Market Development and Policy for One Belt One Road 0128159715, 9780128159712

Citation preview

MARKET DEVELOPMENT AND POLICY FOR ONE BELT ONE ROAD

This page intentionally left blank

MARKET DEVELOPMENT AND POLICY FOR ONE BELT ONE ROAD Series Volume Editor

ACHIM I. CZERNY Associate Professor, Department of Logistics and Maritime Studies (LMS), Faculty of Business, The Hong Kong Polytechnic University, Hong Kong, China

XIAOWEN FU Professor, Department of Industrial and Systems Engineering, The Hong Kong Polytechnic University, Hong Kong, China

PAUL TAE-WOO LEE Professor and Director, Maritime Logistics and Free Trade Islands Research Center, Ocean College, Zhejiang University, Zhoushan, Zhejiang, China

Elsevier Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States Copyright © 2022 Elsevier Inc. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. ISBN: 978-0-12-815971-2 For information on all Elsevier publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Joseph P. Hayton Acquisitions Editor: Kathryn Eryilmaz Editorial Project Manager: Joshua Mearns Production Project Manager: Kiruthika Govindaraju Cover Designer: Greg Harris Typeset by STRAIVE, India

Contents Contributors

1.

An introductory overview of the Belt and Road Initiative studies

ix

1

Achim I. Czerny, Xiaowen Fu, and Paul Tae-Woo Lee

2.

China's Belt and Road Initiative: Quantifying the causal relationship between maritime connectivity and global trade

9

Tsz Leung Yip, Eve Man Hin Chan, and Danny Chi Kuen Ho

3.

1. The Belt and Road Initiative's role in improving transport connectivity and trade 2. Literature review 3. Extended gravity model for B&R countries’ exports 4. Data analysis and results 5. Discussion and policy implications 6. Conclusions for market development Acknowledgments References

9 11 13 16 18 23 24 24

China's recent railway developments and policy reforms

27

Kun Wang, Wenyi Xia, and Anming Zhang

4.

1. Introduction 2. Recent railway developments in China 3. China's HSR developments 4. Conclusions and future research References

27 28 41 55 57

Effects of the “Belt and Road” initiative on the cruise industry

59

Yui-yip Lau, Meihua Xu, Xiaodong Sun, and Adolf K.Y. Ng 1. Introduction 2. Overview of BRI

59 61

v

vi

5.

Contents

3. Cruise policy in BRI 4. Exploring the cruise market along BRI 5. Identifying and evaluating the possible cruise itinerary along BRI 6. Conclusion References

62 66 68 74 75

The investment efficiency of overseas ports: Three macroscopic factors

77

Dong Yang and Lu Li

6.

1. Introduction 2. Overview of Chinese overseas port investment 3. Three port case studies 4. Conclusion References

77 78 83 95 95

Impacts of air transport subsidies on landlocked developing countries’ connectivity under the “One Belt One Road” initiative

99

Kan Wai Hong Tsui and Hanjun Wu 1. 2. 3. 4. 5. 6. 7.

Introduction The One Belt, One Road initiative: An overview Landlocked developing countries and definitions of landlockedness Key problems and challenges among LLDCs Obligation to support LLDCs’ air connectivity: Five key reasons Air transport improves connectivity and eradicates landlockedness Rationale for supporting the essential air services of in small communities through air transport subsidies 8. A “win-win” outcome for LLDCs and partner countries benefiting from improved air connectivity resulting from air transport subsidies 9. Conclusion Acknowledgments References

7.

Collusive pricing detection in ocean container transport: A case study of Maritime Silk Road

99 100 102 104 106 109 112 115 118 119 119

125

Gang Dong, Jin Li, and Paul Tae-Woo Lee 1. Introduction 2. Literature review

125 127

Contents

3. Collusion detection model 4. Analysis of the collusive pricing solutions 5. Case study 6. Discussion and policy implications 7. Conclusion Appendix A Appendix B Acknowledgments References

8.

An infrastructure investment game: Or, why the belt and road initiative can represent an equilibrium outcome

vii 130 139 141 145 145 146 148 151 151

155

Achim I. Czerny and Se-Il Mun 1. Introduction 2. The model 3. Investment behaviors 4. Oligopoly and import taxes 5. Conclusions References Further reading

9.

High-speed rail and air transport integration in hub-and-spoke networks: The role of airports

155 157 158 163 168 169 169

171

Alessandro Avenali, Tiziana D’Alfonso, Alberto Nastasi, and Pierfrancesco Reverberi 1. Introduction 2. Literature review 3. Incentives to intermodal cooperation 4. The model 5. Benchmark case: No agreement 6. Air-rail cooperation 7. The role of airports 8. Discussion and concluding remarks Appendix References

171 173 175 180 182 183 189 190 193 193

viii

Contents

10. On impacts of Pakistan Railways Main Line 1 on “North China-EU” export transit—Taking export transit between Beijing and UK as an example

197

Ying-En Ge, Mingfeng Mo, Fangwei Zhang, Muhammad Arsalan Khalid, Mengmei Yu, Guanke Liu, and Wenqian Lu 1. Introduction 2. Research tools 3. Probability distribution model and parameter analysis of North China-Europe export channel 4. Empirical analysis of the probability distribution model of the North China-Europe export channel 5. Conclusions and emerging directions Acknowledgments References Index

197 200 201 204 217 219 219 221

Contributors Alessandro Avenali Department of Computer, Control, and Management Engineering Antonio Ruberti, Sapienza University of Rome, Rome, Italy Eve Man Hin Chan Faculty of Design and Environment, Technological and Higher Education Institute of Hong Kong, Hong Kong, China Achim I. Czerny Department of Logistics and Maritime Studies (LMS), Faculty of Business, The Hong Kong Polytechnic University, Hong Kong, China Tiziana D’Alfonso Department of Computer, Control, and Management Engineering Antonio Ruberti, Sapienza University of Rome, Rome, Italy Gang Dong School of Economics and Management, Shanghai Maritime University, Shanghai, China Xiaowen Fu Department of Industrial and Systems Engineering, The Hong Kong Polytechnic University, Hong Kong, China Ying-En Ge College of Transportation Engineering, Chang’an University, Xi’an, China Danny Chi Kuen Ho Department of Supply Chain and Information Management, The Hang Seng University of Hong Kong, Hong Kong, China Muhammad Arsalan Khalid School of Business, National University of Singapore, Singapore Yui-yip Lau Division of Business and Hospitality Management, College of Professional and Continuing Education, The Hong Kong Polytechnic University, Hong Kong, China Paul Tae-Woo Lee Maritime Logistics and Free Trade Islands Research Center, Ocean College, Zhejiang University, Zhoushan, Zhejiang, China Jin Li School of Economics and Management, Shanghai Maritime University, Shanghai, China Lu Li Department of Logistics and Maritime Studies (LMS), The Hong Kong Polytechnic University, Hong Kong, China Guanke Liu School of Business, National University of Singapore, Singapore

ix

x

Contributors

Wenqian Lu Tilburg School of Economics and Management, Tilburg University, Tilburg, Netherlands Mingfeng Mo College of Transport & Communications, Shanghai Maritime University, Shanghai, China Se-Il Mun Graduate School of Economics, Faculty of Economics, Kyoto University, Kyoto, Japan Alberto Nastasi Department of Computer, Control, and Management Engineering Antonio Ruberti, Sapienza University of Rome, Rome, Italy Adolf K.Y. Ng Department of Management, Faculty of Business and Management, BNU-HKBU United International College, Zhuhai, China Pierfrancesco Reverberi Department of Computer, Control, and Management Engineering Antonio Ruberti, Sapienza University of Rome, Rome, Italy Xiaodong Sun School of Business Administration, East China Normal University, Shanghai, China Kan Wai Hong Tsui School of Aviation, Massey University, Palmerston North, New Zealand Kun Wang School of International Trade and Economics, University of International Business and Economics, Beijing, China Hanjun Wu School of Aviation, Massey University, Palmerston North, New Zealand Wenyi Xia Department of Logistics and Operations Management, HEC Montreal, Quebec, Canada Meihua Xu School of Business Administration, East China Normal University, Shanghai, China Dong Yang Department of Logistics and Maritime Studies (LMS), The Hong Kong Polytechnic University, Hong Kong, China Tsz Leung Yip Department of Logistics and Maritime Studies (LMS), The Hong Kong Polytechnic University, Hong Kong, China Mengmei Yu College of Transport & Communications, Shanghai Maritime University, Shanghai, China Anming Zhang Sauder School of Business, University of British Columbia, Vancouver, BC, Canada Fangwei Zhang College of Transport & Communications, Shanghai Maritime University, Shanghai, China

CHAPTER 1

An introductory overview of the Belt and Road Initiative studies Achim I. Czernya, Xiaowen Fub, and Paul Tae-Woo Leec a

Department of Logistics and Maritime Studies (LMS), Faculty of Business, The Hong Kong Polytechnic University, Hong Kong, China Department of Industrial and Systems Engineering, The Hong Kong Polytechnic University, Hong Kong, China c Maritime Logistics and Free Trade Islands Research Center, Ocean College, Zhejiang University, Zhoushan, Zhejiang, China b

Since the Belt and Road Initiative (BRI) first proposed by the Chinese government in 2013, significant changes have taken place along the Asia-Middle East-Europe region. In addition to the investments directly associated with the BRI, the initiative has also triggered policy changes in the affected countries and regions, firms’ perspectives of future developments, and thus their operations and strategic planning. Unlike common foreign direct investments (FDIs), the Chinese government’s support behind the BRI has led to some concerns and even potential political conflicts. These complications have introduced more challenges in assessing the initiative’s consequences, future developments, and interactions with the existing market mechanisms. Although there are many different views concerning the BRI, either positive or negative, it is generally agreed that such an extremely large project is expected to bring long-lasting changes to the industries and markets affected. The initiative encourages policy coordination, trade facilitation, financial integration, and transport connectivity. It covers countries across Asia, Africa, the Middle East, and Europe, involving at least 70% of the global population, 75% of world energy reserves, and 55% of world gross national product. There is a need for policy makers, industry players, and academic researchers to obtain a good understanding of the BRI in a timely manner. Although many papers and studies are now available publicly, they have discussed a wide range of topics and sectors. Among these, the transport and logistics sector is of particular importance to the initiative: it not only directly contributes to the production of transport and logistics services but also provides essential inputs to other sectors such as tourism, trade, Market Development and Policy for One Belt One Road https://doi.org/10.1016/B978-0-12-815971-2.00003-7

Copyright © 2022 Elsevier Inc. All rights reserved.

1

2

Market development and policy for one belt one road

infrastructure investment, and management, which are the core targets of the initiative. Therefore, it is important to more carefully examine the implications of the BRI to the transport and logistics sectors in the region, the best strategies and operation practices that the industry can pursue, and the right government policies that should be implemented in relation to the initiative. To facilitate timely studies of the transport and logistics sectors along the BRI regions, this book engaged researchers to examine the development and performance in multiple countries. Even though not every chapter is dedicated to the implications of the BRI, all of them have analyzed the transport and logistics sectors along the BRI region and thus facilitate further studies more directly or exclusively to BRI. Both theoretical and empirical studies have been included as we believe that in-depth analysis of this important issue needs advancements in both areas in the long term. These chapters’ main contributions and conclusions are summarized as below. In Chapter 2, Tsz Leung Yip, Eve Man Hin Chan, and Danny Chi Kuen Ho conducted an analysis of the BRI’s influences by examining the causal relationship between maritime connectivity and global trade. A number of studies have investigated the effects of improved transport connectivity in promoting international trade. This chapter contributes by considering trade beyond the BRI region. The authors develop an extended gravity model to study textile exports of 16 Asian countries or regions in BRI to United States, with the incorporation of the Liner Shipping Bilateral Connectivity Index (LSBCI) and Logistics Performance Index (LPI). Their analysis suggests that LSBCI and LPI play important roles in textile trade, which demonstrates the changing sectoral trade patterns between BRI and non-BRI economies. Their analysis provides insights into the importance of enhancing both international maritime connectivity and domestic logistics performance for trade facilitation. The recommendation of improving domestic logistics performance to promote international trade is echoed in the analysis in Chapter 3, where Kun Wang, Wenyi Xia, and Anming Zhang discuss China’s recent railway developments and policy reforms. The authors first review China’s recent developments and policy reforms in its rail sector and then further discuss the implications of BRI. The policy discussions in the context of market performance and outcome allow the authors to draw a number of important conclusions. The authors concluded that China has been promoting deregulation in its rail sector, which introduces more changes facilitating the emergence of a more market-oriented rail industry. These changes brought various benefits and significant changes in terms of operational efficiency and

An introductory overview of BRI studies

3

network structure. The deregulation and changing pricing decisions of the rail sector will further facilitate China’s trade by rail with BRI countries. These findings are apparently consistent with the key conclusions in Chapter 2 on the maritime sector. The authors further used the China-Euro Railway Express as an example for promoting transport connectivity and trade linking China, BRI countries in Central Asia, and Europe. Further discussions on China’s domestic operations of high-speed rail (HSR) are also provided. Although HSR has mostly affected the passenger transport market, it has long-lasting effects on the rail industry’s service provision, involvements of private capital, and positive effects on regional developments. Indeed, passenger operation and service are important to the transport’s growth per se, in addition to their roles in facilitating trades, tourism, and regional development. In Chapter 4, Yui-yip Lau, Meihua Xu, Xiaodong Sun, and Adolf Ng examined the effects of the BRI on the cruise industry. Currently, the cruise industry is still at its early stage of market penetration in most of the BRI countries, and most previous studies have examined the cruise market in North America and Europe, which are much larger cruise markets. However, with its huge population and fast-expanding economy, many scholars believe the cruise market in the Asia-Pacific region will have a bright future in the coming decades, with more emerging destinations added and fast-growing passenger volumes. There are however many issues to be solved before the industry can achieve sustained growth. For example, the availability and smooth operations of cruise terminals need to be solved. The industry also needs to better understand the opportunities and challenges of developing the cruise market along the BRI region. To fill these gaps in research, the authors first provide a review of the related policy and industry development in the context of the BRI market. They further examine the promising cruise itinerary in the region and how they could lead and facilitate sector growth in the region. No transport sector can perform well without sufficient and efficient infrastructure invested. In the case of cruise operations, cruise terminal can be essential for high customer satisfaction. For terminal and port operations in general, sufficient investments in infrastructure are of critical importance for the efficient operations of the whole supply chain. Because port investments are often lumpy and significant, efficient port operation is itself an important research topic. In Chapter 5, Dong Yang examined overseas investment efficiency using three port case studies. Such a study aims to identify the factors influencing Chinese enterprises’ investment effectiveness in ports along the BRI region. Specifically, in the case of the Port of Piraeus,

4

Market development and policy for one belt one road

the effects of the Greece debt crisis in 2009 and COSCO Shipping’s efforts in exploiting its supply chain integration power are examined. Political stability and existing investments are studied for the case of Hambantota Port, together with the operation experience of port operators. For the case of Gwadar port, the favorable geopolitical opportunity is also discussed. The last empirical study was carried out by Kan Tsui Wai Hong and Hanjun Wu in Chapter 6, which focused on the effects of air transport subsidies on landlocked developing countries’ (LLDCs) connectivity under the BRI. Because LLDCs lack maritime connectivity, which is often one of the most important trade facilitators, there is more pressing need for them to improve air connectivity. However, the relatively high cost of air transport can be one major challenge to developing countries with a relatively low income. As a result, many countries have resorted to aviation subsidy. The authors examine the effects of air connectivity to landlocked developing countries in terms of trade and economic development, flows of goods and people, and tourism, in the context of the BRI. They argued that the lack of funding is a major obstacle for the growth of air transport. Although subsidy can be an efficient policy, a clear set of subsidy policies should be secured for LLDCs first. Such an effort can help identify the rationale and justifications for air subsidies to LLDCs too. To achieve this objective, the authors tried to draw lessons from the provision of essential air services to small and remote communities, a practice commonly used in many countries. Based on their analysis, the authors argued that improved air connectivity is expected to bring a “win-win” outcome for both LLDCs and partner countries, facilitating economic development and trade in goods and services and people movements between LLDCs and global markets. Both empirical and modeling works have their distinctive advantages and limitations. Chapter 7 combines both approaches. The study, carried out by Gang Dong, Jin Li, and Paul Tae-Woo Lee, offers a case study of collusive pricing detection in ocean container transport along the Maritime Silk Road (MSR). The authors first develop a non-cooperative game theoretic model to detect the collusive pricing solutions between liner shipping companies and the corresponding container terminals. According to the game equilibrium results, a direct and applicable indicator is proposed, that is, the ratio of freight rates between the liner shipping companies, by testing 24 liner routes from the Far East to the Mediterranean region along the MSR. The collusive pricing solution is independent of designed capacity under one pair of colludes, whereas there is positive correlation with designed capacity in double collusion. The finding reveals that the collusive pricing solution would be

An introductory overview of BRI studies

5

most likely to occur between the liner shipping company, under the scenarios of low- or medium-level competition, and the corresponding container terminal with lower designed capacity. The remaining chapters are all analytical in nature. In Chapter 8, Achim I. Czerny and Se-il Mun developed an analytical economic model to explain why the BRI can represent an equilibrium outcome. Specifically, the authors modeled the scenario in which trade costs are determined by trade volume, such as the international trade between China and Europe. This feature aims to depict the fact that unit transport costs are influenced by delays, which in turn are jointly determined by the transport volume and the capacity of transport infrastructures. This is an important issue because transport infrastructure shortage is common in many countries, notably developing countries in Africa or Central Asia. The authors first consider markets as perfectly competitive, and the two trading regions unilaterally decide to invest in the transit countries’ transport capacities. The unit capacity costs are considered to be different, which captures China’s low cost of building transport infrastructure. The analysis results show that unilateral actions such as the BRI in which only the region with the lower unit capacity costs invests in transit countries’ transport capacities can represent an equilibrium outcome. However, equilibrium investments are too low to reach the first-best outcome which maximizes the welfare across all involved regions. That is, BRI’s benefits to participating countries may be further enlarged by attracting investments beyond China. The authors then consider oligopolies in international trade. This part shows that local firms can benefit from import taxes in two ways. First, they increase the local profit by increasing the full price of consuming the non-local product. The second is by discouraging infrastructure development in transit countries, which softens the competition between local and non-local producers. Although the import tax and low-investment scenario effectively protect local firms, they can still produce negative welfare effects for both trading partners, highlighting the ambiguous effects import taxes can have on the local economy. In addition to the general modeling work above, in Chapter 9, Alessandro Avenali, Tiziana D’Alfonso, Alberto Nastasi, and Pierfrancesco Reverberi focused on the interactions between high-speed rail and air transport integration in hub-and-spoke networks. The authors pointed out that air transport and HSR are not simple competitors. Indeed, air and HSR services can be complements on long-haul routes served by connecting flights through a hub airport. This complementarity creates room for cooperation between airlines and HSR operators, particularly relating to international

6

Market development and policy for one belt one road

connecting passengers. Airport managers are also interested in such agreements because they affect, among others, air traffic volumes and the demand for slots on the part of the airlines. The authors thus analytically developed a theoretical model to study transport operators’ incentives to cooperate and the strategic role of airports in facilitating or dampening airline-HSR cooperation via the airport per passenger fee. In the analysis, transport operators cooperate to offer a bundle of domestic HSR and international air services via a multimodal hub airport. The authors show that the scope for cooperation depends on two main factors, that is, the related sunk costs and mode substitution between air and HSR services. Although these two chapters are not limited to a particular market, Chapter 10 is focused on the Pakistan Railways Main Line 1’s effects on the “north China-EU” export transit using the export transit from Beijing-UK as an example. The authors include Ying-En Ge, Mingfeng Mo, Fangwei Zhang, Muhammad Arsalan Khalid, Mengmei Yu, Guanke Liu, and Wenqian Lu. The authors argued that with the development of BRI, Pakistan’s Railways Main Line (ML-1) will play a pivotal role in promoting trade and economic exchange between China and other countries along the routes, especially between China and Europe. Based on the historical trade data between China and Europe over the past 12 years, the authors set up a multi-logit model by introducing variables such as freight rate, time economic value, and the corresponding time preference coefficients. Such a specification emphasizes on exploring the characteristics of the export transport pattern from China to Europe and also the impact of expansion and operation of Pakistan’s ML-1 railway line on the choices of transportation routes, especially on the shipping channel. The authors first analyze the influence of different control variables on the distribution of routes, including time preference coefficients and freight rates of channels in Pakistan and the sea. They then move on to assess the impacts of the Pakistan channel on the freight pattern through the indicators such as probabilistic route assignment and changes in freight volume and provide a comparison between different channels at price levels. Finally, the study applies the freight rate of the Pakistan channel and sea channel as variables to establish the income functions of Singapore and Pakistan and analyzes the pricing game and its Nash equilibrium solution. Needless to say, many more issues and sectors could have been included in this book. That said, we hope that the book, which already covered a number of important issues in the maritime, aviation, and rail sectors, could

An introductory overview of BRI studies

7

serve as a good starting point for readers to appreciate the key benefits and challenges associated with BRI and how advanced studies could contribute to this important topic. We hope that this book could promote and facilitate more advanced studies on BRI in particular and the transport and logistics sectors in general.

This page intentionally left blank

CHAPTER 2

China’s Belt and Road Initiative: Quantifying the causal relationship between maritime connectivity and global trade Tsz Leung Yipa, Eve Man Hin Chanb, and Danny Chi Kuen Hoc a

Department of Logistics and Maritime Studies (LMS), The Hong Kong Polytechnic University, Hong Kong, China Faculty of Design and Environment, Technological and Higher Education Institute of Hong Kong, Hong Kong, China c Department of Supply Chain and Information Management, The Hang Seng University of Hong Kong, Hong Kong, China b

1. The Belt and Road Initiative’s role in improving transport connectivity and trade The Belt and Road Initiative (BRI) is an ambitious program that seeks to connect Asia with Africa and Europe through land and maritime networks along six main economic corridors. It aims at deepening regional integration, promoting trade, and stimulating economic growth. Economic corridor development is a core part of the BRI. The six economic corridors account for a significant part of the global population and capital channeled to support major infrastructure projects along the Belt and Road (B&R). The extensive investment of infrastructures provides powerful financial impetus. As regional cooperation grows and intensifies under the BRI, the inter-regional network connectivity will be expanded (Fang & Nolan, 2019). The success of unimpeded trade facilitated by the BRI lies essentially on the improvement of transport connectivity and logistics of countries connected along the economic corridors and beyond. The BRI improves connectivity across both continents and oceans in different forms and ways to facilitate and enhance cooperation and trade (Sharma & Kundu, 2016). Worth mentioning is the 21st Century Maritime Silk Road (MSR)—the maritime complement to the Silk Road Economic Belt—aiming to improve infrastructure connectivity throughout Southeast Asia, Oceania, the Indian Ocean, and East Africa. Wong and Yip (2019) identified the roles of Market Development and Policy for One Belt One Road https://doi.org/10.1016/B978-0-12-815971-2.00007-4

Copyright © 2022 Elsevier Inc. All rights reserved.

