harwitz (1962) states that the revenues generated from optimal pricing match the costs of providing the optimal capacity. This holds if scale economies in the construction of the infrastructure are absent, and if equi-proportional changes in the number of trips and capacity do not affect travel costs (the user cost function is homogeneous at degree zero). These conditions tend to hold for car travel, but not for public transport, where capacity expansions are likely to affect user costs, meaning that the user cost function is not any longer homogenous at degree zero. This happens as operators may employ larger vehicles, increase service

26  Handbook on transport pricing and financing

frequency, or provide a denser network structure, hence also affecting user costs associated with crowding and waiting time (Mohring, 1972). These economies of scale are yet another argument for optimal fares being lower than the marginal cost of service provision and hence for subsidies for the public transport operators. See Chapter 3 of this Handbook for a more in-depth treatment of the self-financing theorem, as well as Verhoef and Mohring (2009) for a discussion on potential misinterpretations of the self-financing theorem, and Part III of this Handbook for more detailed discussions of financing topics in transport. The pricing of public transport is covered in more detail in Chapter 8 of the Handbook.

2.3 MODELS OF CAPACITY SCARCITY Road traffic causes a number of different negative externalities, with congestion being usually the most costly one, especially in marginal terms. With congestion, everybody incurs time losses while also imposing them on others. The marginal social costs hence increase with the number of road users. Pricing can be used to reduce the level of congestion to a socially optimal level by making individuals internalise the time costs inflicted upon other travellers. Congestion also causes schedule delays to other travellers, implying that they adjust their schedule such that they arrive at their destination earlier or later than desired. Schedule delays are ignored in static models (covered in Section 2.3.1), whereas they form an essential part of dynamic congestion models (Section 2.3.2). Moreover, delays due to congestion are also positively correlated with travel time variability (this relationship is not only evident for car travel, but also for public transport; see Durán-Hormazábal & Tirachini, 2016). Unreliable travel times are a source of disutility to travellers (see for instance the meta-analysis on valuations of reliability by Carrion & Levinson, 2012), causing them to take into account buffer times when deciding on their departure time, and inducing schedule delays relative to the expected arrival time. Taking into account schedule delays and travel time variability as additional external costs when determining optimal prices will ceteris paribus lead to a higher optimal price. As with road congestion, also in the context of public transport services (discussed in detail in Chapter 9 in this Handbook), the marginal social costs are co-determined by the number of users, as users can impose in-vehicle crowding costs on each other. Although less relevant than for road transport, also public transport may cause congestion externalities to other users of the infrastructure. 2.3.1 Static Models This section focuses on the most basic congestion model, which is a static model in which congestion occurs along a uniform stretch of a road (Walters, 1961). At the end of this subsection, static models in the public transport domain will also be discussed briefly. Figure 2.1 illustrates the static congestion model.1 The congestion costs are increasing in the traffic volume denoted by q because each additional car increases the average travel time. This implies that the marginal congestion costs are above the average congestion costs. The users internalise only the average congestion costs in their decisions, as they regard the aggregate traffic conditions and, thus, the average congestion costs as given. The inverse demand curve reflects the willingness to pay for travel, in other words, the road users’ marginal benefits. Therefore, the equilibrium traffic volume q’ is determined by the intersection of the inverse

Transport pricing 

27

Figure 2.1  Optimal pricing in the static congestion model demand curve and the average congestion costs curve. In contrast, the social optimum q* is located where the inverse demand curve intersects the marginal congestion costs curve at which point congestion is not eliminated but reduced to its optimal level. The difference between the marginal congestion costs and the average congestion costs determines the uninternalised part of the marginal congestion costs: the marginal external congestion costs. The social optimum can be achieved by introducing a toll p* that is equal to the marginal external congestion costs evaluated at the optimal traffic volume q*. To derive the optimal toll, a model of traffic congestion – usually based on the speed-volume curve of the fundamental diagram of traffic congestion – needs to be assumed. The behavioural response to the imposed marginal social cost pricing leads to a reduction in traffic from q’ to q*, which causes a reduction in benefits that is smaller than the corresponding reduction in congestion costs and, therefore, leads to a net social gain depicted by the grey area in Figure 2.1. The tolls do not only lead to the optimal volume of traffic, but also to the optimal composition of road users: the scarce infrastructure capacity is allocated to those who derive the largest benefits from using it, hence achieving allocative efficiency. 2.3.2 Dynamic Models Transport activities have an inherent time dimension. The previous section demonstrated that pricing can be studied in a static model in which the time dimension is absent. Adding the time dimension leads to a dynamic approach, which can capture how departure time choices affect traffic volumes and delays over time. This section discusses dynamic approaches and implications for pricing policies. It distinguishes between approaches in which the time dimension is modelled either in fixed intervals or continuously. 2.3.2.1 Discrete time: peak-load pricing The consideration of fixed time intervals is common in so-called peak-load pricing models. These models are used to capture demand variations throughout the day (Crew et al., 1995). For instance, urban transport typically experiences high demand during early and late hours

28  Handbook on transport pricing and financing

of workdays relative to other times of the day and weekdays because of commuting. One way of handling such demand variations would be to install sufficient infrastructure capacity to ensure that travellers can be served without delays, even during periods of high demand. Because capacity is costly, this is unlikely to reflect an optimal solution. Early models focused on independent demands in which the prices charged in one period have no effect on the demands in other periods (Steiner, 1957; Keeler & Small, 1977). Another simplified version of a dynamic model is the static model by Walters (1961) with interdependent demands for two roads in which the average travel time on each road rises proportionally to the quantity of users on the roads. His static model can have a dynamic interpretation in the sense that the two roads could represent the utilisation of a single road in two subsequent periods. If individuals are indifferent to travelling in the first or the subsequent period, then Walters’ result on the importance of equalising the marginal social costs of road use for efficiency also applies to the dynamic interpretation. If individuals prefer travelling in one period over another period, then not only do the (marginal) social costs matter for the allocation of users between peak and other periods, but additionally the higher (marginal) benefits in the preferred travel period induce higher travel volumes and therefore higher marginal social costs and optimal prices in that period. A feature of the above-mentioned approaches is that, by treating congestion as independent of the travel volume in the other period, they implicitly assume that users who start their trip in the peak period will also arrive during the peak period despite the delays associated with congestion. Exceptions are Levinson (2005) and Zou and Levinson (2006) who develop multiperiod models with a finite number of users and Czerny and Zhang (2014) with an infinite number of users. They consider each period’s capacity as limited in the sense that the users’ period of departure and the period of arrival can be pushed apart by congestion creating deviations from preferred and actual arrival times thus leading to schedule delay costs. Their findings are in line with the standard result that optimal prices coordinate travelling decisions such that the total user costs including congestion and schedule delay costs are minimised. 2.3.2.2 Continuous time: the bottleneck model Continuous approaches that do not resort to fixed time intervals can be used to elegantly describe and analyse how queue lengths and congestion develop over time. The most common and influential approach is the so-called bottleneck model which goes back to the seminal paper of Vickrey (1969). Recent overviews of the bottleneck model and its numerous applications can be found in Small (2015) and Li et al. (2020). The bottleneck model considers dynamic congestion as an evolution of the queue behind the bottleneck. Congestion forms if the preferred arrival times are clustered such that the bottleneck capacity cannot accommodate for each traveller to arrive at his/her preferred moment. The bottleneck model is particularly attractive due to its analytical tractability. The bottleneck model is illustrated in Figure 2.2, which closely resembles the figure provided by Arnott, de Palma, and Lindsey (1993). The figure ignores free-flow travel time and assumes that travellers have a preference to arrive at t*. The horizontal axis of Figure 2.2 represents a timeline with time variable t. The figure displays three (solid) lines. The line in the middle represents the cumulative arrivals. Its slope thus represents the capacity of the bottleneck. On the vertical axis, “A” indicates the cumulative number of travellers that can travel through the bottleneck during a time period of length t’ (the time when the first traveller arrives) to t’’ (the time when all travellers have arrived). The line on the left displays the

Transport pricing 

29

Figure 2.2  Illustration of the bottleneck model cumulative departures. This line is steeper than the line representing the cumulative arrivals for the period t’ to t and flatter for the period t to t’’, with t indicating the moment in time where the queue is longest (see the line indicating the queue length at the bottom of the figure). The steeper part reflects times in which the departure rate exceeds the bottleneck capacity and the queue length grows, whereas the flatter part reflects times in which the departure rate is lower than the bottleneck capacity and hence the queue length declines. In the standard bottleneck model, capacity allocation relies on a first-come-first-serve basis and the queue has no spatial dimension (“vertical queue”). The schedule of departures reflects an equilibrium, as no traveller is better off by unilaterally changing the departure time. With the assumption that all travellers have the same preferred arrival time t* as well as identical time and scheduling preferences, all travellers face the same costs in equilibrium, consisting of time costs and schedule delay costs relative to their preferred arrival time. The relative size of these two components depends on the departure time. This can be illustrated by considering three departure moments. The first departure moment considers departure time t’. Queuing is not necessary in this departure moment because departure is early enough to avoid the queue; however, arrival is far from the desired arrival time which leads to scheduling costs. The second departure moment is t. This departure time leads to an arrival time that is exactly equal to the preferred arrival time and because this departure time is particularly attractive, it is acceptable to join the longest queue. The third departure moment is t’’. As in the first department moment, also in this departure moment there is no need for queuing because departure is late enough to avoid the queue; however, arrival is again far from the desired arrival time which also leads to scheduling costs. Observe that the difference between t and t’ is greater than the difference between t’’ and t. This can be an equilibrium constellation only if passengers prefer to arrive early compared to late, which is a common assumption. To install a pricing scheme that decentralises the social optimum, the toll must vary in time such that the users are incentivised to arrive at the bottleneck at a rate equal to the bottleneck capacity. Unlike most static models, optimal pricing fully eliminates congestion in the bottleneck model and does not make users worse off. The optimal price is zero for the first and the

30  Handbook on transport pricing and financing

last traveller, and highest for the traveller who arrives at t*, imitating the patterns of queuing costs in the user equilibrium. The social optimum also entails that the scheduling costs of the first and the last driver need to be equal, implying that – as in the user equilibrium – the first traveller will arrive at t’ and the last traveller at t’’. The optimal toll reflects the marginal external costs, as for a given moment it is equal to the difference between the (time-independent) marginal social costs and the time-dependent private (scheduling) costs for that moment. The classical Vickrey approach considers the total number of users fixed. Arnott, de Palma, and Lindsey (1993) generalise the model to price-sensitive total demand. They find that the average travel time rises proportional to the quantity of users. As highlighted by Small (2015), this links their approach to the earlier model developed by Walters (1961) and implies that the pricing lessons derived by Walters also apply to the bottleneck setting. Congestion pricing has also been investigated in alternative dynamic models, most importantly the flow congestion model of Chu (1995). In Chu’s model, the benefits of congestion pricing are less pronounced compared to the bottleneck model, most importantly because of a longer peak period and a corresponding increase in scheduling costs. The bathtub model builds on the macroscopic fundamental diagram and captures hypercongestion, which is a major empirical and theoretical topic in traffic flow. Fosgerau and Small (2013) consider a variable-capacity bottleneck to model hypercongestion. For public transport, time-dependent pricing schemes have been investigated to reduce invehicle crowding during peak hours and schedule delays. For instance, de Palma et al. (2017) study train travellers who face a trade-off between in-vehicle crowding and arriving earlier or later than desired. They jointly optimise the time-of-day dependent train fare, number of trains, and train capacity. See Chapter 9 in this Handbook for more details on the pricing of public transport services.

2.4 ENVIRONMENTAL, NOISE, AND SAFETY EXTERNALITIES Increasingly, negative environmental externalities are considered in co-determining the optimal price of (car) travel (recent meta-analysis by Sovacool et al., 2021). These include local air pollution, and greenhouse gas emissions.2 Local air pollution causes health damage as well as material damage, and crop and biodiversity losses. The size of the local pollution externality depends on vehicle-specific characteristics, such as fuel type, engine type, and emission class, as well as area-specific characteristics such as population density. The average and marginal costs of local air pollution are often assumed to be equal, as the relationships between air pollutants and damages are fairly linear (Van Essen et al., 2019). Combustion engines also emit substantial amounts of greenhouse gases, which lead to global warming, thereby causing severe impacts on the society and ecosystems (see Chapter 4 in this Handbook for a more detailed overview). In absence of better evidence, marginal and average costs of greenhouse gas emissions are usually assumed to be identical. Not only car travel, but also public transport (in particular trains) are an important origin of noise. Noise causes annoyance, and is associated with negative health and psychosocial consequences. Marginal and average noise costs diverge, as local factors such as population density, existing noise level, and time-of-day influence the marginal impact. While noise costs are external to a large extent, a small share may be internal in the form of nuisance to the vehicle user (Van Essen et al., 2019).

Transport pricing 

31

Another negative externality of transport (in particular road transport) is accidents. They cause injuries, fatalities and material damage. The marginal accident costs are defined as the extra (expected) accident costs associated with an additional vehicle, and depend on a variety of factors, such as vehicle type, road type, traffic volume, speed, weather, etc. The marginal external costs associated with accidents are generally difficult to estimate, in particular because a substantial part of the accident costs is internal (Mayeres et al., 1996). Accidentrelated road tolls would, in principle, have to be differentiated according to driving behaviour, as the latter (co-)determines accident risk. The magnitude of environmental externalities, noise, and accident risks may be affected by the presence and extent of congestion (De Palma & Lindsey, 2011; Theofilatos & Yannis, 2014; Høye & Hesjevoll, 2020; Yin & Lawphongpanich, 2006; Szeto et al., 2008). These interdependencies should be considered in the derivation of optimal prices (Parry & Bento, 2002; Coria & Zhang, 2017).

2.5 MODEL EXTENSIONS There are several other complementary topics which raise interesting methodological aspects and new insights on the pricing of transport services, and that are of relevance to a number of other chapters of this Handbook. A selection of four of them will be briefly discussed. 2.5.1 Atomic Users A common assumption in transport studies is that users are atomistic (or non-atomic) in the sense that they consider the price of travelling as exogenously given. This assumption seems appropriate, especially in the context of road transport and public transport. In other contexts, such as air transport, this assumption seems less appropriate (Borenstein, 1989). The reason is that passengers are served by dominant airlines which often operate substantial shares of flights at airports. Daniel (1995) applies a Vickrey-type approach to analyse how the presence of dominant airlines affects the optimal pricing of congested airports. He highlights that in the presence of a competitive fringe the optimal price may be equal to the price that would be charged in the case of non-atomic users. Others (e.g. Brueckner, 2002; Verhoef & Silva, 2017) find an inverse relationship between airline market shares and the optimal airport charges, which occurs because airlines are expected to internalise the congestion they impose on themselves. Pels and Verhoef (2004) further highlight the need to reduce airport charges as a measure against airline market power. For a more detailed discussion of airport policies and the role of atomistic users, see Chapter 11 in this Handbook. 2.5.2 Tradable Permits Versus Prices Pricing solutions are often difficult to implement whereas quantity-based measures are more common. Licence plate restrictions, perimeter control, and traffic calming in road transport, as well as slot controls in air transport, can serve as examples (de Palma & Lindsey, 2020). For the case of tradable travel permits, it has been shown that in an environment with perfect information and an efficient allocation of permits, the permit scheme can be equally effective in achieving welfare-maximising solutions as pricing. The assessment changes in an

32  Handbook on transport pricing and financing

environment with uncertainty. Weitzman (1974) highlights the importance of the curvatures of the demand functions and correlations between demand and marginal external costs for the assessment of pricing and permit policies. Czerny (2010) and de Palma and Lindsey (2020) extend Weitzman’s analysis to capture congestion externalities. They demonstrate that congestion could lead to steep demands and a negative correlation between demand and marginal external congestion cost functions, both improving the benefits of pricing relative to permit schemes. They further demonstrate that strongly convex marginal external congestion costs and the possibility of non-binding permit constraints tend to improve the relative welfare performance of permit schemes. 2.5.3 Networks Equilibrium user behaviour in road networks had been considered by Knight and Wardrop as early as 1924 and 1952, respectively. They highlighted that users will distribute across alternative routes so that their travel times/costs will be equal across all alternative routes. Since then, numerous studies have investigated marginal cost pricing in a network setting, ranging from highly stylised networks to real-size networks. In a first-best world, where pricing is feasible on all network links, marginal social cost pricing has been shown to extend to networks (Walters, 1961; Beckmann, 1965), even when accounting for stochasticity (Yang & Huang, 1998; Yang, 1999). The resulting tolls maximise the difference between the benefits of the users summed over origin–destination pairs minus the social costs summed over links.3 Also, various second-best cases have been investigated for networks (see also Chapters 3 and 8 in this Handbook). One stylised case corresponds to two routes being substitutes (Verhoef et al., 1996): when two parallel routes exist for a specific OD-pair, out of which only one route can be tolled, the optimal toll is below the marginal external congestion costs for the tolled route in order for congestion on the untolled link to decrease. Similarly, in the case where two links are sequential and hence complements, and only one (the upstream or the downstream) link can be tolled, the optimal toll is above the marginal external congestion costs of that link. For a more indepth discussion of second-best situations in a network setting, see Chapter 3 in this Handbook. 2.5.4 Heterogeneity in Time Valuations Individuals differ substantially with respect to their valuation of travel time and delays. Even for the same individual, time valuations may differ across trip purposes and transport modes (see, for instance, Shires and de Jong (2009) and Wardman et al. (2016) for meta-analyses of time valuations). Cohen (1987) and Hau (1992) highlight that travellers with high time valuations tend to benefit from pricing measures at the expense of low time valuation passengers in the absence of revenue recycling. This has been considered a major barrier to the practical implementation of pricing policies in transport (for example, Starkie, 1986; Small & Verhoef, 2007) because it involves redistributional tensions, which are discussed in more detail in Chapter 7 in this Handbook. Furthermore, the optimal toll could be higher when time valuations are heterogeneous: The toll might discourage travellers with low time valuations from travelling, leading to an increase in the average time valuation of the remaining travellers with the possible consequence that the marginal external congestion costs and hence the optimal toll increases (Fosgerau & Van Dender, 2013). Hall (2021) highlights that a pricing scheme that would involve half of the lanes could lead to a Pareto improvement in such a way that tolling

Transport pricing 

33

would make everyone better off. Czerny and Zhang (2011, 2015) consider the case where travellers with a high time valuation have a lower price elasticity of demand than those with a low time valuation. They show that this reduces the incentives to internalise self-imposed congestion, therefore increasing the importance of pricing policies in the presence of atomic users.

2.6 SOLUTION AND OPTIMISATION METHODS The previous sections highlighted the properties of optimal transport prices without putting much emphasis on the underlying solution and optimisation methods. Studies on optimal transport pricing range from highly stylised models that involve an individual link, a single transport mode, one externality, and one user type to large-scale simulation-based models of multi-modal transport networks with realistic traffic dynamics, and heterogeneous user types accounting for different types of externalities. Reviews of existing methods to derive optimal prices can, for instance, be found in Small and Verhoef (2007, Chapter 4), Tsekeris and Voβ (2009), and de Palma and Lindsey (2011). For a recent review with a focus on transport pricing in the presence of environmental externalities see Wang et al. (2018b). In the case of stylised models, the common aim is to derive closed-form expressions for the socially optimal price. In a first-best situation, this typically involves solving an unconstrained maximisation problem, with social welfare being the target variable. In second-best problems with constrained pricing instruments, the optimal prices are usually derived using Lagrangian techniques. Second-best pricing problems that account for distortions in other parts of the economy require general equilibrium models, often formulated as computable general equilibrium (CGE) models (see recent review by Robson et al., 2018).4 A large body of literature studies optimal tolls within networks. Only for stylised cases of small networks, can closed-form expressions for optimal tolls be derived. For networks, the optimisation problem is usually formulated as a mathematical programme with equilibrium constraints or a bilevel optimisation with the upper-level corresponding to the tolls being set optimally by the (public) authority, and the lower level corresponding to travellers’ decisionmaking (Tsekeris & Voß, 2009; Cheng et al., 2017).5 Deriving the tolls that support the system’s optimum flow pattern then requires solving a traffic assignment problem, in which for each origin–destination pair departure rates, route choices and the sequence of travellers need to be determined. This is particularly challenging when within-day and between-day traffic dynamics are taken into account, yielding tolls that vary by time of the day, or even instantaneously depending on traffic conditions (see the overview of model typologies by Bliemer et al., 2017). Solving dynamic traffic assignment models can either involve an analytical approach – mathematical programming, optimal control, and variational inequalities being the most relevant ones (e.g. Akamatsu, 2000; Friesz & Han, 2019; Long et al., 2016) – or simulations. Simulation-based approaches have gained popularity for being able to reproduce realistic traffic patterns as well as other externalities (Ben-Akiva et  al., 1998, Kaddoura & Nagel, 2018; Agarwal & Kickhöfer, 2018), but have also received criticism due to their “black box” character. Depending on the level of granularity, at which demand and supply are modelled, micro-, meso-, and macroscopic simulators can be distinguished. Well-known mesoscopic and dynamic simulators that endogenise mode, departure time, and route choice and are operational on large-scale networks include METROPOLIS (De Palma & Marchal, 2002;

34  Handbook on transport pricing and financing

de Palma et al., 2005, 2020) and MATSim (Horni et al., 2016). Among the macroscopic simulation approaches is the cell transmission model, which has been proposed by Daganzo (1995) and recently, for instance, been applied by Qin and Wang (2019). It is based on the idea that both road space and time can be discretised into “cells”. Empirically, link-specific prices are difficult to calculate in one step, as their computation requires knowledge of link-specific speed-flow curves, demand functions for each origin–destination pair, and the travellers’ value of time. All these required inputs are usually not readily available and cannot be known precisely. Due to these difficulties, so-called trial-and-error approaches have been suggested, in which the prices are set iteratively and adjusted based on observed (traffic) count data until the system optimum is reached. Most applications of trialand-error approaches refer to road transport (Yang et al., 2010; Zhao et al., 2015; Seo, 2020), but also public transport applications have recently gained interest (Wang et al., 2018a).

2.7 CONCLUSIONS This chapter described the theory of welfare-related pricing in the context of transport and its externalities. It described how standard static and dynamic approaches and their extensions are used to model pricing in transport markets and the different optimisation methods that are used to solve these models. Transport is an essential part of everyone’s life and the theory and methodologies described in this chapter shed light on how the pricing of transport services can be optimised to ensure that transport markets are efficient. Optimising the design of transport pricing schemes in the face of the economic challenges of our times such as climate change and social inequality, as well as disruptive technological developments such as autonomous vehicles, is expected to further advance modelling and solution approaches.

NOTES 1. The average costs of a trip also contain a constant reflecting the costs at free-flow conditions, which are not included in Figure 2.1 for simplification. 2. Emissions related to vehicle production or infrastructure production are ignored here as they are not a direct consequence of travel decisions. 3. The toll vector that decentralises the system optimum is generally not unique. This allows for the consideration of secondary objectives (in addition to welfare optimisation) such as the maximisation of toll revenues (e.g. Hearn & Yildirim, 2002). 4. Exceptions are papers that include the marginal costs of public funds (i.e. the welfare losses associated with raising additional taxes) in partial equilibrium models (e.g. Proost & van Dender, 2008). 5. When the lower-level problem is sufficiently regular, the bilevel optimisation problem can be transformed to a single-level constrained optimisation problem (e.g. Verhoef, 2002).

REFERENCES Adler, M. W., & van Ommeren, J. N. (2016). Does public transit reduce car travel externalities? Quasinatural experiments' evidence from transit strikes. Journal of Urban Economics 92, 106–119. Agarwal, A., & Kickhöfer, B. (2018). The correlation of externalities in marginal cost pricing: Lessons learned from a real-world case study. Transportation 45, 849–873.

Transport pricing 

35

Akamatsu, T. (2000). A dynamic traffic equilibrium assignment paradox.  Transportation Research Part B: Methodological 34(6), 515–531. Anas, A., & Lindsey, R. (2011). Reducing urban road transportation externalities: Road pricing in theory and in practice. Review of Environmental Economics and Policy 5, 66–88. Arnott, R., de Palma, A. & Lindsey, R. (1993). A structural model of peak-period congestion: A traffic bottleneck with elastic demand. American Economic Review 83, 161–179. Beckmann, M. J. (1965). On Optimal Tolls for Highway Tunnels and Bridges. L. Edie, R. Herman, R. Rothery, eds. Vehicular Traffic Science. Ben-Akiva, M., Bierlaire, M., Koutsopoulos, H., & Mishalani, R. (1998). DynaMIT: A simulation-based system for traffic prediction. In DACCORD Short Term Forecasting Workshop, 1–12. Bliemer, M. C., Raadsen, M. P., Brederode, L. J., Bell, M. G., Wismans, L. J., & Smith, M. J. (2017). Genetics of traffic assignment models for strategic transport planning. Transport Reviews 37, 56–78. Borenstein, S., 1989. Hubs and high fares: Dominance and market power in the U.S. airline industry. RAND Journal of Economics 20, 344–365. Brueckner, J. K. (2002). When carriers have market power. American Economic Revenue 92, 1357–1375. Carrion, C., & Levinson, D. (2012). Value of travel time reliability: A review of current evidence. Transportation Research Part A: Policy and Practice 46(4), 720–741. Cheng, Q., Liu, Z., Liu, F., & Jia, R. (2017). Urban dynamic congestion pricing: An overview and emerging research needs. International Journal of Urban Sciences 21, 3–18. Chu, X. (1995). Endogenous trip scheduling: The Henderson approach reformulated and compared with the Vickrey approach. Journal of Urban Economics 37(3), 324–343. Cohen, Y. (1987). Commuter welfare under peak-period congestion tolls: Who gains and who loses? International Journal of Transport Economics 14, 239–266. Coria, J., & Zhang, X. B. (2017). Optimal environmental road pricing and daily commuting patterns. Transportation Research Part B 105, 297–314. Crew, M. A., Fernando, C. S. & Kleindorfer, P. R. (1995). The theory of peak-load pricing: A survey. Journal of Regulatory Economics 8, 215–248. Czerny, A. I. (2010). Airport congestion management under uncertainty. Transportation Research Part B 44, 371–380. Czerny, A. I. & Zhang, A., (2011). Airport congestion pricing and passenger types. Transportation Research Part B 45, 595–604. Czerny, A. I. & Zhang, A. (2014). Airport peak-load pricing revisited: The case of peak and uniform tolls. Economics of Transportation 3, 90–101. Czerny, A. I. & Zhang, A. (2015). Third-degree price discrimination in the presence of congestion externality. Canadian Journal of Economics 48, 1430–1455. Daganzo, C. F. (1995). The cell transmission model, part II: Network traffic. Transportation Research Part B 29, 79–93. Daniel, J. I. (1995). Congestion pricing and capacity of large hub airports: A bottleneck model with stochastic queues. Econometrica: Journal of the Econometric Society, 327–370. De Palma, A. & Lindsey, R. (2020.) Tradable permit schemes for congestible facilities with uncertain supply. Economics of Transportation 21, 1–23. De Palma, A., & Lindsey, R. (2011). Traffic congestion pricing methodologies and technologies. Transportation Research Part C 19, 1377–1399. De Palma, A., Lindsey, R., & Monchambert, G. (2017). The economics of crowding in rail transit. Journal of Urban Economics 101, 106–122. De Palma, A., & Marchal, F. (2002). Real cases applications of the fully dynamic METROPOLIS tool-box: An advocacy for large-scale mesoscopic transportation systems. Networks and Spatial Economics 2, 347–369. De Palma, A., Kilani, M., & Lindsey, R. (2005). Congestion pricing on a road network: A study using the dynamic equilibrium simulator METROPOLIS. Transportation Research Part A 39, 588–611. De Palma, A., Vosough, S., & Lindsey, R. (2020). Pricing Vehicle Emissions and Congestion Using a Dynamic Traffic Network Simulator (No. 2020-09). THEMA (THéorie Economique, Modélisation et Applications), Université de Cergy-Pontoise.

36  Handbook on transport pricing and financing

Durán-Hormazábal, E., & Tirachini, A. (2016). Estimation of travel time variability for cars, buses, metro and door-to-door public transport trips in Santiago, Chile.  Research in Transportation Economics 59, 26–39. Fosgerau, M. & Small, K. A. (2013). Hypercongestion in a downtown metropolis. Journal of Urban Economics 76, 122–134. Fosgerau, M. & Small, K. A. (2017). Endogenous scheduling preferences and congestion. International Economic Review 58, 585–615. Fosgerau, M., & Van Dender, K. (2013). Road pricing with complications.  Transportation 40(3), 479–503. Friesz, T. L., & Han, K. (2019). The mathematical foundations of dynamic user equilibrium. Transportation Research Part B 126, 309–328. Hall, J. D. (2021). Can tolling help everyone? Estimating the aggregate and distributional consequences of congestion pricing. Journal of the European Economic Association 19, 441–474. Hau, T. D. (1992). Economic fundamentals of road pricing - A diagrammatic analysis. Policy Research Working Papers Transport, The World Bank, December 1992, WPS 1070. Hearn, D. W., & Yildirim, M. B. (2002). A toll pricing framework for traffic assignment problems with elastic demand. In Transportation and Network Analysis: Current Trends, 135–145. Boston, MA: Springer. Høye, A. K., & Hesjevoll, I. S. (2020). Traffic volume and crashes and how crash and road characteristics affect their relationship–A meta-analysis. Accident Analysis & Prevention 145, 105668. Hörcher, D., de Borger, B., Seifu, W., & Graham, D. J. (2020). Public transport provision under agglomeration economies. Regional Science and Urban Economics 81, 103503. Horni, A., Nagel, K., & Axhausen, K. W. (Eds.). (2016). The Multi-Agent Transport Simulation MATSim. Ubiquity Press. Kaddoura, I., & Nagel, K. (2018). Simultaneous internalization of traffic congestion and noise exposure costs. Transportation 45, 1579–1600. Keeler, T. E. & Small, K. A. (1977). Optimal peak-load pricing, investment, and service levels on urban expressways. Journal of Political Economy 85, 1–25. Knight, F. H. (1924). Some fallacies in the interpretation of social cost. Quarterly Journal of Economics 38, 582–606. Levinson, D. (2005). Micro-foundations of congestion and pricing: A game theory perspective. Transportation Research Part A 39, 691–704. Li, Z. C., Huang, H. J., & Yang, H. (2020). Fifty years of the bottleneck model: A bibliometric review and future research directions. Transportation Research Part B 139, 311–342. Long, J., Szeto, W. Y., Gao, Z., Huang, H. J., & Shi, Q. (2016). The nonlinear equation system approach to solving dynamic user optimal simultaneous route and departure time choice problems. Transportation Research Part B 83, 179–206. Mayeres, I., Ochelen, S., & Proost, S. (1996). The marginal external costs of urban transport. Transportation Research Part D 1, 111–130. Mayeres, I., & Proost, S. (2001). Marginal tax reform, externalities and income distribution. Journal of Public Economics 79(2), 343–363. Mohring, H. & M. Harwitz (1962). Highway Benefits: An Analytical Framework. Evanston, IL: Northwestern University Press. Mohring, H. (1972). Optimization and scale economies in urban bus transportation. American Economic Review 62, 591–604. Parry, I. W., & Bento, A. (2002). Estimating the welfare effect of congestion taxes: The critical importance of other distortions within the transport system. Journal of Urban Economics 51, 339–365. Parry, I. W., & Small, K. A. (2009). Should urban transit subsidies be reduced? American Economic Review 99, 700–724. Pels, E., & Verhoef, E. T. (2004). The economics of airport congestion pricing.  Journal of Urban Economics 55(2), 257–277. Pigou, A.C. (1920). The Economics of Welfare. London: Macmillan. Proost, S., & Van Dender, K. (2008). Optimal urban transport pricing in the presence of congestion, economies of density and costly public funds. Transportation Research Part A: Policy and Practice 42(9), 1220–1230.

Transport pricing 

37

Qin, Y., & Wang, H. (2019). Cell transmission model for mixed traffic flow with connected and autonomous vehicles. Journal of Transportation Engineering, Part A: Systems 145, 04019014. Ramsey, F. P. (1927). A Contribution to the Theory of Taxation. The Economic Journal 37(145), 47–61. Robson, E. N., Wijayaratna, K. P., & Dixit, V. V. (2018). A review of computable general equilibrium models for transport and their applications in appraisal. Transportation Research Part A 116, 31–53. Rouwendal, J., & Verhoef, E. T. (2006). Basic economic principles of road pricing: From theory to applications. Transport Policy 13, 106–114. Seo, T. (2020). Trial-and-error congestion pricing scheme for morning commute problem with day-today dynamics. Transportation Research Procedia 47, 561–568. Shires, J. D., & de Jong, G. C. (2009). An international meta-analysis of values of travel time savings. Evaluation and Program Planning 32, 315–325. Small, K. A. & Verhoef, E. T. (2007). The economics of urban transportation. Routledge. Small, K. A. (2015). The bottleneck model: An assessment and interpretation. Economics of Transportation 4, 110–117. Sovacool, B. K., Kim, J., & Yang, M. (2021). The hidden costs of energy and mobility: A global metaanalysis and research synthesis of electricity and transport externalities. Energy Research & Social Science 72, 101885. Starkie, D. (1986). Efficient and political congestion tolls. Transportation Research Part A 20A, 169–173. Steiner, P. O. (1957). Peak loads and efficient pricing. Quarterly Journal of Economics 71, 585–610. Szeto, W. Y., Li, X., & O’Mahony, M. (2008). Simultaneous Occurrence of the Braess and Emission Paradoxes. In Traffic and Transportation Studies, 625–634. Theofilatos, A., & Yannis, G. (2014). A review of the effect of traffic and weather characteristics on road safety. Accident Analysis & Prevention 72, 244–256. Tikoudis, I., Verhoef, E. T., & Van Ommeren, J. N. (2015). On revenue recycling and the welfare effects of second-best congestion pricing in a monocentric city. Journal of Urban Economics 89, 32–47. Tsekeris, T., & Voß, S. (2009). Design and evaluation of road pricing: State-of-the-art and methodological advances. NETNOMICS: Economic Research and Electronic Networking 10, 5–52. Van Essen, H. et al. (2019). Handbook on the External Costs of Transport, Version 2019. European Commission. Venables, A. J. (2007). Evaluating urban transport improvements: Cost–benefit analysis in the presence of agglomeration and income taxation. Journal of Transport Economics and Policy 41, 173–188. Verhoef, E. T. (2002). Second-Best Congestion Pricing in General Static Transportation Networks with Elastic Demands. Regional Science and Urban Economics 32, 281–310. Verhoef, E. T., & Mohring, H. (2009). Self-financing roads.  International Journal of Sustainable Transportation 3(5–6), 293–311. Verhoef, E., Nijkamp, P., & Rietveld, P. (1996). Second-best congestion pricing: The case of an untolled alternative. Journal of Urban Economics 40, 279–302. Verhoef, E. T. & Silva, H. E. (2017). Dynamic equilibrium at a congestible facility under market power. Transportation Research Part B 105, 174–192. Verhoef, E. T. & Small, K. A. (2004). Production differentiation on roads. Constrained congestion pricing with heterogenous users. Journal of Transport Economics and Policy 38, 127–156. Vickrey, W. S. (1969). Congestion theory and transport investment. American Economic Review 59, 251–260. Walters, A. A. (1961). The theory and measurement of private and social cost of highway congestion. Econometrica 29, 676–699. Wang, S., Zhang, W., & Qu, X. (2018a). Trial-and-error train fare design scheme for addressing boarding/ alighting congestion at CBD stations. Transportation Research Part B 118, 318–335. Wang, Y., Szeto, W. Y., Han, K., & Friesz, T. L. (2018b). Dynamic traffic assignment: A review of the methodological advances for environmentally sustainable road transportation applications. Transportation Research Part B 111, 370–394. Wardman, M., Chintakayala, V. P. K., & de Jong, G. (2016). Values of travel time in Europe: Review and meta-analysis. Transportation Research Part A 94, 93–111. Wardrop, J. G. (1952). Road paper. some theoretical aspects of road traffic research. Proceedings of the Institution of Civil Engineers 1, 325–362.

38  Handbook on transport pricing and financing

Weitzman, M. L. (1974). Prices vs. quantities. Review of Economic Studies 41, 477–491. Yang, H. (1999). System optimum, stochastic user equilibrium, and optimal link tolls. Transportation Science 33, 354–360. Yang, H., & Huang, H. J. (1998). Principle of marginal-cost pricing: How does it work in a general road network? Transportation Research Part A 32, 45–54. Yang, H., Xu, W., He, B. S., & Meng, Q. (2010). Road pricing for congestion control with unknown demand and cost functions. Transportation Research Part C 18, 157–175. Yin, Y., & Lawphongpanich, S. (2006). Internalizing emission externality on road networks. Transportation Research Part D 11, 292–301. Zhao, B., Bliemer, M., Yang, H., & He, J. (2015). A trial-and-error congestion pricing scheme for networks with elastic demand and link capacity constraints. Transportation Research Part B 72, 77–92. Ziegelmeyer, A., Koessler, F., My, K. B., & Denant-Boèmont, L. (2008). Road traffic congestion and public information: An experimental investigation. Journal of Transport Economics and Policy 42, 43–82. Zou, X. & Levinson, D., 2006. A multi-agent congestion and pricing model. Transportmetrica 2, 237–249.

3. Transport pricing beyond the social optimum1 Erik T. Verhoef

3.1 INTRODUCTION Much of the economics literature on transport pricing assumes that the network operator controlling tolls and capacities is what is sometimes referred to as a “benevolent dictator,” seeking and imposing policies that are designed to maximize overall social surplus. There are good reasons for making this assumption: it is a construct that helps us identify the efficient outcome in which social surplus is maximized, and in that way identifies a natural benchmark for policies. A sizeable amount of literature on second-best pricing has made abundantly clear that, indeed, the first-best optima that really achieve that benchmark are better not thought of as realistic policy scenarios, since the various constraints that are considered in these analyses of second-best pricing are far from exotic theoretical notions, but instead represent restrictions that policymakers face on a daily basis. Think of, for example, pre-existing distortionary taxes, the inability to differentiate tolls in an optimal fashion over all users, and the inability to price all links of or all users on a network, etc. Nevertheless, for such second-best pricing problems, it is still typically the objective of maximizing social surplus that is assumed to apply, irrespective of the additional constraints that are considered. This would still match the situation where the network under consideration is in the hands of the aforementioned benevolent dictator, even though this dictator now faces constraints. In reality, however, the possession of and control over networks may be organized differently. In this chapter, we will consider two important practical deviations from the benevolent dictator model and will look at operators who pursue objectives other than the maximization of overall social surplus, investigating the welfare implications relative to the situation considered so far. To maintain focus, we will consider the case of congestion pricing for road infrastructure in the analytic expositions, but where relevant we will also treat other modes in a verbal, qualitative manner. The first alternative institutional setup that we will consider in fact introduces the fundamental economic question of whether a public authority should set congestion tolls in the first place. As soon as road pricing becomes technically possible, then why not switch to a complete market approach and allow a private company to operate the road? Such privatization of roads is often suggested as an economically attractive way of dealing with congestion. It will, however, be shown why privatization is certainly not a magical cure for inefficiencies in road transport. The second case of interest concerns the situation where there are multiple governments at stake. In particular, we will discuss the theme of tax competition, which can be expected to arise when different governments represent different populations and tax bases overlap. Because transport by definition involves mobile agents, the chances of tax competition becoming relevant is far from imaginary. At the same time, tax competition is a topic that has a much broader relevance than to only transportation markets. So, the issues we discuss and insights 39

40  Handbook on transport pricing and financing

we derive and present are much more widely applicable than for the case of transportation markets alone. The final topic we address in this chapter concerns the case where a single public regulator is – or feels – constrained to heavily weigh in acceptability next to social surplus in the design and implementation of pricing instruments. Indeed, there is extensive literature assessing the paradoxical situation where policies that maximize social surplus – road pricing – meet with social resistance. An important factor in this debate is the simple observation that people dislike the imposition of new or higher taxes, irrespective of the notion that the receipts will or can be used by the government to finance valuable public goods or lower taxes elsewhere in the economy. This has led to literature on rewards, or subsidies, as well as (budget-neutral) tradable permits, as alternative price-based instruments to manage traffic congestion. These we review in Section 3.5.

3.2 THE FIRST-BEST REFERENCE While Chapter 2 in this Handbook introduces the theory of externality pricing in a mostly graphical way, for this chapter we will complement this with an analytical approach, which is better suited for the exposition of the cases we will be discussing. We begin by discussing the first-best case. The insights will be fully consistent with those from the graphical exposition in the previous chapter, but having the analytical expressions for this case will be helpful when comparing the later cases in this chapter. For the matters of interest in this chapter, it is sufficient to centre the discussion around the simplest possible setup where we consider a static model for a single road that is used by homogenous travellers. Obviously, extensions to more elaborate cases are of great academic and societal interest, but understanding the roles played by factors such as heterogeneity, dynamics, uncertainty or network interaction becomes easier if we have a basic case fully developed and understood. 3.2.1 Optimal Road Pricing for Congestion Externalities We begin by specifying the objective for the first-best case, which we assume will be social surplus S defined as the difference between total benefits B from the consumption of road trips and the total costs involved. For completeness, we distinguish two types of cost: variable generalized user cost C, which includes the valuation of travel time and therefore has an upward-sloping average cost when congestion prevails, and the cost of supplying road capacity Ccap, which is assumed to be given in the short run:

S = B - C - Ccap (3.1)

The Marshallian benefit measure that is used to determine social surplus is intimately related to the Marshallian demand function (which for a given income depicts the relation between market price and quantity demanded), and follows as the integral of inverse demand D(N): N



B( N ) =

ò D(n)dn (3.2) 0

Transport pricing beyond the social optimum  41

where N represents the number of travellers and D the marginal willingness to pay (in that way, D is indeed the inverse of the demand function that gives quantity N as the function of generalized price p). In our derivations, it will be convenient to work with the average and also the marginal user cost functions, ac and mc, which by definition satisfy: N



ò

C ( N ) = N × ac( N ) = mc(n)dn (3.3) 0

Note that Equation (3.3) assumes that there are no fixed user costs: with zero travellers, there is no user cost. If fixed user costs were relevant, the analysis would not change, but the interpretation of C would become the total variable user cost. Taking the derivatives of each of the three terms in Equation (3.3), we get another triple equality:

C ¢ ( N ) = ac ( N ) + N × ac¢ ( N ) = mc ( N ) (3.4)

The longer middle expression is central to understanding the essence of congestion pricing. It reflects that when the next user enters a congested road, two types of user costs are created. The first concerns the costs borne by the new user. These are simply equal to the prevailing average costs on the road. Note here that individual road users are treated as infinitesimally small so that average costs can for this purpose be simply evaluated at quantity N. Second, however, the travel time for all other users will increase. And this part of the marginal cost causes this to differ from the average cost. Because these additional costs are not borne by the person who creates them – the last user added – these costs constitute the marginal external cost (mec) of a trip:

mec = mc - ac = N × ac¢ (3.5)

The expression on the right gives the natural interpretation of the marginal external cost: the number of other road users (N), multiplied by the impact of a marginal user on average (per user) cost (ac). As already explained in Chapter 2 in this Handbook, the optimal road price in this case is equal to this marginal external cost. To see why, we can compare the first-order condition for optimizing Equation (3.1):

¶S = D ( N ) - ac ( N ) - N × ac¢ ( N ) = 0 (3.6) ¶N

with the equilibrium condition that with road pricing, the equilibrium is at the quantity where the marginal willingness to pay equals the generalized price or the sum of average cost plus the road price:

D ( N ) - ac ( N ) - r = 0 (3.7)

Subtracting Equation (3.7) from Equation (3.6) shows us that the optimal road use in our model can be realized by setting an optimal congestion charge, equal to the marginal external cost in the optimum:

42  Handbook on transport pricing and financing



r * = N * × ac¢( N * ) (3.8)

where superscripts * reflect that we are evaluating at the optimum (we will be dropping that reminder from this point onwards for brevity). Equation (3.8) tells the same story as the diagram discussed in Chapter 2 in this Handbook: the optimal road price in this setting is equal to the marginal external congestion cost evaluated in the optimum, consistent with the insights from Pigou (1920). If we had included other external costs in the objective in Equation (3.1), notably environmental externalities, by the same logic their marginal value would have become a second component in the optimal road price in Equation (3.8): it would have been part of the first-order condition in Equation (3.6), but not of the equilibrium condition in Equation (3.7), and therefore would have entered the difference between the two on the righthand side of Equation (3.8). Since the road capacity and the costs involved in supply are fixed in the short run, it does not enter our optimal pricing result directly. Indirectly, the choice of capacity does affect the optimal road price: it affects the curvature of ac(N), and the conventional operationalization of capacity in this type of model is such that a larger capacity is associated with a lower and flatter ac function, meaning also that mc will be lower and flatter, so that the optimum would involve a larger N and typically a lower r when capacity is increased. This is the indirect effect: the equilibrium value of r will change, but its optimality condition (or: the optimal tax rule) in Equation (3.8) remains the same. The reader familiar with the literature on the dynamic pricing of traffic bottlenecks (Vickrey, 1969; Arnott, de Palma and Lindsey, 1993) may wonder whether the tolls for the bottleneck model can somehow be reconciled with the Pigouvian toll in Equation (3.7). The answer is affirmative, but the reasoning behind it may not be immediately straightforward. It rests on the notion that in the model’s first-best optimum, the time-invariant generalized price is equal to the time-invariant marginal cost of adding one more user to the dynamic equilibrium. The generalized price consists of two time-varying components: the schedule delay cost, which is the only resource cost that a user incurs as travel delays are eliminated in the optimum, and the time-varying toll, which is not a resource cost but a transfer. This toll is therefore equal to the difference between the cost that is privately incurred, and the marginal cost. And that difference, indeed, is the time-varying marginal external cost. 3.2.2 Optimal Capacity Choice For the long-run problem, road capacity also becomes a variable of choice; we will call it cap. This requires the addition of capacity as an argument in the two cost components in our objective in Equation (3.1). Since the first-order condition in Equation (3.5) and the equilibrium condition in Equation (3.6) will remain unaltered (these are partial derivatives), also the optimal tax rule in Equation (3.8) remains the same. That is, for completeness, we should add the argument cap also in the optimal tax rule, to remind us that the derivative of ac with respect to N also depends on cap:

r = N × ac¢ ( N , cap ) (3.8′)

where for convenience we keep using the prime to denote the partial derivative with respect to N. We do, however, get a second optimality condition, namely for optimal capacity choice:

Transport pricing beyond the social optimum  43



¶ac ( ×) ¶Ccap ¶S = -N × = 0 (3.9) ¶cap ¶cap ¶cap

The interpretation is quite straightforward: the marginal benefits of capacity expansion in the first term of the middle expression consists of a decrease in average user cost multiplied by the number of users. It should be equal to the marginal cost of capacity expansion: the second term, after the minus sign. A perhaps somewhat unexpected result that we will show below for private road supply is that the profit-maximizing choice for optimal capacity boils down to the same first-order condition as the one shown in Equation (3.9). This does not mean that the road capacities chosen for social welfare maximization and profit maximization are the same: the different choice of toll affects the resulting value of N, at which Equation (3.9) is to be evaluated. Nevertheless, the exact match of the first-order condition remains remarkable. Equipped with Equations (3.8’) and (3.9), we can also derive the famous Mohring-Harwitz (1962) result on the self-financing of roads, which states that under certain technical conditions, the revenues from optimal road pricing – according to Equation (3.8’) – are just sufficient to cover the cost of the optimal supply of capacity – according to Equation (3.9). The technical conditions can be represented as follows: (i) capacity can be defined in units such that the congestion technology exhibits constant returns to scale: changing capacity and road use in the same proportion leaves travel times and therefore average generalized user cost untouched; (ii) for this definition of capacity, the supply of capacity takes place against neutral-scale economies: Ccap varies in fixed proportion with cap; and (iii) capacity can be supplied in continuous increments so that its value can be optimized in a marginal sense. The mathematical derivation of the result is provided in various textbooks, including Small and Verhoef (2007). The intuition is pretty straightforward. First, observe that in the long-run optimum the short-run average user cost, so in the firstbest short-run optimum treating the capacity as given, must be equal to the long-run average user cost. Second, observe that the assumed scale-neutrality of road supply in the assumptions in Equations (3.1) and (3.2) above means that the optimized amount of capacity varies in perfect proportion with the optimized number of users. The result is that the long-run average (per user) capacity cost is constant. Third, observe that, as a result, the long-run average user cost is also constant. Also the sum of these two long-run average cost components is therefore constant; and therefore so is the long-run marginal cost, as it is equal to that same sum. The self-financing result then follows from the notation that both the short-run and the long-run marginal cost functions intersect with inverse demand at the same point in the long-run optimum. The first observation above then implies that the toll must be equal to the average (per user) cost of capacity. And this means that, after multiplying both by the number of users, we find that the total toll revenues will be equal to the total cost of capacity. Empirical evidence suggests that the required scale-neutrality for exact self-financing may hold at least approximately in a range of circumstances (Small, 1992). The condition of the continuity of capacity typically does not hold for a single road, because the number of lanes is discrete. But capacity can still be varied by widening lanes, or by resurfacing. And at the scale of a road network, capacity may be almost perfectly divisible. The theorem may thus be highly relevant for practical policymaking. First of all, application of the theorem would help in achieving an overall efficient road system, both in terms of capacities and in terms of pricing. Second, the need to raise tax revenues from other sources

44  Handbook on transport pricing and financing

for the financing of roads may be firmly reduced. Efficiency may then increase even further, when these other taxes are distortionary. Moreover, problems with the public acceptability of road pricing may be reduced, because the resulting scheme may be perceived as “fair” (only the users of a road pay for the capacity, but do not pay anything more than that) and transparent (there are no “hidden” transfers surrounding the financing of roads). And finally, application of the theorem may lead to improved transparency in political decisions on infrastructure expansion. It can easily be demonstrated that if the technical assumptions are fulfilled, the capacity of a road should be expanded whenever currently optimal congestion pricing yields revenues per unit of capacity that exceed the unit (capital) cost of capacity. The market would thus indicate whether or not expansion is socially warranted, which will generally help improve the transparency and credibility of cost–benefit analyses. Up until this point, we have assumed that the network operator controlling tolls and capacities is what is sometimes referred to as a “benevolent dictator,” seeking and imposing policies that are designed to maximize overall social surplus. As stated in the introduction to his chapter (Section 3.1), there are good reasons for making this assumption: it is a construct that helps us identify an efficient outcome, in which social surplus is maximized. Nevertheless, in reality, the possession of and control over networks may be organized differently. In this chapter, we will consider two important practical deviations from the benevolent dictator model and will look at operators who pursue objectives other than the maximization of overall social surplus, investigating the welfare implications relative to the situation considered so far.

3.3 CONGESTION PRICING BY A PRIVATE OPERATOR The consideration of public roads is in accordance with the practice in most modern societies. Nevertheless, the possibility of private road tolls is receiving increasing attention. One practical motivation for this is simply the lack of sufficient public funds for (rapidly) financing the additional road capacity that some people consider necessary for securing an acceptable level of accessibility to major economic centres. A more fundamental motivation might be that roads have traditionally been supplied publicly because they satisfied the theoretical conditions for public goods: non-rivalness and non-excludability. A road’s capacity is non-rival in consumption before congestion sets in, and it is practically non-excludable in consumption unless a toll is charged – which would, however, be inefficient at a zero congestion level (at least, if we ignore other external costs of road transport such as pollution, noise annoyance and accidents). Clearly, when congestion becomes relevant, a road can no longer be classified as a pure public good, and this raises the question of whether it would not be appropriate to privatize the road, and let the free market find its optimum. Surely, a private operator would like to see revenues from its investment in road capacity, and one may wonder whether the required tolls would optimize congestion just as a public toll would do. 3.3.1 Revenue-Maximizing Pricing To compare private tolling to public tolling, it is instructive to start with the same situation as we did when studying public congestion tolls, namely with a single road of a given capacity. Recall that under these conditions, the public regulator would set a toll equal to the marginal external cost in the optimum as given in Equation (3.8). Would a private operator set the same

Transport pricing beyond the social optimum  45

toll, and thus reproduce the same efficient equilibrium? We can answer this question only after specifying the objective that the private operator would pursue. For a fixed capacity, and hence with fixed costs for the private operator, it is to be expected that the objective would be to maximize total toll revenues, because this will maximize total profit when all costs are fixed. If we denote the private operator’s toll with rp, total revenues R can simply be written as:

R = N × rp (3.10)

If N were given because demand was completely inelastic, the solution to the problem of maximizing total revenues would simply be to set the toll infinitely high. Unfortunately – for the private operator – demand will in general not be perfectly inelastic. A higher toll may increase the revenues per road user, but at the same time reduces the number of road users. One way of finding the revenue-maximizing toll under such circumstances is to incorporate the road users’ responses to different toll levels in Equation (3.10). We do this by substituting the condition that the toll must be equal to the difference between marginal benefits (mb = D) and average cost (ac) in the equilibrium with tolling. To remind ourselves that both D and ac will depend on N, we therefore rewrite (3.10) as:

R = N × ( D( N ) - ac( N ) ) (3.10’)

The problem faced by the private operator is then to find that particular N for which Equation (3.10’) is maximized. This means that we should determine its derivative with respect to N, and set it equal to zero:

( D( N ) - ac( N ) ) + N × ( D¢( N ) - ac¢( N ) ) = 0 (3.11)

Equation (3.11) was found by using the so-called “product rule” of differentiation.2 The first term between large brackets will in equilibrium be equal to the toll, and by taking the second term to the right-hand side we can thus find that the revenue-maximizing toll should satisfy:

rp* = N * × ac¢( N * ) - N * × D¢( N * ) (3.12)

(where * now indicates the optimum from the private operator’s perspective). A comparison between Equations (3.12) and (3.8) reveals an interesting result. The private operator in fact does internalize the congestion externality (the first component in the revenuemaximizing toll represents mec), but adds to this a positive term that is related to the slope of the demand function. (In Equation (3.12), note that D′ itself is negative – it is the slope of the inverse demand function – so that the entire second term, including the minus sign shown, is positive.) The appearance of this second term is easily understood once we realize that the private operator is in fact a monopolist supplying road capacity. The second term then corresponds with our micro-economic intuition that the monopolistic price increases as demand becomes less elastic; or: as the absolute value of D′ increases and hence the inverse demand function becomes steeper. Interestingly, the profit-maximizing toll for a congested bottleneck would be consistent with what is found in Equation (3.12), and is equal to the time-varying

46  Handbook on transport pricing and financing

component that also applies for the public operator, plus a time-invariant demand-related markup that is equivalent to the one in Equation (3.12). But why would the private operator also internalize the congestion externality? The intuitive explanation is that the internalization of the externality simply exchanges time losses on the road for toll revenues. And time losses do not add to the private operator’s profits, whereas toll revenues do. A more precise explanation is, however, that the term appears because the overall demand elasticity faced by the monopolist – which is the crucial variable in the determination of the profit-maximizing price – is in this case not only determined by the slope of the demand function, but also by that of the average cost function. And the required correction in the standard monopolistic pricing rule introduces exactly the mec (N·ac′) in the monopolist’s optimal toll expression, in addition to the demand-related term N·D′.3 Equation (3.12) thus shows that a private operator would typically charge a toll above the optimal level, unless demand is perfectly elastic and D′ is equal to zero. Privatization of roads may therefore indeed lead to tolls that internalize the congestion externality, but they typically do more than that because of the introduction of monopolistic power. This means that tolls will be charged that lead to an over-reduction in road use – just as monopolists in general would have an incentive to charge excessively high prices. 3.3.2 Profit-Maximizing Capacity We now turn to the case where the private operator would not only set a profit-maximizing toll, but also a profit-maximizing capacity. The objective then should include the capacity cost, and becomes:

(

)

P = N × D ( N ) - ac ( N , cap ) - Ccap (3.13)

It is now easily verified that the first-order conditions mimic Equation (3.11) for the optimal toll, and Equation (3.9) for the capacity. Again, the toll rule does not change when we move from a short-run to long-run analysis, and the interpretation is as before. Probably more surprising is that the investment rule is the same as for the welfare-maximizing problem. As explained above, because the toll is different this does not imply that the resulting capacities will be the same; on the contrary, that will only be the case when the two tolls are the same and that only occurs in the theoretical situation where demand is perfectly elastic. But why are the capacity rules the same? The interpretation rests on the observation that for any traffic volume N that the firm chooses, the toll it can charge varies on a euro-by-euro basis with the average user cost, as their sum should equal to a marginal willingness to pay at that level of N. Reductions in average user cost therefore translate directly and fully into increases in average revenue. And that means that whichever N the firm chooses, profits will be maximized if the sum is minimized for capacity cost, borne by the firm, and user cost, borne by the user but translating fully into foregone revenue for the firm. For the welfare-maximizing problem in Section 3.2, the task of minimizing the sum of those two cost components, given the level of use chosen, is exactly the same. 3.3.3 Franchising of Private Road Infrastructure Given that the pricing and capacity choice of private operators is generally not in line with social optimality, a relevant question is whether governments should try to intervene through

Transport pricing beyond the social optimum  47

direct regulation of tolls and/or capacities, or whether behaviour can be steered towards more desirable outcomes through the design of auctions that are used to select the firms that are allowed to operate certain toll roads. Contributions by Engel et al. (1997) and Verhoef (2007) consider the efficient design of such auctions from different perspectives and find that this design indeed can have important implications, which is not a surprise for those familiar with auction theory. Engel et al. (1997) focus on ways to avoid renegotiations under demand uncertainty; see also their contribution in Chapter 16 of this Handbook. Verhoef (2007) focuses on the optimization of capacity and toll through auction design and shows that, in an otherwise unpriced network and when the scale-neutrality conditions underlying the Mohring-Harwitz result are satisfied, a competitive auction that assigns the franchise to the bidder that offers the highest traffic volume in fact replicates the second-best optimal road in terms of capacity and toll.4 At the same time, auctions based on other selection criteria, even when sounding plausible at first sight, may be far from the second-best optimum. 3.3.4 Heterogeneous Preferences There is an important, somewhat hidden consequence of our treatment of profit-maximizing pricing in the context of a simple homogeneous-preferences model. As shown by Edelson (1971) and Mills (1981), while for the welfare-maximizing toll it is the average value of time that matters, for the optimal congestion toll, when there is heterogeneity in the values of time, the congestion component of the profit-maximizing toll will then depend entirely on the value of time of the marginal user(s). One could then even construct examples where the public toll would exceed the private one. This underlines that a profit-maximizing supplier cares about congestion only because it affects the marginal willingness to pay for using the road, and – unlike the welfare-maximizer – not because infra-marginal users are affected. 3.3.5 Networks Another context in which the different incentives for profit-maximizers versus welfare-maximizers become clearly visible is through their treatment of the utility of travellers who are not subject to the toll, but who are affected because of network spill-overs. An instructive example is the classic two-routes problem, where only one of two parallel lanes is priced so that the toll would, through the diversion of traffic, have the undesirable by-product of intensifying congestion on the untolled lane. While the public operator would set the second-best toll below the marginal external cost on the tolled lane in an attempt to optimally balance the reduction of congestion on the tolled link and its increase on the untolled link, the profitmaximizer would still fully internalize the marginal external cost on the tolled link and add a demand-related markup to that. The markup would be smaller than what is shown in Equation (3.12) though, because the relevant demand sensitivity is enhanced compared to the term D’ because of the availability of the untolled link. But the markup would remain non-negative (e.g. Verhoef, Nijkamp and Rietveld, 1996). 3.3.6 Oligopolistic Markets and Congested Facilities When more than one firm is present, competition of course affects the market outcomes. One example would be when different operators control different links in the same network. Consistent with the findings of Economides and Salop (1992) on the economic effects of

48  Handbook on transport pricing and financing

integration in network markets, it appears that the way in which different firms are organized in the network is crucial for the equilibrium welfare impacts of their uncoordinated pricing strategies. In a rather stylistic setting, where the welfare impacts are considered by increasing the number of symmetric firms controlling a single corridor in either a purely parallel fashion versus a purely serial fashion, Small and Verhoef (2007) show that for the parallel case, the equilibrium toll would asymptotically approach the socially optimal value as market power evaporates. The intuition, when comparing the outcome with the monopolistic toll of Equation (3.12), is that for each individual firm the demand appears to become more price sensitive with the smaller the portion of capacity it controls. In contrast, for the serial case, the resulting equilibrium toll summed over all operators approaches the prohibitive level at which demand is reduced to zero – a move away from Equation (3.12) in the “wrong” direction from the welfare perspective. The mechanism is similar to that of “double marginalization”: each individual firm has the incentive of fully internalizing the congestion externality of the full trip, and each firm applies a markup rule based on the sensitivity of the aggregate demand, so that the same markup – already too high when charged just once – will be charged multiple times. A quite different setup arises when multiple firms jointly use the same congestible facility. The classic example concerns the joint use of a congested airport by multiple airlines. For example, Brueckner (2002) has shown how in such cases an operator has an incentive to internalize congestion within its own firm, with the consequence that congestion tolls for firms would be lower than the standard toll in Equation (3.8), as only effects on other firms and their passengers remain to be internalized. Chapter 11 in this Handbook discusses the literature on airport congestion pricing in detail. 3.3.7 Price Discrimination A final issue concerns price discrimination. A first intuition may be that when market power leads to welfare losses due to inefficient pricing, a further exploitation of market power through price discrimination would most likely aggravate these losses. This, however, is typically not the case. In fact, under the extreme situation of first-degree price discrimination, where every unit can be sold against the relevant consumer’s marginal willingness to pay, the monopolist has the incentive to expand output up to the point where marginal benefit equals marginal cost – which is also the necessary condition for the optimization of social surplus. The reason is that the possibility of price discrimination implies that the price for infra-marginal consumers does not have to be lowered in order to capture the marginal consumer into the market. Marginal revenues then coincide with marginal benefits. Of course, while the resulting equilibrium is efficient, the distributional implications will often be considered undesirable, as the entire consumer surplus is skimmed off and is turned into profits for the firm. In cases where undifferentiated marginal cost pricing would result in losses because of economies of scale and scope, price discrimination may be more acceptable as it allows coverage of total cost while limiting the distortive consequences of pricing above marginal cost. For instance, for public transport, this may be an interesting possibility, although the required sophistication in the price discrimination may well have to go beyond what is possible with more traditional types of travel passes and season tickets (e.g. Hörcher et al., 2018, find these to be rather inefficient in the face of crowding externalities). In other cases, however, the distributional implications will often be considered as undesirable.

Transport pricing beyond the social optimum  49

For more realistic types of price discrimination, notably second-degree where prices are nonlinear and third-degree where prices can only be varied between groups of consumers, it is not the case that price discrimination would always increase social welfare, but it may very well still do that. Varian (1985), for example, argues that a necessary condition for this to be true is that output increases under price discrimination. Specific to transport, Wang, Lindsey and Yang (2011) establish that second-degree profit-maximizing road pricing, with an access fee on top of a use fee, increases profit and may increase or decrease social surplus, compared to use-based profit-maximizing tolls. This confirms that the welfare impact of allowing price discrimination by profit-maximizers may both be positive and negative, depending on the circumstances. This issue is of importance for transport, as price discrimination is quite common and is likely to become only more important when online marketing gives transport operators increasing knowledge of individuals’ preferences. For public transport, customers can, for instance, easily be separated by age, which suggests that discounts for the elderly – with a typically more elastic demand for public transport – may be part of a profit-maximizing strategy rather than just a friendly gesture. Because transportation products are not storable, it is often also possible to separate by trip motive: inelastic demand by commuting peak travellers can easily be exploited by using discount tariffs outside the peak – although of course not every price differential between peak and off-peak tickets needs to reflect genuine price discrimination, as the marginal cost may also differ when capacities are binding in the peak and non-binding outside the peak. And the use of multiple classes in trains also allows price discrimination. Surely the marginal cost may again differ when first-class seats offer more convenience. But a derived benefit for the profit-maximizing firm is that it will typically be business travellers and higher income groups that will choose first class, and both groups can be expected to be relatively price insensitive. Ticket prices can thus be differentiated more strongly between classes than what would be appropriate on the basis of cost considerations alone. Also, in aviation, price discrimination is extensively practised. The use of multiple classes on offer, for the same reason as for public transport, gives the possibility to separate by income groups and trip purpose. Another possibility arises because airline tickets are usually bought in advance. Tourists tend to plan their trips earlier and to be more price sensitive than business travellers, which is one reason why it makes sense to offer lower tariffs for early bookings. At the same time, few business travellers would like to run the risk of missing an important meeting because a flight is fully booked, which explains why it may also be sensible to sell the last seats cheaply on a last-minute basis – a sub-market that is certainly for some destinations populated primarily by highly price-sensitive back-packers. Furthermore, most tourists would appreciate spending a weekend at the destination of the trip, whereas experienced professional travellers often do not want to lose a weekend at home every time they make a trip. Most airlines therefore find it profitable to charge different prices, depending on whether a Saturday night is spent at the destination.

3.4 TRANSPORT PRICING BY MULTIPLE GOVERNMENTS Another quite different reason why the overall objective pursued by the agency setting the toll may not be social welfare occurs when there are different governments, each representing its

50  Handbook on transport pricing and financing

own constituency. Even if in such a case each government acts so as to maximize social surplus for the inhabitants of its own territory, the combined result of all actions undertaken by all governments is likely not to be maximizing aggregate surplus; that is, social surplus enjoyed over all territories jointly. There can be a variety of mechanisms at work that may cause such failure to achieve the best possible outcome. Different mechanisms may in turn lead to deviations in different directions from what would be chosen by a “global benevolent dictator,” i.e. a global government that maximizes social surplus over all territories combined. For example, taxes may be set too high or too low from the overall social perspective – but, obviously, in either case, the aggregate social surplus will be below what is achieved under globally optimal policies. When such instances occur, there will potentially be a strong case for the coordination of policies between governments, since at least in theory it must be possible to distribute the aggregate welfare gains from optimal coordination in such a way that everybody is better off. Still, governments may continue to deviate from what would be globally optimal, and more strongly so the more the voting behaviour in their territory depends on the extent to which the government succeeds, through its policies, in raising local social surplus enjoyed by these voters. Indeed, it may well be partly the pressure from democratic voting in their own jurisdiction that pushes a local government away from policies that would be more in line with global objectives, towards satisfying more local interests. Quite intuitively, the stronger the links between territories served by different governments, the stronger the chances that the combined result of the policies implemented by the different governments will deviate from what is globally optimal. That is, imagine an extreme situation where territories are completely isolated from each other, in each possible respect. The optimization problems faced by the local governments will then be entirely independent, and independent optimization by each government would produce not only local optima, but also the overall global optimum. But as soon as the optimization problems become mutually dependent, risks of inefficiencies through the absence of policy coordination emerge. Because transport by definition means that agents involved are mobile and may thus enter territories other than their home territory, the chances of inefficiencies of this type occurring in transportation markets are far from imaginary. Still, it is by no means a challenge facing transport markets alone. For example, one of the seminal contributions to the literature on “tax competition” addressed the case where governments compete to attract (internationally) mobile capital to their region, which can subsequently be taxed to finance the supply of a local public good. Setting a lower capital tax rate may then attract so much more capital to the region that the total tax revenues in fact increase (Oates, 1972). This gives an incentives to set a relatively low tax rate, which is reinforced when competing governments do the same, so that a “race to the bottom” may occur, where relatively low taxes are applied, and public goods are in all regions supplied in lower qualities and quantities than what would be socially optimal. Anyone who has driven south from the north of Europe to more sunny holiday destinations is likely to recognize that a similar mechanism appears to be at work in Luxemburg, but then in a transport setting. Fuels in Luxemburg are relatively cheap compared to fuel prices in the neighbouring countries, the reason being that a lower tax rate is applied. Behind this lower tax rate is Luxemburg’s recognition that this will attract so much more traffic that it boosts total tax revenues, even though the tax revenues per litre are somewhat lower. Phrased differently, with a fuel tax equal to that of the neighbouring countries, demand will be so elastic that it is simply too hard to resist the temptation to lower the fuel tax, and in that way raise revenues.

Transport pricing beyond the social optimum  51

At the same time, the neighbouring countries – France, Germany, Belgium – face other objectives with their (nationally determined) fuel taxes, above competing for mobile taxpayers with Luxemburg in one corner of their territories, and therefore this situation appears a logical and stable outcome of the process of tax competition. Indeed, transport is just one example of an economic activity in which tax competition and other types of coordination problems and strategic interactions between different governments can arise, and the literature on tax competition encompasses many different sectors (Wilson, 1999, provides a broad review). At the same time, there are more types of strategic interactions between governments than only the setting of taxes, making the competition between governments in reality often a complex and multidimensional process. Moreover, governments can be organized in different ways and may have different types of hierarchical relations between them. The outcomes of the strategic interactions between governments will generally strongly depend on the type of relation between the governments involved, and on the type of interactions that is at stake. It is therefore useful to present a structured overview of the types of interactions that may exist between jurisdictions, and that is what the next sub-section will do. 3.4.1 Different Types of Policy Interactions: A Taxonomy Table 3.1, adapted from De Borger and Proost (2012), presents the most relevant types of strategic interactions that can occur between governments in the context of transport policymaking, also identifying the likely implications on policy choices. A first important distinction is between horizontally versus vertically ordered governments, where the former refers to governments with non-overlapping areas and populations (e.g. neighbouring countries), and the latter to instances where there is such an overlap (e.g. a national government versus the local government of one of its cities). In general, because with vertical interactions there is at least some overlap in voting populations served, one may expect deviations from globally optimal policies to be larger under horizontal interactions. But this certainly does not always have to be the case. A second distinction is between types of externalities between jurisdictions. A fiscal externality occurs when tax setting by one government affects tax revenues for the other government. An expenditure externality occurs when expenditures such as infrastructure investments by one government affects the well-being of the population served by the other government. And finally, an environmental externality refers to a situation where environmental effects spill over into other jurisdictions. In this way, six different categories are distinguished in Table 3.1. The first of these, the case of a horizontal fiscal externality, can in turn be subdivided into two distinctly different types of interactions. The fact that the likely implications on tax levels are opposite is illustrative of the variety of impacts that strategic interactions between jurisdictions may have, and therefore the care that should be taken when analyzing the economics of such interactions. The first mechanism concerns tax exporting, and this refers to the desire to make foreigners pay taxes. Especially when foreigners are captive and have no alternative, this will lead to an upward bias on taxes. A good example of tax exporting in the context of road pricing is the toll on German roads that was scheduled to be introduced in 2016. Where both foreign and German drivers are required to pay a toll by acquiring a ten-day, two-month or annual pass, German drivers would be fully compensated through a simultaneous reduction in annual vehicle taxes. Effectively, therefore, the revenues from the toll system would be entirely due to foreign

52  Handbook on transport pricing and financing

Table 3.1  Possible types of interactions in transport markets between jurisdictions Type

Source

Transport example

Likely implications

Horizontal fiscal externality

Tax exporting: make outsiders pay taxes Tax competition: attract a mobile tax base

Higher tolls for foreign road users Lower fuel tax to attract foreign drivers (Luxemburg)

Upward pressure on taxes Downward pressure on taxes

Horizontal expenditure externality

Benefit spill-over Expenditure competition

Underinvestment in infrastructure used by foreigners Overinvestment to attract foreign firms

Downward pressure on investments Upward pressure on investments

Horizontal environmental externality

Pollution spill-overs

Trans-boundary pollution, such as global warming from CO2 emissions

Downward pressure on environmental regulations and taxes

Vertical fiscal externality

Overlapping tax bases

National fuel taxes and local parking fees used to raise revenues

Upward pressure on combined tax levels

Vertical expenditure externality

Expenditure interdependence

Spending on local roads raises use and therewith national fuel tax revenues

Downward pressure on such expenditures

Vertical environmental externality

Pollution spill-overs

Trans-boundary pollution between lower-level jurisdictions positioned under the same higher-level government

Dampening the bias from a purely horizontal environmental externality

Source:   adapted from De Borger and Proost (2012).

drivers, and this discrimination is why the plans were foiled by the European Court of Justice (ECJ) in 2019. The second type of horizontal fiscal externality, tax competition, becomes relevant when the tax base (the individuals paying the tax) is mobile and can therefore escape paying the tax by adjusting behaviour. The earlier example of fuel taxes in Luxemburg fits this case, and shows that in contrast to what we saw for tax exporting, there will be a downward bias on the tax level. Also, in the case of a horizontal expenditure externality, there may be opposing effects. With a benefit spill-over, foreigners also use the facility under consideration, and this may lead to a downward bias in the incentives to invest in its quality and maintenance, especially when foreigners have unpriced access. The mediocre or poor quality of untolled roads that are used relatively intensely by foreigners is a case in point. But under expenditure competition, in contrast, jurisdictions try to attract firms or visitors by providing excellent infrastructure, and in this case an upward bias in the incentives to invest in quality and maintenance can be expected. A good example is the oversupply of commercial areas (“bedrijventerreinen”) and associated infrastructures, offered by Dutch municipalities in the hope of attracting firms and therewith employment. Next, with a horizontal environmental externality, emissions caused within a certain jurisdiction also affect environmental quality outside that area. The emission of CO2, affecting

Transport pricing beyond the social optimum  53

global warming, provides an extreme example. To the extent that the local government seeks to promote local social welfare, the environmental effects occurring outside the jurisdiction tend to be overlooked in the formulation of environmental policies. Indeed, for a perfectly global environmental externality such as global warming, the appropriate description of the incentives faced by local governments would be that of a public good (or bad), where the economic incentive would be to free-ride on others’ contributions. More in general, the larger the share of effects of emissions occurring outside the jurisdiction, the larger the incentive for the local government to make less than a globally optimal effort to restrict the externality. The basic mechanisms for the three types of vertical externalities distinguished in Table 3.1 are similar, but, as stated, because populations are partly overlapping with the interests of the two governments are typically somewhat better aligned than for the horizontal externalities just discussed. But there remain inefficiencies. For a vertical fiscal externality, for example, the joint effect of both governments seeking to distract funds from the same individuals through taxation may mean that the aggregate tax becomes so high that a reduction would increase aggregate revenues. For a vertical expenditure externality, a local government may ignore that part of the benefits of induced traffic from an infrastructure investment accrues to the central government in the form of, for example, fuel tax revenues. Counting these as a loss at the local level, rather than as a transfer, would underestimate the aggregate benefits of the project and hence may lead to underinvestment. And finally, with a vertical environmental externality the underlying mechanisms are comparable to those for a horizontal environmental externality, with the difference that the higher-level government may now succeed in dampening the bias from a purely horizontal environmental externality. 3.4.2 Diverging Incentives with Horizontal Externalities: An Example We may finally further illustrate the issues at stake by considering an example with two types of horizontal externalities: fiscal and environmental. Consider a world that consists of two regions only, A and B. There is a road in region A that is used by drivers from both regions, the quantities being NA and NB. For both groups, there is an inverse demand function denoted DA(NA) and DB(NB), and assuming that drivers are otherwise identical we may write the average cost as ac(N) with N=NA+NB; i.e. there is congestion and both groups contribute to it in equal amounts. And finally, there is an environmental effect: every trip causes damage eA in region A, and eB in region B. Note that the superscripts A and B therefore do not refer to the region of the origin of the driver: that is assumed to be immaterial for the environmental externality caused. Finally, we assume that, if desired, tolls can be differentiated between drivers from regions A and B. To see the possible impacts of the horizontal externalities upon public policies, let us first write out the local social surplus SA, and compare it to the global social surplus, S. To start with the latter, this consists of the sum of benefits, over all users, minus user costs for all users, minus all external environmental costs: NA



S=

ò 0

NB

DA (n)dn +

ò D (n)dn - N × ac(N ) - N × (e B

0

A

+ e B ) (3.14)

54  Handbook on transport pricing and financing

The local government, in contrast, maximizes local social surplus SA. This consists of total benefits for local inhabitants, minus user costs incurred by these same drivers, minus locally incurred environmental costs, plus tax revenues extracted from foreign drivers: NA



SA =

ò D (n)dn - N A

A

× ac( N ) - N × e A

0

(3.15)

+ N B × ( DB ( N B ) - ac( N ) ) Maximization of global surplus requires us to take the derivative of Equation(3.14) with respect to both levels of use, and this leads to:



¶S = DA ( N A ) - ac( N ) - N × ac¢( N ) - (e A + e B ) = 0 ¶N A (3.16a) ® rA = N × ac¢( N ) + (e A + e B )



¶S = DB ( N B ) - ac( N ) - N × ac¢( N ) - (e A + e B ) = 0 ¶N B (3.16b) ® rB = N × ac¢( N ) + (e A + e B )

Note that to obtain the toll expressions in the second line, we substitute r = D–ac. Both tolls are set equal to marginal external costs, entirely following the logic outlined in Section 3.2. And because the marginal external costs are independent of the origin of a driver, the tolls are equal, too. The incentives for the local government maximizing SA can be identified by maximizing Equation (3.15):



¶S A = DA ( N A ) - ac( N ) - N A × ac¢( N ) - e A - N B × ac¢( N ) = 0 ¶N A (3.17a) ® rA = N × ac¢( N ) + e A ¶S A = - N A × ac¢( N ) - e A + DB ( N B ) - ac( N ) ¶N B



(

)

+ N B × DB¢ ( N B ) - ac¢( N ) = 0

(3.17b)

® rB = N × ac¢( N ) + e A - N B × DB¢ ( N B ) The tolls that the local government sets reflect that it behaves as the social welfare optimizer from Section 3.2 insofar as local drivers are concerned, and as the profit-maximizer from Section 3.3 towards foreign drivers. Interestingly, this implies that both types of drivers face the full marginal external congestion costs N∙ac′(N). This is a direct consequence of what we

Transport pricing beyond the social optimum  55

found in Section 3.3: also a profit-maximizer perfectly internalizes congestion externalities, albeit for different reasons than those applying for a welfare-maximizer. For foreign drivers, also the profit-maximizing demand-related markup –NB∙D´B that we found in Section 3.3 is again added. And finally, for both groups only the locally incurred environmental cost eA is included, missing out on the damage abroad eB. We see that, as a consequence, the toll for local drivers is below what would be globally optimal, the difference between Equations (3.16a) and (3.17a) being equal to eB. For foreign drivers, the toll may be either below or above the globally optimal one, depending on whether eB is smaller or bigger than –NB∙D´B. This illustrates how the effects of horizontal externalities on tax levels can be varied, and may even change significantly depending on the circumstances. The above example is instructive in that it shows how tax competition may indeed seriously affect outcomes in transport markets, and also how the interpretation of results may benefit by comparing these to what is known from public pricing and private pricing. Still, this is obviously just one, very stylized example. The literature on tax competition in transport is rich and growing, and has considered many cases and situations, each with their own peculiarities. De Borger and Proost (2012) provide a nice overview. As their review makes clear, the outcomes under tax competition depend on many aspects, including the network configuration (notably, serial versus parallel links); the game-theoretic setup (including the distinction between Nash versus Stackelberg behaviour of governments); and the set of policy instruments considered (for example, pricing and/ or capacity choice). So the example just given should be treated in that light: it is just an example.

3.5 PRICING, REWARDING AND BUDGET-NEUTRAL INCENTIVES The notion of marginal cost pricing as the instrument that achieves the highest efficiency and maximum social welfare is well-established in the literature, but in practice things typically do not resemble the stylized assumptions that need to be made to make the result indeed apply. In the foregoing, we have already briefly touched upon various types of second-best pricing, for which it is generally true that given the constraints that apply, a higher social welfare can be achieved when deviations are made from the classic Pigouvian tax rule equal to marginal external cost. The two distinct types of reason for why second-best pricing is so relevant in practice are: (1) inherent constraints on the pricing instrument itself, such as the inability to differentiate taxes optimally over time, place, user types, etc.; or (2) non-optimal prices in markets that are interacting with the transport market under consideration, a good example being the case considered by Parry and Bento (2001) where commuting involves workers who are subject to distortive labour taxation. Another type of constraint, and one that has played an important role in preventing widescale implementation of road pricing in practice, concerns limited social and political acceptability. An important factor behind the limited acceptability is believed to be the general tendency of people to dislike higher or additional taxes, despite the consequence that the revenues could be used to finance useful public expenses or reduce public debts. This has given rise to a search for more acceptable price incentives, two important ones being rewards and tradable permits. Rewards have been tested empirically in a number of experiments in The Netherlands, often under the label Spitsmijden (Peak Avoidance). The general insight drawn from such experiments is that participants seem very sensitive to financial incentives that are designed

56  Handbook on transport pricing and financing

to make them reduce their peak travel behaviour, where reductions in the order of 50% are no exception (e.g. Knockaert et al., 2012). An important caveat is that such projects tend to attract relatively flexible travellers, who expect to be able to earn relatively high rewards, so the results should not be mistaken to be representative. This seems consistent with the findings of Graham et al. (2020), who found relatively small effects from a 25% early peak shoulder reduction of transit fares in Hong Kong. Another caveat is that subsidies to combat external costs are not the first-best solution and in the long run may attract too many travellers to the road and to the scheme. And a third caveat is that a wide-scale and structural application of rewards to steer transport behaviour would require huge public budgets, which would need to be financed from taxes that in itself would be distortionary too. These considerations have spurred research into budget-neutral price incentives, avoiding a net financial flow from the government to travellers, as with rewarding, or in the reverse direction as with pricing. An important example is tradable permits. Originally proposed as an instrument to deal with environmental externalities (e.g. Dales, 1968) and also used as such in the well-known European Emissions Trading Scheme, it has also been considered as a possible instrument in the containment of road traffic externalities (Verhoef et al., 1997). Modelling exercises have focused on aspects such as the implementation of the instrument in a transport network environment (Yang and Wang, 2011), or under conditions of uncertainty (De Palma et al., 2018). The actual implementation would assume that travellers understand the instrument as intended, and that the market for permits can be designed such that it operates in an efficient way – issues that have been tentatively explored in experimental studies (e.g. Brands et al., 2020). It seems too early to make any definitive statements on the prospects for tradable permits in the regulation of transport externalities, but the theoretical properties of combining the efficiency of incentives with budget-neutrality, and the first experiences in experimental settings seem sufficiently promising to justify the continuation of research on this topic.

3.6 CONCLUSION This chapter has reviewed the theory of transport pricing under different objectives than the textbook reference of first-best optimization of social welfare. We have discussed profit maximization pricing, pricing by multiple governments and – albeit briefly – price incentives that are constrained to be positive or budget-neutral. We have emphasized that although different objectives naturally lead to non-optimal prices, there remain strong conceptual links with insights from the theory of first-best pricing. For example, a profit-maximizing operator has an incentive to internalize congestion externalities, while our example of tax competition produced pricing rules that balance elements from surplus-maximizing and profit-maximizing pricing. Despite the beneficial impacts that optimized prices might bring, application to apply these remains scarce, in particular in road transport. Especially then, economic theory can be helpful in identifying ways of providing financial incentives that, under constraints of public and political acceptability, produce maximum efficiency. Budget-neutral instruments such as tradable permits may have an important role to play there. Important challenges for near-future policymaking, but also transport economics research, including the design, evaluation and implementation of innovative price incentives that employ novel technologies in terms of price differentiation and (first-, second- and third-degree)

Transport pricing beyond the social optimum  57

discrimination to better meet objectives and/or constraints in terms of behavioural change, equity and acceptability. Some current developments make this even more pressing than before. An obvious first one is the increasing awareness of, and commitment to counter, global warming. A second one is the predicted rapid growth in Mobility-as-a-Service facilities and sharing-economy-based services, be it from private operators with their own objectives or from public operators, and be it in monopolistic or in strategic oligopolistic supply. This leads to a much more complex and volatile environment in which prices are formed, and in which socially optimal pricing cannot be expected to arise spontaneously. And third is the continuing concern with urban congestion and its likely return to pre-COVID-19 levels after the pandemic has calmed. Carefully designed pricing schemes to address these matters, fully exploiting the possibilities that contemporary pricing technologies offer, are not the only or full answer to combat the grand challenges in urban transportation – but ignoring the importance and potentials of pricing would undoubtedly make that combat much more so an uphill battle.

NOTES 1. The expositions in some sections of this chapter closely follow presentations in an earlier text that is made freely available as a reader for Bachelor students at VU Amsterdam (Verhoef, 2002). Section 3.2 uses the same notation as used in that reader, but presents the material differently. Sections 3.1 and 3.4 are taken directly from that source, with at most only minor textual adjustments. The same texts served as the basis for a textbook that recently appeared in Chinese as Verhoef, Wang and Hu (2020). 2. The product rule of differentiation tells us that the derivative of a composite function f(x) = h(x) ⋅ g(x) with respect to x is equal to h(x) ⋅ g′(x) + g(x) ⋅ h′(x). The reader may test this rule for the simple function f(x) = x2 = x ⋅ x. 3. The standard monopolistic price rule, with the operator’s mc = 0, would be: r · (1 + 1/ε) = 0 (where ε is demand elasticity). The demand elasticity ε is generally defined as dN/drp·rp/N. If ac were constant, the first term dN/drp would be equal to 1/D′ (because D′ would then give drp/dN). With a non-constant ac, however, dN/drp becomes equal to 1/(D′–ac′), because the slope of the ac function makes N less responsive to marginal changes in rp than what would appear from the demand function alone. Put differently: the actual inverse demand function faced by the monopolist, with rp rather than p on the vertical axis, can be constructed as the vertical distance between D and ac and thus reads D–ac, so that its slope is given by D′–ac′. We thus find: ε = 1/(D′–ac′) ⋅ rp/N and, therefore, 1/ε = (D′–ac′)·N/rp. We can then finally rewrite the standard rule as rp + N ⋅ (D′–ac′) = 0. This exactly replicates (12). 4. The fact that the network is otherwise unpriced makes this a second-best situation: the second-best optimal capacity and toll accounts in the best possible way for the distortions occurring elsewhere in the network. Should the network be otherwise optimally dimensioned and priced, then the auction replicates the first-best capacity and toll. That is, in all circumstances does it produce the most efficient solution available.

REFERENCES Arnott, R., A. De Palma and R. Lindsey (1993) “A structural model of peak- period congestion: A traffic bottleneck with elastic demand.” American Economic Review 83 (1), 161–179. Brands, D. K., E.T. Verhoef, J. Knockaert and P.R. Koster (2020) “Tradable permits to manage urban mobility: Market design and experimental implementation.” Transportation Research Part A: Policy and Practice 137, 34–46.

58  Handbook on transport pricing and financing

Brueckner, J.K. (2002) “Airport congestion when carriers have market power.” American Economic Review 92, 1357–1375. Dales, J. (1968) “Land, water, and ownership.” The Canadian Journal of Economics/Revue Canadienne d’Economique 1 (4), 791–804. De Borger, B. and S. Proost (2012) “Transport policy competition between governments: A selective survey of the literature.” Economics of Transportation 1, 35–48. de Palma, A., S. Proost, R. Seshadri and M. Ben-Akiva, M. (2018) “Congestion tolling-dollars versus tokens: A comparative analysis.” Transportation Research Part B: Methodological 108, 261–280. Economides, N. and S.C. Salop (1992) “Competition and integration among complements, and network market structure.” Journal of Industrial Economics 40, 105–123. Edelson, N.E. (1971) “Congestion tolls under monopoly.” American Economic Review 61 (5), 872–882. Engel, E., R. Fisher and A. Galetovic (1997) “Highway franchising: Pitfalls and opportunities.” American Economic Review, Papers and Proceedings 87 (2), 68–72. Graham, D. J., D. Hörcher, R.J. Anderson and P. Bansal (2020) “Quantifying the ex-post causal impact of differential pricing on commuter trip scheduling in Hong Kong.” Transportation Research Part A: Policy and Practice 141, 16–34. Hörcher, D., D.J. Graham and R.J. Anderson (2018) “The economic inefficiency of travel passes under crowding externalities and endogenous capacity.” Journal of Transport Economics and Policy 52 (1), 1–22. Knockaert, J. S. A., Y. Tseng, E.T. Verhoef and J. Rouwendal (2012) “The Spitsmijden experiment: A reward to battle congestion.” Transport Policy 24, 260–272. Mills, D.E. (1981) “Ownership arrangements and congestion-prone facilities.” American Economic Review, Papers and Proceedings 71 (3), 493–502. Mohring, H. and M. Harwitz (1962) Highway Benefits: An Analytical Framework. Evanston, IL: Northwestern University Press. Chapter II: Benefits and the tax system (pp. 70–90). Oates, W. (1972) Fiscal Federalism. New York: Harcourt Brace Jovanovich. Parry, I.W.H. and A. Bento (2001) “Revenue recycling and the welfare effects of road pricing.” The Scandinavian Journal of Economics 103 (4), 645–671. Pigou, A.C. (1920) Wealth and Welfare. London: Macmillan. Small, K.A. and E.T. Verhoef (2007) The Economics of Urban Transportation. London: Routledge. Varian, H.R. (1985) “Price discrimination and social welfare.” American Economic Review 75, 870–875. Verhoef, E.T. (2002) Markets and Governments: Transport Economic Applications. Reader, Vrije Universiteit Amsterdam. Verhoef, E.T. (2007) “Second-best road pricing through highway franchising.” Journal of Urban Economics 62, 337–361. Verhoef, E.T., P. Nijkamp and P. Rietveld (1996) “Second-best congestion pricing: The case of an untolled alternative.” Journal of Urban Economics 40 (3), 279–302. Verhoef, E., P. Nijkamp and P. Rietveld (1997) “Tradeable permits: Their potential in the regulation of road transport externalities.” Environment and Planning B: Planning and Design 24 (4), 527–548. Verhoef, E.T., Y. Wang and Y. Hu (2020) Markets and Governments: Transport Economic Theory and Applications (in Chinese). China: Social Sciences Academic Press. Vickrey, W.S. (1969) “Congestion theory and transport investment.” American Economic Review 59, 251–260. Wang, J. Y., R. Lindsey and H. Yang (2011) “Nonlinear pricing on private roads with congestion and toll collection costs.” Transportation Research Part B: Methodological 45 (1), 9–40. Wilson, J. (1999) “Theories of tax competition.” National Tax Journal 52, 269–304. Yang, H. and X. Wang (2011) “Managing network mobility with tradable credits.” Transportation Research Part B: Methodological 45 (3), 580–594.

4. Pricing and other instruments for climate change mitigation in private transport Henrik Andersson, Davide Cerruti and Cristian Huse

4.1 INTRODUCTION There are various objectives behind the pricing of transport services. For instance, from the perspective of a transport agency responsible for financing infrastructure the main objective may be cost recovery, e.g. how to charge air-traffic passengers for investing in and running airports. Decisions about pricing can also be made from a broader social perspective taking into account distributional aspects. To facilitate less wealthy groups gaining access to job opportunities or social activities, transport services like public transport may be subsidized. Another objective, which is the one that is often the focus of economists, is the use of pricing to achieve an efficient resource allocation. That is, how can the pricing mechanism be used to influence transport users to reach an optimal level of transport service demand? Due to the several externalities linked to personal decisions on how, where, and when to travel, like congestion, pollution, accident risk, and noise, the discussion on pricing and efficiency often is focused on how to internalize such externalities. The pricing objective of interest in this chapter is efficiency, and more specifically how to internalize emissions from transportation related to global warming. The transport sector is only one of several sectors that contribute to global warming, but its contribution is significant. For instance, the International Panel on Climate Change (IPCC, 2014) reported that the main sectors for global greenhouse gas (GHG) emissions were electricity and heat production (25%), agriculture, forestry, and other land use (24%), and industry (21%), followed by transportation (14%). Regarding transportation, much public attention today is on emissions from air-traffic, with both private initiatives and government policies on how to reduce the demand for air-traffic and make people either travel less or choose less emitting modes like trains. Given this attention, it is relevant to put its GHG emission into perspective and the aviation industry contributes to about 2% of global GHG emissions, which puts its footprint on par with the emissions from information and communications technology (data centres, mobile phone networks, etc.) (Jones, 2018). Figure 4.1 shows the emissions from different transport modes in Europe. As shown, the main source of emissions is from road transport. Hence, the potential gain addressing GHG emissions from road transport is much larger than from airtraffic and we will therefore focus on GHG emissions from road traffic in this chapter. Within any decarbonization strategy in the transport sector, passenger cars will play a pivotal role. According to the International Energy Agency (IEA), passenger vehicles account for almost half of worldwide transport sector carbon emissions, and it is the category that requires the largest decline in emissions in the IEA Sustainable Development Scenario (IEA, 2019). Different measures will be necessary to facilitate any transition, but economic instruments have an important role to play. One such instrument is pricing, which will be the main focus of this chapter. 59

60  Handbook on transport pricing and financing

Source:   Authors’ own, data from https://www​.eea​.europa​.eu​/data​-and​-maps​/indicators​/transport​-emissions​ofgreenhouse​-gases​-7​/assessment (Accessed 2021-03-04).

Figure 4.1  Greenhouse gas emission from transport in Europe This chapter is structured as follows. In the following section, we first describe the theory of pricing and the internalization of externalities, and then provide a discussion of a selection of implementation issues. Thereafter in Section 4.3 we provide examples and discuss how to apply different instruments to car usage and ownership. The chapter finishes with some concluding remarks and suggestions for further reading.

4.2 PRICING OF TRANSPORT EXTERNALITIES In the first part of this section, we briefly describe the motives behind and the theory of pricing transport externalities. We then in the second part discuss some issues of relevance when implementing pricing policies. Again, the discussion is brief and only intended to raise important issues. For more details, see Chapter 2 in this Handbook by Achim Czerny and Stefanie Peer, or cited work. 4.2.1 Pricing Motives and Principles Pricing transport usage is at the core of transport economics. To enhance economic efficiency is the primary economic motivation for pricing transport usage, but it is also of interest from the perspective of financing infrastructure and transport services. Therefore, to a large extent theoretical and empirical research have been carried out within the context of optimal infrastructure usage (see, e.g. Rouwendal and Verhoef, 2006, for a discussion). One issue of major relevance that has been given considerable attention among policymakers and researchers is road congestion (see, e.g. Lindsey, 2012). In a congested road network, an additional trip

Pricing for climate change mitigation in private transport  61

taken by a road user will inflict a cost on other users since it will contribute to the congestion and slow down the traffic further. By pricing this marginal externality a more efficient use of the current road network can be achieved, and the revenues collected can be used to finance maintenance or investments in the road network. The focus of this chapter is GHG emissions from road traffic, another transport externality. In an unregulated market, the driver or passenger of a vehicle will only consider his/her private costs when making travel decisions. Other costs, like GHG emissions, local air pollutants, noise, etc., are not borne by the driver or passengers and therefore likely to be ignored. A significant difference between the congestion externality and the GHG emission externality is that the former depends not only on the traffic volume but also on the infrastructure available. That is, in contrast to the congestion that can be reduced, or eliminated, by investing in new infrastructure, GHG emissions will still be present after the expansion of the infrastructure with the same amount of traffic volume. Even if total emissions from cars will depend on the traffic situation, i.e. congestion, they are more correlated with the traffic volume and the type of vehicles used. The common feature shared by congestion and GHG emissions is that both can be addressed using economic instruments such as prices. Examples of economic instruments to address GHG emissions from car usage is indeed the focus of Section 4.3 of this chapter, and hence the focus of the discussion below. The principles of pricing, or taxing, externalities were introduced by Pigou (1920). Figure 4.2 illustrates the basic principles. The downward-sloping inverse demand curve (D (V )) represents the marginal benefit (MB) of undertaking trips (V). The private cost of undertaking the trips is represented by MC, and is assumed to be constant. In an unregulated market, only private benefits and costs will be considered by the individual and the equilibrium is given by U, i.e. where D = MC. Fewer trips would mean that the benefits of undertaking them are higher than the costs, and more trips would mean that the costs are higher than the benefits. Hence, it is optimal for the individual to undertake VU trips. In a situation without any externalities, the VU number of trips would also be the social optimal number of trips, since in that case the marginal social benefit (MSB) equals the marginal social cost (MSC) in U. In this chapter, we focus on GHG emissions from car usage and we now introduce this externality from car trips, E, in Figure 4.2, which for simplicity is also assumed to be constant. The total MSC from undertaking trips is now given by the sum of the private cost and the externality, i.e. MSC = E + MC . Following the reasoning above, the optimal level of trips is when the social benefit of an additional trip equals its social cost. As illustrated in Figure 4.2 the internalization of the externality has increased the price of undertaking trips from PU to PO which results in the new allocation O, where D = MSC, that is at trips VO. As shown, with the externality each trip above VO has a social cost that is higher than the social benefit, and hence, provides a negative net social benefit, represented by the yellow area. The optimal tax (T) on the externality in our scenario (in which the externality is constant) is equal to E, but more generally the tax should be set according to T = MSC (VO ) - MC (VO ) . This level of the tax will internalize the externality and lead to a demand for the number of trips where the MSB of that trip equals the private, and social, marginal cost. 4.2.2 From Theory to Practice The above description of a situation where prices should equal their social marginal costs is often in the economic literature referred to as a “first-best” solution. However, it is well

62  Handbook on transport pricing and financing

Source:   Author’s (Andersson) own lecture notes.

Figure 4.2  Equilibrium in an unregulated market and social optimum established that first-best conditions do not generally hold due to other unpriced externalities, distortionary taxes, etc. (Parry and Bento, 2002). Solutions according to first-best are quite straightforward, requiring only information about the social marginal cost. Second-best solutions are more complicated, since they require input not only on the market failure studied, but also information on related markets and how they interact. This means that the optimal level of the tax may be different from the value of the externality in optimum and instead could be lower or higher. Pricing externalities increases the private costs of users (from PU to PO in Figure 4.2), and hence public support for them is often weak. Due to the revenues collected the public perception may be that the tax is implemented for fiscal reasons, not to address a market failure. From a Pigouvian tax perspective how the revenues are used is not of importance, but it is both from an acceptance perspective and from a fiscal perspective. Regarding the former, acceptance among transport users may increase if revenues collected are earmarked for investments in the transport sector (Leape, 2006). Regarding the latter, “Pigouvian taxes” are favoured by economists since they not only internalize the externality (i.e. the Pigouvian objective) but can also be used to offset distortionary taxes in the economy, the combined effect often referred to as a “double dividend” (Goulder, 1995). That is, the revenues from, e.g. a tax on GHG emissions can be used to reduce the revenues from a distortionary tax such as a tax on labour. Parry (1995) showed, though, using a labour tax as an example of a distortionary tax, and assuming that the market good to be taxed (due to the externality) and leisure being substitutes, that the optimal level of the green tax should actually be lower than the external cost. Hence, even if green taxes can be used to offset distortionary taxes, in the presence of them the optimal green tax may differ from external cost (marginal damage). Related to the acceptance of Pigouvian taxes is the public’s perception of their fairness. On the one hand, based on the “polluter pays principle” (OECD, 1995) it can be viewed as fair that those who inflict harm on others are the ones who should pay for the pollution. On the other hand, environmental taxes may more heavily affect the welfare of poor individuals than richer

Pricing for climate change mitigation in private transport  63

ones, i.e. being regressive (Chiroleu-Assouline and Fodha, 2014). Whether distributional aspects should be considered jointly with or separately from efficiency is an ongoing discussion, but it is important to remember that the primary objective of Pigouvian taxes is efficiency and, ideally, distributional effects can be addressed separately if needed. Chapter 6 of this Handbook covers the underlying trade-offs in more depth. Implementing Pigouvian taxes also requires a monetary estimate of the externality. The standard approach to estimating the social value of goods and services without any easily available prices is to either rely on revealed- or stated-preference methods, or both (see, e.g. Freeman et  al., 2014). An alternative approach is based on the hypothesis that the public’s preferences in a representative democracy are reflected in public decision-making. One issue with this approach is that there is overwhelming evidence that there is a large variation in the estimates of the abatement costs between sectors or industries (Gillingham and Stock, 2018). Other options to estimate the shadow values of these goods and services are to estimate the social cost of environmental degradation, or what costs are necessary to achieve policy targets (see, e.g. Quinet, 2019, for an example and a discussion).

4.3 POLICY INSTRUMENTS FOR PERSONAL VEHICLES 4.3.1 Economic Instruments and Car Usage and Ownership As described, economic agents typically do not take into account the social cost of their choices or actions, that is, the cost they impose on society. When it comes to transport-related choices this often translates, e.g. into using private rather than public means of transportation, larger rather than smaller vehicles, and more rather than less powerful engines. In the previous section, we introduced the concept of the Pigouvian tax, which can be seen as an attempt to discipline economic agents. This arguably simple framework is the starting point of most economic analyses of environmental problems to date, including problems related to car usage and ownership. Traditional policy instruments have been covered extensively in the literature. For instance, Calthrop and Proost (2003) cover environmental pricing in transport whereas Parry et  al. (2007) and Anderson and Sallee (2016) provide a comprehensive overview of automobile externalities. Determining the level of the tax is, however, a non-trivial task for a number of reasons. First, one has to identify the factors generating external social costs. This could include carbon emissions from transportation, but also local pollution generated by combustion engines and tyres, costs of congestion, accidents, noise and visual pollution caused by driving, wearand-tear of roads, etc. In recent years, it has been identified that, in some jurisdictions, some external costs have been given more importance than others, which may benefit one technology to detriment of another. For instance, the European Union (EU) legislation has historically been relatively strict in terms of CO2 emissions but more lenient with regard to local pollutants, which was documented to benefit diesel to detriment of gasoline vehicles (Miravete et al., 2018). Second, one has to correctly quantify each of the above externalities. Recent evidence of the manipulation of driving tests has given credence to the notion that measurement is potentially of first-order importance, especially in diesel vehicles, which led to the recent re-design of test cycles in Europe. When it comes to electric vehicles, externalities are dependent on local

64  Handbook on transport pricing and financing

factors, such as how electricity used to charge the electric vehicles is generated (Holland et al., 2016). Third, one has to assign monetary values to each of the above factors. For instance, local pollution generated today might have long-lasting health effects that current research has not yet been able to identify. Fourth, the passing-on of taxes onto prices depends on both the competitive structure and on the characteristics of the supply and demand profiles of a particular market. In particular, firms may or may not pass on the cost of a tax to consumers. Fifth, it assumes that consumers correctly quantify and take into account in their decisions the tax levied on their activities or the products they purchase. Despite the above limitations, the Pigouvian framework provides the basic toolkit used by economists and regulators in the analysis of environmental problems, also in the transport sector. As described, if carbon emissions are the only external cost of transport imposed on society, then the Pigouvian tax is efficient (or first-best), in that it imposes on consumers the entire external costs their activity generates to society. In particular, a fuel tax is a Pigouvian tax in this specific case, as introducing it tackles both utilization (kilometres driven) and fuel consumption (how many litres a vehicle requires to drive 1 kilometre) of a driver-vehicle combination. An additional factor making taxes appealing is that they are cost-effective, in the sense that they satisfy the least-cost theorem, an important result in environmental economics (Perman et al., 2003). That is, taxes are able to attain a given environmental target at least cost to the society, which results in an economically efficient allocation of resources. In more concrete terms, if a policy instrument is least-cost, then it satisfies the so-called equi-marginal principle, according to which the cost of abating pollution is equalized among all polluters – in particular, economic agents for which abating pollution is cheaper will abate more than those for which abating pollution is more expensive until the cost of abatement of all agents equalizes. If consumers are not able to identify and/or quantify the extent of the taxation they are subject to, then they will undervalue the fuel tax and over-utilize transportation. As a result, this is one channel through which the fuel tax loses its first-best property. Therefore, additional policy instruments are called for, such as standards or vehicle taxes. Although they do not tackle utilization, standards have an impact on the fuel economy (or fuel consumption) of a vehicle, making less efficient vehicles more expensive than their more efficient counterparts. Circulation taxes operate in a similar way, but while standards typically influence only the purchase of new vehicles, circulation taxes are charged yearly and thus affect both the new and second-hand car market. The distinction between new and used vehicle markets is crucial since while the former receives considerably more attention, it is only a small (typically single-digit) share of the latter. Thus, one has always to bear in mind the scope of the policy instruments aiming to tackle carbon emissions at the risk of said instruments not having a meaningful impact. Recent events in the transport sector have brought back attention to local pollutants and shed light on a number of alternative policy instruments tackling pollution in transportation. If local pollutants are not addressed when determining the level of the fuel tax, then alternative policy instruments – especially those targeting diesel vehicles – are called for. These include congestion pricing and driving restrictions, which can occur in various forms. Finally, considerations about optimal policies have to take into account the political constraints that often prevent a straightforward implementation of those measures. While interventions adopted by governments might not represent the most efficient practice from an economic point of view, they might be the only politically feasible solutions; Bruno De Borger

Pricing for climate change mitigation in private transport  65

and Antonio Russo deliver more insights on the political economy of transport pricing in Chapter 7 of this Handbook. In what follows, we briefly describe the main policy instruments available to mitigate carbon emissions and their applications. 4.3.2 Instruments Based on Marginal Cost Pricing 4.3.2.1 Fuel taxes The most common pricing measure addressing CO2 emissions from the transport sector is the fuel tax. Because it is charged on a per-litre or per-gallon basis, it represents a Pigouvian tax on carbon emission conditional on fuel consumption. In particular, fuel taxes help reduce vehicle CO2 emissions in two distinct ways: first, they encourage consumers to choose energy-efficient vehicles, or an electric vehicle; second, they induce drivers to reduce the number of kilometres driven. The presence of two channels of emission reduction generally makes fuel taxes a preferable solution compared with other kinds of measures, such as standards, circulation taxes, or rebates for the purchase of energy-efficient vehicles. Furthermore, the fuel tax helps addressing other types of externalities linked to vehicle use, such as local air pollution. Fuel taxes have been adopted in many countries, including various European states, China, India, the United States, and several emerging economies. According to OECD statistics, fuel taxes vary substantially across countries (OECD, 2018). One common characteristic, however, is that the corresponding price per tonne of CO2 is typically larger for gasoline than for diesel. The effectiveness of the fuel tax in reducing carbon emissions depends on various factors. For instance, drivers might not correctly understand the impact of fuel prices on the lifetime cost of owning a vehicle, i.e. a 1€ change in (discounted) driving costs are not considered equivalent to a 1€ change in the vehicle purchase price. Correct valuation of fuel costs would require consumers to know and correctly combine information on fuel prices, utilization, and vehicle fuel economy. Therefore, the undervaluation of fuel costs might require complementary policies such as rebates for the purchase of energy-efficient vehicles or information campaigns in order to make fuel costs more salient (Huse and Koptyug, 2022). Current evidence on how consumers evaluate fuel economy and fuel costs suggests that undervaluation occurs but is typically moderate (see Huse and Koptyug, 2022, and references therein). On the other hand, fuel taxes tend to be more salient than fluctuations in fuel prices due to supply shocks for a couple of reasons. First, changes in fuel taxes are generally advertised on various media. Second, fuel taxes are more stable, and therefore a forward-looking consumer might be more induced to buy an energy-efficient car due to an increase in fuel tax today than to other fluctuations in fuel prices. If that is the case, people might respond to fuel taxes more strongly than equivalent changes in pre-tax fuel prices. Furthermore, in the way they are currently defined, fuel taxes only affect gasoline- or diesel-fuel vehicles, but not other types of alternative fuel vehicles. Therefore, the expected increase in the market share of plug-in and battery electric vehicles will likely lessen the effectiveness of fuel taxes in the future, especially if a carbon tax is not in place and if fossil fuels still play a relevant role in electricity generation. Further factors working against fuel taxes are their low price elasticity of demand (Brons et al., 2008), obtain short- and long-run elasticities of, respectively, –0.34 and –0.84 in their meta-analysis) and that the demand for alternative fuels seems to be significantly more elastic than those for fossil fuels when both compete directly (Huse, 2018). Taken together, these factors point to a downward trend of fuel

66  Handbook on transport pricing and financing

tax revenues, which is leading some governments to consider utilization charges such as mileage taxes. These, however, do not directly target carbon emissions. Nevertheless, currently fuel taxes generate considerable revenues. Hence, they are an example of the “double dividend” discussed above, where fuel taxes help reduce carbon emissions and other externalities (first dividend), and the revenue would allow to fund public expenditure or reduce other distortionary taxes such as income taxes (second dividend). As described, the presence of this double dividend depends on the characteristics of the general tax system: the interaction between an environmental tax and other pre-existing distortionary taxes can also exacerbate the efficiency loss caused by the latter. These factors have important implications on the actual optimal level of a fuel tax, and whether it is lower or higher than the marginal damage linked to CO2 emissions (Goulder and Parry, 2008). When evaluating the optimal fiscal policy to address transport carbon emissions, it is important to take into account the practical difficulties that might hinder the implementation of the “best” solution. In particular fuel taxes and carbon taxes in general might encounter significant political opposition, for a series of reasons. First, as discussed earlier, fuel taxes and in general fuel price changes are quite salient to the general public compared to other vehicle-related monetary policies such as circulation taxes or sales taxes. Second, there might be considerable variation in the tax burden, and some specific categories might be affected more strongly than others. For instance, people living in rural areas with little public transport availability, or poorer households that cannot afford to buy a more efficient vehicle. As the 2019 protests of the “Gilet jaunes” (“yellow vests”) in France show, the pushback against this type of policy can be stark. The inequalities caused by the implementation of a fuel tax can be mitigated through a tailored redistribution of the tax revenue as a lump-sum transfer towards the most affected categories. 4.3.2.2 Mileage (or utilization) tax A utilization tax is based on how much a vehicle is driven. It has drawn increasing interest over the years due to technological improvements – the GPS technology necessary to implement such a tax allows collecting it at a high frequency – and due to the foreseen decreases in tax revenues due to improvements in the energy efficiency of vehicles. While a standard mileage tax not depending on energy efficiency would only affect kilometres driven, a more sophisticated one could address both margins, namely energy efficiency and utilization. Nevertheless, this policy instrument does not address congestion and its distributional effects may be adverse, since a mileage tax would more heavily affect rural households, who typically have restricted access to public transport. All in all, utilization taxes should be seen as a complement to fuel taxes, standards, and vehicle taxes from a climate policy viewpoint. 4.3.3 Instruments Based on Average Cost Pricing 4.3.3.1 Vehicle taxes Along with fuel taxes, vehicle taxes are a very widespread fiscal policy related to the transport sector. They can be either a tax applied only when the car gets registered for the first time (registration tax), or a tax that drivers must pay on a yearly basis in order to use the vehicle (circulation tax, road tax, or vignette). Initially, the amount of the tax was either the same for any vehicle, or linked to vehicle weight, vehicle segment, or sales price. In the most recent decades though, governments started to set the vehicle tax according also to energy efficiency,

Pricing for climate change mitigation in private transport  67

carbon emission rates and other environmental indicators, thus explicitly providing an incentive to choose energy-efficient/low-emission vehicles. Electric vehicles and other alternative fuel vehicles might benefit as well from lower tax rates. The main difference between a fuel tax and a vehicle tax is that the latter affects only the choice of the vehicle, but not the utilization, making them in general less cost-effective and less efficient than fuel taxes in reducing carbon emissions in the transport sector. However, vehicle taxes can be a complement to fuel taxes in case consumers tend to underestimate vehicle fuel costs. Important distinctions also exist between different types of vehicle taxes: registration taxes have leverage only in the choice of new vehicles, while circulation taxes influence also the decision on the purchase of second-hand vehicles and on the timing of scrappage. Thus, circulation taxes might have a stronger influence on the composition of the vehicle fleet in the short- and medium-term. In the case of a registration tax based on vehicle CO2 emissions, the tax is akin to a standard, the key distinction being that it is an explicit rather than an implicit tax. One advantage of vehicle taxes is that they are easier to implement from a political standpoint, in part because they are less salient to drivers. In particular, they could be used to capture differences in externalities between vehicles with different fuel types when the use of fuel taxes is constrained by political considerations (De Borger and Mayeres, 2007). However, a lack of visibility represents a problem for the effectiveness of these measures, as consumers might not be aware of the existence of fiscal incentives for the purchase of low-emission vehicles. Vehicle taxes directly or indirectly linked to CO2 emissions have played an important role when it comes to technology adoption (gasoline vs. diesel vehicles) in the European car market, partially explaining the market share commanded by diesel vehicles in Europe (EEA, 2015). As for fuel taxes, vehicle taxes are present in several countries, from most European countries to Japan, China, India, and South Africa. 4.3.3.2 Rebates and feebates Governments have often implemented a policy instrument whereby low-emission vehicles are subsidized in the new vehicle market through a rebate (Huse and Lucinda, 2014), sometimes coupled with a fee on high-emission vehicles (Adamou et al., 2014). The combination of both policy instruments gives the name of such programmes, feebate. Feebates are policy measures used to change the relative prices of high- and low-energy efficiency vehicles. Fees are charged to vehicles emitting more than a given threshold emissions level (the pivot point) and generate revenues used to finance the rebates given to vehicles emitting less than said level. Thus, the rationale of a feebate programme is often that it should be fiscally neutral, i.e. self-financing. However, this requires that the problem be well-calibrated ex ante, which is not always the case. While feebate programmes have been adopted in some European countries, noticeably France and Sweden, rebates on low-emission vehicles are more widespread worldwide. Feebates have been the object of many studies due to the values involved, be it on a pervehicle basis or in the aggregate (Adamou et  al., 2014). The main pitfalls in their implementation are free-riding and low pass-through rates. Free-riding occurs when part of the consumers benefiting from the rebate would have chosen a low-emission vehicle regardless of incentives, or when consumers take advantage of political or administrative borders. In a case of a low pass-through rate instead, part of a rebate on the final vehicle price might not

68  Handbook on transport pricing and financing

benefit consumers, as car dealers and manufacturers would increase purchase prices accordingly. Either way, such type of behaviour decreases the cost-effectiveness of environmental programmes. 4.3.4 Other Instruments 4.3.4.1 Congestion pricing and driving restrictions Congestion pricing and driving restrictions (sometimes combined with parking restrictions and pricing) allow local governments to influence the policy arena in the transport sector. Driving restrictions in the form of licence plate bans have been studied by Eskeland and Feyzioglu (1997), Davis (2008), and, more recently, Barahona et al. (2020), whereas Wolff (2014) studied the effect of low-emission zones focusing on the German case. Congestion pricing is more recent, see Lehe (2019) and Chapter 8 of this Handbook for comprehensive overviews. Oftentimes, driving restrictions are imposed due to non-compliance with (local pollution) air quality standards, as in the cases of many EU cities, which are then required to develop a plan for tackling air quality following the Clean Air for Europe Directive (CAFE, 2008/50/ EC). The implementation of such a plan has as a by-product the reduction of congestion and carbon emissions. There are substantial differences between congestion pricing and a low-emission zone (LEZ): congestion pricing charges vehicle owners a fee – which might depend on the time of the day and the type of vehicle – to get access to a specific area, typically a city centre. Instead, a LEZ allows only specific categories of vehicles – typically low-emission or electric vehicles – to enter specific areas. Several cities throughout the world have introduced either vehicle driving restrictions (e.g. Munich, Mexico City, Quito, Shanghai), or congestion pricing schemes (e.g. London, Milan, and Singapore); see Lehe (2019). A variant of such policies, mostly implemented in North America, are high-occupancy vehicle (HOV) lanes (see Caltrans, 2018, and Chapter 23 in this Handbook). These measures restrict access to vehicles based on occupancy, so that only vehicles with at least a minimum number of passengers are allowed access to specific lanes. Economic efficiency arguments favour congestion pricing over driving restrictions, since the price mechanism inherent in the former allows flexibility and theoretically achieves a least-cost outcome. In contrast, the latter can be thought of as being akin to licences, resulting in a number of unintended consequences. The pricing component affects both distance driven and the vehicle stock, and in some cases might induce people to replace their vehicle earlier with a newer, more efficient vehicle with fewer driving restrictions. On the other hand, licence-related policies may well backfire. For instance, restrictions based on the last digit of the licence plate number (odd or even) have prompted households to keep older vehicles longer and in some cases decided to own two vehicles – one ending with an odd number and another ending with an even number – to circumvent the regulation. Given a fixed amount of the household budget, the additional vehicle purchased was often found to be older, with worse energy efficiency (Davis, 2008). When implementing these policies, trade-offs between environmental goals must also be taken into account. For instance, diesel vehicles emit less CO2 than gasoline vehicles, but more local pollutants such as PM10 and NOx (Miravete et al., 2018). For this reason, even if the literature showed that driving restrictions based on emission rates of local pollutants are

Pricing for climate change mitigation in private transport  69

generally more effective, policymakers should also consider the presence of such trade-offs when choosing between different environmental measures. Distributional effects of such policies depend on how the population is distributed in and around a city. Typically, driving restrictions and congestion charges are adopted in city centres and residents are granted free access, or discounts on the charge. Therefore, these policies tend to more heavily affect individuals living in suburbs and commuting to the city. If these tend to be low-income, they will suffer to a greater extent the effects of such policies. In a similar vein, if those policies are based on vehicle emission rates, low-income households who keep their older car for longer, or who cannot afford to replace their vehicle, would be disproportionally affected by the measures. In the particular case of LEZs, low-income individuals owning older vehicles will not be allowed into the LEZ at all. Banning the circulation of certain types of vehicles generates quite high social costs, and for this reason, it is fundamental to evaluate ex ante if the potential gains from lower pollution – in particular improvement in health outcomes – compensate for such costs (Börjesson et al., 2021). 4.3.4.2 Information programmes The effectiveness of the monetary incentives mentioned so far can vary depending on the level of information available to consumers. In fact, there are two main informational hurdles that can prevent consumers from fully internalising the financial incentives towards energyefficient and low-carbon choices. The first hurdle is that consumers are either unaware of the monetary incentives, or not completely informed about them. For instance, they might ignore that the circulation tax is lower for energy-efficient vehicles. If that is the case, the incentive is simply not taken into account by consumers in their choice of vehicle. The second hurdle is that, even when consumers are perfectly informed about the presence of monetary measures favouring energy-efficient choices, they are not able to include these incentives in their decision-making process. For instance, the calculation of the overall fuel costs of a vehicle requires combining information on the fuel price per litre (including the fuel tax), the fuel economy of the vehicle, yearly mileage, and the expected lifetime of the vehicle. If consumers lack the capability to perform these calculations, they might not fully internalize the various monetary incentives when making decisions on which vehicle to buy, or how much to drive. Information measures can help overcome these hurdles in three ways: first, by making individuals aware of the presence of monetary incentives through informational campaigns (e.g. advertising or letters). Second, by directly providing a calculation of the monetary costs of each option – for instance, through an energy label that for each vehicle on sale illustrates the fuel costs per year for a typical mileage, or a website that compares different vehicles and directly calculates the circulation tax amount for each option. Third, by teaching consumers how to correctly calculate the overall costs of each option, and how to incorporate the various monetary incentives into such calculation. 4.3.4.3 Vehicle scrappage schemes Vehicle scrappage schemes usually offer a rebate on the purchase of a new, energy-efficient vehicle if a buyer trades in her old vehicle, subject to satisfying some requirements, which can range from age to energy efficiency (Adda and Cooper, 2000; Mian and Sufi, 2012; Hoekstra et al., 2017). One reason why governments impose conditions on which vehicles can be traded in is the so-called free-riding problem; the short duration of the programme is such that consumers are likely to respond to incentives by anticipating the purchase of a low-emission

70  Handbook on transport pricing and financing

vehicle and pocket the rebate. Thus, the environmental gains from the programme will be likely modest compared to its costs (Mian and Sufi, 2012). Governments have often incentivized the early retirement of old, inefficient vehicles through vehicle scrappage schemes. For instance, several such programmes were introduced following the 2008 recession (e.g. in China, the EU, Japan, and the United States), both as an environmental policy instrument and as a temporary stimulus for the economy (Mian and Sufi, 2012). The effectiveness of these incentives as a measure to reduce vehicle carbon emissions is unclear, since the new vehicle fleet comprises a small share of the total vehicle fleet.

4.4 CONCLUDING REMARKS This chapter has surveyed the policy instruments used in the passenger vehicle sector to tackle the externalities it creates. Given difficulties often associated with the implementation of “first-best” policies, policymakers worldwide have implemented a diverse number of policies, often in combination. Such a menu of policies is also motivated by the different externalities policymakers are aiming to address, from global pollutants such as CO2, to local pollutants, and to congestion. Due to the different local contexts and approaches adopted by governments, in this chapter we offer some general findings on the effectiveness of the various policies from the literature, without focusing on evaluating specific situations. While this chapter is focused on the reduction of carbon emissions, passenger car transport generates several other types of externalities, including some that are particularly relevant. In the EU congestion and accidents represent the most important types of external costs from passenger vehicle traffic, followed by climate and air pollution externalities (Van Essen et al., 2019). How this ranking of different types of externalities would change between different countries and with the evolution of the vehicle fleet is still an ongoing question. An important factor influencing the environmental impact of passenger transport is the type of urban form. In general, the presence of urban sprawl can contribute to worsening environmental externalities, although whether a higher urban density is always desirable is still an open question (Gaigné et al., 2012). For a broader discussion of the relationship between urban form and transportation we refer the reader to Chapter 5 of this Handbook. To achieve the goals of the Paris Agreement, carbon taxes – which would also affect the transport sector – should reach a level between 50 and 100 US$ per tonne of CO2 and should be combined with other policies aimed to reduce carbon emissions. Because these pricing levels are higher than those currently existing, the political acceptability of these measures is a significant challenge. Successful carbon price schemes currently implemented have introduced some transfer schemes to compensate those actors who were affected the most or to make carbon pricing more palatable to voters (Klenert et al., 2018). Technological developments in the transport sector in the coming years are likely to challenge the established view and use of many such policy instruments. Such developments range from the prevailing technology in transport to the business model under which the transport sector will operate, e.g. whether vehicles will be owned or just rented on an hourly basis or accessed via a subscription service. The impact of the diffusion of electric vehicles on carbon emissions, currently exempt from many of the measures described in this chapter, will be tied to the degree of decarbonization of the electricity mix. In sum, the developments are likely to impact the very externalities created by the sector, from pollution to utilization to vehicle size.

Pricing for climate change mitigation in private transport  71

REFERENCES Adamou, A., S. Clerides, and T. Zachariadis (2014). Welfare implications of car feebates: A simulation analysis. The Economic Journal 124(578), F420–F443. Adda, J. and R. Cooper (2000). Balladurette and juppette: A discrete analysis of scrapping subsidies. Journal of Political Economy 108(4), 778–806. Anderson, S. T. and J. M. Sallee (2016). Designing policies to make cars greener. Annual Review of Resource Economics 8, 157–180. Barahona, N., F. A. Gallego, and J.-P. Montero (2020). Vintage-specific driving restrictions. The Review of Economic Studies 87(4), 1646–1682. Börjesson, M., A. Bastian, and J. Eliasson (2021). The economics of low emission zones. Transportation Research Part A: Policy and Practice 153, 99–114. Brons, M., P. Nijkamp, E. Pels, and P. Rietveld (2008, September). A meta-analysis of the price elasticity of gasoline demand. A SUR approach. Energy Economics 30(5), 2105–2122. Calthrop, E. and S. Proost (2003). Environmental Pricing in Transport. In D. A. Hensher and K. J. Button (Eds.), Handbook of Transport and the Environment, Volume 4, pp. 529–545. Emerald Group Publishing Limited. Caltrans (2018). High-occupancy vehicle (HOV) systems. (Link) (Accessed: March 21, 2021). Caltrans, California. Chiroleu-Assouline, M. and M. Fodha (2014). From regressive pollution taxes to progressive environmental tax reforms. European Economic Review 69, 126–142. Davis, L. W. (2008). The effect of driving restrictions on air quality in Mexico City. Journal of Political Economy 116(1), 38–81. De Borger, B. and I. Mayeres (2007). Optimal taxation of car ownership, car use and public transport: Insights derived from a discrete choice numerical optimization model. European Economic Review 51(5), 1177–1204. EEA (2015). Dieselisation in the EEA. (Link) (Accessed: March 21, 2021). European Environmental Agency. Eskeland, G. S. and T. Feyzioglu (1997). Rationing can backfire: The “day without a car” in Mexico City. The World Bank Economic Review 11(3), 383–408. Freeman, M. A., J. A. Herriges, and C. L. Kling (2014). The Measurement of Environmental and Resource Values (3rd ed.). New York: RFF Press, Routledge. Gaigné, C., S. Riou, and J.-F. Thisse (2012). Are compact cities environmentally friendly? Journal of Urban Economics 72(2–3), 123–136. Gillingham, K. and J. H. Stock (2018). The cost of reducing greenhouse gas emissions. Journal of Economic Perspectives 32(4), 53–72. Goulder, L. H. (1995). Environmental taxation and the double dividend: A reader’s guide. International Tax and Public Finance 2, 157–183. Goulder, L. H. and I. W. Parry (2008). Instrument choice in environmental policy. Review of Environmental Economics and Policy 2(2), 152–174. Hoekstra, M., S. L. Puller, and J. West (2017). Cash for Corollas: When stimulus reduces spending. American Economic Journal: Applied Economics 9(3), 1–35. Holland, S. P., E. T. Mansur, N. Z. Muller, and A. J. Yates (2016). Are there environmental benefits from driving electric vehicles? The importance of local factors. American Economic Review 106(12), 3700–3729. Huse, C. (2018). Fuel choice and fuel demand elasticities in markets with flex-fuel vehicles. Nature Energy 3, 582–588. Huse, C. and C. Lucinda (2014). The market impact and the cost of environmental policy: Evidence from the Swedish green car rebate. Economic Journal 124, F393–F419. Huse, C. and N. Koptyug (2022). Salience and policy instruments: Evidence from the auto market. Journal of the Association of Environmental and Resource Economists 9(2), 345–382. IEA (2019). Transport sector CO2 emissions by mode in the sustainable development scenario, 20002030. (Link) (Accessed: March 21, 2021; last updated: November 22, 2019), International Energy Agency (IEA), Paris.

72  Handbook on transport pricing and financing

IPCC (2014). Climate change 2014: Mitigation of climate change. Technical report, Intergovernmental Panel on Climate Change (IPCC): Contribution of Working Group III to the Fifth Assessment Report of the IPCC. Cambridge University Press, Cambridge, UK and New York. Jones, N. (2018). How to stop data centres from gobbling up the world’s electricity: The energy-efficiency drive at the information factories that serve us Facebook, Google and bitcoin. Nature 561, 163–166. Klenert, D., Mattauch, L., Combet, E., Edenhofer, O., Hepburn, C., Rafaty, R., & Stern, N. (2018). Making carbon pricing work for citizens. Nature Climate Change 8(8), 669–677. Leape, J. (2006). The London congestion charge. Journal of Economic Perspectives 20(4), 157–176. Lehe, L. (2019). Downtown congestion pricing in practice. Transportation Research Part C: Emerging Technologies 100, 200–223. Lindsey, R. (2012). Road pricing and investment. Economics of Transportation 1(1), 49–63. Mian, A. and A. Sufi (2012). The effects of fiscal stimulus: Evidence from the 2009 cash for clunkers program. The Quarterly Journal of Economics 127(3), 1107–1142. Miravete, E. J., M. J. Moral, and J. Thurk (2018). Fuel taxation, emissions policy, and competitive advantage in the diffusion of European diesel automobiles. The RAND Journal of Economics 49(3), 504–540. OECD (1995). Environmental principles and concepts. General Distribution OCDE/GD(95)124, The Organisation for Economic Co-operation and Development (OECD), Paris, France. OECD (2018). Consumption Tax Trends 2018: VAT/GST and Excise Rates, Trends and Policy Issues | READ online. Parry, I. W. (1995). Pollution taxes and revenue recycling. Journal of Environmental Economics and Management 29(3), S64–S77. Parry, I. W. and A. Bento (2002). Estimating the welfare effect of congestion taxes: The critical importance of other distortions within the transport system. Journal of Urban Economics 51(2), 339–365. Parry, I. W., M. Walls, and W. Harrington (2007). Automobile externalities and policies. Journal of Economic Literature 45(2), 373–399. Perman, R., Y. Ma, J. McGilvray, and M. Common (2003). Natural Resource and Environmental Economics. London, UK: Pearson. Pigou, A. C. (1920). The Economics of Welfare. London, UK: Macmillan. Quinet, A. (2019). What value can we attach to climate action? Economics and Statistics / Economie et Statistique 510-511-512, 165–179. Rouwendal, J. and E. T. Verhoef (2006). Basic economic principles of road pricing: From theory to applications. Transport Policy 13(2), 106–114. Van Essen, H., L. van Wijngaarden, A. Schroten, D. Sutter, C. Bieler, S. Maffii, M. Brambilla, D. Fiorello, F. Fermi, R. Parolin, and K. El Beyrouty (2019). Handbook on the External Costs of Transport, version 2019. Number 18.4 K83. 131. Wolff, H. (2014). Keep your clunker in the suburb: Low-emission zones and adoption of green vehicles. The Economic Journal 124(578), F481–F512.

5. Urban form and the pricing of transport and parking Sofia F. Franco1

5.1 INTRODUCTION This chapter provides insights on how transport pricing in general, and parking prices and their availability in particular, impact urban structure and urban form. Since the topic of transport pricing is vast, the focus of this chapter is on first- and second-best pricing in the context of congested road traffic, and on the theoretical effects of congestion and parking pricing on the location of jobs and residences within cities. As such, the chapter complements Chapter 2 where the theory of welfare-related transport pricing and its externalities are covered, and Chapter 8 which reviews recent developments on road congestion pricing and capacity provision. This chapter further discusses parking pricing, whereby the costs of onstreet and/or off-street parking are increased. Parking pricing is considered by some to be the second-best alternative to road pricing and congestion tolls (Verhoef et al., 1995). In addition, it discusses how two widely implemented parking practices – minimum-parking requirements (blunt land-use mandate) and employer-paid parking (transport subsidy at work) – affect land consumption and use, car use, and suburbanization levels. While a review of empirical contributions and real-world applications of congestion pricing are beyond the scope of this chapter, when appropriate and for illustrative purposes we rely for most of our evidence on the cases of Los Angeles County, USA, and its cities. 5.1.1 Urban Form, Urban Spatial Structure and Travel Mode Choice Urban form, as defined by job and population densities (dispersed or compact urban settings), centricity (mono versus polycentric urban formats) and city size (spatial physical expansion and total population), reflects the interplay between the demand for space and the demand for proximity, which is shaped by transportation technology (e.g., bike, car, public transit). Urban spatial structure on the other hand reflects land-use arrangements and is shaped by transportation innovations and government intervention (e.g., zoning and land-use controls, parking mandates and fees, fuel taxes, and road pricing). Transportation technology choice is influenced by the amount of transportation infrastructure (roads, highways, subway lines, or parking spaces) and household income.2 The generalized cost of travel which includes travel time and monetary costs is also key to travel mode choice, although other components (traveler comfort and travel reliability) have a strong influence as well. Therefore, travel mode choices, land-use patterns, and urban form are intrinsically related. In addition, less sustainable travel mode choices (e.g., private motorized travel modes such as cars) generate unintended external effects, known as externalities, ranging from traffic congestion, air pollution, climate change, noise, and accidents to name a few. These externalities 73

74  Handbook on transport pricing and financing 

give rise to various external costs like time costs of delays due to traffic congestion or cruising for parking, health costs caused by air pollution, and productivity losses due to fatalities and injuries in traffic accidents, among others (van Essen et al., 2008; Maibach et al., 2008; Shoup, 2005; Calthrop and Proost, 1998; Verhoef, 1994; EC, 2019). These costs affect society at large because they fall on individuals who are not part of the decision resulting in those costs and are not directly borne by the individual who has caused them. These costs then create a wedge between the private costs faced by the decision-maker (e.g., the car user), and the social costs incurred by society. Without policy intervention, transport users, in the presence of these externalities, face incentives that are not socially optimal, and thus do not internalize the external costs. As a result, their travel behavior is inefficient leading in turn to urban forms and urban spatial structures that are also inefficient. 5.1.2 Congestion Pricing in Monocentric Settings Since the seminal works of Alonso (1964), Mills (1967), and Muth (1969), a number of theoretical studies have explored the effects of congestion taxes on city size, population densities, and land-use arrangements within the monocentric city framework. Six key features characterize the traditional framework in which congestion taxes have been studied: (i) firms’ location decisions are exogenously given, with jobs located in a single central business district (CBD), (ii) labor–leisure trade-offs are absent, (iii) congestion is the only externality, (iv) congestion at a location depends only on the number of commuters passing, regardless of the timing (congestion is static), (v) tax revenue recycling is absent, and (vi) there is only one travel mode – the car – to commute to work. In such a framework, the major long-term effects of congestion taxes are achieved through greater population-densification toward the city core and a reduction in city size (so a denser and more compacted city), following the desire of residents to shorten their commutes and move their residences toward the CBD in response to higher travel costs. It is further shown that marginal cost pricing or Pigouvian tolling is a first-best policy, reflecting the wedge between the marginal private and marginal social costs of a trip. For some of these analyses see Solow (1972, 1973), Mills and de Ferranti (1971), Kanemoto (1976), Arnott (1979a, 1979b), Pines and Sadka (1985), Wheaton (1998), Brueckner (2001, 2007), and Kono and Joshi (2012).3 Studies that have relaxed features of the traditional monocentric setup while still considering one exogenous CBD, revealed that the centralizing effect of a congestion tax on residences depends on labor–leisure trade-offs, the degree of substitutability between the car and public transit, preferences for location and lot size and tax revenue redistribution among households. When labor–leisure trade-offs are included, a congestion toll can either contract or expand the residential part of the monocentric city, and either flatten or steepen the rent and density functions. The reason is because the congestion toll also reduces the CBD labor supply causing employers to offer higher wages. So, while the travel cost increase (due to the toll) causes the city to shrink and residential densities to rise, the income increase induced by the toll expands the city and flattens densities. The combined effect of these two influences is thus ambiguous. Assuming a congested car mode and an uncongested public transit mode, Buyukeren and Hiramatsu (2016) show that the modal substitution effect limits the centralizing force of the congestion toll because commuters can save on travel costs by switching to public transit in response to tolling and still travel from farther away. It is further shown that an equal lump sum redistribution of the toll revenues among residents increases housing demand and

Urban form and the pricing of transport and parking  75

causes a decentralizing effect because marginal utility of income is higher in suburban areas.4 Therefore, and in contrast to past findings from congested-city models with a single travel mode and no revenue-recycling, anti-congestion policies may now cause urban expansion. A qualitatively similar result is also found when dynamic bottleneck congestion is added to the monocentric city (Arnott, 1998; Gubins and Verhoef, 2014;  Takayama and Kuwahara, 2017; Fosgerau et al., 2018; Fosgerau and Kim, 2019).5 By eliminating the queuing time, optimal road pricing induces individuals to spend more time at home and to have larger houses, causing a development pattern characterized by dispersed, low-density residential housing (Gubins and Verhoef, 2014). Several studies have further investigated optimal (second-best) congestion tolling in the monocentric city when multiple urban externalities (Arnott, 2007; Verhoef, 2005) or distortions from taxation of labor income (Tikoudis et al., 2015) or from housing market regulations (Tikoudis et al., 2018) exist in addition to traffic congestion.6 In general, the optimal (secondbest) congestion toll requires an adjustment relative to Pigouvian levels when unpriced externalities (Arnott, 2007; Verhoef, 2005) or tax-induced distortions outside the transport system (Tikoudis et al., 2018) are present.7 This adjustment occurs because a congestion toll affects not only traffic congestion but also labor costs, land-use patterns, and rents, and thereby interacts with other unpriced urban externalities or with pre-existing distortions. In particular, a downward adjustment relative to the Pigouvian level is called for if agglomeration economies are reduced when workers travel less to the CBD in response to the congestion toll (Arnott, 2007).8 If agglomeration economies exist, then in addition to having a congestion toll to internalize the congestion externality, we would also need a production or labor subsidy to internalize the agglomeration externality. In the absence of such a subsidy, the congestion toll being the only instrument, it would have to do double duty. So, it would have to be lower. It is also further shown that with a pre-existing labor tax, only a part of the optimal space-varying congestion tax corresponds to road externalities, with the remaining concerning the Ramsey–Mirrlees component (Tikoudis et  al., 2015). Though firms’ land-use decisions are exogenous and a characterization of the urban spatial equilibrium is not provided, the insights from these studies highlight the importance of considering the interplay of multiple urban externalities and types of pre-existing distortions when designing congestion pricing in second-best settings.9 Lastly, it is known that unpriced traffic congestion in monocentric city models causes excessive travel and excessive aggregate land in urban use (known as urban sprawl). In such models, a Pigouvian congestion tax may lead to an optimum city that is more compact and smaller in aggregate urban land use than the one with unpriced congestion. Yet, theoretical models of job dispersion (Anas and Rhee, 2007) and polycentric urban land use (Anas and Pines, 2008) reveal that the optimum urban form is often more spread out than if congestion remained unpriced. For instance, Anas and Pines (2008) examined two monocentric cities, one initially larger than the other (say due to an exogenous amenity). In this setting first-best and secondbest congestion tolling in both cities makes the bigger city smaller in land area while enlarging the smaller city and in such a way that the combined urban land area of the region may actually increase. While congestion tolls induce commuters to move closer to their CBDs (intracity effect), there is an intercity effect resulting from some commuters in the larger and more congested city relocating to the smaller city to avoid their high congestion toll. If the intercity effect is more powerful than the intracity effect, the sum of the two cities’ land areas increases due to congestion pricing.

76  Handbook on transport pricing and financing 

5.1.3 Congestion Pricing in Polycentric Cities Responding also to the criticisms of the monocentric model, some simulation-based general equilibrium models have relaxed the monocentricity assumption and explored the land-use effects of congestion pricing in polycentric cities with the endogenous location of jobs and residents. Some models focus on the interactions between multiple urban externalities and urban structure (Anas and Kim, 1996) and also marginal cost pricing (Zhang and Kockelman, 2014, 2016a), while others examine the efficiency and land-use effects of second-best congestion pricing schemes (Verhoef and Nijkamp, 2008; Zhang and Kockelman, 2016a, 2016b). For example, Anas and Kim (1996) develop a spatial model of polycentric land use to explore the interplay between congestion due to home-to-work commutes and agglomeration externalities for consumers. Though this study does not test congestion toll efficiency, it shows that strong congestion externalities disperse urban form (with more job centers dispersed throughout space), while strong shopping agglomerations favor more compact forms (with fewer and more job-rich centers).10 On the other hand, Anas and Xu (1999) use a spatial general equilibrium model with endogenous location of industry, residents, and traffic congestion to examine how a Pigouvian congestion toll would change the spatial dispersion of jobs and residences. They found that the addition of a congestion toll could disperse producers away from the regional center while centralizing households, thus bringing jobs and workers closer together but not necessarily toward the center.11 However, because this model does control for the agglomeration economies that may cause firms to locate close to one another, it can somewhat still misestimate the congestion pricing effects on job dispersion. Later, Zhang and Kockelman (2014) using a spatial general equilibrium model with both congestion and agglomeration externalities interactions, compare socially optimal land-use patterns to those under a free-market equilibrium. They found that a marginal cost pricing strategy may lead to job decentralization and residential densification. Recently, Zhang and Kockelman, (2016a) extended their (2014) model by exploring the land-use effects of marginal cost pricing and two second-best congestion pricing policies, a distance-based VMT tax and a cordon toll, in both monocentric and duo-centric (two-ring) city settings where household and firm locations and wage rates are endogenously determined. Their simulation results show that marginal cost pricing in the polycentric setting can cause jobs to leave the CBD and relocate to a relatively dense but suburban ring. A distance-based VMT tax generates lower welfare improvements, but stimulates similar land-use effects. A cordon toll imposed in monocentric cities may agglomerate all firms in a smaller CBD or reagglomerate parts of firms in a sub-center ring of development. Finally, Zhang and Kockelman (2016b) develop a spatial general equilibrium model that also includes endogenous congestion and agglomeration externalities, but land-use decisions by firms and households are endogenous across a continuous space. Focusing on the interaction between urban externalities and land-use patterns, the study uses simulations to examine the efficiencies of first-best interventions and second-best pricing and land-use policies, such as an urban growth boundary and firm cluster zoning regulations. Congestion pricing policies alone are shown to improve social welfare only in heavily congestion cities. Urban growth boundaries may partially correct distortions in both transport and labor markets, but may worsen land market distortions via residential rent-escalation effects. In contrast, firm cluster zoning regulations can generate welfare improvement closer to first-best levels by regulating firm’s locations while not resulting in excessive escalation of housing rents.

Urban form and the pricing of transport and parking  77

Even though polycentric models with multiple externalities and endogenous locations of firms and residences provide already good insights into the general effects of congestion pricing on urban structure and urban form, there are still opportunities to make these models more realistic. For example, allowing for alternative travel modes, parking availability, and trip scheduling flexibility is important when discussing congestion toll effects. A model that enables a gradual, dynamic city evolution is also important to explore. The one-shot, static equilibrium is never achieved in practice. In reality, most cities already exist, and population and economic activity regularly expand, in the midst of uncertainty and imperfect information, along with market speculation. 5.1.4 Remaining Chapter Structure The rest of the chapter is organized as follows. Section 5.2 develops a simulation-based twoperiod, perfect-foresight, spatial general equilibrium model that includes economic activity growth, vacant land in the downtown area, alternative travel modes, traffic congestion, and externalities from temporary uses on vacant land awaiting development, all from a spatial perspective in the most basic form. Within this framework, we provide insights into the interaction of traffic congestion and land-use patterns, and the possible effects of first- and second-best congestion tolling. Even though the model abstracts from other important features in cities, it offers a first step to examine the interplay between traffic congestion and urban vacant land, which has been overlooked in past general equilibrium models. While there are several reasons for the existence of urban vacant land, unpriced externalities from temporary uses on vacant land (e.g., surface parking lots) may lead to inefficient amounts of vacant land in city cores, leaving less land for other uses such as commercial and residential uses, with feedbacks on urban form and spatial structure. The results of the numerical simulation illustrating the free-market, first-best, and second-best equilibria are presented and discussed in Section 5.3. Next, Section 5.4 focuses on the pricing of parking and urban form. It starts with a short review of theoretical papers exploring the interconnections between parking prices, parking supply, mode choice and urban form. Graphical analyses are then used to discuss the effects of employer-paid parking and minimum-parking requirements on travel mode choice and urban form.12 Finally, Section 5.5 offers conclusions and avenues for future work.

5.2 ANALYTICAL FRAMEWORK13 5.2.1 Model Setup Consider an open linear city with two transport systems (a congested auto system and a noncongested transit system) and a single endogenous downtown core (denoted as CBD) located on the far left-hand side, and the city boundary located on the farther right-hand side of the line. The only retailing and transport node, denoted as the trade node, is located at the very end of the far left-hand side of the CBD (that is, located at zero). Proximity to this trade node is the only attribute that distinguishes land parcels in this city, with the trade node the point where all city activities position themselves. Thus, a location in the city, x, is described by its distance to the trade node.

78  Handbook on transport pricing and financing 

Possible land uses outside the trade node include irreversible durable uses such as residential and commercial buildings, and reversible temporary uses such as surface parking lots on developable land or agriculture. Demolition of residential and commercial buildings once they are built is assumed to be prohibitively expensive and therefore is not considered. The CBD is the city zone where all the jobs locate and commercial activity occurs from commercial landowners outbidding residential landowners and farmers. The CBD houses a tradable commercial sector that employs all the resident households and produces a market good that is shipped to the trade node, where it is sold to resident households and exported to other cities. The residential zone is the city area over which residential landowners outbid commercial landowners and farmers. Land outside the city boundary is under agricultural use. The linear city setup is illustrated ahead in Figure 5.1. Trade node Vacant land at t =1 |_______________|_____________|___________________|____________|___________ 0 C1 a C2 b R1 c R2 d Downtown core or CBD

Residential Area

Agriculture

Note: Where C1 and C2 represent allocated commercial land in periods 1 and 2, respectively. R1 represents residential land in period 1 and R2 is added residential land in period 2. xa denotes the endogenous outer developed edge of the downtown core and xb is the endogenous inner boundary of the residential zone. Since xa < xb , then xb - xa units of downtown land are vacant in period 1 awaiting commercial development in period 2. Finally, xc and xd are the endogenous outer spatial boundaries of the residential zone (or city boundaries) in periods 1 and 2, respectively.

Figure 5.1  Equilibrium land-use pattern with vacant land in period 1 Resident households combine work and shopping trips. They travel to the trade node from their home sites either by car or public transit, buy the tradable good at the trade node, and then disperse to their work sites located outside the trade node by walking (at cost zero) within the downtown core. If commuting by car, resident households park at the trade node. For simplicity, all non-visitor parking in the downtown core is located at the trade node where there is unlimited provided free underground parking to all city residents.14 Non-resident households, denoted as visitors, reside outside the city and travel to the downtown core solely for tourism purposes by either car or train. Visitor parking at the trade node is exogenously fixed, underground, and limited to N v parking spaces, though temporary visitor parking may be endogenously supplied outside the trade node but within the downtown core as a temporary land use of the CBD vacant land awaiting development, as we will discuss later.15 For simplicity, visitors are indifferent between parking at the trade node or in surface lots outside the trade node. Once at the CBD, visitors disperse throughout the CBD area by walking at cost zero. We examine urban form and urban spatial structure in two time periods, t = 1, 2, of fixed but unspecified length. We also consider that an exogenous growth mechanism from an increase in the export demand for the tradable good exists between the two periods. Landowners are assumed to be absentee and profit-maximizing agents who anticipate this future growth with certainty.16 At the beginning of period 1, each parcel of land must be allocated to a sequence of land uses such that the present value bid of the sequence is the highest. Next, we describe in detail more of the features of the model.

Urban form and the pricing of transport and parking  79

5.2.1.1 Commercial and residential sectors outside the trade node The tradable commercial sector is competitive and produces a tradable good, Q, with a Leontief technology that combines λ units of land and μ units of labor to produce one unit of Q. There is no storage between periods, so some Q is exported and the rest is for local consumption. Regardless of its use, tradable production must be transported to the trade node. Let τ represent the per unit distance shipping cost to the trade node per period. The housing-supply sector uses 1 unit of land to produce single-family fixed-size dwellings with parking. Therefore, one unit of residential land bundles housing services and residential parking, which we assume for simplicity to be time and space invariant. 5.2.1.2 City residents All city residents are identical households and provide a single unit of labor to the tradable commercial sector per period in exchange for a wage wt. We remind the reader that resident households combine work and shopping trips when traveling to the CBD. Once they reach the CBD trade node, they disperse throughout the downtown core by walking. Walking costs are zero within the CBD. Resident households traveling by public transit pay a time-invariant fixed cost Fp and a time-invariant unit distance travel cost of tp. On the other hand, resident households traveling by car incur a time-invariant fixed cost Fc, with Fc > Fp, a time-invariant per unit distance travel cost tc such that tp > tc, and a congestion cost, g N tc . Since road capacity is fixed and invariant in both periods, the congestion cost increases with the number of drivers in the city in a given period, N tc , and therefore is time variant. The difference in the auto variable unit commuting costs relative to transit together with the congestion cost means that it is cost-effective for residents using public transit to live closer to the CBD than car commuting households. As a result, there will be a boundary in the city in each period, denoted as the modal boundary xts, where on the CBD-side of the boundary public transit users live and on the other side car users live (see Figure 5.1 for illustration). The modal boundary is also a location where a resident household is indifferent between the two travel modes for trips to the trade node. For analytical tractability, we abstract from substitutability in consumption and assume resident households consume in each period one unit of the tradable good, one unit of an import good and one unit of residential land, subject to a budget constraint to attain an exogenous utility. The number of resident households in either period, Nt, is endogenous and because the city is open, migration will ensure that every resident household obtains the exogenous utility level. Let p Z be the exogenous time-invariant price of the import good. To ensure that local con1 sumption of Q does not exhaust total tradable production, > 1 is assumed to hold. m

( )

5.2.1.3 Resource balance and growth in export demand Total tradable production in a given period, Qt, satisfies the external and domestic demands and is represented as:

Qt = f ( pt , G t ) + N t * 1, with G1 < G 2 (5.1)

80  Handbook on transport pricing and financing 

implying that the exogenous export demand in period t, f ( pt ,G t ), grows between the two time periods. Γt is an exogenous demand-shift parameter that embeds all the information to predict the price-quantity demand relationship in a given period, and it is known with certainty in period 1. The endogenous unit price at which the tradable good is sold in period t is given by pt. 5.2.1.4 Land market outside the trade node Landowners have perfect foresight of what their land rent will be in both periods for all possible development strategies, that is, developing commercially or residentially in either period, and leaving vacant land (even if for temporary use) in both periods. There are many possible temporary uses for urban vacant land awaiting development, but we assume, following what we see in downtown areas worldwide, that the highest return temporary use is for temporary commercial surface parking. Surface parking is an attractive temporary use because it is easily reversible, generally inexpensive to implement, yet can rapidly generate revenue for landowners. Assuming that each surface parking space takes s units of land (set to 1) and is sold for simplicity at an exogenous time-invariant price θ per parking q space, the return on temporary surface parking per unit of land is f = . s Land beyond the city boundary is used for agriculture, which earns an exogenous spatially increasing but time-invariant return RA ( x ) per unit of land, with RA ( x ) < f inside the CBD, suggesting that land farther away from the CBD has higher-quality soils for agricultural purposes. 5.2.1.5 Vacant land in period 1 and urban form Note that landowners have perfect foresight and know Γ2 in Equation (5.1) with certainty in period 1. Therefore, all development decisions, whether executed in the first or second period, t are made simultaneously in period 1. Provided that > t p is satisfied, residential development l lies more distant from the trade node than commercial development in each period. However, it is not certain that all second-period development will lie beyond the first-period city boundt ary. Depending on the relative magnitudes of and t p, it may be advantageous for landowners l to preserve in period 1 some parcels between the commercial and residential zones for secondt period commercial use. This type of land-use pattern is more likely to occur the greater is l relative to t p. Since we are interested in examining how congestion tolls impact urban form in the presence of road congestion and externalities from temporary land uses of vacant land in t downtown cores, we assume that is sufficiently bigger than t p so that leapfrog development l t emerges in period 1.17 In our simulation, we also calibrate the model so that greatly exceeds l t t p to confine our analysis to this type of land-use pattern. Below it is further shown that and l t p are the absolute value of the slopes of commercial and residential bid land-rent functions. Thus, provided that commercial bid land rents decrease more rapidly with distance from the trade node than residential rents, a land-use pattern with vacant land in period 1 as seen in Figure 5.1 and illustrated numerically in Figure 5.2 may emerge.

Urban form and the pricing of transport and parking  81

Figure 5.2  Land-rent configuration in the free-market equilibrium The land-use pattern equilibrium described in Figure 5.1 allows us to explore temporary land uses of CBD vacant lots that may generate additional externalities to the urban system. For illustration purposes, we explore in this chapter the case of additional city congestion associated with temporary surface visitor parking on vacant land in the downtown core.18 Future work may explore other temporary land uses on vacant land such as green space or billboards.19 5.2.1.6 City visitors Visitors commute to the CBD either by car or public transit. Commuting by train incurs no congestion and costs Fv to visitors in each period. Visitors commuting by car pay a parking fee θ per parking space, and park either underground at the trade node or at temporary visitor parking lots located outside the trade node.20 In contrast to the case of city residents and for analytical tractability, we abstract from issues of visitors’ commuting distance and fixed costs related to car ownership. Thus, the costs of a visitor auto commuting include just a parking fee and a congestion cost, g N tc . If q + g N tc < F v , visitors commute by car as long as visitor parking exists in the downtown area. We assume for analytical tractability that visitor parking at the trade node is always full (regardless of the period) and that this information is known to non-residents in both periods. Visitors also know that in period 1 there is temporary visitor parking outside the trade node. The number of auto commuting visitors parking outside the trade node in period t, Ntv, is defined by the amount of downtown vacant land in period t (equal to xb − xa in period 1 and zero in period 2) divided by the units of land taken by one parking space (that is, s).

( )

( )

5.2.1.7 Traffic congestion externality For expository purposes, the congestion cost is for now a linear function of the number of cars x -x entering the downtown core, g N tc = N tc , with N1c = xc - x1s + N1v + N v = xc - x1s + b a + N v s and N 2c = xd - x2 s + N v . Because temporary surface parking is entirely converted to commercial use in period 2, it follows that N2v = 0. In our numerical exercise, we use a convex congesy tion function, g N tc = b éë N tc ùû with b,y > 0 .

( )

( )

82  Handbook on transport pricing and financing 

5.2.2 Urban Equilibrium Conditions The urban equilibrium conditions with unpriced externalities are presented in Table 5.1 and consist of 18 equations with 18 endogenous variables:

Q1, Q2 , xa , xb , xc , xd , x1s , x2 s , P1, P2 , w1, w2 , R1Q , R2Q , R1HP , R2HP , R1HC , and R2HC .

Table 5.1  Free-market urban equilibrium conditions Period 1

Period 2

1

Q1 = f ( P1,G1 ) + mQ1

2

w1 = R

HC 1

+ P1 + Pz + F

+ xc - x1s +

Q 2 = f ( P2 ,G 2 ) + mQ2 c

xb - x a + Nv s

w2 = R2HC + P2 + Pz + F c + x d - x2 s + N v

3

w1 = R1HP + Pz + P1 + F p

w2 = R2HP + Pz + P2 + F p

4

P1 = mw1 + lR1Q

P2 = mw2 + lR2Q

5

lQ1 = xa

lQ2 = xb

6

mQ1 = xc - xb

mQ2 = xd - xb

7

HC 1

R

- t xc = R A ( xc ) c

xb - x a + N v + xc s 1 + t p - tc

Fc - F p +

8

x1s =

9

R1Q -

10

R1HP - t p xb +

R2HC - t c xd = RA ( xd ) x2 s =

F c - F p + N v + xd 1 + t p - tc

txa =q l 1 1 é Q txb ù é R2HP - t p xb ù = q + R2 ë û l úû 1+ r 1 + r êë

The first equilibrium conditions ensure that in each period export demand plus local demand for the tradable good equals total production of the tradable good. The second and third conditions ensure that in each period household budget balance is in equilibrium for private car and public transit users, respectively. This requires that the household wage, wt, equals expenditures on housing, RtHC or RtHP , on tradable goods, Pt *1, on import goods, PZ *1, and on fixed transport costs including congestion costs for drivers. x -x Congestion costs in period 1 are given by N1c = xc - x1s + N1v + N v (with N1v = b a ), s while in period 2 are N 2c = xd - x2 s + N v . Since there are only two periods, it is not efficient to have vacant land in the downtown core in period 2. As a result, N 2 v = 0 and car users in period 2 are composed of xd - x2 s resident households and N v visitors. Note also that RtHC is the car users bid land rent for a residential site located at the trade node, that is, at x = 0, while RtHP has the same interpretation but refers to public transit users.

Urban form and the pricing of transport and parking  83

As households move away from the trade node (that is, move away from x = 0) their maximum WTP for a unit of residential land decreases in the amount of the variable commuting cost they must incur. Therefore, the rental price of residential land for car users is given by RtHC - t c x and for transit users is given by RtHP - t p x , where the slopes are determined by unit distance travel cost. Since F c + g N tc > F p and t p > t c , it follows that RtHP > RtHC and transit users have steeper bid rents than car users in both periods. The fourth conditions enforce zero profits on the production of the tradable good each period as the market is perfectly competitive. This sets the price of the exported good, Pt, equal to production costs from labor,mw1, and land, lRtQ . RtQ represents the bid rent for a unit of land for commercial use at x = 0 in period t. If RtQ > RtHP > f, land at x = 0 is allocated to commercial use in period t. If the reverse occurs but RtHP > f, land is allocated to residential t use. Note however that the rental price of commercial land is given by RtQ - x , where the l t slope is also determined by unit distance travel cost to the trade node. Provided that greatly l exceeds tp, commercial landowners have steeper bid land rents than residential landowners in the downtown core area. The fifth conditions represent the amount of land for commercial use, lQt , set equal to the boundary of the commercial region, xa, in the first period and xb in the second period. The sixth conditions characterize the amount of residential land, mQt , necessary to house resident households (who also constitute the city labor force) in period t which equals the area of the residential land zone, the difference between xc and xb in period 1 and xb and xd in period 2.21 The seventh conditions state that residential bid land rents at the city boundary adjusted for the variable commuting costs to the trade node must equal the agricultural rent at that location.22 The eighth conditions define in each period equality of transport costs between public transit and private vehicle commuting at the modal boundary, xts , t = 1, 2. That is, at the modal boundary, xts, resident households are indifferent between the two travel modes to the trade node. The ninth condition requires that the bid rent for commercial land at the commercial boundt ary xa in the first period, R1Q - xa , equals the unit return on vacant land from its temporary use l as a surface parking lot for visitors, θ.23 There is no analogous condition for the second period because all downtown vacant land in period 2 is converted to commercial use in period 2. Landowners in the vicinity of xa must decide only whether to develop land for commercial use in period 1 or period 2. If postponing development to period 2, they use their land temporarily as surface parking for visitors. Therefore, provided that commercial land rent is greater than θ in period 1, these landowners develop the land for commercial use.24 Finally, remember that xb defines the boundary separating land uses in the central part of the city (commercial use and temporary surface parking in period 1; commercial use only in period 2) and land uses outside the central core (residential use and agricultural land). Moreover, bid land rents for commercial land decrease with distance from the trade node because of the varit able transport costs (with the unit transport cost of goods per mile equal to ). l The last equilibrium condition then requires that landowners determine xb from the intertemporal relationship between bid land rents at this location. This condition is related to landowners

( )

84  Handbook on transport pricing and financing 

having perfect foresight about the expected export growth in period 2 and that urban development is irreversible. Thus, at location xb landowners must be indifferent between converting the land to residential use in period 1 or leave it vacant under surface parking and then convert it to commercial use in the second period. Let r represent the discount rate. At xb the present discounted value of development for residential use must equal the present discounted value of leaving the land vacant (but temporarily as a parking lot) in period 1, earn return θ per unit, t and then be used for commercial use in period 2, with a land rent of R2Q - xb . l It is not possible to obtain analytical (equilibrium) results for this model. Therefore, the results of a numerical illustration fully consistent with the analytical model are presented next to discuss the effects of road congestion, externalities from temporary uses of downtown vacant land and congestion pricing on urban form, urban structure, land rents and travel mode choices. 5.2.3 Calibration Calibration of the parameter values, as seen in Table 5.2, is set to illustrate qualitative results and a spatial discontinuity in period 1 as seen in Figure 5.1.

