Economic Development under Climate Change: Economy-Wide and Regional Analysis for Ethiopia 3658294124, 9783658294120

Citation preview

Amsalu Woldie Yalew

Economic Development under Climate Change Economy-Wide and Regional Analysis for Ethiopia

Economic Development under Climate Change

Amsalu Woldie Yalew

Economic Development under Climate Change Economy-Wide and Regional ­Analysis for Ethiopia

Amsalu Woldie Yalew Potsdam, Germany Dissertation at Technische Universität Dresden, Germany, 2017

ISBN 978-3-658-29412-0 ISBN 978-3-658-29413-7  (eBook) https://doi.org/10.1007/978-3-658-29413-7 © Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer VS imprint is published by the registered company Springer Fachmedien Wiesbaden GmbH part of Springer Nature. The registered company address is: Abraham-Lincoln-Str. 46, 65189 Wiesbaden, Germany

Acknowledgments This monograph is a PhD dissertation, with the same title, that I submitted to the Technische Universität Dresden in 2017. There are only slight modifications relative to the copy I submitted to the university library. I am, therefore, indebted to my supervisors and my graduate school. I would like express my sincere gratitude to Prof. Dr. Georg Hirte (main supervisor, Technische Universität Dresden), and to Prof. Dr. Hermann Lotze-Campen (co-supervisor, Potsdam Institute for Climate Impact Research and Humboldt Universität zu Berlin) for all of their guidance, and constructive comments and remarks throughout the study period. I am also grateful to Dr. Stefan Tscharaktschiew (mentor, Technische Universität Dresden) who generously provided comments in all stages of the work. I would also like to extend my gratitude to Prof. Dr. Bernhard Müller (Dresden Leibniz Graduate School, Leibniz Institute for Ecological Urban and Regional Development, and Technische Universität Dresden) and Dr. Paulina Schiappacasse (Dresden Leibniz Graduate School and Technische Universität Dresden) who also served in my doctoral Advisory Committee in the study period, and whose remarks in particular helped me to give an interdisciplinary touch to the monograph. I am grateful to the Dresden Leibniz Graduate School (DLGS) for its generous doctoral fellowship and Leibniz Institute of Ecological Urban and Regional Development (IÖR) for its wonderful research facilities. I am grateful to all DLGS fellows of cohort 3, 4, 5, and 6 for keeping the working environment cheerful. At last, but not least, I would like to thank all participants who gave me valuable comments and remarks during my presentations in Nordic Conference in Development Economics (13-14 June 2016, Oslo); International Conference of Economic Modeling (6-8 July 2016, Lisbon), International Conference on Shocks and Development (6-7 October 2016, Dresden); Conference on Energy and Climate Economic Modeling (3-4 November 2016, Prague); and DLGS/IÖR Autumn School (Autumn 2014, 2015, and 2016, Dresden).

Amsalu Woldie Yalew November 2019

Table of Contents Acknowledgments ................................................................................................................ V List of Figures .................................................................................................................... IX List of Tables ..................................................................................................................... XI Acronyms ........................................................................................................................ XIII 1 2

3

4

5

Introduction ...................................................................................................................1 Theoretical and Conceptual Background .....................................................................7 2.1

Introduction ..............................................................................................................7

2.2

Climate and climate change ......................................................................................7

2.3

Climate change, agriculture, and development ..........................................................8

2.4

Climate change, public policy, and development .................................................... 10

2.5

Computable general equilibrium models ................................................................. 11

Overview of the Ethiopian Economy .......................................................................... 15 3.1

Introduction ............................................................................................................ 15

3.2

Geography.............................................................................................................. 15

3.3

Population and labor force ...................................................................................... 16

3.4

Rural-urban divide .................................................................................................. 17

3.5

Administrative regions of Ethiopia .........................................................................18

3.6

Macro-economy ..................................................................................................... 20

3.7

Public finance ......................................................................................................... 21

3.8

Prospects of the economy ....................................................................................... 23

Methodological Framework ........................................................................................ 27 4.1

Introduction ............................................................................................................ 27

4.2

CGE model choice.................................................................................................. 27

4.3

Description of the CGE model ................................................................................ 28

4.4

The CGE model database ....................................................................................... 32

4.5

The CGE model calibration .................................................................................... 43

4.6

CGE simulations and economy-wide analysis ......................................................... 44

4.7

Regional projections and analysis ........................................................................... 45

Impacts of Climate Change ......................................................................................... 51 5.1

Introduction ............................................................................................................ 51

5.2

Climate change, agriculture, and migration ............................................................. 51

5.3

Climate change and Ethiopia .................................................................................. 53

5.4

Materials and methods ............................................................................................ 57

5.5

Economy-wide results and analysis ........................................................................ 63

VIII

6

7

8

9

Table of Contents

5.6

Regional projections and analysis ........................................................................... 68

5.7

Conclusions ............................................................................................................ 70

Costs of Planned Adaptation....................................................................................... 71 6.1

Introduction ............................................................................................................ 71

6.2

Planned public adaptation to climate change ........................................................... 71

6.3

Adaptation to climate change in Ethiopia ................................................................ 72

6.4

Materials and methods ............................................................................................ 74

6.5

Economy-wide results and analysis ........................................................................ 80

6.6

Regional projections and analysis ........................................................................... 84

6.7

Conclusions ............................................................................................................ 85

Public Finance for Adaptation .................................................................................... 87 7.1

Introduction ............................................................................................................ 87

7.2

Adaptation finance in developing countries ............................................................ 87

7.3

Adaptation finance in Ethiopia ............................................................................... 89

7.4

Materials and methods ............................................................................................ 90

7.5

Economy-wide results and analysis ........................................................................ 92

7.6

Regional projections and analysis ........................................................................... 96

7.7

Conclusions ............................................................................................................ 97

Climate-Resilient Development................................................................................... 99 8.1

Introduction ............................................................................................................ 99

8.2

Structural change and climate change ..................................................................... 99

8.3

Structural change in Ethiopia ................................................................................ 100

8.4

Materials and methods .......................................................................................... 101

8.5

Economy-wide results and analysis ...................................................................... 105

8.6

Regional projections and analysis ......................................................................... 107

8.7

Conclusions .......................................................................................................... 108

Conclusions and policy implications ......................................................................... 109

Bibliography ..................................................................................................................... 113 Appendix ........................................................................................................................... 129 Materials and Methods .................................................................................................... 129 Sensitivity analysis.......................................................................................................... 133

List of Figures

Figure 3.1-Topography of Ethiopia ....................................................................................... 16 Figure 3.2-Economic structure of Ethiopia ............................................................................ 20 Figure 3.3-Tax composition in Ethiopia ................................................................................ 21 Figure 3.4-Trends of public expenditure in Ethiopia ............................................................. 22 Figure 3.5-Fiscal balance in Ethiopia .................................................................................... 23 Figure 4.1-Schematic presentation of the production technology nest ................................... 29 Figure 4.2-Schematic presentation of the commodity structure ............................................. 30 Figure 5.1-Households’ welfare effects of climate change .................................................... 68 Figure 6.1-Normal distribution curve of average elasticities.................................................. 79 Figure 6.2-Households’ welfare effects of planned adaptation costs ...................................... 84 Figure 7.1-Households’ welfare effects of public adaptation finance ..................................... 96

List of Tables

Table 3.1-Selected regional socio-economic indicators ......................................................... 18 Table 3.2-Regional shares in selected national indicators ...................................................... 19 Table 4.1-Activity accounts in the SAM ............................................................................... 34 Table 4.2-Factor accounts in the SAM .................................................................................. 36 Table 4.3-Share of market commodities from activities in the SAM ..................................... 37 Table 4.4-Final demand for commodities in the SAM ........................................................... 38 Table 4.5-International trade in the SAM .............................................................................. 38 Table 4.6-Government account in the SAM .......................................................................... 39 Table 4.7-Tax accounts in the SAM ...................................................................................... 40 Table 4.8-Household accounts in the SAM ........................................................................... 41 Table 4.9-The external sector account in the SAM ................................................................ 42 Table 4.10-Savings-Investment account of the SAM............................................................. 42 Table 4.11-Summary of elasticities used in calibration ......................................................... 43 Table 4.12-Regional value-added GDP ................................................................................. 48 Table 4.13-Economic structure of regions ............................................................................. 49 Table 5.1-Climate change effects on grain yields .................................................................. 60 Table 5.2-Macroeconomic effects of climate change............................................................. 63 Table 5.3-Sectoral output effects of climate change .............................................................. 65 Table 5.4-Factor market effects of climate change ................................................................ 67 Table 5.5-Regional effects of climate change ........................................................................ 69 Table 6.1-Summary of elasticities of agricultural productivity .............................................. 77 Table 6.2-Direct costs of planned adaptation in agriculture ................................................... 80 Table 6.3-Macroeconomic effects of planned adaptation costs .............................................. 81 Table 6.4-Sectoral output effects of planned adaptation costs................................................82 Table 6.5-Factor market effects of planned adaptation costs ................................................. 83 Table 6.6-Regional effects of planned adaptation costs ......................................................... 85 Table 7.1-Macroeconomic effects of public adaptation finance ............................................. 92 Table 7.2-Sectoral output effects of public adaptation finance .............................................. 94 Table 7.3-Factor market effects of adaptation finance ........................................................... 95 Table 7.4-Regional effects of public adaptation finance ........................................................ 97 Table 8.1-Economy-wide effects of climate change ............................................................ 106 Table 8.2-Regional effects of climate change with structural change................................... 108

Acronyms ADLI AgMIP AgSS CES CET CGE CRGE CSA EDRI EPIC ERA ETB EV FAO FDI FDRE GCM GDP GGCM GHG HadGEM2-ES HICES IFPRI ILRI IPCC LES LPJmL MoARD MoE MoFED NBE NLFS NPC ODA PHC RCP ROW SAM S-I SNA SNNP UNDESA UNEP UNFCCC WDI

Agriculture Development Led Industrialization strategy of Ethiopia Agricultural Model Inter-comparison and Improvement Project Annual Agricultural Sample Survey of Ethiopia Constant Elasticity of Substitution function Constant Elasticity of Transformation function Computable General Equilibrium Climate Resilient Green Economy strategy of Ethiopia Central Statistics Agency/Authority of Ethiopia Ethiopian Development Research Institute Environmental Policy Integrated Climate model Ethiopian Roads Authority Ethiopian Birr (Ethiopian Currency) Equivalent Variation Food and Agriculture Organization of the United Nations Foreign Direct Investment Federal Democratic Republic of Ethiopia Global Climate (Circulation) Model Gross Domestic Product Global Gridded Crop Model Greenhouse gases Hadley Global Environment Model 2 - Earth System Household Income, Consumption and Expenditure Survey of Ethiopia International Food Policy Research Institute International Livestock Research Institute Intergovernmental Panel on Climate Change Linear Expenditure demand System Lund-Potsdam-Jena managed Land model Ministry of Agriculture and Rural Development of Ethiopia Ministry of Education of Ethiopia Ministry of Finance and Economic Development of Ethiopia National Bank of Ethiopia National Labor Force Survey of Ethiopia National Planning Commission of Ethiopia Official Development Assistance Population and Housing Census of Ethiopia Representative Concentration Pathways Rest of the World Social Accounting Matrix Saving-Investment account/balance System of National Income Accounts Southern Nations, Nationalities, and Peoples regional state of Ethiopia United Nations Department of Economic and Social Affairs United Nations Environment Programme United Nations Convention on Climate Change World Development Indicators

1

Introduction

Economic development refers to the process and policy endeavors that aim to increase the sustenance (the ability to meet basic needs), the self-esteem (the sense of worth and self-respect), and the freedom (the ability to choose among a bundle of alternatives to satisfy ones needs) of a given society (Todaro and Smith, 2012). There are many – global and national – environmental, social, economic, and political challenges impeding on such endeavors in the contemporary developing countries. Particularly, the nexus between environment and economic development in least developed countries (LDCs) is strong. Consequently, many LDCs are susceptible to environmental changes. Climate change is one of such environmental changes that have profound effects on the economic prosperity and development of LDCs (Stern, 2007). Vulnerability of LDCs to climate change accrues to their heavy dependence on nature-based economic activities such as agriculture, the existing environmental conditions (high temperature, erratic rainfall, and degrading environment), and lack of institutional, technical and financial capacities to deal with the change (Cline, 2007; Fankhauser and Burton, 2011). The production process in agriculture, for instance, inherently depends on climate (Padgham, 2009). Therefore, climate change directly and indirectly affects agricultural production through agricultural productivity, area suitable for agricultural production, and agricultural labor supply. These impacts are pronounced in LDCs as their agriculture is rain-fed, traditional, and small-scale with low level and slow rate of adoption of biotechnologies. Because of this, Goal 13 of the Sustainable Development Goals (SDGs) underlines that climate change “presents the single biggest threat to development, and its widespread, unprecedented impacts disproportionately burden the poorest and most vulnerable.” 1 Ethiopia is a case in point since agriculture is the backbone of its economy. Agriculture employs about 75-80% of the country’s labor force (NPC, 2016; NLFS, 2005; 2013; HICES, 2011), and contributes nine out of ten major export items of the country (NBE, 2016). In the last decade, on average, about 75-80% of the total merchandise export earnings is directly from agriculture while about 11% is from agriculture dependent manufacturing goods such as meat products and leather and leather products (NBE, 2016). Agriculture also contributes to about 40% of national GDP (NBE, 2016). Consequently, from the public policy point of view, agriculture is usually regarded as the engine of economic growth leading to poverty reduction (MoFED, 2006; 2010; NPC, 2016). Accordingly, the country’s public expenditure system prioritizes agriculture and rural development over the other sectors. About 15% of the total public budget (roughly 5-6% of GDP) is allocated to agriculture and natural resources making it one of the highest in Africa (MoARD, 2010; FAO, 2015). Despite the public policy measures undertaken in the last four decades, however, agriculture in Ethiopia is yet traditional and subsistence. Agriculture is virtually rain-fed (CSA-FAO, 2014; AgSS, 2014). About 95% of annual agricultural output is produced by smallholder peasants (MoARD, 2010) while about 60% of grain output is used for own household consumption (AgSS, 2014). Land degradation, soil erosion, and deforestation in the country are formidable environmental challenges that need serious action. Illiteracy is high among smallholder peasSustainable Development Knowledge Platform (https://sustainabledevelopment.un.org/sdg13). Accessed on 18 December 2016.

1

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 A. W. Yalew, Economic Development under Climate Change, https://doi.org/10.1007/978-3-658-29413-7_1

2

1. Introduction

ants affecting the adoption rate of modern agricultural practices. Credit and market infrastructure are still formidable constraints (BMGF, 2010). Simply put, the adaptive capacity of the sector is low. The adaptive capacity of the general economy is also low since Ethiopia is a typical low-income country with per capita income of USD 690 (NBE, 2016) where fiscal deficit is still a defining characteristic of the public sector (MoFED, 2014). Taken all together, the Ethiopian agriculture and economy are predisposed to climate change. Therefore, climate change represents a potential threat to the economic prospect of Ethiopia. First, it will influence crop and livestock productivity, area suitable for crop and livestock production, agricultural employment and income. These primary effects in agriculture may ripple into the rest of the economy, strain economic growth and, eventually, are detrimental to the country’s ability to reduce poverty and inequality. Second, as its economic effects get worse, climate change calls for deliberate public action to adapt. This needs to consider the type, extent, timing and way of undertaking measures to underpin the resilience of the agricultural production system to climate change. Planning and mainstreaming adaptation of a single sector (i.e. agriculture) to a single stressor (i.e. climate change) into the overall macroeconomic framework imposes an additional task for policy-and decision-makers. Third, planned public adaptation in agriculture entails incremental budgetary burden to the public sector. Since the highly anticipated international climate finance for adaptation in LDCs is inadequate and unpredictable (Adaptation Watch, 2015), it may be necessary to look for new domestic resources and earmark them for adaptation in agriculture and rural settlements. These, among others, may need to simply scale up the public spending on agricultural and rural development (and prepare to shoulder the resulting fiscal deficit), to divert public resources from other activities, or to raise new finance through different taxes. Usually decisions with regard to fiscal policy in LDCs have non-negligible allocation and distributional effects. Fourth, climate change, adaptation costs, and adaptation finance may have distributional effects among different regions of the country. Thus, in addition to the economy-wide (aggregate) effects, regional effects may matter in adaptation policy- and decision- making process. Fifth, the allocation and distributional effects of public adaptation to climate change may strain the current and future macroeconomic plans of the country. This poses a challenge to policy-and-decision makers. Among others, it requires to figure out to what extent to compromise the general (medium-and long-term) plans for the sake of adaptation in agriculture (which is also justifiable on many grounds). In light of the forgoing discussion, climate change represents new economic development challenge to Ethiopia which thus needs to go beyond the conventional and business-as-usual development planning approach. One, among others, is to start investigating what economic development under climate change would look like. As such, an ex-ante analysis of the economic consequences of climate change and public policy responses to climate change is important. However, there exist only few studies that assess the economic impacts of climate change but barely delve in to planned adaptation costs and finance. In addition, to the best of my knowledge, there is no single study that couples economy-wide analysis with regional projections and analysis. The latter is important since the economic structure of some regions is significantly different from the national average, and the regional governments usually depend on federal government block-grants. This study is motivated by the importance of the topic well as the prevailing gaps in the literature with respect to Ethiopia in particular and to LDCs in general. The general objective of the study is to assess the economy-wide and regional effects of climate change and government

1. Introduction

3

responses to climate change in Ethiopia. It specifically intends to address four research objectives. First, it aims to assess the economy-wide and regional effects of climate change. Second, it intends to examine the economy-wide and regional effects of planned public ‘full’ adaptation in agriculture. Third, which follows from the second objective, it aims to examine the economywide and regional effects of alternative adaptation finance to agriculture. Fourth, the study intends to highlight the contribution of structural change to climate-resilient economic development. The main hypothesis of the study is that climate change as well as planned public adaptation to climate change have substantial general equilibrium effects. Climate change negatively affects Ethiopian agriculture and economy. The influences will be severe in regions where agriculture is the dominant economic activity. Planned public ‘full’ adaptation in agriculture will help to offset the aggregate economic impacts of climate change. However, it bears some residual and indirect effects to the rest of the economy and deteriorates government saving (budget surplus). The urbanized regions of the country will bear the bulk of such indirect effects due to government sponsored adaptation in agriculture. Alternative adaptation finance schemes for agriculture may dampen the effects of public adaptation costs on government saving. Nonetheless, they will have distributional effects and may worsen the indirect effects of planned adaptation on urbanized regions. Last, but not least, modifications in some structural rigidities of the economy may dampen the adverse economic consequences of climate change, and contribute to the general resilience of the economy. The study applies the static IFPRI-CGE model calibrated to the 2005/06 Social Accounting Matrix (SAM) of Ethiopia. Economy-wide and regional analysis is done for all research questions. The CGE model is used to simulate the economy-wide effects of different experiments, i.e., research questions. Particularly, the effects on the macro economy, sectoral output, factor markets, and households’ welfare are analyzed. The economy-wide effects on sectoral output are further mapped into a regional module depicting the economic structure of the eleven regional states (level one administrative units) of the country. This helps to project the effects of each CGE experiments on value-added GDP of different regions. The results show that the economy-wide effects of climate change are profound. The effects of climate change induced productivity and labor supply shocks in agriculture on GDP may reach down to -8%. The household welfare effects measured by equivalent variation (EV) range from -3% to -11%. Climate change hits hard agricultural activities. In particular, it reduces agricultural output and increases agricultural prices. This in turn alters the international trade (import and export) mix of the country. The regional effects of climate change range from -10% in agrarian regions (e.g. Amhara) to 2.5% in urban regions (e.g. Addis Ababa). Planned public adaptation that aims to fully neutralize climate change induced agricultural productivity shocks may require to uplift public budget to agricultural productivity enhancing measures by 25 % to 100%. This results in declining government saving by 32% to 173% which shifts the savings adjustment (to maintain the macroeconomic saving-investment balance) burden to households, eventually, implying decline in households’ welfare (-0.6% to -2.7%). However, as public services employ skilled labor, there are gains to urban households. Planned public adaptation also pulls factors of production, following which nonagricultural sectors’ output

4

1. Introduction

decline. These include manufacturing (-2% to -10%), ‘other’ services (-3% to -13%), and hotels and restaurants (-1% to -6%). The regional projections show that urbanized regions will bear the bulk of the trade-offs of government sponsored adaptation in agriculture. For instance, the regional effects may reach to -3% in Addis Ababa, and to -2% in Dire Dawa. Therefore, full public adaptation may help to avert the aggregate effects but bears residual and indirect effects. In terms of the effects on GDP, the marginal effects of alternative adaptation finance schemes (compared to default option through public deficit scheme) are insignificant. In addition, the aggregate effects vary little across financing schemes. However, the distributional effects as reflected on macroeconomic components, industrial activities, household groups, and regions are considerable. Availability of foreign finance seems in terms of real households’ consumption. Nevertheless, transfers from abroad may appreciate the real exchange rate. Consequently, comparatively, exports become be worse off. Relative to the default option, urban households are slightly worse off under taxing schemes. Raising direct (income) tax rates implies the worst welfare effect for urban households compared to other schemes. Diverting schemes (from general public administration, and public social services) would imply lesser distributional effects between households. Yet, diverting schemes widen the range of regional effects. For instance, diverting from general public administration implies -0.5% for Ethiopia-wide value-added GDP, but, -0.1% for Tigray region, and -1.7% for the city of Addis Ababa. The discussion on adaptation costs and adaptation finance show that planned public adaptation in agriculture imply pressure on government saving and impede on manufacturing and private services, and urban GDP. This especially contradicts with the current macroeconomic plans of the country. Ethiopia is striving to structural change led by government, to reduce fiscal deficit relative to GDP, and to increase domestic savings (see, for example, MoFED, 2010; NPC, 2016). Per contra, structural change itself is regarded as generic adaptive capacity to climate change (Mendelsohn, 2012; Millner and Dietz, 2015). This study tested this view by constructing cost-free exogenous structural change scenarios in labor markets and transaction costs of market commodities. The results suggest that some structural change scenarios have the potential to dampen the adverse consequences of climate change on aggregate GDP and households’ welfare by about 30%. Therefore, structural change contributes to a climate-resilient economic development in Ethiopia. All in all, climate change imposes formidable risk to the economic prospect of Ethiopia. Public action to neutralize the biophysical impacts of climate change, i.e., before they turn to be economic problems, bear trade-offs. Raising new public funds for adaptation implies opportunity costs. These tradeoffs and opportunity costs may impair the current and planned development endeavors of the country. On the other hand, urban regions are less affected by climate change, and arbitrarily constructed structural change scenarios dampen the adverse effects of climate change. This lend us to suggest that the past and current public investments in institutions, transport, energy, human capital, and urbanization have double-dividends as they also contribute to the resilience of the economy to climate change. Yet, given the role of the sector in the current economy, and the potency of the anticipated biophysical impacts, proactive adaptation in agriculture is imperative. The government may reconcile this by attracting more private sector investment into agriculture, and by devising and arranging institutions that enable the majority of agricultural output to enter into markets. In parallel, the government shall promote climate-smart agriculture that may even help to earn mitigation finance for cutting GHG emission from livestock production, land use and land use changes. Such mitigation funds can be

1. Introduction

5

earmarked to undertake agricultural adaptation measures. Foreign aid in the form of biotechnology transfer and debt-relief may help to control the side effects of grants on foreign exchange market and trade balance. The study contributes to the scientific discourse on climate change and adaptation with respect to Ethiopia in particular and to LDCs in general. First, it enlarges the primary effects (entry points) of climate change. It incorporates climate change induced productivity shocks in grain and livestock production, and agricultural labor outmigration. Second, it attempts to gauge uncertainty of climate change impacts through biophysical impact models unlike the dominant approaches which are through emission scenarios or climate models. Third, it applies a method to derive the direct costs of adaptation in agriculture that can be regarded as an alternative to the commonly used “experts’ opinion” approach. The study also attempts to identify the sectors, households, and regions that would gain or lose from planned public adaptation in agriculture and alternative adaptation finance schemes. Fourth, it contributes to the empirical research on the role of structural change in climate-resilient development. Fifth, the study combines economy-wide and regional analysis for each policy experiments which will give clearer picture for adaptation policy-and-decision making process. To the best of my knowledge, in spite of focusing on a single sector, to cover such a range of climate change aspects in a single study is a new attempt in the empirical research for Ethiopia as well as for LDCs. Such approach portrays the problem of climate change clearer, and sheds better light for the adaptation policy-and decision-making process. The rest of the monograph is organized as follows. Chapter 2 presents the theoretical and conceptual background of the study. Chapter 3 presents the Ethiopian economy in a nutshell. Chapter 4 outlines the methodological framework, and discuss the calibration of the CGE model and the construction of the regional module. Chapter 5 assesses the economy-wide and regional effects of climate change. Chapter 6 examines the economy-wide and regional effects of planned public adaptation in agriculture. Chapter 7 builds on Chapter 6 and goes further to assess the economy-wide and regional effects of alternative adaptation finance schemes. Chapter 8 highlights the role that structural change would play to dampen the adverse effects of climate change presented in Chapter 5. Chapter 9 presents the overall conclusions of the study along with their policy implications.

2 2.1

Theoretical and Conceptual Background Introduction

Economic development refers to the process and policy endeavors that aim to increase the sustenance (the ability to meet basic needs), the self-esteem (the sense of worth and self-respect), and the freedom (the ability to choose among a bundle of alternatives to satisfy ones needs) of a given society (Todaro and Smith, 2012). A multitude of environmental, social, economic, and political challenges impinge upon the prospects of the economic development in the contemporary LDCs. The relationship between the environment and the economy in LDCs is complex and strong. Poor countries usually depend on the environment for energy (e.g. fuel-wood, hydropower), for food (e.g. agriculture, fishing), and for exports (e.g. extractives, timber). On the other hand, many LDCs lack proper technical, institutional, and financial capacities to deal with exogenous changes. Particularly, agriculture in LDCs is overwhelmingly small-scale, rain-fed, and with a low level and slow biotechnology adoption rate. As a result, many LDCs are susceptible to environmental shocks. Climate change is one of such environmental changes that have profound effects on the economic prosperity and development of LDCs (Stern, 2007). This chapter presents the theoretical and conceptual foundations of the study. It attempts to discuss how climate change may influence economic development, and hence shall be considered as economic problem. Particularly, it discusses how climate change may pose problems to the production and employment in agriculture, to the public sector, and to the general economic equilibrium of LDCs. By implication, the study is built on the economic theories of production, migration, public policy, and general equilibrium. The remainder of the chapter discusses the concept of climate and climate change (2.2), the interlinkage between climate change, agriculture, and development (2.3), the interlinkage between climate change, public policy, and development (2.4), and the general equilibrium theory and computable general equilibrium models (2.5). 2.2

Climate and climate change

Climate in a narrow sense is referred as an average weather in a specific geographic area over a long period of time, usually 30 years (IPCC, 2014). It also refers to the statistical description in terms of the mean and variability of temperature, rainfall, wind, and other relevant quantities over a period of time (IPCC, 2014). As such, climate change refers to “a change in the state of the climate that can be identified (e.g., by using statistical tests) by changes in the mean and/or the variability of its properties and that persists for an extended period, typically decades or longer” (IPCC, 2014, p.120). Climate change is different from climate variability which refers to variations in the mean state, standard deviations, occurrence of extremes, and the likes of the climate on all temporal and spatial scales beyond that of individual weather events (IPCC, 2014). Scientific evidence shows that the earth climate is changing due to increasing greenhouse gas (GHG) emissions and concentration since the Industrial revolution (IPCC, 2014). The atmospheric CO2 concentration has transgressed 400 particles per million (ppm) up from 315 ppm in

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 A. W. Yalew, Economic Development under Climate Change, https://doi.org/10.1007/978-3-658-29413-7_2

8

2. Theoretical and Conceptual Background

1960s 2. Burning fossil fuels and land use changes are the main sources of GHG emissions (IPCC, 2014). The scientific evidence on global warming is overwhelming (Tanner and HornPhathanothai, 2014). The evidence on climate change impacts on the natural system is comprehensive, and many of the observed changes since the 1950s are unprecedented (IPCC, 2014). The global atmosphere and ocean have warmed, the volume of snow and ice have diminished, and sea level has risen (IPCC, 2014). The current mean global surface temperature is 0.89oC above the pre-industrial levels (Tanner and Horn-Phathanothai, 2014). Unless the GHG emission is abated, relative to 1850-1900 average, the global mean surface temperature in the end of 21st century is likely to exceed 1.5°C (may even be greater than 5oC in high-emission scenarios) (IPCC, 2014). These observed and projected changes in mean and variability of temperature, precipitation, wind and solar radiation, directly or indirectly, affect all basic elements of human life (Stern, 2007). As such, climate change influences all dimensions of sustainable development (i.e., environment, society, and economy), and hence the potential development pathways of human systems in any country (Stern, 2007; IPCC, 2014). In particular, the effects on tropical poor countries are immediate and strong (Stern, 2007; IPCC, 2014). On the other flip of the coin, therefore, climate change calls for deliberate government actions to mitigate it (especially in developed countries) and to adapt to it (especially in developing countries). However, government actions usually have allocation and distribution effects which in turn are linked to (and can influence) economic development especially in LDCs. In this study, the entry point is agriculture. On the one hand, the sector inherently depends on climate and is sensitive to climate change. On the other hand, agriculture plays a pivotal role in the Ethiopian economy in terms of food supply, employment, income, and export earnings. Therefore, in the subsequent sections, I will focus on climate change and agriculture. I will eschew from discussing on the impacts of climate change on other economic activities and their repercussion effects. 2.3

Climate change, agriculture, and development

Production in agriculture uses natural factors (e.g. temperature, rainfall, wind, CO2, soil and water) in combination with socio-economic factors (e.g. labor, fertilizer, pesticides, improved seeds, farm implements, irrigated water, harvested water, and so on) (Adams et al., 1998). As such, agricultural production functions show that climate and hence climate change are key determinants of agricultural production (Adams et al., 1998; Antle and Capalbo, 2010). Climate change affects agricultural production in many ways. It affects area suitable for crop cultivation, crop yield, physiological performance and reproduction of animals, feed and water availability for livestock, the type and number of livestock to be raised per farm; increases the incidence of pests, plant and animal diseases, and the frequency, duration, and intensity of extreme events such as droughts and floods; and exacerbates existing environmental problems (e.g. land degradation and soil erosion) that reinforce the aforementioned problems. The forgoing primary effects of climate change are already being observed especially in low-income countries lying in the tropics (Mendelsohn et al., 2006; Cline, 2007; IPCC, 2014).

NASA Global Climate Change (http://climate.nasa.gov/vital-signs/carbon-dioxide/). Accessed on 12 October, 2016.

2

2.3 Climate change, agriculture, and development

9

The change in agricultural production affects returns to agricultural factors and agricultural prices. Therefore, climate change is also expected to affect agricultural income, and real consumption in LDCs. In addition, with falling agricultural production, agricultural exports decline (Jones and Olken, 2010) and thus alter trade balance and impair economic growth in LDCs (Dell et al., 2012; Alagidede et al., 2016). By coincidence, the countries at higher risk of climate change and its biophysical impacts such as those in the sub-Sahara Africa (e.g., IPCC, 2014) are also at the bottom ring of the development ladder (e.g., WDI, 2016). These countries are at great thirst to increase agricultural productivity, to grow economically, to reduce poverty, to reduce unemployment, to reduce malnutrition, and to increase life expectancy. Therefore, climate change may perpetuate poverty, inequality, and aid dependence of LDCs. On the other hand, rural livelihood and life style are inextricably linked to agriculture while per capita agricultural land is declining due to high population growth rates in LDCs (cf. WDI, 2016). As such, climate change risks on agricultural productivity and production may trigger outmigration from agricultural sector/rural areas. With falling rural/agricultural income, the expected wage difference between agricultural/rural and non-agricultural/urban sectors widens which in turn triggers agricultural labor migration. In addition, agricultural households’ may regard migration as a risk management or coping strategy in order to diversify the family sources of income. The premise is supported by economic theories of migration 3 which more or less agree on the fact that expected income differentials, spatial opportunity disparities, underdevelopment, poverty, hedging against risk, relative deprivation, and sudden disruptive changes in terms of livelihood are main motives of migration (e.g., Lee, 1966; Harris and Todaro, 1970; Stark and Taylor, 1989; Lucas, 1997; Abreu, 2010; de Haas, 2010). Generally, the decision to migrate can be taken by either larger units of people (e.g. by families) or individual labor with an aim to maximize income, to accumulate wealth, or to spread risk across sectors/locations (Abreu, 2010; de Haas, 2010). Nevertheless, the contribution of rural-to-urban or agriculture-to-nonagricultural sectors migration to the overall economic development in LDCs is yet open to debate. Some regard that rural-to-urban migration contributes to optimal allocation of factors of production (e.g. Lewis, 1954), and contributes to remittances that can boost investment and foster development in the emigrants origin (e.g. Keely and Tran, 1989 cited in de Haas, 2010). Accordingly, rural-to-urban migration positively contributes to the overall economic development. In contrast, for some others (cf., Taylor, 2001; de Haas, 2010) rural-tourban migration implies loss of agricultural labor and agricultural productivity, and thus undermines rural economy. As such, rural-to-urban migration only reinforces itself leading to migration syndrome and end with aggravating the problems of underdevelopment in rural areas and rural-urban inequality (de Haas, 2010). In conclusion, climate change impinges upon the economic development process in LDCs. It affects their main economic activity (i.e., agriculture). It also complicates and exacerbates environmental degradation, rural-urban migration, and the imbalance between population and food supply growth. All of these undermine the role of the current and planned development endeavors of the LDCs plus the international support to tackle inequality, poverty, and food insecurity in the LDCs.

3

For review of economic theories of migration see Abreu (2010) and de Haas (2010).

10

2.4

2. Theoretical and Conceptual Background

Climate change, public policy, and development

Governments play allocation, stabilization, and redistribution roles in an economy using public policies (Benassy-Quere et al., 2010). Allocation functions aim at affecting the quantity, the quality, and the sectoral and regional allocation of factors of production; the stabilization functions are more or less similar to the allocation functions with focus on only short-run problems; and the redistribution functions emanate primarily from equity concerns (Benassy-Quere et al., 2010). Climate change appears as a new problem that public policies shall address, at least, in three main ways. First, climate change is a typical externality problem (Stern, 2007). It needs collective action to mitigate as well as to adapt to it. Mitigation requires implementing policies and incentives that promote cleaner production, distribution, and consumption of goods and services to curb GHG emission. Such public policies are expected from the developed countries. Second, especially in LDCs, climate change adaptation calls for government action. The role of government in anticipating, planning, and preparing to climate change adaptation is indispensable (Mendelsohn, 2000; Antle and Capalbo, 2010). In other words, climate change implies a set of new tasks for the public sector. It demands either new climate change /adaptation oriented public services or additional tasks on existing climate change/adaptation relevant public services. Third, climate change and public response to it have economic consequences that may perturb public sector balance. Climate change influences tax bases (i.e., production, exports, imports, distribution, and consumption). Therefore, for a given level of public spending, climate change has a potential to perturb the fiscal balance. Public policies that aim to curb GHG emission need new sets of regulations and incentives which have implications for public revenue and expenditure. Planned public adaptation to climate change bears additional budgetary burden which also impede on government saving for a given level of government revenue. In addition, public actions for mitigation and adaptation may have allocation and distributional effects to the rest of the economy. Simply put, there is a two-way relationship between climate change and the public sector. The direction, nature and strength of the relationship depends on the level of economic development. Public policy is necessary to curb GHG emission in developed countries, and to anticipate, and provoke adaptation in developing countries. Public revenue may be positively or negatively affected due to mitigation policies in the developed world. Climate change impacts may reduce public revenue while adaptation to climate change may increase public spending in LDCs. In order to influence the economy, governments often use a range of policy instruments such as legislations and regulations (e.g. competition policy, employment and price laws), direct provision of goods and services (e.g. public goods), and fiscal (e.g. taxes, expenditure) and/or monetary (e.g. interest rates, money supply) policies (Benassy-Quere et al., 2010). Likewise, governments can place such instruments for the sake of climate change adaptation. However, economic theory tells us that no single policy instrument can achieve two or more objectives simultaneously (Tinbergen, 1952). There are always trade-offs (Tinbergen, 1952; BenassyQuere et al., 2010) which can be inter-temporal, spatial, and sectoral. For instance, planned adaptation in agriculture can be regarded as direct provision of public goods and services aimed at agricultural and rural development. However, this may require diverting public resources from other public sectors such as health and education. It may require to generate additional

2.5 Computable general equilibrium models

11

revenue through taxation that may have allocation and distribution effects. The forgone opportunities of the public spending and the new resources for adaptation, which may be inter-temporal, mounts the concern about the size of the trade-offs. Taken altogether, climate change is a new issue that needs public policies, incentives, measures, and spending to address it. By implication, climate change influences economic development through its effects on the public sector. This is particularly important in LDCs where the share of public investment (and services) in total investment (and GDP) is significant. As such, as any form of government intervention in an economy, the government response to climate change shall be evaluated. Such evaluation, using precise criteria, ought to examine and compare the outcomes of alternative policies (Benassy-Quere et al., 2010). In general, an economic policy can be evaluated using either one or a combination of CGE models, macroeconomic models, statistical models, and microsimulation models (Benassy-Quere et al., 2010). This study pursue a CGE model for the reasons discussed in the next section. 2.5

Computable general equilibrium models

Computable general equilibrium (CGE) models are based on general equilibrium theory that regards an economy as a system of interdependent components. The components may be agents (households, firms), markets (commodity, factor, foreign, or domestic), and commodity types (exports, imports, domestic sales). According to the general equilibrium theory, the concept of equilibrium should consider the interactive simultaneous determination of equilibrium prices across markets (Starr, 2011). The general equilibrium solution, thus, refers to the whole economy rather than to a single market of a commodity or a factor. Accordingly, general equilibrium of an economy is attained if and only if an array of equilibrium prices result in zero excess demand in all commodity markets. 4 Simply put, general equilibrium theory defines equilibrium as “simultaneous price-guided clearing of several good markets” (Starr, 2011, p.1). The theory is based on the fundamental observation that markets in the real world economies are linked and mutually interdependent (Piermartini and Teh, 2005), therefore, everything in the economy depends on everything else in the economy (Burfisher, 2011). As such, relative prices matter as they dictate resource allocation and income distribution within (and across) markets in an economy. Therefore general equilibrium analysis provides a logically consistent way to look at policy issues which usually involves more than one agent (Bandara, 1991). Initially, the focus of general equilibrium theory was on proving the existence and uniqueness of a general equilibrium (Shoven and Whalley, 1984; Bandara, 1991). However, we need operational tools to be able to analyze the real world economic problems. In other words, we need to have empirical economic models consistent with general equilibrium theory. CGE models came with the quest for such numerical and empirically based general equilibrium models (Shoven and Whalley, 1992). CGE models numerically describe “how the economy behaves and reacts to different external shocks while being consistent with standard economic theory” (Andre et al., 2010, p.14). Therefore, any CGE model is a “system of equations that describe an economy as a whole and the interactions among its part” (Burfisher, 2011, p.3). It, therefore, depicts all transactions in the economy “in such a way that it is possible to connect each element The theoretical foundation of general equilibrium analysis is pioneered in the second half of 19th century and later formalized in the 1950s and 1960s. See Bandara (1991) and Starr (2011) for more.

4

12

2. Theoretical and Conceptual Background

of the model with some observed empirical data’’ (Andre et al., 2010, p. 14). The functional forms chosen, of course, must satisfy certain restrictions of general equilibrium theory (Robinson et al., 1999). CGE models presume that an outcome of any change in the economy is determined by all direct, indirect and induced (feedback) effects (Cardenete et al., 2012). Endogenous prices and quantities are allowed to “transmit market information through different sectors of the economy, thereby simulating the workings if a perhaps regulated and intervened, yet absolutely decentralized markets” (Yeldan, 1986, p.6). The effects of a policy/exogenous change on the macro economy, different sectors, and agents are gleaned by comparing the initial (reference, baseline, or benchmark) equilibrium with the counterfactual (or simulated) equilibrium (Shoven and Whalley, 1984; Cardenete et al., 2012). CGE models have many novel features. Their main advantage accrue to their capacity to explain the economy-wide effects of incorporating changes in a particular policy parameter or in a sector’s characteristic in relation to the economy as a whole (Cardenete et al., 2012). CGE models offer “substantial detail for policy makers concerned with feedback effects of policy initiatives directed at specific products or industries” (Shoven and Whalley, 1992, p.2). CGE models also help to capture the combined effects of simultaneously changed policies and shocks, and to learn about their combined effects (Bandara, 1991). Moreover, CGE models provide information from the aggregate macroeconomic variables to households’ welfare level (Bandara, 1991; Piermartini and Teh, 2005; Cardenete et al., 2012). This makes CGE models as “the most suitable tool so far devised for this purpose providing sector specific information on prices and output, with added informational bonuses on welfare, income distribution and macro magnitudes’’ (Cardenete et al., 2012, p.49). CGE models are able to capture market-optimization behavior of individual agents as well as commanding nature of the exogenously determined government policies (Yeldan, 1986; Ginsburgh and Keyzer, 1997). To put alternatively, CGE models are important “to capture the institutional arrangements characterizing a particular country” (Robinson et al., 1999, p.6). The aforementioned comparative advantages have made CGE models very popular in empirical economic analysis (Cardenete et al., 2012) with application to a wide-range of policy questions in a number of fields of economics that include public finance, trade, environment, energy, and regional and national development planning. However, CGE models are not without critiques. CGE models are generally criticized for being pre-determined models such that model results depend on the pre-selected assumptions (e.g. values of elasticities) and specifications (e.g. macro-closures) (Mitra-Kahn, 2008; Burfisher, 2011; Cardenete et al., 2012). In addition, the specification, estimation, and values of elasticities and parameters of the model are usually “based on scant empirical evidence, and what evidence exists is often contradictory” (Shoven and Whalley, 1992, p.5). Despite this, CGE model are usually calibrated so that the solutions to every equations reproduce the values in the initial (reference) equilibrium (Andre et al., 2010; Burfisher, 2011; Cardenete et al., 2012). Since the model parameters are calibrated to usually one year observed data, there are no ways to statistically test the model specifications (Shoven and Whalley, 1992). Calibration method presumes that the CGE database, i.e., the social accounting matrix (SAM), represents the initial equilibrium of the economy. 5 Another point of debate is on closure rules which involve associating a 5 SAM is a database that describe and report the value of all economic transactions in a given economy in a specific period of time (Burfisher, 2011). In other words, SAM is built around all economic agents, all markets, all production activities, and all commodities. Accordingly, SAM is “a logical framework that provides a visual display of the transactions as a circular flow of national income and spending” (Burfisher, 2011, p.44). SAM delineates economic flows across product and factor markets, and provides the statistical underpinnings for a CGE model

2.5 Computable general equilibrium models

13

particular endogenous variable to system constraints (Ginsburgh and Keyzer, 1997; Burfisher, 2011). Since closure rules define the direction of causality and influence the simulation results, alternative closures may even imply different qualitative results (Taylor, 1990). Therefore, the choice between alternative closures for identity equations has been a debating issue in the subject (Mitra-Kahn, 2008; Burfisher, 2011; Cardenete et al., 2012). The appropriateness of the choice of a specific closure rule depends on the economic situations of the economy under investigation, context of the research question and analysis (Lofgren et al., 2002; Hosoe et al., 2010; Burfisher, 2011). Of course, CGE modelers often confront difficulty on how to represent (or treat) the policies themselves in the CGE model (Shoven and Whalley, 1992). All said, the aforementioned limitations cannot undermine the relevance of CGE models. First, CGE models are not exceptions in making assumptions as it is inherent in any economic models (Shoven and Whalley, 1992). Statistical estimation of elasticities and parameters in lieu of calibration is generally possible but often impractical. 6 Statistical estimation would require huge data (e.g., a time-series of input-output matrices) and the time required to estimate all parameters is unmanageable (Mansur and Whalley, 1981; Kehoe and Kehoe, 1994; Cardenete et al., 2012). Moreover, one would encounter three technical problems while using econometric estimation method for CGE model parameters. First, the parameter estimates will only be based on behavioral functions without considering the equilibrium constraints while considering the latter implies that the error terms of different functions are no more independent to each other (Mansur and Whalley, 1981; Ginsburgh and Keyzer, 1997). Second, as the model size increases, the number of parameters to be estimated increases and will eventually exceed the number of available data points (Mansur and Whalley, 1981). Third, having many endogenous variables to be solved simultaneously will imply identification problem (Ginsburgh and Keyzer, 1997). Second, evaluating the performance of CGE models (i.e., testing the reliability of CGE model results) requires comparing the model predictions with actual outcomes after the policy/exogenous changes happened (Kehoe and Kehoe, 1994). For fair evaluation, this requires to go back to the model and account all other external shocks that affected the economy in the meantime (Kehoe and Kehoe, 1994). Generally, this exercise is not easy as many things change in between. On the other hand, to validate CGE results by comparing with actual outcomes requires the model to capture and replicate the evolution of the economy. However, such efforts are ridiculous until we are able to model technical change properly which does not exist so far (O’Rourke, 1995). Yet, as much as the changes in between are captured, CGE models “can accurately predict the changes in relative prices and resource allocation that result from a major policy change” (Kehoe and Kehoe, 1994, p.14). For instance, a study that conducted such exercise to test the reliability of the results of the CGE model used to predict the effects of joining the European Community for Spain in 1985 found a weighted correlation of 0.94 between the model results and the actual changes (Kehoe and Kehoe, 1994, p.14). Alternatively, Piermartini (EDRI, 2009; Burfisher, 2011). Each entry (cell) in a SAM represents payments from the column account to the row account. For every account, the corresponding column (expenses) sums and row (receipts) sums must be equal. All accounts in the SAM are flows which means SAM describes any macroeconomic relationship (government balance, trade balance, and saving-investment balance) in flow terms. A standard SAM has no asset (financial) accounts. As a result, standard CGE models are concerned with resource allocation (Burfisher, 2011). 6 See Mansur and Whalley (1981, pp.3-25) more on calibration (versus the statistical estimation for parameters in CGE modeling) and on how statistical/econometric approaches are not good candidates for estimating parameters for CGE models.

14

2. Theoretical and Conceptual Background

and Teh (2005) suggests that validating CGE model results using econometric models (e.g., gravity models for international trade) although this is always not viable exercise. Third, macroeconomic closures came due to the presence of macroeconomic variables (government, rest of the world, savings, and investment) and constraints defining them (government/public sector, external sector, and saving-investment balances). We need the macroeconomic variables as we want economic models to produce policy relevant information (Ginsburgh and Keyzer, 1997), and the macroeconomic closures to close (make solvable) the model (Dewatripont and Michel, 1987; Taylor, 1990). Therefore, as remedial measure, CGE modelers perform sensitivity analysis to resolve the aforementioned drawbacks of CGE modeling. Sensitivity analysis involves looking at how robust are the results of the model by changing the parameters or specification of the model (O'Rourke, 1995). 7 It can tell us something about the robustness of CGE results with respect to different sets of parameters (Harrison et al., 1993), and hence the influence of the calibrated parameters on model results (Burfisher, 2011). In balance, CGE models provide an ideal framework for appraising the effects of policy and other exogenous changes on resource allocation and for assessing who gains and losses which usually are not covered well by other economic models (Shoven and Whalley, 1992). 8 This is the main reason as to why I pursue a CGE model approach in this study. In addition, CGE approach can be used to address all of the research questions of the study which are production (e.g., climate change induced productivity effects) as well as policy (e.g., public sector response to climate change) problems. Furthermore, some of the research objectives are represented by multiple (and simultaneous) shocks that can be treated best by CGE models.

