Climate Change and Development Series SHOCK WAVES Managing the Impacts of Climate Change on Poverty Stephane Hallegatte, Mook Bangalore, Laura Bonzanigo, Marianne Fay, Tamaro Kane, Ulf Narloch, Julie Rozenberg, David Treguer, and Adrien Vogt-Schilb Shock Waves Managing the Impacts of Climate Change on Poverty Shock Waves Managing the Impacts of Climate Change on Poverty Stephane Hallegatte, Mook Bangalore, Laura Bonzanigo, Marianne Fay, Tamaro Kane, Ulf Narloch, Julie Rozenberg, David Treguer, and Adrien Vogt-Schilb © 2016 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 18 17 16 15 This work is a product of the staff of The World Bank with external contributions. The findings, inter- pretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Hallegatte, Stephane, Mook Bangalore, Laura Bonzanigo,­ Marianne Fay, Tamaro Kane, Ulf Narloch, Julie Rozenberg, David Treguer, and Adrien Vogt-Schilb. 2016. Shock Waves: Managing the Impacts of Climate Change on Poverty. Climate Change and Development Series. Washington, DC: World Bank. doi:10.1596/978-1-4648-0673-5. License: Creative Commons Attribution CC BY 3.0 IGO ­ Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third- ­ party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to the Publishing and Knowledge Division, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@ worldbank.org. ISBN (paper): 978-1-4648-0673-5 ISBN (electronic): 978-1-4648-0674-2 DOI: 10.1596/978-1-4648-0673-5 Cover design: Bill Pragluski, Critical Stages. Library of Congress Cataloging-in-Publication Data has been requested. Contents Foreword. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Abbreviations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Climate change is a threat to poverty eradication. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 By 2030, rapid, inclusive, and climate-informed development can prevent most (but not all) climate change impacts on poverty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Emissions-reduction policies are required to remove the long-term threat from climate change, and need not threaten progress on poverty reduction . . . . . . . . . . . . . . . . . . 23 In conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 1 From Climate Change to Poverty and Back: A Framework. . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Climate change is an obstacle for people to escape poverty. . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Poverty reduction, socioeconomic trends, and non-climate policies affect climate risk . . . . . 40 The road map for our report. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 2 Bad Seed: Climate Change, Agriculture, and Food Security. . . . . . . . . . . . . . . . . . . . . . . . . . 49 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Climate change and climate policies will impact food security. . . . . . . . . . . . . . . . . . . . . . . . 50 Poor people are vulnerable to climate impacts through prices and ecosystems. . . . . . . . . . . . 56 Policies can avoid negative consumption effects and increase incomes. . . . . . . . . . . . . . . . . . 65 In conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 v v i    C O N T E N T S 3 Threat Multiplier: Climate Change, Disasters, and Poor People. . . . . . . . . . . . . . . . . . . . . . . 79 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 Climate change will worsen natural hazards in most regions of the world. . . . . . . . . . . . . . . 80 Poor people are often—but not always—more exposed to hazards . . . . . . . . . . . . . . . . . . . . 83 Poor people lose relatively more to disasters when affected. . . . . . . . . . . . . . . . . . . . . . . . . . . 91 The reasons why poor people are more at risk point to possible policy solutions. . . . . . . . . . 97 In conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 4 Under the Weather: Climate Change, Health, and the Intergenerational Transmission of Poverty. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Disease and poor health contribute to poverty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Climate change magnifies threats to health, especially for poor people . . . . . . . . . . . . . . . . . 116 Health care systems and development pathways play a critical role . . . . . . . . . . . . . . . . . . . 128 In conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 5 Lend a Hand: Poor People, Support Systems, Safety Nets, and Inclusion. . . . . . . . . . . . . . . . 141 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 Saving, borrowing, and insurance help people adapt to changes and cope with shocks, but are not always accessible for poor people . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Social protection schemes are critical for helping people adapt and cope with shocks, but must be flexible and easily scalable. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 Migration and remittances play an increasingly important role and need to be supported by policies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160 Voice and governance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162 In conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 Annex 5A. Case studies of social protection and risk management in Ethiopia, the Philippines, and Pakistan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 6 A Window of Opportunity: Climate-Informed Development and Pro-Poor Climate Policies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 By 2030, climate change will increase; but rapid, inclusive, and climate-informed development can minimize its impact on poverty. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 Pro-poor mitigation policies are needed to reduce the long-term threat of climate change. . . . 191 In conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 Boxes O.1 Agriculture is the key driver for climate change’s impact on poverty. . . . . . . . . . . . . . . . . 14 1.1 Multiple reports explore the complex relationship between development and climate change. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 1.2 A call for zero net CO2 emissions by 2100. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 C O N T E N T S   v ii 2.1  Climate change could pose major hurdles for Africa’s leading cocoa and coffee producers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 2.2 Climate-driven livestock diseases can have high economic costs. . . . . . . . . . . . . . . . . . . 60 2.3 The wider functions of ecosystems and biodiversity in rural livelihoods. . . . . . . . . . . . . 62 2.4 Mitigating losses from the 1998 flood in Bangladesh. . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 2.5 Despite significant benefits, adoption rates of conservation agriculture remain limited . . . . . 69 2.6 Securing local benefits from harnessing the forests to lower emissions . . . . . . . . . . . . . . 71 3.1 Climate change makes extreme weather events more likely or more intense. . . . . . . . . . 82 3.2 Large coastal cities: Wealthier places at risk of floods . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 3.3 In Mumbai, poor people are disproportionately exposed to floods. . . . . . . . . . . . . . . . . 88 3.4  Hidden costs of recurrent hazards for poor people in Mumbai and Ho Chi Minh City . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 3.5 In an uncertain future, developing into the wetlands of Colombo is dangerous . . . . . . 100 3.6 Reversing the degradation of hydrometeorological services. . . . . . . . . . . . . . . . . . . . . . 104 4.1 Getting harder to breathe. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 4.2 Dengue’s future hinges on whether development or climate change prevails. . . . . . . . . 120 4.3 The uncertain triangle of climate change, conflict, and poverty . . . . . . . . . . . . . . . . . . 125 4.4 Universal health coverage: Kenya’s bottom-up strategy . . . . . . . . . . . . . . . . . . . . . . . . . 130 5.1 Developing catastrophe insurance in Turkey through public-private partnerships . . . . . 145 5.2 Food provision and school feeding schemes are commonplace and effective. . . . . . . . . 151 5.3 Indexing as an automatic scale-up mechanism. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 5.4 Private insurance and social protection schemes are complements, not substitutes. . . . . 157 5A.1 How the PSNP helped households cope with Ethiopia’s 2011 food crisis . . . . . . . . . . . 167 6.1 It is possible to inform decision making, even in a context of deep uncertainty. . . . . . . . 181 6.2 Building two scenarios to explore the large uncertainty on the future of poverty. . . . . 182 6.3 Is there a trade-off between climate mitigation and reducing extreme poverty? . . . . . . 197 Figures O.1  Flows in and out of poverty in Andhra Pradesh are larger than their net effect on poverty. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 O.2 Climate change can significantly reduce food availability in poor regions. . . . . . . . . . . . . 4 O.3 Rainfall shocks in Uganda take a big toll on crop income, less so on consumption. . . . . . 5 O.4  Without environmental income, poverty rates could be much higher in (sub)tropical forest landscapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 O.5  Poor people in hotter countries—like Nigeria—live in hotter areas, but less so in cooler countries. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 O.6  When disasters hit in the past, poor people were more likely to be affected (panel a) … and poor people always lost relatively more than nonpoor people (panel b) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 O.7 If it gets too hot, productivity falls. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 O.8  Poor people have less access to financial tools, social protection, and private transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 O.9 Our model for estimating the number of people in poverty due to climate change. . . . . 13 Agriculture is the main sectoral driver explaining higher poverty due to BO.1.1  climate change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 O.10  Drought vulnerability is reduced by agricultural techniques that integrate trees and store carbon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 O.11  In poorer countries, half of all health expenditures are paid out of pocket, unlike in richer ones. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 v i i i    C O N T E N T S O.12 Poorer households need different types of solutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 O.13 Using the revenue from a carbon tax could boost social assistance. . . . . . . . . . . . . . . . . 25 1.1 The bigger the climate change, the bigger the total impact. . . . . . . . . . . . . . . . . . . . . . . . 35 1.2  Flows in and out of poverty in Andhra Pradesh are larger than their net effect on poverty. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 2.1 Climate change could sharply reduce crop yields. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 2.2  Sub-Saharan Africa and South Asia are the most vulnerable to climate-induced increases in agricultural prices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 2.3 Climate change can significantly reduce food availability in poor regions. . . . . . . . . . . . 55 2.4  Ill-designed land-mitigation climate policies could sharply increase agricultural prices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 2.5  Poor households spend a higher share of their expenditure on food than nonpoor households . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 2.6 In poorer countries, agriculture plays a vital role for poorer households’ incomes . . . . . 58 2.7 Food price rises could lead to big increases in extreme poverty in most countries. . . . . . 58 2.8 Rainfall shocks in Uganda take a big toll on crop income, less so on consumption. . . . . 59 2.9  Ecosystem-based incomes explain most rural income in (sub)tropical smallholder systems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 2.10 Without environmental incomes poverty rates could be much higher. . . . . . . . . . . . . . . 62 2.11 Improved cropping technologies increase resilience in the African drylands. . . . . . . . . . 68 2.12 Faster technological progress would dampen long-term increases in food production costs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 2.13 Drought vulnerability is reduced by agricultural techniques that integrate trees and store carbon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 3.1  Poor people in hotter countries—like Nigeria—live in hotter areas, but in cooler countries, less so. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 3.2 When disasters hit in the past, poor people were more likely to be affected . . . . . . . . . . 93 3.3 Poor people always lose relatively more than nonpoor people. . . . . . . . . . . . . . . . . . . . . 93 3.4 Home ownership in Tanzania encourages home investment . . . . . . . . . . . . . . . . . . . . . 102 3.5 Poorer people lack sufficient access to financial instruments. . . . . . . . . . . . . . . . . . . . . 103 4.1  Diarrheal diseases, respiratory diseases, and malaria contribute to record child mortality rates in Sub-Saharan Africa and South Asia. . . . . . . . . . . . . . . . . . . . . . . . . . . 113 4.2  In poorer countries, half of all health expenditures are paid out of pocket, unlike in richer ones. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 4.3 Health and funeral expenses are a major reason why households fall into poverty. . . . . 114 4.4  For poorer countries, access to better sanitation for the bottom 40 percent is much worse than for the top 60 percent. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 4.5 As incomes rise, the prevalence of diarrhea for children under five falls . . . . . . . . . . . . 123 4.6  Stunting projections for 2030 and 2050 suggest that regardless of the socioeconomic scenario, climate change will increase severe stunting among children under 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 4.7  Poor households in Mumbai face multiple stresses, with a key one the risk from floods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 4.8 If it gets too hot, productivity falls significantly. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 4.9 A lot of room to improve the quality and cost of health care in poor countries. . . . . . . 129 5.1 Poorer households need different types of solutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142 5.2  Coverage of poor people is often under 50 percent, and they often receive lower transfer amounts. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 5.3 Poor people in the poorest countries barely covered by social safety nets . . . . . . . . . . . 150 C O N T E N T S   i x B5.2.1 School feeding programs are the most prevalent type of social safety net. . . . . . . . . . . 151 5.4  Providing safety nets in the Horn of Africa and Sahel is affordable, but the cost is very volatile. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157 5.5 Within a country, remittances tend to be higher for the wealthier. . . . . . . . . . . . . . . . . . 162 5A.1  Multiple programs answer different needs in postdisaster contexts in the Philippines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168 6.1  Our model for estimating the number of people in poverty because of climate change. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183 6.2  Agriculture is the main sectoral factor explaining higher poverty due to climate change. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 6.3  One billion people living in the poorest countries emit less than 1 percent of global emissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 6.4  Energy consumption is low when GDP per capita is below $5,000, but then increases fast until $10,000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 6.5 Carbon neutrality is needed by 2100 to achieve climate goals. . . . . . . . . . . . . . . . . . . . 194 6.6  Investments in coal-related infrastructure have created large emission commitments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 6.7 Lower air pollution means lower mortality rates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196 B6.3.1 Energy use keeps rising with GDP even though less energy might be enough for basic human needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198 6.8  Recycling $100 from the global fossil fuel subsidy budget as a universal cash transfer would benefit poor people. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 6.9 Using the revenue from a carbon tax could boost social assistance. . . . . . . . . . . . . . . . 200 Maps O.1 The urban poor are more exposed to river floods in many countries. . . . . . . . . . . . . . . . . 8 O.2  Climate change impacts on poverty vary greatly across scenarios, with Africa and South Asia the most vulnerable. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 B2.1.1 Ghana and Côte d’Ivoire could experience a loss of area suitable for cocoa production by 2050. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 2.1 Risks to food security would be much reduced in a more prosperous future. . . . . . . . . . 66 3.1 Continued high emissions will mean many more “broiling” summer months. . . . . . . . . 81 3.2  With unmitigated climate change, total days under drought conditions will increase by more than 20 percent in most regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 B3.2.1 Most cities with the highest relative coastal flood losses are in South and Southeast Asia. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 3.3  Poor people are more exposed to river floods in many countries, especially in urban areas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 B3.3.1 Mumbai’s poor are over-represented in the Mithi River Basin flood zone. . . . . . . . . . . . 88 3.4  Sub-Saharan Africa’s and Asia’s poor tend to be more exposed to droughts than the nonpoor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 3.5  Poor people in most countries are more exposed to higher temperatures than nonpoor people. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 4.1 Most countries on track for significant declines in the incidence of malaria. . . . . . . . . . 118 4.2  By 2050, socioeconomic development should reduce malaria incidence, even with climate change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 6.1  Sub-Saharan Africa and—to a lesser extent—India and the rest of South Asia are the most vulnerable. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190 x    C O N T E N T S Tables O.1 Climate change threatens to worsen poverty, but good development can help . . . . . . . . 15 O.2  Many targeted actions can lower poor people’s vulnerability to climate change impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 1.1 Households in developing countries face many shocks. . . . . . . . . . . . . . . . . . . . . . . . . . . 37 1.2  Weather shocks hit the poorer populations the hardest in the Middle East and North Africa region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 2.1 Climate change risks for ecosystems and potential livelihood impacts across regions. . . . 63 B3.3.1 Poor people tend to be more exposed to floods in Mumbai . . . . . . . . . . . . . . . . . . . . . . . 88 3.1 Poor people in Mumbai suffered higher relative losses from the 2005 floods . . . . . . . . . 94 B3.4.1 The health of Ho Chi Minh City’s poor is especially vulnerable to flood impacts. . . . . . 95 3.2 Bangladesh’s poor became food-insecure after the 1998 Great Flood. . . . . . . . . . . . . . . 96 3.3 Mumbai’s poor spend a lot to regularly repair their dwelling . . . . . . . . . . . . . . . . . . . . 101 4.1 Many days of work are lost because of malaria episodes . . . . . . . . . . . . . . . . . . . . . . . 115 5.1 Social protection includes safety nets, social insurance, and labor market policies. . . . . 147 5.2  Methods for targeting beneficiaries with social safety nets are more or less appropriate during a crisis or after a disaster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154 B6.2.1 Our optimistic and pessimistic scenarios. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182 6.1  Climate change can have a large impact on extreme poverty, especially if socioeconomic trends and policies do not support poverty eradication. . . . . . . . . . . . . 188 Foreword Ending poverty and addressing climate each year. The poor live in uncertainty, just change are the two defining issues of our one natural disaster away from losing every- time. Both are essential to achieving sustain- thing they have. able global development. But they cannot be We need good, climate-informed develop- considered in isolation. ment to reduce the impacts of climate change This report brings together these two over- on the poor. This means, in part, providing arching objectives and explores how they can poor people with social safety nets and uni- be more easily achieved if considered together. versal health care. These efforts will need to It demonstrates the urgency of efforts to be coupled with targeted climate resilience reduce poverty and the vulnerability of poor measures, such as the introduction of heat- people in the face of climate change. It also resistant crops and disaster preparedness provides guidance on how to ensure that cli- systems. mate change policies contribute to poverty The report shows that without this type of reduction and poverty reduction policies con- development, climate change could force tribute to climate change mitigation and resil- more than 100 million people into extreme ience building. poverty by 2030. But with rapid, inclusive Our studies show that without action, cli- development that is adapted to changing cli- mate change would likely spark higher agri- mate conditions, most of these impacts can be cultural prices and could threaten food prevented. security in poorer regions such as Sub- Over the longer term, we will face the lim- Saharan Africa and South Asia. And in most its of what good development and risk man- countries where we have data, poor urban agement can achieve. Only immediate households are more exposed to floods than emissions-reduction policies can limit the the average urban population. long-term impacts of climate change on the Climate change also will magnify many poor. This report shows that these policies threats to health, as poor people are more need not burden, and can actually benefit, the susceptible to climate-related diseases such as poor, through the use of proven mechanisms malaria and diarrhea. As the report points such as social safety nets to mitigate the out, poverty reduction is not a one-way street. impact of higher energy prices. The interna- Many people exit or fall back into poverty tional community must also support poor xi x i i    F O R E W O R D countries that cannot provide such And ending extreme poverty will be more protection. achievable now—with limited climate change The report combines the findings from impacts—than later, when impacts are likely household surveys in 92 countries that to be larger. describe demographic structures and income The report shows us that the best way for- sources with the most recent modeling results ward is to design and implement solutions to on the impacts of climate change on agricul- end extreme poverty and stabilize climate tural productivity and food prices; natural change as an integrated strategy. Such con- hazards such as heat waves, floods, and certed action, implemented quickly and inclu- droughts; and climate-sensitive diseases and sively, can help ensure that millions of people other health consequences. are not pushed back into poverty by the mul- Based on these findings and results, the tifaceted impacts of climate change. report gives a renewed urgency to the objec- tive of eradicating extreme poverty by 2030 John Roome Ana Revenga while tackling climate change. Development Senior Director Senior Director and poverty alleviation reduce people’s vul- Climate Change Poverty and Equity nerability to the effects of a changing climate. World Bank Group World Bank Group Acknowledgments This book was written by a team led by Kristie Ebi, Chris Field, Michael Stephane Hallegatte and composed of Mook Oppenheimer, and Youba Sokona. Bangalore, Laura Bonzanigo, Tamaro Kane, Detailed comments on drafts of this book Ulf Narloch, Julie Rozenberg, David were provided by Javier Baez, Timothy Treguer, and Adrien Vogt-Schilb under the Bouley, Maurizio Bussolo, Shaohua Chen, supervision of Marianne Fay. Vikas Choudhary, Sarah Coll-Black, Carlo Background papers for this book were del Ninno, Quy-Toan Do, Jane Ebinger, Sam provided by Arild Angelsen, Ed Barbier, Anne Fankhauser, Charles Feinstein, Alan Fuchs, Biewald, Benjamin Bodirsky, Jan Börner, Ruth Hill, Hanan Jacoby, Dean Jolliffe, Thomas Bowen, Nils Brinckmann, Matthew Marie-Agnes Jouanjean, Eeshani Kandpal, Cantele, Michael Carter, Steve Davis, Therese Tom Kerr, Herve Levite, Leonardo Lucchetti, Dokken, Nicklas Forsell, Mykola Gusti, Jonna Lundvall, Olivier Mahul, Bradford Petr Havlík, Mario Herrero, Sarah Janzen, Mills, Mario Negre, Grzegorz Peszko, Jun Kelly Johnson, Brenden Jongman, Nikolay Rentschler, Kanta Rigaud, Carlos Rodriguez Khabarov, David Leclère, Hermann Lotze- Castelan, Karin Shepardson, Kalanidhi Campen, Aline Mosnier, Frederik Noack, Subbarao, Hans Timmer, and Sara Van Wie. Michael Obersteiner, Ilona Otto, Jisung Park, For valuable contributions and advice, the Archana Patankar, Alexander Popp, Evan team thanks Neil Adger, Syud Amer Ahmed, Sandhoefner, Hans Joachim Schellnhuber, Zoubida Allaoua, Margaret Arnold, Jehan Erwin Schmid, Petra Tschakert, Hugo Valin, Arulpragasam, Abigail Baca, Judy Baker, Ted Veldkamp, Philip Ward, Isabelle Weindl, John Balbus, Sushenjit Bandyopadhyay, Hessel Winsemius, and Sven Wunder. Arup Banerji, Diji Chandrasekharan Behr, Guidance was provided by the book’s peer Genowefa Blundo, Paula Caballero, Cyril reviewers. Internal peer reviewers included Caminade, Raffaello Cervigni, Michael Carter Brandon, Richard Damania, Francisco Chaitkin, Daniel Clarke, James Close, Louise H. G. Ferreira, Masami Kojima, Andrea Cord, Christophe Crepin, Marcio Cruz, Liverani, Tamer Rabie, Emmanuel Skoufias, Saurabh Dani, Alejandro De la Fuente, Chris and Mike Toman. External advisors included Delgado, Fionna Douglas, Paul Drummond, Purnamita Dasgupta, Stefan Dercon, Erick Fernandes, Marc Forni, Camille xiii x i v    A C K N O W L E D G M E N T S Funnel, Tam Giang, Ugo Gentilini, Francis The World Bank Group’s Publishing and Ghesquiere, Indermit Gill, Ilmi Granoff, Knowledge Unit managed the editorial ser- Rashmin Gunasekera, Kirk Hamilton, Steve vices, design, production, and printing of the Hammer, Niels Holm-Nielsen, Maddalena book, with Aziz Gökdemir, Patricia Katayama, Honorati, Ede Jorge Ijjasz-Vasquez, Oscar and Nora Ridolfi anchoring the process. Ishizawa, Maros Ivanic, Jan Kwakkel, Peter Others assisting with the book’s publication Läderach, William Lamb, Glenn-Marie included Bruno Bonansea (maps), Honora Lange, Jeffrey Lewis, Simon Lloyd, Patricio Mara (copyediting), Datapage International M a r q u e z , To m M c D e r m o t t , K a t i e (typesetting), Catherine Farley (proofreading), McWilliams, Stacy Morford, Michael and Bill Pragluski of Critical Stages (cover Morris, Rick Murnane, Urvashi Narain, Jim design). The World Bank Group’s General Neumann, Israel Osorio-Rodarte, Anand Services Department managed the printing. Patwardhan, Madhu Raghunath, Angel The team acknowledges the generous sup- Rangel, Narasimha Rao, Maurice Rawlins, port for the preparation of this book of the Rob Reid, Ana Revenga, Christopher Reyer, U n i t e d K i n g d o m ’s D e p a r t m e n t f o r Alex Robinson, Meerim Shakirova, Meera International Development (DFID) and the Shekar, Jagjeet Singh Sareen, Ben Stewart, Global Facility for Disaster Reduction and Anshuman Tiwari, Sailesh Tiwari, Renos Recovery (GFDRR). In particular, the team Vakis, Bernice Van Bronkhorst, Axel van thanks the main counterparts from each Trotsenburg, Nick Watts, Anne Zimmer, organization—Annika Olsson (DFID) and ­ and the International Committee on New Alanna Simpson (GFDRR). Integrated Climate Change Assessment The book was sponsored by the Climate Scenarios (ICONICS) group. Change Cross-Cutting Solutions Area of the The book was skillfully edited by Laura World Bank under the leadership of Rachel Wallace. Kyte and John Roome. Abbreviations 4P(s) Pantawid Pamilyang Pilipino Program (Philippines) AFOLU agriculture, forestry, and other land uses BUL Beneficiary Update List (Philippines) Cat-DDO(s) Catastrophe Deferred Drawdown Option CatMex catastrophe bond (Mexico) CCRIF Caribbean Catastrophic Risk Insurance Facility CDCP Citizen’s Damage Compensation Program (Pakistan) CRW Crisis Response Window DAFAC Disaster Affected Family Assistance Card (Philippines) DANIDA Danish International Development Agency DHS Demographic and Health Surveys DSWD Department of Social Welfare and Development (Philippines) EU European Union FONDEN Natural Disasters Fund (Mexico) FSP Food Security Program (Ethiopia) GDP gross domestic product GEF Global Environment Facility GFDRR Global Facility for Disaster Reduction and Recovery GFEI Global Fuel Economy Initiative GHG greenhouse gas GHI Global Hunger Index GLOBIOM Global Biosphere Management Model GRS Grievance Redress System (Philippines) GTAP Global Trade Analysis Project HABP Household Asset Building Program (Ethiopia) xv x v i    A B B R E V I A T I O N S HDI human development index HSSF Health Sector Services Fund I2D2 International Income Distribution Data Set IEA International Energy Agency IPCC Intergovernmental Panel on Climate Change kWh kilowatt-hour MAgPIE Model of Agricultural Production and its Impact on the Environment MDB multilateral or bilateral development bank MoA Ministry of Agriculture MSME micro-, small, and medium enterprises NADRA National Database Registration Authority (Pakistan) NCDD National Community-Driven Development program (Philippines) NDRRMF National Disaster Risk Reduction and Management Fund (Philippines) NDRRMP National Disaster Risk Reduction and Management Plan (Philippines) NGO nongovernmental organization ODA official development assistance OECD Organisation for Economic Co-operation and Development OSDMA Odisha State Disaster Management Authority (India) PCRAFI Pacific Catastrophe Risk Assessment and Financing Initiative PET Temporary Employment Public Works Program (Mexico) ppm parts per million PPP purchasing power parity PSNP Productive Safety Net Program (Ethiopia) PTSD post-traumatic stress disorder QRF Quick Response Fund (Philippines) RCP Representative Concentration Pathway RFM Risk Financing Mechanism (Ethiopia) SP social protection SRES Special Report on Emissions Scenarios SSP Shared Socioeconomic Pathways TCIP Turkish Catastrophe Insurance Pool tCO2 tons of carbon dioxide TWh terawatt-hour UNEP United Nations Environment Programme UNFCCC United Nations Framework Convention on Climate Change UNICEF United Nations Children’s Fund UNOCHA United Nations Office for the Coordination of Humanitarian Affairs WCFC Watan Card Facilitation Center (Pakistan) WFP World Food Program WHO World Health Organization Overview Introduction acknowledging that ending poverty is an utmost priority. The goal of maintaining Climate change threatens the objective of ­ climate change below a 2°C increase in global sustainably eradicating poverty. Poor people temperature above preindustrial levels—the and poor countries are exposed and vulner- very goal the international community has able to all types of climate-related shocks— committed to—will require deep structural natural disasters that destroy assets and changes in the world economy. These changes livelihoods; waterborne diseases and pests will affect the conditions under which poor that become more prevalent during heat people succeed or fail to escape poverty. waves, floods, or droughts; crop failure from Emissions-reduction policies can increase reduced rainfall; and spikes in food prices energy and food prices, which represent a that follow extreme weather events. Climate- large share of poor people’s expenditures. But related shocks also affect those who are not these same policies can be designed to protect, poor but remain vulnerable and can drag and even benefit, poor people—for instance, them into poverty—for example, when a by using fiscal resources from environmental flood destroys a microenterprise, a drought taxes to improve social protection. decimates a herd, or contaminated water Ending poverty and stabilizing climate makes a child sick. Such events can erase change will be two unprecedented global decades of hard work and asset accumula- achievements and two major steps toward tion and leave people with irreversible health sustainable development—that is, develop- consequences. Changes in ­ climate conditions ment that balances the economic, social, and caused by increasing concentrations of environmental considerations. But these two greenhouse gases (GHGs) in the atmosphere objectives cannot be considered in isolation: can worsen these shocks and slow down they need to be jointly tackled through an poverty reduction. integrated strategy. Ending poverty will not be possible if This report brings together these two ­ climate change and its effects on poor people objectives—ending poverty and stabilizing are not accounted for and managed in devel- ­ climate change—and explores how they can opment and poverty-reduction policies. But more easily be achieved if considered together. neither can the climate be stabilized without It examines the potential impact of climate 1 2    SHOCK WAVES change and climate policies on poverty reduc- Between now and 2030, climate policies tion. It also provides guidance on how to can do little to alter the amount of global create a “win-win” situation so that climate ­ warming that will take place. The only change policies contribute to poverty reduc- option, therefore, is to reduce vulnerability tion and poverty-reduction policies contribute through both targeted adaptation invest- to climate change mitigation and resilience ments and improved socioeconomic condi- building. tions (higher incomes and lower poverty The key finding of the report is that cli- and inequality). mate change represents a significant obstacle Although development and adaptation to the sustained eradication of poverty, but cannot prevent all negative impacts from future impacts on poverty are determined by climate change, by 2030 they can prevent policy choices: rapid, inclusive, and climate- or offset most of its effects on poverty. But informed development can prevent most development must be rapid and inclusive short-term impacts whereas immediate pro- to reduce poverty and provide poor people poor, ­emissions-reduction policies can drasti- with social safety nets and universal health cally limit long-term ones: coverage. It also needs to be climate informed—meaning that investments and • Climate-related shocks and stresses, development patterns do not create new already a major obstacle to poverty reduc- vulnerabilities and account for what we tion, will worsen with climate change. know about future climate conditions. Climate is involved in most of the And it needs to be accompanied by tar- shocks that keep or bring households into geted adaptation (like upgrades in flood poverty—notably, natural disasters (such defenses or more heat-tolerant crops). as floods that cause asset loss and disabil- • Immediate mitigation is required to ity); health shocks (such as malaria that remove the long-term threat that climate results in health expenditures and lost change creates for poverty eradication. labor income); and crop losses and food Mitigation need not threaten short-term price shocks (due to drought or crop progress on poverty reduction provided disease). policies are well designed and interna- Poor people are disproportionately tional support is available. affected—not only because they are often Our ability to manage increasing cli- more exposed and invariably more vul- mate change impacts is limited. To keep nerable to climate-related shocks but also long-term impacts on poverty in check, because they have fewer resources and global temperatures need to be stabilized receive less support from family, commu- at a safe level—which implies that net nity, the financial system, and even social global carbon emissions be brought down safety nets to prevent, cope, and adapt. to zero before the end of the century. Such Climate change will worsen these shocks an ambitious goal requires that all govern- and stresses, contributing to a decoupling ments act now to implement emissions- of economic growth and poverty reduc- reduction policies. These policies will tion, thereby making it even harder to unambiguously benefit poor people over eradicate poverty in a sustainable the long term, thanks to reduced climate manner. change impacts, and they can be designed • In the short run, rapid, inclusive, and not to slow down poverty reduction over climate-informed development can pre- the short term. vent most (but not all) consequences of All countries should pursue options that climate change on poverty. Absent such provide local and immediate benefits (like good development, climate change could less pollution, better health, improved result in an additional 100 million people energy access and efficiency, reduced living in extreme poverty by 2030. energy expenditures, and higher O V E R V I E W   3 agricultural productivity). Governments FIGURE O.1  Flows in and out of poverty in Andhra Pradesh are can protect the poor from the conse- larger than their net effect on poverty quences of those mitigation policies that could impose net costs and create Flows out of poverty 14% per year Drought, trade-offs—notably by strengthening social ­ irrigation failure, protection and cash transfers or reducing or crop disease Net flows Decreasing the flow from 2% per year taxes, possibly using revenues from energy 14% to 13% would halve involved in 44% or carbon taxes or fossil fuel subsidy poverty reduction of the cases Nonpoor removal. In poor countries where domestic Poor Increasing the flow from resources are insufficient to protect poor Weather events keep 12% to 13% would halve people, support from the international people poor through poverty reduction asset and human community is essential. This is particularly capital destruction true for investments with high upfront Flows into poverty costs that are critical to prevent lock-ins 12% per year into carbon-intensive patterns (such as for Source: Based on Krishna 2006. urban transport, energy infrastructure, or deforestation). as death and illness) that are influenced by cli- Climate change is a threat to mate and environmental conditions (like higher rainfall and more malaria outbreaks, poverty eradication or higher temperatures and more frequent Poverty reduction is not a one-way transi- diarrhea). In addition, climate risks affect the tion out of poverty: many people exit or fall behavior of people, who may reduce invest- back into poverty every year. For instance, ments and asset accumulation because of the over a 25-year period, every year an average possibility of losses and select lower-risk but of 14 percent of households in 36 communi- lower-return activities—a rational strategy to ties in Andhra Pradesh, India, escaped avoid catastrophic outcomes, but one that poverty and 12 percent of nonpoor house- ­ can keep them in poverty. holds became poor—resulting in a net The key question then is: How much will 2 ­ p ercent annual decrease in poverty climate change influence the flows in and out (­figure O.1). The fact that, in practice, the of poverty and affect poverty over time? This net flow out of poverty is much smaller than report reviews the evidence and provides new the gross flows in and out of poverty means quantification on the issue. It does this by that a relatively small change in the gross examining the impact of climate change on flows in and out of poverty can significantly three interacting channels that are already affect net flows and overall poverty dynam- affecting the ability of the poor to escape ics. In the India example, if the flow into ­ poverty—agricultural and ecosystem impacts, poverty increased from 12 to 13 percent per natural disasters, and health shocks—and year or the flow out of poverty slowed from then deriving policy implications. Here we 14 to 13 percent per year, the pace of pov- should note that climate change will have erty reduction would be reduced by half. other impacts (for example on tourism or Today, climate conditions or climate events energy prices) that are not reviewed and are already involved in many cases where assessed in this report, and a comprehensive households fall into poverty. They include estimate of all climate change impacts remains price shocks that can be linked to lower agri- out of reach. However, even a subset of all cultural production (as occurred after the possible impacts reveals worrying patterns on Russian droughts in 2010); natural disasters how changes in climate conditions would that destroy poor people’s assets and affect threaten the objective of eradicating extreme health and education; and health shocks (such poverty by 2030. 4    SHOCK WAVES We find that climate change already people through food production impacts, w ­ orsens—and will further exacerbate—­ higher consumption prices, and changes in climate-sensitive shocks and negative trends rural incomes. in the three sectors that we consider, consis- Lower crop yields and higher food prices. tent with recent reports from the World Bank Modeling studies suggest that climate change (2014a) and the Intergovernmental Panel on could result in global crop yield losses as large Climate Change (IPCC 2014; Olsson et al. as 5 percent in 2030 and 30 percent in 2080, 2014). We also show that there will be an even accounting for adaptive behaviors such impact on poverty and inequality because as changed agricultural practices and crops, poor people (i) are more often affected by more irrigation, and innovation in higher these negative shocks or trends (they are yield crops (Biewald et al., forthcoming; more exposed); (ii) lose more when affected, Havlík et al., forthcoming). Over the short relative to their income or wealth (they are term, climate change will also create some more vulnerable); and (iii) receive less sup- benefits, but mostly in cold and relatively rich port from family, friends, and community, countries, while poorer regions will be the and have less access to financial tools or most negatively affected. The expected yield social safety nets to help prevent, prepare for, losses are likely to translate into higher agri- and manage impacts. cultural prices; and climate change will make it more difficult, even with more trade, to ensure food security in regions like Sub- Poor people are more vulnerable Saharan Africa and South Asia. In a world to spikes in food prices and more with rapid population growth, slow economic dependent on agricultural and growth, and high GHG emissions (that is, a ecosystem-related income scenario in which global temperatures Impacts on agricultural production and increase by approximately 4 oC by 2100), prices—triggered by either gradual changes food availability in these regions could pla- in long-term climate trends or more frequent teau at levels far below current levels in devel- and severe natural disasters—will affect poor oped countries (figure O.2). FIGURE O.2  Climate change can significantly reduce food availability in poor regions a. Sub-Saharan Africa b. South Asia 90 90 Daily calories per capita availability Daily calories per capita availability relative to developed countries relative to developed countries 80 80 in 2015 (%) in 2015 (%) 70 70 60 60 2000 2030 2050 2080 2000 2030 2050 2080 No climate change Low emissions High emissions High emissions without CO 2 fertilization Source: Havlík et al., forthcoming. Note: Results are based on simulations from the Global Biosphere Management Model (GLOBIOM) in a scenario with large population growth and little ­economic growth. O V E R V I E W   5 But these estimates come with a high level FIGURE O.3  Rainfall shocks in Uganda take a big of uncertainty. They vary depending on the toll on crop income, less so on consumption type of climate, crop, and economic model applied, as well as on assumptions about CO2 20 fertilization (its presence should mean higher 15 Reduction (%) crop yields)—hence the −30 percent to +45 ­percent range in likely food price changes 10 in 2050 that is reported by the IPCC (Porter 5 et al. 2014). And they do not include local pollution and ozone, pests and crop diseases, 0 Bottom 40% Top 60% food losses along the supply chain, or natural household income household income disasters that could result in temporary, but Crop income Consumption very severe, food price shocks. In addition, emissions-reduction efforts Source: Hill and Mejia-Mantilla 2015. could affect food prices and availability. The Note: Values calculated based on Uganda National Household Survey IPCC concludes that large-scale, land-based UNHS 2005/6, UNHS 2009/10, UNHS 2012/13. Rainfall shock is represented by a 10 percent decrease in the Water Requirement Satisfaction Index, mitigation at the global scale, especially bio- estimated for Uganda. energy expansion, can reduce the availability of land for food production, with implica- tions for food security. In fact, new modeling As for the rural poor, the situation could simulations show that mitigation policies that be mixed. If production shocks are accompa- do not consider food security could have price nied by price rises, agricultural workers and impacts that are larger than those of climate farmers may benefit from higher wages and change (Havlík et al., forthcoming). However, earnings (Jacoby, Rabassa, and Skoufias more carefully designed mitigation policies 2014). So the net effect on income depends could lead to price impacts that are smaller on how food prices react to reduced global than those caused by unmitigated climate production and how demand and diets can change (Lotze-Campen et al. 2014). adjust over the short term and the long term. Changes in consumption and incomes. It also depends on the balance between local Losses in the agricultural sector and spikes in changes (which affect farmers’ production) food prices can push vulnerable consumers and global changes (which affect global food into poverty—take, for example, the 2008 prices). And it depends on institutions—­ food price spike that caused about 100 ­ million especially labor markets—that determine people to fall into poverty, or the 2010–11 how changes in revenues from agriculture are episode that increased poverty by 44 million. distributed between workers, landowners, Part of the problem is that poor people spend and traders. a larger share of their budget on food than the However, even if the net impact on income rest of the population, with nonagricultural is positive, it is unlikely to offset the negative rural households and urban residents the impacts of higher consumption prices on most vulnerable (Ivanic, Martin, and overall poverty. One study of 15 developing Zaman 2012). countries in various regions finds that climate- In addition, farmers would directly suffer induced price rises increase extreme poverty from production shocks that could reduce by 1.8 percentage points (Hertel, Burke, and income and consumption. Data from Uganda Lobell 2010). It also finds that, in parts of between 2005 and 2011 suggest that a Africa and Asia, climate-related price adjust- 10 ­percent reduction in water availability due ments could increase poverty rates for nonag- to a lack of rainfall reduces crop income by an ricultural households by 20–50 percent. average of 14.5 percent—and almost 20 per- Similarly, another study shows that a once-in- cent for the poorest households. Consumption 30-year climate extreme could double the also falls, but less so (­ figure O.3).­ number of poor urban laborers in the most 6    SHOCK WAVES vulnerable countries, including in Malawi, Natural hazards, to which poor people Mexico, and Zambia (Ahmed, Diffenbaugh, are often more exposed and almost and Hertel 2009). Our own simulations reach always more vulnerable, will become similar results (see below). more intense and frequent in many Another complicating factor is that climate regions change—especially when combined with local We are already experiencing an increase in stressors such as pollution and overuse— natural hazards. About 75 percent of the threatens ecosystems, which provide subsis- moderate hot extremes over land and tence production and safety nets for many 18 ­ p ercent of moderate precipitation people in rural areas. Poor smallholder com- extremes are attributable to global warming munities across (sub)tropical landscapes (Fischer and Knutti 2015). Even though depend on the extractive use of ecosystems some positive impacts are expected—such as for up to 30 percent of their income and often fewer cold spells—the frequency and inten- rely on ecosystem resources to keep them- sity of many hazards are expected to increase selves above the poverty threshold (figure in most places: O.4). Even though a precise quantification remains out of reach, a growing number of • Heat waves that are considered excep- studies document how increasing climate tional today will become common. In stress threatens the livelihoods of poor people Europe, the summer of the 2003 heat in a variety of rural contexts and forces them wave, which led to more than 70,000 to pursue new livelihood strategies. Over the deaths, will be an “average” summer at long term, climate change will even make the end of this century under a high-­ some ecosystems (such as small island states emissions scenario (a scenario in which or low-lying coastal areas) completely unin- the global mean temperature has increased habitable, forcing inhabitants to move. by about 4°C by 2100). • The number of drought days could increase by more than 20 percent in most of the world by 2080, and the number FIGURE O.4  Without environmental income, poverty rates could of people exposed to droughts could be much higher in (sub)tropical forest landscapes increase by 9–17 percent in 2030 and (Poverty rate in (sub)tropical smallholder systems) 50–90 percent in 2080. • The number of people exposed to river 60 floods could increase by 4–15 percent 50 in 2030 and 12–29 percent in 2080 W insemius et al., forthcoming), and (­ 40 coastal flood risks can increase rapidly Poverty rate (%) with sea level rise (Hallegatte et al. 2013). 