Table 5.2   Parameters values Description

Parameter

Value

Export demand shift in period 1

Г1

2

Export demand shift in period 2

Г2

3

Export demand exponent

δ

–0.2

Labor per unit produced

μ

0.5

Land per unit produced

λ

0.1

Unit transport cost for firms

t

0.1

Unit transport cost for transit users

tp

0.25

Unit transport cost for auto users

tc

0.1

Car fixed cost

Fc

0.2

Public transit fixed cost

Fp

0.1

Discount rate

r

0.05

Value of agricultural land in both periods

RA

0.25

Value of vacant land under surface parking

θ

0.4

Size of a surface parking space

S

1

Price of imported goods in both periods

Pz

1

Congestion cost per mile

β

0.01

Congestion exponent

ψ

2

Nv

0

Visitor parking at the trade node in both periods

Urban form and the pricing of transport and parking  85

The functional forms used are for: (i) the export demand f ( pt , G t ) = G t ptd and (ii) the cony gestion function b éë N tc ùû . Since the supply of visitor parking at the trade node, N v , is a parameter whose value does not change our qualitative findings, we set N v = 0 . With unpriced traffic congestion, the car commuting costs for a resident household are y y x -x ù é F c + t c x + b ê xc - x1s + b a ú in period 1 and F c + t c x + b éë xd - x2 s ùû in period 2. When s û ë y -1 a congestion tax set to N tcyb éë N tc ùû is charged to each resident household traveling by y x -x ù é c car, car commuting costs become F + t c x + (1 + y ) b ê xc - x1s + b a ú in period 1 and s û ë y F c + t c x + (1 + y ) b éë xd - x2 s ùû in period 2. If visitor parking is a temporary land use for downtown vacant land with no negative exterx -x nal effects, the term b a is suppressed in the previous expressions. If a tax per surface s parking space is charged to landowners awaiting to develop their downtown vacant parcels, y tx then Table 5.1 condition 9 adjusts to R1Q - a = q + yb éë N tc ùû . l In all policy scenarios, tax revenues are equally lump sum distributed to resident households.

5.3 EFFECTS OF A CONGESTION TAX ON TRAVEL MODES, SPATIAL STRUCTURE AND URBAN FORM 5.3.1 When the Temporary Use for Downtown Vacant Land Does Not Create an Externality Let´s consider first the case where the temporary use of downtown vacant land does not generate an externality. Given the existence of traffic congestion from residents commuting by car to the trade node, it should be no surprise that the free-market equilibrium is inefficient. With unpriced traffic congestion (the free-market case) the monocentric city occupies too much urban land. The first-best policy in this scenario consists of a congestion tax charged to each resident commuting by car equal to the marginal external cost. The level of this tax was calculated numerically for both periods. Table 5.3 shows the results of the free-market and of the congestion tax on urban form (city size), travel mode choices and urban structure (land-use arrangements). Figure 5.2 illustrates the spatial pattern of land-use rents in the free-market equilibrium. In contrast to the traditional static monocentric city model, we observe in Figure 5.2 that the urban land-rent profile in each period is not characterized by a continuous function. In both periods there is a discontinuity point at boundary xb. From condition 10 in Table 5.1, these discontinuities offset each other when we discount period 2 land rent. This can be easily verified by plugging in Table 5.1 condition 10 the parameter values provided in Table 5.2 and Table 5.3 equilibrium values.

86  Handbook on transport pricing and financing 

Even though Figure 5.2 illustrates the rent profile for the free-market equilibrium, a similar qualitative land-rent configuration is expected to emerge when a congestion tax is imposed. The reason is because leapfrog development in period 1 results from the intertemporal planning of commercial landowners who anticipate future export growth with certainty.25 However, traffic congestion is a market failure that, when uncorrected, distorts urban form (city size), urban structure (land-use arrangements), and travel mode choices. From Table 5.3, when a congestion tax is charged, the city physical size is reduced by comparison to the unpriced situation. Thus, a congestion tax reduces urban sprawl (that is, the excessive aggregate land area in urban use) in both periods.26 In addition, note also from Table 5.3 that a congestion tax affects the amount of vacant land in the downtown core left for future development in period 1. By increasing the commuting costs by car, the congestion tax motivates resident households to move closer to the trade node. This reduces the demand for residential locations near the city boundary (reducing pressure on land rents at the periphery), and increases the demand for residential locations closer to the trade node (creating upward pressure on land rents in these locations). Increased residential land rents in adjacent areas of the downtown core push the lower edge of the residential area (xb) toward the trade node, reducing vacant land held for future use at the core. In addition, a congestion tax moves the modal boundary outwards in both periods, meaning that more resident households commute by public transit to the trade node in comparison to the free-market case. Thus, a congestion tax not only affects residents’ home location choices, but it also induces changes in their travel mode choice. Furthermore, by reducing city size, a congestion tax affects the number of workers residing in the city. In particular, fewer workers live in the city in period 1 compared to the free-market case. Since tradable firms use labor and land in fixed proportions, firms have also in period 1 lower demand for commercial land. This pushes the outer edge of the commercial area (xa) slightly toward the trade node, increasing the amount of vacant land held for future development. There are then two opposing effects on downtown vacant land from the use of a congestion tax. However, because in our simulation the effect on xb is stronger than the effect on xa, the amount of downtown vacant land held for future development following a congestion tax is smaller compared to the free-market case. Future work could explore how a congestion toll interacts with vacant land speculation in downtown areas and the goal of urban infill development. Finally, the level of the congestion tax is not constant over time. As the city grows from period 1 to period 2, more people live in the city and at farther distances from the trade node (as density is fixed due to fixed land consumption). As a result, the congestion level changes between the two periods, requiring an upward adjustment in the optimal tax level in period 2.

Urban form and the pricing of transport and parking  87

Table 5.3  Free-market equilibrium and the optimal solution when the temporary use of urban vacant land has no external effects Description

Parameter

Free-Market

Congestion tax

Production of tradable good

Q1

3.5364

3.5350

Price of tradable good

P1

1.8514

1.8549

Salary/worker

w1

3.5520

3.5592

Outer edge of the commercial area

xa

0.3536

0.3535

Lower edge of the residential area

xb

0.5201

0.5191

Modal boundary

x1s

0.8120

0.9985

City boundary

xc

2.2883

2.2866

HC 1

0.4788

0.4787

HP 1

0.6006

0.6284

Q 1

R

0.7536

0.7535

Period 1

Residential land rent for a car user at x=0 Residential land rent for a transit user at x=0 Commercial land rent at x=0

R R

Period 2 Production of tradable good

Q2

5.2013

5.1905

Price of tradable good

P2

2.0426

2.0640

Salary/worker

w2

3.8508

3.8777

Modal boundary

x2s

0.9739

1.3145

City boundary

xd

3.1208

3.1143

Residential land rent for a car user at x=0

R2HC

0.5621

0.5614

HP 2

0.7082

0.7586

Q 2

1.1724

1.2515

Residential land rent for a transit user at x=0 Commercial land rent at x=0

R R

Other variables of interest Congestion tax in period 1

0.0332

Congestion tax in period 2

0.0648

Total value of urban land rent (in present value)

206.3052

213.3424

Total amount of vacant land in period 1

0.1665

0.1655

Car users in period 1

83.5%

72.9%

Car users in period 2

82.6%

69.3%

Transit users in period 1

16.5%

27.1%

Transit users in period 2

17.4%

30.7%

5.3.2 When the Temporary Use of Downtown Vacant Land Creates an Externality Let’s consider now the case of two endogenous sources of traffic congestion: resident car commuters and visitor cars. Since N v = 0 , visitor cars are just related to the amount of visitor parking in the downtown core which is provided on vacant lots awaiting development. However, landowners do not account for the effect on traffic congestion of this temporary

88  Handbook on transport pricing and financing 

Table 5.4  The free-market, the first-best and second-best equilibria when the temporary use of urban vacant land as parking lots creates external effects Description

Parameter

Free-Market Second-best

First-best

Production of tradable good

Q1

3.5348

3.5333

3.5371

Price of tradable good

P1

1.8555

1.8596

1.8495

Salary/worker

w1

3.5603

3.5686

3.5553

Outer edge of the commercial area

xa

0.3535

0.3533

0.3537

Lower edge of the residential area

xb

0.5201

0.5190

0.5186

Modal boundary

x1s

0.8403

1.0360

1.0364

City boundary

xc

2.2875

2.2856

2.2872

HC 1

0.47875

0.4786

0.4787

HP 1

0.6048

0.6340

0.6342

Q 1

R

0.7535

0.7533

0.7183

Production of tradable good

Q2

5.2009

5.1900

5.1864

Price of tradable good

P2

2.0435

2.0651

2.0722

Salary/worker

w2

3.8516

3.8787

3.8855

Modal boundary

x2s

0.9739

1.3144

1.3135

City boundary

xd

3.1205

3.1140

3.1118

Residential land rent for a car user at x = 0

R2HC

0.5621

0.5614

0.5612

HP 2

0.7081

0.7586

0.7582

Q 2

1.1767

1.2572

1.2941

Congestion tax in period 1

0.0312

0.0313

Congestion tax in period 2

0.0648

0.0647

Period 1

Residential land rent for a car user at x = 0 Residential land rent for a transit user at x = 0 Commercial land rent at x = 0

R R

Period 2

Residential land rent for a transit user at x = 0 Commercial land rent at x = 0

R R

Other variables of interest

Tax per unit of surface parking land

0.035

Total value of urban land rent (in present value)

206.5574

213.792

214.360

Total amount of vacant land in period 1

0.1666

0.1657

0.1649

Car users in period 1

81.88%

70.73%

70.72%

Car users in period 2

82.55%

69.35%

69.35%

Transit users in period 1

18.1%

29.27%

29.28%

Transit users in period 2

17.5%

30.65%

30.65%

land use of vacant land.27 Note that the free-market scenario in this new setting differs from Table 5.3 because it represents the equilibrium when these two sources of traffic congestion are unpriced. Table 5.4 presents the results in this case with free-market and under two policy scenarios: (i) when a policy mix is implemented (first-best), and (ii) when a congestion tax on residents commuting by car is the only instrument in use (second-best). The first-best pricing policy requires a congestion tax to be charged to resident auto commuters, and a tax per unit of vacant land under surface parking charged to CBD landowners. In the

Urban form and the pricing of transport and parking  89

optimum, the levels of these two taxes should equal the marginal externality. These levels were determined numerically and Table 5.4 shows their corresponding values in each period as well as this policy mix impacts on the city's physical size and land-use arrangements in each period. Compared to the free-market case, both city size (and thus urban sprawl) and vacant land (and thus parking lots) in the downtown core are lower under the policy mix. This occurs because the tax on surface parking reduces the unit return of vacant land as visitor surface parking, while the congestion tax increases residents commuting costs by car. The tax on surface parking pushes xa away from the trade node (so more CBD land is allocated to commercial use in period 1) but xb toward the trade node (so more land is also allocated to residential use near the downtown core in period 1). Our result then suggests that taxing surface parking in the CBD may disincentive excessive land speculation and encourage development to high-value uses downtown. The congestion tax also pushes xb toward the trade node due to its impacts on travel costs by car, incentivizing residents to migrate inwards and reducing the city boundary in both periods. In the absence of a tax on surface parking, the congestion tax is at a level below its first best in period 1, but the reverse occurs in period 2. Since the city size is smaller (and thus population and congestion are also smaller) in period 1 due to excessive vacant land being held at the downtown core, the congestion tax adjusts to a level below its first best. However, because more land is converted to commercial use in period 2 compared to first-best, the city size in period 2 is also larger than what is optimal (so that it accommodates all the new workers), and thus, there is a need for a higher level of the congestion tax (because as city population increases and more workers live farther away from the trade core, traffic congestion worsens).

5.4 PARKING PRICES, PARKING AVAILABILITY, AND URBAN FORM28 Parking accounts for a major part of the social costs of car trips to central cities (Shoup, 2005, 2018). Furthermore, parking is very subsidized by many governments, employers, and businesses. However, parking subsidies are an invitation to drive everywhere because they collectivize the cost of parking. In absence of such subsidies and disregarding the external costs of parking (e.g., air pollution and congestion from cruising), the individualized parking cost that drivers would incur would be quite high as market parking prices can be very high, particularly in urban cores. Yet, employer-paid parking, a popular transportation fringe benefit in the United States and other countries, incentivizes workers to drive to work alone over walking, biking, or taking public transit. Another feature to consider when discussing how parking pricing affects travel mode choice and urban form is parking availability. After all, all car trips start and end in a parking space, and how easy it is to find a parking space is relevant to commuters. Zoning ordinances that contribute to abundant parking supply, especially in areas well served by public transit, such as minimum-parking requirements, then work at cross-purposes with the goals of reducing solo driving and increasing the use of sustainable travel modes. Since the total cost of providing parking varies by location, the shadow cost of blunt-mandated parking is spatially variant, influencing the location and type of parking supplied and ultimately urban form (Brueckner and Franco, 2018). A good understanding of the many ways pricing of parking affects travel mode choice, land use and urban form is then also necessary if a shift away from the private car toward more sustainable methods of transport and land-use patterns are to be achieved. Comprehension of such a matter will result in the development of appropriate land-use planning strategies, transportation reforms and infrastructure provision to increase the use of sustainable transport modes to

90  Handbook on transport pricing and financing 

places of residence, employment and shopping, and therefore more sustainable urban forms in the long-run. This section provides insights into how minimum off-street parking requirements and employer-paid parking, affect land consumption, car use and urban form. 5.4.1 Existing Studies Even though employer-paid parking is an important feature of workplace parking pricing, there is remarkably little analytical work on the effects of this type of parking subsidy. Most theoretical studies on parking pricing explore the effects of parking in downtown areas and how parking pricing policy affects short-run commuter decisions regarding trip scheduling and frequency, transport mode choice, and parking location. See Inci (2015) for a survey. Existing studies focus on the efficiency of second-best pricing of parking in the absence of congestion tolls (Arnott et al., 1991; Verhoef et al., 1995; Arnott and Rowse, 1999; Anderson and de Palma, 2004); the effects of curbside parking fees on cruising for parking in downtown areas (Arnott and Inci, 2006); the effects of underpricing of parking facilities on social welfare (Calthrop and Proost, 2006); and the effects of parking and transit subsidies on the CBD’s size (Voith, 1998). Other analytical studies focus on the effect of employer-paid parking on the level of optimal congestion charges (De Borger and Wuyts, 2009) and on the optimal level of curbside parking capacity in downtown areas when both urban transport and curbside parking are underpriced (Arnott et al., 2015). Only recently, the effects of employer-paid parking have been addressed to understand its effects on mode choice, road investment, and suburbanization (Brueckner and Franco, 2018). Brueckner and Franco (2018) use a simplified two-zone monocentric model with the center and suburbs connected by a congested road and a public transit line. The study studies a decentralized planning solution, which requires employee – rather than employer – paid parking, congestion tolls, and a tax (subsidy) to offset the road capacity deficit (surplus). The analysis further considers the effect of a switch to employer-paid parking, with the burden of parking costs shifting from auto commuters to employers, thus reducing the wage for all workers. This switch increases road usage and road capacity, while reducing the center’s residential land area. The result is greater suburbanization of the population along with an overall increase in the suburban commute flow to the CBD. Building on the numerical results presented in this study, Section 5.4.4 provides insights into how employer-paid parking affects travel mode choices and urban form. Another set of studies focuses on the effects on the urban form of parking requirements and parking supply in either downtown areas or in shopping malls. Anderson and de Palma (2007) analyze a city where the CBD is surrounded by a zone of parking lots, with the outlying residential area comprised of two zones. The study characterizes the socially optimal configuration of the city and shows that the optimum coincides with an equilibrium in which parking is provided by monopolistically competitive lot owners. Brueckner and Franco (2017) present a two-zone spatial analysis of residential as opposed to employee parking, focusing on the choice among different parking technologies (surface, structural and underground) and on how generous parking minimums may result in lower development densities and over-allocation of land to parking space. Franco (2017) on the other hand explores the spatial effects of restrictions in downtown parking supply on urban welfare, travel modal choices, and urban spatial structure using a continuous monocentric city model with air pollution from commuting and downtown congestion. Ersoy et al. (2016) provide an analysis

Urban form and the pricing of transport and parking  91

of the interaction of mode choice and provision of parking in the shopping-center context, with some shoppers accessing the mall by car and others by public transit. Next, a brief description of the main characteristics of the framework provided by Brueckner and Franco (2017) is presented. Section 5.4.3 then provides a graphical discussion, based on Brueckner and Franco’s (2017) simulation results, of the effects of minimum-parking requirements (MPRs) on urban form. For completeness, evidence is also presented to show that this regulation is a binding constraint, and how the cost of satisfying the default city parking requirement increases the total cost of construction in development projects. 5.4.2 Minimum-Parking Requirements29 5.4.2.1 Setting the stage: Brueckner and Franco (2017) model In the model, all households own a car and thus value residential parking space regardless of the parking regime generating the parking area. Off-street parking provides utility to households by offering more convenience and safety than on-street parking. These benefits are assumed to be greater the larger the amount of parking space associated with the dwelling. On-street parking is not priced or regulated. All households work in a single CBD located at point zero, commute to work only by car, incur commuting travel costs and earn a homogeneous income. Parking is freely available at the workplace and assumed to be on-site and underground for simplicity. While road congestion is omitted for simplicity, the household is assumed to experience parking-related congestion in the neighborhood of residence (e.g., due to cruising for off-site parking), which generates an additional neighborhood-level travel cost. With the provision of off-street parking reducing this congestion, the extra travel cost falls as the average off-street parking area per dwelling in the neighborhood increases, a cost-side benefit in addition to the utility gain from a dwelling’s off-street parking area. The rental price a household is willing to pay for a dwelling embodies payments for both dwelling space and parking area, which are bundled by the residential developer. Residential developers use capital to produce floor space, which is divided into dwelling units. Developers also provide parking areas associated with the dwelling.30 Under any of the three parking regimes (surface, structural and underground), the developer maximizes profit by choosing the dwelling size, parking area per dwelling, residential structural density (capital per unit of residential land, an indicator of building height), and parking inputs, taking into account the maximum rental price households are willing to pay for a dwelling in each location.31 Atomistic residential developers are portrayed as ignoring the collective beneficial effects of their offstreet parking choices per dwelling on the neighborhood parking-related congestion. Therefore, residential off-street parking is undersupplied and on-street parking is in too high demand. Next, we provide insights into the location of different parking regimes and on urban forms with and without MPRs based on Brueckner and Franco’s (2017) numerical example. 5.4.2.2 Location of different parking regimes Figure 5.3 provides insight into which land-use/parking regime is chosen at each location in a linear city. Parcels closed to the CBD command higher rents relative to suburban areas. At each location from the CBD, land is allocated competitively to one of three residential uses which differ by the type of parking provided with a dwelling. Land is allocated to the best and most valuable use in a specific location. As a result, a particular parking regime will be present in a given location if developers using that regime bid more for land than developers using

92  Handbook on transport pricing and financing  6.00

Underground Structural Surface

5.00

Land Rent

4.00 3.00 2.00 1.00 0.00

0

5 10 15 20 25 30 35 40 45 50 55 60

Distance to the CBD

Figure 5.3  Land rent under different parking regimes (CBD located at point zero) the other regimes. Therefore, the relative location of the parking regimes can be inferred by considering the heights and the slopes of the land-rent curves described in Figure 5.3. From Figure 5.3, surface parking is observed in the suburbs, with one of the other parking regimes (structural or underground) observed in the central core. Since both non-surface parking regimes conserve land, their use in the central core, where land is expensive, is to be expected. In the simulation, land rent with structured parking is always dominated at each location by the land rent with either underground or surface parking, Thus, the only two parking regimes that will exist in equilibrium are underground and surface parking. Figure 5.4 provides the equilibrium spatial profile of the parking area supplied per dwelling. It is interesting to note that regardless of the parking regime, the amount of parking area increases with distance from the CBD, being much higher in the suburbs than toward the central core. This is consistent with anecdotal evidence and is related to parking areas being more costly to supply in central city areas than in fringe areas where room to expand exists and land values are lower. The spatial behavior of equilibrium parking land is described by the envelope dotted line of the two lines illustrating the spatial behavior of parking area under each parking regime. 5.4.2.3 Urban form with and without MPRs Figures 5.5 and 5.6 illustrate the impacts of parking requirements on parking area and dwelling size. Note that the market equilibrium is not efficient: dwelling sizes are too big and parking area per dwelling is too low when compared to the optimal solution (when the externality is internalized in developers’ decisions). This occurs because in the unregulated scenario (without MPRs) developers do not account for the collective positive effect of on-site parking choices per dwelling on the parking-related congestion in the neighborhood. While the optimal solution requires raising the parking area per dwelling above the private equilibrium value at each city location, cities often impose a uniform minimum-parking

Urban form and the pricing of transport and parking  93 9

Underground Surface Equilibrium

8

Parking Area

7 6 5 4 3 2 1 0

0

5

10

15

20

25

30

35

40

45

50

55

60

Distance to the CBD

Figure 5.4  Equilibrium parking area per dwelling (CBD located at point zero) 9.00 8.00 7.00

Parking Area

6.00 5.00 4.00 3.00 2.00 Equilibrium Parking Area Opmal Parking Area

1.00 0.00

0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60

Distance to the CBD

Figure 5.5  Market (equilibrium) and optimal parking areas in the presence and absence of MPRs requirement everywhere. Under such a blunt policy, when the minimum amount of parking area that needs to be supplied per dwelling far overshoots the optimal value, the underground regime is much more constrained compared to the surface regime because surface parking area per dwelling tends to be higher already both in the equilibrium and optimal cases. As a result, land rent falls by more for the underground regime than for the surface regime in the boundary

94  Handbook on transport pricing and financing  210 205 200

Dwelling Size

195 190 185 180 175 170 165 160

Equilibrium Dwelling Size Opmal Dwelling Size Dwelling Size under MPR 0 3 6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60

Distance to the CBD

Figure 5.6  Market (equilibrium) and optimal dwelling sizes in the presence and absence of MPRs between the parking regimes, leading the surface parking area to expand (that is, more parcels will be developed for residential use with surface rather than underground parking). This over-allocation of land to surface parking land creates sprawling areas inhospitable to nonautomotive forms of transportation. It is also visible from Figure 5.5 that the amount of oversupply of parking area due to MPRs is bigger toward the central core than toward the suburbs. Finally, Figure 5.6 shows that a uniform MPR reduces dwelling sizes everywhere in the city, which in turn has implications for population density levels (urban spatial structure). 5.4.2.4 Any evidence that parking mandates are binding constraints? While MPRs are binding almost everywhere in the numerical example provided by Brueckner and Franco (2017), Cutter and Franco (2012) provide empirical results designed to test whether MPRs in LA County represent binding constraints on developers, using two approaches relying on a non-residential property database. The first approach denoted direct test, compares a building’s available parking area to the area mandated by the MPR. If the two areas tend to be close suggests that the MPR constraint is typically binding. Table 5.5 presents the results of such a direct test for 249 office properties built in Los Angeles County between 1973 and 2006. Note that the difference between the average parking supplied and the estimated parking required is very close to zero suggesting that MPRs are well-enforced and bind for the office properties in the sample. The second approach by Cutter and Franco (2012) estimates the value of additional parking using a hedonic price model, and then compares this value to the cost of providing additional parking. The results show that value is less than cost, suggesting that MPRs lead developers to provide more parking than they would voluntarily, making the constraints binding.

Urban form and the pricing of transport and parking  95

Table 5.5  Average parking supply, parking mandate, and the difference between the two for office properties in different cities in Los Angeles County Cities in Los Angeles County

Building Area (sqft)

Parking Supplied

Required Parking

Difference

Alhambra

1000

2.98

3.88

–0.90

Baldwin Park

1000

4.90

3.75

1.14

Burbank

1000

2.61

2.93

–0.32

Downey

1000

2.79

2.77

0.02

El Monte

1000

3.92

3.64

0.28

Glendale

1000

2.18

2.60

–0.42

Long Beach

1000

3.69

3.74

–0.04

Los Angeles

1000

2.10

1.92

0.18

Pomona

1000

3.63

3.98

–0.35

Inglewood

1000

8.63

3.92

4.71

Santa Monica

1000

1.81

3.26

–1.45

Torrance

1000

3.20

3.29

–0.09

West Covina

1000

3.13

3.29

–0.17

Whittier

1000

4.21

3.26

0.95

Source:   Cutter and Franco (2012).

Stangl (2019) also conducted a similar direct test as in Cutter and Franco (2012) while focusing on a sample of 300 approved residential and mixed-use developments located in the City of LA between 2013 and 2018. Her results show that parking minimums did have a binding effect on the amount of residential parking built, although developers in certain areas built more parking than required. In particular, 58% of developments in her sample built at or just above their default binding minimum, and 42% of developments were built at least 10% above the binding minimum. The study also shows that as developments received larger parking minimum reductions, they built less parking relative to what they would otherwise have been required to build. 5.4.2.5 Cost of mandated parking Table 5.6 presents back-of-the envelop calculations illustrating how the cost of satisfying the default city parking requirement in the city of Los Angeles increases the total cost of constructing 500 square feet of an office building with an underground garage in different areas of the city.32 From Table 5.6, complying with the default parking minimum increases the cost of an office building on average by 48%. It should be noted, however, that a city average can mask spatial variation in the cost of mandated parking as parking construction costs tend to vary within the city. Similar calculations can be done for the case of an apartment building. The city of Los Angeles requires two parking spaces for a two-bedroom apartment. Franco (2016) estimates that the average cost of building an underground parking space is around $50,779 (in 2013 US dollars) in Downtown and $27,776 (in 2013 US dollars) in El Segundo. This suggests that the parking requirement for a multistory apartment building would require a developer to spend $101,558 and $55,552 per dwelling in Downtown and in El Segundo, respectively, to comply with the city´s default residential parking mandate.

96  Handbook on transport pricing and financing 

Table 5.6  Cost of MPRs for office buildings in the City of LA Underground Parking Structure Area Name

Default Mandated Parking

Building Area (sqft)

Parking Area (sqft)

Construction Cost ($/sqft) Parking

Construction Cost ($/sqft) Building

Cost Increase

Downtown Los Angeles CBD Area

1

500

350

145

158

64%

Westside CBD Area

1

500

350

166

158

74%

Beverly Hills* CBD Area

1

500

350

83

158

37%

El Segundo* CBD Area

1

500

350

79

158

35%

Santa Monica* CBD Area

1

500

350

92

158

41%

Marina del Rey* Urban Area

1

500

350

55

158

25%

Westwood CBD Area

1

500

350

90

158

40%

Reseda-van Nuys Urban Area

1

500

350

149

158

66%

East van Nuys Urban Area

1

500

350

81

158

36%

Pasadena* Urban Area

1

500

350

145

158

64%

Sources:   Building construction costs for a Grade A office building measured in 2012 US dollars (Shoup, 2018). Parking construction costs are in 2013 US dollars (Franco, 2016).

It is then not surprising that such high costs of parking likely incentivize developers to limit the density of development as well as the total amount. They also create an incentive for developers to build in low-density suburban areas of the city where land is less expensive and parking requirements can be met through surface parking lots.33 Therefore, policies that require a minimum amount of parking when parking is already difficult to provide can stifle an entire neighborhood´s growth. 5.4.3 Employer-Paid Parking, Parking Charges at the Workplace, and Cash-Out Policies There is also evidence that inexpensive or free parking is an incentive for solo driving (Willson and Shoup, 1990; Willson, 1992; Hess, 2001; Shoup, 2005; Franco and Khardagui, 2021). When commuters are shielded from the cost of parking, they drive alone more often. This implies that parking subsidies such as employer-paid parking incentivize workers to drive to work alone over walking, biking or taking public transit.34

Urban form and the pricing of transport and parking  97

In California, for example, approximately 95% of auto commuters receive free parking (Shoup, 2018). Employer-paid parking is common even in CBDs, where the cost to employers of offering free parking is much higher. One survey of the Los Angeles CBD found that 53% of auto commuters received employer-paid parking (Shoup, 2005). By directly or indirectly subsidizing parking at work, employers reduce the cost of the commute trip by car while requiring the employee to pay only the driving cost. This encourages employees to drive more often to work. Moreover, when such a subsidy is common practice everywhere, the cost of travel within cities is low, potentially encouraging their spatial expansion. Brueckner and Franco (2018) provide insights into the effects on mode choice and level of suburbanization (measured by the level of population in the suburbs) of a switch from employee-paid parking to employer-paid parking. Under such a switch, the burden of parking costs moves from auto commuters to employers, thus reducing the wage for all workers. Figures 5.7 and 5.8 are based on the authors’ numerical simulations. The figures illustrate how mode choices and suburbanization vary as a function of the employee’s parking cost share. The “foot” choice in Figure 5.7 indicates a central residence, which entails the mode choice of walking to work. A switch from employee-paid parking to employer-paid parking increases road usage and road capacity, while reducing the city core’s residential land area. With less central residential land, the result is greater suburbanization of the population along with an overall increase in the suburban commute flow to the CBD. Since all these changes are welfare-reducing, employer-paid parking leads to inefficiently high road usage and capacity investment along with an excessive degree of suburbanization.

% COMMUTING BY A PARTICULAR MODE

0.9000 0.8000 0.7000 0.6000 0.5000 0.4000 0.3000 0.2000 0.1000 0.0000

Transit 0

0.1

0.2

0.3 0.4 0.5 0.6 0.7 % WORKER PAID PARKING COST

Auto 0.8

0.9

Figure 5.7  Travel mode choice with different workplace parking subsidies

Foot 1

98  Handbook on transport pricing and financing  0.9000 0.8000

% POPULATION

0.7000 0.6000 0.5000 0.4000 0.3000 0.2000 0.1000 0.0000

Suburbs 0

0.1

0.2

0.3 0.4 0.5 0.6 0.7 % WORKER PAID PARKING COST

0.8

Central 0.9

1

Figure 5.8  Impact of workplace parking subsidy on suburbanization One possible solution to improve overall economic efficiency is to have parking charges at the workplace. In such a scenario, workers would have to pay for their parking at market rates, which are usually high in large cities and in CBDs, with employers raising their wages accordingly. Yet, eliminating a transportation subsidy such as employer-paid parking that is so popular and that benefits so many workers is likely to raise many political objections. As such, parking cash-out programs have been advocated instead. Under a cash-out program employers who lease (or partially subsidize) parking on behalf of their employees must offer their employees the choice to keep their allotted parking spot or trade it for an equivalent cash payment. This cash payment must equal the parking subsidy, that is, the cost to the employer of renting or leasing a parking space on behalf of the employee. Giving all employees (both drivers and non-drivers) such a choice reveals that free parking is not free after all. Commuters who forgo the cash are therefore spending it on parking and paying a “price” to park at work. Employees who accept the cash pay income taxes on it, but can use the money as they choose. The employer also pays payroll taxes associated with the cash that is provided to employees in lieu of parking. Brueckner and Franco’s (2018) simulations show that in the absence of income taxation this policy restores efficiency to the first-best solution.35 In addition, a parking cash-out policy may even outperform a congestion tax in terms of the effects on welfare and modal shift (De Borger and Wuyts, 2009). The level of effectiveness of cash-out programs is, however, dependent on urban density, alternative travel mode accessibility and size of the cash subsidy. Table 5.7 shows estimates of the daily market parking price workers would have to pay in different locations of LA County if workplace parking subsidies were removed. The table further provides estimates of what would be the cash-equivalent per month that employers would need to offer their employees if the cash offered equals the parking subsidy, and the parking subsidy is valued at the Fair Market Value (FMV).36 All values presented in Table 5.7 are in 2013 US dollars (Franco, 2016).

Urban form and the pricing of transport and parking  99

Table 5.7  The value of parking subsidies Los Angeles County Areas

Exclusion Amount ($/Month)

Market Parking Price ($/Day)

Value of Parking Subsidy ($/Month)

Downtown Los Angeles (223)

245

10

205

Beverly Hills (63)

245

12

247

Santa Monica (53)

245

13

263

Westwood (33)

245

14

284

Pasadena (56)

245

7

143

Burbank (20)

245

3

67

Long Beach (47)

245

10

205

Sources:   The monthly value of the parking subsidy is calculated as an FMV and assumes that employees work five days a week. In parenthesis is the number of off-street commercial parking garages used to calculate the average daily market price in each area of the County.

5.5 CONCLUSIONS This chapter provides insights into how congestion pricing, parking subsidies and parking mandates affect urban form, urban spatial structure, and travel mode choice in urban spatial settings. The chapter reviews existing theoretical studies on these topics, and observes that although many important contributions have been made, there is still scope and need for research. In addition, the chapter offers a simple general equilibrium framework that can be used as a first step to incorporate other features of urban areas that have not been explored yet in the existing literature on transport pricing, urban form and land-use arrangements. For instance, the analyses presented here suggest that urban vacant land and its temporary uses in downtown areas can have interactions with urban externalities, and in turn affect landuse configuration, urban form and the effects of transport pricing. Moreover, a model that enables a dynamic city evolution is important to explore. The one-shot, static equilibrium typical of existing models is never achieved in the real world. Urban population and economic activities expand over time, in the midst of uncertainty and imperfect information along with speculation. Some avenues in which the framework presented here could be extended include the use of production functions that allow for substitutability between labor and land, endogenous land consumption, endogenous parking supply at home and workplaces, pre-existing distortions such as labor taxes, sales taxes, or workplace parking subsidies, temporary uses for vacant lots that can have an impact on sustainable development (such as open space) and the existence of agglomeration externalities. Second-best congestion pricing by design like a VMT tax could also be explored. Future research on transport pricing and urban form should also include the explicit interplay of urban form and urban externalities, and focus on the role of parking pricing strategies as second-best congestion pricing mechanism. Road pricing has proven to be effective at reducing traffic congestion, but faces strong public opposition. For this reason, it is key to understand the effectiveness and potential role of alternative pricing mechanisms, such as parking fees or taxation, as a possible solution to pricing congestion and related externalities.

100  Handbook on transport pricing and financing 

Theoretical models of congestion and parking suggest already that parking fees can provide a second-best pricing mechanism that can be used to reduce congestion (Verhoef et al., 1995). Yet, there is hardly any work exploring the interactions of parking pricing in residential and/or workplace sites and/or parking mandates and urban form in general equilibrium settings with multiple urban externalities or pre-existing distortions. In addition, the increased vehicle travels are responsible for a big chunk of the increased greenhouse emission in urban areas. Households residing in low-density neighborhoods (like those in the suburbs) emit disproportionally air pollution because they drive more but also choose less fuel-efficient cars (Kim and Brownstone, 2013). Thus, urban forms such as urban sprawl will continue to be inefficient (despite all congestion road policies already implemented) unless vehicle emissions are also corrected. Vehicle fuel efficiency can then be an important tool to reduce this type of car use externality. In fact, it is not uncommon to see nowadays energy policies for cars implemented like fuel taxes, vehicle fuel efficiency standards and subsidies and penalties for the purchase of high- and low-efficiency vehicles, respectively. Finally, automated vehicles (AV) are a rapidly developing technology. All over the world technical, political, legal, and economic issues of AV are already being discussed and analyzed. Yet studies on the impacts of AV on urban form and land-use patterns are very scarce. Despite the benefits AV may have on mobility and parking land needs in urban areas (as AV can be parked in remote locations), the widespread use of self-driving cars by a large portion of urban residents is likely to reinforce the current trend toward sprawling cities and excessive vehicle miles traveled. On the other hand, the need for less parking may help increase central core densities because the existing surface and structured parking facilities may be converted into offices, houses and retail or recreation spaces. Moreover, if automated cars are privately owned, an increase in road traffic may also block the time savings typically described in the AV debates. Thus, there is also a need to fully understand how this emerging technology may affect urban space.

NOTES 1. The author thanks Erik T. Verhoef, Daniel Hörcher, and an anonymous referee for their insightful comments. 2. Low-income travelers typically choose a lower speed travel mode when more than one travel mode is available to economize on monetary costs (Jara-Díaz and Videla, 1989), which may explain the city location of different income groups (LeRoy and Sonstelie, 1983; Sasaki, 1990; Gin and Sonstelie, 1992; DeSalvo and Huq, 1996; Glaeser et al., 2008). 3. Some studies have explored urban growth boundaries (UGBs) and floor-to-area regulations as possible second-best policies to congestion tolls when traffic congestion is unpriced in monocentric city models. Examples include Kanemoto (1977), Arnott (1979b), Pines and Sadka (1985), Brueckner (2007), Pines and Kono (2012) and Kono et al. (2012). An UGB specifies a city boundary beyond which development may not take place. It is reasoned that a restrictive UGB can be an alternative to a congestion tax in reducing traffic congestion by shrinking the city’s spatial size, but would do so imperfectly since a UGB does not promote central densification to the same extent as a congestion tax does. Anas and Rhee (2007) and Anas and Pines (2008) have also explored UGBs as a secondbest policy but in a city-suburb system and a two-city system, respectively. When jobs are pre-set to locate at the region’s core, they found that congestion tolls can cause jobs to relocate to the suburbs or another city, in order to shorten average commuting distances. They found that an effective UGB should be expansive rather than restrictive, forming a more sprawling structure through a process of jobs decentralization.

Urban form and the pricing of transport and parking  101

4. Another highlight of this study is that the combined use of public transit subsidies and an urban growth boundary becomes a second-best policy regime in the absence of a congestion tax. When public transport is available in a city, an UGB may fail to be second-best efficient or effective on its own under certain parameters related to consumer preferences and transportation technology because it could fail to mitigate congestion. 5. Li et al. (2020) provide a review of the bottleneck model research since its inception. 6. Congestion pricing schemes that are second-best “by design” (and not because distortions exist in the spatial economy), like cordon charging, have also been studied within the traditional monocentric framework (Mun et al., 2003; Verhoef, 2005; De Lara et al., 2013; Brueckner, 2014; Li and Guo, 2017). In contrast to congestion tolls, cordon pricing requires that a commuter pays a single toll upon entering a zone surrounding the CBD. Like first-best congestion tolls, optimal cordon tolling spatially shrinks the city by promoting short-distance commuting, despite the city be more spread out than under the first-best regime. 7. Tikoudis et al. (2018) show that pre-existing property taxes (a tax-induced distortion in the housing market) invoke congestion toll deviations from the Pigouvian principle, while with pre-exiting building height limits (a command-and-control quantitative restriction) such deviation is not efficiency enhancing. 8. The same qualitatively downward adjustment is expected if agglomeration economies are reduced or eliminated when workers and/or firms adjust their job location to less dense areas following a congestion toll. Verhoef and Nijkamp (2004) also show how second-best taxes or subsidies are lower than the Pigouvian level when firm agglomeration externalities and pollution from commuting exist in the monocentric city model. 9. Parry and Bento (2001) have endogenized multiple urban externalities in an aspatial setting. While their insights on the interaction of congestion taxes with pre-exiting labor distortions may carry over to more sophisticated general equilibrium models, their study fails to examine the interaction between externalities and urban form. 10. Wheaton (2004) uses a general equilibrium model with congestion and center-agglomeration externalities that allows for mixed rather than exclusive land uses at each location. Though his analysis also did not consider the efficiency of a congestion toll, his results show that worse congestion may cause greater job decentralization, and thus interactions between congestion and agglomeration externalities should not be overlooked on land use-congestion studies. 11. Residents in this model derive idiosyncratic utility from choosing their preferred home and work locations and might be less sensitive to increased transportation costs. On the other hand, industry can respond to the toll by moving out of the core, leading to a more decentralized land-use pattern. 12. Employer paid-parking is a tax-exempt fringe benefit provided by employers to workers driving to work and entails either free parking or parking at very low rates in workplaces. Minimum parking requirements mandate new buildings to include a number of off-street parking spaces based on its use or in proportion to building size. 13. The framework presented here builds on Franco and Waxman (2021) who extend Mills (1981) setup to study the interplay of regulatory delay, surface parking lots in CBD areas and road congestion; parts of the presentation of the model in this section draws from theirs. 14. For urban spatial models with endogenous parking prices and supply see Brueckner and Franco (2017, 2018). Here we focus on parking land as a temporary urban land use for CBD vacant lots awaiting development. 15. Future work may explore the case where visitors participate in the city labor market and/or contribute to the city agglomeration economies in terms of consumption or production. 16. Future work may consider resident landowners and allow them to share equally in total land rent. Here absentee landowners implies that land rents are not a component of households’ income. 17. In the context of a closed city model, discontinuous development may emerge when the growth mechanism is an exogenous population increase. 18. Anecdotal evidence of vacant land in downtown areas shows that these parcels, while scattered throughout the area, are mostly concentrated in groups composed of numerous parcels, located either adjacent to each other within the same block, or within an immediate area of several blocks. Often times such vacant land is used for parking, especially near commercial areas. This may occur because the land is valuable for a future use and is not needed in the present, so it is put to a

102  Handbook on transport pricing and financing 

temporary use, with suitable restrictions to prevent it becoming non-adaptable to its future use. For instance, in 2018 the city of Los Angeles had 3,772 parking lots, spanning over 1.8 square miles, according to data from the LA County Assessor’s Office. 19. Early models that have explored leapfrog development as the result of optimal intertemporal decision-making by developers who choose the timing, type, and location of development include Mills (1981), Wheaton (1982), Ohls and Pines (1975) and Brueckner and von Rabenau (1981). Given sufficient growth, developers find it optimal to reserve land located closer to the downtown area for future higher-valued (e.g., as higher-density residential (Ohls and Pines, 1975; Wheaton, 1982) or industrial (Mills, 1981)) development and to first pursue lower-valued development (e.g., lowerdensity residential) in locations that are farther away. The emergence of leapfrog development in these models depends on forward-looking developers that anticipate future prices and optimize over time; a choice between at least two development options that differ in their net returns; and sufficient growth in land values over time. 20. We remind the reader that all non-residential, non-visitor parking is located at the trade node for simplicity. Parking at the trade node is also free only to city residents. 21. There is no unemployment in this model and thus, total population equals total workers. Since each resident consumes 1 unit of residential land and the tradable good production function uses a Leontief technology that requires μ units of labor to produce one unit of Q, total residential land demanded by residents in a given period equals μQt. 22. Since xd ³ xc and RA ( xc ) < RA ( xd ) , from conditions 7 in Table 5.1 R2HC > R1HC , In the simulation RA ( xc ) = RA ( xd ) , suggesting there is a distance from the trade node after which the soil quality is equally good for agriculture. 23. Note that the two land use sequences on the two sides of xa are (production in period 1, production in period 2) and (parking in period 1, production in period 2). These two sequences must provide the tx 1 é Q txa ù 1 é Q txa ù and simplifyR2 R2 =q+ same present value at xa. Setting R1Q - a + l úû l 1 + r êë l úû 1 + r êë ing yields condition 9 in Table 5.1. 24. Using the numerical results provided in Table 5.3 we observe that t x 1 1 é ù é R2HP - t p xa ù < q + R1HP - t p xa + R2Q - a ú and that condition 10 in Table 5.1 is just met û l û 1+ r ë 1 + r êë when evaluated at xb, where xb > xa . 25. Note that on the left-hand-side of n Figure 5.2, we have land in commercial use in period 1, and on the right-hand-side of we have land developed in residential use in period 1. Vacant land between xa and xb is developed to commercial use in the second period. 26. Since land consumption is fixed for simplicity, it is not possible to draw conclusions on the effects of a congestion tax on population and firm densities. Future work can relax this assumption and explore such effects. 27. Future work may explore the case when visitors add to agglomeration economies or are part of the city labor force. 28. This section of the chapter builds heavily on some parts of the author´s ITF discussion paper “Parking Prices and Availability, Mode Choice and Urban Form,” International Transport Forum Discussion Papers, February, No. 2020/03, OECD Publishing, Paris., available at https://www​.itfoecd​.org​/sites​/default​/files​/docs​/parking​-mode​-choice​-urban​-form​.pdf. 29. Reasons put forward to justify such mandates include preventing parking spillovers from new development into surrounding area, reducing illegal parking, and prevent congestion from cruising for vacant on-street parking spaces. 30. Surface parking area is provided via a parking lot, which requires minimal capital, assumed to be zero for simplicity. Structural parking is an adjacent above-ground structure to the residential structure that provides parking area. While capital cost is much higher than under surface parking, structural parking saves on land through use of a multistory structure. Underground parking, by contrast, requires no additional land beyond that used for the residential building. Parking area is provided within an underground structure below the building, which involves higher capital cost than structural parking.

Urban form and the pricing of transport and parking  103

31. The parking inputs are land in the case of surface parking, land and capital in the case of structural parking, and capital alone in the case of underground parking. Under the latter two regimes, capital per unit of parking land is a choice variable of the developer, indicating the height of the aboveground parking structure or the depth of the underground structure. 32. The reader should note that the names on Table 5.4 refer to areas in the City of Los Angeles. For example, the areas designated as Pasadena, Beverley Hills, El Segundo, and Santa Monica in Table 5.4 do not refer to the cities of Pasadena, Beverley Hills, El Segundo, and Santa Monica. They represent the areas in the City of Los Angeles around those cities which are also part of the same name zones that compose Los Angeles County. 33. Oftentimes parcels in downtown areas are very small and irregular and buildings frequently cover the entire parcel. In addition, downtown areas have a lot of architecturally and historically significant buildings that predate widespread car ownership, and thus do not have parking or even the space to add it. In these situations, providing on-site underground parking can be very expensive or even physically impossible. 34. Employers have incentives to provide such type of parking subsidy because it is a qualified fringe benefit that is not taxed. This allows firms to pay lower wages and save on payroll taxes without making employees worse off. Moreover, it is a way to retain employees. The excess supply of parking space due to MPRs also reduces workers´ willingness to pay for it. 35. With income taxation, the wage supplement is taxable, but the tax disappears if the supplement is exchanged for parking. As a result, the additional tax can be viewed as a cost of not using the auto mode. Since this cost will distort mode choice, leading to road usage beyond the first-best level, the cash-out does not have the same efficiency benefit as in a world without income taxes. To restore efficiency, the fringe benefit of employer-paid parking would need to be taxable, in which case there would be no adverse tax effect from not choosing the auto mode. 36. The FMV of parking provided by an employer is based on the cost an individual would have to pay for parking at the same time and site in an arm’s length transaction or, if the employer cannot ascertain this information, in the same or a comparable lot in the general location under the same or similar conditions.

REFERENCES Alonso, W. (1964). Location and Land Use. Harvard University Press, Cambridge, MA. Anas, A. and Kim, I. (1996). General equilibrium models of polycentric urban land-use with endogenous congestion and job agglomeration. Journal of Urban Economics 40: 232–256. Anas, A. and Pines, D. (2008). Anti-sprawl policies in a system of congested cities. Regional Science and Urban Economics 38: 408–423. Anas, A. and Rhee, H. (2007). When are urban growth boundaries not second-best policies to congestion tolls? Journal of Urban Economics 61: 263–286. Anas, A. and Rhee, H.-J. (2006). Curbing excess sprawl with congestion tolls and urban boundaries. Regional Science and Urban Economics 36: 510–541. Anas, A. and Xu, R. (1999). Congestion, land-use, and job dispersion: A general equilibrium model. Journal of Urban Economics 45: 451–473. Anderson, S. and de Palma, A. (2007). Parking in the city. Papers in Regional Science 86: 621–632. Anderson, S. and de Palma, A. (2004). The economics of pricing parking. Journal of Urban Economics 55: 1–20. Arnott, R. (2007). Congestion tolling with agglomeration externalities. Journal of Urban Economics 62: 187–203. Arnott, R. (1998). Congestion tolling and urban spatial structure. Journal of Regional Science 38: 495–504. Arnott, R. (1979a). Optimal city size in a spatial economy. Journal of Urban Economics, 6: 65–89. Arnott, R. (1979b). Unpriced transport congestion. Journal of Economic Theory 21: 294–316. Arnott, R.J., de Palma, A. and Lindsey, R. (1991). A temporal and spatial equilibrium analysis of commuter parking. Journal of Public Economics 45: 301–335.

104  Handbook on transport pricing and financing 

Arnott, R.J. and Inci, E. (2006). An integrated model of downtown parking and traffic congestion. Journal of Urban Economics 60: 418–442. Arnott, R.J., Inci, E. and Rowse, J. (2015). Downtown curbside parking capacity. Journal of Urban Economics 86: 83–97. Arnott, R.J. and Rowse, J. (1999). Modeling parking. Journal of Urban Economics 45: 97–124. Bento, A.M., Cropper, M.L, Mobarak, A.M. and Vinha, K. (2005). The effects of urban spatial structure on travel demand in the United States. Review of Economics and Statistics 87: 466–478. Brueckner, J.K. (2014). Cordon tolling in a city with congested bridges. Economics of Transportation, 3 (4): 235–242. Brueckner, J.K. (2007). Urban growth boundaries: An effective second-best remedy for unpriced traffic congestion? Journal of Housing Economics 16 (3): 263–273. Brueckner, J.K. (2001). Urban sprawl: Lessons from urban economics. Brookings-Wharton papers on urban affairs, Brookings Institution, Washington, DC, 65–89. Brueckner, J.K. and Franco, S.F. (2018), Employer-paid parking, mode choice, and suburbanization. Journal of Urban Economics 104: 35–46. Brueckner, J.K. and Franco, S.F. (2017). Parking and urban form. Journal of Economic Geography 17: 95–127. Brueckner, J.K. and von Rabenau, B. (1981). Dynamics of land-use for a closed city. Regional Science and Urban Economics 11: 1–17. Buyukeren, A.C. and Hiramatsu, T. (2016). Anti-congestion policies in cities with public transportation. Journal of Economic Geography 16: 395–421. Calthrop, E. and Proost, S. (1998). Road transport externalities. Environmental and Resource Economics 11: 335–348. Calthrop, E, and Proost, S. (2006). Regulating on-street parking. Regional Science and Urban Economics 36: 29–48. Cervero, R. (1996). Mixed land uses and commuting: Evidence from the American Housing Survey. Transportation Research Part A: Policy and Practice 30: 361–377. Cervero, R. (1989). America’s Suburban Centers: The Land Use-Transportation Link. London: Allen and Unwin. Cervero, R. and Wu, K. (1997). Influence of Land Use Environments on Commuting Choices: An Analysis of Large U.S. Metropolitan Areas using the 1985 American Housing Survey. Working Paper 683. University of California at Berkeley. Cutter, W.B. and Franco, S.F. (2012). Do parking requirements significantly increase the area dedicated to parking? A test of the effect of parking requirements values in Los Angeles County. Transportation Research Part A: Policy and Practice 46: 901–925. De Borger, B. and Wuyts, B. (2009). Commuting, transport tax reform and the labor market: Employerpaid parking and the relative efficiency of revenue recycling instruments. Urban Studies 46: 213–233. De Lara, M., de Palma, A., Kilani, M. and Piperno, S. (2013). Congestion pricing and long-term urban form: Application to Paris region. Regional Science and Urban Economics 43: 282–295. DeSalvo, J.S. and Huq, M. (1996). Income, residential location, and mode choice. Journal of Urban Economics 40: 84–99. Ersoy, F.Y., Hasker, K. and Inci, E. (2016). Parking as a loss leader at shopping malls. Transportation Research Part B: Methodological 91: 98–112. European Commission (EC) (2019). 2019 Handbook on the external costs of transport. Available from: https://ec​.europa​.eu​/transport​/sites​/transport​/files​/studies​/internalisation​-handbook​-isbn​-978​-92​-79​96917​-1​.pdf Ewing, R. and Cervero, R. (2010). Travel and the built environment: A meta-analysis. Journal of the American Planning Association 76 (3): 1–30. Ewing, R. and Cervero, R. (2001). Travel and the built environment: A synthesis. Transportation Research Record 1780: 87–114. Franco, S.F. (2017). Downtown parking supply, work-trip mode choice and urban spatial structure. Transportation Research Part B: Methodological 101: 107–122. Franco, S.F. (2016). Parking Costs in Los Angeles County. Technical Report 2016-2-June, MRPI program of the University of California, Application of the RELU-TRAN model to the Greater Los Angeles Region, January, Retrieved from https://vcpa​.ucr​.edu/.

Urban form and the pricing of transport and parking  105

Franco, S.F. and Khordagui, N. (2021). Commute mode choices and the role of parking prices, parking availability and urban form: Evidence from Los Angeles county. UCI Working Paper. Franco, S.F. and Waxman, A. (2021). They paved paradise? Vacant land and surface lots in downtown urban areas and the role of regulatory delay in optimal dynamic land use. UCI Working Paper. Fosgerau, M. and Kim, J. (2019). Commuting and land use in a city with bottlenecks: Theory and evidence. Regional Science and Urban Economics 77: 182–204. Fosgerau, M., Kim, J. and Ranjan, A. (2018). Vickrey meets Alonso: Commute scheduling and congestion in a monocentric city. Journal of Urban Economics 105: 40–53. Gin, A. and Sonstelie, J. (1992). The streetcar and residential location in nineteenth century Philadelphia. Journal of Urban Economics 32: 92–107. Glaeser, E.L., Kahn, M.E. and Rappaport, J. (2008). Why do the poor live in cities? The role of public transportation. Journal of Urban Economics 63: 1–24. Gubins, S. and Verhoef, E.T. (2014). Dynamic bottleneck congestion and residential land use in the monocentric city. Journal of Urban Economics 80: 51–61. Hess, D.B. (2001). Effect of free parking on commuter mode choice.  Transportation Research Record 1753: 01-448. Inci, E. (2015). A review of the economics of parking. Economics of Transportation 4: 50–63. Jara-Díaz, S. and Videla, J. (1989). Detection of income effect in mode choice: Theory and application. Transportation Research Part B: Methodological 23: 393–400. Kanemoto, Y. (1977). Cost-benefit analysis and the second-best land use for transportation. Journal of Urban Economics 4: 483–503. Kanemoto, Y. (1976). Optimum, market and second-best land use patterns in a von Thünen city with congestion. Regional Science and Urban Economics 6: 23–32. Kim, J. and Brownstone, D. (2013). The impact of residential density on vehicle usage and fuel consumption: Evidence from national samples. Energy Economics 40: 196-206 Kockelman, K.M. (1995). Which matters more in mode choice: density or income? Institute of Transportation Engineers, Compendium of Technical Papers, 844–867. Kono, T., Joshi, K.K., Kato, T. and Yokoi, T. (2012). Optimal regulation on building size and city boundary: An effective second-best remedy for traffic congestion externality. Regional Science and Urban Economics 42: 619–630. LeRoy, S.F. and Sonstelie, J. (1983). Paradise lost and regained: Transportation innovation, income, and residential location. Journal of Urban Economics 13: 67–89. Levinson, D. and Kumar, A. (1997). Density and the journey to work. Growth Change 28: 147–172. Li, Z. and Guo, Q. (2017). Optimal time for implementing cordon toll pricing scheme in a monocentric city. Papers in Regional Science 96: 163–190. Li, Z.C., Huang, H.J. and Yang, H. (2020). Fifty years of the bottleneck model: A bibliometric review and future research directions. Transportation Research Part B: Methodological, 139: 311–342. Maibach, M., Schreyer, D., Sutter, D., van Essen, H., Boon, B., Smokers, R., Schroten, A., Doll, C., Pawlowska, B. and Bak, M. (2008). Handbook on estimation of external costs in the transport sector. Internalisation Measures and Policies for All external Cost of Transport, Version 1.1., European Commission DG TREN, Delft, CE, The Netherlands. Mills, D.E. (1981). Growth, speculation and sprawl in a monocentric city. Journal of Urban Economics 10: 201–226. Mills, E.S. (1967). An aggregative model of resource allocation in a metropolitan area. American Economic Review 57: 197–210. Mills, E.S. and de Ferranti, D.M. (1971). Market choices and optimum city size. American Economic Review, Papers and Proceedings 61: 340–345. Muth, R.F. (1969). Cities and Housing. University of Chicago Press, Chicago, IL. Mun, S., Konoshi, K. and Yoshikawa, K. (2003). Optimal cordon pricing. Journal of Urban Economics 54: 34 21–38. Ohls, J.C. and Pines, D. (1975). Urban development and economic efficiency. Land Economics 51: 224–234. Parry, I. and Bento, A. (2001). Revenue recycling and the welfare effects of road pricing. Scandinavian Journal of Economics 103: 645–671. Pines, D. and Kono, T. (2012). FAR regulations and unpriced traffic congestion. Regional Science and Urban Economics 42: 931–937.

106  Handbook on transport pricing and financing 

Pines, D. and Sadka, E., (1985). Zoning, first-best, second-best, and third-best criteria for allocating land to roads. Journal of Urban Economics 17: 167–183. Sasaki, K. (1990). Income class, modal choice, and urban spatial structure. Journal of Urban Economics 27: 322–343. Shoup, D.C. (2018). Parking and the City. Planners Press, Chicago. Shoup, D.C. (2005). The High Cost of Free Parking. Planners Press, Chicago. Solow, R.M.  (1973). Congestion cost and the use of land for streets.  Bell Journal of Economics and Management Science 4: 602–618. Solow, R.M. (1972). Congestion, density, and the use of land in transportation. The Swedish Journal of Economics 74: 161–173. Stangl, K. (2019). Parking? Lots! Parking over the minimum in Los Angeles. Technical Report. Institute of Transportation Studies, UCLA. Takayama, Y. and Kuwahara, M. (2017). Bottleneck congestion and residential location of heterogeneous commuters. Journal of Urban Economics 100: 65–79. Tikoudis, I., Verhoef, E.T. and van Ommeren, J.N. (2018). Second-best urban tolls in a monocentric city with housing market regulations. Transportation Research Part B: Methodological 117: 342–359. Tikoudis, I., Verhoef, E.T. and van Ommeren, J.N. (2015). On revenue recycling and the welfare effects of second-best congestion pricing in a monocentric city. Journal of Urban Economics 89: 32–47. van Essen, H., B. Boon, A. Schroten, M. Otten, M. Maibach, C. Schreyer, C. Doll, P. Jochem, M. Bak, and B. Pawlowska. (2008). Internalization measures and policy for the external cost of transport. Technical report. Verhoef, E.T. (2005). Second-best congestion pricing schemes in the monocentric city. Journal of Urban Economics 58: 367–388. Verhoef, E.T. (1994). External effects and social costs of transport. Transportation Research Part A: Policy and Practice 28: 273–287. Verhoef, E.T. and Nijkamp, P. (2008). Urban environmental externalities, agglomeration forces, and the technological ‘Deus ex Machina’. Environment and Planning A: Economy and Space 40: 928–947. Verhoef, E.T. and Nijkamp, P. (2004). Spatial externalities and the urban economy. In Urban Dynamics and Growth: Advances in Urban Economics, eds. Capello, R. and Nijkamp, P. Elsevier, Amsterdam, pp. 87–120. Verhoef, E.T., Nijkamp, P. and Rietveld, P. (1995). The economics of regulatory parking policies: The im(possibilities) of parking policies in traffic congestion. Transportation Research A 29: 141–156. Voith, R. (1998). Parking, transit, and employment in a central business district. Journal of Urban Economics 44: 43–58. Wheaton, W.C. (1982). Urban residential growth under perfect foresight. Journal of Urban Economics 12: 1–21. Wheaton, W.C. (1998). Land use and density in cities with congestion. Journal of Urban Economics 43: 258–272. Wheaton, W.C. (2004). Commuting, congestion, and employment dispersal in cities with mixed land use. Journal of Urban Economics 55: 417–438. Willson, R. (1992). Estimating the travel and parking demand effects of employer-paid parking. Regional Science and Urban Economics 22: 133–145. Willson, R. and Shoup, D. (1990). The effects of employer-paid parking in downtown Los Angeles: A study of office workers and their employers. Working paper, Graduate School of Architecture and Urban Planning, UCLA. Zhang, H. and Arnott, R. (2011). Computing value per square foot of vacant parcels and aggregating to model zone level [Version 1]. MRPI Technical Report. Zhang, M. (2004). The role of land use in travel mode choice: Evidence from Boston and Hong Kong. Journal of the American Planning Association 70: 344–60. Zhang, W. and Kockelman, K.M. (2016a)  Congestion pricing effects on firm and household location choices in monocentric and polycentric cities. Regional Science and Urban Economics 58: 1–12. Zhang, W. and Kockelman, K.M. (2016b). Optimal policies in cities with congestion and agglomeration externalities: Congestion tolls, labor subsidies, and place-based strategies. Journal of Urban Economics 95: 64–86. Zhang, W. and Kockelman, K.M. (2014). Urban sprawl, job decentralization, and congestion: The welfare effects of congestion tolls and urban growth boundaries. Proceedings in the 93rd Annual Meeting of the Transportation Research Board, Washington, DC.

6. Equity and distributional issues in transport pricing Christophe Heyndrickx and Inge Mayeres

6.1 INTRODUCTION Different chapters in this book discuss transport pricing policies to tackle transport externalities, with a focus on road pricing. From economic theory, it is clear that road pricing can contribute to a larger efficiency of the transport system, a better environmental performance, and traffic safety. However, the number of implemented schemes is still limited. A number of cities apply road pricing, using cordon tolls or area licensing schemes, for example, Singapore, London, Milan, Stockholm, Gothenburg, or Valetta. On some corridors, high occupancy toll lanes are implemented, mainly in the United States but also in some other countries (Rotaris et al., 2010; Anas and Lindsey, 2011; Attard and Enoch, 2011; Croci, 2016; Walker, 2018). In other cities and regions, road pricing schemes have been explored, but are not (yet) implemented. The as-yet limited roll-out of road pricing schemes is due to many factors (see, for example, Gaunt et al., 2007; De Borger and Proost, 2012; Eliasson, 2014; Schade, 2017). The distributional consequences of these policies is one of them. The way in which different groups of people are affected, or perceive themselves to be affected, is an important determinant of the political acceptability of transport pricing reforms. In the public debate, the distributional impacts of the change in the monetary costs are often the most prominent. However, the welfare impacts also depend on how people are affected by the change in transport externalities (congestion, environmental impacts, accidents), and on the way in which revenues raised by the pricing reforms are used. Equity can refer to both intra-generational and intergenerational equity. Intergenerational equity is concerned with the distribution of impacts across generations. The way in which intergenerational trade-offs should be evaluated is a complex issue (see, e.g. Portney and Weyant, 1999) with several considerations coming into play, including: economic elements (the economic prosperity of current versus future generations), the degree of inequality aversion, and the possibility of transfers between generations. The focus of this chapter is on intra-generational equity, i.e. the distribution of policy impacts of transport pricing between different groups within a given generation. These groups may differ from each other in terms of income, but also other characteristics such as household composition, location, whether a person is mobility impaired or not, etc. While some of these other perspectives are also discussed, this chapter focuses on equity in terms of income distribution (vertical equity) and on the distribution of costs and benefits across locations (spatial equity). The structure of this chapter is as follows. The first section summarizes the lessons from economic theory on how distributional considerations should be taken into account when determining transport pricing. The second section summarizes the findings of applied studies on the equity effects of transport pricing policies, either for actual road pricing cases, or from modelling studies. It also reflects on the implications of a growing share of electric cars. The 107

108  Handbook on transport pricing and financing

third section discusses the way forward towards more objectivity in the distributional assessment of transport policy.

6.2 ECONOMIC THEORY Economic theory provides rules for optimal taxation and the implications of equity considerations for these rules. These will be discussed first. Next, we consider policy reforms starting from an arbitrary initial equilibrium and summarize the factors that determine whether a tax reform is socially beneficial and how this is affected by equity considerations. 6.2.1 Optimal Tax and Investment Rules What are the rules for the optimal tax and investment policies in the presence of externalities, in a setting with nonidentical individuals? Should equity considerations be taken into account when determining transport prices in order to reduce transport externalities, and if so how should this be done? Here one can build upon an extensive literature in environmental economics about corrective taxation in the broader framework of the general tax system. Such a broader view is important because the government also has other objectives besides controlling the externalities: raising revenue in order to provide public services and goods, as well as attaining its distributional objectives. The literature investigates rules for setting the policy instruments such that social welfare is maximized. Social welfare is a function of individual utilities (Atkinson and Stiglitz, 1980). The social welfare weights indicate how social welfare changes with a change in individual utility and depend on the degree of inequality aversion in society. With income inequality aversion people with a lower income get a higher social welfare weight. We do not make statements about what the level of inequality aversion should be, as this is the outcome of political processes. Rather, the implications of different levels of inequality aversion are analyzed. We consider the case of when the government cannot make use of individualized lump sum taxes. Therefore, we are in a second-best setting. To focus the discussion we assume that only passenger and freight transport cause externalities. Transport users are assumed to regard themselves as infinitely small compared to society and hence to ignore their own impact on the level of the externalities. The main transport externalities are accidents, congestion, and environmental externalities related to climate change, air pollution and noise (European Commission, 2019). Note that for transport many of the externalities do not only depend on demand but also have an impact on demand: for example, the level of congestion and the level of accident risks affect transport choices. Some transport externalities are positive. For public transport there is the so-called Mohring effect: when more people use public transport, this leads to an increase in the frequency of the public transport services and consequently shorter waiting times (Mohring, 1972). Another example are the benefits related to better health by cycling and walking, in as far as they are not internalized (Börjesson and Eliasson, 2012). The lessons from the optimal tax literature depend on the type of tax instruments that are assumed to be available. We first consider the optimal tax rules for transport externalities if the government can make use of indirect taxes on the consumption of all commodities (including passenger transport) and a uniform transfer to households. In the absence of other income than labour income, this corresponds with a linear income tax combined with a full

Equity and distributional issues in transport pricing  109

set of indirect taxes. This setting is considered by, for example, Jacobs and de Mooij (2015), Mayeres and Proost (1997), Pirttilä and Schöb (1999), and Sandmo (2000). Moreover, use can be made of taxes on intermediate inputs (including freight transport) in production and investments in the capacity of the transport infrastructure. The next paragraphs summarize the optimal tax and investment rules, based on Mayeres and Proost (1997), who use a second-best tax model building on the optimal tax literature (e.g. Diamond and Mirrlees, 1971; Sandmo, 1975; Atkinson and Stiglitz, 1976; Bovenberg and van der Ploeg, 1994). First, on the consumption side, it can be shown that, in the absence of restrictions on the transport taxes that the government can use, only tax on goods that cause externalities contains a component that corrects for the externalities. This is added to a revenue-raising or Ramsey component, which is also present for the other commodities that do not generate externalities. This additivity property, which was derived by Sandmo (1975) in the case of identical consumers, also holds with nonidentical consumers. In both components, a trade-off is made between efficiency and equity. Ceteris paribus, if transport goods are consumed relatively more by people with a high social welfare weight, the revenue-raising component will be lower. The externality component includes a term for the impact of the transport externalities on production and on government revenue (via the impact of the externalities on transport demand). It also consists of a weighted sum of the impact on the households of congestion, environmental effects, and safety effects caused by passenger transport. The weights assigned to the impacts on the different households depend on the degree of inequality aversion. It is taken into account that households can be impacted differently by the externalities. They can be exposed to different levels of externalities or can suffer from their consequences to varying degrees. For example, air pollution and noise are worse in some areas than in others. Some people may also be more vulnerable to the health impacts than others. In addition, people may have a different valuation of the externalities. For example, the value of travel time savings increases with income (see, e.g. Batley et al., 2019). If a relatively higher social welfare weight is assigned to groups of people that are exposed more to air pollution and noise or to people who are more vulnerable to pollution, then the externality component of the tax will be higher. The opposite is the case if a larger weight is given to people with a relatively low valuation of the externalities. Second, the optimal tax rules indicate that intermediate inputs in production that do not cause externalities should not be taxed (Diamond and Mirrlees, 1971). However, intermediate inputs that do cause externalities, such as freight transport, should be taxed. The externality tax is similar in structure to that of the externality component for the tax on passenger transport. It therefore also takes into account distributional considerations. Third, transport externalities such as congestion can also be addressed via the capacity of the transport infrastructure. The optimal tax rules that were discussed above hold for any given level of infrastructure provision. However, the level of the taxes will depend on the available capacity. Optimally, infrastructure should be provided up to the point where the costs of additional provision equal the benefits. In practice, this needs to be assessed by social cost-benefit analyses of infrastructure projects. The benefits should take into account the impacts on the different groups and their valuation of these impacts. With inequality aversion, the impacts per group should be weighted by the relative social welfare weights. For the case of homogeneous environmental damages across individuals Jacobs and de Mooij (2015), Pirttilä and Schöb (1999), and Sandmo (2000) also find that optimal corrective taxes should generally take into account income redistribution concerns. However, Pirttilä and

110  Handbook on transport pricing and financing

Schöb (1999) and Jacobs and de Mooij (2015) point out that when preferences are weakly separable and homothetic the corrective tax should be set according to the Pigouvian tax rule and should otherwise not take into account distributional considerations. Jacobs and van der Ploeg (2019) extend the analysis to heterogeneous environmental damages across individuals and show that in general corrective taxes should reflect distributional concerns. However, if Engel curves for the polluting goods are linear, the corrective taxes should follow the Pigouvian tax rule (with equity weighting of the impacts). A number of papers consider non-linear rather than linear income taxes (see, e.g. Cremer et al., 1998, 2003; Micheletto, 2008; and Jacobs and de Mooij, 2015). In that case, corrective taxation is not directly used for redistribution. Besides characterizing optimal policies, the process followed towards this optimum is also important. The introduction of optimal policies often implies substantial policy changes, and not only for transport. Such changes may involve a long political process, for which insights into the consequences of a more stepwise approach can be very useful. The implementation of the optimal tax analysis is also highly data-intensive, requiring information about individual behavioural reactions with the additional complication that this information is also required for situations that can differ a lot from the current situation. Hence the interest in evaluations of gradual policy reforms starting from the current policies. 6.2.2 The Evaluation of Marginal Tax Reforms The so-called marginal tax reform approach builds upon the Ahmad and Stern (1984) model, which studied the equity-efficiency trade-off in an economy without externalities. It was extended by Schöb (1996) to include environmental quality and by Mayeres and Proost (2001) for externalities that are non-separable from the consumption of private commodities, of which traffic congestion is an example. Tirachini and Proost (2021) use a similar model for evaluating transport tax and subsidy reforms in the context of developing countries, recognizing the presence of both formal and informal labour market sectors. The approach compares different tax instruments relating to the marginal cost in terms of social welfare of raising one additional unit of government revenue via the tax instruments. Social welfare, which is a function of individual utilities, can be increased when raising a tax instrument with a low marginal welfare cost and when revenue neutrality is achieved by reducing a tax instrument with a higher marginal welfare cost. A number of insights follow from this literature: ●

●

●

A full assessment of the efficiency and equity impacts of the reforms should take into account not only the effect of the transport policy reform, but also of the measures that ensure budget neutrality for the government. For example, if road pricing is introduced, the impact of using the revenues generated should be included in the evaluation. Alternatively, with transport subsidies it should be considered how these are financed. While the analysis before revenue recycling offers only an incomplete assessment of the welfare impacts, it nevertheless offers important insights to assess the political acceptability of the policy reform and can also be useful to determine whether accompanying measures are required for particular groups. If a transport good is consumed proportionally more by people with a high marginal social welfare weight, then – everything else being equal – it becomes less attractive to increase the tax on that good or more attractive to reduce the tax on that good or subsidize

Equity and distributional issues in transport pricing  111

●

●

the good. In the case of differentiated road pricing, it should be assessed who travels proportionally more by car during peak hours and on congested roads than others. If the reduction in the transport externality is valued proportionally more by people with a higher social welfare weight then – everything else being equal – it is more attractive to increase the tax on the transport good contributing to the externality. For example, as indicated before, empirical studies indicate that the value of travel time savings increases with income (see, for example, Batley et al., 2019). This implies that if a higher social welfare weight is assigned to lower-income groups it becomes less attractive to increase transport taxes in order to mitigate congestion. If people with a high marginal social welfare weight are proportionally affected more physically by the externality, then – everything else being equal – it is more attractive to increase the tax on the transport good that causes the externality. This is for example relevant in the case of air pollution or noise nuisance, where the environmental impacts caused by traffic may differ across neighbourhoods and some groups may be more vulnerable than others to pollution, given their general health status.

Mayeres and Proost (2001) study marginal tax reforms in Belgium in a setting with a linear income tax. When only efficiency matters, it is welfare improving to reduce public transport subsidies, to increase the tax on peak car travel or to reduce the lump sum transfer, and to use the revenue to expand road capacity, to cut the tax on the composite non-transport commodity or that on off-peak car travel. With inequality aversion, it becomes less attractive to reduce the lump sum transfer as this instrument is important for redistribution in a linear income tax system. Only at high levels of inequality aversion does welfare increase with higher public transport subsidies, financed by a higher tax on peak travel. Such conclusions depend on the context that is considered. For example, in their analysis for Santiago, Chile, Tirachini and Proost (2021) find that a revenue-neutral reform within the transport sector would have to increase the car cost and reduce the bus fare during the peak, and vice versa in the off-peak period. However, the inclusion of inequality aversion points to a general reduction in bus fares independent of the time period, while car taxes should be increased. The previous discussion follows what is termed the ‘standard approach’ by Kreiner and Verdelin (2012). This is different from what they call the ‘new approach’ which can be illustrated by the evaluation of a marginal expansion of a public good. The approach looks for an adjustment of the tax function that keeps everyone’s utility level unchanged, which implies that the reform is distributionally neutral. The desirability of the expansion of the public good is then assessed by considering the impact on the government’s budget. If there is a net increase in government revenue, it is possible to make a Pareto improvement and the reform is desirable. The approach relies on the potential possibility of using a flexible non-linear income tax to undo any unintended distributional effects. Note that it does not require that the compensating changes in the income tax are actually implemented. A consequence is that distributional considerations should not play a role in decisions on the provision of public goods, under a number of conditions. According to Fosgerau and Van Dender (2013) applying this reasoning to congestion pricing means that ‘that distributional concerns and other tax distortions are irrelevant to how much reduction in congestion should be provided’. This holds as long as the individual’s income-earning ability and ability to home produce (at a given income) do not depend on his/her ability to benefit from less congested traffic.

112  Handbook on transport pricing and financing

6.3 INSIGHTS FROM EXISTING SCHEMES AND APPLICATIONS What do existing schemes teach us about the distributional impacts of pricing schemes? Here we focus on pricing schemes that differentiate by the time of use and location (as opposed to, for example, fuel taxation). Next to these schemes we also summarize the lessons that are drawn from applied economic studies of schemes that have been explored but are not (yet) implemented (e.g. UK cities, the Netherlands, Flanders and Brussels in Belgium). First, we consider vertical equity or the distributional impacts of road pricing in terms of income. Second, spatial equity is discussed, or the distribution of costs and benefits across regions. Third, the role of the use of the revenue generated by road pricing is highlighted. Finally, we reflect on the implications of a growing share of electric vehicles. 6.3.1 Distributional Impacts in Terms of Income In their literature review, Anas and Lindsey (2011) point out that in most studies the net impact of road pricing (i.e. the impact thanks to less congestion minus the charge that is paid) is found to be regressive in terms of income. In general, people with the highest incomes gain because the value of their time gains is larger than the charge they have to pay. In addition, people in the higher-income classes also have more flexibility (more flexible working hours, more possibilities to work at home etc., rather than fixed schedules, work in shifts, and the requirement to work at a certain location) and make more use of company cars. On the other hand, the reduction in congestion can also benefit public transport users by increasing the speed of public transport that uses the same infrastructure as the vehicles that are charged. On average people in lower-income groups travel relatively more by public transport, so that in this way road pricing can also have a progressive impact (Levinson, 2010). Basso and Silva (2014) look at the case where the supply of public transport is optimally adjusted (in terms of vehicle size, frequency and design of the bus stops), to accommodate the new demand for public transport brought about by road pricing. For Santiago, Chile, they find that road pricing is a progressive measure if such adjustments are taken into account, even before using the revenues from road pricing. The specific design and level of the charge and the presence – if any – of exemptions or discounts for certain groups or vehicles of course have implications for the distributional impacts of road pricing. For example, people using a company car will be affected more when they have to pay the congestion charge for their private travel than when this is included in the taxable benefit value in the labour income tax (Eliasson, 2016). A local charge can be progressive, neutral or regressive. This depends to a large extent on where people with different incomes live, where they work or frequently travel to for other purposes, as well as on the transport modes that they use and that are available to them. This follows from an analysis of theoretical cordon tolls in eight UK towns by Santos (2004). The income distributional impact of such a local system will therefore depend strongly on the specific city or region where it is implemented. The Stockholm congestion charge is an example of a local pricing scheme that is actually implemented. Kristofferson et al. (2017) explore the trade-off between efficiency and equity in the design of the Stockholm charge. They find that the most efficient charges also have the largest impact on the lowest income groups. The reason for this is the location where people belonging to different income groups live and work in and around Stockholm. People with

Equity and distributional issues in transport pricing  113

higher incomes live more centrally and closer to their working locations. In addition, they find that the differences between people within the lower- and middle-income groups in terms of car ownership and car dependency can be larger than the difference between people belonging to different income groups. Eliasson (2016) analyzed the cordon tolls in Stockholm and Gothenburg, as well as theoretical road charges in Lyon and Helsinki. The set-up of these systems differs with implications for the share of the population that pays a high amount in tolls. On average people with a higher income pay more tolls. However, car travel increases less than proportionally with income, so lower-income groups spend a relatively larger share of their income on the toll. The local charges are regressive in the four cities, slightly so in Stockholm and Helsinki, and moderately in Lyon and Gothenburg. He argues that this is not necessarily a problem if the purpose of the charge is to correct for the externality, but that it is a problem if the purpose is to generate revenues, where the tax is set higher than the marginal external costs. In that last case, the lower-income groups contribute more than proportionally to the revenues. In the analysis by Eliasson (2016), the higher-income groups in Helsinki and Gothenburg pay less than the middle-income groups. In the first city, this is the case because the highest income groups drive less by car because they live and work more centrally. In Gothenburg company cars, driven to a larger extent by people with high incomes, did not have to pay the charge at the time of the analysis. The congestion charge was included in the taxable benefit value of the car, which did not depend on the number of kilometres nor on the time of travel. In 2018 these rules were changed (Börjesson, 2018) so that company car users have to pay the congestion charge for their private travel. In an analysis of a local kilometre charge in Washington DC, using a regional computable equilibrium model, the result was progressive (Harrington et al., 2014). This is because the higher-income groups live further from the city centre, travel more by car and have longer commuting distances Managed lanes (high occupancy toll (HOT) lanes) have become more prominent in the United States since the 1990s. Even at the very start of their introduction, they led to considerable debate on fairness and equity. Weinstein and Sciara (2006) look at the practical implications of equity for a number of HOT lanes in operation. They indicate that the use of HOT lanes was more situation-dependent and less income dependent than commonly understood. The main equity consideration should not only focus on the affordability of the HOT lane, but also on the access to electronic payment (credit card), a bank account, or an electronic transponder. Safirova et al. (2003) analyzed three road pricing policies in Metropolitan Washington DC: (i) a conversion of high occupancy vehicle (HOV) lanes into HOT lanes with a toll during peak hours, (ii) a toll on all vehicles using the HOV lanes and on general purpose lanes adjacent to these lanes, with a higher toll on the HOV lanes, and (iii) an extension of the lower toll to all freeway lanes. These policies have different distributional impacts. With the HOT lane policy all income quartiles gain, with the largest gain (as a percentage of income) for the third quartile as many people of that group live in the suburbs and benefit from the improved traffic conditions. The second and third policy options are regressive. In the second case, the highest income quartile has a small welfare gain, and in the third case all income quartiles lose. While road pricing on a region-wide or national scale has not yet been implemented, a number of studies have explored its impacts, including the distributional effects. The research on the possible introduction of road pricing in the Netherlands found that the switch from the fixed car tax to an environmentally differentiated kilometre charge (of

114  Handbook on transport pricing and financing

approximately 6.4 Eurocent per km) in the whole of the country in combination with a peak tariff for particular road sections and time periods would have a positive impact on 60% of the households and a negative impact for 15% of them and that no large differences are expected according to income in these groups (Nederland, Tweede Kamer, 2009). The share of households gaining from the reform is expected to be larger in the higher-income groups, but also the share losing. Among the people in the lowest income groups, there is a large group that is not affected (46%), probably because car ownership is lower in those groups. The average impact over the income groups is expected to be positive because a large share of the charges is shifted to companies (distances travelled by leased cars and business-related travel). The final impact would depend mainly on the age, car ownership, the type of vehicle and the location where the km are travelled and less on the income itself. Heyndrickx et al. (2021) analyze the distributional effects of potential road pricing schemes in Flanders. The road pricing scenarios consist of a variable charge aimed at internalizing the external costs of transport, and the simultaneous abolishment of the car purchase tax and tax on car ownership. Considering the distributional impacts on the income quintiles in Flanders they find that within each quintile there are households who are affected positively as well as households who are affected negatively, as the intensity of car use and the driving patterns varies substantially between households with similar incomes. This finding is in line with that of, e.g. Kristofferson et al. (2017) for Stockholm. Aggregating over income and considering only the average impact per quintile makes these divergent effects less visible. The driving intensity and profile of the car users, more than the income group they belong to, determines the way in which they are affected by road pricing and therefore also their attitude towards the measure. At the same time, when expressed relative to income, lower-income people with high car intensity and travelling a lot in locations and at times with high levels of road pricing are worse off than higher-income people with similar driving profiles. Road pricing can also have a differential effect on different income groups as a function of their home location via its impact on air quality and noise levels (Ecola and Light, 2009). However, the association between environmental quality and socioeconomic status is not uniform across locations, even within the same country. According to Fecht et al. (2015) studies for North America point to higher exposures with lower socioeconomic positions, while the picture is more mixed in Europe. A recent EEA report on the case of air pollution and noise confirms this finding in the European context (EEA, 2018). While the association is casespecific, studies indicate that people of lower socioeconomic status are more vulnerable to the effects of environmental pollution (Brunt et al., 2017). Note that many of the studies adopt a static view of congestion, and do not consider the trip timing decisions and the dynamic build-up of congestion. Under this static setting, the model with homogeneous transport users predicts that a majority, or even all of them, will lose from the imposition of first-best congestion pricing. Only when the revenues are taken into account the pricing policy will lead to a welfare gain. With heterogeneous transport users, some users with a high value of time may gain before revenue is redistributed. With a dynamic model of traffic congestion, these conclusions change. In the bottleneck model, drawing upon Vickrey (1969), congestion takes the form of queuing at a bottleneck and departure time decisions are endogenous. Congestion costs are composed of both queuing costs and schedule delay costs (costs of arriving too early or too late). With homogeneous transport users, first-best pricing involves the implementation of a fine toll which eliminates all queuing. De Borger and Proost (2019) term this ‘supersmart’ road charging. The currently implemented schemes are

Equity and distributional issues in transport pricing  115

not yet of this type. With such supersmart charging the generalized costs of the transport users remain unchanged as is their welfare before revenue recycling and the above discussion on the equity impact would change. Van den Berg and Verhoef (2011) extend this model to account for heterogeneity in the value of time and schedule delays, while also considering price-sensitive demand. They find that a majority of transport users gain in this case, before considering the use of the revenues. The transport users who lose the most are drivers with an intermediate value of schedule delays, and the lowest value of time for that value of schedule delay. 6.3.2 Distributional Impacts across Locations Evidence for the currently existing cordon systems and area licence systems shows that they have a different impact on households within and outside of the charged area (Levinson, 2010). For example, building on the work of Flanders De Ceuster et al. (2020) performed an impact analysis of the SmartMove scheme, a city-wide car charging system that is envisaged by the Brussels Capital Region in Belgium. In this scheme, the time-differentiated kilometre charge is combined with a daily charge that depends on the power of the vehicle (measured by the so-called fiscal horsepower) and is larger when one travels during the peak. This combination offers a small but consistently progressive correction on the kilometre charge. In combination with implementing a zero tariff for private car ownership in the Brussels Region, the scheme is largely progressive for inhabitants of Brussels. However, there are regional equity trade-offs. Brussels attracts commuting from the Flemish and the Walloon region. Car ownership in Brussels is substantially lower than in the other regions, such that car dependence on commuting falls largely along regional lines. So while the overall welfare effect is positive and progressive along income categories, this is not true when considering the regions. In the institutional setting of Belgium, this is hard to remedy by after-tax redistribution, as fiscal aspects of car ownership are also regionalized. Hence it is difficult to develop a distribution scheme that can compensate non-inhabitants of Brussels after the implementation. With local systems, the size of the cordon area or licence area plays an important role. A smaller cordon will affect relatively more the people living inside the area because more of their destinations lie outside the area. People living outside the area are then also less dependent on their destination within the area and can also make use of other travel options. A larger cordon area has the opposite effect and a larger negative effect on people living outside the area. De Borger and Russo (2018) state that in the short run the benefits of an inwards cordon are mainly for the people inside the area. In Harrington et al. (2014) the difference for people living inside or outside of the area were, however, small and the differences were found to be larger in terms of income than where people live. A double cordon makes that more people living outside the city would have to pay. Residents of the neighbouring areas would then partly subsidize the city (Santos, 2004) In region-wide systems, the impact can be larger for people who live in areas served less by public transport or who have to travel distances that are not feasible to travel by bicycle or on foot. Eliasson et al. (2018) find that the share of people that would be worse off with a (flat) km charge would be larger in rural areas than in larger cities. In urban areas, the difference between central cities and satellite cities or suburbs is relevant because in the second case the car dependence is larger. The research for a potential road pricing system in the Netherlands (Nederland, Tweede Kamer, 2009) also found large regional differences. The travel time gains

116  Handbook on transport pricing and financing

are more limited in for example North and East Netherlands because there is less congestion. A regional differentiation of the tariff would therefore lead to higher social gains. The distributional effects can also arise because of other reasons. Car dependence can differ between different individuals or households. People who make a lot of trip chains, people working in industrial sites that are not easily accessible by public transport or people with a physical impairment are more car-dependent and may therefore have fewer possibilities to avoid the charges. Controlling for income, Eliasson (2016) finds for Lyon and Gothenburg that people with children under 18 years old pay more and that in Lyon older people would pay more. 6.3.3 The Role of Revenue Recycling Road pricing generates revenues which can be used in different ways. They can be allocated to transport, for example, by reducing existing vehicle taxes, improving road transport infrastructure, or providing better alternatives for car use (public transport, cycling infrastructure) such that it is easier to adapt to road pricing. While the allocation to transport purposes is often considered to be the most logical option in the general debate on road pricing, the revenue can also be used for more general purposes, including a reduction in distortive or regressive taxation elsewhere in the economy. The latter option is most relevant for road pricing schemes that are applied at a region-wide or nationwide scale. The revenue use will determine to a large extent the overall distributional effect of the reforms and whether this is in line with the distributional objectives of the government. A number of studies have analyzed the distributional impacts of road pricing using computable general equilibrium models. All of these studies consider region-wide or country-wide road pricing schemes. One example is the analysis of Belgium by Mayeres et al. (2004), who compare marginal social cost pricing and average cost pricing in combination with a change in the labour income tax or lump sum transfers. The policy scenarios consider an equal percentage change in these two instruments for all income groups. The analysis finds that average cost pricing based on financial costs leads to welfare losses for all five income quintiles considered. The percentage loss is the highest for the poorest households. This is the case for the two revenue recycling instruments. Marginal social cost pricing increases social welfare with equity impacts depending on how the revenue is used. When society becomes more inequality averse the lump sum transfer, which is more beneficial to the poorer income groups, is preferred. Steininger et  al. (2006) studied the distributional impacts of nationwide road pricing reforms for Austria and Kalinowska and Steininger (2009) for both Germany and Austria. Both studies show that by redistributing part of the road pricing revenues to the households the welfare costs of road pricing can be reduced across households and in particular for the poorer households. The study of Kalinowska and Steininger sheds light on the way in which preexisting differences in settlement structures, car availability, and public transport use across income groups affect the distributional impacts of the policy reforms. In Austria, these factors lead to a strong increase in car ownership and car mileage when income rises. In Germany car mileage also rises with income but less so for the high-income groups. As a result, the combination of road pricing and the redistribution of the revenues is progressive across all income groups in Austria, while in Germany it is progressive first and then regressive for the higher-income groups.

Equity and distributional issues in transport pricing  117

More recently the effects of revenue recycling were analyzed by Heyndrickx et al. (2021), using an innovative combination of a computable general equilibrium model (the EDIP model) with a household micro-simulation model (EUROMOD). This allowed using detailed information on household consumption and income. In their analysis for Flanders, they compare different revenue recycling scenarios, considering fiscal instruments that are available to the Flemish region. If everyone receives the same welfare weight the highest welfare gain can be obtained for society if the revenues are recycled via a proportional reduction of the Flemish surcharges on the personal income tax. If people belonging to lower-income quintiles receive a higher weight – for example with weights that range from 1 for the poorest quintile to 0.57 for the richest quintile – other recycling instruments become preferable, namely a progressive rather than proportional cut in the Flemish surcharges on the personal income tax, an earned income tax credit with a cap as well as a progressive income tax credit. They conclude that the earned income tax credit offers the best trade-off between equity and efficiency. Moreover, in that case people with similar driving characteristics have relatively similar welfare effects. While this analysis compared separate revenue use, the information can be used to design a combination of measures consisting of a general tax reform, improvements in transport services, and possibly targeted interventions for specific vulnerable groups. This approach was proposed for example by Small (1992) with the aim of mitigating possible negative impacts of road pricing, contribution to the broader distributional goals and getting the support of a sufficiently large group of voters. 6.3.4 Road Pricing and Electric Cars In future years, the share of electric vehicles (EVs) is expected to increase, reducing the environmental impacts of road transport. Currently, taxes on electric car use are mostly lower than for internal combustion engine vehicles (ICEVs), which has distributional consequences. Davis and Sallee (2020) estimate that in the United States the annual loss in gasoline tax revenue due to electric cars was $250 million in 2017 (this corresponds with less than half of 1% of total state fuel tax revenues, as the number of EVs was still limited). Electric cars were mainly driven by high-income households. Hence the tax treatment of EVs is regressive: using nationally representative microdata, Davis and Sallee find that households with an annual income of more than $100 000 and $200 000 are responsible for, respectively, more than twothirds and 31% of foregone tax revenues. Taxing EV use would be progressive, would generate revenues and confront EV users with the non-environmental externalities. However, given the low gasoline taxes in the United States, it would discourage the substitution away from gasoline vehicles. The authors indicate that the second-best tax on EV use depends on the relative importance of the different considerations, and on the type of instruments that are available (only a mileage tax, or a two-part tariff). In Norway, where the share of electric cars in the stock was more than 15% in 2020 (EAFO, 2021), a wide range of fiscal incentives are used to promote zero exhaust emission vehicles (Fridstrøm, 2021). According to Fevang et  al. (2020), the pioneers of EVs in Norway had higher incomes, more technical education and lived more in the suburbs of large cities compared to households with an ICEV. This suggests that the distributional impacts of the incentives are also regressive. With the growth of the market share of EVs that Norway has seen, the differences between the two groups remain, but have diminished over time.

118  Handbook on transport pricing and financing

As the variable user costs of EVs are lower than for ICEVs in countries with high fuel taxes, there will be a rebound effect on mileage which will lead to more congestion. This becomes an important concern when a significant part of the fleet becomes electric. For example, Wangsness et al. (2020) find that in Oslo, a city with a high penetration of electric cars, the promotion of EVs has led to higher levels of congestion and a reduction in public transport use. The requirement to keep such rebound effects under control, strengthens the case for efficient pricing of the externalities. The increase in the use of electric vehicles and the pressure this will have on public financing due to lower fuel tax revenue, also creates a unique opportunity for the introduction of innovative charging systems. For the implementation of road charging this may be a now or never moment. Börjesson et al. (2021) calculate the optimal Pigouvian tax for the densest regions in Sweden for 2040, considering a full replacement of the fleet by EVs. They find that the revenue from an optimal Pigouvian tax would roughly equal the present fuel taxes (excluding system costs). While optimal taxes on congested roads would largely satisfy to pay for road capacity, this would also imply low or zero taxes in non-congested rural areas. Implicitly this creates a shift of resources from cities to rural areas. Therefore, one should be aware of a possible inequity in the spatial dimension. The authors conclude that road charging that is limited to big cities and surrounding highways may be much more cost-effective than a national wide charging scheme that include full GPS tracking and enforcement with little impact on actual revenue collected. The equity impacts of a road tax when only EVs are used, of course, depend on the design of the tax. Alternatively, based on the analysis in Section 6.1.2, one could think of a regionwide differentiated road tax that would be composed of a base tariff (revenue-raising component), a charge related to non-environmental externalities (mainly congestion and road safety) and a possible correction for equity impacts. Moreover, also the system costs of road pricing can be expected to change in the future. Many electric cars are already equipped with electronic systems that allow for easy integration with ‘smart charging’ systems and GPS tracking. When such systems become more widely available in the vehicle fleet, this will facilitate the technological leap to road charging and reduce the system costs.

6.4 OVERCOMING IDEOLOGICAL BIAS IN ROAD CHARGING Discussions on the social impact of road charging often contain ideological bias. First, by ignoring inherent inequity in the Business-as-Usual scenario, either in financing infrastructure or in environmental and social impact of current car travel. Second, by not taking into account indirect impacts through revenue recycling. Third, by putting too much emphasis on specific groups with perceived vulnerabilities. Fourth, by disregarding possible alternatives to car travel. And fifth, by overlooking benefits from reduced car travel, which can spill over into public transport, accessibility for active modes and safety. Providing objective counterarguments can be challenging, as the impact of road charging is notably complex. Any change in tax policy comes at cross-points between revenue collection, distributional concerns, and correction of externalities. Since road charging taxes mobility, it may have broad economic impacts. Also, there are groups that can potentially be more affected by road charging. Among those: commuters from rural regions with high car dependence, low-income households with inflexible working arrangements and elderly people or mobility impaired people with high accessibility requirements.

Equity and distributional issues in transport pricing  119

If one wants to overcome inherent ideological biases and gain objectivity, the first step is to analyze the impacts of planned road pricing schemes. Preferably such an analysis should use detailed data. Experience by Heyndrickx et al. (2021) using microdata at the household level, shows that this may offer invaluable insight into the distributional and social impacts. Once this information is available, it can be used in the policy discussion where the relative weights of individuals within the social welfare function will play. Showing the impacts of using different weights will also be instructive. Even adding mild inequality aversion can lead to shifting welfare weights that lead to radically different conclusions concerning road charging. In a typically (quasi) Rawlsian perspective, only the changes for the poorest and most deprived individual(s) in society matter. If equity concerns on the other hand are largely absent, for example in a utilitarian perspective, the losses of any individual can be offset against any other individual(s). We would therefore propose to define clearly what are acceptable distributional weights and if these should be defined on the basis of income, gender, household composition, age or any other characteristic. Before considering the use of road pricing revenues or changes in public transport supply, there is little doubt that the direct impact of road charging is overall regressive with respect to household income, especially when we consider the subgroup of households that own cars (Eliasson, 2016). In the general population, road charging will only be progressive if car use and car ownership are closely correlated with high income levels. This implies that in countries with higher and more evenly distributed levels of car ownership, the regressivity of road charging will generally increase. In a similar way, regions with high levels of public transport services will be less affected than more rural or sprawled regions with higher car dependency. It is probably not a coincidence that successful road charging schemes were implemented in urban regions (Stockholm, London, Singapore, Milan), often with considerable opposition from the periphery. Indeed, the (positive) result of the Stockholm referendum on congestion charges, might have looked different if a larger number of peripheral regions had been included in the vote. At the same time, the vote in the urban region was more favourable than expected, even when road users had lower time benefits and paid higher relative charges. This shows that the social benefits of road charging and the perceived environmental and safety benefits may be underestimated (Eliasson et al., 2009). It is also interesting to note that while car ownership may be a critical value to start from in any distributional impact analysis, it is not the whole story. While theoretically optimal road charging should cover an entire network, road charging in practice targets specific infrastructure and areas. Diaz and Proost (2014), for example, show that the optimal policy in terms of equity and efficiency depends on the type of equilibrium that is generated; whether the wealthier individuals choose the taxed road only or are distributed over the two types of roads. If we limit our analysis of road charging to the transport network and exclude schedule delay costs, the reduction in congestion and other externalities of road transport will generally lead to a net loss for individual car travellers. The reason is that elasticity of substitution for car travel is generally quite low and often lower than one. This brings out several insights and conclusions: 1) the distribution of car use in the initial equilibrium cannot be ignored, 2) the use of tax revenue is a critical element of any social impact analysis of road charging, 3) we may underestimate benefits by excluding ‘bottleneck’ behaviour and the importance of ‘fine tolling’ (Van den Berg and Verhoef, 2011; De Borger and Proost, 2019) To gain momentum in the implementation of road charging, it may be useful to focus more on actual policy design than discussing road charging in general terms. Experience with

120  Handbook on transport pricing and financing

congestion charging in cities as well as managed lanes in the United States shows that users – independent of their affluence – do appreciate choice in (road) transport. Therefore, the way forward for road charging can be to focus more on a message of extending user choices and improving the quality of the transport system, rather than ‘punishing’ the user or removing options. Road charging extends user choices by allowing for a better quality higher speed network, improving the speed and safety of vulnerable road users in cities and extending the financing of both road and public transport alternatives.

ACKNOWLEDGEMENTS The work was supported via the energy transition funds project ‘EPOC 2030–2050’ organized by the Belgian FPS economy, S.M.E.s, Self-employed and Energy. The authors would like to thank the editors and the reviewer for their valuable comments. Any remaining errors remain the responsibility of the authors.

REFERENCES Ahmad, E. and N. Stern (1984), The theory of reform and Indian indirect taxes. Journal of Public Economics 25, 259–298. Anas, A. and R. Lindsey (2011), Reducing urban road transportation externalities: Road pricing in theory and in practice. Review of Environmental Economics and Policy 5(1), 66–88. Atkinson, A.B. and J.E. Stiglitz (1976), The design of tax structure: Direct versus indirect taxation. Journal of Public Economics 6, 55–75. Atkinson, A.B. and J.E. Stiglitz (1980), Lectures on Public Economics. McGraw-Hill International Editions. Attard, M. and M. Enoch (2011), Policy transfer and the introduction of road pricing in Valletta, Malta. Transport Policy 18(3), 544–553. Basso, L.J. and H.E. Silva (2014), Efficiency and substitutability of transit subsidies and other urban transport policies. American Economic Journal: Economic Policy 6(4), 1–33. Batley, R., J. Bates, M. Bliemer, M. Börjesson, J. Bourdon, M.O. Cabral, P.K. Chintakayala, C. Choudhury, A. Daly, T. Dekker and E. Drivyla (2019), New Appraisal values of travel time saving and reliability. Transportation 46: 583–621. Börjesson, M., D. Asplund and C. Hamilton (2021), Optimal kilometre tax for electric passenger cars, VTI working paper 2021:3. Börjesson, M. (2018), Long-Term Effects of the Swedish Congestion Charges. International Transport Forum Discussion Papers. OECD Publishing, Paris. Börjesson, M. and J. Eliasson (2012), The value of time and external benefits in bicycle appraisal. Transportation Research A 46, 673–683. Bovenberg, A.L. and F. van der Ploeg (1994), Environmental policy, public finance and the labour market in a second-best world. Journal of Public Economics 55, 349–390. Brunt, H., J. Barnes, S. Jones, J. Longhurst, G. Scally and E. Hayes (2017), Air pollution, deprivation and health: Understanding relationships to add value to local air quality management policy and practice in Wales. Journal of Public Health 39(3), 485–497. Cremer, H., F. Gahvari and N. Ladoux (1998), Externalities and optimal taxation. Journal of Public Economics 70(3), 343–364. Cremer, H., F. Gahvari and N. Ladoux (2003), Environmental taxes with heterogeneous consumers: an application to energy consumption in France. Journal of Public Economics 87(12), 2791–2815. Croci, E. (2016), Urban road pricing: A comparative study on the experiences of London, Stockholm and Milan. Transportation Research Procedia 14, 253 – 262.

Equity and distributional issues in transport pricing  121

Davis, L.W. and J.M. Sallee (2020), Should electric vehicle drivers pay a mileage tax? Environmental and Energy Policy and the Economy 1, 65–94. De Borger, B. and S. Proost (2012), A political economy model of road pricing. Journal of Urban Economics 71(1), 79–92. De Borger, B. and S. Proost (2019), Superslim rekeningrijden voor dummies (Super smart road pricing for dummies), Leuvense Economische Standpunten 2019/175 (in Dutch). De Borger, B. and A. Russo (2018), The political economy of cordon tolls. Journal of Urban Economics 105, 133–148. De Ceuster, G., I. Mayeres, B. Ons, C. Heyndrickx, T. Truyts, G. Grandjean and V. Sheremeta (2020), SmartMove, Impact analyse, Effecten op de mobiliteit en externe kosten van transport, budgettaire effecten en social-economische effecten (SmartMove Impact Analysis, Effects on mobility and external costs of transport, budgetary effects and socio-economic effects), Study by the Motivity Consortium for the Brussels Capital Region (in Dutch). Diamond, P.A. and J.A. Mirrlees (1971), Optimal taxation and public production I: Production efficiency and II: Tax rules. American Economic Review 61, 8–27 and 261–278. Diaz, A. and S. Proost (2014), Second-best urban tolling with distributive concerns, Economics of Transportation 3(4), 257–269. EAFO (2021), Vehicles and fleet, Passenger cars. https://www​.eafo​.eu​/vehicles​-and​-fleet​/m1. Accessed June 15, 2021. Ecola, L. and T. Light (2009), Equity and congestion pricing, a review of the evidence. Technical Report, RAND Corporation. EEA (2018), Unequal exposure and unequal impacts: social vulnerability to air pollution, noise and extreme temperatures in Europe, EEA Report No 22/2018. Eliasson, J. (2014), The role of attitude structures, direct experience and reframing for the success of congestion pricing. Transportation Research A: Policy and Practice 67, 81–95. Eliasson, J. (2016), Is congestion pricing fair? Consumer and citizen perspectives on equity effects. Transport Policy 52, 1–15. Eliasson, J., L. Hultkrantz, L. Nerhagen and L. Smidfelt-Rosqvist (2009), The Stockholm congestioncharging trial 2006: Overview of the effects. Transportation Research A 43, 240–250. Eliasson, J., R. Pyddoke and J.-E. Swärdh (2018), Distributional effects of taxes on car fuel, use, ownership and purchases. Economics of Transportation 15, 1–15. European Commission (2019), Handbook on the External Costs of Transport. Publications Office of the European Union, Luxembourg. Fecht, D., P. Fischer, L. Fortunato, G. Hoek, K. De Hoogh, M. Marra, H. Kruize, D. Vienneau, R. Beelen and A. Hansell (2015), Associations between air pollution and socioeconomic characteristics, ethnicity and age profile of neighbourhoods in England and the Netherlands. Environmental Pollution 198, 201–210. Fevang, E., E. Figenbaum, L. Fridstrøm, A.H. Halse, K.E. Hauge, B.G. Johansen and O. Raaum (2020), Who goes electric? Characteristics of electric car ownership in Norway 2011–2017, TØI Report 1780/2020, TØI, Oslo, Norway. Fosgerau, M. and K. Van Dender (2013), Road pricing with complications, Transportation 40, 479–503. Fridstrøm, L. (2021), The Norwegian vehicle electrification policy and its implicit price of carbon. Sustainability 13, 1346. Gaunt, M., T. Rye and S. Allen (2007), Public acceptability of road user charging: The case of Edinburgh and the 2005 referendum. Transport Reviews 27(1), 85–102. Harrington, W., E. Safirova, C. Coleman, S. Houde and A.M. Finkel (2014), Distributional consequences of public policies, an example from the management of urban vehicular travel. RFF discussion paper 14–64, Resources for the Future. Heyndrickx, C., T. Vanheukelom and S. Proost (2021), Distributional impact of a regional road pricing scheme in Flanders. Transportation Research Part A 148, 116–139. Jacobs, B. and R.A. de Mooij (2015), Pigou meets Mirrlees: on the irrelevance of tax distortions for the second-best Pigouvian tax. Journal of Environmental Economics and Management 71(May), 90–108.

122  Handbook on transport pricing and financing

Jacobs, B. and F. van der Ploeg (2019), Redistribution and pollution taxes with non-linear Engel curves. Journal of Environmental Economics and Management 95, 198–226. Kalinowska, D. and K.W. Steininger (2009), Distributional impacts of car road pricing: Settlement structures determine divergence across countries. Ecological Economics 68(12), 2890–2896. Kreiner, C.T. and N. Verdelin (2012), Optimal provision of public goods: A synthesis. Scandinavian Journal of Economics 114(2), 384–408. Kristoffersson, I., L. Engelson and M. Börjesson (2017), Efficiency vs equity: Conflicting objectives of congestion charges. Transport Policy 60, 99–107. Levinson, D. (2010), Equity effects of road pricing: A review. Transport Reviews 30(1), 33–57. Mayeres, I. and S. Proost (1997), Optimal tax and investment rules for congestion type of externalities. Scandinavian Journal of Economics 99, 261–279. Mayeres, I. and S. Proost (2001), Marginal tax reform, externalities and income distribution. Journal of Public Economics 79(2), 343–363. Mayeres, I., S. Proost and K. Van Dender (2004), The impacts of marginal social cost pricing. In Nash, C. and B. Matthews (eds.), Measuring the Marginal Social Cost of Transport, Elsevier Science. Micheletto, L. (2008), Redistribution and optimal mixed taxation in the presence of consumption externalities. Journal of Public Economics 92(10–11), 2262–2274. Mohring, H. (1972), Optimization and scale economies in urban bus transportation. American Economic Review 62(4), 591–604. Nederland, Tweede Kamer (2009), Regels voor het in rekening brengen van een gebruiksafhankelijke prijs voor het rijden met een motorrijtuig (Wet kilometerprijs), Kamerstuk. https://www. parlementairemonitor.nl/9353000/1/j9vvij5epmj1ey0/viab81uy7izg (Rules for taking into account a use dependent price for driving a motor vehicle (Law kilometre price))(in Dutch). Pirttilä, J. and R. Schöb (1999), Redistribution and internalization: The many-person Ramsey tax rule revisited. Public Finance Review 27(5), 541–560. Portney, P.R. and J.P. Weyant (1999), Introduction. In: P.R. Portney and J.P. Weyant (eds.), Discounting and Intergenerational Equity, Resources for the Future, pp. 1–11. Washington, DC. Rotaris, L., R. Danielis, E. Marcucci and J. Massiani (2010), The urban road pricing scheme to curb pollution in Milan, Italy: Description, impacts and preliminary cost-benefit analysis assessment. Transportation Research A: Policy and Practice 44(5), 359–375. Safirova, E., K. Gillingham, I. Parry, P. Nelson, W. Harrington and D. Mason (2003), Welfare and distributional effects of HOT lanes and other road pricing policies in metropolitan Washington, D.C. In: G. Santos (ed.), Road Pricing: Theory and Practice, pp. 179–208. Elsevier Publishing, Amsterdam. Sandmo, A. (1975), Optimal taxation in the presence of externalities. Swedish Journal of Economics 77, 86–98. Sandmo, A. (2000), The Public Economics of the Environment, The Lindahl Lectures 1996. Oxford University Press, Oxford. Santos, G. (2004), Urban congestion charging—A second-best alternative. Journal of Transport Economics and Policy 38, 345–369. Schade, J. (2017), Brief review about the public acceptability of road pricing strategies. Reflets et perspectives de la vie économique 2017/2, LVI, 139–148. Schöb, R. (1996), Evaluating tax reforms in the presence of externalities. Oxford Economic Papers 48, 537–555. Small, K.A. (1992), Using the revenue from congestion pricing. Transportation 19, 359–381. Steininger, K.W., B. Friedl and B. Gebetsroither (2006), Sustainability impacts of car road pricing: A computable general equilibrium analysis for Austria. Ecological Economics, Elsevier 63(1), 59–69. Tirachini, A. and S. Proost (2021), Transport taxes and subsidies in developing countries: The effect of income inequality aversion. Economics of Transportation 25, 100206. Van den Berg, V. and E. Verhoef (2011), Winning or losing from dynamic bottleneck congestion pricing? The distributional effects of road pricing with heterogeneity in values of time and schedule delay. Journal of Public Economics 95, 983–992. Vickrey, W.S. (1969), Congestion theory and transport investment. American Economic Review (Papers and Proceedings) 59 (2), 251–260.

Equity and distributional issues in transport pricing  123

Walker, J. (ed.)(2018), Road Pricing, Technologies, Economics and Acceptability. IET Transportation Series 8, The Institution of Engineering and Technology, London, UK (ISBN 978-1-78561-205-3). Wangsness, P. B., S. Proost and K.L. Rødseth. (2020), Vehicle choices and urban transport externalities. are Norwegian policy makers getting it right? Transportation Research Part D: Transport and Environment 86, 102384. Weinstein, A. and G.-C. Sciara (2006), Unraveling equity in HOT lane planning. Journal of Planning Education and Research 26, 174–184.

7. The political economy of transport pricing and investment Bruno De Borger and Antonio Russo

7.1 INTRODUCTION In this chapter, we summarize some insights obtained from studying the recent literature on the political economy of transport decision-making. For many decades, most of the literature on the economics of transport pricing and investment decisions by public authorities took a purely normative approach. The main objective of this literature was to provide welfareoptimal pricing and investment rules that could provide some guidance for policy-makers. Starting from first-best principles covered in Chapter 2 of this Handbook (Pigouvian taxation and investment up to equality between marginal benefit and marginal cost) the literature has provided a large number of careful second-best extensions to deal with complex network structures and the potential impossibility to optimally price all links in a given network, to allow for schedule delay, to introduce concerns of income distribution, to account for the cost of public funds, and so on.1 Confronting the guidelines from the welfare economic approach to pricing and investment decisions with observed transport policies, it becomes clear that many policy prescriptions of the transport economic literature have not been translated into actual policy-making rules. For example, despite much economic support for the introduction of some form of road pricing, cases of successful implementation are scarce.2 There are at least two reasons why this is the case. First, the practical application of welfare-optimal pricing and investment rules on a real-world network with many different types of users requires the solution of a number of difficult technical and organizational problems. Second, the guidelines that come out of the economics literature are just one input into the political process that shapes transport policies. The policies that democratically elected representatives propose take into account ideology, the preferences of politicians’ constituency, the characteristics of the democratic process, etc. Simply stated, economic guidelines are an input in the political decision-making process, but they are not a substitute. Clearly, therefore, economists must pay more attention to the politics of transport decisions if we want to better understand the pricing and investment decisions observed. Fortunately, recently the economics profession has taken a less normative view on decision-making, emphasizing the role of the political process in shaping transport policy. This chapter is an attempt to bring together a number of major insights developed in this fairly recent literature. Since our purpose is to introduce the reader to the political economy of transport policy, we do not exhaustively cover the huge body of existing work on the role of politics in other areas of economic decision-making. Instead, we focus on theoretical and empirical models that economists have used to understand transport policy decisions. The remainder of this chapter proceeds as follows. Section 7.2 provides a brief overview of the different political economy models that have been adopted by transport economists. 124

The political economy of transport pricing and investment  125

Sections 7.3 and 7.4 survey the literature on the political economy of transport pricing, focusing on road pricing and parking charges, respectively. Section 7.5 presents an overview of studies that emphasize the political economy of infrastructure provision. In Section 7.6 we briefly consider the political economy of traffic law enforcement. Finally, Section 7.7 provides concluding remarks and discusses some interesting avenues for future research.

7.2 THEORETICAL MODELS OF POLITICAL DECISION-MAKING This section provides a brief and non-technical overview of theoretical models in political economy and collective decision-making. Some models emphasize the role of voters and the voting process, others focus on the pressure of lobby groups on political decision-makers; still others have provided stylized descriptions of how the legislature operates, analyzing the legislative bargaining process among representatives from different constituencies. For a more comprehensive survey of the political economy literature, we refer to textbooks such as Persson and Tabellini (2000). 7.2.1 Majority Voting Majority voting is arguably the simplest and most common political economy framework. The basic idea is to consider a group of voters deciding on a single policy variable (e.g., a tax rate). The policy adopted in equilibrium is the “Condorcet Winner”, i.e., the policy that obtains most votes in a pairwise contest with any other feasible policy (Marquis de Condorcet, 1875). Downs (1957) showed that if voters’ preferences are single-peaked a Condorcet Winner always exists, and the winning policy (for example, the chosen tax rate) coincides with the policy preferred by the median voter. Majority voting is well suited to explain the adoption of – possibly inefficient – policies that involve some redistribution of welfare that benefits the majority, such as the adoption of fuel subsidies. It has also been used to explain the failure to adopt road tolls, see below. Despite its versatility, the majority voting framework has some important limitations. An equilibrium exists only under fairly restrictive conditions, and there is typically no equilibrium when individuals vote on multiple policy variables jointly. Thus, this framework is not well suited for interpreting the adoption of multi-dimensional policy platforms such as pricing of multiple transport modes, or joint pricing and investment decisions. Furthermore, it ignores the personal preferences and interests of elected officials as well as the influence of special interest groups on implemented policies. This framework also overlooks the fact that, in representative democracies, policy is typically the result of bargaining among representatives of different constituencies and jurisdictions. For these reasons, scholars have developed alternative political economy frameworks that address these limitations. 7.2.2 Probabilistic Voting Models As mentioned above, the standard majority voting framework is unsuitable to study multidimensional policy platforms. Furthermore, this framework fails to account for the importance of “swing” voters in elections, as opposed to the “median” voter. The probabilistic voting

126  Handbook on transport pricing and financing

model overcomes these limitations (for some early references, see Lindbeck and Weibull, 1987; and Coughlin, 1992). The model assumes that otherwise identical voters have not only idiosyncratic preferences for competing candidates (or political parties), but also preferences toward particular policies being proposed. For example, a voter may ideologically be on the left of the political spectrum, while specifically being opposed to a proposed infrastructure project. It is typically assumed that there is a given number of voters, and that within each group voters have identical policy preferences but heterogeneous, smoothly distributed, ideological preferences (e.g., ranging from left to right wing) for the candidates. These preferences are assumed to be private information. Voters decide on which candidate (or party) to vote for based on their ideology as well as on the proposed policies of the candidate. The candidates are only interested in being elected; they decide on the policy platform proposed to the voters to maximize the expected share of the vote. It is shown that the equilibrium platform is such that it maximizes a weighted welfare function with weights corresponding to the size of each group and the share of swing voters, who are indifferent between the two candidates, in each group. For more details, see the references given above. 7.2.3 Citizen–Candidate Models The citizen–candidate framework (originally introduced by Osborne and Slivinski (1996) and Besley and Coate (1997)) assumes that individuals can choose to run for office at a cost and, if elected, they implement their preferred policy. In principle, this policy may be multidimensional. In this framework, voters can either vote “sincerely” (i.e., for the candidate proposing the policy closest to their own preferences) or “strategically” (i.e., taking into account each candidate’s probability of winning). If voters are sincere, an equilibrium exists where the median voter is the only candidate and wins the election, similar to the standard majority voting framework. However, there can also be equilibria with two distinct candidates running for office, each with a positive probability of winning, and with positions that are neither too different, nor too close to each other. Unlike the previous two types of models, therefore, the citizen-candidate model admits equilibria in which voters are presented with different policy platforms from which to choose. 7.2.4 Lobbying by Special Interests Albeit difficult to measure, the influence of lobbies on economic policy is likely to be relevant. One of the most common models capturing such influence is the “protection for sale” model based on the common agency framework of Bernheim and Whinston (1986), and originally introduced by Grossman and Helpman (1994). This model focuses on the incentives of politicians in office, assuming they care for social welfare as well as for collecting transfers from special interest groups (e.g., campaign contributions from large companies or industrial lobbies). Neither the political system nor the lobbying process are explicitly modeled. It is assumed that the decision-maker maximizes a weighted average of two components, viz., social welfare and the transfers received from the special interests; the weights are determined by the importance of transfers in her utility function. Of course, the transfers received from the lobbies depend on the policy the decision-maker implements, i.e., the contribution of the lobby to the policymaker is a function of the policy chosen. Therefore, unless every sector

The political economy of transport pricing and investment  127

of the economy is organized in a lobby, the policy adopted in equilibrium differs from the welfare-maximizing outcome to benefit the special interests. 7.2.5 Legislative Bargaining Transport policy decisions often involve multiple levels of government (e.g., federal and local), and they result from a compromise between representatives of different constituencies. A variety of models have been proposed to incorporate this observed characteristic of the democratic process, assuming that regional representatives are elected to a federal or national parliament. For example, Baron and Ferejohn (1989) develop a model of repeated bargaining over a fixed amount of resources in a legislature. The authors assume that one of the representatives in each period is chosen randomly as the proposer of the “agenda”. This “agenda setter” proposes a policy to the legislative body. Legislators vote by majority voting on whether to adopt the policy proposed. If the proposal gets a majority it is executed; if it is rejected, a new agenda setter is randomly determined and the process is repeated. It is shown that in equilibrium, each agenda setter proposes a policy that maximizes the welfare of his or her constituency, conditional on the policy obtaining a majority against the status quo. Besley and Coate’s (2003) legislative bargaining framework tackles one of the fundamental questions in public finance, i.e., whether spending and taxation powers should be centralized or delegated to regional or local governments.3 It has been quite popular with transport economists, see below. Besley and Coate compare centralized and decentralized (regional) decisions on the provision of a public good in a two-region setting. Regional decisions are the result of majority voting at the regional level. In the centralized scenario, policy decisions result from bargaining in a national (or federal) legislature of elected representatives from each region. The authors highlight two potential sources of inefficiency from centralized policy-making. First, to strengthen their representative’s position at the legislative bargaining table, voters from each region strategically elect representatives that have strong preferences for public spending, leading to excessive spending at the federal level. Second, if decisions in the legislative assembly are made by winning coalitions, spending is skewed toward those regions whose representatives enter the winning coalition. This results in the misallocation of resources and uncertainty about public spending. When choosing between a centralized and a decentralized structure of government, the above inefficiencies must be weighed against the classic inefficiencies caused by regional spill-overs with uncoordinated policy-making (Oates, 1999). 7.2.6 Political Budget Cycles The political budget cycle hypothesis is quite simple: as elections approach, incumbent politicians want to buy voter support by manipulating the public budget, e.g., by raising government spending or cutting taxes. A counterargument is that, if voters are rational and well informed, they should also be wary of the budgetary restrictions that politicians must adopt to balance the budget after the electoral period. Therefore, it is not obvious that voters would reward politicians for engaging in this kind of behavior. Based on this observation, the literature on political budget cycles has developed in two directions (see Drazen, 2008, for a survey of this literature). At the theoretical level, researchers have provided models explaining why politicians would follow political budget cycles with rational voters. The arguments put forward

128  Handbook on transport pricing and financing

include, among others, that manipulating the budget could be a way to signal their competence or their preferences to imperfectly informed voters. At the empirical level, economists have sought evidence of such cycles, focusing on different budgetary components, countries and government levels. Furthermore, scholars have tried to measure the possible knock-on effects on welfare due to tax-induced distortions, or to policymakers not enforcing regulations when close to elections.

7.3 THE POLITICAL ECONOMY OF TRANSPORT PRICING: ROAD PRICING We now turn to the political economy that underlies government pricing decisions in the transport sector. A large part of the literature on this topic focuses on pricing road use; as will become clear below, much less is written about the political economy of public transport fares. In this section, we therefore start with the politics of road pricing in its various formats (electronic road pricing, cordon pricing). The next sections will consider parking pricing policies and models that focus on investment in transport infrastructure. Road congestion is ubiquitous in cities throughout the world. As there is a clear economic rationale for road pricing – based on the classic “Pigouvian taxation” argument (Pigou, 1920) – it is one of the policies that many economists widely support. However, the public as well as elected politicians often strongly oppose this policy; accordingly, few governments have adopted it. Based on economic intuition rather than formal modeling, several transport economists have argued that who gets the toll revenues and how these revenues are used are critical determinants of the political support for implementing road pricing (see, for example, Small, 1992; Goodwin, 1994; King et al., 2007). Moreover, a remarkable consensus seems to be that a substantial share of the revenues should be allocated to public transport. Both Small (1992) and Goodwin (1994) concluded that to get a political majority a useful allocation of the revenues is to use the money to invest in public transport, to improve the quality of the road network, and to reduce other taxes (for example, the distortive tax on labor). Over the past two decades, several authors have proposed more formal political economy models in an attempt to explain the lack of widespread application of road pricing. One group of papers takes a short-run view and treats individuals’ locations as given, ignoring the land and housing markets. These studies focus on the short-run distributional implications of the policy, such as the transfer of resources from those who travel by car (and pay the toll) to those who do not. A second group of papers consider the long-run effects of road pricing policies, allowing for changes in the price of land and housing, and evaluating the implications for the distribution of welfare among voters. In the remainder of this section, we discuss both types of theoretical models in turn. We conclude with a brief overview of empirical studies. 7.3.1 The Short-Run: Fixed Location To the best of our knowledge, the first formal paper trying to explain the lack of support for road pricing is the citizen–candidate model of Marcucci, Marini, and Ticchi (2005). Their analysis concentrates on the diverging interests of three income classes (the poor, the rich, and the middle class) that differ in their value of time and, hence, in their willingness to pay for reducing congestion. The authors show that, unless toll revenues are fully redistributed,

The political economy of transport pricing and investment  129

only the rich may favor a positive toll. Since the rich are typically not the majority, the political equilibrium therefore implies that no toll is implemented. However, some equilibria with a positive toll can emerge if the revenue is used to fund public transport, depending on which mode the poor and middle class use and on how sensitive they are to congestion. The findings of Marcucci et al. (2005) are consistent with the observed reluctance of many governments to implement road pricing and the intuition of earlier authors (see above) that improvements to public transport – often part of policy proposals involving road tolls – enhance their acceptability.4 Apart from allocating a substantial part of the revenues to improve public transport, the few cases of successful implementation of road pricing (London, Stockholm, Milan, Trondheim) shared another remarkable characteristic: public opinion was typically opposing road pricing before it was introduced but became much more favorable after formal implementation. To explain these observations, De Borger and Proost (2012) focused on the role of individual and political uncertainty in determining the attitude of voters toward a road toll. Following Fernandez and Rodrik (1991), the authors assume that car drivers in the no-toll status quo cannot perfectly evaluate their idiosyncratic net cost (or benefit) caused by the introduction of the toll. This net cost takes into account the monetary expenditure (toll payment), the time gain from reduced congestion, and the amount of toll revenues the government lump-sum redistributes to all voters. It is assumed that uncertain voters focus on the expected net cost per driver due to the toll which, under a mild assumption, the authors show to be positive.5 As a result, uncertainty leads even status-quo drivers, who would benefit from the introduction of the toll, to oppose its implementation. Overall, uncertainty makes opposition to the toll more widespread. The model explains several stylized facts, some of which were mentioned above. First, it explains why most of the attempted referenda on the introduction of a road toll in cities have failed. More precisely, cities that organized referenda on road pricing without a trial had their proposal rejected (e.g., Manchester, Edinburgh). However, some cities (e.g., Stockholm) that eliminated uncertainty by organizing a temporary trial of the toll before asking voters whether or not to maintain it, received voter approval for toll introduction. Second, their result is consistent with the observation that opposition to the toll tends to decrease once it is implemented (so that uncertainty is resolved) and voters come to appreciate its effects (Owen et al., 2008). Finally, De Borger and Proost (2012) show that fewer voters oppose the toll when revenues are used to fund public transport which, as mentioned, is consistent with cities tying toll proposals to increased funding for public transport. Recognizing that governments have multiple instruments to tackle congestion in their arsenal, some authors have compared the acceptability of road pricing to that of alternative policies, also considering the possible interactions between them. Russo (2013) studies voting on road tolls as well as parking fees, under the assumption that tolls are decided by a regional government (representing the city as well as the suburbs), whereas parking is controlled by the city government. His main finding is that governments are likely to coordinate on low (or zero) tolls and high(er) parking charges, which is consistent with the casual observation that parking fees are much more pervasive than tolls. The result is due to the combination of tax exporting by the city government, substitutability between the two policy instruments, and the assumption that voters in the city are less car-dependent than suburban voters. Nevertheless, the parking charge may still be below the optimal level if a strong preference for driving is widespread among the city population.

130  Handbook on transport pricing and financing

Interestingly, note that a common conclusion of the papers discussed so far (Marcucci et al., 2005; De Borger and Proost, 2012; Russo, 2013) is that the acceptability of policies that discourage car use increases when tied to subsidies to public transport. This finding is clearly robust with respect to the use of very different models. Recently, Ren and Huang (2020) found similar results in a two-mode setting with dynamic congestion at a bottleneck. De Borger and Glazer (2017) examine the acceptability of road tolls from a different, though somewhat related angle. Drawing from the literature in behavioral economics, the authors assume individuals have reference-dependent preferences characterized by loss aversion (Kahneman and Tversky, 1979). In a nutshell, individuals get some intrinsic (dis) utility from changes with respect to their reference point (the no-toll status quo), and they value any given potential gain less than (the absolute value of) an equal potential loss. The authors show that loss aversion reduces the size of the socially optimal toll and, all else given, it makes opposition to the toll more widespread among current drivers. Furthermore, loss aversion can explain the increase in support for road pricing after the introduction of the policy. Finally, the authors consider lobbying by organized groups (e.g., associations of drivers, pedestrians, and environmentalist groups), finding that loss aversion can either encourage or discourage lobbying against the toll, but it unambiguously discourages lobbying in favor of the toll. Several papers have used a similar framework to compare the political acceptability of road tolls with alternative policies to reduce the negative externalities of road use. First, Fageda, Flores-Fillol, and Theilen (2021) tried to understand why vintage-based driving restrictions, such as low-emission zones (LEZ), are much more widespread than road tolls. The authors explain this stylized fact by showing that, if individuals with the highest willingness to pay for driving own the newest (least polluting) cars, under a low-emission zone they can keep on driving at no extra cost, and they enjoy lower congestion and pollution. By contrast, the net benefit from a road toll to this group would be negative, because the value of the time savings does not outweigh the toll payment. Since LEZs and road tolls have similar effects on non-drivers, the authors conclude that fewer voters support tolls than LEZs; in this sense, tolls are less likely to be adopted than LEZs. Second, De Borger, Glazer, and Proost (2022) compare the political support for road tolls and tradable driving permits, emphasizing the role of strategic behavior by drivers.6 Their model assumes that both permits and toll revenues are allocated to those that were driving in the previous period. They find that – similar to observed strategic behavior by firms prior to the introduction of pollution permits – anticipatory behavior after the policy is announced but prior to its introduction induces rational drivers to drive more than they otherwise would in an attempt to get more permits, or to obtain a larger share of the toll revenues. The authors show that the overall effect of such excessive driving may make both policies welfare-reducing. Moreover, it is found that drivers oppose the policies even when they receive all permits for free, or toll revenues are distributed to drivers only. Consequently, strategic behavior makes it more difficult to get a political majority to support both tolls and tradable permits. Interestingly, in an infinite horizon setting, tradable permits turn out to be superior to congestion tolls. They avoid strategic behavior once the system is implemented, whereas with congestion tolls the steady-state equilibrium implies continuing strategic behavior. One consequence is that it is easier to get a political majority for permits than for tolls: the authors find that drivers will always oppose congestion tolls, but they do support tradable permits as long as they receive a sufficient share of the permits for free.

The political economy of transport pricing and investment  131

7.3.2 The Long-Run: The Role of the Housing Market A number of studies have emphasized the role of the land and housing markets in determining the support of different groups of voters. Several of these studies are based on the Alonso–Muth–Mills “monocentric city” framework.7 In its simplest version, this assumes that all inhabitants of a city commute regularly from their residences to a central business district (CBD). Since commuting costs increase with distance, locations closer to the CBD tend to be more attractive, so that the unit price of land increases with proximity to the CBD. Individuals must therefore trade-off shorter commutes against higher land prices and, hence, smaller and/ or more expensive living spaces. In equilibrium, changes in commuting costs caused by transportation policy are capitalized in the price of land. For this reason, land ownership arrangements matter for the distributional implications of transport pricing policies and the ensuing divergence of interests among various groups. Borck and Wrede (2005) use the monocentric city model to study the political economy of commuting subsidies (or taxes). They consider a single commuting mode and two groups, the rich and the poor; the latter have lower labor income and own a smaller share of the city’s land. The model allows for two spatial income sorting patterns: (i) the “rich-in-city” case, where the rich live closer to the CBD than the poor (common in European Cities); (ii) the “poor-incity” case (common in North American cities), where the opposite holds. The fundamental insight from the analysis is that, since commuting subsidies reduce the price of land, even voters who do not commute at all (e.g., those at the edge of the CBD) may benefit from them, provided these voters do not own any land. Thus, absentee landownership tends to induce a political equilibrium with higher subsidies than socially optimal. However, if city residents are landowners, subsidies reduce their income. Residents who own significant amounts of land (the rich, in particular) may then oppose subsidies. Among others, this depends on their commuting distance. For example, when the landownership shares of the rich and the poor are similar, the rich tend to benefit from subsidies at the expense of the poor in the “poor-in-city” case, due to their longer commute. Borck and Wrede (2008) extend the analysis to consider an additional commuting mode and generate a rather rich set of spatial patterns, including the case where the urban poor commute by public transport and the suburban rich commute by car. In this scenario, the poor may favor subsidies for cars even if they do not receive them, because of the reduced pressure on land prices. Brueckner and Selod (2006) study how cities choose the combination of “speed and cost per mile” of the urban transport system. In their model, a faster transport system entails a higher monetary cost per mile than a slower one. Furthermore, the authors assume that, conditional on a given commuting cost, individuals prefer to live farther from the CBD (for example, this may be due to their preference for green space). Moreover, this preference is stronger the greater an individual’s hourly wage – a proxy for “skill”. If sufficiently intense, the preference for suburban locations results in a spatial equilibrium such that the hourly wage – and the value of travel time – of city residents increases with distance from the CBD (similar to the “poor-in-city” case discussed above). This pattern causes a political distortion in the choice of transport system. High-skill individuals have longer commutes and higher time costs for a given commuting distance. Therefore, the preference for a faster and more expensive transport mode increases at an increasing rate with individual skill level. Thus, since the median individual typically has lower than average skills (wage), the transport system chosen by majority voting is slower and less expensive than the welfare-maximizing one. This difference is even

132  Handbook on transport pricing and financing

more pronounced if residents, rather than absentee landowners, own the land the city is built upon. Overall, the models by Borck and Wrede (2005, 2008) and Brueckner and Selod (2006) suggest that cities tend to adopt transportation policies skewed toward lower monetary commuting costs and higher time costs than is socially optimal. Road pricing implies lower time costs and, typically, higher monetary expenditures, so this insight contributes to our understanding of the widespread opposition to this policy. But a caveat is in order. Despite the important insights they offer, these models were not developed to study road tolls. For this reason, they lack some important ingredients for a thorough political analysis of road tolls: for example, they ignore road congestion, and they assume that travel costs are linear functions of travel distance. These assumptions are at odds with the observation that, in its most common incarnation, urban road pricing takes the form of a cordon around a certain area (typically the center).8 Hence, car users pay the toll if and only if they enter this area, regardless of the distance they travel. De Borger and Russo (2018) tackle the above issues using a stylized version of a “threeisland” urban model (originally proposed by Brueckner, 2015). The three islands or zones capture the central city (encompassing the CBD), a mid-city area and the suburbs. These zones are connected by congested bridges, a central and a suburban one. The cordon toll is captured by a tax for crossing one of the bridges. This setup allows us to study several interrelated sources of friction between different groups of voters. The most important potential conflicts arise between those who live inside the cordon and those who do not (the former do not pay the toll, the latter do), and between homeowners and renters within the cordon (the former see their land/housing value increase due to the toll, the latter face higher rents). Moreover, the toll induces further interrelated conflicts depending on modal choices (car versus public transport users) and income (the poor versus the rich). The authors first consider a short-run situation where individuals’ location and land prices are exogenous. Unless residents within the cordon are the majority, the equilibrium toll resulting from a majority voting process is below the optimal level. Furthermore, while rich car commuters prefer a toll higher than socially optimal, the poor majority prefer a toll below the optimum unless the cost of accessing the public transit system is small. In a long-run setting where the toll capitalizes into land rents within the cordon, the authors show that only voters owning land in the center support it. These voters are typically a minority given the relatively small area encompassed by the cordon. Finally, if cities have some choice as to where to locate the cordon, the model shows that it is easier to get a political majority for a small cordon around the CBD rather than a cordon covering a more extensive area.9 7.3.3 Empirical Evidence The empirical literature on the political economy of road pricing is limited, most likely due to the small number of cities that have seriously attempted, let alone managed, to implement tolls. Arguably, this scarcity of “data points” makes it difficult to conduct quantitative analyses of the attitude of voters and/or policymakers toward this policy across cities. Nevertheless, the empirical literature has identified some important determinants of voter support for road pricing. Several authors have investigated the link between commuting costs and voting in cities that ran or are planning referenda on road tolls.10 Quigley and Hårsman (2010) match precinct

The political economy of transport pricing and investment  133

voting records to resident commute times and costs, and analyze patterns of voting in the 2006 referendum on the Stockholm congestion charge. Not surprisingly, they find that voting behavior was significantly influenced by the changes in commuting costs induced by the toll, and that travel time was a particularly important determinant of voters’ decisions. However, their findings also point to an important role of political ideology and party politics. Finally, the authors document a positive change in attitudes toward road pricing following its implementation, as predicted by some of the theoretical analyses described above. Boggio and Beria (2019) take a similar approach in the analysis of voting on the congestion charge in Milan, which was approved by a referendum at the end of 2011. Their findings point to a relation between acceptance of the toll and proximity of access to the public transport network, particularly for the underground. In a similar vein, and also using data from the Milan referendum, Percoco (2017) finds that heterogeneity in the cost of accessing alternative modes to the car is an important determinant of support for road pricing. Baranzini, Carattini, and Tesauro (2021) rely on a large survey of residents of the metropolitan area of Geneva to analyze the acceptability of road pricing, comparing multiple options differing in such dimensions as perimeter, price, pricing schedules, revenue use, and exemptions. The findings confirm the importance of location and access to public transport, but they also point to the role of exemptions from payment in gaining support from key groups, such as inner-city residents. The results also suggest, quite intuitively, that greater complexity of the scheme may increase efficiency, but it reduces acceptability unless voters are extensively informed. While policy experiments with road pricing in cities are quite rare, it is possible to investigate the response of voters to this policy by means of lab experiments. Janusch, Kroll, and Goemans (2021) propose such an experiment considering a setting with two routes and with referenda before and after the implementation of a toll. The authors implement two alternative treatments: the allocation of toll revenues and the level of information before voting. Overall, the findings confirm two important insights from the theoretical literature: uncertainty is associated with more widespread opposition to road pricing, and the use of toll revenues is key to achieving acceptability. Lastly, Fageda, Flores-Fillol, and Theilen (2021) empirically document the prevalence of quantity measures (such as low-emission zones) over tolls. Moreover, they formally estimate a relation between the existence of a LEZ and several explanatory variables in a sample of cities with populations exceeding 300,000. They find that LEZs are often observed in polluted and high-income cities.

7.4 THE POLITICAL ECONOMY OF PRICING PARKING Parking is an important component of urban traffic policy. Every car trip begins and ends with parking, so the time and monetary costs associated with parking can affect mode choice and the time of travel. In addition, parking occupies a great amount of space in urban areas (Jakle and Sculle, 2004), with important implications for mobility as well as urban form (Brueckner and Franco, 2017, 2018). Inci (2015) provides a survey of the literature on the economics of parking. Moreover, Chapter 5 of this Handbook deals in detail with the normative theory of parking pricing. The literature referred to above has identified several well-documented distortions, including the suboptimal pricing of curbside parking space on congested roads, the inefficient

134  Handbook on transport pricing and financing

allocation of residential parking permits and the provision of subsidized workplace parking by employers. Broadly speaking, the literature has focused primarily on evaluating the welfare losses caused by these distortions and proposing policy corrections. Arguably, the fact that many urban governments do not correct these obvious inefficiencies may be due to the political decision process. Despite this observation, surprisingly few analytical studies have considered the political economy of parking policy.11 The paper on road pricing by Russo (2013), discussed in Section 7.3 above, provides a political economy analysis of curbside parking fees, based on a majority voting framework. In fact, to the best of our knowledge, curbside parking pricing is the only aspect of parking policy studied in formal political economy analyses. Economic theory indicates that curbside prices should be set high enough to eliminate cruising for parking, while ensuring that available capacity is almost fully utilized (Arnott and Inci, 2006). Parking tariffs seem to exceed the socially optimal level on the most popular streets of some cities (e.g., Amsterdam, London), but cheap (or free) curbside parking on busy streets is quite common in many other cities, especially in North America (Shoup, 2005). To provide an explanation for these stylized facts, De Borger and Russo (2017) focus on the role of special interests such as retailers and downtown residents, embedding a simple spatial model in the “protection for sale” framework of Grossman and Helpman (1994). The authors assume that parking tariffs discourage visits by shoppers to the downtown commercial area, and that competing suburban (“big box”) retailers sell goods at a lower margin than downtown stores. In equilibrium, downtown retailers have strong incentives to lobby against parking charges. While suburban retailers would in principle benefit from higher parking tariffs downtown, they have weak incentives to lobby because of the low margin on additional sales. Furthermore, downtown residents may join forces with local retailers and lobby against high parking fees for shoppers coming from outside, to avoid the closure of local shops and associated negative externalities, e.g., urban blight. These contributions are a first step in the study of the political economy of parking policy but ignore some important aspects, such as cruising costs. Although anecdotal evidence of the opposition to parking fees by downtown retailers is abundant, it is not obvious that higher parking tariffs reduce footfall in busy commercial areas if one accounts for search and cruising costs.12 It would therefore be desirable to incorporate such costs in future political economy analyses of curbside parking in commercial areas. Furthermore, there are other instances of parking policy where political considerations seem relevant. An example is the allocation of resident parking permits in the proximity of busy commercial streets. Not rarely, do residents receive these permits at a price not only much lower than the value that outside visitors would be willing to pay, but also lower than residents’ own willingness to pay (van Ommeren et al., 2011). Most likely, this allocation of curbside parking capacity is inefficient.13 However, local governments seem reluctant to eliminate these implicit subsidies to residents, probably for fear of alienating them.

7.5 THE POLITICAL ECONOMY OF INFRASTRUCTURE PROVISION Transport infrastructure absorbs a large share of the public budget. Unsurprisingly, therefore, the economics literature has devoted quite some attention to the funding and financing of

The political economy of transport pricing and investment  135

transport capacity investments (also see Chapters 14 and 15 of this volume). However, the provision of transport infrastructure is often the subject of political controversy. The literature has considered this issue from several angles. For the most part, it has focused on the observation that the benefits and costs of investment in infrastructure (roads, railroad lines, stations, airports) are typically not limited to the area where the investment takes place, but they generate spill-overs to several other regions. Accordingly, in a democracy the provision and financing of large-scale infrastructure entails bargaining among representatives of such regions; as argued above, this may produce distortions with respect to the welfare-maximizing policy. Several authors have proposed theoretical analyses of this decision-making process, based on models of legislative bargaining, discussed in Section 7.2. Some exclusively focus on infrastructure, while others combine infrastructure provision with user pricing. A few empirical papers are also available that used data from the allocation of infrastructure funding in legislatures to test the predictions of such models. In what follows, we first summarize the main theoretical arguments, and then turn to the (limited) available empirical evidence. 7.5.1 The Political Economy of Infrastructure Spending: Theory Knight (2004) developed a legislative bargaining model to study the level of overall spending on infrastructure as well as the allocation of funds across districts. The model considers a federation consisting of a given number of districts. Each district elects one local representative to a federal legislative body (say, parliament). To finance spending on infrastructure, the model assumes equal tax sharing by all districts. Decisions in the legislature are taken as follows. An agenda setter is randomly assigned and prepares a proposal that includes an overall spending level as well as the allocation of infrastructure over the districts. If the proposal is accepted by a majority, it is executed; if it is rejected, a new agenda setter is randomly assigned. To formulate the proposal, the model assumes that the agenda setter maximizes the utility of inhabitants of his constituency subject to the constraint of having a minimum winning coalition for the proposal made. In other words, the agenda setter has to give up some funds for his own district and to allocate sufficient funds to a number of other districts so as to make sure a majority will support his proposal. The paper produces at least three intuitive results. First, it shows that the decision-making process described leads to overall excessive spending on infrastructure. Second, members of the winning coalition are treated much more favorably than non-members. Third, the probability that an individual representative votes in favor of the proposed allocation strongly and positively depends on the allocated funds to his own district, and negatively on the tax cost for the district. The politics of local (or regional) versus central decision-making on infrastructural choices was studied in detail by Xie and Levinson (2009). They consider a two-level (central, local) federation in the provision of road infrastructure that connects different jurisdictions. Unsurprisingly, they first show that residents’ preferences for public spending on road infrastructure strongly depend on the origin and destination of the trips they want to make. At each governmental level (central, local), they then assume decisions are made within the framework of a citizen–candidate setting. The authors develop a two-stage imperfect information game to study the choice between centralized or decentralized spending in a representative democracy, together with the resulting spending decisions. They find that the choice between a centralized and decentralized spending structure depends on road conditions and spill-over

136  Handbook on transport pricing and financing

effects, assuming the latter is exogenous. Variations in spill-overs may therefore imply endogenous changes in governance choice. Although the above models have relevant implications for transport policy, they lack some important ingredients of actual transport decision-making. Exclusively emphasizing spending decisions (on infrastructure), they ignore congestion and the issue of pricing the infrastructure made available. De Borger and Proost (2016a) consider the political economy of joint pricing and investment decisions in countries where decisions can be taken both at a central (national) and local (regional) level. The model assumes a fixed population of users and non-users of the road infrastructure. It allows for spill-overs between regions (residents of a given region make trips to other regions) and it explicitly captures road congestion. Moreover, it allows for demand heterogeneity between regions (in terms of the share of “foreign” users on a region’s road system) as well as within regions (in terms of the share of voters that are road users). It is assumed that regional decisions are taken by majority voting; national decisions are taken by a minimum winning coalition in a legislature of regionally elected representatives. Toll revenues are redistributed to drivers and non-drivers. In a two-region setting, the authors show that when road users form the majority in at least one region, decentralized regional decisionmaking performs better than national decision-making as long as spill-overs are not too large. National decisions may yield higher welfare than decentralization only if road users have a large voting majority and the infrastructure in a given region is intensively used by both local and “outside” users. If non-users form a majority in both regions, centralized and decentralized decision-making yield the same result, but the outcome is socially very undesirable: the majority of voters (that does not drive) will “exploit” the minority of road users, and impose road user prices that are much higher than the socially optimal values. It is further shown that the performance of decentralized supply decisions strongly improves if the national level imposes self-financing rules on regions. Intuitively, this limits the possibility of imposing excessively high charges on road use; it therefore protects users of the infrastructure from being exploited by non-users. In a follow-up paper, De Borger and Proost (2016b) study to what extent imposing institutional constraints on the decision-making process (for example, the requirement of uniform user charges in all regions, the requirement to decide at the national level via bargaining between the different regions) affect the relative performance of regional versus national decision-making. They find that both requiring user prices to be uniform across regions and negotiated decisions at the national level greatly improves the efficiency of centralized decision-making. Federal decisions may easily outperform decentralization, even when the opposite would hold in the absence of the constraints. The model also finds that if regions are symmetric and drivers have a majority in both regions, they will voluntarily transfer power to the federal level, provided the relevant policy restrictions (uniform pricing or legislative bargaining) are constitutionally imposed. However, if drivers have a majority in one region only, the region where non-users have a majority will never agree to transfer decision power to the federal level. The models just described look at road infrastructure and road user charges, but a similar analysis can also be used to analyze decisions of how much to invest in public transport and what fare to charge to users. De Borger and Proost (2015) consider again a hypothetical tworegion federation, where each region has two types of inhabitants: some voters do not own a car and use only public transport, but others (car owners) demand both public transport and car trips. The composition of the voting population in terms of the fraction of car users may

The political economy of transport pricing and investment  137

differ between regions, and public transport travelers in any region include both local people and users from other regions. The user cost of a car trip is assumed to be exogenously given. The authors show that political decisions may lead to very low public transport fares, even if car owners are a large majority of the population. The cost recovery ratio of the public transport system always improves with more outside users. Compared to centralized decision-making, decentralized decision-making leads to higher fares and better cost recovery. Although there are many other reasons for low public transport fares (for example, income distributional concerns), the results may partly explain the lack of opposition to very large public transport subsidies in many countries. Gleaser and Ponzetto (2018) are interested in explaining the evolution of infrastructure spending in the United States during the twentieth century, ranging from a period of overspending on megaprojects in the inter- and postwar years to a period of underspending due to increased opposition of residents (“not in my back yard (NIMBY)” behavior). The authors emphasize the influence of what they call “voter attention” on policy: not all voters are well informed, and the fraction of voters that does pay attention to proposed policies is likely to depend on education and income. In Glaeser and Ponzetto’s model, transport infrastructure within a city benefits not only city residents but also suburban ones; in addition, there are spillovers to the rest of the country. Urban infrastructure projects are funded partly by a tax on urban residents, and partly by a national tax. The infrastructure also has a local inconvenience cost for residents (pollution, noise, etc.). The election process is described by a probabilistic voting model whereby voters are imperfectly informed about the investment proposal. Voter attention then affects transport investment decisions in several ways. First, national funding of infrastructure projects implies that urban residents devote too little attention to the projects’ financial costs, since the tax bill is less salient to residents than the benefits of the infrastructure. By itself, this aspect would cause spending on infrastructure to exceed the socially optimal level. Second, however, the highly localized disadvantages of new infrastructure are salient to voters living nearby. Given that urban residents became better educated and better organized over time, the salience of such disadvantages may have resulted in increasingly stiff opposition to infrastructure projects.14 Finally, the authors incorporate the possibility of adding nuisance abatement investment to the model (e.g., placing infrastructure underground), in response to local opposition. The increasing salience of nuisances to informed voters causes excessive abatement and raises the cost of infrastructure projects, resulting in fewer projects being built in equilibrium. Finally, one study focuses on the interaction between investing in infrastructure and tolling. Westin et al. (2016) study the decision to build a bypass road and the introduction of a toll. Their model of the political decision process incorporates interactions between voters, citizen interest groups, and politicians in a legislative assembly representing different local constituencies (e.g., the central city and surrounding areas), based on the agenda-setter model. Their main conclusion is that, while efficient road pricing can emerge as an equilibrium of the political decision process, there are many other inefficient possible equilibria, each coinciding with the policy preferred by a specific group. 7.5.2 Empirical Results We close this section with a brief review of empirical studies. Some of these have used data from transport infrastructure projects to test some of the basic predictions of political

138  Handbook on transport pricing and financing

economy models. For example, Knight (2004) empirically tests the main predictions derived from his model (see above). Using data on the voting behavior of district representatives on US infrastructure projects, he estimates a probit model that relates the probability of voting in favor of a series of proposals to the funds allocated to their own district, the tax cost of the proposal to the district, and a series of control variables (area, industry structure, etc.). He finds compelling evidence that representatives push the approval of transportation projects in their districts. Moreover, their district’s tax cost reduces representatives’ support for a proposed investment allocation. Lastly, the results confirm that the decision-making process described leads to overall excessive spending on infrastructure. In a related paper, Knight (2005) estimates the value of sitting on legislative committees that formulate proposals on the projects to be included in the transportation budget. His results confirm the predictions of the legislative bargaining model by Baron and Ferejohn (1989), showing that congressional members sitting on the committee secure higher spending allocated to their own district. More recently, Halse (2016) uses data from road construction and maintenance in Norway to show that elected representatives from small and scarcely populated geographic areas have greater incentives to spend on local transport projects than representatives from larger regions. The reason is that small regions pay a smaller share of the cost of such projects. Bilotkach (2018) further confirms this same type of behavior in an empirical analysis that focuses on the potential role of political influence in allocating aviation infrastructure investment in the United States. He finds that states that have members in the Transportation and Infrastructure Committee received substantially more investment funds. Some economists have investigated the role of pure partisan politics in the allocation of infrastructure investment. Albouy (2013) uses data from infrastructure funding by the US Congress, showing that states whose representatives belong to the majoritarian party in the assembly receive greater federal grants, particularly in the case of transportation infrastructure. Along the same lines, Bilotkach (2018) finds that airports located in states carried by a republican at the latest senate election not only showed a higher probability of obtaining funding, but they also obtained larger amounts. Rather than looking at the amount of allocated funding, Selod and Soumahoro (2019) study the geographical allocation of roads in Mexico, using geo-referenced data. Their main finding is that municipalities that voted for the president’s party in legislative races get rewarded with more federally-funded highways, particularly at times when the president has no majority in the legislature. This finding suggests that the central government biases the allocation of infrastructure in an effort to control the Congress. Lastly, although the empirical evidence they present is sketchy and based on aggregate and simple econometric relations, Glaeser and Ponzetto (2018) produce illustrative reduced-form evidence in support of some of their theoretical predictions. If the urban voter is attentive, the salience of nuisances associated with transport infrastructure implies that one expects investment to be targeted more to areas with lower population density. They indeed find that per capita spending on road projects is negatively related to population density. Moreover, assuming that education stimulates political engagement, the authors expect large-scale projects to experience more opposition in areas where people are highly educated; for example, highways will be more concentrated in less educated areas. At the same time, there will be more spending on abatement in areas with a well-educated population. In effect, they find empirical support for this statement: spending per mile rises in education.

The political economy of transport pricing and investment  139

7.6 POLITICAL DETERMINANTS OF THE ENFORCEMENT OF TRAFFIC LAW The enforcement of traffic law – which is typically entrusted to local governments – is an important element of transportation policy. Governments can choose how strictly they enforce certain policies, and exercise discretion regarding which groups are subject to stricter enforcement. To our knowledge, the political economy literature investigating these aspects is purely empirical. This has sought to explain the extent of enforcement with factors that determine the incentives of local politicians, taking stock of the observation that enforcement entails monetary penalties (e.g., speeding or parking tickets) to individuals who may also be voters. These factors include budgetary restraints and the approaching of elections. The literature has also analyzed the welfare effects of enforcement, considering the effects on individual behavior and the frequency and severity of accidents. Garrett and Wagner (2009) use county-level data on traffic enforcement in North Carolina to show that local governments are more likely to issue traffic tickets in the years following fiscal shortfalls. In a similar vein, Makovsky and Stratman (2009) study the issuance of speeding tickets in Massachusetts, providing evidence that local police officers tend to issue more tickets when their municipality faces stricter constraints on other revenue sources, such as property taxes. Furthermore, conditional on speeding, officers are more likely to issue tickets to drivers who are from out of town, particularly if they reside very far away. The authors interpret these findings as evidence of tax exporting (stricter enforcement on drivers who do not vote in local elections) as well as of officers minimizing the risk of having their ticket contested, since the cost of filing an appeal tends to be greater for those who live farther away from the municipality that issued a ticket. To corroborate these findings, the authors also show that state patrol officers (who are not employees of local municipalities) are less likely to show the same bias in enforcement. In a follow-up study, Makowsky and Stratmann (2011) study how enforcement affects the behavior of drivers and, ultimately, the likelihood of road accidents. Following the logic of their previous study, the authors show that traffic tickets are issued more frequently by municipalities facing budgetary shortfalls. In addition, the authors use such shortfalls as instrumental variables for the intensity of enforcement. The results indicate that enforcement causes a significant reduction in the frequency and severity of accidents. More recently, the literature has examined the link between enforcement and the political budget cycle, based on the hypothesis that, to avoid alienating voters, incumbent politicians tend to relax the enforcement of traffic policy as elections get closer. Bracco (2018) uses data from Italian municipalities to show that incumbent mayors who are up for re-election tend to issue fewer tickets as elections approach. The author also shows that the collection of issued fines is weaker in electoral years, which is consistent with the existence of a political budget cycle in enforcement. Bertoli and Grembi (2021) also use data from Italy but focus on the effect of the political cycle on the frequency and severity of traffic accidents. Their findings confirm the existence of an electoral cycle in traffic enforcement and show a significant effect on traffic accidents, in line with earlier findings by Makowsky and Stratmann (2011).

140  Handbook on transport pricing and financing

7.7 CONCLUSIONS AND SUGGESTIONS FOR FURTHER RESEARCH This chapter discussed a number of political economy models that have contributed to a better understanding of transport policy-making. For example, the models suggest that voting outcomes on the introduction of road tolls depend on individual and political uncertainty, on the income distribution, and on the use of toll revenues. Voting on the introduction of cordon charges in urban areas is also strongly affected by idiosyncratic modal preferences, residential location, and the degree of homeownership at different locations. Together, the various models convincingly explain the widely observed opposition to tolls and cordon charges, and the importance of using part of the revenues for public transport (subsidies or investment) to build support for these policies. The models described in this chapter also suggest why low-emission zones and parking fees are politically easier to introduce than congestion charges. The literature further shows how the democratic bargaining process between elected regional representatives leads to excessive investment, and how coalition formation implies a positive bias toward regions politically represented in the coalition. Moreover, it emphasizes the role of information and voter attention. Of course, there are some obvious gaps in the political economy literature on transport pricing and investment. For example, the survey shows that there is very little literature on the pure party politics underlying the overall development of the transport system. To what extent are cities’ choices related to pure ideology, party politics, and coalitions in charge? For example, green parties may emphasize “green” policies, socialist parties may focus on transport policies that redistribute from rich to poor, and center-right parties may typically favor deregulating obstructions to market functioning. The role of political ideologies has insufficiently been studied in the literature. Moreover, no literature is available on transport policy-making by non-democratic regimes. Another interesting area for future work is the analysis of potential political frictions that are likely to arise as transportation systems adapt to new technologies, such as autonomous and electric vehicles. As these new vehicle types replace traditional ones, the impact on the public budget is likely to be significant (Adler, Peer and Sinozic, 2019). For instance, in many countries the taxation of fossil fuels generates much revenue for the national government, and the diffusion of electric vehicles will gradually erode this tax base. On top of this, autonomous vehicles are expected to reduce the demand for parking, and thus potentially erode parking fees as a source of revenue as well for local governments. Furthermore, the diffusion of autonomous and electric vehicles will require massive investment in infrastructure, including the deployment of charging stations and of devices supporting automated driving. Finally, new concepts of travel such as “mobility as a service” (MaaS) are expected to reduce private ownership of vehicles and thus affect the tax revenue from ownership and registration taxes. Conceivably, these changes will require a redesigning of fiscal policy applied to the transport sector, with less reliance on fuel taxation and greater reliance on alternative financing tools, including road tolls. However, as we discussed in Section 7.3, tolls have so far been quite controversial and unpalatable to voters. Will the elimination of traditional forms of vehicle taxation make tolls more acceptable? Furthermore, road pricing in urban areas is typically administered by local or regional governments. Can we expect these governments to implement socially optimal policies without coordination from the national government? And if tolls are to replace fuel and ownership taxes, what share of the revenue should be transferred to the national government? Finally, will possible lags in the adoption and availability of new technologies among

The political economy of transport pricing and investment  141

various groups (e.g., rich and poor, urban and rural voters) generate political frictions and policy distortions of the kind we discussed above? These are just some of the open questions that the diffusion of new transportation technologies will bring to the table in the coming years.

NOTES 1.

2. 3. 4. 5. 6.

7. 8. 9.

10. 11. 12. 13. 14.