Sensitivity analysis can be either systematic or ad hoc approach (see Hermeling and Mennel, 2008 for more). See Shoven and Whalley (1992, pp. 2-5) for more on the new insights from the CGE model results on different policy questions.

7 8

3 3.1

Overview of the Ethiopian Economy Introduction

Ethiopia is a country of rich ethnic, language, culture, topography, and agroecology diversities. The country has entered into its present political orientation since 1995 after which the country is called as Federal Democratic Republic of Ethiopia (FDRE hereafter). The Ethiopian federation comprises nine federal states, one municipal city administration, and one city council (FDRE, 1995). The aim of this chapter is to briefly present about Ethiopia and its economy. The chapter focuses especially on the economic aspects of the country that are relevant for the empirical chapters. It can be regarded as a general prelude for the subsequent chapters. The remainder of the chapter presents the geography (3.2), the population and labor force (3.3), the rural-urban divide (3.4), the federal regions (3.5), the macro economy (3.6), the public finance (3.7), and the economic prospects (3.8) of Ethiopia. 3.2

Geography

Ethiopia is located in East Africa between the latitudes of 3° and 15°N, and the longitudes of 33° and 49°E. It has 1.14 million square kilometers surface area size with highly diversified topography that range from the Danakil depression (125 meters below sea level) to the Ras Dashen mountain (4533 meters above sea level), and that consists of high plateaus and mountain ranges dissected by numerous rivers. The country’s topographic variation highly influences the climatic conditions. The mean annual temperature ranges from 3.9°C (southern highlands) to 31.2°C (eastern lowlands) (NMA, 2007; Evangelista et al., 2013). These topographic and the climatic conditions shape the agro-ecological suitability for crop cultivation and animal husbandry (Hurni, 1998). As agriculture is the main stay, the agro-ecological variation indirectly controls the pattern and density of population settlement. Traditionally, Ethiopia is classified as highlands (central territories) and lowlands (peripheral territories). The Ethiopian highlands, areas above 1500 m above sea level, have mild temperature and predictable rainfall. Sedentary agriculture is the chief source of livelihood in the highlands. In addition, the prevalence of malaria and other vector borne diseases is relatively low in the highlands. Consequently, Ethiopian highlands are densely populated with long history of settlement. The highlands constitute about 45% landmass of the country but nearly 90% of the country’s population (Robinson et al., 2013). As such, the highlands are the economic centers of the country where the bulk of the nation’s output is produced, processed, and consumed. In contrast, Ethiopian lowlands, areas below 1500 m below sea level, have dry climatic conditions – hotter temperature and unreliable rainfall. Less than 10% of the country’s population lives in lowlands though the vast majority of the country’s landmass (55%) is lowland (Robinson et al., 2013). Pastoralism (or agro-pastoralism) is the main livelihood in the lowlands. Therefore, Ethiopian lowlands are sparsely populated with small contribution to the national GDP. The contrast between the highlands and lowlands is illustrated in Figure 3.1. An interested reader can compare and contrast the map with other maps depicting economic contours of the

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 A. W. Yalew, Economic Development under Climate Change, https://doi.org/10.1007/978-3-658-29413-7_3

16

3. Overview of the Ethiopian Economy

country based on gross cellular GDP (e.g., one by Yale G-Econ Project)9 or gridded population maps (e.g., one by Socioeconomic Data and Applications Center-SEDAC) 10. Figure 3.1-Topography of Ethiopia

Source: Author’s illustration 11

Besides, the Ethiopian topography has constrained the development of transport network in the country that in turn contributes to the high transport and trade margins, rigid commodity markets, and low agricultural productivity. Therefore, the Ethiopian topography is one of the bottlenecks for the economic development in the country (Wondemu and Weiss, 2012). In closing, the spatial pattern of the Ethiopian economy is highly influenced by nature, i.e., topography and agro-climatic conditions. 3.3

Population and labor force

Ethiopia has a projected population of 93 million in the year 2016 (CSA, 2013). Its population growth rate (2.7%) is one of the highest in the world (CSA, 2013). The country’s population is projected to double itself by 2050 (FAO, 2015). The country’s demographic structure shows that the population in Ethiopia is predominantly young. About 45% of population is below 15 years old (PHC, 2007; NLFS, 2013). The population growth (high) and demographic structure

See, for example, the country map at Yale G-Econ Project (https://gecon.yale.edu/ethiopia). Accessed on 30 September 2016. See, for example, Gridded Population of the World (GPW) database (https://sedac.ciesin.columbia.edu/data/collection/gpw-v4). Accessed on 20 August, 2019. 11 Illustrated using R Raster package (https://cran.r-project.org/web/packages/raster/index.html). Accessed on 20 July, 2019. 9

10

3.4 Rural-urban divide

17

(young) of Ethiopia have implications for the present and future economic prospects of the country. In the short-term, large young population implies a high dependency ratio. A recent survey shows that the dependency ratio in Ethiopia is about 92% (NLFS, 2013). This, combined with the low level of per capita income, implies rampant child labor (about 16% of total labor force) problem in the country (NLFS, 2005; 2013). In the medium- and long-term, young and growing population implies high resource demand. For instance, the ratio of agricultural land (35 million ha, CSA-FAO, 2014) to rural population (FAO, 2015) declines to 0.29 ha in 2050 down from 0.42 in 2016 and 0.71 ha in 1995. This, among others, will require remarkable improvements in agricultural productivity to ensure that food supply grows on par with population growth. Besides, young people require huge investment in human and physical capital. However, on the flipside of the coin, high and youth population represent cheap labor, and huge commodity market potential. Therefore, the future development impact of the country’s population dynamics hinges on how well the country will manage the population-labor-development nexus. 3.4

Rural-urban divide

Ethiopia is one of the least urbanized nations (only 19% of its total population lives in urban areas), but, with highest urbanization rates (about 4.8% per annum) (CSA, 2013). The high urbanization rate is partly explained by the increasing net rural-to-urban migration. Of the total migrants, the proportion of rural-to-urban migrants rose from 24.3% in 2005 to 32.5% in 2013 (NLFS, 2013). Rural-to-urban migration represents about 33% of total migrants (changed their residence at least once in their lifetime) and 39% of recent migrants (those who changed their residence during last five years) (ICPS, 2012). The trends also show that a shift from rural-torural migration, which has been the dominant form of migration for decades, to rural-to-urban migration (ICPS, 2012; NLFS, 2013). Urban Ethiopia is by far developed than rural Ethiopia in many ways. For instance, 94% of urban households have access to improved water sources whereas only 46% of rural households have access (EDHS, 2014). Likewise, 87% of urban households, but, 6% of rural households have access to electricity utility (EDHS, 2014). The urban literacy rate is as twice as the rural literacy rate (HICES, 2011). Similarly, per capita urban households’ expenditure (USD 526) is as twice as that of rural households’ (USD 250) (HICES, 2011). There is also stark difference in rural and urban labor markets as paid employment is uncommon in rural Ethiopia. Of the total employed rural population 53% are self-employed and 42% are unpaid family workers (ICPS, 2012). This contrasts with 44% self-employed and 8% unpaid family workers in urban Ethiopia (ICPS, 2012). About 83% (12%) of economically active rural (urban) labor is engaged in agriculture, fishing, and forestry (NFLS, 2013). The rural unemployment rate (1%) is by far below than urban unemployment rate (17%) (ICPS, 2012). From the rural-urban labor market contrasts, one can imagine rampant underemployment and disguised unemployment in rural areas. The rural dependency ratio (102) is nearly twice of the ratio in urban areas (58) (NLFS, 2013) implicitly revealing that the average rural household size is larger than urban household size (HICES, 2011; ICPS, 2012; NLFS, 2013). The rural poverty rate (30%) is higher than urban poverty rate (26%) whereas the rural Gini coefficient (0.27) is lower than urban Gini coefficient (0.32) (MoFED, 2012). This shows that the labor skills, sources of employment and

18

3. Overview of the Ethiopian Economy

income in rural areas are relatively comparatively uniform. Put another way, employment and economic opportunities are limited but similar for the majority of rural population. The stark difference rural-urban divide in Ethiopia has three important implications. First, policy priorities in rural and urban areas shall be different. An apparent policy action for rural areas will be to increase agricultural productivity, net farm income, and non-farm income opportunities. In urban areas, the policy need will be more in terms of expanding inclusive employment and economic opportunities which aim to reduce urban unemployment rate, and to deal with increasing rural-urban migration. Second, rural-urban transformation will be the key determinant for economic success in Ethiopia. Third, thus, it needs to account such differences in economic modeling for meaningful policy implications. 3.5

Administrative regions of Ethiopia

Ethiopia is a federation of nine national regional states that are delimited on the basis of “the settlement pattern, language, identity, and consent of the people concerned” (FDRE, 1995, p.17). In addition, there are two federal city administrations, namely, Addis Ababa and Dire Dawa. However, simplicity, both regional states and city administrations are referred as regions in this text. Throughout the monograph regions, regional states, states, or administrative units are used interchangeably. In other words, in this monograph, regions in Ethiopia refer to the subnational entities (administrative units of level one) of the country delineated for the purpose of administration of economic, political, social, and environmental policies. 12 Table 3.1-Selected regional socio-economic indicators Region Tigray Afar Amhara Oromia Somali Benishangul-Gumuz Southern NNP Gambella Harari Addis Ababa Dire Dawa Ethiopia

Urban 20 13 12 12 14 14 10 25 54 100 68 16

Water 73 78 68 56 45 74 69 92 94 83 69

Dependency 97 78 94 107 120 99 104 86 86 41 75 99

Source: Author’s calculations based on various sources Notes: Urban (% total regional population in 2007; PHC, 2007), Water (% households with access to potable water in the total regional households in 2014; NBE, 2016), and Dependency (economically inactive (65) population per 100 economically active population in the region in 2011; HICES, 2011).

Table 3.1 depicts the socio-economic heterogeneity of the Ethiopian regions. For instance, the proportion of regional population living in urban areas ranges between 10% (Southern NNP) and 100% (Addis Ababa) (PHC, 2007). More than half of the population in Dire Dawa and Harari regions lives in urban areas (PHC, 2007; CSA, 2013). The proportion of households that have access to potable water – which represent level of regional development in terms of basic

See also the map at http://www.ethiovisit.com/ethiopia/ethiopia-regions-and-cities.html. Accessed on 30 October 2019.

12

3.5 Administrative regions of Ethiopia

19

services – varies between 45% in Somali and 94% in Addis Ababa (NBE, 2016). The dependency ratio of Somali region (120) is three times larger than the dependency ratio of Addis Ababa (41) (HICES, 2011). The Ethiopian regions also differ in terms their contribution to different aggregate (national) indicators (see Table 3.2). Three main attributes of the regions may explain the variations presented in Table 3.2. The first is related to the size of population and physical geography of the regions which, for instance, explains the regional shares in the national grain output. The second is related to the level of development which, for instance, explains the regional shares in the national gross value of production of small-scale manufacturing industries (GVP-SMI). The population and geographic area of Somali region are larger than those of Tigray column 2 of Table 3.2). 13 But, the share of Tigray region in national GVP-SMI is ten times larger than that of Somali region (see Table 3.2). The third (and most important attribute) is related to geographic location of the regions. Location determines both production (supply) and consumption (demand). For instance, though it is with significant variation, only Southern NNP (or SNNP in short), Oromia, and Gambella regions which produce enset crop. This primarily accrue to their agro-ecological suitability for enset cultivation: Oromia is the second major hub for manufacturing industries which primarily attributes to its geographic position to circle Addis Ababa. The latter is the political, economic, and commercial center of the country. The combination of geographic and population size, geographic location, and level of development influences the tax revenue generated in each regions. Nearly half of the total regional tax revenue is collected in Addis Ababa (MoED, 2015). Table 3.2-Regional shares in selected national indicators Region Tigray Afar Amhara Oromia Somali Benishangul-Gumuz Southern NNP Gambella Harari Addis Ababa Dire Dawa Ethiopia

Population 5.9 1.9 23.4 36.6 6.0 1.1 20.2 0.4 0.2 3.7 0.5 100

Grains 6.5 0.1 32.4 49.5 0.9 1.7 8.8 0.1 0.1 0.0 0.1 100

Enset 0.0 0.0 0.0 31.2 0.0 0.0 68.8 0.02 0.0 0.0 0.0 100

GVP-SMI 11.6 0.3 18.5 34.7 0.9 0.6 11.8 0.3 0.7 19.1 1.6 100

GVP-LMI 9.6 0.5 4.9 36.2 0.0 0.1 3.1 0.0 0.9 42.2 2.4 100

GVP-UDTS 1.7 0.1 7.4 15.7 0.2 0.1 2.6 0.2 1.2 67.8 3.2 100

Tax 8.1 1.1 13.5 17.5 2.1 1.0 9.5 0.6 0.4 45.2 0.9 100

Source: Author’s calculations based on various sources Notes: Regional shares in national: Population (in 2007; PHC, 2007); Grains and Enset (production in tones in 2015; AgSS, 2015), GVP-SMI (gross value of production of small-scale manufacturing industries in 2003; SMIS, 2003), GVP-LMI (gross value of production of large and medium-scale manufacturing industries in 2011; LMIS, 2011), and GVP-UDTS (gross value of production of urban distributive trade enterprises in 2011; UDTS, 2011), and Tax (sum of tax collected from regions in 2011, MoFED, 2015).

In conclusion, Ethiopian regions are heterogeneous socio-economically (see Table 3.1) with significant variation in their contribution to different national (Ethiopia-wide) indicators (see Table 3.2). As such, one may expect that different regions to have different set of socio-economic priorities, and that centrally devised policies and regulations to affect (benefit or cost) See also the map at http://www.ethiovisit.com/ethiopia/ethiopia-regions-and-cities.html. Accessed on 30 October 2019.

13

20

3. Overview of the Ethiopian Economy

the regions unequally. Accordingly, economic policy evaluation exercises shall incorporate the regional effects of centrally made policies and exogenous shocks modeled at the economy-wide level. It is against this background that regional projections and analysis are included in this study. 3.6

Macro-economy

Ethiopia is a low-income country with a per capita income of 690 USD (NBE, 2016). Ethiopian economy starts to take-off only after mid-2000s. In the past decade, the country scored a 10% average annual real GDP growth rate (NBE, 2016; NPC, 2016). The growth is mainly driven by growth in agriculture and services (NBE, 2016). The country has also seen surge in foreign direct investment (FDI) overtime (NBE, 2016; NPC, 2016). In parallel, however, the economy has confronted with price spikes. For instance, relative to 2010/11 (=100), the GDP deflator increased from 2003/04 (=35.3) to 2014/15 (=165) (NBE, 2016). Figure 3.2-Economic structure of Ethiopia 120

SHARE IN GDP (%)

100 80 60

41

42.6

43.7

45.2

45.9

45.5

10.1

10.1

10.2

10.4

11.5

13

13.8

15.2

48.8

47.3

46.1

44.4

43.1

42

40.1

38.8

0

2007/08

2008/09

2009/10

2010/11

2011/12

2012/13

2013/14

2014/15

46.6

Services Industry

40 20

46.6

Agriculture

Source: Author’s illustration based on data from NBE (2016)

The share of industry (manufacturing, construction, mining and quarrying) output staggered around 10% over years but recently starts to increase which attributes to the recent boom in the construction subsector (NPC, 2016). Agriculture remains as the main economic sector in Ethiopia in terms of national income, employment, and export earnings for decades albeit its share in the national GDP is overtaken by the services sector (which include public services) since 2010 (see Figure 3.2). It should be noted here that declining share in the GDP does not mean that the overall role of agriculture in the economy is taken by services (private plus public). Agriculture still employs about 75-80% of the country’s labor force (HICES, 2011; NLFS, 2005; 2013; NPC, 2016), and contributes nine out of ten main export items and 75-80% of merchandise export earnings (NBE, 2016). Altogether, it can be argued that Ethiopia has scored rapid economic growth but unsatisfactory structural change in the last two decades (NPC, 2016).

3.7 Public finance

3.7

21

Public finance

The government revenue sources in Ethiopia can broadly be classified as external (loans and grants) and domestic (taxes, and non-tax revenue such as fees and profits from public corporations). The total government revenue is steadily increasing with average annual growth rate of 12% (2001-2006) and 33% (2007-2012) (MoFED, 2014). The tax revenue is also increasing, for instance, by average annual rate of 13% (1995-2000), 15% (2001-2006) and 36% (20072012) (MoFED, 2014). Along this, the ratio of the tax revenue to the total government revenue is rising overtime. It increased from 64% (1995-2000) to 79% (2007-2012) (MoFED, 2014). Nonetheless, the ratio of tax revenue to the GDP still hovers around 11% (MoFED, 2014: NBE, 2016). The total tax revenue is dominated by import tariffs and duties. Figure 3.3 depicts the tax revenue structure in the period between 2007 and 2012. It shows that tariffs and duties contribute the largest share of total tax revenue (which is also increasing overtime) (MoFED, 2014). Figure 3.3-Tax composition in Ethiopia

Direct Taxes Domestic indirect taxes 34%

42%

Import duties and taxes

24%

Source: Author’s illustration based on data from MoFED (2014)

From the expenditure side, public sector budget are broadly classified as recurrent and capital expenditures (budgets). The capital expenditure is overwhelmingly dominated by expenditures for economic development (agriculture and other industries). For instance, in the period of 2010-2014, about 68% of total capital budget was allocated to economic development compared to 25% to social development (health, education, and social services) (MoFED, 2015). The balance of the capital budget was allocated to general development and miscellaneous expenses. In contrast, the recurrent expenditure was dominated by general services (defense, public security and order), but, recently by social services (education, training, health, and social works). Of the total recurrent budget in the period 2010-2014, about 41%, 39%, and 13% is allocated, respectively, to social services, to general services, and to economic services (agriculture, natural resources, industry and trade, transport, tourism, energy and minerals) (MoFED, 2015). Both recurrent and capital expenditures are steadily increasing with the latter rising at higher rate (see Figure 3.4). This reveals the government’s increasing investment in roads, highways, railways, hydropower, and telecommunication infrastructure projects (MoFED, 2015).

22

3. Overview of the Ethiopian Economy

Figure 3.4-Trends of public expenditure in Ethiopia Annual Growth Rates (%)

35 30 25 Total Reccurent Expenditure

20

Total Capital Expenditure

15 10

Total Expenditure

5 0 2000-2004

2005-2009 Year

2010-2014

Source: Author’s illustration based on data from MoFED (2014)

Looking at the public budget from another perspective, about two third of the total public budget is allocated to public activities aimed at poverty reduction which include education, health, agriculture, water, and roads (MoFED, 2015). The public spending on pro-poor programs and activities aims to spur pro-poor economic growth especially in rural areas where both total and food poverty rates are high (MoFED, 2012). Consequently, agriculture and rural development are at the heart of Ethiopian economic policies and strategies. 14As a result, agriculture and natural resources are the main recipients of recurrent and capital budget for public economic services. For instance, in the period of 2000-2014, 67% of the total recurrent budget for economic services, 36% of the total capital budget for economic development, and 41% of the total public budget for economic sectors was allocated to agriculture and natural resources (MoFED, 2015). In the same time frame, the total budget to agriculture and natural resources accounts for about 8% of the total recurrent expenditure, 24% of the total capital expenditure, and 16% of the total public expenditure (MoFED, 2015). However, in tandem with the dynamics of the macro economy (see Figure 3.3), on average the relative share of agriculture and natural resources in total government budget is declining (MoFED, 2015). This shows that the structural dynamics of public budget and the macro economy go hand-in-hand, and reveals the economic role of the government. By implication, changes in fiscal policies will have profound effects on the macroeconomic dynamics of Ethiopia. On the other hand, the total government revenue has never surpass the total government expenditure in the last four decades (MoFED, 2014; NBE, 2016). Figure 3.4 depicts the ratio of the total fiscal balance to GDP between 1995 and 2012. It shows three important things. First, it reveals that the fiscal deficit is yet a defining characteristic of the public sector balance. Second, the fiscal balance with and without grants is widening in recent years. This indicates that grants are used to smoothen the total budget deficit in the country. Thus, third, it reveals that domestic savings falls short of domestic investment which means the importance of loans and transfers from abroad goes beyond fiscal balance smoothening function.

14 Government of Ethiopia has pursued an Agriculture Development Led Industrialization (ADLI) macroeconomic strategy since 1994 following which each of the five year plans of the country (see, for example, MoFED, 2006; 2010; NPC, 2016) give special focus to agriculture, rural development, and poverty reduction.

3.8 Prospects of the economy

23

Figure 3.5-Fiscal balance in Ethiopia 0.0 Fiscal balance to GDP (%)

1995-2000

2001-2006

2007-2012

-2.0 -4.0

Fiscal Balance -Including Grant

-6.0

Fiscal Balance -Excluding Grant

-8.0 -10.0 -12.0

Year

Source: Author’s illustration based on data from MoFED (2014)

To conclude, both the revenue and expenditure of the government are increasing overtime. Increasing government revenue, especially of tax revenue, attributes to the economic growth and improvement in tax administration (NPC, 2016). The increasing public expenditure shows the increasing government role in the economy. Of course, the Ethiopian government pursues a developmental state approach (NPC, 2016). The government’s spending on infrastructure and poverty reduction programs has far reaching consequences, and shall be taken positively. On the other hand, however, the increasing capital expenditure (i.e., government investment) may crowd out private investment while the increasing recurrent expenditure (i.e. government consumption) has inflationary consequences. In addition, the fiscal balance trend shows that grants from abroad are used to dampen the annual fiscal deficits implying that the government’s role in the economy is partly backed by transfers from abroad. Therefore, changes in transfers from abroad to government have the potential to influence the economic trajectory in Ethiopia. 3.8

Prospects of the economy

The rapid economic growth (see section 3.6), and increasing public capital (investment) expenditure (see section 3.7) in the last two decades have resulted in many palpable improvements in the economy. The road density per 1000 square kilometers has tripled between 2001 and 2015 (NBE, 2016). With improvements in access to agricultural extension services, modern agricultural inputs, and markets, the crop productivity is increasing overtime (CSA, 2015). The headcount poverty ratio has declined by 10 percentage points between 2005 and 2011 albeit it still needs further work (MoFED, 2013). The prevalence of food inadequacy has sharply declined from 81% (1990-92) to 44% (2012-2014) (FAO, 2015). The current fiscal deficit has significantly declined albeit it still remains high (MoFED, 2014). For instance, the ratio of the budget deficit (excluding grants) to GDP was 12% in 2000 and 3.8% in 2014 (NBE, 2015). The domestic sources of finance to the country’s capital expenditure increased from 50% (19992003), to 71% (2004-2008), and to 77% (2009-2013) (MoFED, 2015). The country has also shown notable progress in terms of universal access to education, health services, and reducing child mortality. In the last two decades, Ethiopia has also put its utmost efforts to establish and

24

3. Overview of the Ethiopian Economy

expand physical capital (roads, highways, hydropower dams, and recently railway networks), and human capital (education, and health services). The results are promising with far reaching consequences. However, there are many economic, social, and environmental issues that the country has to seriously take and address yet. Productivity is still low, even, at sub-Saharan Africa standard (NPC, 2016). For instance, Ethiopia is placed at 109th rank among 144 countries in the 2015/16 Global Competitiveness Index (WEF, 2016). 15 Measures to improve partial and total factor productivity in all sectors are crucial (NPC, 2016). The economic success in the last decade is strongly backed by public investment and expenditure. This puts pressure on the fiscal balance. The average investment (domestic savings) to GDP ratio between 1999/2000 and 2014/15 was 29% (13%) (NBE, 2016). This shows that the domestic saving falls short to finance domestic investment (NPC, 2016). Foreign borrowings and grants are used to smoothen this gap, in effect, the international debt stock and service is increasing (NBE, 2016). For instance, between 2012/13 and 2014/15, the external public debt service to GDP ratio increased from 6.6% to 10.1% while the external public debt stock to GDP ratio increased from 24% to 29% (NBE, 2016). The debt accumulation and debt services may influence the future trajectory of the economy. Ethiopia is a net importer with considerable amount of trade deficit. For instance, the net export to real GDP ratio fluctuated between -15% and -23% in the period of 2005-2014 (NBE, 2016). The exports and imports lack diversification. About 75% of merchandise export earnings are from nine agricultural goods (NBE, 2016). Nearly 45% of export revenue is generated only from coffee and gold (NBE, 2016). Nearly 80% of the major imports are finished capital and consumer goods (NBE, 2016), among which, natural gas, petroleum, and fertilizers are entirely imported (EDRI, 2009). Neither the domestic sector is diversified. Only five cereal crops – teff, barely, wheat, maize, and sorghum – account, respectively, for 95%, 73%, and 67% of cereal cropland, temporary cropland, and total cropland cultivated by smallholder peasants (AgSS, 2015). About 92% of the peasants are full-time agricultural workers (CSA-EDRI-IFPRI, 2006). The proportion of crossbred cattle and poultry is hardly 2% (ILRI, 2015). The proportion of irrigated land in total cropland is hardly 2% (CSA-FAO, 2014), and the country hardly uses 5% of its irrigation potential (Awulachew, 2010). Consequently, the smallholder farmers are risk-averse to adopt improved biotechnologies (Wossen et al., 2015). In effect, the improvements in food security, human development, and per capita income are not sufficient particularity compared with the measures undertaken by the government in the agricultural sector (MoARD, 2010; BMGF, 2010). The food supply growth is yet below the population growth rate. Low productivity and poverty are yet defining characteristics of rural areas. Likewise, the manufacturing sector suffers from low productivity and low diversification (NPC, 2016). Food, beverage, and textile related industries account for about 60% of the total large-and medium-scale manufacturing output (LMIS, 2011). The labor market in Ethiopia is underdeveloped yet. About 90% of the total labor force is unpaid self-employee or family worker (EDRI, 2009; HICES, 2011) while about 75-80% of the total labor force is still employed in agriculture (NLFS, 2005; 2013; HICES, 2011; NPC, 2016). Similarly, the labor demand is less diversified. For instance, 99% 15 The Global Competitiveness Index (GCI) report defines competitiveness as “the set of institutions, policies, and factors that determine the level of productivity of an economy, which in turn sets the level of prosperity that the country can earn" (WEF, 2016, p.4).

3.8 Prospects of the economy

25

of the total agricultural labor is employed in crop and livestock farming while 70% of the total employment in manufacturing industries is associated with food, beverage, and textile industries (NLFS, 2013). Yet, due to the stylized facts of the agriculture sector and the rural economy, the Ethiopian economy is predisposed to environmental changes such as climate change and variability (see, for example, World Bank, 2006; Block et al., 2008; Ali, 2012; CSA, 2015). Ethiopia is one of the most vulnerable countries to climate change in the world placed at 39th among 180 countries (ND-GAIN index vulnerability ranking) 16, and 10th among 233 world territories (CGD index vulnerability ranking). 17 With recognition of this, the government has already started planning, and preparing for implementing adaptation policies for the economy in general (NMA, 2007; FDRE, 2011) and for agriculture in particular (FDRE, 2015). Thus, seen from both impact and adaptation perspectives, climate change is an economic development problem in Ethiopia. To conclude, the Ethiopian economy is agrarian and rural yet. Structural change in labor markets, macro economy, and international trade is very slow despite the economy has scored remarkable economic growth in the past decade. Environmental changes such as climate change are yet threats to the country’s economic prospect. The structural dynamics of the macro economy and the public expenditure system go hand-in-hand the role of the public sector in the economy. As such, for instance, fiscal decisions to raise adaptation finance will have profound effects on the rest of economy. In light of this, the policy relevance of assessing the sectoral and general equilibrium effects of climate change and deliberate public adaptation measures will be immense. This is where the present study seeks to contribute its own part.

16 The University of Notre Dame Global Adaptation Index (ND-GAIN) Country Index summarizes a country's vulnerability to climate change and other global challenges in combination with its readiness to improve resilience. URL: http://index.gain.org/country/ethiopia. Retrieved on 2 October, 2016. 17 The Center for Global Development (CGD), Mapping the Impacts of Climate Change ranks world countries based on four dimensions of climate impact - Extreme Weather, Sea Level Rise, Agricultural Productivity Loss, and Overall. URL: http://www.cgdev.org/page/mapping-impacts-climate-change. Retrieved on 2 October, 2016.

4 4.1

Methodological Framework Introduction

The general objective of this study is to assess the economy-wide and regional effects of climate change induced productivity and labor supply shocks and planned public adaptation costs and finance in agriculture. The economy-wide analysis is based on CGE model simulations. The regional analysis is based on a top-down approach that integrates the CGE model results on sectoral outputs with a regional module. This chapter presents the CGE model - choice (4.2), structure (4.3), database (4.4), and calibration (4.5). It also presents how the economy-wide analysis (4.6) is coupled with the regional projections and analysis (4.7). The calibrated CGE model, and the regional module constructed and presented in this chapter are used for all of the empirical questions in the subsequent four chapters. 4.2

CGE model choice

CGE modelling has evolved through time. Today, there are varieties of CGE models. The main difference among them originates from the original purpose and scope. Accordingly, CGE models may be designed for multiple or specific purposes. They may be single- or multi-country, static (single-period) or dynamic (recursive or perfectly forward looking). If a study does not intend build its own CGE model, then, the starting point is to choosing an appropriate CGE model fitting its research objectives, the purpose of the study, and the features of the economy to be investigated. The present study focuses on a single developing country with specific objectives of analyzing the general equilibrium and regional effects of climate change (Chapter 5), of public adaptation costs (Chapter 6), of alternative adaptation finance schemes (Chapter 7), and of climate change while we observe changes in labor markets and transaction costs (Chapter 8). The research questions entail productivity, labor migration, fiscal policies, and transaction cost problems. Among the alternative candidates, I find that the standard IFPRI-CGE model (Lofgren et al., 2002) to fit best to the country and purpose of this study. First, the standard IFPRI-CGE model structure, and behavioral and identity equations are basically built to capture the features of low-income countries. For instance, the model distinguishes between households’ consumption of home (non-market) commodities from market commodities. The model also allows to explicitly account for transaction costs (i.e. trade and transport margins) of market commodities. The model distinguishes commodities from production activities and allows “any activity to produce multiple commodities and any commodity to be produced by multiple activities” (Lofgren et al., 2002, p.2). So doing allows different production technologies for the same commodity and facilitates the treatment of competitive imports (Robinson et al., 1999). Second, the IFPRI-CGE model can easily be mapped to the Ethiopian SAM. The consistency between a model structure and its database is very important in a CGE analysis (Burfisher, 2011). Third, previous economy-wide studies for Ethiopia have widely used the IFPRI-CGE model which shows that the model suitability to the features of the Ethiopian economy (Benson et al., 2014, p.6), but also gives opportunity to compare the results of the present study with those of the previous studies.

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 A. W. Yalew, Economic Development under Climate Change, https://doi.org/10.1007/978-3-658-29413-7_4

28

4. Methodological Framework

The next question is to choose between the static and dynamic version of the model. CGE models are basically static that provide a before-and after-comparison of an economy when a ‘shock’ occurs to the benchmark equilibrium with no description of the path of adjustment (Burfisher, 2011; Cardenete et al., 2012). Dynamic CGE models were conceived to trace the transition (or adjustment process) between the initial and the counterfactual equilibria. I prefer the static to the dynamic model. First, climate refers to average of weather conditions usually over three decades, and hence climate change is represented by anticipated (future) average conditions compared to the present average conditions(IPCC, 2014). Climate change impacts, in this study, are due to these changes in average conditions. Using dynamic models to assess the economic effects of these changes would need parallel projections of the socio-economic system. This makes hard to disentangle the marginal impacts due to climate change from the impacts due to the value of socio-economic system, in effect, it may overstate or understate the economic impacts of climate change (Pielke, 2007). Both may mislead the policy-and-decision making process. If the projected socio-economic scenarios contribute to overstate, it may call for over adaptation efforts. If the projected socio-economic scenarios understate (i.e., mask the detrimental effects of climate changes), it may undermine the imperative of proactive adaptation. Second, in this study, I focus on planned public adaptation (see Chapter 6) that represents deliberate measures in response to anticipated changes in average climatic conditions and their biophysical impacts. Accordingly, decisions regarding what measures to undertake, how much to spend on adaptation, and how to mobilize the adaptation finance start from today which will influence the current economic system. Third, I intend to look at the regional effects of each CGE experiments. It will be a daunting exercise to update economic structure of each regions as the regions hardly have explicit regional policies. Therefore, I apply the standard static IFPRI-CGE model (Lofgren et al., 2002). I find it more sensible to impose the anticipated biophysical impacts of climate change as well as the government decisions with regard to adaptation on the current economic system (which is represented by the 2005/06 SAM). The CGE model or shortly the model hereafter refers to the static IFPRICGE model unless it is mentioned. In the subsequent section, I extract and present the core elements and assumptions of the model from Lofgren et al. (2002). The mathematical details and the GAMS codes of the model can be found in Lofgren et al. (2002). Lofgren (2001) and Robinson et al. (1999) are also important complementary papers to the model applied in this study. 4.3

Description of the CGE model18

The CGE model assumes perfect competition in commodity and factor markets, and Ethiopia as a small-open economy with respect to international trade. The transformation between domestic sales and exports, and the substitution between domestic output and imports are imperfect. This gives a degree of independence for the domestic price system from the international prices which in turn helps to prevent unrealistic export and import responses to shocks. Producers, households, enterprises, government, and rest of the world (ROW) represent decision making nodes in the CGE model. Households, enterprises, government, and ROW are also called institutions (agents that can save) of the model. 18 This section is heavily extracted from Lofgren et al. (2002). The paper and GAMS version of the model are available at URL: https://www.ifpri.org/publication/standard-computable-general-equilibrium-cge-model-gams0. Accessed on 10 July 2014.

4.3 Description of the CGE model17F

29

4.3.1 Production In the model, a producer is represented by an ‘activity’. Producers’ decision with regard to production, supply to domestic and foreign markets, and demand for primary factors and intermediate inputs is dictated by profit maximization goal for a given technology, and output and input prices. 19 The production technology exhibits constant returns to scale. Each producer faces a nested production technology (see Figure 4.1). 20 Similar to other standard CGE models (e.g., Burfisher, 2011; Hertel, et al., 2014), the production function at the trunk of the nest is specified by a Leontief (LEO) function. 21 The composite value-added (QVA) nest is aggregated using a Constant Elasticity of Substitution (CES) function allowing imperfect substitution among primary factors. The prices of primary factors of production are determined in factors market. Each activities use a factor only up to the point where the marginal revenue product of a factor is equal to its wage (or factor price). The composite intermediate input (QINTA) nest is aggregated using Leontief (LEO) function of different intermediate inputs. Intermediate input prices are determined in commodity markets. Figure 4.1-Schematic presentation of the production technology nest

Source: Adapted from Lofgren et al. (2002)

19 For mathematical details on producers’ optimization problem and derivation of the first-order condition, see Varian (1992) for general, and Pauw (2003) and Femenia (2012) for functions similar to those of the present CGE model. 20 This allows a CGE model to describe realistically the different ways that subsets of inputs are combined with each other during the production process (Burfisher, 2011). The selection of input combinations with in each nested process is independent of the contents of other nests, which among others, considerably simplifies the database and the solution of the CGE model (Burfisher, 2011). 21 Of course, it would have also been possible with a Constant Elasticity of Substitution (CES) specification. However, it requires empirical evidence on whether the aggregate mix between aggregate value-added and aggregate intermediate inputs vary (Lofgren et al., 2002).

30

4. Methodological Framework

4.3.2 Commodities Every producer (activity) can produce one or more commodities (c) according to fixed yield coefficients (ϴ). Accordingly, revenue of each activity is determined by the level of the activity (QA), yields, and commodity prices. As shown in Figure 4.2, commodities from an activity are directly consumed at home (homecommodities, valued at activity-specific producer prices) or sold at market (market-commodities, valued at market prices). Market commodities from different activities are then aggregated using a CES function to obtain the aggregate domestic output of a commodity. This acknowledges the varieties of the commodity from different activities are imperfect substitutes, for example, due to differences in timing, quality, and distance between the locations of the activities. Activity-specific commodity prices serve to clear the implicit market for each disaggregated commodity. Figure 4.2-Schematic presentation of the commodity structure

Source: Adapted from Lofgren et al. (2002)

4.3 Description of the CGE model17F

31

The aggregated domestic output for each commodity can be sold in foreign markets or domestic markets. The decision on how much to export and sale domestically is dictated by sales revenue maximization constrained by a Constant Elasticity of Transformation (CET) function and commodity prices. 22 The final demand for a composite commodity comes from households (for consumption), government (for consumption), activities (for intermediate inputs), investment (for capital goods and stock inventories), and transactions services (for distributing market commodities). The composite commodity is composed of domestic and imported varieties – aggregated using a CES function. The total market demand of a commodity is directed to imports (or to domestic output) if there is no domestic production (no imports) for it. 4.3.3 Households Households receive income from factors of production they supply either directly (e.g. labor or land) or indirectly through enterprises (e.g. capital). Households also receive transfers from the government and remittances from the ROW. Conversely, households pay direct taxes to the government, transfer to enterprises and the ROW. Part of households’ income is saved. The leftover income (i.e., after taxes, savings, and transfers to other institutions) is spent on consumption of goods and services. Households are assumed to derive utility only from the consumption of goods and services. Their utility is represented by a Stone-Geary function maximizing which, subject to the budget constraint and commodity prices, gives a Linear Expenditure System (LES) demand. 23 Accordingly, the household demand for a specific commodity is a linear function of the household’s total consumption budget (expenditure). The Stone-Geary utility function acknowledges the need to consume a minimum amount of certain goods (e.g. food items) regardless of their price. As such, households always need to preserve part of their consumption budget in order to meet this minimal consumption. 4.3.4 Enterprises Enterprises are non-government domestic institutions that refer to firms or corporations that “receives capital income, pays corporate taxes, saves (retained earnings), and distributes dividends and profits” (Robinson et al., 1999, p.32). Enterprises also receive transfers from households, government, and the ROW. Enterprises pay direct taxes, saves, and transfer to households, government, and ROW. The payments to and from enterprises are modeled in the same way as households. However, enterprises do not consume, and hence do not have optimization problem.

22 CET function describes the “technological flexibility of producers to transform their product between export and domestic sales” (Burfisher, 2011: 123). 23 For the derivation of LES demand system from the Stone-Geary utility function, calculation of income, ownprice, cross-price, and Frisch parameter of LES demand system see Yeldan (1986), Blonigen et al. (1997), and de Boer and Missaglia (2006).

32

4. Methodological Framework

4.3.5 Government Government (public sector) purchases goods and services (mainly public services) and make transfers to households and the ROW. Government collects taxes and receives transfers from public enterprises and the ROW. The balance between government revenue and expenditure is maintained by government saving. 4.3.6 Rest of the world The ROW (foreign or external sector) interacts with the domestic economy in many ways. The ROW transfers to households and government, and purchases exports. It supplies imports, possess factors of production (usually capital), and receives transfers from households and government. The transactions between the external sector and domestic economy are held in foreign currency. Exchange rate expressed as local currency units to foreign currency unit mediates the transactions between the external sector and the domestic economy. The balance between the income and expenditure of the ROW is maintained by foreign saving. 4.3.7 Saving-Investment Saving-Investment (S-I) balance imposes a constraint on the macro economy. The sum of savings from households, government, and the ROW equals to the investment demand (fixed gross capital formation plus stock changes or inventories) of the economy. The S-I balance is an identity equation with no residual variable. Therefore, either saving or investment demand has to adjust to maintain the balance. Like in many other applied static CGE models, investment in the present model is treated as final consumption that merely represents a demand category. As such, investment has no effect on the supply side in the model (Cardenete et al., 2012). Accordingly, the CGE model has nothing to say about the effects of any policy simulation on the level and composition of investment (Missaglia and de Boer, 2004). 24 4.4

The CGE model database

The CGE model database in this study is the 2005/06 SAM of Ethiopia. 25 The original SAM has 254 total accounts that consists of 98 activity, 93 commodity, 26 factor, 14 household, 17 Modeling investment decisions, which are intrinsically dynamic, in static models is open to debate (Missaglia and de Boer, 2004; Cardenete et al., 2012). Nevertheless, investment has to be modeled somehow if a SAM has to reproduce a picture of a given economy (i.e. for accountancy) (Missaglia and de Boer, 2004) and CGE models have to play an important role in policy analysis (Cardenete et al., 2012). 25 This was the latest publicly available SAM for Ethiopia until April, 2015. Constructing own SAM was not viable in the study period. Of course, using a decade old SAM is common practice in empirical studies dealing with developing countries, see for example, Wiebelt et al. (2015) (for Tunisia), Bezabih et al. (2011) (for Tanzania) and Ahmed and O’Donoghue (2010) (for Pakistan). It is justified partly by the paucity of data and partly because structural change is slow in most LDCs. The policy relevance of the study would apparently have been better with latest SAM. But not as such, as the main purpose of a SAM is to depict the underlying relationship among production sectors and institutions (Ahmed and O’Donoghue, 2010). It suffices to assume that there were no major policy and structural changes in the last ten years that affect the underlying relationships presented in the 2006 SAM. In fact, structural change in Ethiopia has been very slow in the last decade (NPC, 2016; Dorosh and Thurlow, 2011). In addition, the SAM year falls in the range of current climate (i.e. 1980-2010 period). For consistency, one shall assume that as if the shocks representing different research questions (to be discussed in chapters 6 to 8) happened and valued in the 2005/06 fiscal year, and the interpretation of results shall keep in mind the reference to the 2005/06 prices. 24

4.4 The CGE model database

33

tax, and 6 other accounts (EDRI, 2009). 26 However, I modify (aggregate and disaggregate) the SAM into 54 total accounts that consists of 17 activity, 18 commodity, 8 factor, 2 household, 3 tax, and 6 other (enterprise, government, ROW, savings-investment, changes in stock inventory-DSTK, and transport and trade margin) accounts. In modifying the SAM, I took three important aspects of the study into consideration. The first is that to make the SAM more suitable for the study purpose. I find it necessary to aggregate all grains in to a single activity so that the crop productivity effects of climate change are more plausible (Chapter 5). I aggregate tax accounts since the adaptation costs would be too much shock to a specific tax type, say for example, labor income tax (Chapter 7). It is necessary to have a benchmark public budget to ‘agricultural productivity enhancing measures’ (Chapter 6 and 7). There is no specific budget or national income accounts (and hence in the original SAM) representing such purpose and account. Such public services would fall in the ‘public administration, defense, and compulsory social services’ category of the national income accounts and hence of the original SAM (see, for example, MoFED, 2005, p.72; UNDESA, 2008, p.245). Therefore, I split the ‘public administration, defense, and compulsory social services’ (activity and commodity) of the original SAM into public administration (general, 80%) and public administration (agriculture, 20%). To be precise, the public administration (agriculture) shall be referred as public activity/spending on measures relevant to adaptation in agriculture which include irrigation and water management (irrigation, water harvesting, hydrological and metrological information), rural roads, and agriculture (R&D, extensions and trainings, and climate services). 27 The second is to facilitate calibration of the model. I find it necessary to aggregate commodities and households of the original SAM to easily borrow (extrapolate) elasticities of the model (see Table 4.11) from the empirical literature. The third is to keep the level of industrial aggregation in the SAM consistent with the regional module (see Table 4.13). It is hardly possible to construct a regional module (to be discussed in section 4.7) based on the original SAM. The SAM henceforward refers to the modified SAM which I present the details below. The list, notation, and description of all accounts of the SAM are presented in Table A1 in the Appendix.

26 The 2005/06 Ethiopian SAM is constructed by the Ethiopian Development Research Institute (EDRI) in collaboration with the Institute of Development Studies (IDS, University of Sussex), and the International Food Policy Research Institute (IFPRI). The original SAM is constructed from different national data sources mainly from Central Statistical Agency of Ethiopia (CSA), Ministry of Agriculture and Rural Development (MoARD), Ministry of Finance and Economic Development (MoFED), and National Bank of Ethiopia (NBE) (EDRI, 2009). The final balanced SAM is produced using a statistical cross-entropy approach (EDRI, 2009). 27 I split this account based on information from various related policy documents. In 2013/14, nearly 20% of recent public investment in roads pertain to Woreda or rural feeder roads (NBE, 2015). About 23% of planned expenditures on water sector development (2002-20016) are allotted to irrigation (MoWE, 2001). Agricultural extension services and R&D, respectively, takes 22% and 5% of government’s expenditure to agricultural and rural development (Lanos et al., 2014). About 15% of total public expenditure is allotted to agriculture and natural resources (MoARD, 2010). I have also referred other documents such as Eshetu et al. (2014), MoFED (2014; 2015), MoFED (2010), and ReSAKSS (2014).

34

4. Methodological Framework

4.4.1 Activity (A) accounts The SAM consists of 17 activity accounts. AGRAIN is agricultural activity producing grain crops –cereals, pulses, and oilseeds. ACCROP produces high-value agricultural commodities like vegetables, fruits, coffee, chat, cotton, sugar cane, spices, tea, tobacco, and crops not classified elsewhere. AENSET is an agricultural activity that produces enset crop. ALIVST represents all activities related to livestock production–raising animals and production of animal products. AFISFOR is an activity producing fish and forest products. AMINQ refers to minerals and quarrying. ACONS stands for construction activities. AMAN is manufacturing activities which include large, medium, and small-scale, and cottage manufacturing industries plus electricity and water supplies. Public administration (general, APADMN) and public administration (agriculture, APAGRI) are entirely produced by the public sector. ASSER include both private and public social services (education, training, health, and social works). The rest of production activities include trade (ATSER), hotels and restaurants (AHSER), transport and communications (ATRNCOM), financial intermediaries (AFSER), real estate and renting business (ARSER), and ‘other’ services (AOSER). ‘Other’ services (AOSER) include services not classified elsewhere such as community, social, and personal services, private households with employed persons (e.g. babysitters, guards, maids), and extraterritorial originations. Table 4.1 presents the intermediate (QINTA), value-added (QVA), and total production value (QA) of activities of the SAM. Table 4.1-Activity accounts in the SAM Activity AGRAIN ACCROP AENSET ALIVST AFISFOR AMINQ ACONS AMAN ATSER AHSER ATRNCOM AFSER ARSER APADMN APAGRI ASSER AOSER TOTAL

QINTA Value 0.44 0.09 0.04 0.04 0.10 0.01 1.83 1.51 1.37 0.63 0.46 0.14 0.14 0.40 0.10 0.17 0.02 7.48

% QA

15 6 19 2 13 9 75 62 46 68 38 34 11 42 42 21 5 35

QVA Value 2.52 1.40 0.16 2.02 0.66 0.08 0.61 0.93 1.59 0.29 0.73 0.26 1.11 0.55 0.14 0.62 0.39 14.05

Source: Author’s modified SAM from EDRI (2009). Notes: The values are in billion USD at 2005/06 prices.