30 Will poor people bear the brunt of these 20 climatic changes? Poor and nonpoor people 10 settle in risky areas for many reasons. Sometimes, they lack information about the 0 level of risk, or they do not account for this Latin South Asia East Asia Sub-Saharan Total information in their decisions (World Bank America Africa 2013, chapter 2). But at-risk areas are often With environmental income Without environmental income attractive in spite of the risk because they Source: Noack et al., forthcoming. offer economic opportunities, public services Note: Figure shows share of sampled households below the extreme-poverty line. Based on the or direct amenities, and higher productivity Poverty and Environment Network (PEN) dataset, including data from 58 sites in 24 countries. ­Environmental income describes income derived from ecosystem services (such as wood gathering and incomes. In some rural areas, proximity or root and berry picking) rather than agriculture. to water offers cheaper transport, and regular O V E R V I E W   7 floods increase agricultural productivity. FIGURE O.5  Poor people in hotter countries—like Nigeria—live in People settle in risky areas to benefit from hotter areas, but less so in cooler countries opportunities—such as coastal areas with export-driven industries or cities with large a. Poor people in hotter countries live in labor markets and agglomeration spillovers. hotter areas, but in cooler countries less so While these factors apply to rich and poor alike, local land and housing markets (or the 1.5 Poverty exposure bias for temperature availability of land) often push poorer people to settle in riskier, but more affordable, areas. 1.0 To shed more light on this issue, we inves- tigated poverty-specific exposure to flood, droughts, and extreme temperatures within 0.5 52 countries to obtain a first global estimate of the difference in exposure for poor and 0 nonpoor people. Our results show that for drought, most of –0.5 the analyzed population (85 percent) lives in countries where poor people are more exposed to droughts than the average –1.0 (Winsemius et al., forthcoming). Poor people 0 10 20 30 are also more exposed to higher temperatures: Annual monthly temperature, 1961–99 (°C) 37 out of 52 countries (56 percent of the pop- b. Nigeria is a good example ulation) exhibit an overexposure of poor peo- ple, with this bias stronger in hotter countries 70 where high temperatures are more likely to be detrimental (figure O.5). As for river floods, Household wealth index the results are mixed: poor people are more 60 exposed than the average in half of the coun- tries analyzed (60 percent of the population). In Africa, countries in the southwest exhibit a 50 strong overexposure of poor people, as do those with large rivers in west Africa (like 40 Benin, Cameroon, and Nigeria). Focusing on urban households, we find that in most coun- tries (73 percent of the population), poor 30 households are more exposed to floods than 28 30 32 34 the average (map O.1). This might be because Hottest month temperature experienced by household (°C) land scarcity is more acute in urban areas (than in rural areas), creating a stronger incen- Source: Park et al., forthcoming; World Bank 2015a. tive for the poor to settle in risky areas due to Note: Panel a plots country-level poverty exposure bias for temperatures against each country’s lower prices. This higher exposure to flood current climate. The poverty exposure bias is the share of poor people exposed to a hazard, divided by the share of the total population exposed, subtracted by 1. A positive bias means poor people risk for poor urban dwellers is also found are more exposed than the average. Panel b plots household-level wealth index and temperature using higher-resolution data on household within a country—Nigeria. location and flood hazards in Mumbai, India. Given that the dynamics of disasters and exposed to natural hazards is through in-depth poverty occur at a fine scale, studies of expo- case studies, analyzing household survey data sure at the national scale may miss important from disaster victims. Here again we find that mechanisms and small-scale differences, from poor people are generally more exposed, one city block to the next. An alternative way although there are exceptions—such as hurri- to examine whether poor people are more cane Mitch in Honduras (figure O.6, panel a). 8    SHOCK WAVES MAP O.1  The urban poor are more exposed to river floods in many countries (Poverty exposure bias for floods in urban areas) Source: World Bank (IBRD 41902, September 2015) based on Winsemius et al., forthcoming. Note: Exposure was calculated for river floods. FIGURE O.6  When disasters hit in the past, poor people were more likely to be affected (panel a) … and poor people always lost relatively more than nonpoor people (panel b) a. Exposure b. Vulnerability 100 100 households (% of annual income) Assets or income lost for affected Surveyed households affected by natural disaster (% of total) 80 80 60 60 40 40 20 20 0 0 1 2 3 as i lad 1 at 2 Gu a na Ho iti as rth as a r d i l gu or Vi a m ba ba n S pa al No e E ny alp Af t an h, h, h, ng sh, Gu sh, Ha r na Te lvad ur M ica ya du em um um Sa Ne es es es dl Ke nd cig et Ba ade e n ad ad ad M a Ho l gl gl l ng ng n n Ba Ba Ba Ba id M Poor Nonpoor Source: See sources in Chapter 3. Note: Each Bangladesh case represents a unique study. O V E R V I E W   9 by the dependence on ecosystems and the large fraction of their budget dedicated to food. As a result of these differences in exposure and vulnerability, natural disasters increase inequality and may contribute to a decoupling of economic growth and poverty reduction. It is thus not surprising that natural disasters are found to worsen poverty. For instance, between 2000 and 2005, floods and droughts increased poverty levels in affected Mexican municipalities by 1.5 to 3.7 percent (Rodriguez-Oreggia et al. 2013). After Ethiopia’s 1984–85 famine, it took a decade on average for asset-poor households to bring livestock holdings back to prefamine levels (Dercon 2004). Poor people are strongly affected by diseases and health issues that climate change is likely to magnify Climate change will magnify some threats to health, especially for poor and vulnerable people—such as children. The exact impacts As for assets and income, nonpoor people are still highly uncertain in what is still an lose a larger amount in absolute terms emerging research field. Past progress on because they have more assets and higher medical treatment offers hope that some of incomes than the poor. But in relative terms, these issues could be solved over the long poor people always lose more than the non- term thanks to new drugs and better health poor, according to the five surveys that report infrastructure. But short-term impacts could the magnitude of natural disaster losses, dis- still be significant. tinguishing by income classes (figure O.6, Health shocks are important for poverty panel b). And it is these relative losses, rather dynamics and the impact of climate change than absolute ones, that matter most for live- for three main reasons. First, the main dis- lihoods and welfare. eases that affect poor people are diseases that Poor people are losing relatively more to are expected to expand with climate change disasters for two main reasons. First, they (such as malaria and diarrhea). Second, often do not save at financial institutions, and health expenditures are regressive, with poor they hold most of their wealth in vulnerable households largely uninsured—such outlays forms, such as housing for urban dwellers and push an estimated 100 million people per livestock for rural households. Second, the year into poverty—and the loss of income for quality of their assets—and the resistance of the sick or the caregiver can have a large those assets to natural hazards—is often lower impact on family prospects (WHO 2013). than average: typical houses found in a slum Third, children are most vulnerable to these can be completely destroyed in a common shocks and can suffer from irreversible flood whereas modern houses or multifamily impacts that affect their lifetime earnings and buildings are much more resistant. And poor lead to the intergenerational transmission of people’s overall vulnerability is exacerbated poverty. 1 0    SHOCK WAVES Malaria. Even small temperature increases and other vectorborne or waterborne dis- could significantly affect the transmission of eases, but these interactions have not yet been malaria. At the global level, warming of 2°C investigated in the context of climate change. or 3°C could increase the number of people at Also impossible to quantify is the impact on risk for malaria by up to 5 percent, or more mental disorders and stress due to increased than 150 million people. In Africa, malaria risk, disasters, or indirect impacts through could increase by 5–7 percent among popula- physical health, household dynamics, or com- tions at risk in higher altitudes, leading to a munity well-being. And changes in climate potential increase in the number of cases of and environmental conditions will interact up to 28 percent (Small, Goetz, and Hay with local air pollution and allergen distribu- 2003). Further, climate change is projected to tion, exacerbating respiratory diseases. One intensify malaria along the current edges of its estimate is that climate change could cause distribution, where malaria control programs annually an additional 100,000 premature are often nonexistent and people have no nat- deaths associated with exposure to small par- urally acquired immunity against the disease. ticulate matter and 6,300 premature deaths Diarrhea. Climate impacts could increase associated with ozone exposure (Fang et al. the burden of diarrhea by up to 10 percent by 2013). 2030 in some regions (WHO 2003). Indeed, Another concern is that high temperatures higher temperatures favor the development of will reduce labor productivity of those who pathogens, and water scarcity affects water are poorer and often work outside or with- quality and the hygiene habits that can out air conditioning (figure O.7). The impact ­ prevent diarrhea. An estimated 48,000 addi- on labor productivity could be large and tional deaths among children under the age of reduce income by several percentage points. 15 resulting from diarrheal illness are pro- Moreover, this effect is not accounted for in jected by 2030 (Hales et al. 2014). And cli- any of the studies we reviewed on estimates mate change could contribute to outbreaks of of agricultural production, although it could other waterborne diseases such as cholera and magnify food security issues. In addition, schistosomiasis. new research suggests that extreme tempera- Stunting. In part because of its impacts on ture stress in either direction—hot or cold— agriculture (figure O.2), climate change will is suboptimal for economic activity, even increase undernutrition and could sharply when considering only nonfarm activities. increase severe stunting among children. By These results imply that the temperature- 2030, an additional 7.5 million children may related loss in performance observed in be stunted (Hales et al. 2014). Climate change could even lead to an absolute increase in the number of stunted children in some parts of FIGURE O.7  If it gets too hot, productivity falls Africa, with the negative effect of climate (Task performance under different temperatures) change outweighing the positive effect of eco- nomic growth (Lloyd, Kovats, and Chalabi 1.00 2011). And recent evidence suggests that the Relative performance nutritional quality of food (for example, its 0.95 content in terms of micronutrients such as iron, iodine, vitamin A, folate, and zinc) could 0.90 also be affected by climate change, even 0.85 though little is known about potential impacts (Myers et al. 2014). 0.80 Even less is known about the combined 15 20 25 30 35 effects of multiple health stressors. For Temperature (°C) instance, it is well known that undernour- ished children are more vulnerable to malaria Source: Based on Seppänen, Fisk, and Lei 2006. O V E R V I E W   1 1 laboratories and at the individual level may FIGURE O.8  Poor people have less access to be observable at the macroeconomic level, financial tools, social protection, and private transfers and that climate change could hurt overall income through this channel (Deryugina and Hsiang 2014; Heal and Park 2013; a. Access to savings 80 Park et al., forthcoming). Access to savings (% of population) 70 60 Poor people receive less support from 50 friends and family and have more limited access to financial tools and 40 social safety nets 30 20 Many policy instruments exist that could help poor people prevent, adapt to, and 10 cope with climate shocks and changes 0 0 5,000 10,000 15,000 20,000 25,000 (World Bank 2013), but poor people have GDP per capita (US$, PPP 2011) only limited access to them (figure O.8). Take the case of financial inclusion—­ b. Public transfers received Public transfers received (US$/year) meaning access to formal savings, borrow- 3,000 ing, and insurance products (figure O.8, 2,500 panel a). People may lack access to these formal financial tools for several reasons, 2,000 including the cost of bank accounts, dis- 1,500 tance and time to access a financial agent, lack of documentation, or mistrust in 1,000 banks. Some people also prefer to stay in 500 the informal sector, or are not aware of the 0 benefits of using financial tools for risk 0 5,000 10,000 15,000 20,000 25,000 management (Allen et al. 2012). GDP per capita (US$, PPP 2011) Poor people also receive limited support c. Private transfers received from social safety nets, ranging from cash Private transfers received (US$/year) transfers to work programs (figure O.8, 900 panel b). In many countries, social programs 800 cover less than half of the poorest quintile. In 700 addition, even when poor households are cov- 600 ered by social protection schemes, amounts 500 received are often too small to make a big dif- 400 300 ference and prevent negative coping strate- 200 gies. In Bangladesh after the 1998 floods, 100 poor affected households had to borrow an 0 amount equal to six to eight times the level of 0 5,000 10,000 15,000 20,000 25,000 government transfers (del Ninno, Dorosh, GDP per capita (US$, PPP 2011) and Smith 2003). Poor Nonpoor Then, too, migration and remittances play a key role in managing shocks—but migra- Source: World Bank computation based on the FINDEX and ASPIRE tion requires resources and assets that the ­databases. Note: Panel a is based on data from FINDEX. Panels b and c are based poorest lack, and data show that remittances on data from ASPIRE. Each country is represented with two dots. Poor tend to benefit nonpoor people more than people are those in the bottom 20% (ASPIRE) or bottom 40% (FINDEX). Nonpoor people are those in the top 80% (ASPIRE) or top 60% (FINDEX). poor people (figure O.8, panel c). As a result, PPP = purchasing power parity. poor people are disproportionally affected by 1 2    SHOCK WAVES climate change and natural shocks, not only We then introduce into each of these sce- because they are more exposed and vulnera- narios estimates of climate change impacts on ble to them but also because they receive less food price and production, natural disasters, support. and health and labor productivity, based on the reviews and analyses presented in the report. But we do so with two climate change By 2030, rapid, inclusive, and impact scenarios—a low-impact and a high- impact scenario—given that the physical and climate-informed development biological impacts will be highly uncertain, can prevent most (but not all) dependent on (i) how ecosystems adapt and climate change impacts on physical systems (like glaciers and coastal zones) respond and (ii) how sectors spontane- poverty ously adapt (like adopting new agricultural Just how large might these impacts be on practices or improved hygiene habits). poverty by 2030 and how much can devel- We do not attribute probabilities or likeli- opment help? We know that between now hoods to the development and climate impact and then, climate policies will have minimal scenarios because we are not interested in impacts on warming, given the long lag forecasting the future of poverty (it is proba- between the introduction of mitigation poli- bly impossible). What interests us is the con- cies, their impact on emissions, and the effect trast across scenarios rather than the absolute of emissions reductions on the climate sys- numbers. That is why we focus on how the tem (IPCC 2014). This means that, by 2030, impacts of climate change on poverty would the only way to reduce climate change differ if development is rapid and inclusive impacts will be by lowering socioeconomic (“Prosperity”) as opposed to slow and nonin- vulnerability to these impacts—which will clusive (“Poverty”). require climate-informed development and The bottom line is that, even though our specific actions to adapt to climate change. analysis looks only at the short term with lim- ited changes in climate conditions, it still finds that climate change could have a large effect The magnitude of future climate change on extreme poverty: by 2030, between 3 and impacts on poverty depends on today’s 16 million people in the prosperity scenario choices and between 35 and 122 million people in the In this report, we try to get a sense of the poverty scenario would be in poverty because magnitude of future climate change of climate change. impacts—and how this magnitude depends That said, these estimates are likely an on today’s choices—by creating two scenar- underestimate for several reasons. First, we ios for what the future of poverty could be follow a bottom-up approach and sum the by 2030 in the absence of climate change sector-level impacts, assuming they do not (figure O.9). The first one, “ Prosperity,” interact. Second, we consider only a subset assumes that the World Bank’s goals of of impacts, even within the three sectors we extreme poverty eradication and shared focus on. For instance, we do not include prosperity are met by 2030 (in particular, losses in ecosystem services and reduced less than 3 percent of the world population nutritional quality of food; we consider only remains in extreme poverty), and that access consumption poverty, disregarding outcomes to basic services is quasi-universal. The sec- like the nonmonetary effects of disease; and ond scenario, “Poverty,” is much more pes- we do not include secondary impacts of simistic in terms of poverty reduction and disasters (like the potential effect on migrants inequalities (for instance, 11 percent of the and refugees). Third, we cannot assess the world population remains in extreme poverty impact everywhere. Our scenarios poverty). are developed based on a household O V E R V I E W   1 3 FIGURE O.9  Our model for estimating the number of people in poverty due to climate change (A schematic to represent the modeling undertaken to estimate the impact of climate change on extreme poverty in 2030 under different scenarios of future development, and thus in worlds with different levels of exposure and vulnerability) In the absence of climate change, we can imagine two different ways for the world to evolve Prosperity Poverty More optimistic on: Less optimistic on: • Economic growth • Economic growth • Poverty • Poverty • Inequality • Inequality • Basic services • Basic services With climate change, we can be more or less optimistic on the future magnitude of sectoral impacts Low impact High impact There are uncertainties on the impacts, in the short and the long run. By 2030, di erences in the physics (and biology) of climate change and sectoral adaptation to climate impacts may give us di erent outcomes (e.g., on local rainfall patterns and crop yields). By 2080, the level of emissions, and thus development patterns and climate mitigation polices, also matter. We introduce climate change impacts from the low-impact and high-impact scenarios into each scenario without climate change (Prosperity and Poverty). We model what poverty looks like in each scenario and then compare the difference. What development can achieve: Comparing the effect of low-impact climate change on poverty, in a world that would be more or less prosperous in the absence of climate change Uncertainty from the magnitude of climate change impacts What development can achieve: Comparing the effect of high-impact climate change on poverty, in a world that would be more or less prosperous in the absence of climate change Note: Photos © Masaru Goto / World Bank (low impact image) and Arne Hoel / World Bank (high impact image). Further permission required for reuse. 1 4    SHOCK WAVES BOX O.1  Agriculture is the key driver for climate change’s impact on poverty Which poverty channel is the dominant influence impacts (diarrhea, malaria, and stunting) and the in our four scenarios? As figure BO.1.1 shows, labor productivity effects of high temperature have agriculture is the main driver of the impact on pov- a second-order but significant role. Disasters have a erty (in all four scenarios)—although its impact is relatively limited role, but this may be because only possibly an underestimation because all climate the direct impact on income losses was taken into scenarios assume CO 2 fertilization. Then, health account. FIGURE BO.1.1  Agriculture is the main sectoral driver explaining higher poverty due to climate change (Summary of climate change impacts on the number of people living below the extreme poverty threshold, by driver) Prosperity scenario (high impact) Prosperity scenario (low impact) Poverty scenario (high impact) Poverty scenario (low impact) 0 20 40 60 80 100 120 140 Additional people (million) below the extreme poverty threshold by 2030 Agriculture Health Labor productivity Disasters Source: Rozenberg and Hallegatte, forthcoming. database that r­ epresents only 83 percent of (like small islands or in locations affected by the developing world’s population. Some large disasters), the local impact could be high vulnerability countries (such as small massive. islands) could not be included because of Note that the large range of estimates in data limitations, in spite of the large effects our results may incorrectly suggest that we that climate change could have on their pov- cannot say anything about the future impact erty rate. of climate change on poverty. The main rea- Although climate change has a significant son for this wide range is not scientific uncer- impact on poverty up to 2030—working pri- tainty on climate change and its impacts. marily through the agricultural channel I n s t e a d , p o l i c y c h o i c e s d o m i n a t e —­ (box O.1)—it remains a secondary driver, as particularly those concerning development evidenced by the nearly 800 million person patterns and poverty-reduction policies difference between the two socioeconomic between now and 2030. While emissions- scenarios in the absence of climate change reduction policies cannot do much regarding (table O.1). This does not mean that climate the climate change that will happen between change impacts are secondary at the local now and 2030 (because that is mostly the scale: in some particularly vulnerable places result of past emissions), development O V E R V I E W   1 5 TABLE O.1  Climate change threatens to worsen poverty, but good development can help Climate change scenario No climate change Low-impact scenario High-impact scenario Number of people in extreme Additional number of people in extreme poverty due to Policy choices poverty by 2030 climate change by 2030 Prosperity scenario 142 million +3 million +16 million Minimum Maximum Minimum Maximum +3 million +6 million +16 million +25 million Poverty scenario 900 million +35 million +122 million Minimum Maximum Minimum Maximum −25 million +97 million +33 million +165 million Source: Rozenberg and Hallegatte, forthcoming. Note: The main results use the two representative scenarios for prosperity and poverty. The ranges are based on 60 alternative poverty scenarios and 60 alternative prosperity scenarios. choices can affect what the impact of that cli- prosperity scenario in 2030, by 2 percent mate change will be. globally and by up to 20 percent in some Also note that the range of possible impacts African countries due to more migration. is even larger than the one represented by our A smaller population mitigates the impact of four scenarios because there is an infinite climate change on food prices. In addition, number of possible socioeconomic pathways the prosperity scenario assumes more tech- by 2030, even without climate change. To nology transfers to developing countries, assess the robustness of our results, we create which further reduces the agricultural loss 60 alternative prosperity and 60 alternative due to climate change. In the prosperity sce- poverty scenarios. We find that the range of nario, a more balanced economy and better possible impacts on poverty remains limited governance also mean that farmers capture a in the prosperity scenario: development not larger share of the income benefits from only reduces the impacts but also protects us higher food prices. from the uncertainty. In the poverty scenario, At the country and regional level, the on the other hand, the range of possible out- hotspots are Sub-Saharan Africa and—to a comes is extremely large: the worst-case esti- lesser extent—India and the rest of South Asia mate increases up to 165 ­ million, and some (map O.2). Almost all countries are less vul- scenarios show a decrease in global poverty nerable to climate change in the prosperity numbers—these are scenarios where climate scenario, often dramatically: in India, the change impacts remain moderate (low- high-impact climate change scenario brings impact) and where farmers benefit the most 2 million people into poverty in the prosperity from higher agricultural prices. scenario, compared to almost 50 million in The lower vulnerability of the developing the poverty scenario. One exception is the world to climate change in the prosperity sce- Democratic Republic of Congo, where cli- nario comes from several channels. First, peo- mate change is found to bring more people ple are wealthier and fewer households live into poverty in the prosperity scenario. This with a daily income close to the poverty line. occurs because, in the poverty scenario with- Wealthier individuals are less exposed to out climate change, the poverty rate is health shocks such as stunting and diarrhea, extremely high (70 percent): climate change and are less likely to fall back into poverty draws fewer people into poverty than in the when hit by a shock. And with fewer farmers, prosperity scenario only because so many the population is less vulnerable to the nega- people are already in poverty. tive impacts of climate change on yields. Such a result warns us against using a pov- Second, the global population is smaller in the erty headcount as the unique indicator of the 1 6    SHOCK WAVES MAP O.2  Climate change impacts on poverty vary greatly across scenarios, with Africa and South Asia the most vulnerable (Increase in number of extreme poor people due to climate change in the high-impact climate scenario (% of total population)) Source: World Bank (IBRD 41903 and IBRD 41904, September 2015) based on Rozenberg and Hallegatte, forthcoming. O V E R V I E W   1 7 impact of climate change on poverty. Because shocks, including those magnified by climate it does not measure poverty depth, it does not change (table O.2). Of course, each country capture the impact on people who are already can identify its own priorities, based on the poor. For instance, in a high-impact climate impacts of climate change that are expected scenario, climate change reduces the income on its territory, but also on synergies and con- of the bottom 40 percent in 2030 by more vergence with other policy priorities. For than 4 percent in most of the countries in instance, where urban planning is a policy both the prosperity and poverty scenarios. priority, mainstreaming natural hazards and And, in most Sub-Saharan African countries climate change into its design is a low-­hanging and Pakistan, climate change reduces the fruit waiting to be plucked. income of the bottom 40 percent by more Climate-smart agriculture and protected than 8 percent. ecosystems. Climate-smart agricultural prac- tices can increase productivity and resilience (Cervigni and Morris 2015). More produc- Climate-informed development needs tive and more resilient practices, however, to be complemented with targeted require a major shift in the way land, water, adaptation interventions and a more soil nutrients, and genetic resources are man- robust safety net system aged to ensure that these resources are used Rapid and inclusive development can prevent more efficiently (FAO 2013). Crop improve- most of the impact of climate change on pov- ment, smarter use of inputs, approaches to erty, but only if new investments and devel- strengthen crop resistance to pests and dis- opments are climate informed —that is, eases, and reduction of post-harvest losses designed to perform well under changing cli- can contribute to the sustainable intensifica- mate conditions so that they do not create tion of agriculture—thereby leading to new vulnerabilities to climate impacts. For greater food production (Beddington 2010; example, new water and sanitation infra- Tilman et al. 2011). structure can make a big difference for diar- For this to happen, innovation is needed to rhea, but only if it can absorb the more keep increasing yields, and the new techniques extreme rainfall episodes that are expected in that result from innovation must actually be many regions. Similarly, new settlements in broadly adopted, including by poor farmers. safe areas will reduce the long-term vulnera- These two conditions are challenging. First, bility only if the selected areas remain safe in yield increases have plateaued in recent years, spite of sea level rise and accelerated erosion. even exhibiting abrupt decreases in some However, even a rapid, inclusive, and cli- regions (Grassini, Eskridge, and Cassman mate-informed development will not cancel 2013). The low and declining levels of invest- out the need for targeted actions that are ment in agricultural research and develop- aimed at lowering people’s vulnerability to ment in the developing world are a major climate change impacts. Although some of constraint to realize further yield gains in poor them are pure climate change adaptation countries (Pardey, Alston, and Chan-Kang measures (like adapting building norms to 2013). Second, disseminating improved tech- new environmental conditions), others (like nologies and making them accessible to poor increasing financial inclusion) can be seen as farmers is difficult, and even promising inno- “good development” and would make sense vations sometimes have low or no uptake. even in the absence of climate change. High implementation costs, cultural barriers, Our report highlights potential options in and lack of access to information and educa- the three sectors that we focus on (agriculture tion need to be overcome. Agricultural exten- and ecosystems, natural disasters, and health) sion services can help farmers make better and emphasizes the potential of social protec- use of new technologies. In Uganda, exten- tion and financial tools to boost the resilience sion visits increased household agricultural of households and economies to all sorts of income by around 16 percent when new crop 1 8    SHOCK WAVES TABLE O.2  Many targeted actions can lower poor people’s vulnerability to climate change impacts International Sectoral options to reduce vulnerability Private sector Governments community Agriculture, ecosystems, and food security Adopt climate-smart technologies and agricultural practices, with X X support from agricultural extension Develop higher yielding and more climate-resistant crop varieties and livestock breeds, adapted to developing country contexts and climate X X X conditions Develop transport infrastructure and facilitate market access (domestic X X and international) Reduce non-climate stresses on ecosystems, including through X X conservation and ecosystem-based adaptation Natural disasters and risk management Increase financial inclusion and participation in banking to reduce the X X vulnerability of poor households’ assets Improve households’ and firms’ preparedness and ability to act upon X X warnings (contingency plans, regular drills) Improve access to risk information, invest in hydro-meteorological services—for observation and forecasting—and link with early X X warning and evacuation systems, and collect more data on disaster consequences Enact risk-sensitive and enforceable land use regulation and building X norms Improve tenure to incentivize investments in housing quality and X resilience, and enforceability of building norms Invest more and better in infrastructure by leveraging private resources and using designs that account for future climate change and the X X X related uncertainty Health Increase R&D and eradication/control efforts toward health issues that X X X affect poor people and are expected to increase with climate change Invest in health infrastructure and access; train health workers X Implement or strengthen effective surveillance and monitoring X X systems to detect emerging health risks Increase health coverage to lower the share of expenses that are out X of pocket Support systems: financial sector, social protection, remittances, and governance Develop market insurance for the middle class to concentrate public X X resources on poor people Enact well-targeted and easily scalable social safety nets designed to maintain incentives for long-term adaptation investments and grant X portable benefits Manage the government’s formal liability using reserve funds, contingent finance (such as Cat-DDOs), and insurance products, along X X with developing and scaling-up tools to share risks internationally Facilitate flow of remittances and reduce cost burden on remitters X X Improve governance and give a role to poor people in the decision- X making process Note: Cat-DDO = Catastrophe Deferred Drawdown Option; R&D = research and development. O V E R V I E W   1 9 varieties were available (Hill and Mejia- FIGURE O.10  Drought vulnerability is reduced by Mantilla 2015). agricultural techniques that integrate trees and store carbon Poor people can become more resilient to (Reduction in average annual number of drought-affected shocks in agriculture thanks to trade and food people) reserves that can overcome local shortages in times of need, better access of poor farmers to 8 markets, and improved technologies and Number of people (millions) climate-smart production techniques. Access ­ 6 to functioning markets, however, depends on 4 better infrastructure and better institutions. In Ethiopia, the incidence of poverty decreased 2 by 6.7 percent following farmers’ access to all-weather roads (Dercon et al. 2009). In 0 Burkina Faso, maize price volatility is found No trees Low tree density High tree density to be greatest in remote markets (Ndiaye, Agroforestry Drought and heat tolerance Soil fertility management packages Maitre d’Hôtel, and Le Cotty 2015). Investments in transport infrastructure Source: Cervigni and Morris 2015. improve market integration, reduce price uncertainty for farmers, and improve food security. For ecosystem-based income, the main Second, strong institutions are needed to option is to reduce the nonclimate stresses ensure that land use plans are actually on ecosystems to make them better able to enforced, and even the highest-capacity cope with changes in environmental condi- countries struggle to reduce flood exposure. tions. Conservation and ecosystem-based Third, one needs to take into account the strategies are critical for making ecosystems reasons why people decide to live in risky more resilient and for protecting the places, namely a trade-off between safety resources on which many poor people in and access to jobs and services (Hallegatte rural areas depend. Healthy ecosystems are 2012a). In a new survey, poor households in generally quite resilient, so protecting them Mumbai say they would relocate to a safer and restoring degraded lands can increase place but only if they had access to cheap their ability to withstand climate-related dis- transport, health services, schools, and social turbances. Integrating trees in agricultural networks (Patankar, forthcoming). Thus, systems can also reduce vulnerability to land use planning can realistically function drought and increase the store of carbon only if accompanied by investments in trans- (figure O.10). port and other infrastructure to make it pos- Land use regulations and better and sible for people to settle in safe places while more infrastructure for natural hazards. maintaining access to the same (or compa- Land use regulations can ensure that new rable) jobs and services. development occurs in places that are safe, Poor people lack the type of protective or easy and cheap to protect using hard or infrastructure that is common in richer coun- soft infrastructure. But effective implementa- tries. For instance, poor households are often tion of such regulations remains challenging. exposed to recurrent floods due to the lack, or First, it requires appropriate data on risk and poor maintenance, of infrastructure (espe- hazard, which remains limited in low-income cially drainage systems)—even if these environments despite recent progress (includ- events do not attract media and policy ing the Global Facility for Disaster Reduc- maker ­ attention, they can represent a large tion and Recovery’s [GFDRR] Open Data burden on poor people. Solving these prob- for Resilience Initiative that makes risk data lems requires investing more and investing available for governments and the p ­ ublic). better. Around $1 trillion per year would be 2 0    SHOCK WAVES needed in developing countries to close the preparedness—can save many lives at a low infrastructure gap, with about $100 billion cost. When Cyclone Phailin made landfall for Africa alone. Closing this gap is difficult, near Gopalpur, India, in 2013, it killed fewer but it would go a long way toward reduc- than 100 people. While still a significant ing the vulnerability of poor people. loss, it is much smaller than the 10,000 Recommendations typically include leverag- deaths that a similar storm caused in 1999. ing private resources to make the most of More generally, early warning systems are available capital, which involves well-known very cost-effective investments, with each steps like improving the investment climate, dollar invested yielding more than $4 in developing local capital markets, and provid- avoided losses (Hallegatte 2012b). However, ing a pipeline of “bankable” projects (Fay over the past 15–20 years, the situation of et al. 2015). many hydrometeorological services in devel- But infrastructure investments will reduce oping countries has worsened (Rogers and the long-term vulnerability of the population Tsirkunov 2013). As a result, the ability to and contribute to long-term poverty reduc- monitor local climate change and increases tion only if they serve poor people. In particu- in natural risks has eroded, making develop- lar, investing where it is most cost-efficient ing countries less able to detect, anticipate, would risk concentrating resources on and adapt to climate change. wealthier populations at the expense of poor Better health infrastructure and universal communities (Tschakert, forthcoming). New health care. Poor people in low- and lower- infrastructure also needs to be designed to middle-income countries have limited access remain efficient in spite of changes in climate to health care, and face out-of-pocket expen- and environmental conditions. Innovative diture exceeding 50 percent of health methods for managing the uncertain risks of expenses—much higher than the less than climate change and multiple (and sometimes 15 percent that is common in rich countries conflicting) policy objectives can be applied to (figure O.11). But examples show that better meet these challenges (Kalra et al. 2014). health coverage is possible everywhere. In Several World Bank pilot projects using these Colombia, thanks to a multilevel govern- methods have been completed or are under ment scheme and cross-subsidization from way, including on water supply in Lima, flood contributory schemes, the poor are covered risk management in Ho Chi Minh City and against primary care and catastrophic event Colombo, hydropower investment in Nepal, costs—with coverage of the poorest quintile and adaptation of road networks in Peru and up to 47 percent in 1997 from only across Africa. 3–8 ­percent in 1993. In Rwanda, the govern- As discussed earlier, poor people lose a ment invested in universal health coverage larger fraction of their assets and income after the 1994 genocide, and today nearly because their dwelling is often their main 80 percent of its population is insured. asset and because they live in buildings with However, benefits from better access to low resistance to natural hazards. In addition care depend on the quality of care, and in to financial inclusion—which could help peo- most countries parallel efforts are required to ple save in less vulnerable ways—improving develop and improve health infrastructure. tenure security could incentivize investment in Climate change makes this need even more housing, including in risk reduction, to make important. Countries should have strong them more resistant. In Peru, the issuance of monitoring and surveillance systems able to property titles to over 1.2 million urban detect new health issues that will periodically dwellers encouraged households to invest arise in response to changing climate condi- more in their homes, thereby reducing their tions. They also need research and develop- vulnerability (Field 2007). ment on the diseases that affect poor people Early warning systems—combined with and that are expected to increase with climate observation systems and evacuation change. O V E R V I E W   2 1 FIGURE O.11  In poorer countries, half of all health expenditures are paid out of pocket, unlike in richer ones 100 Share of health care expenditure (% of total in 2011) 90 80 70 60 50 40 30 20 10 0 Low income Lower-middle Upper-middle High income Global income income Out-of-pocket expenses External resources Other private expenditure Private health insurance Other government expenditure Social security Source: Watts et al. 2015. Social safety nets and financial tools. A FIGURE O.12  Poorer households need different types of solutions growing body of evidence shows that insur- ance and social safety nets are efficient tools More International aid intense to support poor people when they are events Social insurance affected by natural disasters or environmen- and scaled-up tal and economic shocks. In Mexico, benefi- social safety nets Market insurance ciaries of Prospera, the national cash transfer Government program (previously known as insurance and contingent finance Oportunitades or Progresa), are less likely to Savings, credit, and respond to shocks by withdrawing their chil- Government scaled-up remittances reserve funds dren from the classroom (de Janvry et al. 2006; Fiszbein, Schady, and Ferreira 2009; Smaller Basic social protection, remittances, Gertler 2004). events and revenue diversification To ensure that the financial sector and social safety nets provide instruments relevant Poorer Richer to climate change, governments need to households households design a holistic risk management and climate change strategy, giving a voice to poor people and making their protection a priority. Such a strategy will necessarily include a range of needed. For relatively wealthier households, instruments, targeted to specific disasters or savings and market insurance can offer effi- social groups (figure O.12). cient protection for larger losses. But the Basic social protection and revenue diversi- poorest households have minimal savings, fication can help households at all income and high transaction costs make it difficult ­ levels cope with small and frequent shocks. to offer them private insurance. Instead, the But for larger shocks, additional tools are government needs to provide social safety 2 2    SHOCK WAVES nets that are well targeted and can be scaled instruments such as reserve funds, regional up rapidly after a shock. mechanisms, contingent finance or reinsur- A key challenge is to strike a balance ance products (like the World Bank’s between providing rapid support when Catastrophe Deferred Drawdown Option, or needed and precisely targeting those most in Cat-DDOs), or even international aid if local need. Case studies in Ethiopia and Malawi capacities are exhausted (Ghesquiere and suggest that the cost of a drought to house- Mahul 2010). In response to Cyclone Pam in holds can increase from zero to about $50 March 2015, the Pacific Catastrophe Risk per household if support is delayed by four Assessment and Financing Initiative months, and to about $1,300 if support is (PCRAFI), a regional mechanism, provided delayed by six to nine months (Clarke and Vanuatu with a rapid $1.9 million payment Hill 2013). This rapid increase, which is due that supported the immediate response. to irreversible impacts on children and dis- Social protection schemes also need to tress sales of assets (especially livestock), maintain incentives to invest in long-term helps explain why most postdisaster responses adaptation to economic and environmental have multiple stages. Typically, initial support changes. Poorly designed social safety nets is delivered quickly—even at the expense of can reduce the incentive for people to quickly targeting and ­accuracy—and larger recovery adapt and change occupation or activity when and reconstruction efforts are provided later the first effects of climate change appear with more emphasis on appropriate (Chambwera et al. 2014). This problem is not targeting. new and specific to climate change: efforts are Experience shows that countries at all already under way to ensure that social pro- income levels can implement social safety nets tection is a facilitator of—and not an obstacle to protect their population, even though the to—long-term change and adaptation, for appropriate instruments depend on local instance by facilitating migration (Brown, capacity. Preexisting social protection pro- Zelenska, and Mobarak 2013; Bryan, grams with large and flexible social registries Chowdhury, and Mobarak 2014) or making help provide prompt support to affected peo- benefits more portable if the recipient decides ple so that they do not have to resort to costly to move to capture better opportunities coping strategies. For instance, by using the (World Bank 2015b). preexisting conditional cash transfer system Combining rapid, inclusive, and climate- (the 4Ps), the government of the Philippines informed development with targeted interven- was able to quickly release a total of tions and stronger safety nets would largely P550.5 million (US$12.5 million) between reduce the short-term threat from climate November 2013 and February 2014 in emer- change—and, fortunately, developing countries gency unconditional cash transfers to 4Ps have a window of opportunity to go in that beneficiaries affected by Typhoon Yolanda direction before most of the climate change (Bowen, forthcoming). When droughts in impacts materialize. In parallel, the interna- Ethiopia caused food shortages and famine in tional community can do much to ­ support 2011, the Productive Safety Net Program them. This includes offering resources for cli- expanded its coverage from 6.5 million to 9.6 mate risk analysis and project preparation, million people in two months and increased and ensuring that financial instruments and the duration of benefits from six to nine resources are available for development and months per year (Johnson and Bowen, forth- poverty reduction investments—­ especially coming). These safety nets remain affordable when higher resilience implies higher upfront and reduce the need for costly humanitarian costs. The international community can also interventions. build resilience by strengthening international However, adaptive social protection sys- risk-sharing mechanisms and generalizing tems create an additional liability for govern- access to contingent finance in emergency ments, who may then need to turn to specific situations. O V E R V I E W   2 3 Emissions-reduction policies are internationally agreed climate targets (IPCC 2014). Thus, policies are needed now to make required to remove the long- development and climate change stabilization term threat from climate change, compatible: modern living standards will and need not threaten progress need to be supported in a more efficient and radically less carbon-intensive way, and resid- on poverty reduction ual emissions offset through natural carbon In the absence of mitigation policies, risks sinks like forests (Fay et al. 2015). for development and poverty eradication The first step is for all countries to enact will grow over time and only emissions comprehensive packages of emissions-­ reduction can limit long-term risks (IPCC reduction policies (IPCC 2014)—ranging 2014). While this report proposes options to from carbon pricing and innovation support reduce climate risks, it also points to the lim- to environmental performance standards, its of these options: land use planning faces information labels, financing facilities, and difficult political economy obstacles, finan- land use and urban planning (Fay et al. 2015; cial constraints make it tough to invest in NCE 2014; OECD 2015). Priority should go protection infrastructure, and the provision to implementing the policies and measures of health care in rural areas remains that are urgently needed to prevent irrevers- challenging. And, although social safety nets ­ ibility and lock-ins into carbon-intensive pat- and health insurance help households cope terns (such as those regarding deforestation, with shocks, they do not reduce the direct energy infrastructure, or urban transport). and immediate impact on well-being and These policy packages must be designed in assets and will become increasingly costly— a way that does not threaten the objective of even unaffordable—if shocks become more eradicating poverty by 2030. This can be frequent and intense (Carter and Janzen, done in three complementary ways: (i) build- forthcoming). There are clear limits to what ing on no-regret options and cobenefits; adaptation can achieve, and these limits will (ii) protecting the poor and vulnerable popu- be tested by climate change. lations against potential adverse consequences Moreover, some long-term risks could of emissions-reduction options; and (iii) in the prove catastrophic—such as those related to poorest countries, using support from the the response of ice sheets and ecosystems— international community to offset possible and remain impossible to quantify in terms of trade-offs between poverty reduction and cli- consequences or probability. Uncertainty is mate change mitigation. not a reason to delay climate change mitiga- All countries should embrace the mitiga- tion action. On the contrary, the need for cli- tion policies that generate short-term cobene- mate stabilization arises from both a fits that exceed costs—like lower air pollution risk-management approach that accounts for and higher energy efficiency. Recent studies threats created by long-term impacts and the have found that, in all regions, the benefits for fact that GHG emissions lock us into irrevers- health and agricultural yields from less pollu- ible warming. Indeed, these long-term risks tion alone could exceed the cost of mitigation, largely explain why the international commu- at least until 2030 (Shindell et al. 2012). For nity has committed to the goal of stabilizing example, a pathway leading to lowering CO2 climate change. concentrations would avoid 0.5 million pre- Maintaining global warming below 2°C, mature deaths annually in 2030, 1.3 million or even below 3°C, will require bringing emis- in 2050, and 2.2 million in 2100, compared sions down to zero by 2100, a goal recog- to a scenario with only the progress that can nized by the leaders of the major industrial be expected from the historically observed countries at the 2015 summit of the G7. And uptake of pollution-control technologies. there is a consensus that current develop- Many other cobenefits are likely to ment trends are incompatible with these occur in various sectors (World Bank 2014b). 