9

10

Market development and policy for one belt one road

transport infrastructure in mediating the relationship between institutions and average income in both developed and developing economies and provided new insights based on the institutional theory and factor-mobility theory. The B&R land and maritime corridors have the potential to greatly improve trade, attract foreign direct investment (FDI), and reduce poverty in its participating countries. As reported by the World Bank Group (2019, p. 5), If fully implemented, BRI transport infrastructure can reduce travel times for economies along transport corridors by up to 12 percent, reducing trade costs. In the rest of the world, travel times are estimated to fall by an average of 3 percent, showing that the non-Belt and Road countries will also benefit from access to improved rails and ports in corridor economies. These substantiates are stated in the 13th Five Year Plan (2017), Infrastructure connectivity is the foundation of development through co-operation. We should promote land, maritime, air and cyberspace connectivity, concentrate our efforts on key passageways, cities and projects and connect networks of highways, railways and sea ports. The goal of building six major economic corridors under the BRI has been set, and we should endeavour to meet it (WEF, 2017). Despite the growing importance of improved international transport connectivity and national trade logistics in promoting trade and efficient allocation of resources across geographic boundaries, past studies tend to focus on bilateral trade at the country level among members of the BRI, rather than trade beyond the B&R region. Our understanding of the sectoral trade between the B&R countries and non-B&R countries (e.g., United States) is inadequate, and more research in this area is needed (Ho, Chan, Yip, & Tsang, 2020). To address this gap, this study builds a set of extended gravity models to examine the impact of maritime connectivity between importing and exporting countries (measured by the Liner Shipping Bilateral Connectivity Index (LSBCI) developed by the United Nations Conference on Trade and Development (UNCTAD)) and exporting countries’ national trade logistics performance (measured by the World Bank’s Logistics Performance Index (LPI)) on textile exports of 16 Asian countries/ regions along the B&R to United States from 2008 to 2018. The results show that domestic logistics performance (proxied by LPI) of exporting countries and international connectivity (proxied by LSBCI) between importing and exporting countries per se contributes to bilateral textile trade. More importantly, this study has found that the model with both the terms of LPI and LSBCI has a stronger explanatory power than the model with LPI alone. This finding is important as it shows that domestic

Maritime connectivity and global trade

11

logistics will only be effective in environments where countries have sufficient international transport connectivity. Overall, this study has added value to the BRI trade literature and provided evidence supporting the notion that the BRI has the potential to stimulate trade not only among the B&R countries but also, more importantly, between the B&R and non-B&R countries. Moreover, improved international maritime connectivity and national trade logistics are indispensable drivers behind the trade growth under the BRI. The rest of this chapter is structured as follows: Section 2 presents a literature review with a focus on the link between trade under the BRI, domestic logistics performance, and international maritime connectivity. Section 3 describes the configuration of extended gravity models for this study. Data analysis and results are presented in Section 4. Discussion of findings of the impacts of logistics performance and transport connectivity on trade is presented together with policy implications in Section 5. Conclusions and suggestions for market development are presented in Section 6.

2. Literature review The literature on trade opportunities brought by the BRI is growing. One stream of studies focuses on how transport connectivity expansion and national logistics performance enhancement improve the export performance of countries along the B&R. Most often, these empirical studies developed an extended gravity model with an explanatory variable, which is the LPI developed by the World Bank, to reflect a country’s trade logistics performance (Arvis et al., 2018). For example, Hausman, Lee, and Subramanian (2013) analyzed a World Bank Group dataset containing specific quantitative metrics of logistics performance in terms of time, cost, and variability in time and found that logistics performance was significantly related to bilateral trade volumes among the 80 countries examined. Martı´, Puertas, and Garcı´a (2014), in their study of developing countries, grouped into five regions (Africa, South America, Far East, Middle East, and Eastern Europe), revealed that improvements in any of the LPI’s components can lead to significant growth in a country’s trade flows. Gani (2017) examined trade patterns of 60 countries over four time periods of 2007, 2010, 2012, and 2014 and revealed a statistically significant positive correlation of overall logistics performance with exports and imports and all the six of the logistics specificities with exports. C ¸ elebi (2017) tested a gravity model empirically to assess the extent to which logistics performance constitutes a facilitator to international trade and reported that

12

Market development and policy for one belt one road

logistics excellence increases exports more than imports for low- and lowermiddle-income economies. Wang, Qiu, and Choi (2019) employed a 5-year panel dataset of 49 countries for the years 2007, 2010, 2012, 2014, and 2016 and found that the impact of 49 importing countries’ LPI on China’s exports was higher after than before the B&R initiative implementation with 2013 as the cut-off point. In addition to the study of bilateral total trade at the country level, a few studies focus specifically at the sectoral level. Siddiqui and Vita (2021) employed panel data analysis to assess the impact of logistics performance on trade across the three countries Cambodia, Bangladesh, and India from 2001 to 2016 for the garment sector. Ho, Chan, Gunasekaran, and Yip (2020) examined clothing exports from B&R economies in Asia including Bangladesh, Cambodia, India, Indonesia, Malaysia, Pakistan, the Philippines, South Korea, Sri Lanka, Taiwan, Thailand, and Vietnam to Hong Kong from 2000 to 2017. Their gravity model of trade found a positive impact of LPI on exports. Different from Ho et al. (2020), this study considers textile trade and considers LPI and LSBCI. Mendes dos Reis, Sanches Amorim, Sarsfield Pereira Cabral, and Toloi (2020) found a positive effect of the logistics infrastructure dimension of LPI on soybean exports of Argentina, Brazil, and the United States from 2012 to 2018. Besides LPI, another group of studies have examined the role of maritime connectivity in improving trade. Fugazza and Hoffmann (2017) found that improving transport connectivity can be an important facilitating aspect of bilateral export of containerizable goods using the LSBCI as an indicator of maritime connectivity in a sample of 138 coastal countries. Hoffmann, Saeed, and Sødal (2020) found a positive and significant effect of the LSBCI on South African exports of highly containerizable products to its 142 trading partners from 2008 to 2016. Saeed, Cullinane, and Sødal (2021) examined the ten best-connected countries/regions and their 155 trading partners and found a positive association between different components of maritime connectivity, as characterized by the LSBCI, and trade values. Şeker (2020) found that a 1% increase in the Liner Shipping Connectivity Index (LSCI) provides the increment of 0.21% in the exports in European countries and Turkey. Lin, Kuo, and Chang (2020) revealed a positive spill-over effect of the LSCI of China and Hong Kong on the trade of other Asian countries to varying degrees. Worth mentioning is that very few studies examined both LPI and LSCI in an extended gravity model of trade. Lu, Rohr, Hafner, and Knack (2018) examined the effect of LPI, LSCI, and other transport infrastructure

Maritime connectivity and global trade

13

indicators on trade performance of B&R countries. They found that the LPI of the exporter has a positive impact on trade, whereas the effect of LSCI was not analyzed because it was dropped from the model because of collinearity with other explanatory variables. Chang, Huang, Shang, and Chiang (2020) examined a gravity model with trade data in 2007, 2010, and 2015 and found that importing and exporting countries’ LPI and LSCI contribute to bilateral trade flows. Based on the review of the above studies, it is found that national logistics performance (proxied by the LPI) and international maritime connectivity (proxied by the LSBCI) each has a positive effect on bilateral trade at country and sectoral levels. However, the intricacy of their joint effect on improving trade at the sectoral level under the BRI is less known. An extended gravity model is developed by the current study to examine this combined effect.

3. Extended gravity model for B&R countries’ exports International trade is the underlying essence of the BRI. As living standards improve and higher incomes are earned, it then becomes even more crucial to emphasize the significance of the market. Trade and logistic routes are challenges for many manufacturers because of the lack of accessibility and infrastructure to meet needs which can eventually affect the supply chain. As indicated by the mission of the BRI, the main objective is to create global engagement to improve connectivity between regions that were not connected before and stimulate worldwide economic growth. This objective will eventually create a more logistically connected and technologically advanced supply chain. The largest constraint, however, is that more time and resources are necessary to realize this objective and meet demands. The implications of logistics integration will affect the supply chain in the long run as different means to produce textiles goods will be realized, which can reach out to more potential consumers because of the increase in the speed of transporting products. In the long run, this realization of a more stable transportation system will help textile companies to establish themselves in developing nations. The changes in the member countries will be accompanied by property rights, healthy competition, equal opportunities, and good governance of accountability and transparency. This would eventually mean that the B&R countries would receive more support from emerging economies and inter-governmental institutions to move forward in that direction. The whole idea of the BRI is to create an integrated global community without any trade blocs or

14

Market development and policy for one belt one road

trade agreements that may inhibit the productivity of trade and connectivity among different countries. Providing equal opportunities is one of the objectives of the BRI. The BRI does not intend to launch a new system but gain sight of the larger picture in the ongoing international and mutual negotiations (Martinico & Wu, 2020). Additionally, the BRI might impact international trade cargo volume, transporting speed, and transportation routes as well as increase confidence in the supply chain through transparency. These are the strategic advantages that will be reaped from the BRI and will create more options for importing raw and strategic manufacturing materials worldwide. From a global perspective, it can be said that consumers would greatly benefit from international trade as more choices are available because of competition among manufacturers and an increased number of manufacturers. Nevertheless, it would not reduce the quality of the products as trade blocs and agreements will be made among the member states, thus creating a healthy competitive environment and providing a way for the member states to expand their economy to their full potential. This study examined the textile trade between United States and the Asian countries that are now member countries of the BRI by an extended gravity model of global trade. The model attributes trade between two countries to the size of their economies and the distance between them, hence the “gravity” model (Chan & Au, 2007; Chan, Au, & Sarkar, 2008). Moreover, the model performs well in empirical research (Bayoumi & Eichengreen, 1997; Havrylyshyn & Pritchett, 1991). Apart from its use in empirical studies, the gravity trade model has been used as a means to test different trade theories, such as by Anderson (1979), Bergstrand (1985), and Deardorff (1998), because it can be extended to account for the effect of various factors on international trade. The extended model applied in this study covered factors that have not been considered in previous studies in the textile sector, such as demographic and economic factors, and the characteristics of the business environment. The model made use of panel data over time for each scenario, thus exploring the changes and increasing the manipulation of the data quality and quantity which would otherwise not be possible with the use of cross-sectional or time-series estimation alone. The analysis included the impact of transport and logistics by using the UNCTAD’s LSBCI and the World Bank’s LPI, country-specific and social determinants, and economic indicators influential to textile trade. Because both domestic and international logistics may affect a country’s trade, we use two variables to proxy for domestic and international logistics

Maritime connectivity and global trade

15

in trade modeling. Following Lau, Chan, and Nguyen (2017), we extend the gravity model with two new variables, LPI and LSBCI:

ln EXP ijt ¼ α þ β1 ln ðGDP it Þ þ β2 ln ðPCGDP it Þ þ β3 ln GDP jt


þ β4 ln PCGDP jt þ β5 ln Dij þ β6 ln POP jt þ β7 ðLPI it Þ þ β8 ðLSBCI it Þ þ uijt , where i—the exporting country index; j—the importing country index, United States is the importer in this analysis; t—the year index when export transactions are recorded, t ¼ 2008–2018; ln( )—the natural logarithm of a number; EXP—textile export value in US dollars (millions); α—unobserved time-constant factors on EXP; β—coefficients to be determined; GDP—the real gross domestic product (GDP) of the country in USD (millions); PCGDP—per-capita GDP in USD (millions); POP—the population of the country; D—the geographical distance (in km) between the capitals of two countries; LPI—the Logistics Performance Index of the exporting country; LSBCI—the Liner Shipping Bilateral Connectivity Index between importing and exporting countries; and u—the error term. The dependent variable (EXP) is the logarithm of the textile export value of a B&R exporting country/region (including Brunei, Bangladesh, Cambodia, China, Hong Kong, Indonesia, Laos, Malaysia, Myanmar, Nepal, Pakistan, the Philippines, Singapore, Sri Lanka, Thailand, and Vietnam) in the US market. The independent variables include gross domestic product (GDP), per-capita GDP (PCGDP), distance between countries (D), population of exporters (POP), Logistics Performance Index (LPI), and Liner Shipping Bilateral Connectivity Index (LSBCI). Because the economic size of the exporting and importing countries is usually measured by the GDP, the GDPs of the exporting countries and their textile exporters are considered to represent the economic masses which should positively influence the textile exports of the country. As GDP also indicates the supply capability of the textile-exporting countries, coefficients β1 and β3 are expected to be positive. Based on the gravity principle, the PCGDP of the

16

Market development and policy for one belt one road

exporting country is used as a proxy of capital intensity. As the textile industry is labor-oriented, PCGDP is used to indicate the impact of the monetary conditions on the workforce in countries with apparel exports. Additionally, the income level or purchasing power of importing countries is represented by the PCGDP. Controlling for the GDP, richer countries (in terms of their PCGDP) are likely to demand more choices in different products which may be imported from countries that specialize in the production of such products. Therefore, coefficients β2 and β4 are also expected to be positive. The distance (D) has proven to be one of the most significant variables in previous studies (e.g., Frankel & Rose, 2002). In general, a greater distance involves higher transport costs and therefore would have an adverse impact on textile trade. Therefore, coefficient β5 should be negative. The population size (POP) of the exporting country is included as a determinant of demand for textiles products. Coefficient β6 is expected to be positive. In addition to the above macroeconomic factors, the LPI and LSBCI are newly added as independent variables. The LPI is a composite index compiled by the World Bank. The values of LPI are based on six components reflecting a country’s logistics efficiency and effectiveness. Exporting countries that share similar factor endowments may differ in their logistics performance in terms of custom clearance efficiency, transport and information technology (IT) infrastructure quality, ease of arranging international shipments, the ability to track and trace shipments, domestic logistics costs, timeliness in reaching the destination, and competence of the domestic logistics industry. It is anticipated that the domestic logistics performance of exporting countries contributes to their textile exports. Thus, coefficient β7 is expected to be positive. The LSBCI denotes how well a country is connected to the global liner shipping network. The LSBCI dataset has been compiled by the UNCTAD since 2004. The values of LSBCI are calculated from data related to liner shipping, including the number of container ships, container-carrying capacity, the number of liner services, and the size of the largest container ship. However, although it is expected that LSBCI has a positive effect on global textile trade, the expected impact of LSBCI has not been fully determined yet. It is anticipated that coefficient β8 is positive.

4. Data analysis and results A pooled cross-sectional (PCS) or cross-sectional (CS) ordinary least squares (OLS) is often utilized in the gravity trade model analysis. However, Cheng and Wall (2005) showed that these estimation approaches create biased results. Because there is no heterogeneity allowed in the error term for

Maritime connectivity and global trade

17

standard CS regression equations, the gravity trade model produces overestimated results. In order to solve the problem of using the OLS, the panel data estimation method is used to determine the variables that affect the bilateral trade flows from the B&R countries to United States over time. As noted by Baltagi (2013), this method increases the volume of informative data in variability but with less collinearity among the variables. Moreover, the method has more degrees of freedom and efficiency. The data have been analyzed by using a panel data analysis and EViews, statistical software for econometric analysis (Table 1). Historical trade data at the two-digit Standard International Trade Classification (SITC) level from 2008 to 2018 have been obtained from the United Nations Comtrade Database (comtrade.un.org), whereas data for real GDP, per-capita GDP, and population have been collected from the International Financial Statistics of the International Monetary Fund, Eurostat, and other

Table 1 Summary of data. Variable

Description

Obs

Mean

Standard deviations

EXPijt

Textile export value (USD millions) Gross domestic product of the exporting country (USD millions) Per-capita GDP of the exporting country (USD millions) Gross domestic product of the importing country (USD millions) Per-capita GDP of the importing country (USD millions) Distance between the capitals of two countries (km) Population of the exporting country (number) Logistics Performance Index of the exporting country Liner Shipping Bilateral Connectivity Index of the exporting country

176

8.24  108

2.28  109

176

7.63  1011

2.28  1012

176

10,272

15,783

176

1.7  1013

1.96  1012

176

54,071

4599

176

21,835

1243

176

1.48  108

3.21  108

73

3.013

0.567

154

0.422

0.158

GDPit PCGDPit GDPjt PCGDPjt Dij POPjt LPIit LSBCIit

18

Market development and policy for one belt one road

Table 2 LPI rank and score of BRI countries/regions. 2018

2007

Country/region

LPI rank

LPI score

LPI rank

LPI score

Singapore Hong Kong, China South Korea China Thailand Vietnam Malaysia Indonesia The Philippines Brunei Sri Lanka Cambodia Bangladesh Nepal Pakistan Myanmar

7 (6) 12 (4) 25 (0) 26 (+4) 32 (1) 39 (+14) 41 (14) 46 (3) 60 (+5) 80 (/) 94 (2) 98 (17) 100 (13) 114 (+16) 122 (54) 137 (+10)

4.00 3.92 3.61 3.61 3.41 3.27 3.22 3.15 2.90 2.71 2.60 2.58 2.58 2.51 2.42 2.30

1 8 25 30 31 53 27 43 65 / 92 81 87 130 68 147

4.19 4.00 3.52 3.32 3.31 2.89 3.48 3.01 2.69 / 2.40 2.50 2.47 2.14 2.62 1.86

(0.19) (0.08) (+0.09) (+0.29) (+0.10) (+0.38) (0.26) (+0.14) (+0.21) (/) (+0.20) (+0.08) (+0.11) (+0.37) (0.20) (+0.44)

Remarks: LPI denotes the Logistics Performance Index. () is the change of LPI ranks/scores from 2007 to 2018. Source: The World Bank (https://lpi.worldbank.org/).

relevant sources. LPI and LSBCI data have been collected from the open database of the World Bank (http://data.worldbank.org) and the UNCTAD (https://unctadstat.unctad.org/), respectively. Improving a country’s overall logistics performance could take a long time and great effort. Table 2 shows that the five BRI countries (namely, China, Vietnam, the Philippines, Nepal, and Myanmar) increased both the LPI rank and score in 2018, compared to their performance in 2007 (see Table 2). It is hypothesized that the BRI, which heavily promotes infrastructure investment projects, has helped to enhance the shipping connectivity and to speed up the logistics performance improvement of developing economies and expand their capacity to accommodate the higher export volume.

5. Discussion and policy implications The role of the LPI is significant in determining the success factors of the BRI as the index calculates and formulates to showcase the potential of different countries in terms of their strengths and weaknesses. The use of the

Maritime connectivity and global trade

19

LPI can demonstrate the potential countries that can grow economically through international trade relations and highlight the infrastructure quality for land and sea transportation. According to the Logistic Performance Index of the World Bank (2018), countries such as Germany, Sweden, Belgium, Australia, and Japan are the top five highest LPI-ranking countries that have an LPI score of 4, thus indicating that these countries have the facilities to face challenges and grasp opportunities that are presented to them through trade logistics compared to 160 other countries. This also shows that these countries have a better ability to trace and track different consignments and are effective in terms of completing their shipping consignments on time. Thus, they are ideal for logistics purposes. On the other hand, countries with the lowest LPI score include Afghanistan, Burundi, and Niger, which is on average about 1.70. As such, these countries need much more investment in their logistics infrastructures to create a stable logistic import-export network, and these countries will require much more time to develop into a logistic port in the long run. Most of these countries are still in the developing stages and still have many social and economic issues to address. Therefore, it is important to also keep in mind the logistics performance of these countries and evaluate the amount of investment to bring in a global network of sea and land transport. Considering this, these countries will require more investment, and the output would not be the same as that of the top 10 countries with the highest values of LPI. Hence, the impact of LPI on global trade is crucial as it quantifies the relationships among multiple countries through the quality of their logistics, which is important for the B&R as it can be a medium used to evaluate potential markets for successful trade collaborations. Along with this, LSBCI is an important guide that can affect the BRI and international trade. The LSBCI puts forth the core determination of joint exports. According to UNCTAD (2019), Korea and China are one of the best-paired countries based on LSBCI. The Hong Kong and China connectivity of liner shipping networks has always been rated as one of the top 5 pairs among different paired countries. The connectivity in the tech score showcases their strengths and how they have improved the connectivity scores over time compared to countries that have a lower connectivity index (UNCTAD, 2019). The outcome of the connectivity model of the predefined impacts shows that the interaction of the B&R countries with China helps them to develop. This offers an objective proof of the important effect of connectivity among B&R countries and China on their economic development. The logistics business supply chain has become a major focal point

20

Market development and policy for one belt one road

in the supply chains with the growth of the logistics industry. According to the supply chain model of Ellram, Tate, and Bilington (2006), the management of resources is important in order to meet customer requirements, provide good customer service, manage manufacturer relations, and manage the delivery of services as well as the financial flow statements. Today, logistics and supply chain research mainly focuses on vendors, manufacturers, supplier evaluation, and the framework for a collaborative performance review (Liu, Zhang, Chen, Zhou, & Miao, 2018). As a result, domestic logistics and international transport connectivity provide the foundations for the success of economies. They provide quantifying variables to predict the future of the BRI while keeping LPI and LSBCI in mind to allocate resources more intentionally. We examine the effects of LPI and LSBCI of the country on the change of textile trade using models and report the results in Table 3 (with year dummies omitted for simpler presentation). The values of R-squared, adjusted R-squared, and F-statistic show that the functional specifications and the overall fitness of all the gravity models of textile trade are acceptable. Model 1 (the traditional gravity model) and Models 2–5 (extended gravity model) are considered a good fit. As models are in logarithms, we can interpret the coefficients as elasticities. By comparing columns of Table 3, we can find that the five functional specifications agree with each in general, except for the sign of GDP. Coefficients are statistically significant at the 0.05 level or lower. Consistent with gravity theories of trade, the regression results show that textile exports increase with GDP, PCGDP, population, LPI, and LSBCI of exporting countries. Distance has a negative impact on textile exports. The results suggest that textile trade is expected to rise as a result of growth in the GDP and PCGDP, population, and the improvement of shipping connectivity and logistics performance of the exporting countries. The coefficients of GDP and PCGDP of the importer country (or United States) are not statistically significant and so are not reported here. Therefore, as textiles are an intermediate good, their trade can be explained in terms of the gravity model. We have added two new variables (LPI and LSBCI) in the models in order to test the influences of domestic logistics performance and international transport connectivity, respectively, on the textile trade. The coefficients of LPI and LSBCI agree with a-priori hypothesis. We find a statistically significant positive association between LPI and EXP (export values) and between LSBCI and EXP. Because the definitions of LPI and LSBCI are not consistent, we conclude which factors have a greater

Table 3 Gravity models for LPI and LSBCI.

Dependent variable ln(EXPijt) Independent variables ln(GDPit)

Model 1

Model 2

Model 3

Model 4

Model 5

Basic Model

LPI

LPI x PCGDP

LSBCI

LPI x PCGDP and LSBCI

0.767*** (0.127)

0.899** (0.470)

2.022** (0.791)

2.512*** (0.568)

6.444** (2.787) 0.942*** (0.112)

9.341** (3.632) 2.061*** (0.395) 3.888*** (1.014)

7.784** (3.610) 3.337*** (0.743)

4.779*** (1.483) 3.959*** (1.439)

5.640*** (1.457)

3.313*** (0.512)

13.394*** (1.303) 3.933 (2.977) 0.834 0.817 0.000 154

0.270*** (0.075) 11.261*** (1.719) 12.894*** (3.946) 0.763 0.728 0.000 63

ln(PCGDPit) ln(Dij) ln(POPjt) LPIit LPIit x ln(PCGDPit)

0.407*** (0.112)

LSBCIit Constant R-squared Adj R-squared F-statistic Observations

45.573 (28.543) 0.714 0.691 0.000 176

Remarks: () are standard errors. Year dummies are not reported in the table. **Statistically significant with P-value 0, in the case of symmetric retailers: c ¼ 0. Thus, in this paper, we assumed that the operational cost of both the liner companies’ and the terminals’ are normalized to zero.