% QA

85 94 81 98 87 91 25 38 54 32 62 66 89 58 58 79 95 65

QA Value 2.96 1.48 0.19 2.07 0.76 0.08 2.44 2.44 2.96 0.92 1.19 0.40 1.25 0.94 0.24 0.78 0.41 21.53

One can get the glimpse of economic structure of the country from Table 4.1. In terms of contribution to the total value-added GDP, agriculture (the sum of AGRAIN, ACCROP, AENSET, ALIVST, and AFISFOR) is the dominant economic sector. It accounts for 48 % of the total value-added GDP. However, agriculture’s demand for intermediate goods represents only 10% of the total intimidate goods demand. This reveals the weak interlinkage between agriculture and nonagricultural sectors in the economy. Services, including public services, contributes to 40% of the aggregate value-added GDP. Services are the main (46%) intermediate input purchasers. The contribution of manufacturing activity (AMAN) to the total GDP is 11% which is equivalent to the contribution of the construction activity (ACONS). The contribution from

4.4 The CGE model database

35

mining and quarrying (AMINQ) is less than 0.5%. Table 4.1 also depicts information on factor intensity of each activities. Generally speaking, the Ethiopian economy is primary factor intensive. Aggregate value-added GDP accounts about 65% of the total value of production. The agricultural activities, mining and quarrying, real estate and renting, financial intermediaries, social services, other services, and transport and communications, and public administration (general and agriculture) are factor intensive. Conversely, the construction, manufacturing, and hotels and restaurants are relatively less factor intensive. The construction activity is the least, the livestock and cash crop activities are the most primary factor intensive activity. 4.4.2 Factor (F) accounts The factor accounts dissect the aggregate value-added (QVA) nest. The primary factors of the SAM can simply be classified as labor and non-labor factors. The labor factors are further disaggregated in to five segments. FLAB0 refers to agricultural labor which is employed only in agricultural activities. FLAB1 group includes administrative workers, senior managers and legislators. FLAB2 category refers to professional and technical associates. FLAB3 includes laborers employed in elementary occupations that require no specific skills. FLAB4 includes workers (e.g. clerks, artisans, and salesmen) with some skill acquired through formal and informal trainings or experience. 28 The non-labor factors include cropland (FLND, employed only in crop activities–AGRAIN, ACCROP, and AENSET), livestock (FTLU, employed only in livestock activity–ALIVST) and non-agricultural capital, or simply capital (FCAP). Table 4.2 presents the composition and share of primary factors of production for the 17 activities of the SAM. Generally speaking, the Ethiopian economy is labor intensive. The total labor wage bill accounts for about 50% of the aggregate value-added in the economy. About 36% of the total value-added in the economy pertains to agricultural labor. This reflects the importance of the agriculture sector in the economy in terms of employment. Table 4.2 also shows the factor intensity of different economic activities. About 74% of agricultural value-added is paid to agricultural labor. Fishing and forestry (AFISFOR) is the only agricultural activity that uses non-agricultural capital (FCAP). This may attribute to the fact that the vast majority of agricultural activities uses traditional equipment and instruments (CSA-EDRI-IFPRI, 2006) and hence the value of capital is included in the value of land (EDRI, 2009). The share of labor is also significant in ‘other’ services (56%), hotels and restaurants (50%), and social services (47%). The bulk of labor wage bill of public services (administration and social services) go to administrative workers (FLAB1) and professionals (FLAB2) followed by to skilled labor (FLAB4). Manufacturing and private services are skilled labor (FLAB4) intensive. The skilled labor (FLAB4) share is exceptionally high in hotels and restaurants (48% of its total payments to factors). The share of unskilled labor category (FLAB3) in ‘other’ services (21%) is the highest. The real estate and renting services (99%), and transport and communications (90%) are the two most capital intensive economic activities in the economy.

The original SAM (EDRI, 2009) calls this labor segment as ‘skilled’ labor group. I kept the notations here to easy correspondence.

28

36

4. Methodological Framework

Table 4.2-Factor accounts in the SAM Payments to factors (%) Activity

FLAB0

FLAB1

FLAB2

FLAB3

FLAB4

FLND

FTLU

FCAP

QVA

AGRAIN

82

0

0

-

-

18

-

-

100

ACCROP

67

0

0

-

-

33

-

-

100

AENSET

57

0

0

-

-

43

-

-

100

ALIVST

69

0

0

-

-

-

31

-

100

AFISFOR

91

0

0

-

-

-

-

9

100

AMINQ

-

1

1

8

18

-

-

72

100

ACONS

-

1

2

7

15

-

-

74

100

AMAN

-

0

1

8

28

-

-

62

100

ATSER

-

1

1

0

19

-

-

79

100

AHSER

-

2

0

0

48

-

-

50

100

ATRNCOM

-

1

1

0

8

-

-

90

100

AFSER

-

4

4

1

7

-

-

85

100

ARSER

-

0

0

0

0

-

-

99

100

APADMN

-

8

10

1

12

-

-

69

100

APAGRI

-

8

10

1

12

-

-

69

100

ASSER

-

1

42

1

3

-

-

53

100

AOSER

-

4

12

21

17

-

-

46

100

TOTAL

36

1

3

2

7

7

4

39

100

Source: Author’s modified SAM from EDRI (2009)

4.4.3 Commodity (C) accounts: Supply The SAM has two types of commodities, namely, home- and market-commodities. Home commodities are recorded as payments from households to activities. Such commodities are produced and directly consumed by households, and thus do not enter into the market. Market commodities are sold either in domestic or foreign markets. About 21% of aggregate level of domestic activity can be labelled as home-commodities. Less than half of forest and fish, grain, enset, and real estate output value enter to the market. It is important to note here that about 80% of the population in Ethiopia lives in rural areas where housing markets does not exist. Nearly 50% of livestock and 30% of cash crop outputs are consumed at home. Table 4.3 reveals the structural rigidity of the commodity markets. This is not a surprise as the majority of Ethiopian population lives in rural areas and depend on subsistence agriculture. As the majority of agricultural commodities are consumed at home, ceteris paribus, shocks to agricultural production directly affect real households’ consumption.

4.4 The CGE model database

37

Table 4.3-Share of market commodities from activities in the SAM Activity AGRAIN ACCROP AENSET ALIVST AFISFOR AMINQ ACONS AMAN ATSER

% QVA 44 70 47 52 27 100 100 97 100

Activity AHSER ATRNCOM AFSER ARSER APADMN APAGRI ASSER AOSER TOTAL

% QVA 100 100 100 48 100 100 100 100 79

Source: Author’s modified SAM from EDRI (2009)

4.4.4 Commodity (C) accounts: Demand Intermediate input (QINT), transaction margin (trade) inputs (TRC), government (GOV), households (HH), investment (QINV), and exports (QE) represent the final demand for each commodities. Table 4.4 depicts the final demand mix for commodities (home plus market) of the SAM. According to Table 4.4, the majority of the commodities are supplied to domestic markets. The share of exports in the total final demand of a commodity exceeds 20% only for cash crops (CCCROP) and transport and communications (CTRNCOM). Government purchases virtually all of the public administration services (CPADMN and CPAGRI). Government is also the main purchaser of social services (82%) showing that education and health services are mainly provided by the public sector. The bulk of final demand for crops, livestock, manufacturing, real estate and renting, and ‘other’ services commodities comes from households. Whereas the bulk of demand for minerals, forest products, transport and communications, financial intermediaries, and considerable amount of real estate, manufacturing, and other services comes from intermediate input. Construction and manufacturing commodities are the two major investment goods. Mineral goods are mainly used as intermediate inputs, and inventories (negative investment goods). The transaction cost (TRC) is a special intermediate input demand, i.e., a “trade input” consumed in marketing domestic sales, exports, and imports (Lofgren, 2001; Lofgren et al., 2002). Nearly 90% of the total whole sale and retail trade services considered to be ‘trade’ input. Consequently, the value of trade activities depend on the size of market commodities and trade margins realized on them. The CMMAN commodity represents entirely imported manufactured goods which include fertilizers, coal, natural gas, and petroleum oil. Nighty percent of CMMAN commodity is used as intermediate input.

38

4. Methodological Framework

Table 4.4-Final demand for commodities in the SAM Commodity CGRAIN CCCROP CENSET CLIVST CFISFOR CMINQ CCONS CMAN CMMAN CTSER CHSER CTRNCOM CFSER CRSER CPADMN CPAGRI CSSER COSER

Final demand components (%) QINT TRC GOV 25 14 17 14 51 129 11 35 90 8 0 58 59 45 0 0 4 35

0 0 0 0 0 0 0 0 0 88 0 0 0 0 0 0 0 0

HH

0 0 0 0 0 0 0 0 0 0 0 0 0 0 99 99 82 0

Source: Author’s modified SAM from EDRI (2009)

QINV 64 51 83 72 47 23 0 42 11 3 95 15 35 53 1 1 13 57

0 -1 0 7 0 -53 89 19 -1 0 0 0 0 0 0 0 0 0

QE 11 36 0 8 1 0 0 4 0 1 5 27 6 2 0 0 1 8

TOTAL 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100

4.4.5 Commodity (C) accounts: International trade Table 4.5 shows that Ethiopia is net importer. The majority of Ethiopia’s export comes from agriculture (43%) and transport and communications (31%). 29 Imports are dominated by manufactured goods (that competes with domestic output, CMAN, 56%, and entirely imported, CMMAN, 15%), and transport and communications goods and services (18%). Therefore, agriculture is net exporter while manufacturing is net importer. Public administration services (CPADMN and CPAGRI), and enset crop are entirely non-tradable. Table 4.5-International trade in the SAM Commodity CGRAIN CCCROP CENSET CLIVST CFISFOR CMINQ CCONS CMAN CMMAN CTSER CHSER CTRNCOM CFSER CRSER CPADMN CPAGRI CSSER COSER

Exports (%)

10 28 5 0 0 0 17 2 2 31 1 1 0 3

Imports (%)

4 1 0 0 0 0 56 15 0 1 18 1 0 0 4

Source: Author’s modified SAM from EDRI (2009)

29

Transport and communication services include services from airline and shipping services (MoFED, 2005).

4.4 The CGE model database

39

4.4.6 Government (GOV) account The government account shows the revenue and expenditure patterns of the public sector. The tax revenue (the sum of STAX, MTAX, and DTAX) comprises about 60% of the total government revenue. Non-tax revenue from public enterprises (ENT) and transfers from the ROW make nearly 40% of total government revenue. Import tariff (MTAX) is the single most important source of government revenue (30%). The least contribution to government revenue comes from sales tax (13%) which, partly, is explained by the important but non-commercialized agriculture. Table 4.6-Government account in the SAM

Government Revenue

Category

Item

Account

Non-tax revenue

ENT

23

Tax Revenue

STAX

13

Tax Revenue

MTAX

30

Tax Revenue

DTAX

17

Transfers

ROW

16

Government Expenditure

TOTAL Consumption

CPADMN

Consumption

CPAGRI

Consumption

CSSER

Transfers

RURH

Transfers

URBH

Savings

S-I

Transfers

ROW

TOTAL

Source: Author’s modified SAM from EDRI (2009)

Share (%)

100 35 9 25 3 4 23 2 100

The government expenditure side here represents only recurrent expenditures (EDRI, 2009). Nearly 70% of the total government recurrent expenditure is allocated to consumption of public services (CPADMN, CPAGRI, and CSSERV). Government transfers to rural households (RURH), urban households (URBH), and the ROW make about 10% of the total government expenditure. The transfers to households include food and related relief aid, pensions, and other safety-net programs while transfers to the ROW refers to interest payment for foreign public debts (EDRI, 2009). 30 Government saving (GSAV) is the residual of government revenue and 30 GSAV is different from the total fiscal surplus (deficit) which refers to the difference between total government revenue and total (recurrent plus capital) government expenditure. Capital expenditure represents government investment expenditure. The recurrent budget surplus (GSAV), net domestic borrowing, and net international borrowings are the source of finance for government investment (capital expenditure) (MoFED, 2014; EDRI, 2009).Therefore, the total fiscal surplus (deficit) is the difference between government investment (capital expenditure) and government recurrent surplus (GSAV). In other words, the total government deficit is equal to the sum of net domestic and international borrowings of the government. In 2005/06, government investment accounted for about 35% of total investment in the economy (EDRI, 2009). The total fiscal deficit in 2005/06 was

40

4. Methodological Framework

recurrent expenditure, thus, refers to recurrent budget surplus. GSAV accounts for 23% of the total recurrent government expenditure. 4.4.7 Tax (TAX) accounts Tax accounts are virtual accounts in a SAM (Hosoe et al., 2010). They collect taxes from commodities, households, enterprises, and factors, if any, and then pay to the government (GOV). Table 4.7 presents tax accounts of the SAM. Direct tax (DTAX) includes all income taxes. The bulk of direct tax (64%) is collected from urban households (URBH). The contribution of rural households (RURH) to total direct tax is negligible, especially, when we compare to the rural population size. Table 4.7-Tax accounts in the SAM Institution ENT

DTAX (%) 33

RURH

4

URBH

64

Commodity

STAX (%)

Agricultural

4

MTAX (%) 0.0

Manufactured

60

100

Services

36

-

Source: Author’s modified SAM from EDRI (2009) Notes: DTAX (direct taxes), STAX (sales tax), and MTAX (import tariffs and duties)

About 60% of sales tax (STAX) is collected from manufactured (industrial) goods. The import tariffs (MTAX) are virtually collected from manufactured (CMAN and CMMAN) imports. Combining the tax structure (Table 4.7) and household accounts (Table 4.8) shows that taxation in Ethiopia is progressive. 4.4.8 Households (H) account The household accounts in the SAM depicts the income and spending pattern of rural (RURH) and urban (URBH) households. The source of income for households include factor income (from labor, land, and livestock) and transfers (from government, ROW, and enterprises). Transfers from enterprises to households represent factor income from capital (FCAP). Therefore, overall, about 93% of rural households’ accrue to their factor ownership. Agricultural labor and capital (i.e., transfers from enterprises), respectively, contributes to 45% and 33% of the total rural households’ income. Factor income (non-agricultural labor and capital) and transfers (remittances) from abroad, respectively, accounts for 68% and 29% of the total urban households’ income. About 90% (40% goes to home-commodities) of the total rural households’ expenditure is allocated to consumption of goods and services. The balance is saved (11%), and paid as income taxes (which is nearly null). About 81% (3% goes to home commodities) of the total urban households’ expenditure is allocated to consumption of goods and services. The remaining is saved (13%), paid as income tax (7%), and remitted to abroad (almost nil). Thus, from Table 4.8, one may expect that urban households are relatively sensitive to macroeconomic shocks (government balance, S-I balance, and ROW balance) than their rural counterparts. 0.7 billion USD (equivalent to 4.6% of GDP) (EDRI, 2009). The fiscal deficit presented in Figure 3.5 is total government deficit.

4.4 The CGE model database

41

Table 4.8-Household accounts in the SAM

Income

Category

Account

Item

Factor Income

FLAB0

Factor Income

FLAB1

Factor Income

FLAB2

Factor Income

FLAB3

Factor Income

FLAB4

Factor Income

FLND

Factor Income

FTLU

Transfers

ENT

Transfers

GOV

Transfers

ROW

Expenditure

TOTAL Commodity Consumption Commodity Consumption Transfers

Home

Savings

S-I

Transfers

ROW

TOTAL

Market DTAX

RURH

URBH

45

0

0

2

1

7

0

6

0

26

9

0

6

0

33

27

1

3

6

29

100

100

39

3

50

77

0

7

11

13

-

0

100

100

Source: Author’s modified SAM from EDRI (2009)

4.4.9 External sector (ROW) account Table 4.9 presents the external sector account of the SAM. About 98% of the total payments from Ethiopia to the ROW go to imports. Export earnings accounts for about 35% of the total payments from the ROW to Ethiopia. This implicitly shows that exports fall short to finance imports. Foreign saving (FSAV) is the second largest source of income from abroad (23%).31 In addition, Table 4.9 shows that remittance to households (33%) is four times that of the ROW transfers to government (8%).

31 Foreign saving (FSAV) refers to the current account balance. It refers to the net resource inflow in terms of goods and services (EDRI, 2009). As Ethiopia is net importer, Table 4.5, FSAV represents a current account deficit to Ethiopia but a positive saving to the ROW. Current account deficit (or FSAV) accounted about 8.3% of the total GDP in 2005/06 (EDRI, 2009). It is different from capital account balance which includes transactions (purchase and sale) of foreign assets and liabilities. Mathematically, the current account deficit (FSAV) is the government’s net foreign borrowings plus foreign direct investment (FDI) minus acquisition of net foreign exchange reserves (EDRI, 2009). As such, Ethiopia finances its current account deficit by increasing net foreign borrowing (which will have debt burden), by attracting more FDI, or by depleting its stock of foreign exchange reserves (EDRI, 2009).

42

4. Methodological Framework

Table 4.9-The external sector account in the SAM Category

Account

Share (%)

Income

IMPORTS

98

FCAP

0

GOV

1

URBH

0

TOTAL

100

EXPORTS

35

Expenditure

FCAP

1

GOV

8

RURH

12

URBH

21

S-I

23

TOTAL

100

Source: Author’s modified SAM from EDRI (2009)

4.4.10 Saving-Investment (S-I) account Saving-investment (S-I) account of the SAM can be regarded as “loanable funds market” (EDRI, 2009, p.15). It collects savings (S) from households, government, and external sector. Rural households and the ROW contribute about 70% of the total saving. The balance is contributed by government (17%) and urban households (14%). The fact rural households appear as the main contributor to national savings (35%) accrues to the rural population size. Otherwise, as shown in Table 4.8, their saving relative to their total spending is lower than that of urban households. Table 4.10-Savings-Investment account of the SAM Savings GOV RURH URBH ROW

Share (%) 17 35 14 34

Investment

Share (%)

CCCROP CCONS CMAN DSTK

0 60 28 12

Source: Author’s modified SAM from EDRI (2009)

The accumulated savings are then allocated to gross investment that comprises capital goods and stock change (or inventory investment) (DSTK). 32 The three investment goods, i.e., cash crops, construction, and manufactured goods represent 88% of the total investment demand. Inventory changes account for the remaining 12%.

The investment can either be government or non-government. The information on who owns the capital goods or in which sectors they are installed is not indicated in the original SAM (EDRI, 2009). Investment demand is by sector of origin not by sector of destination. Thus, the SAM does not confer information about changes in sectoral capital stocks or their valuation (EDRI, 2009).

32

4.5 The CGE model calibration

4.5

43

The CGE model calibration

Calibration refers to the process to enable the CGE model (and model functions) to reproduce the SAM as a model solution (Mansur and Whalley, 1981). As such, calibration depends on the model structure (see section 4.3 and Lofgren et al., 2002) and the SAM of the economy under investigation (see section 4.4). Due to the model structure and functions (CES, CET, and LES), we need a set of extraneous data for elasticities of factor substitution, import substitution, export transformation, income, and Frisch parameters. 33 The elasticities can be obtained by applying econometric techniques to a data set similar to the SAM (Andre´ et al., 2010; Burfisher, 2011), by borrowing (and complemented with educated guesses) from the empirical literature on the same economy or on economies in similar level of development (Sanchez, 2004), and/or by assuming (guided by economic theory) (Andre´ et al., 2010; Burfisher, 2011). In practice, the last two approaches are common (see also section 2.5). In similar vogue, I referred the empirical literature related to Ethiopia and other low-income countries to extrapolate a range of plausible values for the aforementioned elasticities which are presented in Table 4.11. Generally speaking, the elasticities can be said low. The factor substitution (income) elasticities increase from agricultural activities (commodities) to service activities (commodities) whereas trade elasticities increases with the tradability of the commodity. Such calibration is consistent with economic features in developing countries (e.g., Roberts, 1994; Hertel et al., 2014; Hertel and van der Mensbrugghe, 2016). Table 4.11-Summary of elasticities used in calibration Elasticities Elasticities of factor substitution (for each activities) Elasticities of import substitution (for each import goods) Elasticities of export transformation (for each export goods) Income elasticities (for each household consumption goods) Frisch parameter (absolute value) (for each household groups)

Range

Main references

0.3–2.0

Balistreri et al. (2003); Hertel et al. (2014) Hertel et al. (2014); Hertel and van der Mensbrugghe (2016); Jomini et al. (1991) Hertel et al. (2014); Hertel and van der Mensbrugghe (2016); Jomini et al. (1991)

0.5-2.0 0.5-2.0 0.7-1.5 1.5-2.0

Diao et al. (2012) Lluch et al. (1977); Sanchez (2004); Nganou (2011); Yeldan (1986)

Factor market closure is another issue in model calibration. All factors are assumed to be fully employed. I obtain the observed employment of each labor types per activity from NLFS (2005). I use AgSS (2006) to allocate the aggregate agricultural labor (from NLFS, 2005) among the five agricultural activities of the SAM, and to compute the tropical livestock unit (TLU).34 For capital, I set the average wage rate equal to one, thus, the observed employment of capital per activity is represented by the payment from the activity to capital in the SAM. For each factors, an economy-wide wage rate is flexible to assure that the sum of factor demands is equal to the fixed (observed) quantity of factor supply. Labor (of each skill types) and Frisch parameter for a household group measures the proportion of the households’ consumption expenditure that is preserved for subsistence consumption. The higher is the share of subsistence in total consumption budget, the larger is the absolute value of the Frisch parameter. The richer is the household group, the lower is the budget share of subsistence consumption. The higher is the marginal propensity to save, the lower is the minimum (floor) consumption budget share. Accordingly, I set the Frisch parameter of (-2) and (-1.5), respectively, for rural households and urban households. 34 1 TLU=1*camels + 0.7*cattle + 0.1*sheep + 0.1*goats + 0.01*poultry 33

44

4. Methodological Framework

land are assumed to be mobile across activities whereas livestock and capital are activity-specific. The factor market closure rules that I apply here are also commonly used in empirical CGE models for developing countries (e.g., Robinson et al., 1999; Lofgren et al., 2002). We shall also specify macro closures in the calibration process. The real exchange rate (EXR) is flexible while the foreign saving (FSAV) is fixed in external sector balance. The closure is appropriate as Ethiopia applies managed floating exchange rate whereas generally in developing countries the availability of foreign saving, for example through FDI, is constrained (Hosoe et al., 2010). As such, ceteris paribus, if the foreign saving is different from the exogenous level, the real exchange rate would correct the situation by simultaneously affecting the spending on imports and earning from exports (Lofgren et al., 2002). In other words, equilibrium in the external sector will be achieved through movements in exchange rate to affect export and import prices relative to domestic prices (Robinson et al., 1999). By implication, the government cannot merely increase its foreign debt to finance its new investments. Export earnings shall increase in order to pay for additional imports. The government saving (GSAV) adjusts to maintain the balance between government revenue and recurrent expenditure. All tax rates and real government consumption of goods and services are fixed. They are exogenous to the model since both are determined by political process than changes in the economic environment. The S-I balance closure is investment-driven. That is the savings from households and government shall adjust towards the fixed level of real investment. The combination of the macro closures that I apply here is commonly referred as the ‘Johansen’ closure type (Lofgren et al., 2002), and is highly commended for simulations that aim to examine the equilibrium welfare changes due to alternative policies using static models (Lofgren, 2001; Lofgren et al., 2002; Hosoe et al., 2010). It will, among others, preclude misleading welfare effects of simulations that may arise from changes in foreign savings, real investment, and government consumption (Lofgren, 2001; Lofgren et al., 2002; Hosoe et al., 2010). The consumer price index (CPI) is the numeraire of the model. The CGE model determines prices relative to this reference price. Accordingly, all simulated changes shall be interpreted relative to the CPI. The model is implemented in General Algebraic Modeling System (GAMS) software and solved as a Mixed Complementarity Problem (MCP) using the PATH Solver (see www.gams.com for more). The calibration is checked to be successful. The SAM values were reproduced. Doubling the numeraire doubles all nominal values (i.e., prices, incomes, and expenditures) but leaves all real quantities unchanged. The solution value of WALRAS variable is zero. By now, therefore, the model is ready to undertake experiments and make comparative static analysis.

4.6

CGE simulations and economy-wide analysis

Each research questions are considered as experiments to the calibrated CGE model. The experiments can be exogenous changes (e.g., productivity effects of climate change) or policy changes (e.g., adaptation policy). The CGE experiments are represented by altering the values of one or more exogenous variables and/or parameters. As such, experiments represent shocks to the calibrated model. Solving the CGE model for each experiments yield a different set of commodity and factors prices relative to the numeraire. Changes in relative prices of commodities alter the households’ consumption expenditure and producers’ revenue. Changes in factor

4.7 Regional projections and analysis

45

prices alter factor incomes (for households) and costs of production (for producers). Consequently, there will be change in utility-maximizing basket and demand for consumption goods (for households), profit-maximizing basket and level of production (for producers), composition and level of exports and imports (affecting the ROW balance), tax revenue (affecting the government balance), and level of savings (affecting the S-I balance). In short, CGE experiments result in new configuration in the economy which is commonly referred as a counterfactual general equilibrium. The CGE model simulates changes to all endogenous variables of the model simultaneously. All simulated results are expressed as percentage change, and shall be interpreted relative to the numeraire of the model. Since CGE model results are based on economy-wide representative agents, the analysis is commonly referred as economy-wide analysis. In practice, CGE or economy-wide analysis focus on selected variables which can offer good picture on the effects of a specific shock. For the economy-wide analysis of this study, I focus on the macro economic effects, sectoral output effects, factor market effects, and households’ welfare effects. Analyzing the macroeconomic effects confers information on the overall effects. Nevertheless, the macroeconomic effects conceal resource allocation effects which are better portrayed through sectoral output effects. Sectoral output effects also help to see the implication of an experiment to structural change of the economy. Resource allocation effects are also reflected in factor markets (demand, real wage rates, and income). Changes in factor incomes along with changes in commodity prices affect households’ welfare. Thus, economy-wide analysis will not be completed without analyzing the effects of an experiment on households’ welfare. Households’ welfare is a function of the availability (production) of consumption goods, and the consumption basket mix. Effects on households’ welfare depend on changes in all prices and quantities in an economy (Burfisher, 2011). Therefore, changes in welfare are calculated using equilibrium values of prices and goods in utility, factor demand, and revenue functions (Cardenete et al., 2012). The households’ welfare effects are measured by equivalent variation (EV) that compares the costs of pre-and post-shock levels of household utility valued at the benchmark prices (Burfisher, 2011).35 Such analysis help us to choose the policy alternative that would imply the highest welfare gain. 4.7

Regional projections and analysis

The CGE model results refers to the economy-wide (Ethiopia-wide) effects of each experiments. However, due to the socio-economic heterogeneity of the Ethiopian regions (see section 3.5), one may expect that the economy-wide effects of some experiments (e.g., climate change) to be unevenly shared while some other experiments (e.g., public adaptation policies) may reallocate resources among regions. In spite of the regional variations discussed in section 3.5, however, there are ample reasons to argue that economy-wide representative agents (markets) fairly represent the regional representative agents (markets). The consumption structure of the representative regional households’ exhibits similar pattern to the national average and hovers around it (see Table A3 in the For deriving equivalent variation (EV) from general utility/demand systems (see Varian, 1992), and from the Stone-Geary utility /LES demand system (see Blonigen et al., 1997).

35

46

4. Methodological Framework

Appendix). For instance, the representative households of all regions as well as Ethiopia spend about 70% of their respective total consumption budget on food, non-alcoholic beverages, housing, water, fuel, and energy (HICES, 2011). The retail prices of the majority of commodities, especially of food items, in different regions hovers around the national average (CSA, 2011; HICES, 2011; NBE, 2016). The federal block-grant comprises about 80-95% of regional governments’ recurrent budget except in Addis Ababa (MoFED, 2009). Tax rates across regions are more or less the same (MoFED, 2009). The agro-climatic conditions in Ethiopia are highly controlled by altitude (Hurni, 1998), and thus the regions themselves are agro-ecologically diversified. As such, the Ethiopia-wide production technology can fairly represent regional production technologies of agricultural activities. I assume the same for manufacturing and services. Therefore, I presume that every production activity of the CGE model exhibit similar production technology regardless of its geographical (regional) location. Given these conditions, we can make regional projections and analysis based on the CGE results without losing the methodological elegance, yet, increasing the policy relevance of the study. Top-down, bottom-up, and hybrid approaches are used in the literature to complement economy-wide analysis with regional analysis. 36 What I pursue here is a top-down approach similar to the ORANI Regional Equations System (ORES) for Australia (see Dixon et al., 1982 for more). 37 However, I find it hard to replicate and adopt the ORES-Australia approach to the Ethiopian context. The ORES-Australia requires non-overlapping classification of regional industries (or activities) as ‘national’ or ‘local’ (Dixon et al., 1982; Higgs et al., 1988).38 Such classification allows no single activity to produce ‘local’ as well as ‘national’ commodities which is hardly practical in the Ethiopian context. For instance, Addis Ababa is the hub for auto spare and repair services which is classified as ‘local’ activity in Leontief et al. (1965) and Dixon et al. (1982). In addition, setting all ‘local’ activities as single commodity producers will compromise the elegance of the IFPRI-CGE model that allows one activity to produce one or more commodities. In addition, in the absence of inter-regional trade data, the classification of activities between ‘local’ and ‘national’ is usually augmented with a priori knowledge of technological and institutional reasons (Dixon et al., 1982). For instance, the ORES-Australia used the fact that the majority of the regional population lives in the regional centers which is a sufficient condition to argue that the main consumers are ‘local’ (Dixon et al., 1982; Higgs et al., 1988). However, given the polygons of Ethiopian regions 39 (see Figure 3.2), it is hard to pinpoint the geographic center of each regions. On the other hand, I allowed and calibrated the model such that households can directly consume from economic activities. Or, simply, production for own consumption is allowed. The portion of home-commodities is large in agricultural and real estate and renting business activities (see Table 4.3) which would have fallen in ‘local’ activities (see, for example, Leontief et al., 1965; Dixon et al., 1982). This partly compensates for not separately having ‘local’ activities for each regions as in the ORES-Australia. 36 See Naqvi and Peter (1996) and Higgs et al. (1988) for more on the relative merits and demerits of the three approaches. 37 The ORES approach itself is based on Leontief et al. (1965) to which effect the ORES is also known as the ORANI-LMPST approach. 38 ‘National’ activities produce commodities tradable among regions of the country. The demand for commodities of ‘national’ activities comes from the whole country regardless of the regional location of the activities. ‘Local’ activities, on the other hand, produce commodities which are non-tradable across regions. What is produced is consumed only within the region where it is produced (Dixon et al., 1982; Higgs et al., 1988; Naqvi and Peter, 1996). 39 See also the map at http://www.ethiovisit.com/ethiopia/ethiopia-regions-and-cities.html. Accessed on 30 October 2019.

4.7 Regional projections and analysis

47

Therefore, I consider all economic activities of all regions to be ‘national’ activities. Irrespective of the sales pattern of its output (i.e., where its output is sold & consumed), a regional activity maintains its share in the aggregate (economy-wide or Ethiopia-wide) output of the same activity. Perhaps, the simplest method is to assume that activities (a) in each regions (r) produce a constant portion (s) of the corresponding economy-wide (e) sectoral output (Q) (Naqvi and Peter, 1996). Mathematically, Qra = 𝑠𝑠ar . Qea ………………….………………………… (4.1)

Accordingly, the regional shares (sar ) are exogenous and fixed. As such, the effects of a specific CGE experiment on sectoral output in every region (qra ) is equal to the economy-wide effect (qea ) (Dixon et al., 1982; Higgs et al., 1988). Mathematically, qra = qea …………………………..……………………… (4.2)

The Ethiopia-wide sectoral output effect (qea ) is obtained from the CGE simulations. Equation 4.2 shows that a 1% decrease in the economy’s aggregate sectoral output, for example, of manufacturing leads to a 1% decrease in the output of manufacturing in all regions (Dixon et al., 1982). Then, for each regions, the regional projections involve taking the Ethiopia-wide sectoral output effects as ‘inputs’; weigh by the shares of each industries in the region-wide GDP; and then aggregate. The procedure yields us the regional effects of the CGE experiment. Simply put, by regional projections, I do mean that the effects of a CGE experiment on GDP (valueadded GDP in my case) of the eleven regions of Ethiopia. Unfortunately, to the best of my knowledge, there is no regional data for industrial and regionwide GDP except for Addis Ababa for selected years. I am aware of limited number of micro surveys that report output of selected manufacturing and trade services (e.g., SMIS, 2003; LMIS, 2011; UDTS, 2011). Nevertheless, they are not sufficient to compute region-wide GDP (and hence to make regional comparison) as they mainly cover urban areas. An alternative approach is to use these different micro surveys, though incomplete and inconsistent, to impute and estimate industrial and aggregate output in each regions. This was infeasible, at least, in my study time frame. 40 Therefore, I am compelled to take a remedial measure. I create ‘synthetic’ sectoral and region-wide value-added GDP, for all regions, directly from the SAM. First, I apply a simple rule to disaggregate the Ethiopia-wide sectoral output to obtain regional sectoral output. I find the employment statistics relatively comprehensive and easily modifiable to map with the SAM. 41 In other words, for each activities, I take a regional share in EthiopiaThe exercise, among others, require to harmonize the level of industrial aggregation across the micro surveys, and then, between the micro surveys and the SAM. It also needs to make assumption on missing data, for example, prices and production quantities in rural areas for many of the activities and commodities. 41 My main source of employment data in each regions per industry is NLFS (2005). I make some more adjustments. I use the population and housing census (PHC, 2007) to control for sampling bias in regional labor force reported in NLFS (2005). I use AgSS (2006) to adjust employment among agricultural activities. I use the government expenditure on agriculture and rural development in each regions (MoFED, 2015) to compute regional shares of aggregate public the administration (agriculture, APAGRI) activity. 40

48

4. Methodological Framework

wide employment as proxy to a regional share in Ethiopia-wide sectoral output (sar ). Second, I use Equation 4.1 to obtain regional output by sector (Qra ). That means, I obtain the output value of each of the 17 activities in every region. Third, summing the regional sectoral outputs (Qra ) computed in Step 2, yields region-wide GDP (QrA ). This is given in the second row of Table 4.12 below. Fourth, for each regions, I use Equation 4.3 to compute the share of each industries (war ) in the region-wide GDP. Qr

war = Qra ………………………………………………… (4.3) A

Table 4.12 presents the value-added GDP (sectoral output) for each regions and Ethiopia. Oromia, Southern NNP, Amhara, and Addis Ababa regions are the four largest economic regions. Amhara is the second largest region in terms of total population (PHC, 2007) and hence aggregate employment (NLFS, 2005). However, in terms of value-added GDP, Amhara region takes the third place following the Southern NNP region. This is explained by the fact that the majority of labor in the latter is employed in the production of high-value cash crops such as coffee, tea, and spices compared to the case of the former where the majority is employed in grain agriculture. The share of Addis Ababa in aggregate employment and total population is about 3% which is only half of Tigray and two-third of Somali region (PHC, 2007). However, the economic share of Addis Ababa (11.2%) is larger than both Tigray (5.4%) and Somali (7.5%). Therefore, I argue that what is presented in Table 4.12 is consistent with economic theory and empirical evidence. The economic position of a region improves as the majority of its labor force is employed in high-value economic activities (e.g. cash crops, manufacturing). Table 4.12-Regional value-added GDP CGE Notation GDPFC %

ETH 14.05 100

TIG 0.76 5.4

AFR 0.34 2.5

AMH 2.44 17.3

ORM 4.71 33.6

SOM 1.06 7.5

BNG 0.14 1.0

SNNP 2.79 19.9

GAM 0.09 0.7

HAR 0.06 0.4

ADD 1.57 11.2

DD 0.11 0.8

Source: Author’s calculation based on various data sources Notes: GDPFC is the CGE model notation for total GDP at factor cost which is the sum of value-added GDP of all sectors (values in billion USD). ETH (Ethiopia), TIG (Tigray regional state), AFR (Afar regional state), AMH (Amhara regional state), ORM (Oromia regional state), SOM (Somali regional state), BNG (Benishangul-Gumuz regional state), SNNP (Southern nations, nationalities, and peoples regional state), GAM (Gambella regional state), HAR (Harari regional state), ADD (Addis Ababa city administration), and DD (Dire Dawa city council).

Table 4.13, which I call it as regional module, depicts the economic structure of each regions. 42 Reading down a column shows the economic structure of a region. For instance, grain agriculture (AGRAIN) is the single most important economic activity in Ethiopia, Tigray, Amhara, Oromia, and Benshangul-Gumuz. In contrast, real estate, renting, and business activities (ARSER) and transport and communications (ATRNCOM) are the two most important economic activities in Addis Ababa. Reading down a column also confers information on economic diversification of a region. For instance, a region can be regarded as relatively diversified if I apply the same procedures using HICES (2005) instead of NLFS (2005) as the main source of employment data. The regional economic structure remains more or less similar except for Tigray (see Table A2 in the Appendix). Since the employment in manufacturing as per HICES (2005) is lower than what is reported in NLFS (2005), the regional module based on the former increases the role of agriculture in Tigray region. Despite this, there are no noticeable differences in the rest of regions. One can use both tables (modules) to project regional effects. Nevertheless, I stick on Table 4.13 for all of the empirical chapters since NLFS (2005) is the source of employment data for building the SAM (EDRI, 2009). 42

4.7 Regional projections and analysis

49

more than two sectors have competing shares such as in Addis Ababa, Dire Dawa, Tigray and Afar. In contrast, Amhara is least diversified and agriculture dominated region. In addition, the information presented in Table 4.13 helps to compare the economic structure of a region with another region or Ethiopia. For instance, grain is the dominant economic activity in both Ethiopia and Amhara region. But, the importance is larger in the latter. Table 4.13-Economic structure of regions Activity

ETH

TIG

AFR

AMH

ORM

SOM

BNG

SNNP

GAM

HAR

ADD

DD

AGRAIN

18

21

7

34

21

8

26

13

11

5

0

3

ACCROP

10

2

7

7

12

7

5

19

15

9

0

3

AENSET

1

0

0

0

0

0

0

5

1

0

0

0

ALIVST

14

5

15

12

20

13

3

20

2

1

0

2

AFISFOR

5

0

3

2

2

25

0

8

0

0

0

0

AMINQ

1

2

0

0

0

1

9

0

2

1

0

1

ACONS

4

14

4

5

3

2

15

1

6

5

8

8

AMAN

7

7

12

9

7

3

9

5

7

4

7

4

ATSER

11

9

12

7

12

14

9

12

18

27

12

27

AHSER

2

2

2

2

3

1

3

2

1

1

2

1

ATRNCOM

5

3

8

3

3

6

1

2

8

10

19

22

AFSER

2

1

5

1

1

2

1

1

3

3

6

4

ARSER

8

7

6

7

5

9

2

4

6

4

25

10

APADMN

4

13

9

2

2

4

7

2

9

14

7

4

APAGRI

1

4

2

1

1

1

2

1

2

1

0

1

ASSER

4

5

5

4

4

3

6

3

7

13

7

5

AOSER TOTAL (GDPFC2)

3

4

3

2

3

2

2

1

3

4

7

6

100

100

100

100

100

100

100

100

100

100

100

100

Source: Author’s calculation based on various data sources

The next step is to project the regional effects of a CGE experiment. Recall that the sum of industrial shares in a region-wide GDP (war ) equals to one. Accordingly, we can rewrite the region-wide GDP: ∑ Qra = QrA ………………………………………………… (4.4)

It follows that the regional effects (qrA ) of a specific CGE experiment: ∑ war . qra = qrA …………………………………………… (4.5)

While war (from Table 4.13) captures the importance of a specific industry in the region, qra (from the CGE simulation) captures the direction and strength of a specific CGE experiment on

50

4. Methodological Framework

the sector’s output. Therefore, other things remaining constant, the regional effect depends on the nature of the CGE experiment and the regional economic structure. 43 The approach I pursue here resolves both data availability and consistency problems. It does not require to modify the CGE model and its database. However, it requires a strong assumption to hold. That is labor intensity (or production technology in general) of a specific industry is the same regardless of the administrative region where it is located. Yet, once again, I want to underline that the industrial and hence the region-wide output here are ‘synthetic’ as the values are directly disaggregated from the SAM which itself has undergone scaling and adjustment procedures (EDRI, 2009).

43 One can check the consistency with the CGE results. First, applying Equation 4.5 to Ethiopia (see Equation 4.6), shall give us the aggregate economy-wide effects (qeA - percentage change for GDPFC) simulated by the GCE model. Second, as in Equation 4.7, the sum of regional effects (qrA ) weighted by regional GDP shares (sAr - the second row of Table 4.12) shall give us back the total economy-wide effects (qeA - percentage change for GDPFC simulated by the CGE model). Mathematically, ∑ wae . qea = qeA ………………………………………………… (4.6) ∑ sAr . qrA = qeA ………………………………………………… (4.7)

5 5.1

Impacts of Climate Change Introduction 44

Global warming is now unequivocal (IPCC, 2014). The biophysical and economic impacts of climate change are immediate, negative, and strong in tropical LDCs (Stern, 2007; Cline, 2007; IPCC, 2014). Ethiopia is a case in point where agriculture represents the backbone of the economy. Agriculture employs 75-80% of the country’s labor force (NLFS, 2005; 2013) and contributes to 75-80% of the export earnings and to around 40% of the GDP (NBE, 2016). Nevertheless, agriculture in Ethiopia is yet traditional and virtually rain-fed (CSA-FAO, 2014; NBE, 2015), and less-diversified and dominated by cereal crops (AgSS, 2006; 2014) with smallholder farmers producing about 90-95% of total agricultural output (AgSS, 2015; MoARD, 2010) and about 60% of smallholder farmers’ cereal crop output used for own (household) consumption (AgSS, 2014). As such, the Ethiopian agriculture and economy are predisposed to climate variability and climate change (Conway and Schipper, 2011; FDRE, 2015). This chapter aims to quantify the economy-wide and regional effects of climate change induced shocks in agriculture. The contribution of the chapter is twofold. First, it is a beginning step to formulate national adaptation plans and strategies which in turn help to mobilize adaptation finance. Second, it contributes to place the issue of climate change among the present and future development endeavors of the country. The remainder of the chapter is organized as follows. Section 5.2 reviews the literature related to climate change, agriculture, and migration in the context of LDCs. Section 5.3 narrows down the scope of the literature review to climate change and Ethiopia. Section 5.4 presents the materials and methods of the chapter. Section 5.5 presents and analyzes the economy-wide effects. Section 5.6 projects and analyzes the regional effects. Section 5.7 generalizes the discussion and concludes the chapter. 5.2

Climate change, agriculture, and migration

Production process in agriculture inherently depends on climate (Padgham, 2009). Climate influences soil moisture, the growing period of the plants, and the reproduction and spread of different pathogens that may affect plant, animal, and human health. Agro-climatic conditions also influence the effectiveness of intermediate inputs, like fertilizers, applied to farmlands. As such, climate (and hence climate change) stands out as a primary determinant of agricultural productivity (Adams et al., 1998). Climate change represents changes in temperature, precipitation, carbon dioxide, and solar radiation (IPCC, 2014). It is also expected to increase the frequency, intensity, and duration of extreme weather events such as droughts and floods (IPCC, 2014). The combination will affect soil moisture, water availability, and the incidence and distribution of plant and animal pests and pathogens (IPCC, 2014). These in turn will affect the growth and development of crops (Adams et al., 1998; Hertel and Lobell, 2014), the physiological performance, mortality, 44 Parts of this chapter are merged with parts of chapter 8 and published as Yalew, A.W., Hirte, G., LotzeCampen, H., and Tscharaktschiew, S. (2018). Climate Change, Agriculture, and Economic Development in Ethiopia. Sustainability, 10, 3464. DOI: 10.3390/su10103464.