2 4    SHOCK WAVES Better public transit would reduce congestion than implement win-win options, sometimes and traffic accidents, and greater energy effi- creating net costs and trade-offs. Fortunately, ciency would bode well for productivity. Yet governments can protect the poorest, using many countries, facing strong financing con- specific instruments or their existing social straints, tend to favor technologies with lower protection systems, possibly strengthened by upfront capital costs, at the expense of higher the resources raised by climate policies. For operation costs—in effect, favoring less instance, climate policies need to ensure that energy-­ e fficient technology and reducing they do not slow down the switch from tradi- overall productivity (World Bank 2012). tional biomass to modern cooking fuels, for Governments need to enact policies to example by subsidizing efficient cookstoves. actively promote the adoption of no-regret This matters greatly because traditional cook- options that reduce GHG emissions and ing fuels not only are unhealthy but also accelerate development. A recent World Bank worsen gender imbalances and affect educa- report reviews market and government fail- tional opportunities, given the time women ures that hamper the adoption of these no- and children often spend collecting wood and regret options—such as incorrect pricing, split other traditional fuels (WHO 2006). incentives, poor enforcement of existing regu- There are many options to make climate lations, lack of information, behavioral fail- policies pro-poor—such as introducing a car- ures, and limits to the financing capacity of bon or energy tax and recycling the revenues stakeholders—and offers solutions to over- through a universal cash transfer that would come them (Fay et al. 2015). The interna- benefit the poor. An analysis of 20 developing tional community can help developing countries shows that for each $100 of addi- countries by providing a combination of tech- tional energy tax collected and redistributed, nical assistance and better access to green the bottom quintile gains $13 while the rich- technologies (for instance to help them imple- est quintile loses $23, and overall the bottom ment performance standards for vehicles, 60 percent would benefit from the measure lighting, and appliances). It can also help (del Granado, Coady, and Gillingham 2012; them mobilize private capital to relax existing Fay et al. 2015). investment constraints and favor technologies Similarly, we can estimate how the with higher upfront costs but better efficiency, resources that could be raised by a carbon tax drawing on innovative financial instruments in one country (or an equivalent reform of or the resources from bilateral and multilat- energy subsidies) compare with current social eral development banks. assistance transfers. Based on current In addition, all countries need to avoid CO2 emissions and without any international negative impacts of mitigation policies on transfer, a $30/tCO2 (tons of CO2) domestic food security, since the resulting effects on carbon tax would raise resources amounting global food prices could have a detrimental to more than 1.5 percent of national GDP in impact on the poor. Promisingly, many land- half of the 87 countries where data are avail- based mitigation options also provide an able (figure O.13, panel a). And in 60 out of opportunity to strengthen the productivity of the 87 countries, a $30/tCO2 domestic carbon agriculture and ecosystems and to boost local tax would provide the resources to more than incomes. They can be implemented through double current levels of social assistance in payments for ecosystem services, which can the country (figure O.13, panel b). Even a low provide a source of income for the poor. An carbon tax at $10/tCO2 would make it possi- estimated 25–50 million low-income house- ble to significantly scale up social assistance holds could be benefitting from them by 2030 or other investments that benefit poor people (Milder, Scherr, and Bracer 2010). (like connections to sanitation and improved But to stay on a pathway compatible with drinking water or access to modern energy). the complete decarbonization of the economy More generally, the impacts of climate mit- before 2100, countries will have to do more igation policies on inequality can be corrected O V E R V I E W   2 5 FIGURE O.13  Using the revenue from a carbon tax could boost social assistance a. Revenue as a fraction of GDP b. Revenue compared to social assistance benefits 14 14 12 12 Carbon revenue (% GDP) Carbon revenue (% GDP) 10 10 8 8 6 6 4 4 2 2 0 0 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 0 2 4 6 8 10 12 14 GDP per capita (US$, PPP 2011) Average social assistance received (% GDP) Source: World Bank calculations using data from WDI and ASPIRE (World Bank 2015c, 2015d). Note: Panel a: Revenue of a $30/tCO2 carbon price expressed as a fraction of GDP. Each dot represents a country. Panel b: How this revenue compares to current social assistance benefits in each country. In 60 out of 87 countries for which data are available, a $30/tCO2 tax would provide the resources to more than double current social assistance transfers (dots above the diagonal line for panel b). Calculations assume unchanged energy consumption. PPP = p­ urchasing power parity; tCO2 = tons of carbon dioxide. using policies specifically designed to redis- climate mitigation could still worsen poverty, tribute income in the economy—such as using because the top quintiles are still in, or close income or consumption taxes to fund cash to, poverty. In these cases, international sup- transfers or social safety net programs port will be essential to offset potential trade- (Borenstein and Davis 2015; Gahvari and offs between poverty reduction and climate Mattos 2007; Lindert, Skoufias, and Shapiro change mitigation. 2006). A World Bank study based on house- This is especially the case for investments hold surveys reveals that countries with GDP that involve high immediate costs—and there- per capita above $4,000 (in purchasing power fore large trade-offs with other investments— parity) have sufficient internal resources to but are urgently needed to prevent irreversibility redistribute poverty away, and thus can pro- and lock-ins into carbon-intensive patterns. tect poor people against the possible negative The typical example is urban transit. While effects of climate mitigation (Ravallion 2010). transit-oriented development may require This is important because around 70 percent higher upfront costs than road-based low-­ of people in extreme poverty live in these density urbanization, there is now a unique countries that are able to protect them. window of opportunity to build efficient tran- But in very poor countries, it may be diffi- sit-oriented cities, because of high urbanization cult for economic, political, or institutional rates in many developing countries and the reasons to accomplish this. In particular, the extended lifetime of urban forms and transit same World Bank study shows that countries infrastructure. with a GDP per capita below $4,000 (in pur- chasing power parity) would find it nearly impossible to rely on internal resources for In conclusion redistribution. In these countries, even if most Bringing together the short-run (up to 2030) of the cost of climate mitigation is paid for by and long-run views, this report emphasizes the wealthier quintiles of the population, how climate change could set back poverty 2 6    SHOCK WAVES eradication efforts—including the risk that Developing Countries.” Background paper for unabated climate change creates for the inter- the World Development Report, World Bank, nationally agreed objective of eradicating Washington, DC. extreme poverty. In parallel, it demonstrates Bryan, G., S. Chowdhury, and A. M. Mobarak. 2014. “Underinvestment in a Profitable that the future is not set in stone. We have a Technology: The Case of Seasonal Migration window of opportunity to achieve our pov- in Bangladesh.” Econometrica 82: 1671–1748. erty objectives in spite of ­climate change by Carter, M. R., and S. A. Janzen. Forthcoming. pursuing both (i) rapid, inclusive, and cli- “Social Protection in the Face of Climate mate-informed development, combined with Change: Targeting Principles and Financing targeted adaptation interventions, to cope Mechanisms.” Background paper prepared for with the short-term impacts of climate this report. change and (ii) immediate pro-poor mitiga- Cervigni, R., and M. Morris. 2015. “Enhancing tion policies to limit long-term impacts and Resilience in African Drylands: Toward a create an environment that allows for global Shared Development Agenda.” World Bank, prosperity and the sustainable eradication Washington, DC. Chambwera, M., G. Heal, C. Dubeux, of poverty. S. Hallegatte, L. Leclerc, A. Markandya, B. A. McCarl, R. Mechler, and J. E. Neumann. 2014. “Economics of Adaptation.” In Climate References Change 2014: Impacts, Adaptation, and Ahmed, S. A., N. S. Diffenbaugh, and T. W. Hertel. Vulnerability. Part A: Global and Sectoral 2009. “Climate Volatility Deepens Poverty Aspects, edited by C. B. Field, V. R. Barros, Vulnerability in Developing Countries.” D. J. Dokken, K. J. Mach, M. D. Mastrandrea, Environ. Res. Lett. 4: 034004. T. E. Bilir, M. Chatterjee, K. L. Ebi, Allen, F., A. Demirgüç-Kunt, L. F. Klapper, and Y. O. Estrada, R. C. Genova, B. Girma, M. S. Martinez Peria. 2012. “The Foundations E. S. Kissel, A. N. Levy, S. MacCracken, of Financial Inclusion: Understanding P. R. Mastrandrea, and L. L. White, 945–77. Ownership and Use of Formal Accounts.” Contribution of Working Group II to the Fifth Policy Research Working Paper 6290, World Assessment Report of the Intergovernmental Bank, Washington, DC. Panel of Climate Change. Cambridge, U.K. and Beddington, J. 2010. “Food Security: New York: Cambridge University Press. Contributions from Science to a New and Clarke, D. J., and R. V. Hill. 2013. “Cost-Benefit Greener Revolution.” Philos. Trans. R. Soc. B Analysis of the African Risk Capacity Facility.” Biol. Sci. 365, 61–71. IFPRI Discussion Paper 01292, International Biewald, A., H. Lotze-Campen, I. Otto, Food Policy Research Institute. N. Brinckmann, B. Bodirsky, I. Weindl, A. Popp, de Janvry, A., F. Finan, E. Sadoulet, and R. Vakis. and H. J. Schellnhuber. Forthcoming. “The 2006. “Can Conditional Cash Transfer Impact of Climate Change on Costs of Food Programs Serve as Safety Nets in Keeping and People Exposed to Hunger at Subnational Children at School and from Working When Scale.” PIK Report 128, Potsdam Institute for Exposed to Shocks?” J. Dev. Econ. 79: 349–73. Climate Impact Research. Background paper del Granado, F. J. A., D. Coady, and R. Gillingham. prepared for this report. 2012. “The Unequal Benefits of Fuel Subsidies: Borenstein, S., and L. W. Davis. 2015. “The A Review of Evidence for Developing Distributional Effects of U.S. Clean Energy Tax Countries.” World Dev. 40: 2234–48. Credits.” NBER Working Paper No. 21437, del Ninno, C., P. A. Dorosh, and L. C. Smith. National Bureau of Economic Research, 2003. “Public Policy, Markets and Household Cambridge, MA. Coping Strategies in Bangladesh: Avoiding Bowen, T. Forthcoming. “Social Protection and a Food Security Crisis Following the 1998 Disaster Risk Management in the Philippines.” Floods.” World Dev., Economic Crises, Natural Background paper prepared for this report. Disasters, and Poverty 31: 1221–38. Brown, J. K., T. V. Zelenska, and M. A. Mobarak. Dercon, S. 2004. “Growth and Shocks: Evidence 2013. “Barriers to Adoption of Products and from Rural Ethiopia.” J. Dev. Econ. 74: Technologies That Aid Risk Management in 309–29. O V E R V I E W   2 7 Dercon, S., D. O. Gilligan, J. Hoddinott, and Causes of Death, 2030s and 2050s. World T. Woldehanna. 2009. “The Impact of Health Organization. Agricultural Extension and Roads on Poverty Hallegatte, S. 2012a. “An Exploration of the Link and Consumption Growth in Fifteen Ethiopian between Development, Economic Growth, and Villages.” Am. J. Agric. Econ. 91: 1007–21. Natural Risk.” Policy Research Working Paper Deryugina, T., and S. M. Hsiang. 2014. “Does the 6216, World Bank, Washington, DC. Environment Still Matter? Daily Temperature Hallegatte, S. 2012b. “A Cost Effective Solution and Income in the United States.” NBER to Reduce Disaster Losses in Developing Working Paper 20750, National Bureau of Countries: Hydro-Meteorological Services, Economic Research, Cambridge, MA. Early Warning, and Evacuation.” Policy Fang, Y., D. L. Mauzerall, J. Liu, A. M. Fiore, Research Working Paper 6058, World Bank, and L. W. Horowitz. 2013. “Impacts of Washington, DC. 21st Century Climate Change on Global Air Hallegatte, S., C. Green, R. J. Nicholls, and Pollution–Related Premature Mortality.” Clim. J. Corfee-Morlot. 2013. “Future Flood Losses Change 121: 239–53. in Major Coastal Cities.” Nat. Clim. Change 3: FAO (Food and Agriculture Organization of 802–06. the United Nations). 2013. Climate-Smart Havlík, P., H. Valin, M. Gusti, E. Schmid, Agriculture Sourcebook. Rome, Italy: FAO. D. Leclère, N. Forsell, M. Herrero, Fay, M., S. Hallegatte, A. Vogt-Schilb, J. Rozenberg, N. Khabarov, A. Mosnier, M. Cantele, and U. Narloch, and T. Kerr. 2015. Decarbonizing M. Obersteiner. Forthcoming. “Climate Change Development: Three Steps to a Zero-Carbon Impacts and Mitigation in the Developing Future. Washington, DC: World Bank. World: An Integrated Assessment of Agriculture Field, E. 2007. “Entitled to Work: Urban Property and Forestry Sectors.” Background paper Rights and Labor Supply in Peru.” Q. J. Econ. prepared for this report. 122: 1561–1602. Heal, G., and J. Park. 2013. “Feeling the Heat: F i s c h e r, E . M . , a n d R . K n u t t i . 2 0 1 5 . Temperature, Physiology and the Wealth of “Anthropogenic Contribution to Global Nations.” NBER Working Paper No. 19725, Occurrence of Heavy-Precipitation and High- National Bureau of Economic Research, Temperature Extremes.” Nature Clim. Change Cambridge, MA. 5: 560–64. Hertel, T. W., M. B. Burke, and D. B. Lobell. 2010. Fiszbein, A., N. R. Schady, and F. H. Ferreira. “The Poverty Implications of Climate-Induced 2009. Conditional Cash Transfers: Reducing Crop Yield Changes by 2030.” Glob. Environ. Present and Future Poverty. Washington, DC: Change, 20th Anniversary Special Issue 20: World Bank. 577–85. Gahvari, F., and E. Mattos. 2007. “Conditional Hill, R., and C. Mejia-Mantilla. 2015. “Welfare Cash Transfers, Public Provision of Private and Shocks in Uganda.” Background Paper Goods, and Income Redistribution.” Am. Econ. prepared for the Uganda Poverty Assessment. Rev. 97: 491–502. World Bank, Washington, DC. Gertler, P. 2004. “Do Conditional Cash IPCC (Intergovernmental Panel on Climate Transfers Improve Child Health? Evidence Change). 2014. Climate Change 2014: Synthesis from PROGRESA’s Control Randomized Report. Contribution of Working Groups I, II, Experiment.” Am. Econ. Rev. 94: 336–41. and III to the Fifth Assessment Report of the Ghesquiere, F., and O. Mahul. 2010. Financial Intergovernmental Panel on Climate Change Protection of the State against Natural [Core Writing Team, R. K. Pachauri and L. A. Disasters: A Primer. Washington, DC: World Meyer (eds.)]. Geneva, Switzerland: IPCC. Bank. Ivanic, M., W. Martin, and H. Zaman. 2012. Grassini, P., K. M. Eskridge, and K. G. Cassman. “Estimating the Short-Run Poverty Impacts of 2013. “Distinguishing between Yield Advances the 2010–11 Surge in Food Prices.” World Dev. and Yield Plateaus in Historical Crop 40: 2302–17. Production Trends.” Nat. Commun. 4. Jacoby, H. G., M. Rabassa, and E. Skoufias. Hales, S., S. Kovats, S. Lloyd, and D. Campbell- 2014. “Distributional Implications of Climate Lendrum. 2014. Quantitative Risk Assessment Change in Rural India: A General Equilibrium of the Effects of Climate Change on Selected Approach.” Am. J. Agric. Econ. aau084. 2 8    SHOCK WAVES Johnson, K., and T. Bowen. Forthcoming. Incomes.” Background paper prepared for this “Ethiopia’s Food Security Program Social report. Protection, Disaster Risk Management and OECD (Organisation for Economic Co-operation Climate Change Adaptation.” Background and Development). 2015. Aligning Policies for paper prepared for this report. a Low-Carbon Economy. Paris, France: OECD. Kalra, N., S. Hallegatte, R. Lempert, C. Brown, Olsson, L., M. Opondo, P. Tschakert, A. A. Fozzard, S. Gill, and A. Shah. 2014. Agrawal, S. H. Eriksen, S. Ma, L. N. Perch, “Agreeing on Robust Decisions: New Processes and S. A. Zakieldeen. 2014. “Livelihoods for Decision Making under Deep Uncertainty.” and Poverty.” In Climate Change 2014: Policy Research Working Paper 6906, World Impacts, Adaptation, and Vulnerability. Part Bank, Washington, DC. A: Global and Sectoral Aspects, edited by Krishna, A. 2006. “Pathways Out Of and Into C. B. Field, V. R. Barros, D. J. Dokken, K. J. Poverty in 36 Villages of Andhra Pradesh, Mach, M. D. Mastrandrea, T. E. Bilir, M. India.” World Dev. , Part Special Issue: Chatterjee, K. L. Ebi, Y. O. Estrada, R. C. 324–404. Corruption and Development: Genova, B. Girma, E. S. Kissel, A. N. Levy, Analysis and Measurement 34: 271–88. S. MacCracken, P. R. Mastrandrea, and L. L. Lindert, K., E. Skoufias, and J. Shapiro. 2006. White, 793–832. Contribution of Working Redistributing Income to the Poor and the Group II to the Fifth Assessment Report of the Rich: Public Transfers in Latin America and the Intergovernmental Panel of Climate Change. Caribbean. Washington, DC: World Bank. Cambridge, U.K. and New York: Cambridge Lloyd, S. J., R. S. Kovats, and Z. Chalabi. University Press. 2011. “Climate Change, Crop Yields, and Pardey, P. G., J. M. Alston, and C. Chan-Kang. Undernutrition: Development of a Model to 2013. “Public Agricultural R&D over the Quantify the Impact of Climate Scenarios Past Half Century: An Emerging New World on Child Undernutrition.” Environ. Health Order.” Agric. Econ. 44: 103–13. Perspect. 119: 1817–23. Park, J., S. Hallegatte, M. Bangalore, and E. Lotze-Campen, H., M. von Lampe, P. Kyle, Sandhoefner. Forthcoming. “Households and S. Fujimori, P. Havlik, H. van Meijl, T. Heat Stress: Estimating the Distributional Hasegawa, A. Popp, C. Schmitz, A. Tabeau, Consequences of Climate Change.” H. Valin, D. Willenbockel, and M. Wise. 2014. Background paper prepared for this report. “Impacts of Increased Bioenergy Demand on Patankar, A. Forthcoming. “The Exposure, Global Food Markets: An AgMIP Economic Vulnerability, and Ability to Respond of Poor Model Intercomparison.” Agric. Econ. 45: Households to Recurrent Floods in Mumbai.” 103–16. Background paper prepared for this report. Milder, J. C., S. J. Scherr, and C. Bracer. 2010. Porter, J. R., L. Xie, A. J. Challinor, K. Cochrane, “Trends and Future Potential of Payment for S. M. Howden, M. M. Iqbal, D. B. Lobell, and Ecosystem Services to Alleviate Rural Poverty M. I. Travasso. 2014. “Food Security and Food in Developing Countries.” Ecol. Soc. 15: 4. Production Systems.” In Climate Change 2014: Myers, S. S., A. Zanobetti, I. Kloog, P. Huybers, Impacts, Adaptation, and Vulnerability. Part A. D. Leakey, A. J. Bloom, E. Carlisle, L. H. A: Global and Sectoral Aspects, edited by C. B. Dietterich, G. Fitzgerald, and T. Hasegawa. Field, V. R. Barros, D. J. Dokken, K. J. Mach, 2014. “Increasing CO 2 Threatens Human M. D. Mastrandrea, T. E. Bilir, M. Chatterjee, Nutrition.” Nature 510: 139–42. K. L. Ebi, Y. O. Estrada, R. C. Genova, B. NCE (New Climate Economy). 2014. Better Girma, E. S. Kissel, A. N. Levy, S. MacCracken, Growth Better Climate: The New Climate P. R. Mastrandrea, and L. L. White, 485–533. Economy Report . Washington, DC: New Contribution of Working Group II to the Fifth Climate Economy. Assessment Report of the Intergovernmental Ndiaye, M., E. Maitre d’Hôtel, and T. Le Cotty. Panel of Climate Change. Cambridge, U.K. and 2015. “Maize Price Volatility: Does Market New York: Cambridge University Press. Remoteness Matter?” Policy Research Working Ravallion, M. 2010. “Do Poorer Countries Have Paper 7202, World Bank, Washington, DC. Less Capacity for Redistribution?” J. Glob. Noack, F., S. Wunder, A. Angelsen, and J. Börner. Dev. 1. Forthcoming. “Responses to Weather and Rodriguez-Oreggia, E., A. De La Fuente, Climate: A Cross-Section Analysis of Rural R. De La Torre, and H. A. Moreno. 2013. O V E R V I E W   2 9 “Natural Disasters, Human Development and A. Hunter, X. Jiang, M. Li, I. Kelman, L. Poverty at the Municipal Level in Mexico.” Liang, M. Lott, R. Lowe, Y. Luo, G. Mace, M. J. Dev. Stud. 49: 442–55. Maslin, M. Nilsson, T. Oreszczyn, S. Pye, T. Rogers, D. P., and V. V. Tsirkunov. 2013. Weather Quinn, M. Svensdotter, S. Venevsky, K. Warner, and Climate Resilience: Effective Preparedness B. Xu, J. Yang, Y. Yin, C. Yu, Q. Zhang, P. through National Meteorological and Gong, H. Montgomery, and A. Costello. 2015. Hydrological Services. Washington, DC: World “Health and Climate Change: Policy Responses Bank. to Protect Public Health.” The Lancet. Rozenberg, J., and S. Hallegatte. Forthcoming. WHO (World Health Organization). 2003. “The Impacts of Climate Change on Poverty Summary Booklet: Climate Change and in 2030, and the Potential from Rapid, Human Health – Risks and Responses. Geneva, Inclusive and Climate-Informed Development.” Switzerland: WHO. Background paper prepared for this report. ———. 2006. Fuel for Life: Household Energy Seppänen, O., W. J. Fisk, and Q. Lei. 2006. Effect and Health. Geneva, Switzerland: WHO. of Temperature on Task Performance in Office ———. 2013. Universal Health Coverage: Report Environment . Lawrency Berkeley National by the Secretariat. Geneva, Switzerland: WHO. Laboratory. Winsemius, H., B. Jongman, T. Veldkamp, Shindell, D., J. C. I. Kuylenstierna, E. Vignati, R. S. Hallegatte, M. Bangalore, and P. J. Ward. van Dingenen, M. Amann, Z. Klimont, S. C. Forthcoming. “Disaster Risk, Climate Change, Anenberg, N. Muller, G. Janssens-Maenhout, F. and Poverty: Assessing the Global Exposure Raes, J. Schwartz, G. Faluvegi, L. Pozzoli, K. of Poor People to Floods and Droughts.” Kupiainen, L. Höglund-Isaksson, L. Emberson, Background paper prepared for this report. D. Streets, V. Ramanathan, K. Hicks, N. T. K. World Bank. 2012. Inclusive Green Growth: Oanh, G. Milly, M. Williams, V. Demkine, and The Pathway to Sustainable Development . D. Fowler. 2012. “Simultaneously Mitigating Washington, DC: World Bank. Near-Term Climate Change and Improving ———. 2013. Risk and Opportunity: Managing Human Health and Food Security.” Science Risk for Development, World Development 335: 183–89. Report 2014. Washington, DC: World Bank. Small, J., S. J. Goetz, and S. I. Hay. 2003. “Climatic ———. 2014a. Turn Down the Heat: Climate Suitability for Malaria Transmission in Africa, Extremes, Regional Impacts, and the Case for 1911–1995.” Proc. Natl. Acad. Sci. 100: 15, Resilience. Washington, DC: World Bank. 341–45. ———. 2014b. Climate-Smart Development: Tilman, D., C. Balzer, J. Hill, and B. L. Befort. Adding Up the Benefits of Actions That Help 2011. “Global Food Demand and the Build Prosperity, End Poverty and Combat Sustainable Intensification of Agriculture.” Climate Change . Washington, DC: World Proc. Natl. Acad. Sci. 108: 20260–64. Bank. Tschakert, P. Forthcoming. “The Ability of the ———. 2015a. Climate Data (No. Climate 4 Poor to Cope.” Background paper prepared for Development). Washington, DC: World Bank. this report. ———. 2015b. The State of Social Safety Nets. Watts, N., W. N. Adger, P. Agnolucci, J. Blackstock, Washington, DC: World Bank. P. Byass, W. Cai, S. Chaytor, T. Colbourn, ———. 2015c. World Development Indicators. M. Collins, A. Cooper, P. M. Cox, J. Depledge, Washington, DC: World Bank. P. Drummond, P. Ekins, V. Galaz, D. Grace, H. ———. 2015d. ASPIRE Database. Washington, Graham, M. Grubb, A. Haines, I. Hamilton, DC: World Bank. From Climate Change to Poverty and Back: A Framework 1 Main Messages • Climate-sensitive events are already a critical • The net impact of climate change on pov- obstacle for people trying to escape poverty erty and well-being will be mediated by and those who are vulnerable to falling back socioeconomic trends (like demography, into poverty. growth, and inequality) and non-climate • There are three major channels through which change policies (like skills development and climate-sensitive events already affect the a ­ bility redistribution). of poor people to escape poverty: (i) agricul- • Over the medium to long term, the impact of tural production, ecosystems, and food security; climate change on poverty will also depend (ii) natural disasters; and (iii) health. All are likely on greenhouse gas emissions and thus on to be significantly affected by climate change. emissions-reduction policies. Introduction At the same time, the post-2015 Sustainable Development Agenda—which Despite substantial progress in reducing pov- builds on the 2015 Millennium Development erty rates, around 700 million people still Goals—has reaffirmed the goal of ending live in extreme poverty (World Bank extreme poverty or, as framed by the World 2015a).1 In addition, hundreds of millions Bank, bringing the number of people living hover close above the poverty line, vulnerable on less than $1.90 per day to under 3 percent to shocks that could send them into poverty, by 2030. Moreover, the agreed goal is not or suffer from other dimensions of poverty— simply to eliminate poverty by 2030, but to such as exclusion, powerlessness, and poor eliminate it for good and ensure gains are not health—even if their consumption is above reversed after 2030. the poverty threshold (Bourguignon and How does climate change fit into this pic- Chakravarty 2003; Ferreira and Lugo 2013). ture? We know that climate change affects the 31 3 2    SHOCK WAVES population, including the poorest, through ending extreme poverty. It also provides guid- changes in environmental conditions and the ance on how to design climate policies such frequency and intensity of extreme weather that they contribute to poverty reduction as events (O’Brien and Leichenko 2000; Olsson well as on how to design poverty-reduction et al. 2014). While the magnitude of climate policies such that they contribute to climate change is likely to be relatively limited change mitigation and resilience building. between now and 2030, compared with what And it contributes to a series of reports that can be expected over the long term, its explore the complex relationship between impacts may still be important in certain loca- development and climate change (box 1.1). tions (like semi-arid areas where precipitation The report aims to answer the following will decrease in response to climate change) questions: or for certain people (like those already living close to subsistence levels). Moreover, antici- • Will the changing climate be a threat to pated warming by 2030 cannot be reduced ending poverty, at what time horizon, and much, as any action taken to reduce green- under which conditions? Would climate house gas emissions takes several decades to change make it impossible to end poverty significantly affect climate change. only in the most pessimistic scenarios, or As we approach the end of this century, in would it be a threat even in more moder- the absence of ambitious climate-mitigation ate or optimistic cases? Will this impact policies, global warming may exceed 4°C, be concentrated in highly vulnerable loca- and impacts are likely to become “severe, tions, or more broadly distributed glob- widespread and irreversible,” threatening ally? Is it a long-term trend, or is it also poverty reduction and development (IPCC relevant over the shorter term? 2014a; World Bank 2015b). But the good • What emissions-reduction policies and news is that we still have a window of oppor- adaptation actions c an reduce this tunity to adopt and implement emissions- threat? In particular, how can we balance reduction policies that could slow down and short-term actions (like reducing vulner- stabilize long-term climate change, and in the ability to floods and droughts) with long- process, usher in a more prosperous world. term goals such as stabilizing the climate? One more wrinkle in this story is that the How should adaptation actions navi- climate change–poverty link is actually two- gate across different options to reduce way. Expected progress in poverty reduction ­ v ulnerability—from poverty reduction and access to basic services has the potential to and increased income to targeted adapta- reduce vulnerability to climate change and tion policies and social protection? How reduce its impacts (Hallegatte, Przyluski, and can adaptation policies prevent lock-ins Vogt-Schilb 2011; Wilbanks and Ebi 2013). into activities and locations that will But the future of poverty also matters for cli- become increasingly unable to sustain ris- mate change. Development and economic ing standards of living? growth, critical elements for reducing poverty, • How should poverty concerns be fac- directly affect energy consumption and access tored into mitigation and adaptation to technologies—which, in turn, affect green- ­p olicies? Are there trade-offs between house gas emissions and the long-term pace mitigation and poverty goals, and if so and magnitude of climate change. how can these be managed? Can com- Against this backdrop—and the limited plementary policies cancel out potential research done so far on this two-way negative effects on the poor? How can relationship—this report sets out to explore ­ we design adaptation policies so that they the potential impact of climate change and benefit poor people and contribute to climate policies on poverty reduction, and poverty reduction? investigate whether climate change can repre- • How should the existence of climate change sent a significant obstacle to the objective of modify poverty reduction ­ s trategies? F rom C limate C hange to P overt y and B ack : A F ramework   33 BOX 1.1  Multiple reports explore the complex relationship between development and climate change This report is just one of many produced by the and Disaster Risk into Development  explores World Bank Group in the past two years on climate solutions for both protecting lives and livelihoods change, mitigation, and development. The scope of and decreasing damages from natural disasters. this flagship report is on climate change and poverty, C limate-Smart Development examines the many ­ and it does not pretend to cover all the dimensions benefits of climate-smart development, including of the complex interplay between climate change and those for health, agricultural yields, and employ- development. In particular, a number of other reports ment. And Enhancing the Climate Resilience of on mitigation, adaptation, and development, along Africa’s Infrastructure assesses the impact of climate with private-sector involvement, provide analyses change on infrastructure development in Africa, that complement the findings of this report. especially in the power and water sectors. On mitigation, the recent report Decarbonizing On the private sector, the report on  Mobiliz- Development: 3 Steps to a Zero-Carbon Future pro- ing Public and Private Funds for Inclusive Green vides policy advice to help put countries on a path Growth Investment in Developing Countries pro- toward decarbonizing development. On Thin Ice: vides guidance on how to scale-up existing innova- How Cutting Pollution Can Slow Warming and tive mechanisms to mobilize private capital for green Save Lives reviews the scientific impacts of meth- growth. And the report on an Enabling Environ- ane and black carbon and strategies to reduce emis- ment for Private Sector Adaptation outlines how sions. And the annual State and Trends of Carbon to create an environment for private sector partici- ­Pricing report reviews the ever-changing landscape pation in promoting climate-resilient development of country-level climate-mitigation policies. paths. On adaptation and development , this report Other organizations have also recently produced relies on the World Development Report 2014, major reports that contribute to the discussion on which focuses on managing risk (including natu- development and climate, including the OECD’s ral risk, and opportunities for poverty reduction), Aligning Policies for a Low-Carbon Economy, the and the three volumes of the Turn Down the Heat Global Commission on the Economy and Climate’s reports, which review the science of global climate Better Growth, Better Climate, the IMF’s Energy change and provide detailed assessments on regional Subsidy Reform: Lessons and Implications , and impacts. Building Resilience: Integrating Climate UNEP’s Adaptation Gap Report. ­ ecognizing that poor people are more R poverty, including the role of socioeconomic vulnerable, does climate change influ- trends (like growth and demography) and ence what should be the priority for pov- patterns (like inequality and governance) in erty reduction? Is there a risk of a “global mediating impacts (Olsson et al. 2014). poverty trap” if people who are still in It explores three main channels through extreme poverty are unable to adapt to which climate has always affected poverty: climate change and thus escape poverty? (i) agricultural production, ecosystems, and And how can we ensure that the goal of food security; (ii) natural disasters; and eliminating extreme poverty by 2030 is (iii) health. It fills in some gaps in the litera- not achieved in a way that creates greater ture and uses data and modeling analyses to vulnerability post-2030? provide quantifications (or at least, orders of magnitude) for some of the qualitative state- This report builds on the existing ments that can be found in the literature. We l ­iterature—especially the review by the combine, for example, large datasets on Intergovernmental Panel on Climate Change household characteristics and location with (IPCC)—on the links between climate and recent global flood and drought modeling to 3 4    SHOCK WAVES examine whether poor people are more Dollar and Kraay 2002). This relationship exposed and vulnerable to natural disasters could of course change in the future if gov- and how this is likely to change with climate ernments implement substantial redistribu- change. We also explore whether lower agri- policies (Ferreira and Ravallion 2009; tive ­ cultural yields will occur in places vulnerable Robalino and Warr 2006). But substantial to hunger, and the role that ecosystems play redistribution is not easy politically, and in reducing both poverty and risk in poor the poorest countries simply lack the communities. resources to eradicate poverty through We then present the results of a new mod- redistribution. Economic growth will thus eling exercise that builds on the collected be needed to bring people out of poverty knowledge in each main channel—­ agriculture, (Ravallion 2010). natural disasters, and health—to explore the At this point, we know that climate potential impacts of climate change on pov- affects economic growth, based on observa- erty in 92 developing countries by 2030, tions of past evolutions. Reduced rainfall in investigating how these impacts are different the 20th century partly explains S ­ ub-Saharan in more or less optimistic scenarios of socio- Africa’s slow growth (Barrios, Bertinelli, and economic development. Finally, we explore Strobl 2010; Brown et al. 2010). And high policy options with an eye on the long term, temperatures in the second half of the 20th and discuss how the stabilization of climate century may have slowed down growth in change can be made compatible with poverty poor countries in both the agricultural and eradication. the industrial sectors (Dell, Jones, and Olken Our assessments—and quantification 2012). One study also finds that every 1°C exercises—cannot be considered a compre- ­ warming reduces income by 1.2 percent in hensive estimate of the impact of climate the short run, and by 0.5 percent in the long change on poverty. However, they are suffi- run (Dell, Jones, and Olken 2009). Other cient to demonstrate that climate change studies have found even larger impacts— poses a significant obstacle to eradicating including a 3.8 percent drop in income in extreme poverty in the days and decades the long run for every 1°C warming ahead. They also stress that we can act and (Horowitz 2009). reach poverty-eradication goals—in spite of But what should countries expect from climate change—by combining rapid, inclu- global warming in the future? Here, the evi- sive, and climate-informed development and dence is inconclusive, with estimates based on targeted interventions (to cope with short- very simple, partial models that vary widely term impacts) with pro-poor mitigation poli- (Pindyck 2013; Stern 2013). Most studies find cies (to avoid long-term impacts). a relatively limited impact on GDP. The latest Synthesis Report of the IPCC states that “incomplete estimates of global annual eco- Climate change is an obstacle for nomic losses for additional temperature increases of [about] 2.5°C above pre-­ people to escape poverty industrial levels are between 0.2 and 2.0% of Given that economic growth plays a critical income” (figure 1.1), but adds that these esti- role in reducing poverty, a key concern for mates are likely to underestimate actual poverty eradication is the impact of climate impacts. In particular, existing estimates have change on growth. Indeed, in the past, the focused on limited warming (below 3°C) and income of the bottom 20 percent of the do not include many sources of impacts that population has increased much more as a may prove consequential—such as the impact result of increases in the average income of ecosystems on economic activity, or the than from increases in the share of the risk from surprises and tipping points. Studies income that goes to the bottom quintiles including these elements, even in a simplistic (Dollar, Kleineberg, and Kraay 2013; manner, have found much larger potential F rom C limate C hange to P overt y and B ack : A F ramework   35 impacts (Stern 2006; Weitzman 2014). As a FIGURE 1.1  The bigger the climate change, the bigger the result, the confidence in these estimates is lim- total impact (Estimates of the total impact of climate change increases with the magnitude of ited, especially for warming that exceeds 2°C climate change) (IPCC 2014a). Based on these macroeconomic estimates 3 of future aggregate impacts of climate change, Impact on welfare (equivalent income change, %) the impacts on poverty via the GDP channel are small—possibly less than 1 percent by 0 2050 (Skoufias, Rabassa, and Olivieri 2011). But of course, climate change does not –3 only affect poor people via economic aggre- gates, it also can affect them directly. In fact, its direct impacts will likely be more signifi- –6 cant than the growth-mediated one, and may not significantly affect aggregate GDP, since –9 poor people represent a very small share of global income. This is because climate change can affect household consumption (and thus –12 consumption-related poverty) through four 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 channels almost independently of aggregate growth: Increase in global mean temperature (°C) Source: IPCC 2014b. • Prices. Consumption in real terms is Note: The magnitude of climate change is proxied by the increase in the global mean surface air driven by price levels and relative prices, temperature (x-axis). Each dot represents one study. which can be affected by multiple trends, shocks, and policies. Spikes in prices of basic goods can have large impacts on • Productivity. Households can increase poverty numbers (Ivanic, Martin, and including labor) the return on their assets (­ Zaman 2012), and climate change may by being more efficient and improving increase the level and volatility of food production processes, although returns prices and thus hurt poor people who are often limited by economic inefficien- usually spend a large share of their income cies (such as corruption, market failures, on food (chapter 2). and inappropriate regulations). Returns • Assets. Households escape poverty by are also affected by changes in the price accumulating assets (Moser 2008). For of what households produce. Climate instance, they acquire education or infor- change can decrease these returns through mation, improve their health, and invest lower agricultural yields (­ chapter 2) and in productive assets such as livestock or labor productivity (­chapter 4), or increase manufacturing equipment. Assets usually them through higher agricultural prices include the financial, physical, human, (chapter 2). social, and natural capital that house- • Opportunities. Households can also holds own. They also include public increase their income by expanding their goods, infrastructure, and institutions range of activities or migrating, typi- that households have access to. Climate cally to cities (Bryan, Chowdhury, and change can affect asset accumulation and Mobarak 2014). This is possible thanks poverty reduction by destroying assets to new opportunities in new sectors and during disasters or affecting people’s activities. But those opportunities are incentive and ability to invest in new often limited by exclusion (for exam- assets (chapter 3), or by affecting people’s ple, based on gender or ethnicity) and health (chapter 4). constraints on mobility. And climate 3 6    SHOCK WAVES change can worsen the situation through Climatic events already affect the more conflicts, increased competition dynamics of poverty, which will be for resources, higher risk aversion, or worsened by climate change poorly designed adaptation policies (such Poverty reduction is not a one-way transi- as greater reliance on nonportable safety tion out of poverty. Recent data show that nets; chapter 5). about a third of urban Indonesian residents Of course, these channels interact closely. moved in or out of poverty in less than a Lower agricultural productivity could lead to decade (Gentilini, forthcoming)—and such higher food prices, which can make it impos- large flows are consistently found in surveys sible for some households to continue saving around the world (Baulch 2011; Beegle, De and build their asset stock. Capturing new Weerdt, and Dercon 2006; Dang, Lanjouw, opportunities requires investments and and Swinkels 2014; Krishna 2006; Lanjouw, sometimes migration, activities that are pos- ­ McKenzie, and Luoto 2011). sible only with an appropriate asset base. Take the case of 36 communities in the And price changes will affect consumption state of Andhra Pradesh in India, over a (for net buyers) and income (for net sellers) 25-year period. As figure 1.2 shows, simultaneously. 