For some relevant references out of a huge literature on optimal pricing, see Vickrey (1969), Small (1982), Arnott, de Palma and Lindsey (1993), Parry and Bento (2002), and Verhoef (2002). On optimal investment decisions see, among many others, Winston (1991) and Mayeres and Proost (1997). Note that Chapter 8 in this Handbook covers the standard congestion pricing literature. See Anas and Lindsey (2011) for a careful discussion of the results of three such cases (London, Milan, and Stockholm). See also Lockwood (2002) for a related model of bargaining on fiscal policy at the legislative level. For instance, the city of London earmarked a share of the revenue from the Congestion Charge to fund the expansion of the public bus fleet. This result is derived under the assumption that all potential transport users would drive if the generalized price of car use were zero. This assumption is sufficient but not necessary for the result. Although their conclusion was not based on a formal political economy analysis, it had been argued by several authors that it may be easier to get political support for tradable driving rights than for congestion tolls (Fan and Jiang, 2013; Raux and Souche, 2003). The intuition is clear. Suppose some permits are provided for free to drivers, then a permit system implies that voters who continue to drive incur a smaller monetary cost than when a toll is imposed. Moreover, initial drivers that give up their trip because of the introduction of the system will get an immediate reward. This is not the case under congestion tolls: drivers will have to wait until the toll revenues are redistributed or used to improve road infrastructure or public transport. See Brueckner (2011) for an introductory overview of the monocentric city framework. This is the case, e.g., of existing congestion charges in London, Stockholm, and Milan, as well as recently rejected schemes in Edinburgh, Manchester, Brussels, and New York. Remarkably, London recently extended the area covered by the congestion charge. An interesting question for further research is whether from a political point of view it is desirable to first get political support for congestion pricing by introducing the policy for a small area and then, in a later stage – when sufficient support has been achieved – extend the area covered by pricing. Also note that this may be consistent with the theory reference-dependent preferences proposed by the behavioral economics literature. An overview of some of these and other studies on voters’ attitudes toward road pricing is provided by Hensher and Li (2013). See Russo (2020) for a more comprehensive survey of the literature on the political economy of parking policy. Hymel (2014) provides evidence that increases in parking meter rates reduce visits to the store when saturation rates are low but increase visits when saturation rates are high. van Ommeren et  al. (2014) show that residential permits are responsible for a 15% increase in parking provision costs in commercial areas in the Netherlands, on average, with a social loss of roughly EUR 275 per permit per year. An example of such sensitivity is the “Freeway Revolt” movement in US cities in the 1960s and 1970s. See Brinkman and Lin (2019) for an analysis of the impact of this movement on infrastructure policy.

REFERENCES Adler, M.W., S. Peer, and T. Sinozic (2019). Autonomous, connected, electric shared vehicles (ACES) and public finance: An explorative analysis. Transportation Research Interdisciplinary Perspectives 2, 100038.

142  Handbook on transport pricing and financing

Albouy, D. (2013). Partisan representation in congress and the geographic distribution of federal funds. The Review of Economics and Statistics 95, 127–141. Anas, A. and R. Lindsey (2011). Reducing urban road transportation externalities: Road pricing in theory and in practice. Review of Environmental Economics and Policy 5(1), 66–88. Arnott, R., A. De Palma, and R. Lindsey (1993). A structural model of peak period congestion: A traffic bottleneck with elastic demand. American Economic Review 83, 161–179. Arnott, R. and E. Inci (2006). An integrated model of downtown parking and traffic congestion. Journal of Urban Economics 60, 418–442. Baranzini, A., S. Carattini, and L. Tesauro (2021). Designing Effective and Acceptable Road Pricing Schemes: Evidence from the Geneva Congestion Charge. Environmental and Resource Economics 79, 417–482. Baron, D., and J. Ferejohn, (1989). Bargaining in legislatures. The American Political Science Review 83(4), 1181–1206. Bernheim, B. and M. Whinston (1986). Common agency. Econometrica 54(4), 923–942. Bertoli, P. and V. Grembi (2021). The political cycle of road traffic accidents. Journal of Health Economics 76, 102435. Besley, T. and S. Coate (1997). An economic model of representative democracy. Quarterly Journal of Economics 65–95. Besley, T. and S. Coate (2003). Centralized versus decentralized provision of local public goods: A political economy approach. Journal of Public Economics 87, 2611–2637. Bilotkach, V. (2018). Political economy of infrastructure investment: Evidence from the economic stimulus airport grants. Economics of Transportation, Special Issue on Political Economy. Boggio, M. and P. Beria (2019). The role of transport supply in the acceptability of pollution charge extension: The case of Milan. Transportation Research Part A 129, 92–106. Borck, R. and M. Wrede (2005). Political economy of commuting subsidies. Journal of Urban Economics 57, 478–499. Borck, R. and M. Wrede (2008). Commuting subsidies with two transport modes. Journal of Urban Economics 63, 841–848. Bracco, E. (2018). A fine collection: The political budget cycle of traffic enforcement. Economics Letters 164, 117–120. Brinkman, J. and J. Lin (2019). Freeway Revolts! Federal Reserve Bank of Philadelphia Discussion Paper. Brueckner, J.K. (2011). Lectures in Urban Economics, Cambridge, MA: MIT Press. Brueckner, J.K. (2015). Cordon tolling in a city with congested bridges. Economics of Transportation 3, 235–242. Brueckner, J.K. and S. Franco (2017). Parking and urban form. Journal of Economic Geography 17 (1), 95–127. Brueckner, J.K. and S. Franco (2018). Employer-paid parking, mode choice, and suburbanization. Journal of Urban Economics 104, 35–46. Brueckner, J.K. and H. Selod (2006). The political economy of urban transport system choice. Journal of Public Economics 90, 983–1005. Coughlin, P. (1992). Probabilistic Voting Theory. New York: Cambridge University Press. De Borger, B. and A. Glazer (2017). Support and opposition to a Pigouvian tax: Road pricing with reference-dependent preferences. Journal of Urban Economics 99, 31–47. De Borger, B., A. Glazer, and S. Proost (forthcoming 2022). Strategic behavior under tradable driving permits and congestion tolls: A political economy analysis. Journal of Urban Economics. De Borger, B. and S. Proost (2012). The political economy of road pricing. Journal of Urban Economics 71, 79–92. De Borger, B. and S. Proost (2015). The political economy of public transport pricing and supply decisions, Economics of Transportation 4(1–2), 95–109. De Borger, B. and S. Proost (2016a). The political economy of pricing and capacity decisions for congestible local public goods in a federal state, International Tax and Public Finance 23, 934–959. De Borger, B. and S. Proost (2016b). Should we leave road pricing to the regions? The role of institutional constraints, Regional Science and Urban Economics 60, 208–222.

The political economy of transport pricing and investment  143

De Borger, B. and A. Russo (2017). The political economy of pricing car access to downtown commercial districts. Transportation Research Part B 98, 76–93. De Borger, B. and A. Russo (2018). The political economy of cordon tolls. Journal of Urban Economics 105, 133–148. Downs, Anthony (1957). An Economic Theory of Democracy. New York: Harper. Drazen, A. (2008). Political budget cycles. In S.N. Durlauf and L.E. Blume, eds., The New Palgrave Dictionary of Economics (2nd edition), London and New York: Palgrave Macmillan. Fageda, X., Flores-Fillol, R. and B. Theilen (2021). Price versus quantity measures to deal with pollution and congestion in urban areas: A political economy approach. Working Paper, University of Barcelona. Fan, W. and X. Jiang (2013). Tradable mobility permits in roadway capacity allocation: Review and appraisal. Transport Policy 30, 132–142. Fernandez, R. and D. Rodrik (1991). Resistance to reform: Status quo bias in the Presence of individualspecific uncertainty. American Economic Review 81(5), 1146–1155. Garrett, T. and G. Wagner (2009). Red ink in the rearview mirror: Local fiscal conditions and the issuance of traffic tickets. The Journal of Law and Economics 52(1), 71–90. Glaeser, E.L. and G.A.M. Ponzetto (2018). The political economy of transportation investment. Economics of Transportation 13, 4–26. Goodwin, P. (1994). Road pricing or transport planning? In: Johansson, B., Mattsson, L. (Eds.), Road Pricing: Theory, Empirical Assessment and Policy. Kluwer Academic Publishers, 143–158. Grossman, G. and E. Helpman (1994). Protection for sale. The American Economic Review 84(4), 833–850. Halse, A.H. (2016). More for everyone: The effect of local interests on spending on infrastructure. European Journal of Political Economy 43, 41–56. Hensher, D.A. and Z. Li (2013). Referendum voting in road pricing reform: A review of the evidence. Transport Policy 25, 186–197. Hymel, K. (2014). Do parking tariffs affect retail sales? Evidence from Starbucks. Economics of Transportation 3, 221–233. Inci, E. (2015). A review of the economics of parking. Economics of Transportation 4, 50–63. Jakle, J. and K. Sculle (2004). Lots of Parking: Land Use in a Car Culture. Charlottesville, VA: University of Virginia Press. Janusch, N., S. Kroll, C. Goemans, T.L. Cherry, and S. Kallbekken, (2021). Learning to accept welfareenhancing policies: An experimental investigation of congestion pricing. Experimental Economics 24, 59–86. Kahneman, D. and A. Tversky (1979). Prospect theory: An analysis of decision under risk. Econometrica 47, 263–291. King, D., M. Manville, and D. Shoup (2007). The political calculus of congestion pricing, Transport Policy 14, 111–123. Knight, B. (2004). Parochial interests and the centralized provision of local public goods: Evidence from congressional voting on transportation projects. Journal of Public Economics 88, 845–866. Knight, B. (2005). Estimating the value of proposal power. American Economic Review 95, 1639–1652. Lindbeck, A. and J. Weibull (1987). Balanced-budget redistribution as the outcome of political competition. Public Choice 52, 273–297. Lockwood, B. (2002). Distributive politics and the benefits of decentralization. Review of Economic Studies 69(2), 313–338. Makowsky, M. and T. Stratmann (2009). Political economy at any speed: What determines traffic citations? The American Economic Review 99(1), 509–527. Makowsky, M. and T. Stratmann (2011). More tickets, fewer accidents: How cash-strapped towns make for safer roads. The Journal of Law and Economics 54(4), 863–888. Marcucci, E., M. Marini, and D. Ticchi (2005). Road pricing as a citizen-candidate game. European Transport 31, 28–45. Marquis de Condorcet (1785). Essai sur l’application de l’analyse à la probabilité des décisions rendues à la pluralité des voix. Mayeres, I. and S. Proost (1997). Optimal tax and public investment rules for congestion-type externalities. Scandinavian Journal of Economics 99, 261–277.

144  Handbook on transport pricing and financing

Oates, W. (1999). An essay on fiscal federalism. Journal of Economic Literature 37(3), 1120–1149. Osborne, M. and A. Slivinski (1996). A model of political competition with citizen candidates. Quarterly Journal of Economics, 111, 85–114. Owen, R., A. Sweeting, S. Clegg, C. Musselwhite, and G. Lyons (2008). Public Acceptability of Road Pricing. Final Report for Department for Transport, London. Parry, I. and A. Bento (2002). Estimating the welfare gains of congestion taxes: The critical importance of other distortions within the transport system. Journal of Urban Economics 51, 339–365. Percoco, M. (2017). Cost distribution and the acceptability of road pricing: Evidence from Milan’s referendum. Journal of Transport Economics and Policy 51(1), 34–46. Persson, T. and G. Tabellini (2000). Political Economics: Explaining Economic Policy. Cambridge: MIT Press. Pigou, A. (1920). The Economics of Welfare. London: Palgrave Macmillan. Quigley, J. and B. Hårsman (2010). Political and public acceptability of congestion pricing: Ideology and self-interest. Journal of Policy Analysis and Management 29(4), 854–874. Raux, C. and S. Souche (2003). An analytical framework of pricing acceptability: Application to four case studies. In: Schade, J. and Schlag, B. (Eds.), Acceptability of Transport Pricing Strategies. Oxford: Elsevier, 153–168. Ren, T. and H.-J. Huang (2020). A competitive system with transit and highway: Revisiting the political feasibility of road pricing. Transport Policy 88, 42–56. Russo, A. (2013). Voting on road congestion policy. Regional Science and Urban Economics 43(5), 707–724. Russo, A. (2020). Political economy of parking policy. In Albalate, A. and Gragera, A. eds., Parking Regulation and Management. London: Routledge. Selod, H. and S. Soumahoro (2019). Highway politics in a divided government: Evidence from Mexico. Policy Research Working Paper Series 8710, The World Bank. Shoup, D. (2005), The High Cost of Free Parking, Planners Press, American Planning Association. Small, K. (1982). The scheduling of consumer activities: Work trips. American Economic Review 72, 467–479. Small, K. (1992). Using the revenues from congestion pricing. Transportation 19, 359–381. van Ommeren, J., J. de Groote, and G. Mingardo (2014). Residential parking permits and parking supply. Regional Science and Urban Economics 45, 33–44. van Ommeren, J., D. Wentink, and J. Dekkers (2011). The real price of parking policy. Journal of Urban Economics 70, 25–31. Verhoef, E. (2002). Second-best congestion pricing in general static transportation networks with elastic demands. Regional Science and Urban Economics 32, 281–310. Vickrey, W. (1969). Congestion theory and transport investment. American Economic Review (Papers and Proceedings) 59, 251–260. Westin, J., J.P. Franklin, S. Proost, P. Basck, and C. Raux (2016). Achieving political acceptability for new transport infrastructure in congested urban regions. Transportation Research Part A: Policy and Practice 88, 286–303. Winston, C. (1991). Efficient transportation infrastructure policies. Journal of Economic Perspectives 5, 113–128. Xie, F. and D. Levinson (2009). Governance choice on a serial network. Public Choice 141, 189–212.

PART II TRANSPORT MODES

8. Road pricing and provision of capacity Se-il Mun and Daisuke Fukuda

8.1 INTRODUCTION This chapter reviews recent developments in research on road pricing and provision of capacity. There are a number of review articles on this topic. Among them, Lindsey (2012) provides an excellent review of the literature on the theory of congestion pricing and the relationship between optimal congestion tolls and optimal road capacity. De Palma and Lindsey (2011) review technological advancements, in conjunction with pricing methodologies. Issues of complications for first-best pricing and coping methodologies for second-best pricing are also discussed by Fosgerau and Van Dender (2013). We updated the earlier reviews by including various developments that have occurred during these past 10–20 years. However, it is beyond our capacity to cover all relevant topics, given the large body of literature on road pricing that numerous research outputs are being produced continuously. Therefore, we focus on the developments in second-best pricing, modeling dynamic congestion, provision of capacities, and empirical evidence. For a discussion of first-best pricing principles, we refer to Chapter 2 of this Handbook, while Chapter 3 of this Handbook is concerned with pricing for other objectives than social welfare. We have seen remarkable growth in the contributions of transportation engineers to congestion pricing since the late 1990s. They developed models for the design of second-best pricing in networks and new models of dynamic congestion, in particular, those in line with macroscopic fundamental diagram (MFD) or bathtub. Economists have also presented new empirical findings based on innovative data collection methods combined with advanced econometric techniques.

8.2 SECOND-BEST ROAD PRICING IN STATIC MODELS Traffic congestion is a result of multidimensional decisions made by trip makers: whether to make trips, where to visit, mode, route, and timing. Trips are generated from many locations; therefore, congestion levels vary across space. The first-best optimal road pricing requires tolling at all links in the road network, and the toll level should vary with the location and time of day, depending on the level of congestion. The implementation of the firstbest policy is infeasible; thus, road pricing in practice will be the second-best; for example in that only part of the network is tolled, or variations of toll levels by location or time are limited. Implementation of such second-best pricing is in some respects simpler, but it is not an easy task for traffic control authorities to design the system, which includes the choice of the toll level and location of toll collection. Second-best pricing should take into account not only the direct effects on the tolled road but also the indirect effects on the road network as a whole. 146

Road pricing and provision of capacity  147

We first describe the second-best pricing problem for simple networks to gain insight into how the indirect effects stated above are controlled. Then, we review recent developments in two types of second-best pricing in practice: cordon pricing and value pricing. We further discuss more sophisticated pricing systems supported by advances in information technology. 8.2.1 Second-Best Pricing in a Road Network The majority of recent research on second-best pricing is based on the network setting. In general, a road network consists of nodes and links, and the trip demand originates from one node and is destined for another node. Path (route) flows, link flows, and origin-to-destination (O–D) flows are distinguished. We provide economic insights into second-best pricing using two examples of simple networks: two routes in parallel and two links in series. Example 8.1: Two routes in parallel This is a classic problem that was first discussed by Marchand (1968) and later elaborated by many researchers such as Verhoef et al. (1996). Suppose a network consists of two nodes, A and B, and they are connected by two links, 1 and 2, as shown in Figure 8.1a. There is a demand for O–D trips between A and B, and the demand function is given by DAB ( pAB ), where pAB is the full price of a trip from A to B, which is the sum of the monetized cost (e.g., fuel), non-monetized user cost (notably time), and – if levied – a toll (which is a transfer, not a societal cost). The number of O–D trips, qAB, is determined by q AB = DAB ( pAB ). There are two routes between A and B: route 1 via link 1, and route 2 via link 2. The full trip price for route j is written as C j ( x j ) + t j , where C j ( x j ) is the user cost required to pass through link j, which is an increasing function with respect to traffic volume, xj, and τj is the toll for link j. A trip maker chooses a route to minimize the full price. User equilibrium of route choice is attained when no driver has the incentive to change the route. This is characterized by Wardrop’s first principle that the full price in all routes actually used is equal and less than those experienced by a single vehicle on any unused route. We consider the case in which two routes are used; thus, the equilibrium conditions are expressed as:

PAB (q AB ) = C1 ( x1 ) + t1 = C2 ( x2 ) + t2 = pAB (8.1a)



x1 + x2 = q AB (8.1b)

where PAB (q AB ) is the inverse of the demand function DAB ( pAB ). For the given toll levels (τ1, τ2), we solve the system of Equations (8.1) for qAB, x1, x2, which is the equilibrium traffic (a)

Two links in parallel Link 1

A

B Link 2

Figure 8.la  Two links in parallel

148  Handbook on transport pricing and financing

assignment. Socially optimal traffic assignment is obtained by maximizing the social surplus, defined as follows:

W=

ò

q AB

0

PAB ( z )dz - C1 ( x1 ) x1 - C2 ( x2 ) x2 (8.2)

Maximizing (8.2) subject to (8.1), we obtain the first-best optimal pricing rule, as follows:

t j = C ¢j ( x j ) x j , for j = 1, 2 (8.3)

The first-best policy requires the toll at each link to be equal to the marginal congestion externality. We should consider the second-best policy when the first-best pricing is infeasible. Suppose that one of the two links, link 1, can be tolled. In this case, the toll at link 2 is zero, that is, τ2 = 0, and the equilibrium condition (8.1a) becomes:

PAB (q AB ) = C1 ( x1 ) + t1 = C2 ( x2 ) (8.4)

The second-best pricing problem is the choice of τ1 to maximize (8.2), subject to (8.1b) and (8.4). The optimal condition is as follows:

¶x



( PAB (q AB ) - C1( x1 ) - C1¢( x1 ) x1 ) ¶t1 1

¶x + ( PAB (q AB ) - C2 ( x2 ) - C2¢ ( x2 ) x2 ) 2 = 0 ¶t1

(8.5)

This condition is consistent with the general rule for determining the second-best policy. The first and second terms are the marginal changes in deadweight losses for links 1 and 2, respectively. The policy variable (toll of link 1 in this case) should be determined such that the sum of the marginal effects is canceled. Using (8.4), the condition above can be rewritten as follows:

¶x

¶x

( t1 - C1¢( x1 ) x1 ) ¶t1 - C2¢ ( x2 ) x2 ¶t2 1

= 0 (8.6)

1

Rearranging terms yields the same expression as in Verhoef, et. al. (1996) with a slight change of notation, i.e.,

é ¢ (q AB ) ù - PAB t1 = C1¢( x1 ) x1 - C2¢ ( x2 ) x2 ê ú (8.7) ¢ ¢ ( ) ( ) C x P q AB AB û ë 2 2

The bracketed part in the second term takes values between 0 and 1 depending on the extent to which traffic is diverted to route 2; thus, we see t1 < C1¢( x1 ) x1, which exemplifies that the second-best toll on link 1 is typically smaller than the congestion externality there. This is

Road pricing and provision of capacity  149

because tolling on link 1 should consider the externality effect on link 2 caused by switching of traffic from links 1 to 2. The second term also includes the effect of tolling on the quantity of O–D flow through a change in the equilibrium full price pAB. Example 8.2: two links in series Trip demands exist between all pairs of nodes in the serial network, as shown in Figure 8.1b. Thus, there are three O-D pairs, i.e., AB, BC, and AC. AB and BC are short trips, whereas AC is a long trip. This setting is closer to general networks than Example 8.1, in that trips of different O–D pairs use the same link. Link flows are the sum of the O–D trips, as follows: x1 = q AB + q AC , x2 = qBC + q AC (8.8)



Suppose that only link 1 can be tolled. For the given τ1, qAB, qBC, qAC is obtained by solving the following system of equations:

PAB (q AB ) = C1 ( x1 ) + t1 (8.9a)



PBC (qBC ) = C2 ( x2 ) (8.9b)



PAC (q AC ) = C1 ( x1 ) + t1 + C2 ( x2 ) (8.9c)

The second-best pricing problem to be solved is



max t1

ò

q AB

0

PAB ( z)dz +

ò

qBC

0

PBC ( z )dz +

ò

q AC

0

PAC ( z)dz

(8.10)

- C1 ( x1 ) x1 - C2 ( x2 ) x2 We have the same optimal condition as in Equation (8.6). The comparative statics for (8.9) ¶x ¶x yields 1 < 0, and 2 < 0 . It follows that t1 > C1¢( x1 ) x1 , which exemplifies that the second¶t1 ¶t1 best toll on link 1 should be larger than the congestion externality there. This is because the toll on link 1 is set to internalize the externality effect of trip AC on the untolled link, that is, link 2. The above condition implies that trip AB are overpriced, and trips AC and BC are underpriced.

(b) A

Figure 8.lb  Two links in series

Two links in series Link 1

B

Link 2

C

150  Handbook on transport pricing and financing

As illustrated above, second-best pricing should take into account the effects of tolling not only on the tolled link but also on other links in the network. Moreover, we need information ¶x regarding the signs of j to set the second-best toll. It is impossible to obtain this informa¶ti tion for a general network, in which each link is used by route flows of many different O-D pairs: links may be complements for some routes and substitutes for other routes. Thus, we obtained analytical insights only for the simple networks described above. The design of the second-best pricing involves not only setting the levels of tolls, but also the choice of links where tolls are levied. Verhoef (2002) developed a heuristic algorithm to obtain the second-best design for general networks and examined the efficiency of the method through simulations. 8.2.2 Cordon Pricing Cordon pricing is adopted in road pricing practices to control area-wide congestion in a city (e.g., Singapore, Stockholm). A typical cordon pricing system is designed as follows. Each vehicle is charged a fixed toll when it passes through the specified cordon surrounding the central area of a city where traffic is most congested. Earlier studies on cordon pricing applied a network simulation model to actual cities to compute the effects of pricing systems on traffic flow patterns and congestion levels. Santos (2004) examined optimal cordon tolls for eight English towns. However, they did not deal with an essential design variable, that is, the size of the cordoned area (or the location of the tolled links in the network). There have been several developments in the method for obtaining the optimal location of tolled links for cordon pricing in a network. Note that the optimization problem involves both discrete (tolled link) and continuous (toll level) variables. In this case, the problem is generally non-convex; therefore, standard solution methods are not applicable to solve the problem. Moreover, cordon pricing imposes a constraint on the set of tolled links that the tolled area should be completely cordoned, that is, there must be no untolled link where cars can cross the cordon. To address this issue, May et al. (2002), Zhang and Yang (2004) and Sumalee (2004) used a genetic algorithm for the simultaneous determination of toll levels and cordon locations on networks. Ekström et al. (2012) proposed an approximation of the optimal design problem using a mixed-integer linear program, which can be solved to its globally optimal solution. Although network models are useful for practical applications, they are not suitable for investigating the general properties of a problem, due to the results being dependent on the network structures specified for calculations. In this regard, continuous space models are useful for the theoretical analysis of the qualitative properties of traffic patterns and resource allocation, providing some policy insights. Mun et al. (2003) presented a model that explicitly dealt with the spatial patterns of trip-making behavior and traffic congestion in a continuous one-dimensional space. They examined the effect of cordon pricing in a monocentric city where all trips are destined for the central business district (CBD) and showed that trips originating from locations inside the cordon are underpriced, those just outside the cordon are overpriced, and those near the urban fringe are underpriced. A notable result is that cordon pricing attains an economic welfare level very close to the first-best optimum.

Road pricing and provision of capacity  151

Cordon pricing induces land use changes through location choices of households and firms. Verhoef (2005) examined the effects of cordon pricing on residential locations based on the general equilibrium model of a monocentric city. The model endogenously determines the spatial variation in population density, land rent, labor supply, congestion level, and wage level. The numerical simulations demonstrate that the earlier result on the welfare effect still holds, with a more realistic model setting. De Lara et al. (2013) incorporated land use for road space into a monocentric city model and examined the effect of cordon pricing on land use patterns in the Paris region. Anas and Hiramatsu (2013) evaluated land use and welfare effects of cordon pricing in the Chicago region, based on the general equilibrium model which incorporates the location choices of firms together with residential location. Their findings suggest that restrictive cordons around Chicago’s CBD may decentralize jobs, whereas cordons around the inner suburbs may centralize jobs. While studies on the land use effect of cordon pricing rely on a numerical approach, Brueckner (2014) analytically investigated the land use effects of cordon pricing in a simplified model in which the city consists of three zones. Zones are connected to each other by mid-city and suburban bridges, and cordon pricing is represented by tolling at one of two bridges, the mid-city bridge. Brueckner proved the propositions that earlier works have numerically shown. Several extensions have been made, taking into account interactions with public policies in other sectors of the economy. Kono and Kawaguchi (2017) discussed how optimal land use regulations (on floor area ratio and urban growth boundary) should be designed in the presence of optimal cordon pricing. Tikoudis et al. (2015) examined the use of toll revenue from cordon pricing, to replace distortionary labor tax and subsidize public transit. Area pricing, similar to cordon pricing, was also implemented in real practices (in London, and initially in Singapore).1 Under area pricing, cars driving in a specified area must pay a fixed toll. The area toll is imposed on a daily basis, regardless of the number of times it crosses the boundary (cordon) of the toll area. Maruyama and Sumalee (2007) compared cordon pricing and area pricing based on the network equilibrium model with a trip chain. They applied the model to Utsunomiya City, Japan. Their main results were as follows: (i) the optimal toll under area pricing is higher, and (ii) the social welfare under optimal area pricing is slightly larger than that under optimal cordon pricing. Fujishima (2011) compared the land use effects of cordon pricing and area pricing based on the polycentric city model by Anas and Xu (1999) and showed that both schemes can lead to population centralization and job dispersion. 8.2.3 Value Pricing Value pricing has been adopted on many highways in the United States. In value pricing, a roadway is divided into two types of lanes: toll lanes, and free lanes. In most cases, high-occupancy vehicles (HOV) can use the toll lanes for free, so they are called HOT (HOV or toll) lanes. HOT lanes have been built by converting HOV lanes that allow the use of HOVs only. The conversion of HOVs to HOTs has the following effects: Individuals with a higher value of time tend to choose HOT lanes, which induces a shift of users from free lanes to HOT lanes, thereby reducing congestion on the free lanes. On the other hand, the HOT policy discourages car-pooling owing to increased congestion in the HOT lanes compared with the case of HOV only. Nevertheless, the HOT policy makes majority of users better off. Small and Yan (2001), and Verhoef and Small (2004) introduced driver heterogeneity in the value of time to analyze the welfare performance of value pricing. Yang and Huang (1999)

152  Handbook on transport pricing and financing

considered combinations of congestion pricing and HOV lanes, assuming that each commuter incurs an identical cost to form a carpool. In contrast to the above works that rely on the numerical approach, Konishi and Mun (2010) theoretically investigated the welfare effects of HOV and HOT policies and alternative pricing policies. They introduced heterogeneity in carpool organization costs among road users and described decision-making regarding carpooling or driving solo. They showed that converting HOV lanes to HOT lanes can improve social welfare, but in some cases this is not effective.2 Optimal allocation requires imposing tolls on all solo drivers and differentiating toll levels by lane. 8.2.4 More Sophisticated Pricing Schemes The Singapore Land Transport Authority (LTA) is planning to install a second-generation Electronic Road Pricing (ERP) system using satellite positioning technology instead of gantries in the current system. This new system will enable distance-based congestion charging (Singapore LTA, 2013).3 It is well known that distance-based pricing is more effective than cordon pricing (e.g., May and Milne, 2000; Verhoef, 2005). Distance-based pricing discourages longer trips that cause more congestion externalities than shorter trips. Fuel tax has the same effect as distance-based pricing, but it is not applicable for congestion management because it is infeasible to adjust the tax rate depending on congestion levels. Owing to the new technology, distance-based pricing has become a practical instrument and has received increased attention in recent years. Meng et al. (2012) considered the problem of obtaining the optimal distance toll, which is approximated by a piecewise-linear function of distance. A distance toll was applied to the cordoned area. They numerically showed that the slope of the toll function is steeper as the trip distance is longer, and the social welfare is significantly larger than the kilometer charge, that is, the linear distance toll. Lawphongpanich and Yin (2012) also developed an algorithm to solve the optimization problem of nonlinear road pricing, similar to Meng et al., in which the toll function consists of a constant and linear distance term. The implementation of distance-based pricing requires installing automated vehicle identification sensors or GPS (Global Positioning System) devices which collect vehicle location information for the purpose of identifying paths (and distance traveled) and charge them accordingly. Zangui et  al. (2015) developed mathematical models for the sensor-location problem and investigated their properties. They considered the following: (i) locating a minimum number of sensors to implement a given pricing scheme, (ii) designing an optimal path-differentiated pricing scheme for a given set of sensors, and (iii) finding a pricing scheme to induce a given target link-flow distribution while requiring a minimum number of sensors.

8.3 DYNAMIC CONGESTION PRICING In this section, we present a review of dynamic congestion pricing schemes, with a focus on development in line with Vickrey’s (1969) bottleneck model, MFD and bathtub models. In dynamic congestion pricing, the primary objective in the first-best setting is to eliminate the queuing delay. A recent comprehensive review of bottleneck models can be found in Small (2015) and Li et al. (2020).

Road pricing and provision of capacity  153

8.3.1 Bottleneck Model 8.3.1.1 Basics of bottleneck model Assume that (1) there is a fixed number of identical travelers for their morning commute; (2) they have the same preferred arrival time (PAT) and homogenous piecewise-linear scheduling preferences with respect to travel time per se, schedule delay, and early with respect to PAT; and (3) there is a single bottleneck with a fixed capacity on the road for commuting. Arnott, de Palma, and Lindsey (hereafter referred to as ADL) (1990) derived the Nash equilibrium of travelers’ departure-time choice for different congestion pricing schemes under these settings as follows: [a] Under non-toll equilibrium, (i) the queue linearly increases until there is no scheduling cost for the driver and then linearly decreases; (ii) the slope of queue evolution/ devolution is parameterized by traveler scheduling preferences and road capacities; (iii) there is a deadweight loss due to the negative externality of the queuing time at the bottleneck; and [b] the first-best time-varying (fine) congestion pricing, which eliminates queuing, is given by a triangular-shaped piecewise-linear tolling, where the first-best social cost is exactly half of the one for the non-toll equilibrium. ADL (1993) further elaborated on the framework for the case in which the total traffic demand is elastic. 8.3.1.2 Second-best congestion pricing in bottleneck models First-best dynamic tolling may rarely be implemented, and thus the peak period coarse tolling problem on a single bottleneck in which a flat toll is charged during part of the morning rush hours has been studied. ADL (1990) showed that the queues at the bottleneck are of zero length not only at the starting and ending moments of the entire peak period, but are also eliminated at the starting and ending moments of the smaller time window during which the toll applies, while queuing occurs at all other moments. They also showed that the efficiency of such an optimal single-step coarse toll (the toll equals a fixed value during the center of the peak, whereas outside this period, it is zero) would be above 50% with respect to the first-best fine toll. Xiao et al. (2011) analytically solved the optimal one-step coarse tolling scheme for a single bottleneck with commuter heterogeneity in the value of travel time and found that the optimal coarse toll is Pareto-improving because the delay costs for all classes of travelers are reduced under this tolling scheme. Van den Berg (2014) analyzed the effects of various travelers’ heterogeneous preferences on optimal coarse (single-step) tolling. A straightforward extension of coarse tolling is multistep tolling. This was initiated by Laih (1994), who found that the amount of n/(n + 1) for the total queuing time can be eliminated by introducing optimal n-step tolling. Lindsey et al. (2012) extended the model of ADL (1990) to multistep tolling and introduced the so-called braking behavior of travelers, who may have the incentive to slow down or stop before reaching the tolling point and wait. Chen et al. (2015) further elaborated on this problem by considering general user heterogeneity and formulated the optimal design of multistep tolling as a mathematical program with equilibrium constraints. Coarse tolling is less efficient than first-best fine tolling. Knockaert et al. (2016) studied the efficiency change in coarse tolling by differentiating its level and timing across groups of travelers. Through a numerical analysis, they confirmed that there would be a welfare gain of 69% for the first-best cases and a gain of 53% for generic coarse tolling. Xu et al. (2019) proposed an optimization framework for coarse tolling under proportional user heterogeneity

154  Handbook on transport pricing and financing

with the upper bound constraints of the toll level and tolling duration, and compared their results with ADL (1990), Laih (1994) and Lindsey et al. (2012). 8.3.1.3 Urban spatial structure and dynamic congestion pricing The bottleneck model is extended to investigate the effect of congestion pricing on the urban spatial structure. Arnott (1998) first incorporated urban spatial structure into the standard bottleneck model with homogenous traveler preferences. In the two region (downtown and suburb) settings, he found that optimal tolling does not have an effect either on urban structure (i.e., migration equilibrium in the long run) or on commuting costs (departure time equilibrium in the short run) when the revenues are not redistributed. It is because the generalized price does not change when optimal tolling is introduced in the standard bottleneck model. Gubins and Verhoef (2014) employed travelers scheduling utility proposed by Vickrey (1973) due to being at home as well as being at the workplace in the framework of a closed monocentric city with the consumption on housing size. They found that congestion pricing promotes for commuters to stay more at home and to have larger houses thereby leading to the population dispersion toward suburbs. Fosgerau et al. (2018) studied dynamic congestion in an open and monocentric city where commuters have Vickrey’s (1973) scheduling preferences. They found similar conclusions to Gubins and Verhoef (2014) that a city with optimal congestion tolling is less dense in the center and denser in the suburb. Takayama and Kuwahara (2017) extended the model by Arnott (1998) by incorporating commuter heterogeneity in the residential location model with bottleneck congestion and showed that optimal time-varying toll may cause more dispersed land use. This is a similar conclusion to the ones by Gubins and Verhoef (2014) and Fosgerau et al. (2018) and different from the standard results based on static flow congestion. Takayama (2020) further extended the framework by introducing positive income elasticity of land demand and concluded that optimal congestion toll makes the city denser when richer commuters are more flexible in the schedule while it makes the city less dense when richer commuters are less flexible; and that optimal congestion pricing helps rich commuters but hurts poor commuters. 8.3.2 More Sophisticated Pricing Schemes In the last two decades, analyses aimed at discovering more socially acceptable and implementable congestion pricing schemes have been conducted, which take the rapid development of information and communication technologies into consideration. 8.3.2.1 Pareto-improving toll Daganzo and Garcia (2000) explored the potential of dynamic tolling to benefit all drivers, even if the collected revenues are not returned to them in a single-point queue bottleneck framework with inelastic traffic demand. Their analysis framework is multi-day-based, which states that daily commuters can be classified as either “free” or “paying.” Daganzo and Garcia (2000) confirmed that, over an extended period of time, the toll level, its duration and the share of free commuting can be chosen in a way that it will benefit everyone (that is, Pareto-improving). Fosgerau (2011) proved that the fast lane scheme is always Pareto-improving when demand is inelastic. In the fast lane scheme, travelers are assigned to a priority group, and a more than

Road pricing and provision of capacity  155

proportional share of road capacity is reserved for the priority group. If the reserved capacity is not used for the priority group, it is available for the nonprioritized group. Fosgerau (2011) also argued that the fast lane scheme can reproduce the equilibrium arrival pattern of the optimal coarse toll given by ADL (1990). Furthermore, Van den Berg and Verhoef (2011) showed that for heterogeneous drivers pricing all the lanes does not necessarily make the majority worse off and it is possible to achieve Pareto improvement. Hall (2018) also obtained a similar result based on the assumption that queuing reduces the throughput of the bottleneck. 8.3.2.2 Tradable bottleneck permit Akamatsu et al. (2006), and Akamatsu and Wada (2017) explored the concept of a “Tradable Bottleneck Permit” (TBP) scheme whereby queuing can be eliminated if the number of tradable permits in auction markets issued per unit of time is equal to the bottleneck capacity. Within Vickrey’s (1969) bottleneck model framework, Akamatsu et  al. (2006) proved that there is a Pareto improvement with the TBP scheme for both travelers and road administrators and that the equilibrated travel pattern is socially optimal. Akamatsu and Wada (2017) extended the TBP scheme to include general networks. It is possible to provide incentives for travelers to change their travel behavior by providing opportunities for positive monetary gains, such as introducing money-like gifts or rewards. Such a reward scheme can also be studied within the dynamic model framework. Sun et al. (2020) investigated the impact of rewards and the market penetration rate for both homogeneous and heterogeneous commuters and showed that the queue would be eliminated if a sufficiently large budget is available for rewarding. 8.3.3 Hypercongestion, MFD, and the Bathtub Model Empirical studies show that travel speed at a single point on a road section is a decreasing function of traffic density, and that when the impact of speed upon throughput dominates, traffic throughput can actually become lower for higher densities than for lower densities. This phenomenon is called hypercongestion, and the functional relationship between speed and density is well known as the fundamental diagram of traffic flow (Greenshields, 1935). The point-queue assumption in standard bottleneck models discussed above cannot generate hypercongestion in the sense that the queue is spaceless and speed in the queue is therefore zero. Verhoef (2001, 2003) proposed a dynamic model of road traffic that accommodates hypercongestion based on car-following theory and found that a naive application of a tolling schedule as found by Vickrey (1969) may worsen welfare. A fundamental diagram of traffic flow, which has been mostly studied at the single road link level, also applies at the urban neighborhood network level. The concept of MFD (Daganzo, 2007; Geroliminis and Daganzo, 2008) in the engineering literature, also called bathtub congestion (Vickrey, 1991, 2020; Arnott, 2013; Fosgerau, 2015) in the transportation economics literature, has been well studied since the late 2000s. Recently, Loder et al. (2019) reported comparison results of MFD estimates from more than 40 cities and confirmed the tendency of network-level critical density. MFD/bathtub congestion technologies can be incorporated into models of the morning commute problem (i.e., with endogenous demand patterns), and may have different implications for dynamic congestion pricing. Although it is difficult to analytically derive the pricing

156  Handbook on transport pricing and financing

scheme compared to standard bottleneck models, some studies considered pricing in conjunction with hypercongestion, MFD, and bathtub models. We note that the derived conclusions about the pricing principle are diverse and that some findings may be incompatible with each other. These models cannot be solved without making further assumptions. The difficulty to derive the equilibrium departure and arrival patterns of travelers lies in the complex relationship between drivers’ arrival patterns and the speed and density at every instant before arriving at the destination. As will be seen below, there are two major approaches for modeling traffic dynamics in an MFD/bathtub. According to the classification by Mariotte et al. (2017), the first may be called “accumulation-based MFD model” and the second may be called “tripbased MFD model.” 8.3.3.1 Accumulation-based models In the accumulation-based MFD model, the dynamics of a single bathtub are represented by a conservation equation where the outflow is determined by the network exit function (NEF) (Small and Chu, 2003; Daganzo, 2007; Gonzales and Daganzo, 2012), which relates the trip-completion rate to the density. A common specification of the NEF is that the tripcompletion rate is given by the ratio of the production measured as vehicle-miles per unit time and the average trip length of travelers. This can be viewed as the approximation by Little’s formula (Little, 1961) at the homogenous neighborhood scale. To ensure that this approximation is valid, the three assumptions need to hold: (i) undifferentiated roads, (ii) MFD, and (iii) time-invariant negative exponential distribution for trip lengths to ensure that the average trip length is the constant (Vickrey, 1991, 2020; Jin, 2020). Within this framework, Vickrey’s (1969) scheduling preference of travelers was combined with accumulation-based models by Small and Chu (2003), and Geroliminis and Levinson (2009). Small and Chu (2003) addressed the intractability issues in non-toll equilibrium analysis rooted in the accumulation-based model by approximating that the travel time completed at time t is determined not by the integral of speed over the trip duration but by the instantaneous speed at time t. However, they did not analyze the impact of congestion tolling. Using similar assumptions on travel time approximation with Small and Chu (2003), Geroliminis and Levinson (2009) combined the departure-time choice model with MFD-type congestion dynamics. Due to a MFD being assumed to exist in an urban downtown neighborhood, a cordon-based dynamic congestion pricing scheme can be harmonized with travelers’ departure-time choice equilibrium analysis. Because the throughput of the entire network is time-varying (i.e., the number of travelers who reach their destination decreases as congestion levels become high), the optimal fine tolling would be different from that obtained from the standard bottleneck model in the following way: (i) the start time of tolling is later, (ii) the end time of tolling is earlier, and (iii) the timing when the toll level is the largest is different. Arnott (2013) developed a bathtub model in which the travel speed is negatively and linearly related to the traffic density and the outflow is proportional to the product of density and speed. This follows the assumption that each trip would be completed stochastically, with no information about trip lengths that lead to late trip completion which is simply and approximately given by Little’s formula, implying that drivers do not know their trip lengths when they depart. Arnott (2013) showed that when demand is high relative to capacity, applying an optimal time-varying toll generates benefits that may be considerably larger than those obtained from standard models and exceed the toll revenue collected.

Road pricing and provision of capacity  157

Although the accumulation-based model is simple and tractable, inconsistent lags for information propagation between boundaries may be observed in the numerical analyses by Mariotte et al. (2017), which implies that outflow may overreact to sudden demand surges, resulting in inconsistent information propagation between opposite perimeter boundaries. One reason is that accumulation-based models assume that the travel time for a trip can be approximated by the trip length divided by the instantaneous travel speed at the destination and that it is substantially different from the experienced travel time (Small and Chu, 2003). We note further that another assumption is implicitly considered that all travelers have the same trip lengths in Small and Chu (2003) and Geroliminis and Levinson (2009). 8.3.3.2 Trip-based models Trip-based MFD models explicitly consider the heterogeneity of trip length among drivers, and resolve the inconsistencies owing to approximation in the accumulation-based model discussed above. For example, Leclercq et  al. (2015), and Ge and Fukuda (2019) proposed alternative models of traffic dynamics that can handle heterogeneous trip lengths of travelers within the MFD framework. Recent studies on the trip-based MFD models incorporated endogenous departure patterns. Without employing NEF, Fosgerau (2015) presented a simple bathtub model belonging to a class of trip-based approaches that computes the time of trip completion as a function of departure time and speed, assuming that drivers choose departure time optimally by knowing their trip length and speed prior to their travel. By employing the general scheduling preferences proposed by Vickrey (1973) that can accommodate the effects of trip length, Fosgerau (2015) derived two interesting conclusions. First, in non-toll user equilibrium, the trips are sorted by travelers’ trip distances; more specifically, the short trip is carried out within the duration of the long trip. The second finding is that the nontoll user equilibrium is also a social optimum. The first argument states that the departure and arrival orders of travelers are last-in-first-out. The second argument is also important because it implies that any policy which changes driver’s departure time would only influence drivers negatively. In contrast, Lamotte and Geroliminis (2018) also investigated the impact of trip length heterogeneity by employing Vickrey’s (1969) scheduling preference within the framework of the trip-based approach. They derived different conclusions that state that at non-toll user equilibrium with trip-based MFD traffic dynamics, departures and arrivals of travelers follow a first-in-first-out order. Furthermore, they found that under rapid demand variations, the social optimum may exhibit hypercongestion.

8.4 PROVISION OF CAPACITY 8.4.1 Basic Theory The provision of road capacity means investment to reduce trip costs. This section deals with the relationship between capacity choice and road pricing. First, we consider the case of a single road where the trip demand is homogenous. We define user cost as C(q, k), where q is the number of trips and k is the road capacity. The user cost increases with q and decreases with k. The equilibrium number of trips is determined by P(q) = C (q, x ) + t: the marginal benefit

158  Handbook on transport pricing and financing

is equal to the full price of a trip, sum of user cost, and toll. The provision cost of capacity is given by G(k ) , which is an increasing function of k. We consider the capacity choice of the road authority for a given toll schedule. The objective function is defined as follows:

max k

ò

q

P( z)dz - C (q, k )q - G(k )

0

The optimality condition is:

-Ck q + éë P(q) - C (q, k ) - Cq q ùû

dq = G¢(k ) (8.11) dk

where Ck and Cq are the partial derivatives of C(q, x) with respect to k and q, respectively.4 The first term on the left-hand side of (8.11) is the reduction in user cost by capacity expansion multiplied by the number of users (trips), which represents the direct benefit. The bracketed part in the second term is the difference between the social marginal benefit and social marginal cost of trips. Thus, the second term represents the indirect effect of capacity expansion, i.e., the change in market distortion through a change in the number of trips caused by capacity expansion. The right-hand side is the marginal cost of capacity expansion. This condition is rewritten, using the equilibrium condition, as follows:

-Ck q + éë t - Cq q ùû

dq = G¢(k ) (8.12) dk

When the road toll is set optimally, i.e., t = Cq q , the above condition is reduced to:

-Ck q = G¢(k ) (8.13)

The direct benefit of capacity expansion is equal to the marginal cost. In other words, the optimal capacity choice is consistent with the benefit-cost rule under optimal pricing. Mohring and Harwitz (1962) showed that, under certain technical conditions, the revenue from the optimal road toll is equal to the cost of the optimal capacity. This is called “the self-finance theorem” or “the cost-recovery theorem.” The conditions for this theorem are as follows: (i) the user cost is a function of the volume-capacity ratio, represented by æqö C (q, k ) = C ç ÷ ; (ii) the capacity cost is linear in k, represented by G(k ) = gk , where γ is a èkø 2 æ q öæ q ö positive constant. In this setting, the condition for optimal capacity becomes C ¢ ç ÷ ç ÷ = g, è k øè k ø æqöq and the optimal toll is t = C ¢ ç ÷ . Then we have tq = gk . èkøk Conditions (i) and (ii) require constant returns to scale in congestion technology and in road construction, respectively. Verhoef and Mohring (2009) discussed whether these conditions are reasonable in the real world. Subsequent researches show that, as long as the conditions of constant returns to scale are satisfied, the theorem remains valid in more complicated (realistic) situations, such as in a network setting (Yang and Meng, 2002), when road users are

Road pricing and provision of capacity  159

heterogeneous (Arnott and Kraus, 1998), when there is uncertainty in demand and capacity (Lindsey, 2009), and in the dynamic model of bottleneck congestion (ADL, 1990, 1993). 8.4.2 Capacity Choice under the Second-Best Pricing The capacity choice following the benefit-cost rule, as in (8.13), is applicable only when the toll is set optimally. The cost‑benefit rule is widely adopted in policy practices. However, the naive application of this rule leads to an excessively large capacity if the road toll is underpriced, i.e., t < Cq q . In this case, the planner should follow the second-best capacity choice rule, as in Equation (8.12), which includes the second term, that is, the effect of the change in distortion. Wheaton (1978) and Wilson (1983) investigated the relationship between secondbest and first-best capacities. Their results suggested that the second-best capacity is likely to exceed the first-best capacity. Small and Verhoef (2007) considered the pricing and capacity choice of a single link in two-link networks (two links in parallel and in series). The second-best pricing rules are the same as those in the two examples in Section 8.2.1. The formula for the second-best capacity of the tolled link is the same as that for the first-best, that is, the benefit-cost rule. However, the capacity level is different from the first-best due to the number of trips in the formula being second-best. Toll revenue is not equal to capacity cost. Arnott and Yan (2000) discussed the two-mode problem in which two routes are imperfect substitutes, as in the case of roads and rails. They consider the second-best pricing of a mode (e.g., rail) when another mode (e.g., road) is underpriced and the capacity choices of both modes. For the mode in which the price is optimized, the capacity choice rule is formally identical to the benefit-cost rule. This result is consistent with that obtained when the two routes are perfect substitutes. On the other hand, the capacity choice appears similar to Equation (8.12) for the mode in which the price is not optimal (e.g., an underpriced road). There have been new developments in capacity provision with mode choice. Zhang et al. (2016) investigated the condition that the Downs-Thomson paradox occurs. They introduced the transit operator’s responses (choice of fare and frequency) to highway expansion and showed that the paradox does not occur if the transit operator chooses fare and frequency to optimize its objective (profit or social surplus). However, imposing self-financing constraints creates the possibility of a paradox. In the large body of literature on cordon pricing and value pricing, relatively few studies have dealt with capacity choice. De Lara et  al. (2013) endogenized the capacity choice by computing land area allocated to road space. They showed that road capacity under cordon pricing is larger than that under first-best pricing. Brueckner (2014) obtained some analytical results concerning capacity choice in a setting where cordon pricing is represented by tolling at one bridge. Brueckner showed that the toll revenue exceeds the cost of the optimal capacity for a tolled bridge. The capacity rule for an untolled bridge under cordon pricing leads to a capacity larger than the first-best level. Guo et al. (2017) numerically analyzed the effects of self-financing constraints on the optimal design of cordon pricing and capacity choice. For their parameter setting, the toll revenue does not cover the cost for the second-best capacity; therefore, self-financing requires a higher toll and a larger cordoned area. For value pricing, Light (2009) dealt with capacity provisions in various forms, such as the division of fixed capacity between free lanes and toll lanes, the addition of new capacity for toll lanes, and so on. Light investigated the welfare effects of value pricing on the distribution

160  Handbook on transport pricing and financing

of welfare among users with heterogeneity in values of time. Konishi and Mun (2010) discussed lane division in the presence of car-pooling. When the lane is divided optimally, there are two types of lanes, car-pooling lanes, and toll lanes, resulting in no solo drivers in the carpooling lane. In other words, HOT lanes (a mix of carpoolers for free and solo drivers paying tolls) are not supported under the optimal lane division. 8.4.3 Decentralized Provision of Capacity It is widely the case that the road network spans multiple jurisdictions, in each of which the local government is responsible for providing the road infrastructure. The construction of new roads by private firms is also increasing in recent decades. In these situations, there are interactions among the decisions made by different agents (governments or firms) through traffic flows on the network. Recent literature has dealt with this problem based on simple network models, such as parallel or serial networks. Suppose a serial network, as shown in Figure 8.1b, where links AB and BC are located in Regions 1 and 2, respectively. Each link is used not only for local trips but also for transit trips. Following the definitions of the variables in Section 8.2, qAB represents local trips in Region 1, and qAC represents transit trips. Transit trips may originate from another region, such as Region 2, or other regions. Each local government decides on the pricing and capacity of the road to maximize regional welfare in the jurisdiction, which is the sum of the local user surplus and toll revenue. An increase in transit trips has both positive and negative effects on regional welfare: an increase in toll revenue and an increase in congestion level. Local governments decide on pricing and capacity, taking into account the effects of changes in transit trips. The problem is that the government has the incentive to increase toll revenue from transit trips, similar to tax exporting or tax competition. Consequently, road capacities may be larger or smaller than the optimal level. The results vary depending on the network structure, relative size of transit demand, and other factors. Ubbels and Verhoef (2008) and De Borger and Proost (2012) provided a thorough review of the literature on this issue. Solutions to this problem have recently been presented. Brueckner (2015) considered a congested bridge that is built and operated by a local government, however, users travel from other jurisdictions. If the operating government adopts a budget-balancing toll in which the toll revenue exactly covers the capacity cost, a decentralized capacity choice attains an efficient allocation.5 De Borger and Proost (2016) obtained the same result for a situation in which two governments provided two facilities, each in its own territory. This result implied that the self-financing theorem of Mohring-Harwitz can be extended to the case in which capacity choice is decentralized. The question is whether the government has an incentive to adopt a budget-balancing toll that is not optimal from the viewpoint of the local government seeking to maximize regional welfare. We often observe budget-balancing policies in the real world, so there must be other social or political factors for the government to make such a decision. 8.4.4 Capacity Choice in Dynamic Models Capacity provision has also been studied in dynamic settings. Under the standard bottleneck model with a fixed demand, ADL (1990) investigated the optimal capacity with a linear capacity-expansion cost for different tolling regimes and found that the finer the toll, the lower the congestion and the lower the marginal benefit of capacity expansion. The

Road pricing and provision of capacity  161

self-financing principle of Mohring-Harwitz also applies to first-best (i.e., fine) tolling. ADL (1993) further showed that the self-financing principle holds irrespective of tolling regime (i.e., fine, and coarse tolls), in contrast to the corresponding result for the static model mentioned in Section 8.4.1, which states that the self-financing principle holds only when tolling is at marginal external cost; which is true both for fine tolls and for flat tolls in the basic bottleneck model. The similarity between static and dynamic models is that the self-financing principle applies when tolls are equal to marginal external costs under the usual technical conditions. For heterogenous traveler cases, Arnott and Kraus (1995) investigated the relation between the “anonymous” pricing scheme (i.e., the tolling is independent of driver type) and the optimal investment rules. They found that marginal cost congestion tolling is self-financing (feasible), even if users differ and is observationally indistinguishable when a completely flexible toll is introduced. Takayama (2020) found with a joint bottleneck and residential location choice model that expanding capacity financed by the revenue from congestion pricing could be regressive in a closed and monocentric city where richer commuters are less flexible about their schedule. However, for the tradable permit pricing scheme, Akamatsu et  al. (2006) showed that the self-finance theorem also applies to the tradable permit scheme for a single bottleneck with a fixed demand for homogenous travelers with Vickrey’s (1969) scheduling preferences. Akamatsu and Wada (2017) further showed that the theorem applies to general networks. We did not find any studies on the capacity provision or feasibility (i.e., self-financing) of tolling under MFD or bathtub congestion technologies. As we reviewed in Section 8.3.3, regarding the theoretical findings of congestion tolling with MFD or bathtub models, it does not appear that rigorous conclusions on the property of optimal tolling have yet been reached. Another possible reason for this is that both the MFD and bathtub models generally deal with the traffic network or an urban neighborhood as a single reservoir, and it is difficult to associate the expansion of the network capacity with actual transportation investments.6

8.5 EMPIRICAL STUDIES ON ROAD PRICING AND PROVISION OF CAPACITY Practical implementations of congestion charging, especially in London and Stockholm, both of which were realized in the 21st century, have stimulated a number of empirical works on traffic congestion and the effects of road pricing in a real city. Lehe (2019) provides a comprehensive review of the practice of congestion toll purchasing schemes of supplementary licenses to downtown areas of five major cities. In this section, we review recent empirical studies on computing the cost of traffic congestion and/or the impact of congestion tolling. Earlier studies used network simulations, and most recent empirical studies employed novel identification strategies and innovative data collection technology. 8.5.1 Estimating Congestion Costs and/or Welfare Effects Santos (2004) used SATURN7 to examine the effects of cordon tolls on actual road networks in eight English towns. The optimal tolls for a single cordon and double cordons were calculated, and the welfare gain was evaluated based on changes in social surplus.8 The ratios of

162  Handbook on transport pricing and financing

gain from the double cordons to those from single cordon are 1.1 to 2.1. And the ratio of the gain from the first-best toll to the double cordons are 1.05 to 3.62, in most cases, below 2.0.9 Based on these results, it can be concluded that cordon tolling performed well. The distributional effects vary across towns. Safirova et al. (2004) used a spatial model developed for metropolitan Washington DC, which computed the effects of converting existing HOV lanes to HOT lanes. They showed that although all income groups benefit from the policy on aggregate, wealthier households benefit considerably more than the lowest-income households. Small et al. (2006) estimated a discrete-choice model for travelers who have the option of traveling solo on the general (free) lanes, paying a toll to use the HOT express lanes, or forming a carpool to use the express lanes at a discount. The estimated model was used to evaluate the welfare effect of alternative pricing schemes in the corridor of California State Route 91 in Orange County. It was shown that a differentiated pricing policy for both general and express lanes generates a larger gain of social surplus than the HOT policy, while a substantial fraction of people incur sizable losses. Shiroma and Fukuda (2020) estimated drivers’ toll price elasticity by considering the 2016 toll reform scheme for the Tokyo metropolitan area. Using panel data of interchange-tointerchange aggregate traffic with a spatial fixed-effect model, they found that toll elasticity is distributed around a mean of –0.236 with a standard deviation of 0.0224. They further found that the change in total traffic demand was small, but the causal impact on the route shift from via CBD to via the outer ring road of Tokyo was significantly large. Fosgerau and Small (2012) presented an econometric model that is capable of estimating the dynamic marginal costs of traffic congestion on expressways. They analyzed consecutive expressway links across which queues may spill back. By using the traffic data of Danish motorways, dynamic econometric models were estimated, and it was found that the marginal congestion cost is largely affected by the dynamic properties of travel times and traffic flows. Recently, Hall (2021) showed that a time-varying toll on a portion of highway lanes may make all road users better off, even without recycling revenue. Using survey and travel time data, combined with a structural model of traffic congestion, he estimated a modified version of the bottleneck model which incorporates heterogeneity over three dimensions: value of time, schedule inflexibility, and desired arrival time. The model also assumes that queuing reduces the throughput of the bottleneck, which is a factor reducing queue by tolling to lead to Pareto improvement. However, Anderson and Davis (2020) provided evidence that rejects the reduction in throughput (capacity drop in their terminology). They conducted an event study to estimate the average flow by time before and after queue formation and showed that the traffic flow downstream of the bottleneck is essentially flat throughout, with no discontinuous change near the moment the queue forms. Their result is in sharp contrast to the large empirical literature that documents capacity drops at highway bottlenecks. 8.5.2 Estimating the Impacts on Other Types of Externalities An increasing number of empirical studies dealt with not only traffic congestion but also other types of externalities, such as accidents and pollution. Gibson and Carnovale (2015) explored an unexpected suspension of road pricing policy in Milan, Italy, as a natural experiment, and

Road pricing and provision of capacity  163

estimate drivers’ responses to road pricing and its effects on air pollution. Green et al. (2016, 2020) examined the effects of the London Congestion Charge (LCC) on traffic accidents and pollution, respectively. They found that there was a substantial and significant reduction in the number of accidents in the charged zone for charged vehicles and times relative to a synthetic control group. They also found significant reductions across a range of pollutants (e.g., NO2 and PM10) in comparison to control cities, and that these reductions are substantially larger than what would be expected from the reduction in traffic flows by itself. By integrating the difference-in-differences method with a generalized linear model, Li et al. (2012) presented the result that LCC reduces the total number of car accidents but is associated with an increase in two-wheeled vehicle accidents. Currie and Walker (2011) studied the effect of the introduction of an electronic toll collection technology called E-ZPass in the northeastern United States and the associated sharp reductions in local traffic congestion on the health of infants born to mothers living near toll plazas with the difference-in-differences method. They found that prematurity and low birth weight among mothers within 2 km of a toll plaza were reduced by 10.8% and 11.8%, respectively. 8.5.3 Estimating the Impacts on Housing Markets Several studies have explored the impacts of congestion pricing on housing markets. Percoco (2014) investigated the effect of the EcoPass in Milan on housing prices by examining average property values across 192 micro-zones between 2006 and 2009 and found that prices are 1.2% to 1.8% lower when congestion tax is introduced. Agarwal et al. (2015) estimated the impacts of Singapore ERP on real estate prices using fixed-effect models and showed that the increase in congestion toll rate caused a 19% drop in retail real estate prices within the cordon ERP areas relative to outside the ERP areas. To estimate the willingness to pay to avoid traffic congestion, Tang (2021) recently explored the effects of LCC on house prices using local traffic conditions as instruments by which the analysis is limited to properties close to the congestion charge boundary. He found that the estimated elasticity of housing values with respect to traffic was approximately –0.30 implying that less traffic due to LCC has generated an aggregate windfall of around £3.8 billion for homeowners in the zone relative to those outside the zone. 8.5.4 Pricing Impacts in Developing Economies Traffic congestion in cities of developing countries is more severe than in developed countries. We have seen new evidence of traffic congestion in developing economies in recent years. Akbar and Duranton (2017) estimated the deadweight loss of congestion in Bogotá, Columbia, based on a unique approach that combined Google Maps data with a household travel survey. They developed an econometric method which identifies the supply function separately from the demand function. The main finding was that the elasticity of travel time cost with respect to the number of travelers was very small. This suggests that the deadweight loss from congestion is very small; therefore, the level of congestion toll to achieve optimality should be low. They explained the reason for this result by the existence of local streets, which remain relatively uncongested even at peak hours, thus accommodating the increased traffic without slowing down.

164  Handbook on transport pricing and financing

For India, Kreindler (2020) derived a similar result in a different way. He used a sample of commuters who participated in a field experiment on congestion pricing in Bangalore to estimate the preference parameters in the departure choice model. Detailed travel behavior data were collected using a smartphone app that logged GPS location data from the participants of the experiment. It was shown that commuters were moderately flexible in terms of trip scheduling, however, the welfare gains from optimal congestion charges were negligible. It is still debatable whether their conclusion that road pricing is not effective is widely valid for cities in developing countries. Nevertheless, we should note that similar conclusions were obtained with different approaches and in different cities. 8.5.5 Impacts of Capacity Provisions Regarding the provision of capacity, Winston and Langer (2006) estimated that 1 dollar of government spending on highways in a given year reduces road users’ congestion costs by only 11 cents. Furthermore, the reallocation of highway funds among states to minimize congestion costs improves efficiency to a limited extent. Duranton and Turner (2011) estimated the elasticity of interstate highway vehicle kilometers traveled with respect to lane kilometers, using city-level traffic data in the United States. The results showed that the elasticity was close to 1, which suggests that the expansion of road capacity does not affect traffic congestion. This implies that “The fundamental law of traffic congestion” suggested by Downs (1962) applies. Hsu and Zhang (2014) supported and further strengthened this result using data from expressways in Japan. They showed that the elasticity may even exceed unity, owing to the coverage effect that expansion of the network encourages longer trips.

8.6 CONCLUSION This chapter selectively reviews recent developments in research on road pricing and capacity provision. We summarize the points discussed in each section and suggest topics we think are important for future research. Section 8.2 discusses the theories of second-best road pricing with the most emphasis. The literature on this topic is growing significantly, due to an increasing interest in road pricing in practice. Owing to the development of information technology, more complicated forms of second-best pricing have become more realistic. Accordingly, there are an increasing number of contributions to the design of more sophisticated pricing systems. However, the focus of existing works is biased toward the mathematical formulations and development of algorithms to solve the problem of simple hypothetical networks. It is desirable to deeply investigate how users behave in response to the introduction of a sophisticated pricing system and provide insights into welfare consequences. Section 8.3 addresses dynamic road pricing. We first review the recent progress in the bottleneck model, mainly on pricing policies under the second-best setting, and with the presence of user heterogeneity. The emergence of traffic models based on MFD or bathtub models has revitalized the discussion on hypercongestion, on which there were confusion in the past. Unlike earlier discussions based on static models, MFD models are dynamic, and thus, more promising. While there are various alternative approaches to modeling traffic dynamics based on MFD, relatively few studies have dealt with pricing policy. What is more problematic is that different assumptions yield drastically different results regarding the effects of road pricing. Therefore, it is too early to conclude that we have reached a consensus. This encourages the notion that further development is required.

Road pricing and provision of capacity  165

Section 8.4 discusses the theoretical developments in the provision of capacity. Following the description of the basic results for capacity choice under the first-best pricing, we review the theoretical developments on the second-best capacity choice and decentralized provision of capacity. Compared to pricing, the stock of research on capacity choice is relatively small. In particular, we expect more contributions to capacity choices under the second-best pricing of realistic forms, such as cordon pricing and value pricing. Section 8.5 reviews the empirical results regarding the effects of road pricing. Studies based on network simulations suggest that road pricing is effective in reducing congestion, even with the second-best pricing. However, some economists report different empirical results for megacities in developing countries that road pricing is not effective, even with the first-best one. It is unclear whether their results extend to other cities in developing countries. The accumulation of empirical findings is thus in order.

ACKNOWLEDGMENTS We thank the editors, two reviewers and participants at the workshop for the Handbook on Transport Pricing and Financing for useful comments and suggestions. We are most grateful to Erik T. Verhoef for detailed and rigorous editorial advice. We also thank Takao Dantsuji for valuable input on dynamic congestion modeling. Any shortcomings in the chapter, however, are our responsibility.

NOTES 1. The system in London is in some sense a mix of area pricing and cordon pricing since there are considerable discounts for residents in the toll area. 2. Converting HOV to HOT lanes is not effective when capacity of HOT lanes is relatively small. In this situation, it is optimal to exclude solo-drivers from HOT lanes by imposing prohibitive toll, i.e., the outcome is the same as that under HOV policy 3. There have been discussions concerning the replacement of the current cordon-based congestion pricing scheme by the distance-based pricing scheme. But it is reported recently that the distancebased charging will be on hold. 4. As assumed, Ck  0. 5. The constant returns to scale in congestion technology and capacity production are assumed. 6. MFD models in engineering communities might be more concerned with (optimal) perimeter control of traffic by which the inflow rate into the urban neighborhood is controlled to maximize the system throughputs (e.g., Daganzo, 2007). 7. Simulation and Assignment of Traffic to Urban Road Networks, developed by the Institute for Transport Studies (ITS) at University of Leeds. 8. Only toll levels are optimized. Locations of toll cordons are given exogenously. 9. In other words, the welfare gain of the second-best pricing is more than 50% of the first-best welfare gain.

REFERENCES Agarwal, S., Koo, K. M., Sing, T. F., 2015. Impact of electronic road pricing on real estate prices in Singapore. Journal of Urban Economics 90, 50–59. Akamatsu, T., Sato, S., Nguyen, L. X., 2006. Tradable time-of-day bottleneck permits for morning commuters. JSCE Journal of Infrastructure Planning and Management 62, 605–620.

166  Handbook on transport pricing and financing

Akamatsu, T., Wada, K., 2017. Tradable network permits: A new scheme for the most efficient use of network capacity. Transportation Research Part C: Emerging Technologies 79, 178–195. Akbar, P., Duranton, G., 2017. Measuring the Cost of Congestion in Highly Congested City: Bogotá, Research Department Working Papers 1028, CAF Development Bank of Latin America. Anas, A., Hiramatsu, K., 2013. The economics of cordon tolling: General equilibrium and welfare analysis. Economics of Transportation 2, 18–37. Anas, A., Xu, R., 1999. Congestion, land use and job dispersion: A general equilibrium model. Journal of Urban Economics 45, 451–73. Anderson, M. L., Davis, L. W., 2020. An empirical test of hypercongestion in highway bottlenecks. Journal of Public Economics 187, 104197. Arnott, R., 1998. Congestion tolling and urban spatial structure. Journal of Regional Science 38, 495–504. Arnott, R., 2013. A bathtub model of downtown traffic congestion. Journal of Urban Economics 76, 110–21. Arnott, R., de Palma, A., Lindsey, R., 1990. Economics of a bottleneck. Journal of Urban Economics 27, 111–130. Arnott, R., de Palma, A., Lindsey, R., 1993. A structural model of peak-load congestion: A traffic bottleneck with elastic demand. American Economic Review 83, 161–179. Arnott, R., Kraus, M., 1995. Financing capacity in the bottleneck model. Journal of Urban Economics 38, 272–290. Arnott, R., Kraus, M., 1998. When are anonymous congestion charges consistent with marginal cost pricing? Journal of Public Economics 67, 45–64. Arnott, R., Yan, A., 2000. The two-mode problem: Second-best pricing and capacity. Review of Urban and Regional Development Studies 12, 170–199. Brueckner, J. K., 2014. Cordon tolling in a city with congested bridges. Economics of Transportation 3, 235–242. Brueckner, J. K., 2015. Decentralized road investment and pricing in a congested, multi-jurisdictional city: Efficiency with spillovers. National Tax Journal 68, 839–854. Chen, H., Nie, Y. M., Yin, Y., 2015. Optimal multi-step toll design under general user heterogeneity. Transportation Research Part B: Methodological 81, 775–93. Currie, J., Walker, R., 2011. Traffic congestion and infant health: Evidence from E-ZPass. American Economic Journal: Applied Economics 3, 65–90. Daganzo, C. F., 2007. Urban gridlock: Macroscopic modeling and mitigation approaches. Transportation Research Part B: Methodological 41, 49–62. Daganzo, C. F., Garcia, R.C., 2000. A Pareto improving strategy for the time-dependent morning commute problem. Transportation Science 34, 303–311. De Borger, B., Proost, S., 2012. Transport policy competition between governments: A selective survey of the literature. Economics of Transportation 1, 35–48. De Borger, B., Proost, S., 2016. The political economy of pricing and capacity decisions for congestible local public goods in a federal state. International Tax and Public Finance 23, 934–959. De Lara, M., De Palma, A., Kilani, M., Piperno, S., 2013. Congestion pricing and long term urban form: Application to Paris region. Regional Science and Urban Economics 43, 282–295. de Palma, A., Lindsey, R., 2011. Traffic congestion pricing methodologies and technologies. Transportation Research Part C: Emerging Technologies 19, 1377–1399. Downs, A., 1962. The law of peak-hour expressway congestion. Traffic Quarterly 16, 393–409. Duranton, G., Turner, M. A., 2011. The fundamental law of road congestion: Evidence from US cities. American Economic Review 101, 2616–2652. Ekström, J., Sumalee, A., Lo, H., 2012. Optimizing toll locations and levels using a mixed integer linear approximation approach. Transportation Research Part B: Methodological 46, 834–854. Fosgerau, M., 2011. How a fast lane may replace a congestion toll. Transportation Research Part B: Methodological 45, 845–851. Fosgerau, M., 2015. Congestion in the bathtub. Economics of Transportation 4, 241–255. Fosgerau, M., Kim, J., Ranjan, A. 2018. Vickrey meets Alonso: Commute scheduling and congestion in a monocentric city. Journal of Urban Economics 105, 40–53. Fosgerau, M., Small, K. A., 2012. Marginal congestion cost on a dynamic expressway network. Journal of Transport Economics and Policy 46, 431–450.

Road pricing and provision of capacity  167

Fosgerau, M., Van Dender, K., 2013. Road pricing with complications. Transportation 40, 479–503. Fujishima, S., 2011. The welfare effects of cordon pricing and area pricing: Simulation with a multiregional general equilibrium model. Journal of Transport Economics and Policy 45, 481–504. Ge, Q., Fukuda, D., 2019. A macroscopic dynamic network loading model for multiple-reservoir system. Transportation Research Part B: Methodological 126, 502–527. Geroliminis, N., Daganzo, C. F., 2008. Existence of urban-scale macroscopic fundamental diagrams: Some experimental findings. Transportation Research Part B: Methodological 42, 759–770. Geroliminis, N., Levinson, D. M., 2009. Cordon pricing consistent with the physics of overcrowding. In Lam, W. H. K., Wong, S. C., Lo, H. K. (eds.), Transportation and Traffic Theory 2009: Golden Jubilee, 219–240. Gibson, M., Carnovale, M., 2015. The effects of road pricing on driver behavior and air pollution. Journal of Urban Economics 89, 62–73. Gonzales, E. J., Daganzo, C. F., 2012. Morning commute with competing modes and distributed demand: User equilibrium, system optimum, and pricing. Transportation Research Part B: Methodological 46, 1519–1534. Green, C. P., Heywood, J. S., Navarro, M., 2016. Traffic accidents and the London congestion charge. Journal of Public Economics 133, 11–22. Green, C. P., Heywood, J. S., Paniagua, M. N., 2020. Did the London congestion charge reduce pollution? Regional Science and Urban Economics 84, 103573. Greenshields, B. D., 1935. A study of traffic capacity, Highway Research Board Proceedings 14, 448–477. Gubins, S., Verhoef, E. T., 2014. Dynamic bottleneck congestion and residential land use in the monocentric city. Journal of Urban Economics 80, 51–61. Guo, Q., Sun, Y., Li, Z.-C., Li, Z., 2017. An integrated model for road capacity choice and cordon toll pricing. Research in Transportation Economics 62, 68–79. Hall, J. D., 2018. Pareto improvements from Lexus lanes: The effects of pricing a portion of the lanes on congested highways. Journal of Public Economics 158, 113–125. Hall, J. D., 2021. Can tolling help everyone? Estimating the aggregate and distributional consequences of congestion pricing. Journal of the European Economic Association 19, 441–474. Hsu, W.-T., Zhang, H., 2014. The fundamental law of highway congestion revisited: Evidence from national expressways in Japan. Journal of Urban Economics 81, 65–76. Jin, W.-L., 2020. Generalized bathtub model of network trip flows. Transportation Research Part B: Methodological 136, 138–157. Knockaert, J., Verhoef, E. T., Rouwendal, J., 2016. Bottleneck congestion: Differentiating the coarse charge. Transportation Research Part B: Methodological 83, 59–73. Konishi, H., Mun, S., 2010. Carpooling and congestion pricing: HOV and HOT lanes. Regional Science and Urban Economics 40, 173–186. Kono, T., Kawaguchi, H., 2017. Cordon pricing and land-use regulation. Scandinavian Journal of Economics 119, 405–434. Kreindler, G. E., 2020. Peak-hour road congestion pricing: Experimental evidence and equilibrium implications, unpublished manuscript. Laih, C.-H., 1994. Queueing at a bottleneck with single- and multi-step tolls. Transportation Research Part A: Policy and Practice 28, 197–208. Lamotte, R., Geroliminis, N., 2018. The morning commute in urban areas with heterogeneous trip lengths. Transportation Research Part B: Methodological 117, 794–810. Lawphongpanich, S., Yin, Y., 2012. Nonlinear pricing on transportation networks. Transportation Research Part C: Emerging Technologies 20, 218–235. Leclercq, L., Parzani, C., Knoop, V. L., Amourette, J., Hoogendoorn, S. P., 2015. Macroscopic traffic dynamics with heterogeneous route patterns. Transportation Research Part C: Emerging Technologies 59, 292–307. Lehe, L., 2019. Downtown congestion pricing in practice. Transportation Research Part C: Emerging Technologies 100, 200–223. Li, H., Graham, D. J., Majumdar, A., 2012. The effects of congestion charring on road traffic casualties: A causal analysis using difference-in-difference estimation. Accident Analysis and Prevention 49, 366–377.

168  Handbook on transport pricing and financing

Li, Z.-C., Huang H.-J., Yang H., 2020. Fifty years of the bottleneck model: A bibliometric review and future research directions. Transportation Research Part B: Methodological 139, 311–342. Light, T., 2009. Optimal highway design and user welfare under value pricing. Journal of Urban Economics 66, 116–124. Lindsey, R., 2009. Cost recovery from congestion tolls with random capacity and demand. Journal of Urban Economics 66, 16–24. Lindsey, R., 2012. Road pricing and investment. Economics of Transportation 1, 49–63. Lindsey, R., Van den Berg, V. A. C., Verhoef, E. T., 2012. Step tolling with bottleneck queuing congestion. Journal of Urban Economics 72, 46–59. Little, J.D.C., 1961. A proof for the queuing formula: L = λ W. Operations Research 9 (3), 383–387. Loder, A., Ambühl, L., Menendez, M., Axhausen. K. W., 2019. Understanding traffic capacity of urban networks. Scientific Reports 9, 16283. Marchand, M., 1968, A note on optimal tolls in an imperfect environment. Econometrica 36, 575–581. Mariotte, G., Leclercq, L., Laval., J. A., 2017. Macroscopic urban dynamics: Analytical and numerical comparisons of existing models. Transportation Research Part B: Methodological 101, 245–267. Maruyama, T., Sumalee, A., 2007. Efficiency and equity comparison of cordon- and area-based road pricing schemes using a trip-chain equilibrium model. Transportation Research Part A: Policy and Practice 41, 655–671. May, A. D., Liu, R., Shepherd, S. P., Sumalee, A., 2002. The impact of cordon design on the performance of road pricing schemes. Transport Policy 9, 209–220. May, A.D., Milne, D.S., 2000. Effects of alternative road pricing systems on network performance. Transportation Research A: Policy and Practice 34, 407–435. Meng, Q., Liu, Z., Wang, S., 2012. Optimal distance tolls under congestion pricing and continuously distributed value of time. Transportation Research Part E: Logistics and Transportation Review 48, 937–957. Mohring, H., Harwitz, M., 1962. Highway Benefits: An Analytical Framework. Northwestern University Press, Evanston, IL. Mun, S., Konishi, K., Yoshikawa, K., 2003. Optimal cordon pricing. Journal of Urban Economics 54, 21–38. Percoco, M., 2014. The impact of road pricing on housing prices: Preliminary evidence from Milan. Transportation Research Part A: Policy and Practice 67, 188–194. Rouwendal, J., Verhoef, E. T., Knockaert, J., 2012. Give or take? Rewards versus charges for a congested bottleneck. Regional Science and Urban Economics 42, 166–176. Safirova, E., Gillingham, K., Parry, I., Nelson, P., Harrington, W., Mason, D., 2004. Welfare and distributional effects of road pricing schemes for Metropolitan Washington DC. Research in Transportation Economics 9, 179–206. Santos, G., 2004. Urban congestion charging: A second-best alternative. Journal of Transport Economics and Policy 38, 345–369. Shiroma, H., Fukuda, D., 2020. Effects of the 2016 toll scheme reform on urban expressway traffic demand in the Tokyo metropolitan area. Journal of Japan Society of Civil Engineers, Series D3 (Infrastructure Planning and Management) 76, 180–195. Singapore LTA, 2013. Land Transport Master Plan 2013. Singapore Land Transport Authority. Small, K. A., 2015. The bottleneck model: An assessment and interpretation. Economics of Transportation 4, 110–117. Small, K. A., Chu, X., 2003. Hypercongestion. Journal of Transport Economics and Policy 37, 319–352. Small, K. A., Verhoef, E. T., 2007. The Economics of Urban Transportation. Routledge. Small, K. A., Winston, C., Yan, J., 2006. Differentiated road pricing, express lanes, and carpools: Exploiting heterogeneous preferences in policy design. Brookings-Wharton Papers on Urban Affairs 7, 53–96. Small, K. A., Yan, J., 2001. The value of “Value Pricing” of roads: Second-best pricing and product differentiation. Journal of Urban Economics 49, 310–336. Sumalee, A., 2004. Optimal road user charging cordon design: A heuristic optimization approach. Computer-Aided Civil and Infrastructure Engineering 19, 377–392.

Road pricing and provision of capacity  169

Sun, J., Wu, J., Xiao, F., Tian, Y., Xu, X., 2020. Managing bottleneck congestion with incentives. Transportation Research Part B: Methodological 134, 143–166. Takayama, Y. 2020. Who gains and who loses from congestion pricing in a monocentric city with a bottleneck? Economics of Transportation 24, 100189. Takayama, Y., Kuwahara, M. 2017. Bottleneck congestion and residential location of heterogeneous commuters. Journal of Urban Economics 100, 65–79. Tang, C. K., 2021. The cost of traffic: Evidence from the London congestion charge. Journal of Urban Economics 121, 103302. Tikoudis, I., Verhoef, E. T., van Ommeren, J. N., 2015. On revenue recycling and the welfare effects of second-best congestion pricing in a monocentric city. Journal of Urban Economics 89, 32–47. Ubbels, B., Verhoef, E. T., 2008. Governmental competition in road charging and capacity choice. Regional Science and Urban Economics 38, 174–190. Van den Berg, V. A. C., 2014. Coarse tolling with heterogeneous preferences. Transportation Research Part B: Methodological 64, 1–23. Van den Berg, V. A. C., Verhoef, E.T., 2011. Winning or losing from dynamic bottleneck congestion pricing?: The distributional effects of road pricing with heterogeneity in values of time and schedule delay. Journal of Public Economics 95, 983–992. Verhoef, E. T., 2001. An integrated dynamic model of road traffic congestion based on simple carfollowing theory: Exploring hypercongestion. Journal of Urban Economics 49, 505–542. Verhoef, E. T., 2002. Second-best congestion pricing in general networks. Heuristic algorithms for finding second-best optimal toll levels and toll points. Transportation Research Part B: Methodological 36, 707–729. Verhoef, E. T., 2003. Inside the queue: Hypercongestion and road pricing in a continuous time– continuous place model of traffic congestion. Journal of Urban Economics 54, 531–565. Verhoef, E. T., 2005. Second-best congestion pricing schemes in the monocentric city. Journal of Urban Economics 58, 367–388. Verhoef, E. T., Mohring, H., 2009. Self-financing roads, International Journal of Sustainable Transportation 3, 293–311. Verhoef, E. T., Nijkamp, P., Rietveld, P., 1996. Second-best congestion pricing: The case of an untolled alternative. Journal of Urban Economics 40, 279–302. Verhoef, E. T., Small K.A., 2004. Product differentiation on roads: Constrained congestion pricing with heterogeneous users. Journal of Transport Economics and Policy 38, 127–156. Vickrey, W. S., 1969. Congestion theory and transport investment. American Economic Review 59, 251–261. Vickrey, W. S., 1973. Pricing, metering, and efficiently using urban transportation facilities. Highway Research Record 476, 36–48. Vickrey, W.S., 1991. Congestion in Midtown Manhattan in Relation to Marginal Cost Pricing, Technical Report. Columbia University. Vickrey, W. S., 2020. Congestion in midtown Manhattan in relation to marginal cost pricing. Economics of Transportation 21, 100152 (Co-edited by Richard Arnott and Wen-Long Jin). Wheaton, W. C., 1978. Price-induced distortions in urban highway investment. Bell Journal of Economics 9, 622–632. Wilson, J. D., 1983. Optimal road capacity in the presence of unpriced congestion. Journal of Urban Economics 13, 337–357. Winston, C., Langer, A., 2006. The effect of government highway spending on road users’ congestion costs. Journal of Urban Economics 60, 463–483. Xiao, F., Zhen, Q., Zhang, H. M., 2011. The morning commute problem with coarse toll and nonidentical commuters. Networks and Spatial Economics 11, 343–369. Xu, D., Guo, X., Zhang, G., 2019. Constrained optimization for bottleneck coarse tolling. Transportation Research Part B: Methodological 128, 1–22. Yang, H., Huang, H.-J., 1999. Carpooling and congestion pricing in a multilane highway with highoccupancy vehicle lanes. Transportation Research Part A: Policy and Practice 33, 139–155. Yang, H., Meng, Q., 2002. Highway pricing and capacity choice in a road network under a buildoperate-transfer scheme. Transportation Research Part A: Policy and Practice 36, 659–663.

170  Handbook on transport pricing and financing

Zangui, M., Yin, Y., Lawphongpanich, S., 2015. Sensor location problems in path-differentiated congestion pricing. Transportation Research Part C: Emerging Technologies 55, 217–230. Zhang, F., Lindsey, R., Yang, H., 2016. The Downs–Thomson paradox with imperfect mode substitutes and alternative transit administration regimes. Transportation Research Part B: Methodological 86, 104–127. Zhang, X., Yang, H., 2004. The optimal cordon-based network congestion pricing problem. Transportation Research Part B: Methodological 38, 517–537.

9. Public transport: design, scale, and pricing Sergio Jara-Díaz, Antonio Gschwender and Daniel Hörcher

9.1 INTRODUCTION As stated by Jan Owen Jansson in 2005, “after more than 30 years of discussion and theoretical development, the main (and somewhat embarrassing) question is: why is optimal bus transport pricing applied in hardly any urban area of the world?” This astute question brings forward two interrelated issues that emerge every now and then in the transport economics arena and reflect the political facets that underlie the discussion regarding public transport in general: subsidies and regulation.1 As we believe that part of the (embarrassing) answer to Jansson’s question lies in the failure to recognize that transit design and pricing are intimately linked, in this chapter we want to provide the basic elements behind this relation by exposing a) the technical (engineering) causes of scale economies in the provision of scheduled transit services, and b) the relation between those scale economies and optimal pricing. Noting that this is not intended to be a review of the public transport literature (for a recent review see Hörcher and Tirachini, 2021), in the following section we present the foundations of public transport supply, considering all the resources that are necessary to provide a given flow; the key technical variables are identified, a simple model is presented to show their role in optimal strategic design and scale economies, and extensions are provided to more complex settings. In Section 9.3 we make the connection between design and pricing from both sides: how optimal design determines optimal pricing, and how sub-optimal pricing induces suboptimal design. In Section 9.4, we briefly discuss the impact of the context in which transit is considered by extending the discussion to the wider socio-economic context in which public transport operates, by looking at second-best pricing, considering intermodal substitution, and by discussing the role of public funds, equity, social acceptance, political institutions, and environmental resources. Section 9.5 concludes.

9.2 THE FOUNDATIONS OF PUBLIC TRANSPORT SUPPLY 9.2.1 Operators’ Costs, Users’ Costs, and Decision Variables Public transport provision needs two types of resource. Those provided by the operators (rights-of-way, vehicles, stations, depots) and those provided by users, namely their time, which can be divided into three components: access time, waiting time, and in-vehicle time. The design of a public transport system implies a series of decisions aimed at the adequate provision of a public transport service, involving the amount of resources provided by both operators and users. These design decisions are taken in order, not necessarily the one in which they are presented here. Some of these design variables are closely related and may need to be decided simultaneously. Even more, as we will see later, it is convenient to jointly decide on several of them to achieve the best possible final design. 171

172  Handbook on transport pricing and financing

Technical change has played an important role in the evolution of public transport services as we know them today, where new technologies have become obsolete very quickly (see, for example, Post, 2010), moving from the omnibus – a horse-drawn cart – to the horsecars that took advantage of rails, to the cable-car that replaced the horse with a fixed steam engine and cable system, to the electric streetcars. Nowadays very different vehicle technologies coexist in any given city; buses may coexist with underground or elevated trains, streetcars, commuter rails, trolleybuses, funiculars, cog railways, or ferries, among others, whose specific characteristics and technologies differ (see Grava, 2003). Each vehicle technology has to be matched with an appropriate moving surface (track, road, etc.) and stops or stations. In some cases, the degree of separation from other vehicles as well as pedestrians, which greatly affects performance and investment costs, has to be decided (Vuchic, 2005). For a given vehicle technology, several relevant decisions remain. Together, the size of the vehicle, the number of seats, the space available for standing, and the expected passenger densities yield the vehicle capacity. The number and size of the doors, the specific use of each one of them for boarding, alighting or both, and the payment rules (e.g. inside the vehicle versus prior to boarding) will have an impact on the time that each passenger needs to board and alight and, therefore, on the dwell time of the vehicle at each stop or station. This dwell time, together with acceleration and deceleration rates, with the distance between stops, and with the cruise speed, yield the operating or commercial speed. This speed, together with the route length and time spent at the terminals (for vehicle turning, crew changes, etc.), results in the cycle time that each vehicle needs to complete a round trip. When service frequency (veh/h) is decided for each period to accommodate the demand, it also affects the quality of service in terms of waiting times; as we will see below, this can be done by balancing the resources provided by operators and users according to an optimality criterion. Frequency and cycle time are key aspects, as their product determines the number of vehicles needed for service operation (to which a proportion of vehicles for reserve, maintenance, and repair should be added). Public transport is usually a network of complementary and integrated services. Spatial decisions are needed for each line. The path (streets) followed by each line, the number of stops, and their location affect not only operational costs, but also users’ costs, particularly through their walking time. These spatial decisions, together with the operating hours, yield the space–time coverage of the service. Fare and subsidy levels are key financial decisions. But fare structure offers many possibilities as well. Single tickets and travel cards (allowing unlimited trips in a specific time window) may coexist. Fares may depend on distance or on a zone-based system. Different times of the day may have different fares. And fares can also vary among modes or technologies. Many other decisions exist regarding, for example, operations control to avoid bunching or for safety reasons, operations monitoring or users’ information. As we will see, the way the design variables are combined plays an important role in the optimal pricing of a public transport system. The simple choice between a large fleet of small vehicles and a small fleet of large vehicles, providing the same capacity to move a given flow of passengers, illustrates this idea: the former arrangement is more expensive but has advantages in terms of waiting time, whilst the latter is cheaper but not as good for passengers. This type of trade-off can be resolved by finding the combination that minimizes the total value of the resources consumed for each level of flow, i.e. obtaining the optimal balance of resources in order to find the welfare maximizing (optimal) prices given by the marginal total costs.

Public transport: design, scale, and pricing  173

To do so, the value of the resources consumed will be written as a function of the aggregated parametric demand, the value of each resource (time and others), technological parameters, and the decision variables. When this function is optimized in terms of decision variables, a cost function is obtained, which represents the minimum amount of resources needed to provide service for each parametric demand level, given the prices of the resources and the values of technological factors. Later on in this chapter, this cost function will be shown to be a key element for optimal pricing analysis, where it can interact with the demand function. 9.2.2 A Single-Line Model The path from the design to the cost function can be usefully illustrated using the singleline model (Mohring, 1976; Jansson, 1984) from where relevant elements emerge, considering a parametric hourly demand Y that is homogeneously distributed along a circular line of length L, where each passenger travels a distance l. Following the general model presented by Jara-Díaz and Gschwender (2003a), the value of the resources consumed (VRC) could be optimized over design variables such as the frequency (f), vehicle size (K), and the number of stops (p), as well as over operating decisions such as the effort that the operator puts on punctuality or headway regularity (d), safety (r), and reduction of pollution, and other externalities (x), by solving the following problem:

MinVRCT = VRCop + VRCU + VRC x (9.1)

Yl subject to a capacity constraint that assures that vehicle load, constant and equal to k = fL k due to the simplifying assumptions, is smaller or equal to K., i.e. F = £ 1. The total value of K the resources consumed is the addition of operators (op), users (U), and external (x) expenses. Operators’ expenses can be written as:

VRCop = ( c0 + c1K ) ftc + cr + cd + cx (9.2)

where cr, cd, and cx are the operators’ expenses to reduce risk, improve regularity, and reduce externalities levels, respectively, c0 and c1 are the parameters, and cycle time tc is: t c = tr + t p p + t



Y (9.3) f

with t the average time needed for a passenger to board and alight, tp the additional time that the vehicle needs at each stop because of deceleration and acceleration, and tr the remaining time in motion. Users’ expenditure is:



(

)

VRCU = Pa t a ( p ) Y + Pwt w f , d ( cd ) , F Y

(

)

+ Pv d ( cd ) , F t v ( f , p ) Y + Pmr ( f , cr , K , )

(9.4)

where Pa, Pw, and Pv are the respective values of access, waiting, and in-vehicle times. Average access time (ta) decreases with p, whereas average waiting time (t w) decreases with f and d

174  Handbook on transport pricing and financing

(better punctuality or regularity reduces waiting time), and increases with load factor Φ because of a larger probability of not being able to board the first vehicle. Average in-vehicle time (t v) can be written as (l /L )tc and therefore depends on the same variables and parameters as cycle time. The value of in-vehicle time may increase with the average load factor Φ to recognize the negative effect of crowding on users, and may also decrease with headway regularity d, as this reduces the crowding experience. Pm is the value of accident risk for users (r), which increases with f, and may decrease with cr and K. Other variables and parameters may also affect r, e.g. speed behind tr, deceleration rate behind tp, and even the effort on punctuality or regularity d. Finally, the external expenses can be written for example as:

VRC x = Pm m ( f , cr , K , ) + Px x ( f , cx , K , ) (9.5)

with m representing the accident risk level for non-users, which increases with f, K, and other variables, and Px and x representing the value and level of other externalities, such as noise or air pollution. This last variable may also depend on several other variables of the problem. An analytical solution for such a general model is unlikely to be feasible. But in a simplified version, analytical solutions are possible and useful to derive the cost function. As an example, following Jara-Díaz and Gschwender (2009), let us optimize the value of the resources consumed by operators and users in the isolated circular line over only two decision variables, namely frequency and bus size. Cycle time simplifies to: tc = T + t



Y (9.6) f

where T is time in motion, including acceleration and deceleration times at stops (whose number is not optimized now). Then, the fleet is B = ftc = fT + tY and in-vehicle time is l æ Yö t v = ç T + t ÷ Y . Assuming the average waiting time to be half the headway (perfectly regLè fø 1 ular headways), i.e. te = , the problem is: 2f MinVRCT = VRCop + VRCU = ( c0 + c1K ) ( fT + tY )

é 1 l æ Yö ù + ê Pw Y + Pv ç T + t ÷ Y ú 2 f L fø û è ë

(9.7)

k £ 1. As crowding is not considered in this example, VRCT K only increases with K and the constraint will always be active such that f becomes the basic design variable. Its optimal value follows the “square root formula” (Mohring, 1972; Jansson, 1984), i.e.

subject to the capacity constraint

f* =

Y æ1 l ö Pw + tY ( Pv + c1 ) ÷ and ç Tc0 è 2 L ø -1

l l æ1 ö K* = Tc0Y ç Pw + tY ( Pv + c1 ) ÷ . L L è2 ø

(9.8)

Public transport: design, scale, and pricing  175

By replacing f * and K* in VRCT the minimum total expense to produce Y trips, i.e. the total cost function (CT ), is obtained:

l l æP ö CT = tc0Y + 2 Tc0Y ç w + tY ( Pv + c1 ) ÷ + TY ( Pv + c1 ) L L è 2 ø

, (9.9)

where the first term and one square root is the users’ cost, and the rest is the operators’ cost. The average total cost function (ACT = CT /Y ) is:

l l æP ö ACT = tc0 + 2 Tc0 ç w + t ( Pv + c1 ) ÷ + T ( Pv + c1 ) (9.10) L L è 2Y ø

The fact that ACT decreases with demand Y means that there are scale economies. They exist because, as demand increases, optimal frequency also increases and therefore waiting time is reduced for all passengers (the “Mohring effect”). Technical parameters and resource values have important implications for optimal results and cost function. For example, if t, Pv, or c1 decrease, optimal frequency becomes closer to being proportional to the square root of Y. If they increase, the terms in Equation (9.10) that do not depend on Y increase their relative relevance, reducing the degree of scale economies. Average user cost (ACU ) and average operator cost (ACop) can also be obtained separately. Marginal total cost (mT ), marginal user cost (mU ), and marginal operator cost (mop) can be obtained by deriving the relevant cost terms with respect to Y. 9.2.3 Space and Lines Structures Another source of scale economies is the density of public transport lines. As demand increases, the system may adapt by increasing the number of lines in a given area, which can yield lower access time for users as new “closer” lines appear. The optimization of line density can be approached using the regular representations of a city. For example, Newell (1979) used a grid pattern, Byrne (1975) developed a radial concentric model, and Kocur and Hendrickson (1982) used a rectangular area served by parallel lines. But the spatial design of public transport lines raises more challenges than just the density of the lines. In fact, the detailed spatial design consists of the definition of the sequence of arcs for each line which, collectively, form what we refer to here as a line structure: the organization of vehicles in routes and itineraries to carry passengers from many origins to many destinations (Jara-Díaz, 2007). Some of these are known and can be conceptually described in simple terms, like a general cyclical system, where vehicles are organized following a closed circuit and stop at places close to origins and destinations where people board and alight. Some known arrangements are: the feeder–trunk scheme (FT) where different fleets operate as cyclical systems, some – the feeders – in short circuits that feed and collect passengers from stops served by other fleets – the trunk lines – that move passengers along longer circuits in corridors; the hub-and-spoke (HS) system, where many fleets bring passengers to a certain point (the hub) and collect others to be distributed along the vehicles’ routes; or the simplest system of many exclusive lines (EXC) where every OD pair is served by its own fleet.

176  Handbook on transport pricing and financing

Although the structural organization of transit lines in space is usually addressed through heuristics in real-size cases because of its complexity (e.g. Borndörfer et al., 2007), it is quite useful to analyze line structures using general but simplified representations of an urban area. Under this latter approach, the best structure will be the one that presents the lowest cost for a specific situation given by the combination of demand level (and its distribution), value of time, other inputs and technical parameters, and urban conditions. For example, Jara-Díaz and Gschwender (2003b) use a very simple cross-shaped network to analyze under which conditions direct lines without transfers are more convenient than corridor lines with transfers, finding that direct lines are more likely to be the best structure the larger the demand and boarding/alighting time, and the smaller the ratio between the values of waiting and in-vehicle times. Fielbaum et al. (2016) use a more general city representation, where monocentric, polycentric, and dispersed cities can be represented through the appropriate values of some parameters, to compare four strategic lines structures: HS, FT, EXC, and direct with stops (DIR). Figure 9.1 shows the more convenient line structure in an alpha-Y space, where alpha represents the degree of monocentricity of the city. Figure 9.1 shows that as demand increases, it becomes convenient for line routes to reorganize from systems that have mandatory transfers (HS, FT) to ones that involve neither transfers nor stops (EXC). What happens as patronage increases and the system adjusts optimally is that the number of transfers decreases, the number of stops decreases, and passenger route lengths approach the minimal path (the detour ratio is reduced), increasing what can be called “directness”. In terms of operator and user costs, when more passengers use public transport, inducing an increase in directness, in-vehicle times diminish whilst keeping

Source:   Fielbaum et al. (2016).

Figure 9.1  More convenient transit line structures for different degrees of monocentrism and total passenger flows

Public transport: design, scale, and pricing  177

reasonable waiting times, such that users benefit from lower travel times and operators benefit from lower idle capacity and lower cycle times. The (discrete) change in line structure as demand grows generates positive externalities on all passengers and permits a better use of the fleet (Fielbaum et al., 2020), making directness yet another source of scale economies with an effect on optimal pricing as explained in Section 9.3. Using the same parametric city model, Fielbaum et al. (2021) show that optimal line density (D) increases with patronage along with directness, reinforcing scale economies. When line density is optimized, the evolution of the optimal line structures from transfer-based ones to more direct ones remains, but the transition between optimal line structures occurs for larger flow volumes. 9.2.4 Technological Characteristics That Impact Operations and Optimal Design To obtain a cost function, some decision variables are optimized, but others may be assumed fixed. In the long-term, everything is assumed to be variable and optimizable, but which decision variables are fixed in the short term? There is no general answer for this, as it depends on many factors including the specific technology. For example, in a bus system it is relatively easy to modify routes and frequencies, but a change in vehicle size is hard because buses have to be replaced. On the contrary, in a rail system changing vehicle size is not difficult at all, as more cars can be added to trains. But modifying a route may be practically impossible if it involves changing the infrastructure (tracks, tunnels). And what about fare levels? For instance, is it easier to change fares or frequencies? Probably, the more important challenge for a fare change is not technological, but rather of a political dimension. Technology also imposes constraints on some decision variables. For instance, the maximum vehicle size is much larger on a metro system than on buses. But the maximum frequency of a bus system is larger than that of rail systems, if several bus stops are provided with overpassing lanes and multiple berths. Other technologies – trolleybuses, light railway, cable cars, funiculars, ferries, and so on – have their own constraints on vehicle size and frequencies. These constraints not only depend on the vehicles themselves, but also on the infrastructure associated with the service. For example, the maximum size of a metro train will be constrained by the length of the stations. Multiplying maximum frequency and maximum vehicle capacity yields the capacity of the system, i.e. the maximum number of passengers that can be moved per hour-direction. Speed and cycle time also depend on the technology. For instance, acceleration and deceleration rates may be larger in street modes than in trains running in a completely separated infrastructure. But the number of vehicle stops, and the level of interference with other vehicles, are larger in street services than in those with a separated infrastructure. Cruise speed is constrained by the maximum technical speed (a vehicle design characteristic), but also by a combination of distance between stops (infrastructure) and acceleration/deceleration rates (again, a vehicle design characteristic). When looking at costs, relevant differences among technologies arise as well. Several authors have compared operator cost (e.g. Meyer et al., 1965), total cost (e.g. Tirachini et al., 2010), or both (Allport, 1981) for different technologies as a function of patronage. Figure 9.2 shows a generic example considering all costs (operators and users). All technologies reduce in cost per passenger as patronage grows. The more infrastructure-intensive technologies have higher costs for lower patronage levels, but may become more convenient for larger patronages and/or may become the only available technology for sufficiently large demands.

178  Handbook on transport pricing and financing

Figure 9.2  Representation of average total cost for different technologies

9.3 FROM DESIGN TO PRICING AND THE EFFECTS OF SUB-OPTIMAL PRICING Optimal design and pricing are closely related in the case of public transport. As known, the optimal (first-best) prices are equal to total marginal costs mT, which in this case include those contributed by operators mop and users mU. However, a user is already “paying” the money equivalent to the time he or she spends travelling such that the optimal required money payment is given by the total marginal costs mT Y * minus the average user cost ACU (his/her time value) as in Equation (9.11):

( )

( )

( )

P* = mT Y * - ACU Y * (9.11)



Users’ marginal costs can be written in terms of users’ average costs (Equation 9.12, left side), such that total marginal costs can be written as in Equation (9.12, right side), and the optimal money price can be rewritten as in Equation (9.13). mU =

d (Y × ACU ) dACU = ACU + Y dY dY

(9.12)

dACU mT = mop + ACU + Y dY

P* = mop + Y

dACU (9.13) dY

The first term in Equation (9.13) is the standard marginal cost of the producer (the operator) and the second term says that the optimal money price has to deviate from mop depending on the effect the new traveller has on the time spent travelling by all other users. If ACU decreases (increases) with patronage – i.e. if there are increasing (decreasing) returns to scale – the

Public transport: design, scale, and pricing  179

optimal money price is smaller (larger) than the operators’ marginal cost. This is the link between design, scale economies, and pricing. As seen in the previous section, there are some structural elements that have a distinct influence on scale economies, an important area of analysis where the main question is how optimal design and its associated costs evolve as patronage increases. The first element is the growth of frequency, which reduces waiting costs (the Mohring effect). The second element is vehicle size that increases with a non-negative impact on total boarding–alighting time per vehicle, which makes in-vehicle time increase. The third element is line spacing that also diminishes as patronage grows in space, causing access and egress times to lessen. Finally, we have the evolution of the optimal line structure, which occurs in such a way that directness increases as patronage grows, i.e. the length of the line, the number of stops, and the number of transfers lessen such that in-vehicle time drops. As explained earlier, spacing and directness can be adapted more easily in those transit modes, such as buses, that do not require heavy infrastructure. As a result, under most circumstances, the second term in Equation (9.13) is negative, which makes the optimal price fall short of mop. Note that operator costs exhibit constant returns (conventional buses) or increasing returns to scale (BRT, tram, subway), making mop less or equal to the operators’ average cost. As ACop = ACT - ACU , the optimal money price P* falls short of ACop and a subsidy S* is required for efficiency. The resulting scheme can be represented as in point * in Figure 9.3, where demand equals Y*. Behind the optimal price, there is an optimal design that can be illustrated with the singleline model developed by Jansson (1980) and Jara-Díaz and Gschwender (2009), introduced in Section 9.2.2. There, we showed that both optimal frequency and vehicle size increase with patronage at a decreasing rate (see Equations 9.8) as represented by curves f* and K* in Figure 9.4. From this, the cost function was obtained (Equation 9.9) from which the ACT curve as represented in Figure 9.3 can be obtained. This is how optimal design influences optimal pricing. What about the reverse effect? What would happen if the optimal subsidy was deemed undesirable for whatever reason? In that case, the total inclusive price would have to cover

Source:   Jara-Díaz and Gschwender (2009).

Figure 9.3  Optimal prices under different objectives

180  Handbook on transport pricing and financing

Figure 9.4  Optimal frequency and vehicle size at the different pricing levels the total ACT and the money price would have to be P° in Figure 9.3, reducing both optimal frequency and bus size as represented by points ° in the design spaces of Figure 9.4. If this price was considered too large (e.g. politically unacceptable) and a smaller price P a < P° was imposed, covering costs with only the farebox would be unfeasible under an optimal design, and an inferior design would have to be adopted by further reducing frequency (smaller fleet) and increasing bus size (reducing cost per passenger) as shown by Jara-Díaz and Gschwender (2009). This can be better visualized by introducing the curves f O and K O in Figure 9.4, representing frequency and bus size that would minimize operator cost only, such that the combination (f, K) that makes Pa feasible departs from the optimal towards f O and K O . This is represented by point a in both Figures 9.3 (prices) and 9.4 (design variables). The counterparts of P° and Pa in the design variable spaces reveal the influence of prices on design: sub-optimal prices induce a sub-optimal design. Going back to Equation (9.13), the key element in the previous discussion is the presence of decreasing average user costs. So far this has been linked to design and pricing on the basis of a system where the design elements are frequency and vehicle size only (and where crowding and congestion have been left out of the analysis; see below). As explained in Section 9.2, one key structural component of public transport systems is the organization of transit lines in space, which is yet another strategic element that is subject to accommodation as flows grow. As shown in Figure 9.1, the (optimal) arrangement of lines in space responds to the spatial distribution of demand and to its volume. Let us look at this from the point of view of scale economies, recalling that this is an analysis where products grow along a ray (Baumol et al., 1982), i.e. flows in every OD pair increase by the same proportion. As explained in Section 9.2, the literature on transit networks shows that those line structures involving transfers tend to be appropriate for low levels of overall demand distributed in space, e.g. hub-and-spoke or feeder–trunk systems. As patronage grows, more direct line structures become superior. The change in line structure occurs at specific levels of patronage, such that there are ranges of demand volume where a certain structure remains the best. Within those ranges, “scale economies analysis replicates the case of the single-line models, i.e. frequencies and bus capacities increase with patronage, such that waiting times for users lessen (Mohring effect), and the average cost for operators reduce, which outbalances the diseconomies of scale induced by larger times at bus stops. For synthesis, the degree of scale

Public transport: design, scale, and pricing  181

economies increases locally when line structure changes and lessens afterwards until the next change occurs. And this happens until full directness is achieved; from then on frequencies and vehicle sizes increase until scale economies are exhausted”; this is represented in Figure 9.5 (Fielbaum et al., 2020). When line spacing (density) is added as a design element, the overall picture is the same except that the change in line structure as volume increases occurs at larger volume levels (Fielbaum et al., 2021). The now evident relationship between design, scale, and optimal pricing in public transport has various other dimensions that are necessary to present. One is the introduction of seemingly minor elements in design that might have a potentially large effect on this chain. A good example is boarding–alighting time t. By looking at Equation (9.8), optimal frequency increases and vehicle size decreases with t, and t varies depending both on the payment technology and the number of doors that can be used to board–alight on each vehicle. According to Jara-Díaz and Tirachini (2013), these two factors make t vary from 4.5 to 1 second per passenger; improving (reducing) boarding time induces more frequent but larger vehicles, and the subsidy per passenger increases. Other dimensions that have been studied to different degrees are crowding, differences across periods, and the effect of distance. Accounting for crowding has three effects: a vehicle could be missed because of capacity, and onboard discomfort increases (making Pv a function of the load factor), and boarding–alighting time per passenger also increases in crowded conditions; it is therefore likely that optimal vehicles would be larger. Regarding time periods, the peak period is by definition the one with the largest passenger flow, but usually it also exhibits longer trips and lower speeds due to congestion, which induces large differences in optimal frequencies if a single fleet is used (Jara-Díaz et al., 2017) with yet unclear impact on optimal prices. Pricing by distance can be illuminated by two observations extracted from the singleline model and from the model by Krauss (1991); long-distance users induce larger operator costs but short-distance users impose longer in-vehicle times on the former, such that the two effects operate in different directions regarding optimal prices. Finally, optimal pricing has been presented with single tickets for single trips in mind. The extension to travelcards is far from simple, as shown by Carbajo (1988) and Jara-Díaz et al. (2016),

Source:   Fielbaum et al. (2020).

Figure 9.5  Degree of scale economies as directness increases

182  Handbook on transport pricing and financing

who considered the dual effect of income as purchasing power and car ownership, reaching an optimal card price equivalent to some 30 trips in Santiago, Chile.

9.4 PUBLIC TRANSPORT IN A WIDER CONTEXT 9.4.1 The Role of Modal Competition After presenting the underpinnings of optimal transit pricing based on optimal design and scale economies, it is worth looking at the role of other modes, the urban space and, ultimately, the economy. Let us begin by recalling that (unilateral) marginal cost pricing is a Pareto optimal rule whenever pricing in related markets follows the same rule. If that is not the case, then the optimal price should deviate from the marginal social cost of travelling, following an expression that involves the magnitude of the deviation in the related markets, and the own and cross elasticities as shown in Equation 9.14 (see Jara-Díaz, 2007, pp 123–124, for a comprehensive discussion).



Pk = mk +

å j¹k

¶y j ¶Pk ( m j - Pj ) ¶yk (9.14) ¶Pk

In this second-best pricing formula, mk is the net marginal social cost of public transport use, i.e. the optimal fare under first-best conditions. Similarly, mj represents the marginal social cost in other modes, whilst Pj and yj are their usage fees and equilibrium demand levels, respectively. The large fraction in this equation quantifies the degree of substitution between ¶y mode j and public transport after a marginal adjustment of Pk. Note that k < 0 , so that the ¶Pk corrective term in Equation (9.14) is negative as long as mode j is under-priced, m j > Pj . To grasp the intuition behind Equation (9.14), consider the role of public transport as an alternative to individual car use, as frequently argued in policy debates. What Equation (9.14) implies is that if road pricing is sub-optimal then the marginal social cost of using the car is larger than the implicit price paid by the car users, driving is overconsumed and, to induce the right mode split, pricing transit below its marginal social cost would be justified on efficiency grounds. The strength of adjustment should depend on i. the magnitude of under-pricing, i.e. the gap between the marginal social cost and the actual price of the alternative mode, and ii. the degree of substitution between the two modes. Naturally, the greater the magnitude of the two variables above, the more incentives travellers should receive to trigger their shift towards public transport. Hörcher and Tirachini (2021) provide a recent overview of the methodological toolbox of multimodal pricing. Forming policy recommendations in light of the characteristics and prices of alternative modes requires caution, however, as the magnitude of distortions varies by time, location, and mode. For example, the marginal social cost of car use can be very high in dense cities

Public transport: design, scale, and pricing  183

due to the congestion externality, pushing transit prices down, whilst fuel taxes and road tolls might internalize much of the remaining externalities in rural areas, or off-peak periods, in which cases the second term in Equation (9.14) dissipates. The degree of substitution between car use and public transport is also a context-dependent characteristic of the transport system. A typical value of the empirical literature for the cross-price elasticity of 0.15 implies that a 10% reduction in public transport fares cannot catalyze more than a 1.5% reduction in car use (Börjesson et al., 2017). The disaggregation of passenger groups, in particular with respect to their car ownership status, can help fine-tune the mode choice mechanism of demand models. Cross-price elasticities are indeed not applicable for those who do not own a car, whilst actual car owners’ sensitivity to public transport supply might be higher than what aggregate elasticities would suggest. Regarding other modes such as walking and cycling, the so-called active modes, they are reasonable alternatives to public transport, especially for short distances. Given the advantageous properties of these modes from an environmental externality perspective, they are unlikely to further reduce the right-hand side of Equation 9.14.2 As a conclusion, the combination of the distortions and cross-price effects makes the correction to the optimal (first-best) transit fare very case-dependent. This leaves the first-best price presented in the previous section as the dominant element in the policy discussion. As evident, the best world would be to charge the marginal social cost across all transport modes, correcting for all externalities. This raises other issues that are related to urban transport pricing within the urban system and within the economy as a whole. 9.4.2 Subsidies, the Cost of Public Funds and Design As shown earlier, in the presence of scale economies optimal pricing and capacity provision leads to a financial deficit for transit firms. As subsidies have to be covered by the nation’s public budget, one might wonder whether raising public money through different forms of taxation introduces distortions in the rest of the economy, inducing some form of social cost. Instead of analyzing this issue using general equilibrium models accounting for impacts in all markets, there is a practical (local) approach based upon the so-called marginal cost of public funds (MPCF > 1), a construct that measures the loss incurred by society in raising additional revenues to finance government expenses due to the possible misallocation of resources caused by taxation. This (context-dependent) parameter depends on the level of government the researcher considers, the general structure of the tax system, the pre-existing level of taxes, the value attached to income distribution, and other local characteristics (see Kleven and Kreiner, 2006, for details). The standard modelling practice is that MPCF − 1, which is also called the tax revenue premium, is multiplied by the financial result (net operating revenue) and added to the social welfare objective (see Tirachini and Proost, 2021; Hörcher et al., 2020, for actual implementations). This way the total distortion caused by the tax implications of subsidization is taken into account within the supply problem. If taxation is highly distortionary in an economy, the subsidy-based term is likely to dominate the objective function, and the financially constrained public operator’s supply boils down to a profit-maximizing strategy. If, on the other hand, the MPCF is equal to 1 in general as argued by Jacobs (2008), the level of subsidy follows the idea behind S* in Figure 9.3. It is worth pointing out some specific aspects regarding the use of MCPF in public transport pricing. First, there is a difference between internalizing the positive externalities reflected by economies of scale in transit provision (as explained in Section 9.3), and using transit price to

184  Handbook on transport pricing and financing

counterbalance the reluctance to apply car congestion pricing (as in Equation 9.14). In the first case, not correcting for the positive externality that causes an optimal subsidy would create a distortion that would transmit towards the rest of the economy where transit is a necessary input. In the second case, however, correcting the distortion caused by car congestion by inducing the “right” modal split via prices would cause urban transport to be under-priced everywhere (below social marginal cost), permeating to the rest of the economy as well. A second aspect is the potential use of car congestion prices – aimed at a first-best price for car usage – to contribute to the optimal transit subsidy – aimed at a first-best price for transit usage; this policy within the transport sector would act exactly in the opposite way of a permanently under-priced urban transport as just explained, pointing towards efficient prices everywhere. The third aspect deserves a more detailed explanation. What is the relation between the MCPF and the optimal design? The model behind Figure 9.3 presented in Section 9.3 – based upon Equations (9.7) to (9.10) in Section 9.2 – is quite useful to explain the hidden impact of MCPF. Jara-Díaz and Gschwender (2009) formulated the design problem of a single transit line under an exogenous budget constraint reaching an important result: the budget restriction’s impact on optimal supply is equivalent to a reduction of the user’s travel time valuation by a factor that increases with the exogenous constraint, obtaining: f* =



öö Y æ 1 Pw l æ Pv ç + tY ç + c1 ÷ ÷ (9.15) ÷÷ Tc0 ç 2 (1 + m ) L çè (1 + m ) øø è

and -1



æ 1 Pw öö l l æ Pv K = Tc0Y ç + tY ç + c1 ÷ ÷ (9.16) ÷÷ ç 2 (1 + m ) L L çè (1 + m ) øø è *

where μ is the multiplier of the budget constraint; as explained in Section 9.3, compliance with the restriction implies lower service frequency and larger vehicles than the first-best optimum. If the subsidy S was considered a social cost with a weight (social price) MPCF − 1, the optimal design would correspond to the solution of:

MinVRCop + VRCU + ( MCPF - 1) S. (9.17)

It is relatively simple to realize that the first-order conditions for this problem correspond exactly to the Lagrangean of the problem formulated by Jara-Díaz and Gschwender (2009) with the multiplier μ replaced by MCPF − 1. Then the resulting frequency and vehicle size would be given by Equations (9.15) and (9.16), where the values of time get divided by MCPF, i.e. they lose importance with the size of MCPF > 1, reducing the frequency and increasing vehicle size. This relation with optimal design has never been stated explicitly. For synthesis, complying with the financial expectation of responsible governments is indeed a major force in the political process of public transport pricing. Encapsulating all public finance aspects of subsidies into a transit-specific MCPF parameter that is larger than

Public transport: design, scale, and pricing  185

1 is indeed debatable, it has a profound effect on design, and, as argued above, it is inadequate in the case of the correction of positive externalities. 9.4.3 Social Acceptance and Political Economy As stated in the opening sentence, optimal pricing is not a widely adopted principle in the public transport industry. If alternative financial resources are not available for subsidies, an intuitive approach to full cost recovery would adopt an average operator-cost-based pricing policy. Recalling Figure 9.3 and the discussion in Section 9.3, imposing average operator cost pricing means charging a monetary price P o which means moving to optimal frequency f o < f * and optimal bus size K o < K * in Figure 9.4. In transport production, however, one has to recognize that transport firms produce flows in many OD pairs and in many periods (Jara-Díaz, 2007), such that the total cost depends on the combination of flows in time and space. So the very notion of an average cost figure becomes rather ambiguous as it assumes the existence of a meaningful aggregate in time and space. The number of users, the number of trip legs, the number of passenger-kilometres, or the number of hours, are all aggregates that hide different combinations of flows across OD pairs and periods leading to potentially conflicting average cost pricing rules. On the other hand, the general rule represented by Equation (9.11) survives indeed under a multiple-output view because the notion of marginal total cost is well-defined and the average user cost is the value of time for a user in a given OD pair and period. This is an area in which more research is needed with a multioutput approach considering different flows in time and space in order to obtain optimal prices by distance and period; see recent advances in Hörcher and Graham (2018, 2020) and Jara-Díaz et al. (2017, 2020). The superiority of marginal cost pricing becomes apparent when it is compared against alternative policies with the benchmark metric of social welfare, as illustrated above with the pseudo average cost. Sometimes demand maximization subject to financial and political constraints does appear implicitly in policy debates, including the idea of fare-free public transport (which reduces marginal operator costs by eliminating the fare collection system). In all cases, research results could influence the policy arena by replicating its welfare-oriented modelling exercises with alternative objective functions and showing not only the deadweight loss consequence of loosely specified policy optimization but, perhaps most importantly, the effects on design and, consequently, on users. Finding (and understanding) optimal prices considering space and time is far from trivial when Equation (9.11) is seen rigorously. As pointed out in Section 9.3, looking only at the marginal cost of the operator one could think that short trips should be priced lower, but boarding and alighting by users undertaking short trips induce longer in-vehicle time for those onboard making longer trips, a negative externality that works in the other direction. Equity and income disparity are important elements of pricing analysis when it comes to the investigation of either the distributional effects of public transport subsidies (Börjesson et al., 2020; Matas et al., 2020) or the introduction of travelcards (season tickets) under income and car ownership inequities (Jara-Díaz et al., 2016). Fiscal federalism and political economy offer powerful tools to understand the outcomes of political decisions in a quantitative framework. Public transport policy is influenced by multiple levels of government, and the inconsistency between their objectives and decision-making processes are leading causes of departures from first-best optima. Competition between governments is defined in both horizontal and vertical terms. Horizontally related governments

186  Handbook on transport pricing and financing

do not consider the welfare effect of policies on commuters of neighbouring regions, and they may use pricing as a tool to export the financial burden of service provision to other jurisdictions (see De Borger and Proost, 2015). The conflict between vertically related governments (e.g. a municipality and a national government) stems from overlaps between their tax bases (Proost and Sen, 2006; Boadway and Shah, 2009). For example, public transport fares set by the municipality affect national fuel price revenues, and vice versa. Policy outcomes are also affected by the electoral process and the share of distinct groups of voters, e.g. car owners and captive public transport users, at all levels of government (De Borger and Proost, 2015). 9.4.4 Environmental Considerations and Climate Change Empirical data suggests that public transport is an environmentally friendly alternative to individual car use. Taking the mean occupancy rate of vehicles into account, the sum of air pollution and climate-change-related costs per passenger-kilometre has been estimated to be in a proportion of 1/7/11 for rail, bus, and car users, respectively (European Commission, 2019). When other externalities such as accidents, noise emission, and congestion delays are added to the picture, the proportions are 1/1.3/4.3. These externalities can be included in public transport models, as shown in Equation (9.5), and as has been done for example by Evans and Morrison (1997) for accidents, and Oldfield and Bly (1988) for other externalities. The overall environmental efficiency of public transport is apparent in the comparison of average external costs. However, the burden that individual bus and rail users impose on the environment might differ from this average substantially. Given that most of the environmental costs are linked to the volume of public transport capacity instead of the actual number of users, the external cost of the individual trip depends on its expected impact on capacity provision. Rietveld (2002) shows that the capacity planning policy of the service provider has a crucial impact on the marginal environmental cost of individual journeys. In particular, peak travelling is more likely to trigger capacity expansion as compared to time periods when occupancy rates are low. Rietveld’s approach implies that marginal environmental costs move in parallel with marginal operating costs, as both are determined by the expected impact of individual trips on the available capacity. Thus, differentiating public transport fares based on the occupancy rate of vehicles is justified not only by crowding and operating costs, but also by environmental considerations.

9.5 CONCLUSIONS Pricing in public transport is one of the significant topics of urban policy, both in academic dialogue and in public debates in the media. The price of public transport usage directly determines the surplus that users acquire from this service, and indirectly affects the amount of subsidies that the economy has to provide. From a microeconomic point of view, pricing has a pivotal role in improving the resource efficiency of public transport provision, and microeconomic theory offers a powerful toolbox for deriving a social welfare maximizing fare policy in a transparent and objective optimization process. Given that pricing is often considered as a purely financial aspect in policy debates, our chapter puts an emphasis on the mutual dependency between pricing and the optimization of key technological variables, such as line structure, stop spacing, service frequency, and vehicle

Public transport: design, scale, and pricing  187

size. This interlink is straightforward: pricing affects the ridership of public transport services, and changes in demand may induce adjustment in the supply-side variables above, thus leading to shifts in operating costs on the margin, coupled with system-level efficiency improvement. We also emphasize that scale economies are present in almost every technological attribute of public transport systems: operating and user costs fall as the line structure becomes more direct, the spacing between lines and stops decreases, and the frequency of services increases. The presence of scale economies implies that the optimal price of public transport usage is generally lower than the operating cost per capita, which justifies subsidized service provision. The theory of pricing in public transport has evolved rapidly over the last two decades, but we observe a time lag in the practical implementation of research findings. Whilst several developed and developing countries have implemented clear regulations for transparent infrastructure project appraisal in the form of cost–benefit analysis (CBA), it is remarkable that pricing and subsidy reforms are rarely supported by similar economic analyses in practical decision-making.3 This chapter promotes that pricing policies should come under the same level of scrutiny to unlock the full potential of public transport in cities facing mounting environmental and social challenges.

ACKNOWLEDGEMENTS This research was partially funded by Fondecyt, Chile, Grant 1200157, and Conicyt, Chile, Grant PIA-Basal AFB220003.

NOTES 1.

See the very interesting discussion held by Gwilliam et al. (1985) and Beesley and Gleister (1985) regarding deregulation of buses in Britain, and the more recent theoretical discussion on subsidies by van Reeven (2008) and Basso and Jara-Díaz (2010). 2. Tirachini and Hensher (2012) show in a more complex model in which cost interactions between modes are also present that under strong substitution between active modes and public-transport, the third-best public transport fare might be higher than the second-best one with under-priced car use only. 3. Please refer to Part IV of this Handbook for more specific case studies, including best practices of analytically underpinned pricing policies in Australia, in Chapter 24.

REFERENCES Allport, R.J. (1981). The costing of bus, light rail transit and metro public transport systems. Traffic Engineering and Control, 22, 633–639. Basso, L. & Jara-Díaz, S.R. (2010). The case for subsidisation of urban public transport and the Mohring effect. Journal of Transport Economics and Policy, 44, 365–372. Baumol, W. J., Panzar, J. C., & Willig, R. D. (1982). Contestable Markets and the Theory of Market Structure, New York: Brace. Beesley, M.E., & Gleister, S. (1985). Deregulating the bus industry in Britain - (C) a response. Transport Reviews, 5(2), 133–142. Boadway, R., & Shah, A. (2009). Fiscal Federalism: Principles and Practice of Multiorder Governance. Cambridge University Press.

188  Handbook on transport pricing and financing

Börjesson, M., Eliasson, J., & Rubensson, I. (2020). Distributional effects of public transport subsidies. Journal of Transport Geography, 84, 102674. Börjesson, M., Fung, C.M., & Proost, S. (2017). Optimal prices and frequencies for buses in Stockholm. Economics of Transportation, 9, 20–36. Borndörfer, R., Grötschel, M., & Pfetsch, M.E. (2007). A column-generation approach to line planning in public transport. Transportation Science, 41, 123–132. Byrne, B.F. (1975). Public transportation line positions and headways for minimum user and system cost in a radial case. Transportation Research, 9, 97–102. Carbajo, J. C. (1988). The economics of travel passes: non-uniform pricing in transport. Journal of Transport Economics and Policy, 153–173. De Borger, B., & Proost, S. (2015). The political economy of public transport pricing and supply decisions. Economics of Transportation, 4(1–2), 95–109. European Commission (2019). Handbook on the External Costs of Transport, Version 2019. Policy Paper, Publications Office of the European Union. Evans, A.W., & Morrison, A.D. (1997). Incorporating accident risk and disruption in economic models of public transport. Journal of Transport Economics and Policy, 31, 117–146. Fielbaum, A., Jara-Díaz, S.R., & Gschwender, A. (2016). Optimal public transport networks for a general urban structure. Transportation Research Part B, 94, 298–313. Fielbaum, A., Jara-Díaz, S., & Gschwender, A. (2020). Beyond the Mohring effect: scale economies induced by transit lines structures design. Economics of Transportation 22, 100163. Fielbaum, A., Jara-Díaz, S.R., & Gschwender, A. (2021). Lines spacing and scale economies in the strategic design of transit systems in a parametric city. Research in Transportation Economics, published online. Grava, S. (2003). Urban Transportation Systems: Choices for Communities. New York: McGraw-Hill. Gwilliam, K., Nash, C., & Mackie, P. (1985). Deregulating the bus industry in Britain: B- The case against. Transport Reviews, 5(2), 105–132. Hörcher, D., De Borger, B., Seifu, W., & Graham, D.J. (2020). Public transport provision under agglomeration economies. Regional Science and Urban Economics, 81, 103503. Hörcher, D., & Graham, D.J. (2018). Demand imbalances and multi-period public transport supply. Transportation Research Part B: Methodological, 108, 106–126. Hörcher, D., & Graham, D.J. (2020). The Gini index of demand imbalances in public transport. Transportation, 1–24. Hörcher, D., & Tirachini, A. (2021). A review of public transport economics. Economics of Transportation, 25, 100196. Jacobs, Bas (2008). The marginal cost of public funds is one at the optimal tax system. International Tax and Public Finance, 25, 883–912. Jansson, J.O. (1980). A simple bus line model for optimisation of service frequency and bus size. Journal of Transport Economics and Policy, 53–80. Jansson, J.O. (1984). Transport System Optimization and Pricing. Chichester: John Wiley & Sons. Jansson, J.O. (2005). Bus transport system optimization and pricing. In: Ninth Conference on Competition and Ownership in Land Transport (Thredbo9), Lisbon, Portugal. Jara-Díaz, S.R. (2007). Transport Economic Theory. Netherlands: Elsevier. Jara-Díaz, S., Cruz, D., & Casanova, C. (2016) Optimal pricing for travelcards under income and car ownership inequities. Transportation Research Part A, 94, 470–482. Jara-Díaz, S.R., & Gschwender, A. (2003a). Towards a general microeconomic model for the operation of public transport. Transport Reviews, 23, 453–469. Jara-Díaz, S.R., & Gschwender, A. (2003b). From the single line model to the spatial structure of transit services: corridors or direct? Journal of Transport Economics and Policy, 37, 261–277. Jara-Díaz, S.R., & Gschwender, A. (2005). Making pricing work in public transport provision. In Handbook of Transport Strategy, Policy and Institutions. Emerald Group Publishing Limited. Jara-Díaz, S.R., & Gschwender, A. (2009). The effect of financial constraints on the optimal design of public transport services. Transportation, 36, 65–75. Jara-Díaz, S.R., & Tirachini, A. (2013). Urban bus transport: open all doors for boarding. Journal of Transport Economics and Policy, 47, 91–106.

Public transport: design, scale, and pricing  189

Jara-Díaz, S., Fielbaum, A., & Gschwender, A. (2020). Strategies for transit fleet design considering peak and off-peak periods using the single-line model. Transportation Research Part B, 142, 1–18. Jara-Díaz, S., Fielbaum, A., & Gschwender, A. (2017). Optimal fleet size, frequencies and vehicle capacities considering peak and off-peak periods in public transport. Transportation Research Part A, 106, 65–74. Kleven, H.J., & Kreiner, C.T. (2006). The marginal cost of public funds: Hours of work versus labor force participation. Journal of Public Economics, 90(10–11), 1955–1973. Kocur, G., & Hendrickson, C. (1982). Design of local bus service with demand equilibration. Transportation Science, 16(2), 149–170. Krauss, M. (1991). Discomfort externalities and marginal cost transit fares. Journal of Urban Economics, 29, 249–259. Matas, A., Raymond, J.L., & Ruiz, A. (2020). Economic and distributional effects of different fare schemes: Evidence from the Metropolitan Region of Barcelona. Transportation Research Part A: Policy and Practice, 138, 1–14. Meyer, J., Kain, J., & Wohl, M. (1965). The Urban Transportation Problem. Cambridge, MA: Harvard University Press. Mohring, H. (1972). Optimization and scale economies in urban bus transportation. American Economic Review, 62, 591–604. Mohring, H. (1976). Transportation Economics. Cambridge, MA: Ballinger Press. Newell, G.F. (1979). Some issues relating to the optimal design of bus routes. Transportation Science, 13, 20–35. Oldfield, R.H., & Bly, P.H. (1988). An analytic investigation of optimal bus size. Transportation Research Part B, 22, 319–337. Post, R. (2010). Urban Mass Transit: The Life Story of a Technology. Baltimore, MD: The Johns Hopkins University Press. Proost, S., & Sen, A. (2006). Urban transport pricing reform with two levels of government: a case study of Brussels. Transport Policy, 13(2), 127–139. Proost, S., Van Dender, K., (2008). Optimal urban transport pricing in the presence of congestion, economies of density and costly public funds. Transportation Research Part A: Policy and Practice, 42(9), 1220–1230. Rietveld, P. (2002). Why railway passengers are more polluting in the peak than in the off-peak; environmental effects of capacity management by railway companies under conditions of fluctuating demand. Transportation Research Part D: Transport and Environment, 7(5), 347–356. Tirachini, A., & Proost S. (2021). Transport taxes and subsidies in developing countries: The effect of income inequality aversion. Economics of Transportation, 25, issue C. Tirachini, A., & Hensher, D.A. (2012). Multimodal transport pricing: first best, second best and extensions to non-motorized transport. Transport Reviews, 32(2), 181–202. Tirachini, A., Hensher, D.A., & Jara-Díaz, S.R. (2010). Comparing operator and users costs of light rail, heavy rail and bus rapid transit over a radial public transport network. Research in Transportation Economics, 29(1), 231–242. van Reeven, P. (2008). Subsidisation of urban public transport and the Mohring effect. Journal of Transport Economics and Policy, 42(2), 349–359. Vuchic, V. (2005). Urban Transit: Operations, Planning and Economics. Hoboken, NJ: John Wiley & Sons, Inc.

10. From taxis to ride-hailing: market equilibrium analysis and implications for regulations Xiaolei Wang and Fangfang Yuan

10.1 INTRODUCTION In most cities around the world, taxi service is an indispensable mode in their urban transportation system. As an efficient supplement to public transportation, taxis offer great convenience via door-to-door service and 7/24 availability and provide a flexible alternative when private cars and public transit are not options. In recent years, app-based, on-demand ride services – also known as ride-hailing – have emerged and rapidly grown into popularity around the world and disruptively changed the conventional taxi industry. In 2017, Uber was operating in 84 countries in over 760 cities (Uber Estimator, 2018), with approximately 7 million drivers (Smith, 2018), while Lyft was operating in 200 cities in the United States, with over 315,000 drivers (Li et al., 2019). As of 2020, the global daily trips served by Didi have reached 60 million (BUSINESSWIRE, 2020). Unlike in previous days when people had to stand on the street to hail a taxi, customers can now publish their travel demand to nearby for-hire vehicles through ride-hailing apps, and registered drivers can instantaneously receive the nearby demand information. Starting as solely taxi-hailing platforms that connect customers with taxi drivers, both Didi Chuxing and Uber have evolved to offer a full range of ride-hailing services, e.g., ridesourcing, ride-sharing, and ride-pooling. Except for taxi-hailing services, all the other types of ride-hailing services are served by registered private-car drivers. The distinction is that ride-sourcing and ride-pooling drivers are profit-seekers, while ride-sharing drivers only look for passengers with similar routes and schedules with themselves to share trip costs (Wang et al., 2019). And ride-pooling passengers will be organized to share vehicle space with others during their trips to save vehicle travel miles, while ride-sourcing customers take exclusive riding services (Wang et al., 2021). In nearly all large and medium-sized cities around the world, the traditional taxi markets are subject to strict fare and entry regulations. Local governments set a fixed fare rate that applies uniformly across all taxis. Only licensed drivers are allowed to provide taxi services, and the total number of taxis is usually subject to a rigid quantity limit. Every raise in taxi fare or increase in taxi fleet size would need to go through extreme deference by the reviewing courts. The fast development of ride-hailing services, especially the sourcing service, brings about revolutionary changes to the taxi industry, directly challenges existing regulations toward the taxi industry, and raises the questions of whether the quantity regulation for the taxi industry should be kept or removed, and how should ride-sourcing services be regulated? In an attempt to answer the two questions, this chapter will focus on an economic analysis of the taxi and ride-sourcing markets. We will first review why the traditional taxi industry needs entry and price regulation. With a common agreement on the necessity of regulation, we will then proceed to introduce a classic model for taxi market equilibrium, investigate the 190

From taxis to ride-hailing  191

impacts of taxi fare and taxi fleet size on the taxi market performance, and provide a guide for the design of taxi fare and taxi fleet size for social benefits. With a discussion of the similarities and dissimilarities between the taxi and ride-hailing market, we will then modify the model to portray ride-sourcing market equilibrium, investigate the impacts of the platform’s pricing strategies on ride-sourcing market efficiency, and discuss the necessity of price regulations for the ride-sourcing platform.

10.2 ECONOMIC ANALYSIS OF A TRADITIONAL TAXI MARKET 10.2.1 The necessity of Regulations for the Traditional Taxi Industry Whether the taxi industry should be regulated or not was in hot debate during the 1960s–1970s. Opponents of regulation believe that taxi fare and entry liberalization would generate a more competitive market with higher service quality and lower taxi prices. But the normal forces of competition do not operate in the traditional taxi market. As documented in detail for cities such as San Diego, Seattle, WA, and Sacramento, CA, which deregulated the taxi industry in the 1970s, the overall number of cabs increased by one-third (Schaller, 2007), but the taxi prices did not fall. In some cases, the taxi price even materially rose (Bekken, 2007; Teal and Berglund, 1987). Despite a large number of drivers and customers in the taxi market, each customer meets one driver at once. Suppose drivers set their prices independently, then the best strategy of customers is to accept any price that does not exceed their willingness to pay, because otherwise they have to wait for another uncertain amount of time and be in the same position as they were with the previous drivers. So competition between taxis will not drive the taxi prices down. On the other hand, with a large volume of taxi drivers flowing into the taxi market, taxi drivers’ hourly income was reduced by 34%~48% (Schaller, 2007). Drivers had to work longer to earn sufficient money for their living essentials, and traffic congestion got more severe in urban areas. Frequent driver replacement was observed, and the service quality dropped as new drivers are usually not as efficient in knowing the best route. In the end, frequent driver strikes occurred in these areas, demanding the resumption of taxi license control. So since the 1920s, many US cities went through a full circle from regulation, deregulation, to regulation again (Dempsey, 1996). To design appropriate regulatory policies to fix taxi market failure, continuous efforts by researchers and economists, in particular, have been made to better characterize the economic features of the taxi market at equilibrium, and to investigate the impact of different market configurations and regulatory regimes (e.g., Manski and Wright, 1967; Douglas, 1972; De vany, 1975; Beesley, 1973; Beesley and Glaister, 1983; Schroeter, 1983; Daniel et al., 2003; Salanova et al., 2011, 2014). For a traditional taxi market where vacant taxis cruise on the streets to find customers, it is commonly acknowledged that the taxi market does not fit into the classic Arrow–Debreu model where trade is represented as a costless process and the market is cleared by an equilibrium price (Arrow and Debreu, 1954). Due to information asymmetry and spatial deviation among customers and taxi drivers, there are simultaneously waiting customers and vacant taxis at equilibrium (Lagos, 2000).1 Taxi service efficiency, in terms of taxi–customer meetings during a unit study period, is determined by the number of waiting customers and searching taxis. The more vacant taxis and the more waiting customers at any moment, the higher the meeting rate between customers and taxi drivers. On the other

192  Handbook on transport pricing and financing

hand, the average number of waiting customers in the system is dependent on the taxi demand and the average time that each taxi customer spends on waiting time; and the average number of vacant taxis in the system is determined by the taxi fleet size and demand for taxi service hours. Customers jointly consider the waiting cost and riding cost to decide whether or not to take a taxi service, and the resulting taxi demand, in turn, affects customers’ average waiting time. The aggregate model introduced below is mainly based on Yang and Yang (2011), because it best compromises comprehensiveness and tractability. 10.2.2 An Aggregate Model of Taxi Market Equilibrium Consider a one-hour modeling period for an aggregate taxi market with a given taxi fleet size of N and average taxi price P. Let Nc and Nvt, respectively, be the number of waiting customers and the number of vacant taxis at any moment within the study period. The customer–taxi meeting rate (i.e., the number of meetings during each period), denoted by m, is a function of the number of waiting customers and vacant taxis at any given moment (Yang et al., 2010), i.e.,

(

)

m = M N c , N vt (10.1)



(

)

where ¶M ¶N c > 0 and ¶M ¶N vt > 0 . To enable a more concrete analysis M N c , N vt is usually assumed to follow a Cobb–Douglas-type production function (Varian, 1992) with constant elasticities:

(

)

( ) (N )

M N c , N vt = A N c



ac

vt

at

(10.2)

where A is a positive model parameter that depends on the spatial characteristics of the urban area, and α c and αt, respectively, represent the constant elasticity of meeting rates with respect to the number of waiting customers and available taxis: ac =



¶M N c ¶M N vt t , , a = ¶N c M ¶N vt M

with 0 £ a c , a t £ 1. The meeting function in Equation (10.2) is homogeneous of the degree a c + a t , and it exhibits increasing, constant, or decreasing returns to scale for cases of a c + a t > 1, a c + a t = 1 and a c + a t < 1, respectively. By analyzing the taxi data in Minneapolis, US, and Hong Kong, Schroeter (1983) and Yang et al. (2014) both report the meeting functions to be increasing returns to scale with the degrees varying from 1.13 to 1.16. And similar observations are reported by Nourinejad and Ramezani (2020) based on simulation data with a high goodness of fit. So the meeting rate is generally expected to increase faster than linearly with proportionate increases in the number of waiting customers and vacant taxis. Let wc be the average customer waiting time, and Q be the total taxi demand during the unit study period. By Little’s law, the number of waiting customers can be given by N c = w cQ . On the other hand, with a number of Nc customers waiting for taxis at any moment and meeting taxis at a rate m, the average customer waiting time wc satisfies:

(



)

wc =

Nc 1 Nc = m A

( ) (N ) 1- ac

vt

- at

.

From taxis to ride-hailing  193

Inserting N c = w cQ into the above equation, we can write wc into a function of Q and Nvt: w =A c



-

1 ac

(Q )

1- ac ac

(N ) vt

-

at ac

. (10.3)

When a c = a t = 1, Equation (10.3) reduces to the waiting time function as assumed in Douglas (1972), where the impact of demand on customer waiting time is ignored; and when a c = 1, a t = 0.5, Equation (10.3) reduces to the model derived in Arnott (1996). Let f(u) be the strictly decreasing demand function that relates taxi trip demand Q with the full price of taxi service u, i.e., Q = f(u) with f ¢ ( u ) < 0 and f ( u ) Î éë0, Q max ùû for any u Î R . Taxi customers consider not only taxi prices but also waiting time. Let l be the average taxi riding time per trip and β be the customers’ value of time. Then given taxi price P, the full price of taking a taxi service for a typical customer can be given by u = b w c + l + P , and the total taxi trip demand satisfies:

(

((

)

) )

Q = f b w c + l + P (10.4)



Usually, the taxi price is composed of flag-drop charge F0 and a per-mile charge τ, i.e., P = F0 + tl . The structure of taxi price has an obvious impact on the distribution of taxi trip length, so a deliberately designed taxi price structure can be used as an effective tool to avoid the long taxi queue at the airport and the long customer queue downtown (Scheroeter, 1983). Here, we take P as a whole as all trips are assumed to take the same length. In a cruising taxi market, taxis are either occupied or vacant. By Little’s law, with a total taxi demand of Q and an average taxi riding time of l, the average number of occupied vehicles at any moment is Ql. Then with a taxi fleet size N, the total number of vacant taxis in Equation (10.3) at any moment thus can be given by: N vt = N - Ql. (10.5)



Inserting Equation (10.5) into Equation (10.3), we can rewrite wc into a function of taxi demand and taxi fleet size, i.e.,

w = W ( Q, N ) = A c

c

-

1 ac

(Q )

1- ac ac

( N - Ql )

-

at ac

. (10.6)

And substituting W c ( Q, N ) into Equation (10.4) yields:

((

) )

Q = f b W c ( Q, N ) + l + P . (10.7)

Provided a taxi price P and taxi fleet size N, we can then obtain the taxi demand Q ( P, N ) at equilibrium by solving Equation (10.7). Suppose N - Q maxl > 0, then W c ( Q, N ) is a continuous function of Q on éë0, Q max ùû ; and since f(u) is a continuous function with f ( u ) Î éë0, Q max ùû for all u Î R , the existence of solutions to Equation (10.7) when N - Q maxl > 0 can be concluded based on Brouwers’ fixed point theorem.

194  Handbook on transport pricing and financing

10.2.3 Impacts of Taxi Fare and Taxi Fleet Size on the Taxi Market Performance Now suppose N - Q maxl > 0. By taking the derivative with respect to N on both sides of Equations (10.6) and (10.7), we have:

¶Q ¶W c = ¶N ¶N



¶w c ¶W c = ¶N ¶N





¶Q = ¶P

f ¢b > 0 (10.8) ¶W c 1 - f ¢b ¶Q 1 1 - f ¢b

¶W c ¶Q

< 0 (10.9)

f¢ < 0 (10.10) ¶W c 1 - f ¢b ¶Q

¶w c ¶W c ¶Q = < 0 (10.11) ¶P ¶Q ¶P

where all the inequalities hold because 0 < a c , a t £ 1, f ¢ ( u ) < 0 and from Equation (10.6),

¶W c 1 - a c w c at wc = +l c > 0 (10.12) c ¶Q a Q a N - Ql



¶W c at wc =- c < 0. (10.13) ¶N a N - Ql

So increasing taxi fleet size will lead to a shorter customer waiting time and higher taxi demand, while increasing taxi price will lead to a shorter customer waiting time but lower taxi demand. Let p ( P, N ) be the profit of each driver, i.e., p ( P, N ) = PQ - c . From Equation (10.8), taking N the partial derivative of π with respect to N leads to:

¶p = ¶N

1- ac - at ac

f ¢bw c - Q P (10.14) ¶W c N2 1 - f ¢b ¶Q

From Equation (10.14), if the meeting function shows decreasing or constant returns to scale (i.e., a c + a t £ 1), then the increase of taxi supply will always go against drivers’ benefit (i.e., ¶p ¶N < 0). However, if the meeting function shows increasing returns to scale (i.e., a c + a t > 1), then it is possible for taxi drivers to be better off (i.e., ¶p ¶N > 0) from the

From taxis to ride-hailing  195

addition of taxi supply. So a Pareto-improving win–win situation where an increase in taxi fleet size enhances the benefit of both customers (in terms of reduced customer waiting time) and drivers (in terms of increased driver profit) can occur if and only if the bilateral customer– taxi meeting function exhibits increasing returns to scale (Yang and Yang, 2011). On the other hand, taking the partial derivative of π with respect to P leads to: é ù ú ¶p Q P ¶Q 1 ê Pf ¢ ú (10.15) = + = êQ + ¶W c ú ¶P N N ¶P N ê ¢ 1 f b ê ¶Q úû ë



As implied by Equation (10.15), increasing taxi fare is not always beneficial to taxi drivers.

(

c

When P > Q b ¶¶WQ income.

1 f¢

) , a further increase in the taxi fare rate would reduce drivers’ hourly

10.2.4 System Optimal Taxi Fare and Taxi Fleet Size Let S ( P, N ) be the social welfare under taxi price P and taxi fleet size N, i.e., S ( P, N ) =



ò

Q

0

f -1 ( z ) dz - bQw c - bQl - cN (10.16)

where the first term is the total gross benefit of taxi customers from traveling, the second and third terms are the total time cost of traveling, and the last term is the total operating cost for taxi drivers. Regarding Q and wc as functions of N and P, we have: ¶S ¶Q ¶w c æ ¶W c =P - bQ = ç P - bQ ¶P ¶P ¶P è ¶Q



¶S æ ¶W c = ç P - bQ ¶N è ¶Q



ö ¶Q (10.17) ÷ ø ¶P

ö ¶Q ¶W c - bQ - c (10.18) ÷ ¶N ø ¶N

Let Pso and Nso, respectively, be the system’s optimal taxi fare and taxi fleet size that maximizes P so , N so Î arg max S ( P, N ) . Then from Equations (10.17), (10.18), and (10.13), the first-order conditions i.e., ¶S ¶P = 0 and ¶S ¶N = 0 imply that:

(



)

P so = bQ so

¶W c ¶Q

(

Q so , N so

)

(10.19)

and

( )

cN so = cQ sol + bQ so w c

so

at (10.20) ac

196  Handbook on transport pricing and financing

From Equations (10.19) and (10.20), the total profit of the taxi industry at the system’s optimal taxi fare and fleet size is given by:

(

)

P P so , N so = P soQ so - cN so

é = Q so êb w c êë

( )

(

so

ù (10.21) æ 1 - ac at ö N so cl ú ç ÷ c a c N so - Q sol ø úû è a

)

Since N so N so - Q sol > 1, we have:

(

)

( )

P P so , N so < Q sob w c

so

æ 1 - ac - at ö so ç ÷ - cQ l (10.22) ac è ø

From Equation (10.22), if the meeting functions exhibit constant or increasing scale (i.e., a c + a t ³ 1), then the total profit of the entire taxi industry is negative. The taxi industry thus needs to be subsidized if the government wants to achieve maximal social welfare. At the end of this section, it is worthwhile to point out that despite the deliberate models for the taxi market as described above, regulating the taxi market for a city with an appropriate taxi price and taxi fleet size is not an easy task. From the above model, to determine an optimal taxi fare and taxi fleet size, one needs to have accurate information on the elasticity of taxi demand with respect to taxi cost, the customers’ value of time, and the elasticity of meeting rates with respect to the number of waiting customers and available taxis. Even the best-intentioned regulators are very likely to make bad decisions on taxi fleet size and taxi fare rate due to the lack of the means or resources to obtain the necessary information (Beesley and Glaister, 1983). Furthermore, with the development of economics, these parameters will change with time, so the optimal taxi fare rate and taxi fleet size need to be adjusted from time to time. However, every increase in taxi fleet size or decrease in taxi fare rate may arouse severe strikes of taxi drivers, while an increase in taxi fare rate would be strongly resisted by the public. Therefore, as we observed for many cities in China, despite the fast increase of taxi demand with economic growth, the taxi fleet size grows very slowly. This led to prevalent complaints about the difficulty of hailing a taxi and triggered the growth and popularity of ride-hailing apps.

10.3 ECONOMIC ANALYSIS OF A RIDE-SOURCING MARKET The difficulty of hailing a taxi on street and the easiness of app-hailing led to the popularity of ride-sourcing services all around the world. Smartphone-based ride-sourcing apps enable almost all private-car drivers to provide taxi-like door-to-door transport services, without a need to know the road network very well (with the aid of GPS). In the absence of any regulation, a large number of private-car drivers flow into this market, and brings about significant challenges to the taxi industry. Ride-sourcing service providers, e.g., Uber and Lyft, claimed they would supplement public transit, reduce car ownership, and ease traffic congestion. But more and more evidence is reporting the services as a great contributor to the growing traffic

From taxis to ride-hailing  197

congestion (Tirachini, 2020; Schaller, 2017a, 2017b). Instead of helping people get rid of cars, the services attracted people away from public transit, bicycles, and walking, and generated a substantial number of trips that would not have been made in the absence of a ride-hailing option (Clewlow and Gouri, 2017; Rayle et  al., 2016; Graehler et  al., 2019; Henao and Marshall, 2018; Agarwal et al., 2019). So how to regulate the ride-sourcing market has become a big concern for local governments. While the necessity of price regulation for the taxi industry mainly arises from the information barriers between customers and taxi drivers, the necessity of price regulation for the ride-sourcing market is caused by the existence of a monopoly platform. Unlike in traditional taxi services where customers pay fully and directly to taxi drivers, in ride-sourcing services, the transactions between customers and drivers are through the platform, and the customer’s payment for and driver’s reward from each order is centrally determined by the platform. The platform’s pricing strategies on the customer and driver sides significantly impact the demand and supply of ride-sourcing services and the urban transportation system. In this section, we first introduce a model to depict the ride-sourcing market equilibrium under a given price and reward rate and then discuss how to regulate the platform’s pricing strategies in the case of monopoly. 10.3.1 An Aggregate Model of Ride-Sourcing Market Equilibrium The well-established equilibrium model for the traditional taxi market laid a good foundation for us to characterize the ride-sourcing market equilibrium. However, consider the same market where customers’ average trip time is l, and their value of time is β. Let (p, r) be the platform’s pricing strategy, with p being the average trip price that the platform charges customers and r being the platform’s average payment to drivers for each trip. The ride-sourcing demand Q is also assumed to be strictly decreasing with the full price u, following the demand function f(u), i.e., Q = f ( u ), with f ¢ ( u ) < 0 and f ( u ) £ Q max for any u ³ 0. And with trip cost p and trip riding time l, the customers’ full cost of taking a ride-sourcing service is u = b w c + l + p , where wc represents customers’ waiting time cost. The vehicle supply N is endogenously determined by the drivers’ hourly income rQ N . With an hourly cost of c, additional drivers will flow into the market to provide ride-sourcing service until the hourly income equals c, i.e., rQ N = c . To determine the ride-sourcing demand and supply under a given pricing strategy, now the remaining job is to evaluate the customers’ waiting time wc. Unlike in traditional taxi services where customers wait for taxis by the roadside and drivers search for customers through cruising, the matchings between customers and drivers in ride-sourcing services are implemented online by a ride-sourcing platform. The main advantage of such online matching technology is the higher matching rate given the same number of vacant taxis and waiting customers. But the spatial deviation between customers and taxi drivers cannot be eliminated by information technology.2 So in ride-sourcing service, the customer waiting time wc consists of two components: waiting time for matching (wm) and waiting time for pick-up (wpk), i.e.,

(



)

w c = w m + w pk

The modeling of wm and wpk in literature is dependent on the assumptions of the platform’s matching algorithm. There are generally two types of matching algorithms, i.e.,

198  Handbook on transport pricing and financing

first-come-first-match (FCFM) and batch matching. In FCFM, every newly arrived riding request is assigned immediately, if possible, to the nearest vacant vehicle. So in this case, the matching time is usually perceived as negligible, while the pick-up time is usually assumed as a decreasing function of the number of idle drivers, i.e.,

( )

w m » 0, w pk = g N vt (10.23)



( )

with g¢ N vt < 0 for all N vt ³ 0 (e.g., Ke et al., 2020a, 2021; Dong et al., 2021). It is worthwhile to mention that when driver supply is scarce relative to demand, vacant taxis would be always sent to pick up distant customers under FCFM strategies. Such a phenomenon is called Wild Goose Chase (WGC) (Arnott, 1996). WGC decreases the number of available vehicles both directly by occupying cars for a longer pick-up time, and indirectly as drivers exit the market in the face of reduced income. So the system would collapse with a steep decline in welfare in a vicious circle (Castillo et al., 2017). In the case of batch matching, the platform conducts a bipartite matching after every time interval, and customers and drivers wait for one or several time intervals to get matched. As discussed by Yang et al. (2020), the matching radius (i.e., the maximal allocable pick-up distance within which customers and drivers can be matched or paired) and the matching time interval play an important role in balancing the waiting time for matching and waiting time for pick-up. If the searching radius is set small (big), then it would take longer (shorter) for both drivers and customers to get matched, but once getting matched the pick-up time is short (long). And a similar effect can be achieved by extending the matching time interval to accumulate passengers and vacant vehicles. Different models have been proposed to depict the matching time and pick-up time in batch matching. Some of them are derived by analyzing the matching results based on detailed assumptions of the customer arrivals, driver distributions, and matching rules (e.g., Xu et al., 2017; Yang et al., 2020). Models derived in this way are usually precise, but too complex for economic analysis. Another type of model, as in Schroeter (1983), Zha et al. (2016), and Wang et al. (2016), formulates the customer waiting time into functions of waiting customers and vacant vehicles, based on assumptions of the relationship between matching efficiency and the number of customers and the number of vacant vehicles. Despite the much more efficient matching technology, the matching rate of the ride-sourcing platform is largely determined by the number of waiting customers and vacant vehicles. So the same as for the traditional taxi market, the matching rate me of the ride-sourcing platform can be written as a meeting function of the number of waiting customers Nc and the number of vacant vehicles Nvt:

(

)

m e = M e N c , N vt



where ¶M e ¶N c > 0 and ¶M e ¶N vt > 0. Nourinejad and Ramezani (2020) simulated the matching process when vehicles are in random cruising and greedy dispatching mode, respectively, and estimated the bilateral meeting functions based on the simulation data. Their results show that the meeting function in the e-hailing case can also be well captured by the following Cob–Douglas function:3

(

)

( ) (N )

M e N c , N vt =Ae N c

ace

vt

ate

(10.24)

From taxis to ride-hailing  199

with Ae being a positive constant which depends on the details (e.g., matching radius, matching time interval) of the matching technology adopted by the platform, and 0 £ a ce £ 1 and 0 £ a te £ 1, respectively, represent the constant elasticity of ride-sourcing meeting rates with respect to the number of waiting customers and available taxis. Furthermore, as reported by Nourinejad and Ramezani (2020), the same as that for the traditional taxi service, the meeting function for the ride-sourcing service exhibits increasing returns to scale as well; and the superior efficiency of the online matching technology is portrayed in a larger value of Ae in comparison with A. But the asymmetric impacts of customers and drivers on the matching rate are different in the taxi and ride-sourcing markets. For the traditional taxi market, the addition of a driver to the system has a higher impact on the meeting rate than adding a passenger (i.e., a t > a c ), because vehicles are moving agents searching for stationary passengers. But in the ride-sourcing market, the opposite is observed (i.e., a te < a ce ). With the meeting function in Equation (10.24), following the same procedure as for Equation (10.3), the matching time can be given by:

(

) ( )

w m = W m Q, N vt = Ae



-

1 ace

1- ace

(Q ) a

c e

( ) N vt

at - e ace

. (10.25)

Note that in Equation (10.25) the total number of vacant vehicles Nvt waiting for matching is not N - Ql anymore, because in the ride-sourcing market not all unoccupied vehicles are available for matching – some of them are on the way to pick up passengers. By Little’s law, with an average pick-up time of wpk, there is an average number of Qwpk vehicles in pick-up trips at each instant. The number of vacant taxis available for matching in Equation (10.25) thus is N vt = N - Ql - Qw pk . In the case of batch matching, the customer waiting time for pick-up wpk is dependent on the average distance between each matched passenger and driver. Intuitively, the more waiting customers and vacant vehicles, the shorter the pick-up distance. So wpk follows:

(

)

w pk = W pk N c , N vt , (10.26)



(

where W pk N c , N vt

(

)

)

is a continuous function with ¶W pk ¶N c < 0, ¶W pk ¶N vt < 0 , and

W pk N c , N vt £ w pk for any N c ³ 0 and N vt ³ 0, with w pk being the maximum pick-up time which is usually dependent on the searching radius set up by the platform. Readers can refer to Xu et al. (2017) and Yang et al. (2020) for a detailed formulation of W pk N c , N vt in the case of batch matching. From the above discussion, provided there is a pricing strategy (p, r), the ride-souring demand and supply reach an equilibrium when all the following conditions are met simultaneously:

(

((

)

) )



Q = f b w pk + w m + l + p (10.27)



r N = Q (10.28) c



w m = W m Q, N vt (10.29)

(

)

200  Handbook on transport pricing and financing

(

)



w pk = W pk N c , N vt (10.30)



N vt = N - l + w pk Q (10.31)

(

(

)

( )

where W m Q, N vt = 0 and W pk = g N vt W X

(

in the FCFM case, and W m Q, N vt

)

and

( N , N ) follow Equations (10.25) and (10.26) in the case of batch matching. Taking = ( Q, N , w , w , N ) as arguments defined on the feasible set Ω with:

pk

T

)

c

vt

m

pk

vt

ì 0 £ Q £ Q max , ü ï ï W = íX ý , ïî N ³ 0, w m ³ 0, 0 £ w pk £ w pk , N vt ³ 0 ïþ



and let F(X) be a vector of functions defined by the right-hand side of Equations (10.27) to (10.31). Then the ride-sourcing market equilibrium can be written as a fixed point problem F ( X ) = X on Ω. Suppose rc - l - w pk ³ 0 , then by Brouwers’s fixed point theorem, the existence of solutions to the fixed point problem is guaranteed to exist. However, the uniqueness of solutions is generally not guaranteed. 10.3.2 Impacts of Platform’s Pricing Strategies on Ride-Sourcing Market Performance Suppose rc - l - w pk ³ 0 always holds, we now proceed to investigate the impacts of the platform’s pricing strategies on the ride-sourcing market performance based on Equations (10.27) to (10.31). Taking partial derivative on both sides of (10.27) to (10.31) with respect to p and r yields: ¶Q f¢ = (10.32) ¶p 1 - f ¢b H + B (V + 1)





(

é ¶W pk ê Q ¶Q ê ¶N vt = ¶r ê æ ¶W pk ê c ç1 + Q ¶N vt êë è

¶w pk ¶Q = ( H + VB ) (10.34) ¶p ¶p

Q ¶W pk vt ¶w ¶Q ¶w ¶w ¶Q =B - Z, =V +H + c ¶N pk (10.35) ¶W ¶r ¶r ¶r ¶r ¶r 1+ Q ¶N vt m



ù ú f ¢b - Z (V + 1) ú (10.33) ú 1 - f ¢b é H + B (V + 1) ù ö ë û ú ÷ úû ø

¶w m ¶Q =B , ¶p ¶p



)

pk

m

From taxis to ride-hailing  201

¶N r ¶Q ¶N r ¶Q Q = , = + (10.36) ¶p c ¶p ¶r c ¶r c

where

H=



wm

¶W pk ¶N c

+

(

r c

- l - w pk

)

¶W pk ¶N vt

pk

1 + Q ¶W vt

(10.37)

¶N pk

V=





é ¶W m + ¶Q B= ë



Z =-

(

r c

Q ¶W c

¶N ¶W pk ¶N vt

(10.38)

1+ Q

)

- l - w pk 1 + QV

¶W m ¶N vt

- QH

¶W m ¶N vt

m

¶W ¶N vt

ù û (10.39)

m

Q ¶Wvt ¶N

(

pk

c 1 + Q ¶W vt ¶N

) (1 + Q

¶W m ¶w pk ¶N vt ¶w m

)

(10.40)

(

)

Suppose the platform adopts the FCFM matching algorithm, such that w m = W m Q, N vt = 0 and W pk N c , N vt = g N vt , then Equations (10.37)–(10.40) degenerate to:

(



) ( )

H=

(

r c

- l - w pk

)

g¢ 1+ Qg ¢

, V = 0, B = 0, Z = 0 (10.41)

Inserting Equations (10.41) into Equations (10.32) to (10.36) yields:

¶Q ¶Q Q f¢ = = , ¶p 1 - f ¢bH ¶r c



¶N r ¶Q ¶N Q éê = = , 1+ cê ¶p c ¶p ¶r ë



( (

r c

1 f ¢bH , (10.42) - l - w pk 1 - f ¢bH

)

r c r c

-l -w

pk

)

f ¢b H ùú , (10.43) 1 - f ¢b H ú û

¶w pk ¶Q ¶w pk 1 Q g¢ =H = , . (10.44) ¶p ¶p ¶r 1 - f ¢bH c 1 + Qg¢

As can be seen from Equations (10.41) to (10.44), the gradients of the demand function and the pick-up time function play a key role in determining the impacts of p and r on the demand, supply, and service quality of the ride-sourcing service. 10.3.3 Optimal Pricing Strategies of a Ride-Sourcing Platform Let P ( p, r ) be the profit of the platform defined by:

202  Handbook on transport pricing and financing

(

)

P ( p, r ) = Q ( p, r ) ( p - r ) - k Q ( p, r )



where k(Q) is the platform’s cost for maintaining the service for trip demand Q with k¢ ( Q ) ³ 0, and Q(p, r) is the equilibrium demand that satisfies Equations (10.27) to (10.31) under given (p, r). From Equations (10.32) to (10.40), we can derive that:

¶P ¶Q = ( p - r - k¢ ) + Q (10.45) ¶p ¶p



¶P ¶Q = ( p - r - k¢ ) - Q (10.46) ¶r ¶r

(

)

Let p pm , r pm be the pricing strategy that maximizes the platform profit, i.e.,

(p



pm

)

, r pm Î arg max P ( p, r ) , (10.47)

then the first-order conditions imply that: p pm - r pm = ( k¢ )



(

¶Q p pm , r pm



¶p

pm

pm

æ ¶Q ö - çQ ÷ (10.48) ¶p ø è

) = - ¶Q ( p

pm

, r pm

¶r

) . (10.49)

As can be seen from Equations (10.48) and (10.49), at the profit-maximization (PM) pricing strategy p pm , r pm , increasing p by one unit and decreasing r by one unit have the same effect on Q; and the platform’s commission fee from each order (i.e., p pm - r pm ) equals the marginal pm pm cost of the platform ( k¢ ) plus an additional term - Q ¶¶Qp . In the case of FCFM matching, Equations (10.48) and (10.49) imply that at the profit-maximization price, we have:

(

)

(

Q pm = -





p pm - r pm = ( k¢ )

pm

+

c

( g¢) ( c + b ) pm

)

(10.50)

æ 1 ö pm ç ÷ (10.51) b H pm pm ÷ ( g¢) ( c + b ) çè ( f ¢) ø c

On the other hand, let S ( p, r ) be the social welfare defined by:



S ( p, r ) =

ò

Q

0

f -1 ( z ) dz

(

- bQ w + w m

pk

) - bQl - cN - k (Q ) ,



From taxis to ride-hailing  203

where the first term is the customers’ gross benefit from traveling, the second, third, and fourth terms are the total customer waiting and riding time cost, the fifth term is the total driver cost, and the last term is the platform’s operating cost. From Equations (10.32) to (10.40), we can derive that:

¶S é Q ù ¶Q = p - k¢ - r - ú + Q ¶p êë f ¢ û ¶p



¶S é Q ù ¶Q = p - k¢ - r - ú - Q ¶r êë f ¢ û ¶r

(

)

Let p so , r so be the optimal pricing strategy that maximizes social welfare, i.e.,

(p



so

)

, r so Î arg max S ( p, r )

Then the first-order optimality conditions ¶S ¶p = 0 and ¶S ¶r = 0 imply that: p so - r so = ( k¢ ) -

Q so

so



(

¶Q p so , r so



¶p

(

( ¶Q

¶p )

) = - ¶Q ( p

+

so

so

Q so

( f ¢)

, r so

¶r

so

(10.52)

) (10.53)

)

(

)

So the same as at the PM price p m , r m , at the system optimal price p so , r so , increasing p by one unit and decreasing r by one unit have the same effect on Q. However, the platform’s so commission fee from each order (i.e., p so - r so ) equals the marginal cost of the platform ( k¢ ) so so plus an additional term - Q so + Q so . ( ¶Q

¶p )

( f ¢)

In the case of FCFM matching, Equations (10.52) and (10.53) can be further written into: Q so = -



p - r = ( k¢ ) so



so

so

c

( g¢) ( c + b ) so

(10.54)

( )

so

pk r cb c - l - w (10.55) c + b 1 + Q so ( g¢ )so

Suppose r c - l - w pk ³ 0, then Equation (10.55) implies that p so - r so < ( k¢ ) if Q so < - 1 ( g¢ ) , so so and p so - r so > ( k¢ ) if Q so > - 1 ( g¢ ) . So at the system’s optimal price, the platform’s marginal profit from serving an additional order could either be positive or negative, depending on the relationship between the gradient of the pick-up time function and the SO trip demand. so

so

so

204  Handbook on transport pricing and financing

10.4 CONCLUSIONS The fast development of ride-hailing services challenges the existing regulations of the traditional taxi market and raises a new question – how to regulate ride-hailing services in the mobile internet era. In this chapter, we introduce two equilibrium models for the traditional taxi market and ride-sourcing market, respectively, and examine the impacts of different regulatory variables on the performance of the two markets. As can be seen from the above models, while the advanced matching technology can improve the matching efficiency, the relationships among demand, supply, and service quality in the taxi and ride-sourcing markets are similar. The biggest distinction is that for ride-sourcing services, the prices that customers pay and the rewards that drivers get are determined by the platform. In the case of a monopoly ride-sourcing platform, we examined the impacts of the platform’s pricing strategy on the ride-sourcing market performance, and discussed the distinct optimal pricing strategies from the perspective of platform profit and social welfare. A limitation of the models in this study was that we did not consider the impacts of taxi and ride-sourcing services on urban congestion and the competition between taxis and ride-sourcing services. Interested readers may refer to Xu et al. (2021), Ke et al. (2020b), Wang et al. (2020), and A. Gómez-Lobo et al. (2022). And due to space constraints, the discussions in this chapter are limited to taxi and ride-sourcing services only. Different types of ride-hailing services serve different types of drivers and passengers and exhibit different impacts on urban traffic, therefore requiring different considerations in economic analysis. Interested readers may refer to Wang et al. (2019), Li et al. (2020) and Wang et al. (2021a) for the equilibrium model and pricing strategy design of ride-sharing services, Wang et al. (2016) for taxi-hailing services, and Ke et al. (2020a) and Wang et al. (2021b) for ride-pooling services.