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 A. W. Yalew, Economic Development under Climate Change, https://doi.org/10.1007/978-3-658-29413-7_5

52

5. Impacts of Climate Change

growth, reproduction, immunity, and production of animals (Adams et al., 1998; Nardone et al., 2010), and the quality and quantity of animal feed (Adams et al., 1998; Thornton et al., 2009).45 Climate change alters agricultural production for a given set of capital, labor, land, and institutions (markets and government policies). 46 Therefore, climate change is often regarded as an analogous to technological change that increases or decreases the total or partial productivity of factors in agriculture (Adams et al., 1998, p.29). Therefore, climate change effects will be stronger in developing regions where agriculture is still primitive. There are also some threshold climatic conditions below or above which agricultural production is hardly possible. Due to the exiting high temperature, for example, a unit change in temperature affects tropical agriculture stronger than temperate zone agriculture (Mendelsohn et al., 2006; Cline, 2007; Müller et al., 2011; Knox et al., 2012). Likewise, the direction and the magnitude of the effects, and the main channel through which climate change affects livestock production system varies across regions and the current production systems. Livestock production in the tropics and subtropics is highly susceptible to indirect effects (Nardone et al., 2010). Smallholders’ livestock production system is vulnerable than the industrial system (Seo and Mendelsohn, 2008; Nardone et al., 2010). Consequently, the observed and projected climate change impede African agriculture (Kurukulasuriya et al., 2006; Cline, 2007; Seo and Mendelsohn, 2008; Schlenker and Lobell, 2010; Nelson et al., 2010; Müller et al., 2011; Knox et al., 2012; World Bank, 2013; Waha et al., 2013; IPCC, 2014; Müller et al., 2014). The effects on agriculture have further repercussions for exports (and hence trade balance) (Jones and Olken, 2010), for economic growth (Dell et al., 2012; Alagidede et al., 2016), for poverty (World Bank, 2013), and hence for the economic prospect of the continent (Stern, 2007). Climate change has also profound influence on human settlement patterns (McLeman and Smit, 2006; Brown, 2008). Climate change induced migration may perhaps be the greatest impact of climate change on the socio-economic system (Brown, 2008 citing IPCC, 1990). Climate change may induce out-migration especially from rural areas by increasing the incidence of climate sensitive diseases (e.g. malaria), by making some places too hot to live permanently, by degrading the asset base of rural households through extreme events (e.g., droughts), by distressing the cropland, grazing land, and water availability, and by making future rural livelihood prospects less predictable and reliable (Brown, 2008). Therefore, climate change is a risk for rural households in developing countries (Mertz et al., 2011; Knox et al., 2012; Kubik and Maurel, 2016). It may impel rural population either to fight (for scare natural resources) or flight from their usual surroundings (UNCCD, 2014). It may lead to dramatic rural-urban drift (Brown, 2008).47 Climate change induced migration may increase pressure on urban infrastructure and services, undermine economic growth, exacerbate risk of conflicts over natural resources, and disrupt the education and social life of the migrants themselves (Brown, 2008; UNCCD, 2014). See Thornton et al. (2009, pp.115-117) and Nardone et al. (2010, pp.58-63) for the channels through which climate change affects livestock production system; and on how these effects influence animal health, production (meat, milk, egg), and reproduction. 46 The set of these given conditions collectively represent the ‘adaptive capacity’ of the sector to climate change. Adaptive capacity refers to the capacity of the production system to deal with the changes in climatic conditions (Smit and Pilifosova, 2001). Adaptive capacity is a function of the existing infrastructure and institutions and hence level of economic development (Smit and Pilifosova, 2001). 47 To account this, Brown (2008, p.15) uses the term “forced climate migrants’’. 45

5.3 Climate change and Ethiopia

53

Of course, historically, migration has been the common way of adaptation to environmental changes and natural disasters (McLeman and Smit, 2006; Brown, 2008; Mbow et al., 2008; Burnham and Ma, 2015). Globally, environmental migrants reached to 25 million in mid-1990s while projected to rise to 50 million (by 2010) and to 200 million by 2050 (Brown, 2008 and references within). The decision to migrate may be driven by an objective of income earnings (as in neoclassical view) or to spread out risks associated with climate change (as in new economics of labor migration) (McLeman and Smit, 2006). Whether the decision to migrate is made at individual or household level, or for temporary or permanent, the majority of environmental change induced migrants will remain as internal migrants (Brown, 2008; Nawrotzki et al., 2015). All said, however, it should be noted here that the climate change-migration nexus (as impact or autonomous adaptation) is very complex (McLeman and Smit, 2006; Brown, 2008). Non-climate drivers like population growth, poverty, and governance remain as critical determinants of the time and scale of climate change induced migration (Brown, 2008; Mertz et al., 2011). Yet, different institutional (legal) and physical (e.g. distance and cultural values), and financial and social resources set constraints on human mobility (Brown, 2008). In addition, migration as adaptation to climate change depends on the availability of other adaptation options (McLeman and Smit, 2006). The likelihood of climate change induced outmigration is expected to be high from rural areas with less diversified economy (which offers low employment opportunities other than agriculture), and with high incidence of poverty (which makes other coping mechanisms impossible). Therefore, Africa is one of such regions where climate change and other environmental change induced migration is expected to rise (Naude, 2010; Mertz et al., 2011). Fifteen out of 38 cases of environmental migration events in recent years occurred in Africa (Naude, 2010 citing Reuveny, 2007). In Tanzania, Kubik and Maurel (2016) finds that weather shock induced decline in agricultural income by 1% increases the probability of migration, on average, by 13 % within the following year. To sum up, climate change imposes risks to agricultural production, particularity, in the subSaharan Africa. This has repercussions to exports, economic growth, and poverty reduction. As the majority of the rural households in the region depend on agriculture, climate change is also expected to trigger outmigration from agriculture and rural areas (Mertz et al., 2011; SalazarEspinoza et al., 2015). 5.3

Climate change and Ethiopia

The national average temperature in Ethiopia has increased by 1°C since the 1960s (FDRE, 2015). It is increasing by 0.37°C per decade (NMA, 2007). The number of hot days and nights in a year is increasing (World Bank, 2016). In contrast, the national average rainfall remained more or less constant although the inter-seasonal and inter-annual rainfall variability have been defining characteristics (NMSA, 2001; NMA, 2007; Cheung et al., 2008; FDRE, 2015). The observed trends can be summarized as increasing annual temperature and hot days and nights but erratic rainfall. Correspondingly, majority of the farmers in different parts of the country perceive that temperature is rising whereas rainfall is decreasing and/or being erratic in their respective villages (e.g., Bryan et al., 2009; Tesso et al., 2012; Tessema et al., 2013; Hadgu et

54

5. Impacts of Climate Change

al., 2014; Kassie, 2014). In some places, farmers perceive that the onset of the rainy season is delaying (Kassie, 2014) and soil moisture is declining (Tesso et al., 2012). The observed trends of climatic conditions in the past five decades have immensely influenced the socio-economic trajectory of the country. Farmers reported declining crop yields in the last 20-30 years (e.g., Teka et al., 2012; Tafesse et al., 2013; Hadgu et al., 2014). Drought is being frequent and unpredictable phenomenon (FDRE, 2015). About fifteen major droughts stroke Ethiopia since 1950 (Ali, 2012; Robinson et al., 2013). More than half of the households in Ethiopia experienced at least one major drought shock between 1999 and 2004 (UNDP, 2007 cited in Robinson et al., 2013). The economic consequences of droughts are paramount. A 10% seasonal rainfall drop below its long-term average implies -4.4% in food production for the country (von Braun, 1991), and -5% in food consumption for rural households (Dercon, 2004). Droughts degrade rural households’ asset such as livestock (von Braun, 1991; Dercon, 2004), usually turn into famines that trigger temporary and permanent migration (Ezra, 2001; Gray and Mueller, 2012), increase the number of people looking for food aid (Ali, 2012), and increase government expenses for food aid and emergency (Robinson et al., 2013). Recent droughts costs about 1% to 4% of the country’s GDP (FDRE, 2015). Moreover, the country’s economic growth trajectory is highly influenced by rainfall availability and variability (World Bank, 2006; Ali, 2012). For instance, if had there no rainfall variability between 1960 and 2008, average Ethiopian income would have been at least four times higher than what it actually was (Ali, 2012). What we can draw from the past experiences is that Ethiopia has very low adaptive capacity to deal with environmental changes. Despite the fact that the projected temperature and rainfall changes in Ethiopia are sensitive to both GHG emission scenarios (World Bank, 2010a) and climate models (Admassu et al., 2013): There is clear agreement on rising mean annual temperature (NMA, 2007; Conway and Schipper, 2011). Even climate models that disagree on precipitation (e.g. MIROC to increase and CSIRO to decrease) predict a warmer future in Ethiopia (Admassu et al., 2013). The number of hot days and nights will continue to rise (World Bank, 2016) while the rainfall projections are uncertain yet (Conway and Schipper, 2011; FDRE, 2015; World Bank, 2016).48 Nevertheless, rainfall in kiremt season (Ethiopia’s main crop growing period) is likely to decrease (NMSA, 2001; World Bank, 2008). Projections in northern Ethiopia substantiates the same (Hadgu et al., 2014). In balance, evaporation and plant transpiration rates may increase in future. This will eventually distress soil moisture and shorten the length of the growing period for crops and grasses (Cline, 2007; Admassu et al., 2013; Hertel and Lobell, 2014). Therefore, climate change pose palpable risks to the Ethiopian agriculture, and with repercussions for the macro-economy. Nonetheless, compared to the importance of the subject, there exist only few studies on the biophysical and economic impacts of climate change in Ethiopia. 49 Kassie (2014) estimates a range of -2 to -29% effects on maize yield in 2050s in the Central Rift Valley. In a dry case scenario, wheat, maize, sorghum, and barely yields may decline by 2 to 5% in 2050s in Ethiopia (World Bank, 2010a). Wheat yield may decline by 28 to 30% in 2070s (NMSA, 2001). Similarly, in 2050s, Admassu et al. (2013) projects decreasing wheat yield throughout the country, and decreasing maize and sorghum yields in major growing areas. World Bank (2008) extrapolate, from global and regional studies, a range of -2.6 to - 5.8% (for wheat yield), and -3.5 to 48 The observed trends and projected changes of climate conditions in Ethiopia are similar to the case for African continent in general and East African region in particular (see IPCC, 2014 for more). 49 I reported here only the case of dry scenario impacts when the study has considered different scenarios.

5.3 Climate change and Ethiopia

55

-7.3% (for maize and coarse grains yields) by 2030. Sugarcane yield may decline by 9% in 2050s (FDRE, 2015). Climate change affects soil moisture and length of growth period which alter the total area suitable for crop cultivation. Admassu et al. (2013) using DSSAT crop model projects net area loss in major wheat, maize, and sorghum growing areas of Ethiopia in 2050s. The net area suitable for growing teff, maize, sorghum, and barley is projected to decline, respectively, by (11%, 17%), (14%, 25%), (7%, 7%), and (31%, 46%) by (2020, 2050) compared to 2000 (Evangelista et al., 2013).50 In short, even though different emission scenarios, climate models, time horizons, and crop models are used, there is consensus on the thesis that climate change is detrimental for crop yield in Ethiopia, and impacts will get worse overtime. Climate change induced crop yield changes can easily be regarded as productivity shocks (Knox et al., 2012) since the use of biotechnologies and irrigation in the country is unsatisfactory (AgSS, 2014; CSA-FAO, 2014; Dercon and Christiaensen, 2011; Wossen et al., 2015). As such, for instance, crop productivity is projected to fall by 1.3% (Maddison et al., 2007) and 7% (World Bank, 2008) due to climate change. Livestock productivity with climate change may be 30% (World Bank, 2010a) and 50% (FDRE, 2015) below the no-climate change scenario. Agricultural production and income decline with falling agricultural productivity. The production of wild coffee may decline by 40 to 90% in 2080s (FDRE, 2015). In the Nile Basin, net farm revenue per hectare may decline by 10 to 303% by 2050 (Deressa and Hassan, 2009). Agricultural GDP with climate change may be 3 to 30% below the baseline – with no climate change – scenario by 2050 (FDRE, 2015). As a result, climate change may increase the number of people looking for food aid by 30% (FDRE, 2015) and drought expenses by 72% (World Bank, 2010a) in 2050s. Some studies go further to assess the economy-wide effects climate change induced productivity shocks in agriculture. Crop productivity shocks in response to +1°C temperature change on GDP may reach to -6% by 2050 (Ferede et al., 2013) and -10% by 2100 (Mideksa, 2010). GDP with climate change will be below the baseline scenario without climate change by 10% by 2050 (World Bank, 2010a), and by 46% by 2030 (World Bank, 2008). Households’ real consumption is also affected by climate change (World Bank, 2008; Ferede et al., 2013). The effects are negative and strong on rural-poor households in arid lowland production zones (Arndt et al., 2011; World Bank, 2010a). Climate change strains economic growth by 2.7% by 2030 (World Bank, 2008), cuts off a quarter of poor households’ income that would have been obtained from baseline economic growth without climate change (Gebreegziabher et al., 2016), and worsens income inequality (Mideksa, 2010). However, the majority of the studies (Deressa and Hassan, 2009; Mideksa, 2010; Gebreegziabher et al., 2016; Ferede et al., 2013) are based on a Ricardian approach. 51 The presumptions of the Ricardian approach–perfect autonomous adaptation and adjustments in farming decision/implementation–are hardly practical in the Ethiopian context (see, for example, Dercon and Christiaensen, 2011; Wossen et al., 2015 and references within). As such, Ricardian 50 The study uses a Maximum Entropy (Maxent) model which is a species distribution model. Figures reported here are an average of three GCMs of A2a SRES scenario of the study. 51 The Ricardian (spatial-analogous) approach applies a cross-sectional econometric methods to estimate the elasticity of crop farm income with respect to a unit percentage increase in temperature. Majority of the studies used the estimates from Deressa and Hassan (2009).

56

5. Impacts of Climate Change

approach studies may understate the impacts of climate change (Adams et al., 1998). The aforementioned studies also focus only on the crop agriculture. Ethiopian livestock sector is virtually smallholder mixed rain-fed production system (Gebremariam et al., 2010; AgSS, 2014; ILRI, 2015). Grazing land and crop residues contribute more than 85% of animal feed (AgSS, 2014). The sector still lacks attention in terms of animal health, nutrition, and breeding policies, regulations, and strategies (MoARD, 2010; ILRI, 2015). Only 26% of the total afflicted livestock population are treated (AgSS, 2015). Hardly 2% of the total cattle and poultry are crossbred (ILRI, 2015). The livestock sector is also less commercialized (Gebremariam et al., 2010). Given these conditions, one can expect that Ethiopian the livestock sector is susceptible to climate change which needs to be incorporated in economic studies. World Bank (2010a), Robinson et al. (2012), and Robinson et al. (2013) have advanced the topic one step further. They pursue a structural approach that blends biophysical crop model (CliCrop) and hybrid livestock model with economic (CGE) model. 52 Their hybrid livestock model partly depends on Seo and Mendelsohn (2008) which, among others, assumed automatic adjustment in the choice of livestock species and in ecosystem responses to climate change. In other words, Seo and Mendelsohn (2008) is a Ricardian study and hence the aforementioned studies share the limitations of the approach. This class of studies also considered only cereals crops. They also applied a dynamic CGE model making it hard to distinguish whether the sign and magnitude of climate change impacts are due to the presumed socio-economic changes (or their uncertain projections) or climate change (or its uncertain projection) per se (Pielke, 2007; FDRE, 2015).53 At last, but not least, this class of the studies gauge uncertainty of climate change impacts through GCMs using single crop model. However, recent studies (e.g., Rosenzweig, et al., 2014) indicate that uncertainty gauged by crop models is wider than uncertainty gauged by GCMs. To the best of my knowledge, there is no attempt to link climate change and migration at macroeconomic level in Ethiopia. This is however topical to Ethiopia. First, in Ethiopia, rural livelihood is inextricably linked to agriculture. Nearly 90% of the total rural labor is employed in agriculture (NLFS, 2005; 2013). About 99% of agricultural labor is engaged in crop and livestock farming (NLFS, 2013) which are highly susceptible to climate change. About 92% of agricultural labor is full time agricultural worker (CSA-EDRI-IFPRI, 2006). About 60% and 20% of annual crop production go, respectively, to households’ own consumption and own seeds (AgSS, 2014). More than 65% of rural households’ income for consumption is obtained from agricultural activities (HICES, 2005; 2011). Under such conditions, even perceived adverse effects on agriculture can trigger outmigration (Massey et al., 2010). Second, Ethiopian population is projected to double by 2050 (FAO, 2015). Per capita agricultural land availability dwindles as population increases (see sections 3.3 and 3.8). The growth of land area under cultivation is already slowing overtime indicating that agricultural land is already getting scarce (CSA, 2015). Lack of land dissuades autonomous adaptation (Burnham and Ma, 2015; Tessema et al., 2013). Declining agricultural land per caput and falling agricultural productivity imply that rural income growth rate will be lower than the rural population growth rate. This further reduces agricultural labor productivity which may exacerbate rural-urban migration. Simply put, climate change will affect employment in agriculture (Nawrotzki et al., 2015). Such mi-

52 Robinson et al. (2012), and Robinson et al. (2013) are based on World Bank (2010a). The three studies also include hydropower and road sectors in addition to agriculture. 53 See also the arguments I provided in section 4.2 in defense of using static model in this study.

5.4 Materials and methods

57

gration would reduce rural labor and hence rural or agricultural output (Lipton, 1980). Migration also affects the labor market and hence wages in receiving regions or sectors. As such, migration either rural-to-urban or agriculture-to-non-agriculture will have economy-wide implications. Therefore, incorporating climate change-induced agricultural labor outmigration will broaden the spectrum of primary effects due to climate change. The present chapter attempts to contribute its own part to the scientific discourse on the subject, particularly, for Ethiopia. It pursues a structural approach. 54 It broadens the number of impacted agricultural activities in Ethiopia. It attempts to gauge uncertainty of climate change impacts using two crop models. The study also benefits from the latest available data on biophysical (crop and grassland yield) impacts of climate change. It also incorporates the issue of climate change- induced agricultural labor outmigration. Most importantly, it complements the economy-wide analysis with regional analysis to portray the impacts of climate change clearer. As such, to the best of my knowledge, the study design is first in its kind to Ethiopia which, I believe, can easily be adapted to other countries in the sub-Saharan Africa. 5.4

Materials and methods

In this study, climate change refers to the change in mean annual temperature and precipitation in future period (2050s, i.e., 2035-2065) compared to the current period (1990s, i.e., 1980-2010) which is consistent with the literature (see, for example, Admassu et al., 2013; World Bank, 2010a; Nelson et al., 2010). In this section, I will discuss the materials and methods with respect to climate change scenario, primary effects of climate change in agriculture, and mapping the primary effects into the CGE model. Climate change scenarios descriptions start with the presumed future global socio-economic trajectories (SSPs-Shared Socioeconomic reference Pathway) 55 that influence the GHG emissions and concertation scenarios (RCPs-Representative Concentration Pathways). 56 Global Climate Models (GCMs) incorporate the emission scenarios and simulate changes in mean climatic conditions. Biophysical impact models take the changes in climatic conditions, process and simulate the biophysical impacts of climate change, for example, on crop yields. 57 Then, economic models (CGE model in this case) model, process, and simulate the economic effects of climate change induced biophysical impacts. Nevertheless, I obtain a set of already processed biophysical effects of climate change from the Agricultural Model Inter-comparison and Improvement Project or shortly AgMIP (see www.agmip.org for more). The AgMIP project processes the biophysical effects of climate change (for SSP-2 under the four RCPs and five GCMs) using seven Global Gridded Crop Models (GGCMs) from mid of the 20th century to end of the 21st century. The methodology and mechanics behind the SSPs, RCPs, GCMs, and 54 Structural approach in climate change impact assessments blends biophysical models with economic models. See Adams et al. (1998) for more on its relative merits and demerits compared to the Ricardian (spatial-analogues) approach. 55 See O’Neill et al. (2017) for more on the five SSPs currently being used in climate change studies. 56 See Wayne (2013) and Moss et al. (2010) for an overview of RCPs currently being used in climate change studies. 57 See White et al. (2011; p. 357) for general discussion on simulations of climate change impacts using crop models.

58

5. Impacts of Climate Change

GGCMs used in the AgMIP project are beyond the scope of this study. Therefore, in the subsequent paragraphs, I only highlight my choice in each steps. Generally speaking, my choice is influenced and meant to be consistent with the recent empirical literature in the field. The AgMIP project is based on SSP-2 which is considered as the ‘middle of the road’ (O’Neill et al., 2017). There are four emission scenarios (RCPs) - RCP8.5, RCP6, RCP4.5 and RCP2.6. Each RCP plot “a different emissions trajectory (pathway) and cumulative emission concentration in 2100” (Wayne, 2013, p.6).58 I pick the case of RCP8.5 which is regarded as high-end emission scenario. Of course, there is no substantial radiative forcing difference among the RCPs in the near-term like in 2050s (Moss et al., 2010). Among the five GCMs used in the AgMIP project, I picked the case of HadGEM2-ES. 59 From the seven AgMIP-GGCMs, I picked the case of Lund-Potsdam-Jena managed Land model (LPJmL) and Environment Policy Integrated Climate (EPIC).60 LPJmL is a global ecosystem based model and the EPIC is a sitebased crop model, and the two models differ with respect to their model structure, parameters, and calibration (see Rosenzweig, et al., 2014 for more). I choose two crop models to gauge uncertainty in biophysical impacts of climate change. 61 The LPJmL and EPIC are applied for large number of crops compared with other GGCMs which is an advantage for an economywide study. I focus on the biophysical impact scenarios with no carbon fertilization effects as the actual benefit of elevated CO2 is substantially below what is expected in general (Nelson, et al., 2010; White et al., 2011; Rosenzweig, et al., 2014; Hertel and Lobell, 2014; Müller and Robertson, 2014; Brookshire and Weaver, 2015), and in Ethiopia in particular (Robinson et al, 2013). The biophysical impacts are with no-irrigation scenario as irrigation in Ethiopia is negligible while it is unlikely to have ‘full-irrigation’ - the alternative irrigation scenario - in the next few decades (AgSS, 2006; 2014; Awulachew, 2010; NBE, 2015). To sum up, I consider two climate change impact scenarios which, for convenience, I will refer as LPJmL and EPIC scenarios hereafter. Due to the combination of the RCP8.5 and HadGEM2ES, one may regard both of the climate change impact scenarios as ‘dry’ or ‘high-end impact’ scenarios. The next step, which is the central objective of this chapter, is to model and assess the economic consequences of the two climate change impact scenarios. 5.4.1 Climate change effects on crop productivity In line with the literature (e.g., Knox et al., 2012; Müller and Robertson, 2014), I consider crop yield changes simulated by LPJmL and EPIC models as climate change effects on crop productivity. The LPJmL and EPIC models simulate for globally important crops at spatial resolution of 0.5o x 0.5o grid (Rosenzweig et al., 2014; Müller and Robertson, 2014). I use the AgMIP-

The numbers refer to radiative forcing for each RCP. Radiative forcing is “a measure of the influence a factor has in altering the balance of incoming and outgoing energy in the Earth-atmosphere system, measured in watts per square meter’’ (Wayne, 2013, p.15). The changes in radiative forcing of RCPs are relative to pre-industrial conditions (Moss et al. 2010). 59 The five GCMs used in AgMIP are the GFDL-ESM2M, HadGEM2-ES, ISPL-CM5A-LR, MIROC-ESMCHEM, and NorESM-M. 60 See Rosenzweig et al. (2014), Müller and Robertson (2014), and Villora et al. (2014) for more discussion on the AgMIP-GGCMs. 61 I have also checked and find that range of yield changes (for a given crop yield) is wider due to crop models than the range implied by RCPs or GCMs corroborating Rosenzweig et al. (2014). 58

5.4 Materials and methods

59

Tool at GEOSHARE (following the procedures given in Villora et al., 2016) to obtain the aggregate mean annual crop yields from 1980 to 2065 for Ethiopia. 62 Climate change effects on crop yields (∆Y c) are computed as the average yield with the future climate, i.e., 2035-2065 (Yfc) compared to the average yield under the present climate, i.e., 1980-2010 (Ypc). ∆Y c =

Ycf −Ycp Ycp

x 100…………………………………… (5.1)

However, some of AgMIP-GGCM crops (e.g., rice and cassava) are economically less important while some others (e.g., potatoes, sugarcane) are not directly represented in the economic accounts of Ethiopia. On the other hand, Ethiopia produces many crops which are grouped into 12 broad crop activities in the original 2005/06 SAM. Some of them (e.g., teff) are local but economically important. When a crop is not directly simulated by a crop model (s), the common practice is to infer from the effects on ‘similar’ crops (Müller and Robertson, 2014). The similarity can be established by the type of photosynthetic pathway (e.g., C3 or C4 crops), main climate zone suitable for the crop (e.g., temperate or tropical), the susceptibility to drought damage, and the economic valuable part of the crop (Bondeau et al., 2007; Müller and Robertson, 2014; Hertel and Lobell, 2014). Accordingly, I first map the AgMIP-GGCM crops, i.e., those crops simulated by LPJmL and EPIC with the Ethiopian-SAM crops on the basis of their photosynthetic pathway and main climatic zone suitable for the crops. Barley, teff, and wheat are ‘cold’ crops as they grow in areas with mild temperature and reliable rainfall. The AgMIP-GGCM soybean and field pea fall in ‘pulses’ crop of Ethiopian-SAM while the AgMIPGGCM groundnuts, rapeseed and sunflower fall in ‘oilseeds’ crop of Ethiopian-SAM. 63 Then, I compute the correlation coefficient between the yields of the ‘similar’ crops using 20 years yield data from CSA (2015). I find high correlation (r ≥ 0.85) among barley, teff, and wheat yields; between the average yield of soybean and field pea, and the average yield of pulses; and between the average yield of groundnuts, rapeseed, and sunflower and the average of yield of oilseeds. The procedures so far provide us yield changes for 7 grain crops of the 12 crop activities/commodities of the original SAM (EDRI, 2009). Next, I impose upper (+30%) and lower (-30%) limits to yield changes due to climate change. This is done to control the sensitivity of variations and model artifacts in case the crop model simulates with very low reference productivity (Nelson et al., 2010; Müller and Robertson, 2014). Finally, I weigh by the shares of the seven crops in total grain area 64, and aggregate which gives an average “grain” yield changes of -10% (LPJmL scenario) and -26% (EPIC scenario). The weighted average yield changes below are rounded up. Recall from section 4.4 that AGRAIN is grain producing activity in the SAM.

62 AgMIP Tool: A GEOSHARE Tool for Aggregating Outputs from the AgMIP’s Global Gridded Crop Modeling Initiative (Ag-GRID). URL: https://mygeohub.org/resources/agmip. Accessed on 24 November, 2015. 63 The Ethiopian-SAM ‘pulses’ include Faba beans, Field peas, Haricot beans, Chick-peas, Lentils, Grass peas, Soya beans, and Fenugreek, Gibto, and ‘oilseeds’ include Neug, Linseed, Groundnuts, Sunflower, Sesame, and Rapeseed (AgSS, 2006; 2014). 64 The weights are calculated based on 2006 main harvest season (AgSS, 2006). The weights change only slightly in 2014 harvest season (AgSS, 2014).

60

5. Impacts of Climate Change

Grain crops account for about 84% of total crop area and 66% of all agricultural land use by smallholder peasants (AgSS, 2014). 65 Table 5.1-Climate change effects on grain yields

Crop

Weight

Unweighted

Weighted

LPJmL

LPJmL

EPIC

EPIC

Teff

0.23

-18

-30

-4

-7

Barley

0.11

-18

-30

-2

-3

Maize

0.15

+30

-19

+5

-3

Sorghum

0.13

-15

-17

-2

-2

Wheat

0.15

-18

-30

-3

-5

Pulses

0.14

-18

-22

-3

-3

Oilseeds

0.09

-8

-30

-1

-3

-10

-26

AGRAIN

1

Source: Author’s calculation

5.4.2 Climate change effects on livestock productivity Climate change impedes the stock of livestock per farm and the livestock species to be reared at each farm (Seo and Mendelsohn, 2008; Nardone et al., 2010). As such, climate change influences livestock production (animal stocks and their products like meat, egg, and milk) and productivity (the value per unit of inputs). Nevertheless, unlike for crops, there is no publicly available physiological model(s) to link climate change and livestock productivity (Robinson et al., 2013; Weindl, et al., 2015). Consequently, I limit myself to the indirect effects of climate change on livestock production (productivity) through feed availability. I arbitrarily assume that from all of the ways through which climate change may affect livestock production (productivity), 30 percent comes through forage quality and quantity. About 87% animal feed (59% from grazing and 28% crop residues) in Ethiopia (AgSS, 2006; 2014) is directly affected by climate change. Then, I specify that climate change induced livestock productivity effects (ΔYL) depend on climate change induced managed grassland productivity (ΔYG) and grain productivity (ΔYC) effects. Implicitly, I assume that grain productivity changes can represent crop residue productivity changes. Also, other things remaining constant, I assume that proportional relationship among grassland productivity-grass feed quantity-livestock productivity: and, among crop residue productivitycrop residue quantity- livestock productivity. ∆YL = 0.3 ∗ (0.59 ∗ ∆YG + 0.28 ∗ ∆YC )……………………………………. (5.2)

Although it is simplistic, the specification to link livestock with crop reflects the Ethiopian context as about 85% of the peasants exercise mixed (crop and livestock) agriculture (AgSS, 2014). The yield changes for grain crops are as in Table 5.1. For grassland yield change, I use All agricultural land use includes: crop area, fallow land, grazing land, wood land and other land uses (AgSS, 2006; 2014).

65

5.4 Materials and methods

61

the same (-4%, simulated by LPJmL) for both LPJmL and EPIC scenarios as it is missing for the latter in my data source (AgMIP-Tool, Villora et al., 2016). The procedure gives us effects of climate change on livestock productivity -2% (LPJmL scenario) and -5% (EPIC scenario). 5.4.3 Climate change effects on agricultural labor migration The national labor force surveys (NLFS, 2005; 2013) and population censuses (ICPS, 2012) show that rural-to-urban migration is increasing overtime. 66 This attributes to increasing population pressure, environmental degradation, low agricultural productivity, and recurrent droughts and famines in rural areas. The thesis is supported by both macro and micro level studies. The macro evidence shows that rural-urban migration, and economic motives of migration (to search job, due to droughts and land scarcity) are increasing overtime. 67 State sponsored resettlement programs are also common at the time of droughts and famines. For instance, in response to the 1984/85 famine about 600 thousands settlers were moved from drought affected areas into areas with reliable rainfall (Rahmato, 1989 cited in Ezra and Kiros, 2001). Micro evidences on migration bluntly points that environmental change has profound effect on rural-urban migration in Ethiopia. 68 Temporary migration has been common form of risk management and coping strategy during droughts and famines (Ezra, 2001; Dercon, 2004). The proportion of labor migration surpasses the non-labor migration (marriage and other social reasons) in drought periods (Ezra and Kiros, 2001; Gray and Mueller, 2012). Taken together, climate change may exacerbate rural-to-urban, at least, agriculture-to-non-agriculture migration in Ethiopia. Such migration may further be intensified due to the increasing rural population and declining per capita agricultural land. However, it is still hard to explicitly distinguish the number of migrants induced by climate change from those induced by non-climate drivers (Brown, 2008). This is particularly true for Rural-urban migration is recently catching up the rural-rural migration which has been the dominant form of migration for decades. Of all recent migrants (those migrants who change their place of residence at least once within five years before the survey), 35% (down from 46% in 2005) were rural-rural while 33% (up from 24% in 2005) were rural-urban migrants (NLFS, 2013). 67 Droughts (and conflicts) and land scarcity were the main reason of migration for about 0.15 million of the 3.8 million total recent migrants (NLFS, 2013). ‘Search for jobs’ was a reason for about one million recent migrants (NLFS, 2013). A third of total migrants in Ethiopia are rural-urban migrants (ICPS, 2012; NLFS, 2013). Ruralurban migration is increasing over time. Between 1999 and 2013, total recent migrants increased by 67% (NLFS, 1999; 2013). In the same period, the number of rural-urban migrants rose by 130% (NLFS, 1999; 2013).The share of recent migrants motivated by job searches in total recent migrants increased from 17.4% in 1999 to 28.6 % in 2013 (NLFS, 1999; 2013). Between 1999 and 2013, recent migrants increased by 110% in urban areas and by 30% in rural areas (NLFS, 1999; 2013). However, in similar period, migrants that mentioned shortage of land as their main reason of migration increased by 240% in urban areas and by 50% in rural areas (NLFS, 1999; 2013). 68 Male migration rate for employment increases from 1.4% in no drought periods to 2.6% in severe drought periods (Gray and Mueller, 2012). About 40% of male and 31% of the total sample migrants from rural areas to Woldiya town, northern Ethiopia, mentioned that ‘famine, poverty, land shortages, and crop failure’ are their motives of migration (Miheretu, 2011). Another study in Damot Galie district, southern Ethiopia, finds that of the total rural out-migrants, 65% relate to looking for additional income, 40% relate to lack of land to make a living on farming, 16% relate to declining agricultural productivity, and 7% relate to food shortage (Gebeyehu, 2015). Based on the discussion with key informants, Gebeyehu (2015) concluded that the major reasons of rural-urban migration in his study area are “population pressure, shortage of land, food insecurity, drought and lack of nonfarming opportunities” (p.92). 66

62

5. Impacts of Climate Change

Ethiopia where the current regulations make a decision to permanently migrate very hard (Dorosh and Schmidt, 2010; Dorosh and Thurlow, 2011). In addition, sale of agricultural land is prohibited by law, people who move from their rural livelihood would lose their entitlement over their farm lands, and proof and registration is required at areas of destination (Dorosh and Schmidt, 2010). In addition, the scant migration statistics in Ethiopia do not provide (neither it is easy to do so) sectoral migration. Therefore, with reference to the macro and the micro evidences, I arbitrarily extrapolate that LPJmL and EPIC climate change scenarios may, respectively, push out one half and one million labor from agriculture. I consider and model ‘migration’ as moving from agricultural to non-agricultural occupations. In particular, I model it as moving from agricultural occupation (FLAB0) to elementary occupations (FLAB3). In the calibrated CGE model context, the migration from FLAB0 to FLAB3 is equivalent to migration from agricultural to non-agricultural sectors. My approach requires less information with regard to the specific sectors and areas of destination. The migration can be intra-rural, inter-rural, and rural-to-urban. Agricultural workers can stay and work in cottage industries such as weaving, tanning, grain milling, own water supply, and the likes in rural areas. The migration can also be temporary or permanent. 5.4.4 Modeling into the CGE model The next critical question is how to incorporate the foregoing first-order shocks in climate change-induced agriculture (crop productivity, livestock productivity, and labor outmigration) into the CGE model. The exercise involves mapping the primary effects with exogenous parameters or variables of the calibrated CGE model. How the productivity effects of climate change are modeled is critical. Other things remaining constant, climate change is modeled as exogenous technological change leading to change in land productivity (e.g., Bezabih et al., 2011; Bosello et al., 2013) or total factor productivity (e.g., Robinson et al., 2012; Wiebelt et al., 2015). I followed the latter approach as the policy implication of partial productivity changes are not clear (Zepeda, 2001; Benin et al., 2011). Therefore, climate change effects on grain productivity (LPJmL= -10% and EPIC= -26%) and livestock productivity (LPJmL= -2% and EPIC= -5%) are modeled as shocks to the shift (efficiency) parameter of the value-added nest of grain (AGRAIN) and livestock (ALIVST) production activities. As per the SAM, the two activities account for 67% of the agricultural GDP and 32% of the total GDP measured at factor cost (see Table 4.1). There is no migration variable (and hence equation) in the standard IFPRI-CGE model. Thus, I model it as an exogenous phenomenon that reduces agricultural labor (FLAB0) supply but increases unskilled labor (FLAB3) supply by the same physical units of labor. The initial supply of agricultural labor is about 25.4 million while that of unskilled labor is about 1.4 million. This gives us shocks to agricultural labor supply (LPJmL= -2% and EPIC= -4%) and to unskilled labor supply (LPJmL = +36% and EPIC = +73%). Therefore, the general equilibrium effects of migration mainly accrue to changes in wage rates and factor substitution effects. We have now a total of six simulation experiments that include two productivity alone (LPJmLP & EPIC-M), two migration alone (LPJmL-M & EPIC-M), and two productivity plus migration (LPJmL-PM & EPIC-PM) scenarios. For convenience, one can consider LPJmL impacts as ‘mild’ dry and EPIC impacts as ‘high’ dry scenario impacts. The subsequent sections discuss the economy-wide and regional effects of these productivity and labor supply shocks in agriculture.

5.5 Economy-wide results and analysis

5.5

63

Economy-wide results and analysis

In this section, I present the effects of climate change on the macro-economy, sectoral output, factor markets, and households’ welfare. 5.5.1 Macro-economy Table 5.2 depicts the macroeconomic effects of climate change. Climate change induced agricultural productivity and labor supply shocks reduce food production (and supply) while increasing food prices. The combination results in declining real private consumption (PRVCON) in a range of -0.3% to -9% which is expected as agriculture is the main food supplier. It also corroborates Arndt et al. (2011) and Dercon (2004) which find households’ consumption in Ethiopia is highly sensitive to climate change and variability. The effects on GDP may reach down to -8%. The decline in total absorption (-2% to -6.5%) is low compared to the decline in GDP indicating that imports may have a potential to reduce the overall impacts of climate change as also argued in World Bank (2010a) and Robinson et al. (2012). Exports of the country decline that may reach to -7% since agriculture is the main source of exports. This impedes the trade balance of the country and hence imports has to fall to maintain the balance. Table 5.2-Macroeconomic effects of climate change Simulations (% change) Notation

Variable

ABSORP

Absorption

-2.1

-6.3

-0.2

-0.4

-2.3

-6.5

PRVCON

Private consumption

-3.0

-8.9

-0.3

-0.6

-3.3

-9.2

EXPORTS

Exports

-2.8

-7.8

-0.2

-0.5

-2.9

-7.3

IMPORTS

Imports

-1.0

-2.8

-0.1

-0.2

-1.0

-2.6

GDPMP

GDP at market prices

-2.6

-7.7

-0.2

-0.5

-2.9

-8.0

GSAV

Government saving

5.8

16.0

7.2

13.4

12.6

26.1

EXR

Real exchange rate

-0.8

-3.6

0.1

0.2

-0.8

-4.3

LPJmL-P

EPIC-P

LPJmL-M

EPIC-M

LPJmL-PM

EPIC-PM

Source: CGE simulations Notes: LPJmL-P (-10% on grain productivity and -2% on livestock productivity), EPIC-P (-26% on grain productivity and -5% on livestock productivity), LPJmL-M (-2% agricultural labor supply and 36% unskilled labor supply), EPIC-M (-4% agricultural labor supply and 73% unskilled labor supply), LPJmL-PM (LPJmL-P plus LPJmL-M), and EPIC-PM (EPIC-P plus EPIC-M).

Table 5.2 shows favorable changes, 6 to 26%, to the government’s saving (GSAV) which stems from the macroeconomic closures. The S-I balance closure rules that the total savings (from government, households, and the ROW) shall adjust to the fixed level of investment. However, foreign savings are fixed (due to external sector closure) while the households’ marginal propensity to save declines with climate change. As such, the savings adjustment burden rests on government which, here, comes through larger decline in government expenditure than government revenue. As such, the positive GSAV changes shall be interpreted with caution as it only implies additional responsibility to the government, i.e., saving burden.

64

5. Impacts of Climate Change

Table 5.2 also shows that the macroeconomic effects of productivity shocks alone scenarios (LPJmL-P & EPIC-P) are by far larger than the case of migration alone scenarios (LPJmL-M & EPIC-M). The macroeconomic effects of the latter are negligible (the change in many macroeconomic variables hardly reaches to -1%). This partly corroborate Dorosh and Thurlow (2011) which argues that rural-urban migration in Ethiopia may be good to the macro economy. 69 5.5.2 Sectoral output Table 5.3 depicts the effects of climate change on value-added GDP of different activities of the SAM. As one would expect climate change hits hard the agricultural activities. The effects may reach to -26% in grain agriculture under the EPIC-PM simulation. The shocks in grain and livestock activities ripple through the rest of agricultural activities (ACCROP, AENSET, and AFISFOR) which is explained by the increasing competition for cropland and agricultural labor which are fixed in supply but mobile across activities (see section 4.5). Accordingly, adding agricultural labor supply shock worsen the effects on agricultural activities. Because agriculture is labor intensive, the effects of migration alone scenarios (LPJmL-M & EPIC-M) on agricultural output are proportional to the agricultural labor supply shocks. This is in contrast with the findings of previous micro studies which find migration has no ‘lost-labor’ effect on agricultural productivity (de Brauw, 2014) and agricultural output (Wondimagegnehu, 2015). Declining agricultural output may affect production in non-agricultural activities where agricultural commodities are part of their intermediate inputs. These include, for example, hotels and restaurants, and construction activities. We see some negative effects on these activities but only with increasing productivity shocks. Likewise, since agricultural commodities contribute for the total marketed output of the economy, we see repercussion to trade activity that may go down to -5%. Declining agricultural output also imply rising agricultural prices that makes the domestic market attractive, relative to foreign markets, for agricultural producers. As a result, agricultural exports decline while agricultural imports increase. As such, to maintain the fixed trade balance of the economy, it requires to increase (to decrease) exports (imports) of nonagricultural commodities. Both require to increase output from some other activities. This is reflected on manufacturing (2.5-13%), transport and communications (1-6%), ‘other’ services (1-13%), and minerals and quarrying (1-5%) activities. Output from these activities increase partly to fill the export gap (e.g. AMAN, ATRNCOM, and AOSER), and partly to fill the gap in domestic demand as a result of falling import variety (e.g., AMAN, AMINQ). The migration scenarios increase the labor supply and reduce the average real wage rate for FLAB3 (see Table 5.4). Consequently, with migration scenarios, activities where the bulk of unskilled labor is employed (e.g. manufacturing, other services) will expand further. In activities where unskilled

However, this does not mean that rural-to-urban or agriculture-to-non-agriculture migration is always beneficiary. For example, I arbitrarily checked a case with 6 million agricultural labor outmigration. I find it to be deleterious to the macro economy, sectoral output, factor markets, and households’ welfare. By implication, the overall economic effects of rural-to-urban or agriculture-to-non-agriculture hinges on the capacity of the destination areas/sectors to absorb the immigrants and expand their production using it. In similar vein, Henderson et al. (2017) argues that climate change induced migration may benefit urban economies in sub-Saharan Africa if the urban areas produce tradable goods.

69

5.5 Economy-wide results and analysis

65

labor employment is non-negligible (e.g., AMINQ, ATSER, and AHSER), adding migration scenarios offset some of the indirect effects due to productivity shocks. Table 5.3-Sectoral output effects of climate change Simulations (% change) Notation

Activity

AGRAIN

Grain crops

-9.3

-24.0

-1.7

-3.4

-10.8

-25.6

ACCROP

Cash crops

-3.8

-13.1

-1.5

-3.1

-5.5

-15.7

AENSET

Enset crop

-3.4

-10.4

-0.8

-1.7

-4.2

-11.6

ALIVST

Livestock production

-4.0

-11.4

-1.3

-2.7

-5.3

-13.6

AFISFOR

Fishing & forestry

-2.5

-8.1

-1.1

-2.2

-3.6

-10.0

AMINQ

Mining & quarrying

0.8

2.2

1.6

3.0

2.4

5.1

ACONS

Construction

0.0

-0.1

0.0

0.0

0.0

0.0

AMAN

Manufacturing

2.5

7.0

3.5

6.7

6.0

13.2

ATSER

Whole sale & retail

-1.6

-5.0

0.0

0.0

-1.6

-4.8

AHSER

Hotels & restaurants

-0.8

-3.2

0.5

0.8

-0.4

-2.6

ATRNCOM

Transport & comm.

1.2

4.0

0.7

1.4

2.0

5.6

AFSER

Financial intermediaries

0.2

0.5

0.8

1.4

0.9

1.9

ARSER

Real estate & renting

0.0

0.0

0.1

0.2

0.1

0.2

APADMN

Public admin. (general)

0.0

0.0

0.0

0.0

0.0

0.0

APAGRI

Public admin.(agriculture)

0.0

0.0

0.0

0.0

0.0

0.0

ASSER

Social services

0.1

0.1

0.3

0.6

0.4

0.7

AOSER

Other services

1.1

3.3

5.1

9.8

6.2

13.5

GDPFC

Total GDP at factor cost

-2.7

-7.6

-0.2

-0.5

-2.9

-7.6

LPJmL-P

EPIC-P

LPJmL-M

EPIC-M

LPJmL-PM

EPIC-PM

Source: CGE simulations Notes: LPJmL-P (-10% on grain productivity and -2% on livestock productivity), EPIC-P (-26% on grain productivity and -5% on livestock productivity), LPJmL-M (-2% agricultural labor supply and 36% unskilled labor supply), EPIC-M (-4% agricultural labor supply and 73% unskilled labor supply), LPJmL-PM (LPJmL-P plus LPJmL-M), and EPIC-PM (EPIC-P plus EPIC-M).

To conclude, it seems that the negative effects of climate change are contained in agriculture than spreading to the rest of economic sectors. This attributes to the low factor reallocation effects, the weak interindustry linkage, and the low demand elasticities in calibration. 70 Nevertheless, revealing the contribution of the sector, the total GDP at factor prices (GDPFC in Table 70 Factor reallocation effects are apparently weak as the vast majority of agricultural factors of production (FLND, FTLU, and FLAB0) are used only in agriculture. The backward and forward interlinkages between agriculture and non-agricultural activities is weak since the majority of agricultural output is consumed at home while its main intermediate input (i.e. fertilizer) is entirely imported. Nor shall we expect strong effects through relative commodity prices change. The income elasticities are low as Ethiopia is low-income country (see Table 4.11) while

66

5. Impacts of Climate Change

5.3) and GDP at market prices (GDPMP in Table 5.2) are highly shaped by impacts on agricultural GDP. Migration may squeeze production in agriculture but expand production in manufacturing and services. 5.5.3 Factor markets Shocks in agriculture make the two agricultural factors relatively scarce and increase their wage rate and factor income relative to other primary factors (see Table 5.4). The average wage rate and factor income of agricultural labor may, respectively, increase from 5 to 28% and 3 to 17%. The average wage rate and factor income of cropland, respectively, increases from 6 to 21% and 4 to 13%. The case for livestock factor (FTLU) is different and needs further explanation. FTLU is livestock-specific factor, thus, its economy-wide wage rate and employment remains unchanged. Despite this, Table 5.4 shows that FTLU factor income declines due to climate change. This is explained by falling activity-specific wage (or marginal revenue product of the factor) due to falling livestock output (see Table 5.3). In contrast, in general, the average wage rate and factor income for non-agricultural factors decline. This may be explained by declining relative prices of non-agricultural commodities that reduce the marginal revenue product of different non-agricultural factors. Under all simulations, average wage rates of administrative workers, professionals and technical associates, and skilled labor decline. Their demand increase in some activities but decrease in some other activities. In balance, the factor income of the aforementioned factors declines in all simulations and reaching to -25% in high-impact scenario, i.e., EPIC-PM. The unskilled labor market faces similar consequences in productivity alone shock scenarios. However, as it is the occupation of destination for emigrants from agriculture, its economy-wide rate declines further down to 45% under EPIC-PM scenario. In parallel, however, employment for unskilled labor shall increase in order to satisfy the factor market equilibrium constraint. In balance, the total factor income for unskilled labor increases under migration alone scenarios while its decline is dampened when migration is combined with productivity shocks. Capital is activity-specific in demand and fixed in supply. Therefore, the decline in capital’s income (-2 to -22%) in simulations accrue to its declining weighted marginal revenue product of activities employing it. To conclude, climate change alters the relative prices of commodities and value-added output of different activities which in turn affect the wage and demand of different factors. The combined effects, alter income of different factors which will have strong effect on households’ real consumption as discussed in the subsequent subsection.

the LES demand system implies low own-price and cross-price elasticities and assumes commodities are gross complementary (de Boer and Missaglia, 2006).

5.5 Economy-wide results and analysis

67

Table 5.4-Factor market effects of climate change Simulations (% change)

Factor Income (YF)

Economy-wage (WF)

Variable

Notation

Factor

FLAB0

Agricultural labor

FLAB1

LPJmL-P

EPIC-P

LPJmL-M

EPIC-M

LPJmL-PM

EPIC-PM

4.9

17.2

4.5

9.1

9.8

27.6

Administrative labor

-5.9

-18.6

-3.4

-6.5

-9.4

-24.4

FLAB2

Professional labor

-5.8

-18.2

-3.6

-6.9

-9.4

-24.2

FLAB3

Unskilled labor

-5.5

-17.5

-19.9

-32.9

-24.4

-44.8

FLAB4

Skilled labor

-6.8

-21.0

-2.4

-4.7

-9.3

-25.3

FLND

Cropland

6.0

20.9

0.3

0.4

6.4

21.0

FLAB0

Agricultural labor

3.4

11.4

2.5

4.9

6.0

16.8

FLAB1

Administrative labor

-6.0

-18.6

-3.9

-7.4

-9.8

-24.9

FLAB2

Professional labor

-5.8

-18.3

-4.1

-7.7

-9.8

-24.9

FLAB3

Unskilled labor

-4.1

-14.2

7.7

13.8

3.1

-3.4

FLAB4

Skilled labor

-6.2

-19.1

-2.9

-5.5

-9.1

-24.1

FLND

Cropland

3.8

12.6

0.2

0.3

4.1

12.7

FTLU

Livestock Units

-4.7

-15.4

-2.0

-4.2

-6.7

-17.7

FCAP

Capital

-5.8

-18.2

-2.0

-4.0

-7.9

-21.8

Source: CGE simulations Notes: LPJmL-P (-10% on grain productivity and -2% on livestock productivity), EPIC-P (-26% on grain productivity and -5% on livestock productivity), LPJmL-M (-2% agricultural labor supply and 36% unskilled labor supply), EPIC-M (-4% agricultural labor supply and 73% unskilled labor supply), LPJmL-PM (LPJmL-P plus LPJmL-M), and EPIC-PM (EPIC-P plus EPIC-M).

5.5.4 Households’ welfare Climate change effects on commodities supply and prices, and factors demand and wages eventually influence households’ welfare. Figure 5.1 depicts the household welfare effects of climate change measured by EV as percentage of the benchmark consumption spending of each household groups. Due to increase in factor income of agricultural labor and land (see Table 5.4) which constitute more than half of the total rural households’ income (see Table 4.8), the rural households’ total income shall increase. Nonetheless, it is not able to compensate real consumption lost due to increasing agricultural (food) prices. As a result, rural households’ real consumption (welfare) falls down that may reach to -10% in the EPIC-PM scenario. In contrast, the urban households’ wellbeing is affected from both side–declining factor incomes (of FLAB1, FLAB2, FLAB4, and FCAP) and increasing food prices. Consequently, the welfare effects to urban households are worse than that of rural households (see Figure 5.1). In addition, Figure 5.1, shows that adding migration scenarios on productivity shock worsen the welfare loss of both households especially that of urban households.

68

5. Impacts of Climate Change

Figure 5.1-Households’ welfare effects of climate change

EQUIVALENT VARIATION (%)

-8.5

EPIC-P

EPIC-PM

-9 -9.5 -10

RURH URBH

-10.5 -11 -11.5

Source: CGE simulations Notes: EPIC-P (-26% on grain productivity and -5% on livestock productivity), and EPIC-PM (EPIC-P plus -4% agricultural labor supply and 73% unskilled labor supply).