3.1 Uniform pricing In this scenario, there is none of collusion. In the first stage, two container terminals adopt centralized pricing, the container port sets THC by maximizing its own profit, complied by the two container terminals, that is f1 5 f2 5 f. In the second stage, line chooses freight rate for its shippers. LC1 and LC2 maximize their operation profits, respectively, π L1 ¼ ðp1  f Þq1

(10)

π L2 ¼ ðp2  f Þq2

(11)

Shippers surplus function Eq. (1) also gives rise to a linear demand structure,
 2t p1 ¼ 1  τq2  q1 1 + s1
 2t p2 ¼ 1  τq1  q2 1 + s2 According to the above two equations, we can reach the following conclusion, 0 < p1 < 1, 0 < p2 < 1 To ensure that the profit of liner shipping company is greater than zero, THC needs to meet the following conditions, 0 < f < p1 < 1, 0 < f < p2 < 1 By taking the derivation of Eqs. (10) and (11), the first-order conditions are given by,

Collusive pricing detection in ocean container transport

135

∂π L1 ½2ð1 + f Þt + ð1 + f  τ + τp2 Þs2  2p1 ð2t + s2 Þs1 ¼ ∂p1 4t2 + 2t ðs1 + s2 Þ + ð1  τ2 Þs1 s2

(12)

∂π L2 ½2ð1 + f Þt + ð1 + f  τ + τp1 Þs1  2p2 ð2t + s1 Þs2 ¼ ∂p2 4t2 + 2t ðs1 + s2 Þ + ð1  τ2 Þs1 s2

(13)

Similarly, the second-order optimality conditions can be written as, ∂2 π L1 2s1 ð2t + s2 Þ ¼ 2 4t + 2tðs1 + s2 Þ + ð1  τ2 Þs1 s2 ∂p21

(14)

∂2 π L2 2s2 ð2t + s1 Þ ¼ 2 2 4t + 2tðs1 + s2 Þ + ð1  τ2 Þs1 s2 ∂p2

(15)

∂2 π L1 ∂2 π L2 τs1 s2 ¼ ¼ 2 ∂p1 p2 ∂p2 p1 4t + 2tðs1 + s2 Þ + ð1  τ2 Þs1 s2

(16)

Since τ  [0, 1], we have, ∂2 π L1 ∂2 π L2 ∂2 π L1 ∂2 π L2 ∂2 π L1 ∂2 π L2 , < 0, ∗  ∗ ∂p21 ∂p22 ∂p21 ∂p22 ∂p1 p2 ∂p2 p1 ¼

½4t + ð2  τÞs1 ½4t + ð2 + τÞs1 s22 >0 f2tð2t + s2 Þ + s1 ½2t + ð1  τ2 Þs2 g2

(17)

The definite Hessian Matrix is negative, which derives the following solution, p1 ¼

p2 ¼

ð1 + f Þð8t 2 + 4ts1 Þ + 2t½2  τ + f ð2 + τÞs2 + ð1 + f  τÞð2 + τÞs1 s2 16t2 + 8tðs1 + s2 Þ + ð4  τ2 Þs1 s2 (18) ð1 + f Þð8t 2 + 4ts2 Þ + 2t ½2  τ + f ð2 + τÞs1 + ð1 + f  τÞð2 + τÞs1 s2 16t 2 + 8t ðs1 + s2 Þ + ð4  τ2 Þs1 s2 (19)

Therefore, we obtain the derived container demands, q1 ðf Þ ¼

q2 ðf Þ ¼

ð1  f Þð2t + s2 Þf4t ð2t + s1 Þ + ½2t ð2  τÞ + ð2  τ  τ2 Þs1 s2 gs1 ½16t 2 + 8t ðs1 + s2 Þ + ð4  τ2 Þs1 s2 ½4t 2 + 2t ðs1 + s2 Þ + ð1  τ2 Þs1 s2  (20) ð1  f Þð2t + s1 Þf4t ð2t + s2 Þ + ½2t ð2 + τÞ + ð2 + τ + τ2 Þs2 s1 gs2 ½16t 2 + 8t ðs1 + s2 Þ + ð4  τ2 Þs1 s2 ½4t 2 + 2t ðs1 + s2 Þ + ð1  τ2 Þs1 s2  (21)

136

Market development and policy for one belt one road

In the first stage, the container port maximizes its own profit with respect to THC. π B ¼ f ðq1 + q2 Þ

(22)

If α1  1, α2 ¼ 1, in the case of none of the pairs colludes, we can obtain, n h    f ¼ 82 t2 s2 ð2t þ s2 Þ þ s21 82 t2  2t 2 þ τ þ 1 4 þ τ þ τ2 s2 h   i  ð1 þ 1Þ 2 þ τ þ τ2 s22 þ 2ts1 82 1t2  2ð1 þ 1Þtð4 þ τÞs2   io n h  4 þ 1ð2 þ τÞ þ τ þ τ2 s22 = 4 42 t2 s2 ð2t þ s2 Þ     þ s21 42 t2  tð6 þ τð2 þ τÞÞs2  2 þ τ þ τ2 s22  io þ ts1 82 t 2  4tð4 þ τÞs2  ð6 þ τð2 þ τÞÞs22 α1 increases monotonically with respect to f. ∂f s1 ð2t þ s2 Þf2t ½4t þ ð2  τÞs2  þ s1 ½4t ð2  τ  τ2 Þs2 g   >0 ¼  22 ∂α1 4 4 t s2 ð2t þ s2 Þ þ s21 42 t 2 þ t ð6  2τ  τ2 Þs2 þ ð2  τ  τ2 Þs22   þts1 82 t 2 þ 4t ð4  τÞs2 þ ð6  2τ  τ2 Þs22

Especially, if α1 ¼1, through setting the first-order condition to zero, we can obtain, f ¼ 1=2

(23)

Eq. (23) indicates that THC does not change with other parameters under the assumption that the two liners are symmetrical. Because increasing of THC will increase the freight rates, decrease container demand, and further reduce profit of port. Decreasing of THC will decrease the freight rates, increase container demand, the increased benefit from increased container demand cannot compensate for the loss of benefit resulting from reducing THC. This result is the same as that of Nalca and Cai (2018), in which if the retailers are symmetric, the manufacturer’s wholesale price to two retailers is uniform, i.e., w1 ¼ w2 ¼ 1/2. By substituting Eq. (23) into Eqs. (18) and (19), the ratio of freight rates between liners can be rewritten as follows, p1 12tð2t + s1 Þ + ½2t ð6  τÞ + ð6  τ  2τ2 Þs1 s2 ¼ p2 12tð2t + s2 Þ + ½2t ð6  τÞ + ð6  τ  2τ2 Þs2 s1

(24)

Collusive pricing detection in ocean container transport

137

3.2 Only one pair collusive pricing In this scenario, for example, CT2 and LC2 collude. In the first stage, CT2 and LC2 can be regarded as the same entity. CT1 does not change THC set by the container port, f1 ¼ f ¼ 1/2. CT2 change THC to LC2, adopts decentralized pricing. In this chapter, we did not analyze the specific value of f2, but it must be able to ensure that the profit after the container terminal collusion is greater than or equal to the profit before the collusion. In the second stage, LC1 and LC2 set freight rates for the shippers. LC1 and the entity of CT2 and LC2 maximize their operation profit, respectively, π L1 ¼ ðp1  f Þq1

(25)

πM 2 ¼ p2 q2

(26)

The first-order conditions can be given by, ∂π L1 ½2ð1 + f Þt + ð1 + f  τ + τp2 Þs2  2p1 ð2t + s2 Þs1 ¼ ∂p1 4t2 + 2t ðs1 + s2 Þ + ð1  τ2 Þs1 s2

(27)

∂π M ½2t + ð1  τ + τp1 Þs1  2p2 ð2t + s1 Þs2 2 ¼ ∂p2 4t 2 + 2t ðs1 + s2 Þ + ð1  τ2 Þs1 s2

(28)

Since the definite Hessian Matrix is negative, ∂2 π L1 ∂2 π M ∂2 π L1 ∂2 π M ∂2 π L1 ∂2 π M 2 2 2 , < 0, ∗  ∗ ∂p21 ∂p22 ∂p21 ∂p22 ∂p1 p2 ∂p2 p1  2  16t + 16ts2 + ð4  τ2 Þs22 s21 ¼ >0 ð2t ð2t + s2 Þ + s1 ð2t + ð1  τ2 Þs2 ÞÞ2 which leads to the following solutions, p1 ¼

ð1 + f Þð8t 2 + 4ts1 Þ + 2tð2 + 2f  τÞs2 + ½2f + ð1  τÞð2 + τÞs1 s2 16t2 + 8t ðs1 + s2 Þ + ð4  τ2 Þs1 s2 (29)

p2 ¼

8t2 + 4ts2 + 2tð2 + τf  τÞs1 + ½τf + ð1  τÞð2 + τÞs1 s2 16t2 + 8tðs1 + s2 Þ + ð4  τ2 Þs1 s2

(30)

By substituting f ¼ 1/2 into Eqs. (29) and (30), we can obtain the ratio of freight rates between liners, p1 24t2 + 12ts1 + 4tð3  τÞs2 + 2ð3  τ  τ2 Þs1 s2 ¼ p2 16t 2 + 8ts2 + 2t ð4  τÞs1 + ð4  τ  2τ2 Þs1 s2

(31)

138

Market development and policy for one belt one road

Similarly, under CT1 and LC1 collusion pricing, the symmetric results can be written as, p1 16t2 + 8ts1 + 2tð4  τÞs2 + ð4  τ  2τ2 Þs1 s2 ¼ p2 24t2 + 12ts2 + 4tð3  τÞs1 + 2ð3  τ  τ2 Þs1 s2

(32)

3.3 Both pairs collusive pricing In this scenario, two container terminals adopt decentralized pricing. Since the two merges compete directly in freight rate, they maximize their profits with respect to the freight rate, πM 1 ¼ p1 q1

(33)

πM 2 ¼ p2 q2

(34)

First-order conditions for two mergers can be expressed as, ∂πM ½2t + ð1  τ + τp2 Þs2  2p1 ð2t + s2 Þs1 1 ¼ ∂p1 4t2 + 2tðs1 + s2 Þ + ð1  τ2 Þs1 s2

(35)

½2t + ð1  τ + τp1 Þs1  2p2 ð2t + s1 Þs2 ∂πM 2 ¼ ∂p2 4t2 + 2tðs1 + s2 Þ + ð1  τ2 Þs1 s2

(36)

Since the definite Hessian Matrix is negative, ∂2 π M ∂2 π M ∂2 π M ∂2 π M ∂2 π M ∂2 π M 1 2 1 1 2 , < 0, ∗ 22  ∗ 2 2 2 ∂p1 ∂p2 ∂p1 ∂p2 ∂p1 p2 ∂p2 p1  2  16t + 16ts2 + ð4  τ2 Þs22 s21 ¼ >0 ð2tð2t + s2 Þ + s1 ð2t + ð1  τ2 Þs2 ÞÞ2 which causes the following solutions, p1 ¼

8t2 + 4ts1 + 2tð2  τÞs2 + ð2  τ  τ2 Þs1 s2 16t2 + 8tðs1 + s2 Þ + ð4  τ2 Þs1 s2

(37)

p2 ¼

8t2 + 4ts2 + 2tð2  τÞs1 + ð2  τ  τ2 Þs1 s2 16t2 + 8tðs1 + s2 Þ + ð4  τ2 Þs1 s2

(38)

Accordingly, we obtain the ratio of freight rates between liners, p1 8t2 + 4ts1 + 2tð2  τÞs2 + ð2  τ  τ2 Þs1 s2 ¼ p2 8t2 + 4ts2 + 2tð2  τÞs1 + ð2  τ  τ2 Þs1 s2

(39)

Collusive pricing detection in ocean container transport

139

4. Analysis of the collusive pricing solutions NN NC Let pNN 1 and p2 be the freight rates of LC1 and LC2 without collusion, p1 and pNC denote the freight rates of LC1 and LC2 when only CT2 and LC2 2 CN collude, while pCN 1 and p2 express the freight rates of LC1 and LC2 under CC collusion between CT1 and LC1, and pCC 1 and p2 the freight rates of LC1 and LC2 when both pairs collude.

Proposition 1 In the scenarios where none of the pairs colludes and both pairs collude, the freight rate of the liner shipping company depends on the design capacity of the corresponding container terminal. Larger design capacity of the container terminal, higher freight rate of the corresponding liner. Proof Subtracting the numerator from the denominator of Eq. (24), we can obtain the difference between the freight rates of the liners, 12t ð2t + s1 Þ + 2t ð6  τÞs2  12tð2t + s2 Þ  2t ð6  τÞs1 ¼ 2tτðs1  s2 Þ pNN 1 ¼ 1 ! pNN ¼ pNN 1 2 pNN 2 pNN > 0ðif s1 > s2 Þ ! 1NN > 1 ! pNN > pNN 1 2 p2 pNN < 0ðif s1 < s2 Þ ! 1NN < 1 ! pNN < pNN 1 2 p2 ¼ 0ðif s1 ¼ s2 Þ !

Accordingly, subtracting the numerator from the denominator of Eq. (39), we find that, 4t ð2t + s1 Þ + 2t ð2  τÞs2  4t ð2t + s2 Þ  2tð2  τÞs1 ¼ 2tτðs1  s2 Þ pCC CC 1 ¼ 1 ! pCC 1 ¼ p2 pCC 2 pCC CC > 0ðif s1 > s2 Þ ! 1CC > 1 ! pCC 1 > p2 p2 pCC CC < 0ðif s1 < s2 Þ ! 1CC < 1 ! pCC 1 < p2 p2 ¼ 0ðif s1 ¼ s2 Þ !

140

Market development and policy for one belt one road

LC1’s price elasticity of demand, ed1 ¼ 

∂Q1 p1 12  t ð2  t + s1 Þ + ð2  t ð6  τÞ + ð6  τ  2τ2 Þs1 Þs2 ∗ ¼ >1 ∂p1 Q1 4  t ð2  t + s1 Þ + ð2  t ð2  τÞ + ð2  τ  τ2 Þs1 Þs2

LC2’s price elasticity of demand, ed2 ¼ 

∂Q2 p2 12  t ð2  t + s2 Þ + s1 ð2  tð6  τÞ + ð6  τ  2τ2 Þs2 Þ ∗ ¼ >1 4  t ð2  t + s2 Þ + s1 ð2  tð2  τÞ + ð2  τ  τ2 Þs2 Þ ∂p2 Q2

ed1  ed2 ¼ 

2  tτðs1  s2 Þð8  t ð2  t þ s2 Þ þ s1 ð8  t þ ð4  τ2 Þs2 ÞÞ ð4  tð2  t þ s1 Þ þ ð2  tð2  τÞ þ ð2  τ  τ2 Þs1 Þs2 Þ ð4  tð2  t þ s2 Þ þ s1 ð2  tð2  τÞ þ ð2  τ  τ2 Þs2 ÞÞ

Proposition 1 shows that larger design capacity of the container terminal leads to higher freight rate of the corresponding liner. There are two main causes: one is ed1 > 1, ed2 > 1, another is that if s1 > s2, ed1 < ed2; if s1 < s2, ed1 > ed2; if s1 ¼ s2, ed1 ¼ ed2. According to the west economics theory, ed > 1 indicates that the rate of change in demand is greater than the rate of change in price. For goods with ed > 1, lowering the price will increase the sales revenue of the manufacturer. In the case of s1 > s2, ed1 < ed2, CT2 is more flexible than CT1, the increase in sales revenue caused by CT2 lowering the freight rate is higher than the increase in sales revenue caused by CT1 lowering the freight rate. So, in the case of s1 > s2, the freight rate of CT2 is lower than the freight rate of CT1. Thus, larger design capacity of the container terminal leads to higher freight rate of the corresponding liner. Proposition 2 In the scenario where only one pair colludes, the liner shipping company colluding with the corresponding container terminal chooses a lower freight rate than the liner company that does not, no matter what the designed capacity of container terminal is. Proof By subtracting the numerator from the denominator of Eq. (31), there exists, 8t2 + ð4t + 2tτÞs1 + ð4t + 4tτÞs2 + ð2  τÞs1 s2 ¼ 8t2 + 2tð2 + 2τÞs1 + 4t ð1  τÞs2 + ð2  τÞs1 s2 pNC > 0 ! 1NC > 1 ! pNC > pNC 1 2 p2

Collusive pricing detection in ocean container transport

141

Similarly, subtracting the numerator from the denominator of Eq. (32), we have, 8t2 + ð4tτ  4tÞs1  ð2tτ + 4tÞs2 + ðτ  2Þs1 s2 ¼ 8t2  4tð1  τÞs1  2tð2 + τÞs2  ð2  τÞs1 s2 pNC < 0 ! 1NC < 1 ! pNC < pNC 1 2 p2 The collusion between container terminal and liner shipping company leads to a lower freight rates than doesn’t collude. The reason is that collusion eliminates the double marginal effect (see Barbot et al., 2013). Colluding members seek to maximize profits together, rather than each member only considers its own marginal benefits when making price decisions. When there is no collusion between the upstream and downstream members, the unilateral decision will affect the market demand, and thus reducing the profits of each party. Therefore, the vertical collusion between container terminal and liner shipping company leads to a lower freight rate. In addition, there is horizontal competition between liners, which intensifies the decrease of freight rates and leads to lower freight rates.

5. Case study We take a case study to test the liner routes departing from Shanghai Port to the Mediterranean along the MSR. There are two main container terminals in Shanghai Port: Shengdong International Container Terminal (SDCT) and Guandong International Container Terminal (GDCT). According to the official website, the designed capacity of SDCT and GDCT is 4.3 and 5.0 million TEUs, respectively. In addition, the value of containerized cargo waiting cost is usually set to be 22 RMB of Shanghai Port. We take the general handling charge for a foreign trade container as the THC, set by the container terminal to the liner. According to operation lump sum issued by Shanghai International Port Group on March 1, 2016, the value of THC is assumed to be 930 RMB/TEU. Scenario of uniform pricing gives us the value of f as 1/2. Thus, the value of t in this study is set to be 0.01 (1/2 ∗ 22/930). By comparing the liner routes of SDCT and GDCT, we find their major liner routes are to the Mediterranean and Europe. Among which, the dominant liners of SDCT are China Ocean Shipping Group Company (COSCO Shipping), CMA-CGM Group and Evergreen Marine Corp, as well as

142

Market development and policy for one belt one road

OOCL, belonging to OCEAN Alliance. However, the leading liners of GDCT are Maersk Line (MSK) and Mediterranean Shipping Co. (MSC), attached to 2M Alliance. Therefore, the liners selected in our test are COSCO Shipping, ranking the first in China, and MSK at the top of the world. In this study, we test collusion between container terminal and liner through the ratio range of freight rates between liners. Considering the different competitive degree, we substitute the specific value of two container terminals into our model to obtain the ratio range, which is shown in Table 2. As shown in Table 2, we classify the ratio range into three different cases: low-degree competition (LD) with 0 < τ  0.3, medium-degree (MD) competition with 0.3 < τ  0.7, and high-degree competition (HD) with 0.7 < τ < 1. We set the maximum and minimum values in each case as the boundaries of each ratio range. Then, we apply our test to SDCT and COSCO Shipping, and GDCT and MSK, respectively. Since liners have differentiated rates at different sailing times, we collect 24 pairs of freight rates of liner routes from the Far East to the Mediterranean along the MSR (http://shipping.jctrans.com) (see Appendix A, March 20, 2019), calculating the daily average freight rate by simple arithmetic mean and its ratio, which is shown in Table 3. As shown in Table 3, both liners depart from the Port of Shanghai, i.e., COSCO Shipping departs from the container terminal of SDCT, while MSK’s departure is from the container terminal of GDCT. Their routes are from the Far East to the Mediterranean along the MSR. With regard to the test of collusion between container terminal and liner, we use the ratio of freight rates minus the maximum and minimum of the ratio range, respectively, in Table 2, and then get the minimum of the absolute values (see Appendix B). It is important to note that if the ratio of freight rates is within the range, then the value is set to be zero. We analyze the results according to two Table 2 The ratio range of freight rates between liners. Ratio range

0 < τ ≤ 0.3

0.3 < τ ≤ 0.7

0.7 < τ < 1

NN NC CN CC

0.99997–1.00000 1.48259–1.50000 0.66667–0.67442 0.99988–1.00000

0.99990–0.99997 1.48284–1.55854 0.64134–0.67430 0.99945–0.99988

0.99989–0.99990 1.55854–1.98191 0.50392–0.64134 0.95119–0.99945

Collusive pricing detection in ocean container transport

143

Table 3 Daily average freight rate and its ratio between liners. Daily average freight rate (20 ft) Number

Destination

COSCO Shipping (USD/day)

MSK (USD/day)

Ratio of freight rates

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Istanbul Latakia Limassol Venice Beirut Koper Catania Gebze Rijeka Valencia Naples Fos Misrata Livorno Malta Ancona Leixoes Marsaxlokk Ravenna Algecias Barcelona Napoli Skikda Iskenderun

34.47312 36.71875 34.46535 33.59296 46.02041 35.43333 42.00000 37.27273 28.75000 30.52000 24.21875 25.70896 45.84375 32.62626 33.06306 28.70000 36.36364 31.36364 27.28571 40.60606 30.27778 25.90361 37.43750 33.36283

45.37647 49.03846 44.29268 35.01036 35.93197 43.48299 31.35088 34.04255 40.08065 31.48413 30.55556 27.14173 49.62857 29.80583 47.87879 45.73034 32.57143 30.61224 40.64516 33.53125 41.42857 32.82828 44.82895 49.93103

0.75971 0.74877 0.77813 0.95951 1.28076 0.81488 1.33968 1.09489 0.71730 0.96938 0.79261 0.94721 0.92374 1.09463 0.69056 0.62759 1.11643 1.02455 0.67132 1.21099 0.73084 0.78906 0.83512 0.66818

different criteria: the value is set to be smaller than 1% or 3% (see Barbot et al., 2013), which is shown in Table 4. As shown in Table 4, we find that CT1 and LC1 have a greater likelihood of vertical collusion under the low or medium level competition. Since CT2 has a higher designed capacity than CT1, the later has a strong incentive to collude with LC1 in order to deal with their dominant competitors. Compared to CT1 and LC1, the pair of LC2 and CT2 in an advantageous position has no motivation to collude if its weak competitor does not choose vertical collusion; the pair of LC2 and CT2 may choose collusion if its weak competitor choose vertical collusion. It should be noted that there is a greater likelihood of collusion on both pairs if high-degree competition exists in market.

Table 4 Results of collusive pricing between container terminals and liners. Case 1/LD Values

NN

Case 2/MD

Case 3/HD

NC

CN

CC

NN

NC

CN

CC

NN

NC

CN

CC

0 1

0 0

2 3

0 2

0 2

0 0

3 3

0 2

0 2

0 0

1 1

3 5

0 4.2

0 0

8.3 12.5

0 8.3

0 8.3

0 0

12.5 12.5

0 8.3

0 8.3

0 0

4.2 4.2

12.5 20.8

Number of cases

d, this implies that an increase in the total capacity increases the total demand for local and non-local goods.

3.1 Best capacity responses To derive the best responses in terms of investments, it is useful to write their local welfares depending on the total capacity by substituting ki by K  kj. This yields W i ðKÞ ¼ Bi ðDii ðKÞ, Dji ðKÞÞ  Dji ðKÞvTðDji ðKÞÞ  r i  ðK  kj Þ: (9) The first two terms on the right-hand side can be written as Bi ðDii ðKÞ, Dji ðKÞÞ  Dji ðKÞvTðDji ðKÞÞ a2 ð2ðb  dÞ2 ðb + dÞK 2 + 2ðb  dÞðb + dÞKv + bv2 Þ : ¼ 2ððb  dÞðb + dÞK + bvÞ2

(10)

Because the two trading countries’ benefits and transit costs are symmetric, the right-hand side of Eq. (10) representing the difference between country i’s benefits and transit costs in demand equilibrium is equal for the two trading countries. Thus, the only difference in local welfares can arise from differences in the countries investments k1 and k2 and from differences in the countries marginal investment costs r1 and r2. This observation is useful for the derivation of the countries’ best response functions in terms of their own investments depending on the other countries’ investments denoted as kbri ðkj Þ for j6¼i. The best response functions in terms of own investments can be described as  
 br ki ðkj Þ ¼ max 0, arg max W i ðKÞ  kj : (11) K

If the other country invests a sufficiently large capacity, then the best response of the own country will be to invest a zero amount. However, if the other country invests a sufficiently small amount, then the own best response is determined by the first-order condition for an interior capacity solution, W 0i ðKÞ ¼ 0, which can be written as ! ∂Bi 0 ∂Bi 0 D ðKÞ+  vT ðDji Þ  Dji ðKÞvT ðDji Þ D0ji ðKÞ  r i ¼ 0 (12) ∂qii ii ∂qji

160

Market development and policy for one belt one road

The first term on the left-hand side shows the benefit losses from capacity expansion associated with the reduced consumption of local goods. The second term shows the net benefit gain from capacity expansion associated with the increased consumption of non-local goods. The third term captures the marginal capacity cost. The strict concavity of the local welfare function Wi(K) is ensured for K > vb/2(b2  d2). In this sense, the existence of an interior solution for the best response function kbri ðkj Þ, which is unique in the case of existence by the strict concavity, requires sufficiently low marginal investment costs. The solution for this condition, W 0i ðKÞ ¼ 0, is huge and, therefore, omitted. If countries invest into an equal amount of infrastructure capacity, local welfares differ only if the marginal investment costs differ between the trading countries. This observation together with the strict concavity of the local welfare function implies that the optimal total capacity that maximizes the own country’s welfare is higher or lower than the corresponding other country’s optimal total capacity depending on whether the own country’s marginal investment costs are lower or higher, respectively, than the other country’s marginal investment costs. Fig. 1 illustrates best responses for the case in which country 1’s marginal investment costs are lower than country 2’s marginal investment costs, that is, r1 < r2. Country 1’s best responses are depicted by the dotted line, whereas country 2’s best responses are depicted by the dashed line. Consider the dashed line. If k1 ¼ 0, then kbr2 > 0 and reaches its maximum. If k1 increases, then kbr2 decreases by an equal amount. This is because the total

br

k1 k2

br

k2

k1

Fig. 1 The trading countries’ best response functions in terms of investments for r1 < r2. Country 1’s best responses are depicted by a dotted line, whereas country 2’s best responses are depicted by a dashed line.