ACKNOWLEDGMENTS The work described in this study was supported by grants from the National Natural Science Foundation of China under Project No. 72022013, No. 71974146 and No. 72061127003 and the Fundamental Research Funds for the Central Universities.

NOTES 1. Interested readers may refer to Lagos (2000) for spatial model that discusses the raise of search frictions in the taxi market. 2. We note that the spatial deviation can be mitigated by repositioning of vacant vehicles over space to balance demand and supply. 3. For the random cruising mode, the meeting rate function is estimated as M ( N c , N vt ) = 0.005( N c )0.47 ( N vt )0.98

REFERENCES Agarwal, S., Mani, D., Telang, R., 2019. The Impact of Ridesharing Services on Congestion: Evidence from Indian Cities. Working paper. Available at SSRN https://ssrn​.com​/abstract​=3410623. Accessed 2 Sept 2019.

From taxis to ride-hailing  205

Arnott, R., 1996. Taxi travel should be subsidized. Journal of Urban Economics 40, 316–333. Arrow, K., Debreu, G., 1954. Existence of a competitive equilibrium for a competitive economy. Econometrica 22(3), 265–290. Beesley, M.E., 1973. Regulation of taxis. Economic Journal 83(329), 150–169. Beesley, M.E., Glaister S., 1983. Information for regulating: The case of taxis. Royal Economic Society, The Economic Journal 93 (371), 594–615. Bekken, Jon-Terje, 2007. “Experiences with (De-)regulation in the European Taxi Industry,” in OECD/ ECMT. BUSINESSWIRE, 2020. Our Extraordinary Year in DiDi Numbers. https://www​.businesswire​.com​/ news​/ home​/20210128006172​/en​/2020​-Our​-Extraordinary​-Year​-in​-DiDi​-Numbers Castillo, J. C., Knoepfle, D., Weyl, G., 2017. Surge pricing solves the wild goose chase. In Proceedings of the 2017 ACM Conference on Economics and Computation (pp. 241–242). Clewlow, R. R., Gouri, S. M., 2017. Disruptive Transportation: The Adoption, Utilization, and Impacts of Ride-Hailing in the United States. Institute of Transportation Studies, University of California, Davis, Research Report UCD-ITS-RR-17-07. Daniel, F.G., 2003. An economic analysis of regulated taxicab markets. Review of Industrial Organization 23, 255–266. Dempsey, P. S., 1996. Taxi industry regulation, deregulation & reregulation: The paradox of market failure. Transportation LJ 1, 73. De vany, A.S., 1975. Capacity utilization under alternative regulatory constraints: An analysis of taxi markets. Journal of Political Economy 83(1), 83–94. Dong, T., Xu, Z., Luo, Q., Yin, Y.F., Wang, J., Ye, J.P., 2021. Optimal contract design for ride-sourcing services under dual sourcing. Transportation Research Part B: Methodological 146, 289–313. Douglas, G.W., 1972. Price regulation and optimal service standards: The taxicab industry. Journal of Transport Economics and Policy 20, 116–127. Gómez-Lobo, A., Tirachini, A., Gutierrez, I., 2022. Optimal prices for ridesourcing in the presence of taxi, public transport and car competition. Transportation Research Part C 137, 103591. Graehler, M., Mucci, R.A., Erhardt, G.D., 2019. Understanding the recent transit ridership decline in major US cities: Service cuts or emerging modes? In 98th Annual Meeting of the Transportation Research Board (TRB), Washington, DC. Henao, A., Marshall, W.E., 2018. The impact of ride-hailing on vehicle miles traveled. Transportation 46(6), 2173–2194. Ke, J.T., Yang, H., Li, X., Wang, H., Ye, J.P., 2020a. Pricing and equilibrium in on-demand ride-pooling markets. Transportation Research Part B: Methodological 139, 411–431. Ke, J., Yang, H., Zheng, Z. 2020b. On ride-pooling and traffic congestion. Transportation Research Part B: Methodological 142, 213–231. Ke, J.T., Zhu, Z., Yang, H., He, Q.C., 2021. Equilibrium analyses and operational designs of a coupled market with substitutive and complementary ride-sourcing services to public transits. Transportation Research Part E: Logistics and Transportation Review 148, 102236. Lagos, R., 2000. An alternative approach to search frictions. Journal of Political Economy 108(5), 851–873. Li, W., Pu, Z., Li, Y., & Ban, X. J. 2019. Characterization of ridesplitting based on observed data: A case study of Chengdu, China. Transportation Research Part C: Emerging Technologies, 100, 330–353. Li, Y., Liu, Y., & Xie, J. (2020). A path-based equilibrium model for ridesharing matching. Transportation Research Part B: Methodological, 138, 373–405. Manski, C. F., Wright, J. D., 1967. Nature of equilibrium in the market for taxi services. Transportation Research Record: Journal of the Transportation Research Board 619, 11–15. Nourinejad, M., Ramezani, M., 2020. Ride-sourcing modeling and pricing in non-equilibrium twosided markets. Transportation Research Part B: Methodological 132, 340–357. Rayle, L., Dai, D., Chan, N., Cervero, R., Shaheen, S., 2016. Just a better taxi? A survey-based comparison of taxis, transit, and ridesourcing services in San Francisco. Transport Policy 45, 168–178. Salanova, J. M., Estrada, M., Aifadopoulou, G., Mitsakis E., 2011. A review of the modeling of taxi services. Procedia and Social Behavioral Sciences 20, 150–161. Salanova, J. M., Estrada, M., Amat, C. (2014). Aggregated modeling of urban taxi services. Procedia Social and Behavioral Sciences 20, 352–361.

206  Handbook on transport pricing and financing

Schaller, B., 2007. Entry controls in taxi regulation: Implications of US and Canadian experience for taxi regulation and deregulation. Transport Policy 14, 490–506. Schaller, B., 2017a. Empty Seats, Full Streets: Fixing Manhattan’s Traffic Problem. Report, Schaller Consulting. Schaller, B., 2017b. Unsustainable? The Growth of App-Based Ride Services and Traffic, Travel and the Future of New York City. Report, Schaller Consulting. Schroeter, J.R., 1983. A model of taxi service under fare structure and fleet size regulation. The Bell Journal of Economics 14(1), 81–96. Smith, C., 2018. Amazing Uber Statistics, Demographics and Facts. https://expandedramblings​.com​/ index​.php​/uber​-statistics/ (accessed February 10, 2018). Teal, R. F., Berglund, M., 1987. The impacts of taxicab deregulation in the USA. Journal of Transport Economics and Policy, 37–56. Tirachini, A., 2020. Ride-hailing, travel behavior and sustainable mobility: An international review. Transportation 47, 2011–2047. Uber Estimator, 2018. Uber Cities. https://uberestimator​.com​/cities (accessed February 10, 2018). Varian, H.R., 1992. Microeconomic Analysis, 3rd ed. W.W. Norton & Company, Inc., New York. Wang, X., He, F., Yang, H., Gao, H. O., 2016. Pricing strategies for a taxi-hailing platform. Transportation Research Part E: Logistics and Transportation Review 93, 212–231. Wang, X., Liu, W., Yang, H., Wang, D., & Ye, J. 2020. Customer behavioural modelling of order cancellation in coupled ride-sourcing and taxi markets. Transportation Research Part B: Methodological, 132. Wang, X.L., Yang, H., Zhu, D., 2019. Driver-rider cost sharing strategies and equilibrium in a ridesharing program. Transportation Science 52(4), 868–881. Wang, X., Wang, J., Guo, L., Liu, W., Zhang, X. 2021. A convex programming approach for ridesharing user equilibrium under fixed driver/rider demand. Transportation Research Part B: Methodological 149, 33–51. Wang, X., Wang, J., Guo, L., Liu, W., & Zhang, X. 2021a. A convex programming approach for ridesharing user equilibrium under fixed driver/rider demand. Transportation Research Part B: Methodological, 149, 33–51. Wang, J., Wang, X., Yang, S., Yang, H., Zhang, X., & Gao, Z. 2021b. Predicting the matching probability and the expected ride/shared distance for each dynamic ridepooling order: A mathematical modeling approach. Transportation Research Part B: Methodological, 154, 125–146. Xu, Z., Chen, Z., Yin, Y., & Ye, J. (2021). Equilibrium analysis of urban traffic networks with ridesourcing services. Transportation Science, 55(6), 1260–1279. Xu, Z., Yin, Y., Zha, L. 2017. Optimal parking provision for ride-sourcing services.  Transportation Research Part B: Methodological 105, 559–578. Yang, H., Leung, C. W., Wong, S. C., Bell, M. G., 2010. Equilibria of bilateral taxi–customer searching and meeting on networks. Transportation Research Part B: Methodological 44(8–9), 1067–1083. Yang, H., Qin, X., Ke, J., Ye, J., 2020. Optimizing matching time interval and matching radius in on-demand ride-sourcing markets. Transportation Research Part B: Methodological 131, 84–105. Yang H., Yang T., 2011. Equilibrium properties of taxi markets with search frictions. Transportation Research Part B: Methodological 45(4), 696–713. Yang, T., Yang, H., Wong, S.C., Sze, N.N., 2014. Returns to scale in the production of taxi services: An empirical analysis. Transportmetrica Part A: Transport Science 10, 775–790. Zha L., Yin, Y.F., Yang, H., 2016. Economic analysis of ride-sourcing markets. Transportation Research Part C 71, 249–266.

11. The economics of airports’ pricing Tiziana D’Alfonso, Martina Gregori, Hugo E. Silva and Leonardo J. Basso

11.1 INTRODUCTION Airports play a central role in urban development by connecting individuals, businesses, and governments, and spurring economic activity; multiple studies have demonstrated a positive effect of air transport on the economy, such as employment, trade, and GDP (Brueckner, 2003; Gibbons and Wu, 2020; McGraw, 2020). Aviation boosts productivity by expanding markets, facilitating competition, and giving access to a wider pool of skilled labor (Carbo and Graham, 2020). The exchange of knowledge and networking, all facilitated by air-link connections and face-to-face communication, is crucial for scientific collaboration and patenting activities (Hovhannisyan and Keller, 2015) and international financial investments (Chen and Lin, 2020).1 Over the past few decades, airports have accommodated increasing operations to support such regional and national growth. Before the COVID-19 pandemic, air traffic had grown significantly in the United States and Europe, even more rapidly in Asia and Oceania, and, more recently in Africa and Latin America. This growth has developed despite local contractions following 9/11 and the economic downturn in 2008 and 2009. It is also expected to grow again and reach pre-COVID-19 levels in 2024 (IATA, 2020). The need to connect remote regions (Fageda et al., 2019) as well as an increased focus on tourism (supported by national and international travel promotion, millennials’ higher propensity to fly, higher living standards and income in major developing economies) and the emergence of low-cost carriers have also contributed to stimulate the traffic growth. At the same time, airport throughput is limited by the existing infrastructure and operational capabilities. At many of the world’s busiest airports – and despite several capacity expansion projects – traffic growth has outpaced capacity increases. According to pre-pandemic estimates provided by Eurocontrol, 1.5 million flights (8% of the forecasted demand) would not be accommodated in Europe by 2040 due to an imbalance between demand and capacity (Eurocontrol, 2018). London’s Heathrow Airport is probably the best well-known case. According to pre-pandemic ACI forecasts (ACI, 2018), around 16 Heathrow-like airports will operate at capacity by 2040. At airports where airline access is restricted (for example, the overwhelming majority of the busiest European airports), the restrictions can result in demand losses or displacement, e.g., to less preferred times of the day or other airports. At airports with largely unconstrained access (e.g., the vast majority of US airports), this often results in very costly delays.2 While a significant cause of delays is traffic volume relative to airport capacity, other causes vary across countries. For example, the weather is a major source of delays in the United States. In contrast, limited aerospace for civil aviation (versus military aerospace) is a major source for China (see Zhang and Czerny, 2012, for a survey). 207

208  Handbook on transport pricing and financing

Indeed, airports are associated with several externalities, with congestion, pollution, and noise being the most important. The nationwide impact of flight delays in the United States has been estimated at over $30 billion in 2007 (Ball et al., 2010). Recently, the Federal Aviation Administration (FAA) estimates that the airline cost for an hour of delay ranges from about $1,400 to $4,500 and that passenger time valuations range from $35 to $63 per hour (FAA, 2020), with delays affecting ticket prices too.3 These figures confirm that air transport delays are costly, highlighting the importance of airport congestion policies. Other than congestion, air traffic is also associated with environmental externalities. While the aviation industry is getting more fuel-efficient, overall emissions have risen as air travel volume has increased. Excess airplane idling, i.e., residual daily taxi time due to delays, has significantly impacted emissions levels (Schlenker and Walker, 2016). Despite dramatic reductions over the years in the noise produced by individual aircraft, airport noise also remains a critical public policy issue today. Real estate could be negatively affected by noise exposure (McMillen, 2004), justifying measures such as government economic compensation for households (Mense and Kholodilin, 2014). In this chapter, we review the literature about pricing and regulation concerned with dealing with these negative externalities. The literature has devoted significantly more attention to congestion than to other externalities, and, consequently, we focus more on congestion. There are two types of measures to deal with congestion: supply-oriented measures such as expanding airport capacity and demand-oriented measures. Furthermore, demand-oriented policies can be classified as price-based (e.g., congestion tolls) and quantity-based policies (e.g., slot control). Supply-oriented measures include adding infrastructure to existing airports, implementing operational and technological innovation in air traffic control and management, and constructing new airports. The latter involves many stakeholders, often having conflicting objectives. Moreover, airport capacity investments take a very long time from gestation to implementation, and tend to be expensive, lumpy, and subject to significant budget and environmental constraints. As a result, they are less of a solution for decreasing externalities. In the following section, we discuss the literature about demand-oriented price-based measures. We focus on congestion pricing, understood as the welfare-maximizing pricing of the negative externality between flights. We also discuss second-best pricing and other issues related to pricing, such as private airports’ pricing, airport competition, capacity investments, regulation, and the role of non-aeronautical operations. The airport pricing literature is vast, yet the congestion tolls have not been implemented. Section 11.3 discusses the literature on quantity-based approaches to airport congestion management. Unlike pricing, slot control, which is a direct way of limiting the number of flights, has been implemented in several countries. We also briefly summarize the current practice, including slot control, trade, and auctions. Section 4 complements the review by examining how externalities other than congestion, such as environmental externalities (noise, emissions), could be incorporated into airports’ pricing schemes. As this chapter focuses on airports, we do not discuss relevant topics such as airline pricing, price discrimination, yield management, the economics of baggage fees, and network pricing. We conclude in Section 11.5 by providing what we think should be the lines of future research.

11.2 PRICE-BASED DEMAND-ORIENTED MEASURES The current practice of levying takeoff and landing fees based on the aircraft’s weight that are constant throughout the day has been criticized since early contributions by Levine (1969)

The economics of airport pricing  209

and Carlin and Park (1970). Since then, several theoretical and empirical analyses have investigated efficient runway pricing policies and their potential benefits and have argued in favor of their implementation. The difference with the extensive congestion pricing literature developed for road traffic lies in that airlines usually control large shares of traffic at some airports and have market power, in contrast to road drivers who control a single vehicle. Airlines are large players in the aviation markets and, therefore, are expected to internalize the congestion imposed on themselves and their passengers and react differently from road users when facing congestion charges. Brueckner (2002), the first to formally model congestion and Cournot competition together, shows that welfare-maximizing congestion tolls should only account for the congestion cost imposed on other carriers. This became known as the self-internalization hypothesis, which implies that an airline’s congestion charge should be proportional to the aggregate market share of the other airlines, excluding its own flights. This result remains valid under several different settings, including networks, when airport capacity investments are included and when airlines offer differentiated products and schedule delay costs are considered (e.g., Basso, 2008; Brueckner, 2005; Pels and Verhoef, 2004; Zhang and Zhang, 2006). For a more detailed review of the airport pricing literature offering these insights, we refer the reader to Zhang and Czerny (2012).4 On the empirical side, the issue has been more contentious. Daniel (1995) offers the first analysis on the potential for self-internalization by investigating the dynamic scheduling behavior of Northwest Airlines at the Minneapolis-St. Paul airport in 1990. Daniel and Harback (2008) follow a similar approach but consider a much broader database, including traffic patterns and delays from 27 major US airports. These analyses provided evidence against the hypothesis of the internalization of self-imposed delays. Empirical evidence in contradiction to the theory on self-internalization is also offered by Van Dender (2007). Based on a dataset from 55 US airports from 1998 to 2002, he found that delays increase when markets become more concentrated. The argument is that dominant airlines behave as Stackelberg leaders and the rest as a competitive fringe, i.e., a group of aircraft that adjust their operations to fill the capacity and keep the delays constant. Given this behavior, the leader has no incentive to internalize delays. The argument is analytically further developed in Brueckner and Van Dender (2008). Moreover, Brueckner (2002) and Mayer and Sinai (2003) find empirical evidence of the inverse relationship between delays and airline market concentration at an airport and hence internalization of self-imposed congestion effects. Rupp’s (2009) findings are mixed. The analysis shows that excess travel times are reduced by carrier market power, which indicates that self-internalization exists and is in line with the results derived by Mayer and Sinai (2003). He also finds that punctuality is reduced by carrier market power, which may stand in contrast to self-internalization. Ater (2012) finds that periods of high flight volume are more extended when the share of flights operated by the hub-airline is greater, and these more extended periods exhibit shorter delays. These results imply that hub-airlines consider the impact of their scheduling decisions on the congestion that they bear. The internalization of delays is central as it determines the scope and efficiency of congestion pricing at an airport. While the evidence may seem contradictory, Brueckner and Van Dender (2008) were the first to reconcile the two sets of evidence theoretically. They show that a Stackelberg leader’s degree of internalization depends on the substitution pattern between leader and fringe, and it can be anywhere between full self-imposed delay internalization and no internalization at all. Silva and Verhoef (2013) complement the analysis by pointing out that results may be reconciled even in the absence of a competitive fringe or a leader firm.

210  Handbook on transport pricing and financing

They argue that airline behavior is more consistent with a Bertrand competition model with differentiated products. In this case, airlines always internalize less than the self-imposed delays. Also, the equilibrium congestion levels approach the levels consistent with no internalization at all, the closer the airlines’ products are. For example, for a medium degree of substitutability, airlines may internalize less than half of their self-imposed delays, a result consistent with empirical tests that reject internalization and fail to reject no internalization (Daniel, 1995; Daniel and Harback, 2008). However, despite being low, the share of internalized delays always increases with market concentration, validating the empirical findings that support internalization (e.g., Mayer and Sinai, 2003). The airport pricing literature has moved past the internalization debate to develop new ideas and policy insights in the last ten years. The first strand that we review builds on Daniel (1995) and studies the dynamics of airport runway congestion over time. Daniel (2009) and Silva et al. (2014a) study the interaction between a Stackelberg leader and a competitive fringe using the deterministic Vickrey bottleneck model.5 They show that the overall internalization pattern is consistent with results that use static models of congestion, such as Brueckner and Van Dender (2008). Nevertheless, the pricing implications are different. The optimal congestion charges are the same for both leader and fringe and do not depend on the degree of internalization of delays in the untolled equilibrium. Theoretical advances in other models of competition in the dynamic congestion strand of the literature are scarce. According to Silva et al. (2017), the main problem is that the workhorse structural model of dynamic congestion, the Vickrey bottleneck model, cannot be easily extended to simultaneous competition. Indeed, Silva et al. (2017) show that a pure strategy Nash equilibrium (PSNE) does not exist in the simplest bottleneck model in which two identical firms schedule vehicles.6 11.2.1 Second-Best Pricing A second strand of the literature deals with second-best congestion pricing. Among the many second-best problems that exist, we focus here on the existence of financial constraints, limitations on differentiating tolls over firms, and limitations on the number of instruments used. Regarding the financial constraints, it is well known that the self-financing theorem that holds for road pricing (Mohring and Harwitz, 1962) does not hold for airports because of the presence of airline market power. As an alternative, Basso (2008) shows that welfare maximization subject to cost-recovery performs quite well, achieving congestion levels similar to a private unregulated airport without inducing such significant traffic contraction. Verhoef (2017) proposes a regulatory scheme that restores self-financing under specific technical conditions. Importantly, unlike in the original case for road provision, self-financing holds when airport capacity provision exhibits increasing returns to scale, which is most likely the case for airports. We discuss this topic further in the following subsection, where non-aeronautical revenues are considered. Also, Basso and Zhang (2010) show that if airport revenues matter, the choice of price versus quantity-based congestion management mechanisms is material. One of the main contributions of uniform pricing is Morrison and Winston (2007). They develop a simple structural model to estimate the welfare effects of congestion tolls and calibrate it with US national trip data. They simulate the impact of charging the (uniform) marginal congestion damage to all airlines as if they did not internalize congestion at all, the so-called atomistic toll. Their main result is that the atomistic and the optimally differentiated

The economics of airport pricing  211

congestion tolls achieve similar benefits as the share of internalized delays is small. However, the main simplifying assumptions, absence of airline market power, intertemporal substitution, and substitute products are too strong to overlook when generalizing the results. On the other hand, the studies that have looked into the efficiency of atomistic congestion tolls without such strong assumptions are theoretical (as opposed to empirical) and use stylized settings. For example, Silva and Verhoef (2013) consider only one market with symmetric airlines and calibrate the model to reflect as much as possible realistic values. Lin (2019) studies welfare-maximizing uniform airport pricing for asymmetric airlines, but with similar simplifications as Silva and Verhoef (2013). Finally, regarding the limitations of the pricing instruments, Zhang and Zhang (2006) noted that the common assumption of constant load factors implies that per-passenger and per-flight tolls are equivalent. Silva and Verhoef (2013) are among the first to depart from this assumption and study the implications of using one instrument over the other. Intuitively, distortions caused by flight congestion should be corrected through per-flight charges, and the inefficiency caused by airline market power exertion must be corrected with per-passenger subsidies. Naturally, the latter is practically impossible and the relevant question is how well per-flight tolls perform when market power effects are not corrected directly. They find that if the airport is not too concentrated, the relative efficiency of per-flight tolls relative to the first-best with subsidies can be significant. For example, with five symmetric firms with equal market shares, the relative efficiency is between 20% and 60%. Importantly, they show that charging atomistic tolls reaches at least 75% of those benefits. This result is in line with Morrison and Winston (2007), implying that the differentiation of congestion tolls may not be of crucial relevance. Czerny and Zhang (2015) and Czerny et al. (2017) contribute to the topic of per-passenger versus per-flight charges by studying under which conditions airlines and welfare-maximizing airports under a cost-recovery constraint prefer to increase per-passenger charges and decrease per-flight tolls. Their analysis suggests that the current trend of increased per-passenger tolls and reduced per-flight tolls should not concern regulators, as long as congestion delays are relatively unimportant. Finally, several articles deal with airport pricing considering airlines’ networks and route structure choice (e.g., Flores-Fillol, 2010; Silva et al., 2014b; Lin and Zhang, 2017). Although these are outside the scope of this review, a few of them show how the optimal per-flight and per-passenger charges should account for the route structure choice. 11.2.2 Pricing of Private Airports, Airport Competition and Capacity Investments Airports have traditionally been owned and managed as government entities. Yet, starting with the privatization of some UK airports in 1987, more and more airports worldwide have been (partially) privatized. The rationale behind the privatization of airports is that they would implement more efficient congestion pricing schemes and better invest in capacity. The main concern has been that privatized airports are essentially local monopolies and will utilize their market power. But it has been argued that regulation may be unnecessary because a private unregulated airport would not induce large allocative inefficiencies, since price elasticities are low, and because potential collaboration between airlines and airports – or airline countervailing power – would put downward pressure on market power. The theory shows that an unregulated profit-maximizing airport would overcharge for the congestion externality and, compared to the first-best, would induce large allocative

212  Handbook on transport pricing and financing

inefficiencies and deadweight losses. Accordingly, the private airport charge is excessive from a social viewpoint (Zhang and Zhang, 2003, 2006; Basso, 2008).7 As for the effects of privatization on capacity decisions, Oum et al. (2004) show that the capacity investment rule of a private airport is the same as the first-best rule when carriers are atomistic. This is that capacity is added until the marginal costs equate to the marginal benefits due to reduced delays to passengers and airlines. However, Basso (2008) and Zhang and Zhang (2010) show that in the general case of non-atomistic carriers, a private airport, for a given traffic volume, will overinvest in capacity, whether unregulated or regulated. The intuition is that an expansion in the capacity that reduces congestion raises the passengers’ willingness to pay and cuts the carriers’ ticket prices due to the decline of the marginal congestion cost that is internalized by the airlines. Both effects can contribute to a higher airport charge without changing the total demand. Consequently, when carriers have market power, reducing congestion is an effective way to increase the airport’s revenue. For this reason, the result also holds for a public airport under a cost-recovery constraint, as Zhang and Zhang (2010) show. Nonetheless, the growth of air transport demand and low-cost carriers have contributed to increased competition for passengers and carriers between airports. Indeed, there are several multi-airport regions such as Greater London, New York, and the San Francisco Bay Area where airports cannot be considered as monopolies. Basso and Zhang (2007b) are among the earliest theoretical studies on the topic. Using a Hotelling-type model, they show that duopoly facilities have lower charges and invest less in capacity than a monopolist when the facilities first decide on capacities and then on prices.8 This is important as both the airport charge and the capacity investment move toward the socially optimal ones. However, the extent to which competition in multi-airport regions reduces the need for regulation is ultimately an empirical matter. Yan and Winston (2014) study the potential effects of privatization on airport prices and welfare for the San Francisco metropolitan area. They develop an empirically tractable model of competition that considers the three airports (Oakland, San Francisco, and San Jose) and the network of domestic flights. Their main result is that competition can lead to large welfare gains: if there is a competitive environment for airports (i.e., privatized to different owners), this allows airlines to negotiate with airports to prevent excessive airport charges, which stimulates differentiated services by larger and smaller airports. Nevertheless, the gains come at the expense of consumers whose welfare decreases in all studied privatization scenarios. Another relevant strategic interaction between private airports occurs when they are complements in international markets. Early studies by Mantin (2012) and Matsumura and Matsushima (2012), later extended by Lin and Mantin (2015), analyze the pricing decision of international airports that connect two countries and serve as hubs for local markets. There are three types of passengers (national, international hub-to-hub, and international one-stop) and two types of airlines, regional carriers, and hub-airlines. Moreover, they derive the pricing decisions of local welfare-maximizing and private airports and study the governments’ incentives to privatize them. The main insight of the analysis is that a country may benefit from privatizing the hub-airport (alone or together with the local airport) if the international hub–hub market size is sufficiently large with respect to the domestic hub–local market size. The intuition is that airports cannot price discriminate and that the country benefits from extracting surplus to foreign passengers in the international hub–hub market. When the international hub–hub market is large and the other country privatizes its hub, privatizing its own hub-airport is a dominant strategy. Furthermore, they show that when countries privatize,

The economics of airport pricing  213

governments are in a Prisoners‫ ׳‬Dilemma type of equilibrium as both would be better off by keeping all airports under public ownership.9 11.2.3 Regulation and Non-Aeronautical Operations Given the growing pressure on airports to self-finance their operations, airports have increasingly depended on revenues generated by non-aeronautical businesses, including side activities like terminal concessions, advertising, car rentals, car parking, and land rent. For the last three decades, non-aeronautical revenues have grown faster than aeronautical ones; thus, they have become the primary income source for many airports (Zhang and Czerny, 2012). Commercial operations tend to be more profitable than aeronautical operations, owing partly to the “locational rents” enjoyed by an airport and partly to prevailing regulations and charging mechanisms (Starkie, 2021). While aeronautical operations are subject to various forms of regulation – either explicitly or implicitly – commercial operations are usually unregulated. One consequence of this profit disparity is that the profits made from commercial activities may be used to cross-subsidize aeronautical operations, diminishing the need for government aid (see, e.g., Zhang and Zhang, 1997; Zhang and Zhang, 2003). Traditionally, the demand for side services is complementary to the demand for aeronautical services, in that the more people use the airport, the higher the non-aeronautical revenues are. Moreover, most consumers’ decision to fly is based on the full price of the aeronautical service only; they do not consider the price of side activities in their travel decisions.10 With atomistic airlines, Zhang and Zhang (1997) find that the first-best optimal solution (for an unconstrained public airport) involves marginal cost pricing on side services. Since a smaller aeronautical charge increases the demand for aeronautical and non-aeronautical services, the social marginal cost pricing on core activities would have an additional markdown. Conversely, the second-best optimal price (for a budget-constrained public airport) of non-aeronautical services would be higher than the related marginal cost, showing that profits would be made in the non-aeronautical business. Therefore, non-aeronautical operations would subsidize aeronautical operations. If the airport were not allowed to make profits from its non-aeronautical operations but were still asked to self-finance its operations, this would lead to a smaller level of social welfare. The downward pressure on the aeronautical charge elicited by non-aeronautical services is also confirmed in the presence of airline market power (Zhang and Zhang, 2010) and has been discussed in the case of different passenger types (Czerny and Zhang, 2014; D’Alfonso et al., 2013) and terminal congestion (Wan et al., 2015). Notably, in the absence of passenger-type-based price discrimination by airlines, it can be useful to increase the airport charge to protect passengers with a great relative time value. For example, to protect business passengers from excessive congestion caused by passengers with a low relative time value, e.g., leisure passengers. Extensions include price discrimination between business and leisure passengers when operating costs are the same for all passengers (Czerny and Zhang, 2014). Zhang et al. (2010) and Fu and Zhang (2010) deal with the issue of non-aeronautical revenue sharing with multiple airlines and airports and its effects on social welfare, but they focus on non-congested airports. Bracaglia et al. (2014) and D’Alfonso and Bracaglia (2017) discuss the role of service consumption and how it could be considered independent of traveling. This is relevant because it implies that a markdown would not be appropriate. Zhang and Zhang (2003) and Oum et al. (2004) find that, while private airside prices diminish, they decrease less than the prices in a public airport with non-aeronautical activities, and

214  Handbook on transport pricing and financing

that this is the case for both the first-best pricing and second-best pricing. Therefore, non-aeronautical revenues would not be a valid argument for deregulation once an airport is privatized. The intuition of the result is simple: a private airport would care about the extra profits it can make from side activities; a public airport maximizing social welfare, however, would care about non-aeronautical profits and the consumer surplus induced. Consequently, the decrease in the aeronautical charge would be more significant in the public case: non-aeronautical revenues would increase the gap between private and public airside charges. Czerny (2006), Oum et al. (2004), Lu and Pagliari (2004), Yang and Zhang (2011), and Kidokoro et al. (2016) analyze the effects of alternative mechanisms of regulation on the performance of private airports, with a particular focus on how revenues from non-aeronautical services should be dealt with. Oum et al. (2004) conclude that the dual-till rate of return (ROR) induces the airport to invest optimally in capacity, while a price cap (either single-till or dual-till) induces underinvestment in capacity, worsening the congestion problem.11 Indeed, the airport cannot recoup fully from its investment in capacity – which reduces congestion and increases the users’ willingness to pay – because the price is capped. Lu and Pagliari (2004) and Czerny (2006) argue that single-till price-cap regulation is socially more desirable than a dual-till approach at non-congested airports, while the dual-till scheme dominates at congested airports. In principle, under the single-till price-cap regulation, the problem is that the more profit the airport makes from concessions, the smaller the allowed aeronautical charge would be in future revisions of the cap, even if traffic grows and congestion builds. Yang and Zhang (2011) demonstrate that when airport congestion is not a major problem, single-till price-cap regulation dominates dual-till price-cap regulation concerning social welfare. However, they identify that dual-till regulation performs better than single-till regulation when significant airport congestion occurs. Kidokoro et al. (2016) confirm that dual-till regulation yields higher welfare than single-till regulation, as long as the profit from nonaeronautical services is positive.

11.3 QUANTITY-BASED DEMAND MANAGEMENT MECHANISMS: SLOTS An airport slot is an authorization to either take off or land on a particular day during a specified period (usually 15 minutes). The granting of a slot means the airline may use the full range of elementary infrastructure services (both airside and groundside) necessary for operating a flight at a given time. The use of slots to manage airport congestion has been gaining support due to the constraints on airport expansions and the lack of backing for congestion pricing. Currently, around 350 airports are slot coordinated, most being in Europe. The main events related to slot regulation and usage are displayed in Figure 11.1 as a timeline, and Table 11.1 summarizes the use of slot coordination worldwide by region as of January 2022. The intuition and appeal of slot coordination at airports are that congestion externalities can be reduced to an optimal level by fixing a total desired flight volume. However, the practice of allocating slots has not been in line with this target. In the United States, slot administration at coordinated airports generally follows the process laid out by IATA in the Worldwide Slot Guidelines. The method relies heavily on “historical slots,” or slots carriers have operated for a qualifying duration under the FAA rules. The carriers submit their schedules or slot requests, and the FAA responds to their proposal. In the European Union, the Slot Allocation

215

Figure 11.1  Timeline of slot regulation main events

216  Handbook on transport pricing and financing

Table 11.1  Countries with slot-coordinated airports by region (as officially declared to IATA). The number of airports per country is in parenthesis Asia Pacific

Europe

Middle East and Africa

North Asia

The Americas

Australia (8)

Austria (6)

Kosovo (1)

Bahrain (1)

China, P.R. (22)

Bermuda (1)

Bangladesh (1)

Belgium (1)

Lithuania (1)

Egypt (6)

Chinese Taipei (2)

Brazil (13)

Cambodia (2)

Bulgaria (1)

Luxembourg (1)

Ghana (1)

Hong Kong (SAR), China (1)

Canada (5)

India (10)

Croatia (4)

Malta (1)

Kuwait (1)

Macau (SAR), China (1)

Cayman Islands (1)

Indonesia (2)

Cyprus (2)

Netherlands (3)

Morocco (2)

 

Colombia (1)

Japan (6)

Czech Republic (1)

Northern Macedonia (1)

Qatar (2)

 

Cuba (4)

Malaysia (1)

Denmark (2)

Norway (14)

Republic of Cape Verde (3)

 

Mexico (1)

New Zealand (4)

Faroe Islands (1)

Norway/Svalbard & Jan Mayen Islands (1)

Saudi Arabia (13)

 

Peru (1)

Pakistan (4)

Finland (1)

Poland (3)

Seychelles (1)

 

Turks & Caicos Islands (1)

Pakistan (4)

France (8)

Portugal (5)

South Africa (3)

 

United States (7)

Papua New Guinea (1)

Germany (15)

Russian Federation (4)

Sultanate of Oman (1)

 

 

Philippines (1)

Greece (20)

Slovakia (1)

Tunisia (2)

 

 

Republic of Korea (3)

Greenland (3)

Slovenia (1)

United Arab Emirates (3)

 

 

Republic of Maldives (1)

Hungary (1)

Spain (27)

 

 

 

Singapore (1)

Iceland (1)

Sweden (8)

 

 

 

Sri Lanka (1)

Ireland (1)

Switzerland (2)

 

 

 

Thailand (9)

Israel (2)

Turkey (8)

 

 

 

Vietnam (2)

Italy (24)

Ukraine (2)

 

 

 

United Kingdom (18)

 

 

 

Source:   www​.iata​.org​/contentassets​/4ed​e2aa​bfcc​14a5​5919​e468​054d714fe​/wasg​-annex​-12​.7​.xlsx (last retrieved on January 27, 2022).

The economics of airport pricing  217

Regulation (EC Regulation 95/93, as amended by Regulation 793/2004) defines the mandatory rules for coordinated airports (EC, 2004). Although there are no property rights, there are grandfather rights in using slots. If an air carrier has used some slots for at least 80% of the time during a season, it is entitled to use the same slots in the next corresponding season; otherwise, slots become free and may be allocated to new entrants. The outcome of slot allocation mechanisms based on historical rights can be far from economic efficiency. The use-it-or-lose-it rule may induce airlines to slot hoarding – being reluctant to cede slots for fear of a rival’s entry – and inefficient use of slots. Examples are behaviors such as “slot babysitting,” in which airlines only use them for the minimum required amount to retain their grandfather rights (Ball et al., 2018; Dempsey, 2001). In practice, few carriers hold many available slots and operate several flights merely to comply with the use-it-or-loseit rule (Madas and Zografos, 2006, 2008). As a result, market concentration may increase, raising competition concerns. Market mechanisms for slot allocation have been widely discussed in the economic literature (Brueckner, 2009; Button, 2008; Verhoef, 2010). In principle, well-designed auctions may ensure that the resource is assigned to those bidders that will generate the highest value from managing it. Secondary trading may improve efficiency by creating an opportunity cost for holding slots since airlines that use slots forego revenues from selling (the right to use) slots. Despite their benefits, market mechanisms also have some crucial drawbacks. Above all, high private valuations do not necessarily reflect the social value of the resource. While there is the need to achieve economic efficiency in allocating valuable resources (thereby maximizing a weighted sum of consumers’ and producers’ surplus), this goal can hardly be attained in practice as consumers do not participate in auctions. Thus, the outcome of an auction is primarily driven by bidders’ profits rather than social welfare (Klemperer, 2002). Moreover, Hoppe et al. (2006) stress the importance of market structure as a determinant of bidders’ valuations. Finally, potential buyers and sellers do not necessarily meet each other due to a lack of information (Aravena et al., 2019; Avenali et al., 2015). Table 11.2 summarizes the use of secondary markets since their first use. As the last column of Table 11.2 reveals, slots are very valuable to airlines. However, the use of trading and leasing has been limited, and auctions have occurred very rarely. For example, in 2008, the US FAA initiated a proposal to auction off 10% of the slots at New York City’s three major airports, which was met with criticism from airlines and IATA. In 2009, the Obama Administration rescinded the plans for slot auctions after the US Court of Appeals stayed the proposal in December 2008 (IATA, 2010). An obvious question is whether these two mechanisms, pricing versus slot policies, are equivalent. Brueckner (2009) shows that distributing slots and allowing free trade and a slot auction would achieve the first-best outcome in the absence of market power and information asymmetries. On the other hand, slot sales – which are treated as a uniform congestion toll – are inferior to congestion pricing and slot trading/slot auctions except when the airlines are symmetric in size. Brueckner's (2009) discussion parallels Verhoef’s (2010) analysis, which finds that tradable slots may yield the first-best outcome if the congestion externality is relatively important and the market power distortion relatively unimportant. Conversely, Basso and Zhang (2010) show that the equivalence is lost under perfect information if the airport profits matter (either because it is privately owned or because it has to self-finance). If airport profits matter marginally, slot auctions will outperform pricing to achieve a higher objective function

218

Type

Trading

Trading

Trading

Trading

Trading

Trading

Trading

Trading

Trading

Leasing

Leasing Long-term

Leasing

Auction Attempt blocked

Auction GovernmentControlled

Year

2007

2008

2008

2014

2015

2015

2016

2017

2020

1997

2015

2020

2008

2011

Ronald Reagan Washington National Airport (USA), New York LaGuardia (USA)

New York John F. Kennedy (USA) New York LaGuardia (USA) New York Newark Liberty (USA)

Amsterdam Schiphol (EU) London Heathrow (UK)

New York John F. Kennedy (USA) New York Newark Liberty (USA)

London Heathrow (UK)

London Heathrow (UK)

London Heathrow (UK)

London Heathrow (UK)

London Heathrow (UK)

London Heathrow (UK)

London Heathrow (UK)

New York LaGuardia (USA)

London Heathrow (UK)

London Heathrow (UK)

Airport

Air France\KLM

Delta (EWR), United Airlines (JFK)

United Airlines

Air New Zealand

SAS Airlines

Air France/KLM

Air France\KLM

SAS Airlines

Cyprus Airways

GB Airways

GB Airways

Seller/Leaser/Divested

JetBlue for DCA and US Airways LGA, WestJet for LGA Delta Airlines

Delta Airlines

Delta (JFK), United Airlines (EWR)

Lufthansa

Not disclosed

American

Oman Air

Delta Airlines

American Airlines

American Airlines

Southwest Airlines

Continental Airlines

Continental Airlines

Buyer/User

Table 11.2  The use of secondary markets since their first use

$40MM, 16 slots at DCA $49.6MM, 32 slots at LGA

$14MM, 14 slots (EWR) $14MM, 30 slots (JFK)

$27MM, 1 pair

$75MM, 2 slots

$75MM, 1 pair

$276MM, 6 slots

$60MM, 1 slot

$32MM, 2 slots

$7.5MM, 14 slots

$52.3MM each, 4 pairs

$209MM, 1 pair

Price, Number of slots

219

Frankfurt hub (EU) Auction One-off approach, Munich hub (EU) Controlled by Orly Airport (EU) government Milano Linate Airport (EU) Roma Fiumicino hub (EU)

2020

Source:  own elaboration.

2021

2021

Divest Anti-trust

2016

New York John F. Kennedy (USA) Mexico City (MEX)

Shanghai Pudong International Airport (China)

Auction Announced

2016

Guangzhou Baiyun International Airport (China)

Auction Executed

2015

Vueling

Air China, China Eastern, China Southern, Hainan Airlines

ITA Airways, as acquiring Alitalia

Air France\KLM

Lufthansa

Delta Airlines Aeroméxico

15% of Alitalia slots at Linate airport, 57% at Roma Fiumicino hub

18 slots per day

24 slots per day

550 MM RMB, 9 slots

220  Handbook on transport pricing and financing

value. Pricing may be strongly preferred over slot auctions, especially when airlines are asymmetric. Another stream of literature presents variants of Weitzman’s (1974) analysis and compares the performance of quantity and price mechanisms considering more complex information structures. Czerny (2008, 2010) shows that the social benefits of slots relative to congestion pricing depend on the functional forms of passenger demand and congestion cost when passenger demands are uncertain. Notably, Czerny (2010) shows analytically that airport networks (as opposed to single airports) increase the relative welfare benefits of quantity-based mechanisms. Aravena et al. (2019) compare tolls with slots under asymmetric information. In this setting, the equivalence between the two is lost: direct allocation is always ex-post inefficient, and, in some cases, tolls and subsequent quantity delegation is a better alternative social welfare wise. Auctions may be superior or inferior to tolls.12 The empirical evidence about the impacts of slot auctions and trade is mixed. In the UK, trading has had varied effects. It has helped a dominant carrier (British Airways) increase the share of slots at Heathrow airport and other strong carriers, such as Virgin Atlantic, emerge. The UK case study also suggests that the slot trades among partner carriers contributed to slightly increased competition measured in terms of the number of competitors per route. In contrast, the slot trades between rival carriers decreased the number of competitors at the route level (Fukui, 2014). The results suggest that carriers seem to have used the slots obtained from their rivals to strengthen their dominance on their existing routes. Mixed results have also been observed for the secondary trading of slots in the United States. Evidence shows that, while slot-holding carriers did sell their slots to other airlines, including new entrant carriers and rival carriers, the number of slots sold to them was too limited to allow these carriers to increase their presence at airports (Fukui, 2010, 2012). Focusing on the impact of slot restrictions as a determinant of scarcity rent (i.e., the markup on fares resulting from airport capacity shortages), slot restrictions have limited the effect of competition on airfare (Fukui, 2019). The scarcity rent is not transferred to passengers: the average fare at Newark, for instance, decreased by about 2.5–2.6% after the slot removal, and the key driver lowering airfares was carriers other than the dominant carrier (i.e., United Airlines).

11.4 AIRPORT PRICING CONSIDERING ENVIRONMENTAL COSTS In addition to congestion, charges at airports could be set to account for environmental damages (e.g., emissions and noise). Excess airplane idling, i.e., residual daily taxi time due to delays, has been proven to significantly impact residents’ health, primarily driven by increased levels of carbon monoxide (CO) exposure. For instance, Schlenker and Walker (2016) use the variation in daily airport congestion to estimate the population dose-response of health outcomes to daily CO exposure, examining hospitalization rates for asthma, respiratory, and heart-related emergency room admissions. They find a one standard deviation increase in daily pollution levels leads to an additional $540 thousand in hospitalization costs for respiratory and heart-related admissions for the 6 million individuals living within 10 km of the airports in California. The circling in the air waiting for landing also burns extra fuel increasing greenhouse gas (GHG) emissions. Furthermore, the possibility of being held up induces airlines to carry extra fuel in their aircraft, increasing the aircraft’s weight and, consequently, fuel consumption and GHG emissions (see The Economist, 2006).

The economics of airport pricing  221

Despite the dramatic gains in aircraft “quietness” over the jet age, evidence shows growing trends in various airport noise limits, such as operational curfews, noise quotas, and noise surcharges, in the United States. Such measures are even more widespread in Europe, as Girvin (2009) discussed. Indeed, aircraft noise is one, if not the most detrimental environmental effect of aviation, being associated with severe health effects, such as sleep disorders, cardiovascular disorders, and even low birth weight babies (Argys et al., 2020). When charges at airports are set to account for environmental damages, in addition to the airport’s marginal cost and the marginal cost of congestion, each aircraft would have to pay the marginal environmental damage it produces. This should consider that the extra delay a new flight imposes on existing flights increases all flights’ average environmental cost (Carlsson, 2003). The optimal charge would depend on the aircraft types and times of the day. In practice, optimal environmental taxation and regulation are complicated by several factors, such as network effects (Carlsson, 2002; Girvin, 2010; Nero and Black, 1998; O’Kelly, 2012). Higher flight frequency benefits passengers by reducing schedule delays (allowing departures at more convenient times) but generating greater environmental damage (Brueckner and Girvin, 2008). Along with the companion paper by Girvin (2010), Brueckner and Girvin (2008) explore the impact of airport noise regulation on airline service quality and airfares. They also characterize the socially optimal stringency of noise limits, taking both noise damage and the various costs borne by airlines and their passengers into account. They show how noise regulation harms airline passengers by raising fares and potentially reducing service quality. Brueckner and Zhang (2010) explore the effect of airline emissions charges on airfares, airline service quality, aircraft design features, and network structure by analyzing the impact of an increase in the effective fuel price, and the path by which emissions charges will alter airline choices.13 The results show that emission charges will raise fares, reduce flight frequency, increase load factors, and raise aircraft fuel efficiency, not affecting aircraft size.

11.5 CONCLUDING REMARKS We have reviewed the airport pricing literature of the last 10–15 years, emphasizing optimal congestion pricing. We revised the studies that focus on the dynamics of airport runway congestion and those that study second-best pricing problems. We have also provided an overview of the pricing of private airports, airport competition and capacity investments, regulation and the role of non-aeronautical operations. We have discussed quantity-based approaches and their equivalence with congestion pricing and the pricing of externalities such as noise and emissions. We conclude by providing what we believe are the most relevant avenues for future research. First, the discussion on quantity-based regulation and its comparison with pricing abstracts away from the consideration that assigning slots via market mechanisms locally (i.e., at any given airport) cannot fully internalize interdependencies among slots at different airports (for each flight, an airline needs a feasible combination of slots at the origin and destination airports). Czerny and Lang (2019) and Lang and Czerny (forthcoming) analyze the subject and study the conditions under which quantities determined by a local authority can reach the first-best quantity, yet more research is needed. Moreover, the vast majority of the articles about airport pricing and regulation assume perfect information, with the few exceptions of Czerny (2010), Aravena et al. (2019), and de Palma and Lindsey (2020). The lack of studies with private information is most likely due to the added complexity given by the within

222  Handbook on transport pricing and financing

and between-markets negative externality. Vickrey–Clarke–Grooves mechanisms seem to be a promising approach for solving the congestion problem when regulators are not perfectly informed. Second, in most dynamic congestion models, congestion tolls must vary continuously over time to achieve the optimum. In practice, however, only tolls that take different values over discrete time intervals (but constant within each interval) have been implemented. Given the prevalence of these so-called step-tolling schemes, they have received surprisingly little attention in the airport pricing literature.14 Third, the problem of per-passenger-based versus per-flight-based airport charges is somehow under-investigated. Czerny et al. (2017) focus on the issue while extending the monopoly case in Czerny and Zhang (2015) to the oligopoly case. Further topics include discussing passenger types and price discrimination, including transfer passengers, and introducing the investigation in a hub–spoke network. An environment designed to study second-best pricing with endogenous load factors would also be a worthwhile avenue for future research. Studies only recently started to analyze more realistic and complex models with endogenous load factors (e.g., Czerny et al., 2016). On the modeling technique side, we believe structural and empirically tractable models that integrate airport decisions, the vertical relationship between airports and airlines, the competition between airlines, and congestion externalities should be at the center of future research. Some of the components we deem fundamental for future research on airport pricing and regulation include discrete choice models in the spirit of Berry et al. (1995), structural models of airport congestion such as Daniel and Harback (2009), and strategic choices of capacity and prices by airlines. To the best of our knowledge, the only study that incorporates some of these elements is Yan and Winston (2014), who study airport privatization and competition. The airport pricing and quantity regulation literature would greatly benefit from these approaches.

NOTES 1.

Although the positive effects of aviation on the economy seem intuitive, the identification of such as a causal relationship is difficult because of the strong interdependence between the provision of aviation services and regional growth (Blonigen and Cristea, 2015). Campante and YanagizawaDrott (2018) confirm a causal link between long-haul non-stop flights and economic growth, but also show how the flow of investments tends to originate largely in the wealthiest countries towards medium income ones, with low-income countries partially shut out of this process. 2. Whether congestion leads to flight delays is largely under the control of the airlines since they are free to set scheduled flight durations. In other words, the congestion-related lengthening of flight times can be built into airline schedules through a practice known as “schedule padding,” whose recent growth is documented by Forbes et al. (2019) and Brueckner et al. (2021). While airport congestion may make flights longer, this schedule adjustment prevents them from arriving late with respect to the schedule. 3. Forbes (2008) uses data from New York-La Guardia airport and finds an average price reduction per additional minute of delay of $1.42 for direct passengers; this price decrease amounts to $0.77 for connecting passengers. Britto et al. (2012), on a sample of US routes, find that a 10% decrease in delays implies a benefit of $1.50–2.50 per passenger, while the gains for airlines of reducing delays are about three times higher. Recently, Bilotkach and Pai (2020) find that one additional minute of weather delay decreases average US fares by between $4.46 and $6.55, while an extra minute of carrier delay results in a $2.70 to $5.13 price decrease. 4. Before Brueckner’s contribution, it was common to model airports as if they were selling directly a final product; Basso and Zhang (2007a) call this the “traditional approach” and review these papers.

The economics of airport pricing  223

5.

6. 7. 8. 9.

10.

11.

12.

13.

14.

This approach is different to the current common modelling setting of a vertical airport-airlinepassengers structure. Basso and Zhang (2008) show analytically that the traditional approach is valid only if air carriers have no market power and have constant marginal operational costs. The essence of the bottleneck model is that users trade off the costs of queuing delay at the bottleneck with the costs of schedule delay (i.e., arriving/leaving earlier or later than desired). Basso (2008), Brueckner (2004) and Flores Fillol (2010) do not consider a dynamic model of congestion but they include schedule delay costs as the expected gap between passengers' actual and desired departure time, which decreases with frequency. The model assumes α-β-γ preferences. Lindsey et al. (2019) extend the result to multiple firms and shows that with sufficient heterogeneity PSNE can be restored. Bel and Fageda (2010), using data for 100 large airports in Europe, provide evidence that unregulated private airports charge higher prices than public or regulated airports. When the capacity and price decisions are made simultaneously, the duopolists provide the same service quality as the monopolist. Lin (2020) analyzes uniform and discriminatory charges for local-welfare-maximizing international airports that benefit from extracting rents from foreign airlines. Unlike the papers we discuss above, they study capacity investments and show that each airport tends to overinvest under locally optimal pricing rules. On the other hand, a second strand of the literature assumes that consumers may make decisions about buying flights and non-aviation services simultaneously rather than independently. In this case, passengers increase as the price for side services decreases. A reduction in the price for side services can be considered as an increase in airport “quality,” which enhances consumer surplus and thus stimulates travel demand (Czerny, 2006; Flores-Fillol et al., 2018). In this setting, the provision of non-aeronautical services can increase or reduce the private aeronautical charge relative to a situation where side services not exist. Recently, Gomez and Tirole (2018) study the imposition a minimum price on for ancillary services in order to curb the incentive to increase prices for core services. Under the single-till ROR, airport charges are set for cost recovery plus a return on the invested capital. Under the dual-till ROR, the return applies only to aeronautical operations. As of price cap regulation, under a single-till airport revenues from all services should cover the airport’s total costs. Under dual-till price cap, the airport charge is set such that only the aeronautical services’ revenues cover its costs, while keeping non-aeronautical services unregulated. Basso et al., (2017) examine the optimal design of monopoly regulation in the presence of asymmetric information Under decreasing marginal cost and with the possibility of offering menus, a regulator would significantly improve welfare regulating through quantity. As long as the regulator can enforce the actual sale of the product, which is the case, for example, when airport slots are assigned, quantity seems an attractive regulatory instrument. The Emission Trading System (ETS) generates a permit price that becomes part of an airline’s cost structure. With the carrier’s required outlay on emissions permits varying in step with its total fuel consumption, the permit price is effectively added to the price of fuel, even though most of the permits will be freely distributed. Thus, the planned trading system can be viewed as equivalent to a carbon-tax scheme applied to aviation, which would explicitly raise the price of fuel. As a result, regardless of whether policy interventions to limit aviation emissions follow the EU’s cap-andtrade approach or rely on taxation, they can all be depicted as policies that raise the fuel price paid by airlines (Brueckner and Zhang, 2010). Daniel (2009) develops a general step-tolling model for which the no-toll and continuously-varying toll equilibria are the limiting cases as the number of tolling steps goes from zero to infinity. This unified model of tolling demonstrates how the dominant-fringe equilibria interact with the tolling structures.

REFERENCES ACI (2018). The ACI World Airport Traffic Forecasts 2018-2040. Aravena, O., Basso, L. J., & Figueroa, N. (2019). Effects of asymmetric information on airport congestion management mechanisms. International Journal of Industrial Organization, 62, 4–27.

224  Handbook on transport pricing and financing

Argys, L. M., Averett, S. L., & Yang, M. (2020). Residential noise exposure and health: Evidence from aviation noise and birth outcomes. Journal of Environmental Economics and Management, 103, 102343. Ater, I. (2012). Internalization of congestion at US hub airports. Journal of Urban Economics, 72(2–3), 196–209. Avenali, A., D’Alfonso, T., Leporelli, C., Matteucci, G., Nastasi, A., & Reverberi, P. (2015). An incentive pricing mechanism for efficient airport slot allocation in Europe. Journal of Air Transport Management, 42, 27–36. Ball, M., Barnhart, C., Dresner, M., Hansen, M., Neels, K., Odoni, A. R.,... & Zou, B. (2010). Total delay impact study: A comprehensive assessment of the costs and impacts of flight delay in the United States. Ball, M. O., Berardino, F., & Hansen, M. (2018). The use of auctions for allocating airport access rights. Transportation Research Part A: Policy and Practice, 114, 186–202. Basso, L. J. (2008). Airport deregulation: Effects on pricing and capacity. International Journal of Industrial Organization, 26(4), 1015–1031. Basso, L. J., Figueroa, N., & Vásquez, J. (2017). Monopoly regulation under asymmetric information: Prices versus quantities. The RAND Journal of Economics, 48(3), 557–578. Basso, L. J., & Zhang, A. (2007a). An interpretative survey of analytical models of airport pricing. Advances in Airline Economics, 2, 89–124. Basso, L. J., & Zhang, A. (2007b). Congestible facility rivalry in vertical structures. Journal of Urban Economics, 61(2), 218–237. Basso, L. J., & Zhang, A. (2008). On the relationship between airport pricing models. Transportation Research Part B: Methodological, 42(9), 725–735. Basso, L. J., & Zhang, A. (2010). Pricing vs. slot policies when airport profits matter. Transportation Research Part B: Methodological, 44(3), 381–391. Bel, G., & Fageda, X. (2010). Privatization, regulation and airport pricing: An empirical analysis for Europe. Journal of Regulatory Economics, 37(2), 142–161. Berry, S., Levinsohn, J., & Pakes, A. (1995). Automobile prices in market equilibrium. Econometrica: Journal of the Econometric Society, 841–890. Bilotkach, V., & Pai, V. (2020). A Price for delays: Price-quality competition in the US airline industry. Journal of Transport Economics and Policy, 54(2), 151–176. Blonigen, B. A., & Cristea, A. D. (2015). Air service and urban growth: Evidence from a quasi-natural policy experiment. Journal of Urban Economics, 86, 128–146. Bracaglia, V., D’Alfonso, T., & Nastasi, A. (2014). Competition between multiproduct airports. Economics of Transportation, 3(4), 270–281. Britto, R., Dresner, M., & Voltes, A. (2012). The impact of flight delays on passenger demand and societal welfare. Transportation Research Part E: Logistics and Transportation Review, 48(2), 460–469. Brueckner, J. K. (2002). Airport congestion when carriers have market power. American Economic Review, 92(5), 1357–1375. Brueckner, J. K. (2003). Airline traffic and urban economic development. Urban Studies, 40(8), 1455–1469. Brueckner, J. K. (2004). Network structure and airline scheduling. The Journal of Industrial Economics, 52(2), 291–312. Brueckner, J. K. (2005). Internalization of airport congestion: A network analysis. International Journal of Industrial Organization, 23(7–8), 599–614. Brueckner, J. K. (2009). Price vs. quantity-based approaches to airport congestion management. Journal of Public Economics, 93, 681–690. Brueckner, J. K., Czerny, A. I., & Gaggero, A. A. (2021). Airline mitigation of propagated delays via schedule buffers: Theory and empirics. Transportation Research Part E: Logistics and Transportation Review, 150, 102333. Brueckner, J. K., & Girvin, R. (2008). Airport noise regulation, airline service quality, and social welfare. Transportation Research Part B: Methodological, 42(1), 19–37. Brueckner, J. K., & Van Dender, K. (2008). Atomistic congestion tolls at concentrated airports? Seeking a unified view in the internalization debate. Journal of Urban Economics, 64, 288–295.

The economics of airport pricing  225

Brueckner, J. K., & Zhang, A. (2010). Airline emission charges: Effects on airfares, service quality, and aircraft design. Transportation Research Part B: Methodological, 44(8–9), 960–971. Button, K. J. (2008). Auctions - what can we learn from auction theory for slot allocation? In: Czerny, P., Forsyth, D. G., & Niemeier, H.-M. (Eds.), Airport Slots - International Experiences and Options for Reform. Ashgate, pp. 291–309. Campante, F., & Yanagizawa-Drott, D. (2018). Long-range growth: Economic development in the global network of air links. The Quarterly Journal of Economics, 133(3), 1395–1458. Carbo, J. M., & Graham, D. J. (2020). Quantifying the impacts of air transportation on economic productivity: A quasi-experimental causal analysis. Economics of Transportation, 24, 100195. Carlin, A., & Park, R. E. (1970). Marginal cost pricing of airport runway capacity. The American Economic Review, 60(3), 310–319. Carlsson, F. (2002). Environmental charges in airline markets. Transportation Research Part D: Transport and Environment, 7(2), 137–153. Carlsson, F. (2003). Airport marginal cost pricing: Discussion and an application to Swedish Airports. International Journal of Transport Economics/Rivista internazionale di economia dei trasporti, 283–303. Chen, M. X., & Lin, C. (2020). Geographic connectivity and cross-border investment: The belts, roads and skies. Journal of Development Economics, 146, 102469. Czerny, A. I. (2006). Price-cap regulation of airports: Single-till versus dual-till. Journal of Regulatory Economics, 30(1), 85–97. Czerny, A. I. (2008). Managing congested airports under uncertainty. In Airport Slots: International Experiences and Options for Reform. Aldershot, UK: Ashgate, 111–126. Czerny, A. I. (2010). Airport congestion management under uncertainty. Transportation Research Part B: Methodological, 44(3), 371–380. Czerny, A. I., Cowan, S., & Zhang, A. (2017). How to mix per-flight and per-passenger based airport charges: The oligopoly case. Transportation Research Part B: Methodological, 104, 483–500. Czerny, A. I., & Lang, H. (2019). A pricing versus slots game in airport networks. Transportation Research Part B: Methodological, 125, 151–174. Czerny, A. I., Van Den Berg, V. A., & Verhoef, E. T. (2016). Carrier collaboration with endogenous fleets and load factors when networks are complementary. Transportation Research Part B: Methodological, 94, 285–297. Czerny, A. I., & Zhang, A. (2014). Airport congestion pricing when airlines price discriminate. Transportation Research Part B: Methodological, 65, 77–89. Czerny, A. I., & Zhang, A. (2015). How to mix per-flight and per-passenger based airport charges. Transportation Research Part A: Policy and Practice, 71, 77–95. D’Alfonso, T., & Bracaglia, V. (2017). Two-sidedness and welfare neutrality in airport concessions. In The Economics of Airport Operations. Emerald Publishing Limited. D’Alfonso, T., Jiang, C., & Wan, Y. (2013). Airport pricing, concession revenues and passenger types. Journal of Transport Economics and Policy, 47(1), 71–89. Daniel, I. (1995). Congestion pricing and capacity of large hub airports: A bottleneck model with stochastic queues. Econometrica, 63, 327–370. Daniel, J. I. (2009). The deterministic bottleneck model with non-atomistic traffic. Working Papers 09-08, University of Delaware, Department of Economics. Daniel, I., & Harback, K. T. (2008). (When) Do hub airlines internalize their self-imposed congestion delays? Journal of Urban Economics, 63, 583–612. Daniel, J. I., & Harback, K. T. (2009). Pricing the major US hub airports. Journal of Urban Economics, 66(1), 33–56. de Palma, A., & Lindsey, R. (2020). Tradable permit schemes for congestible facilities with uncertain supply and demand. Economics of Transportation, 21, 100149. Dempsey, P. S. (2001). Airport landing slots: Barriers to entry and impediments to competition. Air and Space Law, 26, 1, 20–48. EC-European Commission (2004). European Council Regulation No 793/2004 of April 2004 amending Council Regulation No 95/93 on common rules for the allocation of slots at Community airports. Official Journal of the European Union L 138 (2004): 50–60.

226  Handbook on transport pricing and financing

Eurocontrol (2018). European Aviation in 2040 Challenges of Growth Annex1 Flight Forecast to 2040. Available at: https://www​.eurocontrol​.int​/sites​/default ​/files​/2019​- 07​/challenges​-of​-growth​-2018​ -annex1​_0​.pdf. Last retrieved: February 23, 2021. FAA (2020). Fact Sheet – Facts about the FAA and Air Traffic Control. Available at https://www​.faa​ .gov​/news​/fact​_sheets​/news​_story​.cfm​?newsId​=23315. Last retrieved: March 16, 2021. Fageda, X., Suárez-Alemán, A., Serebrisky, T., & Fioravanti, R. (2019). Air transport connectivity of remote regions: The impacts of public policies. Regional Studies, 1161–1169. Flores-Fillol, R. (2010). Congested hubs. Transportation Research Part B: Methodological, 44(3), 358–370. Flores‐Fillol, R., Iozzi, A., & Valletti, T. (2018). Platform pricing and consumer foresight: The case of airports. Journal of Economics & Management Strategy, 27(4), 705–725. Forbes, S. J. (2008). The effect of air traffic delays on airline prices. International Journal of Industrial Organization, 26(5), 1218–1232. Forbes, S. J., Lederman, M., & Yuan, Z. (2019). Do airlines pad their schedules? Review of Industrial Organization, 54(1), 61–82. Fu, X., & Zhang, A. (2010). Effects of airport concession revenue sharing on airline competition and social welfare. Journal of Transport Economics and Policy, 44(2), 119–138. Fukui, H. (2010). An empirical analysis of slot trading in the United States. Transportation Research Part B, 44, 330–357. Fukui, H. (2012). Do carriers abuse the slot system to inhibit airport capacity usage? Evidence from the US experience. Journal of Air Transport Management, 24, 1–6. Fukui, H. (2014). Effect of slot trading on route-level competition: Evidence from experience in the UK. Transportation Research Part A: Policy and Practice, 69, 124–141. Fukui, H. (2019). How do slot restrictions affect airfares? New evidence from the US airline industry. Economics of Transportation, 17, 51–71. Gibbons, S., & Wu, W. (2020). Airports, access and local economic performance: Evidence from China. Journal of Economic Geography, 20(4), 903–937. Girvin, R. (2009). Aircraft noise-abatement and mitigation strategies. Journal of Air Transport Management, 15(1), 14–22. Girvin, R. (2010). Aircraft noise regulation, airline service quality, and social welfare: The monopoly case. Journal of Transport Economics and Policy, 44(1), 17–35. Gomes, R., & Tirole, J. (2018). Missed sales and the pricing of ancillary goods. The Quarterly Journal of Economics, 133(4), 2097–2169. Hoppe, H., Jehiel, P., & Moldovanu, B. (2006). Licence auctions and market structure. Journal of Economics & Management Strategy, 15(2), 371–396. Hovhannisyan, N., & Keller, W. (2015). International business travel: An engine of innovation? Journal of Economic Growth, 20(1), 75–104. IATA (2010). Airport Slots - The Building Blocks of Air Travel. Available at: https://airlines​.iata​.org​/ analysis​/airport​-slots​-the​-building​-blocks​-of​-air​-travel. Last retrieved: March 15, 2021. IATA (2020). COVID-19 Outlook for Air Travel in the Next 5 Years. Available at: https://www​.iata​. org​/en ​/iata​-repository​/publications​/economic​-reports​/covid​-19​-outlook​-for​-air​-travel​-in​-the​-next​-5​years/. Last retrieved: March 15, 2021. Kidokoro, Y., Lin, M. H., & Zhang, A. (2016). A general-equilibrium analysis of airport pricing, capacity, and regulation. Journal of Urban Economics, 96, 142–155. Klemperer, P. (2002). What really matters in auction design. The Journal of Economic Perspectives, 16(1), 169–189. Lang, H., & Czerny, A. I. (forthcoming). How (grandfathered) slots can be first-best and prices not. Journal of Transport Economics and Policy. Levine, M. E. (1969). Landing fees and the airport congestion problem. The Journal of Law and Economics, 12(1), 79–108. Lin, M. H. (2019). Airport congestion and capacity when carriers are asymmetric. International Journal of Industrial Organization, 62, 273–290. Lin, M. H. (2020). Congestion pricing and capacity for internationally interlinked airports. Transportation Research Part B: Methodological, 142, 126–142. Lin, M. H., & Mantin, B. (2015). Airport privatization in international inter-hub and spoke networks. Economics of Transportation, 4(4), 189–199.

The economics of airport pricing  227

Lin, M. H., & Zhang, Y. (2017). Hub-airport congestion pricing and capacity investment. Transportation Research Part B: Methodological, 101, 89–106. Lindsey, R., De Palma, A., & Silva, H. E. (2019). Equilibrium in a dynamic model of congestion with large and small users. Transportation Research Part B: Methodological, 124, 82–107. Lu, C. C., & Pagliari, R. I. (2004). Evaluating the potential impact of alternative airport pricing approaches on social welfare. Transportation Research Part E: Logistics and Transportation Review, 40(1), 1–17. Madas, M., Zografos, K. (2006). Airport slot allocation: From instruments to strategies. Journal of Air Transport Management, 12(2), 53–62. Madas, M. A., & Zografos, K. G. (2008). Airport capacity vs. demand: Mismatch or mismanagement? Transportation Research Part A: Policy and Practice, 48(1), 203–226. Mantin, B. (2012). Airport complementarity: Private vs. government ownership and welfare gravitation. Transportation Research Part B: Methodological, 46(3), 381–388. Matsumura, T., & Matsushima, N. (2012). Airport privatization and international competition. The Japanese Economic Review, 63(4), 431–450. Mayer, C., & Sinai, T. (2003). Network effects, congestion externalities, and air traffic delays: Or why all delays are not evil. American Economic Review, 93(2003), 1194–1215. McGraw, M. J. (2020). The role of airports in city employment growth, 1950–2010. Journal of Urban Economics, 116, 103240. McMillen, D. P. (2004). Airport expansions and property values: The case of Chicago O’Hare Airport. Journal of Urban Economics, 55(3), 627–640. Mense, A., & Kholodilin, K. A. (2014). Noise expectations and house prices: The reaction of property prices to an airport expansion. The Annals of Regional Science, 52(3), 763–797. Mohring, H., & Harwitz, M. (1962). Highway Benefits: An Analytical Framework. Morrison, S. A., & Winston, C. (2007). Another look at airport congestion pricing. American Economic Review, 97(5), 1970–1977. Nero, G., & Black, J. A. (1998). Hub-and-spoke networks and the inclusion of environmental costs on airport pricing. Transportation Research Part D: Transport and Environment, 3(5), 275–296. O’Kelly, M. E. (2012). Fuel burn and environmental implications of airline hub networks. Transportation Research Part D: Transport and Environment, 17(7), 555–567. Oum, T. H., Zhang, A., & Zhang, Y. (2004). Alternative forms of economic regulation and their efficiency implications for airports. Journal of Transport Economics and Policy, 38(2), 217–246. Pels, E., & Verhoef, E. T. (2004). The economics of airport congestion pricing. Journal of Urban Economics, 55(2), 257–277. Rupp, N. G. (2009). Determinants of fares and operating revenues at US airports. Journal of Urban Economics, 65, 24–37. Schlenker, W., & Walker, W. R. (2016). Airports, air pollution, and contemporaneous health. The Review of Economic Studies, 83(2), 768–809. Silva, H. E., Lindsey, R., De Palma, A., & Van den Berg, V. A. (2017). On the existence and uniqueness of equilibrium in the bottleneck model with atomic users. Transportation Science, 51(3), 863–881. Silva, H. E., & Verhoef, E. T. (2013). Optimal pricing of flights and passengers at congested airports and the efficiency of atomistic charges. Journal of Public Economics, 106, 1–13. Silva, H. E., Verhoef, E. T., & Van Den Berg, V. A. (2014a). Airlines’ strategic interactions and airport pricing in a dynamic bottleneck model of congestion. Journal of Urban Economics, 80, 13–27. Silva, H. E., Verhoef, E. T., & Van Den Berg, V. A. (2014b). Airline route structure competition and network policy. Transportation Research Part B: Methodological, 67, 320–343. Starkie, D. (2021). Two-sided airport markets reprised. Journal of Transport Economics and Policy, 55(1), 1–15. The Economist (2006). Global warming, The Economist, 8 June 2006, Economist Intelligence Unit. Van Dender, K. (2007). Determinants of fares and operating revenues at US airports. Journal of Urban Economics, 62, 317–336. Verhoef, E. T. (2010). Congestion pricing, slot sales and slot trading in aviation. Transportation Research Part B: Methodological, 44(3), 320–329. Verhoef, E. T. (2017). Cost recovery of congested infrastructure under market power. Journal of Urban Economics, 101, 45–56.

228  Handbook on transport pricing and financing

Wan, Y., Jiang, C., & Zhang, A. (2015). Airport congestion pricing and terminal investment: Effects of terminal congestion, passenger types, and concessions. Transportation Research Part B: Methodological, 82, 91–113. Weitzman, M. L. (1974). Prices vs. quantities. The Review of Economic Studies, 41(4), 477–491. Yan, J., & Winston, C. (2014). Can private airport competition improve runway pricing? The case of San Francisco Bay area airports. Journal of Public Economics, 115, 146–157. Yang, H., & Zhang, A. (2011). Price-cap regulation of congested airports. Journal of Regulatory Economics, 39(3), 293–312. Zhang, A., & Czerny, A. I. (2012). Airports and airlines economics and policy: An interpretive review of recent research. Economics of Transportation, 1(1–2), 15–34. Zhang, A., Fu, X., & Yang, H. G. (2010). Revenue sharing with multiple airlines and airports. Transportation Research Part B: Methodological, 44(8–9), 944–959. Zhang, A., & Zhang, Y. (1997). Concession revenue and optimal airport pricing. Transportation Research Part E: Logistics and Transportation Review, 33(4), 287–296. Zhang, A., & Zhang, Y. (2003). Airport charges and capacity expansion: Effects of concessions and privatization. Journal of Urban Economics, 53(1), 54–75. Zhang, A., & Zhang, Y. (2006). Airport capacity and congestion when carriers have market power. Journal of Urban Economics, 60(2), 229–247. Zhang, A., & Zhang, Y. (2010). Airport capacity and congestion pricing with both aeronautical and commercial operations. Transportation Research Part B: Methodological, 44(3), 404–413.

12. Pricing in freight transport Edoardo Marcucci, Valerio Gatta, Michele Simoni and Ila Maltese

12.1 INTRODUCTION This chapter discusses the main issues pertaining to freight transport pricing by focusing on two different yet complementary concepts: (1) price setting by, primarily, the market (charges, tolls, and fares), and (2) influencing the price through additional taxes, levies, or subsidies by the government. Hence, pricing encompasses both the process of determining the price for delivering goods by the service provider and the extra price imposed by public authorities as a mobility management solution (de Palma et al., 2007a). In more detail, one might investigate how pricing strategies are defined given the public policies in place, market structure, competitive pressure, market share, and customers’ preferences for each transport mode or combination thereof. Additionally, one might look at the policies developed to generate revenues for maintaining and expanding the infrastructure (Jiang et al., 2017; Anand et al., 2021; De Borger & Proost, 2021) or, more often, to internalise the external costs of freight transport to determine the most socially efficient market shares for each transport mode. Besides market share, controlling the overall traffic volume can also represent a pricing objective to avoid overconsumption of movements that might produce negative net welfare effects. While separate, we consider it appropriate to discuss both concepts since the results will depend upon their dynamic interplay. While acknowledging the importance of determining a socially optimal pricing structure to ensure sustainability goals, one also has to consider the financial sustainability of the service. Freight transport is essentially a form of derived demand (Wang et al., 2021), and it depends on other economic activities. It is just one component of any given supply chain, and the decisions stakeholders take should always be studied taking into due consideration the broader logistic picture to understand the overall rationality of the decisions taken, which otherwise might appear inconsistent. Other logistic functions (e.g. warehousing, packaging, picking, handling, etc.) need to be investigated, assuming firms maximise their overall profit function. These considerations add another layer of complexity, including the geographical scale that might play a substantial role in determining what happens in a given freight market. For example, the probability that a container, once it has reached the port of destination in a worldwide logistic supply chain, uses either road or rail, does not only depend on the relative price of each mode available to reach the destination (e.g. warehouse/retailer), but also on the consolidation strategy adopted by the third-party logistic provider that can either opt for consolidation in the departure/destination port (Lin et al., 2020). Investigating freight externalities poses peculiar challenges due to the lack of reliable data and geographical scale. Internalising external freight costs is characterised by three peculiar issues. The traditional transport pricing model, relying on marginal cost pricing, is not always used in this context. Private pricing does not typically include environmental nor 229

230  Handbook on transport pricing and financing

users’ external costs, such as, for instance, congestion (Nash et al., 2014). Transport pricing strategies usually account for both service and infrastructure costs. Road transport represents the main source of externalities in freight transport. For this reason, several countries have introduced road tolls aimed at trucks and Heavy Goods Vehicles (HGVs), such as, for instance, distance-based tolls or those relying on a fixed price scheme via a digital pass. Rail freight transport produces non-negligible impacts on the environment. A standard procedure to internalise CO2-related emissions for freight trains foresees imposing a carbon emissions tax. Promoting a shift from road to rail requires overcoming technological/organisational constraints hindering rail freight attractiveness. Research suggests that railway freight subsidies might play a substantial role. Air freight transport is particularly difficult to clearly define due to providers’ heterogeneity (line haul operators; integrated/courier/express; and niche). An increasing proportion of global greenhouse gas emissions is due to the maritime sector. Internalising the external costs of maritime freight transport is responsible for necessarily needs addressing the problem on a regional scale. Research suggests that cargo-based measures including emissions produced throughout the trip are to be preferred to other carbonpricing options. The chapter is structured as follows. Section 12.2 discusses pricing in freight transport policy. Sections 12.3 to 12.6 report, respectively, on road, rail, air, and maritime freight pricing. Section 12.7 concludes.

12.2 PRICING IN FREIGHT TRANSPORT POLICY This section discusses both the internalisation of freight transport externalities and the pricing of freight transport services, with the intent of underlining differences and interactions between them. The results one observes in the market depend on the private agents’ decisions that take place within a given set of rules determined by public agents. 12.2.1 Internalising Freight Transport Externalities As it is widely acknowledged, a government can adopt and implement different measures to correct market failures. This is especially true in cases of negative externalities (Krugman & Wells, 2021), such as environmental pollution or climate change, for which private solutions are not viable. Price-based measures (taxes and subsidies) and quantity-based measures (regulation) are the public tools for internalising external costs (Auerbach et al., 2013; Santos et al., 2010). Freight transport represents the backbone of international trade (Hesse & Rodrigue, 2004; Rodrigue et  al., 2016). It has played a major role in promoting growth on a global scale (Meersman & Van de Voorde, 2013). What is commonly referred to as globalisation would not be possible without reliable, fast, efficient, and inexpensive transportation. Freight shipping activities have greatly contributed to this end, when looking at this from an international perspective, while considerable public investments on highways, rail infrastructures, and other large infrastructural projects have supported this process with respect to an international, national, regional, and local perspective. Freight transport not only contributes significantly to the productivity of the economy, but also produces relevant costs to society. Environmental concerns are rising in current society

Pricing in freight transport  231

(the interested reader is referred to Chapter 4 in this Handbook for further discussion on transport pricing and climate change). Governments and markets now realise the importance of the side effects transportation imposes on society, which are called “externalities”. In 2016 the total external costs of (passenger and freight) transport in EU28 amounted to 6.6% of the total GDP (van Essen et al., 2019), with freight transport being responsible for 31%. Pollution, accidents, congestion, and land use account for 40%, 29%, 27%, and 4%, respectively (van Essen et al., 2019). Externalities can have different magnitudes depending on the transport mode, as it is illustrated in Figure 12.1, which is a summary of different previous evaluations, in terms of monetary values, of the externality impact (right axis) of each transport mode (left axis) provided in a study by Demir et al. (2015). While road freight transport allows for more flexible, fast, and cheaper services, it is characterised by several negative externalities. The most important ones include pollutant emissions, noise, accidents, and congestion (Janic, 2007). This is true for all externalities except for water pollution which is mainly pertinent to maritime transportation. The latter, together with rail and air transportation are not only akin to one another but they have also, on average, a lower impact on health and the environment. It is important to have an accurate measure of marginal social costs (i.e. the sum of internal and external costs) since incorrect pricing and internalisation of external costs might determine a socially sub-optimal use of alternative transport modes or even, put simply, excessive

Note:   Demir et al. (2015) have considered and summarised the results of previous evaluations, in monetary terms, of the externality impact on each transport mode. For more details on these studies see Demir et al. (2015). Source:   Authors’ elaboration on Demir et al. (2015), Table 1, page 99.

Figure 12.1  Relevance of the negative externalities by transport modes

232  Handbook on transport pricing and financing

use of freight transport services. It is important to stress that the knowledge level for external costs is much higher for land modes (rail and road) compared to maritime and aviation, where estimation procedures are not fully defined (Nilsson et al., 2018). Furthermore, calculating the external costs of freight transportation might also prove useful in deploying more efficient pricing systems, optimising research funds allocation so as to focus on the largest external costs internalisation, and promoting cost-benefit analysis so as to optimise transportation modes and infrastructural investments (Demir et al., 2015). Investigating freight externalities poses peculiar challenges to the researcher mainly due to: (i) coherence of methods used; (ii) availability of reliable data; and (iii) relevance and prevalence of geographical scale linked to the existence/non-existence of an administrative body capable of defining and supporting an internalisation strategy. It is important to consider alternative policy options capable of internalising external costs. The available options, depending on the specific context, include cap and trade or taxes (Lindsey & Santos, 2020). The first, also known as “emission trading”, is a market-based policy for controlling pollutant emissions. Among the latter, there are: (i) imposing taxes based on the weight or distance of each shipment; (ii) increasing existing fuel taxes; (iii) implementing a tax on the transport of shipping containers; (iv) increasing current truck tyres taxes. Increasing the cost of a given mode implicitly incentivises using a less environmentally impacting substitute. While this illustrates the basic microeconomic textbook paradigm, there is now consolidated evidence that this does not necessarily apply in practice due to freight transport-specific characteristics (Holguín-Veras, 2011). Nevertheless, the previous approach constitutes the main theoretical underpinning of internalisation policies revolving around the polluter pays principle. Internalising external freight costs is characterised by three peculiar issues, namely: 1) supply chain international dimension – the perimeter within which international transport decisions are taken is not always well-defined given the current absence of adequate administrative/regulatory bodies having both the capabilities and interest in promoting an effective and easily enforceable internalisation approach. Furthermore, in this respect, one should also consider, even if outside of the scope of the current chapter, the relationship between transport and geopolitics. Sometimes, transport infrastructure investment choices and price setting are defined with a geopolitical long-term perspective agenda in mind and not so much within an economic framework. As an example of this last point, one could consider and read the copious literature on the belt and road initiative China is pursuing (Lin, 2019) or the attention posed on arctic logistics (Milaković et al., 2018). While these considerations clearly illustrate the relevance of the international dimension of the supply chains, at the same time suggest that predicting its future impact on freight pricing is, if not impossible, extremely difficult; 2) logistic versus transport costs – companies adopt a system optimisation perspective when considering logistic costs (e.g. transport, handling, warehousing, picking, etc.) where transport ones are just a component. This might have strong implications when trying to define appropriate internalisation policies. If the transport cost, while relevant, is proportionally and structurally less than other cost types (e.g. land) it can only explain a marginal part of the overall rationale used when mixing the various components needed to satisfy the end-consumer preferences which de facto determine the final price of the goods thus also how much transport cost can be incorporated in it. To make things clear one can refer to what is happening in urban freight distribution, which represents a large

Pricing in freight transport  233

and growing market, especially thanks to the skyrocketing of e-commerce. In this case, large warehousing facilities guaranteeing substantial scale economies are moved further away from city boundaries while at the same time requiring higher transportation costs that are nevertheless substantially lower than the costs that logistic operators would have to incur when asking for the same amount of land closer to the city. This well-known phenomenon goes under the name of logistics sprawl (Heitz et al., 2017). Also in this case, while the characteristics and the dynamic implication of the phenomenon described are extremely relevant in determining freight transport pricing, it is also clear that its specific determination will depend on many heterogeneous and interacting geographical characteristics making predictions difficult; 3) modal versus multi-modal perspective – the full internalisation of external costs with respect to a single mode (e.g. shipping) might have modal shifts effects promoting the use of less environmentally friendly modes (e.g. road) which would then have to be dealt with. Internalising external costs with a mode-specific approach can produce counterintuitive and unintended results whenever – this is most likely the case in the great majority of international trade – the third-party logistic provider, freight forwarder, logistic integrator, etc. adopts an end-to-end cost minimisation approach whereas single and, most likely, uncoordinated transport authorities and policymakers look at the problem from a mode-specific perspective. This is, in principle, also aggravated not only by regulatory or policy-related interventions, but also by supply chain agreements and operational decisions performed by importers or wholesalers. Consider the example of an importer from Europe that must decide whether to consolidate in China products purchased in that country, thereby purchasing a freight container shipping service, or to consolidate items at a warehouse in Europe. The two conceivable solutions will likely have a significant impact on the transport mode used once the goods are imported into Europe (most likely, rail in the first case, and road in the second one). 12.2.2 Pricing Freight Transport Services Various pricing models, characterised by peculiarities, are used within the freight sector. The traditional transport pricing model, relying on marginal cost pricing,1 is not always used in this context. While distance-based pricing can be linked to marginal cost pricing and discount on volumes, which pertains to scale economies due to large units, private pricing does not typically include environmental (e.g. pollution) nor users’ external costs (e.g. congestion). Furthermore, distance-based pricing does not account for capacity utilisation nor load limits. The relevant transport market deregulation that took place in the last decades of the past century has favoured, to a certain point, the reversal of these practices which, however, have never been complete. In fact, there is still a persistent use of distance and weight-related pricing policies in the sector. Containerisation at the international scale and standardised packaging at the local level (e.g. e-commerce) have generated substantial pressure on pricing strategies linked to goods value discrimination (Pettersen Strandenes, 2013). Value-based price discrimination proves an inaccurate instrument when trying to capture a higher willingness to pay since other non-monetary characteristics might play an important role in explaining mode choice. For example, Danielis et al. (2005) show the relative importance logistics managers pose on time, reliability, and safety when buying freight transport services.

234  Handbook on transport pricing and financing

Transport pricing strategies usually account for both service and infrastructure costs since they are influenced by investment and capacity at the same time. Both are often characterised by economies of scale and public sector regulation (e.g. speed limits, maximum weight, access restrictions, etc.). First-best marginal cost pricing conditions are not usually fulfilled by private freight suppliers. In fact, social surplus-oriented objectives lead to a pricing rule under which marginal social cost equals the marginal user’s benefit, while profit maximisation equates marginal operating cost with marginal revenue (see Chapter 3 in this Handbook for a deeper theoretical discussion on the pricing implications of these objectives). Indeed, competition between private firms might reduce the price close to marginal operating cost, but due to externalities and imperfect competition, this is still far from marginal social cost pricing. This paves the way to second-best Ramsey pricing which implies the adoption of price markup by those firms that supply transport services on the market. Non-linear pricing models, not assuming proportionality to the amount of the items bought, can also be applied to gain extra revenue by exploiting customers’ heterogeneity that explicitly represents a price discrimination strategy (Varian, 1989; Wilson, 1993). Trying to extrapolate transport cost pricing from the overall logistics strategy a company is pursuing is not only difficult in practice but could also prove counterproductive. The researcher needs to develop a clear understanding of the organisation’s overall strategy adopted whenever taking decisions from a logistic standpoint. Should one eschew adopting such a holistic perspective, the transport mode-related decisions operators take would be difficult to understand or seem irrational. Thus, making sense of pricing strategies in the freight transport market is not an easy task and this should be kept in mind when defining and deploying any internalisation policy. In general, the factors influencing transportation cost and pricing from a logistic perspective can be grouped into two major categories: product-related and market-related factors. The former ones are density, storability, handling ease/difficulty, and liability, while the latter are intramodal/intermodal competition, market location, nature/extent of transport carrier regulation, freight traffic into/out of market balance/imbalance, movements seasonality, domestic/ international transportation (Lambert et  al., 1998). Furthermore, there are some important service factors one should also consider since customer service is a logistic management pillar. Transport represents a fundamental building block to provide the desired customer service levels. Among these, the most customer service impacting ones are service reliability, transit time duration, door-to-door service capabilities, handling flexibility, loss and damages, and provision of ancillary transportation services. One should also recall that on top of the four basic transport modes (road, rail, air, water), shippers might also use intermodal combinations or multi-modal ones. The most popular of the latter is a trailer on a flat car. Inter-modality mingles the cost and/or service advantages of at least two different modes in a single product shipment. Different pricing strategies apply to freight transportation for determining how rates are defined, in general, and applied by a carrier when transporting a given item from origin A to destination B. There are two main methods used to determine a transport price depending either on the cost of producing the service or on its value (Lambert et al., 1998). Costof-service pricing typically sets transportation rates based on carrier costs, both fixed and variable (Dolinayová et  al., 2015), plus a markup. Distance and volume are the most powerful explanatory variables determining transportation costs from a cost-of-service pricing

Pricing in freight transport  235

perspective. This approach is useful in establishing lower rate limits yet intrinsically complex. It rests upon a clear identification of fixed and variable costs which is not always easy due to the daunting difficulties one might encounter when trying to detect relevant cost components and measure them. Furthermore, allocating fixed costs to a specific shipment is difficult given that the fixed cost incidence per unit decreases with a rising number of shipments, and vice versa. This is particularly challenging, especially in those transport segments where demand is highly unpredictable (e.g. freight charter). Fixed cost allocation also depends on shipment volumes. The value-of-service pricing strategy, instead, simply sets the price at the highest level the market will bear and depends on the specific characteristics of demand and competition in each market segment.2 This determines upper rate limits and aims at maximising the difference between revenues and variable costs. Competition levels have a strong impact on determining prices charged in the market (Lambert et al., 1998). These considerations clarify the complex scenario within which freight pricing rates are determined and illustrate possible alternative strategies. An additional layer of complexity derives from the INCOTERMS (International Commercial Terms)3 specified in the transport service contract (e.g. cost insurance freight – CIF, free on-board – FOB). The degree of regulation within the transport sector has widely varied over the years due to economic, environmental, and safety issues, which are the most important areas influenced by public policymaking. Therefore, pricing strategies from a private perspective are greatly impacted by these exogenous regulatory elements.

12.3 ROAD FREIGHT PRICING Worldwide, in 2018 freight transport over land was dominated by road transport, 75% in Europe (Eurostat, 2020) and 73% in the United States (US Bureau of Transportation & Statistics, 2020b). When considering light-duty vehicles (LDVs) and HGVs together, road transport represents the main source of externalities in freight transport. Interestingly, while LDVs and HGVs represent only a smaller fraction of vehicles on road transportation networks, the marginal external costs they produce are significant. For example, trucks accounting for 8% of urban traffic in US cities, provoke delays responsible for about 18% of total congestion costs (Schrank et al., 2015). According to Austin (2015), in the United States the external road freight costs represent approximately 37% of the current price paid for truck transport (5.86 cents per tonne-mile versus 15.6 cents per tonne-mile average price). In 2016, in the European Union alone, the total external costs of freight were 1.3% of GDP, while passenger-generated ones were 4.2% (van Essen et al., 2019). One can promote road freight transport efficiency by adopting different solutions so as to internalise its marginal social costs. The typical instruments deployed include CO2-related taxes/access restrictions, tolls, and fees. While road transport external costs internalisation via taxation is a well-established tool, there is still a research gap with respect to their practical implementation. From a research perspective, significant work has been produced in the past 50 years although mainly focused on passenger transport. Given the complicated interactions among stakeholders (shippers, carriers, logistic service providers, etc.) and sometimes restricted transport options available, one needs to implement ad-hoc policy solutions. For example, tolls can sometimes be transferred to customers without generating changes in the transport behaviour of carriers who often operate in relatively constrained settings (short lead

236  Handbook on transport pricing and financing

times, given fleet). In practice, freight transport internalisation tools are only partially capable of producing the desired results. For example, the trucking industry’s tax contribution in the United States accounts for approximately 10–15% of federal highway network expenditures (US Bureau of Transportation Statistics, 2020a). However, its burden in terms of highway maintenance and repair costs would require substantially higher taxes (FHWA, 2000). In what follows, we provide an overview of the main real-world applications and academic studies of road pricing in Europe and North America from the past 20 years. This focus is adopted due to the more significant information on data and policy applications in these geographical areas. The most popular pricing approach in Europe consists of distance-based tolls for HGVs travelling on the highway and major road networks. Many European countries, such as Germany, Austria, Switzerland, Portugal, Belgium, and the Czech and Slovak Republics, have established distance-based tolls for HGVs based on different technologies (e.g. on-board GPS units, dedicated short-range communications) (Toll EU, 2020). The fares are based on EU emission rules. Other countries like Denmark, Luxembourg, the Netherlands, and Sweden rely on a fixed price scheme based on a digital pass called “Eurovignette”. The vignette applies to HGVs above 12 tonnes on highways and specified routes (mainly highways). The key pricing differences are the validity duration, axle count, and emission requirements. Charges for highways are often collected in the same manner as for passenger vehicles, at toll gates using standard payment methods. The interested reader is referred to Chapter 21 in this Handbook for a more comprehensive discussion of road pricing in Europe. Another interesting solution for tackling road freight externalities consists of congestion charges introduced by cities to reduce the overall congestion problem in their centres. The main concept behind this is to make users pay for the delay costs generated by their trips, particularly during the most congested hours of the day. To date, this travel demand management measure has been successfully implemented in several cities including Singapore, London, Stockholm, Gothenburg, and Milan (Lehe, 2019). Most of these schemes do not seem to discriminate between passenger and freight vehicles except for Singapore where HGVs are subject to different tariffs (Land Transport Authority, 2020). Conceptually, price differentiation for freight vehicles is a more accurate approach since they affect traffic differently from cars due to lower speeds and higher road occupancies (three or more passenger car equivalents). On the other hand, road freight transport might be rather inelastic in comparison to passenger transport because of the limited transport mode alternatives (Allen et  al., 2012; Taniguchi et al., 2016). Along these lines, a growing body of academic research has highlighted the peculiarities of freight transport in terms of logistic operators’ travel choices and constraints. For example, Holguín-Veras et al. (2006) show how carrier behaviour is often tied to customer demand and how this prevents them from fully changing their distribution patterns. For this reason, it is important to better understand the effects of congestion pricing strategies to identify efficient solutions for mixed traffic, similar to what was proposed by Chen et al. (2018) who focus on second-best strategies aimed at freight and passenger movements. Depending on the type of commodity and service provided, some carriers likely transfer toll increases to customers (Zhang et al., 2019). However, other types of behavioural responses involving productivity and facility usage changes can be adopted (Holguín-Veras, 2010). Interestingly, Perera et al. (2021) show that the common practice of charging high urban tolls from freight vehicles may result in their diversion from freeways (toll roads) to highways and arterial roads which negatively impacts the environment and society.

Pricing in freight transport  237

Finally, specific compensatory measures, such as the provision of parking space and parking availability, need to be developed by means of collaborative approaches between policymakers and transport providers (Marcucci et al., 2015). Outside the realm of City Logistics, a few studies have investigated road charging from a freight transport perspective. Mostert et al. (2017) investigate the strategic impacts of national road charges in Belgium. The authors identify only minor impacts on the mode share resulting from an increase in the existing distance-based toll. Similar results were obtained by Gomez and Vassallo (2020) who adopted a dynamic panel data model to evaluate the effects of charges on road freight demand in Europe over a time span of 20 years. The analyses show how road charges have little influence on modal shift. These studies suggest the necessity of further research in this field to better understand the efficiency of current charges and to develop more advanced and effective pricing strategies. For example, Zhang et  al. (2019) investigate carrier ability to transfer toll increases and identify critical differences for the type of commodities transported. Teo et al. (2014) focus on how a distance-based road pricing and load factor control joint scheme can help reduce the impact of urban freight on the road network around Osaka, Japan. More recently, Perera et al. (2020) propose a link-based multi-class toll with promising reductions of total emissions (around 10%) in a mediumsize scenario (part of the Melbourne roadway network). Martin and Thoresen (2015) observe Australian freight road transport, finding out that allowing increased axle loads may turn out to reduce emissions. Finally, Cavallaro et al. (2018) explore the state-wide experiment of a road charging scheme in California aimed at reducing kilometres driven and gasoline consumed by LDVs such as vans and pickup trucks. From a private perspective, the total cost of road freight transportation can be broken down into different major components: capital, operating, and maintenance costs. Operating and maintenance costs can also be categorised as driver-based and vehicle-based costs (Williams & Murray, 2020). The first ones include wages and benefits. According to recent surveys, these costs represent the highest percentage of the cost per mile/kilometre, with values between 30% and 40% (Williams & Murray, 2020; Conrad & Joseph, 2018). The current and expected shortage of workforce in the trucking sector is likely to become the major determinant of this cost (Costello & Suarez, 2015). Vehicle-based costs include fuel, vehicle lease/payment, repair/maintenance, and insurance costs. Fuel costs represent the highest percentage of the cost per mile/kilometre, with varying prices according to the geographical area and year of operation. All costs incurred by a vehicle during its lifespan are typically measured at regular intervals. Important factors including fuel price, load levels, transported commodity, and vehicle used play a crucial role in the overall pricing of the long-haul trucking business. The road freight transportation industry is extremely fragmented, with several global and regional competitors. As a result, the freight transportation industry is extremely competitive and cost-effective. While the freight transportation industry is consolidating, the broad use of subcontracting to smaller businesses is an important element of the market, serving to lower bigger companies’ overheads, expand geographical reach, fulfil peak demand times, and improve supply chain efficiency (Schnell-Lortet & Jacob, 2019). While the road freight sector is under significant pressure from ongoing COVID-19 and supply chain disruption, it is experiencing a period of fast innovation and change as market leaders accelerate their digital transformation.

238  Handbook on transport pricing and financing

12.4 RAIL FREIGHT PRICING Rail freight transport produces non-negligible impacts on the environment. Rail freight costs related to air pollution and climate change in EU28 amount, respectively, to 0.67 (Table 16, page 57) and 0.24 (EC, 2019, Table 25, page 76) billion € per year. These total values might vary substantially depending on the diffusion of different locomotive propulsion systems (i.e. diesel or electric). In the case of electric trains, emissions are linked to the method adopted to produce the electricity used whose consumption primarily depends on train weight while also influenced by speed, aerodynamics and driving patterns (Boulter & McCrae, 2009). Additionally, rail freight transport also produces relevant noise externalities, due both to traction and rolling – 2.5 Billion € per year due to noise (EC, 2019) – with substantial differences depending on whether the noise is produced during day or night. A standard procedure to internalise CO2-related emissions for freight trains foresees imposing a carbon emissions tax. Rail access costs are levied in all EU28 countries, but they are not classified by externality type. Rail freight charges vary between 1 and 1.3 cents/tkm, whereas road transport costs 1.5 cents/tkm (Schroten et al., 2019). Since railway undertakings typically use electric-propelled locomotives, the Emissions Trading System (ETS) has only indirectly been used in this sector. This represents a major issue given the high percentage of electrified tracks (54%) in EU28. Given the overall incidence of freight railway transport, this segment has paid approximately 32 Million € in 2016 (EC, 2019). Additionally, excise duties are heterogeneously levied for diesel trains in most European countries (EC, 2020b). Different types of taxes can be imposed; however, it is important to note that, ideally, taxation should be set so as to internalise marginal social/environmental costs. While acknowledging the higher environmental efficiency characterising rail freight transportation, in terms of tonne-km hauled per unit of energy consumed, one has also to recall that rail is less flexible than road thus limiting its intrinsic attractiveness. Furthermore, rail freight transportation implies non-negligible loading/unloading costs thus making it also financially inconvenient for shipments of less than 300 km. Promoting a shift from road to rail with the intent of abating emissions necessitates not only substantial subsidies but also overcoming those technological/organisational constraints hindering rail freight attractiveness. These considerations are especially relevant for e-commerce-related shipments. Research suggests that railway freight subsidies might play a substantial role (Qu et  al., 2017). The subsidies might be awarded either to shippers or carriers with markedly different impacts. Providing subsidies to shippers reduces their generalised costs when using rail freight services thus inducing CO2 reduction thanks to the use of a less environmentally impacting mode. Whereas subsidising carriers provides them with first-mover advantages allowing for higher profits (Li & Zhang, 2020). The overall impact is higher for medium/long-distance shipments, which are particularly relevant from an international perspective. Governments can financially support operators in various forms. For example, one could favour private company investments when acquiring equipment or constructing infrastructures so as to promote rail freight transportation. This might have different implications within a rail freight service liberalisation process and encourage market competition between road and rail. The overall rail freight liberalisation process can be also favoured by the promotion of freight interoperability capable of stimulating road-rail competition.

Pricing in freight transport  239

Appropriate freight service pricing is fundamental to maximise returns, and freight rail revenue management is crucial to pursue this objective (Ljungberg, 2013; de Palma et  al., 2007b). Pricing strategies are based on two pillars: (1) match rail network capacity to demand; (2) analyse and understand customer preferences and behaviour. This process takes place in a peculiar environment where high fixed costs (e.g. yards/rail lines) coexist closely with low marginal per-load train costs (e.g. fuel and car). This gives rise to a setting where capacity is scarce while marginal costs are low. It is thus fundamental for private companies to ensure high-capacity utilisation while accounting for capacity constraints and network effects. Incremental traffic can be either very profitable (train would depart anyhow – lower average fixed costs) or very expensive (more locomotives – increase average fixed costs). Furthermore, government-controlled infrastructure operators have a significant influence on the price for the usage of tracks. While there is a public push to reduce such fees to incentivise rail over less environmentally friendly modes like road transport, there is also a need for stable revenues. Rail freight pricing is complex due to different reasons among which the most relevant are: (1) mixed traffic/merchandise types using the network; (2) seasonality; (3) patterns/volumes; (4) predetermined/fixed contract prices; (5) differentiated short-term (e.g. intermodal) versus long-term (e.g. coal) strategies/flexibility; and (6) predetermined revenue-splitting agreements due to multiple railroads use. Pricing strategies have jointly addressed only a few of the critical dimensions reported above. Research in revenue management models has developed different approaches. The most relevant ones are: (1) service-based – for instance, Kraft (2002) proposes a method for establishing aggressive yet attainable delivery appointment times for railroad shipments. The paper accounts for individual customer needs as well as predicted available train capacity. In more detail, the paper investigates a stochastic “train segment pricing” process for forecasting future demands; (2) train/block-based – Sibdari et al. (2008), for example, develop a revenue management model allowing passengers to bring their vehicles on the train. The authors propose an algorithm to calculate the optimal pricing strategy yielding maximum revenue. The paper investigates three pricing policies, namely myopic policy, static-price heuristic, and pseudo-dynamic heuristics, providing a relevant contribution to multiproduct revenue management in the Auto Train real case; 3) container-centric yield management – Gorman (2010) develops an integrated production decision support system, based on analytical tools, including forecasting, error distribution analysis, expected value-based heuristics, and optimisation to advance yield management and container allocation for an intermodal freight rail transportation company. Considering liberalisation, at least in Western developed countries, the importance of market-based rail freight pricing will, most likely, increase. Europe’s aspiration to overturn the decline of its rail freight industry has not materialised. The liberalisation process has not produced the expected results so far. A new comprehensive strategy should re-think the overall regulation framework. The current operating model, including a privatised and liberalised market, has stimulated competition on marginal cost in an industry characterised by high fixed costs. The end result being a market segmentation with two main components: highly competitive and profitable segments (e.g. long complete trains, and long distances) and segments whose profitability depends on their feeder nature (e.g. single wagon loads or short distances). The major issues to be addressed are, for example: (i) integrate rail and road infrastructure charges so that stakeholders could develop an integrated rail and road view and financing approach; (ii) dedicate tracks to freight rail only with access to daytime slots, especially close

240  Handbook on transport pricing and financing

to critical nodes where currently it is almost impossible for freight trains; and (iii) enable new technology, such as track and trace, European Rail Traffic Management System/European Train Control System, automatic coupling, and automated terminals (Chapuis et al., 2022). When analysing the last three decades of intercity freight transport in the United States, Bitzan and Keeler (2014) note that deregulation has been beneficial for the rail industry and its users, even with limited evidence related to long-term pricing trends by commodity. In particular, significant increases were observed for some “non-captive” commodities, which suggests that moving toward a more market-based pricing system can enhance railroad viability without harming those with fewer transport options. Interestingly, Merkert and Hensher (2014) investigate the transferability of the European Commission's approach toward openaccess rail organisation to the Australian context from a transaction cost perspective while also focusing on vertical freight integrated railways. It is important to note that the conceptual framework adopted to investigate the public-private rail freight service provision used so far is not necessarily adequate when strong geopolitical issues are at stake, such as in the case of the road-and-belt initiative, based on a railroad network of six major corridors (Lin, 2019). In conclusion, market-based rail freight pricing should be jointly studied along with operation planning given the implications this might have on alternative service frequencies and train route options (Li & Zhang, 2020). In fact, the role revenue management and logistic activities integration play in achieving profit maximisation is important (Crevier et al., 2012).

12.5 AIR FREIGHT PRICING Air transport produces relevant environmental impacts such as emissions and noise. This paragraph investigates air pricing with specific reference to its freight segment, while Chapter 11 in this Handbook provides a literature review on airport pricing literature of the last 10–15 years. One peculiar problem of this mode relates to externality allocation between passenger and freight since the same plane, thus its emissions, are ascribable to both. A relevant part of air freight shipments is performed via passenger belly cargo shipments (Feng et al., 2015). The allocation problem is hard and cumbersome since passenger aircraft movements would take place independently of the presence/absence of freight in the belly. In principle, one should calculate as freight-related externalities only the marginal costs due to the additional cargo transported by passenger planes. All this is not relevant for the full cargo flights such as those express couriers/postal service operators use for overnight flights where the importance of noise is paramount and, in some cases, must be confronted with noise curfews. No wonder, given these complexities, data are scarce, not distinguishing between passenger and freight. The European Commission reports that average air pollution costs for short, medium, and long haul are, respectively, 0.30, 0.13, and 0.06 €-cent/pkm for the 33 selected European airports considered (EC, 2019). The share of global emissions that air cargo produces is influenced by operational and regulatory restrictions (Morrell & Klein, 2018). In fact: (1) scant and low-quality infrastructure availability produces higher emissions due to longer turnaround times; (2) no-fly areas (e.g. military zones) imply using longer than optimal routes (e.g. the case of Belarus in 2021, EASA, 2021a); (3) bilateral air traffic agreements often produce a substantial reduction in service efficiency with multilateral ones that are harder to come by (e.g. the African example,

Pricing in freight transport  241

Abate, 2016; Warnock-Smith & Njoya, 2017). Anser et al. (2021) investigate the role of air freight pricing in mitigating GHG emissions in a 1995–2018 panel of 39 countries all over the world, finding out that efficient environmental regulation would be able to finance technological innovation-based markets and promote sustainable development. The noise problem is considered relevant worldwide. The typical policies deployed to address it are operational restrictions, penalties, charges, and night curfews. More in detail, airports: (1) apply both landing and take-off noise-abating procedures (e.g. designated runways, maximum noise allowances, etc.); (2) levy penalties/fines when planes deviate from flight path causing an excessive nuisance (night/day differentiation); (3) enforce surcharges/ discounts linked to aircraft weight landing fees; (4) adopt night curfews/restrictions on planes activities (Morrell & Klein, 2018). Not all these options, aimed at internalising external costs, can equally/easily be introduced at an international level due to prohibitions linked to Air Services Agreements (ICAO, 2021). Furthermore, from a strictly legal point of view, the situation is somewhat convoluted. European airports can only levy charges related to the use of airport facilities and activities, such as landing, take-off, lighting, aircraft parking, and passengers/freight processing (EC, 2020a) but they have no legal status to do so. At the same time, the state, having this status, has no access to the data needed to perform an accurate job and must refer to the ICAO Aircraft Engine Emissions Databank (EASA, 2021b). The environmental stance of the European aviation strategy rests upon three main pillars: (1) cooperate with ICAO in developing a Global Market-Based Mechanism; (2) complete the Single European Sky; and (3) promote bilateral and comprehensive aviation agreements with the most important foreign partners to promote worldwide green strategies.4 The most relevant actions deriving from the above-mentioned strategies are the following: (i) promote research and development for greener technology; (ii) modernise air traffic management systems; and (iii) support market-based measures heavily relying upon the ETS, in line with the EC Directive 2008/101/EC (EC, 2008) and ICAO A35-5 resolution foreseeing incorporating international aviation into existing trading schemes (EASA, 2019). Furthermore, the ETS revision, in line with the EU’s 2030 climate objectives, will support the implementation of the Carbon Offsetting and Reduction Scheme for International Aviation (CORSIA), foreseen in 2021, which will be included in the broader perspective of the European Green Deal. The pre-COVID-19 air transport industry had gradually increased its global passenger and freight traffic share, with a strong correlation between gross domestic product variations and world air freight traffic. Air freight prices increased substantially since the COVID-19 pandemic outbreak, mainly due to reduced capacity. However, international air freight prices are also changing due to specific market conditions, such as fluctuating oil prices, restrictive G20 government measures, and acute geopolitical concerns (IATA, 2021). Freight transport has greatly benefitted from increasingly sophisticated operations made possible by new technologies that increase reliance, facilitate storage, tracking consignments, and reducing shipment fulfilment times. The perimeter of the air freight market is difficult to locate due to providers’ heterogeneity characterised by different capabilities, logistic skills, and service sophistication. The main air freight operators are: (1) line haul; (2) integrated/courier/express; (3) niche (Reynolds-Feighan, 2017). Line haul operators transport cargo from airport A to airport B relying on freight forwarders/consolidators to deal directly with customers. They can be: (i) all cargo operators (scheduled/non-scheduled); (ii) combination of passenger and cargo operators (using both dedicated freighter aircraft and the belly holds of passenger

242  Handbook on transport pricing and financing

aircraft); (iii) passenger operators using only belly holds in passenger aircraft. Integrated/ courier/express focus on door-to-door shipments with time-constrained delivery services typically operating quite a reliable aircraft compatible with low utilisation levels. Niche operators, instead, relying upon specialised equipment/expertise, focus on extraordinary requirements not compatible with standard operators’ capabilities. This differentiation among air freight operators substantiates the heterogeneity in service, marketing, and sales. Line haul operators only sell a limited part of their cargo space directly to customers with most of their supply sold through general sales agents or freight forwarders which, in turn, sell freight space to final customers. Line haul airlines, in accordance with IATA requirements, publish their cargo tariff conference levels. However, only a few customers pay these tariffs just representing an upper bound for air cargo rates. In fact, volume, density, weight, commodity type, routing, season, shipment regularity, priority, and speed discounts are common. Integrated operators typically offer a set of transport services priced according to delivery weight and speed with discounts linked to volume and regularity. They also offer multi-modal service leveraging on the distance, cost, and time trade-offs characterising the different transport modes. This also comprises, in a non-negligible part, “air trucking” where air cargo shipments take place by road. Niche operators typically adopt an ad-hoc pricing strategy. Actual rates, characterising the forwarder/airline relationships, are typically confidential. Furthermore, air cargo firms apply more sophisticated yield management techniques with respect to passengers, since both weight and volume constraints apply in determining cargo capacity limits (Pak & Dekker, 2004). More recently, Qin et al. (2012) propose an optimisation model of the single-leg air cargo space inventory control and overbooking in high-demand season based on a Markov decision process. Huang and Lu (2015), on the other hand, develop a multi-dimensional dynamic programming model to address a network revenue management problem for air cargoes, providing a decision support tool for airlines overcoming computational challenges. Finally, Wong and Ling (2020) discuss a mathematical/optimisation tool aimed at assisting air cargo load planning so to maximise profits and improve operational efficiency. Freight cargo is not only characterised by heterogeneous handling/packing costs but also by hard-to-predict actual space availability since this becomes known only close to flight departure time due to uncertain passenger baggage (volume/weight) incidence (Pettersen Strandenes, 2013).

12.6 MARITIME FREIGHT PRICING A relevant and increasing proportion of global greenhouse gas emissions is due to the maritime sector (Asariotis & Benamara, 2012). In case appropriate mitigating strategies are not deployed, it is reasonable to expect an additional increase in the emissions this mode will produce in the future (IMO, 2021). Accidents, air pollution, and greenhouse gas emissions are the most frequently studied forms of external costs maritime freight transport produces. The estimated average external accident costs are 318€ per port call and 36,524€ per million tonnes (EC, 2019). Maritime transportation is one of the least polluting modes due to its large transportation capacity. It mainly accounts for global air pollution since a large part of the emissions are generated far away from populated areas (Demir et al., 2015).

Pricing in freight transport  243

Internalising the external costs of maritime freight transport is responsible for necessarily needs addressing the problem from a differentiated geographical scale perspective. Tavasszy et  al. (2014) recall that a major internalisation obstacle relates to the uncertain economic impact this might have on different economic sectors and regions of the world. More specifically, concerning the Dutch economy, they discover significant internalisation effects for specific sectors yet rather small ones on overall trade/transport when performing an investigation of worldwide supply chains considering production/redistribution effects and flow/pattern changes in global transport. It is difficult to ensure proportionality between maritime-trafficgenerated societal costs and internalising policy measures, with the most important problem being the impact on national waters (Vierth & Merkel, 2020). Carbon pricing is an efficient internalisation policy also for this mode (see, for example, IMF-WB, 2011 and Keen et  al., 2013). In principle, one could implement this policy by imposing a tariff on carbon emissions, while, in practice, this internalisation strategy is hard to put in place on a global scale. The hot political debate taking place on such an issue confirms this, while a regional adoption seems feasible. Even in this case, several economic, legal, and political challenges have to be duly taken care of first. Jurisdiction, data availability, environmental effectiveness, avoidance strategies, competitiveness repercussions, developing country exceptions, and incentives are the critical aspects one needs to consider (Dominioni et al., 2018). Research suggests that cargo-based measures including emissions produced throughout the whole trip are to be preferred to other carbon-pricing scheme options, such as a fuel tax or a levy. Shipping is, overall, less exposed to incentivebased regulation than road transport even if there are some noticeable exceptions, such as in Sweden where fairway dues are differentiated according to ships’ environmental performance (Vierth & Merkel, 2020). Port pricing schemes are complex and reflect the different service types/standards offered. The intricacy of this multi-pricing structure is, at times, bewildering. It often produces difficult-to-understand combinations forestalling competition among ports. This depends on the thorny decision ship operators and cargo owners are confronted with when striving to make the most convenient choice (Meersman, et al., 2014a). The main cost categories are port-calling, terminal-handling, and concession pricing. They, respectively, refer to costs incurred for: 1) vessel-related services offered (e.g. access to quay/ terminal, pilotage, bunkering, etc.); 2) services needed to move the cargo both in the port and through the supply chain (e.g. loading/unloading, storage, customs clearance, repacking, forwarding, etc.); and 3) acquiring a dedicated terminal (Meersman et al., 2014b). Ports are heterogeneous. This needs to be considered when investigating port pricing structure/models. There is no generally representative port pricing scheme. All the stakeholders involved focus their attention on their cost structure reflected in their pricing behaviour (Meersman et al., 2014b). Papers investigating port pricing typically adopt a port-operator perspective when studying efficient infrastructure cost coverage. This is relevant since a strict marginal pricing approach, due to relevant economies of scale, would necessitate public subsidies (Talley, 1994; Meersman, et al., 2014a; Acciaro, 2013). In a multi-modal port pricing setting, one has to determine the relative attractiveness of shipping vis-a-vis other competing modes (Haralambides, 2002; Bergantino, 2002; Strandenes & Marlow, 2000). Strandenes and Marlow (2000) categorise port pricing structures as follows: (1) cost-based pricing; (2) cost recovery; (3) congestion pricing; (4) strategic port pricing; (5) quality pricing.

244  Handbook on transport pricing and financing

Additionally, ship owners, when confronted with actual decisions, have to consider other elements such as discounts, adjustments, and rebates, thus making the actual price paid difficult to calculate (Meersman et al., 2014b). Current port pricing schemes adopt a linear approach, ignoring price differentiation and revenue management strategies thus not fully exploiting profit maximisation and efficient capacity allocation opportunities (Meersman, et  al., 2014a). Last but not least, King et  al. (2014) propose an interesting and different “intermodal” perspective to understand how road pricing can affect freight activity and costs in ports. Inland water transport (IWT) not only represents a great opportunity to promote intermodality in landlocked countries, but also provides a viable alternative to road and rail travel, offering a low-energy, low-noise, low-emission transport system (Hanaoka & Regmi, 2011). Due to cheap infrastructure and external expenditures, IWT is frequently the most cost-effective route of inland transport (UNECE, 2011). When it comes to external costs of climate change, IWT already has a significant advantage over heavy goods road transport (40% lower gas emissions, according to Hofbauer and Putz, 2020). However, IWT is commonly underused due to several factors, including a lack of infrastructure investment and supporting policies from local governments (Caris et al., 2014).

12.7 CONCLUSIONS This chapter has investigated freight transport pricing concepts analysing the specific characteristics pertinent to the road, rail, air, and maritime segments. The chapter discusses both the characteristics of the external costs due to freight transport and the main internalisation strategies public authorities might adopt with attention paid to: (1) sector characteristics; (2) geographical dimension; (3) intermodal relationships. Moreover, the chapter illustrates the peculiarities pertaining to each freight transport segment discussing both theoretical and practical issues. The most important peculiarity of freight transport pricing pertains to the relationship between transport and logistic elements where the former is just a component (usually a minor one) of the latter. Comprehending freight transport pricing and forecasting the impacts of changes is thus more complex also due to the intrinsic complexity of the freight transport ecosystem. These considerations suggest that whenever intervening in this realm, caution is necessary due to the presence of heterogeneity (Gatta & Marcucci, 2014), non-linearities (Gatta & Marcucci, 2016), and interaction effects (Marcucci et  al., 2017a).  To produce long-term desirable effects in terms of environmental, economic, and social objectives, it is highly recommended to adopt a well-structured participatory planning approach. This is also testified by the European Commission documents guiding the definition, deployment, and evaluation of Sustainable Urban Mobility Plans, where the Sustainable Urban Logistic component is nonmarginal (ELTIS, 2018).5 Finally, one should underline that freight transport pricing is a complex, heterogeneous, and articulated subject characterised by different hurdles one must overcome when trying to provide a theoretically robust, comprehensive, and well-balanced account of the different issues influencing its current status and future evolutions. Moreover, in the last decades the sector has been witnessing a general shift toward smartness and sustainability (Zamparini & Reggiani, 2021), which deserves additional attention and research. This should adopt an

Pricing in freight transport  245

interdisciplinary stance where economics, finance, management, operation research, regulation, and law (just to mention the most relevant ones) competences should intermingle in a functionally fruitful perspective so as to disentangle the hard-to-solve problems deriving from an interaction between different stakeholders, institutions, geographical scales, transport modes, and business models.

NOTES 1. 2.

For more details, please see Button (2010) and De Borger and Proost (2001). One should note that if the markup is determined on the basis of “what the market will bear”, then the value-of-service and the cost-of-service approaches are almost equivalent due to the endogeneity of the mark-up. 3. According to International Chamber of Commerce (https://iccwbo​.org​/resources​-for​-business​/ incoterms​-rules​/incoterms​-2020/) INCOTERMS stands for International Commercial Terms, i.e. the world’s essential terms of trade for the sale of goods. 4. For further details, please refer to https://ec​.europa​.eu​/transport​/modes​/air​/environment​_en. 5. Under this respect, there are some noticeable contributions pertaining to each of the four steps foreseen within a SUMP. In particular: (i) preparation and analysis (Gatta et  al., 2017, 2019b; Lindenau & Böhler-Baedeker, 2014; Ibeas et al., 2011; Moore & Elliott, 2016); (ii) strategy development (Marcucci et al., 2015, 2018; Polinori et al., 2018; Poplin, 2012; Loukopoulos, & Scholz, 2004; Gabrielli et al., 2014); (iii) measure planning (Le Pira et al., 2017a, 2017b; Gatta et al., 2019a; Macário & Marques, 2008; Okraszewska et al., 2018; Torrisi et al., 2020); (iv) identification and monitoring (Le Pira et al., 2017c; Marcucci et al., 2017b, 2020; Wefering et al., 2014; Morfoulaki & Papathanasiou, 2021; May, 2015).

REFERENCES Abate, M. (2016). Economic effects of air transport market liberalization in Africa. Transportation Research Part A: Policy and Practice, 92, 326–337. Acciaro, M. (2013). A critical review of port pricing literature: What role for academic research? The Asian Journal of Shipping and Logistics, 29(2), 207–228. Allen, J., Browne, M., & Cherrett, T. (2012). Investigating relationships between road freight transport, facility location, logistics management and urban form. Journal of Transport Geography, 24, 45–57. Anand, N., van Duin, R., & Tavasszy, L. (2021). Carbon credits and urban freight consolidation: An experiment using agent based simulation. Research in Transportation Economics, 85, 100797. Anser, M. K., Ali, S., Khan, M. A., Nassani, A. A., Askar, S. E., Zaman, K.,... & Kabbani, A. (2021). The role of sustainable technological innovations in the relationship between freight pricing and environmental degradation: Evidence from a panel of 39 R&D economies. Atmósfera, 35(2), 395–420. Asariotis, R., & Benamara, H. (Eds.). (2012). Maritime Transport and the Climate Change Challenge. Routledge. Auerbach, A., Chetty, R., Feldstein, M., & Saez, E. (eds.) (2013). Handbook of Public Economics, vol. 5. Amsterdam: North Holland. Austin, D. (2015). Pricing freight transport to account for external costs. Working Paper Series Congressional Budget Office, Washington, DC. Bergantino, A. S. (2002). The European commission approach to port policy: Some open issues. International Journal of Transport Economics, 337–379. Bitzan, J. D., & Keeler, T. E. (2014). The evolution of US rail freight pricing in the post-deregulation era: Revenues versus marginal costs for five commodity types. Transportation, 41(2), 305–324. Boulter, P., & McCrae, I. (2009). ARTEMIS: Assessment and reliability of transport emission models and inventory systems: Final report. Technical Report. Berkshire, United Kingdom.

246  Handbook on transport pricing and financing

Button, K. (2010). Transport Economics. Edward Elgar Publishing. Caris, A., Limbourg, S., Macharis, C., Van Lier, T., & Cools, M. (2014). Integration of inland waterway transport in the intermodal supply chain: A taxonomy of research challenges. Journal of Transport Geography, 41, 126–136. Cavallaro, F., Giaretta, F., & Nocera, S. (2018). The potential of road pricing schemes to reduce carbon emissions. Transport Policy, 67, 85–92. Chapuis, R., Delporte, T., Köpke, S., Lotz, C., & Ott, A. (2022). Bold Moves to Boost European Rail Freight. McKinsey & Company. Chen, D., Ignatius, J., Sun, D., Goh, M., & Zhan, S. (2018). Impact of congestion pricing schemes on emissions and temporal shift of freight transport. Transportation Research Part E: Logistics and Transportation Review, 118, 77–105. Conrad, I. V., & Joseph, L. (2018). Costs and challenges of log truck transportation in Georgia, USA. Forests, 9(10), 650. Costello, B., & Suarez, R. (2015). Truck Driver Shortage Analysis 2015. Arlington, VA: The American Trucking Associations. Crevier, B., Cordeau, J.-F., & Savard, G. (2012). Integrated operations planning and revenue management for rail freight transportation. Transportation Research Part B: Methodological, 46(1), 100–119. Danielis, R., Marcucci, E., & Rotaris, L. (2005). Logistics managers' stated preferences for freight service attributes. Transportation Research Part E: Logistics and Transportation Review, 41(3), 201–215. de Borger, B., & Proost, S. (Eds.). (2001). Reforming Transport Pricing in the European Union: A Modelling Approach. Edward Elgar Publishing. de Borger, Bruno, & Proost, Stef (2021). Pricing principles in the transport sector. In: Vickerman, Roger (ed.), International Encyclopedia of Transportation, vol. 3, pp. 95–101. UK: Elsevier Ltd. Demir, E., Huang, Y., Scholts, S., & Van Woensel, T. (2015). A selected review on the negative externalities of the freight transportation: Modeling and pricing. Transportation Research Part E: Logistics and Transportation Review, 77, 95–114. de Palma, A., Lindsey, R., & Proost, S. (2007a). Investment and the use of tax and toll revenues in the transport sector: The research agenda. Research in Transportation Economics, 19, 1–26. de Palma, A., Lindsey, R., Proost, S., & Van der Loo, S. (2007b). Comparing alternative pricing and revenue use strategies with the MOLINO model. Research in Transportation Economics, 19, 111–131. Dolinayová, A., Ľoch, M., & Kanis, J. (2015). Modelling the influence of wagon technical parameters on variable costs in rail freight transport. Research in Transportation Economics, 54, 33–40. Dominioni, G., Heine, D., & Romera, B. M. (2018). Regional carbon pricing for international maritime transport: Challenges and opportunities for global geographical coverage. Carbon & Climate Law Review, 12(2), 140–158. EASA (2019). European Aviation Environmental Report 2019 (last access December 21st, 2021) EASA (2021a). EASA issues Safety Directive calling on Member States to mandate avoidance of Belarus airspace. Retrieved from: https://www​.easa​.europa​.eu​/newsroom​-and​-events​/news​/easa​-issues​-safety​ -directive​-calling​-member​-states​-mandate​-avoidance (last access December 21st, 2021) EASA (2021b). ICAO Aircraft Engine Emissions Databank. Retrieved from: https://www​.easa​.europa​ .eu​/domains​/environment​/icao​-aircraft​-engine​-emissions​-databank (last access December 21st, 2021) EC (2008). DIRECTIVE 2008/101/EC OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 19 November 2008 amending Directive 2003/87/EC so as to include aviation activities in the scheme for greenhouse gas emission allowance trading within the Community. Retrieved from: https://eur​-lex​.europa​.eu​/ legal​-content​/ EN​/ TXT​/ PDF/​?uri​= CELEX​:32008L0101​&from​=EN (last access December 21st, 2021) EC (2019). Transport taxes and charges in Europe An overview study of economic internalisation measures applied in Europe. Transport taxes and charges in Europe - Publications Office of the EU (europa​.​eu) EC (2020a). COMMISSION STAFF WORKING DOCUMENT Full-length report Accompanying the document Report from the Commission to the European Parliament and the Council Updated analysis of the non-CO2 climate impacts of aviation and potential policy measures pursuant to EU Emissions Trading System Directive Article 30(4) {COM(2020) 747 final}

Pricing in freight transport  247

EC (2020b). Excise Duty Tables, Part II Energy products and Electricity, In accordance with the Energy Directive (Council Directive 2003/96/EC), Brussels: European Commission (EC), DirectorateGeneral Taxation and Customs Union. ELTIS (2018). The Urban Mobility Observatory, Planning sustainable urban logistics. Retrieved from: https://www​.eltis​.org​/resources​/tools​/planning​-sustainable​-urban​-logistics (last access December 21st, 2021) Eurostat (2020). Freight Transport Statistics. Retrieved from: https://ec​.europa​.eu​/eurostat​/statistics​ -explained​/index​.php​?title​=Freight​_transport​_statistics​#Modal​_split (last access December 21st, 2021) Feng, B., Li, Y., & Shen, Z. J. M. (2015). Air cargo operations: Literature review and comparison with practices. Transportation Research Part C: Emerging Technologies, 56, 263–280. FHWA, Federal Highway Administration, Addendum to the 1997 Federal Highway Cost Allocation Study Final Report (May 2000). Retrieved from: www​.fhwa​.dot​.gov​/policy​/ hcas​/addendum​.htm (last access December 21st, 2021) Gabrielli, S., Forbes, P., Jylhä, A., Wells, S., Sirén, M., Hemminki, S.,... & Jacucci, G. (2014). Design challenges in motivating change for sustainable urban mobility. Computers in Human Behavior, 41, 416–423. Gatta, V. & Marcucci, E. (2014). Urban freight transport and policy changes: Improving decision makers’ awareness via an agent-specific approach. Transport Policy, 36, 248–252. Gatta, V., & Marcucci, E. (2016). Behavioural implications of non-linear effects on urban freight transport policies: The case of retailers and transport providers in Rome. Case Studies on Transport Policy, 4(1), 22–28. Gatta, V. Marcucci, E., Delle Site, P., Le Pira, M., & Carrocci, C. S. (2019b). Planning with stakeholders: Analysing alternative off-hour delivery solutions via an interactive multi-criteria approach. Research in Transportation Economics, 73, 53–62. Gatta, V., Marcucci, E., & Le Pira, M. (2017). Smart urban freight planning process: Integrating desk, living lab and modelling approaches in decision-making. European Transport Research Review, 9(3), 1–11. Gatta, V., Marcucci, E., Nigro, M., & Serafini, S. (2019a). Sustainable urban freight transport adopting public transport-based crowdshipping for B2C deliveries. European Transport Research Review, 11(1), 1–14. Gomez, J., & Vassallo, J. M. (2020). Has heavy vehicle tolling in Europe been effective in reducing road freight transport and promoting modal shift? Transportation, 47(2), 865–892. Gorman, M. F. (2010). Hub group implements a suite of OR tools to improve operations. Interfaces 40(5), 368–384. Hanaoka, S., & Regmi, M. B. (2011). Promoting intermodal freight transport through the development of dry ports in Asia: An environmental perspective. Iatss Research, 35(1), 16–23. Haralambides, H. E. (2002). Competition, excess capacity, and the pricing of port infrastructure. International Journal of Maritime Economics, 4(4), 323–347. Heitz, A., Dablanc, L., & Tavasszy, L. A. (2017). Logistics sprawl in monocentric and polycentric metropolitan areas: The cases of Paris, France, and the Randstad, the Netherlands. Region, 4(1), 93–107. Hesse, M., & Rodrigue, J. P. (2004). The transport geography of logistics and freight distribution. Journal of Transport Geography, 12(3), 171–184. Hofbauer, F., & Putz, L. M. (2020). External costs in Inland waterway transport: An analysis of external cost categories and calculation methods. Sustainability, 12(14), 5874. Holguín-Veras, J. (2010). The truth, the myths and the possible in freight road pricing in congested urban areas. Procedia-Social and Behavioral Sciences, 2(3), 6366–6377. Holguín-Veras, J. (2011). Urban delivery industry response to cordon pricing, time–distance pricing, and carrier–receiver policies in competitive markets. Transportation Research Part A: Policy and Practice, 45(8), 802–824. Holguín-Veras, J., Wang, Q., Xu, N., Ozbay, K., Cetin, M., & Polimeni, J. (2006). The impacts of time of day pricing on the behavior of freight carriers in a congested urban area: Implications to road pricing. Transportation Research Part A: Policy and Practice, 40(9), 744–766. Huang, K., & Lu, H. (2015). A linear programming-based method for the network revenue management problem of air cargo. Transportation Research Procedia, 7, 459–473.

248  Handbook on transport pricing and financing

IATA (2021). Air Cargo Tariffs and Rules: What You Need to Know. www​.iata​.org/. Ibeas, A., dell’Olio, L., & Montequín, R. B. (2011). Citizen involvement in promoting sustainable mobility. Journal of Transport Geography, 19(4), 475–487. ICAO (2021). Balanced Approach to Aircraft Noise Management. Retrieved from: https://www​.icao​.int​ /environmental​-protection​/ Pages​/noise​.aspx (last access December 21st, 2021) IMF-WB (International Monetary Fund and World Bank) (2011). Market-based instruments for international aviation and shipping as a source of climate finance. Retrieved from: https://www​.imf​ .org​/external​/np​/g20​/pdf​/110411a​.pdf (last access December 21st, 2021) IMO (2021). Fourth IMO GHG Study 2020. Retrieved from: https://www​ .imo​ .org​ /en​ /OurWork​ / Environment​/ Pages​/ Fourth​-IMO​-Greenhouse​-Gas​-Study​-2020​.aspx (last access December 21st, 2021) Janic, M. (2007). Modelling the full costs of an intermodal and road freight transport network. Transportation Research Part D: Transport and Environment, 12(1), 33–44. Jiang, Y., Lu, J., Li, J., & Wang, L. (2017). Dynamic impacts of Harbor Tolls Policy on China’s port economy: The case of Zhanjiang Port. Research in Transportation Economics, 61, 37–43. Keen, M., & Parry, I., & Strand, J. (2013). Planes, ships and taxes: Charging for international aviation and maritime emissions. Economic Policy, 28(76), 701–749. King, D. A., Gordon, C. E., & Peters, J. R. (2014). Does road pricing affect port freight activity: Recent evidence from the port of New York and New Jersey. Research in Transportation Economics, 44, 2–11. Kraft, E. (2002). Scheduling railway freight delivery appointments using a bid price approach. Transportation Research Part A: Policy and Practice, 36, 145–165. Krugman, P. & Wells, R. (2021). Microeconomics. MacMillan, VI ed. Lambert, D. M., Stock, J. R., & Ellram, L. M. (1998). Fundamentals of Logistics Management. Boston, MA: McGraw-Hill International Editions. .lta​ .gov​ .sg​ /content​ / Land Transport Authority (LTA) (2020). Retrieved from: https://onemotoring​ onemotoring​/ home​/driving​/ ERP​.html (last access December 21st, 2021) Le Pira, M., Marcucci, E., & Gatta, V. (2017c). Role-playing games as a mean to validate agent-based models: An application to stakeholder-driven urban freight transport policy-making. Transportation Research Procedia, 27, 404–411. Le Pira, M., Marcucci, E., Gatta, V., Ignaccolo, M., Inturri, G., & Pluchino, A. (2017a). Towards a decision-support procedure to foster stakeholder involvement and acceptability of urban freight transport policies. European Transport Research Review, 9(4), 1–14. Le Pira, M., Marcucci, E., Gatta, V., Inturri, G., Ignaccolo, M., & Pluchino, A. (2017b). Integrating discrete choice models and agent-based models for ex-ante evaluation of stakeholder policy acceptability in urban freight transport. Research in Transportation Economics, 64, 13–25. Lehe, L. (2019). Downtown congestion pricing in practice. Transportation Research Part C: Emerging Technologies, 100, 200–223. Li, L., & Zhang, X. (2020). Reducing CO2 emissions through pricing, planning, and subsidizing rail freight. Transportation Research Part D: Transport and Environment, 87, 102483. Lin, N., Hjelle, H. M., & Bergqvist, R. (2020). The impact of an upstream buyer consolidation and downstream intermodal rail-based solution on logistics cost in the China-Europe container trades. Case Studies on Transport Policy, 8(3), 1073–1086. Lin, W. (2019). Transport geography and geopolitics: Visions, rules and militarism in China’s Belt and Road Initiative and beyond. Journal of Transport Geography, 81, 102436. Lindenau, M., & Böhler-Baedeker, S. (2014). Citizen and stakeholder involvement: A precondition for sustainable urban mobility. Transportation Research Procedia, 4, 347–360. Lindsey, R., & Santos, G. (2020). Addressing transportation and environmental externalities with economics: Are policy makers listening? Research in Transportation Economics, 82, 100872. Ljungberg, A. (2013). Impacts of increased rail infrastructure charges in Sweden. Research in Transportation Economics, 39(1), 90–103. Loukopoulos, P., & Scholz, R. W. (2004). Sustainable future urban mobility: Using ‘area development negotiations’ for scenario assessment and participatory strategic planning. Environment and Planning A, 36(12), 2203–2226.

Pricing in freight transport  249

Macário, R., & Marques, C. F. (2008). Transferability of sustainable urban mobility measures. Research in Transportation Economics, 22(1), 146–156. Marcucci, E., Gatta, V., & Le Pira, M. (2018). Gamification design to foster stakeholder engagement and behavior change: An application to urban freight transport. Transportation Research Part A: Policy and Practice, 118, 119–132. Marcucci, E., Gatta, V., Le Pira, M., Hansson, L., & Bråthen, S. (2020). Digital twins: A critical discussion on their potential for supporting policy-making and planning in urban logistics. Sustainability, 12(24), 10623. Marcucci, E., Gatta, V., Marciani, M., & Cossu, P. (2017b). Measuring the effects of an urban freight policy package defined via a collaborative governance model. Research in Transportation Economics, 65, 3–9. Marcucci, E., Gatta, V., & Scaccia, L. (2015). Urban freight, parking and pricing policies: An evaluation from a transport providers’ perspective. Transportation Research Part A: Policy and Practice, 74, 239–249. Marcucci, E., Le Pira, M., Gatta, V., Inturri, G., Ignaccolo, M., & Pluchino, A. (2017a). Simulating participatory urban freight transport policy-making: Accounting for heterogeneous stakeholders’ preferences and interaction effects. Transportation Research Part E: Logistics and Transportation Review, 103, 69–86. Martin, T. C., & Thoresen, T. R. (2015). Estimation of the marginal cost of road wear as a basis for charging freight vehicles. Research in Transportation Economics, 49, 55–64. May, A. D. (2015). Encouraging good practice in the development of sustainable urban mobility plans. Case Studies on Transport Policy, 3(1), 3–11. Meersman, H., Strandenes, S. P., & Van de Voorde, E. (2014a). Port Pricing: Principles, Structure and Models. NHH Dept. of Economics Discussion Paper, (14). Meersman, H. & Van de Voorde, E. (2013). The relationship between economic activity and freight transport. In Ben-Akiva, M., Meersman, H. & Van de Voorde, E. (Eds.), Freight Transport Modelling, pp. 15–43. Bingley: Emerald Group Publishing Limited. Meersman, H., Van De Voorde, E., & Vanelslander, T. (2014b). Future Challenges for the Port and Shipping Sector. CRC Press. Merkert, R., & Hensher, D. A. (2014). Open access for railways and transaction cost economics– Management perspectives of Australia’s rail companies. Research in Transportation Economics, 48, 227–236. Milaković, A.S., Gunnarsson, B., Balmasov, S., Hong, S., Kim, K., Schütz, P., Ehlers, S. (2018). Current status and future operational models for transit shipping along the Northern Sea Route. Marine Policy, 94, 53–60. Moore, K. R., & Elliott, T. J. (2016). From participatory design to a listening infrastructure: A case of urban planning and participation. Journal of Business and Technical Communication, 30(1), 59–84. Morfoulaki, M., & Papathanasiou, J. (2021). Use of the Sustainable Mobility Efficiency Index (SMEI) for Enhancing the Sustainable Urban Mobility in Greek Cities. Sustainability, 13(4), 1709. Morrell, P. S., & Klein, T. (2018). Moving Boxes by Air: The Economics of International Air Cargo. Routledge. Mostert, M., Caris, A., & Limbourg, S. (2017). Road and intermodal transport performance: The impact of operational costs and air pollution external costs. Research in Transportation Business & Management, 23, 75–85. Nash, C. A., Smith, A. S., van de Velde, D., Mizutani, F., & Uranishi, S. (2014). Structural reforms in the railways: Incentive misalignment and cost implications. Research in Transportation Economics, 48, 16–23. Nilsson–VTI, J. E., Isacsson–VTI, G., Haraldsson–VTI, M., Nerhagen–VTI, L., Odolinski–VTI, K., Swärdh–VTI, J. E.,... & Österström–VTI, J. (2018). The efficient use of infrastructure–is Sweden pricing traffic on its roads, railways, waters and airways at marginal costs?, CTS Working Paper 2018(2), Centre for Transport Studies, Stockholm, SE. Okraszewska, R., Romanowska, A., Wołek, M., Oskarbski, J., Birr, K., & Jamroz, K. (2018). Integration of a multilevel transport system model into sustainable urban mobility planning. Sustainability, 10(2), 479.

250  Handbook on transport pricing and financing

Pak, K., & Dekker, R. (2004). Cargo revenue management: Bid-prices for a 0-1 multi knapsack problem. 2004 (8). Retrieved from: SSRN: https://ssrn​.com​/abstract​=594991 (last access December 21st, 2021) Perera, L., Thompson, R. G., & Wu, W. (2020). A multi-class toll-based approach to reduce total emissions on roads for sustainable urban transportation. Sustainable Cities and Society, 63, 102435. Perera, L., Thompson, R. G., & Wu, W. (2021). Toll and subsidy for freight vehicles on urban roads: A policy decision for City Logistics. Research in Transportation Economics, 90, 101132. Pettersen Strandenes, S. (2013). Freight transport pricing models. In Freight Transport Modelling, pp. 319–331. Emerald Group Publishing Limited. Polinori, P., Marcucci, E., Bollino, C. A., Bigerna, S., & Gatta, V. (2018). Eco-labeling and sustainable urban freight transport: how much are people willing to pay for green logistics? 631–658. Poplin, A. (2012). Playful public participation in urban planning: A case study for online serious games. Computers, Environment and Urban Systems, 36(3), 195–206. Qin, C. R., Luo, L., You, Y., & Xiao, Y. X. (2012). An optimization model of the single-leg air cargo space control based on Markov decision process. Journal of Applied Mathematics, 2012. Qu, C., Wang, G. W., & Zeng, Q. (2017). Modelling port subsidy policies considering pricing decisions of feeder carriers. Transportation Research Part E: Logistics and Transportation Review, 99, 115–133. Reynolds-Feighan, A. J. (2017). Air-freight logistics. In Handbook of Logistics and Supply-Chain Management. Emerald Group Publishing Limited. Rodrigue, J. P., Comtois, C., & Slack, B. (2016). The Geography of Transport Systems. Routledge. Santos, G., Behrendt, H., Maconi, L., Shirvani, T., & Teytelboym, A. (2010). Part I: Externalities and economic policies in road transport. Research in Transportation Economics, 28(1), 2–45. Schnell-Lortet, I., & Jacob, B. (2019). Long Distance Freight Transport. A roadmap for System integration of Road Transport. ERTRAC: Brussels, Belgium. Retrieved from: https://www​.ertrac​. org​/uploads​/documentsearch ​/id56​/ ERTRAC​-Long​-duty​-Freight​-Transport​-Roadmap2019​.pdf (last access December 21st, 2021). Schrank, D., Eisele, B., Lomax, T., & Bak, J. (2015). 2015 Urban mobility scorecard. https://static​.tti​. tamu​.edu​/tti​.tamu​.edu​/documents​/mobility​-scorecard​-2015​-wappx​.pdf Schroten, A., van Essen, H., van Wijngaarden, L., Sutter, D., & Andrew, E. (2019). Sustainable Transport Infrastructure Charging and Internalisation of Transport Externalities: Executive Summary. 2019. Retrieved from: https://op​.europa​.eu​/en​/publication​-detail/-​/publication​/0efedf2c​-a386​-11e9​-9d01​01aa75ed71a1 Sibdari, S., Lin, K.Y., & Chellappan, S. (2008). Multiproduct revenue management: An empirical study of Auto train at Amtrak. Journal of Revenue and Pricing Management, 7(2):172–184. Strandenes, S. P., & Marlow, P. B. (2000). Port pricing and competitiveness in short sea shipping.  International Journal of Transport Economics/Rivista Internazionale di Economia dei Trasporti, 315–334. Talley, W. K. (1994). Port pricing: A cost axiomatic approach. Maritime Policy and Management, 21(1), 61–76. Taniguchi, E., Thompson, R. G., & Yamada, T. (2016). New opportunities and challenges for city logistics. Transportation Research Procedia, 12, 5–13. Tavasszy, L., Harmsen, J., Ivanova, A., & Bulavskaya, T. (2014). Effect of a full internalization of external costs of global supply chains on production, trade and transport. Transport Research Arena, 2, 337–351. Teo, J. S., Taniguchi, E., & Qureshi, A. G. (2014). Evaluation of load factor control and urban freight road pricing joint schemes with multi-agent systems learning models. Procedia-Social and Behavioral Sciences, 125, 62–74. Toll EU (2020). Retrieved from: https://www​.tolls​.eu/ (last access December 21st, 2021) Torrisi, V., Garau, C., Ignaccolo, M., & Inturri, G. (2020, July). Sustainable Urban Mobility Plans: Key Concepts and a Critical Revision on SUMPs Guidelines. In International Conference on Computational Science and Its Applications, pp. 613–628. Springer, Cham. U.S. Department of Transportation, Bureau of Transportation Statistics (2020a). Transportation Statistics Annual Report 2020. Washington, DC. U.S. Department of Transportation, Bureau of Transportation Statistics (2020b). Retrieved from: https:// www​.bts​.gov​/transborder

Pricing in freight transport  251

UNECE (2011). Efficient and Sustainable Inland Water Transport in Europe. White Paper. Retrieved from: https://unece​.org​/ DAM​/trans​/doc​/2011​/sc3wp3​/ ECE​-TRANS​-SC3​-189e​.pdf van Essen, H., van Wijngaarden, L., Schroten, A., Sutter, D., Bieler, C., Maffii, S.,... & El Beyrouty, K. (2019). Handbook on the External Costs of Transport, version 2019 (No. 18.4 K83. 131). Varian, H. R. (1989). Price discrimination. In Handbook of Industrial Organization, Chapter 10, Elsevier, Volume 1, 597–654. Vierth, I., & Merkel, A. (2020). Internalization of external and infrastructure costs related to maritime transport in Sweden. Research in Transportation Business & Management, 100580. Wang, H., Han, J., Su, M., Wan, S., & Zhang, Z. (2021). The relationship between freight transport and economic development: A case study of China. Research in Transportation Economics, 85, 100885. Warnock-Smith, D., & Njoya, E. T. (2017). The development of air service agreements in Africa. In The Economics and Political Economy of African Air Transport, pp. 61–79. Routledge. Wefering F., Rupprecht S., Bührmann S., & Böhler-Baedeker S. (2014). Guidelines Developing and Implementing a Sustainable Urban Mobility Plan. Rupprecht Consult – Forschung und Beratung GmbH. Williams, N., & Murray, D. (2020). An Analysis of the Operational Costs of Trucking: 2020 Update. Wilson, R. B. (1993). Nonlinear Pricing. New York: Oxford University Press. Wong, E. Y., & Ling, K. K. (2020, September). A Mixed Integer Programming Approach to Air Cargo Load Planning with Multiple Aircraft Configurations and Dangerous Goods. In 2020 7th International Conference on Frontiers of Industrial Engineering (ICFIE), pp. 123–130. IEEE. Zamparini, Luca, & Reggiani, Aura (2021). Freight transport policy. In Vickerman, Roger (ed.), International Encyclopedia of Transportation, vol. 3, pp. 29–34. UK: Elsevier Ltd. Zhang, D., Wang, X., Holguín-Veras, J., & Zou, W. (2019). Investigation of carriers’ ability to transfer toll increases: An empirical analysis of freight agents’ relative market power. Transportation, 46(6), 2291–2308.

13. Connected and automated vehicles: effects on pricing César Núñez and Alejandro Tirachini

13.1 INTRODUCTION The definition of autonomous driving capabilities accounts for a spectrum of functionalities that start with basic driver support functions, all the way to the top level, which is complete driving autonomy (SAE International, 2021). The spectrum consists of six levels, where the first three (L0, L1, and L2) are denominated “Driver Support Systems” (which include currently available features such as cruise control and lane keeping), while L3, L4, and L5 are used for actual “Automated Driving Systems”. In this chapter, the terms human-driven vehicle (HDV) and automated vehicle (AV) will be used to refer to the bottom and top three levels of automation, respectively. A distinctive characteristic of the highest levels of AVs is their ability to use wireless technologies such as vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communications, allowing the interchange of information with other entities from their surroundings, such as other AVs, traffic signals, and toll gates, among others. AVs that are equipped with this technology are called connected and automated vehicles (CAVs). This interaction capability is expected to enable better decision-making by the vehicles, improving the capacity and safety of roads. At the same time, expected problems associated with the introduction of AVs include the possibility of software hacking and issues of privacy and liability, among other factors that may affect the choice of AVs over human-driven vehicles (Tscharaktschiew & Evangelinos, 2019). On a broader view, the technology of vehicle automation has myriad economic, social, and environmental effects, affecting transport cost and traveller’s behaviour, energy consumption, land use and urban shape, among others. The relative level of adoption of AVs for individual or shared use is crucial for the future sustainability of mobility under different scenarios of adoption of AVs. This is because the effect of AVs on energy consumption and greenhouse gas emissions crucially depends on whether future mobility is going to be mostly individual, in which we will largely replicate the current car ownership paradigm, or shared, in which people subscribe to mobility services provided by shared vehicles, either on-demand or under the classic structure of fixed-route public transport. As shown by Wadud et al. (2016), the net effect of AVs on environmental outputs depends on a balance between factors that push to reduce energy consumption (such as vehicle platooning, eco-driving, and car-sharing) and factors that increase energy consumption (such as new trips due to induced demand and a reduced cost of travel time). Long-term decisions regarding housing location and car ownership, as well as day-to-day travel decisions on mode and destination choice will be influenced by a key variable: how future transport services will be priced, including tax and subsidy decisions. Current developments suggest that, in the early stages, AVs will be mostly deployed for shared use, given the high initial cost of the technology (Stocker & Shaheen, 2017). However, 252

Connected and automated vehicles  253

as prices decrease over time, AVs will become attractive for personal use. Anticipating this scenario, it is clear that the pricing policy for future AVs and shared transport services based on AVs will have a key role to play, just like today’s pricing and tax incentives shape the transport and mobility landscape. However, the transition from human-driven to automated driving may make the application of optimal road pricing more complex, instead of easier (Tscharaktschiew & Evangelinos, 2019). An updated overview of the topics related to transport pricing and automated vehicles is presented in this chapter. In Section 13.2, we analyse the effect of vehicle automation on the modal attributes. Based on those effects, in Section 13.3 we discuss how automation will modify the transport system, and how pricing could be implemented in a scenario of automated vehicles, taking into account current insights gained in the academic literature. In Section 13.4 a transport pricing model is presented considering three modes: private car, public transport, and an active mode, incorporating the effect of vehicle automation. Finally, in Section 13.5 conclusions and recommendations are discussed, about the future of transport pricing with AVs.

13.2 VEHICLE AUTOMATION AND MODAL ATTRIBUTES In transport policy, pricing arises as a mechanism of behavioural change to deal with negative externalities caused by individual agents. So, to determine the role that pricing will play in an AV-based transport system, it is paramount to understand the role that automation will have in all the components that influence both the cost structure of transport providers and the decisions of users. Operator costs comprise infrastructure, vehicle, fuel, and crew costs. Generalised user cost includes monetary costs (fares, tolls, running costs, and/or parking in the case of cars), as well as time costs (in-vehicle, walking and waiting time) converted to money. In-vehicle time is given by the trip length and its speed, and the latter depends on traffic flow and density. Therefore, it is necessary to analyse how AVs influence each one of the aforementioned elements in order to understand the relationship between automation and pricing. 13.2.1 Effect on Road Capacity The effect of AVs on road capacity has been extensively examined in the literature. On the one hand, AVs would have shorter reaction times, leading to both reduced gaps between vehicles (Dresner & Stone, 2008), and lower road crash rates (Fagnant & Kockelman, 2015; Keeney, 2017), thus increasing the capacity of individual roads. Also, in terms of network efficiency, V2V and V2I communications would allow us to anticipate traffic conditions downstream, diminishing the likelihood of traffic breakdowns, and clearing queues faster in case of occurrence (Hoogendoorn et al., 2014). However, these predictions rely on the strong assumption of an immediate change from conventional cars to AVs, neglecting the transition phase where both technologies coexist and interact in the network. In the latter case, the capacity increase would be relevant only with a high share of AVs (Tientrakool et al., 2011). Furthermore, the introduction of AVs may even reduce the capacity if their share on the roads is low, relative to conventional vehicles (Mena-Oreja et al., 2018; Van Arem et al., 2006). Therefore, the final effect of AVs on road capacity is unclear at this stage, and likely depends on the penetration rate of AVs on the roads.