14 ­percent of households in the 36 communi- If poor people are disproportionally ties escaped poverty every year, while 12 per- affected by climate change impacts through cent of nonpoor households became these channels, the impact of climate change poor—resulting in a 2 percent net annual on poverty will be larger than what is sug- decrease in overall poverty (Krishna 2007).2 gested by the macro impact on GDP—and The fact that the net flow out of poverty is climate change can lead to a decoupling of much smaller than the gross flows in and out economic growth and poverty reduction. To of poverty means that a relatively small explore this question, the report examines the change in the gross flows can have a signifi- impacts of climate change on poor and vul- cant effect on the net flows and thus overall nerable people and draws implications for poverty dynamics. In this example, if the flow future poverty reduction. We start with cur- into poverty increased from 12 to 13 percent rent poverty dynamics to identify the obsta- per year or the flow out of poverty slowed ­ overty reduction that are most likely cles to p from 14 to 13 percent per year, that would to be affected by climate change. cut the pace of poverty reduction by half. Climate events and environmental degra- dation can make people fall in poverty. How FIGURE 1.2  Flows in and out of poverty in Andhra Pradesh are larger than their net effect on poverty can climate change affect these flows in and out of poverty? We know that natural haz- Flows out of poverty ards and climate conditions are involved in 14% per year Drought, many cases where households fall into pov- irrigation failure, erty, notably because of the shocks they Net flows Decreasing the flow from or crop disease 2% per year ­ create or contribute to. Table 1.1 shows the involved in 44% 14% to 13% would halve results of 15 household surveys in six devel- poverty reduction of the cases Nonpoor oping countries—Afghanistan, India, the Increasing the flow from Poor Lao People’s Democratic Republic, Malawi, Weather events keep people poor through 12% to 13% would halve Peru, and Uganda—which ask people poverty reduction whether they have experienced a shock in asset and human capital destruction the past year (Heltberg, Oviedo, and Flows into poverty 12% per year Talukdar 2015; World Bank 2013). Among the six categories of shocks, three of the Source: Krishna 2006. most commonly reported can be directly F rom C limate C hange to P overt y and B ack : A F ramework   37 TABLE 1.1  Households in developing countries face many shocks (Percentage of respondents reporting type of shock) Afghanistan India Lao PDR Malawi Peru Uganda Shocks U R R U R U R U R U R Natural disasters (drought, flood) 10.6 42.2 57.3 5.6 36.0 10.4 47.2 2.6 21.5 19.9 52.1 Price shocks 0.2 3.0 — 4.4 4.9 21.1 42.0 — — 1.7 3.2 Employment shocks 6.4 4.3 — 9.3 3.1 7.7 3.4 6.4 1.5 1.9 0.7 Health shocks (death, illness) 6.9 14.0 30.2 23.2 33.8 10.1 18.0 9.1 8.9 11.8 14.9 Personal and property crime 1.8 6.6 0.9 5.8 1.9 8.5 8.4 3.2 3.1 6.6 8.7 Family and legal disputes — — 1.9 0.0 0.9 1.7 4.3 0.7 0.3 — — Source: World Bank 2013, based on data from household surveys, various years 2005–2011. Note: R = rural; U = urban; — = not available. related to weather events and environmental TABLE 1.2  Weather shocks hit the poorer populations the hardest conditions: in the Middle East and North Africa region (Percentage reporting economic impacts from weather shocks) • Price shocks can be linked to lower agri- Quintiles cultural production—for instance, follow- Percent Poorest Q2 Q3 Q4 Richest All ing a shock like the 2010 Russian Lost income 46 44 43 29 21 37 droughts or as a result of a long-term Lost crops 58 62 62 49 42 55 trend—and they can reduce poor people’s Lost livestock or consumption, pushing them into poverty cattle 24 25 30 23 15 23 and making it even tougher to save and Less fish caught 10 10 9 10 5 9 accumulate assets. Source: Wodon et al. 2014. • Many natural disasters are climate related. Note: Households from five countries in the region are asked to report impacts from weather shocks These include asset and crop or livestock in the last 5 years. loss that can be provoked by natural haz- ards such as droughts and water scarcity, In addition, employment can be affected and floods and storms. Natural disasters by climate events indirectly, because demand can bring people into poverty by destroy- for work can decrease as a response to disas- ing their assets—forcing them to either ters (as when productive capital is destroyed use their savings or borrow to repair or by a storm) (Hallegatte 2008). It can also be replace them—or by impacting health and affected by the degradation of natural education. resources and ecosystems, in which climate • Health issues (death and illness) are can play a role. influenced by climate and environmental Climate change is expected to make many conditions, as shown by the relationship of these shocks worse (for an in-depth dis- between rainfall and malaria outbreaks cussion, see chapters 2, 3, and 4). The effect or the role of temperature in the number of these shocks on poverty is all the more of cases of diarrhea. Moreover, health important given that people in lower income shocks remain the primary reason why quintiles often appear more exposed and people fall into poverty, owing to a com- vulnerable to weather shocks than the rest of bination of health expenditures, reduced the population. In the Middle East and income (for the sick or their caregivers), North Africa, a recent survey of five coun- and long-term consequences on produc- tries found that the bottom three quintiles tivity (like disability). are more exposed to weather shocks than the 3 8    SHOCK WAVES top two, especially in terms of income losses disasters. In Bangladesh, a recent study tells a (table 1.2). Other household surveys show a similar story. Households that receive assets similar story. (such as livestock or a sewing machine, com- Climate-related shocks push people into plete with income support and training) build poverty. In the Andhra Pradesh communities, and diversify their asset portfolio to increase among households falling into poverty, their income and reduce vulnerability to nega- 44 percent cite “drought, irrigation failure, or tive shocks (Barrett et al. 2013). crop disease” as one of the reasons for their Households also escape poverty because descent. A household affected by droughts in they have access to jobs with better wages, the past is 15 times more likely to fall into and in developing countries these jobs are poverty (Krishna 2006). In Bangladesh, of largely created by private micro-, small, and nearly 400 households falling into poverty, medium enterprises (MSMEs), often in the 15 percent cite natural disasters and informal sector. These firms therefore play a 18 ­percent the loss of natural assets as the key role in reducing poverty, especially where main reasons (Sen 2003). Changed weather many young people enter the job market patterns and modified rainfall may increase or (World Bank 2012). But recent studies show decrease the frequency and intensity of such that these firms are particularly ill-equipped weather shocks and therefore change the flow to anticipate and adapt to environmental of households into poverty. change and prepare for natural disasters. In Climatic and environmental conditions Turkey, more than 60 percent of the MSMEs affect the ability of households to escape pov- interviewed in a recent study reported that erty. Climate change may make it more diffi- they did not have enough understanding of cult for poor people to increase their income climate risks and access to resources and and accumulate assets, or may even lead in financial and insurance instruments to man- extreme cases to “poverty traps” (that is, age climate-related risks (IFC and EBRD when people own so little that they cannot 2013). This lack of preparation translates into invest to increase their income). While evi- needless losses in economic activity and jobs, dence on the existence of poverty traps is even in rich countries, as illustrated by case mixed, surveys do suggest that poor people studies after (­climate- or non-climate-related) experience slow income growth and slow disasters in the past (Groen and Polivka 2008; recovery from shocks (Antman and McKenzie Kroll et al. 1991; Tierney 1997). 2007; Carter et al. 2007; Kraay and McKenzie Climate and environmental shocks and 2014; Ravallion and Jalan 2001). And a slow- degradation can also restrict asset accumula- ing down of income growth for the poor tion and slow down poverty reduction (Carter would result in slower rates of poverty et al. 2007; Dercon and Christiaensen 2011). reduction. Poor people who have little other means to At the household level, we know that asset cope with shocks may be forced to sell their accumulation offers a way out of poverty, productive assets, such as distress sales often over several generations. In Guayaquil, of ­ l ivestock during drought periods Ecuador, a study on asset-poor households (Little et al. 2006). They may also be forced found that they start by accumulating housing to overextract environmental resources in a capital through improving their dwelling struggle for short-term survival (Reardon and (Moser and Felton 2007). This improves the Vosti 1995). Such strategies can lead to pov- quality of life, but also helps build human erty traps when they undermine the resources capital through better health, safety, and secu- poor people depend on for future income gen- rity. Next, households consume more durable eration (Barbier 2010; Barrett, Travis, and goods and diversify their asset base by invest- Dasgupta 2011). ing in productive assets (like children’s educa- Climate-sensitive shocks can also have tion and financial capital) to better cope with irreversible impacts on education and health, negative shocks such as illness or natural transmitting poverty from one generation to F rom C limate C hange to P overt y and B ack : A F ramework   39 the next. While households with enough and education for long-term prospects, pro- assets can be expected to smooth consump- ductivity, and income, even a moderate tion following a shock, asset-poor house- impact of climate change on health and edu- holds may smooth assets and destabilize cational achievement could affect poverty vis- consumption in an attempt to preserve the ibly over the long term. And because poor small productive resources they still have households suffer disproportionately from (Carter et al. 2007). This interruption in climate impacts, it would reduce the chance ­ postdisaster consumption can result in irre- for children from poor families to escape pov- versible impacts for children. A review of erty, further harming social mobility and the literature elicits many examples from increasing the intergenerational transmission S ub-Saharan Africa, but also in Asia, ­ of poverty. Latin America, and elsewhere (Baez, de la Increased risk can push poor households Fuente, and Santos 2010; Maccini and into low-risk, low-return strategies that Yang 2009). keep them poor. Natural risk can affect people’s prospects even before a disaster • Following weather shocks in Sub- hits. Household choices on risk-return Saharan Africa, asset-poor households trade-offs depend on their ability to cope typically provide children with lower- with potential negative futures (such as bad quality nutrition (Alderman, Hoddinott, rainfall, reduced consumption, and lower and Kinsey 2006; Dercon and Porter demand). With less steady income, a larger 2 0 1 4 ; Ya m a n o , A l d e r m a n , a n d percentage of total assets exposed, and Christiaensen 2005) and are less likely reduced insurance coverage, poor people to take sick children for medical consul- generally have a lower ­ ability to adapt to tations (Jensen 2000). These behaviors bad outcomes than the rich. As a result, have short- and long-term impacts, espe- low-income households disproportionately cially for children under the age of choose low-risk activities, which are also two—like stunted growth (Yamano, low-return, perpetuating poverty Alderman, and Christiaensen 2005) and (Bandyopadhyay and Skoufias 2013; a greater tendency to get sick (Dercon Dercon and Christiaensen 2011; Elbers, and Porter 2014). Gunning, and Kinsey 2007; Mobarak and • After droughts in Côte d’Ivoire, school Rosenzweig 2013). enrollment rates declined by 20 percent This effect can be as important as the actual (Jensen 2000), and drought-affected impacts of a shock. In Zimbabwe, an agricul- households in Zimbabwe delayed the tural study found that ex ante impacts from start of school for children on average 3.7 increased weather risk explain almost half of months, resulting in children completing the reduction in income due to droughts 0.4 fewer grades (Alderman, Hoddinott, (Elbers, Gunning, and Kinsey 2007). and Kinsey 2006). Such livelihood strategies often entail exces- • In Ethiopia, children younger than sive and costly diversification of activities and 36 months at the apex of the 1984 fam- less productive investments, thereby con- ine were less likely to have completed pri- straining wealth accumulation. This risk mary school, with calculations suggesting stance, in turn, discourages the adoption of this led to income losses of 3 percent per new technologies and lowers incentives to year (Dercon and Porter 2014). invest in productive capital. Risk exposure also reduces credit market Moreover, health challenges are not lim- willingness to lend—in other words, those ited to shocks: malnutrition can be a chronic with uninsured weather risk have limited condition that is expected to worsen in a access to credit and investments. Importantly, future with climate change (see chapter 4). households consider their vulnerability to nat- Considering the importance of child health ural risks such as floods and droughts when 4 0    SHOCK WAVES making risk-related decisions in other information (Adger et al. 2002; Bryan, domains—such as creating a business or Chowdhury, and Mobarak 2014; Moser migrating to a city. Research suggests that and Felton 2007). under fairly general conditions, the higher the background risk (due to floods or droughts), the less individuals are willing to take other Poverty reduction, socioeconomic risks (like innovation or entrepreneurship) (Gollier and Pratt 1996). Empirical evidence trends, and non-climate policies also provides support for this hypothesis in affect climate risk many places (Ahsan 2014; Cameron and So we know that climate events and environ- Shah 2015; van den Berg, Fort, and Burger mental degradation affect poverty reduction 2009), although not everywhere (Bchir and today. But the future impacts of climate Willinger 2013). change will depend on future conditions, Poor people can be protected by the sup- including not only the magnitude and pat- port and tools they have access to. People terns of the change in climate but also the rely on multiple support systems to manage speed and direction of poverty reduction and risks and trends—like their household and future socioeconomic changes (Hallegatte, family, the community, the socioeconomic Przyluski, and Vogt-Schilb 2011). It is not system around them (including the financial hard to imagine that, in a world where system), and the government (World Bank everyone has access to water and sanitation, 2013). How much support they receive will the impacts of climate change on waterborne largely determine the impact of various diseases will be smaller than in a world shocks and stresses on their ­welfare and their where uncontrolled urbanization has led to ability to escape poverty (chapter 5). For widespread underserved settlements located instance, financial instruments (like bank in flood zones. Similarly, in a country whose accounts and insurance contracts) help house- workers mostly work outside or live and holds and firms adapt to climate change, pre- work without air conditioning, the impact of pare for natural shocks, and recover when hotter weather on labor productivity and affected. Protected savings and borrowing income will be stronger than in an also make it possible for households to cope industrialized economy. And a poorer house- ­ with income losses while maintaining con- hold with a large share of its consumption sumption and avoiding detrimental coping dedicated to food will be more vulnerable to measures (like reducing food intake or taking ­ climate-related food price fluctuations than a children out of school). wealthier household. Social insurance and social safety nets The impacts of—and risks from—climate are also efficient tools to support poor peo- change depend on the following three factors: ple when they are affected by natural disas- ters or environmental and economic • Hazard: The physical event or trend (like shocks. When droughts in East Africa a windstorm, a flood, or a trend in tem- caused food shortages and famine in 2011, perature), which is measured using physi- Ethiopia’s Productive Safety Net Program cal metrics (like the maximum wind expanded coverage from 6.5 million to speed, the water level, or the temperature 9.6 million beneficiaries in two months and change over a decade) and is independent extended benefits from six to nine months of any socioeconomic characteristics or per year. In addition, migration and remit- human presence. tances help people manage temporary or • E xposure: T he population and the permanent shocks and escape poverty. amount of assets that are located where Migrants typically benefit, as do their the hazard can occur (like the popula- ­ family and area of origin, from remittances, tion and houses located in a flood plain) enhanced social networks, and better or more generally that are potentially F rom C limate C hange to P overt y and B ack : A F ramework   41 affected by a hazard (like the population different scenarios for the analyses at different working in the agricultural sector and time horizons: thus exposed to reduction in yields). • Vulnerability: The expected amount of • Short-term hazard. In our analyses of the loss, if a hazard occurs. This depends on short-term impacts of climate change, by the physical strength of exposed assets (a 2030, we use two scenarios for the magni- mud house tends to be more vulnerable to tude of climate change impacts ( low- flood than a brick house); the technolo- impact and high-impact scenarios) to gies used (some agricultural techniques represent the scientific uncertainty on are more vulnerable to a decrease in pre- how local climates will change in response cipitation); and the role of exposed assets to global climate change, and how physi- for the community’s well-being (the loss cal and biological systems will respond of a critical bridge typically results in (such as the effects of higher temperatures higher losses than the reconstruction on ecosystems). These scenarios do not value). depend on emissions and extend to 2030. • Long-term hazard. When analyzing the How these three factors evolve is uncer- longer-term, beyond 2050, we use two tain, given our lack of full understanding scenarios for the future of global green- about the climate system and impacts of cli- house gas emissions and climate change mate change (for instance, on ecosystems)— (low-emissions and high-emissions sce- as well as how socioeconomic systems narios), which are driven by develop- evolve. But it is also uncertain because they ment trends and climate policies. Our can be affected by policies and therefore by low-emissions scenario is the Represen- our current and future choices. These two tative Concentration Pathway (RCP)2.6, sources of uncertainty have very different which is consistent with the objective of implications: while the scientific uncertainty stabilizing climate change at 2°C above (for instance on how local climates will preindustrial temperatures (Van Vuuren change and how physical and biological sys- et al. 2011). Our high-emissions scenario tems will respond) is a bad thing, because it is the RCP8.5, which represents a world impairs decision making and creates risks, of high population and economic growth the uncertainty due to our choices is good combined with a growing use of fossil- news, because it shows that our decisions fuel energy (Riahi et al. 2011). These sce- can shape the future of climate change and narios extend to 2100, and are used for its impacts. instance to discuss the long-term impact These two sources of uncertainty also of climate change on agricultural pro- influence these factors differently at different duction and food prices in chapter 2, or time horizons. Emissions-reduction policies, the future intensity of heat waves in even those implemented today, cannot signifi- chapter 3. cantly affect the rate and magnitude of cli- • Socioeconomic scenarios. To represent mate change by 2030, so that the uncertainty the uncertainty on the exposure and about the hazard comes only from the scien- vulnerability, linked to ­ d evelopment tific uncertainty: at this time horizon, reduc- and adaptation policies, we also intro- ing climate change impacts can be done only duce two socioeconomic scenarios, the by reducing exposure and vulnerability. But poverty and prosperity scenarios, which for 2050 and beyond, greenhouse gas emis- describe different possible evolutions of sions and climate policies have a large impact the world until the end of the century, on the climate change hazard: the scientific in the absence of climate change . The uncertainty and the policy uncertainty matter prosperity scenario is optimistic, assum- for the long-term climate change hazard. To ing that the World Bank’s twin goals account for these differences, this report uses of extreme poverty eradication and 4 2    SHOCK WAVES shared prosperity are met by 2030, et al., forthcoming). Another study that ­p opu lation g row th is slow i n examines the global distribution of rural developing countries, education levels poverty in low-elevation coastal areas and labor productivity increase rap- (Barbier, forthcoming). Two additional idly, and the productivity gap between papers ­analyze the climate (rainfall and developing and developed ­ c ountries temperature) sensitivity of subsistence decreases quickly. The poverty scenario and cash incomes, and the relative vul- is pessimistic, assuming high popula- nerability of poor households, using data tion growth, low economic growth, and from 58 sites representing smallholder greater inequalities between and within production systems in (sub)tropical areas countries. Population and GDP growth with good forest access (Angelsen and in these two scenarios are based on Dokken, forthcoming; Noack et al., two socioeconomic scenarios developed forthcoming). by the scientific community to support • Natural disasters: Chapter 3 reviews the climate change research, the Shared changes that are expected in the distribu- Socio­E conomic Pathways (SSPs), and we tion, frequency, and intensity of natural add projections for poverty until 2030 hazards, exploring the exposure and vul- (chapter 6). 3 nerability of poor and nonpoor people to these shocks. While previous studies rely on self-reported shocks (which can The road map for our report be biased), this chapter provides ample evidence to support the finding that poor This report explores the impacts of climate people are more exposed and more vul- change looking not only at varying intensi- nerable. New analyses drawing on a ties of climate change but also at how ­climate variety of data sources (such as people’s impacts will vary depending on progress occupations and livelihoods, expendi- made on poverty, inequality, and access to tures, location, housing types, asset port- basic services and social protection. As such, folios, and the ability to cope with and chapters 2 to 4 present new analyses and react to shocks and economic or environ- provide a review of what we think are the mental change) from three background three major channels through which climate- papers are presented in this chapter. Two sensitive events already affect people’s move- papers investigate the relative exposure ments in and out of poverty, namely the of poor and nonpoor people to floods, following: droughts, and extreme heat in 52 devel- • Agricultural production, ecosystems, and oping countries by combining hazard food security: Chapter 2 looks at how data with household surveys (Park et al., climate change will affect food prices, forthcoming; Winsemius et al., forth- agricultural incomes, and the nonmarket coming). Another paper focuses on the consumption that is provided by ecosys- city of Mumbai, India, and explores the tems, and the consequences these effects exposure, vulnerability, and ability of will have on poverty dynamics. New poor urban dwellers to respond, based analyses from five background papers on ­survey data collected for this report presented here. Two of them explore are ­ (Patankar, forthcoming). the impact of climate change (and poli- • Health: Chapter 4 discusses the effects cies) on agricultural yields, food prices, of health shocks on poverty and explores and food security, using different model- how climate change can magnify already ing approaches and assumptions existing health risks that have conse- (Biewald et al., forthcoming; Havlík quences for people’s ability to escape F rom C limate C hange to P overt y and B ack : A F ramework   43 or stay out of poverty. This chapter different scenarios of future development, and is a review of existing work, bringing thus at worlds with different levels of expo- together the literature on the economic sure and vulnerability (Rozenberg and impact of disease and poor health and Hallegatte, forthcoming). Encouragingly, it what we know about the potential effects finds that, if socioeconomic trends and poli- of climate change on health. cies manage to minimize the exposure and vulnerability of poor people to climate change These three chapters also explore the pol- by 2030, a large fraction of the negative icy options within each of these sectors that impact of climate change on poor people can can help reduce impacts—a discussion that be prevented. These results highlight the win- is expanded upon in chapter 5 to include dow of opportunity to act now to promote cross-sectoral options such as social protec- rapid, inclusive, and climate-informed devel- tion and migration. Chapter 5 investigates opment and reduce the future impacts of various tools and support systems, looking c limate change that cannot be avoided ­ at whether poor people have access to them, through mitigation measures. and reviews recent innovations to make How development is done, and in particu- financial instruments, social safety nets, and lar how low carbon it is, will determine the remittances more efficient and useful for longer-term impacts of climate change on poor people. It also discusses the role of poverty. This means moving quickly now to governance systems in designing these decarbonize development to make poverty instruments and the policies that drive pov- reduction and climate change stabilization erty reduction and risk management and compatible (box 1.2). Multiple reports have adaptation to climate change. New insights been published on mitigation policies. from four background papers are presented The recent IPCC report reviews possible here. The first explores the social inequali- pathways toward zero net carbon emissions ­ ties that shape the ability to cope with and and discusses the policies that can be imple- adapt to climate change, especially for the mented to follow these pathways (IPCC poor and nonpoor (Tschakert, forthcom- 2014c). Three other reports—the World ing). The second models the performance of Bank’s Decarbonizing Development , the different designs of social protection OECD’s Aligning Policies for a Low-Carbon schemes in protecting poor people under Economy, and the Global Commission on the increasing climate change (Carter and Economy and Climate’s Better Growth, Janzen, forthcoming). And two case studies Better Climate (Fay et al. 2015; OECD 2015; examine how social protection can protect NCE 2014)—also explore policy options, poor people against natural hazards and from carbon pricing to innovation, environ- environmental changes, looking at Ethiopia mental performance standards, and land use and its Productive Safety Net Program and and urban planning. the Philippines and its response to Typhoon In this report, we do not present a Haiyan (Bowen, forthcoming; Johnson and detailed discussion of mitigation policies, Bowen, forthcoming). but chapter 2 touches on the impact of Chapter 6 brings all of the findings land-based mitigation policies (like fighting together to highlight the extent to which over- deforestation) on poverty, and chapter 6 all development patterns will condition how explores how to design mitigation policies climate change affects poverty. Based on a that do not slow down poverty reduction— background paper for this report, it presents in particular by combining mitigation poli- the results of a novel modeling exercise that cies with measures to protect poor and examines how climate change would affect vulnerable people against potential nega- extreme poverty in 2030 looking across tive impacts. 4 4    SHOCK WAVES BOX 1.2  A call for zero net CO2 emissions by 2100 The scientific reality is that as long as human soci- Rozenberg et al. 2015). As a reference, the car- (­ eties release CO2 in quantities greater than natural bon intensity of global GDP today is around 360 carbon sinks (such as forests and other vegetation) gCO2 /$. There is no country—except those that can absorb, the climate will continue changing, with outsource emissions (like Singapore or Hong Kong corresponding risks for well-being, development, and SAR, China)—whose carbon intensity of GDP today poverty reduction (IPCC 2014a). This is why the approaches 100 gCO2 /$. This is true even of coun- international community committed to the goal of tries that produce most of their electricity from non- keeping global temperature increases below 2°C, and fossil energy sources (such as Brazil and France). thus to the full decarbonization of the global econ- Moreover, delaying action locks the global econ- omy (UNFCCC 2010; G7 2015). omy into a carbon-intensive pathway—where it Because energy consumption and related emis- becomes increasingly more difficult and expensive to sions are unlikely to decrease by themselves, main- reduce CO2 emissions sufficiently to stay within a low taining global warming below 2°C, or even 3°C, budget of cumulative emissions—because of the long requires immediate action. This requires a radical lifetime and relatively high capital cost of energy infra- change in development patterns. For instance, a structure. Each year that additional mitigation efforts background study for this report finds that it will are delayed (that is, we continue along the current be very difficult to remain below a 2°C warm- path), the required carbon intensity of new production ing if the carbon intensity of new development decreases by 20 to 50 gCO2/$ for a 2°C budget. exceeds 73 gCO2 /$ (grams of carbon dioxide emit- ted per dollar of GDP) on average at the global level Source: Fay et al. 2015; Rozenberg et al. 2015. Notes we can only add poverty ­ projections until 2030 (see chapter 6 for details on the method- 1. Until this year, extreme poverty was defined ology and its limits). using the $1.25 poverty line, based on the 2005 PPP exchange rates. Since the publica- tion of the Global Monitoring Report (World Bank 2015a), the poverty line is defined by References a consumption threshold at $1.90, using the Adger, W. N., P. M. Kelly, A. Winkels, L. Q. Huy, 2011 PPP exchange rates. These two poverty and C. Locke. 2002. “Migration, Remittances, lines are consistent, and the country-level Livelihood Trajectories, and Social Resilience.” poverty headcounts calculated using the two AMBIO J. Hum. Environ. 31: 358–66. definitions are close (Jolliffe and Prydz 2015). Ahsan, D. A. 2014. “Does Natural Disaster 2. Similar findings exist in many countries and Influence People’s Risk Preference and Trust? regions, see for instance data for four other An Experiment from Cyclone Prone Coast of countries in Krishna (2007), for Bangladesh Bangladesh.” Int. J. Disaster Risk Reduct. 9: in Sen (2003), and for South Africa in Carter 48–57. and May (2011). Alderman, H., J. Hoddinott, and B. Kinsey. 2006. 3. The prosperity scenario uses population and “Long-Term Consequences of Early Childhood GDP projections from SSP5 and the poverty Malnutrition.” Oxf. Econ. Pap. 58: 450–74. scenario uses SSP4 (O’Neill et al. 2015). While Angelsen, A., and T. Dokken. Forthcoming. we use the quantification of the SSPs for popu- “Environmental Reliance, Climate Exposure, lation and GDP, we do not retain the narra- and Vulnerability: A Cross-Section Analysis of tives associated to these SSPs, especially regard- Structural and Stochastic Poverty.” Background ing the energy mix. For more information paper prepared for this report. on the SSPs, see https://www2.cgd.ucar.edu​ Antman, F., and D. McKenzie. 2007. “Poverty research/iconics. While the SSPs go to 2100, /­ Traps and Nonlinear Income Dynamics F rom C limate C hange to P overt y and B ack : A F ramework   45 with Measurement Error and Individual Brown, C., R. Meeks, K. Hunu, and W. Yu. 2010. Heterogeneity.” J. Dev. Stud. 43: 1057–83. “Hydroclimate Risk to Economic Growth Baez, J. E., A. de la Fuente, and I. Santos. 2010. in Sub-Saharan Africa.” Clim. Change 106: “Do Natural Disasters Affect Human Capital? 621–47. An Assessment Based on Existing Empirical Bryan, G., S. Chowdhury, and A. M. Mobarak. Evidence.” IZA Discussion Paper No. 5164, 2014. “Underinvestment in a Profitable Institute for the Study of Labor (IZA). Te c h n o l o g y : T h e C a s e o f S e a s o n a l Bandyopadhyay, S., and E. Skoufias. 2013. “Rainfall Migration in Bangladesh.” Econometrica 82: Variability, Occupational Choice, and Welfare in 1671–1748. Rural Bangladesh.” Rev. Econ. Househ. 1–46. Cameron, L., and M. Shah. 2015. “Risk-Taking Barbier, E. B. 2010. “Poverty, Development, and Behavior in the Wake of Natural Disasters.” Environment.” Environ. Dev. Econ. 15: 635–60. J. Hum. Resour. 50: 484–515. Barbier, E. B. Forthcoming. “Climate Change Carter, M. R., and S. A. Janzen. Forthcoming. Impacts on Rural Poverty in Low-Elevation “Social Protection in the Face of Climate Coastal Zones.” Background paper prepared Change: Targeting Principles and Financing for this report. Mechanisms.” Background paper prepared for Barrett, A., R. McIntosh, M. Pritchard, M. Hannan, this report. Z. Alam, and M. Marks. 2013. Asset Values: Carter, M. R., P. D. Little, T. Mogues, and Why Are Some Households Doing Better than W. Negatu. 2007. “Poverty Traps and Natural Others? Chars Livelihood Program. Disasters in Ethiopia and Honduras.” World Barrett, C. B., A. J. Travis, and P. Dasgupta. 2011. Dev. 35: 835–56. “On Biodiversity Conservation and Poverty Carter, M. R., and J. May. 2001. “One Kind of Traps.” Proc. Natl. Acad. Sci. 108: 13,907–12. Freedom: Poverty Dynamics in Post-apartheid Barrios, S., L. Bertinelli, and E. Strobl. 2010. South Africa.” World Dev. 29: 1987–2006. “Trends in Rainfall and Economic Growth Dang, H.-A., P. Lanjouw, and R. Swinkels. 2014. in Africa: A Neglected Cause of the African “Who Remained in Poverty, Who Moved Growth Tragedy.” Rev. Econ. Stat. 92: 350–66. Up, and Who Fell Down? An Investigation of Baulch, B. 2011. Why Poverty Persists: Poverty Poverty Dynamics in Senegal in the Late 2000s.” Dynamics in Asia and Africa. Cheltenham, Policy Res. Work. Pap. No 7141, World Bank, U.K.: Edward Elgar Publishing. Washington, DC. Bchir, M. A., and M. Willinger. 2013. “Does the Dell, M., B. F. Jones, and B. A. Olken. 2009. Exposure to Natural Hazards Affect Risk and “Temperature and Income: Reconciling New Time Preferences? Some Insights from a Field Cross-Sectional and Panel Estimates.” Am. Experiment in Perú.” Working Paper No. Econ. Rev. 99: 198–204. 13-04, LAMETA, University of Montpellier. ———. 2012. “Temperature Shocks and Economic Beegle, K., J. De Weerdt, and S. Dercon. 2006. Growth: Evidence from the Last Half Century.” Poverty and Wealth Dynamics in Tanzania: Am. Econ. J. Macroecon. 4: 66–95. Evidence from a Tracking Survey. Washington, Dercon, S., and L. Christiaensen. 2011. DC: World Bank. “Consumption Risk, Technology Adoption Biewald, A., H. Lotze-Campen, I. Otto, N. and Poverty Traps: Evidence from Ethiopia.” Brinckmann, B. Bodirsky, I. Weindl, A. Popp, J. Dev. Econ. 96: 159–73. and H. J. Schellnhuber. Forthcoming. “The Dercon, S., and C. Porter. 2014. “Live Aid Impact of Climate Change on Costs of Food Revisited: Long-Term Impacts of the 1984 and People Exposed to Hunger at Subnational Ethiopian Famine on Children.” J. Eur. Econ. Scale.” PIK Report 128, Potsdam Institute for Assoc. 12: 927–48. Climate Impact Research. Background paper Dollar, D., T. Kleineberg, and A. Kraay. 2013. prepared for this report. “Growth Still Is Good for the Poor.” SSRN Bourguignon, F., and S. R. Chakravarty. 2003. Scholarly Paper No. ID 2309849. Social Science “The Measurement of Multidimensional Research Network, Rochester, NY. Poverty.” J. Econ. Inequal. 1: 25–49. Dollar, D., and A. Kraay. 2002. “Growth Is Good Bowen, T. Forthcoming. “Social Protection and for the Poor.” J. Econ. Growth 7: 195–225. Disaster Risk Management in the Philippines: Elbers, C., J. W. Gunning, and B. Kinsey. 2007. The Case of Typhoon Yolanda (Haiyan).” “Growth and Risk: Methodology and Micro Background paper prepared for this report. Evidence.” World Bank Econ. Rev. 21: 1–20. 4 6    SHOCK WAVES Fay, M., S. Hallegatte, A. Vogt-Schilb, J. Rozenberg, IFC (International Finance Corporation) and U. Narloch, and T. Kerr. 2015. Decarbonizing EBRD (European Bank for Reconstruction and Development: Three Steps to a Zero-Carbon Development). 2013. Pilot Climate Change Future. Washington, DC: World Bank. Adaptation Market Study: Turkey. Ferreira, F. H. G., and M. Ravallion. 2009. IPCC (Intergovernmental Panel on Climate “Poverty and Inequality: The Global Context.” Change). 2014a. Climate Change 2014: In The Oxford Handbook of Economic Synthesis Report. Contribution of Working Inequality, edited by W. Salverda, B. Nolan, Groups I, II, and III to the Fifth Assessment and T. M. Smeeding. Oxford, U.K.: Oxford Report of the Intergovernmental Panel University Press. on Climate Change. [Core Writing Team, Ferreira, F. H. G. and M. A. Lugo. 2013. R.K. Pachauri and L.A. Meyer (eds.)]. Geneva, “Multidimensional Poverty Analysis: Looking Switzerland: IPCC. for a Middle Ground.” World Bank Res. Obs. ———. 2014b. Climate Change 2014: Impacts, lks013. Adaptation, and Vulnerability. Part A: Global G7. 2015. Leaders’ Declaration G7 Summit and Sectoral Aspects. Contribution of Working Germany. WWW Document. URL http:// Group II to the Fifth Assessment Report of the www.international.gc.ca/g8/g7_germany​ Intergovernmental Panel on Climate Change. _declaration-g7_allemagne_declaration​ [Field, C.B., V.R. Barros, D.J. Dokken, .aspx?lang=eng (accessed 6.14.15). K.J. Mach, M.D. Mastrandrea, T.E. Bilir, Gentilini, U. Forthcoming. Safety Nets in Urban M. Chatterjee, K.L. Ebi, Y.O. Estrada, R.C. Areas: Emerging Issues, Evidence and Practices. Genova, B. Girma, E.S. Kissel, A.N. Levy, Gollier, C., and J. W. Pratt. 1996. “Risk S. MacCracken, P.R. Mastrandrea, and L.L. Vulnerability and the Tempering Effect of White (eds.)]. Cambridge, U.K. and New York: Background Risk.” Econometrica 64: 1109–23. Cambridge University Press. Groen, J. A., and A. E. Polivka. 2008. “The Effect ———. 2014c. “Summary for Policymakers.” In of Hurricane Katrina on the Labor Market Climate Change 2014, Mitigation of Climate Outcomes of Evacuees.” Am. Econ. Rev . Change . Contribution of Working Group 43–48. III to the Fifth Assessment Report of the Hallegatte, S. 2008. “An Adaptive Regional Intergovernmental Panel on Climate Change. Input-Output Model and Its Application to the [Edenhofer, O., R. Pichs-Madruga, Y. Sokona, Assessment of the Economic Cost of Katrina.” E. Farahani, S. Kadner, K. Seyboth, A. Adler, Risk Anal. 28: 779–99. I. Baum, S. Brunner, P. Eickemeier, B. Kriemann, Hallegatte, S., V. Przyluski, and A. Vogt-Schilb. J. Savolainen, S. Schlömer, C. von Stechow, 2011. “Building World Narratives for Climate T. Zwickel, and J.C. Minx (eds.)]. Cambridge, Change Impact, Adaptation and Vulnerability U.K. and New York: Cambridge University Press. Analyses.” Nat. Clim. Change 1: 151–55. Ivanic, M., W. Martin, and H. Zaman. 2012. Havlík, P., H. Valin, M. Gusti, E. Schmid, “Estimating the Short-Run Poverty Impacts of D. Leclère, N. Forsell, M. Herrero, N. Khabarov, the 2010–11 Surge in Food Prices.” World Dev. A. Mosnier, M. Cantele, and M. Obersteiner. 40: 2302–17. Forthcoming. “Climate Change Impacts and Jensen, R. 2000. “Agricultural Volatility and Mitigation in the Developing World: An Investments in Children.” Am. Econ. Rev. 90: Integrated Assessment of Agriculture and 399–404. Forestry Sectors.” Background paper prepared Johnson, K., and T. Bowen. Forthcoming. for this report. “Ethiopia’s Food Security Program: Social Heltberg, R., A. M. Oviedo, and F. Talukdar. Protection, Disaster Risk Management and 2015. “What Do Household Surveys Really Tell Climate Change Adaptation.” Background Us about Risk, Shocks, and Risk Management paper prepared for this report. in the Developing World?” J. Dev. Stud. 51: Jolliffe, D., and E. B. Prydz. 2015. “Global Poverty 209–25. Goals and Prices: How Purchasing Power Parity Horowitz, J. K. 2009. “The Income–Temperature Matters.” Policy Res. Work. Pap. No 7256, Relationship in a Cross-Section of Countries World Bank, Washington, DC. and Its Implications for Predicting the Effects Kraay, A., and D. McKenzie. 2014. “Do Poverty of Global Warming.” Environ. Resour. Econ. Traps Exist? Assessing the Evidence.” J. Econ. 44: 475–93. Perspect. 28: 127–48. F rom C limate C hange to P overt y and B ack : A F ramework   47 Krishna, A. 2006. “Pathways Out of and Into Organization for Economic Co-Operation and Poverty in 36 Villages of Andhra Pradesh, Development. India.” World Dev. , Part Special Issue: Olsson, L., M. Opondo, P. Tschakert, A. Agrawal, 324–404. Corruption and Development: S. H. Eriksen, S. Ma, L. N. Perch, and S. A. Analysis and Measurement 34: 271–88. Zakieldeen. 2014. “Livelihoods and Poverty.” ——— . 2 0 0 7 . “ T h e S t a g e s o f P r o g r e s s In Climate Change 2014: Impacts, Adaptation, Methodology and Results from Five and Vulnerability. Part A: Global and Sectoral Countries.” Reducing Glob. Poverty Case Asset Aspects, edited by C. B. Field, V. R. Barros, D. Accumul. 62–79. J. Dokken, K. J. Mach, M. D. Mastrandrea, T. Kroll, C. A., J. D. Landis, Q. Shen, and S. Stryker. E. Bilir, M. Chatterjee, K. L. Ebi, Y. O. Estrada, 1991. Economic Impacts of the Loma Prieta R. C. Genova, B. Girma, E. S. Kissel, A. N. Earthquake: A Focus on Small Businesses . Levy, S. MacCracken, P. R. Mastrandrea, and University of California at Berkeley. L. L. White, 793–832. Contribution of Working Lanjouw, P., D. McKenzie, and J. Luoto. 2011. Group II to the Fifth Assessment Report of the “Using Repeated Cross-Sections to Explore Intergovernmental Panel on Climate Change. Movements In and Out of Poverty.” Policy Res. Cambridge, U.K. and New York: Cambridge Work. Pap. 5550, World Bank, Washington, University Press. DC. O’Neill, B. C., E. Kriegler, K. L. Ebi, E. Kemp- Little, P. D., M. P. Stone, T. Mogues, A. P. Castro, Benedict, K. Riahi, D. S. Rothman, B. J. van and W. Negatu. 2006. “Moving in Place: Ruijven, D. P. van Vuuren, J. Birkmann, Drought and Poverty Dynamics in South K. Kok, M. Levy, and W. Solecki. 2015. Wollo, Ethiopia.” J. Dev. Stud. 42: 200–25. “The Roads Ahead: Narratives for Shared Maccini, S., and D. Yang. 2009. “Under the Socioeconomic Pathways Describing World Weather: Health, Schooling, and Economic Futures in the 21st Century.” Glob. Environ. Consequences of Early-Life Rainfall.” Change. Am. Econ. Rev. 99: 1006–26. Park, J., S. Hallegatte, M. Bangalore, and Mobarak, A. M., and M. R. Rosenzweig. 2013. E. Sandhoefner. Forthcoming. “Households “Informal Risk Sharing, Index Insurance, and Heat Stress: Estimating the Distributional and Risk Taking in Developing Countries.” Consequences of Climate Change.” Background Am. Econ. Rev. 103: 375–80. paper prepared for this report. Moser, C., and A. Felton. 2007. “Intergenerational Patankar, A. Forthcoming. “The Exposure, Asset Accumulation and Poverty Reduction in Vulnerability, and Ability to Respond of Poor Guayaquil, Ecuador, 1978-2004.” Reducing Households to Recurrent Floods in Mumbai.” Glob. Poverty Case Asset Accumul. 15–50. Background paper prepared for this report. Moser, C. O. N. 2008. Reducing Global Poverty: Pindyck, R. S. 2013. “Climate Change Policy: The Case for Asset Accumulation. Washington, What Do the Models Tell Us?” J. Econ. Lit. 51: DC: Brookings Institution Press. 860–72. NCE (New Climate Economy). 2014. Better Ravallion, M. 2010. “Do Poorer Countries Have Growth Better Climate: The New Climate Less Capacity for Redistribution?” J. Glob. Economy Report . Washington, DC: New Dev. 1. Climate Economy. Ravallion, M., and J. Jalan. 2001. “Household Noack, F., S. Wunder, A. Angelsen, and J. Börner. Income Dynamics in Rural China.” Policy Forthcoming. “Responses to Weather and Research Working Paper, World Bank, Climate: A Cross-Section Analysis of Rural Washington, DC. Incomes.” Background paper prepared for this Reardon, T., and S. A. Vosti. 1995. “Links report. between Rural Poverty and the Environment O’Brien, K. L., and R. M. Leichenko. 2000. in Developing Countries: Asset Categories “Double Exposure: Assessing the Impacts and Investment Poverty.” World Dev . 23: of Climate Change within the Context of 1,495–1,506. Economic Globalization.” Glob. Environ. Riahi, K., S. Rao, V. Krey, C. Cho, V. Chirkov, Change 10: 221–32. G. Fischer, G. Kindermann, N. Nakicenovic, OECD (Organisation for Economic Co-operation and P. Rafaj. 2011. “RCP 8.5—A Scenario and Development). 2015. Aligning Policies of Comparatively High Greenhouse Gas for a Low-Carbon Economy. Paris, France: Emissions.” Clim. Change 109: 33–57. 4 8    SHOCK WAVES Robalino, D. A., and P. G. Warr. 2006. “Poverty Experimental Evidence from Latin America.” Reduction through Fiscal Restructuring: An Presented at the International Association of Application to Thailand.” J. Asia Pac. Econ. Agricultural Economists Conference, Beijing, 11: 249–67. China. Rozenberg, J., S. J. Davis, U. Narloch, and Van Vuuren, D. P., E. Stehfest, M. G. den Elzen, S. Hallegatte. 2015. “Climate Constraints on T. Kram, J. van Vliet, S. Deetman, M. Isaac, the Carbon Intensity of Economic Growth.” K. K. Goldewijk, A. Hof, and A. M. Beltran. Environ. Res. Lett. 10: 095006. 2011. “RCP2. 6: Exploring the Possibility Rozenberg, J., and S. Hallegatte. Forthcoming. to Keep Global Mean Temperature Increase “The Impacts of Climate Change on Poverty Below 2 C.” Clim. Change 109: 95–116. in 2030, and the Potential from Rapid, Weitzman, M. L. 2014. “Fat Tails and the Inclusive and Climate-Informed Development.” Social Cost of Carbon.” Am. Econ. Rev. 104: Background paper prepared for this report. 544–46. Sen, B. 2003. “Drivers of Escape and Descent: Wilbanks, T. J., and K. L. Ebi. 2013. “SSPs from Changing Household Fortunes in Rural an Impact and Adaptation Perspective.” Clim. Bangladesh.” World Dev., Chronic Poverty and Change 1–7. Development Policy 31: 513–34. Winsemius, H., B. Jongman, T. Veldkamp, Skoufias, E., M. Rabassa, and S. Olivieri. 2011. S. Hallegatte, M. Bangalore, and P. J. Ward. “The Poverty Impacts of Climate Change: Forthcoming. “Disaster Risk, Climate Change, A Review of the Evidence.” SSRN Scholarly and Poverty: Assessing the Global Exposure Paper No. ID 1803002, Social Science Research of Poor People to Floods and Droughts.” Network, Rochester, NY. Background paper prepared for this report. Stern, N. 2006. Stern Review: The Economics of Wodon, Q., A. Liverani, G. Joseph, and Climate Change. London, U.K. HM Treasury. N. Bougnoux. 2014. Climate Change and ———. 2013. “The Structure of Economic Migration: Evidence from the Middle Modeling of the Potential Impacts of Climate East and North Africa . Washington, DC: Change: Grafting Gross Underestimation of World Bank. Risk onto Already Narrow Science Models.” World Bank. 2012. Jobs and Development: World J. Econ. Lit. 51: 838–59. Development Report 2013. Washington, DC: Tierney, K. J. 1997. “Business Impacts of the World Bank. Northridge Earthquake.” J. Contingencies ———. 2013. Risk and Opportunity: Managing Crisis Manag. 5: 87–97. Risk for Development, World Development Tschakert, P. Forthcoming. “The Ability of the Report 2014. Washington, DC: World Bank. Poor to Cope.” Background paper prepared for ———. 2015a. Global Monitoring Report this report. 2015/2016. Washington, DC: World Bank. UNFCCC (United Nations Framework Convention ———. 2015b. Turn Down the Heat. Washington, on Climate Change). 2010. 16th Conference of DC: World Bank. the Parties of the United Nations Framework Yamano, T., H. Alderman, and L. Christiaensen. Convention on Climate Change. 2005. “Child Growth, Shocks, and Food Aid van den Berg, M., R. Fort, and K. Burger. in Rural Ethiopia.” Am. J. Agric. Econ. 87: 2009. “Natural Hazards and Risk Aversion: 273–88. Bad Seed: Climate Change, Agriculture, and Food Security 2 Main Messages • Climate change impacts on agriculture are • Climate change adds to the stress on eco- already evident in vulnerable regions, such as systems and makes them even more fragile. Sub-Saharan Africa and South Asia, even as Impacts for poor communities dependent on they remain globally limited. Future impacts ecosystem-based livelihoods can be large— are highly uncertain, but are expected to be cutting their subsistence production and significant—even when accounting for adap- removing one of their safety nets—but are tive behaviors and trade. Land-based mitiga- still impossible to quantify. tion can also bring risks for food production • In the shorter term, food stocks, better access and prices. of poor farmers to markets, improved technolo- • Production and price impacts—whether gies, and climate-smart production practices can triggered by climate change or climate reduce climate impacts. Over the longer term, p olicies—significantly affect poor people. ­ negative impacts can be avoided only through Net-consumers of food products will be mitigation actions, including in agriculture, for- harmed, while those who depend on agricul- estry, and other land uses. These actions can be tural wages and profits will experience mixed designed to avoid negative impacts or even to impacts. benefit local production and incomes. Introduction failures are among the chief culprits. And we know that many of these shocks can be trig- What are the key obstacles hindering poor gered by climate-related events, whose fre- people from escaping poverty and making quency and severity will increase with them vulnerable to falling back into poverty? climate change. In particular, the pace of Household surveys investigating this ques- poverty reduction could be affected by cli- tion reveal that changes in prices, employ- mate change impacts on agriculture and ment shocks, or death of livestock and crop ecosystems. 