In interpreting the households’ welfare effects, one shall also consider the role of the low income, own-price and cross-price elasticities of demand used in the calibration (see Table 4.11). Recall also that I have set Frisch parameters which impose minimum (mandatory) consumption for both households. As such, households’ consumption spending is less sensitive to income and commodity price changes. The households’ welfare effects are also robust to ±25% of the elasticities presented in Table 4.11 (see Table A4 in Appendix). Therefore, implicitly, the results show that Ethiopian households compromise their savings and transfers they make to other institutions for the sake of maintaining their subsistence consumption during harsh times. This corroborates the findings of micro studies (e.g., von Braun, 1991; Dercon, 2004) which find that asset selling is the common method of consumption smoothing during drought periods. 5.6

Regional projections and analysis

The regional projections involve combining sectoral output effects (Table 5.3) with the regional module (Table 4.13). Regional effects presented in Table 5.5 below refer to the weighted effects on the region-wide value-added GDP. The regional effects of climate change are uneven, and range between -4.1% and 1.1% (LPJmL-PM), and between -10.3% and 2.4% (EPIC-PM). Climate change effects are adverse and strong in the three largest agrarian regions–Oromia, Amhara, and Southern NNP. Under each experiments, the effects in these regions are larger than the national effects as well as the effects on the rest of the regions. Conversely, climate change effects are relatively low in urbanized regions like Addis Ababa, Dire Dawa, and Harari. The changes in Addis Ababa city are favorable under all climate change scenarios. This is expected as manufacturing, transport and communication, and other services are the main economic sectors in Addis Ababa (see Table 5.3, Table 4.13, and Table 3.2).

5.6 Regional projections and analysis

69

Table 5.5-Regional effects of climate change Simulations (% change) Region

LPJmL-P

EPIC-P

LPJmL-M

EPIC-M

LPJmL-PM

EPIC-PM

Ethiopia

-2.7

-7.6

-0.2

-0.5

-2.9

-7.6

Tigray

-2.1

-5.7

0.1

0.1

-2.0

-5.1

Afar

-1.4

-3.9

0.3

0.5

-1.1

-3.2

Amhara

-3.8

-10.2

-0.4

-0.8

-4.1

-10.3

Oromia

-3.2

-9.1

-0.4

-0.8

-3.6

-9.4

Somali

-2.2

-6.6

-0.4

-0.8

-2.6

-7.0

Benishangul-Gumuz

-2.6

-6.8

0.0

0.0

-2.5

-6.4

Southern NNP

-3.2

-9.3

-0.6

-1.3

-3.8

-10.1

Gambella

-1.7

-4.9

0.1

0.1

-1.6

-4.5

Harari

-1.0

-3.0

0.3

0.4

-0.8

-2.4

0.2

0.8

0.8

1.6

1.1

2.4

-0.5

-1.4

0.6

1.0

0.1

-0.1

Addis Ababa Dire Dawa

Source: Table 5.3 and Table 4.13 Notes: LPJmL-P (-10% on grain productivity and -2% on livestock productivity), EPIC-P (-26% on grain productivity and -5% on livestock productivity), LPJmL-M (-2% agricultural labor supply and 36% unskilled labor supply), EPIC-M (-4% agricultural labor supply and 73% unskilled labor supply), LPJmL-PM (LPJmL-P plus LPJmL-M), and EPIC-PM (EPIC-P plus EPIC-M).

Another important point to look at Table 5.5 is the difference between the regional effects between productivity alone scenarios (LPJmL-P and EPIC-P) and productivity plus migration scenarios (LPJmL-PM and EPIC-PM) for each regions. Or, alternatively, look at migration alone scenarios (LPJmL-M and EPIC-M) across regions. Accordingly, regions with significant shares of non-agricultural industries (e.g., Addis Ababa, Dire Dawa, Harari, and to some extent Tigray and Afar) may benefit from occupational migration. Conversely, occupational migration may hurt Southern NNP, Amhara, Oromia, and Somali regions. This implies that the size of non-agricultural activities in these regions have no the capacity to absorb and take the opportunity from agricultural labor migration. By implication, as no constraints are imposed on interregional migration, the out-migrants from agriculture in these regions may end in manufacturing and services sectors in urbanized regions. The national migration statistics corroborates my arguments. The net migration rate per 1000 people is positive and high in Addis Ababa (430) and Dire Dawa (289) but negative in Amhara (-64) and Southern NNP (-27) (ICPS, 2012). Therefore, regional economic diversification may help to trap and benefit from climate changeinduced agricultural labor outmigration in the same region. This explains the case of Tigray, Afar, and Harari regions where there are some sectors, other than grain and livestock, which have important contribution to their respective region-wide value added GDP (see Table 4.13). As a result, in these regions, occupational migration tends to dampen productivity alone effects. In conclusion, the results show that the regional effects of climate change depend on the regional economic structure (and relative to the national economic structure). For instance, even though Tigray is an agrarian region, the role of grain is relatively lower than the case of Ethiopia

70

5. Impacts of Climate Change

(see Table 4.13). It follows that value-added GDP effects in Tigray are relatively lower than in Ethiopia. The regional effects also show that economic diversification within a region may help to absorb and benefit from agricultural labor outmigration. Otherwise, climate change may take productive labor into other regions where manufacturing and services are the main economic stays. 71 5.7

Conclusions

This chapter assesses the economy-wide and regional effects of climate change induced productivity and labor supply shocks in agriculture. The economy-wide analysis shows that climate change reduces agricultural output, increases agricultural prices, alters the trade mix, and profoundly affects households’ welfare. The effects of climate change on aggregate GDP (households’ welfare) resemble that of the effects on agricultural GDP (rural households’ welfare). The macroeconomic effects of the presumed migration scenarios are negligible, but, the effects on agricultural production are proportional to the agricultural labor supply shocks revealing that the country’s agriculture is labor intensive. In general, the ripple effects of climate change on non-agricultural sectors are not as expected. This attributes to the structural features of the economy which include weak interindustry linkage, using the bulk of agricultural output for own rural household consumption and seed, low income and price elasticities, almost nil factor competition between agriculture and nonagricultural sectors. The regional analysis shows that regional effects of climate change are uneven. The regional effects are negative and stronger in regions such as Amhara, Oromia, and Southern NNP where agriculture is the main livelihood. In contrast, the regional effects are positive in Addis Ababa and less adverse in other urbanized regions such as Dire Dawa and Harari. The regional effects show that diversifying regional economies may help to turn climate change-induced agricultural labor outmigration into opportunity. Otherwise, outmigration from agriculture may widen regional disparity and impair the economic prospects of both sending regions (loss in productive labor) and receiving regions (decline in real wage rates and pressure on infrastructure). Even though the materials and methods used in the chapter are different, the findings and general conclusions substantiate the findings of previous studies related to the topic in Ethiopia (see, for example, World Bank, 2008; 2010a; Arndt et al., 2011; Robinson et al., 2012). Therefore, it is important to build on their conclusions along with the findings of this chapter, and to underscore that impacts are enough to worry about and to call for public action. Belated adaptation responses in the sector may leave narrow window of opportunity to adapt in later stages (Müller et al., 2014), and hamper the economic development of the country (Millner and Dietz, 2015). Therefore, I examine the economy-wide and regional effects of ‘full’ public adaptation in agriculture (in Chapter 6), and of alternative adaptation finances (in Chapter 7). On the other hand, the economy-wide and regional analysis of this chapter indicates that structural change in the country may dampen the adverse consequences of climate change. This go with Henderson et al. (2017) which concludes that structural transformation would facilitate better response to climate change in Africa. Accordingly, the government’s current economic growth and transformation plans (MoFED, 2010; NPC, 2016) may contribute to the resilience of the economy to climate change. We will come back to this point in Chapter 8. I would like to remind the reader that the direct and indirect costs of migration induced pressure on receiving regions are not accounted here. If such costs are high, which probably may be, climate change induced migration may have negative consequences for both regions of origin and regions of destination.

71

6 6.1

Costs of Planned Adaptation Introduction 72

In Chapter 5, we see that climate change dims the economic prospect of Ethiopia. It may strain the country’s ability to achieve sustained economic growth, and to reduce poverty. Cognizant of this, government of Ethiopia has already put planned public adaptation in agriculture as its primary response to climate change (FDRE, 2011; 2015). This will, however, bring incremental budget burden to the public sector which depends on the type, extent, and timing of adaptation measures. Nevertheless, in this chapter, I will focus on a bundle of proactive adaptation measures that aim to fully neutralize the climate change induced productivity shocks in agriculture (see section 5.4.1 and 5.4.2). The chapter intends to derive the direct costs of this planned ‘full’ adaptation in agriculture, and to examine their economy-wide and regional effects. The rest of the chapter is organized as follows. Section 6.2 briefly discuss the concept and approaches of adaptation with special emphasis to planned public adaptation. Section 6.3 reviews the literature related to climate change adaptation in Ethiopia. Section 6.4 presents the materials and methods of the chapter that include defining the bundle of adaptation measures, deriving the direct costs of adaptation and modeling into the CGE model. Section 6.5 analyzes the economy-wide results followed by section 6.6 that analyzes the regional projections. Section 6.7 concludes the chapter. 6.2

Planned public adaptation to climate change

Adaptation to climate change refers to “adjustments in ecological, social, or economic systems in response to actual or expected climatic stimuli and their effects” (Smit and Pilifosova, 2001, p.879). Adaptation involves changes in process, practices, and structures with the aim to moderate potential damages (or take opportunities, if any) associated with climate change (IPCC, 2007). Adaptation in any sector is important since the global warming over the next three decades is unavertable even if the world is able to curtail its GHG emission today (UNFCCC, 2009; Hertel and Lobell, 2014). Adaptation can be classified on the basis of scope, purpose, form, and time (Smit et al., 1999). Its scope may be local, regional, or national; its purpose may be to accommodate, to prevent, or to restore from climate change induced damage; its form may be institutional, regulatory, financial, structural, or technological; and its temporal scope may be short-term or long-term (Smit et al., 1999; Smit and Pilifosova, 2001). Adaptation can be proactive (undertaken before the impacts are observed) or reactive (undertaken when the impacts are experienced) (IPCC, 2007). It is also classified as autonomous (automatic or spontaneous) or deliberate (planned or policy) adaptation (Smit et al., 1999; IPCC, 2007). Autonomous adaptation refers to a nonconscious response to climate change triggered by ecological, market, and welfare changes (IPCC, 2007). In contrast, planned adaptation is a result of a deliberate policy decision “based 72 Parts of this chapter are merged with parts of chapter 7 and published as Yalew, A.W., Hirte, G., Lotze-Campen, and Tscharaktschiew, S. (2019). The Synergies and Trade-Offs of Planned Adaptation in Agriculture: a General Equilibrium Analysis for Ethiopia. Economics of Disasters and Climate Change, 3:213–233. DOI: 10.1007/s41885-019-00041-3.

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 A. W. Yalew, Economic Development under Climate Change, https://doi.org/10.1007/978-3-658-29413-7_6

72

6. Costs of Planned Adaptation

on an awareness that conditions have changed or are about to change and that action is required to return to, maintain, or achieve a desired state” (IPCC, 2007, p.869). Accordingly, planned adaptation consists of four important elements. First, it involves deliberate policy decision. Second, the policy decision is stimulated by awareness of actual or anticipated climate change. Third, the policy decisions shall turn into actions, i.e., adaptation shall be practical. Fourth, there shall be aim(s) that adaptation shall target. For observed impacts, the aim of adaptation may be to return to the state before the impacts. Whereas, for projected impacts, the aim of adaptation may be to maintain the status quo or move into another desired state. Deliberate and planned actions are required, especially, when and where autonomous adaptation is inadequate to cope up with the anticipated impacts of climate change. This is usually true in LDCs where individual agents lack the information, finance, and technology necessary to deal with climate change and variability (Antle and Capalbo, 2010). Public adaptation primarily aims to fill these missing adaptive capacities. It also aims to aware, trigger, and facilitate automatic adaptation by individual agents. Besides, when there are multiple beneficiaries of the adaptation measures, then, government action is the only efficient way to undertake joint adaptation (Mendelsohn, 2000). Planned public adaptation efforts entail measures that are more or less in economic development toolbox (McGray et al., 2007), therefore, it is inextricable from other endeavors for economic development (Padgham, 2009; Fankhauser and Schmidt-Traub, 2011). To put it differently, public sector responses to climate change are inseparable from responses to other development problems such as population growth, urbanization, or globalization (McGray et al., 2007). As such, adaptation to climate change in LDCs shall be integrated and mainstreamed into their wider development policy and practices (Fankhauser and Schmidt-Traub, 2011; Tanner and Horn-Phathanothai, 2014). By mainstreaming adaptation, we mean that building adaptation measures upon the existing sectoral and macroeconomic institutions, policies, and practices (Padgham, 2009; Kissinger, et al., 2013; Tanner and Horn-Phathanothai, 2014). This eases the process to identify and quantify the conflicts and synergies of adaptation in a single sector with the rest of economic endeavors of a country. 6.3

Adaptation to climate change in Ethiopia

Climate change adaptation and adaptation planning are topical in Ethiopia (FDRE, 2011; 2015; 2016). Ethiopia is one of the few LDCs that submitted, in time, both the Initial National Communication (NMSA, 2001) and National Adaptation Programme of Actions (NAPA) (NMA, 2007) to the UNFCCC. The country launched its Climate-Resilient Green Economy strategy (CRGE) in 2011(FDRE, 2011), and Climate Resilient Strategy for Agriculture and Forestry (FDRE, 2015). The national, regional and sectoral adaptation plans in Ethiopia attempt to assess the country’s vulnerability to the current weather variability and future climate change (NMSA, 2001; NMA, 2007; FDRE, 2015; 2016). They also identify and prioritize adaptation options, and, in some cases, attempt to estimate the direct costs of selected adaptation options (NMA, 2007; FDRE, 2015). The government documents place adaptation in agriculture as a prime and urgent response to climate change in Ethiopia (NMA, 2007; Conway and Schipper, 2011; FDRE, 2011). It is underlined that “a major shift is needed to ensure that climate resilient actions in agriculture are implemented” (FDRE, 2015, p.7). Giving priority to adaptation in agriculture is justifiable, at

6.3 Adaptation to climate change in Ethiopia

73

least, in three main grounds. First, the sector is already vulnerable to the observed climate change trends in the second half of the 20th century (see section 3.8 and 5.3). Second, agriculture is the main source of employment, income, and export earnings (see section 3.3 and section 4.4) to which effect we see that climate change effects on agriculture impede the macro-economy (see Table 5.2 and Figure 5.1). Third, despite the favorable institutional and budgetary support, the growth in the sector has been unsatisfactory in the past decades (BMGF, 2010; Mitik and Engida, 2013). However, farmers in different parts of Ethiopia indicate that lack of climate information, extension services, irrigation, and financial capacity inhibit them to undertake adaptation by their own (see, for example, Hadgu et al., 2014; Tessema et al., 2013; Tafesse et al., 2013). By implication, government intervention is necessary for meaningful adaptation in agriculture. Accordingly, the CRGE Vision (FDRE, 2011, p.4) explicitly puts the government will take the lead in responding to climate change. Among others, government’s role in expanding irrigation, agricultural research and extension, and rural road density is indispensable. In addition, government at different levels can inform, induce, and facilitate autonomous adaptation. Planned public adaptation in agriculture entails costs of planning and implementing. For instance, in the NAPA, the design and implementation of ten adaptation projects in agriculture may cost about USD 767 million (NMA, 2007). Agricultural adaptation that include irrigation, R&D, draining and watershed management requires an extra budget of USD 68–71 million per annum (Robinson et al., 2013). Between 2008 and 2012, adaptation relevant expenses account for about 87% of the climate relevant expenditure (which is around 15% of the total government expenditure and 1.6% of the GDP) (Eshetu, et al., 2014). The bulk of those adaptation relevant expenses are related to adaptation in agriculture (Eshetu, et al., 2014; MoFED, 2016). 73 Delivering forty one agricultural adaptation measures identified and prioritized in FDRE (2015) may require an additional investment of about USD 236 million by 2030 from federal government alone that may require to uplift the baseline sectoral budget by about 18% (FDRE, 2015). Yet, the federal government expense should be supplemented by about USD 367 million per year investment from non-federal government sources (regional governments, donors, and private sector). Taken together, by 2030, Ethiopia requires a sum of USD 600 million (at 2008 prices) budget to build resilient agriculture to climate change on top of the baseline sectoral budget– which is USD 2.9 billion (FDRE, 2015, p.8).74 Of this sum, about 80% of the projected incremental budget is expected from the public sector budget (FDRE, 2015). The opportunity costs of the incremental budget demand for adaptation in agriculture are immense, especially, compared to the country’ level of development, public budget scarcity, and strong drive to expand transport, energy, and urban infrastructure. Agricultural adaptation policy making process is further complicated with the time lag between adaptation costs (present and certain) and adaptation benefits (future and uncertain). Under such conditions, an ex-ante evaluation of public adaptation in agriculture is crucial. In particular, it requires to examine the general equilibrium and regional effects of adaptation costs. So doing helps to get the glimpse of the type, size, and strength of the indirect effects of public adaptation costs in agriculture. It 73 CRGE Fast Track. URL: http://www.mofed.gov.et/English/Featured%20Articles/Pages/TheCRGEFastTrackInvestments.aspx. Retrieved on 06 January, 2016. 74 Even the fifteen urgent adaptation measures require a sum of USD 132 million (at 2008 prices) from federal and regional governments budget (FDRE, 2015).

74

6. Costs of Planned Adaptation

helps policy makers to identify and be prepared for compensation mechanisms in advance. It also hints the possible conflicts and synergies of planned adaptation in agriculture with the current and future economic plans of the country. The questions are topical as the country is at early stage of planning adaptation. However, neither the previous micro case studies nor policy documents attempt to address these appealing questions to Ethiopia. The class of case studies (e.g., Tessema et al., 2013; Hadgu et al., 2014; Berhanu and Beyene, 2015) examine farmers’ behavior and capacity to adapt to climate change and underlined the need of public action. The class of national climate change documents (e.g., NMSA, 2001; NMA, 2007; FDRE, 2011; FDRE, 2015) focus on institutional aspects of adaptation in general and in agriculture in particular. Costs of adaptation in agriculture are scantily researched with the exception of NMA (2007), World Bank (2010a), Robinson et al. (2013), and FDRE (2015). Cost estimates in NMA (2007) are based on limited project type measures which are difficult to integrate with the general macroeconomic framework. FDRE (2015) estimates the direct costs of adaptation in agriculture leaving out the indirect effects to the rest of the economy. The World Bank (2010a), Robinson et al. (2012), and Robinson et al. (2013) attempt to address the general equilibrium effects. However, the studies do not make it explicit as to how to mobilize the required resources which is critical, especially, as the likelihood of using domestic sources of finance is high. I will turn more to this point in Chapter 7. In addition, regional effects may matter since not every region benefit equally from the increased public spending to adaptation in agriculture. This chapter attempts to fill these research gaps. It attempts to derive the direct costs of planned public adaptation in agriculture as a function of anticipated productivity shocks and the elasticity of agricultural productivity with respect to public spending. Such approach helps to gauge uncertainties that stem from climate change impacts as well as adaptation policy effectiveness both of which are important issues, and shall be considered in designing adaptation policy (Heal and Millner, 2014; Tanner and Horn-Phathanothai, 2014). It, then, examines the economy-wide and the regional effects of public spending to fully neutralize the anticipated agricultural productivity effects of climate change. 6.4

Materials and methods

Generally, a list of specific adaptation measures follows from the sign and the size of anticipated biophysical impacts and the aim of adaptation in the sector of interest. This chapter focuses on the case of planned public adaptation that fully offset the anticipated agricultural productivity shocks discussed in section 5.4. 75 Government of Ethiopia is presumed to be in charge of planning and undertaking the set of deliberate adaptation measures. Adaptation in agriculture is assumed to be part of the country’s development endeavors. Therefore, adaptation shall be aligned to the existing agricultural development policies, and public budget account. As such, planned public adaptation measures can be regarded as additional public services to enhance efficiency in agricultural activities.

75 In other words, adaptation in this chapter is in response to anticipated rising temperature and unpredictable rainfall (Conway and Schipper, 2011; IPCC, 2014; FDRE, 2011; 2015; Admassu et al., 2013), rising evapotranspiration and decreasing soil moisture (Admassu et al., 2013), and hence falling crop and livestock productivity (World Bank, 2010a; Chapter 5 of this study) in Ethiopia.

6.4 Materials and methods

75

6.4.1 Adaptation measures I begin with a rigorous literature review to make initial list of agricultural adaptation measures (e.g., Smit and Skinner, 2002; Kurukulasuriya and Rosenthal, 2003; Padgham, 2009). However, not every measure (such as adjustments in sowing and planting dates) can directly be undertaken by the public sector. Then, with reference to the farm level case studies which represent different parts of the country (e.g., Tessema et al., 2013; Kassie, 2014; Berhanu and Beyene, 2015), I identify and focus on measures which are beyond farmers’ autonomous adaptive capacity. However, not all measures can easily be integrated with the rest of the national sectoral and macroeconomic plans. Therefore, in the third step, I review a set of government sectoral and macroeconomic policies in general (e.g., MoFED, 2010; MoARD, 2010) and climate change and adaptation reports/plans in particular (e.g., NMSA, 2001; NMA, 2007; FDRE, 2011; FDRE, 2015). I further review various reports (e.g., various years’ annual agricultural sample survey reports) and research papers (e.g., BMGF, 2010 and background reports for it) for the gaps in implementing agricultural policies in the country. However, the costs and the benefits of all adaptation measures identified using the procedures above cannot be quantified. In addition, we shall take in to account our economic model, which is static CGE model. Therefore, it was necessary to narrow down the list (but broad measures) with which estimating the costs and benefits eases and yet fits to the CGE framework. Following this, in fifth step, I review a set of agricultural economics studies that link agricultural development with public expenditure using econometric techniques (e.g., Benin et al., 2009; Zepeda, 2001; Fan et al., 2000; Evenson et al., 1999). This brings me to the final list of four broad adaptation measures which include irrigation and water management, agricultural research and development, extension services and farmers’ training, and rural feeder roads. The aim of these adaptation measures (or incremental investments) is to collectively offset the productivity lost due to climate change. In the subsequent paragraph, therefore, I briefly discuss how these measures contribute to agricultural productivity. 76 Irrigation and water management enhance productivity of a specific crop, reduce risk of crop failure, and enlarge the production basket and mix of cereals, fruits, vegetables, and root crops (Makombe et al., 2007). Irrigation helps farmers to produce more than once in a year and hence reduce disguised (seasonal) unemployment, and increases the productivity of other inputs like fertilizers. For instance, the marginal productivity of irrigated land (labor) was four (five) times higher than rain-fed land (labor) in Ethiopia (Makombe et al., 2007). Irrigation as an adaptation measure helps to produce the same quantity of output (by letting farmers to produce more than once a year) or value of output (by allowing farmers to cultivate a mix of crops), and to augment labor and land productivity (by compensating soil moisture lost due to dry climate change). Agricultural R&D is crucial investment for agricultural productivity (Pardey et al., 2013). It increases the set of available technologies for agriculture (Zepeda, 2001). Agricultural R&D has tremendously improved farm incomes, reduced costs of production, and reduced the pressure on natural resources base (Alston and Pardey, 2013). The R&D can help adaptation to See, for example, Huang et al. (2006), Diao et al. (2010), Fan et al. (2002), and Thritle et al. (2003) for more on the role these measures in overall agricultural and rural development, food security, poverty reduction, and economic growth; and Nelson et al. (2010) and Lybbert and Sumner (2012) for more on their role in climate change adaptation.

76

76

6. Costs of Planned Adaptation

climate change though producing heat resistant and high yield crop and livestock varieties, and introducing new techniques of production. Agricultural extension services and trainings are complementary to agricultural R&D. They determine whether and how agricultural technology is applied by farmers (Zepeda, 2001), increase farmers’ flexibility to produce and ability to use climate and agricultural information, and help farmers to calculate the costs and returns to alternative technology and marketing strategies (Wiebe et al., 2001). As such, extension services and trainings enable farmers to deal with disequilibria by enhancing the efficiency of farmers to “perceive, to interpret correctly, and to undertake actions that will appropriately reallocate their resources” (Schultz, 1975, p.827). In Ethiopia, for instance, the average cereal yield in areas covered by extension services was 1.4 times the average yield with no extension services (AgSS, 2006). Receiving at least one visit from an extension worker increases annual consumption growth (by 7%) and reduces poverty incidence (by 10%) of rural households (Dercon et al., 2007). In this study, the extension services and trainings are broader than the conventional ones. They include government efforts to build the capacity of its institutions to plan and facilitate adaptation, to improve the effectiveness of climate relevant extension services, and to upgrade farmers’ skill to adopt climate compatible biotechnologies, to seek and understand climate information, to use water efficiently, and to undertake autonomous soil and water management practices. They also include ‘climate services’. 77 Rural feeder roads contribute to agricultural productivity through input and output prices, diffusion and application of biotechnologies (Wiebe et al., 2001; Dercon et al., 2007; Stifel and Minten, 2008; Benin et al., 2009). For example, agricultural households’ access to all-weather roads in Ethiopia increases annual rural consumption by 15%, and reduces the likelihood of being poor by 6-7% (Dercon et al., 2007). Therefore, rural roads can be regarded as indirect climate change adaptation measures contributing to agricultural productivity (Nelson et al., 2010). 6.4.2 Direct benefits of adaptation I start with linking public spending with agricultural productivity for each measures as specified below (see also Nelson et al., 2010). εi =

%∆TFPA %∆Gi

……………….…………………………. (6.1)

Equation 6.1 measures elasticity (εi) of total factor productivity in agriculture (TFPA) with respect to a percentage change in government spending in measure i (Gi). I collect the minimum, mean, and maximum elasticities (εi) from agricultural economics studies which applied econometric techniques (see Table 6.1).

Climate services are measures expected to influence farmers to adjust and make climate-sensitive decisions in crop and livestock farming which include: scheduling (e.g., planting, harvesting operations), tactical crop management (e.g., fertilizer and pesticide use), crop selection (e.g., wheat or sorghum) or herd management, crop sequence (e.g., long or short fallows) or stocking rates, crop rotations (e.g., winter or summer crops), crop industry (e.g., grain or coffee; native or improved pastures), agricultural industry (e.g., crops or pastures), land use (e.g., agriculture or natural systems), and land use and adaptation of current systems (Meinke and Stone, 2005 cited in Tall, et al., 2014). 77

6.4 Materials and methods

77

Table 6.1-Summary of elasticities of agricultural productivity Elasticity Measure

Min

Mean

Max

Main References

Irrigation

0.03

0.09

0.20

Fan et al. (1999), Fan et al. (2000), Mitik and Egida (2013)

Research and development

0.03

0.15

0.44

Criag et al. (1997), Evenson et al. (1999), Fan et al. (1999), Thirtle et al. (2003), Alene and Coulibaly (2009)

Extension, education, and training

0.00

0.10

0.71

Evenson et al. (1999), Fan et al. (1999), Fan et al. (2000), Chen et al.(2008), Felloni et al. (2001), Mitik and Engida (2013)

Rural roads

0.06

0.07

0.08

Fan et al. (1999), Fan et al. (2000)

Agriculture

0.15

0.20

0.24

Benin et al. (2009), Diao et al.(2010), Mitik and Egida (2013)

From the discussion so far, one can see that the collective productivity effect of the adaptation catalog will depend on the total budget available and its allocation among the four adaptation measures, and how effectively is the budget used. However, budget allocation to each investment areas (or adaptation measures in our case) is explicit neither in the public budget nor in the national income accounts of Ethiopia. In contrast, public expenditure principles for effective climate finance delivery requires ‘on-budget’ public spending (Bird et al., 2013). It requires climate change expenditure to be planned and budgeted in the national budget formulation process, and to be executed through government systems during the budget year (Bird et al., 2013). We have to find a way to reconcile the two. 6.4.3 Composite adaptation measure I blend the four measures and treat as single ‘composite’ public service aiming to improve efficiency in agriculture sector. For the sake of economic modeling (i.e., static CGE model) as well as meeting the principles of public expenditure for climate change (Bird et al., 2013), this public service shall be aligned with an existing public sector account in the national income or budget accounts. So doing makes it convenient to impose the additional budget for adaptation as a shock to the CGE model. I have already factored in this research problem and create new public service activity (APAGRI) and new public service commodity (CPAGRI) while I modify the SAM (see section 4.4) and calibrate the model (see section 4.5). The procedure gives us annual government expenditure on ‘adaptation relevant agricultural services’ or simply government consumption of CPAGRI to be around 2.04 billion ETB (or 234 million USD) in 2005/06 prices. This amount is equivalent to 13% of the total government consumption (recurrent) budget, 9% of total government spending, and 2% of GDP in 2005/06 (see section 4.4). 78

The new public budget account relates to agriculture, natural resources, and roads in national budget accounts (see, for example, MoFED, 2014; 2015). The figure seems acceptable compared to the total annual budget for agriculture and rural development which is estimated to be about 15% of total government spending and 5% of GDP (MoARD, 2010).

78

78

6. Costs of Planned Adaptation

6.4.4 Direct cost of adaptation As we are dealing with public adaptation, direct costs of adaptation refer to the incremental public budget on agriculture due to adaptation. Direct costs of adaptation depend on the adaptation policy targets and adaptation policy effectiveness. The elasticity of agricultural productivity with respect to change in public spending captures the adaptation policy effectiveness. As mentioned earlier, I assume that the aim of adaptation is to fully offset the agricultural productivity anticipated to be lost due to climate change. Therefore, the policy (productivity) targets are reciprocals of the anticipated climate change induced agricultural productivity shocks. %∆αava

ε = %∆QG

CPAGRI

……………………………………. (6.2)

Where αava is the shift (efficiency) parameter of value-added nest of agricultural activities. It is the CGE parameter where I imposed the climate change induced productivity shocks (see section 5.4.4). It represents the aim of adaptation. QGCPAGRI is the benchmark government spending in public administration (agriculture). In Chapter 5, we deal with climate change impacts pertaining to grain and livestock activities. Per contra, the elasticities from the literature links the adaptation measures with agriculture as a sector. Therefore, I weigh the grain and the livestock productivity shocks by the shares of the two activities in the total value-added GDP of agriculture (see Table 4.1). The procedure gives us agricultural productivity shocks of -4% (LPJmL) and -11% (EPIC) when agriculture includes ‘fishing and forestry’ or -5% (LPJmL) and -12% (EPIC) when agriculture excludes ‘fishing and forestry’. In other words, climate change induces -4% to -12% productivity shocks to the aggregate agricultural activity. However, neither the impacts nor the benefits of adaptation measures are confined to grain and livestock. Direct climate change effects to other agricultural activities (cash crops, enset crop, and even fish and forest) are not included here only because of lack of an applicable biophysical model. On the adaptation side, for instance, farmers can still use the same irrigation facilities for growing not only grains but also vegetables, fruits, cash crops, and enset. To account these possibilities, I enlarge the anticipated productivity shocks to the whole agriculture to be -5% (minimum), -10% (mean), and -15% (maximum) which are also in range of projections by other studies (see, for example, Müller et al., 2011 and Knox et al., 2012 for impacts on African agriculture). Therefore, the adaptation targets are to increase total agricultural productivity by 5% (minimum), 10% (medium), and 15% (maximum). As we settle with our policy targets, the next step is to gauge uncertainties in policy effectiveness. As per Equation 6.2, we have to compute the elasticity of agricultural productivity with respect to public budget to the composite adaptation measure. Whereas the elasticities from the literature are commonly per measure (see Equation 6.1 and Table 6.1). For the sake of this, first, I generate thousands of random variables (elasticities) between the minimum and maximum elasticities of each measures (εi). Then, I compute the average of the four (ε). The procedure yields average random elasticity (ε) with a minimum (0.05), mean (0.2), and maximum (0.35) values. Figure 6.1 below depicts the normal distribution curve of the randomly generated average elasticity (for N=1000). The average elasticity of productivity, 0.2, with respect to the public spending on the ‘composite’ agricultural adaptation measure is comparable with the elasticity (0.24) with respect to public expenditure in agriculture in Nigeria (Diao et al. 2010), and the elasticity (0.2) with respect to public expenditure irrigation and extension services in Ethiopia (Mitik and Engida, 2013).

6.4 Materials and methods

79

Figure 6.1-Normal distribution curve of average elasticities

7 6

Frequency

5 4 3 2 1 0

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

Average elasticity (ε)

We can now derive the incremental public spending required for adaptation in agriculture or simply direct costs of adaptation in agriculture. If we rewrite Equation 6.2, it will give us: %∆QGCPAGRI =

%∆αva a ε

………………………………… (6.3)

Equation 6.3 explicitly sets out that the incremental public budget for adaptation is a function of the elasticities (representing effectiveness of adaptation) and the productivity targets (representing anticipated impacts). The direct costs are proportional to productivity targets but inversely related to policy effectiveness. However, uncertainty is inherent in climate change impacts as well as the effectiveness of public adaptation policies. There may be ecological, technological, economic, and institutional limits that influence the adaptation policy effectiveness. To deal with this, for a given target of productivity (10%), I derive direct costs PAG1 to PAG3, for each elasticities (0.05, 0.2, and 0.35). To deal with uncertainty in anticipated impacts of climate change, for a certain policy effectiveness (elasticity = 0.2), I derive direct costs PAG4 to PAG6 for each policy targets (5%, 10%, and15%). As such, we get a range of 25-100% incremental budget relative to the benchmark budget (%ΔQGCPAGRI) (see Table 6.2). Then, using Equation 6.4 we can obtain the direct costs of adaptation (ΔQGCPAGRI). %ΔQGCPAGRI =

ΔQGCPAGRI QG0CPAGRI

∗ 100 =

QG1CPAGRI −QG0CPAGRI QG0CPAGRI

∗ 100 ……………………….. (6.4)

Accordingly, the derived costs of adaptation in agriculture per year are in a range of USD 60 to 234 million in 2005/06 prices (see Table 6.2). The larger the anticipated impacts and/or the lower the elasticities, the higher are the direct costs of adaptation.

80

6. Costs of Planned Adaptation

Table 6.2-Direct costs of planned adaptation in agriculture Simulation

Target (%Δαava )

Elasticity (ε)

Budget uplift (%ΔQGCPAGRI )

Costs in USD (ΔQGCPAGRI )

PAG1

10

0.05

100

234

PAG2

10

0.20

50

117

PAG3

10

0.35

29

68

PAG4

5

0.20

25

59

PAG5

10

0.20

50

117

PAG6

15

0.20

75

176

Source: Author’s computation Notes: Costs in USD are in millions.

Simulation PAG2 and PAG5 represent the same cost scenario. They represent the medium costs scenario which is about USD117 million in 2005/06 prices. My medium costs scenario are higher than the costs (USD 68-71 million per year) in Robinson et al. (2013, p.15) but close to the additional public expenditure (USD132 million) required for the urgent 15 adaptation measures in agriculture (FDRE, 2015). Of course, the measures in Robinson et al. (2013) do not include rural roads while the measures in FDRE (2015, p.63) include measures related to biodiversity and forests. Costs in both studies are based on experts’ opinion (Robinson et al., 2013, p.14; FDRE, 2015, p.48). Therefore, despite the fact that the approach I pursue here is different, the derived direct costs are in reasonable range of the two studies. 6.4.5 Modeling into the CGE model Since I assumed ‘full’ adaptation to each anticipated impact scenarios, it only needs to model the incremental costs due to adaptation into the CGE model. The incremental costs relative to the benchmark (%ΔQGCPAGRI) are modeled as shocks to the fixed government consumption of public administration (agriculture) (QGCPAGRI). I model the incremental costs as change in public consumption (recurrent budget) not as change in investment for three reasons. First, government investment is not explicit in the original SAM. Second, some of the expenditures (e.g., for R&D, climate service, extension services, trainings) are recurrent by their nature. Third, the general equilibrium effects of increasing real investment in static CGE models will be vague (Lofgren et al., 2002; Missaglia and de Boer, 2004; Hosoe et al., 2010). The simulation results in the subsequent sections can alternatively be regarded as the economywide and regional effects of increasing public expenditure for the sake of full adaptation in agriculture. I believe that the range of simulations - PAG1 to PAG6 – can gauge the uncertainty in the anticipated impacts and the adaptation policy effectiveness. I do not report the economywide and regional effects of PAG5 simulation as they are redundant of PAG2 (see Table 6.2). 6.5

Economy-wide results and analysis

The macroeconomic, sectoral output, factor market, and welfare effects of alternative adaptation cost scenarios (see Table 6.2) are presented and analyzed below.

6.5 Economy-wide results and analysis

81

6.5.1 Macro-economy The macroeconomic effects show that the total government consumption (GOVCON) increases by 3 to 13% that leads to government saving (GSAV) to decrease by 33 to 173% (see Table 6.3). Falling government saving shifts the saving-adjustment burden, due to S-I balance closure, to households. In effect, private households’ consumption (PRVCON) declines by 0.6 to 2.7%. Simply put, increasing government consumption crowds out private consumption. Table 6.3-Macroeconomic effects of planned adaptation costs Simulations (% change) Notation

Variable

ABSORP

Absorption

-0.6

-0.2

-0.1

-0.1

-0.4

PRVCON

Private consumption

-2.7

-1.2

-0.7

-0.6

-1.9

EXPORTS

Exports

-2.3

-1.0

-0.6

-0.5

-1.6

IMPORTS

Imports

-0.8

-0.4

-0.2

-0.2

-0.6

GDPMP

GDP at market prices

-0.8

-0.3

-0.1

-0.1

-0.5

GOVCON

Government consumption

12.8

6.4

3.7

3.2

9.6

GSAV

Government saving

-173.6

-72.5

-38.7

-32.9

-119.2

EXR

Real Exchange Rate

-1.6

-0.6

-0.3

-0.3

-1.0

PAG1

PAG2

PAG3

PAG4

PAG6

Source: CGE simulations Notes: PAG1 (uplifting ‘adaptation relevant’ budget by 100%), PAG2 (uplifting ‘adaptation relevant’ budget by 50%), PAG3 (uplifting ‘adaptation relevant’ budget by 29%), PAG4 (uplifting ‘adaptation relevant’ budget by 25%), and PAG6 (uplifting ‘adaptation relevant’ budget by 75%).

In parallel, due to the increasing government consumption of ‘adaptation relevant’ public administration services (QGCAPGRI), output from public administration agriculture (APAGRI) expands (see in Table 6.4). This in turn drives up the demand (and wage rates) especially for nonagricultural factors (see Table 6.5). This has two further macroeconomic implications. First, it increases non-agricultural factor income and ultimately, urban households’ real consumption which partly offsets the decline in the total private consumption due to shift in saving-adjustment burden. Second, it increases cost of production in manufacturing and private services, and thus, lowers output and exports from these sectors (see Table 6.4). Consequently, aggregate exports fall by 0.5 to 2.3% following which imports slightly decline to satisfy the external sector closure. In balance, at least at the macroeconomic level, the decrease in private consumption and exports surpass, only slightly, the increase in total government consumption which leaves negligible effects of planned adaptation costs on the GDP and total absorption. 6.5.2 Sectoral output The negligible aggregate effects (e.g., on total absorption and GDP in Table 6.3) do not mean that the indirect costs of planned adaptation are small. We shall further investigate the resource pull effects of the incremental public spending for adaptation.

82

6. Costs of Planned Adaptation

Table 6.4 presents the sectoral output effects. As expected, the adaptation measures increases administration (agriculture) output (APAGRI) by 25 to 100% which is directly proportional to the budget uplifts in Table 6.2. Public activities usually contract construction services. As a result, expansion in APAGRI slightly improves construction output (ACONS). Table 6.4-Sectoral output effects of planned adaptation costs Simulations (% change) Notation

Activity

AGRAIN

Grain crops

-0.2

-0.1

-0.1

0.0

ACCROP

Cash crops

0.4

0.3

0.2

0.1

0.4

AENSET

Enset crop

-1.6

-0.6

-0.3

-0.3

-1.1

ALIVST

Livestock production

0.9

0.3

0.2

0.2

0.6

AFISFOR

Fish and forest

-1.6

-0.6

-0.3

-0.3

-1.1

AMINQ

Mining and quarrying

-3.8

-1.6

-0.9

-0.7

-2.6

ACONS

Construction

0.4

0.2

0.1

0.1

0.3

AMAN

Manufacturing

-9.8

-4.2

-2.3

-1.9

-6.9

ATSER

Wholesale and retail

-1.6

-0.6

-0.3

-0.3

-1.1

AHSER

Hotels and restaurants

-6.3

-2.7

-1.4

-1.2

-4.4

ATRNCOM

Transport and comm.

-3.6

-1.6

-0.9

-0.8

-2.6

AFSER

Financial intermediaries

-2.8

-1.3

-0.7

-0.6

-2.0

ARSER

Real estate and renting

-0.3

-0.1

-0.1

-0.1

-0.2

APADMN

Public admin. (general)

-0.1

0.0

0.0

0.0

-0.1

APAGRI

Public admin. (agriculture)

98.9

49.4

28.7

24.7

74.2

ASSER

Social services

-2.6

-1.3

-0.7

-0.6

-1.9

AOSER

Other services

-13.2

-6.2

-3.4

-2.9

-9.6

GDPFC2

Total GDP at factor cost

-0.7

-0.3

-0.1

-0.1

-0.5

PAG1

PAG2

PAG3

PAG4

PAG6 -0.2

Source: CGE simulations Notes: PAG1 (uplifting ‘adaptation relevant’ budget by 100%), PAG2 (uplifting ‘adaptation relevant’ budget by 50%), PAG3 (uplifting ‘adaptation relevant’ budget by 29%), PAG4 (uplifting ‘adaptation relevant’ budget by 25%), and PAG6 (uplifting ‘adaptation relevant’ budget by 75%).

On the other hand, expanding public services (APAGRI in this case) increases the economywide wage rate, especially, of the two most skilled labor factors, i.e., FLAB1 and FLAB2 (see Table 6.5). As a result, the cost of production in manufacturing and the rest of services will increase and following which value-added output from these sectors fall. In all cost scenarios, the resource-pull effects are vivid on manufacturing (-2% to -10%) and ‘other’ services (-3% to -13%). Hotels and restaurants, transport and communications, and mining and quarrying are other activities that shoulder the indirect effects. Like in the case of macroeconomic effects, the resource-pull effects of planned public adaptation in agriculture will get worse as the direct costs of adaptation increase. High direct costs may even impede other agricultural activities (which have non-negligible value-added contribution from FLAB1 and FLAB2, e.g., AENSET and AFISFOR), and other public services (e.g. ASSER). The resource-pull effects, especially,

6.5 Economy-wide results and analysis

83

on manufacturing and private services imply that government financed adaptation may deter structural transformation in the country towards which the country is thriving for (MoFED, 2010; NPC, 2016). I will turn more to this point in Chapter 8. 6.5.3 Factor markets Public adaptation affects the factor market as it pulls some of the factors towards public activities, and contributes to push up average wage rates of some factors relative to others. The change in relative wages of factors used in the public services relative to others (e.g. agricultural labor) will trigger factor substitution in other activities. It is natural to expect that public adaptation measures to increase demand for administrative workers, professionals and associates, and to some extent, skilled labor. Consequently, the average wage rate as well as total factor income of these factors increase (see Table 6.5). In particular, the effects are notable on the average wage rates of administrative workers (14 to 78%) and professionals and associates (10 to 57%). In contrast, average wage rate and factor income of agricultural factors decline. The relative wage and income declines are stronger in agricultural labor and cropland. Table 6.5-Factor market effects of planned adaptation costs Simulations (% change)

Factor income (YF)

Economy-wide wage (WF)

Variable

Account

Factor

FLAB0

Agricultural labor

FLAB1 FLAB2

PAG1

PAG2

PAG3

PAG4

PAG6

-10.2

-4.3

-2.3

-2.0

-7.1

Administrative labor

78.4

31.6

16.5

13.9

53.0

Professional labor

57.5

23.1

12.0

10.1

38.8

FLAB3

Unskilled labor

31.2

12.1

6.2

5.3

20.7

FLAB4

Skilled labor

13.3

5.0

2.6

2.2

8.7

FLND

Cropland

-10.6

-4.5

-2.4

-2.0

-7.3

FLAB0

Agricultural labor

-10.1

-4.3

-2.3

-2.0

-7.0

FLAB1

Administrative labor

40.9

15.9

8.2

6.9

27.2

FLAB2

Professional labor

42.6

17.2

9.0

7.6

28.8

FLAB3

Unskilled labor

FLAB4

Skilled labor

FLND

Cropland

FTLU

Livestock units

FCAP

Capital

9.5

3.8

2.0

1.7

6.4

23.1

8.7

4.5

3.8

15.1

-10.4

-4.4

-2.4

-2.0

-7.2

-6.0

-2.7

-1.5

-1.2

-4.3

13.0

5.5

3.0

2.5

9.0

Source: CGE simulations Notes: PAG1 (uplifting ‘adaptation relevant’ budget by 100%), PAG2 (uplifting ‘adaptation relevant’ budget by 50%), PAG3 (uplifting ‘adaptation relevant’ budget by 29%), PAG4 (uplifting ‘adaptation relevant’ budget by 25%), and PAG6 (uplifting ‘adaptation relevant’ budget by 75%).

In summary, Table 6.5 shows that planned public adaptation increases (decreases) returns for factors owned by urban (rural) households. As discussed below, this will influence the urban and households’ wellbeing.

84

6. Costs of Planned Adaptation

6.5.4 Households’ welfare Since the government saving falls (Table 6.3), households’ saving shall increase to maintain the macroeconomic saving-investment balance. Other things remaining constant, this reduces households’ consumption budget and hence their real consumption. On the flipside of the coin, however, public adaptation increases the demand (and income) for factors mainly owned by urban households (Table 6.5). In balance, we see decreasing welfare to rural households but increasing welfare to urban households (see Figure 6.2). Figure 6.2 shows that the equivalent variation for rural and urban households may reach down to -5%, and up to 4%, respectively. The welfare effects between the two household groups widen as adaptation costs increase. The households’ welfare effects are also robust to ±25% of the elasticities presented in Table 4.11 (see Table A4 in the Appendix). Figure 6.2-Households’ welfare effects of planned adaptation costs

EQUIVALENT VARATION (%)

6 4 2 0

RURH PAG1

PAG3

URBH

-2 -4 -6

Source: CGE simulations Notes: PAG1 (uplifting ‘adaptation relevant’ budget by 100%), and PAG3 (uplifting ‘adaptation relevant’ budget by 29%).