254  Handbook on transport pricing and financing

13.2.2 Effect on Travel Time and the Value of Travel Time Savings From the user’s perspective, travel time costs depend on the travel time itself and the subjective valuation that each user gives to reductions in travel time. Estimations of the effect of AVs range from slight reductions to significant increases in travel time (Chen & Kockelman, 2016; Childress et al., 2015; Gurumurthy et al., 2019; Hörl et al., 2021). Current simulation models show that schemes that promote higher occupancy of vehicles, such as shared rides or public transport, present the largest travel time reductions (Salazar et al., 2019). In terms of the value of travel time savings (VTTS), it has been anticipated that the disutility of travel time will be reduced for former car drivers that, due to vehicle automation, would be relieved from driving tasks and therefore could experience a less distressed trip. Former drivers could also make a more productive use of their time while travelling, at least for commuting. This intuition has been assumed or estimated via stated-preference surveys in several studies (Childress et al., 2015; Kockelman et al., 2017; Kolarova et al., 2019; Van den Berg & Verhoef, 2016); results show reductions up to 50% of VTTS for car drivers. However, it has been argued that, in actual conditions, AVs might not substantially affect VTTS (Cyganski et al., 2015; Rashidi et al., 2020), or could even increase it, relative to VTTS of driving a conventional car, for instance, if drivers experience discomfort due to not feeling in physical control of the vehicle (Singleton, 2019). Another aspect, in the case of shared rides, is that some people do not feel comfortable sharing the same vehicle with strangers, with no driver to overlook people’s behaviour. Therefore, the impacts of automation in VTTS might be more modest than anticipated, especially when rides are shared (Singleton, 2019). Finally, unlike cars, it is clearer that automation in buses will not have a great impact on VTTS since the change in possible activities to perform while travelling would be little or non-existent. Even, negative attitudes towards automation could increase the subjective VTTS mainly due to a decreased perception of safety and/or security (Guo et al., 2020). Finally, a positive effect of AVs in potentially decreasing the valuation of time is through changes in travel time reliability or predictability. Travellers are willing to pay to reduce the variability of travel time, i.e. there is a value of reliability, which has been estimated in the literature (Börjesson et al., 2012; Li et al., 2010). Therefore, if AVs provide more reliable travel times, the modal utility is increased. In the absence of human intervention, driving times should be more stable, as, for instance, V2V and V2I communication technologies will inform incidents more quickly, proposing alternative routes that would be optimised in real time to prevent bottlenecks. In the case of automated public transport, this effect does not only apply to in-vehicle time but also to waiting time, due to the possibility of applying new strategies to stick to schedules (Cao et al., 2019) and to keep regular headways. 13.2.3 Effect on Parking Availability and Price1 As a result of not needing a driver to operate, an expected consequence of vehicle automation is the reduction of dedicated space for parking, which would be achieved in two ways. First, the promise of a larger adoption of SAVs in replacement of personally-owned cars reduces the total fleet in the system. Narayanan et al. (2020) reviewed the literature finding replacement rates from 1.17 to 11 but emphasising that in real conditions the actual rate would probably be in the lower bracket. Second, AVs would be able to park in compact and conveniently located facilities (Nourinejad et al., 2018), which would be a major shift from HDVs, given

Connected and automated vehicles  255

that vehicles cruising for parking represent a large rate of traffic in several cities (Shoup, 2006). Certainly, parking pricing would also be key for the final output of parking availability in an era of AVs. Analysis of parking demand in an AV-based transport system shows in general a significant reduction in parking requirements, especially when shared AVs are predominant. Zhang et al. (2015) conducted an agent-based simulation to determine the parking requirement variation due to the introduction of shared AVs, obtaining reductions of up to 90%. Nourinejad et al. (2018) propose a new design for parking facilities for AVs, allowing multiple rows of vehicles stacked behind each other, and then develop a mixed-integer non-linear optimisation model to find the optimal car-park layout with minimum relocations, decreasing parking space in 62% to 87% compared to HDV classic layouts.

13.3 EFFECTS OF VEHICLE AUTOMATION ON THE PRICING OF TRANSPORT MODES The central idea behind congestion pricing (CP) is that, in congested settings, individual travellers impose delays on others, therefore they cause a negative externality that can be internalised by an economic penalty (Pigou, 2013; Vickrey, 1969).2 Though marginal cost pricing (MCP) is recognised as the first-best benchmark solution to address urban transport externalities (not only congestion, but also pollution, traffic crashes and noise, among others), it is equally agreed that the practical applicability of this principle is hardly reachable in real conditions, and second-best pricing solutions must be developed to generate feasible tolling mechanisms, such as facility-based tolls, cordon-based tolls, area-based tolls, and distance-based tolls (Verhoef, 2002). But the introduction of AVs could make dynamic pricing strategies feasible in both time and space since CAVs communication channels do not require additional infrastructure (Simoni et al., 2019), and its higher computation capabilities would help to keep pricing schemes understandable and transparent, helping to increase public acceptability of CP (Gu et al., 2018). This would allow to implement pricing mechanisms that are comparable or close to first-best pricing. In what follows, we analyse how vehicle automation could influence the pricing of alternative travel modes. In the Appendix, a summary of selected studies about AV pricing strategies is presented. 13.3.1 Private Vehicles If a reduced VTTS, an increase in capacity, and the ability to self-park due to automation demonstrate to be true, owning and using a private AV will be more attractive. It would offer the possibility to use a car to people unable to drive nowadays, such as people without a driving licence and people with reduced mobility or cognitive limitations. Also, the owners of AVs would be encouraged to use it more intensively, since some advantages of other modes, such as lower cost, more convenient use of time while travelling, and avoidance of parking costs and hassles at their destinations, among others, would melt away with automation (Lutin, 2018), leading to more and longer trips by private AVs. Then, an expected consequence is a worsening in traffic conditions and energy consumption. At this point is where pricing can be used to modify the behaviour of travellers. Two approaches are identified: pricing can be applied either to vehicles or roads.

256  Handbook on transport pricing and financing

In the first approach, Van den Berg & Verhoef (2016) use a bottleneck model to investigate the effects of migration from HDV to AV in congestion for three market organisations (private monopoly, perfect competition, and public supply) and where the share of AVs is endogenous. The introduction of AVs affects congestion via three channels: the resulting increase in capacity due to AVs, the decrease in the VTTS for those who acquire an AV, and the implications of the resulting changes in the heterogeneity of VTTS. Two effects are identified; first, a “capacity” effect, where AVs cause (as expected) a decrease in congestion. There is also a “heterogeneity” effect, caused by the introduction of additional AVs which have lower VTTS than HDV, altering the departure time behaviour of the former and therefore increasing congestion. The authors presented numerical results for the USA and the Netherlands, suggesting that a positive net externality is most likely. But, if buying an AV reduces congestion, MCP tends to lead to under-consumption of AVs, so the public supplier would need to provide a subsidy to attain the second-best optimum. In the opposite case, a corrective tax is needed to prevent over-consumption. In the second approach, Delle Site (2021) analyses a link-based pricing policy only for CAVs in a mixed-traffic network with HDVs. In a conservative assumption, no increase in capacity due to the introduction of AVs is assumed. HDVs behave according to Wardrop’s user optimum, while three behavioural scenarios are considered for CAVs: 1) CAVs driven by selfish users, 2) CAVs managed as a fleet by a monopolist, where the total cost (time + tolls) of CAVs is minimised, and 3) CAVs managed by a social planner that seeks to minimise the total cost of the HDV and CAV fleet. Three pricing schemes are considered: a classical MCP, a minimum expenditure pricing, and a zero net expenditure pricing (i.e. some arcs are charged while others are subsidised). Applying these schemes in the Anaheim network, tolls collected in the minimum expenditure and the toll-and-subsidy schemes are about ten times lower than the optimal tolls from the MCP scheme. Therefore, such pricing alternatives would facilitate the acceptability of pricing schemes among the population without compromising the benefits of the road charging schemes in terms of travel times reductions. 13.3.2 Automated Mobility on-Demand (AMoD) Vehicle automation promises to significantly reduce transport operator costs due to a reduction of driving costs, therefore, the cost advantage of placing many travellers in large vehicles such as buses will be reduced. Empirical estimations of the effects of automation on reducing the costs of motorised shared mobility indicate that the effect is potentially large. In Figure 13.1, the ratio between driver cost and total operator cost is shown for different vehicle sizes, from cars (that could be used for SAV services) to buses of different lengths (that are used for public transport), considering Munich data (Tirachini & Antoniou, 2020). Depending on the asset life assumed, the car-size vehicles present driver costs between 73% and 82% of the total operator cost, while in the case of buses driver costs are in the range of 30–56% of the total operator cost. This is a quantification of an expected effect: the smaller the vehicles, the larger the potential cost benefits due to vehicle automation, and differences between small and large vehicles are significant. Thus, shared-mobility services with smaller vehicles are expected to play a larger role in the world of AVs. The anticipated large cost savings due to automation should be, at least in part, transferred to lower prices, since a strong modal competition between several travel alternatives will cause that shared AVs have to set competitive prices to attract users (Chen & Kockelman, 2016;

Connected and automated vehicles  257

Source:   Reprinted from Tirachini and Antoniou (2020), with permission from Elsevier.

Figure 13.1  Driver cost as a proportion of the total cost, Munich values Gurumurthy et al., 2019; Hörl et al., 2021). This trend is reinforced by an attitudinal change towards car ownership and use among young people (Zhou & Wang, 2019), so users who cannot or do not want to purchase a car, or neither want to travel by car regularly, are able to access to one in case they need it. Narayanan et al. (2020) point out that SAV systems can have different booking time frames; namely, on-demand (vehicle booking in real time) and reservation-based (booked in advance) systems. To avoid confusion, we will use the term AMoD to refer to the first type of operation. As well as with private AVs, an important concern about the massification of AMoD is the worsening of traffic conditions. A lower price of AMoD (relative to a system with humandriven vehicles) coupled with improved access time and/or comfort relative to public transport (PT) and active modes, suggest that AMoD will increase its modal share at the expense of other modes. Dynamic traffic assignment models show that an increase in vehicle kilometres travelled (VKT) and travel time is seen when AMoD is introduced (Hörl et  al., 2021; Kaddoura et  al., 2020; Simoni et  al., 2019), not only because of induced demand but also because of empty VKT (due to picking up passengers, fleet rebalancing, and charging, among other factors). So, to take advantage of AMoD possibilities while constraining the negative externalities of increased motorised traffic, centralised policies such as optimal road pricing will be crucial. The effects of pricing on the performance of SAVs have been analysed with simulation studies. Simoni et al. (2019) study behavioural responses to different congestion pricing schemes and their effects on congestion, considering a capacity increase due to the introduction of AVs in Austin, Texas. “AV-oriented” and “AMoD-oriented” scenarios are designed, and as predicted, a rise in congestion is observed by demand shifts from PT and active modes, in addition to longer trips. VKT increases by 16% and 22%, while total travel delay increases

258  Handbook on transport pricing and financing

between 61% and 87% with respect to the HDV scenario (base). Then, two ”traditional” congestion pricing schemes (distance-based and link-based) and two “advanced” schemes (MCPbased and travel time-congestion-based pricing) are evaluated. The difference between the two types of schemes is that the operationalisation of the former is fairly plausible with the existing technology, while the latter is more complex and requires new technologies (such as those of CAVs) for optimal implementation. All the CP strategies are effective in the relief of congestion by significantly reducing VKT and delays with respect to the unpriced scenarios, but advanced schemes outperform traditional ones in welfare gains. Congestion pricing schemes as a way to counteract the increase in traffic from AMoD are also studied by Kaddoura et al. (2020), who find that it is necessary to price both HDV and AMoD while the presence of AVs is not large in the market. 13.3.3 Public Transport For the case of PT, automation is expected to increase coverage and service frequency, as well as to significantly reduce fares, due to the reduction of operator costs if (at least a part of) vehicles are driverless (Tirachini & Antoniou, 2020; Zhang et al., 2019). The cost advantage of automation in the case of PT in large vehicles is less pronounced than for AMoD, because, as shown in Figure 13.1, the driver cost is a smaller share of the total operator cost for larger vehicles. A potential large reduction in driving costs has been shown to reduce the degree of economies of scale in public transport (Tirachini & Antoniou, 2020). In PT, the existence of crowding externalities is known to increase optimal service frequency and vehicle capacity (Jara-Díaz & Gschwender, 2003) and to increase the optimal PT fare (Tirachini et al., 2014), therefore undermining some of the advantages of automated PT in reducing vehicle size and optimal fare as identified by Tirachini and Antoniou (2020). If there is a reduction in user cost for both private and public transport by using AVs, then this should be reflected in the optimal price of both modes, as we will formally assess in Section 13.4. Previous findings, i.e. potential reductions in the optimal fare and increase in service frequency, have been found analysing PT as a single mode. In practice, if a significant cost reduction of private AV and/or AMoD materialises, the shape of PT will inevitably change, as in all the simulations where AV-related modes and PT coexist the result is the undermining of the latter (Chen & Kockelman, 2016; Gurumurthy et al., 2019; Hörl et al., 2021; Kaddoura et al., 2020; Simoni et al., 2019). The sharpest impacts in terms of travel time and accessibility are seen in low-demand areas and/or periods. Since equity issues and service standards force PT authorities to maintain the coverage even with less ridership, the natural consequence is a change in the network structure, pushing for the adoption of on-demand services in lowdemand markets. Despite these concerns, it is clear that mass public transport will not disappear with automation, as it is not replaceable in high-density settings. Bösch et al. (2018) analyse the cost structure of automated buses, private AVs, and AMoD (with either shared rides or not), concluding that buses will remain as the most effective transport mode in dense areas and corridors. Also, among some user groups there is still a preference for fixed-route lines over dynamic services since it is perceived as more readily available by users reluctant to use technology (White, 2016). Finally, another interesting consequence of the savings in drivers’

Connected and automated vehicles  259

costs, as well as the improved platooning and precision docking of automated buses, is that a light rail transit (LRT) standard service could be offered by automated buses in a similar road width and with the same capacity at significantly lower cost (Lutin, 2018). Hence, a system with the level of service of LRT but the flexibility of a bus could be offered, diminishing unnecessary interchanges and travel time. 13.3.4 Active Modes If PT ridership is affected in longer trips due to the lower cost of AVs, in short trips people could be tempted to walk and cycle less due to the improved accessibility that AMoD will bring at affordable fares. In addition to increased levels of congestion, energy consumption, and pollution (depending on the source of energy used to manufacture and run vehicles), this behavioural change could also represent a public health issue since for a significant fraction of the population most of their physical activity is performed while commuting (Litman, 2017; Sallis et al., 2004), either on door-to-door trips or walking/cycling as to access/egress public transport or other motorised modes. In this context, properly pricing automated motorised transport would have a double effect: on the one hand, it would reduce the attractiveness of motorised modes relative to active alternatives, and on the other hand, if at least a part of the revenues is reinvested in improving conditions for walking and cycling, demand for active modes will be induced. 13.3.5 Intermodal Transport and Mobility-as-a-Service (MaaS) Current estimations of the effects of AVs on quality of service and modal attributes have been made for individual modes. However, new possibilities of integration can be designed with automation by taking advantage of the best characteristics of each mode. The most effective integration envisioned is the AMoD-PT by using the latter in replacement of low-frequency feeder bus routes. Shen et al. (2018) assess the performance of this type of integration in the Tampines area, Singapore, replacing some feeder bus routes with AMoD (with either shared rides or not) to perform the first/last mile. After evaluating different ride-sharing preferences and vehicle sizes, the authors identify scenarios where that reduce waiting times and optimise the utilisation of the bus fleet, while financial sustainability for the AMoD operator is reached. Salazar et al. (2019) analyse the impact of an integrated AMoD-PT scheme in a road-pricing formulation that seeks to maximise social welfare, concluding that such operation implies a reduction in congestion, operator costs, and emissions. From the latter results, if AMoD is thought of as a component of the PT system, then it could represent a path of improved quality of service in terms of waiting times, coverage, and hours of operation without increasing fares nor adding negative externalities. Moreover, integration could go beyond operation to include payment, booking, and trip planning processes, among others, as well as to incorporate other modes such as bike-sharing and car rental, for instance. This bundling of transport services has been called MaaS (Kamargianni et al., 2016), and it could take advantage of vehicle automation and its subsequent drop in operating costs to become more widespread. It remains to be seen if a hypothetical deployment of AVs could help in one of the main struggles of MaaS today: the scalability of such services.

260  Handbook on transport pricing and financing

13.4 A THREE-MODE FIRST-BEST PRICING MODEL 13.4.1 Model Presentation In this section, we synthesise the previous discussion with the analysis of a transport pricing model. We follow Tirachini and Hensher (2012) to develop a three-mode first-best pricing model (see also Chapters 2, 8, and 9 in this Handbook). Consider a single origin-destination pair and three modes: automobile (a), public transport (b), which could be a bus- or rail-based mode, and an active transport mode I, which could be walking or cycling. The attractiveness of walking and cycling is mainly associated with trip distance and factors such as steepness and availability of safe and attractive walking and cycling facilities. Road capacity is assumed fixed, income effects and tax distortions are ignored. The joint demand for the three modes can be obtained from the benefit function B ( qa , qb , qe ), which expresses the consumers’ willingness to pay for a particular combination {qa , qb , qe } of travel by automobile, public transport, and the active mode. The inverse demand function di for mode i is given by:

di ( qa , qb , qe ) =

¶B ( qa , qb , qe ) ¶qi

i Î {a, b, e} (13.1)

Let Ci and ci be the total and average cost functions of mode i, respectively (including both time and operation costs), that is:

Ci = qi ci (13.2)

Let ca ( qa , qb ) and cb ( qa , qb ) be the average cost of car and public transport, respectively. We assume that the cost functions depend on demand qa and demand qb. The relationship between car demand qa and car flow fa is fa = ua qa , where υa is the inverse of the average occupancy rate per car. The relationship between public transport demand qb and frequency f b depends on the frequency rule used in the public transport system. Cost cb includes users’ cost cu (access, waiting and in-vehicle time costs) and operator cost co (which accounts for capital and operating costs):

cb = cu + co (13.3)

We further assume that the travel time associated with the active mode is fixed and independent of the demand or flow of any mode, i.e. the active mode is uncongestible. In equilibrium, the marginal benefit is equal to the generalised price, ca + ta and cu + tb for cars and public transport, respectively (Equation 13.4), where τa is the road use charge for the automobile and τb is the fare for public transport.

¶B = ca + ta ¶qa

¶B = cu + tb (13.4) ¶qi

A social welfare function reflects the level of welfare in a society expressed as a function of economic variables. In transport economics, optimal welfare-oriented pricing decisions ensure that the external costs and benefits of travelling are internalised in the user’s decisions.

Connected and automated vehicles  261

The social welfare (SW) function (13.5) comprises the difference between the benefit function and the total cost associated with travelling by automobile, public transport and active mode.

SW = B ( qa , qb , qe ) - qa ca ( qa , qb ) - qbcb ( qa , qb ) - qece (13.5)

Expression (13.5) is to be maximised. After applying first-order conditions, we find: ¶ca ¶c + qb b (13.6a) ¶qa ¶qa



t a = qa



tb = co + qa



¶ca ¶c + qb b (13.6b) ¶qb ¶qb

te = 0 (13.6c)

Solution (13.6a) is the well-known Pigouvian tax for cars, including in this case the marginal cost of public transport cost due to car demand (second term). Equation (13.6b) is the first-best fare for public transport with congestion interactions (see also Chapter 9 in this Handbook). ¶c ¶c If there is no congestion interaction between cars and public transport, then b = a = 0 in ¶qa ¶qb Equations (13.6a) and (13.6b). Equation (13.6c) states that the price for walking or cycling is zero (the assumed uncongestible mode). Note that the model can be easily generalised to having four modes, including both walking and cycling as separate alternatives, in which case the solution for both would be optimal prices equal to zero, under the no-congestion assumption. For simplicity, we have not included the case of a binding public transport capacity constraint in this model (see Tirachini & Hensher, 2012). A modified version of Equation (13.6a) is presented in Tscharaktschiew and Evangelinos (2019), in a model in which travellers can choose between human-driven or automated vehicles for individual use. In this case, a new term shows up in the optimal toll function, which accounts for the feedback effect of the choice of driving mode on the road capacity. 13.4.2 Changes to the Optimal Fare of Private Cars Due to Automation Next, let us analyse how the first-best pricing rules change after introducing AVs. For simplicity, we are going to assume the case of trains or buses running on segregated railways or busways. We will therefore disregard congestion interactions between cars and public transport, ¶c ¶c which is equivalent to assuming b = a = 0. The average cost for cars can be expressed as: ¶qa ¶qb

ca ( qa ) = Pva ( qa ) t a ( qa ) (13.7)

Where Pva is the value of travel time savings for car users. In Equation (13.7), we have assumed that Pva depends on the actual car demand level, following the empirical evidence that congestion increases the value of travel time savings (for a review, see Wardman & Ibáñez, 2012). Part of this increase in the value of time savings under congested conditions might be explained by a greater unreliability and unpredictability of travel time estimations. Taking the

262  Handbook on transport pricing and financing

derivative of Equation (13.7) with respect to qa, we can re-write the optimal fare in Equation (13.6a) as follows:



II IV é ( I )  ( )  () ù III )  ê  ú    ¶t q ( ¶ P q a( a) va ( a ) ú (13.8) + t a ( qa ) ta = qa ê Pva ( qa ) ê ¶ ¶ qa ú qa ê ú ³0 ³0 ë û

With regard to the value of travel time savings, as analysed in Section 13.2.4, it can either increase or decrease due to automation, although the most likely effect for the case of car drivers switching to their own self-driving car is a reduction in Pva because the driver is released from driving tasks. If there is no congestion, terms (II) and (IV) are zero, and the optimal car toll is zero, as expected. An extended model that includes externalities other than congestion (such as pollution and crashes) would yield a positive fare even in the case of no congestion. If there is congestion, then term (II) is positive and term (IV) might be positive (which is an empirical matter). Assuming a positive value for term (IV), i.e. that car users value travel time ¶Pva ( qa ) reductions more under congested conditions, then AVs may reduce the derivative , if ¶qa more certain travel times are possible with AVs. For instance, the adoption of centralised route assignment strategies with AVs points to providing more certainty in travel times, therefore, we expect, reducing the effects of a demand-induced uncertainty on the value of travel time savings. With regard to term (II), automation, at least with a large penetration rate of AVs, is expected to reduce congestion, and therefore to reduce the marginal time (II). Finally, travel time (III) may be reduced if lower congestion levels are possible with connected and AVs; however, such a scenario will take a long time to materialise, as current pilots with automated minibuses around the world show the opposite: automated minibuses are slower than their human-driven counterparts due to safety considerations in urban environments (interactions with human-driven vehicles, parked cars, pedestrians, and cyclists) and the novelty of the technology. It follows that fully segregated AVs (not running in mixed traffic with human-driven vehicles, pedestrians, and cyclists) are more likely to reach travel time savings. All-in-all, the effect of vehicle automation on the optimal (first-best) car toll cannot be unambiguously determined. On the one hand, we have identified a number of elements that push to reduce the optimal toll in Equation (13.8), which include (i) a potential congestion relief from automation, (ii) a better use of time while travelling that pushes for a reduction in the value of travel time savings, and (iii) the provision of more certain travel times with AVs. If other externalities, such as traffic crashes, are included in the analysis, then the gain in traffic safety from automation increases the difference between optimal fares with AVs versus HDVs. On the other hand, an increase in the optimal toll with AVs is possible if travel time tends to increase instead, and if the VTTS also increases. Therefore, there are strong reasons to suggest a reduction in the optimal toll for private-use AVs, relative to conventional HDVs, at least under full segregation. With mixed traffic, it is not so clear if the optimal toll should be reduced. 13.4.3 Changes to the Optimal Fare of Public Transport Due to Automation Recalling the assumption that cars and public transport do not share the right-of-way in our formulation, optimal public transport fare in Equation (13.6b) can be expressed as:

Connected and automated vehicles  263

( )   ¶ éë Pact ac ( qb ) + Pw ( qb ) t w ( qb ) + Pvb ( qb ) tb ( qb ) ùû II



I) (

tb = co + qb

¶qb

( III )  ¶c + qb o . (13.9) ¶qb

where Pac, Pw, and Pvb are the values of access, waiting and in-vehicle time savings, respectively, and tac, tw and tvb are the access, waiting and in-vehicle times, respectively. If AVs are driverless, a large unit operator cost co is expected, and the effect is larger for smaller vehicles (Section 13.3). This reduction in operator cost pushes to decrease the first-best public transport fare, as analytically shown by Tirachini & Antoniou (2020). Next, we study the marginal effect of demand on user costs (factor II in Equation 13.9). If both the value of waiting and in-vehicle time savings are sensitive to large passenger volumes due to crowding effects on bus stops, train stations, and inside vehicles (Tirachini et al., 2013), then an operation with more reliable travel times and headways due to automation should balance station and vehicle loads more smoothly, reducing the level of crowding in some vehicles and in some headways. Therefore, an improvement in the quality of service as perceived by the users is possible, without even increasing the fleet size, and these more comfortable travel conditions result in reductions of the value of travel time savings, therefore pushing to reduce (II). Waiting time is also reduced if driverless operation is possible, and attached to this there is an increase in service frequency. In general, a waiting time reduction due to increases in demand (because of increases in service headway), translates into making term (III) negative, which is more so if access time and in-vehicle times are reduced as well as a function of demand, or at least stay equal. Changes in in-vehicle time tb depend on how fast the operation of AVs is relative to the operation of HDVs, and therefore a change in this variable, if any, is less predictable. The same can be said about changes, if any, in access time in term (II). Finally, term (III) is likely negative in Equation (13.9), due to the existence of economies of scale in public transport operations (Allport, 1981). The operation of driverless vehicles reduces the level of scale economies in public transport (Tirachini & Antoniou, 2020), therefore making term (III) less negative. All-in-all, this appears to be the only study published about firstbest public transport pricing showing that the decrease in operator cost (I) dominates the other effects and the optimal fare is lower, relative to HDVs (Tirachini & Antoniou, 2020).

13.5 CONCLUDING REMARKS In conclusion, we have shown that cost reductions due to automation tend to push towards lower optimal fares for both private cars and public transport. Therefore, it follows that vehicle automation makes motorised transport more attractive relative to active modes. This may have serious implications for the future pricing of AVs, if not properly addressed with pricing frameworks that include, e.g. the health and environmental benefits from active mobility, which provide the basis for the promotion of walking and cycling in daily life in cities, through a range of policies that include price incentives. A comprehensive policy package aimed at making active modes more attractive might be even more relevant in a scenario of AVs for both private and public transport. These policies should be even more aggressive if reductions in VTTS as well as increases in road capacity result to be significant (which leads to a mostly individual use of AVs). As Soteropoulos et al. (2019) conclude from a review of related

264  Handbook on transport pricing and financing

studies, a large adoption of individual-use AVs leads to a more dispersed urban development and sprawling with all the negative externalities it entails. Conversely, scenarios where shared use of AVs (either AMoD or PT) is predominating implies a long-term urbanisation process, fuelling a virtuous cycle of efficiency in the use of resources in the cities. Pricing will be key to shaping one scenario or the other. It should be noted that most of our analysis is based on the effects of automation on optimal prices. How this should be translated into actual observed prices, which are suboptimal in most cases, is a matter of great uncertainty and a venue for further research. A numerical application of the model developed in Section 13.4, comparing scenarios with and without automation, is also an interesting topic for future research. The model in Section 13.4 was based on congestion externalities only; the formal introduction of environmental and climate change-related externalities is expected to gain increased attention in the coming years and decades, if the promises of automated vehicles materialise.

ACKNOWLEDGEMENTS Support from ANID Chile (Grant PIA/PUENTE AFB220003) is acknowledged. We are grateful for the comments made by Editor Daniel Hörcher and one anonymous reviewer. Any errors are the authors’ responsibility alone.

NOTES 1. The general topic of parking economics is analysed in Chapter 5 (Urban form) and Chapter 7 (Political Economy) in this Handbook. 2. For details on congestion pricing, see Chapter 2 (Theory of externality pricing) and Chapter 8 (Road pricing) in this Handbook.

REFERENCES Allport, R. J. (1981). The costing of bus, light rail transit and metro public transport systems. Traffic Engineering & Control, 22(HS-032 837). Börjesson, M., Eliasson, J., & Franklin, J. P. (2012). Valuations of travel time variability in scheduling versus mean–variance models. Transportation Research Part B: Methodological, 46(7), 855–873. Bösch, P. M., Becker, F., Becker, H., & Axhausen, K. W. (2018). Cost-based analysis of autonomous mobility services. Transport Policy, 64, 76–91. Cao, Z., Ceder, A. A., & Zhang, S. (2019). Real-time schedule adjustments for autonomous public transport vehicles. Transportation Research Part C: Emerging Technologies, 109, 60–78. Chen, T. D., & Kockelman, K. M. (2016). Management of a shared autonomous electric vehicle fleet: Implications of pricing schemes. Transportation Research Record, 2572(1), 37–46. Childress, S., Nichols, B., Charlton, B., & Coe, S. (2015). Using an activity-based model to explore the potential impacts of automated vehicles. Transportation Research Record, 2493(1), 99–106. Cyganski, R., Fraedrich, E., & Lenz, B. (2015). Travel-time valuation for automated driving: A usecase-driven study. Proceedings of the 94th Annual Meeting of the TRB. Delle Site, P. (2021). Pricing of connected and autonomous vehicles in mixed-traffic networks. Transportation Research Record, 0361198120985850. Dresner, K., & Stone, P. (2008). A multiagent approach to autonomous intersection management. Journal of Artificial Intelligence Research, 31, 591–656.

Connected and automated vehicles  265

Fagnant, D. J., & Kockelman, K. (2015). Preparing a nation for autonomous vehicles: Opportunities, barriers and policy recommendations. Transportation Research Part A: Policy and Practice, 77, 167–181. Gu, Z., Liu, Z., Cheng, Q., & Saberi, M. (2018). Congestion pricing practices and public acceptance: A review of evidence. Case Studies on Transport Policy, 6(1), 94–101. Guo, J., Susilo, Y., Antoniou, C., & Pernestål Brenden, A. (2020). Influence of individual perceptions on the decision to adopt automated bus services. Sustainability, 12(16), 6484. Gurumurthy, K. M., Kockelman, K. M., & Simoni, M. D. (2019). Benefits and costs of ride-sharing in shared automated vehicles across Austin, Texas: Opportunities for congestion pricing. Transportation Research Record, 2673(6), 548–556. Hoogendoorn, R., van Arerm, B., & Hoogendoom, S. (2014). Automated driving, traffic flow efficiency, and human factors: Literature review. Transportation Research Record, 2422(1), 113–120. Hörl, S., Becker, F., & Axhausen, K. W. (2021). Simulation of price, customer behaviour and system impact for a cost-covering automated taxi system in Zurich. Transportation Research Part C: Emerging Technologies, 123, 102974. Jara-Díaz, S., & Gschwender, A. (2003). Towards a general microeconomic model for the operation of public transport. Transport Reviews, 23(4), 453–469. Kaddoura, I., Bischoff, J., & Nagel, K. (2020). Towards welfare optimal operation of innovative mobility concepts: External cost pricing in a world of shared autonomous vehicles. Transportation Research Part A: Policy and Practice, 136, 48–63. Kamargianni, M., Li, W., Matyas, M., & Schäfer, A. (2016). A critical review of new mobility services for urban transport. Transportation Research Procedia, 14, 3294–3303. Keeney, T. (2017). The future of transport is autonomous mobility-as-a-service. ARK Invest. Kockelman, K., Boyles, S., Stone, P., Fagnant, D., Patel, R., Levin, M. W., Sharon, G., Simoni, M., Albert, M., & Fritz, H. (2017). An assessment of autonomous vehicles: Traffic impacts and infrastructure needs. University of Texas at Austin. Center for Transportation Research. Kolarova, V., Steck, F., & Bahamonde-Birke, F. J. (2019). Assessing the effect of autonomous driving on value of travel time savings: A comparison between current and future preferences. Transportation Research Part A: Policy and Practice, 129, 155–169. doi: 10.1016/j.tra.2019.08.011 Li, Z., Hensher, D. A., & Rose, J. M. (2010). Willingness to pay for travel time reliability in passenger transport: A review and some new empirical evidence. Transportation Research Part E: Logistics and Transportation Review, 46(3), 384–403. Litman, T. (2017). Reforming Municipal Parking Policies to Align With Strategic Community Goals. Victoria Transport Policy Institute. Lutin, J. M. (2018). Not if, but when: Autonomous driving and the future of transit. Journal of Public Transportation, 21(1), 10. Mena-Oreja, J., Gozalvez, J., & Sepulcre, M. (2018). Effect of the configuration of platooning maneuvers on the traffic flow under mixed traffic scenarios. 2018 IEEE Vehicular Networking Conference (VNC), 1–4. Narayanan, S., Chaniotakis, E., & Antoniou, C. (2020). Shared autonomous vehicle services: A comprehensive review. Transportation Research Part C: Emerging Technologies, 111, 255–293. Nourinejad, M., Bahrami, S., & Roorda, M. J. (2018). Designing parking facilities for autonomous vehicles. Transportation Research Part B: Methodological, 109, 110–127. Pigou, A. C. (2013). The Economics of Welfare. Palgrave Macmillan. Rashidi, T. H., Waller, T., & Axhausen, K. (2020). Reduced value of time for autonomous vehicle users: Myth or reality? Transport Policy, 95, 30–36. doi: 10.1016/j.tranpol.2020.06.003 SAE International. (2021). J3016C: Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles - SAE International. https://www​.sae​.org​/standards​/content​/ j3016​_202104 Salazar, M., Lanzetti, N., Rossi, F., Schiffer, M., & Pavone, M. (2019). Intermodal autonomous mobilityon-demand. IEEE Transactions on Intelligent Transportation Systems, 21(9), 3946–3960. Sallis, J. F., Frank, L. D., Saelens, B. E., & Kraft, M. K. (2004). Active transportation and physical activity: Opportunities for collaboration on transportation and public health research. Transportation Research Part A: Policy and Practice, 38(4), 249–268. Shen, Y., Zhang, H., & Zhao, J. (2018). Integrating shared autonomous vehicle in public transportation system: A supply-side simulation of the first-mile service in Singapore. Transportation Research Part A: Policy and Practice, 113, 125–136. doi: 10.1016/j.tra.2018.04.004

266  Handbook on transport pricing and financing

Shoup, D. C. (2006). Cruising for parking. Transport Policy, 13(6), 479–486. Simoni, M. D., Kockelman, K. M., Gurumurthy, K. M., & Bischoff, J. (2019). Congestion pricing in a world of self-driving vehicles: An analysis of different strategies in alternative future scenarios. Transportation Research Part C: Emerging Technologies, 98, 167–185. Singleton, P. A. (2019). Discussing the “positive utilities” of autonomous vehicles: Will travellers really use their time productively? Transport Reviews, 39(1), 50–65. Soteropoulos, A., Berger, M., & Ciari, F. (2019). Impacts of automated vehicles on travel behaviour and land use: An international review of modelling studies. Transport Reviews, 39(1), 29–49. Stocker, A., & Shaheen, S. (2017). Shared automated vehicles: Review of business models (Working Paper No. 2017–09). International Transport Forum Discussion Paper. https://www​.econstor​.eu​/ handle​/10419​/194044 Tientrakool, P., Ho, Y.-C., & Maxemchuk, N. F. (2011). Highway capacity benefits from using vehicleto-vehicle communication and sensors for collision avoidance. 2011 IEEE Vehicular Technology Conference (VTC Fall), 1–5. Tirachini, A., & Antoniou, C. (2020). The economics of automated public transport: Effects on operator cost, travel time, fare and subsidy. Economics of Transportation, 21, 100151. Tirachini, A., & Hensher, D. A. (2012). Multimodal Transport Pricing: First Best, Second Best and Extensions to Non-motorized Transport. Transport Reviews, 32(2), 181–202. doi: 10.1080/01441647.2011.635318 Tirachini, A., Hensher, D. A., & Rose, J. M. (2013). Crowding in public transport systems: Effects on users, operation and implications for the estimation of demand. Transportation Research Part A: Policy and Practice, 53, 36–52. doi: 10.1016/j.tra.2013.06.005 Tirachini, A., Hensher, D. A., & Rose, J. M. (2014). Multimodal pricing and optimal design of urban public transport: The interplay between traffic congestion and bus crowding. Transportation Research Part B: Methodological, 61, 33–54. doi: 10.1016/j.trb.2014.01.003 Tscharaktschiew, S. and C. Evangelinos (2019). Pigouvian road congestion pricing under autonomous driving mode choice. Transportation Research Part C: Emerging Technologies, 101, 79–95. Van Arem, B., Van Driel, C. J., & Visser, R. (2006). The impact of cooperative adaptive cruise control on traffic-flow characteristics. IEEE Transactions on Intelligent Transportation Systems, 7(4), 429–436. Van den Berg, V. A., & Verhoef, E. T. (2016). Autonomous cars and dynamic bottleneck congestion: The effects on capacity, value of time and preference heterogeneity. Transportation Research Part B: Methodological, 94, 43–60. Verhoef, E. T. (2002). Second-best congestion pricing in general static transportation networks with elastic demands. Regional Science and Urban Economics, 32(3), 281–310. Vickrey, W. S. (1969). Congestion theory and transport investment. The American Economic Review, 59(2), 251–260. Wadud, Z., MacKenzie, D., & Leiby, P. (2016). Help or hindrance? The travel, energy and carbon impacts of highly automated vehicles. Transportation Research Part A: Policy and Practice, 86, 1–18. doi: 10.1016/j.tra.2015.12.001 Wardman, M., & Nicolás Ibáñez, J. (2012). The congestion multiplier: Variations in motorists’ valuations of travel time with traffic conditions. Transportation Research Part A: Policy and Practice, 46(1), 213–225. doi: 10.1016/j.tra.2011.06.011 White, P. (2016). The roles of ‘conventional’ and demand-responsive bus services. In Paratransit: Shaping the Flexible Transport Future. Emerald Group Publishing Limited. Zhang, W., Guhathakurta, S., Fang, J., & Zhang, G. (2015). Exploring the impact of shared autonomous vehicles on urban parking demand: An agent-based simulation approach. Sustainable Cities and Society, 19, 34–45. Zhang, W., Jenelius, E., & Badia, H. (2019). Efficiency of semi-autonomous and fully autonomous bus services in trunk-and-branches networks. Journal of Advanced Transportation, 2019. Zhou, M., & Wang, D. (2019). Generational differences in attitudes towards car, car ownership and car use in Beijing. Transportation Research Part D: Transport and Environment, 72, 261–278.

267

Objective

Gurumurthy et al. (2019)

Determine how fleet size, pricing and fare level change under AMoD with shared rides, from private and societal goals, in the presence of both HDV and AVs

HDV, PT, Active modes, Private AV, AMoD + shared rides AMoD is priced according to Simoni et al. (2019). All major network links are priced in the morning and afternoon peak at $0.05/min

Arc-based tolling applied only to CAVs

Available modes Pricing/tolling scheme

Delle Site (2021) Analyse arc-based HDV, CAV pricing policies for a CAVs fleet in a mixed-traffic network with HDVs

Author(s)

Table 13.1  Selected studies regarding AV pricing

APPENDIX​

Two theoretical networks (twoarc network representative of town bypass and Nguyen-Dupuis network), and Anaheim network

NA/Mode share, Austin, Texas % increase in VKT, Empty VKT, idle hours of AMoDs, average occupation, $/ AMoD/day, Trips/ AMoD/day

Wardrop’s equilibrium in scenario 1, Min total travel time of CAV in scenario 2, Min total travel time of HDV and CAV in scenario 3/total toll paid

Objective Function/ Study area monitoring variables

0. B  ase (without AVs). 1. With AMoD + DRS. 2. With DRS fare 50% discount. 3. With DRS fare 75% discount. * Scenarios 1, 2 and 3 with AMoD availability from 10 to 100.

0. Base (without CAVs) 1. CAVs managed as individual vehicles 2. CAVs are managed as a fleet by a service provider 3. CAVs are managed as a fleet by a social planner. * In all cases, the proportion of CAVs is 50%. * In all cases, tolls are computed according to two schemes: one with positive tolls and minimum toll expenditure, and one with both tolls and subsidies and zero net expenditure.

Scenarios

(Continued)

Modal share of AMoDs is relevant only when fares are moderate to low. Operational balance and system benefits are reached with moderate fleet sizes (one AMoD vehicle for every 25 persons) and competitive prices, especially when road-pricing schemes are applied, in terms of empty VKT and revenue to the operator.

Congestion prices in the min expenditure and in the toll-and-subsidy scheme are significantly lower than classic marginal cost prices (about 10% in the Anaheim network). If time + toll is desired to minimise, then private monopolist is always dominant regarding the selfish agent scenario, if only toll is considered then the least costly option varies depending on the case.

Main results/conclusions

268

Coordination policies for integration between AMoD and PT maximising SW under the assumption of a perfect market with selfish agents

Investigate optimal CP strategies for AMoD and HDV users, where users are able to adjust their mode, departure time and route

Kaddoura et al. (2020)

Salazar et al. (2019)

Objective

Author(s)

Table 13.1   (Continued)

Only Metro in NYC; Metro, S-Bahn and Tram in Berlin, AV(AMoD) in both

HDV, AMoD (only in inner city area), PT, Walk, Bicycle, Ride

PT fares are equal to operational cost, road tolls equal to road congestion multipliers, road prices equal to the sum of AVs operating costs, road tolls, and the origin and destination prices (dual multipliers associated with the vehicle balance constraints)

Marginal cost pricing

Available modes Pricing/tolling scheme

Max welfare, with PT fare and link tolls as decision variables

0. Base (without AMoD) 1. Base + AMoD (without CP) 2. Base + AMoD + CP 3. Case 2 + HDV CP

Scenarios

Manhattan in –E  xogenous road usage New York and from 50% to 200%. the city centre in – Two types of propulsion: Berlin Internal combustion engine (ICEV) and Battery electric (BEV) vehicles. –L  W and SW vehicles.

Max SW/Modal Great Berlin share, Travel time, Traffic volume, air pollution, noise

Objective Function/ Study area monitoring variables

Vehicle size is equally important than propulsion; SU BEV is app. five times more pollutant in CO2 emissions than LW ICEV. Integration with congestion pricing significantly reduce travel time, costs, number of vehicles and emissions.

Implementation of AMoD increases the traffic in the city centre in all scenarios. Congestion pricing only in AMoD slightly reduces travel time and traffic. Only by pricing both AMoD and HDV a significant reduction in travel times, traffic and externalities across the city is reached, with improvements in welfare.

Main results/conclusions

269

Study behavioural responses to different congestion pricing schemes and its effects on congestion in scenarios with a strong market share of AVs and AMoD

To assess the impact of automation on optimal vehicle size, service frequency, fare, subsidy and degree of economies of scale in public transport

Simoni et al. (2019)

Tirachini and Antoniou (2020)

PT

Optimal (first-best) pricing

HDV – PT – – Two “traditional” walk/bike (joint) pricing strategies: – Private AV link-based (LB) – AMoD scheme and distance-based (DB) scheme. – Two “advanced” pricing strategies: Dynamic marginal cost pricing (MCP) at the link level, and Travel time congestion-based (TTC) scheme, which charges users for the delay caused by everyone else Min total cost

NA/Modal share, VMT savings, delay savings, welfare change

Munich in Germany and Santiago in Chile

Austin metro area, Texas

–H  DV (base) – Driver cost saving of 0%, 50% and 100% due to automation regarding the HDV scenario –R  atio running time of AVs/running time of human-driven buses equal to 1 and 0,5.  emand between 100 and –D 4.000 [passengers/h].

1. AV-oriented scenario (90% of car owners of the base scenario have availability to private AV + 1 AMoD every 30 agents). 2. AMoD-oriented scenario (10% of car owners have access to AV + 60% availability of private cars + 1 AMoD every ten agents). * Both scenarios with PT network fixed. For each scenario, the four tolling schemes depicted before.

(Continued)

Automation causes smaller vehicles and more frequent services to be optimal for PT services. There is a reduction in the degree of economies of scale in PT. Optimal fares and subsidies are also reduced. Benefits are reached if a significant fraction (larger than 50%) of driving costs are saved. Benefits are larger in Germany than in Chile due to the higher share of labour costs.

DB scheme seems more effective in the AMoDoriented scenario and in the base scenario, while the LB scheme performs better in the AV-oriented scenario. MCP-based scheme and travel time congestion-based scheme perform better in the AMoD-oriented scenario than in the AV-oriented scenario. In all the scenarios, the TTC scheme presents the largest social welfare improvements.

270

Objective

Tscharaktschiew To assess the and Evangelinos impact of (2019) automation on optimal congestion pricing

Author(s)

Table 13.1   (Continued)

HDV, AV

Optimal (first-best) pricing

Available modes Pricing/tolling scheme Max SW

Scenarios

Highway section – N  o congestion toll, manual driving only. –C  ongestion toll, manual driving only. –C  ongestion toll, manual and automated driving.

Objective Function/ Study area monitoring variables

The choice between manual and automated vehicles is modelled. The marginal social trip costs are no longer strictly increasing in traffic flow. Multiple congestion pricing equilibria may lead to situations without automated vehicles, i.e. with manual vehicles only. Coexistence of manual and automated vehicles add complexity to the setting of optimal road pricing.

Main results/conclusions

271

Van den Berg and Verhoef (2016)

To investigate HDV, AV the effects of congestion of using AVs for three market organisations (private monopoly, perfect competition, and public supply) Subsidies or taxes applied to (des) incentivise buying of AVs Equilibrium or max profit/Total Travel Cost, Total Cost (TTC + car cost), relative efficiency (gain on a policy divided by the gain from case P) Numerical 0. Base (only HDV) examples 1. S  ocially optimal “public” with USA and provision Netherlands data 2. Perfect market (marginal cost provision) 3. Profit-maximising monopolist There is a “capacity” effect, where getting AVs cause a positive externality due to a decreasing of congestion. On the other hand, there is a “heterogeneity” effect, caused by the introduction of additional AVs which have lower VTTS than HDV, altering the departure time behaviour of the former and therefore increasing congestion. Buying an AV reduces congestion, and marginal cost pricing tends to lead to the under-consumption of automated cars. To prevent this and attain the secondbest optimum, the public supplier needs to provide a subsidy. However, if there is a negative externality, a corrective tax is needed to prevent over-consumption. The private monopolist is likely to lead to a large undersupply and welfare loss.

PART III TRANSPORT FUNDING AND FINANCING

14. Transport funding and financing: a conceptual overview of theory and practice José Manuel Vassallo and Laura Garrido

14.1 INTRODUCTION Transport projects need substantial capital investment with the aim of providing a better service to users as well as society as a whole over a long period of time. To conduct such an investment, it is crucial to determine who will ultimately pay for the service, and who will be able to raise the resources to implement it. The first aspect, usually known as funding, has to do with the ability of the project to capture the benefits produced by it and transform them into an ongoing revenue stream over the life of the project. The second one, usually known as infrastructure financing, refers to the capacity of promoters to mobilize the resources needed to raise the investment required to implement the project. Transport investment, especially for infrastructure assets, has been severely constrained after the credit crunch that took place in 2008. The imbalance between financing needs and the limited availability of funds has been widening the infrastructure gap (Vanelslander et al., 2018). Yearly investment needs in economic infrastructure assets—almost 40% of which are transport-related—were estimated at around $3.3 trillion at the global level just to keep pace with the required growth rates (Woetzel et al., 2016). Parts I and II of this Handbook focus to a greater extent on the economic and efficiency aspects of transport pricing. This chapter intends to introduce the role of pricing as an important component of infrastructure funding and provide an overview of the main characteristics and issues currently faced by transport funding and financing approaches. The chapter is structured into four sections. Section 14.2 defines and classifies the terms funding and financing. Section 14.3 focuses on the private financing of infrastructure. Section 14.4 explains the relationship between the social benefits produced by a certain infrastructure and the possibility of capturing them to finance transport investment. To finalize the chapter, Section 14.5 shows the main policy issues of transport infrastructure funding and financing nowadays to point out future research and implementation needs.

14.2 FUNDING AND FINANCING TRANSPORT INVESTMENT 14.2.1 Definition of the Concepts When dealing with transport investment, there are two complementary concepts to consider: funding and financing. Even though the two terms are interrelated, the role they play is different. Funding focuses on who ultimately pays, so it refers to the capacity of the project to capture all kinds of benefits generated by its implementation and, consequently, produce a revenue stream over its lifespan. An investment project may have funding problems for two reasons: first, because the project does not produce sufficiently large benefits; and second, because even though 273

274  Handbook on transport pricing and financing

the project produces benefits, capturing them is difficult for political, technological, or other reasons. For instance, it may not be politically acceptable to establish toll roads in certain countries. Financing, in turn, is the ability to raise resources to finance the capital cost of the project. Investment projects usually require large upfront investment needs to build or acquire the facilities. The difference between funding and financing comes from the time dissociation between the revenue stream and the costs incurred to develop, maintain, and operate an investment project. A certain project may produce enough funding (revenue allocated to cover the project’s costs) over its lifespan but may have problems obtaining financing at the time the capital cost is needed. If bringing resources upfront is not possible, there will be no way to develop the project and, consequently, enjoy the benefits it produces. For the successful implementation of an investment project, funding and financing go hand in hand. First, it is necessary to capture the benefits of the project. These benefits may fall on those users who directly enjoy them, other potential beneficiaries, and the government on behalf of society. And second, it is crucial to ensure that those benefits can be materialized in financing at the time the investment is needed. After introducing the two concepts, the main funding and financing sources usually available for investment projects are described below. Irrespective of their legal nature, funding resources mostly come from three sources: ●

●

●

Prices to direct users, whose perceived benefit is higher than what they actually have to pay. That is the case, for instance, of drivers who pay a toll because their utility associated with that option (in terms of comfort, travel time savings, etc.) makes it more preferable than free alternatives. Charges or taxes to indirect beneficiaries who take advantage of, or reap benefits from, the project in a non-direct way. This is the case, for instance, of landowners or commercial activities that see how their value goes up due to the accessibility increase triggered by the project. Those indirect beneficiaries may be required to pay for those benefits through different methods that will be explained in greater detail in the following subsection. Subsidies from general taxes by government decision. Some transportation benefits, especially environmental and social ones (such as air quality), are not always easy to charge to individuals. However, those benefits are important for the wellbeing of the society and for achieving a balanced sustainable development between current and future generations. Because of that, it is usually the government that is willing to pay for those benefits on behalf of the society through the general budget. Using the general budget implies that taxpayers (current or future ones depending on the model adopted) end up paying for the project fully or partially.

Following up with financing sources, a double classification can be made by distinguishing financing entities willing to invest in the project and financing instruments available to do it. The most important financial entities investing in transport projects are governments, stateowned enterprises, multilateral financial institutions and private investors: ●

Governments provide financing through specific items of the national, regional and local budgets, specifically devoted to investment.

Transport funding and financing  275 ●

●

●

State-owned enterprises (SOEs) are promoted by governments at different levels to develop investment projects, mostly infrastructure ones. Operationally, they work like private corporations, the main difference being that the equity is controlled by the government, which imposes the activities to be carried out by them. Multilateral financial institutions are created by a group of countries that provide financing and professional advice to enhance development. Private investors may invest in both privately owned and government-owned projects.

Financial instruments refer to the specific sources that firms and governments may use to finance infrastructure projects. They can be divided into: ●

●

●

Grants, which are characterized as non-refundable. This is the most common way for governments to provide financing for transportation projects. Debt, which is characterized by being refundable with a cost associated with the interest rate applied. Lenders have no control over the decisions of the firm. Equity contributions, which are composed of shareholders’ contributions to the firm with the possibility of influencing its decisions. Equity refundability and profitability will depend on the ultimate performance of the firm. However, investors will require higher returns as risks become higher.

The combination of the above-mentioned concepts (funding and financing) enables the set-up of a taxonomy of how transport investment projects are financed. For instance, traditional government financing through budgetary items that are paid by the state budget will be classified from the financing point of view as grants provided by the government, while from the funding point of view it will be regarded as revenue from current national taxpayers. A shadow toll highway PPP1 will be classified from the financing point of view as an investment conducted by a private corporation issuing equity and debt, while from the funding point of view will be classified as resources coming from taxpayers over the life of the contract. The table below provides an example of how infrastructure investment projects are mostly financed by the central government of Spain for different transport modes.​

Table 14.1  Funding and financing mechanisms used in Spanish transport projects Funding and financing mechanisms Initial Financing

Public budget

Ultimate funding Taxpayers

Users

High-Speed Rail

Ports/Airports

Conventional Road/ Conventional Rail

State-owned Enterprises (Equity, debt, subsidies) Private corporation (Equity, debt)

Users/taxpayers

Shadow toll and Availability Payment

Toll Concessions

276  Handbook on transport pricing and financing

14.2.2 Funding Sources In this subsection, the different funding sources already mentioned (prices to users, charges to beneficiaries, and subsidies) are described in much greater detail. Beforehand, it is important to note that the previous classification is not always aligned with the legal instrument (taxes, fees, private prices, etc.) used to implement them, basically because legal frameworks differ across countries. For instance, while in Europe, taxes are rarely earmarked—being allocated to the revenue side of the general budget; some countries, such as the United States, allow earmarking of taxes for specific purposes. That happens, for instance, with fuel taxes that are earmarked for transportation activities. Similarly, value capture charges can also be implemented either through charges or taxes. 14.2.2.1 Prices to users Pricing is the most common means of capturing user benefits from transport investment projects (Cambridge Systematics, 2002). However, it is worth noting that prices are not always allocated to the funding of the infrastructure or service that justifies the implementation of the charge, since sometimes users’ revenues are allocated to other goals, such as enhancing environmental objectives or cross-financing other facilities. The greatest amount of user benefits potentially captured through pricing would theoretically be achieved by charging a price equal to the consumer surplus of each user at any time, which would be achieved by charging everyone according to their willingness to pay. That approach is, however, quite difficult for practical and legal reasons, though certain price discrimination strategies are being used by companies (such as airlines) to indirectly differentiate between users with higher or lower willingness to pay. Though there are welfare economic justifications for public services to adopt time-varying pricing, reality shows that flexibility to apply discriminatory and variable user prices depends on the extent to which a given project is controlled by the government. Overall, private institutions managing private projects or services are quite free to charge what they want if they avoid direct discrimination for reasons of age, gender, race, religion, etc. They are free, however, to charge more to someone who buys a ticket the day before or wants to travel on the most favorable schedule. Government-owned projects or services tend to be much more inflexible since prices, in that case, are often regulated through different methods such as price caps or rate of return approaches, thereby constraining the possibility of capturing the full consumer surplus. Recently, some public services are giving more freedom to private companies to set dynamic prices if there is a free alternative available to users. This happens, for instance, with managed lanes in the United States, where prices vary dynamically over time (Federal Highway Administration, 2008). In that case, users always have the option of taking the free alternative if they do not want to pay to save time. User charges set by the government may incorporate markups in some cases to internalize externalities. That is the case of the tolls set to Heavy Duty Vehicles by the Eurovignette Directive in the European Union (EU), where prices incorporate a surge depending on the emission characteristics of the vehicle and the environmental sensitiveness of the area the vehicle is going through (Gomez and Vassallo, 2020). Transport prices can take different forms depending on how they are charged. They can be set as a price for an origin–destination trip, as it happens with many public transport services (airlines, buses, railroads, etc.); they can be set according to the distance traveled

Transport funding and financing  277

by the vehicle, as is the case of tolls or rail usage fees; or they can be set as a fixed-time price that allows the use of a given transport network for a period, which is common in travel passes in many cities or vignettes (stickers) implemented in road networks of some countries. Finally, user charges can be implemented through different legal forms such as private prices, government fees, and even taxes, each with a different level of regulation. Certain taxes, despite their legal nature, may also be considered user charges. That is the case, for instance, of fuel taxes applied in a discriminatory way, and totally or partially earmarked for transport funding purposes, as it happens with the Federal Highway Trust Fund in the United States. 14.2.2.2  Charges to indirect beneficiaries Indirect beneficiaries of transport investment projects can be defined as those who, without being the main users of the projects, end up benefiting from them. Transport investment projects, especially infrastructure ones, produce external costs and benefits falling on people or businesses outside the project’s target. Thus, it seems fair that the project compensates for the costs and captures the benefits directly attributable to it from external agents. This section focuses on the main approaches employed for capturing the benefits falling on indirect beneficiaries for revenue purposes. Most of the external benefits of transport projects are associated with increased land and business value produced by improved accessibility, especially in urban areas (Medda, 2012). The most important land value capture approaches are the land-taking approach, where the government takes the land after compensating the landowners, develops the project, and then sells the service land to capture the whole value increased by the improvement; and the nonland-taking approach, where the government develops the project and then requires landowners to pay the value increase caused by the accessibility improvement through a charge or tax earmarked to the funding of the project; for more discussion, see Gomez-Ibanez et al. (2021) and Section 15.3 of Chapter 15 in this Handbook. Recognizing and accounting for the benefits produced by a certain transport project on landowners is a difficult issue for several reasons. First, because it is not that clear whether providing accessibility is a reward or rather a government obligation. Second, because quantifying external benefits in an objective and standardized way is very difficult. And third, because if landowners are the actual users of the transport service, then the increase in land value is already captured in their consumer surplus. The aforementioned aspects imply that enacting the legal framework to charge for the external benefits from transport projects, especially when the land is already developed at the time of carrying out the project, is a big issue. External beneficiaries are not only landowners. They can also be business activities associated with the project that is being developed (Doherty, 2004). This is the case of the rents paid by shops, restaurants, parking lots, vending machines, and advertising panels in transport nodes such as airports and stations. That revenue source should not be underestimated, as revenues from those activities at some airports may even exceed those from aeronautical fees paid by airline companies (Graham, 2009). Other revenues from non-users may be, for instance, the rents paid by telecom companies for using the surplus capacity of the telecom system installed on transport infrastructure or the rent paid by gas stations located on a highway.

278  Handbook on transport pricing and financing

14.2.2.3  Subsidies from general taxes Subsidies from general taxes are the source used to fund transport projects when the government decides to allocate public money to them. In that case, there is no direct connection between the allocation and the origin of funding, since those who pay more taxes are not necessarily those who mostly enjoy the benefits from the project. Conceptually, general taxes should be used either to pay for those public services whose benefits are difficult to assign to anybody, as might be the case for defense expenditure; to subsidize certain services that produce positive externalities difficult to capture, such as environmental projects; or for equity reasons, such as giving the poor access to essential services. Subsidies can be directed to users (as a substitute for tolls or fares), as it happens with shadow toll and availability payment PPPs; or to the construction, maintenance, and operation of transport facilities. In the former case, the budgetary burden is spread over the project lifespan, while in the latter, the financial burden concentrates on the time the investment is made. Some transport infrastructure facilities (such as airports and ports) are mostly funded by the users, so they rarely receive subsidies. Transit systems in metropolitan areas are often subsidized as a means of discouraging the use of private vehicles and facilitating mobility for the poor (Hess and Lombardi, 2005). Usually, rail and road infrastructure are fully or partially subsidized by governments for historical reasons. In the case of roads, it may have to do with the fact that users already pay a large amount of fuel taxes, which in most countries are not earmarked for transportation purposes. Finally, it is worth mentioning that in some countries, such as some states in the United States, general taxes, such as sale taxes, are earmarked for funding transport facilities (Gomez and Vassallo, 2014). That shows the big variety of funding sources available across countries. 14.2.3 Financing Sources In this subsection, the different options for financing transport infrastructure are classified and described in greater detail. First, the most important entities that ultimately provide resources to the projects are presented. Then, the main instruments that these entities can use to raise the financing of projects are briefly discussed. 14.2.3.1 Financing entities 14.2.3.1.1 Governments Most transport infrastructure is financed by national, regional, and local governments through direct allocations from the public budget. Budgetary resources are raised by governments mainly through taxes and by issuing long-term debt, usually backed by the treasury. Governments can also benefit from international aid in the form of subsidies (i.e., from development funds such as the European Regional Development Fund or the Cohesion Fund in the EU) (Vassallo and Garrido, 2019). General budget appropriations have a long-standing role in the financing of transport investment in many countries. However, the extent to which they have been used greatly varies across countries. China is a very notable example of extensive use of the government budget to finance infrastructure, having spent an estimated average of 13.5% of GDP on urban infrastructure alone since the mid-1990s (Chong and Poole, 2013). In contrast, the United States has traditionally relied to a greater extent on other financing sources (Chan et al., 2009). The scope and application of budgetary financing in different countries have also changed over the

Transport funding and financing  279

past few decades in response to budgetary and financial management reforms, as well as to changes in fiscal policies (Chan et al., 2009). Budget allocations are subject to strict oversight and control by audit authorities, as well as to budgetary constraints arising from each country’s fiscal rules in terms of public debt or deficit limits. That fact implies serious constraints to the timely construction of infrastructure. In addition, the final allocation of funds is at risk of being driven by decisions based on political rather than technical criteria, which may encourage the development of white elephants, and discourage the exploration of alternative funding options such as user charges (Chan et al., 2009; Chong and Poole, 2013). One of the apparent advantages of budgetary financing compared to private financing is that it is cheaper since the cost of government debt is usually lower than the financial cost offered by the private sector. However, that comparison is not as easy as it seems at first glance, since such a lower cost comes at the expense of a higher risk borne by the government, which is not internalized by the investors who acquire public debt (Vassallo et al., 2019; Avdiu and Weichenrieder, 2020). 14.2.3.1.2 State-owned enterprises Governments can also finance transport investment indirectly through SOEs. These are legally independent entities partly or fully owned and supervised by the government with the aim to implement or manage certain infrastructure facilities or services. SOEs have a certain independence from the strict regulatory rules imposed on many governments. That allows them to make investments without the bureaucratic and budgetary constraints to which governments are subject. These enterprises raise financing through several means, such as government’s budgetary allocations in the form of equity contributions or compensations for the services provided, retained earnings from user charges, and access fees, bank loans, and bond issuance (Chan et al., 2009; OECD, 2014). Infrastructure SOEs are among the largest companies in many countries, and they undertake a big share of public investment in infrastructure (International Monetary Fund, 2020). For instance, two-thirds of all public infrastructure investments in emerging markets and developing economies in 2017 were channeled through SOEs, accounting for 55% of all infrastructure project commitments that year (World Bank, 2018). Different worldwide experiences suggest that SOE investments are not always guided by cost–benefit considerations (Akitoby et al., 2007). SOEs often incur persistent losses or become heavily indebted, imposing high costs on governments—to the extent that they often have to be bailed out (Akitoby et al., 2007; World Bank, 2014). This has motivated the implementation of different measures intended to improve its corporate governance, transparency, and accountability to avoid market distortions, with positive results (World Bank, 2014). 14.2.3.1.3 Multilateral institutions International multilateral financial organizations provide financial and non-financial support to develop transport infrastructure projects, mostly in emerging and developing economies. There are many such institutions, diversified in terms of their role, function, mission, investment capacity, and area of activity (Gatti, 2013). They are key to circumventing market failures in countries—usually developing ones—with higher political and regulatory risks, which often find difficulties in raising the necessary funds from their own resources or from the private sector to meet the high upfront capital needs of infrastructure projects. As Chapter 17 in

280  Handbook on transport pricing and financing

this Handbook by José Carbajo shows, multilateral institutions (MIs) contribute to reducing those risks and ease the financing of the project, making it more attractive for the private sector (Gurara et al., 2018). MIs offer a variety of financial products appropriate for supporting both public and private investment in long-term transport assets, such as (repayable) subsidies, concessional and market-based loans with long maturities, equity investments, and guarantee instruments covering political and credit risks. In the last 15 years, MIs support to developing countries has increased from $50 billion to $127 billion (Rillo and Ali, 2017). Also, in response to the increasing need for infrastructure support, new MIs focused on infrastructure investment have been created in recent years (Humphrey, 2018). 14.2.3.1.4 Private investors Different types of private companies, both financial and industrial (private developers, infrastructure funds, pension funds, banks, etc.), can contribute to financing transport projects as either sponsors or lenders. Private corporations can develop, finance, and operate their own projects but also finance infrastructure projects promoted by the public sector through contractual agreements such as concession or PPP schemes. Private corporations can leverage financing to support one or more activities in the infrastructure lifecycle, such as the construction or operation of the facility, or to invest in an equity stake and own and manage the facility, thus being entitled to dividend payments (Deep, 2021). Private banks usually contribute to the financing of projects by providing debt without having control over the firm. The private sector can also receive support from the government or MIs. Governments in both advanced and developing countries are increasingly resorting to the private sector, particularly through PPPs, to invest and manage public infrastructure, especially in the transport sector (Rafaele Della Croce and Yermo, 2013). This trend has been motivated not only by growing public budget constraints that limit governments’ capacity to develop and properly maintain infrastructure, but also by the potential benefits in terms of efficiency gains that the private sector can bring about if the incentives are rightly aligned. However, inadequate risk-sharing and a lack of sufficient competition in the bidding process may undermine those benefits. 14.2.3.2 Financing instruments Each of the aforementioned actors can use different financial instruments to meet the financial needs of transport projects. These instruments can be divided into those that are non-refundable, such as grants, and those that must be repaid with a certain return, with debt and equity being the two most important categories. In practice, the use of grants is almost exclusive to governments, although some MIs and development funds may provide them as well. Different types of instruments have different risk and return profiles depending on their seniority, with senior debt as the lowest-risk, lowest-return instrument and equity as the highest-risk, highest-return type of financing. The most common debt instruments are loans and bonds, which can be structured to have long-term maturities. In large transport projects, debt may be divided into senior and subordinated tranches, whose main difference is the seniority of repayment of the principal. In this respect, subordinated debt ranks behind senior debt and ahead of equity in the seniority of rights to the project’s free cash flows (Gatti, 2013). The subordinated tranche would absorb credit losses before the senior one, thereby enhancing the credit quality of

Transport funding and financing  281

the latter. The greater the risk of default will be reflected in higher yields to the holders of subordinated debt. Finally, equity ranks at the lowest level in terms of seniority. Its return is linked to the ultimate performance of the project, therefore being the riskiest and most expensive source (Gatti, 2013). Equity usually comes from industrial sponsors of the project and third parties such as infrastructure investment funds and pension funds, although the latter tend to invest mainly in brownfield projects or acquire equity of greenfield projects in secondary markets once they are in the operational phase.

14.3 PRIVATE FINANCING OF TRANSPORT INVESTMENT After classifying the funding and financing approaches, the current section provides the most important features of private financing for transport investment by noting the differences between corporate and project finance, explaining the most important mechanisms for financial evaluation, showing the different financial sources, and describing public support mechanisms to private projects. This section places greater emphasis on project finance, given its higher relevance in the financing of transportation projects nowadays. 14.3.1 Project versus Corporate Finance Project finance (“off-balance sheet” finance) and corporate finance (“on-balance sheet” finance) are different financing methods to obtain the funds needed to carry out the upfront investment of a project. The major difference between the two types of financing lies in the security offered to lenders by the sponsors. In corporate finance, the company raises the financing needed against the balance sheet of the main corporation, which will be provided to the lenders as a guarantee or collateral in case of default. In that case, lenders have recourse to all the company assets and cash flows derived from its business activities, which can be claimed to recoup their investment in the event of bankruptcy. In project finance, on the other hand, financing is raised against the project’s assets and capacity to generate resources in the future, which is measured through the expected cash flow of the project. Thus, lenders have non-recourse (or limited, as will be seen later) to the company sponsoring the project.​ Some of the most important characteristics of project finance and its main differences compared to corporate finance are presented below (Gatti, 2013; Yescombe, 2014). ●

●

The owner of the project’s assets is a company constituted ad hoc—usually called a special purpose vehicle (SPV)—whose sole activity will be the management of the project. The shareholders of that company will be the sponsors of the project. Lenders rely on the future value of the project (future cash flows, assets, guarantees, etc.) to pay back the loan and its interest, and have no or very limited recourse against the sponsors, as the risk borne by the sponsors is limited to their contribution (equity and sometimes subordinated debt) in the project. In some cases, the sponsors provide direct guarantees to the lenders of a project finance deal limited to pre-identified events, as is

282  Handbook on transport pricing and financing

CORPORATE FINANCE

PROJECT FINANCE

Investment PROJECT

Cash Flows

SPONSORS Dividends

SPONSORS

Corporate Loan LENDERS

Debt Service

Equity SPV

Cash

Limited guarantees

Debt Service PF Loan

LENDERS

Flows

PROJECT

Figure 14.1  Corporate finance and project finance structures

●

●

the case of the so-called completion guarantees, intended to cover certain risks that may be considered unacceptable by lenders. Project risks are distributed among the stakeholders of the project so that each risk is allocated to the counterparty most capable of controlling and managing it. Leverage is generally very high (usually between 70% and 90% of the capital cost of the project). That is intended to reduce the weighted average cost of capital of the project and limit the amount of resources sponsors have to invest.​

Using project finance as an infrastructure financing mechanism has both advantages and disadvantages compared to corporate finance. Considering advantages, project finance allows sponsors to limit the risk of the project on its corporation, insofar as sponsors’ corporations do not serve as collateral for the financing of the project. Moreover, it enables better risksharing between the different stakeholders and provides greater borrowing capacity, which limits the contributions of the sponsors and increases the return on equity invested. It also has the advantage that an SPV cannot divert financial resources to other purposes, in contrast to a corporation, which can divert the loans for the project to other projects or to solve cash flow issues. Considering disadvantages, it makes the structuring of the project much more complex and expensive, increases the cost of external financing because it has no or very limited recourse to the main corporation, and hinders the restructuring of projects because of the large number of stakeholders involved. The latter makes project finance suitable for capitalintensive projects since otherwise, transaction costs would be too high for the project to be financially viable. Project finance is preferably applied in regulated or monopolistic sectors where the certainty of future flows is greater. The use of project finance in the global market (both the total volume of loans and the number of transactions) has been increasing since the 1990s (Mohammadia, 2020). In just over two decades, the annual volume of project finance loans has increased elevenfold, and more than 13,000 transactions have been financed with a total amount of almost $3.8 trillion. Over the last ten years, the transport sector has accounted for, on average, 20% of the total volume of project finance worldwide (Mohammadia, 2020).

Transport funding and financing  283

Table 14.2  Main differences between corporate and project finance Dimension

Corporate finance

Project finance

Type of capital

Permanent—an indefinite time horizon for equity

Finite—time horizon matches the life of the project

Financial structures

Standardized and easily duplicated structures

Highly tailored structures which cannot generally be re-used

Degree of leverage

Depends on the effects on the borrower’s balance sheet

Depends on the project’s risks—leverage is usually much higher

Transaction costs for financing

Low costs due to competition from providers, routinized mechanisms

Relatively higher costs due to documentation and longer gestation period

Collateral

Assets/cash flows of the borrower company

Limited to project assets/cash flows

Accounting treatment

On-balance sheet

Off-balance sheet (the only effect will be disbursement to subscribe equity in the SPV or for subordinated loans)

Basis for credit evaluation

Overall financial health of corporate entity; focus on-balance sheet and cash flow

Technical and economic feasibility; focus on the project’s assets, cash flow and contractual arrangements

Cost of capital

Relatively lower

Relatively higher

Source:   authors based on Gatti (2013) and Bodnar and Comer (1996).

14.3.2 Financial Evaluation Financial evaluation is the process of identifying and comparing the benefits and costs of different alternatives to select the most suitable one for investors. This evaluation makes it possible to assess whether the profitability of the project is greater or lower than the cost of capital, and define the financial structure of the SPV (leverage, different types of sources, etc.) Overall, a project finance deal is considered ‘financially viable’ if the aggregate value of the expected revenues is greater than the total costs accumulated during the implementation of the project. There are various approaches to measuring the value of a project (Damodaran, 2012), the most commonly used being (i) discounted cash flow (DCF) valuation, which links the value of an asset to the present value of the expected future cash flows of that asset; and (ii) contingent valuation, which relies on real options models to measure the value of assets that share option characteristics. The approach most used for infrastructure valuation is the DCF approach. This method estimates the value of an asset as the sum of the present values of expected cash outflows and inflows over the life of the asset, discounted at a rate representing the cost of capital required to finance the asset under prevailing market conditions (Deep, 2021). That is, the DCF method values a project today in terms of its expected ability to generate revenues in the future, considering both the opportunity cost of money and the economic lifespan of the project.

284  Handbook on transport pricing and financing

The determination of the financial discount rate (cost of capital) is crucial, as it is the key factor to bring the future revenue projections back to the present. The cost of capital will depend on the risk of the estimated cash flows. Consider, for example, a project that is to be financed with a mix of debt and equity. Debt and equity holders will assume different levels of risk in the project, which will be reflected in the cost of their respective financing. To estimate the value of the project, the expected cash flows before any payments are made to the debt or equity holders are discounted at the weighted average cost of capital (WACC), which measures the cost of the different financial sources used in the project weighted according to their respective market value share (Damodaran, 2012). Cash flows and the discount rate must be expressed in comparable terms; thus, nominal cash flows should be discounted at the nominal rate.

RWACC = RE ´

D E + RD ´ (1 - Tc ) ´ (14.1) V V

Where: V = D + E is the total market value of the project, E is the market value of equity, D is the market value of debt, RE and R D, are the cost of equity and debt, respectively, and TC represents the corporate tax rate. Since debt interest are deductible for the calculation of corporate taxes, the greater the share of debt the greater the tax shield (tax savings provided by the financial structure of the project). However, the higher the debt ratio, the higher the probability of getting into financial distress and eventually bankruptcy, given the greater risk that the company will not be able to meet its debt obligations. The cost of equity corresponds to the investor’s opportunity cost of money (opportunity rate), which is the highest return foregone by investing in the project. In other words, the cost of equity is the rate of return required by equity investors to invest in a specific project compared to other available options in the market. The methodology most used for determining the cost of equity is the Capital Asset Pricing Model (CAPM).2 This model estimates the expected return on equity as the sum of the risk-free rate and the expected market risk premium adjusted with the equity beta, which measures the sensitivity of the equity returns to changes in the market returns—that is, the systematic risk of equity. The market risk premium represents the expected return of the market portfolio over the return of the risk-free asset (such as treasury bonds). The systematic risk of equity will vary depending on three factors: (i) the asset beta—also known as the unlevered beta—which measures the market risk of the asset holding the equity, that is, how sensitive the company’s business is to the general market situation; (ii) the corporate tax rate; and (iii) the debt-to-equity ratio, which reflects the higher financing risk assumed by shareholders when the company is leveraged. Under the DCF approach, there are two universally accepted metrics used to value transport projects and thus assess whether the investment is profitable: the Net Present Value (NPV) and the Internal Rate of Return (IRR) (Larrabee and Voss, 2013). These methods succeed in incorporating two core principles of finance into project valuation through the discount rate: value time of money and uncertainty; however, they fail to take managerial flexibility into account. Flexibility refers to the choices that managers can take among alternative options as a response to possible events (Koller et al., 2015), that is, the project’s capacity to adapt to future scenarios. The previous problem can be faced through the use of real options (RO) analysis, which allows to assess and measure the added value that arises from the options that may come up over the life of the project (Brennan and Schwartz, 1985). RO analysis is not a substitute but an extension of the DCF valuation that is especially useful in situations where uncertainty

Transport funding and financing  285

is high. Any valuation using RO takes the underlying asset value estimated using the DCF method (the value of the project without options) and then adjusts it for contingent decisions (the value of the options).

Total NPV = Basic NPV + Value added of the option analysis (14.2)

Types of options can be classified into operating options, investment and divestment options, and contractual options (Amram and Kulatilaka, 1999). The first ones provide flexibility to respond to uncertainty during the operational phase of the asset. These can include alternative uses of the asset, switch up/down products or factors of a productive process or scope up/down the range of activities to cope with a demand increase/decrease, renegotiate the contract, or temporarily shut down the project until conditions improve. The second ones can substantially modify the configuration of the asset or the set of feasible decisions in the future, such as decisions on scale (expand, contract, or giving up the project) and timing (accelerate or defer the investment). Finally, the third ones are contract-specific conditions that change the risk profile to be assumed by the asset owners. Examples include minimum revenue guarantees or contractual compensation mechanisms for early termination in PPP projects. Currently, two main methods are used for the valuation of financial options that can be applied to the analysis of investments involving options: the Black–Scholes (Black and Scholes, 1973) and the binomial (Cox et al., 1979) models. These models estimate the value of an option based on the value of a portfolio of securities traded in the market that have the same return as the option and mimic their fluctuations in value over time.3 Their application in infrastructure projects has been increasing in the academic literature since the 2000s (Martins et  al., 2015), showing the added value of using RO in large public infrastructure investment planning and decision-making (Brand et  al., 2000; Zhao et  al., 2004; Easson, 2016; Zuluaga and Sánchez-Silva, 2020). Many studies show RO analysis applicability to PPP projects regarding risk mitigation strategies and management flexibility versus investment decisions (Rakić and Radenović, 2014). In this respect, several models have been developed that are useful for both private and public partners to better analyze the financial risk of PPP projects under future demand uncertainty and properly design contractual compensation mechanisms, such as revenue risk-sharing mechanisms or early contract termination clauses (Buyukyoran and Gundes, 2018; Pellegrino et al., 2019; Ashuri et al., 2012; Liu et al., 2017). Although duly studied and applied by private sector analysts and practitioners, RO do not yet seem to have received due attention in the decision-making process for public infrastructure investment projects, neither in theory nor in practice (Pizzutilo and Venezia, 2016). 14.3.3 Traditional and Innovative Financial Sources 14.3.3.1 Commercial banks Commercial banks are the largest providers of project finance (Yescombe, 2014). Senior loans for large transport projects are usually arranged through syndicated loans provided by a group of banks. In this case, the project loan is distributed among the different participants, each of which contributes with a defined percentage of the loan, lowering their exposure to the project. The terms and conditions of loans can be tailored to a large extent to the individual needs of each project. They are flexible, as repayment terms can be renegotiated to reflect the actual project’s cash flows in case the project does not perform according to the terms originally agreed (Weber et al., 2016). In some project finance structures, bank loans are often used as a

286  Handbook on transport pricing and financing

bridge or mini-perm to finance the construction phase, in order to refinance it with bonds once the facility is in operation. Loan maturities can range from 5–10 years to 20–30 years (not unusual in PPP agreements). However, due to regulatory reasons, bank appetite for long-term investment is decreasing compared to bonds (Raffaele Della Croce et al., 2015). Generally, the financial contract stipulates that the borrower must pay interest on the outstanding balance of the debt for the period agreed and repay a portion of the principal. These payments take priority over all other debt financing. The annual interest to be paid is the result of multiplying the outstanding principal of the debt times an interest rate equal to the reference rate—e.g., Euribor in the EU and Libor in the United States—plus a spread reflecting the cost of the risk premium, which may be adjusted over the life of the loan depending on contract conditions. 14.3.3.2 Infrastructure funds Infrastructure funds are investment vehicles that gather capital from institutional or private investors with the mandate to invest it on their behalf in infrastructure assets. These funds act as financial intermediaries in the infrastructure investment market, making it accessible to investors that may not have the necessary resources or knowledge to make direct investments in this type of assets, or may not be allowed to invest directly in equity or debt. Typical investors include insurance companies, pension funds, banks, and multilateral development institutions. Some private banks and multilateral institutions are also implementing their own infrastructure funds to promote investment in infrastructure projects. Infrastructure funds can be either publicly traded on a stock market (listed funds) or unlisted private-equity-type funds (unlisted funds) (Bitsch, 2012). Listed funds have an unlimited lifespan, so they can continuously adjust their portfolio and hold some assets indefinitely (Deep, 2021). Unlisted funds have a limited investment horizon, which makes them less attractive for institutional investors such as pension funds and insurance companies. However, they have greater flexibility to decide the instruments and assets they can invest in (Garcia-Kilroy and Rudolph, 2017). Listed funds also provide greater liquidity to investors, which always have the option to sell their shares on the stock market. They also have lower administration fees. These facts make them desirable to a wider investor base, from large institutional investors to retail ones. Investment in infrastructure funds can be materialized through a variety of ways offering different risk-return pairs. That will depend on the type of project, the investment stage— greenfield or brownfield—or the type of capital committed—equity, mezzanine, or debt. In order to bypass intermediaries and have a higher influence on fees and investment teams, some investors are promoting collective investment platforms to make direct investments in infrastructure projects (Garcia-Kilroy and Rudolph, 2017). 14.3.3.3 Bonds Project bonds are an interesting complement or substitute to bank financing for large transport projects, especially brownfield ones, requiring long maturities. Like loans, both principal and interests are repaid to bondholders from the cash flows of the project. The maturity and the interest rate (also called coupon rate) of the bond will depend on the quality and requirements of the specific project, but also the current situation of the capital markets and the appetite of investors (Weber et al., 2016). Project bond issuance by a project finance SPV is usually not common until the project comes into operation, due to the high pre-operational risk related to the construction cost and eventual demand for the project. For that reason, a common practice is to finance the

Transport funding and financing  287

construction with a short-term bridge or mini-perm loan until the project is in operation, and subsequently refinance that loan by issuing a long-term bond, which allows for reducing the financial cost for the SPV. Before the financial crisis that took place in 2008, it was a common practice to wrap project bonds with the guarantee of monoline insurance companies to raise bond ratings, often to the maximum, and make them more attractive for institutional investors. After 2008, these guarantees are not available anymore. However, this limitation has not constrained the bond market as much as originally expected. In fact, the global project bond market has been increasingly active during the last decade, even above pre-crisis levels (Crédit Agricole Corporate and Investment Bank, 2020). In Anglo-Saxon countries, such as the United States and the United Kingdom, there is a relatively liquid market for project bond trading, as project financing through bond issuance has been a fairly common practice for a long time (Weber et  al., 2016). In the rest of the world, the use of bond financing has been gradually increasing (Crédit Agricole Corporate and Investment Bank, 2020) as an alternative promoted by governments to circumvent the strict regulation of the banks, as is the case of the Project Bond Initiative in the EU (MosionekSchweda, 2016; Vassallo et  al., 2017). Project bonds are mainly purchased by institutional investors, whose long-term investment needs are well suited to the life of transport infrastructure projects. However, they are also appealing to other investors, such as investment funds, investment banks, commercial banks, and foundations (Gatti, 2013). 14.3.3.4 Asset securitization Asset-backed securities are bonds collateralized by a pool of infrastructure assets that are sold via the capital markets to investors. The securities issued are usually structured in tranches of different credit qualities ranging from senior to equity ones, according to the priority conditions established in the contracts. Institutional investors generally invest in securities with high credit ratings (no lower than investment grade), while hedge funds are interested in greater returns by purchasing lower-rated securities. The tranche with a lower credit rating is, however, difficult to sell, so it is often acquired by project sponsors. Securitization can be used in two different ways. The first, known as indirect securitization, consists of financing a project with a syndicated bank loan that is subsequently securitized by the bank to remove the asset from its balance sheet for regulation purposes. The second one, called direct securitization, consists of directly transferring all or part of the future project’s revenue rights to a securitization fund which, using the assets as collateral, obtains financing by issuing securities with different risk profiles. To strengthen the security of the bondholders, a reserve fund is usually set up to maintain liquidity and ensure that they receive their coupon even if the project goes through a period of crisis. 14.3.4 Public Support to Private Projects The main objective of public support for private projects is to make them bankable when they are economically sound but not financially viable. In some cases, public support may be crucial to overcoming market failures since capital and insurance markets are both imperfect and incomplete (Stiglitz and Rosengard, 2015). It is not always possible to find loans with sufficient maturity, or at the necessary price. Similarly, certain risks may not be insurable at any price. This is mainly due to asymmetries of information between lenders and insurers, on one side,

288  Handbook on transport pricing and financing

and borrowers and policyholders, on the other. Borrowers and policyholders are more aware of the risks they face and the measures they will adopt to minimize them than lenders and insurers. The latter leads to problems both before the contract is signed (adverse selection) and after (moral hazard) (Irwin, 2003). As previously mentioned, transport infrastructure projects may also bring social or environmental benefits that are not captured by the price users are willing to pay for the service (World Bank, 2012). In that case, public support is used to internalize the positive externalities arising from the project. Due to political or social objectives, governments may also want to keep user fees far below consumers’ willingness to pay. As a consequence, projects’ revenues will not be sufficient to recover costs, which may discourage private investment (Irwin, 2003). Public support for private projects can be provided through different mechanisms (Garrido, 2019): (i) subsidies, in the form of cash or in-kind contributions, intended to reduce the capital requirements of the project or to improve the return of an otherwise unprofitable project; (ii) capital contributions of equity or debt, aimed at meeting the liquidity needs of the project; and (iii) guarantees against certain risks, such as construction or demand risk, to enhance the creditworthiness of the project. Both governments and MIs have implemented different types of instruments or programs intended to provide support for transport infrastructure projects. For example, the Transportation Infrastructure Finance and Innovation Act (TIFIA) in the United States offers loans, which can be both senior or subordinated, with favorable terms regarding maturity and repayment schedule. The program also provides credit enhancement instruments such as loan guarantees and standby credit facilities during the first ten years of project operations. In the EU, the European Commission is setting up an investment program called InvestEU that will merge all centrally managed financial instruments intended to leverage private financing available at the EU level to transport infrastructure projects into a single structure (Vassallo and Garrido, 2019).

14.4 SOCIAL BENEFITS STEMMING FROM FUNDING AND FINANCING TRANSPORT INVESTMENT PROJECTS As Georgina Santos, Iven Stead, and Tom Worsley explain in Chapter 15 of this Handbook, an investment project must be realized if the social benefits it produces are higher than its social costs over its life cycle. The financial feasibility of a project—understood as the possibility of getting enough revenue to raise the capital cost to develop, maintain, and operate the project over its lifespan—does not necessarily ensure the socio-economic feasibility of such a project. And the other way around, projects that are good from the social perspective may not be financially feasible. This is something that decision-makers should consider, since some of them may wrongly think that projects financially feasible without government contributions are always good for society. An investment project may be financially viable but not good from the socio-economic point of view when it produces large benefits for the users that can be easily captured through fares or tolls at the expense of bringing along big externalities that are difficult to internalize (van Essen et al., 2020). That might be the case, for instance, of an elevated highway going through the midst of a big city. While it is a very good option for drivers, who will be willing to pay a lot to save time, it does not appear to be a good solution for the city in terms of noise,

Transport funding and financing  289

pollution, and impacts on the built environment. Actually, most cities in developed countries are getting rid of that type of infrastructure facilities by either replacing them with tunnels or promoting cleaner transport means. The other way around, there might be projects that, being good from the socio-economic point of view (their social benefits are larger than their social costs), are not financially feasible without government subsidies. There are some reasons for that (Irwin, 2003). The most typical one is that most of the benefits produced by the project are positive externalities that are difficult to capture, so that the project ends up not generating enough revenue to finance its capital cost. As Chapter 9 in this Handbook on optimal pricing in public transport shows, subsidies may also be necessary if marginal cost pricing is implemented under increasing returns to scale. From the financial perspective, this will not be a problem if the government decides to subsidize the project, given its social value. However, practical experience demonstrates that governments do not always provide funding to socioeconomically feasible projects if they are not politically appealing (as it happens with road maintenance), or simply if the budget does not have enough resources to face all the investment needs the country or region requires (Smith et al., 2015; Salih et al., 2016). Bruno De Borger and Antonio Russo discuss in Chapter 7 in this Handbook what ‘political appeal’ could mean in the context of transport pricing policies. Another reason may have to do with the fact that implicit or explicit charges to users or beneficiaries are not always devoted to funding the facility they are coming from. That is the case, for instance, of fuel taxes for road vehicles since, despite being discriminatory with other modes, they are usually allocated to the general budget (Schroten et al., 2019). Finally, there might be projects that, even producing sufficient revenue, are not able to bring it in upfront to finance the capital cost required to implement the project. The reason for that is usually the weakness of capital markets in certain countries, especially developing ones (Rosales and Vassallo, 2012). Multilateral banking can definitively contribute to bridging that gap. Finally, it seems necessary to make a reflection on cross-subsidization or cross-financing. By cross-subsidization, we mean the fact that the revenue collected by a project is not necessarily devoted to the project that produces it, but rather to another project or activity related or not to the original source. Cross-subsidization is common across economic sectors, and, within the transport sector, across different modes or sections of the network (Gwilliam, 2017). Several questions come up regarding cross-subsidization. Is it efficient and fair to apply it? Does it make sense to employ the resources collected to policies that have little or nothing to do with the project that produced it? Responding to this question goes far beyond the scope of this chapter. However, it is worth pointing out some ideas that may help understand the problem. Cross-subsidization may be justified for several reasons, mostly efficiency and equity ones (Beato, 2000). From a theoretical perspective, the most efficient outcome will be to charge users at their social marginal cost (wear and tear, congestion and scarcity caused by lack of capacity, and the rest of externalities), and allocate the revenue to finance the most efficient projects from the social perspective. According to that, cross-subsidization will be not only good but also necessary in some cases. However, there are some caveats to consider in this respect. First, as part of the social marginal cost is the wear and tear of the facility, that part of the charge should not be allocated to a different purpose rather than preserving the value of the assets. Similarly, it will be reasonable to devote the congestion and scarcity part of the charge to increase the capacity of

290  Handbook on transport pricing and financing

the infrastructure, though in that case cross-subsidies may be justified if the problem of lack of capacity may be solved by building or improving another infrastructure or mode of transportation. Cross-subsidies across modes may also be justified when one of them has positive externalities, and the other has negative ones. It will make sense, for instance, to allocate the externality side of the charge applied to a certain mode (for instance, an urban highway) to cross-subsidize a facility with positive externalities (such as a metro system) that contributes to reducing congestion in the road network. It will also make sense to use the externality side of the charge to finance research projects to produce technologies that contribute to reducing such an externality. Cross-subsidization may also make sense when applied across sections of the same network for efficiency and equity reasons. Efficiency because the network needs capillarity to capture demand, and equity because sections with low traffic are usually those located in peripheral areas where the population has the lowest income levels. Using cross-subsidization to finance other sectors unrelated to the source of the revenue is more difficult to justify on fairness grounds. Using, for instance, the revenue coming from road user charges to finance health expenditure at the expense of poorly maintaining the road does not seem to be the right approach to tackle the lack of resources of that sector. The government should rather think about whether the problem comes from either not being able to capture the benefits of that sector, having low general taxation levels, or not being efficient enough in allocating the budget to different public expenditure policies.

14.5 POLICY ISSUES AND FURTHER RESEARCH NEEDS The last section of the chapter is devoted to pointing out some crucial policy issues regarding the funding and financing of large-scale transport investment projects to identify future implementation and research needs that may contribute to making the process more efficient and equitable. On the funding side, one of the most controversial issues is to clarify whether discriminatory taxation, as it happens with fuel taxes in the road sector, should be considered a price for the sake of leveling the plain field with other transport modes. Nowadays, a great variety of approaches to revenues and taxation are observed, making it somewhat difficult to determine whether they should be labeled as either prices or taxes from a transport funding perspective. The road sector, for instance, is characterized by high fuel taxes—often not applicable to other modes—which are rarely used to fund roads. Should they be considered as prices despite their legal nature? Clarifying this point is essential to ensure fair competition across modes. To make the process more transparent, governments should promote homogenous taxation across transport modes depending on the externalities produced (emission of pollutants, GHG, etc.). For energy products, the taxation should depend on the efficiency of the energy cycle from well to wheel, considering the type of fuel, the characteristics of the engine, etc. Another crucial aspect for transport authorities concerns the use of revenue from transport charges for other purposes different from the funding of the transport facility where prices were originally charged. In practice, some governments use the revenue coming from one mode or section of the network to fund a different mode, section, or even a completely different policy. Despite being a crucial policy challenge, little research has been conducted on that topic from the efficiency and equity perspectives. Moreover, putting into effect mechanisms to

Transport funding and financing  291

capture the positive externalities from indirect beneficiaries of transport investment projects is still a challenge to tackle by policymakers. Despite the wide literature and experiences available, little progress has been made to date regarding their practical implementation. When it comes to transport financing, an important point concerns the imbalance between supply and demand of high-risk securities, which are unavoidable in project finance deals. While large-scale, long-term investors, mostly pension funds and insurance companies, demand low-risk (high-rating) securities, a large share of the financial assets produced by transportation projects are high-risk (low-rating) ones. For some years, the monoline insurance market was able to bridge that gap by raising the rating of securities above investment grade. However, the overall failure of the monoline industry right after the 2008 credit crunch demonstrated that such an industry was neither resilient nor mature enough. Since then, some governments have been promoting different initiatives to bridge that gap. Further research needs in that respect are important at different levels. First, it appears necessary to explore whether the private sector can develop efficient and robust financial insurance markets that facilitate investment from institutional investors. And second, the efficiency of government guarantees and support mechanisms should be investigated in greater detail using empirical information. Finally, it is worth mentioning that the rapid evolution of technology can bring about new mechanisms to make the infrastructure investment market even more liquid and accessible to all kinds of investors. New approaches such as crowdfunding may help to raise funds from a variety of investors. New trends like that deserve to be studied in greater detail in the following years.

NOTES 1. A shadow toll PPP is an arrangement between a public authority and private contractor whereby the design, construction, financing, maintenance and operation of an infrastructure is transferred to the latter for a contractually fixed term, so that its remuneration—partially or fully demandbased—does not come from the users but from the corresponding authority. 2. See, for example, Damodaran (2012) for further insights on the subject. 3. Further information on valuation techniques can be explored in many books, such as Mun (2012) and Larrabee and Voss (2013)

REFERENCES Akitoby, B., Hemming, R., and Schwartz, G. 2007. “Public Investment and Public-Private Partnerships.” Economic Issues 40. International Monetary Fund. Amram, M., and Kulatilaka, N., 1999. Real Options: Managing Strategic Investment in an Uncertain World. Boston, MA: Harvard Business School Press. Ashuri, B., Kashani, H., Molenaar, K. R., Lee, S., and Lu, J. 2012. “Risk Neutral Pricing Approach for Evaluating BOT Highway Projects with Government Minimum Revenue Guarantee Options.” Journal of Construction Engineering and Management 138(4): 545–557. Avdiu, B., and Weichenrieder, A. J. 2020. “Financing Costs and the Efficiency of Public-Private Partnerships.” SAFE Working Paper, No. 295. Leibniz Institute for Financial Research SAFE, Frankfurt a. M. Beato, P. 2000. “Cross Subsidies in Public Services: Some Issues.” Inter-American Development Bank Sustainable Development Department Technical Papers Series. Inter-American Development Bank, Sustainable Development Department Technical Papers Series IFM-122.

292  Handbook on transport pricing and financing

Bitsch, F. A. 2012. “Infrastructure Funds: Their Role in Financing Infrastructure Investments Florian.” Thesis, Technische Universität München. Black, F., and Scholes, M. 1973. “The Pricing of Options and Corporate Liabilities.” Journal of Political Economy 81(3): 637–654. Bodnar, G. M., and Comer, B. 1996. “Project Finance Teaching Note.” The Wharton School of Business. http://finance​.wharton​.upenn​.edu/​~bodnarg​/ml​/projfinance​.pdf. Brand, D., Mehndiratta, S. R., and Parody, T. E. 2000. “Options Approach to Risk Analysis in Transportation Planning.” Transportation Research Record 1706: 54–63. doi: 10.3141/1706-07. Brennan, M. J., and Schwartz, E. S. 1985. “Evaluating Natural Resource Investments.” Journal of Business 58(2): 135–157. http://www​.jstor​.org​/stable​/2352967. Buyukyoran, F., and Gundes, S. 2018. “Optimized Real Options-Based Approach for Government Guarantees in PPP Toll Road Projects.” Construction Management and Economics 36(4): 203–216. doi: 10.1080/01446193.2017.1347267. Cambridge Systematics, 2002. “Community and Social Benefits of Transportation Investment.” NCHRP Project 8-36, Task 22 Demonstrating Positive Benefits of Transportation Investment, Working Paper 3. https://www​.somervillestep​.org​/files​/Com​muni​tyBe​nefi​tsTr​ansInvest​_0202​.pdf. Chan, C., Forwood, D., Roper, H., and Sayers, C. 2009. “Public Infrastructure Financing: An International Perspective.” Productivity Commission Staff Working Paper, March 2009. Productivity Commission Staff Working Paper, March 2009. Chong, S., and Poole, E. 2013. “Financing Infrastructure: A Spectrum of Country Approaches.” 6 Reserve Bank of Australia, Bulletin, September Quarter. Cox, J. C., Ross, S. A., and Rubinstein, M. 1979. “Option Pricing: A Simplified Approach.” Journal of Financial Economics 7(3): 229–263. doi: 10.1016/0304-405X(79)90015-1. Crédit Agricole Corporate and Investment Bank, 2020. “Project Bond Focus: Project Bond Fundamentals.” Croce, Rafaele Della and Yermo, J. 2013. “Institutional Investors and Infrastructure Financing.” OECD Working Papers on Finance, Insurance and Private Pensions, No.36, OECD Publishing. doi: 10.1787/5k3wh99xgc33-en. Croce, Raffaele Della, Paula, J., and Laboul, A. 2015. “Infrastructure Financing Instruments and Incentives.” OECD. https://www​.oecd​.org​/daf​/fin​/private​-pensions​/ Infrastructure​-Financing​-Instru ments​-and​-Incentives​.pdf. Damodaran, A. 2012. Investment Valuation: Tools and Techniques for Determining the Value of Any Asset. Hoboken, NJ: John Wiley & Sons. Deep, A. 2021. “Infrastructure Finance.” In: Gomez-Ibanez and Liu (eds.), Infrastructure Economics and Policy: International Perspectives. Cambridge, MA: Lincoln Institute of Land Policy. Doherty, M. 2004. “Funding Public Transport Development through Land Value Capture Programs.” Institute for Sustainable Futures. Sydney. http://ecotransit​.org​.au​/ets​/files​/ land​_value​_capture​_ mdoherty2004​.pdf. Easson, M. B. 2016. “The Application of a Theory of Real Options (TRO) to the Decision-Making of Mega Urban Transport Projects (MUTPs).” Thesis, University of Melbourne. Essen, H. van, Wijngaarden, L. van, Schroten, A., Sutter, D., Bieler, C., Maffii, S., Brambilla, M., Fiorello, D., Fermi, F., Parolin, R. and Beyrouty, K. El. 2020. Handbook on the External Costs of Transport. Luxembourg: Publications Office of the European Union. https://www​.cedelft​.eu​/en​/ publications​/2311​/ handbook​-on​-the​-external​-costs​-of​-transport​-version​-2019. Federal Highway Administration, 2008. Managed Lanes: A Primer. Washington, DC: US Department of Transportation. Garcia-Kilroy, C. and Rudolph, H. P. 2017. Private Financing of Public Infrastructure through PPPs in Latin America and the Caribbean. Washington, DC: World Bank. Garrido, L. 2019. “Impact of Public Support Mechanisms on the Performance of Public Private Partnerships: An Empirical Analysis of Road Projects in Spain.” Thesis, Universidad Politécnica de Madrid. Gatti, S. 2013. Project Finance in Theory and Practice: Designing, Structuring, and Financing Private and Public Projects, 2nd ed. Waltham, MA/San Diego, CA/London, UK: Elsevier. Gomez-Ibanez, J., Hong, Y.-H. and Huynh, D. 2021. “Land Value Capture for Infrastructure.” In: Gomez-Ibanez and Liu (eds.), Infrastructure Economics and Policy: International Perspectives. Lincoln Institute of Land Policy.

Transport funding and financing  293

Gomez, J. and Vassallo, J. M. 2014. “Comparative Analysis of Road Financing Approaches in Europe and the United States.” Journal of Infrastructure Systems 20(3): 04014008. doi: 10.1061/(asce) is.1943-555x.0000193. Gomez, J. and Vassallo, J. M. 2020. “Has Heavy Vehicle Tolling in Europe Been Effective in Reducing Road Freight Transport and Promoting Modal Shift?” Transportation 47(2): 865–892. doi: 10.1007/ s11116-018-9922-3. Graham, A. 2009. “How Important Are Commercial Revenues to Today’s Airports?” Journal of Air Transport Management 15(3): 106–111. doi: 10.1016/j.jairtraman.2008.11.004. Gurara, D., Presbitero, A. and Sarmiento, M. 2018. “Borrowing Costs and the Role of Multilateral Development Banks: Evidence from Cross-Border Syndicated Bank Lending.” IMF Working Papers WP/18/263. doi: 10.5089/9781484386200.001. Gwilliam, K. 2017. Transport Pricing and Accessibility. Washington, DC: Brookings Institution. Hess, D. B. and Lombardi, P. A. 2005. “Governmental Subsidies for Public Transit: History, Current Issues, and Recent Evidence.” Public Works Management & Policy 10(2): 138–156. doi: 10.1177/1087724X05284965. Humphrey, C. 2018. “Channelling Private Investment to Infrastructure: What Can Multilateral Development Banks Realistically Do?” Working Paper 534, Overseas Development Institute. International Monetary Fund, 2020. “State-Owned Enterprises: The Other Government.” In Fiscal Monitor: Policies to Support People During the COVID-19 Pandemic. Washington, DC: International Monetary Fund. Irwin, T. 2003. “Public Money for Private Infrastructure. Deciding When to Offer Guarantees, OutputBased Subsidies, and Other Fiscal Support.” World Bank Working Paper No 10. Washington, DC: World Bank. Koller, T., Goedhart, M. and Wessels, D. 2015. Valuation: Measuring and Managing the Value of Companies. Hoboken, NJ: John Wiley & Sons. Larrabee, D. T. and Voss, J. A. 2013. Valuation Techniques: Discounted Cash Flow, Earnings Quality, Measures of Value Added, and Real Options. Hoboken, NJ: John Wiley & Sons. Liu, J., Gao, R. and Cheah, C. Y. J. 2017. “Pricing Mechanism of Early Termination of PPP Projects Based on Real Option Theory.” Journal of Management in Engineering 33(6): 04017035. doi: 10.1061/(asce)me.1943-5479.0000556. Martins, J., Marques, R. C. and Cruz, C. O. 2015. “Real Options in Infrastructure: Revisiting the Literature.” Journal of Infrastructure Systems 21(1): 04014026. doi: 10.1061/(asce) is.1943-555x.0000188. Medda, F. 2012. “Land Value Capture Finance for Transport Accessibility: A Review.” Journal of Transport Geography 25: 154–161. doi: 10.1016/j.jtrangeo.2012.07.013. Mohammadia, M. K. 2020. “Trends of Project Finance in the World Market and in Arab Countries.” Finance: Theory and Practice 24(1): 24–33. doi: 10.26794/2587-5671-2020-24-1-23-33. Mosionek-Schweda, M. 2016. “Financing of European Infrastructure Investments through the Bond Market.” International Business and Global Economy 35(2): 186–199. doi: 10.4467/23539496IB.16.056.5637. Mun, J. 2012. Real Options Analysis: Tools and Techniques for Valuing Strategic Investments and Decisions, 2nd ed. Hoboken, NJ: John Wiley & Sons. OECD, 2014. Financing State-Owned Enterprises: An Overview of National Practices. OECD Publishing. Pellegrino, R., Carbonara, N. and Costantino, N. 2019. “Public Guarantees for Mitigating Interest Rate Risk in PPP Projects.” Built Environment Project and Asset Management 9(2): 248–261. doi: 10.1108/BEPAM-01-2018-0012. Pizzutilo, F. and Venezia, E. 2016. “Real Option Analysis Applied To Transport Investment Projects.” Proceedings of the Third International Conference on Traffic and Transport Engineering (Ictte), 825–830. Rakić, B. and Radenović, T. 2014. “Real Options Methodology in Public-Private Partnership Projects Valuation.” Economic Annals 59(200): 91–113. doi: 10.2298/EKA1400091R. Rillo, A. D. and Ali, Z. 2017. “Public Financing of Infrastructure in Asia: In Search of New Solutions.” ADBI Policy Brief 2017-2. Asian Development Bank Institute. https://www​.adb​.org​/publications​/ public​-financing​-infrastructure​-asia​-search​-new​-solutions.

294  Handbook on transport pricing and financing

Rosales, R. R. and Vassallo, J. M. 2012. “Is There Scope for a New Project Bond Market in Europe? A Promising Opportunity for Promoting Transnational Infrastructure.” Journal of Globalization, Competitiveness & Governability 6: 89–99. doi: 10.3232/GCG.2012.V6.N3.05. Salih, J., Edum-Fotwe, F. and Price, A. 2016. “Investigating the Road Maintenance Performance in Developing Countries.” International Scholarly and Scientific Research & Innovation 10(4): 472– 476. doi: 10.5281/zenodo.1123648. Schroten, A., Scholten, P., Wijngaarden, L. van, Essen, H. van, Brambilla, M., Gatto, M., Maffii, S., Trosky, F., Kramer, H., Monden, R., Bertschmann, D., Killer, M., Greinus, A., Lambla, V., El Beyrouty, K., Amaral, S., Nokes, T., and Coulon, A. 2019. Transport Taxes and Charges in Europe: An Overview Study of Economic Internalisation Measures Applied in Europe. Luxembourg: Publications Office of the European Union. Smith, J., Weyber, M.-C. and Harrison, G. 2015. Assessing the Global Transport Infrastructure Market: Outlook to 2025. London: PwC. Stiglitz, J. E. and Rosengard, J. K. 2015. Economics of the Public Sector, 4th ed. New York/London, UK: W.W. Norton & Company. Vanelslander, T., Roumboutsos, A. and Pantelias, A. 2018. “Editorial: Understanding Funding and Financing of Transportation Infrastructure.” European Journal of Transport and Infrastructure Research 18(4). doi: 10.18757/ejtir.2018.18.4.3260. Vassallo, J. M., García, A., Garrido, L., Rangel, T. and Pino, P. del. 2019. “Beneficios Sociales Del Modelo de Concesión En La Gestión de Carreteras.” Madrid: SEOPAN. Vassallo, J. M. and Garrido, L. 2019. “Research for TRAN Committee – EU Funding of Transport Projects.” European Parliament, Policy Department for Structural and Cohesion Policies, Brussels. Vassallo, J. M., Rangel, T. and Baeza, M. A. 2017. “The Europe 2020 Project Bond Initiative: An Alternative to Finance Infrastructure in Europe.” Technological and Economic Development of Economy. doi: 10.3846/20294913.2016.1209251. Weber, B., Staub-Bisang, M. and Alfen, H. W. 2016. Infrastructure as an Asset Class: Investment Strategy, Sustainability, Project Finance and PPP, 2nd ed. Chichester, SXW: John Wiley & Sons. Woetzel, J., Garemo, N., Mischke, J., Hjerpe, M. and Palter, R. 2016. Bridging Global Infrastructure Gaps. McKinsey Global Institute. https://www​.mckinsey​.com​/industries​/capital​-projects​-and​infrastructure​/our​-insights​/ bridging​-global​-infrastructure​-gaps. World Bank, 2012. Best Practices in Public-Private Partnerships Financing in Latin America: The Role of Subsidy Mechanisms. Washington, DC: World Bank. World Bank, 2014. Corporate Governance of State-Owned Enterprises: A Toolkit. Washington, DC: World Bank. World Bank, 2018. “Who Sponsors Infrastructure Projects? Disentangling Public and Private Contributions.” Washington, DC: World Bank. Yescombe, E. R. 2014. Principles of Project Finance, 2nd ed. Oxford, GB/Waltham, MA: Elsevier. Zhao, T., Sundararajan, S. K. and Tseng, C.-L. 2004. “Highway Development Decision-Making under Uncertainty: A Real Options Approach.” Journal of Infrastructure Systems 10(1): 23–32. doi: 10.1061/(asce)1076-0342(2004)10:1(23). Zuluaga, S. and Sánchez-Silva, M. 2020. “The Value of Flexibility and Sequential Decision-Making in Maintenance Strategies of Infrastructure Systems.” Structural Safety 84(2020): 101916. doi: 10.1016/j.strusafe.2019.101916.

15. Investment appraisal: links between finance and economics Georgina Santos, Iven Stead and Tom Worsley

15.1 INTRODUCTION Appraisal is a tool that supports decision-making. It can be defined as the process of assessing, in a structured way, the case for starting, renewing, or expanding a project, a policy, a proposal, or a programme and the extent to which it would result in measurable benefits and/or costs to society. The results of the appraisal help to inform decision-makers, but, as we explain below, are not binding on them. Many other elements come into play in government decisions. This contribution of the Handbook creates a link between the economics of transport pricing and subsequent chapters on project funding and financing in the transport sector. We start by discussing the methodological developments in investment appraisal over the last half-century. We do this in Section 15.2. Government appraisal of transport projects has been centred around a cost–benefit analysis in most countries since the 1960s. The methods of cost–benefit analysis have been developed over time as its widespread application made its shortcomings apparent and this has given room to a number of adjustments that fine-tune and extend the process. This includes the requirement to combine the results of cost–benefit analysis with the decision-maker’s judgement when evidence on the monetised values of some costs or benefits, or even of their quantities, is lacking. These complex procedures now also attempt to estimate the wider economic benefits, some of which accrue to land and property owners. Capturing all or part of these benefits can sometimes help fund the project itself, a point that is welcome when government departments are strapped for cash. We discuss appraisal and finance, including value capture, in Section 15.3. Another adjustment that has become mainstream in countries like the United Kingdom is estimating and allocating risk, and correcting for the well-known optimism bias. We discuss these in Section 15.4. In Section 15.5 we turn our attention to the problem of public acceptance and political influence in investment appraisal, highlighting that investment decisions, although subject to rigorous appraisal, are ultimately taken in the context of a wider set of a government’s strategic policy objectives.

15.2 METHODOLOGICAL DEVELOPMENTS IN INVESTMENT APPRAISAL AND COST–BENEFIT ANALYSIS There are many high-quality existing summaries of the history of transport appraisal methods, for example, Boyce and Williams (2015), Worsley and Mackie (2015), Mackie et al. (2014), and Mouter (2020). We do not repeat those here but offer a brief overview instead. Early cost–benefit analyses (CBA) focused on travel time savings and accident costs, often using fixed trip matrix appraisal, whereby for any given origin–destination flow the total amount of travel demand for a given mode is assumed to be fixed, and only choice of route 295

296  Handbook on transport pricing and financing

is explicitly modelled (Mackie, 1996). In response to growing levels of road traffic, in the 1980s and 1990s there was increasing recognition of the importance of “induced demand” for transport appraisal (Standing Advisory Committee on Trunk Road Assessment, SACTRA, 1994). This is where improvements to the transport network can encourage changes to the mode, destination, time of day, or frequency of travel. So-called “variable demand modelling techniques” (Department for Transport, 2020a; Ortúzar and Willumsen, 2011), which accommodate these behavioural responses, are now standard practice, and as a result, the estimated impact of induced traffic can be included within transport appraisal calculations. In parallel, the economic valuations applied within transport appraisal have improved significantly, in both breadth and depth, since those early appraisals. Broadly, these fall into three categories: direct user benefits, environmental and social externalities, and wider economic impacts. Values of time remain of prime importance in most transport appraisals, with direct transport user benefits representing the lion’s share of benefits for most transportation projects. However, significant advances have been made in the use of stated preference methods to estimate values of travel time savings (VTTS) in most Western European countries, including the United Kingdom, the Netherlands, Denmark, Sweden, Norway, and Germany (De Jong and Kouwenhoven, 2020). Government guidance tends to be periodically updated to both take these methodological developments into account and to update the recommended values themselves. These studies generally find a large degree of heterogeneity in the VTTS, with significant variations according to journey purpose, income, trip distance, and mode. Despite this, many countries’ CBA guidelines use average VTTS for leisure and commuting travel. In the case of the United Kingdom, this is due to concerns that it would be inequitable if some user groups were assigned a higher VTTS than others in the appraisal (Department for Transport, 2017a). In other countries, varying degrees of segmentation by mode and distance are adopted. Generally, the idea of segmenting by income in CBA is not supported, mainly for equity reasons, and when there is segmentation by income, such as in the Netherlands, it is often used for distributional impact analysis of road pricing projects rather than CBA of transport infrastructure projects (Mouter, 2016). For business travel, variations across all of these dimensions are seen as acceptable, as the benefits of quicker business trips are taken out as an efficiency gain in terms of lower prices and/or higher productivity. Importantly, in all cases user benefit calculations account for monetary costs such as fares, tolls, fuel costs, and motoring taxes, in addition to the value of travel time, using the notion of generalised cost (McIntosh and Quarmby, 1970) whereby time and money are aggregated into a measure of the overall disutility of travel. One of the challenges for user benefit estimation is how to model changes in land use, where not only the costs of travel but also the attractiveness of trip origins and destinations might vary in response to a policy intervention, including, for instance, a new transport infrastructure project (Geurs et al., 2006; Börjesson et al., 2014; New Zealand Institute of Economic Research, 2013). In response to this challenge, some recent advances have taken place in the academic sphere, which will hopefully be translated into practice before too long. One such example is the development of simulation models that take into account changes in residential location, with embedded endogenous land prices and demand for residential land (Anas and Liu, 2007; Eliasson et al., 2020). Project ranking and selection may change when endogenous changes in residential location are modelled (Eliasson et al., 2020), which opens a new area in practice, especially for policymakers attempting to select transport infrastructure projects, with the ever-present budget constraints.

Investment appraisal 

297

In addition to the above, Spatial Computable General Equilibrium (SCGE) models can have location-dependent elements in household utility functions. SCGE models build on microeconomic agents’ behaviours, thus including firms and households, but computations can be very costly (Bröcker and Mercenier, 2011). Many transport infrastructure projects will trigger changes in land use (Laird and Venables, 2017) and this can indeed be reflected in SCGE models. Due to these modelling techniques still being on the research frontier, and computations being so costly, they have not been widely applied to transport appraisal. There have been, however, reviews and reflections on the scope to borrow these ideas from urban economics, such as McCartney et al. (2013), but the key challenge is to find a simple and consensual approach that fits into the existing appraisal framework. The 21st century has seen an increasing monetisation of environmental externalities within transport appraisal, moving beyond impacts historically valued such as accidents and noise to the inclusion of air quality, carbon, physical activity, and landscape impacts. This reflects both political ambitions to reduce greenhouse gas emissions, where transport is now the single largest contributing sector in a number of countries, including the United Kingdom (Department for Business, Energy and Industrial Strategy, 2021) and the increasing evidence base around the adverse health effects of exposure to nitrogen dioxide and fine particulate matter (Committee on the Medical Effects of Air Pollution, 2018). These impacts are captured using a mixture of revealed and stated preference methods, with increasingly strong links between transport and health economics through the use of Quality Adjusted Life Years (QALYs), the Value of a Prevented Fatality (VPF), and the Value of a Life Year (VOLY) to value mortality and morbidity effects (HM Treasury, 2020). A range of other social and environmental impacts, such as community severance, biodiversity, and townscape are much harder to monetise using the available evidence. The area of transport appraisal with the most significant development in the past 20 years has been capturing so-called “wider economic benefits”, an area where UK guidance is most developed. The SACTRA report (1999) advised the UK government to look beyond conventional appraisal practice, rooted in assumed perfect competition but with allowances for key social and environmental externalities, to consider other market imperfections such as agglomeration externalities, planning constraints, market power, and taxes on labour income. This advance was informed by ongoing developments in the spatial economics literature, such as research on agglomeration economies (Glaeser, 1998; Krugman, 1998). As a consequence of developments instigated by SACTRA (1999), current UK guidance focuses on four wider impacts: productivity gains through greater agglomeration (Graham et  al., 2009), land value uplift from new developments which are dependent on transport interventions (Department for Communities and Local Government, 2016), tax revenue from increased or better-paid employment, and the mark-up between marginal cost and price under conditions of imperfect competition (Department for Transport, 2018). This has been termed the “CBA-plus” approach (Mackie et  al., 2018). A key focus for new work in this area is making more robust use of “supplementary economic models”, such as land-use transport interaction and SCGE models, to better capture the impact of transport schemes on the spatial distribution of activity within cost–benefit analyses.

15.3 APPRAISAL AND FINANCE The focus of investment appraisal differs between private and public sector organisations in two principal dimensions: the scope of impacts considered, and the approach taken to risk.

298  Handbook on transport pricing and financing

Most developed economies now consider a full range of social, environmental, and economic impacts with CBA (Mackie et al., 2014) for publicly funded projects. However, commercial operators focus on cashable benefits such as fares and operating costs, rather than consumer surplus or external benefits which would otherwise tend to dominate the CBA calculations. The general consensus amongst professional economists is now in favour of relatively low discount rates (Drupp et al., 2018). These represent required social rates of return on risk-free assets. There is a greater divergence in allowances for risk. Some countries such as France and the Netherlands recommend explicit uplifts to the discount rate to account for risk (Freeman et  al., 2018), whereas others such as the United Kingdom do not (HM Treasury, 2020). At the heart of this debate is the so-called “equity premium puzzle”, whereby observed rates of return on private sector investment cannot be easily reconciled with economic theory (Mehra and Prescott, 1985). Private firms can be expected to discount project returns at the relevant cost of capital, which may be project specific, and generally much higher than the discount rates recommended by finance ministries for public sector projects (Boardman et al., 2018). The opportunity cost of public financing can be handled implicitly using benefit–cost ratios (BCRs) to allocate scarce funds, or explicitly, via either increasing the discount rate or uplifting publicly funded costs in the appraisal (Spackman, 2020). Sometimes, private and public financing models coincide, in Public Finance Initiative (PFI), as they are called in the United Kingdom, or similar types of arrangements. This has been the subject of lively political debate in the United Kingdom, with no clear consensus on the best approach to include PFI costs in appraisal (National Audit Office, 2013), despite general recognition that the required rate of return to PFI investors is a relevant cost to include in the appraisal. Historical approaches have included using the government borrowing rate (HM Treasury, 1973; Office of Management and Budget, 2017), and using the social time preference rate (STPR) (HM Treasury, 2020). Each of these can impart a slight bias for or against private finance depending on whether the risk-free borrowing rate is more or less than the STPR. An alternative, more robust procedure is to profile private finance costs over time and then discount back at STPR (Ofgem, 2011; NERA, 2007). This can be done for government borrowing costs as well, in order to provide an assessment of which financing method delivers the lowest present value of cost for society. There are two equivalent approaches to undertaking CBA – using either resource costs or willingness to pay (Sugden, 1999). Under the first approach, user charges such as tolls and fares, as well as any taxes on transport or labour, are not explicitly included. Instead, using this approach, appraisal focuses on pure resource cost, to include those elements which have a genuine opportunity cost and are not merely transfers of money from one group to another. Using willingness to pay is more useful, as under this approach the changes in economic surplus earned by each sector of the economy – government, firms, and consumers – are evaluated separately. For example, in UK transport appraisal, fares paid appear as offsetting entries in “user benefits” and “operator revenues” (Department for Transport, 2021a). While not affecting the NPV, disaggregating the calculation in this manner makes the links between funding and social welfare clearer as the potential impacts of fare and taxation revenue, for example, on project affordability can be explored. The project itself, however, is likely to increase land and property value in its proximity. This increase in value can be recovered by the government, at least in part. Time savings are the major source of benefits in the appraisal of most transport schemes. Time savings are easy to measure, being part of the scheme design and, as noted in Section 15.2

Investment appraisal 

299

above, there are well-established methods of deriving VTTS. The use of travel time savings as a measure of value, however, has been challenged, as there is some evidence of travel time budgets being constant (Ahmed and Stopher, 2014; Stopher et al., 2017). In this sense, it has been argued, transport users do not actually “save” time, they simply re-allocate travel time. This can be achieved by reallocating travel time from one purpose to another or by changing the location of the activity, i.e. where people live and work (see, for example, Beck et al., 2017). The assumption behind value capture is that investment in public infrastructure, especially transport infrastructure, results in increased land and property prices in the area it serves. There is consistent evidence that this is indeed the case. This has been extensively documented for many public transport projects around the world (Gupta et al., 2020; Shima et al., 2017; Mathur, 2020a, 2020b, Bao and Mok, 2020; Mulley et al., 2017; Mulley and Tsai, 2016; Perk et al., 2013, to name just a few examples), although there is also some evidence that being located too close to a public transport station or along a public transport line or main road can also have negative impacts, such as traffic congestion, noise, and air pollution and this can negatively impact property prices (Zhang and Liu, 2015; Higgins and Kanaroglou, 2016, p. 618). In general, in addition to accessibility, often measured by proximity to the new transport infrastructure, most studies employ a number of other variables likely to influence land and property values (Higgins and Kanaroglou, 2016, p. 616). Practitioners of cost–benefit analysis in the transport sector rarely consider the question of how a project might be funded; a comprehensive coverage of funding and financing instruments is provided in Chapter 14 in this Handbook. In practice, it is generally assumed that the project is funded by the public sector, typically the “Transport Budget” (which can be national, state, provincial, or municipal, depending on the government level in question), perhaps financed through a private–public partnership (see Chapter 16 in this Handbook by Engel, Fischer, and Galetovic), or with some offsetting revenues paid by users. Constraints on public expenditure, as highlighted above, or the effect of the distortions caused by raising taxation, are accounted for by a shadow price for public funds or by a target benefit–cost ratio in excess of unity, which any scheme needs to meet. For reasons of both efficiency and equity, there are good reasons for seeking alternative sources of funding. Capturing the benefits through land value capture can be seen as an equitable example of beneficiaries paying. Value capture entails the public sector recovering some or all of the financial benefits that land and/or property owners and private developers perceive as a result of public investment in transport and other infrastructure. By capturing part of the benefits that accrue from public investments, the costs of the investments themselves can be recovered, at least in part. Different approaches have been used to capture the increase in land values that results from a scheme. Some of these involve contributions, either voluntary, negotiated, or compulsory, from developers of new or refurbished buildings close to the project. Such contributions can be seen as both equitable and efficient, in that the developer’s decision to build takes account of the costs imposed on the transport network. Examples include, but are not limited to, a land value tax, tax increment financing, a public transport levy, a special assessment tax (Yen et al., 2020), obligatory contributions, also called impact fees, and community infrastructure levy. The land value tax is a tax that only considers the value of the land, and not the properties built on it. The value of a specific piece of land typically increases when there is a public infrastructure investment nearby, especially a public transport infrastructure investment, and this increase in value can therefore be captured with a land value tax. This approach is used in Australia, amongst other countries (Yen et al., 2020).

300  Handbook on transport pricing and financing

Tax increment financing consists in financing the investment in (public transport) infrastructure with the revenues from a tax on the anticipated increase in property value (Yen et al., 2020). The key is that all future property tax increases in the area are earmarked to fund the project. One example of such an approach is the funding of the MetroRail Red Line, in Texas, United States (Yen et al., 2020). A public transport levy is an annual charge levied on properties likely to increase their value thanks to their proximity to an infrastructure project. The Gold Coast City and the Sunshine Coast City in Australia were partly funded in this way (Yen et al., 2020). A special assessment tax is a tax levied on properties located within a designated geographic area, often referred to as a special assessment district (SAD), to pay for the (public transport or other) infrastructure project, if they will benefit from it due to their proximity, through an increase in value. Washington, DC’s New York Avenue Metro Station and Seattle’s South Lake Union Streetcar project in the United States were both partly financed with a special assessment tax (Zhao and Larson, 2011). Developer contributions, which include obligatory contributions or “planning obligations” and the community infrastructure levy, are widely used in the United Kingdom. The obligatory contributions or planning obligations are contributions that private developers make to the local authority to help it fund infrastructure to support the development and mitigate the impact of development. These are negotiated between the Local Planning Authorities and the developers on a case-by-case basis (Lord et al., 2020). In other countries, they are often referred to as impact fees, for example by the World Bank (2015) and by the American Public Transportation Association (2015). Payment or commitment to payment is a pre-condition for planning permission (World Bank, 2015; Acts of Parliament, 1990, Section 106). The payment may be up-front (American Public Transportation Association, 2015), or periodical, either indefinitely or for a specified period (Acts of Parliament, 1990, Section 106). The community infrastructure levy is a locally determined fixed charge that can be imposed by local authorities (Ministry of Housing, Communities and Local Government, 2020). It applies to new developments, for example, housing developments. The revenues are then used to fund the infrastructure needed in the new (housing) development, whether this is a library, or a leisure centre, or a transport infrastructure project. To give an idea of the importance of developer contributions in England, planning obligations totalled almost £1.3 billion, and community infrastructure levy totalled over £1 billion in 2018/2019 (Lord et al., 2020). The distribution across regions was unequal. The incidence and value of planning obligations can vary according to location, scale of development, market conditions and local planning authority activity (Lord et al., 2020). One prime example of how a very large transport infrastructure project can be partly funded through value capture is Crossrail in London. This megaproject, which has a target opening date of 2022, is being partly financed by planning obligations, a community infrastructure levy and by a special supplement to the business rate (a local property tax) paid by all larger London firms (Greater London Authority, 2016). This Business Rate Supplement (BRS) was originally introduced in 2010, and has undergone a number of changes. Since 2017, the Crossrail BRS applies to non-domestic properties with a rateable value greater than £70,000. The BRS is 2 pence, so a property with a rateable value of £100,000 is liable to pay £2,000 per year under the concept of BRS (Mayor of London and London Assembly, 2020). What emerges from this section, is that there is no doubt that value capture is justified, and that there are a number of channels in which value can be captured. Discussing the pros

Investment appraisal 

301

and cons of each channel falls outside the remit of this section, and in any case, these are very likely to be specific to each country or region and the way in which they have been implemented. We now turn our attention to the issue of risk allocation, already advanced in Section 15.2.

15.4 ALLOCATION OF RISK AND OPTIMISM BIAS The appraisal process of transport projects entails the estimation of costs and benefits, and any other impacts that cannot be monetised, or in some cases, quantified. The basic costs of a transport project are the costs before any adjustment for risks and optimism bias (Department for Transport, 2017b). Once the basic costs have been estimated, there needs to be an adjustment for risk and optimism bias. Adjusting for risk entails a number of steps. First, all the factors that could affect the delivery and operation of the project should be identified. Second, the impacts of risk should be assessed. Third, the likelihood of the impacts of risk should be estimated. Fourth, the expected value of risk should be calculated (Department for Transport, 2017b). The expected value of risk can be calculated using multi-point probability analysis, real options analysis, or even Monte Carlo analysis (HM Treasury, 2020). For transport projects in England the most commonly used and recommended method is multi-point probability analysis. Multi-point probability analysis takes the weighted average of the (additional) costs of all possible outcomes, with the weights being the different probabilities of those outcomes occurring (Department for Transport, 2017b). In England, adjusting for risk is required for all transport projects with a base cost greater than a given threshold, which as of 2021 is £5 million in 2010 prices, and needs to take into account the impact of delays and potential cost increases (Department for Transport, 2017b). It is common practice in many countries to record and allocate all the identified risks in a risk register (Department for Transport, 2017b), a standard practice in risk management. Some risks may be transferable through insurance (Department for Transport, 2017b). In general, risk allocation is the process by which the different risks are allocated to the party best able to manage or minimise them. Being responsible for managing or minimising some risks may itself provide an incentive to take measures to mitigate the risk, especially if the risk materialising would translate into financial costs for the party that owns the risk. Risks regarding delays in construction or cost increases, or even problems with the workers or machinery/ equipment, may be allocated to the contractor, as the contractor may be the party best suited to mitigate these risks. Risks associated with funding may be allocated to the relevant local authority/government department promoting the project. Allocating risks to the party best able to manage them is fundamental (Molenaar et al., 2006) because it both minimises the likelihood of them being realised and also ensures that when something goes wrong, there is a party responsible for mitigating them. In countries such as France and the Netherlands, these risks are introduced in the discount rate, as already explained in Section 15.3. In addition to risks, there is a tendency for the sponsor of any project to be optimistically biased, and underestimate costs and project duration and overestimate benefits (HM Treasury, 2020, p. 9; Næss et al., 2015). Therefore, basic costs not only need to be adjusted for risk but also for optimism bias (Flyvbjerg et al., 2004; Oxford Global Projects, 2020).

302  Handbook on transport pricing and financing

These adjustments can be carried out by applying a factor to increase the estimated costs of the project (Flyvbjerg et al., 2004; Oxford Global Projects, 2020). If the project is at a very early stage in the appraisal project, and there is still uncertainty regarding some of the costs, then the factors should be higher than if the project is at a later stage in the appraisal project, and the costs have been refined on the basis of better data (Department for Transport, 2017b). The Transport Appraisal Guidance in the United Kingdom goes as far as providing the uplift factors for road, rail, and other projects at different stages of the appraisal (Department for Transport, 2017b), based on estimates by Flyvbjerg (2004), and more recently updated on the basis of Oxford Global Projects, (2020), which was also led by Bent Flyvbjerg. They also recommend performing a sensitivity analysis at every stage of the life of the project to understand the impact that using lower or higher uplift factors would have on the adjusted costs (Department for Transport, 2017b). The discussion above highlights the importance of acknowledging and adjusting for risk and optimism bias. Allocating risk is also essential, so that if any risk materialises, there is a party that owns it and can manage it. Risk and optimism bias cannot be ignored as doing so could jeopardise and even prevent the completion of a transport project. Having discussed the evolution of appraisal methods through time, along with the current state of the art, with a focus on value capture, risk allocation, and optimism bias, we now turn to appraisal in the context of public acceptability and political influence. As already advanced in the introduction, the results of the appraisal exercise only help, rather than determine, the decision-making process.

15.5 PUBLIC ACCEPTANCE AND POLITICAL INFLUENCE IN INVESTMENT APPRAISAL 15.5.1 The Development of a Strategy In most countries, transport projects are developed in the context of a strategy which sets one or more objectives that are less specific than the problem that an individual scheme is intended to resolve. For example, Highways England’s Roads Investment Strategy is intended to provide “a network that is safe, reliable and efficient for everyone” (Department for Transport, 2020b, p. 1). In Australia, infrastructure investment is guided by the government’s strategy to reduce congestion, better connect the different regions, improve road safety, and meet their freight challenges (Department of Infrastructure, Transport, Regional Development and Communications, 2020). In Spain, it is guided by the strategic vision of “making mobility a right, an element of social cohesion, and of economic growth”, under the “Strategy of Safe, Sustainable and Connected Mobility 2030” (Ministry of Transport, Mobility and Urban Agenda, 2020, p. 5). In Canada, all transport investment is guided by their strategic plan, which includes, amongst other things, “provide greater choice, better service, lower costs …”, “build a safer, more secure transportation system”, “reduce air pollution”, improve the performance of the transport system to get products to markers and grow Canada’s economy (Transport for Canada, 2019). The Estonian government’s Transport and Mobility Masterplan sets sustainability, safety, and mode shift as its key objectives (Ministry of Economic Affairs and Communications, 2020). These are just some examples, but there are many more, as most countries around the world set a transport strategy every few years, which supports and guides investment plans.

Investment appraisal 

303

The overall policy stance of a strategy is determined by broad political considerations. Half a century ago transport policy was largely concerned with the provision of road capacity to accommodate the rapid growth in car ownership and the land-use changes that the spread of car-based mobility facilitated. As the environmental costs of transport and other activities became apparent, the direction of policy changed and the strategy was modified in recognition of these impacts. In many countries, the recognition of cities as the engines of economic growth resulted in a further shift in strategic objectives to ensure that this source of economic benefit formed part of the strategy. Strategies are rarely the subject of an economic appraisal prior to being adopted. They represent a non-negotiable government commitment: the 1997 UK government did not offer people a choice between an integrated transport policy (Department of the Environment, Transport and the Regions, 1997) and some different option, since the policy was part of the programme of change to which it was committed when elected. Moreover, the objectives for a strategy generally command public acceptance since they focus on outcomes such as those for the environment, for the economy or for (human) health, which are uncontroversial and widely recognised as meriting action. But there remains a risk of strategies being aimed at dealing with past problems rather than facing up to the future challenges of environmental degradation or to uncertainty about travel demand. In some cases, a major project becomes a strategy, and this obviously casts some doubt on the efficiency outcome, because such practice essentially enables decision-makers to prioritise projects which meet other policy objectives over those which deliver value for money. Perhaps the best example is the case of England’s HS2 project for a new high-speed railway from London to Birmingham with extensions to Manchester and to Sheffield and Leeds. The failure to articulate clearly and consistently the strategic objectives of the scheme and its effectiveness in meeting those objectives or to deliver economic returns comparable with less ambitious schemes has been a cause of much criticism and has contributed to a lack of public support for the project. 15.5.2 Programmes and Projects – Option Generation and Scheme Selection A strategy is delivered through a programme of interventions, such as a combination of new infrastructure, improved traffic management, and inducements to behavioural change. In most countries, cost–benefit analysis has a key role to play in the selection of the projects that make up the programme. Many of the inputs needed by the engineers responsible for designing the scheme, such as expected traffic flows, are generated by the transport models that underpin and interact with the economic appraisal. The incremental costs of carrying out an economic appraisal are therefore low and its use has become widely accepted and is promulgated in national and international guidance (Grant-Muller et al., 2001; International Transport Forum, 2011; Department of Infrastructure, Transport, Regional Development and Communications, 2021; Department for Transport, 2019). A senior UK official, when interviewed for a study of appraisal’s role, noted that if the Department for Transport’s appraisal methods had not been developed, some other framework to inform decision-makers would have had to have been invented (Worsley and Mackie, 2015). The case for any specific intervention forming part of the programme involves a combination of interests. The economic appraisal provides a technical view of the case, showing the extent to which the project is in the national interest, defined as representing the interest of the

304  Handbook on transport pricing and financing

notional typical taxpayer. But while the values of the outcomes of the scheme, in terms of time savings, safety, and environmental impacts, might represent those of the average citizen, those who are directly affected by the project might be atypical, for example by placing on their local environment a higher value than the respondents to the surveys on which the appraisal values are based. Most countries have experience of public inquiries into transport schemes in which objectors have questioned the basis of the appraisal methods and values or the objectives of the strategy. Political interests, aimed at ensuring the strategy is seen to be a success and that specific local demands are met, have a further influence on the case for a scheme. The decision-making process requires the reconciliation of these conflicting interests. The technocratic solution, whereby the analysts are responsible for selecting schemes ranked according to their benefit–cost ratios until the budget is exhausted, fails to meet the objectives of the other parties. Objectors, if their views become accepted, have influenced both decisions and policy, for example in the shift away from urban road building in the 1970s. In the political arena, the conflict arises between the ministers responsible for the governance of public spending and other politicians who seek to prioritise a different set of local objectives, in addition to those included in the cost–benefit welfare economics framework. 15.5.3 Economic Appraisal and Political Influence A potential for political conflict arises because of the difference in interests between the government department responsible for funding transport investment and the politician responsible for delivering such schemes. The minister responsible for finance has the objective of ensuring the efficient spending of public money: other ministers often seek to maximise their own allocations of these funds. In the United Kingdom, this potential for conflict is managed through a requirement imposed on the senior departmental official to ensure that the department’s use of resources achieves good value for money (HM Treasury, 2019), as set out in the appraisal guidance (Department for Transport, 2017c). In the event of a minister deciding on an option for which the benefits, after accounting for non-monetised impacts (as explained below in the section on value-for-money guidance) fail to exceed the costs, the senior official is required to inform the minister that the decision fails to represent value for money when compared with the alternatives and to notify the Treasury Minister of their minister’s decision. The development of the wider economic impacts appraisal methods (Department for Transport, 2018) provides an example of the policy conflict between the objective of maximising the benefits of a programme of schemes and the wider political considerations of place and equity. The increase in the benefits of a scheme in a dense, high-productivity city when agglomeration impacts were included in the appraisal was greater than in a less productive, less successful urban area. The technocratic solution was to shift investment to these higher productivity conurbations, and the higher tax revenues paid by the more productive workforce would, it was assumed, trickle down to the residents of less productive places. But for governments with an objective of geographical equity, the benefit–cost ratio basis for prioritisation in such cases was politically unacceptable. The 2020 update to HM Treasury’s Green Book (HM Treasury, 2020) extends the guidance to include place-based analysis. In addition to the conventional appraisal, interventions focused on a specific part of the country should identify the local impacts of a scheme in terms of productivity and employment effects after allowance has been made for leakages,

Investment appraisal 

305

substitution, and displacement. Decision-makers should be presented with appraisals of the impacts both at a national and at a local level. Such information should help to inform decisions about priorities which inevitably require political judgement. Evidence of the trade-offs that decision-makers make when approving projects for the programme to be delivered is rarely made public. Surveys of decision-making (Mackie et  al., 2014) suggest that the role of the benefit–cost ratio is rarely decisive, although it helps to ensure that projects with low benefit–cost ratios are sifted out well before the options are seen by ministers. The UK’s Department for Transport is more open than most to publishing a value-formoney framework (Department for Transport, 2017c), which sets out the criteria for including impacts which lack a well-based monetary value or are less certain or difficult to quantify. It recommends the use of switching values to establish the potential magnitude of the unquantified impacts that would be needed to bring the results of the appraisal up to the benefit–cost ratio required of the more typical scheme that ministers approve. Decision-makers are then advised to consider how likely it is that unquantified or uncertain benefits (or costs) of the magnitude needed to bring the scheme up to an acceptable BCR will materialise and, by means of a logic map or other reasoning, to provide the grounds for making such an assumption. Measures of value for money derived from the appraisal of a transport scheme are open to political challenges when such schemes are intended as the agent of transformational change. Politicians at both national and local levels of government claim that the transport appraisal system prevents them from delivering their vision for the community that they serve (see, for example, Transport for the North, 2016). Models of the impact of changes in accessibility on the level and location of economic activity and methods of estimating the associated incremental change in economic welfare are less well established than the estimation of the first-round transport user benefits. There is thus the scope for a greater element of political judgement in decisions on schemes aimed at transformational change. Cost–benefit analysis is not standard practice in all countries. For example, many states in the United States use economic models to estimate the impact on regional GDP. But politicians’ freedom to take decisions without accounting for them is constrained by public opinion and by demands for evidence of their value by the funding authority. The ability to demonstrate that a project is in the national interest provides an incentive to politicians in countries with well-informed voters to adopt cost–benefit analysis as an aid to decision-making. Where external funding is provided, the use of cost–benefit analysis can provide both the national government and the funder with some reassurance about the efficient allocation of resources (International Transport Forum, 2020).

15.6 CONCLUSIONS This chapter covers developments in cost–benefit analysis, its place in the decisions made by government ministers. and discusses some of the links between the economic appraisal and the funding of a project and the risks involved in its delivery. Time savings have formed the main source of the benefits attributed to a scheme and the values are updated from time to time as new evidence emerges. But the increase in land values provides an alternative metric of people’s willingness to pay for at least some of the benefits of a scheme.

306  Handbook on transport pricing and financing

Even if the uplift in land and property values is an incomplete measure, if captured through a tax or a levy, or some other contribution it can help fund the cost of a project and reduce the requirement for more traditional public sector funding. A variety of methods have been used to realise these gains in value, ranging from contributions from developers of property close to the project to a special tax imposed on all properties in the city region served by the new transport project. The scope of environmental and other health-related impacts of transport that are included in economic appraisal has been expanded, with values often shared with government departments responsible for health care and for the environment. More recent developments have been in the assessment of the so-called wider economic impacts, which include the impact of changes in transport-related accessibility on productivity. Identification of the potential risks inherent in a project is an essential part of a well-conducted cost–benefit appraisal. Decision-makers need to know the magnitude of such risks and how best these risks might be allocated. Scheme sponsors tend to adopt a more favourable outlook than experience would suggest is merited and it is practised in some countries and in the case of larger projects to adjust initial estimates to account for optimism bias. In other countries, a risk factor is incorporated into the discount rate. Despite all these methodological developments, most of which can be seen as extensions of basic cost–benefit analysis, project appraisal continues to be a tool that can support rather than determine government decisions. The projects which are the subject of an economic appraisal are generally part of a politically determined set of strategic objectives. In this respect, a democratically elected government can expect a fair degree of public support for the overall strategy and for the welfare economic framework on which the appraisal methods are based. Individual schemes, however, involve trade-offs between the interests of the silent majority, those whose interests are adversely affected, and political and other parties who see local concerns as taking priority over the national interest. The extent to which political influence can play a role in determining priorities depends in part on the influence of the minister responsible for funding transport projects. In the United Kingdom, the Treasury requires ministers in spending departments to deliver value for money, as defined in the country’s appraisal guidance. The concept of value for money extends beyond the impacts from which money values can be derived. But for a project to be approved, the implicit value of any unquantified impact must be of a magnitude sufficient to make this implicit benefit–cost ratio comparable to the ratio typical of other schemes in the programme. Thus, a project with a high BCR may not be funded and a project with a low BCR may be funded if these unquantified but important impacts tip the balance and change the overall value for money. Thus, in the United Kingdom, the result of the most detailed monetised methods can be adjusted by impacts that are only appraised qualitatively, and these can, at least in theory, be appraised within the context of a politically determined set of strategic objectives. This is probably the case in most countries, although in the United Kingdom it is an explicit procedure consistent with the value-for-money guidance. Looking into the future, as we move towards further quantification and monetisation techniques, to include impacts not currently monetised or improve on those that are, it is worth keeping in mind that political decision-making in a democracy, when drawing on public funds, is not solely based on the results of the economic appraisal, but on government strategies, often set out during electoral campaigns, and so supported by voters.

Investment appraisal 

307

REFERENCES Acts of Parliament (1990), Town and Country Planning Act 1990, Section 106. https://www​.legislation​. gov​.uk​/ukpga​/1990​/8​/section​/106 Ahmed, A., Stopher, P. (2014), Seventy minutes plus or minus 10 – A review of travel time budget studies. Transport Reviews, 34(5), 607–625. doi: 10.1080/01441647.2014.946460 American Public Transportation Association (2015), Value Capture for Public Transportation Projects: Examples. https://www​.apta​.com ​/wp​-content ​/uploads​/ Resources​/resources​/rep​orts​andp​ubli​cations​/ Documents​/APTA​-Value​-Capture​-2015​.pdf Anas, A., Liu, Y. (2007), A regional economy, land use, and transportation model (RELU-TRAN): Formulation, algorithm design, and testing. Journal of Regional Science, 47(3), 415–455. Ashley, D. B., J. E. Diekmann, and Molenaar, K. R. (2006). Guide to Risk Assessment and Allocation for Highway Construction Management. Federal Highway Administration, U.S. Department of Transportation, Office of International Programs, Washington, D.C. https://international.fhwa.dot.gov/pubs/pl06032/guide_ to_risk_assessment_allocation_for_highway.pdf Bao, H.X.H., Mok, D. K. C. (2020), A link between east and west: How the Guangzhou-Shenzhen-Hong Kong express rail link affects property prices in Hong Kong. International Real Estate Review, 23(3), 993–1021. Beck, M., Hess, S., Ojeda Cabral, M., Dubernet, I. (2017), Valuing travel time savings: A case of shortterm or long term choices? Transportation Research Part E: Logistics and Transportation Review, 100, 133–143. doi: 10.1016/j.tre.2017.02.001 Boardman, A. E., Greenberg, D. H., Vining, A. R., Weimer, D. L. (2018), Cost-Benefit Analysis: Concepts and Practice, Fifth Edition, Cambridge University Press. Börjesson, M., Jonsson, R. D., Berglund, S., Almström, P. (2014), Land-use impacts in transport appraisal. Research in Transportation Economics, 47, 82–91. Boyce, D.E., Williams, H.C. (2015), Forecasting Urban Travel: Past, Present and Future. Edward Elgar Publishing. Bröcker, J., Mercenier, J. (2011), General equilibrium models for transportation economics. In: de Palma, A., Lindsey, R., Quinet, E., Vickerman, R. (eds.), A Handbook of Transport Economics, Chapter 2, pp. 22–66. Edward Elgar Publishing. Committee on the Medical Effects of Air Pollution (2018), Nitrogen Dioxide: Effects on Mortality. https://www​.gov​.uk ​/government ​/publications​/nitrogen​-dioxide​-effects​-on​-mortality De Jong, G., Kouwenhoven, M. (2020), Value of travel time and travel time reliability. Advances in Transport Policy and Planning, 6, 43–74. [Volume 6: Standard Transport Appraisal Methods, edited by Mouter, N., Elsevier Book Series]. doi: 10.1016/bs​.atpp​.2020​.07​​.010 Department for Business, Energy and Industrial Strategy (2021), Final UK Greenhouse Gas Emissions National Statistics. https://www​.gov​.uk​/government​/collections​/final​-uk​-greenhouse​-gas​-emissions​ -national​-statistics. Department for Communities and Local Government (2016), Appraisal Guide, December 2016. https://www​.gov​.uk ​/government ​/publications​/department​-for​- communities​-and​-local​-government​appraisal​-guide Department for Transport (2013), Transport Analysis Guidance. https://www.gov.uk/guidance/ transport-analysis-guidance-tag Department for Transport (2017a), TAG Unit A1.3: User and Provider Impacts. https://www.gov.uk/ government/publications/webtag-tag-unit-a1-3-user-and-providerimpacts-march-2017 Department for Transport (2017b), Transport Analysis Guidance, TAG Unit A1.2 Scheme Costs, July. https://www.gov.uk/government/publications/webtag-tag-unit-a1-2-schemecosts-july-2017 Department for Transport (2017c), DfT Value for Money Framework. https://www.gov.uk/government/ publications/dft-value-for-money-framework Department for Transport (2018), TAG Unit A2.1: Wider Economic Impacts. www.gov.uk/government/ publications/webtag-tag-unit-a2-1-wider-economic-impacts-may-2018 Department for Transport (2019b), Appraisal and Modelling Strategy. https://www.gov.uk/government/ consultations/transport-appraisal-and-modelling-strategyinforming-future-investment-decisions Department for Transport (2020a), TAG Unit M2.1: Variable Demand Modelling. https://www.gov.uk/ government/publications/tag-unit-m2-1-variable-demand-modelling

308  Handbook on transport pricing and financing

Department for Transport (2020b), Highways England Roads Investment Strategy 2 2020-2025. www​. gov​.uk ​/government ​/publications​/road​-investment​-strategy​-2​-ris2​-2020 ​-to​-2025 Department of the Environment, Transport and the Regions (1997), A New Deal for Transport: Better for Everyone. https://webarchive​.nationalarchives​.gov​.uk​/20100304074834​/ http:/​/www​.dft​.gov​.uk​/ print​/about​/strategy​/whitepapers​/previous​/ane​wdea​lfor​tran​spor​tbet​terfo5695 or https://www​.open​ .edu​/openlearn ​/ocw​/pluginfile​.php​/630978​/mod​_ resource​/content​/1​/new​_deal​_for​_transport​.pdf Department of Infrastructure, Transport, Regional Development and Communications (2020), Infrastructure Investment Program. https://investment​.infrastructure​.gov​.au/ Department of Infrastructure, Transport, Regional Development and Communications (2021), Australian Transport Assessment and Planning. https://www​.atap​.gov​.au/ Drupp, M.A., Freeman, M.C., Groom, B., Nesje, F. (2018), Discounting disentangled. American Economic Journal: Economic Policy, 10(4), 109–134. Eliasson, J., Savemark, C., Franklin, J. (2020). The impact of land use effects in infrastructure appraisal. Transportation Research Part A: Policy and Practice, 141, 262–276. Flyvbjerg, B., Glenting, C., Rønnest, A. (2004), Procedures for Dealing with Optimism Bias in Transport Planning: Guidance Document. London: Department for Transport. https://assets​.publishing​.service​. gov​.uk​/government​/uploads​/system​/uploads​/attachment​_data​/file​/191523​/ Procedures​_for​_dealing​_ with​_optimism​_bias​_in​_transport​_planning​.pdf Freeman, M., Groom, B., Spackman, M. (2018), Social Discount Rates for Cost-Benefit Analysis: A Report for HM Treasury. https://assets​.publishing​.service​.gov​.uk​/government​/uploads​/system​ /uploads​/attachment ​_ data ​/file​/935551​/Social ​_ Discount ​_ Rates ​_ for​_Cost​-Benefit ​_ Analysis ​_ A ​_ Report​_for​_HM​_Treasury​.pdf. Geurs, K. T., van Wee, B., Rietveld, P. (2006), Accessibility appraisal of integrated land-use - transport strategies: methodology and case study for the Netherlands Randstad Area. Environment and Planning B: Planning and Design, 33(5), 639–660. Glaeser, E. L. (1998). Are cities dying? Journal of Economic Perspectives, 12(2), 139–160. Global Projects (2020), Updating the Evidence Behind the Optimism Bias Uplifts for Transport Appraisals. https://www​.gov​.uk ​/government ​/publications​/tag​-updated​-evidence​-for​-optimism​-bias​uplifts Graham, D. J., Gibbons, S., Martin, R. (2009). Transport investment and the distance decay of agglomeration benefits. Report to the Department of Transport. https://personal​.lse​.ac​.uk​/gibbons​/ Papers​/Agglomeration​%20and​%20Distance​%20Decay​%20Jan​%202009​.pdf Grant-Muller, S., Mackie, P., Nellthorp, J, Pearman, A. (2001), Economic appraisal of European transport projects: The state-of-the-art revisited. Transport Reviews, 21(2), 237–261. doi: 10.1080/01441640119423 Greater London Authority (2016), Crossrail Funding: Use of Planning Obligations and the Mayoral Community Infrastructure Levy. https://www​.london​.gov​.uk​/sites​/default​/files​/crossrail​_funding​_ spg​_updated​_march​_2016​_final​.pdf Gupta, A., Van Nieuwerburgh, S., Kontokosta, C. (2020), Take the Q Train: Value Capture of Public Infrastructure Projects, NBER Working Paper No. 26789. www​.nber​.org​/papers​/w26789 Higgins, C., Kanaroglou, P. (2016), Forty years of modelling rapid transit’s land value uplift in North America: Moving beyond the tip of the iceberg. Transport Reviews. doi: 10.1080/01441647.2016. 1174748 HM Treasury (1973), Use of Discounted Cash Flow and the Test Discount Rate in the Public Sector. HM Treasury (2019), Managing Public Money. www​.gov​.uk ​/government ​/publications​/managing​-public​money HM Treasury (2020), The Green Book: Central Government Guidance on Appraisal and Evaluation. https://assets​.publishing​.service​.gov​.uk ​/government ​/uploads​/system ​/uploads​/attachment ​_data ​/file​/ 938046​/ The​_Green​_Book​_2020​.pdf International Transport Forum (2011), Improving the Practice of Transport Project Appraisal, ITF Round Tables, No. 149, OECD Publishing, Paris. doi: 10.1787/9789282103081-en International Transport Forum (2020), The Future of Passenger Mobility and Goods Transport in Estonia: Input Study for the Estonian Transport and Mobility Master Plan, International Transport Forum Policy Papers, No. 78, OECD Publishing, Paris. https://www​.itf​-oecd​.org​/sites​/default​/files​/ docs​/passenger​-mobility​-goods​-transport​-estonia​.pdf Krugman, P. (1998), What’s new about the new economic geography? Oxford Review of Economic Policy, 14(2), 7–17.

Investment appraisal 

309

Laird, J., Venables, A. (2017), Transport investment and economic performance: A framework for project appraisal. Transport Policy, 56, 1–11. Lord, A., Dunning, R., Buck, M., Cantillon, S., Burgess, G., Crook, T., Watkins, C., Whitehead, C. (2020), The Incidence, Value and Delivery of Planning Obligations and Community Infrastructure Levy in England in 2018-19, Ministry of Housing, Communities and Local Government. https:// assets​.publishing​.service​.gov​.uk​/government​/uploads​/system​/uploads​/attachment​_data​/file​/907203​/ The​_Value​_and​_Incidence​_of​_Developer​_Contributions​_in​_ England​_201819​.pdf Mathur, S. (2020a), Value capture to fund public transportation: The impact of warm springs BART station on the value of neighboring residential properties in Fremont. The Journal of Planning Education and Research. doi: 10.1177/0739456X19898737 Mathur, S. (2020b), Impact of transit stations on house prices across entire price spectrum: A quantile regression approach. Land Use Policy, 99. doi: 10.1016/j.landusepol.2020.104828 Mackie, P.J. (1996), Induced traffic and economic appraisal. Transportation, 23(1), 103–119. Mackie, P., Worsley, T., Eliasson, J. (2014), Transport appraisal revisited. Research in Transportation Economics, 47, 3–18. doi: 10.1016/j.retrec.2014.09.013 Mackie, P., Batley, R., Worsley, T. (2018), Valuing transport investments based on travel time savings—A response to David Metz. Case Studies on Transport Policy, 6(4), 638–641. Mayor of London and London Assembly (2020), Paying for Crossrail: Business Rate Supplement. https://www​.london​.gov​.uk​/what​-we​-do​/ business​-and​-economy​/promoting​-london ​/paying​-crossrail​business​-rate​-supplement McCartney, P., Lowe, S., Worsley, T., Hall, K., Mackie, P. (2013), Assessment of Methods for Modelling and Appraisal of the Sub-National, Regional; and Local Economy Impacts of Transport, Report to the Department for Transport, September. https://assets​.publishing​.service​.gov​.uk​/government​/ uploads​/system ​/uploads​/attachment​_data ​/file​/252114​/sub​-national​-impacts​-dft​- 017​.pdf McIntosh, P. T., Quarmby, D. A. (1970), Generalised Costs, and the Estimation of Movement Costs and Benefits in Transport Planning. London: Department of the Environment. https://trid​.trb​.org​/view​/92399. Mehra, R., Prescott, E.C. (1985), The equity premium: A puzzle. Journal of Monetary Economics, 15(2), 145–161. Ministry of Economic Affairs and Communications (2020), Estonia’s Vision for Transport and Mobility (English language summary). www​.mkm​.ee​/en​/news​/estonias​-vision​-transport​-and​-mobility​-more​people​-centred​-greener​-and​-smarter​-infrastructure Ministry of Housing, Communities and Local Government (2020), Community Infrastructure Levy. https://www​.gov​.uk​/guidance​/community​-infrastructure​-levy Ministry of Transport, Mobility and Urban Agenda (Ministro de Transportes, Movilidad y Agenda Urbana) (2020), Estrategia de Movilidad Segura, Sostenible y Conectada 2030, September. https:// cdn​.mitma​.gob​.es​/portal​-web drupa​l /esm​ovili​dad/2​02009​17_Dd​ebate​_(dob​le_p)​vf2​.p​​​df Mouter, N. (2016), Value of travel time: To differentiate or not to differentiate? Transportation Research Record, 2597(1), 82–89. Mouter, N., 2020. Standard Transport Appraisal Methods. Academic Press. Mulley, C., Sampaio, B., Ma, L. (2017), South Eastern busway network in Brisbane, Australia: Value of the network effect. Transportation Research Record: Journal of the Transportation Research Board. doi: 10.3141/2647-06 Mulley, C., Tsai, C.H.P. (2016), When and how much does new transport infrastructure add to property values? Evidence from the bus rapid transit system in Sydney, Australia, Transport Policy, 51, 15–23. National Audit Office (2013), Review of the VFM Process for PFI. https://www​.nao​.org​.uk​/wp​-content​/ uploads​/2014​/01​/ Review​-of​-VFM​-assessment​-process​-for​-PFI1​.pdf Næss, P., Andersen, J., Nicolaisen, M., Strand, A. (2015), Forecasting inaccuracies: A result of unexpected events, optimism bias, technical problems, or strategic misrepresentation? Journal of Transport and Land Use, 8(3), 39–55. https://www​.jstor​.org​/stable​/26189165 NERA (2007), Discount Rates for Rail Safety Scheme Appraisals Final Report for the Office of Rail Regulation. https://www​.orr​.gov​.uk ​/sites​/default ​/files​/om ​/cnsltrep​-NERA​_disc​_ rates​.pdf New Zealand Institute of Economic Research (2013), Appraising Transport Strategies that Induce Land Use Change. https://nzier​.org​.nz​/static​/media​/filer​_public​/c8​/ b3​/c8b3e8c5​-afea​- 464b​-b619​9219fe2f4282​/nzier​_wp2013​- 04_-​_appraising​_transport​_strategies​.pdf Ofgem (2011), Discounting for CBAs Involving Private Investment, but Public Benefit. https://www​. ofgem​.gov​.uk​/sites​/default​/files​/docs​/ 2011​/10​/discounting​-for​- cost​-benefit​-analysis​-involving​private​-investment​-but​-public​-benefit​.pdf

310  Handbook on transport pricing and financing

Office of Management and Budget (2017), Memorandum M-17-21: Guidance Implementing Executive Order 13771, “Reducing Regulation and Controlling Regulatory Costs.” Accessed February 18, 2021. https://www​.whitehouse​.gov​/sites​/whitehouse​.gov​/files​/omb​/memoranda​/2017​/ M​-17​-21​-OMB​.pdf Ortúzar, J.D., Willumsen, L. G. (2011), Modelling Transport. Oxford: John Wiley & Sons. Perk, V., Bovino, S., Catalá, M., Reader, S., Ulloa, S. (2013), Silver line bus rapid transit in Boston, Massachusetts: Impacts on sale prices of condominiums along Washington street. Transportation Research Record: Journal of the Transportation Research Board. doi: 10.3141/2350-09 SACTRA (1999), Transport and the Economy. http://webarchive​.nationalarchives​.gov​.uk​/ 20050301192906​/ http:​/dft​.gov​.uk ​/stellent ​/groups​/dft​_econappr​/documents​/pdf​/dft​_econappr​_pdf​_ 022512​.pdf. Accessed 18/02/2021. Shima, H., Kittrell, K., Ewing, R. (2017), Value of transit as reflected in U.S. single-family home premiums: A meta-analysis. Transportation Research Record 2543. doi: 10.3141/2543-12. Spackman, M. (2020), Social discounting and the cost of public funds: A practitioner’s perspective. Journal of Benefit-Cost Analysis, 11(2), 244–271. Standing Advisory Committee on Trunk Road Assessment (SACTRA) (1994), Trunk Roads and the Generation of Traffic: Report of the Standing Advisory Committee on Trunk Road Assessment. http:// webarchive​.nationalarchives​.gov​.uk/+​/ http:​/www​.dft​.gov​.uk ​/pgr​/economics​/rdg ​/nataarchivedocs​/ trunkroadstraffic​.pdf​.Accessed 18/02/2021. Stopher, P., Ahmed, A., Liu, W. (2017), Travel time budgets: New evidence from multi-year, multi-day data. Transportation, 44, 1069–1082. doi: 10.1007/s11116-016-9694-6 Sugden, R. (1999). Developing a consistent cost-benefit framework for multi-modal transport appraisal. University of East Anglia, UK. Report to Department of Transport. https://research​-portal​.uea​.ac​.uk​/ en ​/publications​/developing​-a​-consistent​-cost​-benefit​-framework​-for​-multi​-modal​-tr Transport Canada (2019), Transportation 2030: A Strategic Plan for the Future of Transportation in Canada - Infographic. https://tc​.canada​.ca ​/en ​/corporate​-services​/transportation​-2030 ​-infographic Transport for the North (2016), DfT Modelling and Appraisal Strategy: Coordinated Northern Response. https://tra​nspo​r tfo​r thenorth​.com ​/reports​/dft​-modelling​-appraisal​-strategy​-co​-ordinated​northern​-response/ World Bank (2015), Developer Exactions and Impact Fees. https://urban​-regeneration​.worldbank​.org​/ node​/14 Worsley, T., Mackie, P. (2015), Transport Policy, Appraisal and Decision-Making, RAC Foundation. www​.racfoundation​.org​/research ​/economy​/transport​-policy​-appraisal​-decision​-making​-research Yen, B., Mulley, C., Zhang, M. (2020), Equity in financing public transport infrastructure: Evaluating funding options. Transport Policy, 95, 68–77. Zhang, M., Liu, Y., 2015. An exploratory analysis of bus rapid transit on property values: a case study of Brisbane’s south east busway. Presented at the State of Australian Cities Conference, Gold Coast, Australia, December 9–11, 2015. http:// soacc​​onfer​​ence.​​com​.a​​u ​/wp-​​conte​​nt​/up​​loads​/​/02/​​​Zhang​​.pdf​ Zhao, Z. J., Larson, K. (2011), Special assessments as a value capture strategy for public transit finance. Public Works Management & Policy, 16(4), 320–340.