An infrastructure investment game

161

capacity that maximizes country 2’s welfare is unchanged by country 1’s investment. If k1 is sufficiently large, then country 2’s best response is to invest a zero amount as represented by the horizontal part of the dashed line. Consider the dotted line. The shape of this best response function is similar to the dotted line. If k2 ¼ 0, then kbr1 > 0 and reaches its maximum, whereas if k2 increases, then kbr1 decreases by an equal amount. If k2 is sufficiently large, then country 1’s best response is to invest a zero amount as represented by the vertical part of the dotted line. The main difference is, however, that country 1 has lower marginal investment costs, which translate into a higher optimized value of the total capacity and, therefore, a higher best response in the sense that kbr1 ðxÞ > kbr2 ðxÞ for x sufficiently small to ensure kbr2 ðxÞ > 0.

3.2 Equilibrium investments Fig. 1 implies that it is useful to distinguish between two cases, r1 ¼ r2 and r16¼r2. The first case, r1 ¼ r2, implies that the best response functions of the two countries are overlapping in the sense that kbr1 ðxÞ ¼ kbr2 ðxÞ for x sufficiently small to ensure kbri ðxÞ > 0. In this case, both countries have the same optimized value for the total capacity and all cases in which the sum of best responses are equal to this optimized value represent equilibrium solutions, that is, there is a set of equilibrium investments which is defined by the non-negativity of capacity investments and by the optimized sum of investments. The second case, r16¼r2, is of special interest because marginal investment costs are low in China due to overcapacity. In this case, the optimized total capacity of the country with the relatively low-investment costs exceeds the corresponding value of the country with the relatively high-investment costs. Without loss of generality, consider r1 < r2, which is the case depicted in Fig. 1. In this case, there is only one intersection and this intersection is defined by kbr2 ðkbr1 ð0ÞÞ ¼ 0 with kbr1 ð0Þ > 0. Thus, in this case, there is a unique (Nash) equilibrium solution and in this solution the country with the lower investment costs provides all the capacity and, correspondingly, the other country provides zero capacity. (Infrastructure investments might be associated with economies of scale. The presence of economies of scale could be another reason for equilibrium investments being carried out by only one country.) Consider the viewpoint of the aggregate economy in which investments are chosen to maximize the aggregate welfare of the two trading countries. In this case, the country with the lower marginal investment costs should still provide all the capacity, that is, k2 ¼ 0 in optimum. However, the optimal

162

Market development and policy for one belt one road

Fig. 2 Equilibrium and first-best total capacity depending on r1 for parameters: a ¼ b ¼ v ¼ 1, d ¼ 0, and K > 1/2, where the latter inequality ensures the concavity of the local welfare functions.

capacity that maximizes the aggregate welfare of the two trading countries is greater than the capacity that maximizes the local welfare of the country with the lower marginal investment costs. This is because an increase in the capacity increases the welfare of both countries, whereas the local welfare-maximizing country ignores the welfare gains that can be achieved by the other country given by dW2(K)/dK > 0 for k1 ¼ K. Fig. 2 illustrates the differences in equilibrium and first-best total capacities depending on r1 for parameters a ¼ b ¼ v ¼ 1, d ¼ 0, and K > 1/2, where the latter ensures the concavity of the local welfare functions. The figure shows that the firstbest capacity can be substantially higher than the equilibrium capacity. Mun (2019) highlights that transport capacity supply can be considered as a voluntary supply of a public good when the utilization is free of charge. The underinvestment in transport capacity is therefore consistent with the findings of Bergstrom, Blume, and Varian (1986), Cornes and Sandler (1996), Andreoni (1998), and Batina and Ihori (2005), among others, who showed that the voluntary provision of a public good will be too low from the viewpoint of the aggregate economy. The present analysis assumes that market sizes as measured by the parameters of the benefit function a and b are equal for the two trading countries, which imply equal volumes of imports for them. If instead one country would have a larger market in the sense that import volumes would be higher for given capacities, then this country would benefit more from capacities and this would increase the investment incentives. However, this would not reverse the equilibrium investment strategies if the differential in marginal investment costs is sufficiently high.

An infrastructure investment game

163

4. Oligopoly and import taxes Import taxes are common in international trade. The above-considered framework does not provide reasons for imposing import taxes. This is because the markets are assumed to be perfectly competitive with producer surplus equal to zero. Import taxes, which are typically used to protect the local industry play no role in such an environment. To analyze the capacity investments in an environment in which import taxes could play a strategic role, this section considers oligopolistic market structures. Assume that there is one firm in each country and that this firm produces goods for local consumption, qii, and for exports, qij. Producers charge prices pii and pij for their products with j6¼i. Only country 1 invests in capacity (i.e., K ¼ k1) in anticipation of the corresponding pricing equilibrium, which therefore involves a sequential game structure, whereas country 2 uses a given import tax denoted by t2.

4.1 Demand functions Demands are determined by the equilibrium conditions ∂Bi ∂B1 q ¼ pii for i ¼ 1, 2,  p21  v 21 ¼ 0 ∂qii ∂q21 K   ∂B2 q and  p12 + t2 + v 12 ¼ 0: ∂q12 K

(13)

The first condition captures that the consumers save transport costs and taxes if they buy local products. The second condition captures that the consumers in country 1 need to pay full price including transport costs for the consumption of the non-local good. Because country 2 uses an import tax, the customers in country 2 pay full price which equal to the sum of transport costs and the import tax for the consumption of the non-local good. Simultaneously solving the equilibrium conditions yields country 1’s demand for local goods, which can be written as  v  v a bd +  b+ p + dp21 K K 11 D11 ðp11 , p21 Þ ¼ : (14) v b2  d 2 + b K The right-hand side is decreasing in the local good’s price, p11, and increasing in the non-local good’s price, p21. Country 1’s demand for non-local goods can be written as

164

Market development and policy for one belt one road

aðb  dÞ  bp21 + dp11 : (15) v b2  d2 + b K The right-hand side is increasing in the local good’s price, p11, decreasing in the non-local good’s price, p21, and increasing in the transport capacity, K. Country 2’s demand for non-local goods can be written as D21 ðp11 , p21 Þ ¼

aðb  dÞ  bðp12 + t2 Þ+dp22 : (16) v b2  d2 + b K The right-hand side is increasing in the local good’s price p22, decreasing in the non-local good’s price and the import tax, p12 and t2, respectively, and increasing in the transport capacity, K. Country 2’s demand for local goods can be written as   v v a bd +  p22 b + + dðp12 + t2 Þ K K : (17) D22 ðp12 , p22 Þ ¼ v b2  d2 + b K The right-hand side is decreasing in the local good’s price, p22, and increasing in the non-local good’s price and import tax, p12 and t2, respectively. D12 ðp12 , p22 Þ ¼

4.2 Equilibrium prices Profits denoted by π i can be written as π i ðp11 , p12 , p21 , p22 Þ ¼ pii Dii ðpii , pji Þ+pij Dij ðpjj , pij Þ

(18)

for i ¼ 1, 2 and j6¼i. Best responses are determined by the first-order conditions ∂π i/∂pii ¼ Dii + pii∂Dii/∂pii ¼ 0 and ∂π i/∂pij ¼ Dij + pij∂Dij/∂pij ¼ 0. Solving these first-order conditions simultaneously yields country 2’s equilibrium price for the non-local good v aðb  dÞð2b + dÞ  ð2b2  d 2 Þt2 + ð2bða  t2 Þ  adÞ K : (19) pN 12 ðt 2 Þ ¼ v 2 2 4b  d + 4b K The right-hand side is decreasing in the import tax. Taking the derivative of the right-hand side with respect to the import tax shows that the reduction in the import price associated with a marginal tax increase is strictly smaller than 1/2 in absolute values. An increase in the import tax, therefore, increases the non-local good’s full price despite the associated price reduction.

An infrastructure investment game

165

Simultaneously solving the first-order conditions for the local price p22 yields v aðb  dÞð2b + dÞ+bdt2 + 2ab N K : (20) p22 ðt2 Þ ¼ v 4b2  d 2 + 4b K The right-hand side is increasing in the import tax. This is because the import tax increases the non-local good’s full price, which increases the demand for the local good. The corresponding equilibrium prices for country 1 can be obtained by replacing the import tax t2 by zero because N N N pN 11 ¼ p22 ð0Þ and p21 ¼ p12 ð0Þ. One can show that all equilibrium prices are decreasing in the transport capacity. An increase in capacity reduces the full prices of non-local goods, which increase their demands for given prices. As a result, one might actually expect that the equilibrium prices are increasing in the capacity. However, this is not the case. The reason can be understood by considering the absolute value of the semi-price elasticity of country 2’s demands for local goods, which  can be written as D22 =ð∂D22 =∂p22 Þ ¼ a  p22  ða  p12  t2 Þ= b + Kv . The right-hand side is increasing in the transport capacity. In this sense, an increase in transport capacity leads to a more aggressive response from local producers to import prices, which intensify competition and altogether reduces equilibrium prices. (That congestion can soften competition in the case of congested transport infrastructure has been shown before by, for example, De Borger and Van Dender (2006).) Using equilibrium prices, country 2’s demand for non-local goods can be written as 0 1 1 aðb  dÞ  bt 2bða  t Þ+ad 2 2 N @ A D12 ðpN + 12 ðt 2 Þ, p22 ðt 2 ÞÞ ¼ v : (21) 2 2 3 b2  d2 + b v 4b  d + 4b K K An increase in the import tax reduces the demand for a given price of the nonlocal good despite the negative effect of the import tax on the equilibrium price for the non-local good because it increases its full price. Using equilibrium prices again, country 2’s demand for local goods can be written as  v  v b+ ðaðb  dÞð2b + dÞ+bdt2 Þ+2ab N K : K D22 ðpN 21 ðt 2 Þ, p22 ðt 2 ÞÞ ¼ v  2 v 2 2 2 b d +b 4b  d + 4b K K (22)

166

Market development and policy for one belt one road

The right-hand side is increasing in the import tax because the tax increases the non-local good’s full price. Equilibrium demands are displayed for country 2 only. To obtain the corresponding demand results for country 1, use the results for country 2 and replace the import tax t2 by zero again, N N N N N which leads to D11 ðpN 11 , p21 Þ ¼ D22 ðp21 ð0Þ, p22 ð0ÞÞ , and D21 ðp11 , p21 Þ ¼ N N D12 ðp11 ð0Þ, p21 ð0ÞÞ.

4.3 Welfare assessment Consider country 1. This country’s welfare evaluated at equilibrium prices, denoted by W1, can be written as    N N N W 1 ¼ B1 D11 pN 11 , p21 , D21 p11 , p21   N   N N  N N N þ pN D21 pN 11 , p21 p21 þ vT D21 p11 , p21 12 D12 p12 , p22  r 1 K: (23) The first term on the right-hand side reflects the locals’ benefits from the consumption of local and non-local goods. The second term reflects the payments for the non-local goods in terms of the full price, which includes the non-local goods’ price and transport costs. The third term reflects the export profit, whereas the final term captures the capacity costs. For this country, the import tax imposed by country 2 unambiguously reduces local welfare for a given capacity because it reduces both the export price and export demand and, thus, export profits. Consider country 2. This country’s welfare, denoted by W2, can be written as    N N N W 2 ¼ B2 D12 pN (24) 12 , p22 , D22 p12 , p22  N N  N   N N D12 p12 , p22 p12 þ vT D12 p12 , p22  N N  N N þ pN 21 D21 p11 , p21 þ t 2 D12 p12 , p22 : The meanings of the first three terms on the right-hand side are analogous to the corresponding ones in the welfare function of country 1. One difference between welfare functions is that country 2 generates revenue from the import tax, which is captured by the fourth term on the right-hand side. A second difference is that capacity costs are zero. Because the price of local goods and the full price of the non-local goods are both increasing in the import tax, the import tax unambiguously reduces consumer surplus. Altogether, the effect of import taxes on welfare can therefore be positive or negative.

An infrastructure investment game

167

0.20 p22q22 0.15 0.10 p12q12

0.05

t2 0.05

0.00

0.10

0.15

0.20

0.25

Fig. 3 Equilibrium profits.

0.50

W2

0.45 0.40 0.35 0.0

W1 0.1

0.2

0.3

0.4

0.5

0.6

0.7

t2

Fig. 4 Equilibrium welfares.

Figs. 3 and 4 are used to illustrate the effect of the import tax on profits and welfares, respectively. Parameters a ¼ b ¼ v ¼ 1, d ¼ 1/2, and r1 ¼ 1/20 are used to produce these figures. Consider the solid lines for which capacity is given and equal to K ¼ 1. Fig. 3 displays equilibrium profits from local demands for local goods and non-local goods in country 2 depending on the import tax. Equilibrium profits are simply denoted by p22q22 and p12q12, respectively, to save notation. The figure shows how the local producer gains from the import tax, whereas the non-local producer is hurt by the import tax. Fig. 4 displays the corresponding welfares of countries 1 and 2. The figure illustrates an ambiguous effect of the import tax on the local welfare of country 2. The local welfare, W2, first increases in the import tax and then decreases in the import tax. The increase in the local welfare is based on two effects. The first effect is related to the increase in local profits as illustrated in Fig. 3. The second effect is related to congestion management. The import tax is equivalent to an infrastructure charge, which reduces congestion and the consumers’ full price of the consumption of

168

Market development and policy for one belt one road

the non-local product. Thus, a moderate level of the import tax can actually increase local profit and local welfare because it helps optimizing the use of the congested transport infrastructure. The welfare of the country which abstains from raising an import tax, W1, clearly always decreases in the other country’s import tax because of the reduction in trade profit. Consider the two-stage game in which country 1 one chooses capacity to maximize local welfare anticipating the competition with the local firm in country 2. Using Eq. (23), the capacity is determined by the first-order condition ∂W1/∂K ¼ 0. One can show that the optimal transport capacity is decreasing in the import tax. This is intuitive because an increase in the import tax reduces the profit of the non-local firm and, therefore, the exporting countries incentives to invest in transport capacity. Consider the dashed lines in Figs. 3 and 4 which display the results when capacity is endogenously determined by country 1. Comparison of the solid and dashed lines in Fig. 3 shows how the profit effects of the import tax are amplified by the presence of endogenous capacity choices. This is because an increase in the import tax reduces capacity, which increases the full price of imports, reduces the export profit, and increases the local producer’s profit. Finally, comparison of the solid and dashed lines in Fig. 4 shows how the positive effect of the import tax for the local economy can be substantially reduced by the corresponding reduction in transport capacity because it reduces country 2’s consumer surplus and export profits.

5. Conclusions This study considered three countries two of which are trading with each other, whereas the third country is located between the two trading countries and used for transit purposes only. The two trading countries independently decided about investing in the transit country’s transport capacity to facilitate trading. This setup captures that infrastructure investments such as those encompassed by the Chinese government’s Belt and Road Initiative are not relying on agreements of the main trading partners but can rather be implemented in a unilateral fashion. The scenario with perfectly competitive markets revealed that a situation in which the country with the lower capacity construction costs does all the investments in transport capacity can reflect an equilibrium investment strategy although equilibrium investments will be too low from the viewpoint of all involved regions. This demonstrates that coordinative actions among trading partners are required to increase infrastructure investments beyond the level achieved without

An infrastructure investment game

169

coordination. The scenario with oligopolistic markets revealed that import taxes reduce the incentives to invest in transport capacity. The import tax and low-investment scenario effectively reduces competition between local and non-local firms, increases the profits of local producers, and can even increase local welfare by reducing congestion. The overall effect can, however, still produce negative welfare effects for both trading partners. This suggests that import taxes should be considered with caution because they might reduce regional welfare. The present study concentrated on the finished products. Considering intermediate products and supply chains would require a closer look and, possibly, a separate study although we speculate that our results are robust in the sense that the consideration of intermediate products would not reverse our results qualitatively.

References Andreoni, J. (1998). Toward a theory of charitable fund-raising. Journal of Political Economy, 106, 1186–1213. Batina, R. G., & Ihori, T. (2005). Public goods: Theories and evidence. Berlin, Germany: Springer. Bergstrom, T., Blume, L., & Varian, H. (1986). On the private provision of public goods. Journal of Public Economics, 29, 25–49. Cornes, R., & Sandler, T. (1996). The theory of externalities, public goods, and club goods. Cambridge, USA: Cambridge University Press. De Borger, B., & Van Dender, K. (2006). Prices, capacities and service levels in a congestible Bertrand duopoly. Journal of Urban Economics, 60, 264–283. Dixit, A. (1979). A model of duopoly suggesting a theory of entry barriers. Bell Journal of Economics, 10, 20–32. Head, K., & Spencer, B. J. (2017). Oligopoly in international trade: Rise, fall and resurgence. Canadian Journal of Economics, 50, 1414–1444. Heiland, I., Moxnes, A., Ulltveit-Moe, K. H., & Zi, Y. (2019). Trade from space: Shipping networks and the global implications of local shocks. CEPR Discussion Paper No. DP14193. Mun, S. (2019). Joint provision of transportation infrastructure. Economics of Transportation, 19, 100–118. Mun, S., & Nakagawa, S. (2010). Pricing and investment of cross-border transport infrastructure. Regional Science and Urban Economics, 40, 228–240.

Further reading Brander, J. A., & Spencer, B. J. (1981). Tariffs and the extraction of foreign monopoly rents under potential entry. Canadian Journal of Economics, 14, 371–389. Sarsenbayev, M., & Veron, N. (2020). European versus American perspectives on the Belt and Road Initiative. China & World Economy, 28, 84–112.

This page intentionally left blank

CHAPTER 9

High-speed rail and air transport integration in hub-and-spoke networks: The role of airports Alessandro Avenali, Tiziana D’Alfonso, Alberto Nastasi, and Pierfrancesco Reverberi

Department of Computer, Control, and Management Engineering Antonio Ruberti, Sapienza University of Rome, Rome, Italy

1. Introduction European Union (EU) leaders have endorsed the objective of developing a EU-wide multimodal TEN-T (Trans European Network-Transport) by 2030, which will connect major airport hubs to the high-speed rail (HSR) network by 2050, when the majority of medium-distance passenger transport should go by rail (European Commission (EC), 2011, pp. 9, 19). This is because of the projected increase in demand for flights, which makes capacity at hub airports scarce and the impact of aviation on the environment a growing concern (ICAO, 2014). In this framework, air transport and HSR are not simple competitors. Indeed, air and HSR services can be complements on long-haul routes served by connecting flights through a hub airport. This complementarity creates room for cooperation between airlines and HSR operators, which indeed have signed many intermodal agreements worldwide, particularly relating to international connecting passengers. Airport managers are also interested in such agreements since they affect, among others, air traffic volumes and the demand for slots on the part of the airlines. A useful list of examples comprises partnerships subscribed in: (i) Asia, e.g., between China Railway High-Speed and China Eastern Airlines from Shangai Hongqiao International Airport (China Eastern Air-Rail Service), Taiwan High Speed Rail and EVA Air/China Airlines/China Eastern Airlines from Taoyuan International, Taipei Songshan, Kaohsiung International and Taichung airports; (ii) United States, e.g., between Amtrak and United Airlines from Newark Liberty International Airport (Acela Express); (iii) Europe, e.g., between Deutsche Bahn and Lufthansa/American Airlines/Emirates from Market Development and Policy for One Belt One Road https://doi.org/10.1016/B978-0-12-815971-2.00004-9

Copyright © 2022 Elsevier Inc. All rights reserved.

171

172

Market development and policy for one belt one road

Frankfurt Airport (AIRail), Deutsche Bahn and Singapore Airlines from Dortmund Airport (Rail&Fly), Comboios de Portugal and TAP Portugal from Lisbon, Porto and Faro airports (Rail&Fly Portugal), Austrian Federal € Railways OBB and Air Moldova from Vienna International Airport (Air & Rail Austria), SNCF and Air France from Paris Charles De Gaulle Airport (AIR&RAIL or TGVAir), Thalys and Air France from Brussels Midi Airport (AIR&RAIL), Swiss Federal Railways SBB and Swiss International Air Lines from Zurich Airport (Airtrain). In this chapter, we develop a theoretical model to study transport operators’ incentives to cooperate, and the strategic role of airports in facilitating or dampening airline-HSR cooperation. In our model, transport operators cooperate to offer a bundle of domestic HSR and international air services via a multimodal hub airport. We consider a three-stage game. At stage one, transport operators decide whether to cooperate (thereby incurring sunk costs to make their services compatible) or not. At stage two, the airport company sets the landing fee. Finally, at stage three the airline and the HSR operator set quantities in relevant markets (either cooperatively or not). We show that the scope for cooperation depends on two main factors, that is, the related sunk costs and mode substitution between air and HSR services. On the one hand, huge investments (see also Section 2) are often required to make cooperation effective, which can be a major barrier to intermodal agreements. Indeed, transport operators have to jointly incur considerable sunk costs to ensure that passengers perceive the multimodal trip as a real alternative to the connecting flight. Clearly, an intermodal agreement is incentive-compatible as long as all the involved players achieve benefits in excess of the (relevant share of the) sunk costs of cooperation. Thus, complementarity between transportation modes derives from compatibility, and compatibility is a strategic decision.a On the other hand, mode substitution affects traffic volumes and the mode share both in connecting markets, where air and HSR services mainly act as complements, and in short-haul markets, where they are competing a

Interestingly, some attempts of cooperation have not found fulfillment, e.g., the 2008 Air France talks to Veolia Transport about a rail venture (Financial Times, 2008). Air France was looking at commissioning its own high-speed trains because it had been unhappy with the quality of rail-air connections when it had bought space on existing operators’ trains (Financial Times, 2008). Within the venture, Veolia would have run trains under the Air France brand from the airline’s hub at Paris’s Charles de Gaulle airport to destinations across Europe.

High-speed rail and air transport integration

173

for passengers. Therefore, the impact of the intermodal agreement on transport operators’ profits depends on the degree of substitutability between air and HSR services, which also has an impact on the airport company that collects revenues from the landing fee charged to airlines. We find that transport operators’ incentives to cooperate may not be aligned with the airport company’s attitude toward intermodal agreements. Compared to the benchmark case of competition, under cooperation transport operators’ total profits increase whereas the airport company’s profit decreases with the degree of substitutability between air and HSR services. As to transport operators this is because, as products are stronger substitutes, the profit of the multiproduct monopoly arising after cooperation is increasingly higher than the total profit in a competitive duopoly. As to the airport company, if mode substitution is stronger than the landing fee has to be significantly reduced to avoid a massive shift of passengers from air to HSR services, particularly after an intermodal agreement. Thus, when the sunk costs are as high as transport operators would not cooperate, the airport company may find it profitable to facilitate airlineHSR cooperation, provided that mode substitution is weak. In such a case, airport managers may decide to participate in the infrastructure investments needed to make cooperation effective. On the other hand, when mode substitution is strong, the airport company may prefer to dampen airline-HSR cooperation, for instance, by influencing strategically the ease of transfer between the rail station at the airport and air terminals. This chapter is organized as follows. Section 3 discusses the incentives to airline-HSR cooperation. Section 4 introduces the model. Sections 5 and 6 deal, respectively, with the benchmark case of competition and with airlineHSR cooperation. Section 7 discusses the strategic role of airports in influencing cooperation. Section 8 contains concluding remarks and directions for future work.

2. Literature review This chapter contributes to the growing body of literature that investigates airline-HSR cooperation.b b

A burgeoning literature has investigated airline-HSR competition, both theoretically (see e.g., D’Alfonso, Jiang, & Bracaglia, 2015, 2016; Jiang & Zhang, 2016; Yang & Zhang, 2012) and empirically (see e.g., Fu, Oum, & Yan, 2014; Wan et al., 2016).