49 5 0    SHOCK WAVES This chain of events is expected to occur large and well-designed mitigation policies are because climate change will seriously affect the only way to reduce longer-term risks. land and water productivity. This, in turn, will increase food prices and affect wages and incomes, especially in the agriculture sector. Climate change and climate The impact of such changes in prices and earnings on households depends on their con- policies will impact food security sumption basket and income sources. Food Agriculture is one of the most important eco- dominates the consumption basket of poor nomic sectors in many poor countries. It is households, whose livelihoods—particularly also directly critical to households’ food in poor countries—are often derived from security. Unfortunately, it is also one of the agriculture and other environmental goods most sensitive to climate change given its and services. dependence on weather conditions, both This chapter explores how climate change directly and through climate-dependent and climate policies could affect poor people stressors (pests, epidemics, and sea level through these channels. Importantly, these rise). In fact, some of the most severe pov- effects can be triggered either by short-lived erty impacts of climate change are expected natural disasters or by more gradual, long- to be channeled through agriculture. term changes in climate conditions. While In its latest report, the Intergovernmental natural disasters are discussed at length in Panel on Climate Change (IPCC) notes that chapter 3, this chapter focuses on long-term there is high confidence that all aspects of changes and trends that threaten poor people. food security will be negatively affected by But the line is often blurred between shocks climate change (IPCC 2014). However, the and trends, and the two interact closely. For extent to which yields and prices will be instance, tight agricultural markets—due affected is hard to predict. Uncertainty notably to limited food stocks—make it more about future precipitation and temperature likely that a relatively small event in one change are compounded by uncertainties region (such as a drought in a major export- regarding the likely response of crop ing country) translates into a food price crisis, growth to changes in climate conditions. as illustrated by the consequences of the 2010 Further, the magnitude of CO2 fertilization Russian drought. (that is, the process through which higher We begin by reviewing the possible impacts CO 2 concentrations directly accelerate of climate change and mitigation policies on plant growth) is still unknown. As a result, food security—in terms of agricultural yields, climate change impacts on crop yields and food prices, and food availability—and the prices are highly uncertain and vary across consequences for poor households of such regions, crops, and adaptation scenarios, changes and the associated impacts on ecosys- depending on crop models, economic mod- tems the poor depend on. We then examine els, and underlying assumptions (Nelson how socioeconomic development, improved et al. 2014). technologies, better infrastructure and market This chapter adds to the rapidly growing access, and well-designed land mitigation knowledge on climate change and agriculture ­ policies could help. by presenting the results of new modeling Our main message is that climate change exercises undertaken for this report—one on can have significant impacts on agricultural hunger vulnerability (Biewald et al., forth- yields and food prices, as well as ecosystem- coming) and one on food prices (Havlík et al., based livelihoods, which will particularly affect forthcoming). These exercises combine a crop poor consumers and rural people. However, model for biophysical yield impacts with an impacts will remain globally limited (but economic and trade model to calculate pro- locally significant) until 2030 and can be duction and price impacts under a variety of largely managed through good policies during scenarios, including with and without CO2 this period. Beyond 2030, impacts can become fertilization, and distinguishing between B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    51 short-term (until 2030) and longer-term (until FIGURE 2.1  Climate change could sharply reduce crop yields 2080) impacts.1 (Change in yields compared to no climate change) 5 Climate change could reduce yields in 0 many places, especially the poorest Change in yields (%) –5 As shown in the latest IPCC assessment, food –15 production will be directly affected by changes in climatic conditions (Porter et al., –25 2014): Crop yields and harvest quality are susceptible to extreme events and changing –35 precipitation and temperature. Livestock Low High Low High emissions emissions emissions emissions production can be impacted by grazing land productivity and quality, heat stress, and 2030 2080 water availability. The distribution and abun- GFDL-ESM2M MIROC-ESM-CHEM IPSL-CM5A-LR dance of aquatic species will change with NorESM1-M HadGEM2-ES HadGEM2-ES without negative impacts expected for fish production CO2 fertilization in developing countries in tropical areas. Notwithstanding high uncertainties, the Source: Havlík et al., forthcoming. Note: Results from Global Biosphere Management Model (GLOBIOM) based on 18 species IPCC concludes with high confidence that aggregated on a dry matter yield basis. Climate change impacts are shown for 5 climate crop production will consistently and nega- ­models (­HadGEM2-ES, IPSL-CM5A-LR, GFDL-ESM2M, MIROC-ESM-CHEM, NorESM1-M) with CO2 ­fertilization and HadGEM2-ES without CO2 fertilization. tively be affected by climate change in the lon- ger term and in low-latitude countries (IPCC 2014). Projected impacts vary across crops, regions, and adaptation scenarios with posi- Climate-induced yield reductions are not tive and negative impacts being equally pos- homogenous. Climate change will benefit sible before 2050. Beyond that, risks of some cold regions in the short run, but these negative impacts become more severe. regions are relatively wealthy. In contrast, it Accordingly, Havlík et al. (Forthcoming) will hit other regions especially hard, particu- show that climate change is likely to have a larly the poorest ones. By 2080, the average detrimental impact although impacts remain yield declines estimated from all climate mod- limited in the shorter term. As figure 2.1 els could be as severe as 23 percent for South shows, in a high-emissions scenario, declines Asia, 17 percent for East Asia and the Pacific, in average crop yields remain limited in 2030 15 percent for Sub-Saharan Africa, and 14 once the positive effect of CO2 fertilization is percent for Latin America—even with CO2 accounted for, and do not exceed 10 percent fertilization (Havlík et al., forthcoming). But even without CO2 fertilization. Beyond 2030, if CO2 fertilization effects do not materialize, the severity of the damage will depend on overall impacts can be more severe, with all the actions countries take to decrease their regions experiencing negative yield changes. greenhouse gas (GHG) emissions. In a low-­ Using another crop and economic model emissions scenario, overall crop yield losses brings similar results (Biewald et al., forth- could be stabilized at less than 8 percent by coming): its finer spatial resolution shows 2080 compared to a situation without climate that yield impacts will also vary greatly within change (and with CO2 fertilization, net gains countries, with some regions benefiting while are even possible). But if GHG emissions con- others lose much more than the regional or tinue to increase on an uncontrolled path country average. Climate change could even (high-emissions scenario), yields could make agricultural areas unsuitable for cultiva- decrease by up to 20 percent (on average 14 tion of key crops, resulting in large economic percent across climate models) in 2080 even impacts for poor economies that are highly with CO2 fertilization—and up to more than dependent on a few agricultural commodities 30 percent without CO2 fertilization. (box 2.1). 5 2    SHOCK WAVES BOX 2.1  Climate change could pose major hurdles for Africa’s leading cocoa and coffee producers Climate change can modify the suitability of particu- History suggests that specialized economies lar locations for agriculture production, which could find it difficult to recover from the collapse of their be a major challenge for agriculture-dependent econ- main activity. For instance, one study shows that omies. In Ghana and Côte d’Ivoire, areas suitable regions affected by the Dust Bowl in the United for cocoa production could be negatively affected by States in the 1930s never returned to their pre- climate change—a major problem for two countries drought production levels, with most adjustment whose economies depend critically on employment taking place through outmigration toward other and revenues from this export crop (map B2.1.1). areas (Hornbeck 2012). Since production of such Similarly, in Uganda, the leading coffee producer in cash crops will be affected slowly, a well-prepared Africa, significant declines in most areas suitable for transition toward other production—in agricul- coffee production are possible (Jassogne, Läderach, ture, manufacturing, or services—would be well and van Asten 2013). advised. MAP B2.1.1  Ghana and Côte d’Ivoire could experience a loss of area suitable for cocoa production by 2050 Source: Läderach et al. 2013. Note: Maps show suitability predicted by Maximum entropy (MAXENT) model, which incorporates crop-environment interactions based on the current climatic conditions in cocoa growing areas (panel a) and changes in the climatic conditions according to climate change projections (panel b). B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    53 Biophysical impacts affect prices and and the productivity gap between developing food availability and developed countries decreases quickly. The poverty scenario is more pessimistic, The biophysical impacts on crop yields will assuming high population growth, low eco- also trigger changes in production and food nomic growth, and greater inequalities prices. The impact of these changes will between and within countries (see chapter 1 depend on how farmers and countries adapt for a discussion of the scenario approach to them. Farmers can adjust input use (fertil- used in this report). izer and irrigation) and cultivated area to Over the long term, the risks of harmful compensate for some of the yield losses; and price impacts could be high, especially with- countries can buffer their production deficit out CO2 fertilization. The simulations carried by increasing their food imports. (These are out for this report (figure 2.2) suggest that in the adaptation options used in the back- scenarios with continued high emissions and ground modeling for this chapter.) Food no CO2 fertilization, climate change would prices will evolve depending on the interplay increase world agricultural prices by 4 to of demand and supply of food, which may 5.5 percent in 2030. Over time, these impacts be affected by public policies regarding food increase and could be as large as 30 percent stock management and trade policies as well by 2080.2 With CO2 fertilization, impacts on as management of losses during storage and global prices remain limited even with high transport. emissions. What happens to agricultural prices is par- Regions are affected very differently by ticularly critical for poverty given their impact ­ climate change induced price changes, with on poor households’ budget and income. Sub-Saharan Africa and South Asia the most According to the IPCC, changes in tempera- severely impacted (figure 2.2). Impacts on ture and precipitation are likely to result in prices could be as high as 12 percent in 2030 higher food prices by 2050, but the magni- and 70 percent by 2080 in Sub-Saharan tude remains highly uncertain with increases Africa in a worst-case scenario (poverty and ranging from 3 to 84 percent without CO2 high emissions without CO2 fertilization). In fertilization effects, and between decreases of the same worst-case scenario, prices would 30 percent and increases of 45 percent with rise by 5 percent by 2030 and 23 percent by CO2 fertilization (IPCC 2014). These esti- 2080 in South Asia, 4 percent and 9 percent mates may be optimistic in that they do not in East Asia and the Pacific, and 3 percent and account for impacts of pests, diseases, interac- 12 percent in Latin America. Even with CO2 tion with local pollution, and extreme fertilization, in 2080 prices would be 29 per- weather events. Extreme events in particular cent higher in Sub-Saharan Africa and 16 per- could put global food systems at risk (U.K.– cent in South Asia. U.S. Task Force 2015). As for food availability, unmitigated cli- Impacts also depend on socioeconomic mate change has the potential to cancel out a trends, including technological change and large fraction of the food security gains from population growth. This is why all simula- technological change and economic growth tions are performed under two socioeco- and can be a threat in regions that already nomic scenarios—labeled prosperity and have low levels of food intake per capita and poverty—that represent different evolutions will experience high levels of population of the world’s population and economy. The growth. prosperity scenario is optimistic, assuming Sub-Saharan Africa and South Asia are that the World Bank’s twin goals of extreme particularly vulnerable. In those regions, poverty eradication and shared prosperity without climate change, per capita food avail- are met by 2030, that population growth is ability could increase at least to a level equiv- slow in developing countries, education lev- alent to 80 percent of the current food els and labor productivity increase rapidly, availability in developed countries by 2080, 5 4    SHOCK WAVES FIGURE 2.2  Sub-Saharan Africa and South Asia are the most vulnerable to climate-induced increases in agricultural prices (Climate change impacts on agriculture prices) a. World 40 Change in agriclutural prices (%) Poverty and high emissions w/o CO2 fertilization 20 Poverty and high emissions Poverty and low emissions Prosperity and high emissions w/o CO2 fertilization 0 Prosperity and high emissions Prosperity and low emissions –20 2000 2030 2050 2080 b. Sub-Saharan Africa c. South Asia 40 40 Change in agriclutural prices (%) Change in agriclutural prices (%) 20 20 0 0 –20 –20 2000 2030 2050 2080 2000 2030 2050 2080 d. East Asia and Pacific e. Europe and Central Asia 40 40 Change in agriclutural prices (%) Change in agriclutural prices (%) 20 20 0 0 –20 –20 2000 2030 2050 2080 2000 2030 2050 2080 f. Latin America and Caribbean g. Middle East and North Africa 40 40 Change in agriclutural prices (%) Change in agriclutural prices (%) 20 20 0 0 –20 –20 2000 2030 2050 2080 2000 2030 2050 2080 Source: Havlík et al., forthcoming. Note: Results are based on simulations from Global Biosphere Management Model (GLOBIOM). The figure shows climate shock impact on agricultural prices by comparing the price level in the different emission and development scenarios with the price level without climate change for each year. B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    55 even in a poverty scenario (figure 2.3). It can contribute by reducing direct non- However, in the high-emissions scenario, cal- CO2 emissions of existing agricultural pro- ories per capita grow until 2030 and then duction, and, even more important, by stagnate, remaining 7 to 10 percent lower avoiding deforestation, increasing carbon than in the no climate change world even sequestration through afforestation, and pro- when adaptation and trade possibilities are ducing biomass for energy generation. The accounted for. What is particularly worrisome IPCC estimates that land-based mitigation, about these simulations is that in Africa and including bioenergy, could contribute 20 to South Asia, the increases in daily calories pla- 60 percent of total cumulative abatement up teau at levels far below those of developed to 2030, and 15 to 40 percent up to 2100 countries today. In other regions and in the (Smith and Bustamante 2014). Distributing prosperity scenario, impacts are much smaller mitigation efforts in a cost-efficient way and appear more manageable because of the across sectors could require a decrease in much higher baseline levels. emissions from AFOLU by 64 percent in 2030, compared to their 2000 level (Havlík et al., forthcoming). Ill-designed land-based mitigation Yet such mitigation efforts could affect policies can also be a threat for food food prices and availability given the critical security role of land in agricultural production. The In the longer term, countries will need to IPCC concludes that large-scale, land-based look into well-designed land-based mitiga- mitigation at the global scale, especially bio- tion policies to stabilize climate change (see energy expansion, can reduce the availability chapter 6 for a discussion of mitigation of land for food production, with implica- needs and policies). Here, the agriculture, tions for food security (Smith and Bustamante forestry, and other land uses (AFOLU) sector 2014). Many studies show that regional will be a key player, in addition to fulfilling and local commodity prices (like food, tim- its already critical role in providing food and ber, and energy) could rise as a result (Chen supporting rural livelihoods. et al. 2011; Golub et al. 2013; Kuik 2013). FIGURE 2.3  Climate change can significantly reduce food availability in poor regions (Impact on daily calories per capita) a. Sub-Saharan Africa b. South Asia 90 90 Daily calories per capita availability Daily calories per capita availability relative to developed countries relative to developed countries 80 80 in 2015 (%) in 2015 (%) 70 70 60 60 2000 2030 2050 2080 2000 2030 2050 2080 No climate change Low emissions High emissions High emissions without CO 2 fertilization Source: Havlík et al., forthcoming. Note: Results are based on simulations from Global Biosphere Management Model (GLOBIOM). The figure shows the aggregate daily calories per capita available in the Poverty scenario per region as percentage of the projected level of daily calorie consumption in developed countries in 2015. Data on ­calorie levels in developed countries in 2015—estimated at 3,390 kilocalories—are from Alexandratos et al. (2012). 5 6    SHOCK WAVES Food price impacts could be severe for large- Poor people are vulnerable to scale bioenergy deployment—especially when combined with protecting natural ­ f orests climate impacts through prices (Calvin et al. 2013; Popp et al. 2011; Wise and ecosystems et al. 2009). How are the poor affected by climate-related The simulations carried out for this report impacts on food production and prices? show that ­ mitigation policies implemented Increased agricultural prices will affect con- through a uniform global carbon price, which sumers through higher expenditure for basic does not account for food production impli- goods, but will benefit net sellers and those cations, would hurt crop and livestock pro- who earn wages from agricultural employ- duction and result in lower food availability ment. However, the increase in prices may be compared to a hypothetical baseline without the result of a decline in productivity that ­ climate change and climate policies (Havlík would reduce returns from farm activities et al., forthcoming).3 Such policies could even and agricultural wages, so that the net effect have price impacts that are larger than those on sellers will depend on the interplay of climate change (figure 2.4). Other, more between prices and output. In addition, carefully designed, policies, however, could many poor people—especially those in rural lead to price impacts that are smaller than areas—do not buy or sell consumption those caused by unmitigated climate change goods in markets, but produce them from (Lotze-Campen et al. 2014). Nevertheless, ecosystems to meet subsistence needs (for these findings warn against climate policies example, nontimber products provided by that would not be sensitive to food security forests, production of crops, or small fishery issues, and emphasize the need to protect catches), which may be extremely climate poor people against the negative side effects sensitive. of land-mitigation policies. Poor people will be affected by higher FIGURE 2.4  Ill-designed land-mitigation climate policies could agricultural prices as consumers and sharply increase agricultural prices producers (Impact on agricultural prices) Because climate change and climate policies are likely to affect the prices of basic goods— Limited biomass particularly food—they will alter the pur- Climate policies chasing power and real income of households. All technologies However, the overall impact on poverty will High energy efficiency depend on both how much the prices of these basic goods are affected and whether the High emissions w/o CO2 fertilization households are net buyers or net sellers. Climate change High emissions Poor consumers are highly vulnerable to food price hikes. The higher share of income Low emissions that poor people spend on food makes them −20 0 20 40 60 80 particularly vulnerable to rising prices or Price increase (%) price volatility on food items. Across the 2080 2050 2030 developing world, the poorest households Source: Havlík et al., forthcoming. spend between 40 percent and 60 percent of Note: Results are based on simulations from Global Biosphere Management Model (GLOBIOM). The figure shows difference of agricultural prices between a baseline scenario (that is, no climate their income on food and beverages change and no climate policy) and different emissions and climate policy scenarios assuming a ­ compared to less than 25 percent of wealth- uniform global carbon price. “All Technologies” is a mitigation scenario in which all available tech- nologies enter into the solution portfolio according to their relative competitiveness. “High Energy ier households (figure 2.5). In some African Efficiency” involves the same technologies but assumes higher investments in energy efficiency countries, such as Burundi, Chad, the leading to final energy demand lower by 20 to 30 percent in 2050 and by 35 to 45 percent in 2100. “Limited Biomass” puts a limit on industrial biomass use for energy at 100 EJ/yr. All price impacts are Democratic Republic of Congo, Malawi, shown for the poverty scenario. Results are similar under the prosperity scenario. and Tanzania, food consumption of the B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    57 FIGURE 2.5  Poor households spend a higher share of their expenditure on food than nonpoor households (Expenditure for food and beverages) 70 60 Share of total expenditure (%) 50 40 30 20 10 0 Sub-Saharan Europe and South Asia Middle East and East Asia and Latin America and Africa Central Asia North Africa Pacific Caribbean Poorest Poor Middle Wealthier Source: World Bank Global Consumption Database. http://datatopics.worldbank.org/consumption/sector/Food-and-Beverages. Note: Calculated based on total consumption value in 2010 ($PPP [purchasing power parity] Values) in developing countries. Consumption groups defined based on global income distribution data: poorest = $2.97 per capita a day; poor = between $2.97 and $8.44 per capita a day; middle = between $8.44 and $23.03 per capita a day; wealthier = above $23.03 per capita a day. poorest households amounts to over 70 Price increases can benefit agricultural percent of their total expenditure. Poor producers and laborers. A rise in agricul- people in urban areas often have even tural prices is likely also to lead to an higher food expenditure than rural people, increase in agricultural profits and wages. as the latter can also self-produce some of The poverty impacts of price-induced their food needs. changes in earnings depend on how vulnera- How poor people are affected by a par- ble households make a living. In Sub-Saharan ticular food item’s price rise depends on their Africa, a recent study shows that almost ability to respond to price changes by modi- all rural households are self-employed in fying their diets toward cheaper foods. farm activities, earning 63 percent of their Overall, demand response for basic food income from these sources compared to only products has been found to be limited, espe- 33 ­percent in non-African developing coun- cially in the case of a generalized increase in tries (Davis, Di Giuseppe, and Zezza 2014). food price levels (Ivanic and Martin 2008), More generally, the share of poor people or for staple foods (such as tortillas in who declare agriculture as their main income Mexico) (Wood, Nelson, and Nogueira source is highest in poorer countries, though 2012). But poorer households are more with substantial variation (­ figure 2.6). In likely to reduce food consumption in the face India, agriculture remains the main preserve of higher prices, with a 10 percent increase of the unskilled and disadvantaged people in food price levels translating into a reduc- (Lanjouw and Murgai 2009). These house- tion in daily food intake by 301 kilojoules holds will be highly sensitive to any change (72 kilocalories) in low-income countries in agricultural profits or wages. (Green et al. 2013). Such impacts could lead In spite of the benefits for farmers who are to undernutrition, with potentially severe net sellers of food, existing studies using mul- health impacts, especially on children (see ticountry samples tend to agree that, in the chapter 4). absence of changes in production and wages, 5 8    SHOCK WAVES a rise in food prices increases poverty rates in rise to a 5.8 percentage point poverty increase; most countries due to the negative impacts on and a 100 percent price rise to a 13 percent- consumers. Simulations suggest that a 10 per- age point poverty increase (Ivanic and Martin cent price rise in food prices—with no change 2014). The severity of poverty impacts will in agriculture productivity—leads on average vary among countries—with Guatemala, to a 0.8 percentage point increase in extreme Pakistan, Sri Lanka, Tajikistan, and Yemen poverty headcount rates; a 50 percent price being the worst hit (figure 2.7). Accordingly, food price spikes in the past already had significant poverty impacts. In a FIGURE 2.6  In poorer countries, agriculture plays a vital role for sample of 28 developing countries the global poorer households’ incomes (Agricultural participation of the bottom 20% and country GDP) price spikes between June and December 2010, which increased food prices by an aver- age 37 percent, increased the number of peo- 70 ple in extreme poverty by 44 million (Ivanic, Share of bottom 20% in agriculture (%) 60 Martin, and Zaman 2012). The 2008 food price shock, resulting in price increases of 50 over 100 percent, even resulted in an addi- 40 tional 100 million people in extreme poverty (Ivanic and Martin 2008). 30 Poverty impacts will also depend on the 20 specific food crops for which prices increase. A 10 percent price increase for rice would 10 increase poverty in Bangladesh by 0.67 0 percentage points and in Côte d’Ivoire by ­ 0 2,000 4,000 6,000 8,000 10,000 0.42 percentage points, but reduce poverty Income per capita (2007 US$ PPP) by 1.37 percentage points in Cambodia and 0.29 percentage points in Vietnam (Ivanic, Source: World Bank calculation based on the International Income Distribution Data Set (I2D2). Martin, and Zaman 2012). Similarly, as seen Note: Values based on the number of people indicating agriculture as their main income-earning activity (including farm self-employment and wage-employment). PPP = purchasing power parity. ­ gure 2.7, a 10 percent price increase in all in fi crops could lower poverty rates in Albania, FIGURE 2.7  Food price rises could lead to big increases in extreme poverty in most countries (Poverty change under 10%, 50%, and 100% food price increases) 35 Change in poverty rate (%) 25 15 5 0 5 –15 . P a ov a Rw ma Be a N i da Za ria V i bi a Ta nam ia d a or e s h ra a si a Pa ala Sr stan en a ji k n ep at ia Ug e ng ire e do ne Ch a A l i na A r ni a Ma ia Ec eru Mo dor Cô cara r Ne i P l d te gu m nk Mo est liz er goli Ni ge law di Ta ista pa an Gu Ind en an an ,R Ba ’Ivo In Leo ge ne em na ba m bo Tim lad Ye i La ld ua Ni -L nz i et m k n m Ca Si 10% 50% 100% Source: Ivanic and Martin 2014. Note: Based on microeconomic simulations specified with the Global Trade Analysis Project (GTAP) general equilibrium model. Simulations measure the short-term impacts on pov- erty (without any supply or wage adjustments) of a uniform change in all food prices for 10, 50, and 100 percent (without productivity effects). Poverty is defined by the percentage of households living on less than $1.25 per day. B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    59 Cambodia, China, and Vietnam, while more wealthier households, and by 20 percent in extreme price shocks would have a poorer ones (Hill and Mejia-Mantilla 2015 ­ poverty-reducing impact only in Cambodia and figure 2.8). Interestingly, however, and China. Modest price increases could lift a the average consumption decrease was the group of net-selling farmers out of poverty, same in both sets of households—about while larger price rises would drop other 4 ­percent—suggesting the poorer ones made groups into poverty. significant efforts to preserve consumption. In the longer run, wage and production Nevertheless, even small shocks can also adjustments are possible, but the net impact push vulnerable households below the pov- depends on productivity impacts. In the lon- erty line. ger run, rising food prices will increase mar- In addition, chapter 4 discusses the impact ginal returns from agriculture, which raises of high temperature on labor productivity. agricultural wages. While the effects of com- These effects are not accounted for in the modity price changes on rural wages could agriculture models that are used in this take some time to materialize, several studies ­ chapter, but they could have significant effects have actually found that these adjustments on returns from all outdoor occupations, and could mediate poverty impacts of increased especially agriculture. consumption prices (Devarajan et al. 2013; The net effect of climate change on pov- Ivanic and Martin 2014; Jacoby, Rabassa, erty, which is the combination of impacts on and Skoufias 2014). The wage response productivity, consumption prices, and depends on the elasticity of the agricultural incomes, is likely to be negative in many wage with respect to land productivity. The countries. In the 15-country sample used by more limited the supply of labor to agricul- Hertel, Burke, and Lobell (2010), climate- ture, the more responsive the agricultural induced price rises increase extreme poverty wage to price shock, as seen in India (Jacoby, by 1.8 percentage points—driven mostly by Rabassa, and Skoufias 2014). Such responses the negative impacts in countries with large may be observed if workers cannot easily populations (including Bangladesh and coun- move between sectors and if spatial mobility tries in Sub-Saharan Africa, such as Malawi, is limited. Mozambique, Uganda, and Zambia). Moreover, in the longer run, farmers respond to changing prices by adjusting their production. In some cases, they can switch land toward producing those items whose FIGURE 2.8  Rainfall shocks in Uganda take a big prices have risen relative to others or by toll on crop income, less so on consumption (Impact of rainfall shock on income and consumption) increasing overall agricultural production through expansion of agricultural land or increases in other outputs. With these long- 20 run adjustments and no change in productiv- 15 Reduction (%) ity, extreme poverty could even decrease globally by as much as 1.4 percentage points 10 under a 10 percent price rise and 8.7 percent- 5 age points under a 100 percent price rise (Ivanic and Martin 2014). However, these 0 Bottom 40% Top 60% options are not always available to farmers. household income household income And farmers are also directly affected by Crop income Consumption changes in land and labor productivity. For instance, data from Uganda between 2005 Source: Hill and Mejia-Mantilla 2015. and 2011 suggest that a 10 percent reduction Note: Values calculated using data from Uganda National Household Survey UNHS 2005/6, UNHS 2009/10, UNHS 2012/13. Rainfall shock is rep- in water availability due to a lack of rainfall resented by a 10 percent decrease in the Water Requirement Satisfaction reduced crop income by 14.5 percent in Index, estimated for Uganda. 6 0    SHOCK WAVES Climate change modifies not only average The analysis in chapter 6 also suggests climate conditions but also climate variabil- that—even in the long run—the negative ity, with potential effects on poverty. One impacts through consumption prices and study estimates that by 2080, an increase in yields dominate the positive impacts on agri- the intensity of extreme dry events would lead cultural incomes in almost all scenarios, and to a rise of extreme poverty by 0.53 percent- the net effect of climate change is very likely age points in 16 developing countries with to increase global poverty. A net decrease in Bangladesh (1.35 percentage points), Mexico poverty is not impossible, though, if many (1.76 percentage points), and Zambia (4.64 poor households stay in agriculture (the case percentage points), most severely affected in our poverty scenario) and if the impact of (Ahmed, Diffenbaugh, and Hertel 2009). climate change remains moderate. That said, Other significant effects would affect poor it occurs only in a small fraction of the farmers through impacts on livestock, which ­ scenarios—and only if institutions and labor are not only a source of income (box 2.2) but markets are such that the extra revenue from also used as an asset by poor households. higher prices is distributed fairly across work- Nonagricultural households, which are net ers and landowners. These results parallel buyers of food—especially those in urban those of a study on India, in which higher areas—will be the worst affected. A climate- agricultural wages help poor households and, induced rise in food prices could increase pov- if the negative impacts of climate change on erty rates of nonagricultural households by agricultural productivity impacts remain low, 20 to 50 percent in parts of Africa and Asia could even benefit them (Jacoby, Rabassa, (Hertel, Burke, and Lobell 2010). Similarly, a and Skoufias 2014). But with greater produc- once-in-30-year climate extreme would most tivity impacts, climate change will be increas- severely affect urban laborers with a dramatic ingly likely to contribute to a global increase increase in poverty rates within this group in in poverty. Bangladesh (31 percent), Malawi (111 per- cent), Mexico (95 percent), the Philippines The poorest will be directly affected by (32 percent), and Zambia (102 percent) impacts on ecosystems (Ahmed, Diffenbaugh, and Hertel 2009). With increasing urbanization rates, food price Besides higher agricultural prices, the poor- increases could have an even more severe est people will also be directly affected by poverty impact. climate change through its impacts on BOX 2.2  Climate-driven livestock diseases can have high economic costs Livestock plays a vital role in the economies of In Somalia, Rift Valley fever epidemics prevented many developing countries. Globally it accounts 8.2 million small ruminants, 110,000 camels, and for 40 ­ percent of agricultural production, employs 57,000 cattle from being exported, corresponding 1.3 ­b illion people, and creates livelihoods for to economic losses for the livestock industry esti- 1 ­ billion of the world’s poor. Livestock products pro- mated at $109 million in 1998–99 and $326 mil- vide one-third of human protein intake. lion in 2000–02. In Kenya, the annual cost of East Climate change affects many of the environmen- Coast fever is estimated at $88.6 million; in Malawi, tal variables that can lead to livestock diseases—and $2.6 million; in Tanzania, $133.9 million; and in climate-sensitive livestock diseases already have Zambia, $8.8 million. high economic impacts (via income losses), as well as costs to prevent and control disease outbreaks. Source: Bouley and Planté 2014. B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    61 livelihood activities that fulfill subsistence poverty line—most of them in Sub-Saharan and other needs. In the absence of func- Africa (Noack et al., forthcoming). Out of tioning capital, labor, and land markets the 424 million people living in the African and with very few assets on hand, many drylands, 23 percent (or 40 ­ percent of the poor rural households depend on access 240 million agriculture-dependent individu- to ecosystems (Barbier 2010). They use als) are estimated to be below the extreme them to produce or extract goods for self-­ poverty line (Cervigni and Morris 2015). consumption (like crops, timber, and fish) In these areas, ecosystem-based activities and to smooth income shocks. Climate provide poor people with incomes from inten- change can add to the stress on ecosystems sive ecosystem management (like crop cultiva- and reduce their ability to support tion and livestock) and from the extraction of livelihoods. noncultivated ecosystem goods (like timber, Many poor rural people depend on plants, animals, and fish). These ecosystem- ­ e cosystems for their base income and as based incomes made up 55 to 75 percent of safety nets. Many of them live in low-­ incomes in a cross-section of 58 sites repre- productivity and fragile ecosystems, making senting smallholder systems, with 15 to them highly vulnerable to climate risks, 32 percent coming from forests or other non- including natural disasters (see chapter 3). cultivated ecosystems (figure 2.9).4 Among Data collected for a global comparative study these typically poor households, the wealthier of tropical and subtropical smallholder households derive a slightly lower share of ­ systems show that about 27 percent of the their income from ecosystems and have a households included fall below the extreme lower dependence on subsistence-based FIGURE 2.9  Ecosystem-based incomes explain most rural income in (sub)tropical smallholder systems (Income shares across income quintiles and regions) 80 70 60 Share of total income ( %) 50 40 30 20 10 0 Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5 Latin America South Asia East Asia Sub-Saharan Africa Livestock income—Cash Livestock income—Subsistence Crop income—Cash Crop income—Subsistence Forest income—Cash Forest income—Subsistence Other environmental income—Cash Other environmental income—Subsistence Source: Noack et al., forthcoming. Note: Figure shows average share for households across income quintiles. Q1 is the lowest income quintile and Q5 is the highest. Income calculated as PPP (purchasing power parity) 2005 USD adult equivalent units. Based on Poverty and Environment Network (PEN) dataset, including data from 58 sites in 24 countries. 6 2    SHOCK WAVES incomes. Similar to findings in Damania et al. households that have little other means to (2015), ecosystems boost incomes of wealth- secure subsistence needs. ier households, for example through cash These subsistence incomes from ecosystems crops or work on commercial plantations, help to keep a considerable share of rural peo- and provide last resort incomes for poor ple above the poverty line. If smallholding households in (sub)tropical forest landscapes could not complement their income through FIGURE 2.10  Without environmental incomes poverty rates could forests and other environmental resources and be much higher were unable to find alternative income (Poverty rate in [sub]tropical smallholder systems) sources, an additional 14 percent of all sam- pled households would fall below the extreme 60 poverty line (figure 2.10). Besides providing 50 goods that directly contribute to cash or sub- sistence incomes, ecosystems can also indi- 40 rectly benefit poor people (box 2.3). Poverty rate (%) Poor households have strategies to deal 30 with climate variability. During the survey 20 years, the poorest households in (sub)tropical smallholder systems were not subject to more 10 weather or other livelihood shocks (such as income, asset, or labor losses), despite living 0 in more extreme climatic conditions (Angelsen Latin South Asia East Asia Sub-Saharan Total America Africa and Dokken, forthcoming). In addition, only With environmental income Without environmental income a few of those households that were exposed to weather anomalies or shocks experienced a Source: Noack et al., forthcoming. decline in income. As the survey years did not Note: Figure shows share of sampled households below the $1.25 per day poverty line based on incomes calculated as PPP (purchasing power parity) 2005 USD adult equivalent units. Based on cover any major extreme events, this finding Poverty and Environment Network (PEN) dataset, including data from 58 sites in 24 countries. suggests that rural households can manage at BOX 2.3  The wider functions of ecosystems and biodiversity in rural livelihoods Ecosystems and biodiversity can support humans sheds support hydrological services that benefit local through nonprovisioning services (such as regulating, water supply for domestic use or agricultural activi- supporting, and cultural services) that are of critical ties (Klemick 2011; Pattanayak 2004). importance for poor people in rural areas (MEA 2005). In addition, ecosystems provide an important role Studies that assess the links between such nonprovi- in protecting livelihoods against climate risks. Trees sioning ecosystem services and poverty have been very on steep slopes protect rural villages from landslides rare (Roe et al. 2014; Suich, Howe, and Mace 2015). when heavy rains fall, and mangroves provide pro- Ecosystem services can support incomes and tection to coastal livelihoods during storm surges livelihood activities in many indirect and often (Badola and Hussain 2005; Das and Vincent 2009). invisible ways. For instance, the establishment of a Moreover, ecosystems can contribute to climate sta- protected area in Costa Rica has increased tourism- bilization. For example, higher forest cover helps based incomes and accounted for two-thirds of the reduce the occurrence of droughts (Bagley et al. poverty reduction achieved in the area (Ferraro and 2013; Davidson et al. 2012) Hanauer 2014). Forests provide pollination services A new research agenda is emerging to better that benefit nearby farming activities (Olschewski understand the services provided by ecosystems and et al. 2006; Ricketts et al. 2004). And intact water- who benefits from them (Bennett et al. 2012). B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    63 least modest weather and livelihood shocks et al. 2015). Unfortunately, such responses, without falling deeper into poverty. as well as the general overextraction and One way to smooth income volatility is by overuse of ecosystems resources, can lead to extracting ecosystem resources. Incomes from increasing degradation of ecosystems, resource stocks that grow continuously over which undermines the sustainability of such years (like timber and fish) are less sensitive to strategies. weather fluctuations than those that depend Climate change adds to the stress on eco- on annual cycles (like crops). Hence forest systems and undermines rural livelihoods. resources can help smooth consumption Despite the importance of ecosystems for between seasons and years (Locatelli, poor people, the link between climate- Pramova, and Russell 2012). Although envi- induced ecosystem changes and poverty has ronmental extraction was the primary coping not been systematically assessed. This is due strategy for only one out of 12 households in in part to the difficulty in disentangling eco- (sub)tropical smallholder systems, forest system changes driven by climate change incomes are a substitute for agricultural from those driven by other factors. The incomes and can stabilize total income existing peer-reviewed literature suggests when weather anomalies hit (Noack et al., that the ecosystem-related impacts of climate forthcoming). change will affect poor people mostly As rural households are used to dealing through hazard regulation, soil and water with climate variability, weather anomalies regulation in low elevation coastal zones, are probably not perceived as shocks that and dryland margins (Howe et al. 2013). In require emergency responses, but as tacit many regions, there are a number of risks changes in production conditions that that are relevant for the rural poor—such as require marginal adjustments of existing local warming in the Sub-Saharan drylands, strategies. Thus, in India, households that which can trigger shorter growing seasons have lower than predicted income, indicat- and shifts in areas suitable for rain-fed ing a bad year, derive a higher income share ­agriculture (table 2.1). Moreover, although from environmental resources (Damania current knowledge does not allow for a TABLE 2.1  Climate change risks for ecosystems and potential livelihood impacts across regions Examples of long-term climate Potential livelihood impacts of relevance Region Ecosystem change risks for poor people Sub-Saharan Drylands local warming, amplified by dry shorter growing season and shift in areas Africa conditions leads to expansion of suitable for rain-fed agriculture will have arid areas in Southern Africa and negative impacts on farmers Western Africa Grasslands higher CO2 concentration make trees reduced availability of food for grazing animals better grow in savanna areas reducing with negative impacts for pastoralists grasslands Forests more extreme temperature and rainfall limited availability of timber resources for conditions increase tree mortality in forest communities evergreen forests and woodlands Freshwater depletion of freshwater resources and decrease area for flood recession agriculture, wetlands grazing for livestock and availability of freshwater fish for rural communities Coastal sea-level rise increases coastal flooding loss of land for coastal communities, increased salinization Oceans warming and ocean acidification lead to decrease in catch potential for coastal bleaching of coral reefs and changes in communities and fishery jobs fish species distribution table continues next page 6 4    SHOCK WAVES TABLE 2.1 (continued) Examples of long-term climate Potential livelihood impacts of relevance Region Ecosystem change risks for poor people Middle East & Drylands warmer and drier climate shifts increased water stress affecting livestock and Northern Africa vegetation to the north and triggers crop production with negative impacts for desertification and soil salinization smallholding farmers and herders Coastal sea-level rise accelerates salinization damage to crop production with negative of groundwater in the Nile Delta impacts of smallholder farmers Europe & Forest shift of boreal and temperate forests with decline in timber harvest endangers jobs of Central Asia heat waves, water stress, forest fires, and workers in the forestry sector tree mortality in boreal forests in Russia Drylands expansion of arid areas and more loss of area for rain-fed crop production and frequent and intense droughts pressure on livestock with negative impacts increase desertification for smaller agricultural producers Mountains retreat of glaciers causing increased increased water stress affecting irrigated seasonal water variability and a significant agriculture with negative impacts for workers water shortages in the long run on commercial farms East Asia & Freshwater rising temperatures, salinity intrusion, damage to fish production with negative Pacific and increasing tropical cyclone intensity impacts for aquaculture farmers exceeds tolerance for farmed fish species Coastal sea-level rise and increased tropical loss of land for rice production in the Mekong cyclone intensity leading to increased Delta affecting farmers, reduced protection coastal erosion and saltwater intrusion of coastal settlements and limited availability in river deltas and other low-lying areas, of mangrove forest resources for coastal and loss of mangroves communities Oceans warming and ocean acidification lead reduction in fish catch potential with to bleaching of coral reefs and changes negative impacts for fishery jobs and coastal in fish species distribution communities South Asia Mountains glaciers loss in the Himalayas and Hindu reduced food production within river basins Kush causes increased seasonal variability with negative impacts for smallholder farmers of water flows in glacier-fed river systems Terrestrial changes in monsoonal precipitation increased water stress affecting crop lands increase river floods, number of dry days, production with negative impacts for severity of droughts, and reduction of smallholder farmers groundwater resources Coastal sea-level rise and increasingly intense reduction in land that can be used for tropical cyclones accelerate salinity agriculture with negative impacts for farmers intrusion in deltaic regions and wetlands Latin America Drylands expansion of dryland areas and more increased water stress with negative impacts & the Caribbean extreme drought periods in Mexican on local water supply and livestock and crop dry subtropics and northeastern Brazil production affecting smallholder farmers Forest dry season length, extreme drought, reduced availability of forest resources for and forest fire cause tropical forest indigenous people and other forest dwellers degradation and forest-fringe communities Mountains retreat of glaciers and changing negative impacts on water supply and crop snowfalls cause increased seasonal production affecting smallholder farmers and variability of river stream flows indigenous communities in the Andes Coastal/Islands sea-level rise, storm surge and tropical loss of land and damages of tourism-related cyclones affecting small island states activities with negative impacts for coastal and low-lying coastal zones and island communities Oceans ocean acidification and warming lead to decrease in catch potential in most waters coral bleaching in the Caribbean and fish with negative impacts for local fishermen species shift toward higher latitudes Source: Based on World Bank 2014b and World Bank 2014c. B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    65 quantification of these impacts, studies sug- rural people are the most vulnerable because gest severe livelihood impacts: of their reliance on traditional systems built on ecosystems resources and their exclusion • Climate-related events have been shown from formal policy making (Hoffman and to threaten livelihoods of poor people in a Grigera 2013; Kronik and Verner 2010). In variety of rural contexts—such as precipi- the high Andes, indigenous farmers face tation variability in the Peruvian Andes institutional marginalization and land scar- (Sietz, Choque, and Lüdeke 2011); floods city compounded by delayed rainfalls, which in Senegal (Tschakert 2007); drought in lead to disputes over access to water and the West African Sahel (Sissoko et al. land (McDowell and Hess 2012). Similarly, 2010) and in Northwest China (Li et al. after the 1998 floods, herders in northern 2013); and cyclone-related saltwater Kenya tried to recover herd losses by raiding intrusion in coastal Bangladesh (Rabbani, neighboring farms (Little, Mahmoud, and Rahman, and Mainuddin 2013). Coppock 2001). • Poor people can also be indirectly affected when climate change affects ecosystems that support their livelihoods, such as Policies can avoid negative coastal and near-shore habitats (like wet- lands, mangroves, coral and oyster reefs, consumption effects and and sea grasses) (Barbier, forthcoming). increase incomes In Bangladesh, increased salinity linked While climate change is quite likely to hurt to sea level rise reduced the suitability agricultural yields and raise food prices, of land for rice farming (Dasgupta et al. resulting negative impacts on poor people 2014). In one Bangladeshi site, 70 percent can be offset, at least partially, through of farmers partially or fully abandoned socioeconomic development, better infra- agriculture because of saline soils over a structure and markets, improved farm prac- period of 10 years (Shameem, Momtaz, tices and technological progress, and the and Rauscher 2014). preservation and strengthening of ecosys- • In the most extreme form of livelihood tems. As for land-based mitigation policies, impacts, climate change could make eco- they can be designed to be pro-poor. systems completely inhabitable, forcing out inhabitants—notably in small island states, some of which are at risk of dis- Negative impacts on food security appearing before the end of the