16. The regulation of public–private partnerships Eduardo Engel, Ronald Fischer and Alexander Galetovic*

16.1 INTRODUCTION Governments face three challenges in the provision of transport infrastructure: deciding what to build and when, building cost-effectively, and ensuring proper maintenance and service quality once the project is built. This chapter focuses on highways, since they represent the largest fraction by far of worldwide investment in transport PPPs.1 Until recently, highways were considered public goods and were built by the state, with funding from budget appropriations, and were managed by ministries or public agencies. In the last few decades, many countries have introduced public–private partnerships (PPP) as a new alternative to provide roads, bridges, and tunnels, among other types of infrastructure. A PPP bundles finance, construction, and operation into one long-term contract between the government and a standalone firm – the special purpose vehicle (SPV; Figure 16.1, panel a), which we also denote as the concessionaire. The SPV builds and operates a legally and economically self-contained project for periods lasting up to several decades. On the funding side, the PPP contract pledges the cash generated by the infrastructure, which can be generated by tolls or by government payments, to pay back equity investors and debt. The SPV’s narrow focus leaves no room to divert funds to other businesses, and PPP deals are often highly leveraged. On the production side, the SPV hires firms to build, operate, and maintain the facility. When the contract ends, the infrastructure reverts to the government. PPPs can be contrasted to traditional provision, where the government deals directly with financiers, the builder, and the operator (Figure 16.1, panel b). Under traditional provision, the project is financed with public debt and budget appropriations; a government agency hires the builder and then the operator. This basic structure has many variations, often influenced and determined by country, regional, and city laws and institutions. Sometimes the whole process is led by a single public institution (for example, a central government ministry or a city authority), or the tasks are split among many agencies, layers of government, or even within branches of one government institution. Indeed, PPPs often inherit many of its shortcomings from existing infrastructure programs. The promise of PPPs is that private firms will be more efficient than traditional provisions. One reason is that the entire process is integrated into a single package. When finance, design, construction, operation, and maintenance are bundled, the concessionaire will aim for a design and a construction method that reduces life-cycle costs (Hart, 2003). Moreover, the one-to-one match between infrastructure projects and management teams creates focused firms, which substitute for a governance structure determined by the internal organization of the government. In addition, if the state does not guarantee private debt, the project will also face a

* We are sad to announce that our coauthor Alexander Galetovic died after this chapter was finished. We will miss him as a long-time research partner and friend. 311

312  Handbook on transport pricing and financing (a)

PPPs Government

Service contract

Financiers

Debt & equity

Special purpose vehicle (SPV)

Builder Building contract

O&M contract

(b)

Operator

Conventional provision Builder

Financiers

Public debt & budget appropriations

Government

Building contract

O&M contract

Operator

Figure 16.1  Contracting under PPPs and conventional provision market viability test. Moreover, PPPs can be funded with tolls, which can be used to optimally manage congestion and externalities.2 We argue in this chapter that the promise of PPPs cannot be realized without skillful government planning and regulation. PPPs are a means to procure public infrastructure, and thus governments must still plan and decide which projects to build and when. For PPPs to work, governments must impose and enforce contractual quality and performance standards. Finally, competition among roads is not sufficient to prevent private firms from exploiting market power, which governments must therefore regulate. The optimal PPP contract allows the government to toll optimally and to adjust tolls over time, while remunerating the PPP project.

16.2 THE CONTRACT 16.2.1 BOT Contracts 16.2.1.1 Public and private contracting Governments regulate PPPs through their contracts with concessionaires. The standard is a build, operate, and transfer (BOT) contract. Under this contract, the concessionaire builds the

The regulation of public–private partnerships  313

infrastructure, and then operates and maintains it for a limited term (typically between 10 and 30 years). The project then returns to the state. In the case of DBOT contracts, the government asks the concessionaire to design the project. As shown in Figure 16.2, in a typical PPP contract the concessionaire forms a single purpose vehicle, an independent company with its own balance sheet and whose sole purpose is to undertake the project. On the one hand, the SPV contracts with the government under public law, assuming the responsibility of delivering the infrastructure and its services. In the contract, the government imposes a regulatory and incentive scheme that forces the concessionaire to fulfill minimum quality standards. This scheme sometimes includes a conflict resolution mechanism. On the other hand, the SPV is a private specialized firm whose contracting relations with employees, other firms and financiers are governed by private law. This improves incentives, because during the term of the PPP the concessionaire can manage the infrastructure as a private entity. By contrast, management practices in government are less flexible, and public agencies are constrained by annual budgets. A public manager cannot use the earnings of her organization to reward employee’s performance nor freely allocate factors of production. Regulations imposed by the legislature and the administration constrain hiring, purchasing, contracting and organizational structures.3 Moreover, by creating a specialized firm, the SPV’s scope is clearly defined and bounded, and the project gets a dedicated management team, which answers to the SPV’s board. By

Construction contractor

Sponsors

Debt holders

Equity finance

Building contract

Contract enforcement

Procuring authority

Service fees & subsidies

Debt finance

Special Purpose Vehicle (SPV) Insurance companies

Rating agencies

User fees

Debt insurance

Debt rating

Users of the infrastructure and service

Service contract

O&M contractor

Service & quality delivered

Source:   Engel et al., 2014, op. cit.

Figure 16.2  PPPs as a web of contracts (from: Engel et al., 2014, op​.ci​t.)

314  Handbook on transport pricing and financing

contrast, public agencies in charge of infrastructure projects (e.g., ministries of public works) tend to have multiple objectives and are accountable to multiple principals. Moreover, a ministry or public agency manages a large number of tasks and projects. Scale and scope are likely to be well beyond what is efficient, which blunts incentives. Note that PPPs have finite terms – the infrastructure reverts to the government in finite time by design. While finite terms are not necessary for PPPs (for example, privatized utilities can be thought of as non-ending PPPs), they allow governments to keep control over planning. This is important because network effects may appear over time and these require government planning and authority to execute the projects. More generally, a PPP is not a privatization but a different way of procuring public infrastructure. 16.2.1.2 Bundling A PPP bundles financing, construction, operation, and maintenance into a single contract. Bundling differentiates PPPs from traditional provision and is a key feature of the PPP model. As Hart (2003) shows, when building the project, the concessionaire considers the total life-cycle costs of the project. Moreover, bundling allows the firm to propose a design with features that increase demand in a cost-efficient way. However, as mentioned by Hart (2003), these advantages can be at the expense of quality. Iossa and Martimort (2012) show that if the design and operating efforts of an infrastructure PPP are both contractible, and all shocks are perfectly insurable, then bundling or unbundling is immaterial. However, when there is incomplete contracting there is a difference: if productivity (say traffic demand or operations and maintenance (O&M) costs) is observed ex post, bundling is preferable so long as uncertainty is limited. Iossa and Martimort (2015a) show that when service or infrastructure quality and operating costs are contractible and project externalities are positive, the preference for bundling continues to hold.4 Highways (and tunnels and bridges) have the advantage that quality standards can be contracted and verified easily. This leads to one of the main advantages of highway PPPs: the fact that continuous maintenance is ensured, in contrast to the costlier stop-and-go maintenance of road infrastructure in most countries.5 Another advantage is that user fees (tolls) can reduce the congestion externality while collecting revenue that pays for all or part of the project. The fact that many European countries’ highways have availability payments is a political compromise in response to opposition to tolls, rather than an efficient choice. 16.2.1.3 Funding Just like traditional projects, PPPs can be funded with budget appropriations or tolls. The standard budget-funded PPP is an availability contract where the government pays the SPV for capital (CAPEX) and operating expenses (OPEX) for a fixed number of years, subject to the SPV meeting the contractual quality standards and keeping the infrastructure available. For example, availability contracts have been used to procure roads in the United Kingdom (see Villalba-Romero and Liyanage, 2016) and the United States (see Poole, 2017). Some governments have used shadow tolls, a per-vehicle piece rate that emulates standard tolls, but which are paid by the government, not by drivers. Shadow tolls were used on many roads in the United Kingdom (see Villalba-Romero, and Liyanage, 2016) and in Portugal (see Sarmento and Reneboog, 2015). Note that while both availability payments and shadow tolls are funded from the budget, shadow tolls create demand risk both for the concessionaire and the budget (see Engel, Fischer

The regulation of public–private partnerships  315

and Galetovic, 2014, Table 5.1). By contrast, an availability contract creates a fixed and largely riskless stream of government disbursements for the concessionaire. Many PPPs are funded with tolls collected by the concessionaire. In most cases, toll-funded concessions are fixed-term, which implies that the concessionaire bears substantial exogenous demand risk. An alternative is present-value-of-revenue (PVR) contracts. In a PVR-type PPP, the concessionaire asks for a predefined revenue in present-value terms, and the contract lasts until the amount is collected. The term is shorter when demand is high and longer when demand is low, which transfers demand risk to the government/user.6 PVR contracts were first used in the mid-1980s in the United Kingdom to concession the Second Severn Crossing and the Queen Elizabeth II bridges. More recently, Chile has auctioned PVR contracts worth several billion dollars (Engel et al., 2021).7 16.2.1.4 Risk allocation The basic principle of risk allocation is to assign each risk factor to the party best able to manage it. Irwin (2007) applies this principle to infrastructure projects and focuses on three dimensions of “risk management”: i. Influencing the risk factor. ii. Reducing the sensitivity of the project’s value to the risk. iii. Bearing the risk. This means that, in general, construction cost risk should be assigned to the private party, except in those cases, such as tunnels, in which the costs risks are too high and uncontrollable (because of geology). In those cases, the government usually bears the excess risk. The private party also is assigned operating and maintenance cost risks, and availability and performance risks. Regarding policy risks, the distinction between general policies and those aimed specifically at the project is important. In the first case, which includes the interest rate and currency devaluations, the risk should be borne by concessionaires as is the case for firms in other sectors. In contrast, specific policy measures that affect the costs of the project, such as an increase in safety standards or new environmental requirements, should be borne by the public. Similarly, regulatory risk taking should also fall on the public. In the case of demand risk in tolled PPPs, the risk is usually large and exogenous (given by macro or regional economic factors). It should be assigned to the public, as tolls represent a small proportion of national income, in contrast to its proportion of the revenues for the SPV. Finally, the finance risk should fall on the private party. 16.2.1.5 Awarding the contract PPPs are awarded in competitive auctions or direct assignments, perhaps after a negotiated procedure. In general, direct awards are associated with a higher frequency of contract renegotiation and corruption. Knack, Biletska, and Kacker (2017), using data from 88 countries, show that open auctions reduce the size of bribes and increase participation. Tran (2008) analyzes internal information from a bribe-paying firm and finds that there is less corruption when bidding is competitive, and contracts are awarded following transparent criteria such as the lowest price. Huang (2019) shows that subjective scoring functions, which sometimes include subjective measures of quality, foster corruption.

316  Handbook on transport pricing and financing

These results are consistent with Bosio, Djankov, Glaeser, and Shleifer (2020), who show that in well-governed countries discretion in procurement is to be preferred. In contrast, in countries with weaker governance, less discretion reduces corruption and leads to more efficient outcomes in procurement. Thus, in countries with little corruption, contracts can be awarded on the basis of subjective measures of quality. For example, in New Zealand the government decides the amount of funding for a PPP availability project and an expert panel chooses the best design. Auctions must be tailored to the type of contract. In the case of availability contracts, the government sets the service standards that the infrastructure must meet, and firms compete on the lowest annual payment. When the PPP is funded with tolls, and the term is fixed, the government sets service standards and firms may bid on the toll schedule or the length of the contract term. As a final option, firms bid on the present value of revenues and the contract lasts until the winning bid is collected. 16.2.1.6 Renegotiations It is well known that PPPs tend to be routinely renegotiated. Many renegotiations occur while projects are under construction or even just after signing. An early influential book by Guasch (2004) studied almost 1,000 PPP projects in Latin America. He showed that more than half (54.4%) of the projects in the transport sector had been renegotiated at least once. More recently, Engel, Fischer, and Galetovic (2019) examined 59 PPP highways in Chile, Peru, and Colombia. There had been 535 contract renegotiations, leading to yearly increases over the initial planned investment of 9.5% in Colombia, 3.6% in Peru, and 1.3% in Chile.8 This resulted in accumulated renegotiations of 85.1% of initial investments in Colombia, 13.7% in Peru, and 16.5% in Chile. Renegotiations generate incentives for bidders to lowball their offers in the expectation of profits in future renegotiations. In turn, this leads to an “adverse selection” problem, as it provides an advantage at the auction to firms that have a comparative advantage in renegotiation vis-a-vis firms that are relatively better in technical aspects. Moreover, renegotiations of construction contracts (including PPPs) are intimately linked to bribes and corruption, as shown by the Odebrecht corruption case in Campos et al. (2021). Renegotiations of Odebrecht contracts where there was no evidence of bribe payments amounted to 5.9% of the initial costs, compared with 70.8% for contracts where there was evidence of bribe payments. An additional driver is that incumbents use renegotiations to elude spending limits, burdening future administrations and enhancing their reelection probabilities, as shown in Engel, Fischer, and Galetovic (2019). Aguirre (2015) found that Peruvian transport PPPs are renegotiated with higher frequency during election years. There are two types of contract renegotiations, in the terminology of Laffont and Martimort (2002). There is renegotiation when both parties agree to renegotiate; and breach of contract, when either party breaks the rules of the contract to its advantage. In the case of PPPs, this distinction is often not clear. The reason is that the government may respond to political aims that diverge from the welfare of society, as analyzed by Engel, Fischer, and Galetovic (2019). The government may break the contract either to expropriate the concessionaire or to benefit it at the expense of the public. Alternatively, renegotiation can improve the profitability of projects for the concessionaire ex post (see Menezes and Ryan, 2015; Beuve et al., 2018). Renegotiations are possible during

The regulation of public–private partnerships  317

the construction phase of a traditional project, but a PPP contract adds the possibility of renegotiation at later stages. The firm may pressure the government to change the terms of the contract to its benefit when demand is lower than predicted or when O&M costs are high. If the concessionaire can pressure the government to renegotiate contracts, the advantages of PPPs may disappear. A firm need not design a project to reduce life-cycle costs or be careful about construction costs nor in the estimation of the profitability of the project if it can renegotiate away its mistakes. Indeed, Valero (2015) shows that the ability of governments to pre-commit to a long-term contract is necessary to ensure the realization of the efficiency gains that PPPs promise (see also Spiller, 2013). Thus, renegotiations are at odds with the efficient regulation of PPPs. 16.2.1.7 Unsolicited proposals Unsolicited proposals are schemes for new PPP projects brought to the attention of the Public Works Authority (PWA). Since they involve a measure of Intellectual Property (IP), the proponents of successful proposals should be rewarded. Often it is firms specialized in PPPs that make these proposals, which need to be assessed, while subject to lobbying by the firm. It is necessary to have a system to filter unsolicited proposals so that only projects that are novel, useful, and non-obvious are accepted. The rewards usually take the form of an advantage in the bidding scores for awarding the project, and this can dissuade competition. A better alternative (see Engel et al., 2014) is to have a periodic competition for projects, in which the PWA chooses the proposals it wants to include in its plans and rewards the proponents with a fixed prize. The PWA becomes the owner of the project, eliminating the IP problems of unsolicited proposals. 16.2.2 Financing of PPPs 16.2.2.1 Project finance As shown in Figure 16.2, a key component of a PPP is the relation between the SPV and financiers. Project sponsors invest equity and debt holders lend to the SPV. Because the infrastructure has no alternative use, both provide finance against the cash flows of the project, rather than against the assets of the SPV – a financing technique known as project finance. Lending against cash flows is feasible because the SPV’s focus is narrow, and funds cannot be diverted toward other uses (see Yescombe, 2007; Engel et al., 2014, ch. 5). In their evaluation, financiers focus on the riskiness and the potential profitability of the project. A PPP project normally involves a large initial investment and a payback period of decades. The initial investment requires management capabilities to subcontract and supervise the contracts with building companies (see Figure 16.2).9 Once the project has been built, the SPV contracts with firms that operate and maintain the infrastructure and ensures that the service contract with the government is fulfilled. Once the facility is built, the PPP is often sold to institutional investors, or the short-term construction loans are converted to long-term bonds and loans. 16.2.2.2 Leverage An interesting issue with PPPs is the leverage that is feasible, which depends on how much risk the SPV bears. In the case of an availability contract with annual CAPEX and O&M payments, the only risks are construction risks, O&M risks, and availability risks, i.e., that the project

318  Handbook on transport pricing and financing

delivers services complying with the standards of the PPP contract. The SPV can subcontract construction and O&M with firms that provide full guarantees. The periodic payments from the state pay for construction debt, O&M, and the profits of the SPV. Since construction and O&M costs are guaranteed there is no risk, except for contractual risk. This means that the SPV can obtain financing with little equity. If the PPP contract is a PVR contract, the reduction in demand risk also allows for high leverage. In fact, the Queen Elizabeth II and Second Severn bridges were fully financed with debt and no equity (ITF, 2013). On the other hand, loading risks on the PPP reduces the leverage attainable by the SPV, but can provide incentives for effort. For example, in a sample of 124 transport PPPs in Latin America, Moore, Straub, and Dethier (2014) find that leverage depends on the power of the incentive contract – strong incentives are associated with lower leverage. 16.2.2.3 Risk and guarantees When remuneration is based on availability payments there is no demand risk, though there is a residual risk associated with failing to comply with the agreed service quality. When the PPP is funded by toll revenue and the term is fixed, the government often provides a minimum income guarantee to make the project bankable. Otherwise, demand risk may make it impossible for the concessionaire to obtain loans at a reasonable cost. 16.2.2.4 Valuation of PPPs As a private investment, prospective investors, and lenders need a means of valuing PPP projects. Given that PPP projects confront several risks, including construction cost risk, political risk, and demand risk, a straightforward net present-value (NPV) approach is inappropriate. Two approaches have been used: calculating the PPP’s Value at Risk (VaR) and a real options approach. A VaR calculates the distribution of the NPV of the project through Monte Carlo simulations of the distributions of the key variables. The VaR is the lowest NPV that occurs with a frequency higher than the given confidence level, for instance, 95%. This approach was used to value PPPs used in the 1990s and early 2000s (see Ye and Tiong, 2000), but has become less common since the financial crisis of 2008, since it ignores outcomes deemed as unlikely. Real options were pioneered by Brennan and Schwartz (1985). In general, this approach consists of adding the real options value component associated with different characteristics of the PPP contract, see Trigeorgis (2005). The literature has concentrated on the valuation of aspects such as management options, minimum traffic guarantees (Lara Galera and Sánchez Soliño, 2010; Brandao and Saraiva, 2008), exchange rate guarantees, concession extensions, toll adjustments, contract renegotiation, road expansion (Krüger, 2012), competition restrictions (Liu et al, 2014), and others (Buyukyoran and Gundes, 2017; Power at al., 2016; Lu et al., 2017). 16.2.3 PPP or Traditional Provision? A precondition for successful PPPs, as with infrastructure provision in general, is proper government planning, skilled project design, and efficient project procurement. It is important to note that planning and project design are government tasks under both PPPs and traditional provisions. Therefore, a key decision is whether to procure a project as a PPP or with traditional provision.

The regulation of public–private partnerships  319

Many countries use value-for-money tests to make this decision. These tests compare costs under both types of procurement. Ideally, the analyst would estimate the costs of the project under each option and choose the least-cost solution. In practice, governments issue guidelines that are often – as in the case of the United Kingdom – politically motivated.10 This implies that whether a project is procured in the traditional way or as a PPP will often depend on the policy cycle.11 An additional shortcoming of value-for-money tests is that they eschew quantification of the benefits wrought by the project, which may vary with the type of procurement. For example, if PPPs provide better incentives for highway maintenance than public provision, but these incentives are not incorporated in the comparison, a value-for-money approach will be biased against the PPP option. An alternative to value-for-money tests is to classify infrastructure into broad classes –e.g., airports, seaports, highways – taking into account the incentives created by the different types of infrastructure contracts. Next, one can examine whether the benefits of bundling dominate the costs of PPPs for a particular class of infrastructure assets. Ideally, the best organizational form for providing a specific infrastructure should be determined by the project’s physical and economic characteristics.12 Whether a PPP or traditional procurement is better may depend on the institutional capabilities of a country, so different types of infrastructure may be preferred as PPPs in different countries.13 This has been the approach, for example, in Chile, where airports, seaports, and toll roads are routinely procured as PPPs. That said, the case for highway PPPs seems particularly compelling. First, and as mentioned above, incentives to maintain are inadequate under traditional provisions, since it may take years for the effects of insufficient maintenance to emerge. Moreover, repairing a very deteriorated road has political visibility, providing another incentive to lower the allocation of resources to routine maintenance. This leads to a stop-and-go approach to highway maintenance in most developing and some developed countries, which raises maintenance costs (as mentioned before, it can triple the cost of timely maintenance, according to some estimates) and lowers average quality.14 On the other hand, if the PPP contract specifies quality standards, the concessionaire has strong incentives to maintain the road because routine maintenance is so much less costly than recovering a deteriorated road. Bundling also provides incentives for the concessionaire to design the project to minimize life-cycle costs.

16.3 OPTIMAL CONTRACTING PPPs are regulated via contract, and theory offers guidance about appropriate contractual structures. The following approach to the characteristics of the optimal PPP contract is derived from Engel, Fischer, and Galetovic (2001, 2013). Our model assumes that the main driver of risk is exogenous demand for the project (see Trujillo, Quinet and Estache, 2002), and that construction, operational, and maintenance cost risks, while important, are under the control of the private party and should be assigned to the private party. 16.3.1 A Simple Model Assume that a risk-neutral benevolent social planner hires a concessionaire to finance, build, and operate a road that costs I > 0. For simplicity, we ignore maintenance and operation costs and assume that the up-front investment does not depreciate. The concessionaire is a

320  Handbook on transport pricing and financing

risk-averse expected utility maximizer, with preferences represented by the increasing, strictly concave utility function u. Demand uncertainty is summarized by a probability density over the present value of toll revenue that the infrastructure can generate over its entire lifetime, f(v). This density is common knowledge and is bounded from below by vmin and from above by vmax. For simplicity, we assume that demand can finance the road in all demand scenarios, that is, that vmin ³ I .15 We assume that v is equal to the discounted private willingness to pay for the project’s services (independent of toll and congestion in this section but see Section 16.3.2 for other cases). The PPP can be funded with tolls and subsidies. Let R(v) be the present value of toll revenues received by the concessionaire during the term of the PPP in state v, and let S(v) be the present values of subsidies paid by the government in state v. Further, let PS(v) be producer surplus in state v and CS(v) be consumer surplus in state v. The planner then chooses R(v) and S(v) to solve the following program:

ì í î

ü

ò éëCS ( v ) + PS ( v )ùû f ( v ) dv ýþ

Subject to the participation constraint:

ò u éëPS ( v )ùû f ( v ) dv ³ u ( 0 ) (16.1)

with

PS ( v ) = R ( v ) + S ( v ) - I



0 £ v £ R ( v )

Moreover, when choosing R(v) and S(v), the planner has two concerns. One is that subsidies must be funded with distortionary taxation. Thus, we let 1 + λ > 1 be the cost of public funds. Second, subsidy funding is affected by red tape, so we assume that achieving 1 dollar of useful spending costs 1 dollar if funded with tolls collected directly by the concessionaire, but 1 + ζ dollars if funded with a subsidy. If subsidies are monetary transfers from the government to the concessionaire, then ζ > 0 means resources are wasted in the process, perhaps because of agency problems facing the budgetary authority when monitoring the government agency in charge of spending. The concessionaire receives R(v) ≤ v, so the government can use v−R(v) to reduce distortionary taxation. Then:



CS ( v ) = éë v - R ( v ) - (1 + l ) (1 + z ) S ( v ) ùû + l éë v - R ( v ) ùû = (1 + l ) éë v - R ( v ) - (1 + z ) S ( v ) ùû .



Replacing these last two equations into the planner’s objective function yields an equivalent program where the planner minimizes:

The regulation of public–private partnerships  321



ò {lR ( v ) + (1 + l ) (1 + z ) S ( v )} f ( v ) dv

Subject to the constraints:



ò u éë R ( v ) + S ( v ) - I ùû f ( v ) dv ³ u ( 0 ) , 0 £ R ( v ) £ v, S ( v ) ³ 0.



In the particular case where ζ = 0, it is apparent that the optimal contract is independent from λ, the cost of public funds. This might be surprising, but it follows from the fact that toll revenue can substitute for distortionary taxes in the government’s budget. That is, at the margin any dollar transferred to the concessionaire, be it via tolls or funded by taxes, costs 1 + λ. It is also apparent that in the general case with ζ > 0, tolls are a more efficient means of compensating the concessionaire than subsidies, hence S(v) = 0 for all v. The reason is that the cost to society of 1 dollar in tolls is 1 + λ, while a subsidy costs (1 + λ) (1 + ζ).16 Hence the optimal contract is funded with tolls. Last, note that R(v) = I for all v is feasible, satisfies the participation constraint with equality, and, because u is concave, minimizes the expected cost of meeting constraint (PC). Thus, the optimal contract provides full insurance to the concessionaire. One implication is that the optimum can be implemented with a present-value-of-revenue contract in an auction. Under such a contract, the term is variable and adjusts to demand realizations – with higher demand the term is shorter. Firms bid the present value of revenue and the lowest bid wins the PPP. The term of the PPP ends when the concessionaire collects the present value of revenues that it bid in the auction. Another implication is that a standard fixed-term contract is inefficient. The problem is that the concessionaire is made to bear risk (of too much or too little revenue, depending on the realization of v) and that does not satisfy the conditions for efficient risk sharing. These conditions are that risk should be transferred to those that can manage it or best bear it. This inefficient transfer of risk creates a cost that, in the case of roads in high demand, is shown in Engel, Fischer, and Galetovic (2001) to be:

æ CV A / 2 çç è 1 - CV A / 2

ö ÷÷ × I , ø

where CV is the coefficient of variation of f(v) and A is the coefficient of relative risk aversion of the concessionaire. 16.3.2 Tolling and the Optimal Contract Thus far we have assumed that v is a fixed quantity in each state. In practice, revenues, usage of the highway, and congestion vary with the toll. Moreover, as is well known, tolls play a

322  Handbook on transport pricing and financing

key allocative role in managing congestion. The remarkable result is that under the optimal contract the government always charges the optimal toll, both during the PPP and after it ends. Moreover, at the margin, tolling substitutes for distortionary taxation, and the optimal toll is adjusted accordingly. Under the optimal contract, tolling is separable from funding – there is no tension between funding and tolling. It follows that governments can choose their tolling policy independently of whether the highway is procured under traditional provision or with a PPP. This can be seen in a simple model based on Engel, Fischer, and Galetovic (2007). If the level of tolls is a choice variable, the present value of toll revenue depends on the toll being charged. Thus, assume that the toll p is charged both during and after the PPP. We denote by q(p, θ) the demand for the highway in state θ when the toll is p, by g the probability density of θ, and by k(q) the private cost of using the highway when q vehicles are using it. Let r be the discount rate and assume that the term of the PPP is T(θ). Then the present value of toll revenue generated during the PPP is:

ò



T ( q) 0

(

p × q ( p, q ) × e -rt dt = 1 - e

- rT ( q )

) p × q (r p, q)

= L ( r , q ) P ( p, q )

(

where L º 1 - e

- rT ( q )

) and P º p × q ( pr, q) .

Next, let CS(p, θ) denote the discounted consumer surplus if the toll is p in state θ. and assume that the planner gives weight η  0 is not a sufficient argument against subsidizing projects, for the project’s social value may exceed I, and toll revenue may be insufficient to compensate the concessionaire in low-demand states. See Engel, Fischer and Galetovic (2013). This comparison considers all highway PPPs with at least eight years of operation, see Engel et al. (2021). The numbers remain similar if airport PPPs are also included. However, it is not clear if this is the optimal risk assignment. Sharing some risk between the government and the concessionaire may well be a more efficient allocation, though difficult to implement in practice. As mentioned above, a shadow toll contract is one in which the Public Authority pays a specified amount per user of the PPP project. See Verhoef and Mohring (2007) for a restatement of the result and a synthesis and a review of the literature. The UK provides an example of a program that lost public support because of a combination of excessive promises and some very public failures, such as the underground PPPs (NAO, 2018).

REFERENCES Aguirre, J., “Electoral Cycles and Renegotiations of Transport Infrastructure in Concession Contracts,” in Arnold Picot, M.F., Grove, N., eds., The Economics of Infrastructure Provision. CES-Ifo Seminar Series. MIT Press, 2015. Arnott, R., A. de Palma, and R. Lindsey, “A Structural Model of Peak-Period Congestion: A Traffic Bottleneck with Elastic Demand,” American Economic Review, 1993, 83(1), 161–179. Arnott, R., and M. Kraus, “When Are Anonymous Congestion Charges Consistent with Marginal Cost Pricing?,” Journal of Public Economics, 1998, 67(1), 45–64. Beuve, Jean, Aude Le Lannier, and Zoé Le Squeren, “Renegotiating PPP Contracts: Opportunities and Pitfalls,” in Julie de Brux and Stéphane Saussier, eds., The Economics of Public-Private Partnerships: Theoretical and Empirical Developments. Springer, 2018, pp. 135–162. Ball, Rob, Maryanne Heafey, and Dave King, “The Private Finance Initiative in the UK,” Public Management Review, 2007, 9(2), 289–310. Bitran, Eduardo, Sebastián Nieto-Parra, and Juan Sebastián Robledo, “Opening the Black Box of Contract Renegotiations: An Analysis of Road Concessions in Chile, Colombia and Peru,” Working Paper 317, OCDE April 2013. Bosio, Erica, Simon Djankov, Edward L. Glaeser, and Andrei Shleifer, “Public Procurement in Law and Practice,” Working Paper 27188, NBER 2020. Brandao, Luiz Eduardo T., and Eduardo Saraiva, “The Option Value of Government Guarantees in Infrastructure Projects,” Construction Management and Economics, 2008, 26(11), 1171–1180.

328  Handbook on transport pricing and financing

Brennan, Michael, and Eduardo S. Schwartz, “Evaluating Natural Resource Investments,” The Journal of Business, 1985, 58 (2), 135–157. Buyukyoran, Faruk, and Selin Gundes, “Optimized Real Options-based Approach for Government Guarantees in PPP Toll Road Projects,” Construction Management and Economics, 2017, 36(4), 203–216. Campos, Nicolás, Eduardo Engel, Ronald Fischer, and Alexander Galetovic, “Renegotiations and Corruption in Infrastructure: The Odebrecht Case,” Journal of Economic Perspectives, 2021. De Vany, Arthur, and Thomas R. Saving, “Competition and Highway Pricing for Stochastic Traffic,” The Journal of Business, 1980, 53(1), 45–60. Engel, Eduardo, Ronald Fischer, and Alexander Galetovic, “Least-Present-Value-of Revenue Auctions and Highway Franchising,” Journal of Political Economy, 2001, 109(5), 993–1020. Engel, Eduardo, Ronald D. Fischer, and Alexander Galetovic, “The Basic Public Finance of PublicPrivate Partnerships,” Journal of the European Economic Association, February 2013, 11(1), 83–111. Engel, Eduardo, Ronald D. Fischer, and Alexander Galetovic, “The Basic Public Finance of PublicPrivate Partnerships,” Working paper 13284, NBER, July 2007 Engel, Eduardo, Ronald D. Fischer, and Alexander Galetovic, Public-Private Partnerships: A Basic Guide. Cambridge University Press, 2014. Engel, Eduardo, Ronald D. Fischer, and Alexander Galetovic, “The Joy of Flying: Efficient Airport PPP Contracts,” Transportation Research Part B: Methodological, 2018, 114, 131–146. Engel, Eduardo, Ronald D. Fischer, and Alexander Galetovic, “Soft Budgets and Renegotiation in Public-Private Partnerships,” Economics of Transportation, August 2019, 17, 40–50. Engel, Eduardo, Ronald D. Fischer, and Alexander Galetovic, “When and How to Use Public-Private Partnerships in Infrastructure: Lessons from the International Experience,” in Edward L. Glaeser and James M. Poterba, eds., Economics of Infrastructure Investment. University of Chicago Press, 2021. Feng Xiao, Hai Yang, and Deren Han, “Competition and efficiency of private toll roads,” Transportation Research Part B: Methodological, 2007, 41, 292–308. Guasch, José Luis, Granting and Renegotiating Infrastructure Concessions: Doing It Right, Washington, DC: The World Bank, 2004. Guasch, Jose Luis, Jean Jacques Laffont, and Stephane Straub, “Concessions of Infrastructure in Latin America: Government-led Renegotiations,” Journal of Applied Econometrics, 2007, 22(10), 1267–1294. Guasch, Jose Luis, Jean Jacques Laffont, and Stephane Straub, “Renegotiation of Concession Contracts in Latin America,” International Journal of Industrial Organization, 2008, 26, 421–442. Grout, Paul, “Value-for-Money Measurements in Public Private Partnerships,” EIB Papers, European Investment Bank (EIB), Luxembourg, 2005, 10(2), 33–56. Hart, Oliver, “Incomplete Contracts and Public Ownership: Remarks and an Application to PublicPrivate Partnerships,” Economic Journal, 2003, 113, C69–C76. Hodge, Graeme A., Carsten Greve and Anthony E. Boardman, International Handbook on PublicPrivate Partnerships, Edward Elgar, UK, 2010. Huang, Yangguang, and Jijun Xia, “Procurement Auctions under Quality Manipulation Corruption,” European Economic Review, 2019, 111, 380–399. Krüger, Niclas A., “To Kill a Real Option – Incomplete Contracts, Real Options and PPP,” Transportation Research Part A, 2012, 46, 1359–1371. International Transport Forum, “Better Regulation of Public–Private Partnerships for Transport Infrastructure,” 2013. Iossa, Elisabetta, and David Martimort, “Risk Allocation and the Costs and Benefits of Public-Private Partnerships,” RAND Journal of Economics, 2012, 43, 442–474. Iossa, Elisabetta, and David Martimort, “The Simple Microeconomics of Public-Private Partnerships,” Journal of Public Economic Theory, 2015, 17(1), 4–48. Irwin, Timothy C., Government Guarantees: Allocating and Valuing Risk in Privately Financed Infrastructure Projects. The World Bank, Direction is Development 39497, 2007. Knack, Stephen, Nataliya Biletska, and Kanishka Kacker, “Deterring Kickbacks and Encouraging Entry in Public Procurement Markets: Evidence from Firm Surveys in 88 Developing Countries,” Policy Research Working Paper 8078, The World Bank, Washington, DC, 2017. Laffont, Jean-Jaques and David Martimort, The Theory of Incentives: The Principal Agent Model. Princeton University Press, 2002.

The regulation of public–private partnerships  329

Lara Galera, Antonio L., and Antonio Sánchez Soliño, “A Real Options Approach for the Valuation of Highway Concessions,” Transportation Science, 2010, 44(3), 416–427. Liu, Jicai, Xibing Yu, Charles Yuen, and Jen Cheah, “Evaluation of Restrictive Competition in PPP Projects Using Real Option Approach,” International Journal of Project Management, 2014, 3(2), 473–481. Lu, Zhaoyang and Qiang Meng, “Analysis of Optimal BOT Highway Capacity and Economic Toll Adjustment Provisions under Traffic Demand Uncertainty,” Transportation Research Part E: Logistics and Transportation Review, 2017. Menezes, F., and M. Ryan, “Default and Renegotiation in Public-Private Partnership Auctions,” Journal of Public Economic Theory, 2015, 17, 49–77. Mohring, Herbert, and Mitchell Harwitz, Highway benefits: An analytical framework, Evanston, IL: Northwestern University Press, 1962. Moore, Alexander, Stéphane Straub, and Jean-Jacques Dethier, “Regulation, Renegotiation and Capital Structure: Theory and Evidence from Latin American Transport Concessions,” Journal of Regulatory Economics, 2014, 45, 209–232. National Audit Office, UK, “PF1 and PF2,” Report by the Comptroller and Auditor General. Ordered by the House of Commons to be printed on 17 January 2018. Newbery, D.M., “Cost Recovery from Optimally Designed Roads,” Economica, 1989, 56, 165–185. Poole, Robert, “Availability Payment or Revenue-Risk PPP Concessions? Pros and Cons for Highway Infrastructure,” Reason Foundation, 2017. https://reason​.org​/policy​-study​/availability​-paymentor​revenue​-risk​-public​-private​-partnership​-concessions​-pros​-and​-cons​-for​-hig​hway​infr​astr​ucture/ Power, Gabriel J., Mark Burris, Sharada Vadali, and Dmitry Vedenov, “Valuation of Strategic Options in Public-Private Partnerships,” Transportation Research Part A, 2016, 90, 50–68. Sarmento, Joaquim Miranda, and Luc Renneboog, “Portugal’s Experience with Public-Private Partnerships,” in Akintola Akintoye, Matthias Beck, and Mohan Kumaraswamy, eds., Public Private Partnerships: A Global Review. Routledge, UK, 2015. Small, K.A., “Economies of Scale and Self-Financing Rules with Noncompetitive Factor Markets,” Journal of Public Economics, 1999, 74, 431–450. Small, K.A., C.M. Winston, and C.A. Evans, Road Work: A New Highway Pricing and Investment Policy. Washington, DC: Brookings, 1989. Spiller, Pablo T., “Transaction Cost Regulation,” Journal of Economic Behavior & Organization, 2013, 89, 232–242. Tran, Anh, “Can procurement Auctions Reduce Corruption? Evidence from the Internal Records of a Bribe-Paying Firm,” Job-market paper, Kennedy School, Harvard, 2008. Trigeorgis, Lenos, “Making Use of Real Options Simple: An Overview and Applications in Flexible/ Modular Decision Making,” The Engineering Economist, 2005, 50(1), 25–53. TRIP, “Bumpy Roads Ahead: America’s Roughest Rides and Strategies to Make our Roads Smoother”, October 3, 2013, Washington DC. Trujillo, Lourdes, Emile Quinet, and Antonio Estache, “Dealing with Demand Forecasting Games in Transport Privatization,” Transport Policy, 2002, 9(4), 325–334. Valero, V., “Government Opportunism in Public Private Partnerships,” Journal of Public Economic Theory, 2015, 17, 111–135. Verhoef, E., and H. Mohring, Self-Financing Roads, International Journal of Sustainable Transportation, 2007, 3, 293–311. Villalba-Romero, Felix and Champika Lasanthi Liyanage, “Implications of the Use of Different Payment Models: The Context of PPP Road Projects in the UK,” International Journal of Managing Projects in Business, 2016, 9(1), 11–32. Wilson, James Q., Bureaucracy: What Government Agencies Do and Why They Do It. New York: Basic Books, 1987. Ye, Sudong, and Robert L. K. Tiong, “NPV-at-Risk Method in Infrastructure Project Investment Evaluation,” Journal of Construction Engineering and Management, 2000, 126(3), 227–233. Yescombe, Edward R., Public-Private Partnerships: Principles of Policy and Finance. ButterworthHeinemaan, 2007.

17. Financing sustainable transport infrastructure in emerging markets and developing economies José C. Carbajo1

17.1 INTRODUCTION In emerging markets and developing economies (EMDEs), infrastructure is an essential component of growth strategies, poverty reduction programs, and environmental sustainability policies. Infrastructure investments facilitate market competition, improve productivity, create jobs, and contribute to the clean energy transition (Bhattacharya et al., 2015). Transport infrastructure provides the urban mobility that people need to go to work, shop, take part in leisure activities, and to access basic services, such as health, education, and finance. Transport infrastructure connects economic networks and facilitates foreign direct investment (FDI) and trade flows, within and across borders, helping EMDEs to integrate into regional and global economies, and enhance access to social and economic opportunities for the poorer segments of the population. Transport infrastructure, on the other hand, can also create significant social, physical, and environmental impacts, such as congestion, accidents, and local air pollution and greenhouse gas (GHG) emissions. Transport is the fastest growing source of GHG emissions, therefore investments in transport decarbonization, both technology- and policy-driven, are required, especially in EMDEs, to achieve GHG emissions reductions in the face of increasing motorization and travel demand, and the need to reduce the connectivity gap (OECD, 2020). Multilateral development banks (MDBs), including the World Bank Group (WBG) and regional development banks, finance substantial amounts of transport infrastructure and services in EMDEs. They use a range of standard financing instruments (loans, equity), complex project finance arrangements such as public–private partnerships (PPPs), and policy and project advisory services, working with governments and private clients. Their common objective is to finance transport infrastructure and services to reduce poverty and promote growth while achieving and maintaining economic, financial, and environmental and social sustainability. This chapter describes how the WBG finances transport infrastructure and the type of development outcomes achieved by its transport interventions based on ex-post independent evaluation evidence. Section 17.2 describes the instruments the WBG uses to finance transport investments. Section 17.3 illustrates the development outcomes and success factors associated with the financing of urban transport and of transport infrastructure PPPs, topics that are at the crossroads of highly relevant development policy agendas: urbanization, mobilizing private finance, and climate change. Section 17.4 contains a brief discussion of the climate finance challenges associated with the decarbonization of the transport sector. Section 17.5 concludes with a reflection on the factors that will influence future transport financing trends, especially in EMDEs. 330

Financing sustainable transport infrastructure  331

This chapter complements several other chapters in the Handbook, including Chapter 14 on transport finance theory (Vasallo and Garrido), Chapter 15 on the links between transport finance and economic appraisal (Santos, Stead and Wormsley), and Chapter 16 on the theory of PPPs (Engel, Fischer, and Galetovic).

17.2 TRANSPORT FINANCING INSTRUMENTS The WBG finances public and private transport infrastructure projects in EMDEs using various instruments, including loans, guarantees, equity investments, and political risk insurance. It also provides policy advice, to help build credible and stable transport regulatory frameworks, and offers corporate advice to private clients, for example, to help them adapt their capacity to meet environmental, social, and governance (ESG) standards. This section provides a brief overview of the main financial instruments offered by the WBG’s four institutions: the International Bank Reconstruction and Development (IBRD) and the International Development Association (IDA), which are together referred to as the World Bank; the International Finance Corporation (IFC); and the Multilateral Investment Guarantee Agency (MIGA). It describes the basic financial and contractual structures of the WBG transport financing instruments using real-life project examples. The range of WBG financing instruments and credit enhancement products is illustrated by the entries in Table 17.1. Transport financing instruments vary according to client eligibility. The World Bank works mainly with governments and public sector clients at different levels of government, national and sub-national, for example. The IFC and MIGA only work with private-sector clients. 17.2.1 IBRD Loans and Guarantees The World Bank (IBRD plus IDA) provides financing for transport sector projects through loans and guarantees for middle-income EMDEs (IBRD) and poorer developing countries (IDA). These projects always enjoy the financial backing of the client country’s government. For example, IBRD can provide a loan directly to a public transport company but the loan must be guaranteed by the host country’s government. Alternatively, IBRD can provide a loan to the host government, which on-lends the funds to a transport company. This alternative typically involves a ‘project agreement’ between IBRD and the transport company or operators implementing the project. Figure 17.1 shows the different financial instrument categories used by the World Bank. A combination of factors might explain the declining trend in the magnitude of transport commitments by the World Bank, including the limited borrowing capacity of developing country governments and a change in investment priorities in those countries leading to changes within the World Bank infrastructure lending programs. Investment Project Finance (IPF) represent the most utilized instrument to finance IBRD transport projects. IPFs consist mostly of loans but some IPFs might combine a loan and a guarantee. For example, in 2017 the World Bank extended a loan of EURO 22 million to Croatia for the modernization and restructuring of the road sector, to strengthen its institutional effectiveness, enhance its operational efficiency and increase its debt service capacity.2 The IPF package included a credit guarantee of repayment to private lenders for EURO 350 million to support the debt optimization strategy of the government for the road sector.

332  Handbook on transport pricing and financing

Table 17.1  Main WBG transport financing instruments IBRD

IDA

IFC

Middle-income country governments or subnational entities with government guarantees

Low-income country governments

Private sector

Grant

IFC A-loan

Credit

IFC B-Loan (third parties)

Financing IBRD flexible loan

MIGA

Single currency Parallel loans lending (third parties) program

Equity finance

  Development Policy Financing (DPF)

Local currency

  Program for Results (PforR) Credit  Partial Credit Guarantee (Policy-based & Enhancement Project-based)   Partial Risk Guarantee (PRG)

Full/partial credit guarantee

Political Risk Insurance

Credit-linked guarantee Risk sharing facilities

Credit guarantee (Non-honoring of financial obligations)

7,000

45

6,000

40 35

5,000

30

4,000

25 20

3,000

15

2,000

10

1,000 0

Number of Projects

USD million

Source:  World Bank.

5 2012

2013 IPF

2014

2015

PforR

2016 DPL

2017

2018

Guarantee

2019

2020

2021

0

# of Projects

Source:   World Bank.

Figure 17.1  World Bank transport commitments by financing instrument (2012–2021) World Bank guarantees are intended to help clients mobilize private-sector financing and/ or mitigate government payment risk. IBRD has two types of guarantee for private lenders: (i) a partial credit guarantee (policy-based or project-based) covering certain debt service payments against all risks; and (ii) a partial risk guarantee, protecting lenders against payment defaults arising from breaches of sovereign contractual obligations to a project. All IBRD

Financing sustainable transport infrastructure  333

guarantees must be counter-guaranteed by the host government where the transport project takes place. An IBRD or IDA guarantee is specifically tailored to either the circumstances of the transport project and transaction being guaranteed (project-based guarantees), or the borrowing transaction of a government to meet fiscal needs (policy-based guarantees). Project-based guarantees can be either loan guarantees or payment guarantees. With World Bank loan guarantees, the debt must be extended by commercial banks. World Bank guarantees can be provided with or without an associated World Bank loan or credit. The World Bank uses these financing instruments in transaction structures that extend maturities for borrowers, achieve spread savings, or often leverage additional volume of private capital. 17.2.2 Development Policy Financing and Program for Results Development Policy Financing (DPF) is an instrument that the World Bank uses to respond to client government requests for budgetary support. DPFs provide policy advice to help shape and back a reform program rapidly disbursing non-earmarked budget financing. Transport DPFs support specific policy and institutional actions, assessed by World Bank expertise, that drive transport sector reforms. DPFs contain so-called ‘prior actions’ that the client country must undertake, which are essential to achieving reform objectives and must be completed prior to any World Bank disbursement of financing. The World Bank provides a DPF only when it has determined that a country’s macroeconomic policy framework is adequate. Policies supported by a DPF are reviewed with the objective of identifying whether they are likely to have any significant poverty, social, environmental, or natural resources adverse impacts. Between 2006 and 2019, the World Bank approved 93 DPFs with a transport policy component. Of those, six DPFs led by the World Bank’s Transport Global Practice (TGP) focused on urban transport, while only one DPF targeted the railways sub-sector. The policy components of TGP-led DPFs in urban transport included a large variety of reforms such as, for example, the setting up of public transport institutional frameworks; awarding bus contracts and urban (transport) planning reforms; the setting up of vehicle inspection system and legislation for congestion pricing and electronic vehicle identification. Other non-urban transport policy components include, for example, railway company restructuring; the setting up of a road safety agency; or road agency reform. The TGP-led DPFs only achieved a 60% success rate, that is, programs rated ex-post as moderately satisfactory or above in terms of the development outcomes achieved. This may partly explain the relatively moderate use of the DPF instrument compared with the more standard IPF instrument. Success of transport DPFs depends on a combination of factors: the borrower’s commitment and ownership of the reform program; the quality and depth of the analysis informing the sector reforms; the identification of critical risks; and the coordination of the public agencies responsible for the implementation of the reforms. Program for Results (PforR) is another financing instrument that links the disbursement of funds directly to the delivery of defined results, helping WBG client countries improve the design and implementation of their own development programs by strengthening institutions and building capacity. The proceeds of a PforR financing are disbursed upon the achievement of verified results specified as disbursement-linked indicators. Such disbursements do not depend on, nor are they associated with, individual projects or expenditures. An example is a US$150 million Morocco urban transport program3 approved in 2020 to strengthen

334  Handbook on transport pricing and financing

the capacity of urban transport institutions to plan, implement, and monitor urban transport infrastructure and services, and to improve the level of service of urban transport in targeted corridors in the program area. 17.2.3 IFC Loans and Equity Investments The IFC provides corporate loans (at market commercial rates) to private companies with no direct government guarantees of repayment. The IFC finances transport projects and companies through loans from its own account, typically for 7 to 12 years, but can also make loans to intermediary banks, leasing companies, and other financial institutions for on-lending. The IFC also has an extensive loan syndication program (B-Loans). When an IFC loan includes financing from the market through the B-Loan structure, the IFC retains a portion of the loan for its own account (the ‘A-Loan’) and sells participations in the remaining portion to participants (the ‘B-Loan’). The borrower signs a single Loan Agreement with the IFC, and the IFC signs a Participation Agreement with the participants. The IFC is the sole contractual lender for the borrower. While The IFC is the lender of record, the participants’ involvement is known to the borrower, and is included in any transaction and publicity. The A-/B-Loan structure allows participants to fully benefit from the IFC’s status as a multilateral development institution. All payments including principal, interest, and fees gain the advantages of the IFC’s preferred creditor status, namely, the preferential access that IFC loans have to foreign currency in the event of a country’s foreign exchange crisis. The IFC commits to the participants to allocate payments pro-rata between the A- and B-Loans. As a result, the IFC cannot be paid in full until all participants are paid in full. Similarly, a default to a participant means a default to the IFC. In response to international banks cutting back lending in EMDEs during the 2008 global financial crisis, the IFC began syndicating parallel loans to development finance institutions and other ineligible B-Loan participants. Under this approach, the IFC acts as a loan arranger using its existing syndication platforms, deal-structuring expertise, and global presence to identify investments, perform due diligence, and negotiate loan documents in cooperation with parallel lenders. Parallel lenders are expected to benefit from cost and time savings and borrowers are expected to benefit from enhanced access to financing and time and cost savings throughout the life of the loan. An example is the IFC’s US$45 million corporate loan in 2011 to a strategic regional partner that established port and logistics operations in Iraq, part of the reconstruction efforts in the country.4 The loan financed a port PPP project comprising the operation and management of two container terminals in the southern Iraq port of Umm Qasr; it also included the development of an inland container depot and logistics center near the port; and the development of warehousing and transport facilities for the oil and gas industry in Kurdistan. The IFC can also invest directly in transport companies’ equity and through privateequity funds. Equity investments are intended to provide developmental support and longterm growth capital. The IFC generally invests between 5% and 20% of a company’s equity, encouraging the companies to broaden their share ownership through public listings, thereby deepening local capital markets. Another way in which the IFC takes a capital-type risk is by investing through profit-participating loans, convertible loans, and preferred shares. Following a sharp drop in business activity in 2009–2010 in the aftermath of the 2008 financial crisis, the IFC’s transport commitments have been volatile, averaging

Financing sustainable transport infrastructure  335 1600

20 18

1400

USD million

14

1000

12

800

10 8

600

6

400

4

200 0

Number of projects

16

1200

2 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 Loan

Equity

Loan & Equity

Guarantee

Risk Management

B loan

Fee collecting project

Project Count

0

Source:  IFC.

Figure 17.2  IFC transport commitments by financing instrument (2005–2020) US$600 million annually (excluding 2011 and 2016), mostly dominated by loan instruments (Figure 17.2). 17.2.4 MIGA Guarantees The MIGA offers two kinds of risk mitigation products (guarantees). They are political risk insurance and credit enhancement cover. The MIGA’s political risk insurance (PRI) covers four risks: transfer restriction and currency inconvertibility; expropriation; war and civil disturbance; and breach of contract. PRI provides political risk mitigation primarily for private foreign direct investors. The eligible investments include equity, debt, loan guarantees, and non-equity direct investments. The MIGA-insured parties (guarantee holders) include equity investors, private (or public) lenders, bilateral development institutions, public insurers, etc. The MIGA’s credit enhancement cover is expected to provide insurance against the risk of non-honoring of sovereign financial obligations (NHSFO) by sovereign and sub-sovereign governments or by state-owned enterprises (NHFO-SOEs). This instrument applies to public sector projects, normally involving loans. It covers the nonpayment risk of financial obligations to private-sector lenders by public sector borrowers. An example is a NHFSO guarantee that the MIGA issued in 2012 to cover a $250 million loan to the government of Panama from a bank consortium led by Citibank. The purpose of the loan was to finance an engineering, procurement, and construction (EPC) contract for

336  Handbook on transport pricing and financing

the construction, supply, and installation of Panama Metro’s Line One, supporting a modern and integrated mass-transit system.5 The guarantee included interest and other financing costs coverage, where payment of the syndicated loan was guaranteed by the government of Panama, acting through the Ministry of Economy and Finance. Over the last 11 years, the MIGA has issued a total of 20 guarantees related to transport infrastructure projects with half a dozen projects since 2015 (Figure 17.3). They are split evenly by financing instrument (PRI and NHFO) although the volume of cumulative gross issuance of the non-honoring (NH) type guarantee at US$2.5 billion is almost three times that of the political risk insurance guarantee (US$0.88 billion). About 83% of the gross issuance relates to road and railway projects; the rest relates to air and water transport projects. 17.2.5 Additional Financing Instruments The set of WBG infrastructure financing instruments presented in Table 17.1 is not exhaustive. The WBG has other sophisticated contingent financing instruments, such as deferred drawdown options (World Bank), and risk management instruments such as hedging products managed by the Bank Group’s Treasury Department. These include, for example, currency swaps and interest rate swaps (offered by the World Bank and the IFC); disaster risk financing, such as catastrophe bonds (World Bank), or weather hedges (IFC). The WBG financing of transport infrastructure is often accompanied by advisory services and analytics (ASA) provided by the World Bank or the IFC. These are expert services that support clients in designing or implementing better policies, strengthening institutions, building capacity, and informing strategies or operations. ASA can be financed through the World 1000.00

6

900.00

USD million

700.00

4

600.00 3

500.00 400.00

2

300.00 200.00

1

100.00 0.00

2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 NH

PRI

0

Guarantees

Source:  MIGA.

Figure 17.3  Transport-related guarantee issuance by the MIGA (2011–2021)

Number of guarantees

5

800.00

Financing sustainable transport infrastructure  337

Bank’s or the IFC’s own budget, donor funding or reimbursed by the clients. The World Bank Treasury provides advisory services on asset and debt management to sovereign and subsovereign clients and public sector institutions. 17.2.6 Public–Private Partnerships PPPs are one way the private sector gets involved in financing, building, and operating transport projects. Given their relatively complex nature and how prominent they have become over recent years, PPPs play a special role in the financing of transport infrastructure. A transport infrastructure PPP is a long-term contract between a private party and a government entity, to provide a transport infrastructure asset or service, in which the private party bears significant risk and management responsibility, and its remuneration is linked to performance (World Bank, 2017b). More details on the economic theory of PPP contracts can be found in Chapter 16 of the Handbook by Engel, Fischer, and Galetovic. Three characteristics are at the core of a PPP contract: (i) it bundles together multiple project phases or functions, including design, build (or rehabilitate), finance, and maintenance operations; (ii) a special purpose vehicle (SPV) provides the services, separating the assets and liabilities associated with such provision; and (iii) a payment mechanism, at the core of the risk allocation between public and private parties, remunerates the private party according to performance. Transport PPPs have a long history, having pioneered infrastructure PPP programs in many countries.6 According to the World Bank’s PPI database, over the last 30 years MDBs have financed 880 infrastructure PPPs with a combined investment value of US$260 billion. Of those, 167 (or 19%) are transport PPPs with a combined investment value of US$68.4 billion, or 26% of total infrastructure PPP financing in EMDEs. (Figure 17.4).

Source:   World Bank PPI Database.

Figure 17.4  Transport PPPs financed by MDBs in EMDEs (1990–2020)

338  Handbook on transport pricing and financing

17.3 TRANSPORT FINANCING EXAMPLES Financing transport infrastructure is a means to an end. The immediate goal of financing transport infrastructure is to build (or improve) transport infrastructure assets and provide (or enhance) the transport services they provide. Intermediate outcomes include travel time savings, lower operating costs, reduced accident risks, and improved environmental performance. In addition, with the financing of transport projects, the WBG also seeks to contribute to reducing poverty and promoting shared prosperity in EMDEs, to support private-sector development (IFC), and to facilitate foreign direct investment (MIGA). A permanent challenge faced by MDBs is to monitor and evaluate whether the investment projects they finance achieve their expected results. Ex-post evaluations promote two objectives: (i) accountability for delivering on each MDB’s institutional mandate (through the assessment of performance and results); and (ii) learning, feedback, and knowledge sharing (based on the outcomes and lessons of experience drawn from the evaluation themselves) with a view to improving performance and results over time (World Bank, 2019). The Independent Evaluation Group (IEG) of the WBG is responsible for regularly assessing the development effectiveness of the activities and programs financed by the WBG across sectors and regions. This section summarizes the main findings and lessons of experience regarding the WBG support to urban transport (World Bank, 2017a), to illustrate transport financing in action, that is, the type of development results that can be achieved by financing transport in EMDEs. The transport evaluation experience of other MDBs can be found, for example, in ADB (2020); AfDB (2014, 2021); and IADB (2014). The section also includes a brief reference to key lessons of experience from the MDB financing of infrastructure PPPs, including transport PPPs, traditionally the largest share of infrastructure PPP financing across MDBs. 17.3.1 Financing Urban Transport Urban transport systems in EMDEs facilitate the movement of people and goods and provide access to economic and social opportunities. They also connect the urban poor to job opportunities and other services and can facilitate safe accessibility for women, the disabled, and the elderly. With adequate planning and investment in public transit, integrated services and ticketing can improve affordability and convenience so that low-income users can manage their longer, and often more complex, journeys to their destinations. While rapid motorization puts city transportation systems and the environment under pressure, urban transport can help mitigate the negative consequences of congestion, pollution, safety risks, and poor security associated with unplanned city growth.7 WBG investment in urban transport is small compared to the overall volume of urban transport investment globally. The World Bank Group’s engagement in urban transport throughout 2007–2016 involved a project portfolio of close to $25 billion in financing, mainly loans by the World Bank and the IFC, a handful of DPFs (World Bank) and a few guarantees (MIGA). Of 297 investment projects, the World Bank was responsible for 93% (worth $23.5 billion), while the IFC delivered over half a billion dollars and the MIGA guarantees totaled $1.1 billion. During the IEG evaluation review period 140 projects were completed and 104 were evaluated. Urban roads and metro investments were responsible for the lion’s share of World Bank Group financing urban transport commitments over the 2007–2016 period (Figure 17.5).

8000

160

7000

140

6000

120

5000

100

4000

80

3000

60

2000

40

1000

20

0

Roads only

Conven onal buses

Bus Rapid Transit

Total commitment (USD million)

Metro

Upstream & Other

Number of projects

USD million

Financing sustainable transport infrastructure  339

0

# of projects

Source:   World Bank (2017a).

Figure 17.5  World Bank Group Urban Transport Portfolio (2007–2016) The ex-post evaluation of the WBG urban transport financing activities focused on their impact in three areas: (a) urban mobility and social inclusion (especially that of the poor, women, and disabled persons); (b) the financial and environmental sustainability of service delivery; and (c) the institutional development of urban transport providers. 17.3.1.1 Mobility and social inclusion The WBG’s main approach to improving urban mobility is through an increased supply of urban transport services, which has generally been effective to increase service quality and access, contributing to improved mobility. However, demand management objectives (such as reducing travel demand through integrated transport and land use planning, and explicit measures to shift transport from private cars to public transport) are not yet broadly supported by WBG operations. The mobility achievements of WBG demand management interventions are generally limited to specific transport systems or areas. Projects with both demand management and supply-side measures tend to achieve more mobility improvements than those that focused only on either supply or demand measures. Mobility improvements for the disadvantaged (the poor, women, disabled, and elderly persons) were pursued in WBG urban transport projects primarily through targeted interventions (e.g., connecting poor neighborhoods to urban transport services). While such interventions led to improved access, these measures are not applied broadly yet in WBG urban transport projects. While the affordability of urban transport services can seriously limit the mobility of the poor, it is seldom a clear priority in WBG’s urban transport portfolio. Only 9% of the 104 evaluated projects addressed affordability concerns. In those few cases, affordability was addressed through policies for fare integration or subsidies, or the preparation of affordability studies.

340  Handbook on transport pricing and financing

Factors behind successful WBG urban transport projects included careful design of targeted interventions and strong implementation efforts. Participatory consultations with representatives from disadvantaged groups and civil society, for example, were effective in designing and planning interventions tailored to support targeted beneficiaries and to promote social inclusion. 17.3.1.2 Financial and environmental sustainability Over-optimism about project financial viability is common at the design stage and the quality of financial sustainability analysis varies widely across WBG urban transport projects. A quarter of completed and closed projects (projects at the end of their loan disbursement period) experienced cost overruns; while three quarters experienced time delays. If the investment is successful, private-sector participation in urban transport projects has the potential to improve financial sustainability through resource mobilization, stable contracting that secures finance, and improved capacity yielding operational efficiencies not always available through a public sector delivery model. The WBG has a high level of support for private-sector participation (PSP) in urban transport, as found in 66% of dedicated projects in the sector. Mass-transit projects with private-sector participation, for example, were more likely to achieve financial sustainability of operations than projects utilizing only the public sector. Urban transport interventions in support of environmental sustainability included both upstream (policy and regulatory framework) and downstream (operational) instruments, but downstream mitigation support was the most common feature. More than 70% of closed and evaluated projects tracking environmental benefits contributed to reduced emissions of greenhouse gases or air pollutants. However, modal shifting from private vehicles to public transit or non-motorized transport did not materialize in any significant way, which confined the environmental benefits to the vicinity of the project area or to the specific transport mode supported by the World Bank Group. Broader and sustained environmental benefits were better achieved in projects with a comprehensive approach that included both upstream and downstream measures. 17.3.1.3 Institutional development The World Bank Group supports urban transport institutional development primarily through IBRD/IDA lending. During the 2007–2016 evaluation period, the World Bank provided four types of institutional development support for urban transport: (i) general institutional capacity building including management and human resources; (ii) fiscal capacity enhancement of urban transport agencies or municipalities responsible for urban service, including urban transport service delivery; (iii) strategy, regulation, and policy design; and (iv) structural change of urban transport institutions, including the setting up of a new urban transport agency. An 80% share of WBG urban transport projects included institutional development components, usually focused on urban transport agencies or on city governments’ fiscal, administrative, and service delivery capacity. Project components included training, study tours, procedural improvements, and system upgrades. Evaluation findings suggest that support for institutional development is more likely to succeed when the WBG is continuously engage with the sector over time. Among evaluated projects, institutional development components were fully successful in projects with continuous WBG engagement. This is partly because successive projects could build on previous

Financing sustainable transport infrastructure  341

institutional development achievements. Success was also likely when WBG engagement met the client’s priorities and the client showed strong commitment and leadership. By contrast, a key shortcoming of sustained institutional development is the predominance of ‘one-time’ projects. Of about 100 cities supported by the WBG, only nine cities in six countries had such continuous engagements through dedicated urban transport projects leading to successful institutional development outcomes. 17.3.2 Financing PPPs Multilateral development banks have extensive experience financing transport PPPs in EMDEs (Section 17.2.6). Transport PPPs account for a large share of the MDB’s network infrastructure PPP portfolios, which include energy, telecommunications, transport, and water and sanitation projects. Over the last decade, MDBs have assessed the performance of infrastructure PPP projects (including transport PPPs), drawing lessons, and identifying key factors of success or failure (see ADB, 2009; EBRD, 2014, 2015; EIB, 2016; IADB, 2017; IFC, 2013; World Bank, 2014). A comprehensive synthesis of the findings and lessons of experience from those evaluations and assessments of infrastructure PPP financing is beyond the scope of this chapter.8 The studies cited above identify success factors in transport PPPs linked to several policy and project dimensions. These include, for example, the enabling environment: the degree of long-term political commitment; the quality and stability of the legal and regulatory frameworks, the communication policy carried out by the public contracting authority and private contractors; and the public acceptability of the PPP arrangement. Project preparation is another crucial dimension of the PPP contract; it includes the valuefor-money assessment; the risk allocation between the public and private parties; the ability and capacity of the public and private-sector stakeholders; and the quality, openness and transparency of the procurement practices and project approval processes. Finally, project funding and project financing, the focus of this chapter, is a third key area of any transport PPP arrangement. 17.3.2.1 Project funding and financing PPP funding and PPP financing are distinctive challenges. While there are many instruments for financing the upfront investment costs of infrastructure PPP projects – plain budget financing, long-term loans, bonds, equity – there are essentially only two sources of funding the full cost of infrastructure – i.e., either from the public budget (through taxation) or from the users of the transport infrastructure asset (through direct charges). Where user charges are not feasible or can only provide limited revenues in comparison to the capital and operating costs of the transport project, public budget funding becomes the only source to pay-back for the costs of the transport infrastructure. The funding of projects tends to be a significant obstacle to PPPs (possibly the biggest) since public budgets are severely limited and transport users’ unwillingness to pay for public services is such that the long-term affordability of PPP projects is frequently challenged. Governments may favor the PPP option to overcome fiscal constraints which limit their ability to mobilize the amount of public financing required for the transport infrastructure investment and/or to achieve fiscal targets, especially if the transport PPP investment can be treated as ‘off-balance’ from the public national accounts. However, an excessive focus on the

342  Handbook on transport pricing and financing

off-balance sheet treatment of a transport PPP can be at the expense of sound project preparation and may push procuring authorities to use PPPs where it is not appropriate. Also, when a PPP project is off-balance sheet, there is a risk that the fiscal liabilities that arise from it are not managed properly (e.g., recognition of government financial commitments, whether firm or contingent, explicit, or implicit). Knowledge of the local market and the ability to provide local currency financing (thus avoiding a currency mismatch) are two substantial advantages when local financial markets and institutions co-finance local transport PPP projects.

17.4 SUSTAINABLE TRANSPORT FINANCING A recent analysis of infrastructure investment needs finds that US$6.3 trillion per year is needed until 2030 to meet the Sustainable Development Goals (SDGs), with an additional $300 billion needed to make those investments compatible with the goals of the Paris Agreement (OECD, World Bank, and UNEP, 2018). Achieving net-zero emissions by 2050 requires an estimated US$50 trillion in investments worldwide (Morgan Stanley, 2019). The high share of global GHG emissions caused by the transport sector and the potential impacts of climate change on transport infrastructure suggest the need to explore innovative solutions to finance sustainable transport. Transport is responsible for 65% of global oil demand (IPCC, 2018). This translates into almost a quarter of global CO2 emissions from fossil fuel combustion, while demand for transport is projected to grow rapidly in the coming decades, as low- and middle-income countries continue their economic development and urbanization.9 Climate change, on the other hand, can affect transport systems directly, for example through damage to existing infrastructure, and indirectly, for example through changes in trade flows, agriculture, and energy use. A future with frequent extreme weather events can damage port facilities, roads, railways, and bridges through flooding or sea level rise (Mukhi et al., 2020). Financing sustainable transport is, therefore, essential to foster inclusive growth, expand access to essential services, and address climate change. 17.4.1 Decarbonizing Transport Decarbonizing transport in EMDEs presents multiple challenges; two of them are (a) to identify the policies and investments that support decarbonization (‘what to do’); and (b) how to finance those policies and investments (‘how to pay’ for it). A good example of the policy debate regarding what to do can be found in the World Bank’s Climate Change Action Plan 2021–2025 (World Bank, 2021b). It identifies three areas to focus the proceeds of climate finance and support transport decarbonization in EMDEs. First, mobility and access: cities and urban areas need to invest in planning, developing, and managing integrated transport systems, including high-quality public transit to replace private vehicles – and fragmented informal urban transport services – supporting active mobility through non-motorized modes and the use of electric vehicles, given their potential to reduce GHG emissions, air pollution and associated health impacts. The second area is freight and logistics: investment efforts to decarbonize the freight sector and deliver competitive logistics comprise a range of interventions, including optimizing

Financing sustainable transport infrastructure  343

delivery networks; the re-engineering of supply chains; changes in inventory practices; reducing the fragmentation of production; a shift to lower-carbon transport modes and to energyefficient and low-carbon vehicles across modes, including maritime transport. The third and final area is building resilient transport systems: this involves financing resilient infrastructure solutions, including, in addition to policy and regulatory measures, investments in physical infrastructure and new technologies, and in post-disaster risk and recovery support, so that climate change risk and resilience are integrated into any potential transport infrastructure rebuilding efforts. Transport is, therefore, a key infrastructure sector in EMDEs that must be transformed to achieve a low-carbon future. But the transition efforts to support mobility and access, freight and logistics, and resilient transport systems demand substantial amounts of climate finance, which is a second key challenge to decarbonize transport, requiring the mobilization of private capital. 17.4.2 Climate Finance Instruments to Mobilize Private Capital Climate mitigation and transport decarbonization in EMDEs rely on the financial support of governments, MDBs, and the private sector. Given the high levels of public debt and limited fiscal space of EMDE countries, public financing is largely insufficient to address decarbonization goals and multiple and competing public policy priorities, from health and education to transport and water and sanitation, among many. MDB finance for climate change mitigation is similarly limited. For example, the World Bank Group’s total annual commitments are on average about US$60 billion. Mobilizing private capital to invest in transport decarbonization and climate-smart technologies is, therefore, essential to address the climate financing gap. Three financing instruments that can help attract private capital to decarbonize transport are green bonds, climate funds, and blended finance.10 17.4.2.1 Green bonds Green bonds are fixed-income instruments that can mobilize private capital with proceeds earmarked exclusively for projects that have environmental benefits. The IFC apply the Green Bond Principles (GBP) which regulate the use of proceeds (e.g., loan financing portion of eligible projects selected from the IFC’s climate-related loan portfolio); the process for project evaluation and selection; the management of proceeds; and the reporting according to a framework11 which aims at ensuring the integrity of the market through increased transparency and to avoid ‘green washing’. Green bond issuers typically designate the use of proceeds for specific projects that would contribute to environmental objectives. In the period 2012–2020, 43 emerging market economies have issued green bonds, registering cumulative issuance of US$226 billion. Cumulatively, over the same period, the largest share of the use of proceeds has been designated for renewable energy (35%), while transport projects rank second (29%) and water and building projects, third and fourth, respectively (11% and 9%) (Amundi and IFC, 2020). An example of an urban transport project financed with proceeds from a green bond is IFC’s US$30 million loan to Bogotá Distrito Capital to finance ‘TransMiCable’, an aerial cableway connecting the low-income mountain neighborhood of Ciudad Bolívar to the City of Bogotá’s mass-transit system.12 The result sought is a reduction in congestion and an increased frequency of service. Users are expected to save about 6.2 million travel hours per year which

344  Handbook on transport pricing and financing

translates into savings for users of about $14.7 million per year, and an estimated annual reduction of 134 tons of CO2 equivalent (IFC, 2020). 17.4.2.2 Climate funds Climate funds are designed to support investment projects that reduce GHG emissions. An example is the Green Climate Fund (GCF),13 the world’s largest dedicated climate fund supporting the Paris Agreement. During the initial mobilization stage (2015–2018) the GCF raised US$10.3 billion with an additional US$10 billion pledged during its first replenishment (2020–2023). Most of the climate finance provided by the GCF flows through international institutions such as the UN Development Program (UNDP) or the World Bank. The aim of the GCF is to achieve a 50:50 split between mitigation and adaptation investments with, for example, a geographical focus on small island states and African countries for climate change adaptation investments. At the end of September 2021, the GCF supported a portfolio of 190 projects in eight strategic results areas, including transport, although transport is currently the smallest investment area by number of projects and committed finance. An example of a GCF-approved urban transport project consists of an 85 km of the double-track electric light rail transit system in San José’s Greater Metropolitan Area in Costa Rica. Transport contributes more than 50% of the country’s GHG emissions.14 The project is expected to encourage behavioral change toward an increased use of public transport. The total project cost is US$1.9 billion, 14.5% of which is GCF’s funding consisting of a US$250 million loan and a US$21.3 million grant. The Central American Bank for Economic Integration (CABEI) is the project’s accredited entity. 17.4.2.3 Blended finance Blended finance is the combination of commercial finance and subsidies to de-risk projects. Subsidies may be provided by Trust Funds.15 An example of a recent blended finance instrument is the IDA’s Private-Sector Window (PSW). It was designed to support private investments in the IDA-only and IDA-eligible Fragile and Conflict Situation (FCS) countries where the private sector faces many risks and challenges. The motivation to use concessional funds (subsidies) is to improve the risk–reward profile of private investments (with IFC and MIGA involvement in the case of the PSW) that are unable to proceed on strictly commercial terms. This can be done through first loss guarantees or co-investment with the IFC (including subordinate loans) or by providing solutions to mitigate local currency risk, for example. The result is a credit enhancement of the project through risk mitigation (mitigating the risk of possible losses to investors) and limiting IFC and MIGA risk exposure. This leads to a reduction in pricing that makes the project commercially viable and perhaps to an increased size of the investment. Established in 2017, the PSW has experienced a slow start and has yet to support a transport infrastructure project (World Bank, 2021a). An example of a transport project using blended finance is the IFC’s Zaporizhzhia Smart City project.16 The project provides long-term financing to the City of Zaporizhzhia in Ukraine to back a comprehensive package of initiatives to improve energy efficiency, transport logistics, municipal infrastructure development, and the deployment of innovative and sustainable technologies to improve urban transport. With blended finance, the City of Zaporizhzhia can access long-term financing at affordable rates, without a sovereign guarantee. The level of subsidy provided by the blended

Financing sustainable transport infrastructure  345

concessional finance co-investment is estimated to be 1.5% of the total project cost of EUR 35 million. Concessional funding comes from the Clean Technology Fund (CTF), a multidonor trust fund established in 2008 to provide emerging economies with scaled-up financing for the demonstration, deployment, and transfer of low-carbon technologies with a significant potential for long-term greenhouse gas (GHG) emission savings.

17.5 CONCLUSIONS Financing the substantial transport investments implied by the SDG and Paris Agreement agendas will remain a serious challenge in the future, especially in EMDEs where most governments have limited public debt capacity. Consequently, private capital mobilization will also remain a necessary condition to finance the large transport infrastructure gaps worldwide. The World Bank Group is likely to remain a relevant player in EMDEs, alongside other MDBs, continuing to support transport infrastructure investments and services, using the menu of standard financing instruments (loans, equity, guarantees) described in this chapter. Mobilizing private capital, however, will test the WBG’s ability, and that of other MDBs, to develop new instruments and approaches aligned with the evolving needs of private investors; to help identify robust pipelines of bankable transport projects; and to increase collaboration among MDBs, three necessary conditions to scale up transport infrastructure investments in EMDEs (World Bank, 2020b). A special emphasis should be on financing instruments and approaches to decarbonize the transport sector, such as those discussed in this chapter, including green bonds, climate funds, and blended finance options. Other transport infrastructure financing methods, such as land value capture (see Chapter 14 of this Handbook), is likely to be used more often given their potential revenue-raising properties. Transport financing is a necessary condition but not sufficient to achieve sustainable development outcomes. MDBs will need to improve the monitoring and evaluation of the performance of their transport projects, to learn from experience and to adapt their financing instruments and approaches so that they become increasingly relevant in crowding in private capital, and increasingly effective in achieving development outcomes.

NOTES 1. Director, Financial, Private Sector, Infrastruture and Sustainable Development Department, Independent Evaluation Group, The World Bank Group (position held until October 2021). I am grateful to the editors of the Handbook and an anonymous referee for their comments on the first draft of the chapter; and to Daniel Benítez, Nicolás de León de María, Elizabeth Goller, Clive Harris, Hiro Hatashima, Kavita Mathur, Aurelio Menéndez, Nadia Paola Ramírez Abarca, Andrew Stone, Ichiro Toda, and Fang Xu (World Bank Group); and to Professor Antonio Estache (Université Libre de Bruxelles) for their helpful suggestions and data sharing. Any errors or misinterpretations are solely mine. 2. https://documents.worldbank.org/en /publication /documents-repor ts/documentdetail / 297551535982626811/croatia-modernization-and-restructuring-of-the-road-sector-p155842-ibrdloan-8749-audit-report-on-financial-statements-of-hc-hrvatske-ceste-2017-tr-eng 3. https://projects​.worldbank​.org​/en ​/projects​-operations​/document​-detail ​/ P173048​?type​=projects 4. https://disclosures.ifc.org/project-detail/SII/40784/umm-qasr 5. https://www.miga.org/project/panama-metro-line-one

346  Handbook on transport pricing and financing

6. See Overview of PPP experience, in the Public-Private Infrastructure Advisory Facility’s Toolkit for PPPs in Roads and Highways https://ppiaf​.org​/sites​/ppiaf​.org​/files​/documents​/toolkits​/ highwaystoolkit​/6​/pdf​-version​/1​-21​.pdf 7. Section 17.3.1 draws primarily from World Bank (2017a). 8. World Bank (2017b; Table 1.2; page 26) contains non-MDB sources that review the experience of transport PPP projects worldwide, including case studies on (a) bridges and highways from the United Kingdom, Europe, Australia, China, India, Israel, and Argentina; and (b) 12 light rail systems in the United Kingdom, Malaysia, the Philippines, Thailand, Canada, and South Africa. 9. See also International Energy Agency overview of the transport sector: www​.iea​.org​/topics​/ transport 10. World Bank (2020b) offers a recent and comprehensive assessment of the World Bank Group approaches to private capital mobilization. 11. Harmonized Framework for Impact Reporting www​.ifc​.org​/wps​/wcm​/connect​/3deee5d3​-9073​4eff​-99fb​-b061d7137ff6​/ Handbook​-Harmonized​-Framework​-for​-Impact​-Reporting​-220420​.pdf​? MOD​=AJPERES​&CVID​=nx6alip 12. https://disclosures.ifc.org/project-detail/SII/39772/bog-transmicable 13. See www​.greenclimate​.fund 14. https://www.greenclimate.fund/project/fp166 15. World Bank (2020a) offers a synthesis of recent IFC experience with blended finance instruments based on a cluster of project evaluations. 16. https://disclosures.ifc.org/project-detail/SII/43181/zaporizhzhia-smart-city

REFERENCES African Development Bank, 2014. Transport in Africa: The African Development Bank’s Intervention and Results for the Last Decade. Independent Development Evaluation, Abidjan: AfDB. African Development Bank, 2021. Evaluation of AfDB Road and Port Projects (2012-2019). Cluster Evaluation Report, Abidjan: AfDB. Amundi Asset Management and International Finance Corporation, 2020. Emerging Market Green Bonds Report 2020 – On the Road to Green Recovery. Washington, DC: Amundi and IFC. Asian Development Bank, 2009. Special Evaluation Study on ADB Assistance for PPPs in Infrastructure Development – Potential for More Success. Independent Evaluation, Manila: ADB. Asian Development Bank, 2020. Sector-wide Evaluation: ADB Support for Transport. Independent Evaluation, Manila: ADB. Bhattacharya, Amar, Jeremy Oppenheim and Nicholas Stern, 2015. Driving Sustainable Development through Better Infrastructure: Key Elements of a Transformation Program. Global Economy and Development Program Working Paper, Brookings Institution. Washington, DC. European Bank for Reconstruction and Development, 2014. Private Sector Participation in Municipal and Environmental Infrastructure Projects – Review and Evaluation. Evaluation Department, London: EBRD. European Bank for Reconstruction and Development, 2015. IPPF- Contribution by EBRD to the G20 Infrastructure Focus (mimeo). London: EBRD. European Investment Bank, 2016. Hurdles to PPP Investments A Contribution to the Third Pillar of the Investment Plan for Europe. European PPP Expertise Centre (EPEC), Luxembourg: EIB. Inter-American Development Bank, 2014. Background Paper: Transport Sector (input for the evaluation Climate Change and the IDB: Building Resilience and Reducing Emissions). Office of Evaluation and Oversight. Washington, DC: IADB. Inter-American Development Bank, 2017. Evaluation of Public-Private Partnerships in Infrastructure. Office of Evaluation and Oversight. Washington, DC: IADB. International Finance Corporation, 2013. A Winning Framework for Public-Private Partnerships: Lessons from 60-Plus IFC Projects. Washington DC: IFC. International Finance Corporation, 2020. Green Bond Impact Report - Financial Year 2020. Washington DC: IFC.

Financing sustainable transport infrastructure  347

Intergovernmental Panel on Climate Change (IPCC), 2018. Global Warming of 1.5°C. IPCC Special Report. . Morgan Stanley. 2019. Decarbonization: The Race to Zero Emissions. New York. https://www​ morganstanley​.com​/ideas​/investing​-indecarbonization Mukhi, Neha, Suneira Rana, Sara Mills-Knapp and Eskedar Gessesse, 2020. World Bank Outlook 2050: Strategic Directions Note. Supporting Countries to Meet Long-term Goals of Decarbonization. Washington, DC: World Bank OECD Environment Policy Committee, 2020. Exploring the Impact of Shared Mobility Services on CO2. Working Party on Integrating Environmental and Economic Policies. Paris: OECD. OECD, World Bank, and UNEP, 2018. Financing Climate Futures. Paris: Organisation for Economic Co-operation and Development, World Bank, and United Nations Environment Programme. https:// doi​.org​/10​.1787​/9789264308114​-en World Bank, 2014. World Bank Group Support to PPPs: Lessons from Experience in Client Countries, FY02-12, Independent Evaluation Group. Washington, DC: World Bank. World Bank, 2017a. Mobile Metropolises: Urban Transport Matters, Independent Evaluation Group, Washington, DC: World Bank. World Bank, 2017b. Public–Private Partnerships – Reference Guide Version 3, International Bank for Reconstruction and Development, Washington, DC: World Bank. World Bank, 2019. World Bank Group Evaluation Principles. Washington, DC: World Bank. World Bank, 2020a. The International Finance Corporation’s Blended Finance Operations: Findings from a Cluster of Project Performance Assessment Reports. Independent Evaluation Group. Washington, DC: World Bank. World Bank, 2020b. World Bank Group Approaches to Mobilize Private Capital for Development. An Independent Evaluation. Independent Evaluation Group, Washington, DC: World Bank. World Bank, 2021a. The World Bank Group’s Experience with the IDA Private-Sector Window. An Early-Stage Assessment. Independent Evaluation Group. Washington, DC: World Bank. World Bank, 2021b. Climate Change Action Plan 2021-2025 – Supporting Green, Resilient and Inclusive Development. Washington, DC: World Bank.

18. Transport financing and regional development Javier Asensio and Anna Matas

18.1 INTRODUCTION Transport is a key element of regional development policies. The traditional objectives that governments try to achieve with transport policies are the promotion of regional growth and the guarantee of good accessibility levels. Governments rely on transport infrastructure investment for the former and on pricing policies and regulations for the latter. More recently, sustainability has been considered as an additional objective of transport policy at the regional level. This chapter will review different issues related to the funding of transport policies from a regional perspective. The second section deals with the role of infrastructure investment as a regional policy instrument. In Section 18.3, we review the role that accessibility can play to improve labour market results and the implications for transport policies. Section 18.4 is devoted to one of the core problems when financing regional infrastructures or pricing transport services: the interactions between different levels of government. We then deal with the long-debated issue of fighting poverty using transport policy instruments, before a final concluding section.

18.2 TRANSPORT POLICY IN NATIONAL POLICY AGENDAS The relevance of transport policy on government agendas reflects the widespread idea that improving transport infrastructures or services has a positive effect on economic development. Thus, investing in transport has usually been considered a priority by national, regional, and local governments around the world.1 Given their potential for economic growth, transport improvements are expected to contribute to reducing regional economic disparities and increasing territorial cohesion. From a policy viewpoint, providing equal access to citizens regardless of where they are located is an attractive goal. Hence, a significant amount of resources are being devoted to increasing the region’s infrastructure stock. One example is the European Union’s Cohesion Fund, which has devoted 52% of its available resources during the period 2014–2020 to network infrastructure investment (in transport and energy sectors).2 The positive relationship between infrastructure and growth is supported by both economic theory and empirical evidence. However, from a theoretical viewpoint, a consensus has not been reached about the mechanisms underlying such a relationship, while the difficulties plaguing the empirical estimations explain the wide range of results obtained in the literature. From a theoretical perspective, the production function approach considers that physical, human, and public capital are complementary factors. In this way, as infrastructure investment increases the stock of public capital, it contributes to rising productivity and, consequently, to output growth. See Calderón and Servén (2014) for a review of this analytical framework. From a different perspective, the New Economic Geography focuses on the role that transport 348

Transport financing and regional development  349

costs play on the location of activities in the context of economies of scale under monopolistic competition and horizontal product differentiation. By reducing travel costs, transport investments improve accessibility to input and output markets. As a consequence of having access to broader markets, firms can take advantage of economies of scale, which generates agglomeration economies and results in higher productivity (Fujita et al., 1999) Regarding the empirical evidence, a debate exists about the impact of transport infrastructure on economic development. The available literature provides results that range from almost negligible impacts to substantial positive effects. This diversity can be explained by different reasons, among which the difficulties underlying the econometric strategy stand out as the most relevant ones. First, a regression in levels between non-stationary variables may face a problem of spurious regression. Second, when estimating the impact of infrastructure on economic activity, reverse causation from the dependent variable to public investment may generate an upward bias in the estimated coefficient. If this is the case, the estimated impact of transport infrastructure on economic activity would reflect the fact that investment in infrastructure accrues to those regions with higher economic dynamism and higher potential for growth. Redding and Turner (2015) also point out the difficulties of an econometric model – usually a reduced form equation – to properly account for the complex path of impacts of transport infrastructure on growth and on the geographic distribution of economic activities. Additionally, the differences in the results can also be explained by the alternative methods used in the measurement and definitions of what constitutes transport infrastructures. Traditional monetary and physical measures are only crude proxies for changes in the stock of infrastructure. A better way to approximate their impact is to rely on accessibility measures that directly reflect the changes in a location’s attraction after a reduction of transport costs. Taking the previous issues into account, recent evidence confirms that improvements in transport can foster economic development by contributing to an increase in productivity. Nonetheless, the magnitude of the impact is below the one found in earlier studies and a high level of dispersion remains (See Melo et al., 2013; Calderón and Servén, 2014; Berg et al., 2017). Overall, the impacts will depend on a set of factors at the local level, among which the following stand out: the existence of a sufficient endowment of other production inputs, such as human capital and innovation capacity; the quality of the government and institutions; the productive structure of the region and its geographical characteristics; the characteristics of the project and the type of infrastructure in which the investment takes place.3 Regarding this last issue, returns are higher for road investments than for rail and airport investments. Additionally, there is evidence of a gradual decrease in output elasticity with respect to the stock of infrastructure over time (Fernald, 1999). A possible interpretation is that returns on investment are high in the early stages of development when infrastructures are scarce and progressively decline as the basic networks are completed (De la Fuente, 2010). However, Calderón et al. (2015), using a large cross-country dataset, find a high level of homogeneity in long-run output elasticities with respect to a set of variables such as countries’ populations, their level of income or their infrastructure endowments. In a context of an aggregate production function, these authors interpret that observed cross-country variations in the marginal productivity of infrastructure are driven by the variation in the ratio of infrastructure to output. Redding and Turner (2015), in a review of the literature that analyses the impact of transportation infrastructure on the organisation of economic activity, also conclude that such effects are similar across ranges of countries and levels of development. Given the previous

350  Handbook on transport pricing and financing

reasoning, the returns on investment will not depend so much on the endowment of infrastructure, but on its stock relative to the level of production. Therefore, positive effects can be expected wherever the investment contributes to effectively solving accessibility or bottleneck problems. The expected positive effects of investment in transport infrastructures have generated high levels of hope, and even enthusiasm, by governments trying to foster economic growth. However, investment and funding of transport projects are also linked to a series of potential inefficiencies and problems, so that too frequently their overall impact on the whole economy has not been as positive as initially expected. In what follows, we discuss the following issues: impacts can be merely a reallocation of economic activity; infrastructure investment does not necessarily contribute to regional convergence and efficiency losses may appear as a consequence of redistributive objectives. The first issue is related to the geographical scope of the analysis: although a reduction in transport costs as a result of improved infrastructure can induce local positive outcomes, this may be merely the result of the relocation of activities from another area. In that case, the net gain for the economy will be null. Empirically identifying whether the creation of new activities is a net increase or the result of displacement of activities from other zones has proved to be difficult and, hence, evidence is very scarce and conclusions are not definitive. Gathering the available evidence, Redding and Turner (2015) report, first, that for investments in within-city highways relocation of economic activity is at least as important as the generation of a new one. Second, in the case of investments in intercity highways, the primary effect seems to be to attract economic activity at the expense of more remote areas. Nevertheless, as Proost and Thisse (2019) point out, more research is needed to obtain clearcut answers to this issue. A second problem is that transport infrastructure investment may not effectively contribute to regional convergence, which is often regarded as a key objective from a regional perspective. The controversial issue is whether transport cost reductions result in a diffusion of economic activity to peripheral regions or, on the contrary, they reinforce the concentration of production in space. Improving accessibility between a peripheral region and a well-developed area can enhance the market size advantages of the latter leading to firms relocating to it. Faini (1983) finds that the reduction of transport costs between the North and the South of Italy in the 1950s contributed to accelerating the South’s deindustrialisation. On the other hand, infrastructures can contribute to reducing regional disparities if they facilitate firms’ relocation to developing regions with much lower input costs, improved access to markets or if, as a result of their construction, knowledge is more easily diffused and therefore contributes to reducing cost asymmetries between regions (Ottaviano, 2008). Whether the overall effect of improved transport infrastructure is to increase or decrease economic concentration will depend on both the project’s characteristics and those of the economic environment (Puga, 2002). A clear example and frequent subject of study are the effects of the EU’s infrastructure policy. The EU has devoted large sums of funds to territorial development granting financial support to lagging regions with inconclusive results in terms of the reduction of regional disparities. Teixeira (2006) shows that the substantial increase in investment in road transport in Portugal did not reduce regional imbalances. On the other hand, De la Fuente (2002) finds positive effects of the same policies for convergence among Spanish regions. More recently, Crescenzi and Guia (2020) report that the EU policy has achieved on the aggregate a positive effect on output and employment. However, the diversity of the results obtained suggests that

Transport financing and regional development  351

its effectiveness depends on a large set of local conditions, as already pointed out. Besides, these authors show that an average positive effect at the EU aggregate level can mask an uneven distribution of such effects for regions in different countries. As Fratesi and Wishlade (2017) emphasise, more research should focus on ‘the conditioning factors that explain where, when and how policy is effective’.4 The third type of problem is the potential efficiency losses of infrastructure projects that can arise when they are selected on pure distributional grounds. This is the case, for example, of infrastructure investments that would have higher economic benefits in more developed areas, but instead, take place in other regions, which are selected to allocate investment funds on distributional grounds. In these instances, the output of the whole economy will increase by less than it would have otherwise increased had the investment taken place in a more developed area. This reasoning does not imply that the returns on investment are necessarily higher in more developed regions; rather, infrastructure investment will lead to economic benefits only for those projects that relieve pressure due to bottlenecks and/or connect strategic parts of the network. In general, these criteria are more frequently satisfied in more dynamic areas. As previously explained, what matters is the relative level of infrastructure stock to output, which can be lower in more developed regions. The evidence of excessive redistribution in Spain, as highlighted by Solé-Ollé (2010) and De la Fuente (2004), alerts us about the risks of using public funds to finance pro-development policies. However, the existence of efficiency costs does not imply that infrastructure investment has to be ruled out as an instrument of regional policy. Lall et  al. (2014) argue that although transport investment in existing agglomerations will generate higher economic returns than in remote areas, investment in rural areas is more beneficial for the poor. As usual, achieving the right balance is the challenge. The inefficiencies observed in transport infrastructure investment cannot be disentangled from the way projects are financed. Assuming that infrastructure investment projects are welfare-improving, the existence of economies of scale in their supply justifies the use of public funds. Even if economies of scale are not significant, it may be the case that for those projects provided purely on distributional grounds the total willingness to pay for the project is lower than construction costs. Hence, public funds are necessary, but given that they generate an efficiency loss derived from the marginal cost of the required tax increases, the distortionary effects of such taxation in the economy have to be accounted for in the evaluation of any project. This requires specific assessments taking into account the broader context, as the marginal cost of public funds is higher in developing countries and rises as the budget constraints become tighter. If such cost is not considered, it can lead to overcapacity. De Rus (2017) provides a discussion of the elements that need to be assessed in a proper evaluation of infrastructure investment with a special emphasis on pricing and investment decisions in a multimodal context. The difficulties of raising funds to finance infrastructure arises as a major concern as budget constraints are tightened. In developing countries, where the lack of infrastructure can deter economic development, obtaining funds is particularly difficult. Hence, there may be a case for user financing. However, setting prices above social marginal costs will reduce demand and give rise to the corresponding welfare losses which have to be compared to the social cost of public funds. When user financing is an alternative, private involvement in the construction, financing and maintaining the infrastructure is common. If this is so, the necessary conditions

352  Handbook on transport pricing and financing

for a successful private involvement need to be settled. This issue is dealt with in Chapter 16 in this Handbook on the regulation of public-private partnerships. Finally, other inefficiencies may arise from potential political interferences in the project selection process (Knight, 2004; Kemmerling and Stephan, 2008; Burgess et al., 2015). Although they are more frequently linked to the use of public funds, the risk does not vanish when procurement is in private hands.

18.3 TRANSPORT INFRASTRUCTURE AND JOB ACCESSIBILITY The analysis of the relationship between accessibility and labour market outcomes, known as the ‘spatial mismatch hypothesis’, has its origin in Kain (1968), who studied the decentralisation of employment in US cities and showed how it led to a disconnection between residential and job locations. This was particularly problematic for Afro-Americans who continued to live in the inner city as housing market discrimination prevented them from relocating closer to their jobs. This research showed how the physical disconnection from jobs is a key factor in explaining persistent unemployment. Since then, many authors have extended the analysis by looking at the effects of residential and access characteristics on labour market outcomes. The different theoretical mechanisms that explain how disconnection from jobs results in poor labour market outcomes can be grouped into those related to the supply or the demand side (Gobillon et al., 2007). Supply-side explanations focus on job search efficiency and the impact of accessibility on reservation wages. Job search efficiency decreases with distance to jobs since information on job opportunities is lower for distant jobs. Additionally, the unemployed incur in higher transport costs as distance increases and, accordingly, may restrict their search to accessible areas. This effect may be particularly important for those dependent on limited public transport. Accessibility also has an effect through its impact on reservation wages, as potential employees will refuse job opportunities with commuting costs that are too high in relation to their offered wage. From the demand side, the impact would be due to workers commuting long distances being less productive and having a higher rate of absenteeism, which would result in employers being less willing to hire them. There have been many attempts to provide empirical evidence for the spatial mismatch hypothesis. As in other contexts, research in this area needs to deal with endogeneity issues that may be caused by simultaneity or by unobserved individual characteristics. Andersson et al. (2018) use very rich data on workers searching for jobs after mass lay-offs in different US states adjacent to the Great Lakes. Their results support the spatial mismatch hypothesis: better job accessibility significantly decreases the duration of unemployment among low-medium-income workers. The effect is especially important for blacks, women and older workers. One overall conclusion that emerges from research in this area is that transport policy can improve labour market outcomes by increasing accessibility, but not any infrastructure project will do so. Moreover, positive effects can be expected from investment in private transport infrastructure, but also by improving public transport provision. This last result is important in light of the negative externalities generated by the use of the private car and the fact that disadvantaged workers will benefit most from public transport, since they are usually more constrained to local labour market opportunities.

Transport financing and regional development  353

18.4 TRANSPORT PRICES IN A HIERARCHY OF REGIONAL GOVERNMENTS The optimal decisions related to the pricing of transport services and the provision of infrastructure capacity are to a large extent determined by the need to internalise the external effects of transport activities, such as congestion or environmental damages, but also by providing equity in accessibility. When transport networks are not limited to the responsibility of a single government, externalities may arise as a consequence of each authority’s taxation or spending actions. In this context, fiscal or expenditure externalities arise whenever decisions made by one government level have welfare effects on residents from another jurisdiction. Depending on the hierarchical structure of the governments involved, the externalities can be characterised as horizontal (between governments at the same level, each one responsible for a different region or country) or vertical (when there is a hierarchical relation between governments, as in a federal structure). De Borger and Proost (2012) provide four possible cases combining the horizontal/vertical governmental structure with the fiscal/expenditure nature of the externalities (for details, see Table 3.1 in Chapter 3 in this Handbook): ●

●

● ●

Horizontal fiscal externalities arise as taxation decisions have impacts on residents from other regions. Horizontal expenditure externalities imply that non-residents benefit from expenditure in other regions’ transport projects, as would be the case of an infrastructure providing accessibility benefits to non-residents. Vertical fiscal externalities imply that tax bases overlap between different governments. Vertical expenditure externalities appear when spending on a type of transport project, such as, for instance, road infrastructure, generates additional revenues for another government.

Identifying the optimal decisions to be taken in each case in the framework of a precise analytical model is particularly difficult due to the wide range of dimensions on which certain modelling assumptions would have to be made. An assessment of the effects of pricing and investment decisions in transport networks would have to consider the structure of the network (which can be parallel, serial or a combination of both), the intensity and geographical reach of congestion and environmental externalities, the interaction between local and through traffic, the available policy instruments (users’ tolls and/or investment in additional capacity, in the simplest cases), the aims of public agents (social welfare or net revenue maximisation), the kind of strategic interactions that take place between decision-making agents (typically resulting in Stackelberg or Nash equilibria) and even the complexity of the organisational structure of operators sharing a network: from a simple one with atomistic users to a multi-layered one with many different agents interacting in specific ways (Verhoef, 2008; De Borger and Proost, 2012). The problem, moreover, can be looked at from either a positive or a normative perspective. In this context, the normative question focuses on how sharing responsibilities among different governments can result in optimal outcomes, while the positive one analyses and explains potentially suboptimal outcomes emerging from a particular decision-making process or institutional mechanisms observed in public choices.

354  Handbook on transport pricing and financing

A good example of the range of different possibilities that tax exporting can take is provided by De Borger et  al. (2005), who model a network with two parallel roads, each one tolled by a different government. Users are a mix of local and through traffic, and governments try to maximise the sum of local users’ surplus plus toll revenues. When different tolls can be imposed on local and through traffic, tax exporting takes place in an obvious way: the toll charged to non-residents exceeds the one charged to residents. However, even if tolls are required to be equal for both user types, some kind of tax export also arises as tolls are then set above the marginal local external cost, the difference rising with the relative importance of through traffic. Finally, if only local traffic can be charged, the toll level is set below the marginal external cost generated by total traffic. This would be a subtle form of tax exporting, as it increases congestion above its optimal level and makes the use of the infrastructure less attractive for through traffic. In the case of a network with a serial structure, if no externalities are generated across jurisdictions the problem of setting tolls is equivalent to that of double marginalisation by concatenated monopolies in vertical markets. Tax externalities arise as travel takes place across borders, making it possible for governments to engage in tax exporting. One example is provided by international air travel, which is subject to airport charges set by different regulators. Benoot et al. (2013) show how non-cooperative behaviour by regulators results in too high airport charges, this being a potentially more important distortion than the one due to imperfect competition. Another example of tax exporting is provided by Levinson (2001), who finds that reliance on tolls by US states is directly related to the magnitude of demand for transport originating in other states. One area of interest to study tax competition in transport economics are the fuel taxes imposed by governments, which can be used to attract demand from neighbouring jurisdictions. Rietveld and van Woudenberg (2005) analyse the determinants of differences for diesel and gasoline taxes in a sample of 100 countries, but they find evidence of tax competition only in the case of small European countries, which charge lower taxes than their larger neighbours. Decker and Wohar (2007) and Nelson (2002) find no evidence of competition between US states when setting fuel taxes. A particularly interesting case where tax exporting may arise in transport is that of congestion charging schemes that limit access to urban centres. From a political economy perspective, De Borger and Russo (2018) model a city surrounded by a region whose residents differ by income, location and ability to access the city by private or public transport.5 One key result of this model is that support for road pricing depends on the use given to revenues, as well as on the geographical distribution of residents. When most people live outside the city and revenues are distributed uniformly across the whole population, the acceptable charge will be inefficiently low. However, as a higher share of toll revenues are devoted to subsidising public transport, the chosen toll level increases and tends to its optimal level. Empirical evidence on road pricing acceptability shows the difficulties that arise in obtaining support from suburban residents. In the case of Edinburgh, in 2005 the city council proposed a system based on two cordon tolls, with no charges for driving without crossing the cordons. Although revenues would be earmarked for improving the transport system and shared with neighbouring councils in proportion to the origins of paying trips (Laird et al., 2007), those councils considered that the cordon design unfairly favoured city residents and the proposal was rejected in a referendum. Stockholm provides another example of the different views held by city and suburban residents. There, after a one-year trial implementation, a road pricing scheme was approved in

Transport financing and regional development  355

2007 in a local referendum organised by the City of Stockholm and around half of the neighbouring municipalities that constitute the County of Stockholm. While a majority of voters in the city supported the scheme, those in the neighbouring municipalities mostly voted against it (Börjesson et al., 2012). The literature on tax competition shows that governments can benefit from tax exporting practices on private car users, such as tolling through traffic or collecting fuel taxes from non-residents. However, in the case of public transport services, which are most frequently provided at a loss, the results can be different. Hörcher et al. (2020) assess the effects of different institutional arrangements to provide public transport to suburban commuters who access the city centre. Although the service is assumed not to be available to urban residents, they indirectly benefit from its existence as it reduces car congestion and provides for higher levels of urban employment, which contributes to productivity improvements by means of agglomeration economies, resulting in higher wages. The authors show that, compared to the optimal provision levels of a federal welfare-maximising government, both the regional and the urban governments would prefer to provide public transport at higher fares and lower quality levels, as they only consider the interests of their local residents and need to keep costs under control. Low Emission Zones (LEZ) provide an example of horizontal expenditure externalities. These are typically urban areas to which access by high-polluting vehicles is restricted. They can be a source of externalities since their effects are potentially felt beyond their borders. Such effects can be positive, as they may induce the substitution of polluting vehicles, or negative, if the LEZ simply diverts traffic and pollution to the surrounding areas. In an assessment of the impact of LEZs in German cities, Wolff (2014) shows that the negative impacts have been negligible, while the impacts in terms of purchasing lower-emitting vehicles are felt beyond their borders. Börjesson et al. (2021) assess the welfare impacts of a proposed LEZ in Stockholm and observe that the costs imposed on light vehicle users outweigh the benefits, in the form of health benefits from air quality improvements.

18.5 TRANSPORT FUNDING AND URBAN AND RURAL POVERTY A controversial topic in transport economics is whether subsidies should be used to increase transport affordability to low-income groups. It could be argued that transport affordability is part of a wider problem of poverty and, as such, it should be addressed through income transfers funded with general taxation. Nonetheless, due to the significant limitations faced by first-best policies, it is a reality that transport is not affordable to all income groups. Therefore, subsidies need to be considered.6 Making transport available to low-income groups can improve their opportunities in the labour market but also their opportunities in terms of access to education and healthcare. Such effects can result not only in individual welfare gains, but also in higher economic development by increasing human capital and labour productivity. Moreover, it is often the case that efficient prices are not pro-poor with the corresponding implications on income distribution. Subsidies for equity reasons are present all over the world, but they are more common in developed and rich countries, although the lack of transport affordability is more acute in low-income countries. This may be because subsidies are costly and taxation is politically difficult to implement (Berg et al., 2017). Equity subsidies are used in both urban and rural contexts. Although it is not always the case, given that public transport use decreases with the level of income, subsidies favour

356  Handbook on transport pricing and financing

low-and-medium-income groups. In rural areas, the problem is the lack of sufficient demand to guarantee commercially viable services. Given that governments deem that transport accessibility has to be guaranteed, the services’ operators may receive support under different forms, such as direct subsidies, tax breaks, preferential loans or cross-subsidisation from other profitable services. Given the level of government preference for equity,7 the distributional impact of subsidies will depend on who actually benefits – to whom the subsidy is targeted – and on who pays – how is the subsidy funded. In what follows we review which are the main options to channel the subsidies and the most common sources of revenue used, together with the corresponding implications on redistribution and efficiency. Essentially, there are two policy instruments to channel subsidies: subsidising operators providing transport services (supply-side subsidies) and subsidising specific groups of passengers (demand-side subsidies). Subsidising public transport operators for redistributive reasons is justified by the assumption that the poor use public transport more intensively than the rich. This is usually the case in cities in developed countries where it is common for public transport to absorb a higher share of household expenditures as the level of income decreases. However, the correlation between public transport use and income is far from being perfect, as sometimes the subsidies accrue more than proportionally to well-off people. This problem is more severe for developing countries where the relationship follows an inverted U-shape curve: very low-income households spend less on transport than low-or middle-income ones because for them transport is not affordable and walking is their predominant mode (Serebrisky et al., 2009). Available evidence suggests that in cities in more developed countries supply-side subsidies have progressive or mildly-progressive impacts, whereas evidence from poorer countries indicates that there is scope for policies to be redesigned so as to make the poorest better off. Venter et al. (2018) analyse the equity impacts of Bus Rapid Transit (BRT) systems in Africa, Asia, and Latin America, and conclude that although they offer significant benefits for poorer segments of the population, benefits are often skewed towards medium-income users. This result is explained mainly by their insufficient spatial coverage and inappropriate fare policies. The Transmilenio BRT system implemented in Bogotá is particularly interesting as it extended its coverage of trunk services into peripheral low-income areas so that it provided equal access to the local population. However, according to Teunissen et al. (2015), lack of affordability remained an issue for poor groups. There is widespread evidence of the undesired redistributive effects of subsidies granted to expensive rail or metro investments. It is often the case that these transport modes are used by relatively high-income groups. Examples can be found in Mexico City (Serebrisky et al., 2009); Santiago and Buenos Aires (Gwilliam, 2017); Oslo (Fearnley and Aarhaug, 2019); and in US cities (Taylor and Morris, 2015), among others. Subsidies can be more effective if they are targeted at beneficiaries (demand-side subsidies), as they can be made available to specific groups (senior citizens, students, unemployed) or linked to some income variable. Recent literature surveyed by Gandelman et al. (2019) suggests that equity policy can be improved by moving to demand-side narrowly targeted subsidies. Smart card technology makes it easier to differentiate fares according to income groups, specific locations, or trip categories. Although the most common cases are those based on individual categories, it is increasingly frequent to use some kind of means-tested instrument

Transport financing and regional development  357

to select the potential beneficiaries. Examples can be found both in urban and interurban contexts, as explained in what follows. Brazil’s ‘Vale Transporte’ (VT) is a programme which was made compulsory for all companies in 1987, with the objective of incentivising commuting by public transport. Employees receive a monthly transport voucher for the work trip, for which employers are entitled to deduct 6% of the worker’s earnings. An interesting feature of this system is that employees can opt-out, and since higher income earnings have the incentive to do so as 6% of their salary will be more than their commuting costs, the policy effectively targets low-income workers. The system is basically progressive, though the very poorest, self-employed, or employed in the informal sector do not receive the subsidy (Gwilliam, 2017). A relevant example of a means-tested subsidy is the pro-poor public transport system implemented in Bogotá aimed at solving the problem of lack of affordability mentioned above in relation to Transmilenio. As fares for public transport are designed to cover all operating costs, they become too expensive for poor citizens. In 2014, the local government introduced a pro-poor public transport subsidy with potential beneficiaries targeted according to a scoring scheme. Guzman and Oviedo (2018) show that the subsidy, which is delivered through a personalised smart card providing beneficiaries with a discount ranging from 50 to 60%, is progressive. A specific type of demand subsidy is targeted at residents in isolated or remote areas. Such support can complement the imposition of public service obligations programmes on the transport operators, which are common in many countries. This case illustrates that demandside subsidies can also generate distortions. Focusing on the air transport subsidies addressed to island residents in France, Greece, Italy, Spain and Portugal, Fageda et  al. (2017) find that prices in routes where only island residents benefit from subsidies are higher than those where subsidies do not discriminate between residents and non-residents (such as the ones that apply to connections to Sardinia from the Italian mainland outside the summer season). However, the market distortions depend on the intensity of competition in the market and the specific design of the subsidy. Valido et al. (2014) compare the impact of ad valorem versus unit subsidies in a context in which airlines exert market power. They show that the impact of each design type depends on the proportion of passengers who benefit from the subsidy and their relative willingness to pay for the trip. Using data from the Canary Islands market, they conclude that a unit-based subsidy would increase welfare more than the existing ad valorem one. Any analysis of the impact of subsidies also needs to take into account the revenue sources used to fund them, among which the most common is general taxation. In this case, the net redistributive effect will depend on the degree of progressivity of the tax system. In developing countries tax systems tends to be regressive, which reinforces the undesired negative effects of supply-side subsidies. Additionally, the assessment of the subsidy has to account for the efficiency loss derived from raising additional taxes, as discussed in Section 18.1. For an in-depth discussion of equity issues see Chapter 6 in this Handbook. Given increasingly tighter budget constraints, transport authorities have been forced to draw from alternative taxation sources. Among them, financing public transport using congestion toll revenues has become an attractive option for those regions where such tolls exist. There is some research on the efficiency and equity consequences of using toll revenues to reduce labour income tax that compares it with using revenues to subsidise public transport. Mayeres (2001) assesses the changes in welfare and equity from alternative transport policies

358  Handbook on transport pricing and financing

simulating the conditions of urban areas in Belgium (where no congestion pricing exists) and she shows that it will always be preferable to use congestion toll revenues to reduce labour income tax than to increase public transport subsidies. This result is explained by the high distortionary costs of taxes on the economy. However, the deadweight loss from labour taxes can be partially counterbalanced by the reduction in commuting costs which in turn might increase the supply of labour. Some models have considered the labour market effects of congestion with inconclusive results. Parry and Bento (2001) conclude that public transport subsidies incentivise labour supply and reduce the deadweight loss of labour taxes, but using congestion toll revenues to provide more public transport subsidies is less efficient than directly lowering the labour tax. On the contrary, under different assumptions, De Borger and Wuyts (2009) and Tikoudis et al. (2015) show that recycling congestion tax revenues to subsidise public transport can be more efficient than using them to compensate for reductions in labour taxes. It seems, therefore, that the final effect will be context-specific and will depend, among other factors, on (i) how substitute transport modes are priced, (ii) the effect of subsidies on congestion, and (iii) the wage elasticity of labour supply. A second mechanism for alternative taxes being used to pay for public transport costs is a payroll tax levied on salaries. One relevant example is the French Versement Mobilité (previously named Versement Transport), which consists of a local payroll tax levied on companies with 11 or more employees, whose revenue is directed to local transport authorities and used to finance public transport. As a payroll tax, it might reduce total employment with the consequent efficiency loss and a probable regressive effect. However, such an effect can be counterbalanced if better public transport enlarges the size of the labour market and encourages employment creation. In any case, the net redistributive effects are unclear. It has to be pointed out that this tax on labour is used to finance trips for any purpose. A third alternative to fund loss-making services is to cross-subsidise them with revenues obtained from commercially profitable ones. Cross-subsidies are a mechanism used to support transport routes considered to be of general interest. Consumers of profitable services end up paying prices above their long-run marginal cost in order to sustain the losses of unprofitable ones. Compared to the marginal cost pricing alternative, such pricing policy does not increase overall welfare (Beato, 2002), even if it pursues a redistributive aim in regional terms. Until 1997, Norway allowed airlines connecting remote regions to operate as monopolies and apply cross-subsidies to the different routes (Fageda et al., 2018), a system similar to the one that is currently applied by Spanish intercity buses. The four maritime cargo companies allowed to serve the Azores Islands also operate cross-subsidies, as did Greek ferry operators which until 2015 levied a 3% surcharge on fares of profitable routes in order to cover the losses of non-profitable ones (ITF, 2021). A particular form of cross-subsidy is the use of a flat fare scheme, which increases accessibility for those users located on the outskirts of the cities. With such a system, short trips pay a price higher than their operating cost, whereas long-distance trips receive a subsidy.8 Over and above the welfare losses for those travelling short distances, flat fares can incentivise suburbanisation and excess commuting. The final redistributive effects will depend on the location of the population and the pattern of travel by income groups. Another mechanism for funding transport services is known as ‘value capture’, which consists of transferring to the transport system part of the increases in property market values that arise as an area becomes more accessible (Yang et al., 2019). Earmarking such property taxes

Transport financing and regional development  359

as public transport subsidies is relatively common in North American cities, where its implementation subject to voter approval makes it explicit that the alternative is to use general tax revenues (Ubbels et al., 2001). A variant of this idea is to apply development levies when new areas are being developed, so that they contribute to funding the transport services that will be demanded by the new residents. Value capture is discussed in more detail in Chapter 15 in this Handbook.

18.6 CONCLUSIONS Transport policies contribute to several dimensions of regional development, among which three stand out. First, by improving network connectivity, less developed areas may attract new investment and economic growth. Second, better transport accessibility can have significant labour market effects for disadvantaged groups and, at the same time, contribute to improving their access to education and healthcare. And third, it should be borne in mind that besides providing accessibility, transport should also be affordable. Therefore, subsidies directly granted to targeted groups can be defended both on efficiency and distributional grounds. However, the effectiveness of transport policies cannot be taken for granted. Too frequently the results of transport policies are far from what was expected. Resources devoted to the promotion of lagging areas may yield businesses relocation or concentration of activities in the most prosperous ones. The final results depend greatly on local factors and specific project characteristics. In the case of subsidies, a correct design of their transfer mechanisms is crucial to limit inefficiency costs. All the distortions generated by policy interventions should be included in the project evaluation stage. The opportunity cost of public funds to finance transport projects or subsidies increases as budget constraints are tightened. Besides, the use of public funds may create the opportunity for political or interest groups interferences in the project selection. As a result, there may be overinvestment or selection of projects that favour other aims than efficiency or distributional goals. No simple solution exists for this problem, although earmarking revenues (such as congestion tolls) can be an alternative under some circumstances. Charging users is another option, but it will reduce demand and, consequently, it translates into a welfare loss. The different alternatives need to be assessed on a case-by-case basis. An additional problem arises as more than one government level is involved in the design and execution of transport policy. Tax competition and tax exporting become relevant issues when transport infrastructure has to be priced or financed by different governments. The fact that transport networks are increasingly integrated and used by residents and firms from different jurisdictions shows the need to coordinate policies at a broader level. Economic principles should guide transport policy decisions within a broad enough framework that considers all the trade-offs between costs and benefits. However, this requires an institutional design with the right incentives to guarantee an independent and rigorous evaluation.

NOTES 1.

Given that major transport improvements originate from investment in infrastructure, we identify improvement with infrastructure investment. Nevertheless, the same effects can be reached by improvement in services.

360  Handbook on transport pricing and financing

2. https://cohesiondata​.ec​.europa​.eu​/funds​/cf [last accessed Feb 22nd 2022]. 3. See, for instance Crescenzi and Rodriguez-Pose (2012). 4. On this topic see the papers included in the special issue ‘European Cohesion Policy in Context’, Regional Studies, 51, 6, 2017. 5. See Chapter 7 in this Handbook on the political economy of road pricing. 6. For a more in-depth analysis of affordability in Latin America see Chapter 22 in this Handbook. 7. Basso and Silva (2014) show that if distributional concerns are sufficiently high, optimal transit subsidisation becomes an imperfect means to achieve equity objectives. 8. However, when time costs are taken into account, the effects of a flat fare are not so clear-cut. For instance, consider a bus route that gets more crowded as it approaches the city centre, increasing the time costs for all bus users. In this case, it may be efficient to set a higher fare for short trips in order to discourage them (we thank Andrés Gómez-Lobo for raising this issue).

REFERENCES Andersson, F., Haltiwanger, J. C., Kutzbach, M. J., Pollakowski, H. O. & Weinberg, D. H. (2018). Job displacement and the duration of joblessness: The role of spatial mismatch. Review of Economics and Statistics, 100(2), 203–218. Basso, L. J. & Silva, H. E. (2014). Efficiency and substitutability of transit subsidies and other urban transport policies. American Economic Journal: Economic Policy, 6(4), 1–33. Beato, P. (2002). Cross-subsidy prices in public utilities. In Beato, P. & Laffont, J-J. (eds.), Competition Policy in Regulated Industries. Approaches for Emerging Economies. Washington, DC: InterAmerican Development Bank, 199–218. Benoot, W., Brueckner, J. K. & Proost, S. (2013). Intercontinental-airport regulation. Transportation Research Part B: Methodological, 52, 56–72. Berg, C., Deichmann, U., Liu, Y. & Selod, H. (2017). Transport policies and development. The Journal of Development Studies, 53, 465–480. Börjesson, M., Eliasson, J., Hugosson, M. B. & Brundell-Freij, K. (2012). The Stockholm congestion charges—5 years on. Effects, acceptability and lessons learnt. Transport Policy, 20, 1–12. Börjesson, M., Bastian, A. & Eliasson, J. (2021). The economics of low emission zones, Transportation Research Part A, 153, 99–104. Burgess, R., Jedwab, R., Miguel, E., Morjaria, A. & Padró i Miquel, G. (2015). The value of democracy: Evidence from road building in Kenya. American Economic Review, 105(6), 1817–1851. Calderón, C. & Servén, L. (2014). Infrastructure, growth, and inequality: An overview. Policy Research Working Paper, 7034, World Bank Group, Washington DC. Calderón, C., Moral-Benito, E. & Servén, L. (2015). Is infrastructure capital productive? A dynamic heterogeneous approach. Journal of Applied Econometrics, 30, 177–198. Crescenzi, R. & Rodríguez‐Pose, A. (2012). Infrastructure and regional growth in the European Union. Papers in Regional Science, 91(3), 487–513. Crescenzi, R. & Giua, M. (2020). One or many Cohesion Policies of the European Union? On the differential economic impacts of Cohesion Policy across member states. Regional Studies, 54(1), 10–20. De Borger, B. & Proost, S. (2012). Transport policy competition between governments: A selective survey of the literature. Economics of Transportation, 1(1–2), 35–48. De Borger, B. & Russo, A. (2018). The political economy of cordon tolls. Journal of Urban Economics, 105, 133–148. De Borger, B. & Wuyts, B. (2009). Commuting, transport tax reform and the labour market: Employerpaid parking and the relative efficiency of revenue recycling instruments. Urban Studies, 46(1), 213–233. De Borger, B., Proost, S. & Van Dender, K. (2005). Congestion and tax competition in a parallel network. European Economic Review, 49(8), 2013–2040. De la Fuente (2002). The effects of structural funds spending on the Spanish regions: And assessment of the 1994-99 Objective 1 CSF. CEPR Discussion Paper, 3673.

Transport financing and regional development  361

De la Fuente, A. (2004). Second-best redistribution through public investment: A characterization, an empirical test and an application to the case of Spain. Regional Science and Urban Economics, 34(5), 489–503. De la Fuente (2010). Infrastructures and productivity: An updated survey. Working Papers, BBVA Research, 10/18. De Rus, G. (2017). The economic evaluation of major infrastructure projects. In Bel, G. & Albalate, D. (eds.), Evaluating High-Speed Rail: Interdisciplinary Perspectives, 7–22. London: Routledge. Decker, C. S. & Wohar, M. E. (2007). Determinants of state diesel fuel excise tax rates: The political economy of fuel taxation in the United States. The Annals of Regional Science, 41(1), 171–188. Fageda, X., Jiménez, J. L. & Valido, J. (2017). An empirical evaluation of the effects of European public policies on island airfares. Transportation Research Part A: Policy and Practice, 106, 288–299. Fageda, X., Suárez-Alemán, A., Serebrisky, T. & Fioravanti, R. (2018). Air connectivity in remote regions: A comprehensive review of existing transport policies worldwide. Journal of Air Transport Management, 66, 65–75. Faini, R. (1983). Cumulative processes of de-industrialisation in an open region: The case of Southern Italy, 1951–1973. Journal of Development Economics, 12(3), 277–301. Fearnley, N. & Aarhaug, J. (2019). Subsidising urban and sub-urban transport–distributional impacts. European Transport Research Review, 11(1), 1–10. Fernald, J. (1999). Roads to prosperity? Assessing the link between public capital and productivity. American Economic Review, 89(3), 619–638. Fratesi, U. & Wishlade, F. G. (2017). The impact of European Cohesion Policy in different contexts. Regional Studies, 51, 817–821. Fujita, M., Krugman, P. R. & Venables, A. (1999). The Spatial Economy: Cities, Regions, and International Trade. MIT Press. Gandelman, N., Serebrisky, T. & Suárez-Alemán, A. (2019). Household spending on transport in Latin America and the Caribbean: A dimension of transport affordability in the region. Journal of Transport Geography, 79, 102482. Gobillon, L., Selod, H. & Zenou, Y. (2007). The mechanisms of spatial mismatch. Urban Studies, 44(12), 2401–2427. Guzman, L. A. & Oviedo, D. (2018). Accessibility, affordability and equity: Assessing ‘pro-poor’ public transport subsidies in Bogotá. Transport Policy, 68, 37–51. Gwilliam, K. M. (2017). “Transport Pricing and Accessibility.” Moving to Access/Brookings, Washington DC. Available at: https://www.brookings.edu/wp-content/uploads/2017/07/pricing-andaccessibility-paper_web.pdf Hörcher, D., De Borger, B. & Graham, D. J. (2022). Subsidised transport services in a fiscal federation: Why local governments may be against decentralised service provision. Working paper, Imperial College London. ITF (2021). Connecting Remote Communities: Summary and Conclusions, ITF Roundtable Reports, 179, OECD, Paris. Kain, J. F. (1968). Housing segregation, negro employment, and metropolitan decentralization. The Quarterly Journal of Economics, 82(2), 175–197. Kemmerling, A. & Stephan, A. (2008). The politico-economic determinants and productivity effects of regional transport investment in Europe. EIB Papers, 13(2), 36–60. Knight, B. (2004). Parochial interests and the centralized provision of local public goods: Evidence from congressional voting on transportation projects. Journal of Public Economics, 88(3–4), 845–866. Laird, J., Nash, C. & Shepherd, S. (2007). Cordon charges and the use of revenue: A case study of Edinburgh. Research in Transportation Economics, 19, 161–187. Lall, S. V., Schroeder, E. & Schmidt, E. (2014). Identifying spatial efficiency–equity trade-offs in territorial development policies: Evidence from Uganda. The Journal of Development Studies, 50(12), 1717–1733. Levinson, D. (2001). Why states toll: An empirical model of finance choice. Journal of Transport Economics and Policy, 35(2), 223–237. Mayeres, I. (2001). Equity and transport policy reform. KU Leuven CES; Leuven Working Papers, 1–38. Melo, P. C., Graham, D. J. & Brage-Ardao, R. (2013). The productivity of transport infrastructure investment: A meta-analysis of empirical evidence. Regional Science and Urban Economics, 43(5), 695–706.

362  Handbook on transport pricing and financing

Nelson, M. A. (2002). Using excise taxes to finance state government: Do neighboring state taxation policy and cross‐border markets matter? Journal of Regional Science, 42(4), 731–752. Ottaviano, G. I. (2008). Infrastructure and economic geography: An overview of theory and evidence. EIB Papers, 13(2), 8–35. Parry, I. & Bento, A. (2001). Revenue recycling and the welfare effects of road pricing. The Scandinavian Journal of Economics, 103, 645–671. Proost, S. & Thisse, J. F. (2019). What can be learned from spatial economics? Journal of Economic Literature, 57(3), 575–643. Puga, D. (2002). European regional policies in light of recent location theories. Journal of Economic Geography, 2(4), 373–406. Redding, S. J. & Turner, M. A. (2015). Transportation costs and the spatial organization of economic activity. Handbook of Regional and Urban Economics, 5, 1339–1398. Rietveld, P. & van Woudenberg, S. (2005). Why fuel prices differ. Energy Economics, 27(1), 79–92. Serebrisky, T., Gómez‐Lobo, A., Estupiñán, N. & Muñoz‐Raskin, R. (2009). Affordability and subsidies in public urban transport: What do we mean, what can be done? Transport Reviews, 29(6), 715–739. Solé-Ollé, A. (2010). The determinants of the regional allocation of infrastructure investment in Spain. In N. Bosch, M. Espasa, & A. Solé-Oller (eds.), The Political Economy of Inter-Regional Fiscal Flows, 297–319. Cheltenham: Edward Elgar. Taylor, B. D. & Morris, E. A. (2015). Public transportation objectives and rider demographics: Are transit’s priorities poor public policy? Transportation, 42(2), 347–367. Teixeira, A. C. (2006). Transport policies in light of the new economic geography: The Portuguese experience. Regional Science and Urban Economics, 36(4), 450–466. Teunissen, T., Sarmiento, O., Zuidgeest, M. & Brussel, M. (2015). Mapping equality in access: The case of Bogotá's sustainable transportation initiatives. International Journal of Sustainable Transportation, 9(7), 457–467. Tikoudis, I., Verhoef, E. T. & Van Ommeren, J. N. (2015). On revenue recycling and the welfare effects of second-best congestion pricing in a monocentric city. Journal of Urban Economics, 89, 32–47. Ubbels, B., Nijkamp, P., Verhoef, E., Potter, S. & Enoch, M. (2001). Alternative ways of funding public transport. European Journal of Transport and Infrastructure Research, 1(1), 73–89. Valido, J., Socorro, M. P., Hernández, A. & Betancor, O. (2014). Air transport subsidies for resident passengers when carriers have market power. Transportation Research Part E, 70, 388–399. Venter, C., Jennings, G., Hidalgo, D. & Valderrama Pineda, A. F. (2018). The equity impacts of bus rapid transit: A review of the evidence and implications for sustainable transport. International Journal of Sustainable Transportation, 12(2), 140–152. Verhoef, E. (2008). Toll competition in transport networks. Introduction. Journal of Transport Economics and Policy, 42(3), 361–366. Wolff, H. (2014). Keep your clunker in the suburb: Low‐emission zones and adoption of green vehicles. The Economic Journal, 124(578), F481–F512. Yang, L., Chau, K. W. & Chu, X. (2019). Accessibility-based premiums and proximity-induced discounts stemming from bus rapid transit in China: Empirical evidence and policy implications. Sustainable Cities and Society, 48, 101561.

PART IV REGIONAL PERSPECTIVES

19. Road transport pricing and financing in Africa Leonard Mwesigwa, Moez Kilani and Matti Siemiatycki

19.1 INTRODUCTION Although roads are the predominant mode of transport in Africa, carrying at least 80% of goods and 90% of passengers (AfDB, 2014), major deficits exist in transportation infrastructure throughout the continent. The road network is largely unpaved, with insufficient coverage that isolates people from accessing basic services and economic opportunities. In a period of rapid advancement worldwide, Africa’s transport infrastructure needs remain significant and efficient road pricing policies will go a long way in meeting some of these needs. Road pricing, which has been implemented in several countries around the world, is guided by at least two motivations: to raise critical funds for road infrastructure construction and operations based on the user-pay principle, and to manage demand in congested areas (Newbery et al., 1988; Lishman, 2013; Van Rensburg & Krygsman, 2019). With 54 countries that are highly diverse with respect to their social heritage, economic activities, and stock of road networks and infrastructure, Africa presents a unique context for road pricing. It is home to an estimated total population of 1.3 billion people, representing 16% of the world’s population. According to a United Nations report (UN, 2019), the population of sub-Sahara Africa is projected to double by 2050. Urbanization and a rising middle class are trends that are picking up speed in many African countries, yet poverty remains deeply rooted, which challenges the equity dimensions of charging for road mobility. To date, road pricing has not been widely used across the continent even when it is known to improve overall economic efficiency as it is used as a method of resource allocation. The provision of transport services involves costs that need to be reflected in the prices charged for these services. This is however a thorny problem of political economy and the need to measure and understand its distributional impacts across different groups is critical (see Chapter 6 in this Handbook on equity distributional issues of road pricing). In terms of transport financing, the public sector has been the main source of financing new transport infrastructure and maintenance in Africa (Wentworth & Makokera, 2015), followed by Official Development Finance (ODF) from the Organisation for Economic Co-operation and Development (OECD) and non-OECD countries, especially China, with the private sector contributing the least amount (Quium, 2019). It is worth noting here that how infrastructure is priced, financed, funded, and governed is integral to explaining socially and spatially uneven infrastructural provision and its development ramifications (O’Brien et al., 2019). This chapter examines how the theory of road pricing applies in the African context, examines the experiences with road pricing to date, and explores the equity concerns of the current road pricing mechanisms used in Africa.  It discusses different transport financing mechanisms including public-sector financing, official development financing from multilateral banks, Chinese development financing, and public–private partnerships (PPPs). It further looks at the current situation of informal public transportation and how the sub-sector can in 364

Road transport pricing and financing in Africa  365

part be supported from the financing perspective as well as building institutional capacity to efficiently manage the sector.

19.2 ROAD PRICING Road pricing performs three key functions, that is, it rations and allocates the use of scarce resources; it provides a signal on the need for, and viability of, investment; and it helps in generating funds for the development of the related sectors. Road pricing studies are very limited in many African countries except in South Africa (Lishman, 2013; Van Rensburg & Krygsman, 2019; Venter & Joubert, 2014). In a study to understand the principles of fair and efficient road user charges in South Africa by Van Rensburg and Krygsman (2019), they observed that while the potential development role of road infrastructure is well understood, funding for roads is controversial, faces many conflicting viewpoints, and is notoriously complex. This situation is amplified in Africa as many countries continue to face numerous developmental needs, limited revenue opportunities, and a relatively small road user base (Van Rensburg & Krygsman, 2020). The difficulty to set efficient prices for road use stems mainly from their consideration as public goods and their very nature of indivisibility. Measuring individual use of a road is very difficult and the full cost is often underestimated due to gross mispricing of many externalities such as road crashes, congestion, noise pollution, particulate matter emissions, and greenhouse gas emissions. Road pricing is widely promoted as a tool to reduce these externalities. It is worth noting that many governments in Africa have for a long time relied on fuel levies (Newbery et al., 1988) and recently road tolls to generate funds for road infrastructure development and maintenance. 19.2.1 Fuel Levy In South Africa, a fuel levy was introduced in 1935 with the creation of the National Road Fund, as a way to fund the construction of national roads (Van Rensburg & Krygsman, 2016). Similarly, other African countries have used fuel levies to not only fund road transport infrastructure but also other developmental needs, such as education, security, and health services. There are variations in the amount of fuel levy charged across the continent. For example, in Uganda, fuel tax is about 35% of the pump fuel costs of about $1.18 per liter of gasoline and $92 cents per liter of diesel oil (KPMG, 2021). This generates slightly over half a billion US dollars in total revenue each final year according to the Uganda Revenue Authority database. While in South Africa, the general fuel levy is about 25% of the pump fuel costs of about $1.20 per liter of gasoline. Regardless of the variations in percentages, these taxes form a significant component of the direct road-generated revenue across the continent. According to Lishman (2013), the fuel levy remains popular due to its simplicity and costefficiency as tax avoidance is difficult and administrative costs are low. He further notes that users pay it for every kilometer traveled; the charge varies (to a limited extent) with the nature of the vehicle; it varies with the speeds at which vehicles travel and the manner in which they are driven; and it is easy to differentiate the charge between petrol and diesel-driven vehicles, which allows for the proxying, to a degree, of passenger vehicles and heavy vehicles respectively (citing Pienaar, 2005). Proponents of fuel levies argue that there is a relationship between fuel tax and road usage, as well as public acceptability since it is paid in small

366  Handbook on transport pricing and financing

quantities at frequent intervals thus providing an attractive basis for a road-use charging system (Lishman, 2013). However, the above-claimed relationship ignores the appropriate pricing of externalities. According to Van Rensburg and Krygsman (2016), a kilometer-based road user charge (KBRUC) system provides a viable alternative that addresses many of the problems associated with the fuel levy. They further note that the system could entail an onboard global positioning system (GPS) enabled device to be fitted in a road user’s vehicle where vehicle movement data can be collected to generate a road-use invoice at a set charge per kilometer traveled. Furthermore, Van Rensburg and Krygsman showed that a vehicle tracking experiment configured on a kilometer-based road user charge is operationally feasible on a small scale. Therefore, the continuing reliance on the fuel levy to generate sufficient income is being questioned due to a decrease in the average amount of fuel sold per vehicle per annum (Lishman, 2013; Van Rensburg & Krygsman, 2016). This is attributed to vehicle technological advancements that include more fuel-efficient vehicles, the introduction of electric and hybrid vehicles and alternative fuels as well as societal trends including telecommuting. 19.2.2 Road Tolling Many African countries have announced plans and enacted new legislation geared toward road tolling in recent years. This is intended to generate the necessary funds to repay capital costs and provide adequate maintenance of the infrastructure. Road pricing can significantly help in boosting domestic resource mobilization. Zhou and Chilunjika (2013) noted, while studying the effectiveness of tollgate systems in Zimbabwe, that there is a global shift to tolling systems as alternative means of mobilizing domestic revenue for the development and maintenance of road infrastructure. They further note that prudent management of the tolls collected to avoid leakages and dedicating such funds to road improvement motivates road users to pay.1 It is again South Africa that has the most kilometers of toll roads. According to Venter and Joubert (2014), 19% of 16,421 km of South Africa’s national intercity roads are operated as toll roads, mostly as concessionary schemes with private-sector partners (public–private partnerships are discussed in Chapter 16 of this Handbook). Some of these schemes include the 185-km Gauteng Freeway connecting Johannesburg and Pretoria and the N4 Toll Route (630 km) running from Pretoria to Maputo, the capital of Mozambique. Other countries with operational road toll schemes include Nigeria, Senegal, Tunisia, Uganda, and a few others. Compared with fuel taxes, electronic tolling is considered more progressive in regards to income and vehicle class because such tolling transfers costs from private to commercial vehicles, in line with the greater pavement damage caused by trucks (Venter & Joubert, 2014). With the increased participation of private-sector players in building road infrastructure in Africa, more road toll schemes are likely to be implemented by many countries across Africa. With the advancements in electronic tolling technology coupled with increased congestion levels in big cities in Africa, road tolling schemes are more likely to succeed. Those opposed to toll road schemes have insisted that charging users for road access amounts to paying the state for an existing resource that has, ostensibly, already been funded using fuel levies and vehicle licensing fees. As a result, the introduction of a road toll on an existing road is viewed as a tax increase or “double payment” (Lishman, 2013). There has been strong criticism and protests in Senegal over the tolls on the 20.4 km Dakar–Diamniadio toll road and similar protests led to the termination of the Lekki toll road concession in Lagos, Nigeria (ITS, 2016).

Road transport pricing and financing in Africa  367

In Africa, road users pay various taxes and charges as a result of owning and operating a vehicle. According to Van Rensburg and Krygsman (2019), these charges can be divided into direct or indirect fees. Direct fees refer to charges that are the result of the actual use of the network and include fuel levies and toll fees. Indirect fee sources are general taxes related to vehicle ownership and vehicle sales (number of vehicles sold and imported). In South Africa, road-generated income sources include fuel levies, road accident funds, fines/fees, permits (license fees), toll fees, carbon emissions fees, Pipeline levies, value added tax on vehicle sales, import duties, custom and excise levy, and demand-side management levies (DSML – a fuel levy introduced in 2006 to curtail the use of Unleaded Petrol 95 in the inland market) (Van Rensburg & Krygsman, 2019). 19.2.3 Road Pricing Equity Concerns According to Venter and Joubert (2014), as user charging (e.g., tolling) increasingly supplements taxation (fuel levies) as a transport pricing mechanism, the need to measure and understand its distributional impacts across affected groups grows more critical. Using the 185-km Gauteng Freeway Improvement Project as a case study to assess the equity impacts of tolling, Venter and Joubert found that fuel tax is more regressive (in regards to income) than electronic tolling; lowest-income users pay 20% of taxes while contributing only 13% of freeway kilometers. While drivers in the highest-income bracket pay 18% of taxes while contributing 23% of freeway kilometers. But they further noted that fuel taxation is progressive with respect to vehicle classes. For example, heavy commercial vehicles as a class, being far less fuel-efficient than light vehicles, contribute a greater share of tax revenue than their share of freeway usage demands. This higher charge might be justified, however, if pavement damage is taken into account, as trucks contribute far more to pavement damage than do light vehicles (Venter & Joubert, 2014). Although fuel tax is the most common fiscal policy instrument to raise government revenues, it further contributes to reducing traffic congestion and emissions thereby benefiting low-income public transport users. It is therefore paramount that all transportation policies, including pricing regulations, be checked for equity and distributional impacts across different categories of people, especially the most disadvantaged poor. Road tolling schemes should be assessed for equity effects early in the planning process especially for communities that are directly affected by the scheme, for example, through displacements as well as pollution to the neighborhoods within the proximity of the scheme. In cases in which unacceptable equity effects are found, mechanisms such as public transport investment or toll exemptions are universally accepted as a means of compensating affected populations (Venter & Joubert, 2014). In South Africa (Gauteng Freeway), the welfare impact on the poor has been partly addressed by providing exemptions for public transport modes like buses and taxis (Lishman, 2013). In the future, some innovative schemes are worth exploring for the case of African countries. In particular, and instead of road pricing, tradable transport permits can be considered. With this scheme, road users are given a permit allowing them to drive some distance per unit of time. It is an initial endowment that may be the same for all users or depend on some attributes. The user can either totally use the permit or sell a part of it, to those who need to travel longer distances, at a market price where the permits are traded. These schemes were initially proposed by Goddard (1997) and Verhoef, Nijkamp and Rietveld (1997) and were extensively studied in the last decade (see Li & Robusté, 2021, for a review). As shown by

368  Handbook on transport pricing and financing

Wadud et al. (2008), the main benefit of tradable permits with respect to equity issues is that if the trading market works well, it produces monetary transfers from the high-income users (high willingness to pay for long-distance trips) to the low-income users (small willingness to pay for large distance trips). This may be more acceptable to low-income groups since it provides an opportunity to increase their income by giving up part of their permits. Tradable permits have not been yet examined through real case studies, but this may change as web technologies can provide convenient platforms for the implementation of the trading market.

19.3 TRANSPORT FINANCING There is an important interdependency between road pricing, investment, and financing and any discussion of road pricing policy is incomplete without considering financing policy. As already highlighted, transportation infrastructure needs in Africa are very large with an estimated annual requirement of between $130–170 billion according to the African Development Bank (2019) and much of it has historically been financed by national governments and multilateral institutions. With the continued narrowing of fiscal space, many countries have turned to “non-traditional” sources to close the investment and funding gap. Private investment and bilateral concession loans from emerging countries, especially China have played a major role in filling the gap (Mengsitu, 2013). According to the World Bank, meeting the transport infrastructure needs of the region requires not only efficient domestic resource mobilization from governments but also innovative solutions to crowd-in private financing (Calderon et al., 2018). Efficient infrastructure planning by the public sector remains critical to the region’s infrastructure development efforts. It is however important to note that if the private sector is to participate in financing, building, and operating the infrastructure, then projects must be structured in such a way as to generate a private profit. From a societal perspective, private participation in infrastructure can best be justified if it motivates investors to select the most efficient projects over those that are politically expedient, manage project risks, and set prices to manage demand and cover capital and operating costs. In practice, private involvement in transportation projects globally has had mixed results. The above notwithstanding, public-sector financing from African national governments remains the biggest source of transport sector financing. For example, African governments public-sector spending on transport reached $20.1 billion in 2017, up by 23% from $16.3 billion in 2016 (Infrastructure Consortium for Africa, 2017) as shown in Figure 19.1. This amount is by far larger than public spending in all other sectors. Increasing transport infrastructure spending plays a key role in reducing the poverty levels of many households and has significant positive impacts on economic growth, income, employment, equity, and social inclusion (Quium, 2019). Kapindula and Kaliba (2019) observed in their study that increasing the stock of infrastructure assets positively affects economic growth and income inequality declines with higher infrastructure quantity and quality. Furthermore, infrastructure spending improves the investment climate for foreign direct investment (FDI) by subsidizing the cost of total investment by foreign investors and thus raising the rate of return (Khadaroo & Seetanah, 2009). For example, according to a World Bank study (Teravaninthorn & Raballand, 2009), a 10% drop in transport costs increases trade by 25% in Africa.

Road transport pricing and financing in Africa  369

Source:   (Infrastructure Consortium for Africa, 2017). ICA members: The G8 countries, the World Bank Group (WBG), the African Development Bank Group (AfDB), the European Commission (EC), the European Investment Bank (EIB), the Republic of South Africa, and the Development Bank of Southern Africa (DBSA).

Figure 19.1  Total infrastructure commitments by sector and source 19.3.1 Public-Sector Financing According to International Monetary Fund (IMF) estimates, African countries finance about 65% of their infrastructure expenditures from the public-sector budget, that is, almost $60 billion or about 4% of gross domestic product (GDP) for the sub-Saharan region and this excludes financing from multilateral institutions (Gutman et  al., 2015). The Infrastructure Consortium for Africa (ICA) report shows that $20.1 billion was committed from publicsector budgets for the transport sector alone (Infrastructure Consortium for Africa, 2017). It is important to note that capturing accurate data about public-sector financing is challenging as budgeting and spending happens at both national and sub-national levels. The ICA report highlights the importance of developing methodologies that adequately capture this type of data to ensure that double counting is avoided. Most transport infrastructure projects in Africa give financial returns that are insufficient to attract the private sector but yield large social benefits making them more feasible for public-sector financing. This further highlights the need for efficient domestic resource mobilization by governments to raise more domestic finance to meet the infrastructure gap. There exists a wide variation of tax revenue to GDP ratio across the continent ranging from 25% in South Africa to 2.8% in the Democratic Republic of the Congo (Gutman et al., 2015) and this illustrates partly the inefficiencies in domestic revenue mobilization. In addition to road pricing policies, international sovereign bonds provide another source of mobilizing more

370  Handbook on transport pricing and financing

public-sector financing. Although African debt and capital markets are still in the early stages of development and often ill-equipped to finance the continent’s overwhelming transportation needs (Hara et al., 2008), many countries in sub-Sahara Africa have increasingly accessed international capital markets with 13 countries issuing $15 billion worth of international sovereign bonds since 2006 (Gutman et al., 2015). With increased innovation and sophistication in infrastructure financing, there is a new form of “blending” public-sector financing with private financing, financing from development finance institutions (DFI) and/or donor (development partner) funding. According to Wentworth and Makokera (2015), this blending combines concessionary loans with debt financing from international financial institutions (IFIs), allowing for “grant loan” elements to keep the service tariff affordable. PPP road projects in Africa have attracted this form of blended financing including the 20.4 km of Dakar–Diamniadio toll road in Senegal and the planned Kampala–Jinja Expressway in Uganda. 19.3.2 Official Development Finance from Multilateral Banks For decades, traditional multilateral and OECD bilateral assistance has represented the principal external funding source for infrastructure in developing countries with the World Bank and the regional development banks playing a central role. For example, in 2012, the World Bank lent $4.3 billion, and the AfDB increased its share to $2.6 billion, thereby together contributing to about 70% of the total ODF financing portfolio of over $10 billion a year (Gutman et al., 2015). Multilateral banks have consistently supported transport investment as shown in Figure 19.2 in the form of development grants or low-interest, long-term loans because often the financial returns for many transportation projects such as roads and railways are insufficient to attract private-sector financing but with large (positive) social externalities (Gutman et  al., 2015; Hara et al., 2008). 14000 12000 10000 8000 6000 4000

Energy

Telecom

Transport

Water Supply & Sanitation

Source:   Gutman et al. (2015).

Figure 19.2  ODF infrastructure investment commitments in Sub-Sahara Africa, by sector, 1990–2012, in $ millions

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1990

0

1991

2000

Road transport pricing and financing in Africa  371

However, there has been a notable shift in development finance over a decade with more emphasis on leveraging private capital to finance infrastructure. According to Van Waeyenberge (2015), private flows are now projected as a superior substitute for aid flows, which had traditionally been considered as the more suitable form through which to engage in development finance. This is evidenced by the increasing appetite of multilateral banks toward public–private partnerships toll roads financed through “blending” private capital with concessional and non-concessional finances. Therefore, developing countries including African nations need to develop comprehensive road pricing and tolling instruments and policies in the wake of this new reality. 19.3.3 Chinese Development Finance China is Africa’s largest bilateral trading partner and the biggest developing country investing in the continent. China’s presence in Africa partly relates to China’s need for natural resources, its intention to expand markets for Chinese exports and professional services, and its desire to contribute to world development, and this is integral to its Belt and Road Initiative (BRI) launched in 2013 (Ogwang & Vanclay, 2021). According to Carrai (2021), BRI is China’s global development project and investment plan, aimed at increasing world connectivity by building a network of rail, highways, bridges, and ports. As part of its BRI, China is increasingly investing abroad and has become an important country for development financing (Carrai, 2021; Lisinge, 2020). For Africa, cooperation with China on infrastructure development started much earlier than the launch of BRI. For instance, China was involved in the construction of the TAZARA railway linking Tanzania and Zambia in the 1970s (Lisinge, 2020). Over the years, China has played a major role by filling the infrastructure financing gaps that are not met by either the private sector or official development financing. As noted by Carrai (2021), China has offered opportunities where there were none, diversifying development financing in Africa. The impact of Chinese development financing varies by industry and country, but also by enterprise type, and the governance model behind the BRI is thus constantly reshaped by interactions between Chinese and host-country economic and social actors, producing highly context-specific results (Carrai, 2021). Getting reliable data on the amount of money invested in infrastructure by China is very difficult. However, according to a US-based China–Africa Research Initiative at the Johns Hopkins University (Acker & Brautigam, 2021), Chinese financiers have committed $153 billion to African public-sector borrowers between 2000 and 2019, and at least 80% of these loans financed economic and social infrastructure projects, mainly transport, power, telecoms, and water. Gutman et al. (2015) observed that besides investing in countries that are not receiving major private-sector investments, China is involved in key sectors like transport, particularly railways and roads as shown in Figure 19.3. These are also sub-sectors in which Chinese firms have particular experience and they have successfully competed for large contracts financed by multilateral banks and aid agencies (Gutman et al., 2015; Ogwang & Vanclay, 2021). The much-talked-about experience of Chinese firms in transport infrastructure development is through traditional procurement methods and not PPPs. This is because PPPs in the Chinese domestic market are still in their early stages of development compared to the competitors from Western countries. According to Qin (2016), the Chinese Ministry of Finance (MoF) has been increasingly promoting infrastructure investment via PPPs and this led to the

372  Handbook on transport pricing and financing 5%

34% 53%

Energy Telecom Transport Water Supply & Sanitation

8%

Source:   Gutman et al. (2015).

Figure 19.3  Chinese infrastructure investment commitments in Sub-Sahara Africa, by sector, 2005–2012, proportions introduction of a series of regulations and guidelines on PPPs in 2014 and 2015. The same Ministry has established the China Public–Private Partnerships Centre (CPPPC), which aims to conduct policy research, promote international exchanges, as well as provide consultancy and training related to PPP (Chu & Muneeza, 2019). It is anticipated that once Chinese firms have gained more PPP experience in their domestic market, then Chinese-PPP policy transfer to Africa will grow rapidly. It is observed that most toll roads in China were financed by Province and County-level governments raising funds by borrowing against future toll revenues (Duncan, 2007a cited in Qin, 2016). The Chinese development financing model has been largely criticized by the Western powers arguing that Chinese investments have been concentrated in countries with commodities to extract and that infrastructural investments have aimed at facilitating China’s extractive needs rather than supporting a balanced regional growth strategy (Siu, 2019). However, many studies (Lisinge, 2020; Ogwang & Vanclay, 2021) have indicated that the Chinese model of infrastructure financing has increased competition for development projects, allowing developing countries to bargain more effectively for better economic returns with Western countries. Furthermore, BRI is not associated with conditions that Western countries set in exchange for support in infrastructure development such as democracy, transparency, rule of law, and human rights (Carrai, 2021; Lisinge, 2020). China’s model of not setting similar conditions to Western powers poses particular challenges for urban road infrastructure development in Africa. China’s state-owned enterprises (SOEs) often negotiate mega urban transport projects with national governments without involving city or metropolitan-level governments and this happened with the Kampala– Entebbe expressway (toll road) in Uganda as well as the light rail transit in Addis Ababa, Ethiopia. Goodfellow and Huang (2020) point out that this model of working has led to haphazard outcomes that are poorly integrated with broader urban planning aspects of these cities. OECD donors usually have a significant, if not decisive impact on how projects are designed, built, and managed. While Chinese-financed projects often do not comprehensively analyze the negative consequences and future risks that these megaprojects create especially in the

Road transport pricing and financing in Africa  373

operations and maintenance phase. For example, Kampala–Entebbe toll road has attracted minimal traffic demand even before tolling commences partly due to limited access points while the old Kampala–Entebbe road is still experiencing heavy traffic congestion throughout the day. Further, road infrastructure loans from China are tied to the use of Chinese firms against the local public procurement laws (Goodfellow & Huang, 2020). The common methods for financing public infrastructure in developing economies under the Chinese development financing model are Resources-for-Infrastructure (R4I) and Resource-Financed Infrastructure (RFI), usually involving Chinese financial institutions and Chinese construction companies (Lisinge, 2020; Ogwang & Vanclay, 2021). According to Mengsitu (2013), the main goal of the resource-backed loan is to secure natural resources, while providing loans with competitive interest rates to countries that would otherwise not be able to obtain such loans. Here the Chinese bank negotiates an agreement with the recipient country whereby it lends the country funds for building an infrastructure facility and the country repays the loan through natural resource exports or by providing preferential access to natural resource exploitation rights. For example, in Uganda, expectations around future revenue from oil extraction have led to many infrastructure projects being commissioned, mostly funded under RFI arrangements (Ogwang & Vanclay, 2021). According to Ogwang and Vanclay (2021), R4I/RFI deals are used to reduce the risks of doing business in countries with weak institutions, a lack of transparency, low credit ratings, and high risks of loan nonrepayments. It should however be noted that not all loans from China are tied to resources. 19.3.4 Public–Private Partnerships African governments are increasingly looking to public–private partnerships as a tool to finance transportation projects to address their lack of public resources. According to a World Bank study, around 50% of infrastructure needs in developing countries can be financed on a commercial basis (Calderon et al., 2018). This however requires good project preparation to attract private investment, as well as considerations about the equity implications of charging user fees. According to the African Development Bank, Africa’s infrastructure gap exists in part due to an insufficient number of bankable projects being developed (African Development Bank Group, 2019). Therefore, there is a need for a sufficient dedicated fund at a regional level to help in preparing transport infrastructure projects to reach bankability levels. The existing funds such as the NEPAD Infrastructure Project Preparation Facility (IPPF) housed at the African Development Bank are currently not adequate considering the demand for bankable projects in the region. It has often been observed that a considerable amount of international private finance is deterred because of a lack of bankable projects and that African infrastructure is encumbered by an array of political and organizational impediments that raise perceived risk to unacceptable levels (Collier & Cust, 2015). Under a PPP structure (see Chapter 16), the private sector provides financing and expertise in exchange for the ability to share collected revenues, which usually come in the form of tolls or other user-paid fees (Hara et al., 2008). PPP can be implemented through several contractual relationships, like build, operate, and transfer scheme (BOT) or operate and maintenance (O&M) scheme (see Biau et al., 2008). Financing may come from different sources. According to Calderon et  al. (2018), Public-sector financing may include (a) governments providing part of a project’s upfront capital costs through grants or viability gap funding; (b) SOE investing equity; and (c) state-owned banks extending loans. While private-sector

374  Handbook on transport pricing and financing

financing may include equity through the project’s developer or project finance debt through private lenders, which can be either commercial banks or institutional financiers like DFIs. For example, in Uganda, the African Development Bank worked closely with the government to develop a business model with a viability gap funding component for the proposed Kampala–Jinja highway project (African Development Bank Group, 2019). In its report, the bank notes that the project was not considered commercially viable under the traditional user-fee-supported model, and thus it was structured with a viability gap funding (VGF) component to address affordability and financial viability (African Development Bank Group, 2019). Africa is among the developing regions with few PPP infrastructure projects, including in the transport sub-sector. The situation is however changing due to sustained economic growth in many countries. According to the World Bank’s Private Participation in Infrastructure (PPI) project database, the value of projects with private-sector participation reaching financial close in 2017 totaled $5.2 billion, an increase from the $3.6 billion reported in 2016 and of this, $2.3 billion (44.8%) was privately financed (Infrastructure Consortium for Africa, 2017). The most active countries in PPP infrastructure projects in the region are South Africa, Nigeria, Kenya, and Uganda (Calderon et  al., 2018). But most of these PPP projects are in energy and telecommunication and very few in the roads sub-sector. The limited private participation in transport infrastructure is a result of high real and perceived country and project risks including a history of high debt levels, exchange rate risk, and high inflation. Uncertainties regarding the African political climate have consequently discouraged private-sector infrastructure investors, in a context where revenue-supported projects are rare, and thus long-term state payments are required to make projects financially viable. The World Bank further noted that many economies in Africa are small, and others have weak legal and regulatory frameworks to procure and implement PPPs (Calderon et al., 2018). As a result, those projects that have managed to attract private-sector capital have either been small scale or have benefited from significant risk mitigation mechanisms (Hara et al., 2008) like viability gap funding.