174

Market development and policy for one belt one road

Zhang, Wan, and Yang (2019) examine modal complementarity and airHSR intermodal services, together with the network feature of airline business and the impacts of HSR on airports and regional economies. The main insights include: (i) HSR can have a traffic redistribution effect on airport traffic; in particular, some primary hub airports with good air connectivity may gain traffic while others may lose traffic; (ii) to mitigate congestion at hub airports, policymakers may consider diverting some traffic to regional airports by promoting air-HSR intermodal services; (iii) similar to the impacts on airport traffic, spatial disparity of economic activities may also rise after the introduction of HSR. In general, the disparity tends to rise between the cities with HSR and those without HSR, as the former gets better accessibility. Jiang and Zhang (2014) find that airline-HSR integration improves social welfare when mode substitution in overlapping markets is sufficiently low (so that the adverse effect on competition is small), or the hub airport capacity is tight (so that integration may alleviate congestion). Xia and Zhang (2016) assume that air and HSR services are vertically differentiated. They find that after integration, when the hub capacity is tight, the airline withdraws from the market where it has less competitive advantage over HSR. They also find that airline-HSR integration is likely to improve welfare when the hub airport is capacity constrained. However, both papers do not model transport operators’ incentives to cooperate since they assume that there are no sunk costs related to airline. Moreover, they do not consider the strategic role of airports that follows from the vertical relationship with airlines. Jiang, D’Alfonso, and Wan (2017) investigate different air-rail cooperation schemes between a rail operator and either a domestic or a foreign airline in a single transportation market. Partners incur a fixed cost to make the partnership effective and such a cost is an increasing function of the cooperation level, which is endogenous in the model. The authors find that the cooperation level is lower when both partnerships coexist than when only one partnership exists. Avenali, Bracaglia, D’Alfonso, and Reverberi (2018) study the strategic formation of airline-HSR partnerships, depending on the sunk costs necessary to make cooperation effective and on transport operators’ bargaining power in negotiating agreements. The authors investigate the conditions under which either a capacity purchase or a joint venture agreement improves consumer surplus and social welfare, depending on the level of congestion at hub airports and on mode substitution between air and HSR services. Both papers, do not model the vertical relation

High-speed rail and air transport integration

175

between airports and airlines, and thereby they ignore the role of airports in facilitating (or dampening) airline-HSR cooperation. Recent contributions consider a multiple-airport system (MAS) where independent airports are linked by the HSR. Takebayashi (2015) proposes a numerically calibrated model to investigate the impact of the HSR linkage on competition between two airports in the case where airlines and the HSR also compete. He finds that the degree of airport-HSR connectivity affects the market shares of the two airports. Then, Takebayashi (2016) discusses the possibility of cooperation between airports and the HSR. He shows that cooperation between the HSR and the smaller-demand airport can reduce congestion at the larger-demand airport. Finally, Takebayashi (2018) studies the efficiency of an MAS with the HSR connecting the airports. The numerical analysis shows that reducing airport charges at uncongested airports is effective for improving social welfare. All of these papers do not consider transport operators’ incentives to cooperate. Xia, Jiang, Wang, and Zhang (2019) develop a revenue-sharing mechanism between the airline and the HSR in an MAS. The authors find that social welfare increases with the HSR linkage, and that airline-HSR cooperation serves effectively as a way to divert passengers from capacityconstrained to unconstrained airports. The authors do not model the vertical relation between airports and airlines, and thereby they ignore the role of airports in facilitating (or dampening) airline-HSR cooperation. We contribute to this body of literature as we explicitly take into account the role played by per-passenger fees charged by the airport infrastructure manager in shaping transport operators’ incentives to cooperate; moreover, we highlight cases where the airport company may adopt a “sabotage” strategy to delay and/or deter a cooperation agreement between the airline and the HSR operator, as well as cases where the airport company may be willing to share the sunk cost, so that transport operators can manage to cooperate and provide the combined service in the connecting market.

3. Incentives to intermodal cooperation There are a number of key drivers and barriers to intermodal agreements, some of which are specific to the major players involved, that is, airlines, HSR operators, and airport companies, while others are common to such players. These drivers and barriers have been widely discussed in the literature (Chiambaretto & Decker, 2012; Eurocontrol, 2005; Givoni & Banister, 2006; Vespermann & Wald, 2011). We summarize the main arguments

176

Market development and policy for one belt one road

in Table 1, while in Fig. 1 we report some examples of intermodal products with varying degrees of seamlessness and intensity of cooperation.c In what follows, we briefly discuss the most important points, and we highlight some issues that may deserve further attention. One of the main reasons for intermodal cooperation on the part of airlines is that they can provide wider access (i.e., from a larger number of cities in a country) to their services from the hub airport to international destinations. For instance, Qatar Airways or Etihad Airways have increased their market share in France with the “tgvair” combined product, by which they sell rail trips to 19 cities in France from Paris-CDG airport (see also Fig. 1). After an intermodal agreement, airlines can also divert part of the shorthaul traffic to HSR, thereby making the relevant airport slots available for routes that are more profitable. This is particularly important in hub-andspoke networks with congested hubs, where slots are scarce and expensive.d In such a case, the downside of intermodal agreements is that airlines may lose control of feeder routes, to the benefit of either HSR operators or competing airlines. This often leads airlines to maintain short-haul feeder flights on routes where intermodal agreements are in place. For instance, after Lufthansa and Deutsche Bahn have agreed to offer the “AIRail Service” combined product (see Fig. 1), multimodal passengers can take either flights or HSR trains from Frankfurt to Stuttgart, while flights from Frankfurt to Cologne are no longer available. As to HSR operators, they benefit from cooperation in that it increases their load factor and market share on short-haul routes, with no significant c

d

The reader may refer to the AIR&RAIL intermodal agreement between Thalys and Air France. The intermodal product is handled by Air France. It is included in the Air France booking system and made available to passengers that wish to travel from/to Brussels-Midi and Paris-Charles de Gaulle (CDG). The air carrier will forecast and confirm their traffic volumes to Thalys on an annual basis in order to determine the booking of one or two carriages per journey (first class, “comfort 1”). In addition, Air France has the possibility to book additional seats on an ad-hoc basis subject to availability on trains (e.g., in case of higher volumes of traffic than expected as it is the case with the booking of traveler groups). Thalys schedule of trains to/from CDG have been adapted to suit Air France departure/arrival timetables. Dedicated luggage hold is made available by Thalys to Air France passengers on Brussels-CDG trains. This is an exclusive facility for Air France passengers. Finally, at the Brussels check-in counter, Air France passengers get the boarding cards for both the Brussels-CDG link and for the CDG-Air France destination flight. The ticketing aspect is thus integrated (Eurocontrol, 2005). EU policymakers promote a revision of airport slot regulation to efficiently manage capacity at congested hubs (see Avenali et al., 2015, for a theoretical study).

High-speed rail and air transport integration

177

Table 1 Drivers and barriers to intermodal agreements. Player

Driver

Barrier

Airports

•

•

• • • Airlines

• • •

Increase of airside capacity at the airport Expansion of catchment area Meeting customer needs for both transport and commercial activities Addressing environmental and landside congestion issues/targets Substitution of feeder flights by trains to free slots for more profitable routes Wider access to air services from the hub Reduction of operating costs

• • • • • •

Rail operators

• • •

Improving load factor Improving rail market share Improving the image of highspeed rail

• • •

Partial reallocation of freed slots Catchment area overlap with other airports Car parking revenue loss Loss of control of the feeder routes Benefits to competing airlines Sunk cost of the intermodal agreement Difficulties in selling the intermodal product Capacity and financial issues Sunk cost of the intermodal agreement Benefits to competitors (air transport)

Source: Eurocontrol, 2005. Potential airport intermodality development, CARE II: MODAIR, version 1.1; Givoni, M., & Banister, D. (2006). Airline and railway integration. Transport Policy, 13, 386–397; Vespermann, J., & Wald, A., 2011. Intermodal integration in air transportation: Status quo, motives and future developments. Journal of Transport Geography, 19(6), 1187–1197; Chiambaretto, P., & Decker, C. (2012). Air–rail intermodal agreements: Balancing the competition and environmental effects. Journal of Air Transport Management, 23, 36–40.

increase in marginal costs once the rail infrastructure is deployed. On the other hand, rail infrastructure is a possible barrier to intermodal agreements as long as platform capacity and HSR slots are not sufficient to feed airlines’ hubs. Airport companies are interested in airline-HSR agreements since they can be a means to redistribute the use of slots from short-haul flights to more profitable long-haul flights (that are subject to higher landing fees), to expand their catchment areas and effectively compete with adjacent airports, to promote some commercial activities in airport terminals, and to address environmental targets. The other side of intermodal cooperation for airports

Common online ticket distribution

Integrated Ticketing

Integrated Loyalty programs

End to end check-in

AIRail

available

available

available

available

available

available

available

Brussels-Midi Strasbourg

Air&Rail

available

available

available

available

available

available

available

Paris-CDG Airport

Angers - St-Laud Avignon TGV Champagne-Ardenne Le Mans Lille Europe Lorraine TGV Lyon Part-Dieu Nantes Nimes Poitiers Rennes St-Pierre-des-Corps Toulon Valence TGV

TGVAir

available

available

available

available

available

available

Swiss

Zurich Airport

Basel

Airtrain

available

available

available

available

available

ÖBB

Austrian Airlines

Vienna Int. Airport

Linz

Austrian AIRail

available

available

available

Deutsche Bahn

Lufthansa United Airlines TAP Portugal Cathay Pacific Emirates + others

Frankfurt Airport

Basel Berlin Dresden Holzwickede + others

Rail&Fly

available

HSR

Airline

From Airport

To City

Deutsche Bahn

Lufthansa American Airlines Emirates

Frankfurt Airport

Stuttgart Cologne Siegburg/Bonn/Kassel WilhelmshÖhe/Karlsruhe + others

Thalys/SNCF

Air France American Airlines Brussels Airlines Jet Airways KLM

Paris-CDG Airport

SNCF

Air France

SCC-CFF-FFS

Brand

Schedule Delay/connection coordination assistance

Baggage Handling

available

Fig. 1 Examples of air-rail agreements. (Source: Personal elaboration on data provided by Eurocontrol, 2005 and operators’ websites.)

High-speed rail and air transport integration

179

is revenue losses from a partial (instead of a complete) reuse of freed slots, particularly at uncongested airports, and from stronger airport competition due to overlapping catchment areas. A major barrier to intermodal agreements that is of interest to all players is the huge investments required to make cooperation effective, which result in considerable sunk costs. As a baseline, transport operators should enable passengers to purchase a single ticket for the entire multimodal trip (i.e., to buy a bundle of domestic HSR and international air services). This requires operators to integrate their information technology and computer reservation systems. Moreover, it requires operators to coordinate schedules between air and HSR services, thereby taking the risk of possible delays on one segment of the journey, and providing passengers with proper warranties. Operators may also offer coordinated baggage handling (so that passengers should not care about baggage transfer at the intermediate stop), and/or supplementary services on HSR trains similar to those offered on short-haul flights (e.g., dining) (see also Fig. 1). The most important obstacle to cooperation is perhaps the absence of the HSR station at the hub airport to enable passenger intermodality. Indeed, the cost of deploying the HSR link to the hub airport may be significantly (and sometimes prohibitively) high. Since the benefits of cooperation accrue to all the involved players, then the infrastructure cost should ideally be split among beneficiaries. This means that the private (usually the airlines, and possibly airport companies) and the public (usually rail operators, and possibly airport companies) sectors are called to jointly take on the responsibility to make cooperation effective. As long as intermodal agreements are beneficial to society as a whole, policy makers are also called to play an active role and pave the way for such agreements to become feasible, by providing the involved players with suitable (e.g., fiscal) incentives and/or direct funding (subject to budget constraints) to support infrastructure investments. An additional factor (somewhat overlooked in the abovementioned literature) that may help explain the attitude of the involved players toward intermodal agreements is the degree of substitutability between air and HSR services. Indeed, mode substitution affects the impact of the agreement on traffic volumes and on the mode share. On the one hand, substitutability is essential to determine whether long-distance passengers may consider the airline-HSR combined product as an effective alternative to the connecting flight. On the other hand, substitutability measures the strength of competition for passengers in short-haul markets served by both the airline and the HSR operator.

180

Market development and policy for one belt one road

It is worth noting that assessing these effects is relevant not only to airlines and to HSR operators, but also to airport companies, whose revenue depends primarily on air traffic. Thus, in order to calibrate landing fees, airport managers should anticipate how mode substitution affects the impact of intermodal agreements on traffic volumes and mode shares. In this framework, it may happen that transport operators and airport companies have conflicting interests on intermodal agreements. For instance, transport operators might be more inclined to sign an agreement in response to a high degree of substitutability because, among other factors, they can expect to achieve high traffic volumes for the airline-HSR combined product. Instead, if mode substitution is high then airport managers might have a less favorable attitude toward the agreement. This is because under cooperation, when mode substitution is stronger, transport operators have the incentive to induce passengers to travel by train instead of flying, thereby reducing the costs related to paying the landing fee. Therefore, the airport company has to significantly reduce such a fee to preserve air traffic volumes, which are the only source of revenues for the company. In the next section, we discuss these issues more formally by means of a theoretical model.

4. The model We consider three players, an airline a, a HSR transport operator r, and an airport company h, which run a transportation network of three nodes (i.e., cities) illustrated in Fig. 2. The airline operates the short-haul (e.g., domestic) route between city O and city H. In the same market (also called overlapping market), the HSR transport operator offers a direct ride. The airline also serves two long-haul (e.g., international) routes, that is, market HD with a direct connection and market OD (also called connecting market) with a one-stop trip via city H.e

O

H

D

Fig. 2 Structure of the network.

e

We borrow from Jiang and Zhang (2014) the topology and market structure of the network.

High-speed rail and air transport integration

181

City H can serve as a multimodal hub, since there is a HSR station at the airport. Travelers from city O to city D could, in principle, transfer from O to H by HSR and then fly from H to D. Given that market OD is covered by a single-carrier service, we assume that multimodal trips do not occur unless the airline and the HSR cooperate to offer a bundle of domestic HSR and international air services.f Let M ¼ {OH, HD, OD} be the set of markets m, and T ¼ {A, R, AA, AR} the set of transportation products t, where t ¼ A (t ¼ R) stands for a direct flight (HSR ride), t ¼ AA stands for a connecting flight via hub H, and t ¼ AR stands for the multimodal trip via hub H. Let Tm be the subset of transportation products available in market m. Given the network in Fig. 2, we have that TOH ¼ {A, R}, THD ¼ {A}, and TOD ¼ {AA} when the airline and the HSR do not cooperate, or TOD ¼ {AA, AR} when they cooperate. Thus, there are at most five travel choices for passengers. We consider the following inverse demand curves in relevant marketsg: pAOH ¼ α  qAOH  γqROH pROH ¼ α  qROH  γqAOH pAHD ¼ α  qAHD pAA OD pAR OD

¼α ¼α

qAA OD qAR OD

(1)  

γqAR OD γqAA OD

where ptm and qtm respectively are prices and traffic volumes for transportation products t  Tm in market m  M (note that product AR is supplied only in the case of airline-HSR cooperation, and qAR OD ¼ 0 in the absence of cooperation). Parameter α (α > 0) measures the maximum willingness to pay (hereafter, wtp) for product t  Tm, while γ (0 < γ < 1) measures the degree of substitutability between transportation products, where the lower γ, the lower the substitutability. For simplicity, we assume that α and γ are the same in all markets. We now turn to the supply side. As to transport operators, for simplicity we assume that: (i) the size of vehicles is the same for each transportation mode in each market and (ii) the relation among passengers, seats, and f

This assumption is made for simplicity, but is not essential for the results (see the discussion in Avenali et al., 2018). g These demand curves can be obtained by assuming that the representative passenger in each market has a strictly concave quadratic utility function a` la Singh and Vives (1984).

182

Market development and policy for one belt one road

flights/HSR rides is of the fixed proportions type (see e.g., Basso (2008)), such that the product between the size of vehicles and the load factor is constant for all services in all markets (we normalize such product to unity). Thus, prices per flight/HSR ride, per seat, and per traveler are equivalent. We assume that the operating cost per flight/HSR ride, per seat, or per traveler is constant and, for simplicity, normalized to zero. However, the airline has to pay the airport company a price per flight (or landing fee) w. If transport operators decide to cooperate, then they should jointly incur sunk costs F (F > 0) for cooperation to become effective (see the discussion in Section 3). As to the airport company, for simplicity we assume that the operating cost per flight is constant and normalized to zero. Furthermore, we focus on the case where the hub is not capacity constrained. The timing of the game is as follows. At stage one, transport operators decide whether to cooperate (thereby incurring sunk costs to make their services compatible) or not. At stage two, the airport company sets the landing fee. Finally, at stage three the airline and the HSR operator set quantities in relevant markets (either cooperatively or not).h We solve the game backwards. In doing so, we restrict attention to parameter values for which equilibrium quantities are strictly positive.i

5. Benchmark case: No agreement Consider first the benchmark case where transport operators do not cooperate. Thus, the airline is a monopoly in markets OD and HD, while the HSR operator and the airline compete a` la Cournot in market OH. At stage three, the HSR operator and the airline, respectively, solve: max qROH π r 5pROH qROH

h

i

(2)

This assumption reflects that both airport and the HSR platform capacities cannot be easily increased. Indeed, capacity adjustments are slower and more costly to implement than price adjustments, since the former require lumpy and irreversible investment. In this sense, we are taking a short-run rather than a long-run perspective. We thus follow a number of relevant papers, such as Jiang and Zhang (2014), D’Alfonso et al. (2015, 2016), and Xia and Zhang (2016). pffiffiffiffiffi  This means that γ < 12 41  5 holds. Under this assumption, equilibrium prices and profits are also strictly positive.

High-speed rail and air transport integration

max qAOH ,qAHD ,qAA π a ¼ pAOH qAOH + pAHD qAHD + pAA qAA OD OD  A  OD A AA  w qOH + qHD + 2qOD

183

(3)

where ptm (t  Tm, m  M) are as in Eq. (1). By the first-order conditions for R,N both profits, we find the vector of quantities qN(w) ¼ (qA,N OH (w), qOH (w), AA,N qA,N HD (w), qOD (w)) that solves problem (2)–(3), where N stands for no agreement. At stage two, the airport company chooses w in order to maximize profit:   A,N AA,N max w π h 5w qA,N (4) OH ðw Þ + qHD ðw Þ + 2qOD ðw Þ By the first-order condition for Eq. (4), we obtain: wN ¼

αð3γ 2 + 2γ  16Þ 10γ 2  48

(5)

By plugging wN in qN(w), we find the equilibrium vector qN, that is: 1 0 B αð5γ 2 + 3γ  16Þ B B ðγ + 2Þð5γ 2  24Þ , B B B αð7γ 2  8γ  48Þ B   , R,N A,N AA,N 2 ¼B qN 5 qA,N OH , qOH , qHD , qOD B 2ðγ + 2Þð5γ  24Þ B αð7γ 2  2γ  32Þ B , B B 20γ 2  96 B @ αðγ 2  γ  4Þ 5γ 2  24

C C C C C C C C C C C C C C A

(6)

Then, by plugging qN in Eq. (1), we find the equilibrium vector of R,N A,N AA,N prices pN ¼ (pA,N OH , pOH , pHD , pOD ) for transportation products. Finally, by plugging qN, pN, and wN in their respective profit functions, we find N the equilibrium profits for transport operators, π N a and π r , and for the airN port company, π h (see Appendix).

6. Air-rail cooperation In this section, we consider the case of full-scale cooperation between the airline and the HSR operator to provide the combined service in the connecting market. In such a case, the two firms set quantities in relevant markets to maximize their joint profit.

184

Market development and policy for one belt one road

6.1 Traffic volumes Assume that transport operators have decided to cooperate at stage one. Therefore, at stage three, they act as a merged firm ar that solves: A A R R A A AA AA AR π ar ¼ p max qAOH ,qROH ,qAHD ,qAA OH qOH + pOHqOH + pHD qHD + pOD qOD  OD ,qOD AR A A AA AR + pAR OD qOD  w qOH + qHD + 2qOD + qOD

(7) By the first-order conditions for Eq. (8), we find the vector qI(w) , that is:  ,I  R, I ,I ,I AR, I ðwÞ, qOH ðw Þ, qAHD ðw Þ, qAA qI ðwÞ5 qAOH OD ðw Þ, qOD ðw Þ 1 0 α  αγ  w B , C C 2Þ B 2 1  γ ð C B C B α  αγ + wγ C B , 2 C B 2ð1  γ Þ C B αω C ¼B , C B 2 C B B α  αγ + wγ  2w C B ,C C B 2Þ 2 1  γ ð C B @ α  αγ + 2wγ  w A 2ð1  γ 2 Þ

(8)

where I stands for integration. Based on Eq. (8), we can study the impact of w on traffic volumes. Remark 1 summarizes the comparative statics of qI(w) with respect to the landing fee.j Remark 1. Assume that the airline and the HSR operator cooperate. In equilibrium, as the landing fee increases air passengers decrease in all markets (i.e.,

∂qA,I OH ðw Þ ∂w

< 0,

∂qA,I HD ðw Þ ∂w

< 0, and

∂qAA,I OD ðw Þ ∂w

< 0), and we also have that:

(i) in market OH, HSR passengers increase (i.e., R,I ∂ðqA,I OH ðw Þ + qOH ðw ÞÞ fic decreases (i.e., < 0); ∂w

j

∂qR,I OH ðw Þ ∂w

> 0) and total traf-

We can prove that the same qualitative results hold in the benchmark scenario, namely, we ∂ðqA,N ðwÞþqR,N ðw ÞÞ ∂qA,N ðw Þ ∂qAA,N ðwÞ ∂qR,N ðw Þ ∂qA,N ðw Þ find that OH∂w < 0, HD∂w < 0, OD∂w < 0, OH∂w > 0, and OH ∂w OH < 0. For brevity, all proofs are omitted and are available upon request from the authors.

High-speed rail and air transport integration

185

(ii) in market OD, passengers on the airline-HSR service decrease if and ∂qAR,I ðw Þ

only if transportation products are weak substitutes (i.e., OD∂w < 0 ∂ðqAA,I ðw Þ + qAR,I ðw ÞÞ iff γ < 0.5), and total traffic decreases (i.e., OD ∂w OD < 0). Intuitively, since the landing fee is a cost to the airline, then as w increases the merged firm finds it profitable to reduce the number of air passengers in all markets, while increasing the number of HSR passengers in market OH (where total traffic decreases). An increase in w also reduces total traffic volumes in market OD. Since the cost of a connecting flight (2w) is twice the cost of the air-rail combined service (w), then the merged firm tends to substitute air-rail services for connecting flights. This particularly occurs if transportation products are strong substitutes (γ > 0.5), in which case passengers on the airline-HSR service increase in response to an increase in w.

6.2 Landing fee At stage two, the airport company chooses w in order to maximize profit:   A,I AA,I AR,I max w π h 5w qA,I (9) OH ðw Þ + qHD ðw Þ + 2qOD ðw Þ + qOD ðw Þ By the first-order condition for Eq. (9), we obtain: wI ¼

αðγ 2 + 4γ  5Þ 2ðγ 2 + 4γ  7Þ

(10)

Remark 2 summarizes the impact of mode substitution on wI.k Remark 2. Assume that the airline and the HSR operator cooperate. In equilibrium, as the degree of substitutability between products increases I the landing fee decreases (i.e., ∂w ∂γ < 0). Clearly, when mode substitution is stronger passengers are ready to travel by train instead of flying. Therefore, the airport company has the incentive to reduce the landing fee in order to preserve air traffic volumes, which are the only source of revenues for the company.

k

Qualitatively, the same result holds in the benchmark scenario, namely, we can prove that ∂wN ∂γ < 0.