century can be reduced by development and (Burkett 2011). The low-lying Sunder- poverty reduction bans, a coastal area between India and Food insecurity will be determined not only Bangladesh, are becoming a more dif- by yield declines but also by the socioeco- ficult place to live for its mostly poor nomic conditions that make countries better population, increasingly exposed to sea prepared to respond to and adapt to such level rise, salinization of soil and water, declines. Biewald et al. (Forthcoming) cre- cyclonic storms, and flooding (World ated a hunger vulnerability index, based on Bank 2014a). the Global Hunger Index (GHI) from von Conflict over environmental resources Grebmer et al. (2011).5 This index represents may exacerbate existing vulnerabilities of in particular the fact that, as people get poor ­p eople. In Latin America, climate richer, they have better access to food change and environmental degradation, markets and can buy their food instead of ­ along with the rapid growth of mining, could producing it. By 2030, development and lead to greater competition for land and increasing incomes reduce vulnerability to water resources. Smallholding farmers, hunger in both the prosperity and poverty indigenous communities, and other poor scenarios, with much greater progress in 6 6    SHOCK WAVES the  prosperity scenario, thanks to its rapid limited yield declines until 2030, while the reduction in poverty. opposite is true in the Middle East and North Socioeconomic vulnerability to hunger Africa. In India, where yield declines are mod- combined with climate-induced yield declines est, the socioeconomic vulnerability to hunger create potential food insecurity by 2030, is serious in the poverty scenario, but especially in the poverty scenario. Map 2.1 decreases in parts of the country to moderate shows where yields are expected to decrease, levels in the prosperity scenario. suggesting a reduced ability to produce The implication then is that development locally, and where a high level of poverty may and poverty reduction can prevent some of make it challenging to rely on food markets the worst impacts of climate change on food and imports. Most of the hotspots are in Sub- security but do not replace targeted interven- Saharan Africa. For example, most of tions to increase the resilience of the food pro- Madagascar, Sudan, and Yemen suffer alarm- duction and distribution system. ingly high levels of yield decline and socioeco- nomic vulnerability to hunger in the poverty Better infrastructure and market access scenario. Angola and the Democratic help cope with production shocks Republic of Congo experience high socioeco- nomic vulnerability to hunger but only As observed during the 2010 food price cri- sis, countries where global price increases are matched by high rises in local prices also experienced higher poverty increases (Ivanic, MAP 2.1  Risks to food security would be much reduced in a more Martin, and Zaman 2012). Countries can prosperous future insulate domestic markets from global prices by reducing import protection or increasing export restraints. But by doing so they both increase average domestic food prices and contribute to global volatility—for example, export bans, such as those imposed by a number of large exporters in 2008, reduced availability on international markets and contributed to higher prices (Anderson, Ivanic, and Martin 2013). Well-functioning markets can help coun- tries cope with production shocks, although the ability to rely on markets depends on many socioeconomic conditions—especially institutional barriers (like trade barriers) and transportation costs. Rural road development offers a strong potential to lower transport costs and spur market activity. In Ethiopia the incidence of poverty decreased by 6.7 percent after farmers gained access to all-weather Source: Biewald et al., forthcoming. Note: Risk to food security is indicated by the combination of potential yield declines and hunger roads (Dercon et al. 2009). vulnerability in a Poverty (upper panel) and Prosperity scenario (lower panel). Lighter colors indicate A productivity shock at the local level can lower risks and darker colors represent higher risks. Results for yield decline are based on the Lund- Potsdam-Jena managed Land (LPJmL) model and results for hunger vulnerability are based on the lead to much greater price fluctuations if local spatial Vulnerability-to-Hunger Index (VHI) adapted from von Grebmer et al. 2011. Yield decline markets are isolated. For example, a recent results compare a scenario with climate change to a scenario without climate change. The legend values are as follows: EI = Extremely high impact: decline of 10 percent or more, HI = High Impact: study examines the statistical effect of road −10 to −2 percent, SI = Strong Impact: −2 to 0 percent; NNI = no negative impact: >0 percent. quality and distance from urban consumption The Hunger Vulnerability index ranks spatial units on a 100-point scale. Zero is the best score (no hunger), and 100 is the worst: EA = Extremely alarming: ≥30, A = Alarming: 20 to 29.9, S = Serious: centers on maize price volatility in Burkina Faso 10 to 19.9, LM = Low to moderate: below 9.9. (Ndiaye, Maitre d’Hôtel, and Le Cotty 2015). B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    67 It finds that maize price volatility is greatest in increases on consumers, like food stamps remote markets, suggesting that enhancing (Larson et al. 2013). road infrastructure would strengthen the links Thus, one solution appears to lie in coordi- between rural markets and major consump- nating measures aimed at limiting price vola- tion centers, thereby also stabilizing maize tility with actions to make farm practices prices in the region. more resilient. Together, these actions can sig- In addition, between 2006 and 2008, low nificantly mitigate the impact of a shock, as food stocks may have exacerbated food price occurred in Bangladesh in 1998 (box 2.4). volatility (Gilbert and Morgan 2010), Countries can also implement social protec- although an empirical analysis only confirms tion schemes to ensure that vulnerable people this effect for wheat since 2000. The absence are protected against food price volatility of impact for other crops suggests that food (see chapter 5). stocks alone do not have a direct impact on price volatility but can amplify the effects of Improved farm practices and other factors (Tadesse et al. 2014). But while technologies can mediate negative adequate food stocks can help to reduce price impacts volatility and food insecurity, they can be costly and difficult to manage. And building Climate-smart agricultural practices can food stocks can lead to increased grain scar- increase productivity and make agricultural city and thus even higher prices in the short production more resilient. In countries most run. In the case of a large importing region— exposed to climate variability and change, the Middle East and North Africa—one study disaster preparedness and resilient and shows that a strategic storage policy at the diverse farming systems go hand in hand regional level could smooth global prices but (World Bank 2011). For instance, Vietnam is is much more costly than a social protection improving its water resource management to policy that dampens the effects of price make its cropping and aquaculture regimes BOX 2.4  Mitigating losses from the 1998 flood in Bangladesh The large-scale floods that hit Bangladesh in 1998 In addition to infrastructure, government policies had the potential to cause a major food security encouraged private-sector participation in the grain disaster, but short-term and long-term policies played market. In the early 1990s, the liberalization of rice key roles in preventing such a disaster. Starting in and wheat imports enabled private sector imports the 1980s, public sector investment in agricultural to quickly supply domestic markets and stabilize research and extension, combined with private sec- prices following the floods. Other policies—such as tor investments in small-scale irrigation, substantially the removal of the import tariff on rice in 1998 and increased yields of wheat and boro rice. These invest- better port clearance of private sector food grain ments reduced vulnerability to floods by increasing imports—also provided clear signals of government total food grain production, reducing the length of support for the private grain trade. These private time between major crops from 12 months to 6, sector imports proved a far less costly way of main- and shifting away from flood-susceptible cultivation taining food grain availability than the distribution practices. of government commercial imports or public stocks. Furthermore, long-term investments in public And the inclusion of the private sector in general infrastructure (including roads, bridges, electricity, greatly increased food supply and stabilized food and telecommunications) made agricultural markets prices in the aftermath of the shock. more efficient and enabled traded grains to reach markets throughout the country after the floods. Source: Adapted from del Ninno et al. 2001. 6 8    SHOCK WAVES better adapted to increasing flood risk and Although many studies have shown that salinity levels. But more productive and technological progress can limit climate more resilient practices require a more effi- change impacts on yields and food produc- cient use of land, water, soil nutrients, and tion costs (see f ­igure 2.12), they usually genetic resources (FAO 2013). assume sustained and very rapid yield gain, Better technologies will also be needed to at odds with past yield trends. Empirical evi- tackle future food security challenges (FAO, dence reveals plateaus or even abrupt IFAD, and WFP 2014). These might include decreases in the rate of yield gain—for rice improvements in crop varieties, smarter use in eastern Asia and wheat in northwestern of inputs, methods to strengthen crop resis- Europe (Grassini, Eskridge, and Cassman tance to pests and diseases, and reduction of 2013). Further, another study points to a postharvest losses (Beddington 2010; Tilman general decrease in the growth rate of yields et al. 2011). Improved crops and better use of for maize, wheat, rice, and soybeans at a water and soil can increase both farmers’ global level, although with more optimistic incomes and their resilience to climate shocks trends in poorer countries (Alston, Beddow, (figure 2.11). and Pardey 2010). The low and declining One key way to make agricultural systems levels of investment in agricultural research more climate resilient is by developing and and development in the developing world adopting higher yielding and more climate- can be a major constraint to realize further resistant crop varieties and livestock breeds yield gains in poor countries (Pardey, Alston, (Tester and Langridge 2010). In a random- and Chan-Kang 2013; Pardey and Pingali ized control trial in Orissa, India, a recent 2010). study shows the benefits of using a new, flood-resistant variety of rice, which offers a 45 percent yield gain relative to the current FIGURE 2.12  Faster technological progress most popular variety (de Janvry 2015). After would dampen long-term increases in food a first good experience, these varieties moti- production costs vated farmers to take more risks and adopt (Food production costs under varying assumptions regarding technological progress) more profitable techniques. But the overall potential of technology 250 to increase resilience remains uncertain. 200 FIGURE 2.11  Improved cropping technologies Difference (%) increase resilience in the African drylands 150 (Households made resilient) 100 50 50 45 40 0 Share of households (%) 35 –50 30 Sub-Saharan Middle East South Asia 25 Africa and North Africa 20 Slow Medium Fast 15 Source: Biewald et al., forthcoming. 10 Note: The values show the regional average changes in food production 5 costs for Poverty and High Emissions Scenario without CO2 fertilization 0 compared to a no climate change scenario for the year 2030 based on Soil Drought- Agro- Heat tolerant results from the Model of Agricultural Production and its Impact on the fertility tolerant forestry crops Environment (MAgPIE). Food production costs used in MAgPIE account management crops (low density) for all costs, including infrastructure (irrigation systems) and research, and thus can differ from the prices faced by consumers. Slow, medium, and fast in the legend refer to the speed of technological progress, which is Source: Cervigni and Morris 2015. assumed to translate directly into increase in yields. B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    69 Moreover, disseminating improved tech- in the Amazon, self-amplifying feedbacks nologies and making them accessible to poor between reduced forest cover and extreme farmers will be critical for the gains from such droughts resulting from a combination of technologies to materialize. Adoption of new global warming and forest cover loss puts technological packages is often slow and lim- the forest at risk for large-scale dieback ited (box 2.5). Such technologies can be costly (World Bank 2014c). Similarly, fishing- or difficult to access. For instance, in Africa, dependent countries in South and East Asia fertilizer application remains low because of (such as Bangladesh, Cambodia, India, high transport costs and poor distribution Indonesia, the Philippines Sri Lanka, and systems (Gilbert 2012). Furthermore, cultural Vietnam), where fishing activities are poorly barriers, lack of information and education, regulated, are very vulnerable to the com- and implementation costs need to be over- bined impact of climate change and overex- come. Agricultural extension services can ploitation (Barange et al. 2014). help to make better use of new technologies. This interdependence means that reduc- In Uganda, extension visits coupled with the ing nonclimate stresses on ecosystems can introduction of new crop varieties increased make them better able to adapt to climate household agricultural income by around change and continue to support liveli- 16 percent (Hill and Mejia-Mantilla 2015). hoods. Consequently, any measures that Secure tenure rights, smart subsidies, and reduce or avoid land or forest degradation, access to long-term finance can also provide depletion of natural resource stocks (such farmers with incentives to adopt climate- as fish), or pollution of water and soils can smart technologies and practices (World protect the ecosystems poor people depend Bank 2012). upon and increase their resilience to cli- mate change. Targeted measures to foster ecosystem- Conservation and ecosystem-based based adaptation are a critical way to help adaptation increase the resilience of ecosystems and poor people better prepare ecosystems for climate change. They seek to strengthen The vulnerability of ecosystems to climate ecosystem processes and services, as well as change impacts will also depend on noncli- the human systems that maintain them, in mate human-made impacts. For example, order to make them more resilient to climate BOX 2.5  Despite significant benefits, adoption rates of conservation agriculture remain limited Technological packages, which have been successfully Nevertheless, despite these benefits, adoption is tested in demonstration fields, are available to help farm- still uneven. In Brazil, millions of hectares are cul- ers better prepare for climate change. For instance, con- tivated with conservation agriculture (Triplett and servation agriculture has been promoted to address poor Dick 2008). But in Morocco, after two decades of agricultural productivity and environmental degrada- demonstration, only 5,000 hectares are cultivated tion, particularly in semi-arid areas that are characterized with these techniques—almost exclusively on large by frequent droughts and dry spells (Giller et al. 2009; farms (ICARDA 2012). Low adoption rates are Kassam, Derpsch, and Friedrich 2014). Because of its the case for most countries in the Middle East and potential to increase yields, reduce labor requirements, North Africa and Sub-Saharan Africa, where the and improve soil fertility, it could be a powerful adap- available evidence suggests virtually no uptake of tation strategy (Kassam, Derpsch, and Friedrich 2014). conservation agriculture (Giller et al. 2009). 7 0    SHOCK WAVES stress and other environmental degradation. silvopastoral systems) (Harvey et al. 2014; Such approaches include better protection FAO 2013; World Bank 2011). For exam- and management of natural habitat or vegeta- ple, the inclusion of trees in the farming sys- tion, such as restoring and protecting tem dramatically increases the potential to ­ mangroves and dunes in coastal areas; man- store carbon, while increasing yields—as agement of flood plains in larger river basins; shown in Africa, where farmers who have managing forests sustainably through selec- adopted evergreen agriculture are reaping tive logging, forest buffers, and fire preven- impressive productivity gains of up to 30 tion; and farming systems that integrate percent without the use of costly fertilizer natural vegetation through fallow systems or (ICRAF 2012). In Ethiopia, the Humbo agroforestry (McKinnon and Hickey 2009). Assisted Natural Regeneration Project has Many of these strategies also provide carbon helped restore 2,700 hectares of biodiverse sequestration benefits. native forest, which has boosted carbon sequestration benefits and income genera- tion based on forest products.7 Improved Land-mitigation policies can be tree coverage also reduced drought vulnera- designed to benefit local incomes bility (figure 2.13). Land-mitigation policies can be designed to Some land-mitigation options can be avoid—or at least minimize—harmful implemented through payments for ecosys- impacts on agricultural production and food tem services, which compensate land users security and poor people. Although calibrat- for any forgone production benefits and pro- ing general mitigation policies to local con- vide them with financial incentives for pre- texts or introducing complementary serving or increasing carbon stocks in soil or measures may increase the overall cost of forests. More than 300 such payment mitigation (possibly because more efforts schemes have been established worldwide to would be required in nonagricultural sec- support carbon sequestration, biodiversity, tors), it could actually bring significant ben- watershed services, and landscape beauty efits in terms of improved local livelihoods (NCE 2014). and ecosystems. Even if trade-offs are likely to occur Careful land use planning—such as using between social goals and project efficiency, designated degraded or less-productive areas many programs target poorer land users. for storing and sequestrating carbon stocks— could minimize negative impacts on food pro- duction and even result in more productive FIGURE 2.13  Drought vulnerability is reduced by agricultural techniques that integrate trees and landscapes. Restoring degraded forestlands store carbon and landscapes, as called for by the 2014 (Reduction in average annual number of drought-affected New York Declaration on Forests, 6 could people) yield net benefits in the general order of $170 8 billion per year from watershed protection, Number of people (millions) improved crop yields, and forest products if 6 an area of 350 million hectares was restored by 2030 (NCE 2014). And most of the poten- 4 tial for energy crop production on degraded land is located in developing regions (Nijsen 2 et al. 2012). Adaptation and mitigation benefits 0 No trees Low tree density High tree density can be reaped at the level of the plot Agroforestry Drought and heat tolerance (through reduced tillage), farm (soil terrac- packages Soil fertility management ing combined with tree management), and landscape (thanks to agroforestry and Source: Cervigni and Morris 2015. B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    71 BOX 2.6  Securing local benefits from harnessing the forests to lower emissions Increasing international attention and funding have and environmental risks are minimized and benefits been raised for reducing emissions from deforestation enhanced (UN–REDD 2013a). and forest degradation and for other land-related mit- Fair benefit-sharing systems that allow poor peo- igation activities. At the 2010 Cancún meeting of the ple to receive direct (like monetary gains) and indi- United Nations Framework Convention on Climate rect benefits (like better governance infrastructure Change (UNFCCC), member countries agreed to provision) from REDD+ play a key role in operation- establish an international mechanism, whereby devel- alizing safeguards (Brockhaus et al. 2014). A pre- oped countries would pay low-income and middle- condition for such benefit sharing is tenure security income countries in the tropics for five types of and clarity that ensure access to ecosystems and par- forest-related mitigation activities, called REDD+: (i) ticipation rights for local people, which has become reducing emissions from deforestation, (ii) reducing a key element of many national policies (UN–REDD emissions from forest degradation, (iii) conservation 2013b). Under the Terra Legal program, Brazil has of forest carbon stocks, (iv) sustainable management started a formal process of recognizing indigenous of forest, and (v) enhancement of forest carbon. lands and granting land titles to about 300,000 Concerns about negative social and environmen- smallholders conditional on compliance with the tal impacts—such as restricting access for local peo- Brazilian Forest Code (Duchelle et al. 2014). The ple to forests and harming biodiversity—led to the number of studies that illustrate how REDD+ can establishment of REDD+ “safeguards” within the increase benefits for poor people is rapidly growing UNFCCC decisions. These safeguards require coun- (Groom and Palmer 2012; Berry, Harley, and Ryan tries to put procedures in place to ensure that social 2013; Luttrell et al. 2013). Brazil’s Bolsa Floresta program offers a health—but often clean energy solutions are monthly payment to low-income households not affordable for poor people. Financing if they commit to zero deforestation and efficient cookstoves, introducing cash trans- enroll their children in school. Guatemala fers, and enhancing social protection schemes offers forest incentive programs, which aim can help to protect poor people against some to support forestry activities by poor small- of the negative impacts in the short term holders without land title. In Colombia and (chapter 5). Nicaragua, mainly poor land users receive If done properly, land-based mitigation payment for ecosystem services (Pagiola, actions can create direct employment benefits Rios, and Arcenas 2008 and 2010). In Costa for local people. Climate-smart agriculture Rica, payments for reducing deforestation practices, land restoration, selective logging, go mainly to poor areas (Pfaff et al. 2007). If and forest protection are labor intensive and participation constraints (like a lack of land thus can provide jobs and revenues to poor titles, high transaction costs, or poor con- rural households (see box 2.6). One study nections to government institutions) can be finds that over half of some 40 REDD+ proj- overcome, such payments could be an ects reviewed provided benefits in terms of important livelihood element for poor peo- employment and incomes for local communi- ple (Bremer, Farley, and Lopez-Carr 2014; ties, although these benefits have been modest Jindal et al. 2013; Zbinden and Lee 2005). (Lawlor et al. 2013). In other cases, mitigation actions need to be accompanied by complementary mea- sures. Promoting modern energy for cooking In conclusion can reduce the use of traditional wood This chapter has shown how climate change ­ b iomass—which contributes to emissions can affect agricultural production systems through forest degradation and poor with significant implications for food prices 7 2    SHOCK WAVES and food security. While these impacts may Notes remain limited until 2030, they can become 1. These papers explore possible impacts in considerable in the longer term—especially two socioeconomic scenarios: (i) a prosperity in poorer regions. The only way to avoid this scenario (low population growth, high GDP longer-term ­ outcome is with mitigation poli- growth, and low poverty and inequality) cies that include agriculture, forestry, and and (ii) a poverty scenario (high population other land uses, which are often the main growth, low GDP growth, and high poverty sources of emissions in poor countries. But and inequality). See chapter 1 for a discussion these policies must be designed in a way that of the scenarios and how they were c ­ hosen. avoids negative impacts, or that even bene- The papers also explore possible impacts in fits local production or incomes—or these two climate scenarios: (i) a low-emissions policies risk adding pressure on food pro- scenario (likely to result in a warming of 2°C duction and prices. by the end of the century compared to pre- industrial levels with limited impacts) and While increases in agricultural prices— (ii) a high-emissions scenario (likely to result triggered by climate change impacts or miti- in a warming of 4°C with high impacts). The gation policies—will hurt all consumers, poor low-emissions scenario is consistent with nonagricultural and urban households will be the Representative Concentration Pathway hit the hardest. In rural areas, farmers and (RCP) 2.6 of the IPCC, while the high-­ agricultural laborers will experience mixed emissions scenario is based on the RCP8.5. impacts. On the one hand, production shocks 2. Impacts were calculated by comparing the result in a direct drop in incomes. On the agricultural price level (including crops and other hand, over the longer term, higher animal products) in the different emission prices and wages could increase earnings, scenarios—and thus with different magni- although this effect is more likely to reduce tudes of climate change—with the prices that would occur without climate change. the negative impacts than to reverse them, 3. In this model prices per tCO2 (ton of carbon especially if impacts are large. As for the dioxide) increase to between $17 and $84 resulting impact on poverty, our calculations in 2030 and are between $200 and $1,000 in (discussed in chapter 6) suggest that, in almost 2080. all possible scenarios, the net effect of the 4. Based on the first globally comparable data agricultural impacts of climate change will be from 58 sites representing smallholder sys- to increase global poverty. tems in (sub)tropical landscapes. These data At the same time, climate change can add over-represent sites with high forest cover to existing stresses on ecosystems, which and low population densities so that these could undermine subsistence production, a numbers are not representative for all rural critical safety net for the rural poor. Although areas, but they provide an estimate for some of the most marginal environments. such impacts remain difficult to quantify, case 5. The GHI is computed using three equally studies from various contexts show that cli- weighted indicators that are combined in one mate stress coupled with ecosystem degrada- index, namely the proportion of people who tion forces households to alter their livelihood are undernourished, the prevalence of under- strategies. weight children younger than five, and the The bottom line is that long-term ­ climate mortality rate of children younger than five. change trends are likely to affect agriculture All three index components are expressed in and ecosystems, with severe consequences for percentages and weighted equally. Higher poor people and their livelihoods. These GHI values indicate more hunger; lower val- threats will be further amplified by the ues indicate less. increasing frequency and severity of natural 6. http://www.un.org/climatechange/summit​ /­w p-content/uploads/sites/2/2014/07/New​ disasters, which will exacerbate production -York-Declaration-on-Forest-%E2%80%93​ and price shocks in the short term, and could -Action-Statement-and-Action-Plan.pdf. sharply increase poverty in the longer run— 7. h t t p s : / / w b c a r b o n f i n a n c e . o r g / R o u t e r​ the topic of the next chapter. .cfm?Page=Projport&ProjID=9625. B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    73 References 2012. “Effects of the 1997–1998 El Niño Episode on Community Rates of Diarrhea.” Ahmed, S. A., N. S. Diffenbaugh, and T. W. Hertel. Am. J. Public Health 102: e63–e69. 2009. “Climate Volatility Deepens Poverty Berry, N. J., R. Harley, and C. M. Ryan. 2013. Vulnerability in Developing Countries.” “Enabling Communities to Benefit from Environ. Res. Lett. 4: 034004. REDD+: Pragmatic Assessment of Carbon Alexandratos, N., J. Bruinsma, G. Bodeker, S. Broca, Benefits.” Carbon Manag. 4: 571–73. and M. Ottaviani. 2012. World Agriculture Biewald, A., H. Lotze-Campen, I. Otto, N. towards 2030/2050. Rome, Italy: Food and Brinckmann, B. Bodirsky, I. Weindl, A. Popp, Agriculture Organization of the United Nations. and H. J. Schellnhuber. Forthcoming. “The Alston, J.M., J. M. Beddow, and P. G. Pardey. Impact of Climate Change on Costs of Food 2010. “Global Patterns of Crop Yields and and People Exposed to Hunger at Subnational Other Partial Productivity Measures and Scale.” PIK Report 128, Potsdam Institute for Prices.” In The Shifting Patterns of Agricultural Climate Impact Research. Background paper Production and Productivity Worldwide, prepared for this report. 39–61. Ames, IA: Iowa State University, The Bouley, T., and C. Planté. 2014. Reducing Midwest Agribusiness Trade Research and Climate-Sensitive Disease Risks. World Bank Information Center (MATRIC). Report Number 84956-GLB. Washington, DC: Anderson, K., M. Ivanic, and W. Martin. 2013. World Bank. “Food Price Spikes, Price Insulation, and Bremer, L. L., K. A. Farley, and D. Lopez-Carr. Poverty.” Policy Research Working Paper 2014. “What Factors Influence Participation 6535, World Bank, Washington, DC. in Payment for Ecosystem Services Programs? Angelsen, A., and T. Dokken. Forthcoming. An Evaluation of Ecuador’s SocioPáramo “Environmental Reliance, Climate Exposure, Program.” Land Use Policy 36: 122–33. and Vulnerability: A Cross-Section Analysis of Brockhaus, M., G. Wong, C. Luttrell, L. Loft, Structural and Stochastic Poverty.” Background T.T. Pham, A. E. Duchell, D. Assembe-Mvondo, paper prepared for this report. and M. Di Gregorio. 2014. “Operationalizing Badola, R., and S. A. Hussain. 2005. “Valuing Safeguards in National REDD+ Benefit-Sharing Ecosystem Functions: An Empirical Study on Systems: Lessons on Effectiveness, Efficiency the Storm Protection Function of Bhitarkanika and Equity.” CIFOR REDD+ Safeguards Mangrove Ecosystem, India.” Environ. Brief no. 2, Center for International Forestry Conserv. 32: 85–92. Research (CIFOR), Bogor, Indonesia. Bagley, J. E., A. R. Desai, K. J. Harding, P. K. Burkett, M. 2011. “The Nation Ex-Situ: On Snyder, and J. A. Foley. 2013. “Drought Climate Change, Deterritorialized Nationhood, and Deforestation: Has Land Cover Change and the Post-Climate Era.” SSRN Scholarly Influenced Recent Precipitation Extremes in the Paper No. ID 2372457, Social Science Research Amazon?” J. Clim. 27: 345–61. Network, Rochester, NY. Barange, M., G. Merino, J. L. Blanchard, Calvin, K., M. Wise, P. Kyle, P. Patel, L. Clarke, and J. Scholtens, J. Harle, E. H. Allison, J. I. Allen, J. Edmonds. 2013. “Trade-Offs of Different Land J. Holt, and S. Jennings. 2014. “Impacts and Bioenergy Policies on the Path to Achieving of Climate Change on Marine Ecosystem Climate Targets.” Clim. Change 123: 691–704. Production in Societies Dependent on Cervigni, R., and M. Morris. 2015. “Enhancing Fisheries.” Nat. Clim. Change 4: 211–16. Resilience in African Drylands: Toward a Barbier, E. B. 2010. “Poverty, Development, and Shared Development Agenda.” World Bank, Environment.” Environ. Dev. Econ. 15: 635–60. Washington, DC. Barbier, E. B. Forthcoming. “Climate Change Impacts Chen, X., H. Huang, M. Khanna, and H. Önal. on Rural Poverty in Low-Elevation Coastal Zones.” 2011. “Meeting the Mandate for Biofuels: Background paper prepared for this report. Implications for Land Use, Food and Fuel Beddington, J. 2010. “Food Security: Prices.” Working Paper No. 16697, National Contributions from Science to a New and Bureau of Economic Research. Greener Revolution.” Philos. Trans. R. Soc. Damania, R., J. Russ, A. Barra, and A. Joshi. 2015. B Biol. Sci. 365, 61–71. “India’s Forests – Stepping Stone or Millstone Bennett, A., L. D. Epstein, R. H. Gilman, V. Cama, for the Poor?” World Bank Mimeo. C. Bern, L. Cabrera, A. G. Lescano, J. Patz, Das, S., and J. R. Vincent. 2009. “Mangroves C. Carcamo, C. R. Sterling, and W. Checkley. Protected Villages and Reduced Death Toll 7 4    SHOCK WAVES during Indian Super Cyclone.” Proc. Natl. Insecurity in the World 2014: Strengthening the Acad. Sci. 106: 7,357–60. Enabling Environment for Food Security and Dasgupta, S., M. Hossain, M. Huq, and D. Wheeler. Nutrition. Rome, Italy: FAO, IFAD, and WFP. 2014. “Climate Change, Soil Salinity, and the Ferraro, P. J., and M. M. Hanauer. 2014. Economics of High-Yield Rice Production in “Quantifying Causal Mechanisms to Coastal Bangladesh.” Policy Research Working Determine How Protected Areas Affect Poverty Paper 7140, World Bank, Washington, DC. through Changes in Ecosystem Services and Davidson, E. A., A. C. de Araújo, P. Artaxo, Infrastructure.” Proc. Natl. Acad. Sci. 111: J. K. Balch, I. F. Brown, M. M. Bustamante, 4,332–37. M. T. Coe, R. S. DeFries, M. Keller, M. Longo, Gilbert, C. L., and C. W. Morgan. 2010. “Food J. W. Munger, W. Schroeder, B. S. Soares-Filho, Price Volatility.” Philos. Trans. R. Soc. Lond. C. M. Souza, and S. C. Wofsy. 2012. “The B Biol. Sci. 365: 3,023–34. Amazon Basin in Transition.” Nature 481: Gilbert, N. 2012. “African Agriculture: Dirt 321–28. Poor.” Nature 483: 525–27. Davis, B., S. Di Giuseppe, and A. Zezza. 2014. Giller, K. E., E. Witter, M. Corbeels, and “Income Diversification Patterns in Rural Sub- P. Tittonell. 2009. “Conservation Agriculture Saharan Africa: Reassessing the Evidence.” and Smallholder Farming in Africa: The Policy Research Working Paper 7108, World Heretics’ View.” Field Crops Res. 114: 23–34. Bank, Washington, DC. Golub, A. A., B. B. Henderson, T. W. Hertel, de Janvry, A. 2015. “Quantifying through Ex P. J. Gerber, S. K. Rose, and B. Sohngen. 2013. Post Assessments the Micro-Level Impacts “Global Climate Policy Impacts on Livestock, of Sovereign Disaster Risk Financing and Land Use, Livelihoods, and Food Security.” Insurance Programs.” Policy Research Working Proc. Natl. Acad. Sci. 110: 20,894–99. Paper 7356, World Bank, Washington, DC. Grassini, P., K. M. Eskridge, and K. G. Cassman. del Ninno, C., P. A. Dorosh, L. C. Smith, and 2013. “Distinguishing between Yield Advances D. K. Roy. 2001. “The 1998 Floods in and Yield Plateaus in Historical Crop Bangladesh Disaster Impacts, Household Production Trends.” Nat. Commun. 4:2918. Coping Strategies, and Response.” Research Green, R., L. Cornelsen, A. D. Dangour, R. Turner, Report No. 122, International Food Policy B. Shankar, M. Mazzocchi, and R. D. Smith. Research Institute. 2013. “The Effect of Rising Food Prices on Dercon, S., D. O. Gilligan, J. Hoddinott, and Food Consumption: Systematic Review with T. Woldehanna. 2009. “The Impact of Meta-regression.” BMJ 346: f3703. Agricultural Extension and Roads on Poverty Groom, B., and C. Palmer. 2012. “REDD+ and and Consumption Growth in Fifteen Ethiopian Rural Livelihoods.” Biol. Conserv. 154: 42–52. Villages.” Am. J. Agric. Econ. 91: 1007–21. Harvey, C. A., M. Chacón, C. I. Donatti, E. Garen, Devarajan, S., S. D. Go, M. Maliszewska, I. Osorio- L. Hannah, A. Andrade, L. Bede, D. Brown, Rodarte, and H. Timmer. 2013. “Stress- A. Calle, J. Chará, C. Clement, E. Gray, Testing Africa’s Recent Growth and Poverty M. H. Hoang, P. Minang, A. M. Rodríguez, Performance.” Policy Research Working Paper C. Seeberg-Elverfeldt, B. Semroc, S. Shames, 6517, World Bank, Washington, DC. S. Smukler, E. Somarriba, E. Torquebiau, Duchelle, A. E., M. Cromberg, M. F. Gebara, J. van Etten, and E. Wollenberg. 2014. R. Guerra, T. Melo, A. Larson, P. Cronkleton, “Climate-Smart Landscapes: Opportunities J. Börner, E. Sills, S. Wunder, S. Bauch, P. May, and Challenges for Integrating Adaptation and G. Selaya, and W. D. Sunderlin. 2014. “Linking Mitigation in Tropical Agriculture.” Conserv. Forest Tenure Reform, Environmental Lett. 7: 77–90. Compliance, and Incentives: Lessons from Havlík, P., H. Valin, M. Gusti, E. Schmid, D. REDD+ Initiatives in the Brazilian Amazon.” Leclère, N. Forsell, M. Herrero, N. Khabarov, World Dev. 55: 53–67. A. Mosnier, M. Cantele, and M. Obersteiner. FAO (Food and Agriculture Organization of Forthcoming. “Climate Change Impacts and the United Nations). 2013. Climate-Smart Mitigation in the Developing World: An Agriculture Sourcebook. Rome, Italy: FAO. Integrated Assessment of Agriculture and FAO (Food and Agriculture Organization of the Forestry Sectors.” Background paper prepared United Nations), IFAD (International Fund for for this report. Agricultural Development), and WFP (World Hertel, T.W., M. B. Burke, and D. B. Lobell. 2010. Food Programme). 2014. The State of Food “The Poverty Implications of Climate-Induced B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    75 Crop Yield Changes by 2030.” Glob. Environ. the 2010–11 Surge in Food Prices.” World Dev. Change, 20th Anniversary Special Issue 20: 40: 2,302–17. 577–85. Jacoby, H.G., M. Rabassa, and E. Skoufias. Hill, R., and C. Mejia-Mantilla. 2015. “Welfare 2014. “Distributional Implications of Climate and Shocks in Uganda.” Background Paper Change in Rural India: A General Equilibrium prepared for the Uganda Poverty Assessment. Approach.” Am. J. Agric. Econ. aau084. World Bank, Washington, DC. Jassogne, L., P. Läderach, and P. van Asten. 2013. Hoffman, M., and A. I. Grigera. 2013. Climate “The Impact of Climate Change on Coffee Change, Migration, and Conflict in the Amazon in Uganda: Lessons from a Case Study in the and the Andes: Rising Tensions and Policy Rwenzori Mountains.” Oxfam Policy Pract. Options in South America. Washington, DC: Clim. Change Resil. 9: 51–66. Center for American Progress. Jindal, R., J. M. Kerr, P. J. Ferraro, and Hornbeck, R. 2012. “The Enduring Impact of the B. M. Swallow. 2013. “Social Dimensions American Dust Bowl: Short- and Long-Run of Procurement Auctions for Environmental Adjustments to Environmental Catastrophe.” Service Contracts: Evaluating Tradeoffs Am. Econ. Rev. 102: 1,477–1,507. between Cost-Effectiveness and Participation Howe, C., H. Suich, P. van Gardingen, A. Rahman, by the Poor in Rural Tanzania.” Land Use and G. M. Mace. 2013. “Elucidating the Policy, Themed Issue 1—Guest Editor Romy Pathways between Climate Change, Ecosystem Greiner, Themed Issue 2—Guest Editor Davide Services and Poverty Alleviation.” Curr. Opin. Viaggi 31, 71–80. Environ. Sustain., Terrestrial Systems 5: Kassam, A., R. Derpsch, and T. Friedrich. 2014. 102–107. “Global Achievements in Soil and Water ICARDA (International Center for Agricultural Conservation: The Case of Conservation Research in the Dry Areas). 2012. Conservation Agriculture.” Int. Soil Water Conserv. Res. 2: Agriculture: Opportunities for Intensified 5–13. Farming and Environmental Conservation in Klemick, H. 2011. “Shifting Cultivation, Forest Dry Areas: Farmer Experiences and Potential Fallow, and Externalities in Ecosystem for Uptake in Iraq, Syria, Morocco and Tunisia. Services: Evidence from the Eastern Amazon.” Research to Action 2 Series. J. Environ. Econ. Manag. 61: 95–106. ICRAF (International Center for Research in Kronik, J., and D. Verner. 2010. Indigenous Agroforestry). 2012. Creating Evergreen Peoples and Climate Change in Latin America Agricultue in Africa . Nairobi: World and the Caribbean. Directions in Development Agroforestry Centre. Series. Washington, DC: World Bank. IPCC (Intergovernmental Panel on Climate Kuik, O. 2013. “REDD+ and International Leakage Change). 2014. Climate Change 2014: via Food and Timber Markets: A CGE Analysis.” Impacts, Adaptation, and Vulnerability. Part Mitig. Adapt. Strateg. Glob. Change 1–15. A: Global and Sectoral Aspects. Contribution Läderach, P., A. Martinez-Valle, G. Schroth, of Working Group II to the Fifth Assessment and N. Castro. 2013. “Predicting the Future Report of the Intergovernmental Panel on Climatic Suitability for Cocoa Farming of the Climate Change. [Field, C.B., V.R. Barros, World’s Leading Producer Countries, Ghana D.J. Dokken, K.J. Mach, M.D. Mastrandrea, and Côte d’Ivoire.” Clim. Change 119: 841–54. T.E. Bilir, M. Chatterjee, K.L. Ebi, Y.O. Estrada, Lanjouw, P., and R. Murgai. 2009. “Poverty R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, Decline, Agricultural Wages, and Nonfarm S. MacCracken, P.R. Mastrandrea, and Employment in Rural India: 1983–2004.” L.L. White (eds.)]. Cambridge, U.K. and New Agric. Econ. 40: 243–63. York: Cambridge University Press. Larson, D. F., J. Lampietti, C. Gouel, C. Cafiero, Ivanic, M., and W. Martin. 2008. “Implications of and J. Roberts. 2013. “Food Security and Higher Global Food Prices for Poverty in Low- Storage in the Middle East and North Africa.” Income Countries.” Agric. Econ. 39: 405–16. World Bank Econ. Rev. ———. 2014. “Short- and Long-Run Impacts Lawlor, K., E. M. Madeira, J. Blockhus, and of Food Price Changes on Poverty.” Policy D. J. Ganz. 2013. “Community Participation Research Working Paper 7011, World Bank, and Benefits in REDD+: A Review of Initial Washington, DC. Outcomes and Lessons.” Forests 4: 296–318. Ivanic, M., W. Martin, and H. Zaman. 2012. Li, Y., D. Conway, Y. Wu, Q. Gao, S. Rothausen, “Estimating the Short-Run Poverty Impacts of W. Xiong, H. Ju, and E. Lin. 2013. “Rural 7 6    SHOCK WAVES Livelihoods and Climate Variability in Ningxia, Noack, F., S. Wunder, A. Angelsen, and J. Börner. Northwest China.” Clim. Change 119: 891–904. Forthcoming. “Responses to Weather and Climate: Little, P. D., H. Mahmoud, and D. L. Coppock. A Cross-Section Analysis of Rural Incomes.” 2001. “When Deserts Flood: Risk Management Background paper prepared for this report. and Climatic Processes among East African Olschewski, R., T. Tscharntke, P. C. Benítez, Pastoralists.” Clim. Res. 19: 149–59. S. Schwarze, and A.-M. Klein. 2006. “Economic Locatelli, B., E. Pramova, and A. Russell. 2012. Evaluation of Pollination Services Comparing Using Forests to Enhance Resilience to Coffee Landscapes in Ecuador and Indonesia.” Climate Change: What Do We Know about Ecol. Soc. 11: 7. How Forests Can Contribute to Adaptation? Pagiola, S., A. R. Rios, and A. Arcenas. 2008. Washington, DC: World Bank. “Can the Poor Participate in Payments for Lotze-Campen, H., M. von Lampe, P. Kyle, Environmental Services? Lessons from the S. Fujimori, P. Havlik, H. van Meijl, Silvopastoral Project in Nicaragua. Environ. T. Hasegawa, A. Popp, C. Schmitz, A. Tabeau, Dev. Econ. 13: 299–325. H. Valin, D. Willenbockel, and M. Wise. 2014. ———. 2010. “Poor Household Participation in “Impacts of Increased Bioenergy Demand on Payments for Environmental Services: Lessons Global Food Markets: An AgMIP Economic from the Silvopastoral Project in Quindío, Model Intercomparison.” Agric. Econ. 45: Colombia.” Environ. Resour. Econ. 47: 371–94. 103–16. Pardey, P. G., J. M. Alston, and C. Chan-Kang. Luttrell, C., L. Loft, M. Fernanda Gebara, 2013. “Public Agricultural R&D over the D. Kweka, M. Brockhaus, A. Angelsen, and Past Half Century: An Emerging New World W. D. Sunderlin. 2013. “Who Should Benefit Order.” Agric. Econ. 44: 103–13. from REDD+? Rationales and Realities.” Ecol. Pardey, P. G., and P. Pingali. 2010. “Reassessing Soc. 18. International Agricultural Research for Food McDowell, J. Z., and J. J. Hess. 2012. “Accessing and Agriculture.” In Report Prepared for the Adaptation: Multiple Stressors on Livelihoods Global Conference on Agricultural Research in the Bolivian Highlands under a Changing for Development, Notes to pp. 188–90. Climate.” Glob. Environ. Change, Adding Insult Pattanayak, S. K. 2004. “Valuing Watershed to Injury: Climate Change, Social Stratification, Services: Concepts and Empirics from Southeast and the Inequities of Intervention 22: 342–52. Asia.” Agric. Ecosyst. Environ. 104: 171–84. McKinnon, K., and V. Hickey. 2009. Convenient Pfaff, A., S. Kerr, L. Lipper, R. Cavatassi, B. Davis, Solutions to an Inconvenient Truth: Ecosystem- J. Hendy, and G. A. Sanchez-Azofeifa. 2007. Based Approaches to Climate Change . “Will Buying Tropical Forest Carbon Benefit Washington, DC: World Bank. the Poor? Evidence from Costa Rica.” Land MEA (Millenium Ecosystem Assessment). 2005. Use Policy 24: 600–10. Ecosystems and Human Well-Being: Synthesis. Popp, A., J. P. Dietrich, H. Lotze-Campen, Washington, DC: World Resources Institute. D. Klein, N. Bauer, M. Krause, T. Beringer, NCE (New Climate Economy). 2014. Better D. Gerten, and O. Edenhofer. 2011. “The Growth Better Climate: The New Climate Economic Potential of Bioenergy for Climate Economy Report . Washington, DC: New Change Mitigation with Special Attention Climate Economy. Given to Implications for the Land System.” Ndiaye, M., E. Maitre d’Hôtel, and T. Le Cotty. Environ. Res. Lett. 6: 034017. 2015. “Maize Price Volatility: Does Market Porter, J. R., L. Xie, A. J. Challinor, K. Cochrane, Remoteness Matter?” Policy Res. Work. Pap. S. M. Howden, M. M. Iqbal, D. B. Lobell, 7202, World Bank, Washington, DC. and M. I. Travasso. 2014. “Food Security and Nelson, G. C., D. Mensbrugghe, H. Ahammad, Food Production Systems.” in Climate Change E. Blanc, K. Calvin, T. Hasegawa, P. Havlik, 2014: Impacts, Adaptation, and Vulnerability. E. Heyhoe, P. Kyle, and H. Lotze-Campen. Part A: Global and Sectoral Aspects, edited 2014. “Agriculture and Climate Change in by C. B. Field, V. R. Barros, D. J. Dokken, Global Scenarios: Why Don’t the Models K. J. Mach, M. D. Mastrandrea, T. E. Bilir, Agree.” Agric. Econ. 45: 85–101. M. Chatterjee, K. L. Ebi, Y. O. Estrada, Nijsen, M., E. Smeets, E. Stehfest, and R. C. Genova, B. Girma, E. S. Kissel, A. N. Levy, D. P. Vuuren. 2012. “An Evaluation of the S. MacCracken, P. R. Mastrandrea, and Global Potential of Bioenergy Production on L. L. White, 485–533. Contribution of Working Degraded Lands.” Gcb Bioenergy 4: 130–47. Group II to the Fifth Assessment Report of the B ad S eed : C limate C hange , A gric u lt u re , and F ood S ec u rit y    77 Intergovernmental Panel on Climate Change. Tschakert, P. 2007. “Views from the Vulnerable: Cambridge, U.K. and New York: Cambridge Understanding Climatic and Other Stressors in University Press. the Sahel.” Glob. Environ. Change 17: 381–96. Rabbani, G., A. Rahman, and K. Mainuddin. U.K.–U.S. Task Force. 2015. Extreme Weather 2013. “Salinity-Induced Loss and Damage to and Resilience of the Global Food System, Final Farming Households in Coastal Bangladesh.” Project Report. Taskforce on Extreme Weather Int. J. Glob. Warm. 5: 400–15. and Global Food System Resilience, The Global Ricketts, T. H., G. C. Daily, P. R. Ehrlich, and Food Security Programme, United Kingdom. C. D. Michener. 2004. “Economic Value of UN–REDD (United Nations Collaborative Tropical Forest to Coffee Production.” Proc. Programme on Reducing Emissions from Natl. Acad. Sci. 101: 12,579–82. Deforestation and Forest Degradation in Roe, D., M. Fancourt, C. Sandbrook, M. Sibanda, Developing Countries). 2013a. “Putting A. Giuliani, and A. Gordon-Maclean. REDD+ Safeguards and Safeguard Information 2014. “Which Components or Attributes of Systems Into Practice.” UN-REDD Programme Biodiversity Influence Which Dimensions of Policy Brief No 3. Poverty.” Environ. Evid. 3. ———. 2013b. “Tenure and REDD+ Developing Shameem, M. I. M., S. Momtaz, and R. Rauscher. Enabling Tenure Conditions for REDD+.” 2014. “Vulnerability of Rural Livelihoods to UN-REDD Programme Policy Brief No 6. Multiple Stressors: A Case Study from the von Grebmer, K., M. Torero, T. Olofinbiyi, Southwest Coastal Region of Bangladesh.” H. Fritschel, D. Wiesmann, Y. Yohannes, Ocean Coast. Manag. 102: 79–87. L. Schofield, and C. von Oppeln. 2011. 2011 Sietz, D., S. E. M. Choque, and M. K. B. Lüdeke. Global Hunger Index: The Challenge of 2011. “Typical Patterns of Smallholder Hunger: Taming Price Spikes and Excessive Vulnerability to Weather Extremes with Regard Food Price Volatility . Washington, DC: to Food Security in the Peruvian Altiplano.” International Food Policy Research Institute. Reg. Environ. Change 12: 489–505. Wise, M., K. Calvin, A. Thomson, L. Clarke, Sissoko, K., H. van Keulen, J. Verhagen, V. Tekken, B. Bond-Lamberty, R. Sands, S. J. Smith, and A. Battaglini. 2010. “Agriculture, Livelihoods A. Janetos, and J. Edmonds. 2009. “Implications and Climate Change in the West African Sahel.” of Limiting CO2 Concentrations for Land Use Reg. Environ. Change 11: 119–25. and Energy.” Science 324: 1,183–86. Smith, P., and M. Bustamante. 2014. “Chapter Wood, B. D., C. H. Nelson, and L. Nogueira. 2012. 11: Agriculture, Forestry and Other Land “Poverty Effects of Food Price Escalation: The Use (AFOLU).” In Climate Change 2014: Importance of Substitution Effects in Mexican Mitigation of Climate Change. Contribution Households.” Food Policy 37: 77–85. of Working Group III to the Fifth Assessment World Bank. 2011. Climate-Smart Agriculture: Report of the Intergovernmental Panel on Increased Productivity and Food Security, Climate Change. Cambridge, U.K. and New Enhanced Resilience and Reduced Carbon York: Cambridge University Press. Emissions for Sustainable Development. Suich, H., C. Howe, and G. Mace. 2015. “Ecosystem Washington, DC: World Bank. Services and Poverty Alleviation: A Review of the ———. 2012. Inclusive Green Growth: The Empirical Links.” Ecosyst. Serv. 12: 137–47. Pathway to Sustainable Development . Tadesse, G., B. Algieri, M. Kalkuhl, and J. von Washington, DC: World Bank. Braun. 2014. “Drivers and Triggers of ———. 