19.4 INFORMAL PUBLIC TRANSPORTATION In absence of an efficient publicly owned and financed public transportation system in many African countries, transport regulation is very weak and uncoordinated. Passenger transport is dominated by “informal” public transport (paratransit) modes that include mini-buses, pickups, small saloon cars, and commercial motorcycles commonly referred to as “Boda” in East Africa and “Okada” in many West African countries. These modes are owned and operated by largely unregulated private-sector players and this has led to uncompetitive practices and pricing higher than competitive market levels. These have contributed to a user experience that is often one of low service quality, erratic availability, high prices, and poor road safety (Venter et al., 2014). Kilani and Houassa (2018) discussed several transport reforms for the city of Sousse (Tunisia), where congestion costs and emissions are high. It is shown how road pricing can significantly reduce these negative externalities. At the same time, they notice that it is very important that road pricing be accompanied by the significant development of public transport services to limit the market share of paratransit modes. This requires raising further funds and

Road transport pricing and financing in Africa  375

may seem paradoxical, but the main point is that the transport question should be addressed through a global view within the whole economic development question. Unlike in the urban areas, the conditions of rural roads in Africa affect the quantity and quality of public transport services provided, as well as the fares charged to passengers. The paratransit operators tend to organize into informal associations or unions, which exercise de facto control over fares, route allocation, and the quantity of service. This kind of selfregulation often evolves as a way of allocating economic rights among competing operators (Venter et al., 2014). The paratransit industry in Africa is very complex, political, and often not researched. With the current shortage of knowledge especially how the industry is organized; who are the key players; how decisions at the route level are made; what drives the entry decision for individual operators; what kind of vehicle to use; at what level to set fares and so on; reforming or improving the industry will remain a big challenge. The attempts by many African governments to “replace” informal public transport operators with high-capacity public transport systems, especially bus rapid transit (BRT) systems in many capital cities have given mixed results. The “replacement” option may not be a feasible policy direction as paratransit operations also have some inherent advantages with respect to demand responsiveness and service innovation. In many parts of Africa, especially remote areas, paratransit provides cheap, accessible, and flexible transport solutions for the poor and it has employed many Africans including those with no formal education. A hybrid system comprising both paratransit and formally planned public transport modes in African cities is very possible through the integration of modes as well as strengthening regulation of the existing paratransit services. This has been tried both in Cape Town and Johannesburg, South Africa where paratransit modes operate feeder routes of the BRT system. As observed by Scorcia and Munoz-Raskin (2019), in Johannesburg, BRT buses are operating with the unprecedented involvement of the incumbent operators after a facilitated negotiation process and existing paratransit complementing the BRT system. We thus believe paratransit systems can be supported to complement the “modernization” efforts of many governments wishing to introduce higher-capacity bus systems like BRT. One of the biggest challenges in the paratransit industry is the lack of financing which results in the use of very old second-hand imported vehicles largely from Japan. These poorly maintained vehicles have contributed to many road fatalities in Africa. In an effort to tackle financing challenges, the African Development Bank has provided advisory support to develop private-sector lending and participation through lending to a transit financial solutions provider in South Africa that on-lends to small and medium enterprises (SMEs). According to the bank report, this non-sovereign funding will help paratransit operators to acquire fuel-efficient, environmentally-friendly, quality vehicles (minibus taxis) and the project will catalyze funding for approximately 8,000 commuter transit vehicles across South Africa (African Development Bank Group, 2019). Such innovative instruments will go a long way to improve the sector. However, the sector further requires public funding in form of subsidies to operate efficiently. More research is needed to fully understand the ever-growing paratransit industry in Africa so that planned improvements or reforms are evidence-based. 19.4.1 Institutional Capacity in Improving Public Transportation Addressing transport infrastructure challenges in Africa involves among other things correcting the problems of weak public institutional capacities, underinvestment, and poor

376  Handbook on transport pricing and financing

maintenance (Newbery, 1988; Newbery et al., 1988; Collier & Cust, 2015). For example, the fragmentation of transport functions and responsibilities across a myriad of agencies makes it difficult to implement urban transport reforms. Especially in large metropolitan areas which contain several local governments (Goodfellow & Mukwaya, 2021; Pojani & Stead, 2018). Lack of adequate institutional structures has been identified by Walters (2014), who concluded that implementation of efficient public transport policy is greatly assisted if properly structured and capacitated formal transport authorities are established with the necessary capacity to plan, implement and monitor transport plans. By and large, many cities in Africa have outgrown their original boundaries of city administration and this requires a metropolitan-level kind of governance for proper coordination of different reforms in the metropolitan area. In cities where BRT implementation has been reported as successful, national governments created independent agencies responsible for transport planning across the metropolitan region, for example, Lagos Metropolitan Area Authority in Nigeria, Dar Rapid Transit Agency in Dar-es-Salaam, Tanzania. These agencies have legal mandates to plan and implement public transport reforms across entire metropolitan areas (Hidalgo & Graftieaux, 2008; Manwaring & Wani, 2021; Nguyen & Pojani, 2018). Many studies have reported that the lack of a proper BRT implementation agency that is in charge of all transportation reforms at the metropolitan-level results in delays due to cumbersome administrative procedures and inconsistency of planning (Kumar et al., 2012; Nguyen et al., 2019; Otunola et al., 2019). Therefore, central governments need to empower transportation agencies with adequate expertise and experience (Mitchell & Walters, 2011) so that they can play a more significant role in transport policy-making in their respective countries.

19.5 CONCLUSIONS The transport infrastructure shortage in Africa will continue to hinder rapid economic growth if not addressed with utmost urgency. Road pricing presents an opportunity to raise funds for transport infrastructure investment in the region. It is through such efforts that African countries will reap the benefits of the recently commenced Africa Continental Free Trade Area (AfCFTA) whose immediate objective is to increase participation in cross-border supply chains by reducing trade costs through regional integration (African Development Bank, 2019). Optimal pricing will be critical to ensure that there is a balance between economic efficiency, equity, and transaction costs. In terms of public transportation, there has been relatively slow progress in transforming public transportation systems in Africa partly due to a lack of funding and weak institutional capacity (Walters, 2013, 2014). Road pricing provides an opportunity to raise revenues that can in part be used to fund public transport infrastructure. To address institutional capacity issues, it is necessary to view public transportation reforms as part of the long-term governance reform process that requires re-orientation of government as well as of the entire transport system to achieve better results. Transport policy integration both at the national and regional levels is an effective strategy to raise more external finances for infrastructure development, especially cross-border transport infrastructure projects. Instead of individual country bilateral negotiations with external funders like China, African institutions such as the African Union (AU) or regional blocks such as the East African Community (EAC) can jointly negotiate for better loan agreement

Road transport pricing and financing in Africa  377

terms to finance regionally strategic infrastructure like the Standard Gauge Railway (SGR) that is connecting several countries in East Africa including Uganda, Rwanda, and the Democratic Republic of Congo to the port of Mombasa in Kenya. This would increase trade volumes and in part, challenge the common belief that Africa trades “too little” both with itself and with the rest of the world.

NOTE 1.

This is shown to be important in more general contexts. For example, Odeck and Kjerkreit (2010) show, for the case of Norway, how users’ opposition to road pricing is significantly reduced when they observe that collected funds have been used to develop useful infrastructure.

REFERENCES Acker, K., & Brautigam, D. (2021). Twenty years of data on China’s Africa lending. In Briefing Paper (Issue 4). African Development Bank. (2014). Tracking Africa’s Progress in Figures, 1–39. http://www​.afdb​.org​/ fileadmin​/uploads​/afdb​/ Documents​/ Publications​/ Tracking​_ Africa’s​_ Progress​_in​_ Figures​​.pdf African Development Bank. (2019). African Outlook Economic 2019. African Development Bank Group. (2019). Infrastructure and superstructure. In Travel and Tourism Management. doi: 10.1007/978-1-349-17946-6_6 Biau, C., Dahou, K., & Homma, T. (2008). How to Increase Sound Private Investment in Africa’s Road Infrastructure: Building on Country Successes and OECD Policy Tools. NEPAD-OECD Africa Investment Initiative Roundtable, Kampala, Uganda. Calderon, C., Cantu, C., & Chuhan-Pole, P. (2018). Infrastructure development in Sub-Saharan Africa: A scorecard. Infrastructure Development in Sub-Saharan Africa: A Scorecard, May. doi: 10.1596/1813-9450-8425 Carrai, M. A. (2021). Adaptive governance along Chinese-financed BRI railroad megaprojects in East Africa. World Development, 141. doi: 10.1016/j.worlddev.2020.105388 Chu, J., & Muneeza, A. (2019). Belt and road initiative and Islamic financing: The case in public private partnership infrastructure financing. International Journal of Management and Applied Research, 6(1), 24–40. doi: 10.18646/2056.61.19-002 Collier, P., & Cust, J. (2015). Investing in Africa’s infrastructure: Financing and policy options. Annual Review of Resource Economics, 7(1), 473–493. doi: 10.1146/annurev-resource-100814-124926 Duncan, T. (2007a). Retrospective analysis of the road sector 1997–2005. Asian Development Bank Evaluation Document. Manila: Asian Development Bank. Export-Import Bank of India. (2018). Connecting Africa: The Role of Transport Infrastructure. Goddard, Haynes C. (1997). Using tradeable permits to achieve sustainability in the world’s large cities: Policy design issues and efficiency conditions for controlling vehicle emissions, congestion and urban decentralization with an application to Mexico City. Environmental and Resource Economics, 10(1), 63–99. Goodfellow, T., & Huang, Z. (2020). Contingent infrastructure and the dilution of ‘Chineseness’: Roads, rail and repurposing in Kampala and Addis Ababa. Environment and Planning A. Goodfellow, T., & Mukwaya, P. (2021). The Political Economy of Public Transport in Greater Kampala: Movers, Spoilers and Prospects for Reform (Issue March). Gutman, J., Sy, A., & Chattopadhyay, S. (2015). Financing Africa Infrastructure (Issue March). https:// www​.brookings​.edu​/wp​-content​/uploads​/2016​/07​/AGI​Fina​ncin​gAfr​ican​Infr​astr​ucture​_ FinalWebv2​ .pdf Hara, J., Weygandt, N., & Machulak, J. (2008). Financing Transportation Projects in Africa. May, 2006–2008.

378  Handbook on transport pricing and financing

Hidalgo, D., & Graftieaux, P. (2008). Bus rapid transit systems in Latin America and Asia Results and difficulties in 11 cities. Transportation Research Record, 2072, 77–88. doi: 10.3141/2072-09 Infrastructure Consortium for Africa. (2017). Infrastructure Financing Trends in Africa – 2017. ITS. (2016). Varying Acceptance of Tolling in Africa, 1–6. Kapindula, M. G., & Kaliba, C. (2019). The effects of external debt servicing on infrastructure Spending: A case of Zambia. International Journal of Construction Management, 3599. doi: 10.1080/15623599.2019.1615753 Khadaroo, J., & Seetanah, B. (2009). The role of transport infrastructure in FDI evidence from Africa using GMM estimates. Journal of Transport Economics and Policy, 43(3), 365–384. Kilani, M., & Houassa, F. (2018). Urban transportation reforms under the presence of semi-collective modes: The case of the city of Sousse. Revue d’Economie Régionale et Urbaine, No. 4, 805-828. KPMG. (2021). Budget Brief Uganda (Issue June). https://assets​.kpmg​.com​/content​/dam​/ kpmg​/ ke​/pdf​/ tax ​/ kenya​-budget​-brief​-2017​.pdf Kumar, A., Zimmerman, S., & Agarwal, O. P. (2012). The Soft Side of BRT: Lessons from Five Developing Cities. The International Bank for Reconstruction and Development/The World Bank Group. Li, Siyu, & Robusté, Francesc. (2021). From urban congestion pricing to tradable mobility credits: A review. Transportation Research Procedia, 58, 670–677. Lishman, D. (2013). A Critical Evaluation of Road Pricing in South Africa. 11. http://uctscholar​.uct​.ac​ .za​/ PDF​/99242​_ Lishman​_ D​.pdf Lisinge, R. T. (2020). The Belt and Road Initiative and Africa’s regional infrastructure development: implications and lessons. Transnational Corporations Review, 12(4), 329–342. doi: 10.1080/19186444.2020.1795527 Manwaring, P., & Wani, S. (2021). Informal Transport Reform in Kampala Learning from CrossCountry Experience (Issue August). Mengsitu, T. M. (2013). Emerging Infrastructure Financing Mechanisms in Sub-Saharan Africa. 1–163. http://www​.rand​.org Mitchell, M., & Walters, J. (2011). Efficacy of recent transport policy making and implementation in South Africa. Journal of Transport and Supply Chain Management, 5(1), 241–263. doi: 10.4102/ jtscm.v5i1.76 Newbery, D. M. (1988). Road damage externalities and road user charges. Econometrica: Journal of the Econometric Society, 295–316. Newbery, D. M., Hughes, G. A., Paterson, W. D. O., & Bennathan, E. (1988). Road Transport Taxation in Developing Countries: The Design of User Charges and Taxes for Tunisia. World Bank Discussion Papers No. 26. Nguyen, M. H., Ha, T. T., Tu, S. S., & Nguyen, T. C. (2019). Impediments to the bus rapid transit implementation in developing countries–a typical evidence from Hanoi. International Journal of Urban Sciences, 23(4), 464–483. Nguyen, M. H., & Pojani, D. (2018). Why do some BRT systems in the global south fail to perform or expand? Advances in Transport Policy and Planning, 1(September), 35–61. doi: 10.1016/bs​.atpp​.2018​. 07​​.005 O’Brien, P., O’Neill, P., & Pike, A. (2019). Funding, financing and governing urban infrastructures. Urban Studies, 56(7), 1291–1303. doi: 10.1177/0042098018824014 Odeck, J., & Kjerkreit, A. (2010). Evidence on users’ attitudes towards road user charges—A crosssectional survey of six Norwegian toll schemes. Transport Policy, 17(6), 349–358. Ogwang, T., & Vanclay, F. (2021). Resource-financed infrastructure: Thoughts on four Chinese-financed projects in Uganda. Sustainability (Switzerland), 13(6). doi: 10.3390/su13063259 Otunola, B., Harman, O., & Kriticos, S. (2019). The BRT and the danfo: A case study of Lagos’ transport reforms from 1999–2019. International Growth Centre, 1–28. Parry, I. W., & Small, K. A. (2009). Should urban transit subsidies be reduced? American Economic Review, 99(3), 700–724. Pienaar, W.J. 2005. Road Cost Allocation and Recovery. Working paper WP 1/03. Department of Logistics, Stellenbosch University. Pojani, D., & Stead, D. (2018). Policy design for sustainable urban transport in the global south. Policy Design and Practice, 1(2), 90–102. doi: 10.1080/25741292.2018.1454291

Road transport pricing and financing in Africa  379

Qin, Y. (2016). China’s transport infrastructure investment: Past, present, and future. Asian Economic Policy Review, 11(2), 199–217. doi: 10.1111/aepr.12135 Quium, A. S. M. A. (2019). Transport Corridors for Wider Socio–Economic Development. Schalekamp, H. (2017). Lessons from building paratransit operators’ capacity to be partners in Cape Town’s public transport reform process. Transportation Research Part A: Policy and Practice, 104(November 2015), 58–66. doi: 10.1016/j.tra.2017.08.002 12 Schalekamp, H., & Behrens, R. (2010). Engaging paratransit on public transport reform initiatives in South Africa: A critique of policy and an investigation of appropriate engagement approaches. Research in Transportation Economics, 29(1), 371–378. doi: 10.1016/j.retrec.2010.07.047 Scorcia, H., & Munoz-Raskin, R. (2019). Why South African cities are different? Comparing Johannesburg’s Rea Vaya bus rapid transit system with its Latin American siblings. Case Studies on Transport Policy, 7(2), 395–403. doi: 10.1016/j.cstp.2019.01.010 Siu, H. F. (2019). Financing China’s engagement in Africa: New state spaces along a variegated landscape. Africa, 89(4), 638–661. doi: 10.1017/S0001972019000834 Teravaninthorn, S., & Raballand, G. (2009). Directions In Development Infrastructure Transport Prices and Costs in Africa: A Review of the International Corridors. UN. (2019). World population prospects 2019: Highlights. Department of Economic and Social Affairs, Population Division, 141, 49–78. http://www​.ncbi​.nlm​.nih​.gov​/pubmed​/12283219 Van Rensburg, J., & Krygsman, S. (2016). A kilometre-based road user charge system: proof of concept study. Satc, 300–313. Van Rensburg, J., & Krygsman, S. (2019). Funding for roads: Understanding the South African road funding framework. Journal of Transport and Supply Chain Management, 13, 1–11. doi: 10.4102/ jtscm.v13i0.453 Van Rensburg, J., & Krygsman, S. (2020). Funding for roads in South Africa: Understanding the principles of fair and efficient road user charges. Transportation Research Procedia, 48(2019), 1835– 1847. doi: 10.1016/j.trpro.2020.08.218 Van Waeyenberge, E. (2015). The private turn in development finance. FESSUD Working Papers, 266800, 1–54. Venter, C. J., & Joubert, J. W. (2014). Tax or toll? GPS-based assessment of equity impacts of large-scale electronic freeway tolling in Gauteng, South Africa. Transportation Research Record, 2450, 62–70. doi: 10.3141/2450-08 Venter, C. J., Molomo, M., & Mashiri, M. (2014). Supply and pricing strategies of informal rural transport providers. Journal of Transport Geography, 41, 239–248. doi: 10.1016/j.jtrangeo.2014.10.001 Verhoef, Erik, Peter Nijkamp, & Piet Rietveld. (1997). Tradeable permits: their potential in the regulation of road transport externalities. Environment and Planning B: Planning and Design, 24(4), 527–548. Wadud, Zia, Robert B. Noland, & Daniel J. Graham. (2008). Equity analysis of personal tradable carbon permits for the road transport sector. Environmental Science & Policy 11(6), 533–544. Walters, J. (2013). Overview of public transport policy developments in South Africa. Research in Transportation Economics, 39(1), 34–45. doi: 10.1016/j.retrec.2012.05.021 Walters, J. (2014). Public transport policy implementation in South Africa: Quo vadis? Journal of Transport and Supply Chain Management, 8(1), 1–10. doi: 10.4102/jtscm.v8i1.134 Wentworth, L., & Makokera, C. G. (2015). Private sector participation in infrastructure for development. South African Journal of International Affairs, 22(3), 325–341. doi: 10.1080/10220461.2015.1081568 Zhou, G., & Chilunjika, A. (2013). Mobilising Domestic Revenue through Toll Gate Systems in Zimbabwe. International Journal of Business and Social Science, 4(7), 188–204.

20. A review of selected transport pricing, funding and financing issues in Asia Wei Liu, Fangni Zhang, Xiaolei Wang and Yili Tang

20.1 INTRODUCTION Asia is the largest continent in the world in terms of both area size and population. The transport infrastructure systems support the movements of a population of more than 4.6 billion as of March 2021,1 and a substantial volume of freight (e.g., 83 billion parcels in 2020 for China 2 and 4.84 billion door-to-door deliveries in 2020 for Japan3). Similar to many other regions, the transport sector helps provide economic and social opportunities and benefits to Asian people. At the same time, it is a leading contributor to fossil fuel consumption and greenhouse gas emissions. Transport pricing, funding, and financing issues are critical to ensure an efficient and sustainable transport system, which helps drive regional economic growth while maintaining effective resource consumption and reducing the environmental externality of travel activities. As reported in ESCAP (2017),4 road transport is the most commonly used mode for passenger travel and freight transport in major Asian countries. The rail mode (including metro and subway systems) also contributes a significant share to passenger transport in countries like China, India, Japan, Singapore, and South Korea. In the Asian context, where the population is often densely distributed, this chapter will highlight and discuss a few unique and/or important transport pricing, funding, and financing issues related to road traffic and rail transport, including (i) road pricing, (ii) public transit systems funding and pricing, and (iii) high-speed railway financing. This survey does not aim to cover all Asian countries but aims to highlight the unique and/or important issues with some Asian examples. Moreover, this survey focuses on the Asian experiences regarding transport infrastructure policies, but does not involve much on the wider economic effects of these policies, except a few studies in relation to the Chinese high-speed rail development since the major argument for such huge investments in both infrastructures and operations is the positive effect on the connectivity, accessibility, economic growth, and social welfare. The rest of the survey is organized as follows. Section 20.2 reviews road pricing as either a congestion management tool or a funding source. Section 20.2 also briefly discusses alternatives to road pricing, such as road space rationing and tradable travel credit schemes, for congestion management. Section 20.3 focuses on the funding and pricing issues of public transit systems, with an emphasis on transit fare and subsidy strategies, including the “Rail plus Property” model in Hong Kong. Section 20.4 discusses the high-speed railway financing issue, with an emphasis on the Chinese high-speed railway development and its economic impacts, given that China now has the longest high-speed railway track network in the world, which is longer than that of all other countries combined. 380

Transport pricing, funding and financing issues in Asia  381

20.2 ROAD PRICING Road pricing is to levy tolls on travelers for their usage of roads, tunnels, and/or bridges. The theoretical foundations of road pricing are covered by Se-il Mun and Daisuke Fukuda in Chapter 8 of this Handbook; major contributions include Arnott et al. (1990), Yang and Huang (2005), Van den Berg and Verhoef (2011), Meng et al. (2012), and Daganzo and Lehe (2015). Road pricing is used either to discourage the use of private cars (in order to reduce road traffic congestion, emissions, and air pollution) or to generate revenues for infrastructure system financing. Road pricing as a congestion management tool, i.e., congestion pricing, has been advocated by many researchers but has only been implemented in a very limited number of cities (e.g., Singapore, Stockholm, London) due to objection from the public (De Borger and Proost, 2012).5 In contrast, road pricing as an infrastructure funding source has been adopted in many cities (e.g., the highway/expressway tolling system in China and Japan). 20.2.1 Road Pricing as a Congestion Management Tool When road pricing is used as a congestion management tool, it is called congestion pricing. In this subsection, we will introduce the successful application of congestion pricing in Singapore and a few initiatives in major cities/regions of China. Two alternatives to road pricing for congestion management, i.e., road space rationing and tradable travel credit schemes, will also be discussed. In 1975, in order to discourage road usage during peak hours, the Singapore government implemented one of the earliest forms of congestion toll, the Singapore Area Licensing Scheme, in Singapore’s central business district (Seik, 2000). This pricing scheme was then converted to the Electronic Road Pricing (ERP) system in 1998. The ERP system consists of ERP gantries installed at all road links connecting to Singapore’s Central Area, and all the major roads with congested traffic. There are cameras attached to the ERP gantries in order to identify vehicle license plate numbers. In 2018, there was a total of 93 ERP gantries in Singapore. In 2007, variable tolling/pricing based on road traffic speeds and congestion levels was further introduced by Singapore’s Land Transport Authority (LTA), and the ERP rates are reviewed regularly.6 Global Navigation Satellite System (GNSS) will be further integrated with Singapore’s ERP system to facilitate congestion pricing, as announced by LTA on September 8, 2020,7 while the congestion pricing framework will remain similar. As reported in the study of Phang and Toh (2004), the implementation of ERP and congestion pricing successfully controlled road congestion in the central business district and increased traffic speed, and also yielded a modal shift from driving to public transport (such as rail and bus). The congestion toll in Singapore is also coupled with high new vehicle registration fees and substantial subsidy/investment in public transport. These management strategies effectively encouraged the modal shift to public transport. However, in the early stages, at some locations the congestion charges might be too high, resulting in a waste of road capacity. And some travelers may reschedule their entry to the central business district, which created additional implicit schedule delay costs for them. In order to check whether the toll level in Singapore is approximately optimal, Li (1999) was among the earliest to adapt the marginal cost pricing principle (Pigou, 1920). In particular, according to the marginal cost pricing principle, the optimal toll should be:

382  Handbook on transport pricing and financing



t ( q ) = MC ( q ) - AC ( q ) = q ×

dAC ( q ) (20.1) dq

where q is the link flow, AC(q) is the average cost (as a function of q), and MC(q) is the marginal cost. For the earliest form of congestion pricing, i.e., the Area Licensing Scheme, in Singapore, the system can have two observations, i.e., the flow level qN for the non-toll period and the flow level qY for the toll period, as well as the corresponding average costs ACN and ACY for the non-toll and tool periods, respectively. The first-order approximation of τ(q) based on the two observations can be written as:

t (q ) = q ×

AC N - ACY (20.2) qN - qY

dAC ( q ) AC N - ACY is used to approximate the derivative . As qN - qY dq introduced by Li (1999), the average cost (AC) in the above can be approximated by the travel time cost, which consists of (i) the moving time (i.e., the driving time) and (ii) the waiting time (i.e., the total time stopped at road intersections). Given the current observed flow level, one then can check whether the current toll is equal to that calculated toll based on Equation (20.2). Li (1999) concluded that in 1990 the three Singapore dollar fee might be on the low side, and was at least not too high. The above toll estimation can be further improved if one has multiple rounds of observations, and Li (2002) further proposed an iterative process to locate the optimal tolling level. While the study of Li (2002) is inspiring, Wang and Yang (2012) pointed out that the bisection-based iterative procedure in the study of Li (2002) may not converge under the standard single-link travel demand-supply equilibrium framework. They therefore provided a modified version of the bisection method for congestion price adjustment and established its convergence. In 2010, the World Bank suggested that congestion pricing should be adopted in Beijing in order to avoid hyper-congestion.8 In September 2011, local authorities initiated a pricing proposal, and raised heated debate among the public. Due to the wide and strong objection from the public, this pricing proposal failed to materialize (refer to the source in footnote 5). In December 2015, the Beijing Municipal Commission of Transport announced that congestion pricing was planned to be introduced in 2016. But this plan was further postponed due to public objection. The possible introduction of a congestion charge is also mentioned in Beijing’s vehicle emission and pollution control plan 2013–2017, where a variable tolling scheme based on real observations of traffic and emissions is planned, and the toll level will be vehicle-type-dependent, time-dependent, and district-dependent. In view of the public objection toward congestion pricing, Beijing implemented a road space rationing scheme that restricts driving based on the last digits of vehicle license plates9 since Beijing hosted the 2008 Summer Olympics. Several studies have examined the efficiency of license-plate-based traffic rationing schemes (Wang et  al., 2010; Han et  al., 2010), and combined it with pricing (Liu et  al., 2014). By June 2016, there are in total 12 cities in China that have adopted similar driving restriction schemes.10 where the numerical slope

Transport pricing, funding and financing issues in Asia  383

Besides Beijing, similar proposals for congestion pricing appeared in other major cities in China (e.g., Shanghai, Nanjing, and Guangzhou), but none of these initiatives have been implemented. In early 2010, the city of Guangzhou in the south of China started a debate in the public on whether the city should launch congestion pricing schemes. A survey was conducted online by two major local news outlets, and it turned out that 84.4% of participants in the survey opposed introducing the congestion toll (refer to the source in footnote 5). Between 1983 and 1985, Hong Kong tested the ERP system and obtained positive outcomes (Hau, 1990). However, similar to other major cities in China, public opposition stalled its further implementation. In April 2001, the Hong Kong government announced that the congestion pricing plan will not be implemented in the next ten years, because according to estimates the road congestion situation in Hong Kong will not deteriorate significantly and there will be several other measures to improve air quality. Nevertheless, a number of road tunnels are tolled in Hong Kong,11 including three cross harbor tunnels, namely, the Western Harbor Crossing (WHC), the Cross Harbor Tunnel (CHT), and the Eastern Harbor Crossing (EHC), as shown in Figure 20.1. The tolls at these tunnels have been revised in order to manage congestion and road usage. Existing studies indicated that the tunnel demand in Hong Kong is price-sensitive and tolls at these tunnels are effective in managing flow (Hau et al., 2011). While the road space rationing scheme rather than congestion pricing was implemented in cities of China (e.g., Beijing), some studies have utilized the observed changes in traffic density induced by the traffic rationing scheme to quantify the relationship between traffic density, traffic speed, and optimal charges (e.g., Yang et al., 2020a). They found that congestion pricing may increase speed by about 11% in the city center of Beijing and can yield a

Source:   Adopted from Woo et al., 2015.

Figure 20.1  Congestion at Hong Kong Cross Harbor Tunnel: red line denotes the queuing observed during peak hours on roads connecting to the tunnel

384  Handbook on transport pricing and financing

welfare gain of 1.5 billion Chinese Yuan from reduced congestion and a revenue of 10.5 billion Chinese Yuan per year. Recent studies also pointed out potential issues associated with the license-plate-based traffic rationing schemes (Wang et al., 2010; Nie, 2017a), while it seems to face less objection from the public. It is reported that: (i) the traffic rationing scheme is not a first-best strategy, although it may help mitigate congestion under certain conditions; (ii) a very restrictive traffic rationing scheme may potentially worsen congestion conditions when imposed on a rich population that may bypass the strategy via owning additional cars; and (iii) the wealthier travelers are likely to benefit more from the traffic rationing scheme, especially when they can have additional cars. In this context, Nie (2017b) further proposed potential remedies for the license-plate-based traffic rationing scheme. In particular, the first strategy combines traffic rationing with a vehicle quota scheme that controls vehicle ownership. The other two strategies convert the driving permit into a tradable commodity, inspired by the tradable travel credit scheme (Yang and Wang, 2011), which can maintain revenue neutrality. The tradable travel credit scheme for network traffic management is proposed by Yang and Wang (2011) as an appealing alternative to congestion pricing. In particular, a social planner allocates travel credits to eligible travelers, and then there are road link-specific credit charges to travelers. Moreover, free trading of travel credits among travelers is allowed. The advantages of the tradable credit scheme when compared to congestion pricing are mainly twofold. First, it does not involve direct out-of-pocket money from travelers to the management/planning authority and can be revenue-neutral, which expects to receive less objection from the public. Second, an appropriately designed tradable credit charge scheme can derive the same first-best outcome as congestion pricing, while as mentioned earlier the road space rationing scheme cannot derive the first-best solution. A series of studies have further examined tradable credit scheme design under heterogeneous travelers (Wang et al., 2012; Zhu et al., 2015), with non-zero transaction costs (Nie, 2012), for dynamic traffic management (Nie and Yin, 2013), in the absence of demand functions (Wang and Yang, 2012; Wang et al., 2014). Recent developments and wide applications of smartphone and GNSS (e.g., as mentioned earlier, GNSS will be further integrated with Singapore’s ERP system to facilitate congestion pricing while at the moment the pricing framework remains similar) create new opportunities for realizing more refined congestion pricing schemes and tradable travel credit schemes. For instance, smartphone apps and online platforms can facilitate the transaction process of credits, while GNSS will facilitate the implementation of refined charge schemes. A system that integrates the GNSS, vehicle license, and travelers’ smartphones will largely ease operations that are considered infeasible in the past. 20.2.2 Road Pricing as a Funding Source While congestion pricing faces strong objection,12 China has a nationwide tolling system for almost all expressways. The majority of Chinese expressways are not directly owned by the state. Instead, they are often owned by for-profit organizations or corporations (through various levels of joint public and private ownership). These for-profit organizations or corporations usually borrow from banks or securities markets based on projected toll revenue from the planned expressways. According to the National Highway Toll Standard in 2021, the highway toll is vehicle-type dependent (more expensive for larger vehicles) and varies from 0.33 to 3.77 Chinese Yuan per kilometer across different highways.

Transport pricing, funding and financing issues in Asia  385

It is quite common that Asian countries such as China, Malaysia, Thailand, and Indonesia have utilized road tolls to generate revenue (Chung, 2002). The idea of “maintaining road by road” in China is due to the lack of capital for transport development. A branch of studies examined “build–operate–transfer” (BOT) scheme for expressway/highway projects. In the eighteenth century, the BOT scheme was first introduced in Europe for infrastructure projects. Under the BOT scheme, a highway is built, operated, and maintained by a private firm, and the firm will collect the highway toll revenue in return during a concession period (Yang and Meng, 2000; Tan et al., 2009).13 A series of studies have examined the BOT scheme for highway projects under demand uncertainty (Chen and Subprasom, 2007; Tan and Yang, 2012; Lu and Meng, 2017), asymmetric cost information (Shi et al., 2016), ceiling prices (Hoang-Tung et al., 2021), and cross-border transport infrastructures (Mun and Nakagawa, 2010). Government guarantees are often used to attract private firms to participate in BOT road projects. Some studies examined the impact of government guarantees on toll level, road quality, and capacity (e.g., Feng et al., 2015). It has been found that in general: (i) a minimum traffic guarantee increases the toll level while lowering road quality; (ii) a price compensation guarantee can decrease the toll level and improve road quality and capacity. The positive social and economic outcome of road improvement is the main argument for new road projects even if BOT schemes have to be introduced. However, as reported by Chung (2002), the tolls used to fund the road improvement can reduce traffic, and thus diminish or postpone the benefits of road improvement. This dilemma of tolling and reduction in traffic indicates that there can be a long “time lag” between road improvement and economic growth. In this context, China’s tolling of expressways has been criticized for setting excessively high toll levels.14 Reforms of the tolling system of roads and expressways were planned by the National People’s Congress in order to reduce the costs of using roads and bridges.15

20.3 PUBLIC TRANSIT SYSTEMS FUNDING AND PRICING In the densely populated large cities in Asia, public transit systems (including buses, trams, rails, and ferries, etc.) serve a huge number of travelers on a daily basis. For instance, in Hong Kong, in the first half of 2020, every day around 8.9 million passenger journeys are made through the public transport system; and by the end of June 2020, the entire Mass Transit Railway (MTR) system in Hong Kong has transported around 3.39 million passengers per day (this number can be above 4.50 million per day pre-COVID-19).16 It is also noteworthy that in Hong Kong, public transport is the dominant mode of commuting, which accounts for about 90% of daily trips. A very high ridership during peak hours is also observed in other densely populated cities in Asia, such as Tokyo, Seoul, Shanghai, and Beijing. Different public and/or private operators may operate public transit systems (Calimente, 2012), where multiple operators often exist in the same city and may manage different types of public transit services (e.g., buses and metro/subway) and/or services in different areas within the same city. For instance, there are more than five bus operators in Hong Kong that operate different lines covering different parts of the city, and different types of buses such as minibuses, double-deck buses, etc. There is also an operator for the MTR system. The main funding sources for public transit systems include ticket revenue from passengers, subsidies from government, as well as advertising income. A certain amount of revenue may also be generated from land development activities, rental of stores and vendors, and parking

386  Handbook on transport pricing and financing

fees. Services might be profit-driven and under limited regulation (e.g., red minibus in Hong Kong), or can be subsidized by local authorities but also regulated (green minibus, doubledeck bus, and the metro system in Hong Kong).17 20.3.1 Fare and Ticketing Contactless smart cards have been widely used in the public transport systems in Asia, including the Octopus card in Hong Kong, T-money in South Korea, the Suica and Pasmo cards in Japan, the NETS FlashPay and EZ-Link cards in Singapore, and the Beep card in Manila. A large number of different smart cards have been used in different cities across countries such as China and India. However, the “China T-Union” card is replacing other smart transit cards in multiple cities in China. The smart cards can often be used in multiple different public transit modes, such as buses, trams, railways or light rails, and ferries. The use of smart cards largely eases the payment process such that a range of different public transit fare schemes that incorporate spatial and/or temporal factors, as well as passenger groups, have been considered. The public transit fare can be either dependent on the traveled distance and/or based on zones. For instance, origin–destination specific fares are often adopted for bus, subway, or metro systems in cities such as Beijing, Shenzhen, Hong Kong, Tokyo, and Seoul. The fare information can often be checked through operators’ official digital trip planners. For example, Seoul Metro used a distance proportional system to determine the fares, which is based on the shortest distance between the two metro stations.18 Some cities have introduced time-dependent variable pricing that can involve peak-hour surcharges, off-peak discounts, and combinations of both. It has been found that metro passengers are sensitive to metro fares (as well as peak-hour surcharges), e.g., in Taipei as reported in the study of Lan et al. (2010). A major consideration of the above time-based differential pricing is to offer monetary incentives/disincentives to spread the traffic over time. However, the transit operator may encounter significant profit loss and a substantial subsidy from the transport authorities might be required for the operator. Following this, revenue-neutral farereward schemes have been proposed in recent studies (Yang and Tang, 2018). Such a scheme rewards travelers with one discounted or free trip during the shoulder periods after the travelers have paid for a designated number of trips during the peak. By appropriately designing the pricing levels and the reward ratio, the scheme can be revenue-neutral and Pareto-improving. Yang and Tang (2018) evaluated their approach based on three metro service lines in Hong Kong MTR, as summarized in Table 20.1. Their study showed that the optimal reward ratios (i.e., the proportions of free rides per day) under the fare-reward scheme for the three service lines under realistic demand conditions should be between 32% and 68%. More recently, Tang Table 20.1  Three metro service lines in Hong Kong (based on data reported by Yang and Tang, 2018) MTR service line

Total trip distance

Trip fare for the entire line

Ma On Shan Line

11.4 (km)

7.5 (HKD)

Island line

16.0 (km)

11.0 (HKD)

West Rail line

35.4 (km)

23.0 (HKD)

Transport pricing, funding and financing issues in Asia  387

et al. (2020a) extended the study of Yang and Tang (2018) to the case of user heterogeneity, and Tang et al. (2020b) further incorporated the transit crowding effects. Besides discounted or zero fares at off-peak hours, it is quite common that discounted fares are offered in various jurisdictions to particular social groups such as the elderly, children, students, and the disabled. To mention a few in Asia, elderly people and persons with disabilities in Hong Kong can travel on designated transit services at a concessionary fare per trip of two Hong Kong dollars.19 Moreover, Seoul Metro applied a 20% discount to teenagers and a 50% discount to children, and the metro is free for those aged above 65 (refer to the source in footnote 16). In addition, The Tokyo Metropolitan Government issues Silver Passes to elderly residents aged 70 and older to take public transit free. While the above fare-reward or discounted fare schemes have been proposed or used in Hong Kong, some empirical studies suggested the low price responsiveness of the public transport ridership, which means that reducing public transport fares will likely have a very limited impact on Hong Kong public transport ridership (Woo et al., 2020). This is also consistent with additional findings regarding statistically significant but very limited effects of the Early Bird Discount in the form of earlier departure times in Hong Kong MTR (Anupriya et al., 2020). Such a strong demand (under the dense population), together with the “Rail plus Property” model (as a form of indirect subsidy to be discussed below) indeed allows the Hong Kong MTR to earn a positive profit, e.g., 11.93 billion Hong Kong dollars in the year 2020.20 20.3.2 Subsidy Public transit systems, as an alternative to private car mode, are of large-capacity and are expected to help to reduce road congestion, fuel consumption, as well as vehicular emissions. In addition, public transit systems enable people to travel where they cannot use an automobile to travel, e.g., due to parking restrictions or road restrictions (Litman, 2010). Moreover, the Mohring Effect exists in public transport systems (Mohring, 1972; Parry and Small, 2009; Basso and Jara-Díaz, 2010), where a demand increase for public transport can yield a decrease in the waiting time costs for travelers when the demand increase is accommodated by an increase in the service frequency. Therefore, it is common that governments (including local transport authorities) subsidize public transit systems for the aforementioned social, environmental, and economic benefits. Interested readers may refer to Chapter 9 of this Handbook, which explains the core theoretical justification for subsidized service provision under scale economies. As mentioned in Section 20.2.1, Singapore indeed coupled its congestion pricing strategy with heavy investment/subsidy in the public transit system in order to manage modalsplit and reduce congestion. Subsidies are not limited to the form of direct payments from different levels of authorities, but can also include indirect subsidies such as usage of public infrastructures and services at a lower cost or even zero cost, and land development rights (especially the land and property spatially adjacent to the public transit system in operation, i.e., “Rail plus Property” model). Direct payments for unprofitable public transit services are very common across Asian countries. For example, in Mumbai, India, the rail and bus systems are directly subsidized by the government using the revenue from the electric system, while the rail fare is also subsidized partially by the funds from property taxes. Such a highly subsidized system benefits a large working population who takes the bus or rail to commute (Cropper and Bhattacharya, 2012).

388  Handbook on transport pricing and financing

Land development rights can be rewarded to public transit operators. A successful example in Asia is the MTR Corporation Limited in Hong Kong, which adopts the “Rail plus Property” model and generates substantial profits from land development programs to cover parts of the construction and operation costs of the rail system (Verougstraete and Zeng, 2014). As reported in Verougstraete and Zeng (2014), properties of more than 5 million square meter floor area were awarded to MTR between 1995 and 2010. While most metro systems worldwide depend heavily on direct public financial support, MTR is indeed highly profitable from its real estate business associated with the metro lines.21 However, it should be noted that the success of the MTR’s “Rail plus Property” model has its uniqueness, which relies on (i) a state leasehold system, (ii) very high urban density, (iii) entrepreneurial local authorities and transport agency, (iv) an appropriate legal framework, and (v) well-developed operating procedures (Murakami, 2015). It is also reported in the literature that institutional barriers (e.g., inefficient governance, rigid financial regulation, and unsupportive planning) may prevent effective transit-oriented development, even if the urban density is high (Wang et al., 2019).

20.4 HIGH-SPEED RAILWAY FINANCING The past two decades have witnessed the substantial development of high-speed rail (HSR) systems in China. At the end of 2020, China had about 38,000 km of high-speed rail track network,22 which was around 56 times of that (672 km) in the year 2008 when the first HSR line (Beijing from/to Tianjin) was put into use (Jiao et al., 2020). Nearly all HSR tracks and trains are owned and managed by the China State Railway Group Company, Ltd. (i.e., China Railway), which was formerly known as the China Railway Corporation (2013–2019), and as a part of the Ministry of Railways (before 2013). While the HSR system greatly eases inter-city movements within the country and brings economic benefit (Yang et al., 2020b), the investment required is substantial and intensive. As reported in the study by the World Bank (Lawrence et  al., 2019), the financing model for the HSR system in China has shifted from “direct investment from government (through the Ministry of Railways in China)” before 2004 to the “equity model” with multiple funding sources since 2004. Note that the Ministry of Railways (MOR) was dissolved in 2013. Under the traditional “direct investment” model, MOR provided the capital, and the regional administration was involved and responsible for implementation. In the “equity model”, China Railway (previously known as the China Railway Corporation (2013–2019), and as a part of the Ministry of Railways before 2013) often forms a joint venture with the provincial government, and occasionally may also involve a limited amount of third-party funding. Table 20.2 summarizes estimations of China HSR unit revenues and costs as of 2016 from the World Bank (Lawrence et al., 2019). As can be seen, the revenue from the fares can cover the train operating cost. However, this net gain is far from covering the infrastructure maintenance cost and infrastructure investment cost. The local regional administration and provincial government face significant financial challenges, while they are indirectly subsidized by the China Railway Corporation (CRC), which covers part of the debt service (investment cost). Economists have pointed out that existing HSR constructions and operations require substantial direct central subsidy. Some of them suggested lowering HSR fares to increase HSR usage and revenues. Overall, HSR ridership is growing (e.g., 1414 billion passenger kilometers in 2018) as the coverage of the HSR network continues to grow, which helps generate a

Transport pricing, funding and financing issues in Asia  389

Table 20.2  World Bank estimations of revenues and costs (Yuan: Chinese currency) in China HSR for two types of trains in 2016 Units

200–250 km/hr train

300–350 km/hr train

Passengers/train

Passengers

390

825

Revenue

Yuan per passenger km

0.28

0.50

Train operating cost

Yuan per passenger km

0.19

0.23

Infrastructure maintenance cost

Million Yuan per km

1.80

2.30

Infrastructure investment cost

Million Yuan per km

110

130

Source:   Based on data reported by Lawrence et al., 2019.

growing revenue. In addition, the technology standardization for the HSR system might help reduce investment and operation costs. However, it is expected that the above is still insufficient to cover the substantial infrastructure maintenance cost and infrastructure investment cost. While additional financial solutions should be explored for the Chinese HSR development and operation, the social and economic benefits from the HSR development help justify the current substantial investment/subsidy on HSR in China. As reported in the literature, the improvement in connectivity, accessibility, and economic growth is significant, even if these positive impacts differ across various regions of the country (Chen and Haynes, 2017; Chong et al., 2019; Jiao et al., 2020; Yang et al., 2020b; Zhang et al., 2020). In particular, Tao et al. (2011) carried out a cost–benefit analysis of the Hong Kong to mainland high-speed railway line. Their results show that this high-speed railway project has a positive net present value (NPV) that can be up to 2,068.49 million US dollars.

20.5 CONCLUSION AND DISCUSSION This survey highlights a few unique and important transport pricing, funding, and financing issues in Asia, including road pricing, public transit systems funding and pricing, and highspeed railway financing. We briefly summarize them below. (i) Road pricing as a congestion management tool is implemented in Singapore, which is the earliest application of congestion pricing in the world, while road pricing as an infrastructure funding source is widely adopted in the large-scale Chinese expressway system. (ii) Public transit systems are critical for supporting movements of densely distributed populations in Asia, while they are often substantially subsidized by governments and local authorities. Hong Kong MTR provides a successful example of the “Rail plus Property” model and indirect public transit subsidy, where the land development associated with the metro system generates substantial revenue for the operator and direct public transit subsidy from the government becomes unnecessary. (iii) The HSR system in China is the largest in the world and by far larger than HSR systems in other countries, which requires significant government subsidies that might be justified by positive social and economic impacts. While congestion pricing has been adopted in Singapore, it has received strong public objection in most cities. Instead, road space rationing schemes have been implemented in

390  Handbook on transport pricing and financing

many large cities in China. However, road space rationing schemes can be less efficient, especially when a rich population may buy additional cars. Researchers have proposed and theoretically examined the tradable travel credit schemes for congestion management, which does not involve direct out-of-pocket money from travelers to transport authorities. The development of smartphones, GPS, the internet, and platform technologies provide new opportunities to realize novel schemes such as tradable travel credit. It is expected that the application of such new technologies to allow better traffic management will likely be seen first in Asia (such as in Singapore and China). Road pricing has been used as a funding source for road projects. Many have studied the BOT schemes for road projects. While BOT schemes provide additional financial opportunities for new transport development, which expects to further support economic growth, it is found that the tolls on the road can delay or even eliminate those positive economic outcomes, especially for the rural areas. Therefore, tolls might have to be much lowered for social good. Also, more options and innovations should be explored to support future transport infrastructure projects. In Asia, due to the extreme urban density, public transport systems in mega-cities (such as Hong Kong, Tokyo and Beijing) are serving huge numbers of passengers, and generating substantial ticket revenue. However, an additional subsidy is still required for most public transport systems. The successful example of Hong Kong MTR regarding profitability relies on its unique “Rail plus Property” model, which might not work in most cities in Asia and in the rest of the world due to low urban density and institutional barriers. The HSR system in China is the largest in the world and requires significant investment from central and local governments. Many empirical studies have justified these investments by the positive social and economic outcomes. Different financial models for HSR system construction, operation and maintenance might be further explored, including different combinations of public and private funding sources.

NOTES 1.

World  Population  Review,  https://wor​ldpo​pula​t ion​r eview​.com​/continents​/asia​-population, accessed on March 10, 2021. 2. China’s courier industry ships 83 billion packages in 2020 (People’s Daily), http://en​.people​.cn​/n3​/ 2021​/0115​/c90000​-9809843​.html, accessed on March 10, 2021. 3. Based on data released by the Ministry of Land, Infrastructure, Transport and Tourism of Japan (www​.mlit​.go​.jp​/en​/index​.html) 4. ESCAP: Economic and Social Commission for Asia and the Pacific (United Nations) 5. Will Congestion Pricing Relieve Traffic Jams?. www​.bjreview​.com​.cn​/forum​/txt​/2010​- 05​/31​/content​_275536​.htm, accessed on March 10, 2021. 6. Revised ERP Rates from April 12, 2021, www​.lta​.gov​.sg​/content​/ ltagov​/en​/newsroom​/2021​/4​/ news​-releases​/ Revised​_ ERP​_ rates​_from​_12Apr21, accessed on April 30, 2021. 7. Installation of On-Board Units for Next-Generation ERP System to Commence in Second Half of 2021, www​.lta​.gov​.sg​/content​/ ltagov​/en​/newsroom​/2020​/9​/news​-releases​/installation​-of​-on​-board​ -units​-for​-next​-generation​-erp​-system​-to​.html, accessed on March 10, 2021. 8. Time to fix traffic in Beijing, www​.worldbank​.org​/en​/news​/opinion​/2010​/12​/20​/time​-fix​-traffic​ -beijing, accessed on March 10, 2021. 9. Beijing mulls congestion charge, www​.chinadaily​.com​.cn​/china​/2015​-12​/03​/content​_22621398, accessed on March 10, 2021. 10. The great crawl, www​.economist​.com​/china​/2016​/06​/16​/the​-great​-crawl​?frsc​=dg​%7Ca, accessed on March 10, 2021.

Transport pricing, funding and financing issues in Asia  391

11. Toll Rates of Road Tunnels, HKSAR Transport Department, www​.td​.gov​.hk​/en​/transport​_in​_ hong ​_ kong ​/tunnels ​_ and ​_ bridges ​/toll ​_ rates ​_ of ​_ road ​_ tunnels ​_ and ​_ lantau ​_ link ​/ index​.html, accessed on March 10, 2021. 12. See discussions in Section 20.2.1 of this chapter and also early explanations of this phenomenon by Bruno De Borger and Antonio Russo in Chapter 7 in this Handbook on the political economy of transport pricing. 13. Please refer to Chapter 16 in this Handbook for a comprehensive and in-depth literature review on Public-Private Partnerships. 14. People’s Daily asks three questions about the new situation of expressway tolls, https://news​.sina​. com​.cn​/c​/2020​- 01​- 04​/doc​-iihnzahk1961862​.s, accessed on March 10, 2021. 15. Deepen the reform of the toll road system and reduce the cost of crossing roads and bridges, www​. chinahighway​.com​/news​/2018​/1163263​.php, accessed on March 10, 2021. 16. Hong Kong: The Facts Transport, www​.gov​.hk​/en​/about​/abouthk​/factsheets​/docs​/transport​.pdf, accessed on March 10, 2021. 17. Public Transport, HKSAR Transport Department, www​.td​.gov​.hk​/en​/transport​_in​_ hong​_ kong​/ public​_transport​/index​.html, accessed on March 10, 2021. 18. Subway fares are calculated by unifying all metropolitan subway lines based on a distance proportional system, www​.seoulmetro​.co​.kr​/en​/page​.do​?menuIdx​=348, accessed on March 10, 2021. 19. Government Public Transport Fare Concession Scheme for the Elderly and Eligible Persons with Disabilities, HKSAR Transport Department, www​.td​.gov​.hk ​/en ​/gov​_public​_transport​_fare​_concession​/index​.html, accessed on March 10, 2021. 20. Annual Report 2020 MTR Corporation Limited. https://www​.mtr​.com​.hk​/en​/corporate​/investor​/ 2020frpt​.html, accessed on March 15, 2022. 21. The “Rail plus Property” model: Hong Kong’s successful self-financing formula, www​.mckinsey​.com​/ business​-functions​/operations​/our​-insights​/the​-rail​-plus​-property​-model#, accessed on March 10, 2021. 22. China’s high-speed rail lines top 37,900 km at end of 2020, www​.xinhuanet​.com​/english​/2021​- 01​/ 09​/c​_139654709​.htm, accessed on March 10, 2021.

REFERENCES Arnott, R., De Palma, A. and Lindsey, R. 1990. Economics of a bottleneck. Journal of Urban Economics, 27(1), 111–130. Basso, L. J. and Jara-Díaz, S. R. 2010. The case for subsidisation of urban public transport and the Mohring effect. Journal of Transport Economics and Policy (JTEP), 44(3), 365–372. Calimente, J. 2012. Rail integrated communities in Tokyo. Journal of Transport and Land Use, 5(1), 19–32. Chen, A. and Subprasom, K. 2007. Analysis of regulation and policy of private toll roads in a buildoperate-transfer scheme under demand uncertainty. Transportation Research Part A: Policy and Practice, 41(6), 537–558. Chen, Z. and Haynes, K. E. 2017. Impact of high-speed rail on regional economic disparity in China. Journal of Transport Geography, 65, 80–91. Chong, Z., Chen, Z. and Qin, C. 2019. Estimating the economic benefits of high-speed rail in China. Journal of Transport and Land Use, 12(1), 287–302. Chung, H. 2002. Some socio-economic impacts of toll roads in rural China. Journal of Transport Geography, 10(2), 145–156. Cropper, M. and Bhattacharya, S. 2012, Public transport subsidies and affordability in Mumbai, India. Urban Studies Research, 865972. Daganzo, C. F. and Lehe, L. J. 2015. Distance-dependent congestion pricing for downtown zones. Transportation Research Part B: Methodological, 75, 89–99. De Borger, B. and Proost, S. 2012. A political economy model of road pricing. Journal of Urban Economics, 71(1), 79–92. ESCAP (United Nation). 2017. Review of developments in transport in Asia and the Pacific 2017.

392  Handbook on transport pricing and financing

Feng, Z., Zhang, S. B. and Gao, Y. 2015. Modeling the impact of government guarantees on toll charge, road quality and capacity for Build-Operate-Transfer (BOT) road projects. Transportation Research Part A: Policy and Practice, 78, 54–67. Anupriya, Graham, D. J., Hörcher, D., Anderson, R. J. and Bansal, P. 2020. Quantifying the ex-post causal impact of differential pricing on commuter trip scheduling in Hong Kong. Transportation Research Part A: Policy and Practice, 141, 16–34. Han, D., Yang, H. and Wang, X. 2010. Efficiency of the plate-number-based traffic rationing in general networks. Transportation Research Part E: Logistics and Transportation Review, 46(6), 1095–1110. Hau, T. D. 1990. Electronic road pricing: Developments in Hong Kong 1983–1989. Journal of Transport Economics and Policy, 24(2), 203–214. Hau, T. D., Loo, B. P., Wong, K. I. and Wong, S. C. 2011. An estimation of efficient time-varying tolls for cross harbor tunnels in Hong Kong. The Singapore Economic Review, 56(04), 467–488. Hoang-Tung, N., Kato, H. and Le Binh, P. 2021. Modeling ceiling price for build-operate-transfer road projects in developing countries. Economics of Transportation, 28, 100235. Jiao, J., Wang, J., Zhang, F., Jin, F. and Liu, W. 2020. Roles of accessibility, connectivity and spatial interdependence in realizing the economic impact of high-speed rail: Evidence from China. Transport Policy, 91, 1–15. Lan, L.W., Wen, C.H. and Lee, H.Y. 2010. Effects of temporally differential fares on Taipei metro riders’ mode and time-of-day choices. International Journal of Transport Economics, 37(1), 97–118. Lawrence, M., Bullock, R., and Liu, Z. 2019. China’s High-Speed Rail Development. International Development in Focus. Washington, DC: World Bank. doi: 10.1596/978-1-4648-1425-9. Li, M. Z. 1999. Estimating congestion toll by using traffic count data—Singapore’s area licensing scheme. Transportation Research Part E: Logistics and Transportation Review, 35(1), 1–10. Li, M. Z. 2002. The role of speed–flow relationship in congestion pricing implementation with an application to Singapore. Transportation Research Part B: Methodological, 36(8), 731–754. Litman, T. 2010. Evaluating public transit benefits and costs: Best Practices Guidebook. Liu, W., Yang, H. and Yin, Y. 2014. Traffic rationing and pricing in a linear monocentric city. Journal of Advanced Transportation, 48(6), 655–672. Lu, Z. and Meng, Q. 2017. Analysis of optimal BOT highway capacity and economic toll adjustment provisions under traffic demand uncertainty. Transportation Research Part E: Logistics and Transportation Review, 100, 17–37. Meng, Q., Liu, Z. and Wang, S. 2012. Optimal distance tolls under congestion pricing and continuously distributed value of time. Transportation Research Part E: Logistics and Transportation Review, 48(5), 937–957. Mohring, H. 1972. Optimization and scale economies in urban bus transportation. The American Economic Review, 62(4), 591–604. Mun, S. I. and Nakagawa, S. 2010. Pricing and investment of cross-border transport infrastructure. Regional Science and Urban Economics, 40(4), 228–240. Murakami, J. 2015. Rail plus property program, Hong Kong SAR, China. Suzuki, et al. (Eds.), Financing Transit-Oriented Development with Land Values: Adapting Land Value Capture in Developing Countries, World Bank Publications, Washington, DC. Nie, Y. M. 2012. Transaction costs and tradable mobility credits. Transportation Research Part B: Methodological, 46(1), 189–203. Nie, Y. M. 2017a. Why is license plate rationing not a good transport policy? Transportmetrica A: Transport Science, 13(1), 1–23. Nie, Y. M. 2017b. On the potential remedies for license plate rationing. Economics of Transportation, 9, 37–50. Nie, Y. M. and Yin, Y. 2013. Managing rush hour travel choices with tradable credit scheme. Transportation Research Part B: Methodological, 50, 1–19. Parry, I. W. and Small, K. A. 2009. Should urban transit subsidies be reduced? American Economic Review, 99(3), 700–724. Phang, S. Y. and Toh, R. S. 2004. Road congestion pricing in Singapore: 1975 to 2003. Transportation Journal, 16–25. Pigou, A. C. 1920. The Economics of Welfare. London: Macmillan.

Transport pricing, funding and financing issues in Asia  393

Seik, F. T. 2000. An advanced demand management instrument in urban transport: Electronic road pricing in Singapore. Cities, 17(1), 33–45. Shi, S., Yin, Y. and Guo, X. 2016. Optimal choice of capacity, toll and government guarantee for build-operate-transfer roads under asymmetric cost information. Transportation Research Part B: Methodological, 85, 56–69. Tan, Z. and Yang, H. 2012. Flexible build-operate-transfer contracts for road franchising under demand uncertainty. Transportation Research Part B: Methodological, 46(10), 1419–1439. Tan, Z., Yang, H. and Guo, X. 2009. Build-operate-transfer schemes for road franchising with road deterioration and maintenance effects. In Transportation and Traffic Theory 2009: Golden Jubilee, 241–261. Boston, MA: Springer. Tang, Y., Jiang, Y., Yang, H. and Nielsen, O. A. 2020b. Modeling and optimizing a fare incentive strategy to manage queuing and crowding in mass transit systems. Transportation Research Part B: Methodological, 138, 247–267. Tang, Y., Yang, H., Wang, B., Huang, J. and Bai, Y. 2020a. A Pareto-improving and revenue-neutral scheme to manage mass transit congestion with heterogeneous commuters. Transportation Research Part C: Emerging Technologies, 113, 245–259. Tao, R., Liu, S., Huang, C. and Tam, C. M. 2011. Cost-benefit analysis of high-speed rail link between Hong Kong and Mainland China. Journal of Engineering, Project, and Production Management, 1(1), 36–45. Van den Berg, V. and Verhoef, E. T. 2011. Congestion tolling in the bottleneck model with heterogeneous values of time. Transportation Research Part B: Methodological, 45(1), 60–78. Verougstraete, M. and Zeng, H. 2014. Land value capture mechanism: the case of the Hong Kong mass transit railway. United Nations Economic and Social Commission for Asia and the Pacific. Wang, J., Samsura, D. A. A. and van der Krabben, E. 2019. Institutional barriers to financing transit-oriented development in China: Analyzing informal land value capture strategies. Transport Policy, 82, 1–10. Wang, X. and Yang, H. 2012. Bisection-based trial-and-error implementation of marginal cost pricing and tradable credit scheme. Transportation Research Part B: Methodological, 46(9), 1085–1096. Wang, X., Yang, H. and Han, D. 2010. Traffic rationing and short-term and long-term equilibrium. Transportation Research Record, 2196(1), 131–141. Wang, X., Yang, H., Han, D. and Liu, W. 2014. Trial and error method for optimal tradable credit schemes: The network case. Journal of Advanced Transportation, 48(6), 685–700. Wang, X., Yang, H., Zhu, D. and Li, C. 2012. Tradable travel credits for congestion management with heterogeneous users. Transportation Research Part E: Logistics and Transportation Review, 48(2), 426–437. Woo, C. K., Cheng, Y. S., Li, R., Shiu, A., Ho, S. T. and Horowitz, I. 2015. Can Hong Kong pricemanage its cross-harbor-tunnel congestion? Transportation Research Part A: Policy and Practice, 82, 94–109. Woo, C. K., Liu, Y., Cao, K. H. and Zarnikau, J. 2020. Can Hong Kong price-manage its public transportation’s ridership? Case Studies on Transport Policy, 8(4), 1191–1200. Yang, H. and Huang, H. J. 2005. Mathematical and Economic Theory of Road Pricing. New York: Elsevier Science Inc. Yang, H. and Meng, Q. 2000. Highway pricing and capacity choice in a road network under a build– operate–transfer scheme. Transportation Research Part A: Policy and Practice, 34(3), 207–222. Yang, H. and Tang, Y. 2018. Managing rail transit peak-hour congestion with a fare-reward scheme. Transportation Research Part B: Methodological, 110, 122–136. Yang, H. and Wang, X., 2011. Managing network mobility with tradable credits. Transportation Research Part B: Methodological, 45(3), 580–594. Yang, J., Purevjav, A. O. and Li, S. 2020a. The marginal cost of traffic congestion and road pricing: Evidence from a natural experiment in Beijing. American Economic Journal: Economic Policy, 12(1), 418–453. Yang, Z., Li, C., Jiao, J., Liu, W. and Zhang, F. 2020b. On the joint impact of high-speed rail and megalopolis policy on regional economic growth in China. Transport Policy, 99, 20–30. Zhang, F., Yang, Z., Jiao, J., Liu, W. and Wu, W. 2020. The effects of high-speed rail development on regional equity in China. Transportation Research Part A: Policy and Practice, 141, 180–202. Zhu, D. L., Yang, H., Li, C. M. and Wang, X. L. 2015. Properties of the multiclass traffic network equilibria under a tradable credit scheme. Transportation Science, 49(3), 519–534.

21. Transport pricing in Europe Chris Nash and Heike Link

21.1 INTRODUCTION Historically Europe has seen a wide diversity of approaches to transport pricing. Some countries have relied on high fuel taxes as the main way to charge for the use of roads, whilst others have toll roads. Some countries saw rail and other public transport as largely a commercial business, with relatively high-cost recovery levels, whilst others saw them as a social service with low fares funded by government subsidies. With the growth of the European Union, there has been pressure for a more uniform approach to transport pricing. There was a risk that, particularly in the freight market, subsidies could be used to provide unfair competition with products from other countries, whilst transport operators could gain from being based in a low-tax country and buying their fuel there even for international journeys. In this chapter, we will concentrate on the attempts of the European Commission to establish a consistent approach to transport pricing throughout the member states. Their approach has been heavily grounded in the economic principle of marginal social cost pricing (see theoretical overviews in Chapters 2 and 3 of this Handbook), so one aim of this chapter is to provide a case study of the practical problems that arise in implementing such a principle. The two areas in which the Commission has been most concerned to achieve a compatible approach in all member states have been rail track access charges and charging for the use of roads, particularly for heavy goods vehicles, and these have been chosen as case studies for more detailed examination. The next section expands on the background. Section 21.3 explains the basis of European transport pricing policy. Sections 21.4 and 21.5 consider in turn rail track access charges and charging for the use of roads. Section 21.6 draws conclusions.

21.2 BACKGROUND Dating back to the Treaty of Rome in 1957, it has been recognised that European integration requires a common transport policy to reduce the barriers to trade between member countries. In the early years, the emphasis was on harmonising regulation and avoiding national discrimination, for instance through freight rates that favoured domestic ports over ports in neighbouring countries. Major efforts to establish a common transport policy only followed the 1985 European Court of Justice ruling that the Commission had failed in its Treaty of Rome obligation in this regard (Vickerman, 2021). In 1992, the Commission published its White Paper on the future of the common transport policy (EC, 1992). This was to be based on competition, with the removal of regulation and barriers to entry wherever this was feasible. Competition would ensure fair and efficient 394

Transport pricing in Europe  395

pricing. But it was realised that transport infrastructure is often a natural monopoly, and could not be subject to this approach, whilst growing concern about transport externalities regarding congestion, safety, environment, and accidents needed also to be reflected in infrastructure charges (EC, 1995). This approach was reiterated in a series of further transport White Papers (EC, 2001; EC, 2011), although even in 2019, a study (Van Essen et al., 2019) found that little progress had been made, with no mode of transport covering marginal social cost from its charges, although rail came nearest. The reason for the lack of progress is predominantly opposition from member states. In particular, raising charges for freight transport is seen as potentially damaging to the economy, particularly by states on the periphery of Europe. The political economy of opposition to urban road pricing is discussed in Chapter 6 of this Handbook, whilst raising the price of air travel is no more popular. The Commission has been particularly active in developing policy on infrastructure charging in two areas – rail and road freight vehicles. The reasons for regarding these as priorities are as follows. First, whilst in other areas of transport infrastructure, such as ports and airports, there is a reasonable degree of competition so the need for government intervention is less (although of course main locations such as Heathrow Airport or the Port of Rotterdam do undoubtedly possess monopoly power), in rail and road infrastructure most often each country has a single monopoly supplier. Second, the policy of introducing competition within the rail sector is based on opening access to the monopoly infrastructure to new entrants. The Commission believed that this required introduction of track access charges based on marginal social cost, and ensuring that these are charged in a way that does not favour incumbents requires that such charges are applied equally to incumbents and new entrants and also that both receive the same quality of service. In order to reduce the risk of discrimination, the Commission has been anxious to achieve at least a degree of separation between infrastructure and the train operations of the incumbent. It has also been policy to encourage the growth of rail as a relatively environmentally friendly mode so it has wished to see low charges where possible. At the same time, more than half of rail freight is international. Achieving some degree of coherence in charges between neighbouring countries is important to encouraging this traffic; see more details on the theory of tax competition between governments in Chapters 3 and 18 in this Handbook. In road freight, the issues are somewhat different. Traditionally road hauliers have paid annual licence fees to the government of the country in which they are based, and fuel tax wherever they bought their fuel. The levels of these charges have varied greatly. In addition, some countries have long had tolls on motorways; often associated with lower licence fees and/or fuel tax. The result is that when the market was opened to road hauliers to compete throughout Europe no matter in which country they were based, there was nothing like a level playing field between hauliers based in low-tax countries and those in high. The obvious solution is a charging system which charges vehicles according to infrastructure and external costs in the country in which they are running rather than the level of charges in the country in which they are based. Development of such a charging system has been encouraged by a series of directives concerning the EuroVignette. In the next section, we consider the approach taken by the European Commission to transport infrastructure pricing.

396  Handbook on transport pricing and financing

21.3 TRANSPORT INFRASTRUCTURE PRICING POLICY The Green Paper (EC, 1995) started with the argument that very different pricing policies were followed in different member states and for different modes. Harmonisation was needed for the efficient working of the internal market. Until the 1995 Green Paper, there had been little discussion of the relevance of external costs to European transport infrastructure pricing policy. In general, regulation had been seen as the most appropriate approach to dealing with environmental and safety problems. But that paper argued strongly for the use of pricing policy as well, since this would give users incentives to reduce externalities whenever it was cost-efficient to do so, rather than simply to meet the imposed standards. It was argued that pricing needed to be highly differentiated, to reflect the external costs imposed in different locations, at different times and by different types of vehicle. This would give the correct incentives, not simply to provide the right overall transport volume, but also to use appropriate routes, times of day and types of vehicle. The paper was clear that the most appropriate approach to efficient transport pricing was to base prices on marginal social cost. But it recognised from the first the conflict with the ‘user-pays’ principle that users should meet all the costs of transport infrastructure provision and use. It also concluded that full cost recovery was desirable, as otherwise subsidies would be needed from elsewhere in the economy, and raising the money for these subsidies would require distorting taxes elsewhere in the economy. Whilst it has been shown that, given constant returns to scale and zero indivisibilities, a combination of short-run marginal social cost pricing and optimal capacity will exactly recover total cost (see Chapter 3 in this Handbook), these assumptions do not hold in practice. For instance, a double-track railway has some four times the capacity of a single-track one, whilst costing little more than twice as much (Nash, 2017). On the other hand, the marginal cost of expanding an existing system, particularly in built-up areas, might considerably exceed the average cost of building the infrastructure when it was built many years ago. The following White Paper (EC, 1998) made clear for the first time that it had in mind charging according to short-run marginal social cost, rather than long, not by arguing through the issue, but by the examples it used. It argued that marginal cost can include: ● ●

●

● ●

Operating costs. Infrastructure damage costs: maintenance costs, wear and tear of the infrastructure, reflected by such costs as resurfacing of roads, rails and runways. Congestion and scarcity costs: The cost of time delays to other users or non-users, resulting from congested traffic flows (on roads, queues for airports or railway stations). Moreover, a transport operator’s use of infrastructure may prevent another operator from using it (e.g. an airport runway). Environmental costs: air, water, and noise pollution. Accident costs: Costs in terms of material damage, pain and suffering and production losses.

Notice the absence of any reference to capital costs or financial charges. In general, where member states had supported the principle of social marginal cost pricing, they had been more likely to support long-run marginal cost pricing than short (Quinet, 2005). This may give better incentives to users regarding decisions which will have lagged impacts on demand (such

Transport pricing in Europe  397

as home or work location, sourcing of inputs, location of distribution depots, etc.), and also on producers, who will be directly compensated for any investments they undertake. On the other hand, indivisibilities make the estimation of meaningful estimates of long-run marginal cost difficult in the transport sector and the long life of assets means that there are severe time-lags in adjusting capacity to demand, so that long-run marginal cost pricing may lead to under or over the use of assets for prolonged periods of time. An alternative which may be better in some cases is to charge short-run marginal cost but to smooth out fluctuations caused by changes in capacity. There was a strong attack on the principle of marginal cost pricing from a number of member states that had traditionally supported full cost recovery (Rothengatter, 2003). Amongst these are the second-best arguments from the theory that marginal social cost pricing is only optimal for a particular good if the same principle is adopted for all goods, and that marginal cost pricing is not optimal if it leads to a deficit, which has to be financed by distorting taxes elsewhere in the economy. As will be seen later, when it comes to the application of marginal social cost pricing for instance to rail track access charges, the Commission has recognised these arguments and allowed for some adjustment to allow for them. At the practical level, the most important arguments are the complexity of measurement of marginal social cost and of administering pricing systems that are able to differentiate according to the circumstances between which marginal social cost differs. On the first of these issues, the Commission has sponsored major research projects, results of which have been incorporated in the Handbook of Transport Externalities, the latest edition of which was published in 2019 (EC, 2019). On the second, practical application has allowed for a reasonable degree of averaging, so that more is not spent on complex pricing systems than is justified by the benefits. Research (see Matthews and Nash, 2004 for a summary) has suggested that whilst universal marginal social cost pricing in transport might lead to a fairly high level of cost recovery for the transport sector as a whole, there would be important surpluses and deficits for particular modes and traffic types. Congestion pricing on roads in congested urban areas would lead to large surpluses, whilst marginal cost pricing on roads in rural areas and rail infrastructure wherever it was not heavily congested would require subsidy. These sorts of transfers between urban and rural areas and between rail and road are highly controversial and many argued that it was unfair for anyone to pay more than the cost of providing the infrastructure they use. However, implementing full cost recovery at such a detailed level would require very highly differentiated pricing systems and arbitrary allocations of costs between different types of traffic (for example, passenger and freight). Equity principles in economic decision taking usually rely more on identifying the characteristics of those bearing the costs and benefits and identifying their incomes and whether the results appear to be discriminatory in other characteristics. Such transfers may be made easier by the merging of the infrastructure managers for road and rail, such as took place in Sweden with the establishment of Trafikverket in 2010; Finland has since followed suit. Whilst the Commission’s approach is based on a relatively simple prescription from transport economics, in reality complexities inevitably crowd in. This does not mean, however, that a totally different theoretical approach to pricing policy needs to be adopted, or that full cost recovery as a principle is a good starting point. For implementation, the questions then turn to what to do in the face of these complexities. The following sections consider these issues in turn for rail infrastructure charges and for road transport infrastructure.

398  Handbook on transport pricing and financing

21.4 RAIL TRACK ACCESS CHARGES Starting with some types of international freight in Directive 91/440, European rail policy has progressively opened the rail market to new entries. There is now open access for commercial freight and passenger services and the process will be completed with compulsory competitive tendering of public service contracts in 2023, although with provision for exceptions. Whilst the same company was responsible for both infrastructure and all train operations, there was no need for a system of track access charges, but as soon as the market was opened to new entry this became essential for the implementation of marginal social cost pricing. It was recognised from the beginning that the track access charging system should be non-discriminatory, and that required at the least that the incumbent implement separate accounts for infrastructure and operations with transparent payments for use of the infrastructure. Current legislation requires that infrastructure and operations should each have their own management. Many European countries have placed the infrastructure in a separate company from any train operator; elsewhere (as in Germany, Italy, Austria, and France) they are separate subsidiaries of the same holding company (Nash et al., 2018). Legislation (Directive 2001/14) required that charges should be based on the direct cost of running the train service; the definition of direct cost makes clear that it is essentially a short-run marginal cost. But the likelihood that there would be a need for a higher level of cost coverage than this would achieve, is allowed for. Non-discriminatory mark-ups are permitted to contribute to the funding, provided that these do not exclude any market segment able to pay the direct cost. The mark-ups may be differentiated according to various market segments, including freight commodities and types of passenger train. Charges are also permitted for scarcity and environmental costs, but environmental charges must not raise the overall cost of rail transport unless equivalent charges were levied on other modes. There is also the possibility of allowing for divergences between price and marginal social cost on road haulage by the provision of a time-limited subsidy. However, there remained a major divergence in terms of how the directives were interpreted. Some countries, for instance, Germany, which saw the infrastructure manager as a commercial body, which should be charging train operators the full cost of providing and maintaining the infrastructure, essentially interpreted direct cost as average cost less those costs borne by the state budget. Others, including Sweden, saw the infrastructure manager as a public agency, providing the infrastructure out of general tax revenue and charging for its use on a marginal social cost basis (ITF, 2008). When defining rules for track access charges (TAC), the latest statement is to be found in the recast of the first railway package, Directive 2012/34/EU, and the regulation on calculating infrastructure charges, Regulation (EU) 2015/909. Direct costs may include additional maintenance and renewal costs and operating costs where additional staff are required. It was recognised that charges based on direct cost would only give the right incentives regarding the choice of rolling stock if they were differentiated according to the characteristics of rolling stock (such as mass or axle weight), and also that direct cost may vary according to the characteristics of the track (such as curvature). Nevertheless, averaging is permitted to avoid excessive complications. It is clear from the way direct costs are defined that these are intended to reflect shortrun marginal social costs – that is to say, they exclude capital costs of expanding capacity. However, if the infrastructure manager behaves commercially, the presence of scarcity charges and the failure to charge for capacity provides infrastructure managers with an incentive to

Transport pricing in Europe  399

keep capacity scarce. This perverse incentive is countered in the legislation by a requirement for infrastructure managers (IMs) to prepare and appraise capacity enhancement plans where capacity is declared scarce. In any event, all the main European rail infrastructure bodies are publicly owned and may be expected to use social cost-benefit analysis in assessing the case for infrastructure enhancement. Some authors also argue that marginal cost pricing in general fails to promote efficiency by failing to give both infrastructure managers and train operators sufficient incentive to control total costs (Rothengatter, 2003). Again, this is tackled by a separate measure requiring member states to ensure that infrastructure managers have the incentive to reduce costs either through the regulatory system or through a multi-annual contract with the state. The use of performance schemes is required to compensate train operators for poor performance on behalf of the infrastructure manager (and vice versa). Such schemes may also have the effect of avoiding incentivising the infrastructure manager to achieve cost savings at the expense of quality of service. In practice the most important costs reflected in track access charges are wear and tear costs; mark-ups are also becoming a very substantial element of track access charges and in some countries such as Germany they are the dominant element. Congestion or scarcity costs probably should be more often levied as traffic grows. Environmental costs are rarely a part of charges, except that charging for noise costs for freight trains is strongly promoted by the legislation, in order to encourage the use of wagons with quieter braking systems. 21.4.1 Wear and Tear Costs The European Commission interprets wear and tear costs as including both additional maintenance costs resulting from an increase in traffic and accelerated renewals (Nash et al., 2018). There are essentially two approaches to quantify wear and tear costs (Wheat et al., 2009). The first is a mix of accounting and engineering approaches, which identifies different categories of costs such as maintenance and renewal of track, ballast, sleepers and signalling, and uses engineering models or judgement to divide these into fixed and variable costs. Marginal cost is then approximated as average variable costs. The second is econometric. Data is collected on costs and traffic levels for a large number of sections of track, and costs regressed on various measures of traffic levels and characteristics of the track. Marginal cost is then found directly from the estimated coefficients. There seems good reason to prefer the econometric approach as it provides evidence on what actually happens rather than theoretical models or judgement. But there are challenges. First, it is necessary to ensure that the data on costs reflect the actual costs rather than some sort of allocation of costs measured at a higher level of aggregation. Second, marginal costs actually vary with the detail of rolling stock characteristics (gross weight, axle weight, type of suspension, etc.) and track (rail weight, curvature, type of sleeper, etc.). It is not usually possible to get sufficient data to allow for all these effects in the econometric estimation, so a mixture of econometric estimation and engineering models and judgement is needed to be able to differentiate charges sufficiently to give correct incentives on the type of rolling stock and the route needed. Most countries have limited differentiation in their charges with respect to these factors, which means a serious lack of encouragement for track-friendly vehicles. A third problem with the econometric approach applies particularly to renewals. Renewals generally only take place at intervals of several years, and are determined by cumulative

400  Handbook on transport pricing and financing

traffic volumes over those years. However, adequate data to model this are rarely available and often it is necessary to use data for a single year or small number of years, and assume that traffic data for those years reflects long-run levels. Moreover, in a single year many sections of track will have no renewals. Techniques for dealing with these problems have been developed (e.g. Odolinski et al., 2020) and include duration approaches and corner solution models. It appears that econometric estimates systematically tend to give higher estimates of marginal costs than engineering/accounting approaches. This may be because the former reflect what actually happens and the latter what is estimated should happen under optimal maintenance and renewal conditions. 21.4.2 Congestion and Scarcity The second major element of direct cost is congestion and scarcity. The distinction between congestion and scarcity costs is as follows (Johnson and Nash, 2008). Congestion costs reflect the additional delays that arise when systems are loaded close to capacity. It is sometimes disputed whether congestion costs exist at all on a rail system as, unlike roads, it operates on a timetable designed to ensure that unplanned conflicts do not occur on the track. However, it is widely observed that the reliability of rail services tends to reduce as track capacity approaches its maximum and the margin for recovery from delays reduces. Econometric estimates of this impact may be made, and were used for some years in estimating the so-called capacity charges for Britain (Gibson et al., 2002). However, these charges have now been abandoned, as the billing system did not allow for differentiating them by time of day, and with most passenger services in Britain operating under franchises, the scope for varying the train service in response to track access charges was very limited. Such charges would be more important in countries where a higher proportion of services is run on a commercial basis. Scarcity costs are the costs of being unable to accommodate some operators on the system at the time and place they would want because of a lack of capacity. It has often been argued that these would be best accommodated by auctioning paths. However, such auctions would be complicated. Track capacity may be used in many different ways to link different origins and destinations with different stopping patterns. Moreover, the willingness to pay for one path will depend on the other paths the operator (and its rivals) obtain and the consequent ability to construct an efficient and attractive timetable. Nilsson (2002) has put forward a suggested approach, but it is complex and involves iteration. In fact, auctioning has been little used; where scarcity charges have been levied they have typically been based on a pragmatic approach of introducing low charges and if they are not effective gradually raising them, or model-based calculations of the opportunity cost of paths (Johnson and Nash. op cit). There have been a number of recent attempts to quantify the scarcity value of paths in Sweden (Ali et al., 2020; Börjesson et al., 2021). 21.4.3 Mark-Ups Determining the appropriate magnitude of mark-ups is also complicated. Where the markup will typically simply be passed on to the customer in the form of a higher price, then the impact of the change in the final price can be estimated from final demand elasticities. For freight traffic, this will typically relate to the commodity carried and the length of haul (CEPA, 2017a). Then Ramsey pricing may be implemented. Where the impact will be partly

Transport pricing in Europe  401

or wholly on the frequency of service, as is often the case for passenger services, it is more complicated. If the profitability of each individual train may be estimated, then the possible markup for each train in the timetable will be known. The trains that will cease to run as a result of any particular pricing change may then be estimated (CEPA, 2017b). For passenger services run under public service contracts, mark-ups are largely a political decision, determining the degree to which costs are borne by central or local government. Nevertheless, there is an argument that these services should at least cover their avoidable fixed costs (e.g. tracks used only by them), and that these should be charged for by a fixed element in a twopart tariff. Austria and Germany are examples of countries that use the Ramsey–Boiteux rule to determine mark-ups. Both countries assume that changes in track access charges are fully passed on to final customers both in long-distance and regional rail passenger services (with the latter usually being operated under public service contracts) and freight. Germany bases the mark-ups on its own elasticity studies (BVU et al., 2016; TNS Infratest et al., 2015; KCW et al., 2018). Austria has used a meta-analysis of international studies for rail passenger transport and a scoring study for rail freight based on studies for the United Kingdom (see SCK, 2020). 21.4.4 Final Comment It is clear that track access charges following these principles could be very complex, varying with the characteristics of the train and of the track, as well as by time of day. Wear and tear costs should be based on vehicle or gross tonne-kilometres, differentiated by vehicle characteristics, whereas charges for congestion and scarcity costs should be based on train kilometres, since it is the number of trains run which determines the level of capacity utilisation of the track. In practice, most European countries have much simpler systems, in some cases based solely on train kilometres. A degree of simplification may be warranted (the British system is by far the most complex, with separate charges for literally hundreds of different types of vehicle), but it is important in a vertically separated system that the correct incentives are given to train operators regarding the number and type of trains they run, and the route and time at which they run them. Clearly, charges based on these principles will differ between countries both because of differences in the traffic handled and because of differences in required mark-ups. Nevertheless, the scale of the differences is so large as to suggest that there remain differences also in the measurement of marginal social cost (Nash et al., 2018).

21.5 CHARGING FOR THE USE OF ROADS As discussed in Section 21.2, it is widely recognised that the necessary conditions to apply first-best marginal cost prices will seldom be met in reality (see the theoretical foundations of second-best pricing problems in Chapters 3 and 8 in this Handbook). Apart from imperfectly priced commodities in related markets (for example, the housing market) or even within the same market (for example charging only parts of traffic such as heavy goods vehicles (HGV) or only parts of the network), charging first-best prices in road traffic faces specific challenges compared with rail. Optimal charging mechanisms to set highly differentiated charges for all types of road users and at all links of the whole road network would be required. Road users

402  Handbook on transport pricing and financing

would not only need to have perfect information on these highly differentiated charges but also the cognitive capability to process this information. Furthermore, the mix of individual and commercial users of road infrastructure, the dynamics of congestion and the spatial dimension, e.g. interurban versus urban networks, make a textbook application of first-best prices in the road sector even more complicated than in rail. Therefore, second-best solutions have increasingly gained attention and a rich body of theoretical work has become available, as discussed in Chapters 2 and 3 of this Handbook. Given this, the question arises why in practice it is hard to find tolling schemes that were designed on the grounds of this research. Apparently, HGV charges in Europe are based on average cost as stipulated by EU legislation (see Section 21.5.1). Urban road pricing schemes which are under the responsibility of member states, have evolved from schemes that were purely motivated to raise financing to schemes aimed at curbing congestion (see Section 21.5.3). However, the level of charges as well as the location and shape of tolling cordons are rather driven by practical and acceptability issues than derived from sophisticated algorithms to determine second-best prices. In what follows, we give an overview of legislation and practice of interurban road pricing, with a focus on HGV charging. This is followed by a review of urban congestion charging schemes. 21.5.1 EU Legislation Charging HGVs for the use of roads has been the priority area of the European Union’s charging policy due to their central role in cross-border transport and thus the functioning of the European market. Attempts to set common rules for levying HGV charges, aimed at achieving a fair level playing field for hauliers transporting goods in the EU, started in the 1990s of the last century with EU Directives 1993/89/EEC and 1999/62/62 EC. Subsequent amendments of these directives (notably Directives 2006/38/EC and 2011/76/EC) were aimed at gradually moving from time-based to distance-based schemes and from exclusively infrastructure cost-related tolls to charges that reflect (parts of) environmental, noise, and congestion costs. The most recent directive is based on an average cost principle. It stipulates that charges shall be related to the cost of construction, operation, maintenance, and development of the roads concerned. They may also include a return on capital and/or a profit margin that is based on market conditions. Furthermore, HGV charges are allowed to vary according to air pollution, noise, and congestion costs, but these variations are capped: air pollution charges shall not exceed the charge in the cleanest emission class by more than 100%. Congestion charges shall not exceed the maximum level of the weighted average charge by more than 175% and shall only be charged for a maximum of five hours of peak periods. Already existing tolling schemes are not subject to the rules of this directive. As indicated at the beginning of this chapter, there is a discrepancy between the HGV charging directive and the EU’s goal to charge transport users the social marginal cost they cause. HGV charges are tied to average costs instead of (short-run) social marginal costs, and the extent to which the overall charge level can reflect environmental, noise, and congestion costs is capped by the aforementioned ceilings. Amongst the reasons for adopting an average cost principle for HGV charging in the EU were concerns of member states that a marginal cost scheme, based on construction, maintenance, and operation costs, will not yield sufficient revenue to cover costs. In addition, member states suspected that national governments might have incentives to charge transit traffic above marginal costs causing inefficient detouring

Transport pricing in Europe  403

as well as distortions to production and distribution decisions. Furthermore, calculating and supervising the adequacy of average cost-based charges was assumed less complicated and estimates easier to replicate than marginal cost calculations involving more complexity. It may be questioned whether it is fair or efficient to change HGVs for use on the roads but not other vehicles. Based on general equilibrium models, Calthrop et al. (2007) examine the second-best problem of charging only freight transport but not passenger cars and estimate appropriate prices. Starting in 2017, the Commission has sought to extend the legislation on heavy goods vehicle charging to cover all vehicles. No agreement has been reached on this yet, but the policy of encouraging the complete replacement of petrol and diesel-fuelled road vehicles by electric vehicles makes such a development important. Since electric vehicles do not pay fuel taxes, and there is currently generally no replacement charge, as it stands electric vehicles generally pay nothing towards the marginal social cost of their use of the road system. 21.5.2 Overview of HGV Charges in Europe Charging for road use in Europe can broadly be divided into three types of schemes: 1) tolling on motorways that are operated by private and/or public concessionaires 2) time-based schemes (vignettes) for motorways operated by concessionaires and/or by the state, (the original vignette was a supplementary licence posted on the window of the truck showing the time period for which charges had been paid) 3) distance-based charging schemes for motorways and other parts of networks that are under responsibility of the state To the first group belong countries such as France, Italy, Spain, Portugal, and Norway that have a long tolling tradition. In these countries, concession agreements between the government and concessionaires were already closed in the 1950 and 1960s of the last century, motivated by an interest in a faster development of infrastructure. These countries gave (within different institutional arrangements) concessionaires the right and obligation to build, maintain, and operate motorways as well as specific parts of the network such as tunnels and bridges, and to charge users by tolls (see PWC, 2014 for an overview). After the fall of the Iron Curtain a number of Eastern and South Eastern European countries, such as Hungary, Croatia, Serbia, Slovenia, and Poland also used the concession framework to meet the backlog demand for infrastructure (Carpintero, 2010). Despite differences in the concession schemes, the common characteristic is that they are aimed at the recovery of construction, maintenance, and operation costs as well as remunerating the operator, e.g. focused on charging for pure infrastructure costs excluding other types of usage-related costs such as air pollution, noise, and congestion. The mostly distance-based charges are differentiated by type of vehicles (e.g. motorcycles, passenger cars, and HGVs/buses) whereby HGVs are defined as vehicles with a maximum gross vehicle weight of 3.5 tonnes and more. The ‘historical’ tolling schemes do not fall under the rules of the EU Directive on HGV charging, however, many of the schemes use different tolls for vehicles with a different number of axles, sometimes tolls also vary by the emission standard of the vehicle. Frequent users of motorways are often eligible for discounts that are granted within rental or purchase of OBUs.​ The second group of countries has been applying time-based charges (vignettes), often both for passenger cars and HGVs. This group comprised originally more countries than

404

35 30

5

4***

4

Poland

Croatia

3.5 tonnes

3.5 tonnes

3.5 tonnes

3.5 tonnes

3.5 tonnes

3.5 tonnes

3.5 tonnes

3.5 tonnes

3.5 tonnes

Threshold of vehicle weight

2 classes

2 classes

No

No

No

No

yes

2 classes

2 classes

Weight

No

Yes**

3 classes

2 classes

2 classes

No

2 classes

2 classes

2 classes

Axles

Differentiation of tolls by

No

4 classes**

3 classes

No

No

No

No

No

No

Emissions

OBU Toll gates

OBU (ViaToll)

OBU (HU-GO) Manual login

OBU Pre-paid Toll Card

OBU Toll gates

Automatic AutoPass*

OBU Toll gates

OBU Toll gates

OBU Toll gates

Method of charging

* Linked to the car registration number. Can also be used on bridges and ferries in DK and S. **) On motorways A1, A2, and A4. Differentiation by emission classes only for use of state-owned motorways. *** Apart from four concessionaires there are also state-owned motorways. Notes:   OBU stands for On Board Unit. Hungary has since moved first to a system of vignettes and then to distance-based charges. Sources:   PWC, 2014. Own research.

30

21

Hungary

30

–

24

38

Norway

32

Portugal

32

Spain

30

5

24

Italy

30

Average duration of concessions

Greece

22

France

Number of concessionaires

Table 21.1  Tolling schemes for HGV on concessionary motorways for HGV in Europe (as of 2012)

Transport pricing in Europe  405

now, amongst them Austria and Switzerland where motorway vignettes have had a long tradition (since 1985 in Switzerland and 1997 in Austria).1 The core cluster forms the so-called EuroVignette countries, originally consisting of Belgium, Denmark, Germany, Luxembourg, and the Netherlands. These countries have signed an international agreement, based on Article 8 of the EuroVignette Directive, which established a common system of vignettes valid within the participating countries.2 Sweden joined the agreement in 1997. The EuroVignette for HGVs is based on the rules set in the above directive, differentiated by the number of axles, the emission standard, and duration (yearly, monthly, weekly, daily). It can be booked online and entitles the hauliers to use the motorways of the respective countries. Apart from the EuroVignette countries, also some Eastern and South Eastern countries such as Romania, Bulgaria,3 and the Baltic states have been charging vignettes for HGV.​ The third group comprises countries where the state, which remains responsible for constructing, maintaining, and operating the roads, has introduced distance-based HGV charges, sometimes with a concession agreement for toll collection by a private company. Switzerland was the European pioneer in introducing a distance-based HGV charge for vehicles from 3.5 tonnes max GVW upwards, the so-called LSVA. It was followed by Austria in 2004 and Germany in 2005. In 2010, the Czech Republic and Slovakia introduced distance-based HGV charges, and Belgium followed in 2016, 2011, and 2013, respectively; the concessionaires in Poland and Hungary moved from vignettes to distance-based schemes, and most recently Bulgaria has replaced its vignette scheme with distance-based charges. Based on the rules given in the EuroVignette Directive, HGV charges in these countries are all differentiated by number of axles, in some countries combined with weight classes, and in all countries by emission classes. Austria, Germany, Switzerland, and the Czech Republic have used the option to charge air pollution and noise charges on top of infrastructure costs. Furthermore, Austria and the Czech Republic also apply different charges for certain time periods (Austria day/night, in the Czech Republic a special charge is applied between 15:00–20:00 on Fridays). Whilst all countries of this group apply charges to HGV defined as vehicles with 3.5 tonnes and more, Germany does not charge vehicles from 3.5 to 7.5 tonnes.​ Despite being tied to average costs, these schemes have characteristics that come closer to marginal costs than the duration-based schemes and also than the traditional tolling schemes in the first group of countries. These are the linking of charges to the distance travelled together with the variation of charges by number of axles and (partly) by weight classes, that enable a closer reflection of road damage. Furthermore, charge variation by emission classes, the explicit consideration of air pollution charges on top of pure infrastructure charges and, in some rare examples also the differentiation of charges by noise, move HGV charging schemes in these countries closer to marginal costs. The average cost principle as fixed in the EU charging directive implies that infrastructure costs (e.g. the cost of operation, maintenance, and renewals) are overcharged4 whilst air pollution costs (considering NOx, PMs, SO2, NMVOC, and NH3) are undercharged because of the maximum percentage of charge variation requested.5 Furthermore, there is a failure to reflect congestion appropriately and therefore, at least on highly congested parts of the network overall charges will be too low. However, it should be noted that including congestion costs in charging would require charging all vehicles including passenger cars for congestion, and remains a future task. There is evidence that the differentiation of charges by emission classes have led to a shift of truck fleets towards less polluting engines, both for countries with vignettes and with

406

All roads

Source:  Own research.

Online Booking

Online Booking Electronic Vignette Electronic Vignette Electronic Vignette Electronic Vignette

Yearly Monthly Weekly Daily Yearly Monthly Weekly Daily

Method of charging

4 classes

2 classes Yearly Monthly Weekly Daily

2 classes

3 classes

–

2 classes

3.5 tonnes

Motorways

2018

Estonia

Yearly Monthly Weekly Daily

Yearly Monthly Weekly Daily

d) Duration

2 classes

3 classes

3.5 tonnes

Motorways A1–A18

2005

Lithuania

2 classes

7 classes

c) Emissions

3 classes

2 classes

3 tonnes

Main roads

2014

Latvia

Yearly Monthly Weekly Daily

2 classes for vehicles above 12 tonnes

2 classes

b) Axles

Weight

12 tonnes

All roads

2014

UK

No

4 classes

a) Weight

Differentiation by

Threshold of vehicle 12 tonnes weight

3.5 tonnes

2005

Motorways

Types of road

Netherlands Denmark Romania Luxemburg Sweden (EuroVignette)

Implementation date 1995

Country

Table 21.2  Time-based charging schemes for HGV in Europe (as of 2021)

407

3 classes

5 classes

Day/Night

6 routes with special charges

c) Emissions

e) Time

f) Spatial

–

b) Axles

a) Weight

Differentiation by

Weight threshold 3.5 tonnes

Motorways Schnellstraßen

Types of road

Bulgaria

Czech Republic Infrastructure Air pollution1) Noise1)

2010

–

5 classes

2 classes

2 classes

3.5 tonnes

no

Friday, 15.00– 20.00, all other days and times

4 classes

3 classes

–

3.5 tonnes

Motorway class Motorways I roads

Infrastructure

2020

3 regions (Wallonia, No Flanders, Brussels)3)

–

5 classes2)

2 classes

3 classes

3.5 tonnes

Motorways, some other roads

Infrastructure

Infrastructure Air pollution Noise

Cost categories included

Belgium

2016

Austria

2004

Country

Since

Table 21.3  Distance-based charging schemes for HGV in Europe (as of 2021) Germany

No

–

6 classes

2 classes for vehicles with 18 tonnes and more

3 classes

7.5 tonnes

Motorways Federal roads

Infrastructure Air pollution Noise

2005

Slovakia

Switzerland Infrastructure Air pollution Noise

2001

No

–

No

–

(Continued)

3 classes

No 4 classes for vehicles with 12 tonnes and more 3 classes

Per tonne of vehicle

3.5 tonnes 4 classes

3.5 tonnes

All roads Motorways, Express roads, Parallel Class I roads

Infrastructure

2010

408

Motorways

Go box, microwave system DSRC

Use of revenues

Method of charging

OBU or prepaid route pass

No information available

No information available

Bulgaria

Germany

OBU Pre-pay Post pay

Transport infrastructure

Transport infrastructure

Discounts for mileage driven in excess of the predefined limit4)

Slovakia

OBU Dual system: OBU (HU-GO) and manual login

Road

HGV Discounts (up to 13%) for vehicles innovation paying 2800 Euro programme and more

Czech Republic

OBU identification card for non-equipped vehicles

2/3 rail, 1/3 cantonal budgets

Increase in permitted gross vehicle weight for trucks on Swiss roads

Switzerland

Notes: 1) Since 2021. 2) In Wallonia four classes. 3) Within Brussels, motorways and inner-city roads are charged. 4) 3% to 11% for goods vehicles up to 12 tonnes, 3% to 9% for goods vehicles with 12 tonnes and more.

OBU

Road infrastructure and general budget

Reduction of No information vehicle tax to available the level of 2000

Accompanying measures

Belgium

Austria

Country

Table 21.3  (Continued)

Transport pricing in Europe  409

distance-based schemes. This change in vehicle fleet is more pronounced in countries with distance-based charges that have more charge categories to reflect emission standards (see Vierth et al., 2017). Furthermore, the price increase due to HGV charging has led to higher efficiency in road haulage. In Austria, Germany, and Switzerland empty runs have declined during the first years after introducing HGV charges until a saturation level was reached (see BAG, 2019; ÖIR, 2007; ARE, 2007). The impacts on modal shift are less clear: In both Austria and Germany an increase in rail freight was observed, however, at least in Germany the shares of both road and rail have increased at the expense of the share of inland waterway transport and the positive trend of rail freight can thus not directly be attributed to HGV tolling. In Switzerland where the introduction of HGV charges (LSVA) was combined with using revenues for rail investment, the number of HGVs crossing the Alps has decreased from 2000 to 2016 by around 40%.6 Rail freight tonnes have increased by 39% leading to a slight increase of rail’s share in total tonnes from 70% in 2000 over 63% in 2010 to 71% in 2016 (see UVEK, 2017). 21.5.3 Urban Road Pricing Schemes The European Commission has encouraged the development of urban road pricing schemes but has generally seen these as the responsibility of member states rather than legislating on them. Norwegian cities, such as Bergen, Trondheim, and Oslo, have a long tradition of charging vehicles entering the city, originally aimed at generating sufficient revenue to finance roads. Growing problems with urban road congestion and worsened air quality have motivated various European cities to introduce some form of congestion charges, starting with London (2003), Stockholm (2006), and Milan (2008); Se-il Mun and Daisuke Fukuda cover the theory of urban road congestion in Chapter 8 of this Handbook. These three cities are the largest ones with congestion charges, other examples include Gothenburg, Durham, and Valetta. All of them have installed cordon pricing schemes whereby London also charges intra-cordon trips. London and Milan raise a daily entrance charge allowing for unlimited entry, exit, and travel during the charging period. The remaining schemes are entry-based, partly with a charge cap (in Stockholm a daily cap, in Norwegian cities with a monthly charge cap). Meanwhile, almost all schemes are characterised by time-of-day differentiated charges. This applies also to the Norwegian schemes, which have moved from pure financing schemes to some type of congestion charging schemes. An interesting case is Milan where a pollution charge (labelled EcoPass) was introduced in 2008 and transformed into a congestion charging scheme in 2012, together with a ban of the most polluting vehicles (Euro 0–3 diesel vehicles) from entering the charging area. In most cities, introducing congestion charges was accompanied by improvements in public transport such as increased bus frequencies and, for example, in Milan, longterm improvements such as the extension of the underground system. Apart from the aforementioned cities, there are further urban areas with plans to introduce congestion charges. One example is Brussels, where the Brussels regional government plans to introduce a road charge for private cars (called SmartMove) from 2022 onwards. The toll will be charged distance-based within Brussels’ low-emission zone during peak times with lower charges off-peak and its level also depends on the engine type and size. The Brussels road charge is planned to form part of tax reforms so that Brussels will abolish the annual car tax to compensate for the new charge.7​

410

2013

1990

Gothenburg

Oslo

1986

2006 2007

Stockholm

Bergen

2003

London

Since

18

64

City centre

30

21

Area (sqkm)

Financing

Financing

Reduce congestion

Reduce congestion

Reduce congestion

Goal

Mon-Fri: 6:00 to 18:00 Rush hour: 6:30–9:00, 14:30–16:30

Mon-Fri: 6:00 to 18:00 Rush hour: 6:30–9:00 15:00–17.00

6:00 to 18:29

6:00 to 18:29

7:00 to 22:00

Tolling time

Table 21.4  Urban Road Pricing schemes in European cities

Cordon Per entry One passing/hour Monthly ceiling rate Time of day

Cordon (3 toll rings) Per entry Monthly ceiling rate Time of day

Cordon Per entry Time of day

Cordon Per entry (with daily cap) Time of day

Cordon Per day Intra-cordon trips

Charging type and differentiation

Vehicles < 3.5t:c) NOK 25-30 outside rush hour, NOK 51-56 in rush hour Vehicles > 3.5t:b) NOK 37-70 outside rush hour, NOK 75-123 in rush hour

Vehicles < 3.5t:a) Inner ring: NOK 4-17 outside rush hour, NOK 8-21 in rush hour Outer ring/City boundary: NOK 5-21 outside rush hour, NOK 10-29 in rush hour Vehicles > 3.5t:b) NOK 53-86 no rush hour NOK 61-101 rush hour

6:00–6:29, 8:30–14:59, 18:00–18:29: SEK 9 All other times between SEK 16 and SEK 22

SEK 20 SEK 60 at max.

£15

Charge level

411

2008 2012

Milan

8

50

Reduce air pollution Reduce congestion

Financing

7:30 to 19:30 7:30 to 18:00

Mon-Fri: 6:00 to 18:00 Rush hour: 6:30–9:00 15:00–17:00

Cordon Per day No charge for electric and hybrid vehicles, mopeds, motorcycles

Cordon Per entry One passing/hour Monthly ceiling rate Time of day Pollution charge (until 2012): 2/5/10 € Congestion charge (since 2012): 3 to 6€

Vehicles < 3.5td): NOK 11-15 outside rush hour, NOK 14-30 in rush hour Vehicles > 3.5t: NOK 24-36 outside rush hour, NOK 33-72 in rush hour

Notes: a) Charges differ by engine type (petrol/PHEV, Diesel, Electric), no charges for hydrogen cars. Discount of 20% for vehicles < 3.5 tonnes with AutoPass. b) Charges differ by emission class (Euro V and older, Euro 6), with no charges for zero-emission vehicles. c) Charges differ by engine type (petrol/PHEV, Diesel, Electric; toll for zero-emission and fuel cell vehicles vary between NOK 5 and NOK 10. Discount of 20% for vehicles < 3.5 tonnes with AutoPass. d) Motorcycles exempted. Discount of 20% for vehicles < 3.5 tonnes with AutoPass.

1991

Trondheim

412  Handbook on transport pricing and financing

In general, the envisaged impacts of urban road pricing schemes are the reduction of car trips into and/or within the charged city area together with an extended use of public transport, walking, and biking, and through these changes a reduction of congestion and air pollution. There is evidence that congestion charges have indeed been successful in reducing the number of trips. In Stockholm, from 2006 until 2011 traffic was reduced at a minimum by 18% (2009) and at a maximum by 21% (2006, see Börjesson et al., 2012), in Milan trip reduction during the EcoPass scheme varied between 11% (in 2011) and 21% (in 2008) and within the congestion charging scheme between 38% (2012) and 37% (2014, see Croci, 2016). In London, the congestion charge reduced traffic by 14% in 2003 and 21% in 2008. There is some evidence on the type of responses to charging such as choice of departure time (peak spreading, see, for example, Karlström and Franklin, 2009 for the Stockholm congestion charge and Gibson and Carnovale, 2015 for Milan), route choice (see Gibson and Carnovale, 2015 for the Milan case), and modal shift as observed in London, Stockholm, and Milan (see Croci, 2016). In general, it has to be considered that the effect of congestion charging is difficult to disentangle from other effects such as fuel price increases. Ait Bihi Ouali et al. (2021) is one of the rare studies that estimate traffic response to congestion charging by including also location characteristics of the congestion charging area (in their study London), such as fuel retail price, borough population, and employment rates. With a difference-indifferences (DiD) approach that accounts for the above characteristics, they show that the implementation of London’s Western Expansion Zone led to a 4.9% decline in road traffic, and that this decline was persistent until 2012 (the end date of their study) even after removal of the scheme in 2010. An important policy issue is the magnitude of the price elasticity and the potential difference between the short-term and the long-term elasticity, e.g. the question of whether responses are stable over time. Apparently, the short-term elasticities of European urban congestion charging schemes are in a range of around –0.45 for Norwegian urban motorways (Odeck and Bråthen, 2008), –0.47 for London (Transport for London, 2008), and between –0.46 and –0.66 for Milan (elasticity calculated for different classes of vehicles, see Croci, 2016). Börjesson et al. (2012) estimate a short-term elasticity of –0.7 with an increase to –0.85 in subsequent years which appears to be higher than in the other cases. Börjesson and Kristoffersson (2018) confirm the increase of price elasticity over time in Stockholm, but interestingly the same study finds a decreasing price elasticity over time for the Gothenburg charging scheme (2018). Finally, the urban road pricing schemes reviewed here are mostly motivated by congestion and air pollution problems, however, their levels, the spatial design of cordons, and the consideration of distortions in other markets (in particular housing and labour markets) are hardly derived from the available broad theory base briefly exemplified at the beginning of this chapter. In contrast to road freight, there is no legal obligation to charge average costs. As Sumalee et al. (2005) stated, the driving forces in designing cordon pricing schemes have been issues of practical (including technical) feasibility and judgement of the involved professionals than principles of efficient design. Furthermore, acceptability problems and compromises to achieve acceptability are crucial in particular in an urban context (see Jaensirisak et  al., 2005; Raux and Souche, 2004; Eliasson, 2014). This also adds the issue of complexity, especially in an urban context. This plays not only a role in defining and differentiating efficient prices, but raises also the question of whether and to what extent car drivers are capable of processing highly differentiated charges such that the envisaged price signal leads to the expected response (see Bonsall and Lythgoe, 2009; Franke and Kaniok, 2013 and Link, 2015b).

Transport pricing in Europe  413

21.6 CONCLUSIONS European policy on transport infrastructure charging rests on the foundations of short-run marginal social cost pricing, although with a recognition that mark-ups may be needed for financial reasons. However, when it comes to application, many issues arise. Pricing according to this principle would require a high degree of differentiation by time, place, and vehicle type. In practice pricing systems are far simpler. Charging for congestion (and scarcity in the case of rail) seems particularly complex and is only exceptionally the case (for instance, in the small number of cities that have implemented urban road pricing schemes). Inclusion of environmental costs in charges also remains patchy despite the emphasis given on this in successive transport White Papers. Moreover, there is a degree of inconsistency between the emphasis on marginal cost in legislation on rail infrastructure charges and the continued use of average cost in charging for the use of roads. It is also the case that whilst the rail legislation requires compliance, member states can still choose whether to implement charges (over and above fuel tax and annual licence duty) for the use of roads. In short whilst some progress has been made towards introducing an efficient and harmonised system of charging for the use of transport infrastructure in Europe, there is still a long way to go. Opposition to measures which would raise transport prices remains strong, whilst there is still a failure to completely resolve the tension between marginal cost pricing and full cost recovery. For instance, in the latest version of the Handbook on the External Cost of Transport (European Commission, 2019), the main text only reports total and average costs, with marginal costs relegated to an appendix. The environmental crisis and adoption of the target of net-zero carbon emissions for the EU by 2050 makes progress more urgent. The European Green Deal was announced on 11 December 2019. (EC, 2020). On transport pricing, it says ‘The Commission will therefore pursue a comprehensive set of measures to deliver fair and efficient pricing across all transport modes’. Thus whilst the stricter target for carbon emissions will clearly affect the prices actually set by raising the shadow price of carbon, in terms of principles this document simply restates what has been policy for many years.

NOTES 1. 2. 3. 4.

5. 6.

7.

In Switzerland, vignettes for HGV with GVW of 3.5 tonnes and more were replaced by distancebased tolls in 2001, Austria followed in 2004. Germany and Belgium left the scheme in 2005 and 2016 respectively and introduced a distancebased HGV charge. Bulgaria has changed HGV charging from Vignettes to distance-based charges in 2020. As a review of studies on marginal infrastructure cost shows, marginal costs make up around 30% to 58% of average costs for maintenance and between 40% and 80% of renewal costs (Link, 2015a). Studies for road operation costs are rare, but a Swedish study indicates that marginal costs are negligible for this category (Haraldsson, 2006). For example, in Germany only between 76% and 94 % of air pollution costs can be charged during the charging period 2018–2020 (see Alfen et al., 2018). The goal of the Swisss Law on Freight Modal Shift (Güterverkehrsverlagerungsgesetz – GVVG) to reduce the number of HGVs crossing the Alps to one million in 2011, was achieved with a five year’s delay in 2016. The figure of 975000 HGV trips in 2016 is still too high to meet the goal of 650000 HGVs set for 2018. Opposition comes from the Flemish and Walloon regions because their inhabitants would not enjoy that benefit.

414  Handbook on transport pricing and financing

REFERENCES Ait Bihi Ouali, L., Musuuga, D., and Graham, D.J. (2021). Quantifying responses to changes in the jurisdiction of a congestion charge: A study of the London western extension. Plos One 16(7): e0253881. Alfen, Aviso, BUNG (2018). Berechnung der Wegekosten für das Bundesfernstraßennetz sowie der externen Kosten nach Maßgabe der Richtlinie 1999/62/EG für die Jahre 2018-2022. Gutachten im Auftrage des Bundesministeriums für Verkehr und digitale Infrastruktur, Weimar, Leipzig, Aachen, Münster, Köln. Ali, A. A., Warg, J., and Eliasson, J. (2020). Pricing commercial train path requests based on societal costs. Transportation Research Part A: Policy and Practice, 132, 452–464. ARE (2007). Auswirkungen der LSVA mit höherer Gewichtslimite. Schlussbericht. Available at www​. are​.admin​.ch. Accessed on March 13, 2021. BAG (2019). Marktbeobachtung Güterverkehr. Jahresbericht. https://www.balm.bund.de/SharedDocs/ Downloads/DE/Marktbeobachtung/Jahresberichte/Jahr_2019.pdf?__blob=publicationFile&v=1. Accessed on 25 February 2023. BFS (2019). Alpenquerender Güterverkehr 2019. Available at: https://www​.bfs​.admin​.ch​/ bfs​/de​/ home​/ statistiken​/mobilitaet​-verkehr​/gueterverkehr​/alpenquerend​.html. Accessed on March 13, 2021. Börjesson, M., Eliasson, J., Hugosson, M. B., and Brundell-Freij, K. (2012). The Stockholm congestion charges-5 years on. Effects, acceptability and lessons learnt. Transport Policy, 20, 1–12. Börjesson, M., and Kristoffersson, I. (2018). The Swedish congestion charges: Ten years on. Transportation Research Part A: Policy and Practice, 107, 35–51. Börjesson, M., Rushid, A. R., and Liu, C. (2021). The impact of optimal rail access charges on frequencies and fares. Economics of Transportation, 26, 100217. Bonsall, P., and Lythgoe, B. (2009) Factors affecting the amount of effort expended in responding to questions in behavioural choice experiments. Journal of Choice Modelling, 2(2), 216–236. BVU, TNS Infratest, KIT (2016). Entwicklung eines Modells zur Berechnung von modalen Verlagerungen im Güterverkehr für die Ableitung konsistenter Bewertungsansätze für die Bundesverkehrswegeplanung. Freiburg, Karlsruhe. Calthrop, E., De Borger, B., and Proost, S. (2007). Externalities and partial tax reform: does it make sense to tax road freight (but not passenger) transport? Journal of Regional Science, 47(4), 721–752. Carpintero, S. (2010). Toll roads in central and eastern Europe: Promises and performance. Transport Reviews, 30(3), 337–359. CEPA (2017a). Market-can-bear analysis: Freight services, report by CEPA for ORR, London. CEPA (2017b). Market-can-bear analysis: Passenger services, report by CEPA and Systra for ORR, revised version, London. Croci, E. (2016). Urban road pricing: A comparative study on the experiences of London, Stockholm and Milan. Transportation Research Procedia, 14, 253–262. Eliasson, J. (2014). The role of attitudes structures, direct experience and reframing for the success of congestion pricing. Transportation Research Part A, 67, 81–95. European Commission (1992). White Paper “Future of the Common Transport Policy” Brussels. European Commission (1995). Green paper: Towards fair and efficient pricing in transport policy options for internalizing the external cost of transport in the European Union, Brussels. European Commission (1998). White paper: Fair payment for infrastructure use: A phased approach to a common transport infrastructure charging framework in the European Union, Brussels. European Commission (2001). White Paper. European Transport Policy for 2010. Time to Decide, Brussels. European Commission (2011). White Paper. Roadmap to a Single European Transport Area - Towards a competitive and resource efficient transport system. Brussels. European Commission (2019). Handbook on the External Costs of Transport. Brussels. European Commission (2020). Sustainable and Smart Mobility Strategy – putting European transport on track for the future. Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions. COM/2020/789 final. European Parliament and Council of the European Union (1993). Council Directive 93/89/EEC of 25 October 1993 on the application by member states of taxes on certain vehicles used for the carriage

Transport pricing in Europe  415

of goods by road and tolls and charges for the use of certain infrastructures. Official Journal of the European Union. 12 November 1993. https://eur​-lex​.europa​.eu​/ legal​-content​/ EN​/ TXT/​?uri​= CELEX​ %3A31993L0089. Accessed on March 12, 2021. Franke, A., and Kaniok, D. (2013). Responses to differentiated road pricing schemes. Transportation Research Part A, 48(c), 25–30. Gibson, M., and Carnovale, M. (2015). The effects of road pricing on driver behavior and air pollution. Journal of Urban Economics, 89, 62–73. Gibson, S., Cooper, G., and Ball, B. (2002), Developments in transport policy. The evolution of capacity charges on the UK rail network. Journal of Transport Economics and Policy, 36 Part 2, 341–354. Haraldsson, M. (2006). Marginal Cost for Road Maintenance and Operation – A Cost Function Approach. Annex 1.2BI to Deliverable D3, Marginal cost case studies. GRACE (Generalisation of Research on Accounts and Cost Estimation). EU-Project funded by Sixth Framework Programme. ITS, University of Leeds, Leeds. ITF (2008). Charges for the Use of Rail Infrastructure. OECD, Paris. Jaensirisak, S., Wardman, M., and May, A.D. (2005). Explaining variations in public acceptability of road pricing schemes. Journal of Transport Economics and Policy, 39(2), 127–153. Johnson, D.H., and Nash, C.A. (2008), Charging for Scarce Capacity: A Case Study of Britain's East Coast Main Line, Review of Network Economics, Vol. 7. Karlström, A. and Franklin, J. P. (2009). Behavioral adjustments and equity effects of congestion pricing: Analysis of morning commutes during the Stockholm Trial. Transportation Research Part A: Policy and Practice, 43(3), 283–296. KCW, StatisticEye and HTC (2018). Gutachten zur Bestimmung der Elastizität der Nachfrage der Eisenbahnverkehrsunternehmen. On behalf of the German rail Regulator (BNetzA), Bonn. Link, H. (2015a). Road and rail infrastructure costs. In Nash, C. (Ed.), Handbook of Research Methods and Applications in Transport Economics and Policy. Edward Elgar Publishing. Link, H. (2015b). Is car drivers’ response to congestion charging schemes based on the correct perception of price signals? Transportation Research Part A: Policy and Practice, 71, 96–109. Link, H. (2021). On the difficulties to calculate infrastructure charges for heavy goods vehicles: A review of 15 years’ experience in Germany. Journal of Transport Economics and Policy, 55(2), 141–162. Matthews, Bryan and Chris Nash (2004). Implementing Pricing Reform in Transport – Effective Use of Research on Pricing in Europe FINAL REPORT FOR PUBLICATION. Institute for Transport Studies, University of Leeds. Nash, C. (2017). Railway finance in Europe. Review of Network Economics, 16(2), 67–88. Nash, Chris et al. (2018). Track Access Charges: Reconciling Conflicting Objectives. CERRE, Brussels. Nilsson, J. E. (2002). Towards a welfare enhancing process to manage railway infrastructure access. Transportation Research, Part A. Odeck, J., and Bråthen, S. (2008). Travel demand elasticities and users attitudes: A case study of Norwegian toll projects. Transportation Research Part A: Policy and Practice, 42(1), 77–94. Odolinski, K., Nilsson, J. E., Yarmukhamedov, S., and Haraldsson, M. (2020). The marginal cost of track renewals in the Swedish railway network: Using data to compare methods. Economics of Transportation, 22, 100170. ÖIR (2007). Lkw Road Pricing – Trends und Ausbaumöglichkeiten. Report by Österreichisches Institut für Raumplanung (ÖIR) on behalf of Kammer für Arbeiter und Angestellte Wien, Wien. Available at: https://wien​.arbeiterkammer​.at. Quinet, E. (2005). Alternative pricing doctrines. In Christopher Nash and Bryan Matthews (eds.), Measuring the Marginal Social Cost of Transport. Elsevier, Amsterdam. Raux, C., and Souche, S. (2004). The acceptability of urban road pricing: A theoretical analysis applied to experience in Lyon. Journal of Transport Economics and Policy, 38(2), 191–215. Rothengatter, W. (2003). How good is first best? Marginal cost and other pricing principles for user charging in transport. Transport Policy, 10, 121–130. PWC (2014). Evaluation and Future of Road Toll Concessions. Brussels. Available at: www​.asecap​.com. SCK (2020). Marktaufschläge gemäß §67d Abs6 EisbG für die Fahrplanperioden 2018 und 2019. Available at: www​.schienencontrol​.gv​.at​/files​/1​-Homepage​-Schienen​-Control​/1b​-Wet​tbew​erbs​regu​ lierung​/ Veroeffentlichungen​/ Bescheide​%202021​/21​- 03​-11​%20Bescheid​_Wegeentgelt​_ geschw​%C3​ %A4rzt​.pdf. Accessed on March 13, 2021.

416  Handbook on transport pricing and financing

Sumalee, A., May, T., and Shepherd, S. (2005). Comparison of judgmental and optimal road pricing cordons. Transport Policy, 12(5), 384–390. TNS Infratest, ETH, IVT (2015). Ermittlung von Bewertungsansätzen für Reisezeiten und Zuverlässigkeiten auf der Basis eines Modells für modale Verlagerungen im nicht-gewerblichen und gewerblichen Personenverkehr für die Bundesverkehrswegeplanung. UVEK (2017). Bericht über die Verkehrsverlagerung vom November 2017.Eidgenössisches Departement für Umwelt, Verkehr, Energie und Kommunikation. Available at: https://ub​.unibas​.ch​/digi​/a125​/ sachdok​/2018​/ BAU​_1​_006647814​_2017​.pdf. Accessed on March 13, 2021. Van Essen, H., et al. (2019). State of Play of Internalisation in the European Transport Sector. European Commission Brussels. Vickerman, R. (2021). European Transport Policy 60 Years on. Journal of Transport Economics and Policy, 55(2) 85–103. Vierth, I., Mandell, S., and Schleussner, H. (2017). Road freight transport policies and their impact: A comparative study of Germany and Sweden. Road Freight Transport Policies and Their Impact: A Comparative Study of Germany and Sweden, 213–234. Wheat, P.E., Smith, A.S.J., and Nash, C.A. (2009). CATRIN D8 Rail Cost Allocation for Europe. University of Leeds, Leeds.

22. Pricing urban transport in Latin America Andrés Gómez-Lobo and Tomás Serebrisky

22.1 INTRODUCTION Latin Americans are city dwellers. In 2020 it was estimated that 81.2% of the region’s inhabitants lived in cities, a share expected to increase to 87.8% by 2050.1 The region also boasts not only six “megacities” but also five of the world’s largest 30 cities.2 In this context, urban mobility is critical for access to health, educational, and economic opportunities, as well as labor productivity and general economic performance. Furthermore, rising incomes are increasing private motorization rates, creating severe congestion, safety, and environmental hazards. If these threats are not adequately addressed, they will exacerbate the already complex urban mobility challenges. In this chapter we review urban transport services and their pricing in Latin America, focusing on how successful the arrangements are at allocating resources and shaping incentives. Several features in the region impinge on urban mobility. First is the extreme income inequality seen throughout Latin America. Because transport services are a necessary good and an important household expenditure, distributive issues have become paramount, most evident in transit pricing. Fare increases have sparked widespread riots over the past decade: Brazil in 2013 and Chile in 2019. Distributive issues also touch on fuel taxes and congestion pricing now that lower and middle-income households own automobiles and motorcycles. Extreme income inequality creates tension between the allocative efficiency and distributive impacts of transport pricing. But the region is beginning to address or ameliorate this tradeoff with several novel approaches. In addition to its marked income inequality, the region has relatively weak public institutions. Inadequate regulation and market organization lead to deficient public transport, allowing the now-pervasive informal transport services to fill the void (e.g., paratransit and informal motorcycle taxis),3 further limiting the reach of regulatory frameworks and pricing structures. The recent emergence of ride-hailing applications has added even more complexity to the regulation and pricing of urban mobility services. Despite the noteworthy mobility challenges Latin America faces, the region has pioneered transit reforms with the advent of high-quality Bus Rapid Transit (BRT) systems, first in Curitiba, Brazil, and then Bogotá, acting as a policy catalyst for reform in other Latin American cities and the world at large. The implications of these services for optimal pricing are of interest to policymakers in developed and developing countries alike. Another innovation is the use of a proto-type congestion pricing scheme recently introduced in two cities in Colombia. As concerns transport infrastructure, the region has also embraced public–private partnerships (PPPs) to develop airports, port terminals, highways, and other transport infrastructure. Urban tollways are also relevant to the present discussion. Urban mobility involves complex externalities and general-equilibrium effects that reach into many other markets and policies. Here we focus on pricing without considering the longrun effects of urban design and land-use planning. We do not address economic development, 417

418  Handbook on transport pricing and financing

labor market impacts, or land values.4 While drawing mostly on empirical studies from the academic literature, we make references to some conceptual issues but leave aside the growing literature on the estimation of transport externalities and the design of transit reforms. We will begin by addressing allocative efficiency. The focus is on first- and second-best fare determination, fuel taxes, and ride-hailing services. The discussion moves on to nonpricing regulations for congestion and environmental externalities that generate shadow pricing effects or modify the generalized travel cost of different transport modes. Next, we look at the social and distributive issues, including novel approaches in the region to reduce the tradeoffs between allocative efficiency and distributive impacts of transport pricing. We conclude the chapter with a summary of results and an agenda for further research.

22.2 ALLOCATIVE EFFICIENCY Are urban transport services priced efficiently? Answering this question is difficult. In addition to operational costs, one needs to consider the numerous congestion, environmental, and road safety externalities. The Mohring effect (Mohring, 1972) implies socially optimal fares below marginal operating costs (see also Chapter 9 in this Handbook on public transport pricing). There are additional issues of optimal fare structures by time of day, trip distance, and other determinants of operational costs. Therefore, calculating socially optimal fares (i.e., first- and second-best) for a particular city is not a trivial matter.5 In the absence of a road-pricing mechanism―a norm the world over as well as in Latin America―many externalities caused by private motorized vehicles are not reflected in the private cost of using this mode. We are therefore in a second-best world where available pricing instruments should consider the effects of different price levels on the demand for services in related transport markets (either through substitution or complementary effects) where externalities are significant and unchecked. In Latin America, emphasis should also be placed on the interrelation of pricing instruments with walking and cycling behavior as these are two of the main transport modes, particularly for lower-income individuals. Second-best pricing of urban transport services will need to consider two issues. The first relates to travel by private automobile or motorcycle. Fuel taxes are the main policy instruments to modify user costs for these modes. The region’s growing rates of motorization make taxes an urgent issue to evaluate. The second relates to transit fares—how are these fares set? Are their structures efficient? Are their levels reasonable given the diverse externalities present in urban markets? These two issues, transit pricing and taxes, will be treated in turn, beginning with transit services, both in liberalized and regulated markets. We then turn to private travel costs and fuel taxes, which have generated the most academic research in the region. We conclude by touching on ride-hailing services and what they imply for transit pricing. Nonprice regulations to address congestion and pollution as well as subsidy and equity issues of transit pricing will be discussed in subsequent sections of this chapter. 22.2.1 Transit Fares As in most parts of the world, transit subsidies for formal services are ubiquitous in Latin America. Rivas et  al. (2020) estimated that these subsidies represent on average 0.23% of

Pricing urban transport in Latin America  419

GDP in the region, with wide variations by country and city. At the high end of this range were Chile (0.4% of GDP) and Argentina (0.8%). The percentage of operating costs covered by these subsidies in 2019 ranged from 26% in Bogotá to 69% in Buenos Aires (Rivas et al., 2020). Formal transit systems in Mexico City, Buenos Aires, Panama City, Santiago, Bogotá, São Paulo, Montevideo, and many BRT systems in medium-sized cities of Colombia are subsidized (Basso and Silva, 2014; Gómez-Lobo, 2020; Rivas et al., 2020). In some cases, subsidies are not transparent. In Quito, for example, fares have not been raised in the last 20 years, despite rising inflation (Gómez-Lobo and Barrientos, 2018). The revenue shortfall is made up through fuel subsidies and other, often opaque, transfer mechanisms. Studies are scarce on optimal transit fares and subsidies. Parry and Timilisina (2010) use a simulation model to estimate optimal fuel taxes, formal transit fares, and toll levels for private transport in Mexico City. They found that transit fares (bus and rail, subsidized at around 50% of operating costs) were nearly optimal. In other words, altering fares would provide few welfare gains because of the relatively low modal share of mass transit as compared with private automobiles or privately operated microbuses.6 Tolls-per-mile traveled by automobiles was found to be the most promising welfare-improving policy. Hiking gasoline taxes was another option (the optimal level was close to 16 times the existing tax), although these authors recognize the potential for evasion.7 Tirachini and Proost (2020) also studied optimal transit fare/subsidy, road pricing, and fuel taxes. But they did so in a general equilibrium setting that considered the general tax structure, informal labor markets, and income-distribution welfare weights. They applied their model to Santiago and found that, ignoring income-distribution, peak-period travel costs for private vehicles should rise while peak-period transit fares should fall. For the off-peak period, however, they found the opposite: car costs should fall while bus fares rise. Taking distributional weights into account changes the results, with car costs increasing while bus fares decline for both periods. This result indicates that distributional concerns are important for optimal pricing. Although not their main research question, Basso and Silva (2014) estimated optimal bus subsidies for Santiago (Chile) and found they were at 55% of operational costs. This result does not differ much from observed subsidy levels—40% in 2013, 44% in 2014, and rising to 51% in 2019.8 In a related vein, Avner et al. (2017) studied the effect of transit subsidies on urban sprawl in Buenos Aires, using a model simulating land rent and location. Transit subsidies covered over two-thirds of operational costs, generating a $5.4 billion annual funding requirement (or 0.7% of GDP) in 2012 for a country with chronic fiscal deficits. According to their results, eliminating these subsidies would hardly hinder urban sprawl until 2050. Gómez Gelvez (2021) recently applied the model of Parry and Small (2009) to estimate optimal transit subsidies for Bogotá. He found that for the BRT system, optimal subsidies were 73% and 72%, respectively, of operating costs in the peak and off-peak periods. In the nonBRT mixed-traffic transit, optimal subsidies reached 36% and 46% of operating costs in each period. These results imply that current fares in the BRT system should be radically reduced. Interestingly, the results suggest a 14% increase for peak-period non-BRT services9 owing to the congestion externalities generated by buses in mixed traffic. An incipient discussion in regional policy circles is whether transit should be free. To this end, Bull et al. (2020) undertook a randomized controlled trial in Santiago to study the shortrun impact of free travel cards. They found that total trips increase, but only in off-peak

420  Handbook on transport pricing and financing

periods; automobile trips are unaffected. This last finding calls into question transit subsidies as a second-best instrument to correct for externalities generated by private transport, although these short-run effects do not account for longer-term shifts in vehicle ownership.10 In all, the findings of Bull et  al. (2020) would suggest that contrary to results for London (Parry and Small, 2009; Basso and Silva, 2014), free transit fares would not be optimal for Santiago, Chile. Although the evidence is patchy, it appears that, given resource allocation criteria, subsidies in the formal transit systems are either too low or not too far from optimal. Except for one case (services in peak-period mixed traffic in Bogotá), there is no evidence to support higher fares to fund these systems. An interesting question is whether informal and semi-formal services that abound in Latin American cities, and that must recover all costs through fare income, are correctly priced. On the one hand, informal services benefit from several veiled subsidies, especially for the thriving motorcycle taxi sector. For instance, the labor costs of informal operators are largely subsidized through health assistance programs and other government subsidies. Furthermore, they do not meet requirements posed by insurance and licensing regulations, nor do they pay all taxes as compared to formal services. On the other hand, cost recovery may be inefficient owing to the Mohring effect and unpriced externalities. In addition, unregulated competitive transit services do not necessarily produce competitive fares. In liberalized markets transit competes more on frequency than price (GómezLobo, 2007).11 However, profit-maximizing informal operators may provide optimal or above-optimal frequency.12 The pricing performance of informal or paratransit services in Latin America is still an open question. Meanwhile, even less is known about the efficiency of transit fare structures. Distancebased fares in the major cities of the region are rare. In part this is due to spatial segregation. Poorer households reside in the outskirts, making distance-related fares regressive. For services linking rural and semi-rural areas with city centers, one can often find distance-based fares. However, we are unaware of studies analyzing the efficiency or equity of the resulting fare structure. Peak-load pricing for the region’s urban transit services is also uncommon. Except for the Santiago Metro, we are not aware of other peak-load pricing schemes.13 22.2.2 Mechanisms to Set and Readjust Transit Fares A common political problem in Latin America is the readjustment of transit fares for input price inflation, particularly fuel prices. In the case of Quito, mentioned above, fares have not been readjusted in 20 years. Many formal contracts have input price indexing formulas to maintain the real value of payment to operators (technical fare), but the decision to raise fares paid by users (public fare) is a political one.14 Given the political difficulty in adjusting fares, an interesting institutional mechanism was introduced in the Chilean national transit subsidy law approved in 2009.15 Together with a large subsidy for the Santiago public transport system, a special three-member panel was created. Its mandate was to adjust fares to reflect input cost inflation and to guarantee the system’s financial sustainability. To this end, the panel raises fares if input cost inflation is above a threshold level each quarter or if projected fare revenue plus subsidies approved in the national budget are below projected costs for a given year. The members of this panel are chosen by

Pricing urban transport in Latin America  421

the minister of transport and telecommunications working from three shortlists, two created by the public service agency and one by the deans of the Faculties of Engineering, Business and Economics of national universities. Once chosen, the panel members serve for six years, and their fare decisions are binding for the political authority. Although common in Chile for other regulated sectors of the economy (such as electricity and water provision), the use of a technical panel to set transit fares, in lieu of a political authority, is used in no other country in Latin America.16 22.2.3 Fuel Taxes and Prices With rising motorization rates, traffic accidents along with other environmental and congestion externalities caused by private mobility are also increasing in Latin America.17 Of particular importance is the increasing use of motorcycles and the safety issues it creates.18 One (second-best) instrument to curb these externalities is fuel taxes. As shown in Figure 22.1, gasoline prices in the region are relatively low, compared with prices in developed

Note:   Dark bars identify LAC countries. Light bars correspond to non-LAC OECD countries. Source:   Statista​.co​m. Last accessed December 12, 2020.

Figure 22.1  Gasoline prices for selected countries, May–June 2020 ($/liter)

422  Handbook on transport pricing and financing

countries, apart from the United States and Canada. This is particularly so for oil-producing nations that subsidize fuels, such as Venezuela, Mexico, Ecuador, Argentina, and, until 2010, Colombia.19 In some cases, these subsidies are a direct transfer of resources allocated from the national budget. In other cases, they are veiled subsidies through price stabilization funds, which inject resources to keep prices low when international fuel prices rise. Given this situation, several studies have estimated the optimal fuel tax in Latin American countries. Parry and Timilisina (2010) found that the optimal gasoline tax for Mexico City was 16 times higher than the level prevailing at the time of their research. Antón and Hernandez (2014) also estimated the optimal gasoline tax for Mexico, finding that for 2011 it was $1.90 per gallon, below the $2.72 per gallon benchmark estimated by Parry and Timilisina (2010) but still implying a much higher gasoline price for consumers than the prevailing level at that date. Anton and Hernandez (2019) estimated the optimal gasoline tax for Guatemala, finding it would be optimal to increase this tax by 71% (40% due to congestion, and the rest due to accidents and pollution). Parry and Strand (2012) estimated optimal second-best fuel taxes to offset congestion along with environmental and accident externalities by light vehicles and trucks in Chile. They found that the optimal gasoline tax is 60% higher than the prevailing rate in 2006. For diesel fuel, the optimal tax is somewhat lower than the optimal gasoline tax ($2.09 versus $2.35 per gallon in 2006), but since the prevailing tax for this fuel is only 25% of the tax for gasoline, this tax should be increased substantially more than in the case of gasoline.20 Rizzi and de la Maza (2017) provide an alternative estimate along the lines of Parry and Strand, but for the Metropolitan Region of Santiago. Therefore, in general, research tends to find that optimal second-best fuel taxes are much higher than current levels in the region. This is consistent with the relatively low gasoline prices reported in Figure 22.1 for most countries compared to OECD averages. One difference between developed and developing countries is that the incidence of fuel taxes may be progressive in the region. Moderate progressivity has been found in Chile (Agostini and Jimenez, 2015), for Mexico (Antón and Hernandez, 2014), and Costa Rica (Blackman, Osakwe, and Alpizar, 2009). In the case of Costa Rica, the direct effects of fuel taxes were found to be progressive, while the indirect effects (through bus fares) were regressive, although in both cases impacts are relatively small. Feng et al. (2018) found that fuel price increases are regressive in Argentina, Bahamas, Barbados, and Jamaica; progressive in the case of Costa Rica, Ecuador, Nicaragua, and Paraguay; and relatively neutral in Chile, Guatemala, and Uruguay. They also found that indirect effects―impacts on the price of other goods and services in the economy―are regressive and tend to outweigh the progressive direct effect of fuel taxes. Fuel taxes in the region have been used mostly as an instrument to generate revenue for governments not to allocate resources. The rise of electromobility may erode this tax base in the future, forcing governments to rethink taxes in the transport sector. Some countries are discussing distance-based user charges in response to this erosion of the excise tax base for conventional fuels (e.g., for electric vehicles in Australia). In New Zealand, a distance-based user charge already applies to diesel vehicles and vehicles powered by fuels not taxed at source.21 This will be an emerging issue as electric vehicles increase their share in the total vehicle stock. To date, there has been scant discussion in the region of the tax revenue implications of this trend and of alternative transport tax options.

Pricing urban transport in Latin America  423

22.2.4 New Challenges The emergence of ride-hailing services or NTCs (network transport companies) has added an additional element of complexity to the optimal pricing of urban mobility. Based on a survey in Santiago, Chile, Tirachini and Gómez-Lobo (2018) show that it is very unlikely for these services to reduce congestion.22 The fundamental reason is that close to 40% of ride-hailing trips would have been taken by transit, cycling, walking or not taken at all in the absence of these services, thus they transfer trips from sustainable and efficient modes to less efficient modes. The congestion effect of ride-hailing services may justify a Pigouvian tax to internalize this externality. Still, applying a tax to just one of several modes causing congestion―and one that represents only a small share of overall trips—is debatable (OECD/ITF, 2016). In any case, taxing ride-hailing services would be a second-best policy. Several cities in the region have introduced special charges for ride-hailing services, although how they compare with optimal second-best charges is still an open question. Examples include São Paulo, Mexico City (1.5% of the fare for each trip), and Porto Alegre (a charge of R$73–$15 at the end of 2020 per vehicle per month). The scheme in São Paulo, introduced in 2016, is interesting given that there is a base charge (R$0.10–$0.02 per kilometer), but the regulator can increase this charge if infrastructure use is high, and decrease it outside the city center, in off-peak periods, if the vehicle is environmentally friendly and if the driver is a woman (Alonso Ferreira et al., 2018).

22.3 NONPRICE REGULATIONS TO TACKLE CONGESTION AND POLLUTION Nonprice regulations to tackle congestion and pollution have often been used in Latin America. These regulations create a shadow price effect for some transportation modes; under some circumstances, they can be efficient. One such policy is segregated bus lanes, with BRT systems being a prime example. Latin America has been a pioneer in BRT after its introduction by Curitiba in the 1970s, and Bogotá’s influential TransMilenio reform in 2000. By devoting scarce road space to buses, commuters and other passengers gain speed and reliability, while private transport is “taxed” with less access to road infrastructure.23 Basso and Silva (2014) showed that the welfare consequences of reassigned road space is almost as significant as introducing an optimal road tax on automobiles. Lanes devoted solely to highoccupancy buses are a close policy substitute for a first-best pricing alternative. While a road tax changes incentives and behavior through the monetary costs of using different modes, a segregated bus infrastructure policy changes behavior through the users’ time cost component of the generalized cost of travel. In 2020, 56 cities in Latin America had either a BRT or some sort of dedicated bus infrastructure, reaching 1,840 km in total.24 These transit systems moved more than 20.8 million passengers per day, or 61.4% of the total world BRT patronage. Despite their prevalence, not all BRT systems have succeeded, at least from a mobility perspective (Gómez-Lobo, 2020). Some reforms have implicitly funded the higher costs of these formal high-quality systems by reducing fleet sizes, network coverage and frequency, negatively affecting users’ travel experience and time costs, and ultimately resulting in reduced patronage.

424  Handbook on transport pricing and financing

BRT and similar investments are usually funded through government grants, bestowing an implicit subsidy on transit users. But car infrastructure involves many more resources and, from this perspective, is another type of subsidy.25 Another widely used measure to curb congestion and other externalities is the use of driving restrictions. This type of regulation limits the number of cars on roads on certain days of the week—usually with reference to the last digit of the license plate number—and are seen in Santiago, Mexico City, São Paulo, Quito, La Paz, and most Colombian cities. These restrictions introduce a shadow price for drivers, but the policy may also be counterproductive. Users now have an incentive to purchase an additional vehicle—often older and more polluting (Eskeland and Feyzioglu, 1997; Davis, 2008; Gallego, and Salas, 2013a, 2013b; and Cantillo and Ortuzar, 2014). But short-run effects may be positive (Rivera, 2019). A more useful policy variation would be to exempt newer, cleaner vehicles from restrictions. Montero, Gallego, and Barahona (2020) showed that these vintage-conditioned car restrictions can speed fleet renewal, lower emissions, and avoid the problem posed by a second-car purchase. Driving restrictions are politically popular because they are easily understood by citizens and are also perceived to be fair. However, they are counterproductive regarding congestion and pollution, as we saw in the discussion just above. In 2016, Chile saw some interesting legislation that would exempt a vehicle from restricted days by combining a license plate-based restriction with a paid daily pass. Akin to a “taking turns” congestion charge (proposed by Daganzo, 2000), this scheme charges drivers only the days their vehicle is restricted.26 This option could solve the principal political-economy problem regarding conventional congestion charging—not compensating drivers for their diminished “consumer surplus.” In a conventional scheme, drivers are charged every day but do not always receive a benefit from revenues recycled, for example through higher public transport subsidies. With a “taking turns” policy for pricing access to roads, drivers benefit directly from less congestion and higher speeds on the very days they are not restricted and do not have to pay to use the roads. In addition, as Daganzo (2000) has noted, this scheme may have more favorable distributive properties. The effects of such a scheme were simulated for Santiago by Basso et  al. (2020). Unfortunately, the authors show that it is not Pareto improving. That is, the time saved on less-congested no-restriction days does not compensate lower-income individuals for the toll exacted on restricted days. But if revenues are used to subsidize public transport fares or to improve transit quality, then all income groups gain. In 2017, the city of Cali introduced an exemption charge for its car-restriction policy. Although designed to fund the BRT’s deficit, the program has similarities to Daganzo’s (2000) proposal analyzed by Basso et  al. (2020). Because the license plate restriction was already in place when Cali introduced the exemption, the effects on congestion are unclear. Nevertheless, continuing discussions in Cali about adding even more daily digits to the license plate restrictions may produce the required bite that would make this policy a “taking turns” congestion charge. A similar scheme was introduced in Bogotá in September 2020. It will be interesting to track the impact of these programs as the number of restricted digits rises after an exemption charge was introduced. The exemption charge in Cali and Bogotá is on the one hand a novel approach. On the other hand, to date, the approach has a key weakness: they are not marginal user prices. Payments are monthly, biannual, or annual. Once the user pays the charge, it becomes a sunk fixed cost for the duration of the payment period. As such it may promote more intense vehicle use and discourage modal shifts.

Pricing urban transport in Latin America  425

In 2018, Colombia implemented yet another noteworthy nonpricing regulation to promote sustainable transport, giving public servants a half-day off work for every 30 days they cycle to work (Tirachini, 2019). We are unaware of an evaluation of the effectiveness of this scheme. Finally, some cities—among them Santiago and Cartagena—have built urban toll roads, mainly for intracity travel. In Santiago, tolls rise as congestion hits portions of the roadways (measured by vehicle speeds in the previous semester)—a limited scheme for congestion charging. As with Colombia’s 2018 experiment with biking public servants, the effects of these urban tolls on congestion and emissions have not been evaluated.

22.4 SOCIAL OR DISTRIBUTIVE ISSUES The share of household consumption spent on transport services depends on prices and intensity of use. Household expenditures on transportation include private transportation (purchase of vehicles, fuels and lubricants, and maintenance) as well as on public transportation fares. A regional comparison (Rivas et al., 2018) shows that the share of total consumption expenditure allocated to transportation by LAC households is the highest among all developing regions. (LAC households allocate 11% of their consumption budget to transport services, while the shares stand at 10% in Eastern Europe, 9% in Sub-Saharan Africa, and 5% in South Asia.) Transport expenditures in the region are very heterogeneous (Figure 22.2), with household survey data (Gandelman et al., 2019) showing expenditures ranging from 12.8% in Brazil to 4.6% in Nicaragua. When expenditure is analyzed by income level―key to assessing affordability and distributional consequences of transport pricing―the data show that the share of transport in total expenditure of high-income households more than duplicates the share of poor households (Figure 22.3). As income grows, the driver of higher shares is private transportation. Gandelman et al. (2019) found that public transportation is a necessity for the average household, with an expenditure elasticity of 0.5. Private transportation, on the other hand, is a luxury, with an expenditure elasticity of 2.6. Public transportation is a necessity for the bottom four expenditure quintiles and an inferior good (when a household’s expenditure budget increases, absolute spending on public transportation falls) for the top quintile. Private transportation is a luxury for the bottom four expenditure quintiles and a necessity for the top quintile. Toro-Gonzalez et al. (2020) find that public transport is an inferior good in Colombia. Affordability for low-income groups may not, however, be revealed through measuring and reporting household expenditures, which reflect actual household consumption. For poor households, these expenditures could be lower than what they would demand where disposable income not a binding constraint, as transport competes with other needs. In simple terms, expenditure reported by household surveys does not account for foregone consumption or trips suppressed for lack of money. The regional evidence indicates that foregone consumption is severe, as walking represents about 40 to 45% of all trips made by poor individuals, a percentage that falls to 10 to 20% for higher-income groups (Cavallo et al., 2020). Rivas et  al. (2018) calculated an indicator that overcomes the limitations of household expenditure data to reveal the extent of transport affordability. This indicator reports for several cities the share of income (using average income in the population and the average income of the bottom-income quintile) that would need to be allocated for baskets of 60 and 45 public

426  Handbook on transport pricing and financing 14.0% 12.0%

12.6%

11.6%

10.8%

10.0%

9.8%

9.7%

9.6%

9.0%

8.8%

8.3%

8.0%

7.8%

7.4% 6.0%

6.0%

4.6%

4.0% 2.0%

ra

gu a

he ,T as

am Ba h

Ni ca

ur

as

r

nd

do

Ho

ica aR

ua Ec

y

st Co

ra

gu a

y Pa

ug ua

Ur

ex ico M

ru Pe

ia liv Bo

ile Ch

a m na

Pa

Br a

zil

0.0%

Note:   Sample includes countries for which household surveys include a comparable transport module. Year of information varies between 2010 and 2014 according to household survey availability. Source:   Authors’ elaboration based on Gandelman et al. (2019).

Figure 22.2  Transport expenditure in LAC as share of total household expenditure

Source:   Authors’ elaboration based on Gandelman et al. (2019).

Figure 22.3  Transport expenses as a percentage of total household expenses, by expenditure quintile (in selected countries, 2014) transit trips a month (Figure 22.4). The results for the bottom-income quintile (Figure 22.4, panel b) point to an acute affordability problem.27 Davis (2020) studied notable fare hikes for public transit in Mexico City and Guadalajara and a fare holiday in Monterrey. He found implicit price elasticities that range from –0.24 to –0.33, signaling the existence of affordability problems for a necessary service for a large group of users.

Pricing urban transport in Latin America  427

Source:   Rivas et al. (2018) Published with permission from the Inter-American Development Bank.

Figure 22.4  Share of income allocated to public transit for 45 and 60 trips a month: Selected cities, 2018 We have documented above that transit subsidies are ubiquitous throughout Latin America and the Caribbean. Policymakers, likely aware of the affordability problem facing low-income groups, have tended to justify subsidies with reference to social objectives; they rarely state that their formal justification is to incentivize a modal shift from private to public transportation to reduce externalities (Serebrisky et al., 2009). Recognizing that transportation is an important service for individuals is not enough to justify transit subsidies since this argument could also justify subsidies for other goods and services (Serebrisky et al., 2009). Monetary transfers could be a better way to help the poor

428  Handbook on transport pricing and financing

than sectoral subsidies (Gwilliam, 2017). However, there may be special reasons to subsidize certain goods or services. For example, to improve access to education or health services, overcome poorly functioning welfare systems that hinder the administration of transfers, alleviate constraints on intrahousehold resource allocation so students have sufficient transit funds, or improve resource allocation in a second-best setting. Transport services fall under these special reasons. Transportation subsidies can be divided into supply-side subsidies (channeled to transportation suppliers) and demand-side subsidies (channeled directly to beneficiaries). They can be classified according to their mode of distribution or their funding mechanism (i.e., general tax, specific taxes, and cross-subsidies). Their choice and use should be guided by the desired objective. If the policy objective is to incentivize the use of public transit by changing the relative price of public versus private transportation, then a supply-side subsidy may be justified.28 An extreme form of a supply-side subsidy is free public transit. But when the policy objective is to address social concerns by improving affordability, the most effective instrument is a demand-side subsidy, which targets only those in need (minimizing errors of exclusion) and disregards the ineligible. This latter effect avoids the error of inclusion of universal supply-side subsidies, like free transport, because they benefit all riders, regardless of their income level or need. A pervasive lack of transparency obscures information about the size, time evolution, and types of transport subsidies provided in Latin American cities. Some countries, like Chile and Colombia, are exceptions—at least with respect to direct subsidies provided through various instruments. But even for these countries, information is scant about indirect subsidies for private transport (for instance, through road infrastructure or insufficient pricing of negative externalities), apart from the few studies reviewed above that sought to quantify these negative features. Supply-side subsidies appear to account for most support to public transit. They include government transfers to fund infrastructure or to service providers to cover operating costs. Conditional supply-side subsidies are linked—partially or totally—to performance indicators of transportation operators (number of passengers carried, kilometers traveled). Direct transfers to operators in Buenos Aires are an example of this type of subsidy. Unconditional supplyside subsidies are not linked to performance or the fulfillment of other system objectives. Supply-side subsidies are less targeted than demand-side subsidies because transportation operators do not identify different types of users, except for subsidies conditioned on performance targets or specific services, such as nonviable rural services.29 In remote areas of Chile, for example, the main objective of supply-side subsidies is to provide communities with better territorial, economic, and social integration (DTPR, 2018a). Supply-side subsidies also serve the purpose of providing resilience to demand shocks—for example, with COVID-19 where lockdowns produced steep drops in mobility (in some instances of over 80%) and social-distance requirements imposed low occupancy rates in the provision of transit services. Latin America has made extensive use of demand-side subsidies—most of them categorical (for example, to students, seniors, people with disabilities, and so on) or geographical (free trips for poor residents in some areas, or flat rates or integrated fares in cities where the lowincome passengers reside in the outskirts). These subsidies are most often funded through cross-subsidies from other users. Other schemes, although much less common, have relied on selection mechanisms outside the transport sector. An example is the Vale Transporte

Pricing urban transport in Latin America  429

program in Brazil (employers withhold up to 6% of employers’ gross salary in return for vouchers to cover commuting transit fares). Another type of subsidy in formal systems is one where operators are paid from a central fund, where fare revenues accrue, according to conditions set out in concession contracts (often tendered). If resources in the centralized fund are insufficient to pay operators, then governments often inject additional resources to keep public fares from rising. This would then amount to a demand-side subsidy that allows a change in relative prices between private and public transport but is universal in its targeting properties. This subsidy scheme is used in BRT systems throughout Colombia and in Santiago. The evidence of the targeting and distributional properties of demand-side subsidies has been, at best, mixed (Serebrisky et  al., 2009; Fay et  al., 2017). In part, this is due to lack of access by the poor to formal transport services (they tend to rely on informal services like vans and microbuses). Gómez-Lobo (2009) adds the funding dimension of subsidies, a dimension usually ignored in analysis of the distributional incidence of transport subsidies. This author showed that, at a first glance, the preferential fare policy for students in the bus system in Santiago, Chile, seemed to be progressive. But when the funding source is included (cross-subsidy from normal fares), the policy lost progressivity. Poor households are negatively affected by the surcharge on normal fares as they pay a higher proportion of the surcharge than do wealthier households. There is reason for hope in the design and implementation of demand-side subsidies. A new group of subsidies has recently emerged, supported by smart card technologies. In 2014 the local government in Bogotá introduced a pro-poor public transportation subsidy using a social policy targeting mechanism called SISBEN to classify potential beneficiaries and allocate subsidies. SISBEN uses several socioeconomic characteristics of individuals and households to assign a score between 0 and 100, which is used as a proxy for relative socioeconomic level. Households below the threshold of 30.56 are issued a smart card with discounted fares on a set number of trips per month (SITP, 2018). In 2015 the subsidy represented a 50% discount on TransMilenio trunk services and a 60% discount on zonal-component trips for 40 trips per month. The subsidy increased the number of trips beneficiaries took by 56% when compared with trips when the program was unavailable (Rodríguez Hernández and Peralta-Quiros, 2016). Guzman and Oviedo (2018) found that this subsidy had the desired targeting properties as most of the benefits accrue to poor users. Due to budgetary restrictions, in 2017 the fare discount was reduced to 25% in trunk services and to 27.5% in zonal services, and the number of trips with a discount reduced to 30 per month. The metropolitan area of Buenos Aires implemented a similar subsidy, relying on the smart card SUBE. The selection mechanism identifies beneficiaries via a mix of categorical characteristics (for instance, “single head of household”) and recipients of social programs. Eligible individuals receive a 55% fare reduction in bus and rail services. The number of beneficiaries has risen over time to stem the social effects of rising tariffs. In 2018, 2.3 million beneficiaries were in the metropolitan area of Buenos Aires (Domínguez et al., 2020). An evaluation of the targeting properties of this mean-tested subsidy is not yet available. Finally, in some countries, formal sector workers receive an income supplement for transport expenditures. For example, in Colombia, workers who earn less than two times the minimum wage receive a fixed monthly payment in addition to their wages. Although this supplement is targeted (at least as far as formal sector workers is concerned), the fact that the

430  Handbook on transport pricing and financing

additional payment can be used for any expenditure makes this scheme akin to a labor-income subsidy rather than a transport subsidy.

22.5 CONCLUSIONS Latin America and the Caribbean faces enormous urban mobility challenges in the coming decades. These challenges call for policy action that sets proper prices for the various urban transport services. Pricing will be fundamental to promote sustainable and socially desirable decisions around individual mobility. As reviewed in this chapter, new pricing approaches will probably imply higher fuel taxes. Emerging electromobility trends may, however, erode this tax base and prompt governments to introduce kilometer-based transport taxes. Such taxes would be an improvement over current fuel taxes and, further, may open the door for more innovative taxing structures able to internalize the diverse externalities present in urban mobility. Transit subsidies for formal systems, while ubiquitous in the region, may also have to be adjusted. But the available evidence suggests that current subsidy levels―at least in cities with available studies―are not too far off from their optimal levels. More research on social equity and efficiency implications of these subsidies is warranted. Given the region’s motorization rates, some type of road pricing for private vehicles is becoming an urgent need in view of urban congestion and other negative externalities. Doing this while addressing distributional concerns will require creativity, political will, and public funding. The possibility of combining license plate restriction policies, used throughout the region, with an exemption payment from these restrictions, may provide a promising way forward. Future research should analyze the effects and acceptability of these “taking turn” road-pricing mechanisms recently introduced in Cali and Bogotá. Also, the use of more targeted social benefits to make public transport affordable to poorer households could also make road-pricing schemes more acceptable from a distributional point of view. The recent use of means-tested transit subsidies through smart card technology in Bogotá and Buenos Aires offers interesting experiences in the region, together with the older Vale Transporte case in Brazil for formal sector workers. Another policy that helps to set proper prices for urban transport services is to continue the expansion of exclusive infrastructure for buses (together with other mass transit investments such as metros and light rail). The region has been a pioneer in BRT-type systems as well as simpler bus-priority infrastructure. Research has shown that this infrastructure could be a good policy substitute for road pricing. Continuing to invest in these infrastructures, which improve the users’ time component in the generalized costs of travel, implicitly generates the correct price signals for users among the various urban transport modes. Finally, several other topics merit further research. Among the most pressing are the pricing consequences of informal modes, particularly the growing motorcycle taxi services popping up in several cities. The regulation and pricing of such services will be important challenges for the region.

NOTES 1. UN Department of Economic and Social Affairs, Population Dynamics, World Urbanization Prospects: The 2018 Revision. https://population​.un​.org​/wup​/ Download/.

Pricing urban transport in Latin America  431

2. By “megacity” we refer to cities with over 10 million people: São Paulo, México City, Río de Janeiro, Buenos Aires, Bogotá, and Lima. UN Department of Economic and Social Affairs, Population Dynamics, World Urbanization Prospects: The 2018 Revision.  https://population​.un​. org​/wup​/ Download/. 3. Tun et al. (2020) estimate that more than half of households’ public transport trips in the region are undertaken in informal or semiformal services. 4. Some of these issues are dealt with in Embarq, Berg, Deichmann, Liu, and Selod (2017), Bryan, Glaeser, and Tsivanidis (2019), and Yañez-Pagan et al. (2019). 5. There are some efforts to determine and set optimal second-best transit fares in practice. One example is the approach taken by the Independent Pricing and Regulatory Tribunal (IPART) in New South Wales, Australia (see Chapter 24 in this Handbook on transport pricing and financing in Oceania). 6. As the authors recognize, however, the simulations do not consider the Mohring effect on users’ time cost. 7. In addition, the political economy of raising fuel prices is difficult, as shown by the protests sparked by the proposal to eliminate fuel subsidies in Ecuador in 2019. The idea was subsequently scrapped. 8. DTPM (2019). See also Gwilliam et al. (2015). 9. Gómez Gelvez (2021) assumes the demand increase from lowering fares in the BRT system is matched by a proportional increase in supply. This may be unrealistic for a cash-strapped system where the BRT infrastructure is used at full capacity in peak periods. Changing this assumption would reduce the optimal subsidy calculated for the BRT services, particularly in the peak period. 10. In this section, we discuss a means-tested transit subsidy used in Bogota that had an important demand effect. It is not clear, however, whether it reduced private automobile trips. 11. This generates considerable safety externalities as drivers compete for passengers on the road. In Lima, Peru, nearly 10% of respondents had been involved in an accident while using informal or semi-formal public transport during the six months prior to a survey (Gomez, 2000, cited in Gwilliam, 2001). In Santiago, on average one person died every three days in an accident involving a bus prior to the 2007 Transantiago reform (Estache and Gómez-Lobo, 2005). 12. On the theoretical possibility of above optimal frequency provision in a liberalized transit market, see Van Reeven (2008) and Gómez-Lobo (2014). This would require, however, market power by transit operators. 13. In the case of the Santiago Metro, fares are differentiated according to three periods during normal working days: peak, off-peak and super off-peak. One trip fare is CLP$ 800 ($1.05), CLP $720 ($0.94) and CLP $640 ($0.84) for each of these periods, respectively (www​.metro​.cl, fares from November 26, 2020, and exchange rate average for November 2020). 14. An analysis of bus concession contracts in two Latin American cities can be found in Gómez-Lobo and Briones (2014). 15. Law 20.378. It can be found in www​.bcn​.cl​/ leychile​/navegar​?idNorma​=1005871. 16. After the widespread riots sparked by a transit fare hike in October 2019 the panel setting transit fares in Santiago has become ineffective in practice, since a law was passed whereby the government can fund any financial deficit in the system. Fares have not changed since then. 17. According to Rivas et al. (2019) the motorization rate increased from 127 vehicles per 1,000 inhabitants to 201 in the region between 2005 and 2015, an increase of 63%. Although trips in the region are shorter than trips in developed countries, they take 13 minutes more (Cavallo et al., 2020). 18. Hidalgo and Huizenga (2013) report that in Brazil and Colombia the growth in motorcycle ownership was more than double of that of cars. The motorcycle fleet increased by 275% in Brazil between 2002 and 2013, by 329% in Argentina between 1997 and 2009, by 400% in Colombia between 1997 and 2009 and 448% in Venezuela between 2007 and 2013 (Rodriguez et al., 2015). See Hagen et al. (2016) for a qualitative analysis of motorcycle use in several cities and Rodriguez et al. (2015) for a complete analysis of the motorcycle phenomenon in five cities of the region. 19. See Mendoza (2014) and Carlino and Carlino (2015) for a discussion of fuel subsidies in the region. These last authors report that, according to the IMF, fuel subsidies represented close to 1% of regional GDP in the 2011 to 2013 period. 20. A new “green” tax, introduced in 2014 for light vehicles, would be an indirect way to correct for the price distortion seen in Chile (a distortion caused by diesel taxes that are far lower than those for

432  Handbook on transport pricing and financing

21. 22. 23. 24. 25.

26. 27.

28.