186

Market development and policy for one belt one road

By plugging wI in qI(w), we find the equilibrium vector qI, that is: 1 0 B αð9  9γ  2γ 2 Þ B B 4ð7  γ 2  4γ Þð1 + γ Þ , B B B αð14  3γ  γ 2 Þ B B 4ð7  γ 2  4γ Þð1 + γ Þ ,   B B R,I A,I AA,I AR,I ¼ B αð9  4γ  γ 2 Þ , q , q , q , q qI 5 qA,I OH OH HD OD OD , B B 4ð7  γ 2  4γ Þ B B αð4  5γ  γ 2 Þ B , B B 4ð7  γ 2  4γ Þð1 + γ Þ B @ αð9 + γ Þ 4ð7  γ 2  4γ Þð1 + γ Þ

C C C C C C C C C C C C C C C C C C C A

(11)

Based on Eq. (11), we can study the impact of mode substitution on traffic volumes. Remark 3 summarizes the results.l Remark 3. Assume that the airline and the HSR operator cooperate. In equilibrium, as the degree of substitutability between products increases, we have that: ∂qA,I (i) in market HD, air passengers increase (i.e., ∂γHD > 0); ∂qA,I

(ii) in market OH, air passengers decrease (i.e., ∂γOH < 0) whereas HSR passengers decrease if and only if transportation products are weak sub∂qR,I OH ∂γ

< 0 iff γ < γ , where γ ffi 0:461). Moreover, total traf∂ðqA,I + qR,I Þ fic in market OH decreases (i.e., OH∂γ OH < 0); (iii) in market OD, passengers on the connecting flight decrease  AA,I  ∂qOD < 0 whereas passengers on the airline-HSR service decrease ∂γ stitutes (i.e.,

if and only if transportation products are very weak substitutes ∂qAR,I

OD ( ∂γ < 0 iff γ < γ , where b γ ¼ 0:207). Moreover, total traffic in market OD increases if and only if products are strong substitutes (i.e., AR,I ∂ðqAA,I OD + qOD Þ > 0 iff γ > γ_ , where γ_ ¼ 0:548). ∂γ

l

Qualitatively, the same results hold in the benchmark scenario, namely, we can prove that R,N R,N ∂ðqA,N ∂qA,N ∂qOH ∂qA,N OH þqOH Þ OH HD < 0, with the only exception of ∂γ > 0, ∂γ < 0, ∂γ < 0 iff γ < 0.851, and ∂γ ∂qAA,N

passengers on the connecting flight, for which we find that OD ∂γ > 0 . Since γ < pffiffiffiffiffi  ∂qR,N 1 OH 41  5 (see Footnote (g)), ∂γ < 0 always holds in the feasible parameter region. 2

High-speed rail and air transport integration

187

The rationale for Remark 3 is as follows. From Remark 2, if mode substitution is stronger than the airport company has the incentive to reduce w. I Then, consistent with Remark 1, qA, HD increases since the supply of flights is less costly. In market OH (respectively, OD), when γ increases flights (respectively, connecting flights) and HSR rides (respectively, air-rail services) become stronger substitutes. Therefore, the merged firm has the incentive to substitute HSR rides (air-rail services) for flights (connecting flights), since the former are less expensive than the latter. If γ is sufficiently high, then the supply of HSR rides (air-rail services) increases with γ. Note that the same result holds for total traffic in market OD. By plugging qI in Eq. (1), we find the equilibrium vector of prices R,I A,I AA,I AR,I pI 5 (pA,I OH, pOH, pHD, pOD , pOD ) for transportation products. Finally, by I I plugging q , p , and wI in their respective profit functions, we find the equilibrium profits for the merged firm, π Iar, and for the airport company, π Ih (see Appendix). We find that the airport company’s profit decreases as long as the degree of substitutability between products increases. Generally, this is because (from Remark 2 and Footnote (i)), the landing fee decreases with the degree of substitutability.m Remark 4. Assume that the airline and the HSR operator cooperate. In equilibrium, as the degree of substitutability between products increases, the airport company’s profit decreases (i.e.,

∂π Ih ∂γ

< 0).

6.3 Cooperation or competition? At stage one, transport operators decide whether to cooperate (thereby incurring the sunk costs F) or not, by anticipating the landing fee at stage two, and passenger traffic volumes at stage three. Proposition 1 illustrates transport operators’ decision. Proposition 1. If the sunk costs are sufficiently low, that is, if F is such that N 0 < F  F1 holds, where F1 ¼ π Iar  (π N a + π r ), then the airline and the HSR operator find it profitable to cooperate. Proposition 1 simply follows from the fact that, under cooperation, the merged firm realizes a profit that is higher than the sum of the profits accrued to transport operators in the benchmark case of competition. Thus, as long as m

We can prove that the same qualitative result holds in the benchmark scenario, namely, we find that

∂π N h ∂γ

< 0.

188

Market development and policy for one belt one road

the sunk costs of cooperation are sufficiently low (more precisely, lower than the difference in industry profits between the cooperative and the competitive scenarios), transport operators can manage to incur such costs and provide the combined service in market OD. Note that the value of F1 can be taken as a measure of the scope for cooperation. Thus, the higher F1 the greater the number of cases where transport operators find it profitable to incur the sunk costs of cooperation. Less integrated forms of agreement are similar to traditional interlining agreements, in which an airline is authorized to sell rail tickets, without any further integration of the products. In this case, sunk costs due to cooperation are low. In contrast, more integrated intermodal agreements can involve integration of IT systems, guarantees that are offered to passengers in the case of delays on one segment of the journey. Deeper forms of integration can take the form of coordination of through-baggage handling and other dedicated services such as separate first and business class dining facilities on trains. In this case, sunk costs due to cooperation are high. Generally speaking, the higher the degree of seamlessness and intensity of cooperation the higher the magnitude of F (see also Fig. 1 for related examples of services offered by intermodal products). Remark 5 shows that cooperation is generally easier as long as mode substitution is stronger (except for some extreme cases where the degree of substitutability between products is very low). Indeed, if products are stronger substitutes, then industry profits increase more in the presence of a multiproduct monopolist (i.e., the integrated firm) than of two independent firms. N Remark 5. Let F1 ¼ π Iar  (π N a + π r ). Then, F1 increases with the degree of 1 substitutability between products, unless γ is very low (i.e., ∂F γ, ∂γ > 0 iff γ > e where e γ ffi 0:047).

Table 2 compares equilibrium quantities before and after cooperation. Thus, total traffic in market OH declines after cooperation due to reduced competition between transport operators. Instead, total traffic in Table 2 Comparison of traffic volumes before and after cooperation.

Market HD Market OH

Market OD

0  γ < 0.334 0.334  γ < 0.595 A,N A,I A,I qA,N qHD > qHD HD < qHD A,N A,I qOH > qOH R,I qR,N OH > qOH R,N A,I R,I qA,N OH + qOH > qOH + qOH AA,N AA,I AR,I qOD < qOD + qOD AA,I qAA,N OD > qOD

0.595  γ

R,I qR,N OH < qOH

High-speed rail and air transport integration

189

market OD increases after cooperation due to the supply of the new transportation product, that is, the air-rail combined service.

7. The role of airports Let us now focus on the role of the airport company. For this purpose, we compare the scenario where transport operators cooperate with the benchmark case of competition, first in terms of the level of the landing fee (Proposition 2), and then in terms of the airport company’s profit (Remark 6). We show that, in both circumstances, the results obtained depend on mode substitution. Proposition 2. Assume that the airline and the HSR operator cooperate. In equilibrium, the airport company sets a higher landing fee compared to the benchmark case of competition if and only if transportation products are weak substitutes. Formally, wI > wN iff γ < γ, where γ ffi 0:334. ∂w ∂w From Remark 2 and Footnote   (i), we have that ∂γ < 0 and ∂γ < 0. We  N  I can also prove that ∂w∂γ  < ∂w ∂γ . As transportation products become stronger I

N

substitutes, the airport company loses more revenues from setting a high landing fee under cooperation than under competition. Indeed, in the former case revenues from air traffic would be lower not only in market OH, where passengers move to HSR, but also in market OD, where passengers move from connecting flights to the airline-HSR combined service. Therefore, the airport company sets a higher charge under cooperation than under competition if and only if the degree of substitutability between products is sufficiently low. As to the airport company’s profit, we find that it is higher under cooperation than under competition if and only if the degree of substitutability between transportation products is sufficiently low (Remark 6). Indeed, from Proposition 2, in such a case the airport company can maintain a higher landing fee in the cooperative scenario. However, in some circumstances where the landing fee is lower under cooperation, the airport company can still benefit from the supply of the air-rail combined service on the part of transport operators. Remark 6. Assume that the airline and the HSR operator cooperate. In equilibrium, the profit of the airport company is higher than in the benchmark case of competition if and only if transportation products are weak substitutes. Formally, π Ih > π N γ , where €γ ffi 0:377. h iff γ < €

190

Market development and policy for one belt one road

Fig. 3 Players’ strategies and model outcomes.

Fig. 3 outlines the behavior of transport operators as well as of the airport company, and the consequent outcomes of the model. Depending on F and γ, we can draw a number of different scenarios. Let N us first consider the case where F F1 ¼ π Iar  (π N a + π r ). In this case, the airline and the HSR operator would not find it profitable to integrate. However, when γ < 0.377, the airport company would benefit from integration between transport operators. It follows that, as long as 0 < F  F1 < π Ih  π N h, the airport company is willing to share the sunk costs, and thereby in equilibrium transport operators can manage to integrate and provide the combined service in market OD. Conversely, if F  F1 > π Ih  π N h > 0, the outcome of the three stage game is the competing scenario. N Let us now consider the case where F < F1 ¼ π Iar  (π N a + π r ). In this case, the airline and the HSR transport operator integrate to provide the air-rail combined service. When γ < 0.377, we have that π Ih > π N h , that is, the airport company also benefits from integration between transport operators. Instead, when γ 0.377, firms’ incentives are not aligned, since the airport company would prefer transport operators to compete rather than cooperate. Therefore, the airport company might adopt a “sabotage” strategy to delay and/or deter a cooperation agreement between transport operators.

8. Discussion and concluding remarks Air transport and HSR have begun to act not only as simple competitors, but also as complementary modes. In recent years, several airline-HSR

High-speed rail and air transport integration

191

agreements have been signed worldwide. Policy makers, especially in Europe, encourage cooperation as a means to abate transaction costs and promote intermodality in passenger transport. In this chapter, we have assumed that the airline and the HSR operator cooperate to offer a bundle of domestic HSR and international air services via a multimodal hub airport. We have developed a theoretical model to study how transport operators’ incentives to cooperate and the airport company’s attitude toward cooperation depend on the sunk costs necessary to make cooperation effective and on mode substitution between air and HSR services. As long as the sunk costs of cooperation are sufficiently low (more precisely, lower than the difference in industry profits between the cooperative and the competitive scenarios), transport operators can manage to incur such costs and provide the combined service in the connecting market. This occurs more easily as transportation products become stronger substitutes because, in such a case, industry profits increase more under cooperation than in a competitive duopoly. However, if mode substitution is stronger than the airport company loses a large amount of revenues from setting a high landing fee under cooperation than under competition. Indeed, in the former case air traffic revenues would be lower not only in the short haul market where passengers would shift from flights to HSR rides, but also in the connecting market where passengers would prefer the airline-HSR combined service to the connecting flight. Therefore, the airport company sets a higher landing fee under cooperation than under competition if and only if the degree of substitutability between air and HSR is sufficiently low. Generally, this reflects in a higher profit for the airport company under cooperation than under competition when mode substitution is weak. It follows from the foregoing statements that firms’ incentives may not be aligned. First, this happens when the sunk cost of cooperation is low (such that transport operators find it profitable to cooperate) and mode substitution is strong (such that the airport company would prefer transport operators to compete rather than cooperate). In such a case, the airport company may adopt a “sabotage” strategy to delay and/or deter a cooperation agreement between the airline and the HSR operator. For instance, the airport company may influence strategically the ease of transfer between the rail station at the airport and air terminals. Indeed, the literature discusses different factors affecting users’ perception of air-rail product quality relating to physical and logical ease of transfer at the airport (European Commission

192

Market development and policy for one belt one road

(EC), 2006; IATA, 2003; ITS, 2009; Janic, 2011; Riley & Kumposˇtova´, 2010). Physical barriers include the walking distance of the connection between the airport rail station and air terminals, the number of level gaps in the walking transfer (stairs, ramps, escalators, lifts), the design adaptation for disabled passengers making the transfer, as well as other comfort issues in the transfer path (weather protection, lighting, cleanliness, corridor design, supply of shops, and facilities). On the other hand, logical barriers include the availability of real-time information at the airport, personalized information services (e.g., by mobile phone) or information desks about the connection between the airport rail station and air terminals, as well as the general perception of security along the transfer path at the airport. A misalignment of incentives can also take place when the sunk cost of cooperation is high (such that transport operators would not find it profitable to cooperate) and the degree of substitutability between air and HSR services is low (such that the airport company would benefit from cooperation). In such a case, the airport company may be willing to share the sunk cost, so that transport operators can manage to cooperate and provide the combined service in the connecting market. For instance, airport managers may decide to participate in the infrastructure investments needed to make cooperation effective. For instance, Lufthansa, Deutsche Bahn, and Fraport (the company managing Frankfurt airport) have agreed to share the costs related to deploying the luggage system for the intermodal product. In this framework, the impact of airline-HSR agreements on passengers’ well-being and on social welfare is not clear-cut. Actually, intermodal cooperation increases product variety, but raises competition concerns as long as it involves some coordinated pricing in the hub-and-spoke network. Even with limited or no price coordination between transport operators, one should consider the role of congestion at hub airports. If a hub airport is capacity constrained then traffic volumes, and thereby the consumer surplus, in each relevant market are affected by the airline’s decision on how to allocate scarce slots among markets. In turn, this decision primarily depends on the landing fee set by the airport company. Despite the relevance of airline-HSR agreements, competition authorities have so far devoted little attention to considering the interplay among airlines, HSR operators and airport companies. Further developments of this work should be developed in this direction. First, in this chapter we have focused on incentives to cooperation, while a proper welfare analysis should be carried to evaluating the effects of these agreements both under capacity-constrained and unconstrained airports.

High-speed rail and air transport integration

193

Second, policy intervention in facilitating airline-HSR cooperation should certainly be contextualized within the debate of regional inequality in economic development. HSR may facilitate the exploitation of agglomeration economies and, consequently, attract more business activities at a few major cities where primary hub airports are usually located. In turn, the increased economic activities further stimulate air travel demand at such airports. Relatedly, policies favoring the regional airports that are negatively affected by HSR may be useful to achieve a better traffic distribution among airports. In this context, given that a good air-HSR intermodal linkage (e.g., airport on-site HSR station, airline-HSR cooperation) tends to make airports suffer a less traffic drop or enjoy a more traffic gain, policy makers may consider this as a tool to achieve a better airport traffic distribution (Zhang et al., 2019). For example, air-HSR intermodal service can be encouraged at small airports located in cities with high enough income and growth potential as a tool to provide a potential for connection and traffic movement (Zhang et al., 2019).

Appendix Equilibrium profits Benchmark case: No agreement

Airline

α ð9216 + 4608γ2336γ 1728γ 136γ πN a ¼ 16ðγ + 2Þ2 ð5γ 2 24Þ2

High-speed rail

α ð48 + 8γ7γ Þ πN r ¼ 4ðγ + 2Þ2 ð5γ 2 24Þ2

Airport

α ð48 + 8γ7γ Þ πN h ¼ 8ðγ + 2Þð5γ 2 24Þ

2

2

3

4

+ 200γ 5 + 65γ 6 Þ

2 2

2

2 2

2

Air-rail cooperation

Merger

+ 7γ65Þ π Iar ¼ α16ððγγ ++19γ Þðγ 2 + 4γ7Þ

Airport

Þðγ + 5Þ π Ih ¼ 8ðαγ +ðγ1 1Þðγ 2 + 4γ7Þ

2

3

2

2

2

References Avenali, A., Bracaglia, V., D’Alfonso, T., & Reverberi, P. (2018). Strategic formation and welfare effects of airline-high speed rail agreements. Transportation Research Part B: Methodological, 117, 393–411. 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.

194

Market development and policy for one belt one road

Basso, L. J. (2008). Airport deregulation: Effects on pricing and capacity. International Journal of Industrial Organization, 26(4), 1015–1031. Chiambaretto, P., & Decker, C. (2012). Air–rail intermodal agreements: Balancing the competition and environmental effects. Journal of Air Transport Management, 23, 36–40. D’Alfonso, T., Jiang, C., & Bracaglia, V. (2015). Would competition between air transport and high-speed rail benefit environment and social welfare? Transportation Research Part B: Methodological, 74, 118–137. D’Alfonso, T., Jiang, C., & Bracaglia, V. (2016). Air transport and high-speed rail competition: Environmental implications and mitigation strategies. Transportation Research Part A: Policy and Practice, 92, 261–276. Eurocontrol. (2005). Potential airport intermodality development, CARE II: MODAIR, version 1.1. European Commission (EC). (2006). Air and rail competition and complementarity. Retrieved from http://ec.europa.eu/transport/rail/studies/doc/2006_08_study_air_rail_competition_en. pdf. European Commission (EC). (2011). Roadmap to a single European transport area – Towards a competitive and resource efficient transport system. White paper. COM 144, Brussels. Financial Times. (2008). Air France talks to Veolia about rail venture. Available from http:// www.ft.com/cms/s/0/ce8e15aa-4933-11dd-9a5f-000077b07658. html#axzz4KMCZB1qX (Last retrieved: 16/09/2016). Fu, X., Oum, T. H., & Yan, J. (2014). An analysis of travel demand in Japan’s intercity market: Empirical estimation and policy simulation. Journal of Transport Economics and Policy, 48(1), 97–113. Givoni, M., & Banister, D. (2006). Airline and railway integration. Transport Policy, 13, 386–397. IATA. (2003). Air/rail intermodality study. Prepared by IATA Air Transport Consultancy Services for the European Commission, Airports Council International, Air Transport Action Group, the Community of European Railway and Infrastructure Companies (CER) and the International Union of Railways. ICAO. (2014). Annual report of the Council. Retrieved from https://www.icao.int/annualreport-2014/Pages/default.aspx. ITS. (2009). Kite: A knowledge base for intermodal passenger travel in Europe. Deliverable D14: Guideline for seamless intermodal interchanges. Janic, M. (2011). Assessing some social and environmental effects of transforming an airport into a real multimodal transport node. Transportation Research Part D: Transport and Environment, 16(2), 137–149. Jiang, C., D’Alfonso, T., & Wan, Y. (2017). Air-rail cooperation: Partnership level, market structure and welfare implications. Transportation Research Part B: Methodological, 104, 461–482. Jiang, C., & Zhang, A. (2014). Effects of high-speed rail and airline cooperation under hub airport capacity constraint. Transportation Research Part B: Methodological, 60, 33–49. Jiang, C., & Zhang, A. (2016). Airline network choice and market coverage under highspeed rail competition. Transportation Research Part A: Policy and Practice, 92, 248–260. Riley, P., & Kumposˇtova´, A. (2010). Recommendations of the LINK project: European forum on intermodal passenger travel. In 2010 proceedings of the European Transport Conference. Singh, N., & Vives, X. (1984). Price and quantity competition in a differentiated duopoly. The RAND Journal of Economics, 546–554. Takebayashi, M. (2015). Multiple hub network and high-speed railway: Connectivity, gateway, and airport leakage. Transportation Research Part A: Policy and Practice, 79, 55–64. Takebayashi, M. (2016). How could the collaboration between airport and high speed rail affect the market? Transportation Research Part A: Policy and Practice, 92, 277–286.

High-speed rail and air transport integration

195

Takebayashi, M. (2018). Managing airport charges under the multiple hub network with high speed rail: Considering capacity and gateway function. Transportation Research Part A: Policy and Practice, 112, 108–123. Vespermann, J., & Wald, A. (2011). Intermodal integration in air transportation: Status quo, motives and future developments. Journal of Transport Geography, 19(6), 1187–1197. Wan, Y., Ha, H. K., Yoshida, Y., & Zhang, A. (2016). Airlines’ reaction to high-speed rail entries: Empirical study of the Northeast Asian market. Transportation Research Part A: Policy and Practice, 94, 532–557. Xia, W., Jiang, C., Wang, K., & Zhang, A. (2019). Air-rail cooperation and multiple-airports system: A revenue-sharing mechanism between air and rail sectors. Transportation Research Part B: Methodological, 121, 304–319. Xia, W., & Zhang, A. (2016). High-speed rail and air transport competition and cooperation: A vertical differentiation approach. Transportation Research Part B: Methodological, 94, 456–481. Yang, H., & Zhang, A. (2012). Effects of high-speed rail and air transport competition on prices, profits and welfare. Transportation Research Part B: Methodological, 46(10), 1322–1333. Zhang, A., Wan, Y., & Yang, H. (2019). Impacts of high-speed rail on airlines, airports and regional economies: A survey of recent research. Transport Policy, 81, A1–A19.

This page intentionally left blank

CHAPTER 10

On impacts of Pakistan Railways Main Line 1 on “North China-EU” export transit—Taking export transit between Beijing and UK as an example Ying-En Gea, Mingfeng Mob, Fangwei Zhangb, Muhammad Arsalan Khalidc, Mengmei Yub, Guanke Liuc, and Wenqian Lud a College of Transportation Engineering, Chang’an University, Xi’an, China College of Transport & Communications, Shanghai Maritime University, Shanghai, China c School of Business, National University of Singapore, Singapore d Tilburg School of Economics and Management, Tilburg University, Tilburg, Netherlands b

1. Introduction The European Union (EU) is one of the largest trading partners and largest exports market of China. The trade volume and trade scope between the EU and China have both expanded significantly over the years. After the financial crisis of 2008, the global economy ushered in a “new mediocrity and greater differentiation” economic development situation. As China started to act on the “One Belt, One Road” initiative in 2013, China and the EU have deepened cooperation in many aspects, in particular working together for a long-term Reconstruction and Development efforts with introduction of China-EU 2020 Strategic Plan and constantly explore the ways to ensure the smooth and rapid development of China-EU trade cooperation under the new economic situation. The “One Belt, One Road” initiative started at the end of 2013, and has attracted a global attention, especially in the countries along the Belt or Road. In March 2015, the National Development and Reform Commission issued the “Vision and Action for Promoting the Construction of the Silk Road Economic Belt and the 21st Century Maritime Silk Road” (The Ministry of Commerce Comprehensive Department, 2018). This document proposes “the construction of China-Pakistan, BCIM Economic Corridor, involving China, India, Bangladesh and Myanmar, and aims to promote Market Development and Policy for One Belt One Road https://doi.org/10.1016/B978-0-12-815971-2.00005-0

Copyright © 2022 Elsevier Inc. All rights reserved.

197

198

Market development and policy for one belt one road

regional cooperation to achieve greater progress.” On April 20, 2015, during the visit of Chinese President Xi Jinping to Pakistan, the National Railway Administration of the People’s Republic of China and the Ministry of Railways of the Islamic Republic of Pakistan signed the official documents of “Reconstruction of the No. 1 main line railway (ML1)” and “Construction of Havelian Dry port and its Feasibility Study Framework” (National Railway Administration, 2018). As stated in the agreement, China will help Pakistan upgrade ML-1 Railway line and extend it northward, channeling through the China-Pakistan border port, Hongqi Lapu and finally reaching Kashgar, China. At present, China’s trade with Europe is mainly realized on the new Eurasian Continental Bridge and maritime shipping. The successful reconstruction of Pakistan’s ML-1 line will add a searail combined transport channel for China’s export to Europe. Wang and Meng (2011) have studied on the impact of land-bridge transportation on the market share of the ports, the results indicate that the economic value of time has a significant influence on the volume of shipping. There is a significant difference in freight rate and the travel time between shipping route and CHINA RAILWAY Express (CR Express) routes. Thus, the export volume arrangements between shipping and CR Express has been polarized. However, the channel through Pakistan’s ML-1 line will offer shippers a mid-range transportation alternative, on its strength of shorter transport time and lower freight rate. Moreover, the shipping freight rate fluctuates greatly. In North China, there are densely populated ports, which are located in the proximity of central and western channels of CR express. The travel time and freight cost between the transport alternatives are different with each other. Therefore, ML-1’s freight rate and the fluctuations of the shipping freight rate will have a significant impact on the allocation of different export routes in North China. To ensure the accuracy of the data, this chapter collects and compares the export trade volumes of various provinces and province-level municipalities to 27 European countries, including Austria, France, Germany, Spain, and the Great Britain. Figs. 1 and 2 demonstrate the export volume of Beijing and Inner Mongolia to European countries over 12 years (from 2005 to 2016), respectively. The trend is generally stable except for the year of 2009, the export volume fell due to the effect of financial crisis. The red curve is the amount of exports from different provinces in North China to the United Kingdom. In general, the trend of exports to Europe has shown a relatively stable growth during the post-crisis period. The export situation of the other Chinese regions such as Shanxi, Hebei, and Tianjin

Impacts of Pakistan Railways Main Line 1 (ML-1)

199

300000 200000 100000 0

2005

2006

2007

Austria Czech Republic Germany Latvia Poland Spain

2008

2009

Belgium Denmark Greece Lithuania Portugal Sweden

2010

2011

2012

Bulgaria Estonia Luxembourg Luxembourg Romania UK

2013

2014

2015

Cyprus Finland Ireland Malta Slovakia

2016

Croatia France Italy Holland Slovenia

Fig. 1 Changes in exports of Beijing city to major European countries in 2005–16.