2014a. Building Resilience for Sustainable International Food Price Spikes and Volatility.” Development of the Sundarbans: Strategy Food Policy 47: 117–28. Report. Washington, DC: World Bank. Tester, M., and P. Langridge. 2010. “Breeding ———. 2014b. Turn Down the Heat II: Climate Technologies to Increase Crop Production in a Extremes, Regional Impacts, and the Case for Changing World.” Science 327: 818–22. Resilience. Washington, DC: World Bank. Tilman, D., C. Balzer, J. Hill, and B. L. Befort. ———. 2014c. Turn Down the Heat III: 2011. “Global Food Demand and the Confronting the New Climate Normal . Sustainable Intensification of Agriculture.” Washington, DC: World Bank. Proc. Natl. Acad. Sci. 108: 20,260–64. Zbinden, S., and D. R. Lee. 2005. “Paying Triplett, G. B., and W. A. Dick. 2008. “No-Tillage for Environmental Services: An Analysis of Crop Production: A Revolution in Agriculture!” Participation in Costa Rica’s PSA Program.” Agron. J. 100: S–153. World Dev. 33: 255–72. Threat Multiplier: Climate Change, Disasters, and Poor People 3 Main Messages • Natural disasters push people into poverty share of their assets and income when hit by and prevent poor people from escaping a disaster, natural disasters increase inequal- poverty. ity and may contribute toward a decoupling • An increase in natural hazards is already of economic growth and poverty reduction. observed and will worsen in the next • There are many options to reduce risks for decades. Some events considered exceptional poor people; and, although none are easy to today will become frequent in the long term, implement, they do help reduce poverty and threatening current living conditions. make the population more resilient to cli- • These changes in hazards will affect poor mate change. Examples include risk-sensitive people and our ability to eradicate poverty. land use regulation, more and better infra- Because poor people are often more exposed structure, better housing quality and for- to natural hazards than the rest of the pop- mal land tenure, air-conditioning, financial ulation, and almost always lose a greater inclusion, and early warning and evacuation. Introduction • In Peru over the 2003 to 2008 period, one extra disaster per year increased Across the globe, on top of stresses on agricul- poverty rates by 16 to 23 percent at the ture and ecosystems, shocks from natural provincial level (Glave, Fort, and ­ h azards—from droughts to floods and Rosemberg 2008). storms—are a major reason why people • At the municipal level in Mexico, floods become and stay poor. Indeed, evidence sug- and droughts increased poverty levels gests that disasters increase poverty (Karim and by 1.5 to 3.7 percent between 2000 and Noy 2014). Consider the following examples: 2005 (Rodriguez-Oreggia et al. 2013). 79 8 0    SHOCK WAVES • In Bolivia, the poverty incidence rose poverty. An increase in the frequency or inten- 12 percent in Trinidad following the 2006 sity of natural disasters is expected because of floods (Perez-De-Rada and Paz 2008). climate change—which is likely to push more • For coastal communities in the subdis- people into poverty and increase poverty trict of Shyamnagar in the southwest of headcounts. Another key channel, agriculture ­ Bangladesh, after Cyclone Aila hit in 2009, and ecosystems, was covered in chapter 2, and unemployment skyrocketed (from 11 to a third, health, will be explored in chapter 4. 60 percent between 2009 and 2010) and How can we explain the specific impacts of per capita income decreased sharply (from natural hazards on poor people? And what $15,000 before the storm to $10,000 policy options are available to reduce this vul- af ter). T he pover t y headcount rate nerability of poor people, especially in light of increased from 41 to 63 percent between climate change? Keep in mind that we are 2009 and 2010 (Akter and Mallick 2013). referring to all climate-related natural disas- ters, regardless of whether they are caused by Moreover, recovery is not straightforward natural climate variability or man-made emis- for poor people. After Ethiopia’s 1984–85 fam- sions, as the two are closely linked and call ine, it took a decade on average for asset-poor for integrated risk management strategies that households to bring livestock holdings back to account for climate change. prefamine levels (Dercon 2004). While a pro- This chapter tries to shed light on these longed shock such as a drought can have long- questions. It begins with a short review of term impacts, so too can temporary shocks on how climate change will affect natural haz- human capital and poverty (Rentschler 2013). ards globally, then explores how these changes In Mexico, once children have been taken out will impact poor people and affect the evolu- of school, even just for a temporary shock such tion of poverty. It draws on results from origi- as a flood, they are 30 percent less likely to pro- nal studies that investigate—for the first time ceed with their education compared to children at the global level—the exposure differential who remain in school (de Janvry et al. 2006). between poor and nonpoor people, looking at Temporary spending adjustments by low- droughts, floods, and high temperature. It income households can result in permanent also reviews several case studies on past disas- shifts—at the expense of the child’s human ters, deriving insights on the greater exposure capital and future productivity. and vulnerability of poor people, along with Further, poor households exposed to unin- policy options to reduce this vulnerability. sured weather risk have been shown to reduce Our main message is that the measures investment in productive assets and select low- and policies that could be mobilized to help risk, low-return activities, perpetuating pov- poor people manage natural risks in a chang- erty (Cole et al. 2013; Elbers, Gunning, and ing ­climate amount to “good development,” Kinsey 2007; Shenoy 2013). In terms of the which would make sense even in the absence impact of disaster risk on poverty, these ex of climate change—with the important caveat ante impacts can dominate ex post impacts, that the design of such measures and policies that is, the losses caused by a disaster (Elbers, needs to take into account climate change Gunning and Kinsey 2007). This link from and the uncertainty it creates. natural hazard exposure to poverty may create a negative feedback loop, in which poor house- holds have no choice but to settle in ­ at-risk Climate change will worsen zones (with cheaper rents) and as a result face natural hazards in most regions increased challenges to escaping poverty. Natural disasters are thus one of the critical of the world channels through which climate-sensitive The large-scale changes in temperature, events already affect, and can increasingly ­ p recipitation, and other meteorological affect, the ability of poor people to escape variables that models project as a result of ­ T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   81 climate change suggest that all extreme in non-​ air-conditioned facilities during the events related to these variables (droughts, summer. While no single heat wave or floods, heat waves, and cold spells) will be extreme event is “caused” by climate change, affected. These impacts have been reviewed the effect of climate change on heat waves’ in many studies recently, including in the lat- frequency and intensity is already detectable est report of the Intergovernmental Panel on and is growing over time (box 3.1). Climate Change (IPCC 2013) and in the Droughts. Water availability depends on three volumes of the Turn Down the Heat more than just precipitation—seasonal cycles, reports (World Bank 2015a). This section snow packs, and evaporation rates also mat- briefly reviews the changes in hazards to be ter. Because of the variability in local changes expected. in climate, the evolution of droughts will vary Heat waves and cold spells. Climate depending on location. Overall, however, change will not make all extremes worse in droughts are likely to become more common—​ the future. Cold spells have serious conse- and it is likely that in many locations where quences and are expected to decrease both in droughts already are an issue the situation will frequency and in intensity. However, it is worsen (like the Mediterranean basin, almost certain that heat waves will become Southeast Europe, North Africa, Southern more frequent and intense in the future in Africa, Australia, South America, and Central most regions of the world. For instance, in America). A background paper for this report North Africa, temperatures considered excep- estimates that under a high-emissions scenario, tional today (and that have significant harm- the number of people exposed to droughts ful effects on ecosystem function and people’s could increase by 9 to 17 percent in 2030 and well-being) will become a new normal under 50 to 90 percent in 2080 (Winsemius et al., a high-emissions (4°C) scenario (World Bank forthcoming). Even so, some regions that cur- 2014) (map 3.1). In Europe, the summer rently experience regular water stress (like East 2003 heat wave, which led to more than Africa) are expected to see an improvement in 70,000 deaths, would become an “average” water resources (map 3.2). summer at the end of this century under a Tropical and extratropical storms. With high-emissions scenario—meaning that by higher temperatures, atmospheric circulations 2100, every other summer would be warmer are modified, influencing winds and storms than the 2003 one. globally. But tropical storms (present in the In many regions, such massive changes tropics, the strongest of which are referred to would threaten everyday living conditions— as hurricanes in the North Atlantic and like the ability to work outside or typhoons in the Pacific) and extratropical MAP 3.1  Continued high emissions will mean many more “broiling” summer months (Percentage of summer months with extreme temperatures by 2100, for a low-emissions (left) and high-emissions (right) scenario) Source: World Bank 2014. Note: Extreme temperatures are defined as temperatures that occur today less than once every 700 years. 8 2    SHOCK WAVES BOX 3.1  Climate change makes extreme weather events more likely or more intense Granted, individual events can never be fully attrib- human-caused climate change was a contributing uted to climate change—even the most dramatic factor, even though natural fluctuations also were key events of recent years would have been possible in a (Peterson et al. 2013). Heat waves, like the one that climate with no human influence, simply due to the affected the U.S. midwest and northeast in July 2012, natural variability of the climate (Hulme 2014). But are now four times as likely because of climate change. recent trends in extreme temperature and precipita- As for the current increase in disaster losses, they tion can now be linked with climate change. can be explained by socioeconomic evolutions​ — One recent study estimates that about 75 percent especially by the increase in population and wealth of the moderate daily hot extremes over land and 18 located in coastal and other at-risk areas. While percent of moderate daily precipitation extremes are an impact of climate change on economic losses already attributable to warming (Fischer and Knutti p robably exists, for now it remains undetectable ­ 2015). And the probability of the occurrence of the (IPCC 2012). 2003 European heat wave is estimated to have been Given the close interplay between natural climate doubled by the human influence on climate (Stott, variability and man-made climate change, disaster Stone, and Allen 2004). About half the analyses risk management cannot be separated from climate of extreme events in 2012 find some evidence that change adaptation. MAP 3.2  With unmitigated climate change, total days under storms (present in the mid- to high-latitude drought conditions will increase by more than 20 percent in most regions) may be impacted differently. regions For tropical storms, a best guess today is (Change in the number of days under drought conditions by 2100 under a high-emissions scenario) that their overall number may decrease, even as the most intense storms may become more frequent, especially in the North Atlantic (IPCC 2013; Knutson et al. 2010; Ranson et al. 2014). In addition, tropical cyclones may start affecting new regions that are likely to be less prepared and more vulnerable (Hallegatte 2007; Kossin, Emanuel, and Vecchi 2014). A review of 11 studies con- cludes that economic losses from tropical cyclones could increase from 9 to 417 percent by 2040, depending on the region and the methodology applied (Bouwer 2013). For extratropical storms, there is little agreement on how they will evolve, although models suggest that their intensity will increase and their mean trajectory will shift Source: Prudhomme et al. 2014. toward higher latitudes (Ranson et al. 2014). Notes: Drought days are defined as days during which the river runoff is below 10 percent of the A review of seven studies finds that economic 1976–2005 average. Regions in white are those that experience very low runoff today and in the future. losses from extratropical cyclones could increase from 11 to 120 percent by 2040, depending on region and methodology (Bouwer 2013). And these studies do not account for sea level rise, which could make T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   83 extratropical storms even more destructive changes (like land subsidence for coastal risk), (Hallegatte et al. 2013). socioeconomic changes (like higher popula- Coastal floods. Climate change will cause tion and wealth in at-risk areas), and disaster a global rise in sea level, with widespread preparedness. These changes will have large consequences on coastal risks. This global impacts and are likely to dominate the effect rise combined with local mechanisms— of climate change on economic losses from including changes in water currents and natural hazards in the next few decades local geological dynamics—will increase (Bouwer 2013; Hallegatte et al. 2013; land loss from erosion, water salinization, Mendelsohn et al. 2012). and flood risks from storm surges in most coastal areas (World Bank 2014). Coastal flood risks are already large, as illustrated Poor people are often—but by the disastrous consequences of Tropical not always—more exposed to Cyclone Sidr in Bangladesh in 2007 or the destruction caused by Hurricane Katrina in hazards New Orleans in 2005, and they will only Areas at risk of natural hazards have always increase over time (box 3.2). Multiple stud- attracted people and investment. Globally, ies have shown how even a limited rise in there is a trend toward increased risk-taking: sea level can significantly increase the likeli- from 1970 to 2010, the world population hood of very destructive coastal floods grew by 87 percent, while the population in (Jongman, Ward, and Aerts 2012; Wong flood plains increased by 114 percent and in et al. 2014). cyclone-prone coastlines by 192 percent. Heavy precipitation and floods. As pre- Further, the GDP exposed to tropical cipitation changes, so will river runoff—a cyclones increased from 3.6 percent to development that may have large conse- 4.3 percent of global GDP over the same quences on flood risk, with increases in some period (UNISDR 2011). The same trends places and decreases in others. Averages are hold at the country-level (Jongman et al. weak proxies for changes in risk, as extreme 2014; Pielke et al. 2008). rainfall or river runoff can rise even in a At-risk areas may be more attractive—in region where average precipitation and runoff spite of the risk—when they offer economic fall. Climate change is likely to exacerbate the opportunities, public services or direct ame- most intense precipitation events (Min et al. nities, and higher productivity and incomes 2011), with serious consequences for urban (Hallegatte 2012a). In some rural areas, flash floods. Under a high-emissions scenario, proximity to water offers cheaper transport, the number of exposed people could increase and regular floods increase agricultural pro- by 4 to 15 percent by 2030 and 12 to 29 per- ductivity (Loayza et al. 2012). People may cent by 2080, according to a background settle in risky areas to benefit from opportu- paper for this report (Winsemius et al., forth- nities with industries driven by exports in coming). Under current vulnerability levels, coastal areas (box 3.2). Agglomeration the total number of global fatalities may well externalities may attract people to cities, double between now and 2080, based on the even if cities are more exposed than rural latest projections of climate change, popula- areas and newcomers have no choice but to tion, and GDP (Jongman et al. 2015). settle in risky places. In a background paper Economic losses from river floods could prepared for this report, households in increase by 7 to 124 percent by 2040, depend- flooded areas in Mumbai, India, report that ing on the methodology applied and region they are aware of the flood risks but accept considered (Bouwer 2013). them because of the opportunities offered by While climate change matters for the future the area (such as access to jobs, schools, of these hazards, nonclimate factors will also health care facilities, and social networks) affect future risks. These include physical (Patankar, forthcoming). 8 4    SHOCK WAVES BOX 3.2  Large coastal cities: Wealthier places at risk of floods The world’s 136 largest coastal cities are examples changes—­ c limate change and land subsidence— of relatively wealthier places with large flood risks. would make losses soar rapidly if present protec- A World Bank and Organisation for Economic tion is not upgraded. Even a moderate sea level rise Co-operation and Development (OECD) study (20 cm) combined with subsidence would make it (Hallegatte et al. 2013) estimates that average global necessary to invest massively in protection to avoid flood losses today are about $ 6 billion per year, losses that could otherwise quickly reach levels of despite existing flood defenses. Even though these more than $1 trillion per year. Existing protection cities host pockets of deep poverty in slums and can rapidly prove ill-adapted to changing environ- informal settlements, they are usually wealthier than mental conditions and generate very high risks, the rest of the country, thanks to a concentration of which are invisible until a disaster happens. export-led industries and skilled services. But they But even if adaptation investments (like higher dikes are also hotspots for flood risks, with widely varying and seawalls) keep the probability of coastal floods protection levels. While cities in rich countries have constant, subsidence and sea level rise could increase the largest levels of risk in absolute terms, cities in global flood losses by 2050 to $60–63 ­ billion per developing countries experience higher relative risk year. Further, since more population and assets would levels (in percentage of local GDP), largely driven depend on protection, the consequences of a dike by lower protection. As map B3.2.1 shows, the 20 failure or of an event that exceeds protection design cities with the highest relative risks are almost all would become much higher. While better protection located in developing countries, especially in South can reduce risk, it also raises the potential for larger- and Southeast Asia. ­ verwhelmed scale disasters if protections fail or are o Current trends in urbanization and economic by an exceptional event—making it essential to growth alone are expected to increase flood losses develop early warning and evacuation systems, crisis- in these cities, from $6 billion per year today management preparedness, and reconstruction plans to $52 ­billion per year by 2050. Environmental (Hallegatte 2012b; Hallegatte et al. 2013). MAP B3.2.1  Most cities with the highest relative coastal flood losses are in South and Southeast Asia (Average annual losses from coastal floods (relative to local GDP) in the 20 riskiest cities in the world) Source: World Bank (IBRD 41909, September 2015) based on Hallegatte et al. 2013. T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   85 Within a country or region, the attractive- drought hazard data come from a global ness of risky places means that people living model (GLOFRIS), which produces gridded there need not be poorer than the rest of indicators of inundation depth (for flood, the population. For instance, urban dwellers 1 km resolution) and water scarcity (for are, on average, wealthier than their rural drought, 5 km resolution). For temperature, countrymen. Since many cities are more we use observed spatial data on the maxi- exposed to floods than are rural areas, the mum monthly temperature for each grid cell urban-rural divide may make poorer people (at the 1 km resolution) from the Climatic less exposed to floods than the wealthier Research Unit of the University of East urban population. However, at a more local Anglia, which provide gridded estimates of scale and especially in urban areas, land and temperature extremes from 1960 onward. housing markets often push poorer people to This state-of-the-art hazard data were settle in riskier areas. Where markets factor combined with spatially explicit poverty data in hazard risks, housing is cheaper where risk using a global dataset of household surveys in is higher. And, because poorer people have 52 countries from the Demographic and fewer financial resources to spend on housing Health Surveys (DHS). These surveys contain and a generally lower willingness and ability data on each household’s location and wealth to pay for safety, they are more likely to live status. By calculating the flood, drought, and in at-risk areas. temperature indicator at the household level, The bottom line is that the “opportunity it is possible to examine whether and how effect” attracts both rich and poor people to this exposure is different for poor and non- risky areas, even though land markets push poor households. Poor people are defined as poor people into riskier areas within a city. those in the lowest quintile of the population Whether poor people are more or less exposed in terms of the “wealth index” provided in than nonpoor people is an empirical question the surveys, which is a measure of the assets on which so far there has been little research. that a household owns. That is why this report explores the differen- Combining hazard and socioeconomic tial exposure of poor and nonpoor people, data, a poverty exposure bias can be used to drawing on national studies and local measure whether poor people are more surveys. exposed to a hazard. For a given area, the One of our background papers examines poverty exposure bias is the share of poor poverty-specific exposure to floods and people exposed to a hazard, divided by the droughts in 52 countries (Winsemius et al., share of the total population exposed, sub- forthcoming). It provides new insights by tracted by 1. A positive bias means poor peo- assessing if and where poor people are more ple are more exposed than average; a exposed, and how this may change with a negative bias implies poor people are less changing climate. Using the same socioeco- exposed than average. With this definition in nomic data, another background paper exam- hand we ask whether poor people are more ines the exposure of poor people to extreme exposed to floods, droughts, and high tem- temperatures (Park et al., forthcoming). peratures within the 52 countries for which To understand whether poor people are we have data. more exposed to floods, droughts, and Floods. For river floods at the country- extreme temperatures we need “geo-­ level, we find mixed results as illustrated in referenced” information (where people live, map 3.3, panels a, b, and c, which show the their income levels) and hazard maps—which poverty exposure bias for floods with a return have only recently become available at the period (or 10 percent annual probability of global level and at high resolution (Ward occurrence) of 10 years (other return periods et al. 2013 and Winsemius et al. 2013 for show similar results). In Latin America and the floods; Prudhomme et al. 2014 and Schewe Caribbean and Asia, no pattern emerges: some et al. 2014 for drought). Our flood and countries exhibit a positive bias (poor people 8 6    SHOCK WAVES MAP 3.3  Poor people are more exposed to river floods in many countries, especially in urban areas (Poverty exposure bias for floods at national level (top) and in urban areas only (bottom)) Source: World Bank (IBRD 41905 and 41902, September 2015) based on Winsemius et al., forthcoming. Note: Exposure was calculated for the 10-year return period (results are similar for other return period events for floods). T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   87 more exposed than average) and others exhibit no bias or a negative one (poor people less exposed than average). But in Africa, regional patterns appear. In the southwest, countries exhibit a strong overexposure of poor people, as do those with larger rivers in the west (like Benin, Cameroon, and Nigeria). Among the countries analyzed, about half (representing 60 percent of the analyzed population) live in countries where poor people are more exposed to floods than average. What if we focus only on urban house- holds? Land scarcity is more acute in urban areas (compared to rural areas), and thus might create a stronger incentive for poor people to settle in risky areas due to lower prices. The results for urban households demonstrate a clear difference between the exposure of poor and nonpoor people, as can be seen in panels d, e, and f of map 3.3. In most countries (about 73 percent of the analyzed population), poor urban house- holds are more exposed to floods than the average urban population. There is no such pattern for rural households, suggesting that land scarcity is a driver of flood risk in urban areas. This phenomenon of high exposure to flood risk for poor urban dwell- ers is also found using micro-level data on household location and flood hazard in Mumbai, India (box 3.3). Droughts. Results for droughts at the country level show a more prominent pov- erty exposure bias, as illustrated in map 3.4. In most Asian countries and in southern and eastern Africa, poor households are more exposed to droughts (the definition of drought here is based on surface flows only and does not include groundwater and artifi- cial water storage). In western Africa, coastal countries (Benin, Cameroon, Ghana, Nigeria, and Togo) exhibit a positive bias, with the exception of Niger. In Latin America, poor people appear underexposed in Bolivia and Peru, but overexposed in Colombia, Guyana, and Honduras. Importantly, a number of Sub-Saharan African countries show a positive poverty exposure bias for both droughts and floods. When examining the total population, the 8 8    SHOCK WAVES BOX 3.3  In Mumbai, poor people are disproportionately exposed to floods In July 2005, Mumbai experienced an unprecedented almost completely absent from at-risk areas. Second, flood, causing 500 fatalities and direct economic more households overall are likely to be exposed damages of $2 billion (Ranger et al. 2011). The flood to flood risks under the cli­m ate change scenario. took a toll on low-income and marginalized people— Third, the distribution of expo­ s ure across poor with their losses estimated at about $245 million, and nonpoor people is similar for both scenarios: of which almost $235 million came from household additional exposure (of climate change) has the same asset losses and the rest from informal business losses distribution as current exposure. (Hallegatte et al. 2010). While these impacts are large in and of themselves, they are likely an underesti- mate. Actual impacts on marginalized populations, MAP B3.3.1  Mumbai’s poor are over-represented in especially health impacts and out-of-pocket expendi- the Mithi River Basin flood zone tures, were probably much larger. Are Mumbai’s poor people more exposed than nonpoor people to current and future floods? To answer these questions we explore the exposure of poor and nonpoor people to similar floods in the Mithi River Basin flood zone, drawing on a city-level household survey (containing each household’s loca- tion and income) and two flood maps (one based on today’s climate and the other based on the climate projected in a high-emissions scenario by 2080), as illustrated in map B3.3.1. T hree results stand out. First, under both scenarios, households in lower-income levels are disproportionately exposed, with 75 percent of those Note: Dots represent households (with associated monthly income in Indian exposed reporting a monthly income of 7,500 rupees rupees), overlaid with flood extent (blue is historical; purple is climate change or less (table B3.3.1)—and the richest households impacts for 2080). TABLE B3.3.1  Poor people tend to be more exposed to floods in Mumbai Household income Share of population Share of population exposed Share exposed with 2080 Rs./month in survey (%) in 2005 (%) climate (%) < 5,000 27 44 43 5,000–7,500 28 33 34 7,501–10,000 22 16 17 10,001–15,000 12 5 5 15,001–20,000 6 1 1 > 20,000 6 1 1 Source: Calculations based on Baker et al. 2005 and Ranger et al. 2011. Thanks to Risk Management Solutions for production of flood maps. Household income shown in Rs. (rupees). The share of population exposed in 2005 and 2080 is based on modeling exercises. T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   89 poverty exposure bias is more evident: most important mechanisms and small-scale dif- people (85 percent of the analyzed popula- ferences, from one block to the next. tion) live in places with an overexposure of Another way to examine whether poor peo- poor people to droughts. ple are more exposed to natural hazards is Temperatures. We find that poor people through in-depth case studies of actual past are often more exposed to higher tempera- events, analyzing household survey data. tures: 37 out of 52 countries (representing While many studies of disaster impacts are 56 percent of the population) exhibit a posi- available, only a few look at the exposure of tive bias (map 3.5). In Africa, most countries poor and nonpoor people separately. We have a positive poverty exposure bias, with provide the first systematic review of their regional patterns similar to those found for findings. floods and droughts, with the positive bias At the local scale, poor people seem much particularly strong in western Africa (Benin, more likely to be affected by natural hazards. Cameroon, and Nigeria) and southern Africa In Bangladesh, after Cyclone Aila hit in (Angola, Namibia, and Zambia). In Asia, the 2009, a postdisaster survey of 12 villages on results for temperature are regionally consis- the southwest coast finds that 25 percent of tent, with most countries exhibiting zero or poor households in these villages were negative bias; in Central America, results are exposed to the cyclone while only 14 percent again sporadic. of nonpoor households were (Akter and Also worrying is that many of the 37 Mallick 2013). In Vietnam, a similar pattern ­ countries that exhibit a poverty exposure emerges for the Mekong Delta: 38 percent of bias for temperature are already hot. If we the region’s poor but only 29 percent of the plot the poverty exposure bias against a region’s nonpoor live in frequently flooded country’s average annual temperature from areas (Nguyen 2011). 1961 to 1999 (to represent average climate), However, this pattern is not universal. we find that hotter countries have a higher A postdisaster survey after the 1998 Great exposure bias (figure 3.1, panel a). At the Flood in Bangladesh finds similar exposure: same time, cooler countries exhibit a smaller 75 percent of poor people and 71 percent of bias, and in some cool countries, a negative nonpoor people were affected (del Ninno bias. This occurs because, in these cool coun- et al. 2001). After the 2011 floods in Kenya, tries, nonpoor people tend to settle in areas almost everyone in the Bunyala District was with higher temperatures because they are affected (Opondo 2013). In the Middle East climatically more desirable. and North Africa, a study of five countries The results for temperature suggest a sort- finds that the percentage of households ing of the population into desirable and less- reporting being affected by a disaster in the desirable areas within a country, with last five years is high at 90 percent, but does wealthier households typically living in not vary based on poverty status (Wodon et desirable areas and poorer households al. 2014). And in at least one documented in less-desirable ones. This is investigated in case, poor people were less exposed: after Nigeria, one of the hottest countries in our Hurricane Mitch struck Honduras in 1998, sample. We run a regression to estimate a more than 50 percent of nonpoor households household’s wealth index conditional on the were affected, but only 22 percent of poor hottest monthly temperature a household households were (Carter et al. 2007). experiences. Including socioeconomic and Our conclusion is that most studies climatic controls, we find a clear signal find that poor people are more exposed that poorer households within Nigeria f igure  3.2). However, the relationship (­ tend to live in hotter (less desirable) areas between poverty and disaster exposure is con- (figure 3.1, panel b). text specific and depends on the type of haz- One problem with studies of exposure at ard, local geography, institutions, and other the national scale is that they may miss mechanisms. 9 0    SHOCK WAVES MAP 3.4  Sub-Saharan Africa’s and Asia’s poor tend to be more exposed to droughts than the nonpoor (Poverty exposure bias for droughts at national level) Source: World Bank (IBRD 41906, September 2015) based on Winsemius et al., forthcoming. Note: Exposure was calculated for the 100-year return period (results similar for other return period events for droughts). MAP 3.5  Poor people in most countries are more exposed to higher temperatures than nonpoor people Source: World Bank (IBRD 41907, September 2015) based on Park et al., forthcoming. T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   91 Poor people lose relatively more to disasters when affected Poor people are often—but not always— more exposed to natural hazards. But what about vulnerability? Do poor people lose more as a result of a disaster? Answering these questions is challenging because of data limitations. While global data are sufficient for examining exposure, they cannot provide an estimate of vulnerability since that also depends on asset portfolios and livelihoods. However, out of the 13 local case studies that examine exposure to a disas- ter by poverty status, five (on Bangladesh, Honduras, and Mumbai) also examine losses for poor and nonpoor people separately (cal- culated as income losses, asset losses, or both) and provide insight on the difference in vulnerability. The results show that in absolute terms, wealthier people lose a larger amount of assets or income because of a flood or storm, which is expected as they have more assets and higher incomes. But in relative terms, poor people always lose more than nonpoor people from floods and storms (figure 3.3). It is these relative losses, rather than absolute numbers, that matter more for livelihoods and welfare. In Bangladesh, one study surveyed 700 floodplain residents living without protec- tion along the Meghna River (Brouwer et al. 2007). The authors collected data on the average flood damage experienced because of floods for households above and below the poverty line. In absolute terms, house- holds above the poverty line lost more: $240 per year, compared to $191 for those below the line. However, poor people lost much more in relative terms: 42 percent of house- hold income compared to 17 percent for nonpoor people. In Honduras, following Hurricane Mitch in 1998, a study investigated losses across wealth quartiles based on a survey of 850 rural households (Carter et al. 2007). Affected households in the bottom quartile lost nearly three times as much in relative terms as other households: 31 percent of their assets for 9 2    SHOCK WAVES FIGURE 3.1  Poor people in hotter countries—like Nigeria—live in reconstructing their homes (Patankar and hotter areas, but in cooler countries, less so Patwardhan, forthcoming). A survey of 1,168 households shows that, while nonpoor people a. Poor people in hotter countries live in had higher absolute losses, poor people lost hotter areas, but in cooler countries less so more as a percentage of income, across all three loss categories (table 3.1). When com- 1.5 bining income, asset, and repairs, the total Poverty exposure bias for temperature losses from the event reached 85 percent of the average annual income of the poorest 1.0 people. These impacts obstructed the ability of households to recover in the aftermath— 0.5 not least because the loss of assets meant many poor households found themselves 0 unable to borrow or repay previous loans (Rentschler 2013). Why is it that poor people lose relatively –0.5 more? For asset loss, poor people hold lower-quality assets and keep the assets in a –1.0 more vulnerable form. For income loss, poor 0 10 20 30 people tend to be more dependent on lower-­ Annual monthly temperature, 1961–99 (°C) quality infrastructure and natural capital to earn an income. They also are vulnerable to b. Nigeria is a good example food price rises, and women and children are especially vulnerable to health impacts. 70 We review each in turn. Household wealth index 60 Poor people hold more vulnerable and lower-quality assets 50 The typical asset portfolio of a poor and a nonpoor person are very different. Poor 40 people tend to have less diversified portfo- lios: they hold a larger percentage of their 30 assets in material form and save “in kind.” The first “savings” of poor urban dwellers 28 30 32 34 are often through investments in their Hottest month temperature experienced by household (°C) home, which are very vulnerable to floods (Moser 2007), while many rural poor use Source: Park et al. forthcoming; World Bank 2015b. livestock as savings in spite of their vulner- Note: Panel a plots country-level poverty exposure bias for temperatures against each country’s current climate. Panel b plots household-level wealth and temperature within a coun- ability to droughts (Nkedianye et al. 2011). try—Nigeria. Nonpoor people, with higher financial access, are able to spatially diversify and save in financial institutions, and their sav- the poor compared to 11 percent for the ings are thus better protected from natural nonpoor. hazards. In Mumbai, the 2005 floods not only In addition, the quality of assets owned by caused direct losses to households’ assets but poor people is lower. Take for instance hous- also meant that the inhabitants lost income ing stock. Households living in slums or and spent large amounts on repairing or informal settlements made out of wood, T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   93 bamboo, and mud on steep slopes will suffer FIGURE 3.2  When disasters hit in the past, poor people were more more damage compared to individuals in likely to be affected (Percentage of poor and nonpoor people affected by a disaster) housing made out of stone or brick. In coastal communities in southwest Bangladesh, fol- Exposure lowing Cyclone Aila, 76 ­ percent of house- holds in “kacha” houses (traditional homes Surveyed households affected by 100 natural disaster (% of total) built with mud and bamboo) reported struc- 80 tural damage—far above the 47 percent for those in “pucca” houses (built with concrete 60 and wood). In terms of economic damage, the 40 average for kacha houses, $400, was also well above the $133 for pucca ones. Further, 20 households in kacha houses were significantly 0 more likely to experience fatality or physical lad 1 at 2 Gu a na Ho iti as rth as a r d i l gu or Vi a m ba n S pa al No e E ny alp Af t an ng sh, Gu sh, injury—on average, 0.28 people per kacha Ha na Te lvad ur M ica ya em um Sa Ne dl e nd cig et Ba ade e K a house were injured or killed from the cyclone, l ng Ba id compared to 0.13 per pucca house. M Poor Nonpoor Source: Based on del Ninno et al. 2001 for Bangladesh (1) and Akter and Mallick 2013 for Bangladesh Poor people depend on fragile (2); Tesliuc and Lindert 2003 for Guatemala; Pelling 1997 for Guyana; Fuchs 2014 for Haiti; Carter et infrastructure and are not well al. 2007 for Honduras; Opondo 2013 for Kenya; Wodon et al. 2014 for MENA; Baker et al. 2005 and Ranger et al. 2011 for Mumbai; Gentle et al. 2014 for Nepal; Fay 2005 for San Salvador and Teguci- protected galpa; and Nguyen 2011 for Vietnam. Note: Each study has a different definition of “poor” and “nonpoor” people; further, exposure differs Besides private income and asset losses, based on the type of hazard and context in which it occurs. natural disasters cause significant disrup- tion to public infrastructure. While all peo- FIGURE 3.3  Poor people always lose relatively more than nonpoor ple, to some extent, depend on electricity, people working roads, and running water to earn (Percentage of assets or income lost for poor and nonpoor people after a disaster) a living, poor people tend to be less able to protect themselves from the consequences Vulnerability of disruptions in infrastructure services. 100 households (% of annual income) Assets or income lost for affected And poor people often rely on more fragile 80 or undermaintained infrastructure—such as unpaved roads that are impractical dur- 60 ing the rainy season, or drainage systems that are insufficient or clogged by solid 40 waste. 20 Another important issue is how infra- structure investments are distributed 0 ­ s patially (Fay 2005; Olsson et al. 2014; ,1 ,2 ,3 s i ba ra sh sh sh u um nd Tschakert 2007). Too often, investments are e e e ad ad ad M Ho l l l ng ng ng directed toward relatively wealthier places, Ba Ba Ba at the expense of poorer neighborhoods. Poor Nonpoor This effect can amplify the exposure gap Sources: del Ninno et al. 2001 for Bangladesh (1); Brouwer et al. 2007 for Bangladesh (2); Rabbani, between poor and nonpoor households and Rahman, and Mainuddin 2013 for Bangladesh (3); Carter et al. 2007 for Honduras; and Patankar generate pockets of high risk. Progress along and Patwardhan, forthcoming, for Mumbai. Note: Each study has a different definition of “poor” and “nonpoor” in its sample. Vulnerability this dimension requires appropriate gover- depends on the type of hazard and context in which it occurs; even within the same country (Ban- nance mechanisms, including giving poor gladesh), vulnerability measures vary greatly based on location and severity of flooding. The first three studies use percent of income loss as a metric, while the Honduras case uses asset loss and making people a voice in investment decision-­ the Mumbai case uses asset, income, and repair loss. For Honduras, the graph reflects asset losses processes (chapter 5). Poor households relative to total assets. 9 4    SHOCK WAVES TABLE 3.1  Poor people in Mumbai suffered higher relative losses from the 2005 floods Income loss Asset loss Repair loss Total loss Average annual Number of Total loss income (Rs.) households (Rs.) as a % of yearly income < 60,000 192 51,000 16 29 40 85 120,000 806 62,000 10 20 22 52 270,000 124 83,000 6 13 12 31 450,000 15 143,000 3 7 21 32 > 540,000 10 104,000 8 4 8 19 Source: World Bank calculation based on Patankar and Patwardhan, forthcoming. Around the time of the flood, 50 Rupees was equivalent to about US$1. Numbers rounded for clarity. sometimes spend a lot of time and effort lob- Chapter 2 discussed the fact that poor peo- bying local authorities to invest in their com- ple, especially in rural areas without func- munities to provide basic infrastructure tioning markets, are highly dependent on (such as roads, piped water, and sanitation) agricultural income and ecosystems, and are (Moser 2007, 85). Nevertheless, households therefore vulnerable to the impacts of cli- with little social capital will be unable to mate change on yields and ecosystems’ “invest” in public goods and improve their health and functioning. Here we focus on quality of life. how this dependency translates into a higher In Mumbai, impacts from a lack of vulnerability to natural hazards. appropriate infrastructure can be pervasive Large-scale events can wreak havoc on (Patankar, forthcoming) (box 3.4). Many natural capital. In 2008, Cyclone Nargis hit low-lying and reclaimed areas across the southwest Myanmar, killing an estimated city get flooded, especially when heavy rains 140,000 people, and recovery is still far from combine with high tide or storm surges, complete (World Bank 2015c). A major rea- with the added difficulties due to unsanitary son is the damage to embankments and methods of solid waste and sewage disposal streams from the cyclone, which resulted in a and problems with the drainage systems. reinforcing chain of events for affected farm- After flooding, more than 75 percent of sur- ers. Erosion and destroyed embankments veyed households report electrical disrup- made fields more prone to flooding. Further, tions, a lack of local transport and clean the duration of daily and monthly tides drinking water, and sewage and garbage in became longer after Nargis, making fields their homes—all of which magnified the more saline and prone to pest infestation. impacts of floods (Patankar and Without funds for repair, affected farming Patwardhan, forthcoming). Although many villages became more prone to these external people were affected, poor people were the events—flooding, saline intrusion, and pest ones with fewer options to cope with infra- infestation. As a result, yields decreased, as structure damage. did income. Households have attempted to borrow money but this has led only to more indebtedness. Many poor people depend on Furthermore, natural capital often serves agricultural and ecosystem incomes as a safety net after a disaster, when not that are particularly vulnerable to depleted (Barbier 2010). In Bangladesh after hazards Cyclone Aila hit in 2009, households living Another source of vulnerability is the reli- closest to the coast, while more exposed and ance on agricultural and ecosystem incomes. vulnerable to the storm (and poorer), had a T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   95 BOX 3.4  Hidden costs of recurrent hazards for poor people in Mumbai and Ho Chi Minh City Large-scale events make the news, but repeated small One implication is that households lose workdays: adverse events such as regular floods often have seri- on average 2.5 per year because of poor infra- ous implications for poor people, affecting their live- structure (more than 50 percent cite unavailabil- lihoods and their ability to accumulate assets. To get ity of transport or flooded roads)—implying a loss a better sense of the “hidden costs” of such events, of income and productivity and sometimes jobs. take the following two cases. Second, almost 40 percent of households report Recurrent floods in Mumbai. Mumbai is prone someone in the family experienced health impacts to recurrent floods during the monsoon season, with from diarrhea yearly due to floods, with this fig- significant impacts on poor people (see a background ure rising to 64 percent for malaria and 86 percent paper for this report, Patankar, forthcoming). Based for viral fever. Between 2001 and 2011, the num- on the experience of recurrent floods, the authori- ber of reported cases of malaria has increased by ties have identified 40 chronic flood spots (low-lying 217 ­p ercent, mainly due to lack of sanitation in areas) and 200 localized flood spots (where water- slums and water accumulation during the monsoon logging is due to inadequate drainage and poor land season (Public Health Department 2015). use planning). When we combined this spatial data Recurrent floods in Ho Chi Minh City. A survey with land use maps (Planning Department 2015), we of three flood-prone districts in Ho Chi Minh City found that land use in the flood-prone wards sug- finds health impacts to be pervasive (World Bank gests an unplanned mix of residential, commercial, and Australian AID 2014). Regular floods in a heav- and industrial activities coexisting without clear ily polluted environment have led to many a­ ilments— zoning. As a result, recurrent floods expose a large including skin and intestinal diseases, rheumatism, number of residents, including those in the many bronchitis, and chronic coughing, especially among low-income slum settlements, who report floodwa- children under five. Every year, more than two- ters entering their houses many times during the thirds report that they are suffering from health monsoon season. issues, with more than half suffering from a water- A survey of 200 households yields two key borne (55 percent) or respiratory disease (52 ­percent) insights. First, households regularly report prob- directly related to local flood conditions. These lems with transport, drinking water, power supply, impacts also take a significant toll on employment and food and fuel availability because of the floods. and income, especially for poor people (table B3.4.1). TABLE B3.4.1  The health of Ho Chi Minh City’s poor is especially vulnerable to flood impacts Indicator Total Poor (n = 36) Nonpoor (n = 210) % households whose health was affected 68 86 64 % households whose employment was affected 58 69 56 % households whose income was affected 44 67 40 Source: Based on World Bank and Australian Agency for International Development 2014. more resilient income because the proximity paper for this report (Noack et al., forthcom- to mangrove reserves offered higher income- ing), climate change impacts on these ecosys- generation opportunities than for inland tems may impair their ability to serve as a inhabitants (Akter and Mallick 2013). As safety net and to smooth consumption in the stressed in chapter 2 and in a background face of shocks. 