29.

gasoline). This green tax is applied at the sale of a new vehicle—a function of vehicle NOx emissions and sales price. Because diesel motors produce higher NOx emissions, this tax implicitly counters the incentive to buy a diesel vehicle due to differences in the pump price of each fuel. See Chapter 24 for more details. They may provide other benefits, however. Lagos et al. (2020) provide evidence that the irruption of these platform-based transport services reduced traffic accidents in the Santiago Metropolitan region. Interestingly, this effect is significant only for women. Durán-Hormazábal and Tirachini (2016) documented the higher reliability of bus services in dedicated bus lanes compared to mix traffic conditions for Santiago, Chile. Brtdata​.o​rg last accessed December 12, 2020. Parking on public roads is another type of subsidy for private cars because, in most cases, street parking is not priced. In several Latin American countries, motorcycles also have veiled subsidies through cross-subsidies or exceptions in insurance premiums, toll and tax exemptions, or the lack of strict emission standards. If the number of restricted license plate digits increases to ten per day, then this system converges to a classical congestion charging scheme. However, as argued by Gwilliam (2017), the interpretation may well be that the problem faced by these households is incomes that are too low rather than the price of transit fares. The policy implications differ, since under the first interpretation monetary transfers to poor households would be the policy alternative rather than an intervention in the public transport sector. Supply-side subsidies may, however, create productive inefficiencies that subtract from the effectiveness in reducing relative prices. In their review of efficiency studies, De Borger et al. (2002) employ frontier techniques to conclude that “subsidies tend to worsen the performance of urban public transport in a variety of ways: higher costs, fewer revenue-passengers, excessive wage growth, and technical inefficiency.” See Serebrisky et al. (2009).

REFERENCES Agostini, C.A., and J. Jiménez. 2015. “The Distributional Incidence of the Gasoline Tax in Chile.” Energy Policy 85 (October): 243–52. Alonso Ferreira, M., F. T. Salva Rocha, A. Giuli, and F. de Mello Franco. 2018. Politics, polity and policy of ride-sourcing regulation in São Paulo. Urban Transport in the Sharing Economy Era: Collaborative Cities. CIPPEC, Argentina. https://www​.cippec​.org​/wp​-content​/uploads​/2018​/09/ UrbanTransport​-completo​-web​_CIPPEC​​.pdf. Antón, A., and F. Hernández. 2014. “Optimal Gasoline Tax in Developing, Oil-Producing Countries: The Case of Mexico.” Energy Policy 67: 564–71. Anton, A., and F. Hernandez. 2019. “Internalizando externalidades: el impuesto a las gasolinas en Guatemala.” Revista Mexicana de Economía y Finanzas Nueva Época 14(1): 1–20. Avner, P., S.R. Mehndiratta, V. Viguié, and S. Hallegatte. 2017. “Buses, Houses, or Cash? SocioEconomic, Spatial, and Environmental Consequences of Reforming Public Transport Subsidies in Buenos Aires.” Policy Research Working Paper 8166. The World Bank, Washington, DC. Basso, L.J. and H.E. Silva. 2014. “Efficiency and Substitutability of Transit Subsidies and Other Urban Transport Policies.” American Economic Journal: Economic Policy 6(4): 1–33. Basso, L.J., J.P. Montero, and F. Sepulveda. 2020. “Taking Turns: Testing Daganzo’s Hybrid DrivingRestriction Scheme.” Mimeo. Universidad Católica de Chile and Universidad de Chile. Berg, C.N., U. Deichmann, Y. Liu, and H. Selod. 2017. “Transport Policies and Development.” Journal of Development Studies 53(4): 465–480. Blackman, A., R. Osakwe, and F. Alpízar. 2009. “Fuel Tax Incidence in Developing Countries: The Case of Costa Rica.” Policy Brief. Environment for Development Initiative. Bryan, G., E. Glaeser, and N. Tsivanidis. 2019. “Cities in the Developing World.” NBER Working Paper 26390. National Bureau of Economic Research, Cambridge, MA. http://www​.nber​.org​/papers​/ w26390.

Pricing urban transport in Latin America  433

Bull, O., J.C. Muñoz, and H.E. Silva. 2020. “The Impact of Fare-Free Public Transport on Travel Behavior: Evidence from a Randomized Controlled Trial.” Regional Science and Urban Economics https://doi​.org​/10​.1016​/j​.regsciurbeco​.2020​.103616. Cantillo, V., and J.D. Ortuzar. 2014. “Restricting the Use of Cars by License Plate Numbers: A Misguided Urban Transport Policy.” Dyna 81(188): 75–82, Universidad Nacional de Colombia, Medellín, Colombia. Carlino, H., and M. Carlino. 2015. “Fossil Fuel Subsidies in Latin America: The Challenge of a Perverse Incentives Structure.” IDDRI Working Paper 15/15 November. Fundación Torcuato Di Tella. Cavallo, E., A. Powell, y T. Serebrisky, eds. 2020. “De Estructuras a Servicios: El Camino a Una Mejor Infraestructura en América Latina y el Caribe.” Banco Interamericano de Desarrollo, Washington, DC. Daganzo, C.F. 2000. “Taking Turns: Rx for Congestion.” Access 17: 14–19. Davis, L. 2008. “The Effect of Driving Restrictions on Air Quality in Mexico City.” Journal of Political Economy 116: 38–81. Davis, L. 2020. “Estimating the Price Elasticity of Demand for Subways: Evidence from Mexico.” Energy Institute WP 307, University of California, Berkeley. De Borger, B., K. Kerstens, and A. Costa. 2002. “Public Transit Performance: What Does One Learn from Frontier Studies?” Transport Reviews 22(1): 1–38. Domínguez González, K., A. L. Machado, B. Bianchi Alves, V. Raffo, S. Guerrero, and I. Portabales. 2020. Why Does She Move? A Study of Women’s Mobility in Latin American Cities. Washington, DC: World Bank. DTPR. 2018. “Informe Subsidios al Transporte Público Remunerado en Zonas Aisladas.” Glosa 03 Ley no. 21.053 Ley de Presupuestos del Sector Público. División de Transporte Público Regional, Santiago, Chile. DTPM. 2019. Informe de Gestion 2019: Directorio de Transporte Público Metropolitano. Ministerio de Transporte y Telecomunicaciones, Chile. http://www​.dtpm​.cl​/descargas​/memoria/ InformeGestion​_ 2019​_ DTPM​​.pdf Durán-Hormazábal, E., and A. Tirachini. 2016. “Estimation of Travel Time Variability for Cars, Buses, Metro, and Door-to-Door Public Transport Trips in Santiago, Chile.” Research in Transportation Economics 59: 26–39. Eskeland, G., and T. Feyzioglu. 1997. “Rationing Can Backfire: The ‘Day Without A Car’ in Mexico City.” World Bank Economic Review 11: 383–408. Estache, A. and A. Gómez-Lobo. 2005): “The Limits to Competition in Urban Bus Services in Developing Countries.” Transport Reviews 25(2): 139–58. Fay, M., L. A. Andres, C. Fox, U. Narloch, S. Straub, and M. Slawson. 2017. “Rethinking Infrastructure in Latin America and the Caribbean: Spending Better to Achieve More.” World Bank, Washington, DC. Feng, K., K. Hubacek, Y. Liu, E. Marchán, and A. Vogt-Schilb. 2018. “Managing the Distributional Effects of Energy Taxes and Subsidy Removal in Latin America and the Caribbean.” IDB Working Paper 00947. Inter-American Development Bank, Washington, DC. Gallego, F., J.P. Montero, and C. Salas. 2013a. “The Effects of Transport Policies on Car Use: Evidence from Latin American Cities.” Journal of Public Economics 107: 47–62. Gallego, F., J.P. Montero, and C. Salas. 2013b. “The Effects of Transport Policies on Car Use.” Energy Economics 40(1): S85–S97. Gandelman, N., T. Serebrisky, and A. Suárez Alemán. 2019. “Household Spending on Transport in Latin America and the Caribbean: A Dimension of Transport Affordability in the Region.” Journal of Transport Geography 79. Gomez, L. 2000. “Gender Analysis of Two Components of World Bank Transport Projects in Lima, Peru.” Mimeo, The World Bank, Washington DC. Gómez Gelvez, J. 2021. “Subsidios al Transporte Público en América Latina desde una Perspectiva de Eficiencia.” Manuscrito, noviembre. Banco Interamericano de Desarrollo, Washington, DC. Gómez-Lobo, A. 2007. “Why Competition Does Not Work in Urban Bus Markets: Some New Wheels for Some Old Ideas.” Journal of Transport Economics and Policy 41(2): 283–308. Gómez-Lobo, A. 2009. “A New Look at the Incidence of Public Transport Subsidies: A Case Study of Santiago, Chile.” Journal of Transport Economics and Policy 43(3): 405–25.

434  Handbook on transport pricing and financing

Gómez-Lobo, A. 2014. "Monopoly, Subsidies and the Mohring Effect: A Synthesis.” Transport Reviews 34(3): 297–315. Gómez-Lobo, A. 2020. “Transit Reforms in Intermediate Cities of Colombia: An Ex-Post Evaluation.” Transport Research Part A 132: 349–364. Gómez-Lobo, A., and J. Briones. 2014. “Incentives in Bus Concession Contracts: A Review of Several Experiences in Latin America.” Transport Reviews 34(2): 246–265. Gómez-Lobo, A., and R. Barrientos. 2018. Evaluación Ex-Post de Experiencias de BRT (Bus Rapid Transit) en Ciudades Latinoamericanas, Informe Final. Septiembre. Banco Interamericano de Desarrollo, Washington, DC. Guzman, L.A., and D. Oviedo. 2018. “Accessibility, Affordability and Equity: Assessing ‘Pro-Poor’ Public Transport Subsidies in Bogotá.” Transport Policy 68: 37–51. Gwilliam, K. M. 2001. “Competition in Urban Passenger Transport in the Developing World.” Journal of Transport Economics and Policy 35(1): 99-118. Gwilliam, K. 2017. Transport Pricing and Accessibility: Moving to Access Initiative. Washington, DC: The Brookings Institution. Gwilliam, K., D. Hidalgo, and J.M. Velásquez. 2015. “Estudio de Evaluación Externa al Sistema de Transporte Público Remunerado de Pasajeros de la Provincia de Santiago y de las Comunas de San Bernardo y Puente Alto.” EMBARQ www​.embarq​.org. (http://www​.dtpm​.gob​.cl​/descargas/ estud​ios/ I​​nform​​e​%20E​​mbarq​​%20Ve​​rsi​%C​​3​%B3n​​%20Co​​r regi​​da​%20​​20​15-​​01​-21​​.pdf.​) Hagen, J. X., C. Pardo and J.B. Valente. 2016. “Motivations for Motorcycle Use for Urban Travel in Latin America: A Qualitative Study.” Transport Policy 49: 93–104. Hidalgo, D., and C. Huizenga. 2013. “Implementation of Sustainable Urban Transport in Latin America.” Research in Transportation Economics 40(1): 66–77. Lagos, V., A. Muñoz, and C. Zulehner. 2018. “Entry of Uber, Alcohol-Related Traffic Accidents, and Differences by Gender: Empirical Evidence from Chile.” Mimeo: Telecom ParisTech. Mendoza, M.A. 2014. “Panorama Preliminar de los Subsidios y Los Impuestos a las Gasolinas y Diésel en los Países de América Latina.” Estudios del Cambio Climático en América Latina, Documento de Proyecto, CEPAL. Santiago. Mohring, H. 1972. “Optimization and Scale Economies in Urban Bus Transportation.” American Economic Review 62: 591–604. Montero, J.P., F. Gallego and N. Barahona. 2020. “Vintage-specific driving restrictions.” Review of Economic Studies 87(4): 1646–1682. Parry, I.W.H., and K.A. Small. 2009. “Should Transit Subsidies Be Reduced?” American Economic Review 99(3): 700–744. Parry, I.W.H., and J. Strand. 2012. “International Fuel Tax Assessment: An Application to Chile.” Environment and Development Economics 17(2): 127–144. Parry, I.W.H., and G.R. Timilisina. 2010. “How Should Passenger Travel in Mexico City Be Priced?” Journal of Urban Economics 68: 167–182. Rivas, M.E., J.P. Brichetti, and T. Serebrisky. 2020. “Operating Subsidies in Urban Public Transit in Latin America: A Quick View.” IDB Monograph 786. Inter-American Development Bank, Washington, DC. Rivas, M.E., T. Serebrisky, and A. Suárez Alemán. 2018. “How Affordable Is Transportation in Latin America and the Caribbean?” IDB Technical Note 1588. Inter-American Development Bank, Washington, DC. Rivas, M.E., A. Suárez-Alemán, and T. Serebrisky. 2019. “Hechos Estilizados de Transporte Urbano en América Latina y el Caribe.” Nota Técnica del BID 1640. Banco Interamericano de Desarrollo, Washington, DC. Rivera, N.M. 2019. “Air-Quality Warnings and Temporary Driving Bans: Evidence from Air Pollution, Car Trips, and Mass-Transit Ridership in Santiago.” Mimeo. University of Alaska. Rizzi L.I., and C. de la Maza. 2017. “The External Costs of Private Versus Public Road Transport in the Metropolitan Area of Santiago, Chile.” Transportation Research Part A Policy and Practice 98: 123–140. Rodríguez, D.A., M. Santana, and C.F. Pardo. 2015. La Motocicleta en América Latina: Caracterización de su Uso e Impactos en la Movilidad en Cinco Ciudades de la Región, Banco de Desarrollo de América Latina, CAF.

Pricing urban transport in Latin America  435

Rodríguez Hernández, C.R., and T. Peralta-Quiros. 2016. “Balancing Financial Sustainability and Affordability in Public Transport: The Case of Bogotá, Colombia.” International Transport Forum Discussion Paper 2016/16. OECD, Paris. Serebrisky, T., A. Gómez‐Lobo, N. Estupiñán, and R. Muñoz‐Raskin. 2009. “Affordability and Subsidies in Public Urban Transport: What Do We Mean, What Can Be Done?” Transport Reviews 29(6): 715–39. SITP. 2018. Beneficios de Transporte para Personas Sisbenizadas. Sistema Integrado de Transporte Público de Bogotá. http://www​.sitp​.gov​.co​/publicaciones​/ beneficios​_de​_transporte​_para​_personas​_ sisbenizadas. Tirachini, A. 2019. “South America: The Challenge of Transition.” In J. Stanley and D. Hensher, eds., pp. 118–125, A Research Agenda for Transport Policy. Edward Elgar. Tirachini, A., and A. Gómez-Lobo. 2020. “Does Ride-Hailing Increase or Decrease Vehicle Kilometers Traveled (VKT)? A Simulation Approach for Santiago, Chile.” International Journal of Sustainable Transport 14(3): 187–204. Tirachini, A., and S. Proost. 2020. “Transport Taxes and Subsidies in Developing Countries: The Effect of Income Inequality Aversion.” Economics of Transportation 25(C). Toro-González, D., V. Cantillo, and V. Cantillo-García. 2020. “Factors Influencing Demand for Public Transport in Colombia.” Research in Transportation Business & Management 36: 100514. Tsivanidis, N. 2019. “The Aggregate and Distributional Effects of Urban Transit Infrastructure: Evidence of Bogotá’s TransMilenio.” Dartmouth College, unpublished. Tun T.H., B. Welle, D. Hidalgo, C. Albuquerque, S. Castellanos, R. Sclar, and D. Escalante. 2020. “Informal and Semi-formal Services in Latin America: An Overview of Public Transportation Reforms.” IDB Monograph 839. Inter-American Development Bank in collaboration with the World Resources Institute and the Global Environmental Facility, Washington, DC. Van Reeven, P. 2008. “Subsidisation of Urban Public Transport and the Mohring Effect.” Journal of Transport Economics and Policy 42(2): 349–359. Yañez-Pagan, P., D. Martinez, O.A. Mitnik, L. Scholl, and A. Vazquez. 2019. “Urban Transport Systems in Latin America and the Caribbean: Lessons and Challenges.” Latin American Economic Review 28(15): 1–25.

23. Road pricing applications in North America1 Mark Burris, John Brady and Sruthi Ashraf

23.1 INTRODUCTION This chapter discusses transport pricing and funding in the United States and Canada, with a focus on road pricing in the United States. The chapter begins with a brief history of road pricing in the United States from the earliest days of priced travel until 1995 when the first innovative, priced managed lane projects opened. Now a common part of highway networks in major metropolitan regions, these projects have been extensively studied. This chapter briefly discusses the growth in innovative funding and financing through those priced facilities before focusing on more recent insights into these innovative projects. We review three key issues with initiating pricing projects: (1) the ability of the priced facility to garner support and meet traveler expectations, (2) the industry’s ability to predict travel on, and revenue from, potential priced facilities, and (3) the choice of a dynamic or variable pricing mechanism. These three issues have gained increased importance as priced facilities have moved from infancy to a high growth stage.

23.2 HISTORY OF ROADWAY PRICING 23.2.1 From Tolling to Taxes … The earliest toll roads in the United States were born of necessity. In 1792, the Commonwealth of Pennsylvania wanted to expedite the flow of goods between the farmers of Lancaster County and the city of Philadelphia but lacked the funds to construct a 70-mile road between the two. By an act of the Philadelphia Assembly, the Commonwealth formed the Company of the Lancaster and Turnpike Road, allowing it to raise funds, construct, and operate the first turnpike in the United States (Klein and Majewski, 2008). The road was a financial and political success, connecting the two cities and spurring development along its length. As the demand for long-distance travel grew through the 1800s and into the early 1900s, the burden of building and maintaining road infrastructure increasingly fell to the federal government and then to state governments (FHWA, 2017). To build new roads, turnpike companies were formed by acts of legislature and still others were established by private investors hoping to stimulate development and enhance access to their land. A shift away from tolls began in 1916 with the Federal Aid Road Act which indicated any roads built using this act had to be toll-free (Rusch, 1984). Shortly thereafter, in 1919, Oregon was the first state to collect taxes on gasoline to pay for roads (Chokshi, 2014). The federal government followed suit, setting a one-cent per mile tax on gasoline as part of the Revenue Act of 1932, although the revenues were not earmarked for transportation (June 6, 1932, ch. 209, 47 Stat 169). In 1956 the Federal Aid Highway Act established the federal Highway Trust Fund, financed by state and federal gas taxes, to fund a new interstate highway network. The new fund and its 436

Road pricing applications in North America  437

rules against using federal dollars to support the construction of tolled facilities curtailed and then effectively eliminated the need to construct new toll roads through the 1960s and early 1970s (Lee and Miller, 2015). The responsibility for interstate highways and state roads fell to the states with most, if not all, funding coming through the Highway Trust Fund and state gas taxes. Local roads stayed the responsibility of local governments and are generally funded through local taxes such as sales tax and property tax. 23.2.2 … and Back to Tolling By the 1980s infrastructure built under the Federal Aid Highway Act was in need of significant maintenance and reconstruction. As gas taxes were insufficient to meet both the cost of maintenance and the demand for new roadways, states turned to local toll authorities to fund new, greenfield, infrastructure. As with past tolls, drivers were tolled at a flat rate at fixed points along the route, with the tolls used to cover costs. Tolling was politically preferable to increased taxes, as only users of the facility would pay and the development process had fewer requirements imposed by the federal government than when using federal dollars to fund a new highway (Lee and Miller, 2015). In the early 1990s, a turning point in tolling occurred when the California Department of Transportation (Caltrans) empowered the California Private Transportation Company to construct the nation’s first express toll lanes along State Route 91. These lanes opened in 1995. It was the first privately-funded tollway built since 1940, but more importantly, SR-91 was the first road to promise that its tolls would ensure a congestion-free trip on the express lanes in addition to covering the cost of construction and maintenance (OCTA, 2021). Dozens of similar tolled express lanes have been constructed in the United States since (TRB, 2021), we explore these lanes in greater detail in the following sections. 23.2.3 Meanwhile, in Canada … Transport funding in Canada is divided among the federal government, provincial government, and local government.2 Funding for transportation infrastructure is taken from general funds so there is a weak, indirect link between gas taxes and transportation funding compared to the Highway Trust Fund in the United States. Taxes on fuel are considerably higher in Canada than in the United States,3 which may partly explain why tolling is less common in Canada than in the United States. In addition, there are very few examples of pricing mechanisms other than flat rate tolls which are constant across all times of day and days of the week. This chapter’s later discussion of priced infrastructure therefore focuses on example projects from the United States and briefly discusses the Canadian experience. Most tolled facilities in Canada are bridges and most of those are bridges crossing between Canada and the United States. There are a handful of other tolled bridges and highways within Canada with almost all of those having a single toll price. The two notable exceptions occur in Canada’s largest city, Toronto. This area has the highest levels of congestion and is home to two notable examples of innovative pricing to help manage demand. The most notable example is Highway 407 (407 Express Toll Road), which began operations in 1997 as an all-electronic, open-access tolled highway. It is one of the earliest examples of both variable tolling and all-electronic operations, meaning no toll booths were used. Tolls vary with the distance traveled and time of day. Another notable example is the

438  Handbook on transport pricing and financing

Queen Elizabeth Way (QEW), which began a high-occupancy toll (HOT) lane in 2016. The QEW offers a limited number of travelers the option to use the lanes as a single-occupant vehicle for a toll. This is a couple of decades after HOT lanes appeared in the United States, but it still faced significant political opposition despite the success of HOT lanes in many US states. 23.2.4 Today’s Need for Priced Infrastructure Like in the recent past, today’s need for funding is the sum of the need to maintain the current infrastructure and the desire to improve and expand what we have. Annual spending in the United States on highway capital investment at the federal and state levels was a combined $105.2 billion in 2016, including federal stimulus funds. The Federal Highway Administration’s model suggests that $90 billion per year would be sufficient to maintain the existing infrastructure at status-quo conditions, but to improve highway conditions and performance, the same model estimates that $142.5 billion per year is needed to implement all known highway improvement projects whose benefits are greater than their costs (Poole, 2018). Comedian John Oliver quipped that such numbers are like, “having the state of our nation’s tennis balls assessed by the American Society of Golden Retrievers,” but observers on the left and right acknowledge the country’s need for infrastructure outstrips our nation’s ability or willingness to fund them (Oliver, 2015). In response, a minority of states have increased gas taxes in the past decades but political compromise has led to the diversion of some of that revenue to items such as education and healthcare (NCSL, 2020). More notably, the federal gas tax has not increased since 1993. Accounting for the eroding effects of increased fuel efficiency and inflation, federal gas tax collected in 2019 was about 10% greater in real terms than in 1995 despite a 150% increase in Vehicle Miles Traveled (FHWA, 2019; Scheinberg, 1996; CBO, 2020). All this has left states in a tight fiscal position. States today are spending more and more of their annual revenues on upkeep, leaving less to support new projects. For example, the Texas Department of Transportation spent $4.9 billion on maintenance while spending $3 billion on construction in the fiscal year 2019.4 In addition, as urban areas densify and our expectations for safety and environmentally-sound solutions grow, new projects have become increasingly expensive. To construct these high-cost new projects, states have turned to a modern combination of financing, partnerships, and high-tech tolling. 23.2.5 Public–Private Partnerships States often look to the private sector for innovation and funding when faced with risky, highcomplexity mega-projects. Like the Lancaster Pennsylvania Toll Road Company before them, today’s public–private partnerships (PPP or P3) are often charged with designing, building, financing, operating, and maintaining the new project (Poole, 2018). Innovations at all steps of the process by the private sector can make projects more feasible or possible without state subsidy. When funding a large project, the private sector packages a combination of equity, government loan programs like TIFIA, and public activity bonds (PABs). The up-front investment is often repaid via an availability payment – fixed amounts paid out over time – or via an agreement for the private sector to keep a share of future toll revenues.

Road pricing applications in North America  439

23.2.6 Managed Lanes/HOT Lanes/Express Lanes No piece of infrastructure more fully explores the modern intersection between pricing and infrastructure than managed lanes (MLs). MLs is the broader term for lanes within a freeway that are physically separated from the general purpose lanes (GPLs) and managed through a mix of policy and pricing to ensure a fast and reliable trip for those who use the MLs. Most of this chapter will be examining MLs that were called express lanes (ELs) by their owner/ operator and so this chapter primarily uses the term ELs. MLs are also sometimes called HOT lanes to indicate the common policy where High-Occupancy Vehicle (HOV) lanes allow single occupancy vehicles to pay to use the lanes while HOVs continue to drive for free. ELs have become a popular solution for states looking to create sustainable improvements in mobility along dense, already-congested urban corridors. Often, one or two lanes are added to the freeway with the additional capacity reserved for the ELs. The new lanes are often added at great expense, which can be mitigated through innovative design and financing, and the cost is recovered through thoughtful pricing of the facility. Several examples of innovative ELs have been developed with the private sector in the last decade, all of which have championed pricing as an efficient way to deliver an otherwiseinfeasible piece of infrastructure (Poole, 2018). As shown in Table 23.1, these projects are generally expensive to construct, revenue positive, and even showed continued resilience through the COVID-19 pandemic. While they often provide mobility near city centers (which tend to vote Democrat), they are built in states governed by both parties. The apolitical appeal of ELs springs from their novelty and promise of a financially self-sufficient, sustainable means of congestion relief. As evidenced by the bond ratings in Table 23.1, a well-placed and well-operated ELs largely fund themselves, sometimes with excess revenues going to the public. Well-operated ELs can break the “problem” of induced demand by increasing the capacity of a freeway and providing congestion relief on the ELs. Public opinion is often initially skeptical, concerned that high prices will be needed to keep the ELs flowing during rush hour, when demand for the lanes would exceed supply at a typical toll rate. Research confirms that public opinion is generally positive after the ML is operational. From an operational perspective, the most notable aspect of an EL is its use of pricing. Some facilities use a fixed schedule of rates while others allow prices to fluctuate in real time in response to traffic conditions on the roadway. Regardless of how the rate is determined, the trade-off between the price and the premium service offered by the ML offers a great opportunity to explore how travelers’ value that service. The remainder of this chapter discusses the operational elements of a ML, derived from the growing body of empirical research and experience with these operational facilities. It examines the challenges of forecasting traffic on the ELs, designing effective policies to ensure reliable day-to-day congestion relief, and predicting driver behavior on the facilities.

23.3 A MODERN PERSPECTIVE ON PRICING The preceding sections of this chapter were only an overview of how priced roadway facilities in the United States have matured to date. Many papers, including the many cited above, are available for readers looking for details on specific projects/areas of interest. Given the

440

$1.4 B $0.2 B $1.4 B $0.2 B $0.5 B

Private Public Public Public Private

12

Colorado HPTE C-470

Note: *I-35E only TIFIA Springing Lien.

US-36 Plenary Roads, Denver (CO)

I-35E (TX)

SR-91 (CA) 11

9,10

395/95 (VA)

$1.6 B

Private

13,14

$2.5 B

Private

NTE 35W (TX)8

North Tarrant Express (TX)

$2.7 B

Private

7

LBJ TEXpress (TX)6

Construction Cost

Public or Private?

Express Lanes

Table 23.1  Notable ELs in the United States5

BBB–

BBB

BBB*

A+

BBB

BBB–

BBB

BBB–

Post-COVID Fitch Senior Lien RatingA

Stable

Negative

Positive

Stable

Stable

Stable

Stable

Stable

D

D

D

D

D

R

R

D

Post-COVID Political Fitch Outlook Affinity of Area

D

D

R

D

D

R

R

R

Political Affinity of State

Yes

*

*

*

Yes

Yes

Yes

Yes

Revenue Sharing Agreement in Place?

Road pricing applications in North America  441

relatively mature state of priced facilities, the remainder of this paper focuses on recent research into three key aspects facing priced express lanes. 1. Predicting the use of ELs. Generally, traffic and revenue estimates have overestimated the amount of traffic on priced facilities (Flyvbjerg et  al., 2005). These overestimates have (usually) not been so bad that the facility defaults financially. However, new research shows that although the aggregate estimates are not catastrophically bad, some fundamental assumptions behind these estimates are incorrect. Thus, models of individual traveler use of ELs are missing key aspects of travel behavior and we discuss progress toward improved models below. 2. Traveler sentiment. One key aspect of getting a pricing project “off the ground” is public and political support. Recent research investigated if ELs were delivering on their promise of superior performance. It also examined how public sentiment toward ELs may vary with the ability of the EL to deliver on any promised performance measures. 3. Pricing mechanism. ELs have used both dynamic and variable pricing to manage demand for the lanes. Dynamic pricing adjusts the toll rate based on real-time traffic conditions. Variable pricing adjusts the toll rate on a set schedule throughout the day. Thus, in theory, dynamic pricing should be superior as it can change more rapidly to changing traffic conditions. However, there are downsides to dynamic pricing so the two forms of pricing were compared for several EL facilities across the United States. 23.3.1 Predicting Surprising Travel Behavior on Freeways with ELs The priced facility must offer travelers something of value to encourage some to pay for its use. In the case of traditional toll roads, this could be both a more direct route and a shorter travel time between origin and destination. Thus, toll roads could compete with non-tolled roads partially based on location. However, given the vast network of roads in the United States, the majority of toll roads did have nearby parallel non-tolled routes, often congested surface streets with intersections. So, the primary advantage of the toll road was simply travel time and this was the key variable when estimating if a traveler would pay a toll for using the toll road. This assumption was even more “obvious” when estimating the use of ELs. These lanes were on the same highway as the non-toll lanes and thus both had similar origins and destinations plus both had uninterrupted flow. So, it seemed clear that travelers were choosing ELs based on their travel time savings, plus added travel time reliability, versus the toll cost. This assumption is still in use today despite the empirical evidence from ELs that contradicts this. Good empirical evidence on travel behavior on these roadways is rare. Clearly, toll authorities know when someone used the toll road/lane, but generally do not know when that same traveler used a non-tolled option for a similar trip. Thus, it was impossible to know the choice made by the traveler on days they were not observed on the ELs. Were they choosing the GPLs or were they not traveling on/near the facility? This was particularly problematic for analysis of EL use as many travelers use the ELs on an infrequent basis (Burris and Appiah, 2004; Burris et al., 2012.; Wood et al., 2014) Traditionally, the assumption was that travelers were alternating between the GPLs and the ELs based on their travel needs for a particular trip. This assumption was made despite the lack of information on non-EL travel. However, empirical evidence from two toll facilities in

442  Handbook on transport pricing and financing

Texas shows that travelers are much more “set in their ways” than expected and many are not alternating between the ELs and GPLs. The Katy Freeway in Houston and the North Tarrant Express (NTE) freeway in Dallas have both ELs and GPLs and have transponder readers situated to record travelers on both types of lanes. Most roadways only have these transponder readers on the ELs. This provided a unique opportunity to observe traveler behavior regardless of the lane chosen. There were several interesting findings, the key one being the surprising inertia to switching lanes (Burris and Brady, 2018). On the Katy Freeway, travelers must have a transponder to use the ELs. Estimates are that approximately half of the vehicles on the Katy Freeway have a transponder. Therefore, half of the travelers always use the GPLs because they lack the necessary payment device to use the ELs. These transponders are fairly easy to obtain and TxDOT has repeatedly offered them for free, all that was required was money for prepaid tolls. So, for most of these travelers, it was a deliberate choice not to get a transponder and always use the GPLs (not a financial constraint). That leaves the other half of the travelers who have a transponder and can easily use either set of lanes anytime they are on that section of Katy Freeway. During the months of March, April, and May of 2014, over 1.1 million different transponders were read on the section of the Katy Freeway with ELs. Over 84% only used the GPLs and over 3% only used the ELs, leaving only 12.3% who used each lane type at least once during that time (see Table 18.2). The numbers were not much different when a longer time frame was examined. Thus, most travelers were not making a day-to-day decision on whether to use the ELs based on traffic conditions, toll rate, their travel, etc. They have chosen a type of lane to use and stick with it. Which lane to choose was now one less decision that they needed to make and they have made it an automatic choice rather than a conscious decision (Kahneman, 2011).​ Another clear indication that use of the ELs is based on a great deal more than just travel time savings and tolls has occurred as part of the COVID-19 Pandemic. During the pandemic, travel in the United States plummeted as citizens elected to stay home, governments imposed restrictions on mobility and commerce, schools stopped in-person instruction, and employers encouraged work remotely. In the intervening months, congestion was nearly non-existent (Lee, 2020). As attitudes changed and mobility restrictions were eased, overall Vehicle Miles Traveled in the typical US city began to recover (Wegscheider, 2020). Despite the precipitous drop in congestion and free-flowing GPLs, traffic and revenue on ELs did not drop to 0, as shown in Figure 23.1 (supplied by CDM Smith). With no congestion on the GPLs, it could be expected that EL traffic and revenues would drop to near 0, but this was clearly not the case.​ Unfortunately, the data discussed above offers no insight into the reasons for this behavior. Some of the most likely reasons why travelers use only one lane type include: 1. This is one less decision a person needs to make. We are inundated with decisions and like to move those to the automatic category of thinking if possible. In addition, it is one less bad decision that someone could make. Choosing to pay for the ELs when they are no faster than the GPLs, or choosing the GPLs when the ELs were much faster and cheap could both be considered bad decisions. Someone who makes either of these choices may feel that it is much worse than making no decision at all, pushing their use of lanes into the automatic category. This might be labeled inertia as well – the unwillingness to change.

Road pricing applications in North America  443

Table 23.2  Percentage of travelers choosing express lanes by frequency of travel Roadway

Katy

NTE

Share of Trips on the MLs

Frequency of Freeway Travel Infrequent (1–3 trips per month)

Sometimes (4–12 trips per month)

Commuter Level 1 (13–29 trips per month)

Commuter Level 2 (30 or more trips per month)

TOTAL

0%

89.3%

68.5%

51.9%

55.8%

84.4%

1% to 99%

6.8%

30.6%

47.1%

43.5%

12.3%

100%

3.9%

0.9%

1.0%

0.7%

3.3%

Number of Tags

894,408

152,390

51,684

9,520

1,108,002

0%

52.9%

35.1%

28.8%

33.3%

47.5%

1% to 99%

26.9%

61.9%

69.8%

65.9%

37.2%

100%

20.2%

3.0%

1.4%

0.8%

15.3%

Number of Tags

644,106

137,566

62,701

40,070

884,443

Notes: Katy: March, April, and May of 2014. NTE: August, September, and October of 2015.

Figure 23.1  Performance of express lanes during the COVID-19 pandemic 2. Added travel time reliability on the ELs. In a recent survey of 2800 EL users performed by the authors (discussed in more detail in the next section), reliability was the second highest-rated reason for using ELs. Thus, the travel time savings on a given EL trip may be small or even negative, but the overall reliability causes some travelers to always choose ELs. 3. Greater feeling of safety on the ELs. This was the third highest-ranked reason to use ELs in the survey mentioned above. 4. Unsure of the entry or exit points on the ELs. This was the second highest-ranked reason for not using ELs – behind the toll.

444  Handbook on transport pricing and financing

5. The toll is irrelevant to them. Someone else, such as the employer, may be paying the toll. Or, even though the traveler pays the toll it is done electronically and the cost only appears as an occasional charge on the credit card. Thus, the actual toll on a given trip is ignored. Note that this survey of 2800 travelers was taken in the summer of 2020 while congestion was minimal to non-existent. Approximately 1/3 of them had used the ELs during the previous week despite the lack of congestion. Of those that did use ELs, almost one-fourth said they just used the lanes that they always use. Many others noted travel time savings and travel time reliability of the ELs as reasons they used the lanes. Based on empirical data and the lack of congestion, we feel it is a much higher percentage of travelers who are simply using the lanes that they always use. Understanding and modeling this travel behavior requires a much better understanding of travelers and their choices (or lack thereof). Behavioral economics (BE) is a field of study that attempts to unravel and understand this complex decision-making process. It is not always rational and can be influenced by a variety of internal and external factors. Although there is little research directly on EL decision-making processes from a BE perspective, the research in other domains can be applied to this issue. In 2020, research examined whether people who think about which lane to use (chooser) versus people who always select a specific lane type (non-chooser) may have different psychological traits (individual differences). Being able to identify these travelers will improve EL use modeling by assigning non-choosers to their appropriate lane and then focusing on predicting and modeling the lane use of choosers. Research subjects participated in a BE experiment and completed a survey that collected information on their lane decision-making process, individual differences, and socio-demographic characteristics. During the BE experiment, travelers repeatedly chose between the main road and a side road to travel from point A to point B (Selten et al., 2004, 2007). Their objective was to select a road that had the shortest travel time which maximized their payoff. The travel time, and thus payoff, depended on the decisions made by all the subjects they were playing with. If many participants selected a specific road, then that road became congested and the payoff for choosing that road in that period was very small while the payoff for the other road was large. The control group played 100 periods where each subject knew only the details of their own road choices and payoffs. The treatment group received extra information in the last 50 periods about the travel time on the road which was not chosen in the previous round. This BE experiment intended to replicate the repetitive thought process behind the real-world EL choice scenario. At the end of the BE experiment, the participants completed a 30-minute survey that captured their real-world use of ELs and GPLs, their lane choice decision-making process, and details of their previous trip to EL facilities. The survey also collected information on sociodemographic characteristics and six individual differences. The individual differences examined were conscientiousness, need for cognitive closure, cognitive reflection, maximization, risk choice, and general mental ability. The first phase of this research was conducted on college students. This was a cost-effective way to adjust and finalize the BE experiment and the survey that was to be administered to drivers who travel on EL facilities. During the second phase of this study, the finalized BE experiment and survey were conducted on EL facility users to determine the factors that influenced their road selection process.

Road pricing applications in North America  445

Participants were classified as choosers and non-choosers based on the frequency of their road changes during the first 50 periods of the BE experiment. Participants with their number of road changes in the top 20% for each session were considered choosers. According to this definition, 20% of students were choosers, which was similar to the real-life data as seen in Burris and Brady (2018). Students with high scores for cognitive reflection and low scores for maximization tended to be non-choosers. Cognitive reflection is the tendency of someone to reflect on a question rather than reporting the first response that comes to their mind. Maximizers are individuals who strive to make an intelligent, informed decision that provides them with the maximum benefit. Travelers with high scores for the above two individual differences were hypothesized to be more likely to be choosers due to their ability and interest in analyzing the toll versus travel time trade-off. However, the results found the opposite for cognitive reflection scores. This might be because their previous trade-off analyses in the BE experiment did not result in a positive outcome and chasing past results was the intuitive response. Students who took more time to complete a section of the survey and took longer to verify their answers before submitting them were more likely to be choosers. This makes intuitive sense, participants spending a little more time looking over their answers before clicking submit are more likely to choose between lanes and not simply speed through the experiment selecting the same lane. Similar results were obtained when the above experiment and survey was conducted among drivers in Dallas, Texas. There also, choosers took more time to verify their responses before submitting them. Participants were assumed to have adopted a strategy to navigate through the BE experiment (except for the ones who remained in the same lane for much of the experiment). These strategies were classified into two: (1) direct response and (2) indirect response. Direct responders choose the road that had a higher payoff last period and indirect responders do the reverse. Since subjects do not perfectly play direct or indirect responses, an index was developed to classify students and travelers into the above two response categories. Based on this index, results showed that direct responders tend to be choosers. Both students and drivers tended to switch lanes less often as the experiment progressed. This seems to mimic real life as most travelers stick to one type of lane over time. During the last half of the BE experiment (last 50 periods), randomly selected participants were provided additional information (the treatment group). The additional information was the travel time for the road that they had not chosen in the previous period. This additional information (nudge) made non-choosers choose more. 23.3.2 Traveler Sentiment toward the Premium Service As discussed above, travel time savings and toll rate are just two of many potential factors influencing travelers use of ELs. Based on data from Katy Freeway and NTE there must be many other factors influencing the use of the lanes. For example, approximately 10% of Katy EL trips occurred when the EL speeds were equal to, or less than, the GPL speeds. The authors undertook research to identify key EL characteristics that impact the use of the lanes. Approximately 2800 EL travelers in four cities were surveyed about their use of ELs and their perception of many aspects of the lanes. The survey took place in the summer of 2020. The respondents were from a pool of paid survey participants from the survey firm Lucid. They

446  Handbook on transport pricing and financing

were at least 18 years old, live in a ZIP code near the ELs and had to drive on an EL in the months prior to the pandemic. The results were surprisingly uniform across survey locations, frequency of EL usage, age, income, and gender. When examining all respondents, key findings included that respondents were pleased with the ELs, as shown by: ● ● ●

82% felt the lane was worth the price paid 85% definitely or somewhat glad there are ELs 76% want to see more ELs in the area

The top reasons why they used ELs were because the ELs (1) saved them travel time, (2) their travel time was more reliable, and (3) they felt safer. When not using ELs, it was because (1) they avoid tolls, (2) they were not worth the cost, and (3) they had the flexibility to travel at uncongested times of the day. The changes that would most encourage them to use the ELs more often were (1) charging less and (2) more entrances and exits to the ELs. Respondents also rated how important they felt the EL pavement quality, ease of access on and off the ELs, safety, speed, and travel time reliability of the ELs were. Approximately 55% rated each aspect very important and approximately 35% rated each as important – except for pavement quality which received slightly lower importance. Responses by various groups of travelers were remarkably consistent regardless of location or user characteristic. There were a few differences in responses between groups, including: ●

● ●

High-income travelers were more likely to be glad that there are ELs and that more reliable travel times are important. Older travelers were less likely to want to see more ELs in the area. Those that receive information about the ELs were more likely to rate the EL characteristics higher, were more likely to be glad there are ELs, and more likely to want more ELs in the area.

These findings are representative of results from similar surveys taken by researchers and toll road operators around the United States (Mahandra, et al., 2011; Buckeye, 2014). Commercial vehicle owners selected improved road conditions because of (1) less congestion, (2) safe road conditions, and (3) as the primary reasons for using ELs (Cherry and Adelakun, 2012). Based on these findings it is clear that travelers do appreciate the ELs. Although this is usually much more evident once the lanes are functional. Prior to the opening of these lanes they often face stiff political opposition. One issue that often comes up in public opposition is the aspect of having a toll that varies by time of day or, even worse, congestion level. The merits of these two options are discussed in the next section. 23.3.3 Pricing Mechanism: Dynamic versus Variable Tolling Two common means of pricing ELs are to vary tolls based on time of day or to vary them dynamically based on real-time congestion. To keep the lanes uncongested the toll rate increases during periods of greater demand. This toll rate can be set in advance based on the time of day. For example, the toll might be $8 at 7:30 am and only $1 at 3:30 am. This is known as variable pricing. The other option is for the toll price to change dynamically based

Road pricing applications in North America  447

on prevailing traffic conditions. As traffic increases so does the toll rate. Similarly, as traffic decreases the toll rate declines. This method of tolling is known as dynamic pricing. With dynamic pricing, the exact toll a traveler may have to pay is unknown until they approach the lane and observe the price sign. Both dynamic and variable pricing systems have been proven to work well on multiple EL facilities across the United States. Which of the two tolling options is more effective in regulating EL usage is unclear. In theory, dynamic pricing has an advantage over variable pricing as dynamic pricing can react to current traffic conditions with higher granularity, accuracy, and adaptability. However, dynamic pricing generally faces additional political and public opposition because the exact toll rate at any given time is unknown. This concept is unfamiliar to many people and garners additional opposition. Also, dynamic pricing requires accurate information on traffic conditions at all times and the development of a pricing algorithm. Finally, there are some instances where a high-price signal has the unintended effect of attracting people to use the lane. Travelers observe a high price on the ELs and assume congested traffic conditions on the GPLs, thereby deciding to pay the toll for the ELs (Janson and Levinson, 2014). Putting together all of these issues with recent research showing many travelers are not influenced by price when making their lane choices (Burris and Brady, 2018), calls into question whether dynamic pricing is superior to variable pricing. A recent study examined large datasets of toll prices, vehicle travel speeds, traffic volumes, and densities to assess the effects of the two different congestion pricing strategies on traffic conditions (https://nicr​.usf​.edu​/2020​/12​/11​/4​-1​-pricing​-mechanisms​-for​-managed​-lanes/). The impact of the two pricing approaches on traffic conditions was evaluated to assess if the traffic management benefits from dynamic pricing are commensurate to its more resource-intensive implementation compared to variable pricing. The study used seven metrics to measure the effectiveness of the toll to keep traffic flowing on the ELs. Two of these were developed specifically for the study while the rest are wellknown traffic performance measures. These included: 1. Travel time savings. The average travel time saved by users of the ELs. This was often done assuming the person traveled the entire length of the ELs as detailed trip origin and destination information was not available. 2. Variability benefit. This is defined as the standard deviation of travel times on the GPLs divided by the standard deviation of travel times on the ELs. The larger this ratio, the less variable the travel times on the ELs are compared to the GPLs. 3. Planning time benefit. This metric also measures the reliability of the lanes. It is the difference between the planning time index of the GPLs minus the planning time index of the ELs. The planning time index is the ratio of the 95th percentile travel time divided by the free flow travel time. Thus, it is how much longer (in percent) it takes to travel that route on a congested day versus uncongested travel. 4. The ability of the toll to impact congestion. This measure examines the elasticity of demand on the ELs. It is calculated by first grouping the toll and traffic data into bins based on the toll amount. For example, finding all the EL flows when the toll was anywhere between $1 and $1.99. Then finding the average toll and average flow for that particular toll bin. With that data, the elasticity of demand from one bin to the next was calculated by dividing the percentage change in flow by the percentage change in toll.

448  Handbook on transport pricing and financing

5. Speed threshold. The average travel speed of vehicles was found and the percentage of the time that the average speed exceeded a given threshold was determined. One of these thresholds was 45 mph based on FHWA criteria that HOT lanes should exceed this speed during 90% or more of peak periods. The EL agency may have an additional threshold and that would also be examined. 6. Speed graphs. This is a visual representation of several of the above metrics. The graph simply has a single point for the speed on both the GPLs and ELs for each time interval. 7. Scoring Index. This is a new metric that simply gives the EL a score from –3 to +3 based on the speed and flow of the ELs and GPLs. The larger positive numbers indicate that the ELs are operating at a high level with flow near or exceeding capacity while speeds are also high, while GPLs are congested. This indicates the toll is functioning well. Negative scores indicate the ELs are functioning poorly. The research examined traffic data from two variably priced ELs (SR-91 and I-25) and four dynamically priced ELs (I-35W, I-394, I-35E, and MoPac). Based on the metrics above, both the variably priced and dynamically priced lanes were functioning well (see Tables 23.3 and 23.4). There were too few lanes and too little difference in their scores to definitively say one type of pricing was clearly better than the other. However, the dynamically priced lanes appear to have a slight edge in most metrics. Future research could add ELs to this analysis, and that may provide evidence for a clearly dominant pricing mechanism. Table 23.3  Performance measures from dynamically priced EL facilities Performance Measure

Dynamic Pricing I-35W

I-394

I-35E

MoPac

Average

TTS (minutes)

1.44

2.74

1.03

1.68

1.72

Variability benefit

2.08

1.87

2.26

1.14

1.84

PTI benefit

0.34

0.44

0.36

0.51

0.41

TTI benefit

0.09

0.24

0.17

0.15

0.16

BTI benefit

0.09

0.11

0.13

0.27

0.15

Ability of the toll to impact congestion

Small positive numbers at higher prices.

Small negative numbers at higher prices.

Large negative numbers at higher prices.

Small positive at lower prices. Slightly larger positive/negative at the highest prices.

–

Achieve 45 mph goal (percentage of time)

Yes (97%)

Yes (99%)

Yes (100%)

No (87.5%)

Yes (96%)

Achieve internal speed goal (percentage of time)

Yes (94%)

Yes (92%)

Yes (99%)

Yes (95%)

Yes (95%)

Scoring index

0.75

0.88

0.95

0.86

0.86

Road pricing applications in North America  449

Table 23.4  Performance measures from variably priced EL facilities Performance Measure

Variable Pricing SR-91 (Orange County Portion)

I-25 North Express Lanes

Average

TTS (minutes)

0.45

2.56

1.51

Variability benefit

1.53

2.26

1.90

PTI benefit

0.11

0.56

0.34

TTI benefit

0.01

0.19

0.10

BTI benefit

0.1

0.29

0.20

Ability of the toll to impact congestion

Positive values at all prices. Very small (0.04) at maximum price.

Small negative/positive at higher prices. Large positive at lower prices.

–

Achieve 45 mph goal (percentage of time)

No (82%)

Yes (91.9%)

No (87%)

Achieve internal speed goal (percent of time)

Yes (90%)

N/A

Yes (90%)

Scoring index

0.45

0.97

0.71

23.4 CONCLUSIONS Transport pricing in the United States has had an interesting history with several distinct eras. This chapter has focused on the most recent EL era and the unique experiences the United States has had with ELs. Despite the wealth of research on the subject, we have shown that there is also still much to learn about ELs and a great deal that ELs will teach us about travelers and travel behavior. The future of ELs is difficult to predict. It would have been hard to predict the continued use of the ELs during the pandemic when the GPLs were congestion free. Equally difficult to predict is what will happen when there is a proliferation of connected or automated vehicles on the freeways? Will ELs flourish as the preferred lane for such vehicles? Or will travelers avoid toll facilities altogether once the burden of doing the driving is removed? Another unknown is the potential politicization of ELs. When a new EL is proposed there is often discussion about the equity of the lane. In the current climate in the United States, where items quickly become extremely politicized, can that create insurmountable hurdles for ELs? Despite these potential issues, we feel ELs will play a role in the transportation system for a long time to come. They have been successful to date and adaptable to varying situations – providing them increase flexibility for whatever the future brings.

NOTES 1.

Mexico is officially part of North America but is covered in Chapter 22 in this Handbook on Latin America.

450  Handbook on transport pricing and financing

2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

https://tc ​ . canada ​ . ca ​ /en ​ /corporate ​ - services ​ / policies ​ /government​ - expenditures ​ - revenuestransportation). www​.nrcan​.gc​.ca ​/our​-natural​-resources​/domestic​-and​-international​-markets​/transportation​-fuel​s​/fuel​-consumption​-taxes​-canada​/18885 https://ftp​.txdot​.gov​/pub​/txdot​-info​/fin​/reporting​/distribution​-coin​-chart​-2​.pdf www​.fitchratings​.com ​/research ​/infrastructure​-project​-finance​/fitch​-takes​-rating​-actions​- on​-us​managed​-lanes​-amid​-coronavirus​-disruptions​-30 ​- 03​-2020 www​.lbjtexpress​.com​/faq​-page​/83 www​.ntetexpress​.com​/ faq​-page#:~​:text​=What​%20is​%20the​%20scope​%20of​,corridor​%20in​ %20Northeast​%20Tarrant​%20County www​.txdot​.gov​/government ​/partnerships​/current​-cda ​/north​-tarrant expre​​ss​.ht​​ml#:~​​:text​​=The%​​20​ %24​​1​.6​%2​​0bill​​ion​%2​​0NTE%​​20​​%7C​​%20I,​​LLC​%2​0(NTE​MP3)%​20and​%20Tx​DOT. 395 https://csengineermag​.com​/construction​-begins​-virginias​-395​-express​-lanes- exten​sion/​#:~:t​ ext=T​he%20​proje​ct%20​is%20​antic​ipate​d%20t​o,95%​2FI%2​D395%​20cor​r idor​. 95 www​.p3virginia​.org​/projects​/95​-express​-lanes/ originally  $150M https://www​.fhwa​.dot​.gov​/ipd ​/project​_profiles​/ca ​_91expresslanes​.aspx extension into riverside $1.2B www​.roadtraffic​-technology​.com​/projects​/the​-35express​-project​-dentondallas​-texas/ www​.codot​.gov​/projects​/archived​-project​-sites​/ US36ExpressLanes​/project​-overview https://plenaryroadsdenver​.com ​/wp ​- content ​/uploads​/2019​/08​/ PRD ​-US​-36 ​- Construction​-Fact​Sheet-​.pdf

REFERENCES Buckeye, K. R. (2014). Express lanes performance evaluation: Interstate 35W in Minnesota. Transportation Research Record, 2450(1), 36–43. Burris, M. W. and Appiah, J. (2004). “An examination of Houston’s QuickRide participants by frequency of QuickRide usage.” Transportation Research Record 1864, TRB, National Research Council, Washington, D.C, 22–30. Burris, M. and Brady, J. (2018). “Unrevealed preferences: Unexpected traveler response to pricing on managed lanes.” Journal of the Transportation Research Board. doi: 10.1177/0361198118796928 Burris, M., Nelson, S., Kelly, P., Gupta, P. and Cho, Y. (2012). “Willingness to pay for HOT lanes Empirical analysis from I-15 and I-394.” Journal of the Transportation Research Board, TRR # 2297, 47–55. doi: 10.3141/2297-06 Cherry, C. R. and Adelakun, A. A. (2012). Truck driver perceptions and preferences: Congestion and conflict, managed lanes, and tolls. Transport Policy, 24, 1–9. Chokshi, N. (2014). “A (very) Brief History of the State Gas Tax on its 95th Birthday.” Washington Post, March, 2021. www​.washingtonpost​.com​/ blogs​/govbeat​/wp​/2014​/02​/25​/a​-very​-brief​-history​-of​the​-state​-gas​-tax​-on​-its​-95th​-birthday/ Congressional Budget Office (CBO). (2020). “Reauthorizing Federal Highway Programs: Issues and Options.” May. www​.cbo​.gov​/system​/files​/2020​- 05​/56346​-CBO​-Highway​-Reauthorization​.pdf Federal Highway Administration (2017). “Toll Facilities in the United States.” March 2021. www​.fhwa​ .dot​.gov​/policyinformation​/tollpage​/ history​.cfm Federal Highway Administration (FHWA). “Historical Vehicle Miles Travelled Report (ANA7)”, Travel Monitoring, published November 8, 2019. www​.fhwa​.dot​.gov​/policyinformation​/travel​_monitoring​/ historicvmt​.pdf Flyvbjerg, B., Mette K. Skamris Holm and Søren L. Buhl. (2005). How (In)accurate are demand forecasts in public works projects?: The case of transportation. Journal of the American Planning Association, 71(2), 131–146. doi: 10.1080/01944360508976688 Janson, M. and D. Levinson. (2014). HOT or not: Driver elasticity to price on the MnPASS HOT lanes. Research in Transport Economics, 44, 21–32. doi: 10.1016/j.retrec.2014.04.008 Kahneman, D. 2011. Thinking, Fast and Slow. New York: Farrar, Straus and Giroux.

Road pricing applications in North America  451

Klein, D. and Majewski, J. (2008). “Turnpikes and Toll Roads in Nineteenth-Century America.” EH.Net Encyclopedia, edited by Robert Whaples. March, 2021. http://eh​.net​/encyclopedia​/turnpikes​-and​-toll​roads​-in​-nineteenth​-century​-america/ Lee, J. “Coronavirus: What It Looks like When the World Stands Still”, published April 19, 2020. www​.bloomberg​.com​/opinion​/articles​/2020​-04​-19​/coronavirus​-what​-it​-looks​-like​-when​-we​-all​stop​-driving Lee, C. and Miller, J. (2015). “Lessons Learned from the Rise, Fall, and Rise of Toll Roads in the United States and Virginia.” March 2021. www​.p3virginia​.org​/wp​-content​/uploads​/2015​/06​/ Rise​-Fall​-and​Rise​-of​-Toll​-Roads​-John​-Miller​-2015​.pdf Mahendra, A., Grant, M., Higgins, T. and Bhatt, K. (2011). “Road Pricing: Public Perceptions and Program Development.” NCHRP Report 686, National Academy of Science, Washington, DC. National Congress of State Legislatures (NCSL). (2020). “Recent Legislative Actions Likely to Change Gas Taxes”, Transportation, published August 12, 2020. www​.ncsl​.org​/research​/transportation​/2013​and​-2014​-legislative​-actions​-likely​-to​-change​-gas​-taxes​.aspx Oliver, John. (2015). “Infrastructure: Last week tonight with John Oliver,” YouTube video, 21:13. March 2, 2015. www​.youtube​.com​/watch​?v​=Wpzvaqypav8. The Orange County Transportation Authority (OCTA). (2021). “General Info”, SR-91 Express Lanes, accessed April 9. www​.91expresslanes​.com ​/general​-info/ Poole, Robert. (2018). Rethinking America’s Highways: A 21st-Century Vision for Better Infrastructure. Chicago, IL: The University of Chicago Press. Rusch, W. A. (1984). “Toll Highway Financing.” National Cooperative Highway Research Program, Synthesis of Highway Practice #117, National Research Council. Washington, DC. Scheinberg, Phyllis F. (1996). “Highway Trust Fund: Financial Status and Outlook”, Government Accountability Office (GAO), May 16. www​.govinfo​.gov​/content​/pkg​/GAOREPORTS​-T​-RCED​-96​169​/pdf​/GAOREPORTS​-T​-RCED​-96​-169​.pdf Selten, R., Chmura, T., Pitz, T., Kube, S. and Schreckenberg, M. (2007). Commuters route choice behaviour. Games and Economic Behavior, 58(2), 394–406. Selten, R., Schreckenberg, M., Chmura, T., Pitz, T., Kube, S., Hafstein, S.F., Chrobok, R., Pottmeier, A. and Wahle, J. (2004). “Experimental investigation of day-to-day route-choice behaviour and network simulations of autobahn traffic in North Rhine-Westphalia”. In Schreckenberg, A. and Selten, R., editors, Human Behaviour and Traffic Networks, Springer, Berlin Heidelberg, 2004, pp. 1–21. Transportation Research Board (TRB). (2021). Standing Committee on Managed Lanes, “Projects Open to Traffic”, Matthew MacGregor National Express Lanes Practices, accessed April 9, 2021. https://managedlanes​.wordpress​.com ​/2021​/01​/05​/projects​-database/ Wegscheider, Augustin K. (2020). “4 Patterns for Vehicle Miles Travelled by State in 2020”, published July 24. www​.streetlightdata​.com ​/4​-patterns​-vmt​-vehicle​-miles​-traveled​-by​-state​-2020/​?type​=blog/ Wood, N., Burris, M. and Danda, S. (2014). “Examination of paid travel on the I-85 express lanes.” Journal of the Transportation Research Board, TRR # 2450, 44–51. doi: 10.3141/2450-06

24. Transport pricing and financing in Oceania John Stanley and David A. Hensher

24.1 CONTEXT In Australia and New Zealand, as in many other countries, transport pricing has primarily been something to discuss, rather than an agenda for actually delivering improved resource allocation and/or larger governmental revenue flows. Road pricing reform (often referred to as road user charges), in particular, has generated a huge Australian literature over many years but with very little to show in terms of changing pricing systems. There has, however, been some innovative work on public transport fare settings, particularly in New South Wales. Transport financing is another area where Australia can also lay claim to innovation in delivery, particularly through its early and long-standing engagement with public–private partnerships, mainly on toll-road concessions. We examine current transport pricing and financing practices in Oceania,1 focusing mainly on examples of what we see as international good practice (humility prevents us from calling it international best practice!) in the road and public transport sectors. The primary focus is Australia, but the chapter also draws attention to some practices in New Zealand, a country well known for punching above its weight.

24.2 AUSTRALIAN TRANSPORT PRICING 24.2.1 Road Pricing and Expenditure Australia was a relatively early instigator of cost-driven road pricing for heavy vehicle road use, through the 1991 heavy vehicle charging reforms introduced by the (then) National Road Transport Commission, working with the federal and state transport ministers. In this charging regime, fuel excise paid to the federal government by heavy vehicle (> 4.5 tonnes) road users is designated as a road user charge, with registration fees paid to state and territory governments also designated as road use charges. The model used for the latest update of these charges is described in NTC (2018, 2021), which describes cost attribution procedures across vehicle classes and axle configurations (the pricing focus is on cost recovery in contrast to user benefit attribution); 26.4c/L of the federal fuel excise (of 44.1c/L, as of August 2021) paid by heavy vehicles is designated as a road user charge, the difference being a fuel tax credit for users. Light vehicles pay 44.1c/L fuel excise in Australia but this is not designated as a road use charge. Overall, heavy vehicle charge levels are based on fully recovering expenditures attributable to heavy vehicle road use (but not externalities, such as delay costs), linked to vehicle axle classes, configurations (e.g., trailers) and associated on-road performance (e.g., typical loaded weights). However, NTC (2021) indicates that charge levels have not kept pace with increases in attributable road expenditure in recent years, transport ministers moderating the 452

Transport pricing and financing in Oceania  453

rate of charge increase. Charge levels would need to increase by over 13% (as of mid-2021) to achieve full cost recovery for heavy vehicles in aggregate (NTC, 2021). An increase in this level is unlikely to be approved by the Transport and Infrastructure Council of federal and state ministers (the approval mechanism). NTC (2021) includes alternative charging options that move toward full cost recovery, across all heavy vehicle classes, while ensuring that each particular class of heavy vehicle covers its attributable costs and makes a contribution to joint costs. However, uncertainties about the parameters most suited to attribute particular road costs, such as pavement damage, to individual vehicle classes remains a challenge for the development of a fair charging regime. For example, proposed changes in the cost allocation factors for periodic surface maintenance of sealed roads used in NTC (2021) lead to a huge jump in charges proposed on two-axle buses of more than 12 tonnes in mass. Such costing challenges confront all road charging regimes and point to the importance of ongoing research into cost drivers. The idea of incorporating externalities such as air pollution, enhanced greenhouse gas emissions (mainly CO2) and congestion costs into charges for road use, to make users more accountable for the wider costs of their travel decisions, has made little progress in Australia, for any class of vehicle, albeit that numerous academic and governmental inquiries have proposed so doing, as discussed later in this section. External costs such as air pollution, noise, and accident externalities are typically handled in Australia and New Zealand, as elsewhere, via legislative or regulatory means, such as vehicle emission standards, rather than via direct pricing of the externality. For Australian roads in 2017–2018, the federal government raised over $6 billion more from road-related charges than it spent on roads, whereas state and territory governments spent around $5 billion more than they raised in road-related charges, requiring other own-source revenue streams (including general purpose revenue grants from the federal government) to fund the difference. When road spending and revenue streams by federal, state, and local governments are combined, road expenditure has exceeded revenues for the last couple of years. Furthermore, the rate of expenditure growth on roads averaged over 6% annually in current prices from 2002–2003 to 2018–2019, with road-related revenues growing at only about half this rate (Figure 24.1), suggesting a growing ‘deficit’. When estimated external costs are added, Stanley and Hensher (2016) concluded that Australian road users have been in substantial deficit for some years. Using the Parry and Small (2005) model to estimate social marginal costs of road use, with updated parameter values, they showed that (then) current road user charges for light vehicles, on average, fall substantially short of what would be needed to fully internalize external costs. While recognizing that fuel taxes are not an ideal way to charge for road use, since most of the external costs of such use do not vary as closely with fuel use as they do with distance traveled, they showed that Australian fuel taxes in 2016 would need to have been increased by 15–25c/L to fully recover external costs from light vehicle use. External costs (on average) for car use were estimated at 35.9c/L for congestion, 13.6c/L for accidents, 11.1c/L for GHG emissions, and 4.8c/L for noise and air pollution (2015 prices). Longer term, because external costs of road use relate more closely to distance traveled than to fuel use in free flow conditions, Stanley and Hensher argued that a distance-based charging mechanism should be introduced, with mass and location components to better reflect, for example, road damage and congestion impacts. The trend toward more fuel-efficient vehicles, albeit slow in Australia, accentuates pressure

454  Handbook on transport pricing and financing 35000 30000 25000 20000 15000

20

02 20 03 03 20 04 04 20 05 05 20 06 06 20 07 07 20 08 08 20 09 09 20 10 10 20 11 11 20 12 12 20 13 13 20 14 14 20 15 15 20 16 16 20 17 17 20 18 18 -1 9

10000

Total public sector road expenditure ($m 2017-18) Government use-related charges, inc. tolls ($m 2017-18) Note:   *Revenue is only available to 2017–2018. Source:   Data is from various tables in www​.bitre​.gov​.au​/sites​/default​/files​/documents​/BITRE​_2019​_ YEARBOOK​.pdf.

Figure 24.1  Australian governments' road expenditure and revenue: 2002–2003 to 2018–2019* for such a shift in the way road use is priced, because of the revenue impact on the federal government budget. The major focus of discussions about transport pricing reform in Australia has been on road pricing (often called road usage charges), prompted particularly by the high deadweight external costs of traffic congestion. These were estimated at $16.5 billion in 2015 by BITRE (2016), and projected to reach $30 billion by 2030, comprising a substantial part of the under-recovery of road use costs attributable to light vehicles identified by Stanley and Hensher (2016).2 At the federal level, road pricing reform has received endorsement by a number of investigations, such as the Henry Tax Review (2010), the Harper Competition Policy Review (Harper et al., 2015), the Productivity Commission (2017), and Infrastructure Australia (Davies, 2017). Harper et al. (2015), for example, argued that roads are the least reformed of all Australian infrastructure sectors and that cost-reflective pricing reform should be implemented, using new technologies for vehicle use recognition and charging. Revenue neutrality was proposed as a key aspect of such reform and is also assumed by most other governmental reports on transport pricing reform. All seem mainly concerned about strengthening links between road expenditure, road funding, and user charging, in a way that gives road users a greater say over resource allocation decisions on roads. This is a worthwhile intention but only part of the story, economic efficiency suggesting that substantial increases in charges might be required, albeit that implementation may be more difficult if significant price increases are mooted. The peak body representing road users, the Australian Automobile Association (2017), recognizes the need to reform the way road use is priced and how this links to funding allocations, also being a firm believer in hypothecation.3 Heavy vehicle charging is seen as a possible pathway to implementation. The AAA interest, however, seems to be driven more by concern about declining federal fuel tax receipts, the main source of federal road financial

Transport pricing and financing in Oceania  455

assistance, than by making road users accountable for the wider costs of their travel choices. Pricing for demand management is seen as a longer-term reform to complement more fundamental forms of direct charging (AAA, 2017, p. 8) (i.e., charging that reflects attributable road costs as part of a closed-loop expenditure–pricing–funding system). Political unpalatability has been key to this lack of action on the reform of road user charging. By way of illustration, the federal government made a commitment in November 2016 to establish and fund a study, chaired by an eminent Australian, to investigate the potential impacts of road user charging reform on road users (AAA, 2017). However, this inquiry never got off the ground and was shelved in the lead-up to the subsequent federal election (Tillett, 2018). A positive note on the outlook for road pricing reform in Australia is provided by the CEO of Infrastructure Partnerships Australia who argues that the extensive discussions that have taken place on this topic have demonstrated wide stakeholder support of the need for change and that electrification of the vehicle fleet provides an opportunity to transition to a new charging regime. Electric vehicles (EVs) pay no fuel excise, so an alternative means of charging for their road use is needed, to ensure equity with other vehicles/road users. Hensher and Mulley (2014) are among many economists who have proposed that distance-based charges be levied in general on all cars; however, there is political support for doing this initially on EV road use, with charges set to be similar to those levied on road users who pay fuel excise. The revenue so raised should be hypothecated to a road account. IPA (2018) suggests that the Commonwealth Government may have constitutional difficulties raising such a charge and that it provides a compelling revenue-raising opportunity for states and territories. Cooperation across jurisdictions would be needed in charge setting, to ensure fair pricing. A beneficial legacy of the Australian history of road pricing discussions is broad agreement around the endpoint, which should assist the eventual achievement of the necessary reforms. This idea has subsequently been taken up by the South Australian State Government. The state’s 2020–2021 Budget, released in November 2020 (Government of South Australia, 2020), announced the government’s intention to introduce a road user charge for electric and zeroemissions vehicles, expected to commence on 1 July 2021. The charge will include a fixed component, similar to registration, and a distance-based charge, with the State Government indicating its intention to consult with other jurisdictions about the charge. Victoria and South Australia, and more recently NSW, have proposed a 2.5 cents per km user charge on kilometers of electric vehicle usage. Although this is seen by some as an environmentally bad policy, it nevertheless has been passed in legislation in Victoria4 as a 2.5c/km charge on electric and other zero-emission vehicles, including hydrogen vehicles, and a 2.0 cent/km charge on plugin hybrid-electric vehicles. In deciding if this is a good idea, there are two conflicting agendas – one that argues that those who use the roads should pay for the benefit received, and the other to incentivize a young industry to grow its product given its environmental advantages over ICEs. Overlaying these two agendas is a growing concern that a switch out of petrol and diesel increasingly threatens the flow of fuel excise revenue collected by the federal government and is a tax that is avoided by BEVs (Hensher et al., 2021b). The electric vehicle lobby is not happy about this charge, arguing that there is a case for lower charges on EVs because of their low environmental impact (unless their electricity source is brown coal). A road use charge is seen as discouraging the purchase of EVs and hence being counter-productive in terms of lowering fleet emissions. There is merit in this

456  Handbook on transport pricing and financing

argument, the way forward being to not simply apply a new charge on electric cars but to revise all charges and align them with user-pays, including externalities such as greenhouse gas emissions and air pollution, where EVs have an edge when fueled by clean energy. In the shorter term, hypothecation of revenues raised from distance charges on EVs to fund initiatives such as the roll-out of EV recharging stations is one way to recognize the external benefits of (clean) EVs. Incentives to encourage electrification are also now being offered, in response to environmental criticism. For example, the Victoria Government recently announced a subsidy of $3000 for 4000 buyers of electric and hydrogen cars in Victoria, for models costing less than $68,740 before on-road costs, with two further funding rounds taking this up to 20,000 zero-emissions vehicle purchases to receive the subsidy over the next three years. However, despite the Victorian road user tax of 2c/km being imposed on plug-in hybrid models, the subsidy does not extend to those vehicles. More broadly, Hensher et al. (2021a) has argued that, without road pricing reform, under a scalable model of growth of electric cars with lower purchase prices and usage costs, there will be a proliferation of purchases beyond ICE replacement, with concerning consequences for congestion and public transport patronage, reinforcing the case for mass-distance-location charging. 24.2.2 Australian Public Transport Pricing Land-based public transport services are primarily a state/territory responsibility in Australia, with the federal government providing substantial but ad hoc financial assistance, including for some new major infrastructure projects. Australian public transport systems typically recover 20–30% of their operating costs through user charges (fares), the resulting deficits being funded through state/territory budgets. Perhaps the best example of Australian public transport pricing (fare setting), which stands comparison with the best international examples, is that applied by the Independent Pricing and Regulatory Tribunal (IPART) in NSW. The IPART approach was reviewed in workshop discussions at the 2015 and 2017 Thredbo International Conferences on Competition and Regulation in Public Transport, being recognized as a best practice fare-setting model (Stanley and Levinson, 2016; Stanley and Ljungberg, 2018). The discussions at those workshops provide a basis for assessing the IPART approach. Thredbo workshop participants developed the following formula, to enable estimation of the level of cost recovery that should be sought from public transport fares, based on an economic welfare maximization framework (Stanley and Ljungberg, 2018).5 Amount to be recovered by user fares = MSC – PTEB – MSLC where: ● ●

●

MSC = (efficient) marginal social cost of PT service (in a network sense) PTEB = public transport net external benefits (including system-external benefits, such as agglomeration economies, which should be funded by beneficiaries if possible, and system-internal benefits, especially network economies from the Mohring effect, which should be funded by the government) MSLC = minimum (safety net) service level cost (recognizing social inclusion benefits, which should be funded by government).

Transport pricing and financing in Oceania  457

Taking account of the potential scale of externalities and the MSLC, cost-recovery rates for urban PT of well under 50% (from fares charged to users) are to be expected in an Australian urban setting, with rural fare cost-recovery rates lower than urban. In relation to Sydney’s route bus services, IPART calculates that the level of costs that should be met from fares is as follows (IPART, 2013, Table 1.1): ● ●

● ● ● ● ● ●

Calculate the total efficient costs of the benchmark operator Less the efficient cost of providing school services (deducted to derive the costs of route bus services) Less non-fare revenue Equals net efficient costs of the benchmark operator Less the value of external benefits for the benchmark operator Equals Revenue requirement Less government subsidy for concession fares Total amount that should be funded by passengers (from fares).

Estimation of external benefits is a major focus of the IPART fare-setting procedure. Externalities assessed by IPART are congestion, emissions, accidents, health system benefits from active travel, and frequency benefits (the Mohring effect). The resulting IPART calculation suggests that fares should pay for about 40% of the efficient service costs for route bus services in Sydney, with government meeting about 60%, because of the external benefits from bus services (largely congestion cost savings) and government’s desire to support social inclusion, encompassing both concession fares and supported base service levels on low patronage services. Thredbo’s discussion focused on two aspects of this approach: ●

●

service costs used by IPART were seen as arguably higher than marginal social costs because they include depreciation and return on capital, which are part of the contractual payments to bus operators (about 19% of efficient costs as they are calculated) but not part of short-run marginal costs for efficient fare setting, and the treatment of social inclusion and agglomeration benefits.6

IPART recognizes that social inclusion benefits include elements that are internal to users (e.g., as estimated by Stanley et al., 2011) and others that are externalities (e.g., reduced crime). Only the externalities would apply in the IPART fare-setting framework but IPART reasonably asserts that there is no quantitative evidence of relevant magnitudes of such externalities. IPART recognizes that social inclusion benefits that are internal to transit users are relevant to governmental decisions about service funding levels for low patronage services to support inclusion and should be paid by the governmental contribution, in recognition of the policy interest in inclusion in setting service levels, rather than from fares. This is very much in line with conclusions reached in Thredbo Conference discussions about social inclusion benefits of public transport (Stanley and Ljungberg, 2017). One major benefit of the IPART approach is that it highlights such nuances. Agglomeration benefits are excluded by IPART because, while their existence is recognized, quantification of the externality component is seen as difficult and the relationship to fare changes tenuous (IPART, 2014). The point here is that the wider economic impacts as

458  Handbook on transport pricing and financing

a class of potential benefits for public transport improvements are dominated by productivity gains linked to agglomeration (improved accessibility), but the link with fare changes, in contrast to new infrastructure, is tenuous and minimal. Indeed, even with major investments in public transport such as a new rail line, Australia struggles beyond a 17% markup for wider economic benefits (Hensher et al., 2012). More recently, IPART has been involved in fare adjustments to lift overall PT cost-recovery rates from the current level of around 25%, fares having risen more slowly than service costs in recent years, while patronage had been growing strongly pre-COVID. Their main focus has been on altering peak/off-peak fare balances and better aligning fares across modes, adjusting fare concessions and adding quantity discounts. IPART (2020) concluded that small changes to fares would not make much of a difference to the level of community benefits, but large increases in fares could result in worsening congestion (see Hensher and Ho (2020) for evidence on direct and cross-fare elasticities for various trip segments and journey lengths). The general approach adopted by IPART is useful for transparently illustrating how net external benefits and government-assessed social benefits can be taken into account, via fare concessions and a government payment in the fare-setting process, to net out the proportion of costs that should be sought from users. Research should be undertaken on the scale of the external costs of social inclusion, which are relevant to decisions about base service levels, and the funding of these services. Infrastructure Victoria (IV) has been a leader in proposing (politically unpopular) pricing reform. In presenting its Victoria’s 30 Year Infrastructure Strategy to the Victoria Government, IV (2016) made transport road pricing reform one of its three main policy proposals. Its focus on transport pricing was a useful extension to the usual Australian concentration on road pricing. Infrastructure Victoria (2020a) has extended this work, developing a proposal for how transport network pricing reform might operate in Melbourne, subject to a (rather arbitrary) constraint of revenue neutrality. The proposed arrangements include distance-based charging on all modes, demand management tolls on new freeways, bridges, and tunnels, a CBD cordon charge and dynamic pricing, incorporating cost-reflective charges (including social marginal costs of congestion, pollution, and road trauma). Subsidies and fare concessions are included for more vulnerable and disadvantaged groups in one version, to deliver what IV sees as a fair solution, in which up to 85% of people could pay less. Ease of implementation is supported by primarily applying the pricing model at the system scale, rather than prices varying by link, time of day, etc. For those paying more because they travel longer distances, travel speeds are predicted to improve and reliability increase. IV has proposed that the Victorian Government implement trial programs to test its recommendations. Proposed PT fares vary by mode and by peak/off-peak, reflecting varying costs of service provision. Modeling suggests that, while road pricing increases the use of all PT modes, relative bus patronage gains are largest, particularly in the off-peak, stimulated by lower bus fares. Buses mainly service middle and outer suburbs, where income levels tend to be lower. Cordon pricing is projected to lead to a 40–50% reduction in road travel within the cordon, with inner-urban residents most affected. These people have the largest range of PT and active travel options available. The more extensive analysis of the proposed PT fare changes in Infrastructure Victoria (2020b) suggests benefits of $520 million (including road congestion cost savings, reduced PT peak crowding and lower GHG emissions), little or no change in cost-recovery levels and fare revenues, with 71% of PT users paying less, which are powerful supporting arguments.

Transport pricing and financing in Oceania  459

A weakness of the IV approach is that, while it recognizes the importance of marginal social costs for transport network pricing, it has not provided any estimates of such costs nor sought to optimize price signals across modes (IV, 2020a). IV (2020b) goes further with external costs in terms of setting PT fare levels, following the IPART approach, but CIE (2020) makes it clear that there is no quantification of agglomeration economies or treatment of social inclusion benefits from PT service provision, which is more explicitly addressed by the IPART approach (see Section 24.2.2 above). The primary Infrastructure Victoria intention seems to be to develop a transport network pricing system that might be politically acceptable and will nudge mode shares in directions that support more sustainable travel choices, with refinement deferred to the demonstration study stage. 24.2.3 Pricing of What in Land Transport? The dominant feature of Australia’s transport landscape in recent years has been the growth in capital spending. For example, NSW has had rapid growth in its infrastructure spending, with the annual average growth rate for the four-year budget period from 2021–2022 being nearly double that from 2013–2014 to 2016–2017 (New South Wales Government, 2021). Annual average capital expenditure by the Victoria State Government over the four-year budget period from 2021–2022 is growing even faster and will be more than four times the ten-year average to 2014–2015 (Victoria Government, 2021), with transport the major component. Transport accounts for two-thirds of the NSW capital budget and urban transport spending has been a critical focus of infrastructure spending growth. Nationally, the total value of road and rail projects being built across Australia exceeded $120 billion for the first time in March 2020, having fluctuated around $40–60 billion between 2007 and 2016 (Terrill, 2021). Terrill notes that most of the work is now being done on projects of $1 billion or more, with the average project size having doubled in the last decade. Such spending programs have been partly about overcoming backlogs associated with rapid population growth. Melbourne, for example, added 475,000 (+12%) to its population size between the 2011 and 2016 census dates, and Sydney added 430,000 (+10%). These are high population growth rates for cities of 4–5 million population with a high level of economic development. However, the large and growing transport infrastructure expenditures in Australia are frequently not supported by publicly available economic assessments of alternatives, including assessments that explore different ways of achieving intended outcomes. In Victoria’s case, this concern has recently been highlighted by the state’s Auditor-General, who found that: The absence of a transport plan as required by the [Transport Integration] Act, during a decade of unprecedented investment in transport infrastructure, creates risks of missed opportunities to sequence and optimise the benefits of these investments to best meet Victoria's transport needs. (VAGO, 2021, p. 1)

The Auditor-General could also have added that there is little or no consideration given to the role that charging of users and other beneficiaries should play in funding such massive growth in the transport asset base and in influencing the size of that base. Funding is clearly not a constraint on transport capital spending at present, with Australia (for example) having vast levels of funding available from its compulsory superannuation program and governments showing a willingness to borrow heavily. However, funding for public transport service

460  Handbook on transport pricing and financing

additions is not nearly so readily available, raising questions of distorted resource allocation in the sector, as between capital and services. Better connecting transport policy and planning, transport expenditures (including the transport asset base), pricing/charging models and sectoral governance is becoming increasingly urgent, an issue that has been regularly raised in pricing reviews (e.g., Harper et al., 2015).

24.3 NEW ZEALAND TRANSPORT PRICING Unlike Australia, New Zealand (NZ) has an explicit road user charging system (NZTA, 2020, p. 3): The cost of using New Zealand’s roads is recovered from road users via levies in the price of some fuels or through road user charges (RUC) … The revenue collected from road user charges is dedicated to the National Land Transport Fund.

Most transport funding in New Zealand is sourced from its National Land Transport Fund, providing a charging-expenditure connection that is lacking in Australia. Of the $NZ19.95 billion provided for in the National Land Transport Program over the four years between 2018 and 2021, some $NZ13.01 billion was sourced from the National Land Transport Fund. The main revenue flows into that fund are fuel excise duty on petrol vehicles ($NZ6.61 billion), road user charges on diesel vehicles, which is mainly heavy vehicles ($NZ5.07 billion) and motor vehicle license and registration charges ($NZ685 million). Within the National Land Transport Fund, the current NZ National Government has increased the expenditure share going to public transport and reduced that going to roads, with state highway improvement and maintenance allocations falling in current dollar terms, compared to the 2015–2018 period. This has increased the focus on the issue of road funding reform in New Zealand (Carvalho, 2019). As in Australia, however, this has been more productive of paper than of pricing reform. NZ’s most notable transport pricing initiative is its road user charge (RUC), which extends back to 1977. All diesel-powered vehicles and other vehicles powered by a fuel not taxed at source, must pay RUC. Any liable vehicle must be fitted with a distance recorder and, if the vehicle is > 3.5 tonnes, must be fitted with an approved hubodometer. The user charge is a function of distance traveled and vehicle weight, where the latter is the lesser of the vehicle’s Gross Vehicle Mass or maximum allowable mass, with separate charges applying to unpowered vehicles (e.g., trailers), overweight specialist vehicles and some other categories of vehicle. Distance licenses are purchased in units of 1000 km and displayed on the windscreen. This charging model has many similarities to the Australian heavy vehicle charging system but is more finely tuned with respect to its distance component but perhaps less so for mass, which is based on potential carrying capacity rather than on typical average loaded mass, as in the Australian case. The explicit recognition of excise payments for petrol vehicles as a road use charge and hypothecation of revenues raised therefrom to the National Land Transport Fund is also notable, compared to Australia which recognizes no such link. NZ and Australia both recognize shortcomings of heavy vehicle charging methods that depend on averaging processes, across road types and vehicle classes, and are examining ways in which refinements can be introduced. For example, NZ transport consultant David Greig expects that, as truck telematics improve, the NZ charging system will make use of

Transport pricing and financing in Oceania  461

these advances to incorporate charges based on actual weights and distances traveled. This is likely in Australia too. In both countries, however, it is the recovery of road construction, maintenance, and related costs from heavy vehicles that are the main focus of attention, rather than reforming charging regimes to incorporate charges for external costs such as congestion, air pollution, greenhouse gas emissions and noise across the entire vehicle fleet. Incorporation of external costs remains largely a debate among the academic community and road user member associations, albeit with increasing interest from governmental infrastructure advisory bodies and a growing air of inevitability as fuel excise collections come under increasing pressure. Introducing time and place into charging systems is getting closer!

24.4 FINANCING 24.4.1 Financing of Transport Investments At an institutional jurisdictional level, transport infrastructure in Australia falls into the domains of particular groups as follows: ●

●

●

●

State/territory governments: for enabling infrastructure for a public good (via consolidated revenue, often supported by Commonwealth Government funding?7) and increasingly delivered under a public–private partnership, or PPP8 Local governments: for enabling ‘local’ infrastructure delivering a public good (usually funded via property rates or grants from higher government) Landowners/developers whose proposed development gives rise to – or benefits from: ● Uplift in land value – and increased future development prospects as a result of the provision of new infrastructure capacity ● Related state/territory and local infrastructure needs (via state development contributions and local infrastructure contributions) ● An impact that requires mitigation (via conditions of consent, offsets and development contributions) Other users via user charging (via tolls, parking fees, access fees etc.).

24.4.2 Private–Public Partnerships and Other Approaches In terms of financing transport infrastructure, Australia has led the world in many respects in developing and refining the public–private partnership (PPP) model, especially for toll roads and, in more recent years, across all modes of transport and associated infrastructure. In the PPP model of public infrastructure delivery in Australia, the government calls for tenders for a contract for a single infrastructure project. These contracts commonly take the form of a Design–Build–Finance–Maintain–Operate (DBFMO) contract. The contract gives the successful private consortium responsibility for all aspects of project financing, delivery, and operation for periods often spanning decades, and typically up to 30 years for toll-road concessions (Hensher, 2018). The contract sets out how the consortium receives revenue: either from the government in the form of periodic ‘availability payments’, and/or direct from users (e.g., as tolls). The latter case is associated with demand risks, with patronage risk the greatest of all risks. The expected efficiency gains in the PPP approach are primarily

462  Handbook on transport pricing and financing

determined at a single point in time, by competition between the bidders for the contract (ex-ante). The allure of PPPs has been captured by the discipline of project finance, in that PPPs force a project to service any financial debt from the revenue streams derived from the project itself without recourse to public funding (cf., Debande, 2002; Li et al., 2005b).9 One apparent benefit of transferring financial risk to the private sector is that risks are subject to the scrutiny of commercial practice and extensive due diligence, related to the quantification and allocation of risks that private-sector risk-takers carry out on projects (Corner, 2006).10 In Australia, PPPs have always been considered to be on-balance sheet, with state treasuries allocating capital to fund the capital recovery element of future availability payments. The prime motivator has been the pursuit of value for money, as driven by optimizing competitive tension between bidders and ultimately tested against a public sector comparator. State treasuries increasingly stress that PPPs should be used as a deliverer of infrastructure services, with the benefits of fully funded whole-of-life maintenance, innovative solutions, improved risk transfer, and introduction of financial discipline from private financiers. Central to the philosophical commitment to PPPs is identifying and structuring financial risk, defined as the variability in returns that an asset is expected to earn. It is typically affected by factors such as market confidence, public perceptions, consumer attributes, environmental threats, and perceptions of misconduct (Asenova and Beck, 2003; Beck and Hensher, 2015). Australia has supported the private sector in managing risk, since private finance at risk will harness the private sector’s risk management skills; and finance cost is the most expensive item, such that private consortia are motivated to find ways to drive cost down. The approach prescribed by Infrastructure Australia requires discount rates to reflect systematic risk. An underlying risk-free interest rate (reflecting government bond rates) is the benchmark to which is added a margin for risk. So, for example, in the case of the Melbourne Metro PPP (a major underground rail link currently under construction), the discount rate to determine the public sector comparator benchmark was 3.09%, while the discount rate for the private sector was 5.7%. This means, in effect, that the increased cost of private financing will result in higher nominal payments over the duration of the contract, but that this increased cost reflects an adjustment for pricing risk in financial markets. The decision rule to enter into a concession depends on whether the project yields a positive risk-adjusted net present value. This condition is contingent on the degree to which commercial risks can be mitigated contractually upfront. The private sector has access to a wide range of financial products in the international market, and these resources have facilitated the formulation of the best financial packages, with the benefit that the capital market has on offer various sophisticated financial instruments, such as infrastructure bonds, stapled securities, fixed-rate loans, mezzanine loans, hedging, and insurance to cope with financial risk. Some of these financing instruments were developed by Australian businesses, with Macquarie Bank being the best example (see Hahn and Mack, 2005). States in Australia do not favor co-investment by way of government equity, preferring public sector capital to be contributed as a non-repayable funding source, designed to reduce future availability payments by paying down expensive private-sector capital. For example, under the $6 billion Melbourne Metro PPP, the Victoria Government is set to make a $2.5 billion construction contribution, followed by a $1.5 billion payment upon provisional acceptance, reducing the private financing to be repaid through the availability payments11 to $2 billion. The use of the capital contribution has helped to deliver improved value for money by

Transport pricing and financing in Oceania  463

mitigating the higher cost of private finance, while still ensuring the risk allocation and payment mechanism retains strong financial incentives to perform. Equity gain share has not been included as a standard provision of Australian PPP contracts. Initial equity investors in recent Australian PPPs, who have opted to sell down postconstruction, have been able to realize substantial capital gains as a reward for overcoming greenfield project risks. Sharing of gains in standard Australian PPP contracts to date relate to debt refinancing on more favorable terms and upside revenue sharing on user-pays infrastructure such as toll roads. The manner and form of the risk allocation for a PPP project are the key drivers of the financial and contractual structure of the project (Arndt, 2000, p. 58). A rule of thumb is that private equity normally bears the risks that cannot be, or are too costly to be, mitigated, because equity has greater risk tolerance – it shares a project’s upside gains, a benefit that is not open to debt financiers. The logic entails that lenders are more conservative and thus require a much narrower band for risk errors, particularly so in new roads. This requirement inevitably drives up the cost of finance, and hence equity is preferred. Asenova and Beck (2003) note that finance companies prefer that those risks that are difficult to mitigate remain with the consortia, to be supported by equity rather than debt. The Australian public sector prefers a proponent with a strong balance sheet, who can lower the cost of capital as well as sustain an investment in the long haul. But despite this, the private sector is wary of government’s approach to evaluating private proposals, in which the focus is only attended on capital costs, without giving sufficient attention to life-cycle cost savings. Fortunately, this has not been a major issue in Australia to date. Such an approach pressures the private sector not to price the risk premium into project cost, and may threaten the project’s long-term financial viability. Despite the financial turmoil with the Cross City Tunnel (CCT) and the Lane Cove Tunnel (LCT) in Sydney (cf. Chung, 2008; Chung and Hensher, 2015), market participants remain optimistic about the future of PPP, especially in the toll-road context. They are all cognizant of the fact that motorists value the comfort of driving in private cars, and hence the demand for tollways is likely to remain strong. Further, toll-road investment has had a strong appeal to superannuation fund managers, because it offers investment opportunities that have a similar term to maturity (Malone, 2005). With the concept of user-pays starting to gain greater acceptance, if risk allocation is managed equitably, there should be a growing market for PPP tollways. There are, however, other practitioners and advisers to government and investors more broadly who offered a more pessimistic view of these possibilities.12 They suggest that there is no evidence at present that private financiers are willing to take on the greenfield patronage risk. The last deal done on this basis was Brisbane Airport Link/Brisconnections, which failed, following the failures of Clem 7, Lane Cove Tunnel, and Cross City Tunnel. The new North East Link (NEL) project in Victoria is an availability PPP. WestConnex was initially developed as a government SPV, Sydney Motorway Corporation, later partly sold to Transurban; some 95% of toll revenue on WestConnex is regarded as brownfield. The NSW Government is now proceeding with M6 and Western Harbour Tunnel but both are wholly government financed and are not PPPs. The WestConnex is a ‘Sale of Business’ model and it is not clear why government used it and why the Victoria Government opted not to use it on NEL, preferring the availability of PPP with intent for later toll revenue securitization. This is the major current issue. The outlook for use of toll-road PPPs for greenfield projects is thus gloomy. In December 2020 the

464  Handbook on transport pricing and financing

NSW Government decided to self-deliver the Western Harbour Tunnel, which is a major blow to PPP activity. Value capture (VC) is increasingly mentioned in the literature on financing and funding infrastructure (see Lee and Locke, 2020) and is worthy of a brief comment. It includes (i) passive value capture – where government secures increased revenues from an infrastructure project without taking any further action. For example, where rail projects drive increases in property values along the rail corridor or near stations, this will increase government revenues from income tax, stamp duty, and capital gains tax; and (ii) active value capture – where it is planned that additional revenue will be generated from the main project. For example, if a new faster rail line is built by the government or a company, and it also owns property in the corridor it can directly benefit from increased property values. Alternatively, a government can sell development rights around a new rail station or implement a new tax or levy linked to the precinct. Although value capture is mainly a funding issue, it has subtle links to financing. If the public sector was to make a major commitment to value sharing, with government(s) taking a much higher share of value increase from major infrastructure projects than happens at present by measures that were not developed for value capture purposes, it would increase the beneficiary pays revenue streams captured by government (funding) and hence potentially make government upfront financing an easier option to sell. Using VC as an instrument to achieve this, however, has many challenges including a sense that existing taxes cover much of the value uplift (although not hypothecated). If VC has merit as a way of changing the financing model, the most sensible message is that before making any announcement of an infrastructure project, that governments value the land on which they wish to make a charge (based on some knowledge of the value uplift unambiguously linked to the project), since after the announcement it is too late. After almost 30 years of market involvement in financing transport infrastructure, and especially PPPs, Australia has had many ‘successes’ (e.g., motorway tollway projects in Sydney, Melbourne) and also some notable ‘failures’13 (e.g., Spencer Street Railway Station in Melbourne and the CCT in Sydney). Since the Global Financial Crisis in 2008, there have been questions asked associated with a need for a more evolved form of finance than one carried by sentiment and rampant self-interest. The current PPP market in Australia has recently seen the loss of equity in some high-profile projects (Forward and Aldi, 2017), the failure of revenue streams to meet debt servicing requirements, considerable write-down in the share price or market capitalization of some more recent economic infrastructure project companies, and lack of readily available funds from the private sector for infrastructure building. Forward and Aldis (2017) reviewed the experience with PPPs in Australia and concluded that a new view is required, given that many recent toll-road PPPs have experienced unachievable forecasting of revenue and a ‘ramp-up period’ longer than planned, resulting in equity value loss. Consortiums of investment banks and contractors have used inflated revenue estimates on toll-road projects to enhance their bids, and then have shifted the revenue risk to uninformed equity through IPOs. Much of this concern centers on financial risk and the suggestion that the sharing of risk must be revisited, with the next-generation PPPs restructured to introduce the providers of debt and equity funds into a risk-sharing arrangement similar to that used in the type of alliance arrangement.14 Specifically, the financiers need to become a partner with government for the long term, as well as the design, construction and service delivery partners. The risks can be shared, rather than allocated and the contract managed as

Transport pricing and financing in Oceania  465

an open book arrangement, where all parties to the alliance have access to cost revenue and contractual performance information. We recognize however that from a risk perspective and contract perspective this may end up being an ideal that will not be achieved. The approach is similar to ideas proposed by others whereby government appoints a service provider to lead the bid on behalf of, or with, government. Components of the bid are unbundled so that, through a competitive process, separate offers are sought for the financing, construction, and maintenance of the infrastructure. In this arrangement, the best components of the bid are selected ensuring we obtain the best value for money for each and all contributions to the successful bids.15 A model often discussed in this context is either the inverted bid model or the building Australia model promoted by IFM. There is a risk, however, that this might be self-serving and simply seeks to cement a role for foundation equity. Equity is not a homogeneous commodity and has to be priced and committed according to the specific contract framework and the specific contract counterparties.16 An alternative approach is a risk-sharing partnership, often referred to as ‘cap and collar’, designed to provide protection against uncertainty of just patronage revenue risk in which some of the principles of alliance contracting are adapted to all elements including financing, with the provision of infrastructure in a partnership between the private and public sector involving shared risks rather than allocated risks. Under the risk-sharing partnership approach, within the normal range of operations, gains and losses are shared between the parties within the partnership. Above an agreed internal rate of return, governments collect super-normal profits or share them with the private sector on an agreed basis. Losses below an agreed IRR are underwritten by government. This arrangement provides a strong incentive for government to actively review the patronage forecasts and other risks that may impact on the IRR. Similar arrangements are sometimes used in public transport service delivery contracts. The Regulatory Asset Base (RAB), more common in the energy and water sector, is a system of long-term tariff design aimed at encouraging investment in the expansion and modernization of infrastructure. Although the RAB model is contentious in that it has worked poorly in Australia for water and energy utilities, there has been interest in the transport sector in Australia. It is potentially an alternative to a PPP where a private (or corporatized state-owned) company or consortia act as the infrastructure manager: they own, invest in, and operate infrastructure assets. The infrastructure manager receives, through charges, revenue from users and/or subsidies to fund its operations and recoup investment costs. The infrastructure manager would behave much like a natural monopoly in the absence of regulation – setting prices too high in an attempt to earn ‘super-normal’ profits. An economic regulator is therefore established to provide efficiency incentives and to cap prices, revenue or rates or return received by the infrastructure manager to improve social welfare. Efficiency incentives aim to mimic the incentives that would be faced by the market if it were competitive. The efficiency gains in this case arise primarily through the interaction between the regulator and the infrastructure manager ‘during the contract’. Broadly then, its objective is to help balance the interests of consumers and investors. As a regulated utility model, it is generally viewed as a vehicle for the operation and maintenance of existing assets, but it is actually a lot more flexible than that and has applications for greenfield development as well. Privatesector organizations in the main in the RAB system provide a secure payback and return on investment sufficient to service loans and generate profits. The RAB is an accumulation of the value of investments that a service provider has made in its network, and includes assets of various useful lives. Most of these assets depreciate in value, although a small number (such

466  Handbook on transport pricing and financing

as easements and land) do not. This differs from the traditional PPP in that it delivers a low cost of capital compared with other private-sector procurement models, and is also far more flexible than a traditional PPP. It provides for long-term private-sector ownership while also allowing for reassessment and revision throughout the asset life cycle. One of the criticisms of RAB is the creeping risk of excessive capital expenditures designed to inefficiently increase the asset base on which returns are calculated. This can be circumvented and generally is, in practice, through controlling, through regulatory monitoring, for information asymmetry between a regulator and infrastructure owner. In the context of roads, the RAB approach can help deal with the loss of network control that arises when a PPP contracted business, such as Transurban controls, through a long-term concession (typically 30 years), the motorway network once the growing number of deals are signed. In Sydney and Melbourne, this increasingly is a significant amount of the motorway network and, with tolls preferred by such entities over serious road pricing reform, can be a blockage to government delivering future road pricing reform. This loss of network control makes transport policy and planning a bit like one hand clapping. The availability payments model would have assured greater control by government over revisionary price setting (even with shadow tolls) as traffic levels change, which opportunity is effectively denied under the current PPP Australian toll-road model. The RAB approach can also deal with the uncertainty of future changes in road pricing rather than locking in a fixed tolling regime for the long term.

24.5 CONCLUSIONS Oceania, and particularly Australia and New Zealand, have a long history of discussing road pricing reform but with little to show in terms of changed pricing systems, other than for some aspects of charging heavy vehicles for their road costs. However, the long history of discussion has created a context where change is increasingly likely. For example, no, or very few, major new high-capacity urban roads will be delivered that are not toll roads in the major capital cities along the Australian Eastern Seaboard. There is now a commitment from two Australian states, and an interest by a third, to implement distance-based charges on electric vehicles in 2021. Implementation of such charges for EVs highlights the dilemma of distorting the playing field for road user charging relative to ICEs, since EVs get no recognition for their cleaner environmental performance if externalities are not priced. This distortion will accelerate pressure for a broader approach to road use charging that includes key externalities. Heavy vehicle road use charging in New Zealand and Australia is likely to develop on a mass/distance base in the near future, as improved telematics facilitate such an approach. However, Hensher et al. (2021a) warn of the economic and environmental efficiency dangers of not treating both light and heavy vehicles in an integrated charging scheme. Also, the growing interest at the State Government level in efficient PT fare setting, in the context of externalities, increases the prospects that pricing reform will become more broadly focused on transport pricing, not just road pricing. We finish with three questions designed to challenge contributions to the debate on pricing and financing transport investments and which still remain at the core of any financing considerations:

Transport pricing and financing in Oceania  467

1. How can better alignment be achieved between transport policy and planning, transport expenditures, charging, and sectoral governance? Is there merit in taking long-term infrastructure projects out of the political sphere and creating an independent body that has bipartisan support, with a brief to propose (inter alia) priorities and funding arrangements? Could Infrastructure Australia’s role be broadened to perform this more integrative function? Can such a model be reconciled with political accountability for making major societal trade-off choices in a democracy? 2. Will this allow a more complete pipeline of projects to be developed, beyond those currently going through Infrastructure Australia processes, giving clear signals to the private sector about infrastructure priorities and future plans? 3. Is there scope, or even support, for governments to ‘bail out’ infrastructure projects that are failing in the short-term and, in the long term when they typically become profitable, the government is paid back for the ‘support’ it has provided, taking a slice of the profits as a reward for intervention? The alternative is a hybrid model that can be developed to ensure that the private sector can sensibly manage the risks it is taking on. Patronage risk is the biggest uncertainty addressed. The uncertain risks of construction are now subject to the impact of collaborative contracting principles and it is a question of including provisional sums around a private-sector financial model.

ACKNOWLEDGMENTS We thank a referee and Alejandro Tirachini for comments on an earlier draft and Martin Locke for reviewing the chapter and providing detailed comments on PPPs.

NOTES 1.

The focus of the latter part of the chapter is on financing, in contrast to funding of, transport investments. Funding refers to the primary stream of money to deliver a need (i.e., to how infrastructure is paid for) with two main sources of funding for infrastructure being government investment or direct user charges; in contrast, financing is the means by which these funding streams are manipulated to make money available when needed, and more specifically the way in which debt and/or equity is raised for the delivery and operation of an infrastructure project. 2. In a recent paper Hensher et al. (2021b) have shown for the Greater Sydney Metropolitan area that six months before the COVID-19 pandemic began, there was a reduction of travel time costs from $10bn to $5b, which amounts to significant benefit gains in congested traffic. 3. The AAA (2017) policy paper was intended as a position statement for the promised (2016) federal inquiry into road user charging, which never happened. 4. See www​.innovationaus​.com​/victorian​-parliament​-passes​-evs​-tax/. It is expected to raise about $30 million over four years, with owners to pay an average of $300 per year. Electric vehicle owners will have to keep a log of their driving, with this information used to calculate how much extra they will have to pay when they renew their car registration. Photos of a vehicle’s odometer will be provided to the state government via an online portal. 5. It should be noted that this welfare maximizing fare setting formula differs from that set out by Proost (2018), since the latter ignores agglomeration economies and social inclusion benefits (and may also ignore Mohring network economies). 6. Agglomeration benefits may be less relevant for many settings in the future given that digital platforms now offer very efficient and proven ways for many (not all) meetings to take place and hence the physical connectivity is less relevant.

468  Handbook on transport pricing and financing

7. The Commonwealth government specifically funds railways, national highways and the new Sydney airport. 8. See Chapter 16 in this Handbook on PPPs 9. With occasional situations when developers go back asking for more because they bid too low – even in privately originated initiatives, like the Western Distributor in Melbourne, and when the politicians are too weak to resist. Lessons now learnt make this less likely. 10. The recently published UK National Audit Office report on the UK Private Finance Initiative (PFI) (HM Treasury, 2018) highlights a concern about the accounting treatment of UK PFI, indicating that the transactions are still categorized as off-balance sheet. Despite the tightening of financial reporting obligations with the latest accounting standards, specific tweaks to the PFI model were made to ensure off-balance sheet treatment persisted, in order to avoid a downgrading of the UK credit rating. Moreover, the future liability to pay availability payments under PFI is unfunded by central treasury, leaving departments to find the cash to repay the capital component in addition to the operating costs and stay within budget. 11. An availability PPP is generally used for social infrastructure projects that are fully or partially privately financed and where Government pays the private sector payments for service availability. This is effectively a way of transferring patronage risk from the equity investors to government. It is more common in public transport where their commercial proposition is typically absent. 12. Martin Locke (personal communication 8 December 2020) has offered this somewhat pessimistic view. 13. The meaning of ‘success’ and ‘failure’ is dependent on which stakeholder is being referred to. For example the CCT was a disaster for equity investors but a great success for users who saved travel time, as well as Transurban who subsequently purchased the CCT concession after it went into administration for a basement prices. 14. Note however that the PPP model based on the principles of project finance is inconsistent with the uncertain cost outcomes of alliancing with a preference for fixed price debt and capital. This implies government needs to be willing to contribute additional capital. 15. An idea suggested many years ago by Dr Alastair Stone, Chair of the Board of Advice of the Institute of Transport and Logistics Studies and separately, by Chris Selth, a member of that Board. 16. From a theoretical perspective, Hensher and Johnstone (2021) set out a coherent economic decision model under which investors, sometimes in risk sharing partnerships, can rationally opt to firstly build and then possibly sell, or partly sell, the completed project. A remarkable conclusion is that investors are more attracted to buying equity in the completed risky infrastructure when its perceived operating risk is higher. That result occurs because the endogenous equity price decreases with higher risk at a rate that makes acquisition more attractive to investors when the financial risk is higher. By assuming a rationally priced infrastructure equity market, Hensher and Johnstone (2021) set out an internally consistent theoretical model that explains infrastructure investment at all stages, starting from construction and culminating in sale or floatation of shares on the open market. This model should be of interest to institutional investors but more so to governments and policy makers who wish to gain a deeper understanding of ‘how’ and ‘why’ investment in toll roads, airports and other major infrastructure is undertaken, and how it can be more productive.

REFERENCES Arndt, R. H. (1998). Risk allocation in the Melbourne City Link project. Journal of Project Finance, 4(3), 11e24. Arndt, R. H. (2000). Getting a fair deal: Efficient risk allocation in the private provision of infrastructure. PhD thesis, The University of Melbourne. Asenova, D. and Beck, M. (2003). The UK financial sector and risk management in PFI projects: A survey. Public Money and Management, 23(3), 195–202. Australian Automobile Association (2017). AAA Policy Position on Transport Market Reform. www​. aaa​.asn​.au ​/wp ​-content ​/uploads​/2018​/03​/AAA​-policy​-position ​_Transport​-market​-reform ​_ Jul​-2017​. pdf. Viewed 28 September 2020.

Transport pricing and financing in Oceania  469

Beck, M. and Hensher, D. A. (2015) Finding long-term solutions to financing 21st century infrastructure needs – a think piece. Roads and Transport Research, 24(3), 57–61. Bureau of Infrastructure Transport and Regional Economics (2016). ‘Traffic and congestion cost trends for Australian capital cities’. Information Sheet 74, Commonwealth of Australia. https://bitre​.gov​.au​/ publications​/2015​/files​/is​_074​.pdf. Viewed 28 September 2020. Chung, D. (2008). Private provision of motorways: Tolling our way into the future. In Seminar paper presented to School of Accounting, Economics and Finance of Deakin University, 12 September 2008. Chung, D. and Hensher, D. A. (2015). Risk management in Public Private Partnerships: The case of M4 motorway in Australia. Paper presented at the 13th International Conference on Competition and Ownership of Land Passenger Transport (Thredbo 13), Oxford, September 15–19 2013. Australian Accounting Review, 72 (25), 13–27. Chung, D. and Hensher, D. A. (2016). Risk sharing in Public-Private-Partnerships: A contractual economics perspective. In: Michiel C. J. Bliemer, Corinne Mulley and Claudine Moutou (eds.), Handbook on Transport and Urban Planning in the Developed World, Chapter 14, pp. 254–273. UK: Edward Elgar Publishing. Davies, P. (2017). ‘Transport reform: The national opportunity in front of us’. Speech to the Roads Australia Summit 2017. http://inf​rast​r uct​urea​ustralia​.gov​.au​/news​-media​/speeches​-presentations​/ transport​-reform​-opportunity​.aspx. Viewed 6 June 2017. Debande, O. (2002). Private financing of transport infrastructure. Journal of Transport Economics and Policy, 36(3), 355–387. Forward, P. and Aldis, R. (2017). Towards a New Public Private Partnership Model. Sydney: Evans and Peck. Government of South Australia (2020). State Budget 2020-21. Budget Statement: Paper Number 3, Adelaide: Author. Available at www​.treasury​.sa​.gov​.au/_ ​_data ​/assets​/pdf​_file​/0004​/311458​/2020 ​-21​Budget​-Statement​-non​-laid​-Web​.pdf. Hahn, N. and Mack, T. (2005). A banking perspective on transport. In: Button, K. and Hensher, D.A. (eds.), Handbook of Transport Strategy, Policy and Institutions, Chapter 18, pp. 299–310. Oxford: Elsevier. Harper, I., Anderson, P., McLusky, S. and O’Brien, M. (2015). Competition Policy Review Final Report, March 2015. Canberra: Commonwealth of Australia. http://com​peti​tion​poli​cyreview​.gov​.au​/files​/ 2015​/03​/Competition​-policy​-review​-report​_online​.pdf. Viewed 6 June 2017. Hensher, D. A. (2018). Toll roads – a view after 25 years. Transport Reviews, 38(1), 1–5. doi: 10.1080/01441647.2017.1330850. Hensher, D. A. (2019). Editorial: Road pricing reform – another attempt at getting started! Case Studies on Transport Policy, online June 2018, 7(4) December, 677–678. https://protect​-au​.mimecast​.com​/s​/ FWo​NCoV​zGQi​pglO​0h19bMT​?domain​=doi​.org Hensher, D. A. (2020). Electric cars – they may in time increase car use without effective road pricing reform and risk lifecycle carbon emission increases, Transport Reviews Editorial Series, 40(3), 265– 266. doi: 10.1080/01441647.2020.1709273. Hensher, D. A. and Chung, D. (2011). Road Infrastructure and Institutional Reform - Tolling and Pricing. In: Finger, M. and Künneke, R. (eds.), International Handbook of Network Industries: The Liberalisation of Infrastructure, Chapter 15, pp. 252–268. UK: Edward Elgar Publishing. Hensher, D. A. and Ho, C. (2020). Obtaining direct and cross fare elasticities using Opal data in Sydney, Australia, (release permission granted from IPART, 3 February 2020). Journal of Transport Economics and Policy, 54(4), 1–27. Hensher, D. A. and Johnstone, D. (2021). The rational economic attraction of risk in infrastructure investment. Submitted to Transportation Research Part B, 22 March 2021. Hensher, D. A. and Mulley, C. (2014). Complementing distance based charges with discounted registration fees in the reform of road user charges: The impact for motorists and government revenue. Transportation, 41(4), 697–715. Hensher, D. A., Truong, T. P., Mulley, C. and Ellison, R. (2012). Assessing the wider economy impacts of transport infrastructure investment with an illustrative application to the North-West Rail Link project in Sydney, Australia. Journal of Transport Geography, 24, 292–305.

470  Handbook on transport pricing and financing

Hensher, D. A., Wei, E., Beck, M. J. and Balbontin, C. (2021b). The impact of COVID-19 on the time and monetary cost outlays for commuting - the case of the Greater Sydney Metropolitan Area after three months of restrictions. Transport Policy, 101, 71–80. Hensher, D. A., Wei, E., Liu, W., Ho, L. and Ho, C. Q. (2021). Development of a practical aggregate spatial road freight modal demand model system for truck and commodity movements in Australia with an application of a distance-based charging regime. Transportation, 12 July 2021. Hensher, D. A., Wei, E. and Liu, W. (2021a). Battery Electric Vehicles in Cities: Measurement of some impacts on traffic and government revenue recovery. Journal of Transport Geography, 94, 103121. HM Treasury, National Audit Office, PFI and PF2, January 2018. Independent Pricing and Regulatory Tribunal (2013). Maximum fares for metropolitan and outer metropolitan buses from January 2014. Author: Sydney. www​.ipart​.nsw​.gov​.au​/files​/sharedassets​/ website​/trimholdingbay​/final​_ report​-maximum​_fares​_for​_ metropolitan​_ and​_outer​_ metropolitan​_ buses​_from​_ january​_2014​.pdf. Viewed 29 September 2020. Independent Pricing and Regulatory Tribunal (2014). Review of external benefits of public transport. Transport – draft report, December. Author: Sydney. Independent Pricing and Regulatory Tribunal (2020). Maximum Opal fares 2020-2024. Final Report, February. www​.ipart​.nsw​.gov​.au​/files​/sharedassets​/website​/shared​-files​/pricing​-reviews​-transport​services​-publications​-opal​-public​-transport​-services​-to​-june​-2024​/opal​-public​-transport​-services​to​-june​-2024​-final​-report​-publications​/final​-report​-maximum​-opal​-fares​-2020 ​-2024​-february​-2020​. pdf. Viewed 30 September 2020. Independent Pricing and Regulatory Tribunal (n.d.). External benefits and costs: Draft information paper 8. www​.ipart​.nsw​.gov​.au​/files​/sharedassets​/website​/shared​-files​/pricing​-reviews​-transport​services​-publications​-review​-of​-public​-transport​-fares​-in​-sydney​-from​-july​-2016​/external​_benefits​_ and​_costs_-​_public​_transport​_fares​_draft​_ report_-​_ip​_8​.pdf. Viewed 29 September 2020. Infrastructure Partnerships Australia (2018). Road use charging for electric vehicles. Sydney: Author. https://infrastructure​.org​.au​/wp​-content​/uploads​/2019​/11​/Road​-User​-Charging​-for​-Electric​-vehicles​-1​. pdf. Viewed 5th November 2020. Infrastructure Victoria (2016). Victoria’s 30 Year Infrastructure Strategy, December 2016, Melbourne: Infrastructure Victoria. www​.inf​rast​r uct​urev​ictoria​.com​.au​/sites​/default​/files​/images​/ IV​%2030​ %20Year​%20Strategy​%20WEB​%20V2​.PDF. Viewed 5 June 2017. Infrastructure Victoria (2020a). Good move: Fixing transport congestion, March 2020. Melbourne: Infrastructure Victoria. Available at www​.inf​rast​r uct​urev​ictoria​.com​.au​/wp​-content​/uploads​/2020​/ 03​/Good​-Move​-fixing​-transport​-congestion​-Infrastructure​-Victoria​.pdf. Infrastructure Victoria (2020b). Fare move: Better public transport fares for Melbourne, September 2020. Melbourne: Infrastructure Victoria. Available at: www​.inf​rast​r uct​urev​ictoria​.com​.au​/wp​content ​/uploads​/2020​/09​/ Fair​-Move​-Better​-Public​-Transport​-Fares​-for​-Melbourne​-FINAL​.pdf. Lee, C. L. and Locke, M. (2020). The effectiveness of passive land value capture mechanisms in funding infrastructure. Journal of Property Investment & Finance. doi: 10.1108/JPIF-07-2020-0084 Li, B., Akintoye, A., Edwards, P. and Hardcastle, C. (2005a). The allocation of risk in PPP/PFI construction projects in the UK. International Journal of Project Management, 23(1), 25–35. Li, B., Akintoye, A., Edwards, P. J. and Hardcastle, C. (2005b). Perceptions of positive and negative factors influencing the attractiveness of PPP/PFI procurement for construction projects in the UK: Findings from a questionnaire survey. Engineering Construction and Architectural Management, 12(2), 125–148. Malone, N. (2005). The evolution of private financing of government infrastructure in Australia – 2005 and beyond. Australian Economic Review, 38(4), 420–430. Melbourne Metro Rail Authority/Victoria State Government, Metro Tunnel Project Summary, February 2018. National Transport Commission (2018). PAYGO model user manual – version 2.2. Available at www​.ntc​ .gov​.au​/sites​/default​/files​/assets​/files​/ PAYGO​%20model​%20user​%20manual​%20version​%202​.2​.pdf. National Transport Commission (2019). Heavy vehicle charges consultation report - December 2019. Melbourne: Author. Available at www​.ntc​.gov​.au​/sites​/default​/files​/assets​/files​/ Heavy​%20Vehicle​ %20Charges​%20Consultation​%20Paper​.pdf. National Transport Commission (2021). Heavy vehicle charges determination: consultation Regulation Impact Statement, June 2021. Melbourne: Author. Available at Heavy Vehicle Charges Determination 2021 - Regulation Impact Statement​.p​df (ntc​.gov​​.au).

Transport pricing and financing in Oceania  471

New South Wales Government (2021). Budget paper No. 3 - Infrastructure Statement 2021-22, Sydney: Author. Budget Paper No.3 - Infrastructure Statement - Budget 2021-22 (nsw​.gov​​.au). New Zealand Transport Agency (2020). Road user charges. Wellington: NZ Government. Available at www​.nzta​.govt​.nz​/assets​/resources​/road​-user​-charges​/docs​/road​-user​-charges​-handbook​.pdf. Parry, I. W. H. and Small, K. (2005). Does Britain or the United States have the right gasoline tax. American Economic Review 95, 1276–1289. Proost, S. A. (2018). Reforming private and public urban transport pricing, International Transport Forum Discussion Papers. Paris: OECD Publishing. Terrill, M. (2021). How to get better bang for transport bucks. Submission to House of Representatives Standing Committee on Infrastructure, Transport and Cities Inquiry into Procurement Practices for Government-funded Infrastructure, Melbourne: Grattan Institute. How​-t​​o​-get​​-bett​​er​-ba​​ng​-fo​​r​-tra​​ nspor​​t​-buc​​ks​-s​u​​bmiss​​ion​.p​​df (grattan​.edu​​.au). The Centre for International Economics and Jacobs (2020). Estimating the social marginal cost of public transport in Victoria, June. Canberra: Centre for International Economics. Available at https://static1​. squarespace​.com ​/static​/5df ​9aa0​7864​2f94​3ece6a0b3​/t ​/5f7​1365​3fb5​87d1​1461f68c3​/1601255003545​/ Estimating​-the​-social​-marginal​_cost​-of​-public​-transport​-CIE​.pdf. Tillett, A. (2018). Deputy PM Michael McCormick shelves inquiry into road pricing, Australian Financial review, 5th Octobewww​.afr​.com​/politics​/deputy​-pm​-michael​-mccormack​-shelves​-inquiry​into​-road​-pricing​-20​181004​-h1688d. Viewed 28 September 2018. VAGO (Victorian Auditor General Office) (2021). Integrated Transport Planning. August 2021. Melbourne: Author. Integrated Transport Planning (audit​.vic​.gov​​.au). Victoria Government (2021). Budget paper 4 - State Capital program 2021-22. Melbourne: Author. Budget papers | Victorian Budget 21/22 | Victorian Budget.

Index

A-/B-Loan structure 334 accessibility 348–50 job 352 transport 359 accident externalities 422 accident-related road tolls 31 active modes 259 advisory services and analytics (ASA) 336 aeronautical services 213 Africa mode of transport in 364 road pricing 365 fuel levy 365–6 road pricing equity concerns 367–8 road tolling 366–7 rural roads in 375 transport financing 368–9 Chinese development finance 371–3 official development finance from multilateral banks 370–371 public–private partnerships 373–4 public-sector financing 369–70 transport infrastructure 364 transport infrastructure projects in 369 Africa Continental Free Trade Area (AfCFTA) 376 African Development Bank 368, 373–5 agglomeration 457–8, 467n6 agglomeration externality 75 air cargo firms 242 air freight 230, 240–242 air pollution 30, 330 air traffic 207, 208 airport 207 privatization of 211, 212 airport competition, pricing of 211–13 airport pricing considering environmental costs 220–221 literature 210 price-based demand-oriented measures 208–10 pricing of private airports, airport competition and capacity investments 211–13 regulation and non-aeronautical operations 213–14 second-best pricing 210–211 quantity-based demand management mechanisms 214–20

airport slots 214–20 allocative efficiency, Latin America 418 fuel taxes and prices 421–2 new challenges 423 transit fares 418–21 AMoD see automated mobility on-demand (AMoD) anonymous pricing scheme 161 appraisal defined as 295 and finance 297–301 see also investment appraisal area-based system 15–16 Arrow–Debreu model 191 Asia 380 public transit systems 385 public transport systems in 386 transport pricing, funding, and financing issues in 389 see also China; Singapore asset securitization 287 atomic users 31 Australian transport pricing heavy vehicle road use charging in 466 land transport 459–460 public transport pricing 456–9 road pricing and expenditure 452–6 Automated Driving Systems 252 automated mobility on-demand (AMoD) 256–8 automated vehicles (AVs) 100, 252, 253 studies regarding pricing 267–71 see also vehicle automation automation 5–6 average cost pricing, instruments based on 66–8 bathtub model 155–6 behavioral economics (BE) 444 Belgium, distance-based HGV charge 405, 413n2 Belt and Road Initiative (BRI) 371 benefit spill-over 52 benefit–cost ratios (BCRs) 298, 304 benevolent dictator 39, 44 blended finance 344–5 bonds 286–7 bottleneck model 28–30 basics of 153 second-best congestion pricing in 153–4 urban spatial structure and dynamic congestion pricing 154 472

Index 

braking behavior of travelers 153 Brazil, Vale Transporte program in 357, 428–9 BRI see Belt and Road Initiative (BRI) budget-neutral incentives 55–6 build, operate, and transfer (BOT) contracts 385 awarding contract 315–16 bundling 314 funding 314–15 public and private contracting 312–14 renegotiations 316–17 risk allocation 315 unsolicited proposals 317 Bulgaria, distance-based HGV charge 405, 413n3 bundling 314 bus rapid transit (BRT) systems 356, 375, 417, 419, 423, 424, 431n8 Business Rate Supplement (BRS) 300 calibration, of parameter values 84–5 California Department of Transportation 437 California Private Transportation Company 437 Canada transport funding in 437–8 transport pricing and funding in 436 cap 42 capacity choice in dynamic models 160–161 relationship between road pricing and 157 under second-best pricing 159–60 capacity investments, pricing of 211–13 capacity scarcity models 26 bottleneck model 28–30 dynamic models 27 peak-load pricing models 27–8 static models 26–7 CAPEX 317, 325 Capital Asset Pricing Model (CAPM) 284 carbon emissions 59, 63–8, 70 Carbon Offsetting and Reduction Scheme for International Aviation (CORSIA) 241 carbon pricing 243 carbon taxes 70 car-restriction policy 424 cars GHG emissions from 61 Pigouvian tax for 261 urban road pricing schemes reduction for 412 usage and ownership 63–5 cash-out policies 96–9 CAVs see connected and automated vehicles (CAVs) CBA see cost–benefit analyses (CBA) Central American Bank for Economic Integration (CABEI) 344 central business district (CBD) 74, 75, 78, 131, 150, 151

473

China high-speed rail systems 388–9 high-speed railway financing issue 380 idea of “maintaining road by road” in 385 on infrastructure development 371–3 state-owned enterprises 372 China Public–Private Partnerships Centre (CPPPC) 372 China Railway Corporation (CRC) 388 “China T-Union” card 386 Chinese development financing model 372 circulation taxes 64 citizen–candidate models 126 City Logistics 237 Clean Air for Europe Directive 68 Clean Technology Fund (CTF) 345 climate change 1–2, 186, 342 climate finance instruments 343 climate funds 344 coarse tolling 153 cognitive reflection 445 commercial banks 285–6 completion guarantees 282 computable general equilibrium (CGE) models 33 “Condorcet Winner” 125 congested road network 60–61 congestion 2, 26 nonprice regulations to tackle 423–5 rail track access charges 400 congestion costs 26, 114, 161–2 marginal vs. average 27 congestion externalities 46 optimal road pricing for 40–42 congestion pricing (CP) 30, 47–8, 68–9, 75, 255, 257–8, 381 franchising of private road infrastructure 46–7 heterogeneous preferences 47 in monocentric settings 74–5 networks 47 oligopolistic markets and congested facilities 47–8 in polycentric cities 76–7 price discrimination 48–9 by private operator 44 profit-maximizing capacity 46 congestion taxes 74 on travel modes, spatial structure and urban form 85–9 congestion tolling 75 connected and automated vehicles (CAVs) 252, 255, 256 contactless payment methods 5 container-centric yield management 239 contracts

474  Handbook on transport pricing and financing

BOT contracts 312–17 awarding contract 315–16 bundling 314 funding 314–15 public and private contracting 312–14 renegotiations 316–17 risk allocation 315 financing of PPPs leverage 317–18 project finance 317 risk and guarantees 318 traditional provision 318–19 valuation of 318 cordon pricing 150–151 corporate finance versus project finance 281–3 corporate loans 334 cost–benefit analysis (CBA) 187, 295, 305, 306 methodological developments in 295–7 practitioners of 299 cost-of-service pricing 234–5 cost recovery ratio, of public transport 137 costs logistic versus transport 232–3 of mandated parking 95–6 of public funds 183–5 wear and tear 399–400 COVID-19 pandemic 2, 5, 18, 20, 207 Cross City Tunnel (CCT) 463, 468n13 cross-price elasticities 183 cross-subsidies 358 cross-subsidization/cross-financing 289–90 CTF see Clean Technology Fund (CTF) debt 275, 280, 283 decarbonizing transport 342–3 decentralized planning 90 decision variables 171–3 decision-making process 304, 444 institutional constraints on 136 demand management, through pricing incentives 3–4 demand-oriented price-based measures 208 demand-side management levies (DSML) 367 demand-side subsidies 428 design and implementation of 429 targeting and distributional properties of 429 design 183–5 Design–Build–Finance–Maintain–Operate (DBFMO) contract 461 development policy financing (DPF) 333–4 Didi Chuxing 190 difference-in-differences (DiD) approach 412 direct fees 367 direct securitization 287 discounted cash flow (DCF) valuation 283, 284

distance-based charging schemes, for HGV 403, 405, 407–8 distance-based pricing 152 distance-based VMT tax 76 distributional impacts across locations 115–16 in terms of income 112–15 distributional tensions, transport policy 3–4 diverging incentives, with horizontal externalities 53–5 double dividend 66 double marginalization 48 DPF see development policy financing (DPF) Driver Support Systems 252 driver-based costs 237 driving restrictions 68–9 dual-till ROR 223n11 dynamic congestion pricing 152 accumulation-based approach 156–7 bottleneck model 153–4 hypercongestion, MFD, and bathtub model 155–6 Pareto-improving toll 154–5 Tradable Bottleneck Permit scheme 155 trip-based approach 157 dynamic models 27 dynamic pricing 446–9 econometric approach 399–400 economic appraisal 303–6 economic instruments 63–5 economic theory 107, 108, 134 evaluation of marginal tax reforms 110–111 optimal tax and investment rules 108–10 EcoPass scheme 163, 412 electric cars 117–18 electric vehicles (EVs) 64, 117, 118, 455–6 Electronic Road Pricing (ERP) system 152, 381 electronic toll collection systems 15 emerging markets and developing economies (EMDEs) 330, 338, 345 climate mitigation and transport decarbonization in 343 decarbonizing transport in 342 emission trading 232 Emission Trading System (ETS) 223n13, 238 emissions 30, 34n2 employer-paid parking 90, 96–9, 101n12 enforcement of traffic law, political determinants of 139 engineering, procurement, and construction (EPC) contract 335–6 environmental considerations 186 environmental externalities 30–31, 33, 51, 208, 422

Index 

environmental sustainability, transport financing 340 equi-marginal principle 64 equity 107, 275, 281, 283 and income disparity 185 road pricing concerns 367–8 “equity premium puzzle” 298 equity subsidies 355–6 equity-efficiency trade-off 110 Europe 394 charging for road use 401–2 EU legislation 402–3 HGV charges 403–9 urban road pricing schemes 409–12 rail track access charges 398–9 congestion and scarcity 400 mark-ups 400–401 wear and tear costs 399–400 transport infrastructure pricing policy 396–7 transport modes in 59, 60 European aviation strategy 241 European Commission 288, 394, 395, 397, 399, 409 European Court of Justice 394 European Emissions Trading Scheme 56 European rail policy 398 European Union (EU) 276 Cohesion Fund 348 growth of 394 infrastructure policy 350 European Union (EU) legislation 63 charging for use roads 402–3 expenditure competition 52 expenditure externality 51 export demand, resource balance and growth in 79 ex-post evaluations 338 of WBG urban transport financing 339 express lanes (ELs) 439, 440 external costs 13–14, 230, 453 of freight transportation 232 internalisation of 232–3 of maritime freight transport 243 externalities 24, 33, 162–3, 231, 326n4, 453 environmental 30–31, 33 fiscal/expenditure 353 general-equilibrium effects 417–18 E-ZPass 163 Fair Market Value (FMV) 98, 103n36 fares public transit systems 386–7 structure of 172 Federal Aid Highway Act 436, 437 Federal Aid Road Act 436

475

Federal Aviation Administration (FAA) 208 feebates 67–8 feeder–trunk scheme (FT) 175 financial crisis 287 financial evaluation, transport investment 283–5 financial options valuation 285 financial sustainability, transport financing 340 financing 1, 467n1 of HSR systems 380 of PPPs leverage 317–18 project finance 317 risk and guarantees 318 valuation of 318 financing entities, transport investment governments 278–9 multilateral institutions 279–80 private investors 279–80 state-owned enterprises 279 first-best pricing model, three-mode 260–262 first-best pricing policy 88 first-degree price discrimination 14, 48 fiscal federalism 185 fiscal/expenditure externalities 51, 353 fixed cost allocation 235 fixed-term contracts 321, 324 flat fare scheme 358 floor-to-area regulations 100n3 foreign direct investment (FDI) 330, 368 free-market equilibrium 85, 86 free-riding problem 69 “Freeway Revolt” movement 141n14 freight externalities 229 freight transport 229, 241 freight transport pricing 229, 244 fuel levy 365–6 fuel taxes 64, 65–6, 152, 277, 366, 418, 453 and prices 421–2 vs. vehicle tax 67 full cost recovery 396, 397, 413 funding source 467n1 BOT contracts 314–15 road pricing as 384–5 Gauteng Freeway Improvement Project 367 GBP see Green Bond Principles (GBP) GCF see Green Climate Fund (GCF) general purpose lanes (GPLs) 439, 441, 449 Germany air pollution costs 413n5 distance-based HGV charge 413n2 HGV charges 409 infrastructure manager as commercial body 398 Global Financial Crisis 464

476  Handbook on transport pricing and financing

Global Navigation Satellite System (GNSS) 381 global positioning system (GPS) 366 globalisation 230 government financing 274, 278–9 government guarantees 385 government-owned projects/services 276 grants 275, 280 Green Bond Principles (GBP) 343 green bonds 343–4 Green Climate Fund (GCF) 344 Green Paper 396 green tax 431n20 greenhouse gas (GHG) emissions 59, 220, 330 from road traffic 61 tax on 62 from transport 59, 60 gross domestic product (GDP) 369 guarantees financing of PPPs 318 government 385 MIGA 335–6 for private lenders 332–3 heavy goods vehicles (HGV) 230, 235, 236 charging for use roads 403–9 heterogeneous preferences, congestion pricing 47 high occupancy toll (HOT) lanes 113 high-occupancy vehicle (HOV) lanes 68, 151, 439 high-speed rail (HSR) systems, financing of 380, 388–9 Highway Trust Fund 436–7 Hong Kong Mass Transit Railway system in 385 MTR Corporation Limited in 388 horizontal environmental externality 52–3 horizontal expenditure externality 52 horizontal fiscal externality 51, 52 HOT lanes 151, 160, 165n2, 439 Hotelling-type model 212 household expenditures, on transportation 425–6 housing markets impacts on 163 role of 131–2 housing-supply sector 79 hub–spoke network 222 human-driven vehicle (HDV) 252 hypercongestion 155–6 IBRD see International Bank Reconstruction and Development (IBRD) ICA see Infrastructure Consortium for Africa (ICA) IFC see International Finance Corporation (IFC) IMF see International Monetary Fund (IMF)

income, distributional impacts in 112–15 income inequality 417 income taxation 98, 103n35 Independent Evaluation Group (IEG) 338, 345n1 Independent Pricing and Regulatory Tribunal (IPART) 456–8 indirect securitization 287 informal public transportation 374–5 institutional capacity in improving 375–6 Infrastructure Consortium for Africa (ICA) 369 infrastructure funds 286 infrastructure investment 135, 138, 140 infrastructure managers (IMs) 399 Infrastructure Partnerships Australia 455 infrastructure spending 134–7 Infrastructure Victoria (IV) 458, 459 inland water transport (IWT) 244 innovative financial sources, transport investment asset securitization 287 bonds 286–7 commercial banks 285–6 infrastructure funds 286 institutional development transport financing 340–341 Intellectual Property (IP) 317 intergenerational equity 107 intermodal substitution 171 intermodal transport 259 internal combustion engine vehicles (ICEVs) 117, 118 Internal Rate of Return (IRR) 284 International Bank Reconstruction and Development (IBRD) loans and guarantees 331–3 International Energy Agency (IEA) 59 International Finance Corporation (IFC) loans and equity investments 334–5 international financial institutions (IFIs) 370 International Monetary Fund (IMF) 369 International Panel on Climate Change (IPCC) 59 intra-generational equity 107 in-vehicle kilometres travelled (VKT) 257 InvestEU 288 investment appraisal 305–6 allocation of risk and optimism bias 301–2 methodological developments in 295–7 public acceptance and political influence in development strategy 302–3 economic appraisal and political influence 304–5 option generation and scheme selection 303–4 see also appraisal Investment Project Finance (IPF) 331

Index 

IPART see Independent Pricing and Regulatory Tribunal (IPART) IPF see Investment Project Finance (IPF) job accessibility, transport infrastructure and 352 Kampala–Entebbe toll road 373 Katy Freeway 442, 445 KBRUC system see kilometer-based road user charge (KBRUC) system kilometer-based road user charge (KBRUC) system 366 kilometer-based transport taxes 430 labor–leisure trade-offs 74 land-based public transport services 456 land transport 459–460 Land Transport Authority (LTA) 381 land value tax 299 land-use pattern equilibrium 78, 81 land-use/parking regime 91–2 Lane Cove Tunnel (LCT) 463 Latin America allocative efficiency 418 fuel taxes and prices 421–2 mechanisms to set and readjust transit fares 420–421 new challenges 423 transit fares 418–20 nonprice regulations to tackle congestion and pollution 423–5 pricing in 417 social/distributive issues 425–30 leapfrog development 80, 86, 102n19 legislative bargaining 127 liberalisation 239 license-plate-based traffic rationing schemes 384 life-cycle costs 311, 314, 317, 319, 324, 325 light rail transit (LRT) 259 light-duty vehicles (LDVs) 235, 237 line haul operators 241, 242 link-based pricing policy 256 link-specific prices 34 listed funds 286 loan syndication program 334 lobbying 126–7 London Congestion Charge (LCC) 163 low-emission zones (LEZ) 68, 69, 130, 133, 355 low-income travelers 100n2 macroscopic fundamental diagram (MFD) 155–6, 164, 165n6 Mainland China, road pricing 382–4 majority voting 125 managed lanes (MLs) 439

477

mandated parking, cost of 95–6 marginal accident costs, defined as 31 marginal cost pricing (MCP) 12–13, 55, 74, 76, 185, 255, 256, 396, 413 instruments based on 65–6 land-use effects of 76 principle 381 marginal costs of public funds 34n4 marginal external cost (mec) 40 marginal infrastructure cost 413n4 marginal revenues 48 marginal tax reforms, evaluation of 110–111 maritime freight pricing 242–4 maritime transportation 242 market equilibrium, aggregate model of ridesourcing 197–200 market mechanisms, for slot allocation 217 market-based rail freight pricing 240 mark-ups 400–401, 413 non-discriminatory 398 Mass Transit Railway (MTR) system 385 mass-transit projects 340 matching algorithms, types of 197–8 MDBs see multilateral development banks (MDBs) means-tested transit subsidies 430, 431n10 Melbourne Metro PPP 462 MFD model 156–7 MIGA see Multilateral Investment Guarantee Agency (MIGA) mileage (utilization) tax 66 minimum-parking requirements (MPRs) 91–96 urban form with and without 92–4 mobility-as-a-service (MaaS) 19, 140, 259 facilities 57 modal competition 182–3 modal versus multi-modal perspective 233 Mohring effect 108, 387, 418, 420 Mohring-Harwitz result/theorem 43, 47, 160, 161 monoline insurance market 291 Monte Carlo analysis 301 motor taxes 12 MPCF 183–4 multilateral development banks (MDBs) 289, 330, 341, 345 official development finance from 370–371 multilateral financial institutions, transport financing 275, 279–80 Multilateral Investment Guarantee Agency (MIGA) guarantees 335–6 multiple urban externalities 75, 76, 100, 101n9 National Highway Toll Standard 384 National Land Transport Fund 460

478  Handbook on transport pricing and financing

National Land Transport Program 460 national policy agendas, transport policy in 348–52 National Road Transport Commission 452 negative environmental externalities 30 negative externalities 24–5, 31, 230, 231, 257, 352, 374 NEPAD Infrastructure Project Preparation Facility (IPPF) 373 net present value (NPV) 284, 318, 389 ‘net zero’ transport sector 5 network exit function (NEF) 156 network traffic management, tradable travel credit scheme for 384 networks 32 congestion pricing 47 net-zero carbon emissions 413 New Economic Geography 348–9 New Zealand transport pricing distance-based user charge 422 financing of transport investments 461 heavy vehicle road use charging in 466 private–public partnerships and other approaches 461–6 transport funding in 460 noise 30–31 non-aeronautical operations 213–14 revenues 213, 214 services 213, 214 non-discriminatory mark-ups 398 non-excludability 44 non-honoring (NH) type guarantee 336 non-honoring of sovereign financial obligations (NHSFO) 335 non-linear pricing models 234 nonprice regulations 418 to tackle congestion and pollution 423–5 non-resident households 78 non-rivalness 44 non-surface parking regimes 92 nonviable rural services 428 non-visitor parking 78, 102n20 North America history of roadway pricing managed lanes/HOT lanes/express lanes 439 need for priced infrastructure 438 public–private partnerships 438 tolling to taxes 436–7 transport funding in Canada 437–8 North East Link (NEL) project 463 North Tarrant Express (NTE) freeway 442, 445 novel price-setting mechanisms 4–5 NTC 452, 453

official development finance (ODF) 364 from multilateral banks 370–371 off-street parking 91 oligopolistic markets, congestion pricing 47–8 O&M 317–18, 325 operators’ costs 171–3 optimal capacity choice 42–4 optimal charging mechanisms 401–2 optimal contracting 319 optimal pricing 141n1, 185 for congestion externalities 40–42 of public transport system 172 in static congestion model 26–7 strategies, of ride-sourcing market 201–3 optimal tax 108–10 optimal transit fares, and subsidies 182, 419 optimal transport systems 13–14 optimism bias 301–2 Organisation for Economic Co-operation and Development (OECD) 364 parallel loans 334 Pareto-improving toll 154–5 Paris Agreement 70, 344 parking availability 89–98 vehicle automation effect on 254–5 parking charges, at workplace 96–9 mandates 94–5 policy, political economy of 134 pricing 73, 89–98 subsidies, value of 98, 99 tariffs 134 partial credit guarantee 332 partial risk guarantee 332 peak-load pricing models 27–8, 420 peak-period transit fares 419 periodic surface maintenance of sealed roads 453 personal vehicles, policy instruments for congestion pricing and driving restrictions 68–9 economic instruments and car usage and ownership 63–5 information programmes 69 instruments based on average cost pricing 66–8 instruments based on marginal cost pricing 65–6 vehicle scrappage schemes 69–70 Pigou, contribution of 11–12 Pigouvian taxes 11, 13, 15, 62, 63, 64, 74, 76, 110, 128 for cars 261 planning obligations 300 policy instruments, for personal vehicles

Index 

congestion pricing and driving restrictions 68–9 economic instruments and car usage and ownership 63–5 information programmes 69 instruments based on average cost pricing 66–8 instruments based on marginal cost pricing 65–6 vehicle scrappage schemes 69–70 policy interactions, types of 51–3 political budget cycles 127–8 political decision-making, theoretical models of 125 citizen–candidate models 126 legislative bargaining 127 lobbying by special interests 126–7 majority voting 125 political budget cycles 127–8 probabilistic voting models 125–6 political economy analysis of 141n6 fiscal federalism and 185 of infrastructure provision 134–8 of pricing parking 133–4 social acceptance and 185–6 of transport decision-making 124 of transport policy 124 of transport pricing 128–33 political influence, in investment appraisal development strategy 302–3 economic appraisal 304–5 option generation and scheme selection 303–4 political risk insurance (PRI) 335 political unpalatability 455 “polluter pays principle” 62 pollution, nonprice regulations to tackle 423–5 port pricing 243–4 schemes of 243 positive externalities 24, 25 PPP schemes 275, 278, 280, 285, 291n1 see also public–private partnerships pre-COVID-19 air transport industry 241 pre-existing distortionary taxes 39, 66 pre-existing property taxes 101n7 preferred creditor status 334 pre-pandemic ACI forecasts 207 present-value-of-revenue (PVR) contracts 315 advantages of 324 price discrimination 14–15 congestion pricing 48–9 price-based demand-oriented measures 208–10 pricing of private airports, airport competition and capacity investments 211–13

479

regulation and non-aeronautical operations 213–14 second-best pricing 210–211 priced infrastructure, need for 438 pricing 55–6 to direct users 274, 276–7 in freight transport policy 230–235 fuel taxes and 421–2 impacts in developing economies 163–4 principles of 61 of private airports, airport competition and capacity investments 211–13 in public transport 186 studies regarding automated vehicles 267–71 of transport externalities 60–63 of transport modes, vehicle automation 255–9 of transport services 59 vehicle automation effect on 254–5 see also transport pricing pricing instruments, limitations of 211 pricing mechanism 441 dynamic versus variable tolling 446–9 pricing parking, political economy of 133–4 pricing schemes distributional impacts across locations 115–16 distributional impacts in terms of income 112–15 road pricing and electric cars 117–18 role of revenue recycling 116–17 pricing strategy, ride-sourcing services 197, 199, 202–4 private airports, pricing of 211–13 private automobiles/privately operated microbuses 419 private contracting 312–14 private financing, of transport investment 281 financial evaluation 283–5 project versus corporate finance 281–2 public support to private projects 287–8 traditional and innovative financial sources 285–7 private investors, transport financing 275, 280 private lenders, guarantee for 332–3 Private Participation in Infrastructure (PPI) project database 374 private projects, public support for 287–8 private road infrastructure, franchising of 46–7 private road tolls 44 private transportation 425 private vehicles changes to optimal fare of 261–2 road pricing for 430 vehicle automation 255–6 private–public partnerships (PPPs) 468n11 in New Zealand 461–6

480  Handbook on transport pricing and financing

see also public–private partnerships private-sector participation (PSP) 340 Private-Sector Window (PSW) 344 privatization 39 of airports 211, 212 of roads 46 probabilistic voting models 125–6 “product rule” of differentiation 45 profit-maximization (PM) pricing strategy 202 profit-maximizing capacity 46 price 46 toll 47 program for results (PforR) 333–4 Project Bond Initiative 287 project finance versus corporate finance 281–3 financing of PPPs 317 principles of 468n14 pro-poor public transport system 357 “protection for sale” framework 134 public acceptance, in investment appraisal development strategy 302–3 economic appraisal and political influence 304–5 option generation and scheme selection 303–4 public activity bonds (PABs) 438 public contracting 312–14 Public Finance Initiative (PFI) 298 public funds costs of 183–5 marginal cost of 351 opportunity cost of 359 role of 171 public goods 44 public provision 16–17 of infrastructure 326 public roads, parking on 432n25 public service 16–17 public service contracts 401 public transit subsidies 101n4 public transit systems, funding and pricing 385–6 fare and ticketing 386–7 subsidy 387–8 public transport (PT) 25, 49, 108, 300, 376, 425 changes to optimal fare of 262–3 cost recovery ratio of 137 environmental considerations and climate change 186 pricing in 186 revenues for 140 role of modal competition 182–3 social acceptance and political economy 185–6 subsidies, the cost of public funds and design 183–5

time-dependent pricing schemes 30 vehicle automation 258–9 pricing, in Australia 456–9 services 26, 355 supply, foundations of operators’ costs, users’ costs, and decision variables 171–3 single-line model 173–5 space and lines structures 175–7 technological characteristics impact operations and optimal design 177–8 Public Works Authority (PWA) 317 public–private partnerships (PPPs) 311, 330, 364, 417, 438 Africa transport financing 373–4 contracting under conventional provision and 311, 312 efficient regulation of 326 financing of 341 leverage 317–18 project finance 317 risk and guarantees 318 traditional provision 318–19 valuation of 318 free public funds 326n2 governments fund 327n7 transport financing instruments 337 as web of contracts 313 public-sector financing 373 Africa transport financing 369–70 pure strategy Nash equilibrium (PSNE) 210 Quality Adjusted Life Years (QALYs) 297 quantity-based demand management mechanisms 214–20 Queen Elizabeth Way (QEW) 438 rail freight pricing 238–40 transport 230 rail infrastructure charges 397, 413 “Rail plus Property” model 380, 387–90 rail track access charges 398–9 congestion and scarcity 400 mark-ups 400–401 wear and tear costs 399–400 Ramsey pricing 25 Ramsey–Boiteux pricing 12–13 Ramsey–Boiteux rule 401 Ramsey–Mirrlees component 75 rate of return (ROR) 214 real options (RO) analysis 284 rebates 67–8 redistribution 351, 356 regional development 348, 359 regulation 17–18

Index 

and non-aeronautical operations 213–14 price 191, 197 private road competition substitute for 325–6 strict fare and entry 190 for traditional taxi industry 191–2 Regulatory Asset Base (RAB) 465, 466 renegotiations 316–17 residential sectors, outside trade node 78–9 Resource-Financed Infrastructure (RFI) 373 Resources-for-Infrastructure (R4I) 373 revenue management models, research in 239 revenue neutrality 454 revenue recycling 116–17 revenue use 116, 117 revenue-maximizing pricing 44–6 rewarding 55–6 ride-hailing industry 4 ride-hailing services 190, 191, 204 congestion effect of 423 emergence of 423 ride-sourcing market, economic analysis of 196–7 aggregate model of equilibrium 197–200 impacts of platform’s pricing strategies on performance 200–201 optimal pricing strategies of 201–3 ride-sourcing services 190 meeting function for 199 popularity of 196 transactions between customers and drivers 197 on urban congestion 204 risk allocation 301–2 BOT contracts 315 financing of PPPs 318 for PPP project 463 risk assignment 327n18 risk-sharing partnership 465 road capacity 260 vehicle automation effect on 253 road capacity provision basic theory 157–9 capacity choice in dynamic models 160–161 capacity choice under second-best pricing 159–60 decentralized provision of capacity 160 empirical studies on 161–4 road charging, overcoming ideological bias in 118–20 road congestion 128 road freight 395 externalities 236 pricing 235–7 road networks equilibrium user behaviour in 32 second-best road pricing in 147–50

481

road pricing 107, 112–14, 116–18, 128, 365, 376, 380, 381, 390 in Australia 452–6 charging for use 401–2 EU legislation 402–3 HGV charges 403–9 urban road pricing schemes 409–12 as congestion management tool 381–2 empirical studies on 161–4 fuel levy 365–6 as funding source 384–5 history of managed lanes/HOT lanes/express lanes 439 need for priced infrastructure 438 public–private partnerships 438 transport funding in Canada 437–8 Mainland China and Hong Kong 382–4 for private vehicles 430 road pricing equity concerns 367–8 road tolling 366–7 role of housing market 131–2 road pricing mechanism 418 road pricing reform 452 road safety externalities 418, 431n11 road space rationing scheme 383 road tolling 366–7 road transport 59 externalities source in freight transport 230 investment in 350 negative externalities 24–5 positive externalities 24, 25 road user charge (RUC) 460 road-use charging system 366 rural poverty, transport funding and 355–9 SACTRA report 297 safety externalities 30–31 Santiago public transport system 420 SATURN 161 scale economies 171, 175, 177, 180, 181, 183, 187 scarcity, rail track access charges 400 SCGE models see Spatial Computable General Equilibrium (SCGE) models secondary markets, use of 217–19 second-best congestion pricing in bottleneck model 153–4 policies 76 second-best equilibria 88 second-best pricing 39, 76, 90, 208 capacity choice under 159–60 price-based demand-oriented measures 210–211 in static models 146–7 cordon pricing 150–151 in road network 147–50

482  Handbook on transport pricing and financing

sophisticated pricing schemes 152 value pricing 151–2 of urban transport services 418 second-best Ramsey pricing 234 securitization 287 self-finance theorem/cost-recovery theorem 25, 26, 158, 160–161 self-internalization hypothesis 209 Seoul Metro 386, 387 service-based revenue management models 239 shared-mobility services 256 sharing economy, rise of 4–5 sharing-economy-based services 57 shifting modal priorities 5–6 simulation-based approaches 33 simulation-based general equilibrium models 76 Singapore Area Licensing Scheme in 382 congestion pricing in 381 congestion pricing strategy 387, 389 ERP system 384 toll level in 381–2 Singapore Area Licensing Scheme 381, 382 single-line model 173–5 single-till ROR 223n11 SISBEN 429 Slot Allocation Regulation 214, 217 slot babysitting 217 slots 214–20 small and medium enterprises (SMEs) 375 smart card technology 356 smartphone-based ride-sourcing apps 196 social acceptance and political economy 185–6 social and distributive issues 418 social inclusion, transport financing 339–40 social optimum 29–30 social welfare 108, 111, 148, 162, 202, 261, 465 benchmark metric of 185 social/distributive issues 417, 425–30 Solé-Ollé, A. 351 Soteropoulos, A. 263 Soumahoro, S. 138 South Australian State Government 455 Spanish transport projects, funding and financing mechanisms 275 Spatial Computable General Equilibrium (SCGE) models 297 spatial density, balancing negative and positive externalities of 2–3 spatial economics 2 spatial equity 107, 112 spatial general equilibrium model 76 spatial heterogeneity 3 spatial mismatch hypothesis 352 special assessment district (SAD) 300

special assessment tax 300 special purpose vehicle (SPV) 281, 282, 311, 313, 337 standard monopolistic price rule 57n3 state-owned enterprises (SOEs) in China 372 transport financing 275, 279 static models 26–7 step-tolling schemes 222, 223n14 Stockholm congestion charge 112, 119, 133 stop-and-go approach 319 structural parking 103n31 SUBE 429 sub-optimal pricing, effects of 178–82 subsidies 24–6, 183–5, 355–6 and fare concessions 458 funding and pricing 387–8 from general taxes 274, 278 optimal transit fares and 419 subway fares 391n18 supply chain international dimension 232 supply-oriented measures 208 supply-side subsidies 428, 432n28 surface parking 80, 88, 92–3, 102n30 surge pricing 4 sustainability financial 5 for transport projects and services 1 sustainable development 1–2 Sustainable Development Goals (SDGs) 342 sustainable transport financing 342 blended finance 344–5 climate finance instruments to mobilize private capital 343 climate funds 344 decarbonizing transport 342–3 green bonds 343–4 Switzerland distance-based HGV charge 405, 409 vignettes for HGV 413n1 TAC see track access charges (TAC) tax competition 39, 51, 52, 55 literature on 355 and tax exporting 359 tax exporting 51 tax externalities 354 tax increment financing 300 tax reforms 108, 117 tax revenue premium 183 taxi fare impacts of 194–5 system optimal 195–6 taxi fleet size impacts of 194–5 system optimal 195–6

Index 

taxi services 190 taxi–customer meetings 191, 192 theory of congestion pricing 13, 146 theory of externality pricing 40 third-degree price discrimination 14–15 Thredbo International Conferences on Competition and Regulation in Public Transport 456, 457 “three-island” urban model 132 three-mode first-best pricing model changes to optimal fare of private 261–2 changes to optimal fare of public transport 262–3 model presentation 260–261 ticketing, public transit systems 386–7 time savings 298–9 time valuations, heterogeneity in 32–3 time-based charging schemes, for HGV 403, 405, 406 time-based differential pricing 386 time-dependent pricing schemes 30 time-varying components 42 toll reform scheme 162 toll roads 15 tolling 321–3, 466 dynamic versus variable 446–9 on motorways 403 roadway pricing 436–7 schemes for HGV on concessionary motorways 404 track access charges (TAC) 398 Tradable Bottleneck Permit (TBP) scheme 155 tradable permits 40, 55, 56 versus prices 31–2 tradable travel credit scheme 384 trade flows 330 traditional financial sources, transport investment asset securitization 287 bonds 286–7 commercial banks 285–6 infrastructure funds 286 taxi market, economic analysis of aggregate model of equilibrium 192–3 impacts of taxi fare and taxi fleet size on performance 194–5 necessity of regulations for 191–2 system optimal taxi fare and taxi fleet size 195–6 traffic enforcement 139 traffic performance measures 447–8 train segment pricing 239 train/block-based revenue management models 239

483

transit fares 418–20 mechanisms to set and readjust 420–421 transit pricing and taxes 418 transit subsidies 419 for formal systems 430 transit systems 278 TransMiCable 343 Transmilenio BRT system 356 transport appraisal 295–8, 305 Transport Appraisal Guidance 302 transport decarbonization 343 transport externalities 108 pricing motives and principles 60–61 reduction in 111 from theory to practice 61–3 transport financing 338, 345 Africa 368–9 Chinese development finance 371–3 official development finance from multilateral banks 370–371 public–private partnerships 373–4 public-sector financing 369–70 financial and environmental sustainability 340 financing PPPs 341 financing urban transport 338–9 institutional development 340–341 mobility and social inclusion 339–40 project funding and financing 341–2 public sector 364 transport financing instruments 331 additional financing instruments 336–7 development policy financing and program for results 333–4 IBRD loans and guarantees 331–3 IFC loans and equity investments 334–5 MIGA guarantees 335–6 public–private partnerships 337 transport funding in Canada 437–8 and urban and rural poverty 355–9 Transport Global Practice (TGP) 333 transport infrastructure 330 challenges in provision of 311 investment for 348 and job accessibility 352 positive effects of investment in 350 transport infrastructure pricing policy, Europe 396–7 transport infrastructure systems 380 transport investment 273 financing in New Zealand 461 financing of 278 definition of 273–5 entities 278–80 instruments 280–281

484  Handbook on transport pricing and financing

funding of charges to indirect beneficiaries 277 definition of 273–5 prices to users 276–7 sources of 276 subsidies from general taxes 278 private financing of 281 financial evaluation 283–5 project versus corporate finance 281–3 public support to private projects 287–8 traditional and innovative financial sources 285–7 social benefits from funding and financing 288–90 transport markets pricing objectives in 24 types of interactions in 51, 52 transport modes 229, 231, 233, 234, 236, 242 vehicle automation effects on pricing 255–9 transport policies 1, 127, 359, 376 equity and distributional tensions in 3–4 in national policy agendas 348–52 transport pricing 1, 9–10, 19–20, 276–7, 380, 389, 396, 413 affordability and distributional consequences of 425 in Australia 452 demand management through 3–4 distribution of policy impacts of 107 diverging incentives with horizontal externalities 53–5 in hierarchy of regional governments 353–5 historical developments contribution of Pigou 11–12 Marshallian approach 10–11 nineteenth-century origins 10 implementing ideal prices 18–19 by multiple governments 49–51 paying for roads 15–16 political acceptability of 107 political economy of 128–33 in practice external costs, pricing, and optimal transport systems 13–14 marginal-cost pricing 12–13 price discrimination 14–15 public provision and public service 16–17 regulation 17–18 strategies 230, 234 theory and methodologies 24 corrective taxes and subsidies 24–6 environmental, noise, and safety externalities 30–31 model extensions 31–3

models of capacity scarcity 26–30 solution and optimisation methods 33–4 types of policy interactions 51–3 vehicle automation 253, 260 transport projects 273 appraisal process of 301 external benefits of 277 government appraisal of 295 transport taxes, absence of restrictions on 109 transportation fuel tax and over-utilize 64 internalize emissions from 59 public means of 63 Transportation Infrastructure Finance and Innovation Act (TIFIA) 288 transportation subsidies 428 transportation technology 73 travel behavior data 164 travel mode choice 73–4 travel modes, congestion tax on 85–9 travel time, vehicle automation effect on 254 traveler expectations 436 traveler sentiment 441 toward premium service 445–6 Treaty of Rome (1957) 394 trip-based MFD model 157 two-zone monocentric model 90 Uber 190 UK Private Finance Initiative (PFI) 468n10 UN Development Program (UNDP) 344 uncongested public transit mode 74 United Kingdom (UK) Department for Transport 305 Public Finance Initiative in 298 rail franchising 18 Transport Appraisal Guidance in 302 value for money guidance 306 VTTS 296 United States evolution of infrastructure spending 137 express lanes in 439, 440 external road freight costs 235 infrastructure investment in 138 intercity freight transport in 240 slot administration 214 transport pricing and funding in 436 transport pricing in 449 unlisted funds 286 unpriced externalities, urban equilibrium conditions with 82–4 unpriced traffic congestion 75 unsolicited proposals, BOT contracts 317 untolled roads 325 up-front investment 438

Index 

urban equilibrium conditions 82–4 urban form 73–4, 89–98 congestion tax on 85–9 research on 99 with and without minimum-parking requirements 92–4 urban growth boundaries (UGBs) 76, 100nn3, 4 urban infrastructure projects 137 urban mobility 417–18 urban poverty, transport funding and 355–9 urban road pricing 15–16, 395 schemes of 409–12 urban spatial equilibrium 75 urban spatial models 101n14 urban spatial structure 73–4 congestion tax on 85–9 urban transport 13, 27–8 financing of 338–9 private-sector participation in 340 second-best pricing of 418 urbanization 364 US-based China–Africa Research Initiative 371 use-it-or-lose-it rule 217 ‘user-pays’ principle 396 users’ costs 171–3 Vale Transporte program, in Brazil 357, 428–9 Value at Risk (VaR) 318 value capture (VC) 295, 299, 300, 358–9, 464 Value of a Life Year (VOLY) 297 Value of a Prevented Fatality (VPF) 297 value of travel time savings (VTTS) 254–6 value pricing 151–2, 159–60 value-based price discrimination 233 value-for-money tests 319 value-of-service principle 11, 12 values of travel time savings (VTTS) 296 variable demand modelling techniques 296 variable pricing 446–9 vehicle automation 253 effects on pricing of transport modes 255 active modes 259 AMoD 256–8 intermodal transport and MaaS 259 private vehicles 255–6 public transport 258–9

485

and modal attributes 253 effect on parking availability and price 254–5 effect on road capacity 253 effect on travel time and value of travel time savings 254 Vehicle Miles Traveled 442 vehicle scrappage schemes 69–70 vehicle taxes 66–7 fuel taxes vs. 67 vehicle-based costs 237 vehicle-to-infrastructure (V2I) communication 252 vehicle-to-vehicle (V2V) communication 252 vertical environmental externality 53 vertical equity 107, 112 vertical expenditure externality 53 vertical externalities 53 vertical fiscal externality 53 viability gap funding (VGF) 374 Vickrey bottleneck model 210 Vickrey model 30 Vickrey–Clarke–Grooves mechanisms 222 Victoria Government 456, 462, 463 voting analysis of 133 majority voting 125, 127, 131, 132, 136 probabilistic voting models 125–6, 137 probability of 138 wear and tear costs, rail track access charges 399–400 weighted average cost of capital (WACC) 284 welfare economics 10, 11, 24, 124 welfare effects 161–2 welfare-maximizing toll 47 welfare-optimal pricing and investment rules 124 welfare-related pricing 24, 34 wider economic benefits 297 Wild Goose Chase (WGC) 198 workplace, parking charges at 96–9 World Bank 330, 331, 336, 345, 368, 370, 382 estimations of revenues and costs 389 investment in urban transport 338 PPI project database 374