20000

10000

0 2005

2006

Austria Czech Republic Germany Latvia Poland Spain

2007

2008 Belgium Denmark Greece Lithuania Portugal Sweden

2009

2010

2011

Bulgaria Estonia Luxembourg Luxembourg Romania UK

2012

2013 Cyprus Finland Ireland Malta Slovakia

2014

2015

2016 Croatia France Italy Holland Slovenia

Fig. 2 Changes in exports of Inner Mongolia to major European countries in 2005–16.

are almost the same, so we won’t go into details here. The study will assume the future export trade volume of North China to Europe based on the changes in export volume over the past 12 years. Based on the trade data, this study aims to quantify the impact of Pakistan’s ML-1 line on the choice of transportation routes and shipping freight in North China. The first section is a brief review of the discrete choice model. Then we establish the probability distribution model of the North China-Europe export channels with the description of relevant parameters and indicators, including price elasticity. The third part is the empirical analysis of the above models. Section 4.1 is a preliminary analysis of the basic data; Sections 4.2–4.4 quantify the path assignment results under the influence of different control variables, incorporating the impact of time preference coefficients λ and the freight rate of ML-1 channel on route allocation, the impact of ML-1 channel on the shipping freight volume as well as

200

Market development and policy for one belt one road

the substitution and complement relationship between these two Channels. In Section 4.5, we illustrate the impact of ML-1 freight rate and the fluctuations of shipping freight rate on the income model for Pakistan and Singapore, specially. The last part of this chapter includes a set of concluding remarks and offers prospective research directions.

2. Research tools The discrete choice model was originated in 1970s and is used mostly to analyze and predict the choice of transportation modes. Unlike the aggregate model, which is based on the traffic zone, the discrete choice model is based on the theory of random utility maximization while observing single decision maker (Shi & Yin, 2018). In this model, the utility of one choice limb is divided into a determined part, defined by observable variables, and a random utility part, determined by unobservable variables (Williams, 1977). The utility function of the discrete choice model can be expressed as Uin ¼ Vin + εin where the subscripts i, n represent the mode of travel of traveler, respectively; Uin indicates the utility of n while he chooses the mode i; Vin is the measurable utility while the εin is the random part (Train, 2003). Multinomial logit model can be expressed as pi ¼ X

exp ðU i Þ exp ðU i Þ iM

,

(1)

where pi is the probability of the traveler n choosing mode i, Ui is the utility of travel mode i and M is the set of all the travel modes. Luce (1959), while establishing the multi-logit formula, specified the feature of “independence of irrelevant alternatives," which means that the multi-logit model gets the feature of IIA. In fact, as the travel distance and time increase, the deviation of the model results may become even larger (Tversky, 2004). Moreover, the random entries εi of the utility function in the model are assumed to obey the same Gumbel distribution, which also limits the applicability of the logit model. Therefore, different scholars have extended the multi-logit model in different directions and classified the extended model which relaxed the restriction on random entries into non-IIA models, including the Nested Logit (NL), Generalized Extreme Value model (GEV), Heteroscedastic Extreme Value Model (HEV), and Mixed Multinomial Logit Model (MMNL). The NL model assumes that the random entries of different selective limbs are distributed identically

Impacts of Pakistan Railways Main Line 1 (ML-1)

201

but not independently. The model selects the alternatives in the set as a subset and uses conditional probability to calculate the probability when an actual selected limb is chosen. Thus, NL model overcomes the disadvantages of the IIA feature of the MNL by applying such assumption. Additionally, for some specific research problems, the random entries among the selected limbs do not even satisfy the distribution of the same variance. To reflect this feature, Bhat (1995) developed a discrete choice model for heteroscedasticity. The MMNL simultaneously relaxes the assumptions of both independent and identical distributions between random selected limbs. It was initially established by Mcfadden and Train (2000), assuming the coefficients in the determined entries in the MNL model were no longer certain constants, but randomly distributed variables.

3. Probability distribution model and parameter analysis of North China-Europe export channel 3.1 Distribution probability of each export channel At present, containerized goods shipped to Europe from North China are mainly transported by sea and rail. The two modes of transport have significant differences in travel time and transportation cost. Since 2013, the CR express has started to operate in varied regions of China. With its short transportation distance, fast speed and high safety, the CR express has become an alternative freight transport corridor of China’s export to Europe. However, as a whole, the CR express transport market is still immature. In comparison, shipping rates are low and freights can avoid repeated customs declarations. Therefore, shipping is still the most important mode of transportation between China and Europe. The construction of the China-Pakistan Economic Corridor, led by Pakistan’s ML-1 line, will directly provide China one new channel to the Indian Ocean, and significantly shorten the distance between China, Mediterranean countries, and western Europe. The existing export routes of CR express include the Eastern, Central, and Western Channels. The Eastern channel starts in Southeastern coastal areas of China and leaves the Manhole pass (Suifen River), the Central channel leaves through Erenhot channel and the Western channel leaves through the Alataw channel. As the shipping channel, we just take the channel through the Strait of Malacca and enters the Mediterranean Sea via Suez Canal before finally connecting to Europe into consideration (without considering the Cape of Good Hope route and the Arctic route). After the completion of the Pakistani ML-1 extension, it will participate in the

202

Market development and policy for one belt one road

China-Europe transportation in the form of sea-rail combined transport. To sum up, this study takes the operation of Pakistan’s ML-1 railway line as the background, considering five optional transportation Channels for export trade and transportation. The sea-rail combined transit route based on Pakistan’s ML-1 is referred to the Pakistan channel. The export freight volume of the five transit Channels is distributed as follows: !1 5 X ki U i ki U i Qi ¼ Q  e  e , (2) i¼1

i represents the exports route from North China to Europe; Qi is the freight volume corresponding to the route i; Q is the predicted value of the initial freight volume; Ui represents the measurable utility of the route i; ki is the parameter calibration; e ¼ 2.71828. It is assumed that the shipper’s time preference coefficient in the choice of transit route is λ: i.e., the importance attached to the time cost of the shipper is λ times the freight rate, and fi is the unit freight of the route. D(ti) ¼ d  ti represents the economic value of time, where ti is the transit time of the route i, and d is demurrage fee for a single day, thus, fi + λD(ti) represents the overall cost of the route i. Simplify the calculation by dividing the original utility value of each route by the utility value of the Manzhouli channel, m

λDðti ÞÞ according to utility function used U i ¼ ðfðf i++λD ðt 1 ÞÞm , the probability distri1

bution of the corresponding Channels of different OD points is as follows: pODi ¼ 0 B @

ðfi + λDðti ÞÞm ki  ðf + λDðt ÞÞm 1 1 e ðf1 + λDðt1 ÞÞm k1  ðf + λDðt ÞÞm 1 1 e

ðf2 + λDðt2 ÞÞm k2  ðf + λDðt ÞÞm 1 1 +e

ðf4 + λDðt4 ÞÞm k4  m +e ðf1 + λDðt1 ÞÞ

ðf3 + λDðt3 ÞÞm k3  ðf + λDðt ÞÞm 1 1 +e

ðf5 + λDðt5 ÞÞm k5  m + e ðf1 + λDðt1 ÞÞ

1:

(3)

C A

Among the expression, i ¼ 1, 2, 3, 4, 5 represent the Manzhouli channel, the Alataw channel, the Erenhot channel, the Maritime channel and the Pakistan channel respectively. When the change of the freight rate fi causes the average cost F of transportation to change, the transit demand of the P export also changes to maintain Q  F ¼ C, for F ¼ 5i¼1 pi f i, C is a constant. The demand for goods after the change of the freight rate is Q0 ¼ Q  F  (F0 )1.

Impacts of Pakistan Railways Main Line 1 (ML-1)

203

3.2 Demand price elasticity of Pakistan channel and its cross-price elasticity to the sea channel In the study of goods and services, price elasticity is often used to describe the sensitivity of goods to price changes, i.e., price elasticity of demand is a measure of the change in quantity demanded of a product in relation to its price change ( Jiang, Livingston, Room, et al., 2016). The market demand for a certain commodity changes with its price, and the price elasticity of demand is the sensitivity of the demand in response to price change ( Jinliang, 1995). Set the price elasticity coefficient of the freight demand of the Pakistan channel ε5 ¼  (ΔQ5/Q5)  (Δf5/f5)1, i.e., the ratio of the rate of change in the freight volume carried by the channel to the rate of change in the freight rate. This indicator is used to reflect the impact of the price change on the freight demand and revenue of the channel. ε ¼ 1 indicates that the fluctuation of the freight rate and the fluctuation of the freight volume offset each other, while 1 < ε  ε indicates that the freight volume change caused by the freight rate change is greater than the freight rate change, i.e. the freight demand of the channel is flexible, and the shipper’s transit demand was sensitive to the freight rate. When the range of ε is from 0 to 1, the freight demand of the channel was lack of elasticity. Similar to price elasticity of demand, the cross-price elasticity of a commodity is the sensitivity of the demand to the price change of its substitutes. Suppose that the cross-price elasticity coefficient of the sea channel to the Pakistan channel freight rate is ε45 ¼ (ΔQ4/Q4)  (Δf5/f5)1, i.e. the fluctuation of the freight volume carried by the sea channel is affected by the change of the freight rate of the Pakistan channel, the cross-price elasticity of the Pakistan channel will be the ratio of its volume fluctuation to the freight rate changes of the sea channel ε54 ¼ (ΔQ5/Q5)  (Δf4/f4)1. When εij ¼ 0, there is no substitute relationship between channel i and channel j. When εij > 0, an increase in price of channel j would lead to an increase in freight volume of channel i, and there is substitute relationship between these two Channels. The larger the value of εij, the higher the degree of substitution of channel j being be replaced by the channel i. When εij < 0, an increase in price of channel j would result in the decline in the freight volume of the channel i, and the two Channels are in a complementary relationship.

204

Market development and policy for one belt one road

4. Empirical analysis of the probability distribution model of the North China-Europe export channel 4.1 Basic export data of North China-Europe This study mainly analyzes the route choices, for export to Europe, of five provinces, Beijing, Tianjin, Hebei, Shanxi and Inner Mongolia in North China. 27 EU member states including the United Kingdom were selected for research. To facilitate the calculation, the European logistics node is set in Hamburg, i.e., the goods are transported to Hamburg, and then distributed to various countries through the European railway network assuming that the transport distance between Hamburg and other countries is the same for five transit routes. Because of the large data set, the following five countries including the United Kingdom, Spain, France, Greece, and Germany will be targeted for analysis. The data collected on distances, time, and freight rates from various provinces in North China to European countries for transportation via CR express come from the official website of CR express (Central European Train, 2017). The freight rate and time required for one TEU from North China to the three exit ports was collected through the Railway Customer Service Center (China Railway Customer Service, 2017). The distance data is from China Railway Network (2017). The freight rate, time, and distance from the exit port to Hamburg are calculated based on the above data. In summary, the data on freight rate cost and time of goods, that have been exported to several European countries via three land ports, was generated. The coastal ports in North China are dense, and most of the export goods are shipped to the destination by sea. This study selects Tianjin Port, the largest port in North China, as the outbound node of various provinces in North China area. Similarly, the transport node in Europe is selected in Hamburg, and goods are transported by rail to European countries from Hamburg. The time and the distance between the Tianjin Port and the Port of Hamburg can be collected through the retrieval software BLM-shipping, and the freight rate is from 100allin Global Freight Network, 2017. The data of railway transportation from Hamburg to other countries is the same as that of land transportation. According to the results shown in Tables 1 and 2, shipping takes the longest time in all transportation methods. Due to the advantages of sea-rail transportation, the elapsed time and cost of Pakistan channel are between those of CR express transportation and shipping. The demurrage rate is 170 yuan/day according to the demurrage standard of each shipping company provided by the Sinotrans&CSC shipbuilding website (Sinotrans-csc Network, 2018).

Impacts of Pakistan Railways Main Line 1 (ML-1)

205

Table 1 Freight rate from North China to Europe (without Pakistan channel) yuan/TEU. Freight rate

Beijing-United Kingdom Tianjin-France Hebei-Greece Shanxi-Spain Inner MongoliaGermany

Manchuria Pass

Alataw Pass

Erenhot Pass

Sea

44960.23

62329.70

42426.93

13280.24

44352.97 53420.21 50621.68 43268.73

61428.66 73516.04 67368.80 56680.23

41859.14 51862.60 47369.17 38816.74

11435.80 23482.22 19826.26 11642.74

Table 2 Transportation time from North China to Europe (including Pakistan channel) / day.

Beijing-England Tianjin-France Hebei-Greece Shanxi-Spain Inner MongoliaGermany

Manchuria Pass

Alataw Pass

Erenhot Pass

Sea

Pakistan Pass

13.11 12.90 15.14 14.16 12.70

17.15 17.26 20.01 18.55 16.16

16.19 16.02 18.95 17.82 15.20

48.52 48.18 51.23 50.03 47.98

32.80 32.84 35.29 33.99 31.84

Considering the lack of statistical data on freight volume data exported to Europe from provinces and cities in North China, freight volume is calculated based on the correlation between freight volume and export value. Data collected from the statistical yearbooks of various provinces of North China, from 2005 to 2016, was used to run the regression analysis on the export volume of 27 countries in Europe. Table 3 shows the estimated exports volume in 2034 to the five selected countries from each province in North China. Taking export freight from Beijing to the United Kingdom as an example, assuming the exchange rate of USD to RMB as 6.7659, value of TEU as 10,000 US dollars or 67,659 Yuan, the freight volume from Beijing to the United Kingdom is 5955190.8TEU. The parameter ki is the adjustment coefficient of the logit model. Based on the shipping time, freight rate and export volume ratio of sea and three existing land transport Channels, we set the existing seaborne export volume ratio is more than 85%, suppose when the freight rate of Pakistan channel is zero, the freight ratio of the channel is greater than 85%. When freight rates of Pakistan channel are higher than shipping but not greater than full land

206

Market development and policy for one belt one road

Table 3 The forecast export value from North China to European countries/0.1 million dollars. Export volume

United Kingdom France Germany Greece Spain

Beijing

Tianjin

Hebei

Shanxi

Inner Mongolia

5955190.8

529030.0

8993379.6

362575.9

18124.8

6279790.4 145378.8 13591.2 405526.4

6093792.2 154559.0 23979.5 657350.7

3502507.6 6254722.8 3309.5 41567.5

5357.5 82629.1 139048.6 10621.9

4594.6 14492.6 61716.9 423762.4

freight, its freight allocation falls somewhere between the two. On these foundations, the adjustment coefficients of different transit routes can be obtained by least squares method, n ¼ 4, !1 !2 n n X X 1 ki U i ki U i min z ¼ Qi  Q  e  e : (4) i¼1

i¼1

The adjustment coefficients of the above five routes are available as follows k1 ¼ 4.65, k2, k3 ¼ 5.00, k4 ¼ 4.44, k5 ¼ 3.62, m ¼ 2.

4.2 Influence of different time preference coefficients λ on route assignment Taking the export from Beijing to the United Kingdom as an example, the freight rate of Pakistan channel in this section is calculated based on the proportion of its distance allocated to shipping and land transport, with Manzhouli channel and shipping as a reference, we finally achieve the freight rate f5 ¼ 25763.89. The distribution ratios of each channel are as follows, with pBU1,1–pBU5,1representing Manzhouli channel, Alataw channel, Erenhot channel, sea, and Pakistan channel, respectively. 4:65

psum, 1 ¼ e

ð44960:23 + 17013:11λÞ2 ð44960:23 + 17013:11λÞ2

5:00

+e

ð62329:70 + 17017:15λÞ2 ð44960:23 + 17013:11λÞ2

ð42426:93 + 17016:19λÞ2 ð13280:24 + 17048:52λÞ2 5:00 4:44 2 ð 44960:23 + 17013:11λ Þ ð44960:23 + 17013:11λÞ2 +e +e ð25763:89 + 17032:80λÞ2 3:62 ð44960:23 + 17013:11λÞ2 : +e

(5)

Impacts of Pakistan Railways Main Line 1 (ML-1)

207

Proceeding to the next step, ð44960:23þ17013:11λÞ2 ð44960:23þ17013:11λÞ2

4:65

pBU1,1 ¼ e

ð62329:70þ17017:15λÞ2 ð44960:23þ17013:11λÞ2

5:00

pBU2,1 ¼ e

ð42426:93þ17016:19λÞ2 ð44960:23þ17013:11λÞ2

5:00

pBU3,1 ¼ e

ð13280:24þ17048:52λÞ2 ð44960:23þ17013:11λÞ2

4:44

pBU4,1 ¼ e

ð25763:89þ17032:80λÞ2 ð44960:23þ17013:11λÞ2

3:62

pBU5,1 ¼ e

 p1 sum,1 ,  p1 sum,1 ,  p1 sum,1 ,  p1 sum,1 ,  p1 sum,1 :

According to Fig. 3, when the time preference coefficient λ ¼ 0, due to its absolutely low freight rate, shipping accounts for 68% of the whole export volume transit from Beijing to the United Kingdom, followed by the 30% for Pakistan channel, and the remaining freight transported through three CR express only accounts for less than 2%. When the export market remains sluggish, the sea channel and Pakistan channel have an absolute advantage in China-EU freight transit; in particular, only λ  [0, 3.16], the advantage of low freight rate by shipping makes itself own the largest market share. As the cargo owner’s preference for time increases, the proportion of freight channels other than shipping increases gradually. At λ ¼ 5.18, the proportion of cargo through Pakistan channel will reach to the maximum of 53.7%, and as the λ continues to increase, i.e. if the export market in China-EU is increasingly prosperous, the proportion of freight transit through Pakistan 1

Manchuria Pass Alataw Pass Erenhot Pass Sea Pakistan Pass

0.9 0.8

Probability value

0.7 (5.184,0.537)

0.6 0.5

(3.166, 0.453)

0.4 0.3 0.2 0.1 0

0

10

20

30

40

50

60

Fig. 3 Influence of time preference coefficient on path distribution.

208

Market development and policy for one belt one road

channel will decrease. When λ > 30, the freight volume allocation of the Pakistan channel is nearly zero. In general, the proportion of seaborne exports is greatly affected by time preference; as shown in the graph, at 0 < λ < 10, the proportion of shipping drops from nearly 70% to 0%. The proportion of freight through Pakistan channel is always between that of the CR express and the shipping. With the increasing emphasis on time by shippers, the advantage of the short time in the Manzhouli channel becomes increasingly prominent.

4.3 Impacts of Pakistani channel freight rate on export route selection According to Fig. 3, considering the export market is under the moderate conditions, we set λ ¼ 3.16. when f5 the freight rate of Pakistan channel changes, the export distribution ratio of each channel is as follows: where pBU1, 2–pBU5, 2 representing the probability of allocation to Manzhouli channel, Alataw channel, Erenhot channel, sea, and Pakistan channel, respectively. ð44960:23þ17013:113:16Þ2 ð44960:23þ17013:113:16Þ2

4:65

psum,2 ¼ e

þe

ð42426:93þ17016:193:16Þ2 ð44960:23þ17013:113:16Þ2

þe

5:00

þe

3:62

2 ðf 5 þ17032:803:16Þ ð44960:23þ17013:113:16Þ2

ð62329:70þ17017:153:16Þ2 ð44960:23þ17013:113:16Þ2

5:00

þe

(6)

ð13280:24þ17048:523:16Þ2 ð44960:23þ17013:113:16Þ2

4:44

:

Proceeding to get pBU1, 2 ¼ e

4:65

ð44960:23þ17013:113:16Þ2 ð44960:23þ17013:113:16Þ2 2

pBU2, 2 ¼ e

5:00

ð62329:70þ17017:153:16Þ ð44960:23þ17013:113:16Þ2

2

pBU3, 2 ¼ e

5:00

ð42426:93þ17016:193:16Þ ð44960:23þ17013:113:16Þ2

2

pBU4, 2 ¼ e

4:44

ð13280:24þ17048:523:16Þ ð44960:23þ17013:113:16Þ2 2

pBU5, 2 ¼ e

3:62

ðf5 þ17032:803:16Þ ð44960:23þ17013:113:16Þ2

 p1 sum, 2 ,  p1 sum, 2 ,  p1 sum, 2 ,  p1 sum, 2 ,

 p1 sum, 2 :

According to the freight ratio curve of each channel in Fig. 4, when the freight rate of Pakistan channel is about f5 ¼ 25, 783, it shares the same freight proportion with sea. When f5 < 25, 783, Pakistan channel has the absolute advantage in all channels. When the freight rate decreases to 0, it accounts

Impacts of Pakistan Railways Main Line 1 (ML-1)

1.2

209

Manchuria Pass Alataw Pass Erenhot Pass Sea Pakistan Pass

1

Probability value

0.8

0.6 (25783.410,0.450) 0.4

0.2

0

(41405.530,0.088)

0

1

2

3

(42788.018,0.077) 4 5

6 x104

Fig. 4 The influence of freight of Pakistan channel on route allocation.

for 90% of the total export volume of Beijing-United Kingdom. However, the cargo volume of the CR express only exceeds that of Pakistani channel at f5 > 41, 405. Due to the fact that the Alataw channel stretches across many countries and has the highest freight rate, its existing freight transit ratio is less than 3%. Thus, when the Pakistan’s ML-1 is operated, the proportion of the freight transit through Alataw channel can almost be negligible. According to the probability distribution curve of shipping and Pakistan channel, there is a clear substitution between the two.    1 ε5 ¼  ΔQ5  Q1  Δf 5  f 1 : (7) 5 5 Fig. 5, derived from the freight rate demand elasticity coefficient of the Pakistan channel through Eq. (7), shows that when the transportation conditions of other channels are unchanged and the freight rate of the Pakistan channel is between [0, 13045], the cargo owner’s demand for freight through this channel is inelastic. Considering f5 ¼ 25,763 as reference, under this pricing level, the freight demand of the channel is flexible and a percentage decrease in freight rates would bring it a higher freight increase; when f5 is lower than 13,045 yuan/TEU, continuous price reduction will not bring the channel with the same level of increase in the freight volume. Overall, Pakistan channel has a large competitive advantage. For any f5 > 13,045, as long as the price reduction happens, it would be attractive for cargo’s owners, this threshold is much smaller than the reference price 25,763 yuan/TEU. As is shown in Fig. 6, the trend of the freight volume of the Pakistan is about the same with total volume, i.e., they both decreases with the increase of the freight rate of the Pakistan channel. With the increase in freight rate

210

Market development and policy for one belt one road

6

Elastic coefficient

5

4

3

2 (13045.454545,1.010444)

1

0

0

0.5

1

1.5

2

2.5

3

3.5

4 x104

Fig. 5 Price elasticity coefficient of Pakistan channel.

4.5

x107

Export Volume

Manchuria Pass

4

Alataw Pass Erenhot Pass

3.5

Sea Pakistan Pass Total volume

3 2.5 2 1.5 1 0.5 0

0

0.5

1

1.5

2

2.5

3

3.5

4 x104

Fig. 6 Freight volume allocation under the change of Pakistan channel’ freight rate.

from 0 to 32,000, the proportion of the export volume allocated to Pakistan channel drops from 87.1% to 27.4%. The proportion of sea transportation decreases first with the decline of total freight volume. However, when the freight rate of the Pakistani channel rises to 15,000, the advantage of

Impacts of Pakistan Railways Main Line 1 (ML-1)

211

Table 4 Changes in export volume affected by the freight rate of Pakistan channel. 0

Manzhouli channel Alataw channel Erenhot channel Sea Pakistan channel

4000

8000

12,000 16,000 20,000 24,000 28,000 32,000

0.013 0.015 0.019 0.023

0.030

0.038

0.049

0.060

0.071

0.000 0.000 0.000 0.000

0.000

0.000

0.000

0.000

0.001

0.011 0.013 0.016 0.020

0.025

0.032

0.041

0.050

0.059

0.106 0.127 0.157 0.198 0.871 0.845 0.809 0.759

0.253 0.692

0.324 0.605

0.410 0.501

0.504 0.386

0.596 0.274

the low freight rate of the shipping channel is much more remarkable and accompanied by the increase in freight volume, which comes from the reduction of freight volumes in the Pakistan channel as a result of its higher freight rates. According to Table 4, the total freight volume of the three land transport Channels remained between [2%, 13%], and as the freight rate of the Pakistan channel increases, does not change significantly compared with shipping.