It should be noted here that the welfare effects of planned adaptation on rural households are by far lower than the effects due to climate change (see Figure 5.1). 6.6

Regional projections and analysis

The regional projections and analysis in this section combines the Ethiopia-wide sectoral output effects (Table 6.4) with the regional module (Table 4.13). Recall that the regional effects depend on the CGE experiment and the economic structure of the regions. The CGE experiments of this chapter directly and favorably affect the public administration (agriculture) (APAGRI) sector. In contrast, the simulations indirectly and negatively affect the manufacturing (AMAN) and many services (e.g., AHSER, ATRNCOM, and AOSER). The regional effects of planned public adaptation costs are presented in Table 6.6. The table confirms our hypothesis. Urbanized regions like Addis Ababa (-0.6% to -3%) and Dire Dawa (-0.4% to -2%) will confront with the bulk of indirect effects of public adaptation costs in agriculture. The decline in the region-wide GDP of these two federal city regions are larger than

6.7 Conclusions

85

the national average (-0.1% to -0.7%) as well as in each of the other regions. Addis Ababa is entirely urbanized region with negligible contribution from agricultural activities. As a result, public administration (agriculture) service (APAGRI) in the city is almost nil. Simply put, expanding Ethiopia-wide APAGRI sector will benefit Addis Ababa little. Per contra, manufacturing and services whose output fall due to public adaptation in agriculture (see Table 6.2) represent about 85% of Addis Ababa’s value-added GDP (see Table 4.13). In contrast, government spending for adaptation in agriculture bears favorable effect to Tigray region’s GDP (0.7% to 2.3%). This is because the share of APAGRI in Tigray region, which is 4%, is larger than the shares in the national average, and in other regions (see Table 4.13). Negative effects are indicated for the rest of the regions which may get worse with increasing adaptation costs. Table 6.6-Regional effects of planned adaptation costs Simulations (% change) Region Ethiopia

PAG1

PAG2

PAG3

PAG4

PAG6

-0.7

-0.3

-0.1

-0.1

2.3

1.3

0.8

0.7

1.8

Afar

-0.8

-0.3

-0.1

-0.1

-0.5

Amhara

-0.5

-0.2

-0.1

0.0

-0.3

Oromia

-0.7

-0.2

-0.1

-0.1

-0.4

Somali

-0.7

-0.3

-0.1

-0.1

-0.5

Benishangul-Gumuz

-0.3

0.0

0.0

0.0

-0.1

Southern NNP

-0.5

-0.2

-0.1

-0.1

-0.3

Gambella

-0.5

-0.1

0.0

0.0

-0.3

Harari

-0.8

-0.3

-0.1

-0.1

-0.5

Addis Ababa

-2.9

-1.3

-0.7

-0.6

-2.1

Dire Dawa

-2.0

-0.8

-0.4

-0.4

-1.4

Tigray

-0.5

Source: Table 6.4 and Table 4.13 Notes: PAG1 (uplifting ‘adaptation relevant’ budget by 100%), PAG2 (uplifting ‘adaptation relevant’ budget by 50%), PAG3 (uplifting ‘adaptation relevant’ budget by 29%), PAG4 (uplifting ‘adaptation relevant’ budget by 25%), and PAG6 (uplifting ‘adaptation relevant’ budget by 75%).

In general, the regional effects of adaptation costs in urban regions (Table 6.6) are worse than the effects of climate change on them (Table 5.5). The converse is true for the agrarian regions like Amhara, Oromia, and Southern NNP. This is expected as the increased public spending is meant to prevent climate change from affecting agriculture. 6.7

Conclusions

In this chapter, I consider the case of planned adaptation that aims to fully neutralize the productivity shocks induced by climate change by undertaking a composite of four broad measures known to improve agricultural productivity. I employ a method that specifies direct costs of adaptation as a function of anticipated productivity effects of climate change and elasticity of

86

6. Costs of Planned Adaptation

agricultural productivity with respect to public spending. The derived direct costs of adaptation ranges between 60 and 234 million USD, and may require to uplift budget by 25% to 100% depending on the anticipated productivity shocks and the presumed value of elasticity. Next, I analyze the economy-wide and regional effects of this incremental budget demand. The CGE simulations show that the effects on national GDP and total absorption of the incremental budget demand are negligible. However, increasing government spending shifts the saving adjustment burden to households following which the total households’ real consumption decline. Expanding public adaptation services also squeezes manufacturing and private services. The effects on manufacturing and private services may deter structural transformation in the country which has already been unsatisfactory in the last two decades (Dorosh and Thurlow, 2011; NPC, 2016). Government of Ethiopia regards its role in structural change as indispensable (NPC, 2016). Therefore, deteriorating public saving due to planned public adaptation may strain structural change in the country as it will compete for public budget for sectors crucial to structural transformation such as for railways, hydropower, transport and communications, and industrial parks. As such, the opportunity costs of public spending on agricultural adaptation are high. As output from manufacturing and private services decline, regional effects are negative and strong in Addis Ababa and other urbanized regions. Shrinking output in urban regions may affect the pace of urbanization in the country. Currently, barely 20% of the country’s population lives in urban areas (CSA, 2013). Urban unemployment is about 17% (ICPS, 2012) while the policy discourse on urban agenda is a recent phenomenon in the country (Dorosh and Thurlow, 2011). Therefore, the effects on urban GDP may have far reaching consequences. Comparing the regional effects of adaptation with that of climate change shows that those least affected by climate change will shoulder the largest burden of public adaptation in agriculture. The comparison has also an implication for federal block-grant budget allocation among regions. The existing formula, among others, takes into account population, revenue generating capacity, and infrastructural deficiency of the regions (MoFED, 2009). It may be necessary to think over how to proceed with it in the face of climate change and climate change adaptation.

7 7.1

Public Finance for Adaptation Introduction

The discussion in Chapter 6, implicitly, assumes that the government is committed to allocate extra budget to agriculture and rural development and shoulder the ensuing deficits. However, Government of Ethiopia may not easily take additional budgetary commitments for agriculture (BMGF, 2010). The current budgetary commitments to agriculture are one of the highest in Africa (MoARD, 2010; Lanos et al., 2014) while the country is striving for government assisted structural transformation (NPC, 2016). The unsatisfactory growth of the sector (BMGF, 2010; MoARD, 2010; Mitik and Engida, 2013; ILRI, 2015) itself has triggered policy debate on continuing the priority given to agriculture in public policies and budgets. On the other hand, climate change is an overarching development problem which shall not be regarded as a standalone environmental problem that left to a single government agency (McGray et al., 2007; Fankhauser and Schmidt-Traub, 2011). Seen from both perspectives, it may be necessary to look for new sources of adaptation finance. This chapter is built on Chapter 6. It intends to examine the economy-wide and regional effects of different fiscal schemes to raise and earmark as adaptation finance. The scientific contribution of the chapter is twofold. It gives more comprehensive picture of the economy-wide and regional effects of planned adaptation in a single sector (agriculture in our case). It also contributes to identifying adaptation finance scheme (s) with least indirect effects, particularly, if domestic sources have to be used. The remainder of the chapter reviews the literature related to adaptation finance in developing countries (7.3) and in Ethiopia (7.3), presents the materials and methods (7.4), analyze the economy-wide (7.5) and regional (7.6) effects. Section 7.7 presents the conclusions of the chapter. 7.2

Adaptation finance in developing countries

Adaptation in LDCs is as important response as mitigation to climate change. LDCs ought to identify, prioritize and plan adaptation in their most vulnerable sectors accordingly (UNFCCC, 2009). Climate finance and technological transfers are expected from developed countries to support efforts related to climate change in LDCs (Tanner and Horn-Phathanothai, 2014; UNFCCC, 2009). This arrangement reflects the UNFCCC’s ‘common but differentiated responsibility’ principle as the costs of adaptation in LDCs are beyond their financial capacity (Tanner and Horn-Phathanothai, 2014). As consequence, different initiatives for international climate finance in general and adaptation finance in particular have been established since 2000s (see, for example, CFU, 2016). Nonetheless, these international climate finance initiatives have got a lot of criticisms. First, there is no clear definition on what constitutes climate finance and how it is measured (Tanner and Horn-Phanathanothai, 2014; Adaptation Watch, 2015; Westphal et al., 2015). The concept of additionality 79 and its measurement is still subject to debate (Westphal et al., 2015; 79 Additionality refers that climate finance should be on the top of existing official development assistance (ODA) to LDCs.

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 A. W. Yalew, Economic Development under Climate Change, https://doi.org/10.1007/978-3-658-29413-7_7

88

7. Public Finance for Adaptation

Adaptation Watch, 2015). The international climate finance includes finance to support both adaptation and mitigation activities in both developed and developing countries. However, it is dominated by mitigation finance. For instance, of the total USD 391 billion global climate finance flows in 2014, only USD 25 billion (6.4%) is related to adaptation finance (Buchner et al., 2015). Only 18% the total USD 35 billion Fast-Start Climate Finance (2010-2012) for developing countries is related to adaptation (Nakhood et al., 2013). Similarly, for Africa, AffulKoomson (2015) finds that mitigation funds are three times larger than adaptation funds. As such, reports indicating increasing trends of climate finance conceal the real order of magnitudes of adaptation finance for developing countries. Adaptation finance reported by developed countries is inflated, perhaps, four times of the actual amount (Adaptation Watch, 2015). There still lacks global effort to define what should count as adaptation finance (Adaptation Watch, 2015). The sources, instruments, and channels that should count toward the goal of mobilizing USD 100 billion a year by 2020 is yet ambiguous (Westphal et al., 2015). Second, there are operational problems related to the access and management of international climate finance (Tanner and Horn-Phanathanothai, 2014; Afful-Koomson, 2015). There still lacks clear agreement on who has to contribute and how much (Fenton et al., 2014). The criteria used to distribute the deposited funds are not clear (Tanner and Horn-Phanathanothai, 2014). For instance, the allocation of adaptation finance from the Fast-Start Climate Finance was correlated to the existing ODA patterns rather than vulnerability of the recipient countries (Nakhood et al., 2013). The funding mechanisms and instruments are still open to debate (Tanner and Horn-Phanathanothai, 2014; Afful-Koomson, 2015). Third, the definition and operational problems have bred concerns over the amount, time, and ways of disbursement. Generally, the disbursed amount is substantially below what is even deposited (Tanner and Horn-Phanathanothai, 2014; CFU, 2016). There is formidable time lag between approval and disbursement of funds (Afful-Koomson, 2015; Adaptation Watch, 2015). Ways of disbursements are mixed which include transfers as grants or loans, as project-based or program-based, to non-governmental organizations or directly to governments (Nakhood et al., 2013). Fourth, therefore, adaptation finance from international climate finance is doomed as uncertain and inadequate relative to the costs of adaptation (Afful-Koomson, 2015; Fenton et al., 2014; Buchner et al., 2015; UNEP, 2016). Climate finance from developed countries is uncertain as it is affected by their own domestic issues such as fiscal austerity measures (Afful-Koomson, 2015; Westphal et al., 2015) and other global issues such as financial crisis (Tanner and HornPhanathanothai, 2014; Nakhooda et al., 2013). International climate finance is also inadequate compared to the adaptation costs expected in LDCs. Currently, adaptation costs in developing countries are at least 2 to 3 times higher than international public finance for adaptation (UNEP, 2016). Unless significant progress is made to secure new and additional finance for adaptation, the gap is likely to grow substantially over the coming decades (UNEP, 2016). Fifth, adaptation finance is fragmented and not transparent (Nakhood et al., 2013; Adaptation Watch, 2015). Adaptation finance is usually from limited bilateral sources which guarantees no sustainability (see, for example, CFU, 2016). Even if much of the adaptation finance is supposed to be delivered as grants, the fast-track finance (2010-2012) reveals that developed countries report concessional and non-concessional loans, capital contributions, guarantees, and insurance as climate finance (Westphal et al., 2015). If available, grants are limited to short-term and project type adaptation measures which bear high transaction costs (Afful-Koomson, 2015). Even though it is a viable option, debt-for-nature (debt relief for climate finance swaps) constitutes only 0.3% of the Fast-Start Finance (Fenton et al., 2014).

7.3 Adaptation finance in Ethiopia

89

To conclude, the adequacy and reliability of global adaptation finance for developing countries is questioned and undermined. Developing countries do not know “how much assistance to expect, whether climate funds are simply replacing money previously committed to address other development needs, or whether the funds are being delivered at all” (Adaptation Watch, 2015, p.50). This shows that international adaptation finance, at least, may not cover all adaptation costs that developing countries may incur. Therefore, Afful-Koomson (2015) suggests that African countries shall focus on implementing policies to mobilize adaptation finance from domestic resources. The burden will likely fall on the public finance as the private sector’s willingness to invest in adaptation activities is limited (Mendelsohn, 2012). This apparently will have non-negligible economy-wide and regional effects. This is the central objective of this chapter. 7.3

Adaptation finance in Ethiopia

Ethiopia recognizes climate change as a concrete threat to its economic prospect. The government is planning and preparing for public responses to climate change in different sectors (FDRE, 2011; 2015; 2016). The country has put building agriculture resilience to climate change as its domestic priority (NMA, 2007; FDRE, 2011; 2015). In effect, the government targets national funds for adaptation in agriculture (FDRE, 2015). About 80% of the incremental budget required to build resilience of the agriculture sector to climate change falls on the public sector (FDRE, 2015). Yet, public finance for adaptation in agriculture is only part of the national budgetary resources required for climate change-relevant actions (mitigation, adaptation, or both). Ethiopia may need about USD 440 million per year to deliver its climate resilience strategies that include mitigation and adaptation in many sectors (Eshetu et al., 2014). As such, Ethiopia desperately needs enormous amount of international climate finance especially in grant forms (Eshetu et al., 2014). However, in the past years, international climate finance to Ethiopia was far below what is envisaged in the climate-resilient green economy strategy of the country (Eshetu et al., 2014; CFU, 2016). International support accounted for only 20% of public spending on mitigation and adaptation relevant activities between 2008 and 2012 (Eshetu et al., 2014). In the last decade, only 16% is disbursed from the total USD 123 million approved international climate finance to the country (CFU, 2016). The international climate finance is inclined to mitigation measures (circa 60%), and what is allocated to adaptation measures focuses on building advisory and institutional capacity (e.g., Eshetu et al., 2014; CFU, 2016). Taken together, one may contend that Government of Ethiopia shall prepare itself to mobilize adaptation finance for agriculture from domestic resources. The default option is to shoulder the fiscal deficit, i.e., financing through ‘deficits’. This is what we saw in Chapter 6. However, the government is unlikely to bear additional commitments in agriculture (BMGF, 2010) as the existing public expenditure structure already emphasizes agriculture and rural development (MoARD, 2010; Lanos et al., 2014) while the country desperately needs structural transformation driven by the government (NPC, 2016). In addition, climate change and hence adaptation to climate change is an overarching problem that shall not be left to a specific government agency and public budget account (McGray et al., 2007; Fankhauser and Schmidt-Traub, 2011). As such, the government may seek for new international and domestic resources for adaptation in agriculture

90

7. Public Finance for Adaptation

which, among others, may include diverting from other public budget accounts, seeking for foreign grants, and increasing tax revenue. These require fiscal decisions which usually bear general equilibrium effects which is not addressed well in the literature. The central objective of this chapter is to examine the economy-wide and regional effects of alternative adaptation finance schemes for agriculture. Adaptation finance is a recent topic in the field. The indirect economic consequences of alternative fiscal schemes to raise and finance adaptation in specific or selected sector (s) are barely researched. This chapter contributes its part to the scientific discourse in the topic in two main ways. First, complementing planned adaptation costs with adaptation finance will confer better information on the indirect costs of planned adaptation in a single sector. Second, examining the general equilibrium effects of alternative sources hints on finance scheme (s) that imply relatively least indirect costs. 7.4

Materials and methods

This chapter is built on chapter 6. The adaptation measures and aim remain the same. I draw and focus on the medium adaptation cost case (PAG2 in chapter 6, but, renamed in this chapter as PAG for convenience). The medium adaptation cost scenario pertains to public adaptation to -10% anticipated agricultural productivity shock at 0.2 elasticity of agricultural productivity to public spending. It requires about one billion ETB (or USD 117 million at 2005/06 prices). The spending on public administration (agriculture) services will remain intact. Therefore, the focus of this chapter is how to raise this sum of money. 7.4.1 Adaptation finance schemes I begin with the default option of adaptation finance such that the government increases its spending on public administration (agricultural) services. This can be regarded as financing adaptation through public deficits (PAG simulation). This simulation represents shifting up the benchmark public consumption of agricultural services (CPAGRI) by 50%. This simulation serves us as benchmark. I devise two schemes that involve diverting the same amount of money from the two other public budget accounts of the SAM. These are diverting from public administration (general) services (PAGGA simulation) and public social services (PAGSS simulation). I also include an experiment representing international climate finance (in the form of grants) for adaptation (PAGF simulation). The last three financing schemes (PAGD, PAGS, and PAGT) consider raising tax rates, for a given tax base, to generate government revenue equal to the adaptation finance. The calibrated CGE model (see Table 4.7) presents direct (income) taxes, sales (commodity) taxes, and import tariffs. Direct taxes are collected from rural and urban households, the sales taxes are collected from market commodities, and import tariffs are collected from imports. Taken together, six adaptation finance experiments are designed. These are increasing on public agricultural services (PAG) plus: 1) diverting from public administration services (PAGGA), 2) diverting from public social services (PAGSS), 3) increasing transfers from the ROW to government (PAGF), 4) increasing direct tax rate (PAGD), 5) increasing sales tax rate (PAGS), and 6) increasing import tariffs (PAGT). I believe that this set of experiments fairly gauge the range of the economy-wide and regional effects of adaptation finance for agriculture.

7.4 Materials and methods

91

7.4.2 Modeling into the CGE model By now, adaptation in agriculture is regarded as new public policy represented by changes in government consumption and revenue parameters. The real government consumption of public services as well as the ROW transfer to government are fixed at observed level. The ratio of the incremental budget or direct costs of adaptation (i.e., USD 117 million) to the benchmark values of these exogenous variables is equal to 50% of government consumption of public administration (agriculture) (QGCPAGRI), to 12% of government consumption of public administration (general) (QGCPADMN), to 18% of government consumption of social services (QGCSSERV), to 27% of transfer from abroad to government (TRNSFRGOV, ROW). For a given tax base, the government can generate the required amount by raising the exogenous rates of direct tax (DTAX) by 25% , of sales tax (STAX) by 32%, and of import tariffs and duties (MTAX) by 14%. It is important to note that the shock to government consumption of public administration (agriculture) (QGCPAGRI by 50%) remains the same for all simulations. The default financing scheme, PAG, involves only increasing QGCPAGRI by 50% since no extra source of finance is not assumed. The rest of simulations add one more shock to a fiscal parameter, in the CGE, representing the respective source of adaptation finance. For instance, PAGSS involves increasing QGCPAGRI by 50% and decreasing QGCSSERV by 18% whereas PAGT involves increasing QGCPAGRI by 50% and increasing import tariff rates by 14%. Therefore, the financing approach here is kind of ‘earmark’ financing. 80 That is the incremental revenue collected is only allocated to adaptation in agriculture. The analysis shall also be regarded as ‘balanced-budget’ analysis according to which “any increase in expenditure has to be matched by a decrease in somewhere else or by a new source of tax revenue” (Stiglitz, 2000, p.509). Therefore, the objective here is only to assess and compare the general equilibrium effects of alternative sources of finance. The main hypothesis of the chapter is that compared to the benchmark scheme (PAG simulation), adaptation finance schemes will dampen the effects on government saving (GSAV) but further distress the households’ welfare or pull out resources from other economic sectors. In addition, alternative adaptation finance schemes will have different implications for different sectors, household groups, and regions of the economy. The diverting approaches switch budgets only between budget accounts (or government consumption demand). They may bear the least general equilibrium effects since the pattern of income (government is the main demander) and spending (production structure) of the three public services is very similar. The international adaptation finance scheme brings an additional resource to the Ethiopian economy and shall dampen the effects on households’ welfare. Nevertheless, transfers from abroad may appreciate real exchange rate, and affect the trade balance. Increasing direct tax rates reduce income left for consumption and hence impedes on households’ welfare. Increasing sales tax and import tariffs alter relative prices and may have distortionary effects which depend on own-and cross-price elasticities (Stiglitz, 2000; Bailey, 2002), square of the tax rates (Stiglitz, 2000), and second-best efficiencies (Burfisher, 2011). 81 However, there may 80 Earmark financing approach refers to the case in which the revenues of an earmarked tax, for instance, are “dedicated to a public sector services either in whole or in the form of predetermined and fixed percentage” (Bailey, 2002, p.231). 81 CGE models are able to consider second-best efficiencies, i.e., increasing a specific tax rate may increase or decrease excess burden, if any, due to other tax types in the model (Burfisher, 2011).

92

7. Public Finance for Adaptation

not be such big deadweight losses since the initial tax rates and elasticities in the calibrated model are low. 7.5

Economy-wide results and analysis

As in the previous chapters, the economy-wide effects refer to the effects on the macro economy, sectoral output, factors markets, and households’ welfare. 7.5.1 Macro-economy Table 7.1 presents the macroeconomic effects of the alternative adaptation finance schemes. Relative to the benchmark (PAG simulation), the overall effects of the proposed finance schemes are negligible. The macroeconomic effects of all finance schemes on the aggregate variables (total absorption and GDP) hovers around -0.3%. However, different schemes influence different macroeconomic components differently. The diverting approaches (PAGGA and PAGSS simulations) involve no increase in total government spending. Therefore, their indirect effects on private consumption and exports are relatively small compared to the PAG simulation. From the two diverting schemes, diverting from social services (PAGSS) seems better: At least, it implies relatively lower burden to government saving (-13%). This may attribute to the fact that public social services interact with private social services compared to the public administration (general) which is entirely produced by the public sector. Table 7.1-Macroeconomic effects of public adaptation finance Simulations (% change) Account

Variable

ABSORP

Absorption

-0.2

-0.4

-0.3

0.3

-0.2

-0.2

-0.3

PRVCON

Private consumption

-1.2

-0.6

-0.4

-0.4

-1.2

-1.2

-1.3

EXPORTS

Exports

-1.0

-0.9

-0.6

-3.7

-1.3

-1.3

-2.4

IMPORTS

Imports

-0.4

-0.3

-0.2

0.6

-0.5

-0.5

-0.9

GDPMP

GDP at market prices

-0.3

-0.5

-0.4

-0.3

-0.3

-0.3

-0.3

GOVCON

Government consumption

6.3

0.1

-0.2

6.3

6.3

6.3

6.3

GSAV

Government saving

-70.8

-24.4

-13.0

-54.5

-50.4

-50.8

-55.8

EXR

Real exchange rate

-0.6

-0.4

-0.1

-1.5

-0.6

-1.2

-1.9

PAG

PAGGA

PAGSS

PAGF

PAGD

PAGS

PAGT

Source: CGE simulations Notes: PAG (increasing budget on CPAGRI), PAGGA (PAG + diverting budget from CPADMN), PAGSS (PAG + diverting budget from CSSERV), PAGF (PAG + grants from abroad), PAGD (PAG + increasing direct tax rate), PAGS (PAG + increasing sales tax rate), and PAGT (PAG + increasing tariff rate).

Compared to the benchmark simulation (PAG), the international transfers (PAGF) help to dampen the pressure on public saving (-55%) which will in turn dampen the saving adjustment burden on households. It also increases imports as the exchange rate appreciates. As consequence, compared to others, the PAGF scheme implies the least indirect effects on private consumption (-0.4%). Foreign grants, however, may appreciate real exchange rate (expressed as local currency to foreign currency) that reduce exports (-3.7%) and increases imports (0.6%). Yet, under the PAGF simulation, the effects on GSAV remains high (-55%). This implies that

7.5 Economy-wide results and analysis

93

the resource pull-effects of planned adaptation on other economic activities (e.g., manufacturing and private services) still inhibits the growth of total government revenue. Table 7.1 shows that the resource allocation effects of direct tax (PAGD) and sales tax (PAGS) are more or less similar. Their effects are also similar to the default financing approach (PAG simulation) except on government saving. This mainly accrues to the progressive nature of the tax system (see Table 4.7 and 4.8) and the nature of the simulations. By the nature of the simulations, I do mean that PAGD is a lump-sum tax on households’ total income. PAGS simulation can also be regarded as lump-sum tax on households’ total consumption budget since sales taxes on all commodities are shocked to increase uniformly. Neither of the two tax schemes would cease the negative effects on GSAV which may attribute to the interaction among different tax systems in the economy. For instance, increasing the direct tax rates reduce income left to spend on consumption, in effect, the revenue from sales tax may decline. Of course, the resource pull effects on the rest of the economy (due to PAG simulation) may still reduce the total taxable income or value of market commodities, especially, from manufacturing goods and private services. Compared to the benchmark as well as other tax approaches, the macro-economy is slightly worse off under tariffs scheme (PAGT simulation). Tariffs reduces total imports (0.9%) which, for the sake of trade balance, also pulls down exports (-2.4%). Tariffs are applied only in three commodities of the CGE model, thus, the initial tax rates are high compared to sales and direct tax rates. Thus, PAGT slightly worsen the households’ total consumption (1.3%). In spite of this, we still do not see large effects as one of the imports (CMMAN, representing 15% of total imports) is entirely imported while the substitution elasticities for the other two commodities are low. 7.5.2 Sectoral output As one may expect, compared to the benchmark, diverting budget from other public accounts (PAGGA and PAGSS) imply least indirect effects on the other sectors. This is reflected on AMAN and AOSER in Table 7.2. For the rest of the schemes, Table 7.2 resembles Table 6.4 as we still have public adaptation to pull-out factors from manufacturing and private services. Particularly, the indirect effects on manufacturing (around -4%) and ‘other’ services (around 6%) persist. Foreign grants scheme has no such peculiarities to be discussed compared to the default option. It may, of course, slightly worsen output from economic activities such as cash crops and manufacturing which contribute to exports. In contrast, it may offset some of the effects in activities that produce no or little tradable commodities like enset, and hotels and restaurants. In many of the sectors, there are no visible differences among the effects of the three alternative tax schemes and default financing scheme. The case of manufacturing is an exception. Among the four schemes, the tariff scheme implies relatively smaller effect on manufacturing (-3.4%). Increasing tariffs makes imports (which are mainly manufactured) relatively expensive. As a result, households switch into domestic manufactured goods, and thus, output from AMAN shall increase which, to some extent, offsets the output decline compared to the case of PAG. Taken together, Table 7.2 reaffirms the conclusion we reached in Chapter 6. The factor-pull effects on non-public economic activities are negative and strong. This impinges on the country’s endeavor for structural change.

94

7. Public Finance for Adaptation

Table 7.2-Sectoral output effects of public adaptation finance Simulations (% change) Account

Activity

AGRAIN

Grain crops

-0.1

0.0

0.0

0.0

-0.1

-0.1

-0.1

ACCROP

Cash crops

0.3

0.0

0.0

-1.1

0.0

0.0

-0.2

AENSET

Enset crop

-0.6

-0.2

-0.2

-0.1

-0.4

-0.5

-0.5

ALIVST

Livestock production

0.3

0.2

0.1

0.3

0.3

0.3

0.5

AFISFOR

Fishing & forestry

-0.6

-0.3

-0.2

-0.2

-0.5

-0.5

-0.5

AMINQ

Mining & quarrying

-1.6

-1.0

-0.7

-1.5

-1.7

-1.5

-1.4

ACONS

Construction

0.2

0.0

0.2

0.2

0.2

0.2

0.2

AMAN

Manufacturing

-4.1

-2.4

-1.9

-4.4

-4.0

-4.1

-3.4

ATSER

Wholesale and retail

-0.6

-0.6

-0.3

-0.8

-0.7

-0.7

-0.9

AHSER

Hotels and restaurants

-2.6

-1.3

-1.1

-1.8

-2.4

-2.6

-2.8

ATRNCOM

Transport & comm.

-1.6

-1.0

-0.6

-1.7

-1.5

-1.4

-1.8

AFSER

Financial intermediaries

-1.2

-1.4

-0.1

-1.1

-1.4

-1.1

-1.3

ARSER

Real estate & renting

-0.1

-0.1

-0.1

-0.1

-0.1

-0.1

-0.1

APADMN

Public admin. (general)

0.0

-11.9

0.0

0.0

0.0

0.0

0.0

APAGRI

Public admin. (agriculture)

48.5

48.5

48.5

48.5

48.5

48.5

48.5

ASSER

Social services

-1.2

-1.0

-13.8

-1.1

-1.2

-1.2

-1.2

AOSER

Other services

-6.0

-3.8

-2.0

-5.6

-5.9

-5.8

-5.9

GDPFC2

Total GDP at factor cost

-0.2

-0.5

-0.4

-0.3

-0.3

-0.3

-0.3

PAG

PAGGA

PAGSS

PAGF

PAGD

PAGS

PAGT

Source: CGE simulations Notes: PAG (increasing budget on CPAGRI), PAGGA (PAG + diverting budget from CPADMN), PAGSS (PAG+ diverting budget from CSSERV), PAGF (PAG + grants from abroad), PAGD (PAG+ increasing direct tax rate), PAGS (PAG + increasing sales tax rate), and PAGT (PAG + increasing tariff rate).

7.5.3 Factor markets Following public services (APAGRI) expansion, the average wage rates for most skilled labor categories increase in many of the policy experiments (see Table 7.3). Public service expansion also increases the marginal revenue product of capital, in effect, the factor income of capital increases albeit its average wage rate remains unchanged. The factor market (demand, wage rates, and income) effects of diverting schemes are relatively low as they only shuffle factor demand among public activities. The foreign transfer and direct tax schemes dampen negative effects on agricultural factors. In contrast, the latter scheme, slightly dampen the positive effects on non-agricultural factors. In spite of these, it can overall be said that with the exception of diverting schemes, financing schemes have little (±) marginal effect on average wage and income of different factors compared to the benchmark scheme (see Table 7.3).

7.5 Economy-wide results and analysis

95

Table 7.3-Factor market effects of adaptation finance Simulations (% change)

Factor income (YF)

Economy-wide wage (WF)

Variable

Account

Factor

FLAB0

Agricultural labor

-4.2

-2.0

-1.4

-3.2

-3.6

-4.4

-4.7

FLAB1

Administrative labor

30.8

19.0

19.9

31.0

30.3

28.7

30.0

FLAB2

Professional labor

22.5

13.5

-1.9

22.8

22.0

20.6

21.7

FLAB3

Unskilled labor

11.8

6.9

5.6

12.0

11.4

10.1

11.2

FLAB4

Skilled labor

4.9

2.6

2.4

4.7

4.5

3.0

4.3

FLND

Cropland

-4.4

-2.1

-1.5

-3.3

-3.8

-4.5

-4.9

FLAB0

Agricultural labor

-4.2

-2.0

-1.4

-3.3

-3.6

-4.3

-4.7

FLAB1

Administrative labor

15.5

2.0

5.7

15.7

15.0

13.7

14.8

FLAB2

Professional labor

16.8

7.3

-6.6

17.0

16.3

14.9

16.0

FLAB3

Unskilled labor

3.8

1.3

0.0

3.8

3.4

2.1

3.8

FLAB4

Skilled labor

8.5

3.4

3.7

8.4

8.1

6.7

7.8

FLND

Cropland

-4.3

-2.1

-1.5

-3.8

-3.8

-4.5

-5.0

FTLU

Tropical livestock units

-2.6

-1.3

-0.8

-1.7

-2.3

-3.0

-2.5

FCAP

Capital

5.4

1.9

1.4

5.6

5.0

3.8

4.6

PAG

PAGGA

PAGSS

PAGF

PAGD

PAGS

PAGT

Source: CGE simulations Notes: PAG (increasing budget on CPAGRI), PAGGA (PAG + diverting budget from CPADMN), PAGSS (PAG+ diverting budget from CSSERV), PAGF (PAG+ grants from abroad), PAGD (PAG+ increasing direct tax rate), PAGS (PAG+ increasing sales tax rate), and PAGT (PAG+ increasing tariff rate).

7.5.4 Households’ welfare Diverting approaches do not generate new employment, and hence they imply the least absolute value of total households’ welfare effects (see Table 7.1 and Figure 7.1). There is no new public spending implies two more things. First, factor income changes are relatively low and thus urban households’ welfare gain become smaller. Second, households’ saving burden get lesser and thus the rural households’ welfare loss become smaller. The PAGF simulation reduces the savings burden on households (hence dampen rural households’ welfare loss from -2.1% to -1.3%) and yet creates employment (hence improves urban households’ welfare gain from 1.5% to 2.2%). Relatively, the direct tax scheme implies the worst welfare effect (-0.1%) to urban households. This accrue to the considerable share of income tax in the urban households’ total expenditure (see Table 4.8). As I pointed out in the previous section, PAGS can be considered as tax on households’ total consumption expenditure. Simply put, it can be considered as lump-sum tax on households’ income left after saving and direct tax. Neither its distortionary effects are expected to be high as the calibration is based on low elasticities. As a result, generally, we see little difference between PAG and PAGS in terms of household welfare effects (see Figure 7.1). Urban households’ are slightly worse off under PAGS which may attribute to the distortionary effects due to substitution of domestic varieties by imports. Tariffs scheme (PAGT) bears the worst aggregate households’ welfare

96

7. Public Finance for Adaptation

effect. Only three commodities are subject to import tariffs (CCCROP, CMAN, and CMMAN). Thus, the benchmark tariff rates are higher than the income as well as sales tax rates. Increasing import tariffs increases import prices relative to domestic prices which reduces imports but increases the demand and prices of domestic varieties. The combined effects result in relatively worse aggregate households’ welfare effect. Nevertheless, the distortionary effects are still small as one of the imports (CMMAN) is entirely imported with no domestic substitute while the other two (CCCROP and CMAN) have low substitution elasticities. The households’ welfare effects presented in Figure 7.1 are also robust to ±25% of the elasticities presented in Table 4.11 (see Table A4 in the Appendix). Figure 7.1-Households’ welfare effects of public adaptation finance 2.5 Equivalent Variation (%)

2 1.5 1 0.5

RURH

0 -0.5

PAG

PAGGA PAGSS PAGF

PAGD

PAGS

PAGT

URBH

-1 -1.5 -2 -2.5

Source: CGE simulations Notes: PAG (increasing budget on CPAGRI), PAGGA (PAG + diverting budget from CPADMN), PAGSS (PAG+ diverting budget from CSSERV), PAGF (PAG + grants from abroad), PAGD (PAG + increasing direct tax rate), PAGS (PAG + increasing sales tax rate), and PAGT (PAG + increasing tariff rate).

Figure 7.1 implies that urban households may shoulder the burden of adaptation finance through taxes. In contrast, adaptation finance may slightly dampen the rural households’ welfare loss. Of course, both household groups are better off under all finance schemes if we factor in the avoided welfare damages due to climate change (see Figure 5.1). Results depicted in Figure 7.1 also corroborate with the theoretical and empirical discussion elsewhere. For instance, the marginal welfare costs of sales tax is lower than that of income tax and import tariff in many African countries including Ethiopia (Auriol and Warlters, 2012), and that of import tariffs in other developing countries (Devarajan et al., 2002 cited in Auriol and Warlters, 2012; Burfisher, 2011). 7.6

Regional projections and analysis

For each of the adaptation finance schemes, the regional effects are determined by the sectoral output effects (Table 7.2) and the regional economic structure (Table 4.13). Table 7.4 depicts the regional effects of alternative adaptation finance schemes which, generally, resemble the regional effects of the benchmark scheme (PAG). Diverting from public general administration budget (PAGGA) shows negative effects to all regions. The sales tax (PAGS) scheme implies

7.7 Conclusions

97

relatively narrow range of variation of the regional effects. Diverting public resources from public administration (PAGGA) and social services (PAGSS) will bear the highest costs to the urban regions. For Addis Ababa, the effects of tax schemes (PAGD, PAGS, and PAGT) are four times of the effects on national average. This is expected as nearly half of the total tax revenue from regions is contributed by Addis Ababa (MoFED, 2015). Table 7.4-Regional effects of public adaptation finance Simulations (% change) Region

PAG

Ethiopia

-0.2

-0.5

-0.4

-0.3

-0.3

-0.3

-0.3

1.2

-0.1

1.0

1.2

1.2

1.2

1.3

Afar

-0.2

-0.9

-0.3

-0.3

-0.3

-0.2

-0.2

Amhara

-0.1

-0.2

-0.3

-0.2

-0.1

-0.1

-0.1

Oromia

-0.2

-0.3

-0.4

-0.4

-0.3

-0.3

-0.3

Somali

-0.2

-0.5

-0.2

-0.2

-0.2

-0.2

-0.3

0.0

-0.6

-0.3

-0.1

0.0

0.0

0.0

Southern NNP

-0.2

-0.3

-0.3

-0.3

-0.2

-0.2

-0.2

Gambella

-0.1

-0.9

-0.6

-0.3

-0.2

-0.1

-0.2

Harari

-0.3

-1.7

-1.4

-0.4

-0.3

-0.3

-0.4

Addis Ababa

-1.3

-1.7

-1.4

-1.3

-1.3

-1.2

-1.3

Dire Dawa

-0.8

-0.9

-0.7

-0.9

-0.8

-0.8

-0.9

Tigray

Benishangul-Gumuz

PAGGA

PAGSS

PAGF

PAGD

PAGS

PAGT

Source: Table 7.2 and Table 4.13 Notes: PAG (increasing budget on CPAGRI), PAGGA (PAG + diverting budget from CPADMN), PAGSS (PAG + diverting budget from CSSERV), PAGF (PAG + grants from abroad), PAGD (PAG + increasing direct tax rate), PAGS (PAG + increasing sales tax rate), and PAGT (PAG + increasing tariff rate).

The regional effects discussed in section 6.6 and in this section imply that regional dimension is very important in adaptation policy-and-decision making process. Among others, it may influence and need adjustments in federal block-grant formula that may have further political implications. This substantiates Brown (2008) which contends that environmental change induced economic changes may have political implications. 7.7

Conclusions

Climate change represents a new task for public policy-and -decision makers. Adaptation in any sector is not a stand-alone environmental problem. Consequently, the government may take further steps to devise new adaptation finance schemes for agriculture. It may divert resources from other public budget accounts, seek for foreign grants or increase tax revenue by raising tax rates. Such fiscal moves usually have considerable general equilibrium effects in LDCs like Ethiopia. This is what this chapter attempted to look at.

98

7. Public Finance for Adaptation

The economy-wide and regional effects of alternative adaptation finance schemes are similar to the case of adaptation financed by increasing spending on public agricultural services. It can also be roughly argued that the aggregate effects vary little across alternative fiscal schemes. However, different finance schemes affect different macroeconomic variables, sectors, factors, households, and regions differently. For instance, transfers from abroad may appreciate real exchange rate and hence may affect exports at macroeconomic level. However, compared to other schemes, transfers from abroad imply better outcome to factor markets and households’ welfare. The welfare gain for urban households reaches its maximum under foreign finance scheme but minimum under direct tax rate scheme. Rural households’ welfare loss reaches its minimum when finance is diverted from social services and maximum with tariffs scheme. Diverting schemes relatively worsen the regional effects on urban regions as these regions benefit least from expanding public service (and finance) for agriculture. The regional effects may matter for policy-and-decision makers as the main source of regional governments’ budget is federal block-grant (MoFED, 2009; Eshetu et al., 2014; Lanos et al., 2014). There will be implications for the formula and criteria used to allocate the federal blockgrant. The first may be related to how to regard adaptation finance for agriculture. Shall it be regarded as capital or recurrent budget? This is critical, at least, in the context of the present budget allocation criteria which focus on infrastructural deficiency (for capital budget) and population size and revenue generating capacity (for recurrent budget) (MoFED, 2009). Second, what set of criteria shall be appropriate to allocate adaptation finance among the regions? Would it be appropriate to align to the exposure (which primarily depends on existing environmental conditions of different regions) or the economic vulnerability of the regions to climate change (which depends on the relative share of agriculture in region-wide GDP)? In the case of economic vulnerability, shall it be based on sectoral (i.e., agricultural GDP) effects or regionwide GDP effects of climate change. The agricultural vulnerability criterion will indicate different ranking from overall economic vulnerability of regions. Third, in using either of the criteria, would it be necessary to attach to the adaptation potential of the regions? For instance, regardless of the sign and magnitude of potential effects, regions with better irrigation potential may receive the largest share of federal adaptation finance as block-grant. Such allocation aims to maximize country-wide macroeconomic benefits of adaptation regardless of distributional effects. When we consider all of these, regional effects may matter even more than the economy-wide effects of adaptation costs and finance. The findings of this chapter along with that of Chapter 6 imply that adaptation costs and finance for agriculture have considerable economy-wide and regional effects. Among others, they impede the key sectors (i.e., manufacturing and private sectors) and key actor (i.e., government saving) of structural transformation in Ethiopia. In other words, agricultural adaptation costs and finance may deter structural and spatial transformations which are at the heart of the country’s recent economic growth and transformation plans (see, for example, MoFED, 2010; NPC, 2016). The policy implications of these tradeoffs are vital, especially, when we consider the opportunity costs of agricultural adaptation finance, and the fact that agriculture is only one of the economic sectors that need public adaptation. But, what would happen if the government prefer structural transformation to adaptation in agriculture. What would be the role of structural change in climate-resilient development? The next chapter attempts to deal with this question.

8 8.1

Climate-Resilient Development Introduction

Climate change induced shocks in agriculture impinge on the Ethiopian economy (Chapter 5). The public adaptation costs to fully neutralize the first-order shocks in agriculture is detrimental to government saving and output of some sectors and regions (Chapter 6). Even though alternative adaptation finance schemes may help to reduce the pressure on government saving, the resource-pull effects persist and influence manufacturing and private services output, and hence urban GDP (Chapter 7). In the Ethiopian context, the opportunity costs of declining government saving are enormous as the government plays critical role in urban, transport, and energy infrastructure development. The resource-pull effects of adaptation may strain the structural and spatial transformation in the economy. The concerns over these trade-offs and opportunity costs mount as we think of the uncertainties with respect to climate change, its biophysical impacts, and adaptation policy effectiveness, and the necessity of public adaptation in other sectors of the economy. Policy decisions with regard to adaptation in any sector shall factor in such uncertainties (Tanner and Horn-Phathanothai, 2014). Adaptation policy decisions shall be able to bring benefits to the society in different climate change regimes including in absence of climate change (Fankhauser and Burton, 2011; Tanner and Horn-Phathanothai, 2014). Moving from ‘adaptation to climate change’ to ‘climate-resilient development’ approach may help in this regard (Fankhauser and Schmidt-Traub, 2011; Mendelsohn, 2012). Climate-resilient development approach, among others, regards structural change as a general form of adaptation (Mendelsohn, 2012; Millner and Dietz, 2015). This chapter aims to investigate this argument. It re-examines the economy-wide and regional effects of climate change under three alternative cost-free exogenous structural change scenarios. The results of the chapter indicate that structural change may contribute to climate-resilient economic development in Ethiopia. The remainder of the chapter presents the scientific discourse on structural change and climate change (8.2), structural change in Ethiopia (8.3), the materials and methods (8.4), the economywide effects (8.5) and regional effects (8.6), and the general conclusions (8.7). 8.2

Structural change and climate change

The continuum of measures to reduce the vulnerability of a given economy to climate change range from ‘pure adaptation’ to ‘pure development’ (McGray et al., 2007; Fankhauser and Schmidt-Traub, 2011; Tanner and Horn-Phathanothai, 2014). Pure adaptation approach requires incremental activities on top of business-as-usual development plans and programs of the sector or economy (Fankhauser and Schmidt-Traub, 2011). It is justified by the fact that development in LDCs is contingent on adaptation (Millner and Dietz, 2015). Such approach include measures which specifically aim to address incremental damages, for instance, decrease in crop and livestock yields due to climate change. As such, financing ‘pure’ adaptation competes for public resources for financing conventional development measures (Fankhauser and Schmidt-Traub, 2011). What I pursue in chapter 6 and 7 of this study fall in such approach. In contrast, ‘development’ as adaptation focus on basic development objectives (McGray et al., 2007; Mendelsohn, 2012). It includes broader measures that can generally reduce vulnerability of an economy to multiple stresses (of which climate change is one). This approach regards

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 A. W. Yalew, Economic Development under Climate Change, https://doi.org/10.1007/978-3-658-29413-7_8

100

8. Climate-Resilient Development

development as the best adaptation measure (Millner and Dietz, 2015). Structural change is one of such ‘development adaptation’ measures as it represents a generic adaptive capacity to climate change. Generally speaking, structural change in an economy refers to the changes in how the given economic system functions. It refers to shifts in the composition and patterns of employment, production, trade, consumption, and national income. An economy is often regarded as structurally transforming when the relative contribution to national employment, GDP, and export earnings move from agriculture to manufacturing, and then to services. Yet, with respect to LDCs, structural change may refer to transformation from traditional, less-diversified and subsistence agriculture to more modern, mechanized, and commercialized agriculture, and then to more urban and diversified manufacturing and service economy. Structural change is usually regarded as vital success in economic development. Investments in human capital, infrastructure, and market institutions are the common strategies to foster structural change in LDCs. As such, structural change is crucial for sustained economic growth, and for building an economy resilient to foreseen and unforeseen demand and supply shocks. 8.3

Structural change in Ethiopia

Ethiopia has demonstrated a remarkable economic growth, 10% per annum on average, in the last twelve years (NBE, 2016; NPC, 2016). The contribution of agriculture to national GDP has started to decline gradually and is overtaken by services in recent years (NBE, 2016; NPC, 2016). Despite this rapid economic growth, however, structural change is unsatisfactory in the country (Dorosh and Thurlow, 2011; NPC, 2016). Domestic savings (exports) fall short to finance domestic investment (imports) (NPC, 2016). The low and slow growth in manufacturing sector and exports are setbacks to structural transformation in the country, and call for aggressive investment in productivity, infrastructure, human capital, and markets (NPC, 2016). Government of Ethiopia regards itself as the main actor in leading the structural transformation in the country. The second Growth and Transformation Plan of Ethiopia (GTP II: 2015-2020) bluntly underlines that the government is “fully committed to mobilize the necessary resources including capacity for implementation of the Plan” (NPC, 2016, p.2). On the other hand, the government aims to keep its fiscal deficit relative to the GDP below 3% (NPC, 2016). This will require to broaden tax base, to increase tax rates, to improve tax administration, and to control misuse of public resources (NPC, 2016). As such, agricultural adaptation costs (see Chapter 6) and adaptation finance (see Chapter 7) bear opportunity costs. It holdbacks structural transformation in the economy and distress fiscal surplus. Therefore, planned public adaptation in agriculture impedes the two main macroeconomic goals of the current growth and transformation plan of the country. This raises a question, among others, whether it is a wise policy decision to engage in planned public adaptation in agriculture while the country is thirsting structural change. On the other hand, as we see in section 8.2, structural change itself may contribute to climate-resilient economic development. The central objective of this chapter is to assess the role of structural change to dampen the adverse economy-wide and regional effects of climate change. I assume that the government gives priority and continues to undertake measures to foster structural transformation. This, among others, may be reflected in labor market structure, and in transport and trade margins in near future. The main hypothesis of the chapter is that such changes may offset some of the

8.4 Materials and methods

101

adverse effects of climate change. The aim of the chapter is to draw some policy implications on the role of structural change as generic adaptation measure to climate change. Yet, it is important to keep in mind that the objective of this chapter is not to compare and contrast the benefits and costs of ‘structural change as adaptation’ and ‘planned public adaptation in agriculture’. So doing does not make sense for three main reasons. First, adaptation in agriculture is with specific aim (to increase productivity) in a specific sector (agriculture) to a specific stress (climate change). In contrast, structural change involves a wide spectrum of measures related to productivity, interconnectivity, flexibility, substitutability, and the likes in many sectors and locations. Structural change aims at multiple foreseen or unforeseen, local or global, supply or demand stresses. Second, explicit costs of structural transformation are hardly traceable. Third, at least in this study, modeling structural change is different from modeling adaptation in agriculture. Therefore, the costs and benefits of structural transformation as generic form of adaptation to climate change shall be seen only from the synergies or conflicts it may have with the current and future national socio-economic goals of the country. 8.4

Materials and methods

It is not an easy exercise to assess the benefits of structural change using a static CGE model. One cannot fully track the process and the costs of structural transformation leading to structural change. In this chapter, I treat structural change as a cost-free exogenous shock to the economy that simultaneously happen with climate change. Structural change may accrue to the rapid economic growth the country is scoring (NBE, 2016), to huge public investment in human capital, transport and communications, energy, institutions, and markets (NPC, 2016; NBE, 2016; MoE; 2016; WDI, 2016), or to the growing attention to micro and small enterprises, urban and regional development (NPC, 2016). On the basis of the observed and planned public investment in human capital, demographic structure of the country, and public investment in transport and energy infrastructure, I construct three reasonable structural change scenarios which I discuss below. 8.4.1 Changes in labor skill Labor markets in poor countries are usually segmented (Fields, 2011). We shall maintain and reflect the labor market segments in LDCs in empirical economic modeling as “some parts of the labor market operate in a qualitatively different manner from others” (Fields, 2011, p.18). The calibrated model consists of five labor market segments which include agricultural workers (FLAB0), unskilled workers (FLAB3), skilled workers (FLAB4), professional and technical associates (FLAB2), and administrative workers (FLAB1). Labor markets are among markets in an economy where structural change manifest itself. Structural change will deploy labor (of course many of the productive factors) from primary (e.g. agriculture) to secondary (e.g. construction), and then to tertiary (e.g. transport and communications) sectors. Structural change is also represented by accumulation of labor skills and increase in labor productivity. I presume a scenario (Scenario A) that represents changes in labor skill that narrows down skill difference across labor segments (occupations). Under this class of scenarios, for a portion of the agricultural labor, working in agriculture would be a matter of preference not of a skill. When the agriculture sector is hit by external shocks, such as climate change, that portion of

102

8. Climate-Resilient Development

agricultural labor would easily ‘migrate’ (switch) in to other labor segments. Climate change induced outmigration decreases agricultural labor supply (FLAB = 25.4 million workers) and increases the labor supply in other labor segments by the same units of workers. Because farmers usually do not possess skills other than farming, what normally would happen is from agricultural (FLAB0) to elementary occupation (FLAB3 = 1.4 million workers). This serves as benchmark occupational migration simulation. However, the observed trends in public spending on education signals that structural change in the labor market is inevitable. The ratio of education expenditure to the total government expenditure has increased in the last two decades (MoFED, 2015; MoE, 2016). For instance, it grew from 15% in the period of 1998-2005 to 23% in the period of 2007-2014 (MoFED, 2015). In parallel, the share of higher education expenditure in the total education expenditure is increasing overtime (MoE, 2016; WDI, 2016) yielding palpable progress in human capital. Between 2004 and 2015, the adult literacy rate (> 15 years old), and the youth literacy rate (15-24 years old), respectively, increased from 36% to 49%, and from 50% to 69% (WDI, 2016). The enrolment rate of alternative basic education (for adults), primary school, secondary school, technical and vocational training colleges, and universities is steadily increasing overtime (MoE, 2016). For instance, the net enrolment rate in Grade 1-8 has tremendously increased from 22% in 1995/96 to 94% in 2014/15 (MoE, 2016). In the period of 2004/05-2014/15, the number of undergraduates and postgraduates from colleges and universities grew at average annual growth rate of 30% and 32%, respectively (MoE, 2016). These observed trends in the education sector may lead to structural change in labor markets in which it would be possible to migrate into the other three labor occupations that require specific skills (FLAB4, FLAB2, and FLAB1). Therefore, depending on the strength of the structural change, I presume three additional occupational migration scenarios. First, with investment in basic literacy and skill trainings, the current agricultural labor (FLAB0) can easily move into ‘skilled’ labor category (FLAB4). Skilled labor (FLAB4 = 4.8 million workers) group represents the second largest group in the total labor force. It includes workers that have some skills obtained through formal education and training, experience or informal training (EDRI, 2009). It includes shop and market sales workers, crafts and related, trade workers, plant machine operators and assemblers, and clerks (NLFS, 2005; EDRI, 2009). Second, with considerable investment in education and trainings, the occupational migration from agricultural occupation (FLAB0) to professional and technical associates occupation may be possible (FLAB2 = 0.5 million workers). Third, especially in the long-run, a portion of agricultural labor today may acquire a set of knowledge and skills that would qualify them as administrative workers (FLAB1 = 0.1 million workers). Migration between occupations keeps the total labor force in the economy fixed at observed level, i.e., 32.2 million labor units. Absence from the labor market at the time of education and training, the period of time required to finish and attain the set of skills to fit to a specific labor market segment, and similar issues are beyond the chapter objective. Of course, the government can influence the future of the current child workers by investing in education and training. Working children (10-15 years old), the bulk of which are in agriculture, account for about 16% of the country’s total labor force (NLFS, 2005; 2013).