4.4 Impacts of freight rate of Pakistan channel on shipping Eqs. (8) and (9) calculate the cross-price elasticity coefficient of the shipping and the Pakistan channel, as shown in Figs. 7 and 8. During 0 < f4 < 8894, ε54 < 0, the increase in the freight rate of the sea channel will lead the transit volume of the Pakistan channel to decline, they have a complementary relationship. Since the shipping freight rate is low at this time, the proportion of the freight volume occupied by sea channel is relatively large and hence the increase in freight rate of shipping will lead to a larger decrease in the average freight rate than the other four Channels and eventually makes the total freight volume decrease. As a result, the export volume of the Pakistan channel will also decrease. Under such circumstances, Pakistan channel and sea channel have a complementary relationship; when f4 > 8894, Pakistan channel’ advantages in freight rate and moderate travel time are beginning to emerge. In such a situation, the increase in the freight rate of shipping will make the freight volume of Pakistan channel increase, and the proportion of sea freight has no longer the initiative in the average market price. As is seen in the Fig. 7, they are in the substitutional relationship. In particular, the sea

212

Market development and policy for one belt one road

0.4

Cross elasticity coefficient

0.3

0.2

0.1

0 (8894.009217,0.000292)

-0.1

-0.2 0

0.5

1

1.5

2

2.5

3

3.5

4 x104

Fig. 7 Cross-price elasticity of sea freight rate on Pakistan channel. 1.4 1.2

Cross elasticity coefficient

1

(32396.313364,1.274636)

0.8 0.6 0.4 0.2 0

-0.2

(12304.147465,-0.002332)

-0.4 -0.6

0

0.5

1

1.5

2

2.5

3

3.5

4 x104

Fig. 8 Cross-price elasticity of Pakistan channel’ freight rate on shipping.

channel is replaced by the Pakistan channel with the highest efficiency when f4 ¼ 22, 500, and the cross-price elasticity is about 0.37. For ε45, when 0 < f5 < 12,304, the freight volume of shipping decreases with the increase in the freight rate of the Pakistan channel. The two have a

Impacts of Pakistan Railways Main Line 1 (ML-1)

213

Table 5 The change of export volume allocations for shipping from Beijing-United Kingdom.

1  20% 1  10% 1 1 + 10% 1 + 20%

0

4000

8000

12,000 16,000 20,000 24,000 28,000 32,000

0.142 0.123 0.106 0.091 0.077

0.168 0.147 0.127 0.109 0.093

0.205 0.180 0.157 0.135 0.116

0.255 0.226 0.198 0.172 0.148

0.320 0.286 0.253 0.222 0.192

0.400 0.362 0.324 0.287 0.252

0.491 0.451 0.410 0.369 0.328

0.585 0.546 0.504 0.461 0.417

0.672 0.635 0.596 0.554 0.509

complementary relationship. The reason is similar to the analysis of ε54. At this freight rate, the Pakistan channel has a large advantage in freight transportation, so the rise of its freight rate will lead to the reduction of total freight volume of the five Channels. Therefore, the graph shows that the two have a complementary relation; when f5 ¼ 32,396, the Pakistan channel is replaced by sea channel, the efficiency is the highest, and the cross-price elasticity coefficient is 1.27.   1 1 ε54 ¼ ΔQ5  Q1 , (8) 5  Δf 4  f 4   1 1 ε45 ¼ ΔQ4  Q1 : (9) 4  Δf 5  f 5 Table 5 shows the change in proportion of shipping freight volume when the freight rates of sea channel and Pakistan channel are both changed. Overall, the freight volume of the sea channel varies with its own freight rate fluctuations and change in freight rate of Pakistan channel, its proportion ranges between [7.7%, 67%]. Accounting for 85% of the current sea transportation ratio, the operation of the Pakistan channel will attract the sea freight transit volume by at least 18%.

4.5 Income analysis of Pakistan and Singapore under the double variable constraint of shipping freight rate and Pakistan channel freight rate A path allocation probability model is established based on the double variables of sea freight rates and that of Pakistan channel, the probability of distribution of each channel is as follows, where f4, f5 are the independent variables.

214

Market development and policy for one belt one road

ð44960:23þ17013:113:16Þ2 ð44960:23þ17013:113:16Þ2

4:65

psum,4 ¼ e

5:00

þe

þe

ð42426:93þ17016:193:16Þ2 ð44960:23þ17013:113:16Þ2

ð62329:70þ17017:153:16Þ2 ð44960:23þ17013:113:16Þ2

5:00

4:44

þe

(10)

2 ðf 4 þ17048:523:16Þ ð44960:23þ17013:113:16Þ2

2

3:62

þe

ðf 5 þ17032:803:16Þ ð44960:23þ17013:113:16Þ2

:

Which leads to ð44960:23þ17013:113:16Þ2 ð44960:23þ17013:113:16Þ2

4:65

pBU1,4 ¼ e

ð62329:70þ17017:153:16Þ2 ð44960:23þ17013:113:16Þ2

5:00

pBU2,4 ¼ e

ð42426:93þ17016:193:16Þ2 ð44960:23þ17013:113:16Þ2

5:00

pBU3,4 ¼ e

2 ðf 4 þ17048:523:16Þ ð44960:23þ17013:113:16Þ2

4:44

pBU4,4 ¼ e

2 ðf 5 þ17032:803:16Þ ð44960:23þ17013:113:16Þ2

3:62

pBU5,4 ¼ e

 p1 sum,4 ,  p1 sum,4 ,  p1 sum,4 ,

 p1 sum,4 ,  p1 sum,4 :

When f4, f5 take the reference price 13280.24 and 25763.89, respectively, the average price of the five Channels is set as follows: 0

4:65ð44960:23þ17013:113:16Þ2

1

5:00ð62329:70þ17017:153:16Þ2

2 þ 62329:70  e ð44960:23þ17013:113:16Þ2 þ C B 44960:23  e ð44960:23þ17013:113:16Þ B C 2 5:00ð42426:93þ17016:193:16Þ 4:44ð13280:24þ17048:523:16Þ2 C B B 42426:93  e ð44960:23þ17013:113:16Þ2 þ 13280:24  e ð44960:23þ17013:113:16Þ2 C B C @ A 2 3:62ð25763:89þ17032:803:16Þ

F¼0 B B @

þ25763:89  e 4:65ð44960:23þ17013:113:16Þ2 e ð44960:23þ17013:113:16Þ2

þ

ð44960:23þ17013:113:16Þ2

5:00ð62329:70þ17017:153:16Þ2 e ð44960:23þ17013:113:16Þ2

4:44ð13280:24þ17048:523:16Þ2 þe ð44960:23þ17013:113:16Þ2

þ

þ

5:00ð42426:93þ17016:193:16Þ2 e ð44960:23þ17013:113:16Þ2

3:62ð25763:89þ17032:803:16Þ2 e ð44960:23þ17013:113:16Þ2

1: C C A

(11)

The average price F0 after changing f4, and f5 is F 0 ¼ pBU1,4  f 1 + pBU2,4  f 2 + pBU3,4  f 3 + pBU4,4  f 4 + pBU5,4  f 5 : (12) The sea channel of interest is shipped to Europe via the Mediterranean Sea and the Straits of Malacca. For Singapore, the main source of income in China-EU trade is refueling, transshipment, and related financial services. Estimated proportion of Singapore’s revenue in total transportation costs is 8%. Based on this, Singapore’s income is R4 ¼ 0.08  f4  pBE4, 4  Q0 , where

Impacts of Pakistan Railways Main Line 1 (ML-1)

215

Q0 ¼ (F  Q)/F0 , the initial total freight volume is estimated Q ¼ 5955190.8 TEU, then R4 ¼ 0

0

4:44

2 ðf 4 +17048:523:16Þ

0:08  f 4  Q  e

ð44960:23+17013:113:16Þ2

ð44960:23+17013:113:16Þ2 ð44960:23+17013:113:16Þ2

5:00

ð62329:70+17017:153:16Þ2 ð44960:23+17013:113:16Þ2

4:65

1:

(13)

+e C B e 22 C B ðf 4 +17048:523:16Þ ð42426:93+17016:193:16Þ2 5:00 4:44 C B 2 2 ð44960:23+17013:113:16Þ ð44960:23+17013:113:16Þ C B +e +e A @ 2 ðf 5 +17032:803:16Þ 3:62 2 ð44960:23+17013:113:16Þ +e The setting of the Pakistan channel freight rate will be estimated from the freight rate of the other existing channels. The total land transport length of the channel is 7649 km, of which the domestic transport distance in Pakistan is 2426 km. Referring to the freight rate of Manzhouli channel, the proportion of the cost of Pakistan’s domestic transit is 22%, which earns Pakistan the domestic transportation revenue of R5 ¼ 0.22  f5  Q0  pBE5,4 which yields R5 ¼ 0

0

3:62

2 ðf 5 +17032:803:16Þ

0:22  f 5  Q  e

ð44960:23+17013:113:16Þ2

ð44960:23+17013:113:16Þ2 ð44960:23+17013:113:16Þ2

5:00

4:65

ð62329:70+17017:153:16Þ2 ð44960:23+17013:113:16Þ2

1:

(14)

+e C B e 22 C B ðf 4 +17048:523:16Þ ð42426:93+17016:193:16Þ2 5:00 4:44 C B 2 2 ð44960:23+17013:113:16Þ C ð44960:23+17013:113:16Þ B +e +e A @ 2 ðf 5 +17032:803:16Þ 3:62 2 ð44960:23+17013:113:16Þ +e The results of Eqs. (13) and (14) are depicted in the Figs. 9–11. Under the bivariate constraints, the maximum income achieved for Singapore is obtained at about f4 ¼ 10,000, whereas Pakistan’s maximum income is obtained around f5 ¼ 13,000. According to the pricing rules in game theory, when the two players are in a non-cooperative game, there is a unique Nash equilibrium solution (707,2974), i.e., Singapore’s comprehensive price is 707 yuan and that of Pakistan is 2974 yuan. The total income of the two countries is 1.56  1010 yuan. When they are in completely cooperative games, the pricing goal is to maximize total income of the two countries. Fig. 11 shows the total income surface under different pricing combinations of the two countries when they are in fully cooperative games. When Singapore is priced at 707 yuan, under the system’s optimal situation, Pakistan’s price should be set at 2696.77 yuan. Meantime, the total income of the system is 1.65  1010 yuan. When the freight rate of the Pakistan channel and the sea channel is lower than the land channel freight rate, the alliance still

216

Market development and policy for one belt one road

Singapore’s revenue

x109 6

4

2

0 8 6 4

x104

2 0

0

2

freight of Pakistan pass

4

8

6 x104

freight of sea

Fig. 9 Singapore’s revenue under bivariate constraints.

x1010

Pakistan’s revenue

2 1.5 1 0.5 0 8 6 x104

4 2 0

0

2

4

freight of sea

freight of Pakistan pass

Fig. 10 Pakistan’s revenue under bivariate constrains.

8

6 x104

Impacts of Pakistan Railways Main Line 1 (ML-1)

217

x1010 2

Total revenue

1.5 1 0.5 0 8 6 x104

4 2 0

0

2

4

8

6 x104

freight of sea

freight of Pakistan pass

Fig. 11 The total revenue of Pakistan and Singapore Pakistan channel.

has an absolute advantage in the export market. Thus, they both rise together and the combination of prices is to obtain the optimal value for the alliance.

5. Conclusions and emerging directions 5.1 Concluding remarks With the multi-logit model, we analyze the cargo transportation pattern from the North China to Europe and its change after the completion of the Pakistan ML-1 line. We introduce the time preference coefficient that indicates different export market situations, the Pakistan channel and shipping freight rate as independent variables to study the impact on the results of the route allocation. In addition, this study also estimates the operational impacts of the Pakistan channel on the freight volume allocated by ship. Through the introduction of new variable in the multi-logit model, the proportion of shipping exports is considerably affected by the time preference; as λ rises from 0 to 10, the proportion of sea channel falls from nearly 70% to 0%; the Pakistan channel is moderately time-consuming and attractive in freight rate compared to the CR express. With the increasing concern of shippers on freight time, the advantage of the short time required in the Manzhouli channel becomes more prominent. Under the reference of pricing standard, the transit demand of the Pakistan channel is flexible. Under the

218

Market development and policy for one belt one road

condition in which the freight rate of the channel is lower than the existing shipping freight rate, the freight demand of Pakistan channel becomes inelastic. According to the result, Pakistan channel will attract the transit volume at least 18% of shipping. In addition, the cross-price elasticity of the sea channel and Pakistan channel shows that the two are not absolutely in the relationship of substitution or complementary, but are related to market initiatives in the current situation. When Pakistan and Singapore are in a complete cooperative game, the alliance forms a monopoly pattern, and thus there is a trend of associative rise in the price at both sides. In the longer run, Pakistan’s ML-1 line will serve as the main artery connecting the land and Maritime Silk Road. It has great competitive strength both in terms of international influence and transportation effectiveness. The opening of this channel will not only affect the export and transit of North China to Europe but will also help to make Pakistan a trade connection to the inland regions of central and western China.

5.2 Emerging directions This study takes the export transit of North China-Europe as the research objective. And uses the discrete selection model to analyze the freight volume that can be allocated after the completion of the Pakistan’s ML-1 line, considering the historical data of shipping freight rate fluctuation and the freight rate change of Pakistan channel, and the competition between the five transport alternatives under the influence of these two factors. However, there are still some shortcomings in the model. To some extent, using regression analysis to predict the future freight volume will undermine the diversity and unpredictability of some certain factors, making the regression analysis limited in some cases. In addition, there is a certain calculation deviation between the export volume and its loading method between the actual and the expected one. However, prediction of the export volume doesn’t affect the distribution probability. The discrete model does have some great advantages in analyzing individual choices routes; however, when confronted with different modes of transit, the selection probabilities of various modes are determined only by the difference between the intermode utility functions. The behavioral preference and the random utility items such as the type of goods are independent of each other. There is a certain gap between the above applied methods and the actual situation. In the future research, other disaggregate models such as BL, NL, MMNL can be considered for further study and comparison. Subsequent research can also extend the transit alternatives.

Impacts of Pakistan Railways Main Line 1 (ML-1)

219

Acknowledgments We feel grateful for the support from the Science and Technology Commission of Shanghai Municipality (Grant No.: 17040501800).

References Bhat, C. R. (1995). A heteroscedastic extreme value model of intercity travel mode choice. Transportation Research Part B-Methodological, 29(6), 471–483. China Railway Network. (2017). http://www.12306.cn/mormhweb/. 2017-03-22 (In Chinese). Central European Train. (2017). http://cn.cetrains.com/post/wuliujiage.html. 2017-3-22 (In Chinese). China Railway Customer Service Center. (2017). http://www.12306.cn/mormh web/ hyfw/hyckcx/. 2017-3-22. Global Freight Network. (2017). http://www.100allin.com/fcl/. 2017-03-22 (In Chinese). Jiang, H., Livingston, M., Room, R., et al. (2016). Price elasticity of on- and off-premises demand for alcoholic drinks: A Tobit analysis. Drug & Alcohol Dependence, 163, 222–228. Jinliang, D. (1995). Price of goods, demand elasticity and total income. Journal of Henan Radio & Television University, (Z1). 14–16 +5. (In Chinese). Luce, R. D. (1959). Individual choice behavior. American Economic Review, 67(1), 1–15. Mcfadden, D., & Train, K. (2000). Mixed MNL models for discrete response. Journal of Applied Econometrics, 15(5), 447–470. National Railway Administration. (2018). China and Pakistan jointly signed the cooperation document for the feasibility study of the China-Pakistan Economic Corridor Railway Project. http:// www.nra.gov.cn/xwzx/xwdt/xwlb/201504/t20150422_13225.shtml. 2018-03-22 (In Chinese). The Ministry of Commerce Comprehensive Department. (2018). Promoting the vision and action of CO constructing the Silk Road Economic Belt and the twenty-first Century Maritime Silk Road. http://www.mofcom.gov.cn/article/resume/n/201504/2015040092S9655. shtml. 2018-03-22 (In Chinese). Sinotrans-csc Network. (2018). http://ebooking.sinoagentxm.com/ebooking/EirManage/ TKFreghtEstima-te.aspx. 2018-3-22 (In Chinese). Shi, H., & Yin, G. (2018). Boosting conditional logit model. Journal of Choice Modelling, 26, 48–63. Train, K. (2003). Discrete choice methods with simulation. Publications of the American Statistical Association, 100(469), 351–352. Tversky, A. (2004). Elimination by aspects: A theory of choice. Psychological Review, 79(4), 281–299. Wang, X., & Meng, Q. (2011). The impact of land bridge on the market shares of Asian ports. Transportation Research Part E Logistics and Transportation Review, 47(2), 190–203. Williams, H. C. W. L. (1977). On the formulation of travel demand models and economic evaluation measures of user benefit. Modern Industrial Economy and Informationization, 9(3), 285–344.

This page intentionally left blank

Index Note: Page numbers followed by f indicate figures and t indicate tables.

A Air connectivity, 99 Airline, benchmark case, 182–183 Airline-HSR cooperation, literature review, 173–175 Airports role, 189–190, 190f Air-rail cooperation, 183–189 competition, 187–189, 188t landing fee, 185–187 traffic volumes, 184–185 Air services agreements (ASAs), 109–110 Air transport, connectivity, landlockedness, 109–111 Nepal, 111, 112t Air transport subsidies, 99 Almaty Programme of Action (APoA), 108 Asian Infrastructure and Investment Bank (AIIB), 27, 62 Association of Southeast Asian Nations (ASEAN), 111

B Belt and Road (B&R), 65 Belt and Road Initiative (BRI), 1–2, 27, 60–62, 62t, 77, 155 China, 9 countries, 18 cruise industry, 3, 60–61 cruise market exploration, 66–68, 67t data analysis, 16–18, 17–18t literature review, 11–13 models, market development, 23 models, policy implications, 18–23, 21t Brazil, Russia, India, China, and South Africa (BRICS), 27 BRI, cruise itinerary identification, evaluation, 68–74 BR, international friends circle, 73–74 BR itineraries, heating up, 68–71, 70–71t costa, dream trip, 72–73 Xiamen, BR international cooperation, 72

BRI, cruise policy, 62–66 China’s port construction promotion, 64 opportunities, China’s cruise industry, 63–64 tourism collaboration strengthening, 64–65 tourism visa cooperation, 66 Build-operate-transfer (BOT), 50–51

C Capacity, 155 China–Pakistan Economic Corridor (CPEC), 91 China Railway Corporation (CRC), 33–34 CHINA RAILWAY Express (CR Express), 197–198 China State Construction Engineering Corporation (CSCEC), 92–93 Chinese overseas port investment, 78–83 investors, 80 port investment efficiency factors, 80–83, 81f stages, 78–80, 79f Civil Aviation Administration of China (CAAC), 33 Collusion detection model, 130–138, 131f, 132t both pairs collusive pricing, 138 only one pair collusive pricing, 137–138 uniform pricing, 134–136 Collusion, ocean container transport, 127–130 Collusive pricing solutions, 139–141 Container terminal, 125 CPEC. See China–Pakistan Economic Corridor (CPEC) Critical success factors (CFSs), 82 Cross-sectional (CS) ordinary least squares, 16–17 Cruise tourism, 60, 60f

221

222

Index

Cruise travel, China, 69 CSCEC. See China State Construction Engineering Corporation (CSCEC)

D Deregulation, 36 Difference-in-differences (DID), 37 Discrete choice model, 200

E Empirical analysis, North China–Europe export channel, 204–217 basic export data, North China–Europe, 204–206, 205–206t freight rate, Pakistan channel, 208–213, 209–210f, 211t, 212f, 213t income analysis, Pakistan, Singapore, 213–217, 216–217f time preference coefficients λ, route assignment, 206–208, 207f Equilibrium profits, 193 Essential air services, small communities, 112–115, 113t European Union (EU), 171 Export volume, Beijing, inner Mongolia, European countries, 198–199, 199f Extended gravity model, BR countries, 13–16

F Foreign direct investment(s) (FDI(s)), 1, 10

G Generalized extreme value model (GEV), 200–201 Geographical location, cruise destination, 68–69 Global trade impact logistics performance, 11–12 transport connectivity impact, 9–10 Gravity trade model, 16–17 Gross domestic product (GDP), 15, 27, 61, 105–106

H Hambantota International Port Group (HIPG), 87

Heteroscedastic extreme value model (HEV), 200–201 High-speed rail (HSR), 2–3, 171 Hong Kong Trade Development Council (HKTDC), 61 HSR developments, China, 41–54 construction financing, 50–52, 52f express freight service, 48–50, 49t overall developments, 42–48, 43t, 44–45f, 47t spatial disparities impacts, 52–54 Hub-and-spoke networks, 5–6

I Import tax, 156 Intermodal cooperation, incentives, 175–180, 177t, 178f International Civil Aviation Organisation (ICAO), 107 Investment behaviors, 158–162 best capacity responses, 159–161 equilibrium investments, 160f, 161–162, 162f

L Landlocked developing countries (LLDCs), 4, 102–104, 103f Liner routes, Shanghai port, Mediterranean, 141–144, 142–144t Liner shipping, 125 Liner Shipping Bilateral Connectivity Index (LSBCI), 2, 10 Liquified natural gas (LNG), 89 Liquified petroleum gas (LPG), 89–90 LLDCs’ air connectivity, 106–109 LLDCs, challenges, 104–106 being landlocked, 104–105 higher transport costs, international markets, 105–106 insufficient transport infrastructure, 105 Logistics Performance Index (LPI), 2, 10 Low-cost carrier (LCC), 54

M Maritime connectivity, global trade, 2 Maritime Silk Road (MSR), 4–5, 27, 61 Mid-to-Long-Term Railway Network Plan (MLTRNP), 42

Index

223

Ministry of Railways (MOR), 33 Ministry of Transport (MOT), 33–34, 125 Mixed multinomial logit model (MMNL), 200–201 Model, 157, 180–182, 180f Multiple-airport system (MAS), 175

China–Euro Railway Express (CR Express), 38–41, 39–40f, 41t rail sector developments, major reforms, 29–37, 29–32t, 30f, 34f, 38f Research tools, 200–201

N

Siberian Land Bridge (SLB), 38 Silk Road Economic Belt, 27 Silk Routes, 68–69 Singapore International Port Group (SIPG), 91 Standard international trade classification (SITC), 17–18 State-owned Assets Supervision and Administration Commission of the State Council of China (SASAC), 84 State-owned corporation (SOE), 33–34 Strategic sectors, 33 Sunk costs, 172 Supply chain host, 83 Supply chain integrator, 77–78

National Development and Reform Commission (NDRC), 61, 125 Nested logit (NL), 200–201 New Silk Road Economic Belt, 77, 78f

O Oligopoly, import taxes, 163–168 demand functions, 163–164 equilibrium prices, 164–166 welfare assessment, 166–168, 167f One Belt One Road Initiative (OBOR), 100–102, 101f Ordinary least squares (OLS), 16–17 Overseas investment efficiency, 3–4

P Pakistan’s Railways Main Line (ML-1), 6 Per-capita GDP (PCGDP), 15 Policy implications, 145 Pooled cross-sectional (PCS) ordinary least squares, 16–17 Population of exporters (POP), 15 Probability distribution model, parameter analysis, north China–Europe export channel, 201–203 demand price elasticity, Pakistan channel, 203 distribution probability, each export channel, 201–202 Public-private-partnership (PPP), 82 Public Service Obligation (PSO)/Essential Air Services (EAS), 112–113

R Rail sector, 28 Railway developments, China, 28–41, 29f

S

T TEN-T. See Trans European NetworkTransport (TEN-T) Terminal handling charge (THC), 125 Textile sector, 14 Trade, 155 Trans European Network-Transport (TEN-T), 171 Transit countries, 155–156 Transport, 155 Transport infrastructure, 99 Typical ports, case studies, 83–95 Gwadar port, 86–91, 92–93f Hambantota port, 83–86, 88f port of Piraeus, 83–95, 83t, 86f

U United Nations Conference on Trade and Development (UNCTAD), 10

W Win-win outcome, LLDCs, 115–118

This page intentionally left blank