8.4 Materials and methods

103

8.4.2 Changes in total labor supply We discuss in section 3.3 that Ethiopia is country of young population. The population below 15 years old and above 65 years, respectively, accounts for 44% and 3.5% of the total population (NLFS, 2013). Therefore, every year, the labor entering the market is by far greater than the labor leaving. This is purely due to the demographic structure of the country. However, where the net labor supply is added can represent structural change. Therefore, I design a set of experiments under scenario B to represent the case that the economy gets an extra 0.5 million work force to allocate in ether of the five labor segments. Implicitly, I presume that investments in education and human capital will offer the country the possibility to allocate the extra 0.5 million work force indifferently among the five labor segments. To control the economy-wide effects that pertains to total labor supply increase, the benchmark simulation in scenario B’s class of experiments would be allocating the extra labor to agriculture. Therefore, the hypothesis for this scenario is that allocating the extra labor force to nonagricultural occupations would offset by relatively higher magnitude the economy-wide and regional effects of climate change better than allocating to agricultural occupation. 8.4.3 Changes in transaction costs Marketing margins or transaction costs are usually high in developing countries where transport and marketing infrastructure are underdeveloped breeding market inefficiency (Lofgren, 2001). Ethiopia is a typical example. Transportation problem is still a barrier and main contributor to higher transaction cost and market inefficiencies, especially, in agricultural product markets. About 83 percent of gross grain marketing margins accrue to physical marketing costs related to transport, handling, and other marketing activities (Gabre-Madhin, 2001). Transport costs account for 6-21% of market prices per quintal of maize, sorghum, and millet in rural villages surrounding Atsdemariam town, in northwest Ethiopia (Stifel et al., 2016). With recognition of this, transport and communications are among the sectors in which Ethiopia has remarkably invested during the last two decades. For instance, between 2009/10 and 2013/14, expenditure on roads, transport, and communications accounts for about 38% of the total government capital expenditure (MoFED, 2015). As a result, road density and access to roads have profoundly improved (ERA, 2010; NBE, 2016). Between 2001/02 and 2014/15, road density per 1000 people and road density per 1000 km2, respectively, increased from 0.5 km to 1.2 km and from 30 km to 100 km (NBE, 2016). 82 Due to the Universal Rural Road Access Program (URRAP), the number of rural kebeles (the lowest administrative level in Ethiopia) connected to roads has tremendously increased (ERA, 2010; NBE, 2016). The average annual growth rate of rural roads connecting rural villages and communities in 2007/08-2014/15 period was 162% (NBE, 2016). Consequently, the average distance to reach to a nearby allweather road is declining, for instance, from 21 km in 1997 to 11 km in 2010 (ERA, 2010). The government aims to go further and double the road density by 2019/20 and halve the distance to reach to all-weather roads (see also NPC, 2016). Recently, the country started constructing

82

ERA (2010) reports higher figures as it includes ‘community’ roads.

104

8. Climate-Resilient Development

railway networks in eight corridors of the country. 83 The improvements in road connectivity apparently contribute to integrate the domestic market which lowers transport and trade margins (NPC, 2016) and narrows the regional price disparities (NBE, 2016). Per contra, lower marketing margins and costs can be regarded as indicators of market efficiency. Market efficiency improves with expansion in transport and communications and structural change in an economy. Therefore, in scenario C, I arbitrarily presume that the observed and projected trends of transport and communication networks will reduce transaction costs in the economy by 10% in all market-commodities. 84 Other things remaining constant, reducing transaction cost increases producer revenue for a given demand price or reduce consumer expenditure for a given producer price. For a given set of world export and import prices, decreasing transaction costs increases the domestic receipt from exports but decreases the domestic expenditure to imports. As such, changes in transaction costs have welfare effects. 8.4.4 Modeling into the CGE model Unlike the case of full planned public adaptation in agriculture (see Chapter 6), structural change will neither prevent nor modify the primary impacts of climate change on agriculture. The crop, livestock, and agricultural labor supply shocks remain intact under all structural change scenarios. The hypothesis here is that structural change as adaptation absorbs the primary effects in agriculture but dampen the general equilibrium effects of climate change. As I made it clear earlier, I presume and model as if the structural change and climate change happen simultaneously. Therefore, the climate change induced shocks are complemented with shocks in representing the three alternative classes of structural change scenarios. In scenario A, the total labor supply of the economy remains the same. Structural change is represented by giving the possibility of migration from agricultural to other labor skill segments. The exercise involves negative shock to the shift parameters of grain and livestock activities, and agricultural labor supply (FLAB0) plus, depending on the simulation, positive shock to the supply of unskilled labor (FLAB3), of skilled labor (FLAB4), of professional and technical associates labor (FLAB2), or of administrative labor (FLAB1). In scenario B, the economy gets 0.5 million extra labor force. The strength of structural change is represented by the labor segment in to which the extra labor force is deployed. The exercise involves negative shock to the shift parameters of grain and livestock activities plus, depending on the simulation, positive shock to the supply of agricultural labor (FLAB0), of unskilled labor (FLAB3), of See more at the Ethiopian Railways Corporation (http://www.erc.gov.et/). Accessed on 24 April 2015. Transaction costs refer to the marketing margins. They are marketing costs representing the difference between what the final purchaser pays and the producer receives excluding commodity taxes. Transaction margins usually accrue to trade and transport margins. Trade margins are part of the wholesale and retail activity output (ATSER in this model) that involve in supplying a bundle of goods to “convenient locations and making them easily available for customers to buy” (SNA, 2008, p.113). The value of trade services is measured by “total value of the trade margins realized on the goods they purchase for resale” (SNA, 2008, p.113). According to SNA (2008, p.113) trade margin is “the difference between the actual or imputed price realized on a good purchased for resale and the price that would have to be paid by the distributor to replace the good at the time it is sold or otherwise disposed of”. Transport margins are transportation costs charged and invoiced separately for distributing the trade goods (SNA, 2008). As such, transport margin is recorded as marketing margin increasing the transaction costs of marketed output (Lofgren et al., 2002). Transport margins, thus, are different from transport services which are used either as intermediate inputs in production process (by activities) or final consumption goods (by households, and the rest of the world) (Lofgren et al., 2002; SNA, 2008). 83 84

8.5 Economy-wide results and analysis

105

skilled labor (FLAB4), of professional and technical associates labor (FLAB2), or of administrative labor (FLAB1). Scenario C represents shocks to the scale parameters of grain and livestock activities, occupational migration from FLAB0 to FLAB3, and to transaction costs in all domestic sales, exports, and imports. 85 In the economy-wide and regional analysis in the subsequent sections, I focus and present the case of EPIC climate change impact scenario. To remind the reader, the EPIC impact scenario represents shocks of -26% (grain productivity) and -5% (livestock productivity) and -4% (agricultural labor supply). 8.5

Economy-wide results and analysis

Table 8.1 below presents the macroeconomic, sectoral output, and households’ welfare effects of climate change with alternative structural change scenarios. For a given row, columns depict the role of arbitrarily constructed cost-free structural change scenarios to offset the economywide effects of climate change discussed in section 5.5 (i.e., Scenario 0 in Table 8.1). The first class of experiments (Scenario 0) represent the base scenario effects are EPIC-P (only productivity effects) and EPIC-M3 (productivity shocks plus migration to unskilled labor category, FLAB3). The second class of experiments (Scenario A) are related to migration into occupations with skill requirements while the total labor supply of the economy remains fixed. ‘M’ stands for migration and the numbers represent the destination occupation. For instance, EPIC-M2 represents the migration is from agricultural labor (FLAB0) to professional and technical associate labor (FLAB2). Scenario A simulations shall only be compared against EPICM3. The results show that the ripple effects of climate change on the macro-economy, the rest of economic activities, and households’ real consumption would have relatively been lesser. For instance, if deploying the 4% agricultural labor out-migrants in to the FLAB4, FLAB2, or FLAB1 segment was possible, the effects of climate change on GDP would have been lesser by about 20-30% compared to deploying them into the FLAB3 segment. The offsets are strong under EPIC-M2 simulation. The offsets on the aggregate households’ welfare are also similar. Nevertheless, any form of occupational migration increase the real wage rate and factor income of agricultural labor but decrease that of non-agricultural labors. As a result, Scenario A simulations relatively dampen the rural but worsen the urban households’ welfare (see Table 8.1). The third class of experiments (scenario B) represent the cases where to allocate the extra half million labor (L) at the time of climate change. Again, the number in the simulations represents the labor category. As we added new labor force to the economy, all simulations under scenario B offset part of the adverse effects. Therefore, to control the effects that may accrue to the increase in total labor supply, one shall make the comparison against allocating the labor to agriculture (EPIC-L0). Scenario B simulations show that the economy would better be climate-

85 The standard IFPRI-CGE model dubs marketing margins as transaction services (and hence transaction costs) and treat them as special ‘trade’ inputs for wholesale and retail trade services (Lofgren, 2001; Lofgren et al., 2002). They are treated as special intermediate (trade) inputs used in marketing domestic, export, and import commodities (Lofgren et al., 2002). The 2005/06 Ethiopian SAM records the transaction services (and hence transaction costs) as produced/ received by trade (wholesale and retail) service activity/commodity (ATSER/CTSER) (see Table 4.1 and Table 4.3).

106

8. Climate-Resilient Development

resilient if the next generation of labor force is directed towards professional and technical associate labor (EPIC-L2). The effect on the GDP in the case of allocating the extra labor on FLAB2 (-5.7%) is lower than the allocating the extra labor in to FLAB0 (-6.7%) or FLAB3 (7%). The aggregate households’ welfare effects under EPIC-L2 are lower than EPIC-L0 by 1.5 percentage points. However, allocating the extra labor force into labor segments other than agricultural labor will worsen the urban households’ welfare. Table 8.1-Economy-wide effects of climate change Simulations (% change)

Households

Sectoral output

EPIC -L0

EPIC -L3

EPIC -L4

EPIC -L2

EPIC -L1

EPIC -P

EPIC -M3

EPIC -M1

Scenario C

EPIC -M2

Scenario B

EPIC -M4

EPIC -P

Variable

Scenario A EPIC -M3

Macro-economic

Effects

Scenario 0

ABSORP

-6.3

-6.5

-6.0

-4.9

-5.3

-5.5

-5.7

-5.5

-4.6

-4.9

-4.5

-5.0

PRVCON

-8.9

-9.2

-8.5

-6.9

-7.5

-7.8

-8.0

-7.7

-6.6

-7.0

-6.3

-7.0

EXPORTS

-7.8

-7.3

-4.7

-2.5

-3.0

-6.7

-6.7

-5.4

-3.8

-4.1

-2.6

-2.6

IMPORTS

-2.8

-2.6

-1.7

-0.9

-1.1

-2.4

-2.4

-1.9

-1.3

-1.5

-0.9

-0.9

GDPMP

-7.7

-8.0

-7.4

-6.0

-6.5

-6.7

-7.0

-6.7

-5.7

-6.1

-5.5

-6.1

EXR

-3.6

-4.3

-4.2

-4.1

-4.1

-2.9

-3.3

-3.2

-2.9

-3.0

-3.0

-4.1

AGRAIN

-24.0

-25.6

-25.6

-25.6

-25.6

-21.9

-23.1

-23.1

-23.1

-23.1

-23.2

-25.7

ACCROP

-13.1

-15.7

-15.3

-15.9

-15.8

-10.9

-12.6

-12.3

-12.7

-12.6

-10.3

-14.0

AENSET

-10.4

-11.6

-11.5

-11.0

-11.1

-9.0

-9.8

-9.8

-9.3

-9.4

-9.4

-11.2

ALIVST

-11.4

-13.6

-13.6

-13.6

-13.6

-9.8

-11.0

-11.1

-11.1

-11.1

-11.2

-13.7

AFISFOR

-8.1

-10.0

-9.7

-9.3

-9.3

-6.4

-7.5

-7.4

-7.0

-7.1

-6.9

-9.3

AMINQ

2.2

5.1

6.1

6.8

6.3

1.9

3.5

3.9

4.9

4.5

4.0

7.2

ACONS

-0.1

0.0

0.0

0.1

0.1

-0.1

0.0

0.0

0.0

0.0

-0.2

-0.1

AMAN

7.0

13.2

16.6

18.1

16.8

6.0

9.6

11.0

13.4

12.3

12.0

18.9

ATSER

-5.0

-4.8

-3.4

-3.0

-3.2

-4.3

-4.2

-3.6

-3.1

-3.2

-10.5

-10.5

AHSER

-3.2

-2.6

0.3

0.6

1.1

-2.6

-2.3

-0.9

-0.1

0.1

0.6

0.8

ATRNCOM

4.0

5.6

8.7

12.0

11.9

3.4

4.3

5.7

8.0

7.8

4.6

6.4

AFSER

0.5

1.9

3.2

9.7

12.3

0.5

1.2

1.8

6.0

7.3

1.1

2.6

ARSER

0.0

0.2

0.2

0.5

0.7

0.0

0.1

0.1

0.3

0.4

0.1

0.3

APADMN

0.0

0.0

0.0

0.2

0.2

0.0

0.0

0.0

0.1

0.1

0.0

0.0

APAGRI

0.0

0.0

0.1

0.3

0.3

0.0

0.0

0.0

0.2

0.2

0.0

0.1

ASSER

0.1

0.7

1.0

12.0

3.6

0.1

0.4

0.5

7.4

2.2

0.6

1.2

AOSER

3.3

13.5

11.1

37.6

28.6

2.8

8.1

6.7

21.8

16.6

6.3

17.0

TOTAL

-7.6

-7.6

-7.0

-5.3

-6.0

-6.7

-6.8

-6.5

-5.4

-5.8

-7.1

-7.5

RURH

-9.4

-9.9

-9.1

-7.1

-7.7

-8.2

-8.6

-8.2

-6.6

-7.1

-6.6

-7.6

URBH

-10.1

-11.2

-11.1

-13.4

-12.7

-8.6

-9.3

-9.3

-10.9

-10.4

-9.0

-10.7

TOTAL

-9.6

-10.2

-9.6

-8.6

-8.9

-8.3

-8.8

-8.5

-7.7

-7.9

-7.2

-8.4

Source: CGE simulations

8.6 Regional projections and analysis

107

The importance of reducing transaction costs, scenario C, is straight forward. It would offset the effects on the aggregate output (GDP) and households’ consumption by around 2 percentage points. Reducing transaction costs also facilitates imports and exports which in turn contributes to control domestic to international prices ratio fluctuations. Transaction costs are trade and transport margins needed (and hence incurred) when commodities are marketed. Thus, for a given share of market commodities, the benefits of reduced transaction costs will pass directly to households’ welfare. The effects on rural households’ welfare is clearer than effects on urban households. This may attribute to relatively increased net farm revenue, for instance, from increasing domestic receipts from exports due to reduced transaction costs. This offsets part of rural households’ real consumption lost due to climate change induced fall in agricultural output. One may raise two more question with regard to Table 8.1. First, the effects on grain and livestock activities (and the indirect effects on other agricultural activities) hardly changes under all structural change scenarios. This is expected as structural change would do nothing in preventing or even modifying the direct biophysical impacts of climate change. Second, the offsetting role of structural change scenarios are not big as such. This is because the macroeconomic significance of agricultural output, our most impacted sector, still remains the same. This implies that it is necessary to expand the non-agricultural sectors parallel to upgrading labor skills and reducing transaction costs. The economy-wide analysis, therefore, points that improving labor skill, market connectivity and efficiency contribute to climate-resilient economic development in Ethiopia. This is a cobenefit on top of the expected benefits of structural change under normal circumstance besides being compatible with the current growth and transformation plans of the country. 8.6

Regional projections and analysis

In line with the economy-wide effects, the regional effects (see Table 8.2) also show that migration and allocation to professional labor (EPIC-M2 and EPIC-L2) yield better outcome. Value-added GDP of urbanized regions expand, particularly, in labor market oriented structural change scenarios–Scenario A and B. For instance, compared to the base scenarios (EPIC-P and EPIC-M3), regional effects to Addis Ababa increase by one to five percentage points. This is expected as Addis Ababa and other urban regions of the country mainly depend on manufacturing and services that will benefit from increased skilled labor supply. Migration into the unskilled labor segment (EPIC-M3) worsens the adverse effects in agrarian regions such as in Amhara, Oromia, and Southern NNP (see Table 8.2) whereas migration into other skill categories (EPIC-M4, EPIC-M2, and EPIC-M1) will slightly dampen (or leave unchanged) the productivity shock induced effects. Similarly, adding the extra labor force to nonagricultural labor segments (scenario B) increase the gains, especially, to urban regions. In scenario C, sectoral output of trade activities (ATSER) decline (see Table 8.1). Consequently, scenario C worsens the regional GDP loss in Harari and Dire Dawa regions where the share of ATSER is relatively large (see Table 4.13). Nevertheless, the contribution of Harari and Dire Dawa in the Ethiopia-wide value-added GDP is hardly 1.2% (see Table 4.12). Therefore, the role of reducing transaction costs are still important to the Ethiopia-wide GDP.

108

8. Climate-Resilient Development

Table 8.2-Regional effects of climate change with structural change Simulations (% change)

EPIC -M3

EPIC -P

EPIC -L1

EPIC -L2

EPIC -L4

Scenario C

EPIC -L3

EPIC -L0

EPIC -M1

Scenario B

EPIC -M2

EPIC -M4

EPIC -P

Scenario A

EPIC -M3

Scenario O Region

Ethiopia

-7.6

-7.6

-7.0

-5.3

-6.0

-6.7

-6.8

-6.5

-5.4

-5.8

-7.1

-7.5

Tigray

-5.7

-5.1

-4.6

-2.7

-3.5

-5.1

-4.9

-4.7

-3.4

-3.9

-5.3

-4.9

Afar

-3.9

-3.2

-2.3

0.0

-0.7

-3.4

-3.1

-2.7

-1.1

-1.6

-3.4

-2.8

Amhara

-10.2

-10.3

-9.7

-8.3

-8.9

-9.1

-9.3

-9.1

-8.1

-8.5

-9.4

-9.9

Oromia

-9.1

-9.4

-8.8

-7.4

-8.0

-8.0

-8.3

-8.1

-7.0

-7.4

-8.6

-9.3

Somali

-6.6

-7.0

-6.4

-5.1

-5.5

-5.6

-5.9

-5.7

-4.7

-5.0

-6.4

-7.2

BenishangulGumuz

-6.8

-6.4

-5.7

-4.2

-5.0

-6.2

-6.0

-5.7

-4.6

-5.1

-6.2

-5.9

Southern NNP

-9.3

-10.1

-9.6

-8.6

-9.0

-8.1

-8.6

-8.4

-7.7

-7.9

-8.7

-10.0

Gambella

-4.9

-4.5

-3.6

-1.4

-2.3

-4.3

-4.1

-3.8

-2.2

-2.8

-4.7

-4.5

Harari

-3.0

-2.4

-1.5

1.5

0.2

-2.6

-2.3

-1.9

0.1

-0.8

-3.7

-3.2

0.7

2.4

3.3

7.2

6.1

0.6

1.5

1.9

4.4

3.7

0.9

2.6

-1.4

-0.1

1.0

4.3

3.3

-1.2

-0.6

-0.1

2.1

1.5

-2.1

-0.9

Addis Ababa Dire Dawa

Source: Table 8.1 and Table 4.13

In conclusion, all structural change scenarios tend to offset the effects of climate change on national GDP (see the ‘Ethiopia’ row in Table 8.2). However, the type of structural change may be important for regional effects. The offsetting effects of labor related structural change scenarios lean towards urban regions while that of transaction cost related scenarios lean towards agrarian regions. 8.7

Conclusions

Planned public adaptation costs and finance for agriculture may distress government saving, and deter structural change. The tradeoffs and opportunity costs will increase as we consider the uncertainties in climate change, biophysical impacts of climate change, economic impacts of climate change, and the effectiveness of adaptation measures (Heal and Millner, 2014; Tanner and Horn-Phathanothai, 2014). In light of this, policy makers may be interested to know whether the current and future public investments in education, transport, institutions, and markets will contribute to climate-resilient development. In this chapter, I construct a set of scenarios representing changes in some structural features of labor and commodity markets. In general, the economy-wide and regional effects show that the aforementioned structural change scenarios could underpin climate-resilient economic development in Ethiopia. The labor focused structural change scenarios would help to expand production in urban regions although impinge on the urban households’ welfare. The commodity market (i.e., transaction cost) focused structural change scenario effects are more visible in terms of households’ welfare, especially, to rural households. The qualitative information gleaned from this chapter substantiates the view that structural change underpins the resilience of LDCs to multitude of shocks including climate change.

9

Conclusions and policy implications

Climate change has become undisputable. It imposes formidable risks to the economic prospects of low-income countries where agriculture plays crucial role in employment, income, food supply, and export earnings. Ethiopia is one of such countries where agriculture is the engine of economic growth, and the main source of livelihood, export earnings, and food supply. However, agriculture in Ethiopia is still traditional, virtually rain-fed, and dominated by smallholder peasants that use the majority of their output for own subsistence consumption. These stylized facts of the sector indicate that Ethiopian agriculture is highly sensitive to climate change. Climate change induced agricultural productivity and production shocks reduce rural income and employment opportunities, and hence trigger outmigration from agricultural occupation. This study begins with quantifying the economy-wide and regional effects of climate change induced productivity and labor supply shocks in Ethiopian agriculture. Climate change reduces agricultural output while it increases agricultural prices. Depending the impact scenario, its effects on GDP and households’ real consumption may reach to -8% and to -11%, respectively. Climate change also affects the international trade structure (i.e., the export and import mix) of the country. It decreases (increases) agricultural exports (imports). Economic effects of climate change are unevenly distributed among administrative regions of the country. Agrarian regions like Amhara, Oromia, and Southern NNP regions bear the bulk of the consequences. Climate change-induced effects on agriculture barely propel to affect GDP of urbanized regions like Addis Ababa, Dire Dawa, and Harari. Agricultural labor outmigration may widen the regional disparity as it further decrease GDP of the agrarian regions but increase that of the urbanized regions. To sum up, Chapter 5 of this study finds that climate change induced productivity and labor supply shocks in agriculture will have imminent risks to production, price, income, and consumption in the Ethiopian economy. Government of Ethiopia has already begun planning adaptation in different sectors of the economy. Adaptation in agriculture is its top priority. However, planned public adaptation in agriculture bears costs that may require to increase public spending to agricultural productivity improving measures by 25% to 100%. Chapter 6 of this study finds that the economy-wide and regional effects of this incremental public budget requirement cannot be undermined. Public adaptation costs distress government saving which shifts the saving-adjustment burden (for the sake of S-I balance) to households. This in turn impede on households’ welfare. In parallel, increasing public spending expands public services and relatively rewards nonagricultural factors of production. This has two further implications. First, it improves urban households’ welfare. Second, however, it increases the costs of production in manufacturing and private services and hence squeezes their output. Consequently, regional GDP of urban regions (e.g., Addis Ababa) decline. However, climate change and adaptation to climate change are not stand-alone environmental problems that shall be left to single economic agent, sector, or government agency. Therefore, Chapter 7 of this study, went further to assess the economy-wide and regional effects of alternative adaptation finance schemes. Even though the alternative finance schemes vary little in terms of their aggregate effects, their distributional effects may be important. In many of the cases, adaptation finance reduces the pressure of planned adaptation on government saving.

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 A. W. Yalew, Economic Development under Climate Change, https://doi.org/10.1007/978-3-658-29413-7_9

110

9. Conclusions and policy implications

Nevertheless, with the exception of diverting schemes, the factor-pull effects of public adaptation on manufacturing and services, and urban regions’ GDP persist. Therefore, the findings from Chapter 6 and 7 show that planned public adaptation in agriculture (its costs and finance) impede on the main actor (i.e., government), economic sectors (i.e., manufacturing and private services), and regions (i.e., urban regions) of structural change in Ethiopia. Comparing regional effects of adaptation costs and finance with the regional effects of climate change may have further implication for the allocation of federal block-grants among regional states of the country. The concerns over the tradeoffs due to public adaptation in agriculture mount as one considers adaptation in agriculture is only one among the sectors that need public support to adapt to climate change. Public support for adaptation in water, energy, transport, and health sectors is required yet (NMSA, 2001; NMA, 2007; Robinson et al., 2013; FDRE, 2016). Most importantly, as a low-income country, the forgone opportunities of adaptation finance are high. The government still strives to invest in human capital, transport and communication, and renewable energy as these are critical for driving the Ethiopian economy forward. Of course, the ongoing structural transformation in the country itself may contribute to climateresilient economic development. The discussion in Chapter 8 show that cost-free exogenous structural change scenarios pertaining to labor skills, and transaction costs would dampen indirect economy-wide and regional effects of climate change. The offsets would have been stronger if the economy could experience structural change in the production sectors parallel to changes in the labor market. Taken together, the study concludes that climate change is not a stand-alone environmental problem in developing countries like Ethiopia. It is rather an economic development problem that impairs the economic prospects (through its impacts), and brings new policy challenge (through its adaptation). The policy implications of this study are also relevant to other countries, especially, in the sub-Saharan Africa where smallholder agriculture is important economic sector; where rural livelihood is inextricably linked to agriculture; where regional states vary in their socio-economic development; where public resources are scarce; and where the economic role of government is significant. All said, however, this study is not without limitations. There are ample sources of uncertainties of climate change and its impacts. Nevertheless, this study focuses only on dry climate change scenarios which can be regarded as high-end impacts, and attempts to gauge uncertainty implied only by two biophysical models. It also focuses only on productivity and labor outmigration effects while there are many other climate change related impacts (e.g. frequency of droughts, animal and plant diseases outbreaks, diminishing crop area suitability, and indirect effects through international food prices). In addition, due to lack of applicable biophysical models, it considered only the case of grain crops and livestock. And, yet, the effects on the latter shall still be taken as partial. Migration scenarios are entirely extrapolated due to lack of empirical models and evidence explicitly linking climate change and labor migration. On the other hand, for modeling reason, the adaptation catalog consists of only productivity enhancing measures. The adaptation finance schemes are generic and self-designed. The regional module is synthetically created though it convincingly depicts the regional variations. The discussion with regard to structural change primarily intends to draw qualitative policy implications.

9. Conclusions and policy implications

111

Despite these limitations, however, the findings of this study corroborate the findings of the previous studies in Ethiopia (see, for example, World Bank, 2010a; Arndt et al., 2011; Robinson et al., 2012; FDRE, 2015), and of the studies that contend structural change as a generic and non-regrettable form of adaptation (see, for example, Fankhauser and Burton, 2011; Henderson et al., 2017). Therefore, a set of lessons for policy-and-decision makers can be placed forward. Climate change may impede the economic prospects of Ethiopia. Proactive adaptation in its dominant, but vulnerable, agriculture sector is imperative. Given the inadequacy and unpredictability of the international adaptation finance, policy-and-decision makers shall prepare to devise policy instruments and incentives to generate adequate adaptation finance from domestic sources. In doing so, it is important to keep in mind that alternative finance schemes would mean different for different elements (agents, sectors, and locations) of the economy. Most importantly, policy decisions shall factor in the forgone opportunities of public resources used for adaptation in agriculture. As such, it may be better to devise incentives such that the agriculture sector can be self-financing for its adaptation. For instance, promoting conservation agriculture, reducing emission from land use, land use change, and livestock production may generate considerable amount of international mitigation finance that can be reused as adaptation finance. Policy makers shall also devise incentives to attract more private sector investment in agriculture. This will reduce the share of agricultural output produced by smallholder farmers which is the principal reason for planned public adaptation in the sector. The contribution of international support in terms of agricultural R&D and technology transfer shall be explored further. Future research on the subject shall further explore the climate change-agriculture-migration nexus, and the regional effects of centrally devised policy responses to climate change. Research that fine-tune the sources of adaptation finance, and that explicitly identify and estimate the costs of specific measures that are compatible with structural transformation, yet, contribute to dampen (strengthen) the potential negative (positive) effects of climate change in one or more sectors are highly needed.

Bibliography Abreu, A. (2010). The New Economics of Labor Migration: Beware of Neoclassicals Bearing Gifts. Forum for Social Economics, DOI: https://dx.doi:10.1007/s12143-010-9077-2 Adams, R. M., Hurd, B. H., Lenhart, S., and Leary, N. (1998). Effects of global climate change on agriculture: An interpretative review. Climate Research, 11, pp.19-30. Adaptation Watch. (2015). Toward Mutual Accountability: The 2015 Adaptation Finance Transparency Gap Report. Retrieved from www.AdaptationWatch.org on December 25, 2015. Admassu, G., Getinet, M., Thomas, T. S., Waithaka, M., and Kyotalimye, M. (2013). Ethiopia. Chapter 6 in M. Waithaka, Michael; Nelson, G.C. Nelson; Thomas, T.S. Thomas, and M Kyotalimye (Eds.), East African Agriculture and Climate Change: A Comprehensive Analysis (pp.149-182). Washington, DC: International Food Policy Research Institute. Afful-Koomson, T. (2015). The Green Climate Fund in Africa: what should be different? Climate and Development, 4, pp.367–379 Ahmed, V., and O’Donoghue, C. (2010). External Shocks in a Small Open Economy: A CGE -Microsimulation Analysis. The Lahore Journal of Economics, 15(1), pp.45-90. Alagidede, P., Adu, G., and Frimpong, P. B. (2016). The Effect of Climate Change on Economic Growth: Evidence from Sub-Saharan Africa. Environmental Economics and Policy Studies, 18, pp.417-436. Alene, A. D., and Coulibaly, O. (2009). The impact of agricultural research on productivity and poverty in sub-Saharan Africa. Food Policy, 34, pp.198–209. Ali, N. S. (2012). Climate Change and Economic Growth in a Rain-fed Economy: How Much Does Rainfall Variability Cost Ethiopia? Retrieved from http://dx.doi.org/10.2139/ssrn.2018233 Alston, J. M., and Pardey, P. G. (2013). Agricultural R & D and Food Security of the Poor. Economic Papers, 32(3), pp.289–297. Andre´, F. J., Cardente, M. A., and Romero, C. (2010). Economic Policy Using Applied General Equilibrium Models: An Overview. Chapter 2 in F. J. Andre´, M. A. Cardente, and C. Romero, Designing Public Policies: An Approach Based on a Multi-Criteria Analysis and Computable General Equilibrium Modeling (pp. 9-31). Heidelberg: Springer. Annual Agricultural Sample Survey (AgSS). (2006). Volumes I-V. Central Statistics Agency. Retrieved from www.csa.gov.et. Accessed on 15 November, 2015. Annual Agricultural Sample Survey (AgSS). (2014). Volumes I-V. Central Statistics Agency. Retrieved from www.csa.gov.et. Accessed on 15 November, 2015. Annual Agricultural Sample Survey (AgSS). (2015). Volumes I-V. Central Statistics Agency. Retrieved from www.csa.gov.et. Accessed on 15 November, 2015. Antle, J. M., and Capalbo, S. M. (2010). Adaptation of Agricultural and Food Systems to Climate Change: An Economic and Policy. Applied Economic Perspectives and Policy, 32(3), pp.386–416. Arndt, C., Robinson, S., and Willenbockel, D. (2011). Ethiopia’s growth prospects in a changing climate: A stochastic general equilibrium approach. Global Environmental Change, 21, pp.701-710. Auriol, E., and Warlters, M. (2012). The marginal cost of public funds and tax reform in Africa. Journal of Development Economics, 97, pp.58-72.

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 A. W. Yalew, Economic Development under Climate Change, https://doi.org/10.1007/978-3-658-29413-7

114

Bibliography

Awulachew, S. B. (2010). Irrigation potential in Ethiopia: Constraints and opportunities for enhancing the system. Retrieved from https://agriknowledge.org/downloads/j38606956 Accessed on 20 January, 2016. Bailey, S. J. (2002). Public Sector Economics: Theory, Policy and Practice (2nd Ed.). New York: Palgrave. Balistreri, E. J., McDaniel, C. A., and Wong, E. V. (2003). An estimation of US industry-level capital–labor substitution elasticities: support for Cobb–Douglas. North American Journal of Economics and Finance, 14, pp.343–356. Bandara, J. S. (1991). Computable General Equilibrium Models for Development Policy Analysis in LDCs. Journal of Economic Surveys, 5(1), pp.3-69. Benassy-Quere, A., Coeure, B., Jacquet, P., and Pisani-Ferry, J. (2010). Economic Policy: Theory and Practice. Oxford University Press. Benin, S., Mogues, T., Cudjoe, G., and Randriamamonjy, J. (2009). Public Expenditures and Agricultural Productivity Growth in Ghana. Retrieved from http://ageconsearch.umn.edu/handle/51634. Accessed on 20 January, 2016. Benin, S., Pratt, A. N., Wood, S., and Guo, Z. (2011). Trends and Spatial Patterns in Agricultural Productivity in Africa, 1961–2010. International Food Policy Research Institute. Benson, T., Engida, E., and Thurlow, J. (2014). The Economy-wide Effects of Teff, Wheat, and Maize Production Increases in Ethiopia: Results of Economy-wide Modeling. IFPRI Discussion Paper No. 01366. International Food Policy Research Institute. Berhanu, W., and Beyene, F. (2015). Climate Variability and Household Adaptation Strategies in Southern Ethiopia. Sustainability, 7, pp.6353-6375. Bezabih, M., Chambwera, M., and Stage, J. (2011). Climate change and total factor productivity in the Tanzanian economy. Climate Policy, 11 (6), pp.1289-1302. Bill and Melinda Gates Foundation (BMGF). (2010). Accelerating Ethiopian Agriculture Development for Growth, Food Security, and Equity. Retrieved from https://agriknowledge.org/downloads/h989r326r. Accessed on 20 January, 2016. Bird, N., Tilley, H., Trujillo, N. C., Tumushabe, G., Welham, B., and Yanda, P. (2013). Measuring the effectiveness of public climate finance delivery at the national level. Overseas Development Institute. Block, P. J., Strzepek, K., Rosegrant, M., and Diao, X. (2008). Impacts of Considering Climate Variability on Investment Decisions in Ethiopia. IFPRI Research Brief 15-12. International Food Policy Research Institute. Blonigen, B.A., Joseph E. F., and Kenneth, A. R. (1997). Sector-Focused General Equilibrium Modeling. Chapter 7 in J.F. Francois and K.A. Reinert (Eds.), Applied Methods for Trade Policy Analysis (pp. 189-230). Cambridge, MA: Cambridge University Press. Bondeau, A., Smith, P. C., Zaehle, S., Schaphoff, S., Lucht, W., Cramer, W., . . . Benjamin, S. (2007). Modelling the role of agriculture for the 20th century global terrestrial carbon balance. Global Change Biology, 13, pp.679–706. Bosello, F., Campagnolo, L., and Eboli, F. (2013). Climate Change and Adaptation: The Case of Nigerian Agriculture. Nota Di Lavoro (35). Milan: Fondazione Eni Enrico Mattei. Brookshire, E., and Weaver, T. (2015). Long-term decline in grassland productivity driven by increasing dryness. Nature Communications, 6 (7148). Brown, O. (2008). Migration and Climate Change. IOM Migration Research Series, No.31. International Organization for Migration. Bryan, E., Deressa, T. T., Gbetibouo, G. A., and Ringler, C. (2009). Adaptation to climate change in Ethiopia and South Africa: options and constraints. Environmental Science and Policy, 12, pp.413-426.

Bibliography

115

Buchner, B. K., Trabacchi, C., Mazza, F., Abramskiehn, D., and Wang, D. (2015). Global Landscape of Climate Finance 2015. Retrieved from www.climatepolicyinitiative.org. Accessed on June 2016. Burfisher, M. F. (2011). Introduction to Computable General Equilibrium Models. New York: Cambridge University Press. Burnham, M., and Ma, Z. (2015). Linking smallholder farmer climate change adaptation decisions to development. Climate and Development, 8(4): pp.289-311. Cardenete, M. A., Guerra, A.-I., and Sancho, F. (2012). Applied General Equilibrium: An Introduction. Berlin: Springer-Verlag. Central Statistics Agency (CSA). (2011). Report on Annual Average Retail Prices of Goods and Services. Retrieved from www.csa.gov.et. Accessed on 14 December, 2015. Central Statistics Agency (CSA). (2013). Population Projections of Ethiopia for all Regions at Wereda Level from 2014–2017. Retrieved from www.csa.gov.et. Accessed on 14 December, 2015. Central Statistics Agency (CSA). (2015). Agricultural Sample Survey Time Series Data for National and Regional Level: From 1995/96–2014/15: Report on Crop Area and Production. Retrieved from www.csa.gov.et. Accessed on 20 April, 2016. Chand, R., Kumar, P., and Kumar, S. (2012). Total Factor Productivity and Returns to Public Investment on Agricultural Research in India. Agricultural Economics Research Review, 25(2), pp.181-194. Chen, W., and Ding, Y. (2007). Total Factor Productivity in Chinese agriculture: The role of infrastructure. Frontiers of Economics in China, 2 (2), pp. 212–223. Cheung, W. H., Senay, G. B., and Singh, A. (2008). Trends and spatial distribution of annual and seasonal rainfall in Ethiopia. International Journal of Climatology, DOI: http://dx.doi:10.1002/joc.1623 Climate Funds Update (CFU). (2016). The Latest Information on Climate Funds. Retrieved on from http://www.climatefundsupdate.org/. Accessed on January 18, 2016. Cline, W. R. (2007). Global Warming and Agriculture: Impact Estimates by Country. Washington, D.C: Center for Global Development. Conway, D., and Schipper, E. F. (2011). Adaptation to climate change in Africa: Challenges and opportunities identified from Ethiopia. Global Environmental Change, 21, pp.227237. Craig, B. J., Pardey, P. G., and Roseboom, J. (1997). International Productivity Patterns: Accounting for Input Quality, Infrastructure, and Research. American Journal of Agricultural Economics, 79(4), pp.1064-1076. CSA-EDRI-IFPRI. (2006). Atlas of the Ethiopian Rural Economy. Addis Ababa: Central Statistics Agency. CSA-FAO. (2014). Food and Agriculture Data Network: Ethiopia. Retrieved from http://www.countrystat.org/home.aspx?c=ETH. Accessed on March 11, 2016. De Boer, P., and Missaglia, M. (2006). Estimation of income elasticities and their use in a CGE model for Palestine. Econometric Institute Report, No. 12. Retrieved from http://repub.eur.nl/pub/7753. Accessed on 20 July, 2015. De Brauw, A. (2014). Migration, Youth, and Agricultural Productivity in Ethiopia. Retrieved from http://ageconsearch.umn.edu//handle/189684. Accessed on 24 September 2016 De Haas, H. (2010). Migration and Development: A Theoretical Perspective. International Migration Review, 44(1), pp.227–264.

116

Bibliography

Dell, M., Jones, B. F., and Olken, B. A. (2012). Temperature Shocks and Economic Growth: Evidence from the Last Half Century. American Economic Journal: Macroeconomics, 4(3), pp.66-95. Dercon, S. (2004). Growth and shocks: evidence from rural Ethiopia. Journal of Development Economics, 74, pp.309–329. Dercon, S., and Christiaensen, L. (2011). Consumption risk, technology adoption and poverty traps: Evidence from Ethiopia. Journal of Development Economics, 96, pp.159-173. Dercon, S., Gilligan, D. O., Hoddinott, J., and Woldehanna, T. (2007). The impact of roads and agricultural extension on consumption growth and poverty in fifteen Ethiopian villages. CSAE Working Paper No. 01. University of Oxford. Deressa, T. T., and Hassan, R. M. (2009). Economic Impact of Climate Change on Crop Production in Ethiopia: Evidence from Cross-section Measures. Journal of African Economies, 18(4), pp. 529–554. Dervis, K., de Melo, J., and Robinson, S. (1982). General Equilibrium Models for Development Policy. Washington, D.C.: World Bank. Dewatripont, M., and Michel, G. (1987). On Closure Rules, Homogeneity and Dynamics in Applied General Equilibrium Models. Journal of Development Economics, 26, pp. 6576. Diao, X., Nwafor, M., Alpuerto, V., Akramov, K., and Salau, S. (2010). Agricultural Growth and Investment Options for Poverty Reduction in Nigeria. IFPRI Discussion Paper No. 00954. International Food Policy Research Institute. Diao, X., Taffesse, A. S., Dorosh, P., Thurlow, J., Pratt, A. N., and Yu, B. (2012). Ethiopia. Chapter 5 in X. Diao, J. Thurlow, S. Benin, and S. Fan (Eds.), Strategies and Priorities for African Agriculture: Economy-wide Perspectives from Country Studies (pp.107139). Washington, DC: International Food Policy Research Institute. Dixon, P.B., Parmenter, B.R., Sutton, J., and Vincent, D.P. (1982). ORANI: A Multisectoral Model of the Australian Economy. Amsterdam: North-Holland Publishing Co. Dorosh, P., and Schmidt, E. (2010). The Rural-Urban Transformation in Ethiopia. ESSPII Working Paper, No. 13. Ethiopia Strategy Support Program II. Dorosh, P., and Thurlow, J. (2011). Urbanization and Economic Transformation: A CGE Analysis for Ethiopia. ESSP II Working Paper, No. 14. Ethiopia Strategy Support Program II. EDHS. (2014). Ethiopian Demographic and Health Survey. Central Statistical Agency. Retrieved from www.csa.gov.et. Accessed on 20 April, 2016. Eshetu, Z., Simane, B., Tebeje, G., Negatu, W., Amsalu, A., Berhanu, A., . . . Trujillo, N. C. (2014). Climate Finance in Ethiopia. Overseas Development Institute. Ethiopian Development Research Institute (EDRI). (2009). Ethiopia: Input-output Table and Social Accounting Matrix. Addis Ababa. Ethiopian Roads Authority (ERA). (2010). Universal Rural Road Access Program. Addis Ababa. Evangelista, P., Young, N., and Burnett, J. (2013). How will climate change spatially affect agriculture production in Ethiopia? Case studies of important cereal crops. Climatic Change, 119, pp.855–873. Evenson, R. E., Pray, C. E., and Rosegrant, M. W. (1999). Agricultural Research and Productivity. Research Report, No.109. International Food Policy Research Institute. Ezra, M. (2001). Demographic Response to Environmental Stress in Drought- and Famine Prone Areas of Northern Ethiopia. International Journal of Population Geography, 7, pp.259-279.

Bibliography

117

Ezra, M., and Kiros, G.-E. (2001). Rural Out-migration in the Drought Prone Areas of Ethiopia: A Multilevel Analysis. International Migration Review, 35(3), pp.749-771. Fan, S., Hazell, P., and Thorat, S. (1999). Linkages between Government Spending, Growth, and Poverty in Rural India. Research Report No. 110. International Food Policy Research Institute. Fan, S., Hazell, P., and Thorat, S. (2000). Government Spending, Growth and Poverty in Rural India. American Journal of Agricultural Economics, 82(4), pp.1038-1051. Fan, S., Zhang, L., and Zhang, X. (2002). Growth, Inequality, and Poverty in Rural China: The Role of Public Investments. Research Report No.125. International Food Policy Research Institute. Fankhauser, S. (2010). The costs of adaptation. WIREs Climate Change, 1, pp.23–30. Fankhauser, S., and Burton, I. (2011). Spending adaptation money wisely. Climate Policy, 11, 3, 1037-1049. Fankhauser, S., and Schmidt-Traub, G. (2011). From adaptation to climate-resilient development: The costs of climate-proofing the Millennium Development Goals in Africa. Climate and Development, 3(2), pp.94-113. Federal Democratic Republic of Ethiopia (FDRE). (1995). Proclamation No. 1/1995: Proclamation of the Constitution of the Federal Democratic Republic of Ethiopia. Federal Negarit Gazeta, 1(1). Addis Ababa. Federal Democratic Republic of Ethiopia (FDRE). (2011). Ethiopia's Climate Resilient Green Economy Strategy. Addis Ababa. Federal Democratic Republic of Ethiopia (FDRE). (2012). Ethiopia’s Climate-Resilient Green Economy: Green economy strategy. Addis Ababa. Federal Democratic Republic of Ethiopia (FDRE). (2015). Ethiopia’s Climate Resilient Green Economy: Climate Resilience Strategy - Agriculture and Forestry Sector. Addis Ababa. Federal Democratic Republic of Ethiopia (FDRE). (2016). Ethiopia’s Climate-Resilient Green Economy: Climate Resilience Strategy- Water and Energy. Addis Ababa. Felloni, F., Wahl, T., Wandschneider, P., and Gilbert, J. (2001). Infrastructure and Agricultural Production: Cross-country Evidence and Implications for China. TWP-2001-03. Washington State University. Femenia, F. (2012). Functional Forms Commonly Used in CGE Models. AGRODEP Technical Note TN-02. International Food Policy Research Institute. Fenton, A., Wright, H., Afionis, S., Paavola, J., and Huq, S. (2014). Debt relief and financing climate change action. Nature Climate Change, 4, pp.650-653. Ferede, T., Ayenew, A. B., and Hanjra, M. A. (2013). Agro-ecology Matters: Impacts of Climate Change on Agriculture and its Implications for Food Security in Ethiopia. Nova Science Publishers. Fields, G. S. (2011). Labor market analysis for developing countries. Labour Economics, 18, pp.S16–S22. Food and Agriculture Organization of the United Nations (FAO). (2015). FAOSTAT: Ethiopia. Reterieved from http://faostat.fao.org/. Accessed on April 05, 2015. Gabre-Madhin, E. Z. (2001). Market Institutions, Transaction Costs, and Social Capital in the Ethiopian Grain Market. Research Report No.124. International Food Policy Research Institute.

118

Bibliography

Gebeyehu, Z.H. (2015). Land Policy Implications in Rural-Urban Migration: the Dynamics and Determinant Factors of Rural-Urban Migration in Ethiopia. PhD Dissertation. Technische Universität München. Gebreegziabher, Z., Stage, J., Mekonnen, A., and Alemu, A. (2016). Climate Change and the Climate Change and the Ethiopian Economy. Environment and Development Economics, 21 (2), pp. 205–225. Gebremariam, S., Amare, S., Baker, D., and Solomon, A. (2010). Diagnostic Study of Live Cattle and Beef Production and Marketing: Constraints and Opportunities for Enhancing the System. International Livestock Research Institute. Ginsburgh, V., and Keyzer, M. (1997). The Structure of Applied General Equilibrium Models. Cambridge: MIT Press. Gray, C., and Mueller, V. (2012). Drought and Population Mobility in Rural Ethiopia. World Development, 40(1), pp.134–145. Hadgu, G., Tefaye, K., Mamo, G., and Kassa, B. (2014). Analysis of Farmers’ Perception and Adaptation Methods to Climate Variability/Change in Tigray Region, Northern Ethiopia. Research Journal of Agricultural and Environmental Sciences, 1(1), pp.15-25. Hansen, L. P., and Heckman, J. J. (1996). The Empirical Foundations of Calibration. Journal of Economic Perspectives, 10(1), pp.87-104. Harris, J.R., and Todaro, M. (1970). Migration, Unemployment and Economic Development: A Two-Sector Analysis. American Economic Review, 60 (1), pp.126-42. Harrison, G. W., Jones, R., Kimbell, L. J., and Wigle, R. (1993). How Robust is Applied General Equilibrium Analysis. Journal of Policy Modeling, 15(1), pp.99-115. Heal, G., and Millner, A. (2014). Uncertainty and Decision Making in Climate Change Economics. Review of Environmental Economics and Policy, 8 (1), pp. 120-137. Henderson, J.V, Storeygard, A. and Deichmann, W. (2017). Has climate change driven urbanization in Africa? Journal of Development Economics, 124, pp.60-82. Hermeling, C., and Mennel, T. (2008). Sensitivity analysis in economic simulations-a systematic approach. Retrieved from https://ub-madoc.bib.uni-mannheim.de/2100/1/dp08068.pdf. Accessed on 20 June, 2016. Hertel, T. W., and Lobell, D. B. (2014). Agricultural adaptation to climate change in rich and poor countries: Current modeling practice and potential for empirical contributions. Energy Economics, 46, pp.562-575. Hertel, T. W., and van der Mensbrugghe, D. (2016). Behavioral Parameters. Retrieved from https://www.gtap.agecon.purdue.edu/resources/download/8247.pdf. Accessed on 20 June, 2016. Hertel, T. W., McDougall, R. A., Narayanan, B. G., and Aguiar, A. H. (2014). Behavioral Parameters. Retrieved from https://www.gtap.agecon.purdue.edu/resources/download/7047.pdf. Accessed on 20 June, 2016. HICES. (2000). Households Income, Consumption, and Expenditure Survey. Central Statistics Agency. Retrieved from www.csa.gov.et. Accessed on 14, December 2015. HICES. (2005). Household Income, Consumption, and Expenditure Survey. Central Statistics Agency. Retrieved from www.csa.gov.et. Accessed on 14, December 2015. HICES. (2011). Households Income, Consumption, and Expenditure Survey. Central Statistics Agency. Retrieved from www.csa.gov.et. Accessed on 14, December 2015. Higgs, P. J., Parmenter, B. R., and Rimmer, R. J. (1988). A Hybrid Top-Down, Bottom-Up Regional Computable General Equilibrium Model. International Regional Science Review, 11(3), pp.317-328.

Bibliography

119

Hosoe, N., Gasawa, K., and Hashimoto, H. (2010). Textbook of Computable General Equilibrium Modelling. London: Palgrave Macmillan. Hurni, H. (1998). Agro-ecological Belts of Ethiopia: Explanatory notes on three maps at a scale of 1:1,000,000. University of Bern: Centre for Development and Environment, Soil Conservation Research Programme. ICPS. (2012). Inter-Censal Population Survey. Central Statistics Agency. Retrieved from www.csa.gov.et. Accessed on 14 December, 2015. Intergovernmental Panel on Climate Change (IPCC). (2007). Climate Change 2007: Impacts, Adaptation and Vulnerability. Cambridge: Cambridge University Press. Intergovernmental Panel on Climate Change (IPCC). (2014). Climate Change 2014: Synthesis Report. Geneva. International Livestock Research Institute (ILRI). (2015). Animal breeding and genetics in the Ethiopia livestock master plan. Ethiopian Livestock Masterplan Brief, No.4. Addis Ababa. International Organization for Migration (IOM). (2012). Climate Change, Environmental Degradation and Migration. International Dialogue on Migration, No.18. Geneva. Jomini, P., Zeitsch, J. F., McDougall, R., Welsh, A., Brown, S., Hambley, J., and Kelly, J. (1991). SALTER: A General Equilibrium Model of the World Economy: Volume 1, Model Structure, Database, and Parameters. Canberra: Industry Commission. Jones, B. F., and Olken, B. A. (2010). Climate Shocks and Exports. American Economic Review: Papers and Proceedings, 100, pp.454-459. Kassie, B. T. (2014). Climate variability and change in Ethiopia: Exploring impacts and adaptation options for cereal production. PhD Dissertation. Wageningen University. Kehoe, P. J., and Kehoe, T. J. (1994). A Primer on Static Applied General Equilibrium Models. Quarterly Review, 18(2), pp.2-16. Kissinger, G., Lee, D., Orindi, V. A., Narasimhan, P., King’uyu, S. M., and Sova, C. (2013). Planning climate adaptation in agriculture: Meta-synthesis of national adaptation plans in West and East Africa and South Asia. CCAFS Report No. 10 CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS). Knox, J., Daccache, A., and Hess, T. (2013). What is the impact of infrastructural investments in roads, electricity and irrigation on agricultural productivity? CEE Review, 11-007. Retrieved from www.environmentalevidence.org/SR11007.html. Accessed on 07 August, 2015. Knox, J., Hess, T., Daccache, A., and Wheeler, T. (2012). Climate change impacts on crop productivity in Africa and South Asia. Environmental Research Letters, 7(034032). Kubik, Z., and Mathilde, M. (2016). Weather Shocks, Agricultural Production and Migration: Evidence from Tanzania. The Journal of Development Studies, 52(5). Kurukulasuriya, P., and Mendelsohn, R. (2008). A Ricardian analysis of the impact of climate change on African cropland. African Journal of Agricultural and Resources Economics, 2(1). Kurukulasuriya, P., and Rosenthal, S. (2003). Climate Change and Agriculture A Review of Impacts and Adaptations. Climate Change Series No.91. World Bank. Kurukulasuriya, P., Mendelsohn, R., Hassan, R., Benhin, J., Deressa, T., Diop, M., . . . Dinar, A. (2006). Will African Agriculture Survive Climate Change? The World Bank Economic Review, 20(3), pp.367–388.

120

Bibliography

Lanos, B., Mas, A.A., and Woldeyes, F. (2014). Public expenditures in support of food and agriculture development in the Federal Republic of Ethiopia, 2006/07-2012/13: A preliminary assessment. FAO Technical Notes Series. Retrieved from https://agriknowledge.org/downloads/wh246s16c. Accessed on 10 January, 2016. Lee, E. S. (1966). A Theory of Migration. Demography, 3(1), pp.47-57. Leontief, W., Morgan, A., Polenske, K., Simpson, D., and Tower, E. (1965). The Economic Impact-Industrial and Regional - of an Arms Cut. The Review of Economics and Statistics, XLVII (3), pp.217-241. Lewis, A.W. (1954). Economic Development with Unlimited Supplies of Labour. The Manchester School, 22 (2), pp.131-191. Lipton, M. (1980). Migration from Rural Areas of Poor Countries: The Impact on Rural Productivity and Income Distribution. World Development, 8(1), pp.1-24. Lluch, C., Powel, A. A., and Williams, R. A. (1977). Patterns in Household Demand and Saving. New York: Oxford University Press. LMIS. (2011) Report on Large and Medium Scale Manufacturing and Electricity Industries Survey. Central Statistics Agency. Retrieved from www.csa.gov.et. Accessed 24 November 2015. Lofgren, H. (2001). A CGE Model for Malawi: Technical Documentation. TMD Discussion Paper, No. 70. International Food Policy Research Institute. Lofgren, H., Harris, L. R., and Robinson, S. (2002). A standard computable general equilibrium (CGE) model in GAMS. Microcomputers in Policy Research No. 5. International Food Policy Research Institute. Lucas, R. E. (1997). Internal Migration in Developing Countries. Chapter 13 in M. R. Rosenzweig, and O. Stark (Eds.), Handbook of Population and Family Economics (pp. 721-798). Elsevier Science B.V. Lybbert, T. J., and Sumner, D. A. (2012). Agricultural technologies for climate change in developing countries: Policy options for innovation and technology diffusion. Food Policy, 37, pp.114-123. Maddison, D., Manley, M., and Kurukulasuriya, P. (2007). The Impact of Climate Change on African Agriculture. Policy Research Working Paper No. 4306. World Bank. Makombe, G., Kelemework, D., and Aredo, D. (2007). A comparative analysis of rain fed and irrigated agricultural production in Ethiopia. Irrigation Drainage System, 21, pp.35-44. Mansur, A., and Whalley, J. (1981). Numerical Specification of Applied General Equilibrium Models: Estimation, Calibration, and Data. Department of Economics Working Paper No. 8106. University of Western Ontario. Retrieved from http://ir.lib.uwo.ca/economicscsier_wp/28. Accessed on 20 July, 2015. Massey, D. S., Axinn, W. G., and Ghimire, D. J. (2010). Environmental change and out-migration: evidence from Nepal. Population Environment, 32, pp.109–136. Mbow, C., Mertz, O., Diouf, A., Rasmussen, K., and Reenberg, A. (2008). The history of environmental change and adaptation in eastern Saloum-Senegal: Driving forces and perceptions. Global and Planetary Change, 64, pp.210–221. McGray, H., Hammill, A., and Bradley, R. (2007). Weathering the Storm Options for Framing Adaptation and Development. World Resources Institute. McLeman, R. A. (2014). Climate Change and Human Migration: Past Experiences, Future Challenges. New York: Cambridge University Press. McLeman, R., and Smit, B. (2006). Migration as an Adaptation to Climate Change. Climatic Change, 76, pp.31-53.

Bibliography

121

Mekonnen, A. (2014). Economic Costs of Climate Change and Climate Finance with a Focus on Africa. Journal of African Economies, 23 (2), ii50–ii82. Mendelsohn, R. (2000). Efficient Adaptation to Climate Change. Climatic Change, 45, pp.583– 600. Mendelsohn, R. (2012). The Economics of Adaptation to Climate Change in Developing Countries. Climate Change Economics, 3, 2 (1250006). Mendelsohn, R., Dinar, A., and Williams, L. (2006). The distributional impact of climate change on Rich and Poor Countries. Environment and Development Economics, 11, pp.159–178. Mertz, O., Mbow, C., Reenberg, A., Genesio, L., Lambin, E. F., D'haen, S., . . . Sandholt, I. (2011). Adaptation strategies and climate vulnerability in the Sudano-Sahelian region of West Africa. Atmospheric Science Letters, 12, pp.104-108. Mideksa, T. K. (2010). Economic and distributional impacts of climate change: The case of Ethiopia. Global Environmental Change, 20, pp.278–286. Miheretu, B.A. (2011). Causes and Consequences of Rural-Urban Migration: The Case of Woldiya Town, North Ethiopia. MA Thesis. University of South Africa. Millner, A., and Dietz, S. (2015). Adaptation to climate change and economic growth in Developing Countries. Environment and Development Economics, 20(3), 380-406. Ministry of Agriculture and Rural Development (MoARD). (2010). Ethiopia's Agricultural Sector Policy and Investment Framework: 2010-2020. Addis Ababa. Ministry of Education (MoE). (2016). Educational Statistics 2014/15. Addis Ababa. Ministry of Finance and Economic Development (MoFED). (2005). National Accounts Statistics of Ethiopia: Source and Methods: 1999/2000 Base Year Series. Addis Ababa. Ministry of Finance and Economic Development (MoFED). (2006). Ethiopia: Building on Progress a Plan for Accelerated and Sustained Development to End Poverty: 2005/062009/10. Addis Ababa. Ministry of Finance and Economic Development (MoFED). (2009). Layperson's Guide to the Public Budget Progress at Regional Level. Addis Ababa. Ministry of Finance and Economic Development (MoFED). (2010). Growth and Transformation Plan: 2010/11-2014/15. Addis Ababa. Ministry of Finance and Economic Development (MoFED). (2012). Ethiopia’s Progress towards Eradicating Poverty: An Interim Report on Poverty Analysis Study - 2010/11. Addis Ababa. Ministry of Finance and Economic Development (MoFED). (2014). Government Revenues and Expenditures. Addis Ababa. Ministry of Finance and Economic Development (MoFED). (2015). Public Sector Debt: 2010/11 - 2014/15. Statistical Bulletin, No.15. Addis Ababa. Ministry of Finance and Economic Development (MoFED). (2015). Quarterly Government Finance: National and Regional Budgets. Addis Ababa. Ministry of Water Resources (MoWR). (2001). Ethiopian Water Sector Policy. Addis Ababa. Minten, B., and Barrett, C. B. (2008). Agricultural Technology, Productivity, and Poverty in Madagascar. World Development, 36(5), pp.797–822. Minten, B., Koru, B., and Stifel, D. (2013). The last mile(s) in modern input distribution: Pricing, profitability, and adoption. Agricultural Economics, 44, pp.629–646. Missaglia, M., and de Boer, P. (2004). Food-For-Work versus Cash-For-Work: Emergency Assistance in Palestine. Economic Systems Research, 16(4), pp.367-390.

122

Bibliography

Mitik, L., and Engida, E. (2013). Public investment in irrigation and training for an agriculture-led development: a CGE approach for Ethiopia. Working Paper No. 2013-02. Partnership for Economic Policy. Mitra-Kahn, B. H. (2008). Debunking the Myths of Computable General Equilibrium Models. SCEPA Working Paper 2008-1. The New School for Social Research. Moss, R. H., Edmonds, J. A., Hibbard, K. A., Manning, M. R., Rose, S., van Vuuren, D. P., . . . Wilbanks, T. J. (2010). The next generation of scenarios for climate change research and assessment. Nature, 463(7282), pp.747-756. Müller, C., and Robertson, R. D. (2014). Projecting future crop productivity for global economic modeling. Agricultural Economics, 45, pp.37-50. Müller, C., Cramer, W., Hare, W. L., and Lotze-Campen, H. (2011). Climate change risks for African agriculture. Proceedings of the National Academy of Sciences, 108, (11), pp.4313–4315. Müller, C., Waha, K., Bondeau, A., and Heinke, J. (2014). Hotspots of climate change impacts in sub-Saharan Africa and implications for adaptation and development. Global Change Biology, 20(8), pp.2505–2517. Nakhooda, S., Fransen, T., Kuramochi, T., Caravani, A., Prizzon, A., Shimizu, N., . . . Welham, B. (2013). Mobilizing International Climate Finance: Lessons from the Fast-Start Finance Period. Overseas Development Institute. Naqvi, F., and Peter, M. W. (1996). A Multiregional, Multisectoral Model of the Australian Economy with an Illustrative Application. Australian Economic Papers, 94-113. Nardone, A., Ronchi, B., Lacetera, N., Ranieri, M. S., and Bernabucci, U. (2010). Effects of climate changes on animal production and sustainability of livestock systems. Livestock Science, 130, pp.57-69. National Bank of Ethiopia (NBE). (2015). National Bank Annual Report 2013-2014. Retrieved from http://www.nbe.gov.et/publications/annualreport.html. Accessed on 05 January, 2016. National Bank of Ethiopia (NBE). (2016). National Bank Annual Report 2014-2015. Retrieved from http://www.nbe.gov.et/publications/annualreport.html. Accessed on 12 January, 2017. National Meteorological Agency (NMA). (2007). Climate Change National Adaptation Programme of Action (NAPA) of Ethiopia. Addis Ababa. National Meteorological Services Agency of Ethiopia (NMSA). (2001). Initial National Communication of Ethiopia to the United Nations Framework Convention on Climate Change (UNFCCC). Addis Ababa. National Planning Commission of Ethiopia (NPC). (2016). Growth and Transformation Plan II: 2015/16-2019/20: Volume I. Addis Ababa. Naude, W. (2010). The Determinants of Migration from Sub-Saharan African Countries. Journal of African Economies, 19(3), 330-356. Nawrotzki, R. J., Hunter, L. M., Runfola, D. M., and Riosmena, F. (2015). Climate change as a migration driver from rural and urban Mexico. Environmental Research Letters, 10 (114023). Nelson, G. C., Rosegrant, M. W., Koo, J., Robertson, R., Sulser, T., Zhu, T., . . . Lee, D. (2010). The Costs of Agricultural Adaptation to Climate Change. Development and Climate Change Discussion Paper No. 4. World Bank. Nganou, J.-P. (2011). Estimation of the parameters of a linear expenditure system (LES) demand function for a small African economy. MPRA Paper No.31450. Retrieved from http://mpra.ub.uni-muenchen.de/31450/. Accessed on 07 August, 2015.

Bibliography

123

NLFS. (1999). National Labor Force Survey 1999. Central Statistics Agency. Retrieved from www.csa.gov.et. Accessed on 14 December, 2015. NLFS. (2005). National Labor Force Survey 2005. Central Statistics Agency. Retrieved from www.csa.gov.et. Accessed on 14 December, 2015. NLFS. (2013). National Labor Force Survey 2013. Central Statistics Agency. Retrieved from www.csa.gov.et. Accessed on 14 December, 2015. O’Neill, B. C., Kriegler, E., Ebi, K. L., Kemp-Benedict, E., Riahi, K., Rothman, D. S., . . . Solecki, W. (2017). The roads ahead: Narratives for shared socioeconomic pathways describing world futures in the 21st century. Global Environmental Change, 42, pp.169180. O'Rourke, K. (1995). Computable General Equilibrium Models and Economic History. Retrieved from http://www.gams.de/solvers/mpsge/orourke.htm. Accessed on 10 July, 2014. Padgham, J. (2009). Agricultural Development under a Changing Climate: Opportunities and Challenges for Adaptation. Joint Department Discussion Paper, Issue-1. World Bank. Pardey, P. G., Alston, J. M., and Chan-Kang, C. (2013). Public agricultural R & D over the past half century: an emerging new world order. Agricultural Economics, 44, pp.103–113. Pauw, K. (2003). Functional forms used in CGE models: Modelling production and commodity flows. Background Paper 2003:5. Retrieved from http://ageconsearch.umn.edu/bitstream/15606/1/bp030005.pdf. Accessed on 20 July, 2015. PHC. (2007). Population and Housing Census. Central Statistical Agency. Retrieved from www.csa.gov.et. Accessed on 24 September, 2014. Pielke, R. (2007). Mistreatment of the economic impacts of extreme events in the Stern Review Report on the Economics of Climate Change. Global Environmental Change, 17, pp.302–310. Piermartini, R., and Teh, R. (2005). Demystifying Modelling Methods for Trade Policy. Discussion Paper No. 10. World Trade Organization. Project Catalyst. (2009). Scaling up Climate Finance. Retrieved from http://www.undpcc.org/undpcc/files/docs/publications/climate_finance.pdf. Accessed on 15 November, 2015. Ray, D. K., Gerber, J. S., MacDonald, G. K., and West, P. C. (2015). Climate variation explains a third of global crop yield variability. Nature Communications, DOI: http://dx.doi:10.1038/ncomms6989. Regional Strategic Analysis and Knowledge Support System (ReSAKSS) (2014). Ethiopia: Joint Sector Review Assessment Report. International Food Policy Research Institute. Roberts, B. M. (1994). Calibration Procedure and the Robustness of the CGE Models: Simulations with a model for Poland. Economics of Planning, 27, pp.189-210. Robinson, S., Strzepek, K., and Cervigni, R. (2013). The Cost of Adapting to Climate Change in Ethiopia: Sector-Wise and Macro-Economic Estimates. ESSP II Working Paper, No. 53. Ethiopia Strategic Support Program II. Robinson, S., van Meijl, H., Willenbockel, D., Valin, H., Fujimori, S., Masui, T., . . . von Lampe, M. (2014). Comparing supply-side specifications in models of global agriculture and the food system. Agricultural Economics, 45, pp.21-35. Robinson, S., Willenbockel, D., and Strzepek, K. (2012). A Dynamic General Equilibrium Analysis of Adaptation to Climate Change in Ethiopia. Review of Development Economics, 16(3), pp.489-502.

124

Bibliography

Robinson, S., Yu´nez-Naude, A., Hinojosa-Ojeda, R., Lewis, J. D., and Devarajan, S. (1999). From stylized to applied models: Building multisector CGE models for policy analysis. North American Journal of Economics and Finance, 10, pp.5–38. Rosenzweig, C., Elliott, J., Deryng, D., Ruane, A. C., Müller, C., Arneth, A., . . . Jones, J. W. (2014). Assessing agricultural risks of climate change in the 21st century in a global gridded crop model intercomparison. Proceedings of the National Academy of Sciences, 111 (9), 3268–3273. Salazar-Espinoza, C., Jones, S., and Tarp, F. (2015). Weather Shocks and Cropland Decisions in Rural Mozambique. Food Policy, 53, 9-21. Sanchez, C. M. (2004). Calibration, Solution and Validation of the CGE Model. Chapter 7 in C. M. Sanchez, Rising Inequality and Falling Poverty in Costa Rica's Agriculture During Trade Reform: A Micro-Macro General Equilibrium Analysis (pp. 189-226). Maastricht: Shaker. Schlenker, W., and Lobell, D. B. (2010). Robust negative impacts of climate change on African agriculture. Environmental Research Letters, 5(014010). Schultz, T. W. (1975). The Value of the Ability to Deal with Disequilibria. Journal of Economic Literature, 13(3), pp.827-846. Seo, S.N., and Mendelsohn, R. (2008). Measuring impacts and adaptations to climate change: a structural Ricardian Model of African Livestock Management. Agricultural Economics, 38, pp.151–165. Shoven, J. B., and Whalley, J. (1984). Applied General-Equilibrium Models of Taxation and International Trade: An Introduction and Survey. Journal of Economic Literature, 22(3), pp.1007-1051. Shoven, J. B., and Whalley, J. (1992). Introduction. Chapter 1 J.B. Shoven and J. Whalley, Applying General Equilibrium (pp. 1-36). Cambridge University Press. SMIS. (2003). Report on Small Scale Manufacturing Industries Survey. Central Statistics Authority. Retrieved from www.csa.gov.et. Accessed 24 November 2015. Smit, B., and Pilifosova, O. (2001). Adaptation to Climate Change in the Context of Sustainable Development and Equity. Chapter 18 in J. J. McCarthy, O. F. Canziani , N. A. Leary, D. J. Dokken , and K. S. White (Eds.), Climate Change 2001: Impacts, Adaptation and Vulnerability (pp. 876-912). Cambridge University Press. Smit, B., and Skinner, M. W. (2002). Adaptation Options in Agriculture to Climate Change: A Typology. Mitigation and Adaptation Strategies for Global Change, 7, pp.85–114. Smit, B., Burton, I., Klein, R. J., and Wandel, J. (2000). An Anatomy of Adaptation to Climate Change and Variability. Climatic Change, 45, pp.223-251. Smit, B., Burton, I., Klein, R., and Street, R. (1999). The Science of Adaptation: A Framework for Assessment. Mitigation and Adaptation Strategies for Global Change, 4, pp.199– 213. SNA. (2008). System of National Accounts 2008. European Commission, International Monetary Fund, Organisation for Economic Co-operation and Development, United Nations, and World Bank. Stark, O., and Taylor, J. E. (1989). Relative Deprivation and International Migration. Demography, 26(1), pp.1-14. Starr, R.M. (2011). General Equilibrium Theory: An Introduction (2nd Ed.). Cambridge: Cambridge University Press. Stern, N. (2007). The Economics of Climate Change: The Stern Review. Cambridge University Press.

Bibliography

125

Stifel, D., and Minten, B. (2008). Isolation and agricultural productivity. Agricultural Economics, 39, pp.1–15. Stifel, D., Minten, B., and Koru, B. (2016). Economic Benefits of Rural Feeder Roads: Evidence from Ethiopia. The Journal of Development Studies, 52(9), pp.1335-1356. Stiglitz, J. E. (2000). Economics of the Public Sector (3rd Ed.). New York: W.W. Norton and Company. Tafesse, A., Ayele, G., Ketema, M., and Geta, E. (2013). Adaptation to Climate Change and Variability in Eastern Ethiopia. Journal of Economics and Sustainable Development, 4(9), pp.91-103. Tall, A., Hansen, J., Jay, A., Campbell, B., Kinyangi, J., Aggarwal, P. K., and Zougmoré, R. (2014). Scaling up climate services for farmers: Mission Possible. Learning from good practice in Africa and South Asia. CAFS Report No. 13, CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS). Tanner, T., and Horn-Phathanothai, L. (2014). Climate Change and Development. New York: Routledge. Taylor, J. (2001). Migration: New Dimensions and Characteristics, Causes, Consequences and Implications for Rural Poverty. Retrieved from http://www.fao.org/docrep/003/X9808e/x9808e07.htm#ee. Accessed on 24 October, 2015. Taylor, L. (1990). Structuralist CGE Models. Chapter 1 in L. Taylor (Ed.), Socially Relevant Policy Analysis: Structuralist Computable General Equilibrium Models for Developing World (pp. 1-70). Cambridge: MIT Press. Teka, K., van Rompaeny, A., Poesen, J., Welday, Y., and Deckers, J. (2012). Impact of Climate Change on Small-Holder Farming: A Case of Eastern Tigray, Northern Ethiopia. African Crop Science Journal, 20(2), pp.337–347. Teruel, R. G., and Kuroda, Y. (2005). Public infrastructure and productivity growth in Philippine agriculture, 1974–2000. Journal of Asian Economics, 16, pp.555–576. Tessema, Y. A., Aweke, C. S., and Endris, G. S. (2013). Understanding the process of adaptation to climate change by small-holder farmers: the case of east Hararghe Zone, Ethiopia. Agricultural and Food Economics, 1(3). Tesso, G., Emana, B., and Ketema, M. (2012). Econometric analysis of local level perception, adaptation and coping strategies to climate change induced shocks in North Shewa, Ethiopia. International Research Journal of Agricultural Science and Soil Science, 2(8), pp.347-363. Thirtle, C., Lin, L., and Piesse, J. (2003). The Impact of Research-Led Agricultural Productivity Growth on Poverty Reduction in Africa, Asia and Latin America. World Development, 31(12), pp.1959–1975. Thornton, P. K., van de Steeg, J., Notenbaert, A., and Herrero, M. (2009). The impacts of climate change on livestock and livestock systems in developing countries: A review of what we know and what we need to know. Agricultural Systems, 101, pp.113–127. Tilahun, H., and Schmidt, E. (2012). Spatial Analysis of Livestock Production Patterns in Ethiopia. ESSP II Working Paper No.44. Ethiopia Strategy Support Program II. Tinbergen, J. (1952). On the Theory of Economic Policy. Amsterdam: North-Holland Publishing Company. Retrieved from http://hdl.handle.net/1765/15884. Accessed on 13 October, 2014.

126

Bibliography

Todaro, M. P., and Smith, S. C. (2012). Economic Development (11th Ed.). Boston: AddisonWesley. UDTS. (2013). Report on Urban Distributive Trade Survey 2010/11. Central Statistics Agency. Retrieved from www.csa.gov.et. Accessed 24 November, 2015. United Nations Convention to Combat Desertification (UNCCD). (2014). Desertification: the Invisible Frontline. Retrieved from www.unccd.int. Accessed on 23 October, 2016. United Nations Department of Economic and Social Affairs (UNESDA). (2008). International Standard Industrial Classification of All Economic Activities (ISIC): Revision 4. New York. United Nations Environment Programme (UNEP). (2014). The Adaptation Gap Report: Preliminary Assessment. Nairobi. United Nations Environment Programme (UNEP). (2016). The Adaptation Gap Report: Finance. Nairobi. United Nations Framework Convention on Climate Change (UNFCCC). (2007). Investment and Financial Flows to Address Climate Change. Bonn. United Nations Framework Convention on Climate Change (UNFCCC). (2009). Copenhagen Accord. Conference of the Parties, Fifteenth session, 7-18 December 2009, Copenhagen. Varian, H.R. (1992). Microeconomic Analysis. (3rd Ed.). New York: W. W. Norton & Company, Inc. Villora, N. B., Elliot, J., Müller, C., Shin, J., Zhao, L., and Song, C. (2016). Rapid aggregation of global gridded crop model outputs to facilitate cross-disciplinary analysis of climate change impacts in agriculture. Environmental Modelling and Software, 75, pp. 193201. Von Braun, J. (1991). A Policy Agenda for Famine Prevention in Africa. Food Policy Statement No. 13. International Food Policy Research Institute. Waha, K., Müller, C., Bondeau, A., Dietrich, J. P., Kurukulasuriya, P., Heinke, J., and LotzeCampen, H. (2013). Adaptation to climate change through the choice of cropping system and sowing date in sub-Saharan Africa. Global Environmental Change, 23, pp.130143. Watkiss, P., Downing, T. E., and Dyszynski, J. (2010). Analysis of the Economic Costs of Climate Change Adaptation in Africa. United Nations Environmental Programme. Wayne, G. (2013). The Beginner’s Guide to Representative Concentration Pathways. Retrieved from http://www.skepticalscience.com/docs/RCP_Guide.pdf. Accessed on 13 October, 2014. Weindl, I., Lotze-Campen, H., Popp, A., Müller, C., Havlík, P., Herrero, M., . . . Rolinski, S. (2015). Livestock in a changing climate: production system transitions as an adaptation strategy for agriculture. Environmental Research Letters, 10(094021). Westphal, M. I., Canfin, P., Ballesteros, A., and Morgan, J. (2015). Getting to $100 Billion: Climate Finance Scenarios and Projections 2020. Retrieved from www.wri.org/publication/getting-to-100-billion. Accessed on 13 August, 2016. White, J. W., Hoogenboom, G., Kimball, B. A., and Wall, G. W. (2011). Methodologies for simulating impacts of climate change on crop production. Field Crops Research, 124, pp.357-368. Wiebe, K. D., Soule, M. J., and Schimmelpfennig, D. E. (2001). Agricultural Productivity for Sustainable Food Security in sub-Saharan Africa. Retrieved from http://www.fao.org/docrep/003/x9447e/x9447e06.htm#bb. Accessed on 03 October, 2015.

Bibliography

127

Wiebelt, M., Al-Riffai, P., Breisinger, C., and Robertson, R. (2015). Who bears the costs of climate change? Evidence from Tunisia. Journal of Developing Areas, 49(2), pp.1-21. Wondemu, K. A., and Weiss, J. (2012). Rural Roads and Development: Evidence from Ethiopia. European Journal of Transport and Infrastructure Research, 12(4), pp.417-439. Wondimagegnhu, B. A. (2015). Staying or Leaving? Analyzing the Rationality of Rural-Urban Migration Associated with Farm Income of Staying Households: A Case Study from Southern Ethiopia. Advances in Agriculture, DOI: http://dx.doi.org/10.1155/2015/617373 World Bank. (2006). Ethiopia: Managing Water Resources to Maximize Sustainable Growth. Washington, DC. World Bank. (2008). Ethiopia: A Country Study on the Economic Impacts of Climate Change. Washington, DC. World Bank. (2010a). Economics of Adaptation to Climate Change: Ethiopia. Washington, DC. World Bank. (2010b). Economics of Adaptation to Climate Change: Synthesis Report. Washington, DC. World Bank. (2013). Turn Down the Heat: Climate Extremes, Regional Impacts, and the Case. Washington, DC. World Bank. (2016). Ethiopia: Climate Change Knowledge Portal. Retrieved from http://data.worldbank.org/country/ethiopia. Accessed on 11 July, 2016. World Development Indicators (WDI). (2016). World Development Indicators Retrieved from http://databank.worldbank.org/data/reports.aspx?source=world-development-indicators. Accessed on 23 October, 2016. World Economic Forum (WEF). (2016). The Global Competitiveness Report 2016–2017. Geneva. Wossen, T., Berger, T., and Di Falco, S. (2015). Social capital, risk preference and adoption of improved farm land management practices in Ethiopia. Agricultural Economics, 46, 81– 97. Yeldan, A. E. (1986). A Computable General Equilibrium Model for Development Policy Analysis. Bulletin, 86-1. Department of Agricultural and Applied Economics, University of Minnesota. Zepeda, L. (2001). Agricultural Investment, Production Capacity and Productivity. Retrieved from http://www.fao.org/docrep/003/x9447e/x9447e03.htm#c. Accessed on 03 October, 2015.

Appendix Materials and Methods Table A1-Description of the SAM accounts Account ACTIVITY

COMMODITY

Notation AGRAIN

Description Grains: Cereals, Pulses, Oilseeds

ACCROP

Cash Crops: Coffee, Chat, Tea, Spices, Sugarcane, Cotton

AENSET

Enset

ALIVST

Livestock: raising animals and production

AFISFOR

Fishery and Forestry

AMINQ

Mining and Quarrying

ACONS

Construction

AMAN

Manufacturing, Electricity, Water Supply

ATSER

Trade Services: Wholesale, Retail Trade, Repair services

AHSER

Hotels and Restaurants

ATRNCOM

Transport and Communications

AFSER

Financial Intermediaries

ARSER

Real Estate and Renting Business

APADMN

Public Administration (General)

APAGRI

Public Administration (Agriculture)

ASSER

Social Services: Education, Health Community Services

AOSER

Other Services: services n.e.c.*

CGRAIN

Grains: Cereals, Pulses, Oilseeds

CCCROP

Cash Crops: Coffee, Chat, tea, Spices, Sugarcane, Cotton

CENSET

Enset

CLIVST

Livestock: Raising Animals and Production

CFISFOR

Fishery and Forestry

CMINQ

Mining and Quarrying

CCONS

Construction

CMAN

Manufacturing, Electricity, Water Supply

CMMAN

Fertilizers, Coal, Natural Gas, and Petroleum Oil

CTSER

Trade Services: Wholesale, Retail Trade, Repair services

CHSER

Hotels and Restaurants

CTRNCOM

Transport and Communications

CFSER

Financial Intermediaries

CRSER

Real Estate and Renting Business

CPADMN

Public Administration (General)

© Springer Fachmedien Wiesbaden GmbH, part of Springer Nature 2020 A. W. Yalew, Economic Development under Climate Change, https://doi.org/10.1007/978-3-658-29413-7

130

FACTOR

INSTITUTION

TAX

S-I

ROW

Appendix CPAGRI

Public Administration (Agriculture)

CSSER

Social Services: Education, Health, Community Services

COSER

Other Services: services n.e.c.*

TRC

Transaction Costs (Transport and Trade Margins)

FLAB0

Agricultural Labor

FLAB1

Administrative workers (e.g., Legislators, Senior Officials, Managers)

FLAB2

Professional workers (e.g., Professionals, and Technical Associate)

FLAB3

Unskilled workers (e.g. workers engaged in elementary occupation

FLAB4

Skilled workers (e.g., workers that have some skill)

FLND

Cropland

FTLU

Tropical Livestock Units

FCAP

Nonagricultural Capital

ENT

Enterprises

GOV

Government

RURH

Rural Households

URBH

Urban Households

STAX

Sales (Domestic Indirect) Tax

MTAX

Import Tariffs and Duties

DTAX

Direct (Income) Tax

DSTK

Stock (Inventory) Changes

S-I

Savings-Investment

ROW

Rest of the World

Source: Author’s aggregation and disaggregation from EDRI (2009) Notes: (*) Business Activities, Other Community, Social and Personal Service activities, Private Households with Employed Persons

Unweighted

Weighted

Grains

Cash crops

Enset

Livestock

Fishery and forestry

Mining and quarrying

Construction

Manufacturing, electricity, water supply

Trade services

Hotels and restaurants

Transport and communications

Financial intermediaries

Real estate and renting business

Public administration (General)

Public administration (Agriculture)

Social services

Other services

Regional GDP

GDP1

GDP2

AGRAIN

ACCROP

AENSET

ALIVST

AFISFOR

AMINQ

ACONS

AMAN

ATSER

AHSER

ATRNCOM

AFSER

ARSER

APADMN

APAGRI

ASSER

AOSER

TOTAL

1.00

0.03

0.04

0.01

0.04

0.08

0.02

0.05

0.02

0.11

0.07

0.04

0.01

0.05

0.14

0.01

0.10

0.18

14.05

14.05

ETH

1.00

0.05

0.07

0.01

0.04

0.18

0.03

0.03

0.04

0.10

0.06

0.05

0.01

0.00

0.06

0.00

0.03

0.24

0.71

0.77

TIG

1.00

0.03

0.04

0.02

0.08

0.05

0.02

0.05

0.04

0.12

0.11

0.16

0.00

0.03

0.14

0.00

0.06

0.07

0.37

0.23

AFR

0.02 1.00

1.00

0.04

0.01

0.02

0.06

0.02

0.04

0.02

0.11

0.05

0.02

0.00

0.02

0.21

0.00

0.13

0.22

4.62

5.30

ORM

0.03

0.03

0.02

0.03

0.08

0.01

0.02

0.03

0.07

0.10

0.04

0.01

0.02

0.12

0.00

0.07

0.33

2.45

3.46

AMH

Source: Author’s computation based on HICES (2005), PHC (2007), and AgSS (2006)

Description

Activity

1.00

0.02

0.02

0.01

0.05

0.07

0.01

0.08

0.01

0.15

0.04

0.02

0.01

0.24

0.12

0.00

0.06

0.07

1.08

0.71

SOM

1.00

0.01

0.06

0.02

0.09

0.01

0.03

0.03

0.03

0.07

0.04

0.03

0.00

0.00

0.05

0.00

0.09

0.42

0.10

0.16

BNG

Table A2-Regional economic structure based on employment in HICES (2005) (GDP values are in billion USD)

1.00

0.01

0.04

0.01

0.03

0.04

0.00

0.02

0.02

0.13

0.06

0.01

0.00

0.08

0.19

0.05

0.18

0.13

2.83

2.80

SNNP

1.00

0.03

0.08

0.02

0.08

0.07

0.02

0.08

0.02

0.15

0.07

0.09

0.01

0.00

0.02

0.01

0.14

0.10

0.10

0.07

GAM

1.00

0.03

0.14

0.01

0.11

0.10

0.03

0.11

0.01

0.19

0.05

0.10

0.02

0.00

0.00

0.00

0.06

0.03

0.07

0.03

HAR

1.00

0.07

0.06

0.00

0.09

0.17

0.07

0.18

0.02

0.13

0.06

0.15

0.01

0.00

0.00

0.00

0.00

0.00

1.63

0.47

ADD

1.00

0.05

0.06

0.01

0.05

0.11

0.03

0.17

0.03

0.20

0.08

0.13

0.01

0.00

0.02

0.00

0.03

0.04

0.11

0.05

DD

Appendix 131

5 22 4 1 0 6 11

Clothing and footwear

Housing, water, fuel, and energy

Furnishing, household equipment and maintenance

Health and medical treatment

Education

Unincorporated household enterprise

Other goods & services, recreation, transportation, communication, restaurant/hotel 100

16

8

0

1

5

19

6

2

43

TIG

100

8

2

0

1

5

21

6

5

53

AFR

100

11

10

0

1

4

21

4

3

47

AMH

100

11

6

0

1

5

21

6

3

47

ORM

100

5

2

0

0

5

21

6

12

50

SOM

100

12

9

0

2

5

19

5

3

46

BNG

100

9

4

0

1

4

25

5

4

48

SNNP

100

10

1

0

1

5

22

4

3

53

GAM

100

8

9

1

1

4

23

4

10

41

HAR

100

18

0

2

1

5

29

5

1

39

ADD

100

10

1

1

1

4

27

6

7

44

DD

Source: Author’s computation from HICES (2011) Notes: ETH (Ethiopia), TIG (Tigray regional state), AFR (Afar regional state), AMH (Amhara regional state), ORM (Oromia regional state), SOM (Somali regional state), BNG (Benshangul-Gumuz regional state), SNNP (Southern nations, nationalities, and peoples regional state), GAM (Gambella regional state), HAR (Harari regional state), ADD (Addis Ababa city administration), and DD (Dire Dawa city council).

100

3

Alcohol, tobacco, coffee, tea, chat and buckthorn

Total

45

ETH

Food and non-alcoholic beverages

Item

Table A3-Consumption pattern of representative regional households (values in %)

132 Appendix

Appendix

133

Sensitivity analysis Table A4-Sensitivity of households’ welfare effects (selected simulations) Sensitivity runs (% change)

Household

Simulation

RURAL

EPIC-PM

-9.9

-9.8

-9.8

-10.0

-9.6

-9.6

-9.8

PAG2

-2.1

-2.1

-2.2

-2

-2.5

-2

-2.3

PAGGA

-0.9

-0.9

-0.9

-0.8

-1.2

-0.9

-1.0

PAGSS

-0.6

-0.6

-0.6

-0.5

-0.9

-0.6

-0.7

PAGF

-1.3

-1.3

-1.3

-1.1

-1.7

-1.2

-1.4

PAGD

-1.6

-1.6

-1.6

-1.4

-1.9

-1.5

-1.7

PAGS

-1.9

-1.9

-2

-1.8

-2.3

-1.8

-2

PAGT URBAN

TOTAL

SIM0

SIM1

SIM2

SIM3

SIM4

SIM5

SIM6

-2.1

-2.1

-2.1

-1.9

-2.4

-2

-2.2

-11.2

-9.4

-13.6

-10.6

-12.0

-8.4

-13.0

PAG2

1.6

1.4

1.7

1.3

2.2

1.1

2.1

PAGGA

0.5

0.5

0.6

0.0

1.0

0.3

0.8

PAGSS

0.2

0.1

0.2

0.3

0.5

0.1

0.3

PAGF

2.2

2.1

2.3

1.9

2.8

1.8

2.5

PAGD

-0.1

-0.2

0.0

-0.4

0.5

-0.5

0.3

PAGS

1.0

0.9

1.1

0.7

1.6

0.5

1.4

PAGT

1.1

1.0

1.2

0.8

1.8

0.7

1.7

EPIC-PM

EPIC-PM

-10.2

-9.7

-10.7

-10.1

-10.2

-9.3

-10.6

PAG2

-1.2

-1.2

-1.2

-1.2

-1.4

-1.3

-1.2

PAGGA

-0.6

-0.6

-0.6

-0.5

-0.7

-0.6

-0.6

PAGSS

-0.4

-0.4

-0.4

-0.4

-0.5

-0.4

-0.4

PAGF

-0.4

-0.4

-0.4

-0.4

-0.6

-0.5

-0.4

PAGD

-1.2

-1.2

-1.2

-1.2

-1.3

-1.3

-1.2

PAGS

-1.2

-1.2

-1.2

-1.2

-1.3

-1.3

-1.2

PAGT

-1.3

-1.3

-1.3

-1.3

-1.4

-1.3

-1.3

Source: CGE simulations Note: SIM0 (Base-run), SIM1 (all import substitution and export transformation elasticities increased by 25%)), SIM2 (all import substitution and export transformation elasticities decreased by 25%), SIM3 (all factor substitution elasticities increased by 25%), SIM4 (all factor substitution elasticities decreased by 25%), SIM5 (all Frisch and income elasticities increased by 25%), and SIM6 (all Frisch and income elasticities decreased by 25%).

The CGE results of different simulations in the text are robust to ±25% of the elasticities of factor substitution, import substitution, export transformation, income, and Frisch parameter (see Table 4.11). I present the case for households’ welfare effects for selected simulations in Table A4. Urban households’ welfare is slightly sensitive to many of the perturbations. Urban households are slightly better off (worse off) when elasticities are increased (decreased) by 25%

134

Appendix

compared to the base-run. Compared to other sensitivity runs, urban households’ welfare is relatively sensitive to income elasticities and Frisch parameter (SIM5 and SIM6). In contrast, relatively, rural households’ welfare effects are almost insensitive to ±25% of elasticities. This is expected as 40% of rural households’ consumption accrue to home commodities which will not directly respond to the market demand and supply shocks. Rural households’ welfare effects are relatively sensitivity to changes in elasticities of factor substitution (SIM3 and SIM4). In spite of these, however, one can see that the households’ welfare effects are generally robust to changes to elasticities. There is no major change to the rural and urban households’ welfare effects (across sensitivity simulations – SIM0 to SIM6) and in overall all rankings of adaptation finance schemes.