9 6    SHOCK WAVES Poor people are more vulnerable to agricultural land, decimating production and rising food prices after a disaster sending prices of wheat upward of 50 percent above the preflood level. Another point that was made in chapter 2 is How does poverty fit into the picture? In that poor people in developing countries Bangladesh, after the 1998 Great Flood, a spend on average between 40 and 60 percent study shows that consumption levels differed of their household budget on food—far more based on exposure and poverty status (del than the 25 percent spent by nonpoor peo- Ninno et al. 2001). There was no difference ple. This makes them more vulnerable than in calorie consumption between exposed and the rest of the population to increases in nonexposed households in the top quintile, food prices (although net food producers but in the bottom quintile the difference was could gain, if they can maintain their pro- 11 percent. For those exposed, bottom-­ duction level). Here, we show that this vul- quintile households on average consumed nerability matters in postdisaster situations. 1,400 calories per capita, and 80 percent fell After tropical Storm Agatha struck below the minimum daily caloric require- Guatemala in 2010, per capita consumption ment of 1,800; however, the average calories fell 13 percent, raising poverty by 18 percent; consumed for exposed top-quintile house- in particular, food expenditures fell holds remained above 3,000. In addition, 10 ­percent, accounting for 40 percent of the two-thirds of bottom-quintile households total consumption drop (Baez et al. 2014). spent more than 70 percent of their budget This stemmed from a major loss in food infra- on food. As a result, 48 percent of the house- structure and transport, resulting in a holds in the poorest quintile were deemed 17 ­percent increase in food prices 10 months food insecure after the flood, compared to after the storm. Agatha thus caused a logisti- an average of 16 percent across all quintiles cal problem rather than a decline in domestic (table 3.2). production, since it occurred in the middle of the first planting season, at a benign time with respect to local agricultural cycles. Children are particularly vulnerable to Natural disasters can also result in food indirect impacts through health and price spikes as a result of supply shocks. education Disasters can destroy crops and seed reserves, destroying productive assets in agricultural Building human capital through better health communities and sparking food price shocks, and education is a vital component of escap- as occurred after the unprecedented 2010 ing poverty, but natural disasters can worsen floods in Pakistan (Cheema et al. 2015). The health and education outcomes, especially floods destroyed 2.1 million hectares of for children (chapter 1). TABLE 3.2  Bangladesh’s poor became food-insecure after the 1998 Great Flood (Percentage of affected households reporting food security impacts by expenditure quintile) Quintiles Poorest Q2 Q3 Q4 Richest All Spending more than 70% on food 66 59 66 51 0.4 49 Below minimum caloric requirement 80 50 25 13 0.1 35 Food insecure 48 17 9 0 0.9 16 Source: del Ninno et al. 2001. Note: Numbers rounded for clarity. T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   97 There are acute health effects on children could improve the resilience of poor people. from the direct impact of disasters and lower The benefits of these actions could be signifi- postdisaster consumption, especially after cant, even without man-made climate droughts. In Ho Chi Minh City, in the Thanh change. An increase in the frequency or Xuan Ward of District 12, a majority of intensity of natural hazards due to climate ­ children experience fevers, coughing, and change would make these benefits even flu during a high-tide period (World Bank larger, provided that policies and measures and Australian AID 2014; and box 3.4). are designed to account for climate change Following weather shocks in Sub-Saharan and the uncertainty it creates. Africa, asset-poor households provide chil- dren with lower-quality nutrition and are less Risk-sensitive land use regulations: likely to take sick children for medical Critical but challenging to implement consultations, with long-term impacts on ­ child development and prospects (Alderman, A major reason why poor people often live Hoddinott, and Kinsey 2006; Dercon in riskier areas is cheaper housing. In Ho Chi and Porter 2014; Jensen 2000; Yamano, Minh City, qualitative surveys suggest Alderman, and Christiaensen 2005). flooded areas can be much cheaper than Impacts on education are also prevalent. In nonflooded areas for the same quality of Africa, children affected by droughts are less accommodation (World Bank and Australian likely to complete primary school (Alderman, AID 2014). In addition, recent experience of Hoddinott, and Kinsey 2006; Dercon and a flood can reduce housing prices by around Porter 2014), and similar impacts have been 9 percent (Husby and Hofkes 2015).1 found in Asia, Latin America, and elsewhere In developing countries with informal mar- (Baez, de la Fuente, and Santos 2010; Maccini kets, land scarcity can be particularly acute and Yang 2009). and land markets function poorly (Durand- Moreover, women are particularly vul- Lasserve, Selod, and Durand-Lasserve 2013). nerable as they often take greater responsi- In these places, it may not be the prices that bility for household chores, increasing their push poor people into risky places, but simply hardships during floods. This is in addition the availability of land with appropriate to time taken off work (sometimes for a cou- access to jobs and services. Informal settle- ple of months) to care for children who ments are often located in hazard-prone loca- become sick because of living in flood condi- tions, such as on hill slopes, close to tions, which can be especially punitive for riverbanks, or near open drains and sewers, factory workers. Women also spend more as in Pune, Dhaka, Caracas, Rio de Janeiro, time at home to clean after a flood, making and Mumbai (Lall and Deichmann 2012; them more likely to contract waterborne Lall, Lundberg, and Shalizi 2008; World diseases. Bank 2007). Land use regulations can help by ensuring that new development occurs in places that The reasons why poor people are safe or easy and cheap to protect. They are more at risk point to possible can also avoid unchecked urban development that leaves too little porous green space and policy solutions further increases runoff and flood risk (Lall So, if poor people are disproportionately and Deichmann 2012). But doing so remains affected by disasters—and here the evidence challenging for a number of reasons. is compelling—what can be done to make First, countries need appropriate data on them less exposed or vulnerable? This sec- risk and hazard to identify places that are too tion builds on the insights of the previous risky to develop, or where development is one to identify six examples of policies that possible provided that buildings and 9 8    SHOCK WAVES infrastructure are built following strict rules. are the considerably cheaper rents in risky Unfortunately, access to risk information still areas and proximity to work (usually in facto- varies greatly and is quite limited in low- ries) for late return at night. Thus, to be effec- income environments. To address this issue, tive, flood zoning should be accompanied by the World Bank and the Global Facility for investment in transport infrastructure to Disaster Reduction and Recovery (GFDRR) make it possible for people to settle in safe are investing in risk information. The places while maintaining access to the same GFDRR’s Open Data for Resilience Initiative (or comparable) jobs and services. supports the creation of GeoNode, a web- Fourth, countries need to remember that based open source platform that makes it land use regulations can have unintended easier to develop, share, manage, and publish consequences, particularly for poor people. geospatial data (www.geonode.org). Such ini- Restrictive flood zoning policies can increase tiatives can make a difference locally, by mak- housing costs, making it more difficult for ing risk information freely available not only rural poor people to move to cities and cap- to professionals but also to the public. ture the opportunities of an urban life (like Second, countries need strong institutions better-paying jobs and better health care and that can ensure that land use plans are actu- education). Restrictive policies can also ally enforced. In most of the world today, worsen risks. In Mumbai, because of strict risk-sensitive land use plans face strong politi- regulations, buildings have been held to cal economy obstacles, and are only rarely between a fifth and a tenth of the number of enforced (World Bank 2013, chapter 2). One floors allowed in other major cities (Lall and of the main obstacles is the asymmetry Deichmann 2012). The resulting low-rise between the costs and benefits of risk-­sensitive topography contributes to land scarcity, land use planning. The costs of flood zoning higher housing prices, and slum formation, are immediate, visible, and concentrated, in including in flood zones. the form of reduced land values for landown- ers and higher housing costs for tenants More resilient infrastructure and (Viguie and Hallegatte 2012). In contrast, the protection systems that serve poor benefits occur through avoided losses—which people nobody can see—sometimes in the future, and go to unknown people. In such a context, the Poor people suffer from frequent disasters opponents to flood zoning are usually vocal because they lack the type of protective and well organized while beneficiaries are infrastructure that is common in wealthier absent, making the policies difficult to pass countries. As described in box 3.2, lower and enforce. protection levels are the main reason why Third, countries need to design land use flood risks are higher in relative terms in plans in a way that accounts for the reasons poor than in rich coastal cities (Hallegatte why people decide to live in risky places—­ et al. 2013). And the difference is even more primarily, access to jobs and services. obvious within cities: for instance, poor When asked what it would take to consider households are often exposed to recurrent relocating to a safer, less flood-prone area, floods because of the lack of infrastructure, 44 percent of households in Mumbai cited or its poor condition, especially drainage transport, along with the availability of health systems (box 3.4). Solving this problem services, schools, and social networks requires investing more and investing (Patankar, forthcoming). In Ho Chi Minh better. City, local and migrant households do not Investing more. Governments in both have any plans to move despite high flood developed and developing countries already risk and health impacts (World Bank and struggle to finance infrastructure. Millions of Australian AID 2014). The reasons, accord- people in developing countries still lack access ing to most of the 246 survey respondents, to safe water, improved sanitation, electricity, T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   99 and transport. Even disregarding climate con- The challenge of addressing the long-term cerns, developing countries need substantially risks from climate change in development more infrastructure to grow and address pov- projects is multifaceted. First, there is high erty, inequality, and unemployment uncertainty as to how global climate change concerns. will translate into local changes in environ- Little data exist on how much is being mental conditions, especially for extreme spent on infrastructure, but the World Bank events. Second, climate change is often an Group estimates that at least $1 trillion per exacerbating factor of other development year would be needed in developing countries stressors (such as poverty, urbanization, to close the infrastructure gap, with about water degradation, increasing population, $100 billion for Africa alone. resource use, and existing natural hazards). This lack of infrastructure is an obvious Third, if investment in disaster risk reduction multiplier of natural hazard consequences, or climate adaptation is designed to maximize and one that could be closed through economic returns, it will be concentrated increased investments. But infrastructure toward areas with highest asset values—that does not attract enough capital, especially in is, toward wealthier groups (Füssel 2012; developing countries: long-term, largely illiq- Tschakert, forthcoming). uid investments are not perceived as attrac- Fortunately, there are innovative ways to tive destinations for global capital. Many manage the long-term, uncertain risks of cli- countries are simply too poor to generate mate. These approaches seek to identify domestically the needed pool of savings. robust decisions (those that satisfy decision Many others lack local capital markets that makers’ multiple objectives in many plausible are sufficiently developed to transform local futures and over multiple time frames) rather liquidity into the patient capital that is needed than being optimal in any single best estimate for longer-term investments. Further, public of the future (Bonzanigo and Kalra 2014; spending is limited by a low tax base (10–20 Kalra et al. 2014; Lempert et al. 2013). percent of GDP in many countries) and low Decision making under uncertainty starts debt ceilings. with the options available—from infrastruc- Recommendations typically include lever- ture to early warning systems—and does not aging private resources to make the most of attempt to predict the most likely future(s). available capital, which involves well-known The performance of each option is then tested steps like improving the investment climate against many different possible future condi- (making sure regulations are clear and pre- tions to identify its vulnerabilities. Those dictable and the rule of law and property future conditions include climate, political, rights are enforced), developing local capital and socioeconomic risks. markets, and providing a pipeline of “bank- At that point, it becomes possible to evalu- able” projects (Fay et al. 2015). Official ate the trade-offs among the different options development assistance (ODA) can play a (using different measures of success, like eco- catalytic role in mobilizing additional nomic return, number of people benefiting, resources, but it is constrained by donors’ fis- whether poor or nonpoor people are the main cal constraints and remains limited relative to beneficiaries) and to identify policies that overall needs—at its highest around 2011, it reduce the vulnerability of future investments. reached about $90 billion. Often, these methods favor soft and flexible Investing better. New and additional options over hard ones—including monitor- investments will reduce the long-term vulner- ing systems to make sure risks are systemati- ability of the population—and especially the cally assessed throughout the life of the poorest—only if new infrastructure is project, so that solutions can be adjusted over designed so that it can absorb climate change time. They also encourage decision makers to and remain efficient in spite of changes in cli- look beyond within-sector interventions, and mate and environmental conditions. combine prevention and reactive actions 1 0 0    SHOCK WAVES BOX 3.5  In an uncertain future, developing into the wetlands of Colombo is dangerous Colombo faces recurrent floods that largely affect The analysis determined that, if all urban wet- low-income populations and threaten the city’s long- lands across the Colombo catchment were lost, in term development. Its urban wetlands have been some scenarios the metropolitan area would have identified by local agencies as a critical component to cope with annual average flood losses of about of its flood protection, but wetlands have declined 1 percent of Colombo’s GDP in the near future, with rapidly in recent years because of continuous infill- significant impacts on the poorest populations—a ing, unmanaged development, and land dredging for level of risk considered unacceptable by local deci- lakes. In collaboration with government agencies, sion makers and stakeholders. NGOs, and local universities, a World Bank analysis For long-term strategies, trade-offs between was conducted to examine the value of Colombo’s urban development, lake creation, and wetland con- urban wetlands in the short term and long term servation were weighed, with active management of and to identify the most viable strategies available urban wetlands emerging as the lowest regret option. to increase the city’s flood resilience in an uncertain The analysis also found that, faced with ­ c limate future (in terms of climate change, urban develop- change and fast urban development, wetlands would ment patterns, and the resilience of poor communi- not be sufficient to protect Colombo against severe ties). This involved the use of numerous hydrological floods. This means that proactive urban planning and socioeconomic scenarios as well as the evalua- and land use management are essential to protect tion of some wetlands benefits (such as ecosystem existing wetlands, provide incentives for vulnerable services for the populations who fish in the wetlands, populations to move to safe areas, and reduce future wastewater treatment, or recreational services). exposure of people and assets. within a consistent strategy. Projects follow- the communities they serve more resilient to ing these methodologies that are being piloted natural hazards, with a strong focus on the by the World Bank include water supply in enforcement of building norms. It supports a Lima, flood risk management in Ho Chi Minh safety diagnostic of schools in Lima, Peru, City and Colombo (box 3.5), hydropower and provides technical assistance in investment in Nepal, and road network resil- Mozambique to optimize the delivery of ience in Peru and across Africa. resilient schools at the local level—targeting Finally, better-designed infrastructure and both government and community construc- public investment will translate into reduced tion. Similar actions exist in other sectors. vulnerability only if infrastructure designs are For example, the World Health Organization, respected during the construction phase. the International Strategy for Disaster Studies show that most of the deaths after Reduction, and the World Bank partnered in earthquakes occur in countries with a high 2008 in the “Safe Hospitals” initiative to level of public sector corruption, where build- help health facilities withstand natural ing norms are not enforced, and where public shocks. buildings are often not built according to the designed standards (Ambraseys and Bilhan Improved property rights to incentivize 2011; Escaleras, Anbarci, and Register 2007). resilience investments The same is likely true for climate-related disasters such as floods and storms, although One reason why poor people lose a larger data are not available. share of their assets and income is that they The Global Program for Safer Schools, live in buildings with low resistance to natu- created by the World Bank and the GFDRR, ral hazards. In Latin America, a 1993 inven- aims at making school facilities and tory found 37 percent of its housing stock T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   101 TABLE 3.3  Mumbai’s poor spend a lot to regularly repair their dwelling (Share of households undertaking recurrent measures to protect against flooding, by income group)   Very poor (%) Poor (%) Nonpoor (%) Repairing roof (Rs. 1,300) 50 35 25 Repairs inside house (Rs. 800) 35 22 14 Overhauling vehicle (Rs. 600) 5 6 5 Cleaning house surroundings (Rs. 200) 70 56 68 Cleaning nullah (Rs. 200) 50 48 57 Source: Patankar, forthcoming. Note: Average cost, in Rs. (rupees), of each measure is shown. Very Poor: Rs. 5,000 and less in monthly income; Poor: between Rs. 5,000 and Rs. 10,000; Nonpoor: above Rs. 10,000. Numbers rounded for clarity. A nullah is a stream or waterway. provided inadequate protection against with low levels of household income, is the disaster and illness (Fay 2005). Since then, main reason for underinvestment in housing rising trends in urbanization, settlements in infrastructure, according to a survey of two risky areas, and the low quality of those set- informal settlements without tenure security tlements have likely increased this share (Lall (van Gelder 2010). In contrast, in Tanzania, and Deichmann 2012). as ­figure 3.4 shows, households with home In Mumbai, both poor and nonpoor ownership (and especially those holding households undertake short-term and recur- some form of documentation) invest signifi- rent measures to reduce the intensity of cantly more in their dwelling (Rentschler flooding in their premises (Patankar, forth- 2013). coming). This includes cleaning surround- The lesson here is that better tenure secu- ings and gutters choked with garbage, rity encourages investment in housing, includ- repairing leaking roofs, overhauling vehi- ing risk reduction. In Peru, starting in 1996, cles, and house repairs. But the difference the government issued property titles to over between poor and nonpoor is visible in the 1.2 million urban households, which at the type of action undertaken (table 3.3). Poor time was the largest titling program targeted people undertake repairs for roofs and at urban squatters in low-income countries. houses in larger numbers than the nonpoor, A study on the impact of this program on because their houses are made of more vul- housing renovations found that households in nerable material. Such repairs have to be program neighborhoods invested significantly undertaken annually, are financed without more than those in nonprogram ones (Field much support from the government, and 2007). In addition to better housing, access to end up being more costly than building high- services (water) rose, and crowding was quality roofs in the first place. These reduced as households enlarged their homes expenses place another financial burden on and increased the number of rooms—thereby poor households. stimulating the rental market (Mosqueira The lack of clear and effectively enforced 2003). land and property rights discourages poor households from making more robust and Efficient and sustainable durable—but also costlier—investments. air-conditioning to reduce vulnerability Facing the permanent risk of eviction, they to extreme heat are unlikely to invest in the physical resilience of their homes (like retrofitting to strengthen In May 2015, a major heat wave swept across homes against disasters) (Rentschler 2013). India, with temperatures hitting highs of In Buenos Aires, the fear of eviction, along 118°F (48°C) in some parts. Official statistics 1 0 2    SHOCK WAVES FIGURE 3.4  Home ownership in Tanzania encourages home likely to die compared to people with access investment to air-conditioning (Semenza et al. 1996). In (Average annual expenditure on repairs and improvement of dwelling by ownership fact, the strongest protective factor found was category) air-conditioning: more than 50 ­ percent of the 200,000 deaths related to the heat wave could have been prevented if each home had a working air con- ditioner. And a meta-analysis of heat wave Tanzania shilling (T Sh) 150,000 studies finds working home air-­ conditioning reduces the odds of death by 23 to 34 percent 100,000 (Bouchama et al. 2007). Thus, access to air-conditioning, which 50,000 implies reliable electricity production and dis- tribution, could be a critical tool to reduce 0 health impacts from heat waves—but only if Owner occupied- Owner occupied- Rented Some form of No documentation it reaches the most vulnerable segments of documentation of of ownership the population. In France, programs like the ownership National Heat Wave Plan set up after the Repairs Improvements 2003 heat wave are designed to provide air- conditioning shelters in community centers Source: Rentschler 2013. and in senior citizen homes to reach these Note: Based on household survey data (Living Standards Measurement Surveys). US$1 is equal to about 20,000 T Sh. populations. For those working outdoors, air- conditioning is unlikely to help much. Thus, adaptations such as flexible work hours (like reported more than 1,100 deaths (Al Jazeera not working during direct sunlight) and 2015). Elderly people, as in most heat waves, shorter shifts may become necessary in more were among the most vulnerable, along places. with low-income workers, employed out- One caveat is that increasing air-conditioning doors, in jobs from rubbish collection to is likely to have a significant environmental farming and construction. In the state of cost. First, air-conditioning in buildings Andhra Pradesh, which experienced the increases street temperatures, increasing the greatest impacts from the heat wave, a potential impact on homeless persons or those majority of the 900 reported victims were working outside. Second, air-conditioning elderly or low-income workers (Al Jazeera consumes energy: stabilizing climate change 2015; Vice News 2015). Homeless people, will thus require that air-conditioning equip- who are unable to find shelter, are also ment be extremely efficient and electricity low among the most vulnerable: according to a carbon (zero carbon in the longer term). Delhi-based nongovernmental organization A recent study for Mexico tried to quantify (NGO), of the 186 people who died in the the potential climate impacts from air-­ capital, 80 percent were homeless (Vice conditioning, drawing on household-level News 2015). But these figures may underes- data (Davis and Gertler 2015). It finds that on timate the death toll, since reliable statistics hot days there is a large increase in electricity are difficult to find (The Economist 2015). consumption: above temperatures of 21°C, In Chicago, a lack of air-conditioning was a usage increases nonlinearly; and, for every critical risk factor for death after the 1995 heat additional day above 32°C, usage rises by wave, which resulted in over 700 deaths, con- 3.2 percent. centrated among the poor and elderly The negative impacts of air-conditioning ­ populations (Whitman et al. 1997). People who can be mitigated by additional measures. For did not have a working air conditioner, access instance, urban planning, improved housing to an air-conditioned lobby, or visited an air- quality, highly reflective materials for roads conditioned place were 20–30 percent more and buildings, and irrigated parks can T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   103 minimize heat waves and reduce energy FIGURE 3.5  Poorer people lack sufficient access to financial demand for air-conditioning (Chapman et al. instruments (Fraction of poor and nonpoor people with savings at a financial institution) 2009; Masson et al. 2013 and 2014). 80 Improved financial inclusion and 70 Access to savings (% of population) savings options 60 In most countries, poor people suffer from lower access to finance than nonpoor people 50 (figure 3.5), often forcing them to save 40 “in kind.” Fortunately, in the past decade, 30 an alternative method of extending banking services has developed: mobile money. Most 20 adults in the world today—poor people 10 included—have access to mobile phones: the United Nations estimates that out of 7.3 bil- 0 0 5,000 10,000 15,000 20,000 25,000 lion people, 6 billion have access to these GDP per capita (US$, PPP 2011) devices. Mobile money accounts, by provid- ing more convenient and affordable finan- Poor Nonpoor cial services, offer promise for reaching Source: Data from FINDEX. unbanked adults traditionally excluded from Note: Each dot represents poor people or nonpoor people in one country. PPP = purchasing the formal financial system—such as women, power parity. poor people, young people, and those living in rural areas (Demirguc-Kunt et al. 2015). As such, the expansion of mobile money has people, less than 0.4 percent the death toll the potential to improve parity in financial from the 1999 cyclone. Close to a million inclusion and to make the savings and asset people were evacuated to cyclone shelters, portfolio of poor people less vulnerable to safe houses, and locations inland in Odisha natural hazards. (For a fuller discussion on (around 850,000) and in Andhra Pradesh financial inclusion, see chapter 5). (around 150,000). This success was made possible by years of effort from the Odisha State Disaster Management Authority Better observation systems, early (OSDMA) and the government of Odisha— warning, and evacuation planning thanks to planning, construction of disaster Early warning and disaster preparedness can risk mitigation infrastructure, setting up of save lives and reduce economic losses. evacuation protocols, identification of Weather forecasts enable the anticipation of, potential safe buildings to house communi- and preparation for, extreme events. ties, and, most important, working with The value of preparedness was illustrated communities and local organizations to set when Cyclone Phailin made landfall in the up volunteer teams and local champions State of Odisha, India, on October 12, who knew what needed to be done when the 2013, around 9:15 pm with wind speeds of time came to act. around 200 km/hour.2 The storm that hit Preparing a house before a hurricane (by the same coastline 14 years before, in 1999, shuttering windows, for example) can reduce Cyclone 05B, caused massive devastation, damage by up to 50 percent (Williams 2002). killing more than 10,000 people and After the Elbe and Danube floods in 2002, destroying housing and public infrastruc- studies show that 31 percent of the popula- ture in coastal Odisha. This time around, tion in flooded areas implemented preventive however, the story unfolded differently. measures (Thieken et al. 2007; Kreibich et al. After 72 hours, the official death toll was 38 2005). These measures included moving 1 0 4    SHOCK WAVES goods to the second floor of buildings, mov- subdistrict of Shyamnagar in Bangladesh, ing vehicles outside the flood zone, protecting only 15 percent of nonpoor people and 6 per- important documents and valuables, discon- cent of poor people attend cyclone prepared- necting electricity and gas supplies and ness training (Akter and Mallick 2013). In the unplugging electric appliances, and installing Lamjung district of Nepal, the penetration of water pumps. Warning timing was critical: early warning in flood and landslide-prone businesses that protected their equipment or communities is lower than 1 percent (Gentle inventories were those that received the warn- et al. 2014). In Mumbai, levels of early warn- ing early enough. One study estimates that a ing are also paltry, with only 10 percent of warning emitted 48 hours before a flood the surveyed households reporting receiving enables the overall damage to be reduced by some form of early flood warning. These more than 50 percent (Carsell, Pingel, and shortfalls highlight the challenges and the Ford 2004). opportunities associated with building hydro- Yet, in spite of these large benefits meteorological institutions and systems that (Hallegatte 2012b), early warning and evacu- could produce actionable warnings (box 3.6) ation systems are still underdeveloped. In the (Rogers and Tsirkunov 2013). BOX 3.6  Reversing the degradation of hydrometeorological services Over the past 15–20 years, the situation of many hydromet agencies were more developed. And the hydrometeorological services in developing coun- ability to monitor local climate and increases in nat- tries has worsened, primarily because of underfund- ural risk has eroded, making developing countries ing, low visibility, economic reforms, and in some less able to anticipate and adapt to climate change. instances military conflict. As a result, many hydro- Globally, more than 100 countries—over half meteorological services do not function well, with of which are in Africa—need to modernize their some lacking the capacity to provide even a basic hydrometeorological services. How much will level of service. Observation networks have deterio- modernization cost? A conservative estimate of high- rated, technology is outdated, modern equipment priority investment needs in developing countries and forecasting methods are lacking, the quality exceeds $1.5–$2.0 billion. In addition, a minimum of services is poor, support for research and devel- of $400 –500 million per year will be needed to opment is insufficient, and the workforce has been support operations of the modernized systems eroded. (staff costs plus operating and maintenance costs). In Central Asia, for example, observation systems National governments should cover these recurrent deteriorated dramatically after 1985. In the Kyrgyz costs, but few are ready to do so. Moreover, the Republic, the number of meteorological stations has amount of international support for the national been cut by 62 percent, and in Tajikistan the num- hydrometeorological services is significantly below ber of hydrological stations and posts has been cut what is needed just for the high-priority items. by 41 percent. In both of these countries and Turk- It has been estimated that upgrading all hydrome- menistan, upper-air observations—which are very teorological information and early-warning capacity important for forecasting but are expensive—have in developing countries would save an average of been completely abandoned. These trends are also 23,000 lives annually and would provide between observed in the rest of the world. $3 billion and $30 billion per year in additional eco- As a result, substantial human and financial losses nomic benefits related to disaster risk reduction. have occurred, which could have been prevented if Source: Rogers and Tsirkunov 2013. T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   105 In conclusion role in mitigating the consequences of natural disasters—namely the provision of better The fact that disasters are often followed by a health services and care, and universal health measurable increase in poverty is not a sur- coverage. prise, considering the findings of this chapter: • Poor people are often highly exposed to Notes natural hazards, and at the local level they are often more exposed than their wealth- 1. A review of empirical studies finds that the ier neighbors. range of prices between flood-exposed and non-flood-exposed houses varies widely; a • Poor people lose relatively more from meta-analysis of 37 studies mostly in rich disasters because their livelihood and countries finds a spread of −7 percent to +1 asset portfolio is more vulnerable. percent (Beltran, Maddison, and Elliott 2015). • The measures and policies that could 2. This illustration is from Saurabh Dani and be mobilized to help poor people man- http://www.worldbank.org/en/news/feature​ age natural risks in a changing climate /2013/10/17/india-cyclone-phailin-destruction​ amount to “good development,” with -preparation. the important caveat that the design of such measures and policies needs to take into account climate change and the References ­ u ncertainty it creates. All of these meas- Akter, S., and B. Mallick. 2013. “The Poverty– ures also face significant implementation Vulnerability–Resilience Nexus: Evidence from challenges​ — but with climate change Bangladesh.” Ecol. Econ. 96: 114–24. expected to increase risk in many places, Alderman, H., J. Hoddinott, and B. Kinsey. 2006. it will become even more urgent to make “Long-Term Consequences of Early Childhood meaningful progress with these measures. Malnutrition.” Oxf. Econ. Pap. 58: 450–74. Al Jazeera. 2015. “Poor Bear Brunt as India One important implication of these find- Heatwave Death Toll Tops 1,000.” ings is that the differential in risk levels Ambraseys, N., and R. Bilham. 2011. “Corruption between poor and nonpoor people may create Kills.” Nature 469: 153–55. a decoupling between aggregate growth and Baez, J. E., A. de la Fuente, and I. Santos. 2010. poverty reduction, meaning that the poverty “Do Natural Disasters Affect Human Capital? An Assessment Based on Existing Empirical reduction impacts of growth may decrease. Evidence.” IZA Discussion Paper No. 5164, Indeed, where the wealthier can protect them- Institute for the Study of Labor (IZA). selves against disasters, natural hazards are Baez, J., L. Lucchetti, M. Salazar, and M. Genoni. unlikely to lead to a visible shock on GDP. 2014. Gone with the Storm: Rainfall Shocks However, the impact on poorer households, and Household Wellbeing in Guatemala. which does not appear in aggregate statistics, Washington, DC: World Bank. may nonetheless lock them into poverty traps, Baker, J., R. Basu, M. Cropper, S. V. Lall, and creating or magnifying regional pockets of A. Takeuchi. 2005. “Urban Poverty and poverty. Transport: The Case of Mumbai.” Policy One of the most striking results is the vul- Research Working Paper, World Bank, nerability of children to natural disasters. Washington, DC. Barbier, E. B. 2010. “Poverty, Development, This calls for targeted efforts to protect chil- and Environment.” Environ. Dev. Econ. 15: dren in poor families from disasters and avoid 635–60. irreversible consequences for their lifelong Beltran, A., D. Maddison, and R. Elliott. 2015. prospects. The next chapter expands this dis- “Is Flood Risk Capitalised in Property cussion to look at health shocks that poor Values? A Meta-analysis Approach from the people, and especially children, are exposed Housing Market.” Paper presented at EAERE to, along with measures that could play a big (European Association of Environmental and 1 0 6    SHOCK WAVES Resource Economists), Helsinki, Finland, Exposed to Shocks?” J. Dev. Econ. 79: June 24–27. 349–73. Bonzanigo, L., and N. Kalra. 2014. “Making del Ninno, C., P. A. Dorosh, L. C. Smith, and Informed Investment Decisions in an D. K. Roy. 2001. “The 1998 Floods in Uncertain World : A Short Demonstration.” Bangladesh Disaster Impacts, Household Policy Res. Work. Pap. 6765, World Bank, Coping Strategies, and Response.” Research Washington, DC. Report No. 122, International Food Policy Bouchama, A., M. Dehbi, G. Mohamed, Research Institute. F. Matthies, M. Shoukri, and B. Meanne. 2007. Demirguc-Kunt, A., L. Klapper, D. Singer, and P. Van “Prognostic Factors in Heat Wave–Related Oudheusden. 2015. The Global Findex Database Deaths: A Meta-analysis.” Arch. Intern. Med. 2014: Measuring Financial Inclusion around the 167: 2,170–76. World. Washington, DC: World Bank. Bouwer, L. M. 2013. “Projections of Future Dercon, S. 2004. “Growth and Shocks: Evidence Extreme Weather Losses under Changes in from Rural Ethiopia.” J. Dev. Econ. 74: Climate and Exposure.” Risk Anal. 309–29. Brouwer, R., S. Akter, L. Brander, and E. Haque. Dercon, S., and C. Porter. 2014. “Live Aid 2007. “Socioeconomic Vulnerability and Revisited: Long-Term Impacts of the 1984 Adaptation to Environmental Risk: A Case Ethiopian Famine on Children.” J. Eur. Econ. Study of Climate Change and Flooding in Assoc. 12: 927–48. Bangladesh.” Risk Anal. Off. Publ. Soc. Risk Durand-Lasserve, A., H. Selod, and M. Durand- Anal. 27: 313–26. Lasserve. 2013. “A Systemic Analysis of Carsell, K. M., N. D. Pingel, and D. T. Ford. 2004. Land Markets and Land Institutions in West “Quantifying the Benefit of a Flood Warning African Cities: Rules and Practices—The Case System.” Nat. Hazards Rev. 5: 131–40. of Bamako, Mali.” Policy Res. Work. Paper, Carter, M. R., P. D. Little, T. Mogues, and World Bank, Washington, DC. W. Negatu. 2007. “Poverty Traps and Natural Elbers, C., J. W. Gunning, and B. Kinsey. 2007. Disasters in Ethiopia and Honduras.” World “Growth and Risk: Methodology and Micro Dev. 35: 835–56. Evidence.” World Bank Econ. Rev. 21: 1–20. C h a p m a n , R . , P. H o w d e n - C h a p m a n , Escaleras, M., N. Anbarci, and C. Register. H. Viggers, D. O’Dea, and M. Kennedy. 2009. 2007. “Public Sector Corruption and Major “Retrofitting Houses with Insulation: A Cost- Earthquakes: A Potentially Deadly Interaction.” Benefit Analysis of a Randomised Community Public Choice 132: 209–30. Trial.” J. Epidemiol. Community Health 63: Fay, M. 2005. The Urban Poor in Latin America. 271–77. Directions in Development Series. Washington, Cheema, I., S. Hunt, M. Jakobsen, M. Marzi, DC: World Bank. S. O’Leary, and L. Pellerano. 2015. Citizen’s Fay, M., S. Hallegatte, A. Vogt-Schilb, J. Rozenberg, Damage Compensation Programme: U. Narloch, and T. Kerr. 2015. Decarbonizing Impact Evaluation Report . Oxford Policy Development: Three Steps to a Zero-Carbon Management. Future. Washington, DC: World Bank. Cole, S., X. Gine, J. Tobacman, P. Topalova, Field, E. 2007. “Entitled to Work: Urban Property R. Townsend, and J. Vickery. 2013. “Barriers Rights and Labor Supply in Peru.” Q. J. Econ. to Household Risk Management: Evidence 122: 1,561–1,602. from India.” Am. Econ. J. Appl. Econ. 5: F i s c h e r, E . M . , a n d R . K n u t t i . 2 0 1 5 . 104–35. “Anthropogenic Contribution to Global CRU (Climatic Research Unit), University of Occurrence of Heavy Precipitation and High- East Anglia. 2015. High-Resolution Gridded Temperature Extremes.” Nature Clim. Change Datasets CRU TS v 3.22. 5: 560–64. Davis, L. W., and P. J. Gertler. 2015. “Contribution Fuchs, A. 2014. “Shocks and Poverty in Haiti.” of Air Conditioning Adoption to Future Energy Presentation at Conference “Poverty and Use under Global Warming.” Proc. Natl. Acad. Climate Change Flagship—An Overview of the Sci. 112: 5,962–67. LAC Region,” World Bank, Washington, DC, de Janvry, A., F. Finan, E. Sadoulet, and R. Vakis, September 4. 2006. “Can Conditional Cash Transfer Füssel, H. M. 2012. “Vulnerability to Climate Programs Serve as Safety Nets in Keeping Change and Poverty.” Clim. Change Justice Children at School and from Working When Sustain. 9–17. T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   107 Gentle, P., R. Thwaites, D. Race, and K. Alexander. Intergovernmental Panel on Climate Change, 2014. “Differential Impacts of Climate Change edited by T. F. Stocker, D. Qin, G.-K. Plattner, on Communities in the Middle Hills Region of M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Nepal.” Nat. Hazards 74: 815–36. Y. Xia, V. Bex, and P. M. Midgley, 1–30. Glave, M., R. Fort, and C. Rosemberg. 2008. Cambridge, U.K. and New York: Cambridge “Disaster Risk and Poverty in Latin America: University Press. The Peruvian Case Study.” Research Report, Jensen, R. 2000. “Agricultural Volatility and United Nations Development Programme. Investments in Children.” Am. Econ. Rev . Hallegatte, S. 2007. “The Use of Synthetic 90: 399–404. Hurricane Tracks in Risk Analysis and Jongman, B., E. E. Koks, T. G. Husby, and Climate Change Damage Assessment.” J. Appl. P. J. Ward. 2014. “Increasing Flood Exposure Meteorol. Climatol. 46: 1,956–66. in the Netherlands: Implications for Risk Hallegatte, S., 2012a. A framework to Financing.” Nat. Hazards Earth Syst. Sci. investigate the economic growth impact of 14: 1,245–55. sea level rise. Environ. Res. Lett. 7, 015604. Jongman, B., P. J. Ward, and J. C. Aerts. 2012. doi:10.1088/1748-9326/7/1/015604. “Global Exposure to River and Coastal Hallegatte, S. 2012b. “A Cost Effective Solution Flooding: Long Term Trends and Changes.” to Reduce Disaster Losses in Developing Glob. Environ. Change 22: 823–35. Countries: Hydro-Meteorological Services, Jongman, B., H. C. Winsemius, J. C. Aerts, E. C. Early Warning, and Evacuation.” Policy de Perez, M. K. van Aalst, W. Kron, and P. J. Res. Work. Pap. No 6058, World Bank, Ward. 2015. “Declining Vulnerability to River Washington, DC. Floods and the Global Benefits of Adaptation.” Hallegatte, S., C. Green, R. J. Nicholls, and Proc. Natl. Acad. Sci. 112: E2,271–80. J. Corfee-Morlot. 2013. “Future Flood Losses Kalra, N., S. Hallegatte, R. Lempert, C. Brown, in Major Coastal Cities.” Nat. Clim. Change A. Fozzard, S. Gill, and A. Shah. 2014. 3: 802–06. doi:10.1038/nclimate1979. “Agreeing on Robust Decisions: New Processes Hallegatte, S., F. Henriet, A. Patwardhan, for Decision Making under Deep Uncertainty.” K. Narayanan, S. Ghosh, S. Karmakar, Policy Res. Work. Pap. 6906, World Bank, U. Patnaik, A. Abhayankar, S. Pohit, and Washington, DC. J. Corfee-Morlot. 2010. Flood Risks, Climate Karim, A., and I. Noy. 2014. “Poverty and Natural Change Impacts and Adaptation Benefits Disasters: A Meta-analysis.” SEF Working in Mumbai: An Initial Assessment of Socio- Paper, Victoria University of Wellington. economic Consequences of Present and Climate Knutson, T. R., J. L. McBride, J. Chan, K. Emanuel, Change Induced Flood Risks and of Possible G. Holland, C. Landsea, I. Held, J. P. Kossin, Adaptation Options. Paris: Organisation for A. K. Srivastava, and M. Sugi. 2010. “Tropical Economic Co-operation and Development. Cyclones and Climate Change.” Nat. Geosci. 3: Hulme, M. 2014. “Attributing Weather Extremes 157–63. to ‘Climate Change’: A Review.” Prog. Phys. Kossin, J. P., K. A. Emanuel, and G. A. Vecchi. Geogr. 38: 499–511. 2014. “The Poleward Migration of the Husby, T., and M. Hofkes. 2015. Loss Aversion Location of Tropical Cyclone Maximum on the Housing Market and Capitalisation Intensity.” Nature 509: 349–52. of Flood Risk. Paper presented at EAERE Kreibich, H., A. H. Thieken, T. Petrow, (European Association of Environmental and M. Müller, and B. Merz. 2005. “Flood Loss Resource Economists), Helsinki, Finland, June Reduction of Private Households Due to 24–27. Building Precautionary Measures—Lessons IPCC (Intergovernmental Panel on Climate Learned from the Elbe Flood in August Change). 2012. Special Report on Managing 2002.” Nat. Hazards Earth Syst. Sci. 5: the Risks of Extreme Events and Disasters 117–26. to Advance Climate Change Adaptation: Lall, S. V., and U. Deichmann. 2012. “Density and Summary for Policymakers: A Report of Disasters: Economics of Urban Hazard Risk.” Working Groups I and II of the IPCC. IPCC. World Bank Res. Obs. 27: 74–105. ———. 2013. “Summary for Policymakers.” In Lall, S. V., M. K. A. Lundberg, and Z. Shalizi. Climate Change 2013: The Physical Science 2008. “Implications of Alternate Policies on Basis . Contribution of Working Group Welfare of Slum Dwellers: Evidence from Pune, I to the Fifth Assessment Report of the India.” J. Urban Econ. 63: 56–73. 1 0 8    SHOCK WAVES Lempert, R. J., S. W. Popper, D. G. Groves, Noack, F., S. Wunder, A. Angelsen, and J. Börner. N. Kalra, J. R. Fischbach, S. C. Bankes, B. P. Forthcoming. “Responses to Weather and Bryant, M. T. Collins, K. Keller, A. Hackbarth, Climate: A Cross-Section Analysis of Rural L. Dixon, T. LaTourrette, R. T. Reville, Incomes.” Background paper prepared for this J. W. Hall, C. Mijere, and D. J. McInerney. report. 2013. “Making Good Decisions without Olsson, L., M. Opondo, P. Tschakert, A. Agrawal, Predictions.” Rand Corp. Res. Brief 9701. S. H. Eriksen, S. Ma, L. N. Perch, and S. A. Loayza, N. V., E. Olaberria, J. Rigolini, and Zakieldeen. 2014. “Livelihoods and Poverty.” L. Christiaensen. 2012. “Natural Disasters and In Climate Change 2014: Impacts, Adaptation, Growth: Going beyond the Averages.” World and Vulnerability. Part A: Global and Sectoral Dev. 40 (7): 1,317–36. Aspects, edited by C. B. Field, V. R. Barros, Maccini, S., and D. Yang. 2009. “Under the D. J. Dokken, K. J. Mach, M. D. Mastrandrea, Weather: Health, Schooling, and Economic T. E. Bilir, M. Chatterjee, K. L. Ebi, Y. O. Consequences of Early-Life Rainfall.” Am. Estrada, R. C. Genova, B. Girma, E. S. Kissel, Econ. Rev. 99: 1,006–26. A. N. Levy, S. MacCracken, P. R. Mastrandrea, Masson, V., Y. Lion, A. Peter, G. Pigeon, J. and L. L. White, 793–832. Contribution of Buyck, and E. Brun. 2013. “Grand Paris: Working Group II to the Fifth Assessment Regional Landscape Change to Adapt City to Report of the Intergovernmental Panel on Climate Warming.” Climatic Change 117 (4): Climate Change. Cambridge, U.K. and New 769–82. York: Cambridge University Press. Masson, V., C. Marchadier, L. Adolphe, Opondo, D.O. 2013. “Erosive Coping after the R. Aguejdad, P. Avner, M. Bonhomme, 2011 Floods in Kenya.” Int. J. Glob. Warm. 5: G. Bretagne, X. Briottet, B. Bueno, and C. de 452–66. Munck. 2014. “Adapting Cities to Climate Park, J., S. Hallegatte, M. Bangalore, and Change: A Systemic Modeling Approach.” E. Sandhoefner. Forthcoming. “Households Urban Climate 10: 407–29. and Heat Stress: Estimating the Distributional Mendelsohn, R., K. Emanuel, S. Chonabayashi, Consequences of Climate Change.” Background and Sand L. Bakkensen. 2012. “The Impact of paper prepared for this report. Climate Change on Global Tropical Cyclone Patankar, A., and A. Patwardhan. Forthcoming. Damage.” Nat. Clim. Change 2: 205–09. “Estimating the Uninsured Losses due to Min, S.-K., X. Zhang, F. W. Zwiers, and Extreme Weather Events and Implications G. C. Hegerl. 2011. “Human Contribution to for Informal Sector Vulnerability: A Case More-Intense Precipitation Extremes.” Nature Study of Mumbai, India.” Forthcoming in 470: 378–81. Nat Hazards. Moser, C. 2007. “Asset Accumulation Policy Patankar, A. Forthcoming. “The Exposure, and Poverty Reduction.” In Reducing Global Vulnerability, and Ability to Respond of Poor Poverty: The Case for Asset Accumulation, Households to Recurrent Floods in Mumbai.” edited by Caroline O.N. Moser, 83–103. Background paper prepared for this report. Washington, DC: Brookings Institution Press. Pelling, M. 1997. “What Determines Vulnerability Mosqueira, E. 2003. “Land Titling: The Case of to Floods: A Case Study in Georgetown, Peru.” Paper presented at Mejores Practicas Guyana.” Environ. Probl. 9: 203–26. de Politica Social Conference, Mexico City, Perez-De-Rada, E., and D. Paz. 2008. “Análisis Mexico, May 7–9. de la Relación entre Amenazas Naturales y Nguyen, Van K. 2011. “Building Livelihood Condiciones de Vida: El Caso de Bolivia.” Resilience in Changing Climate.” Presented at Research Report, United Nations Development the Asia Regional Conference, Kuala Lumpur, Programme. Malaysia. Peterson, T. C., M. P. Hoerling, P. A. Stott, and Nkedianye, D., J. de Leeuw, J. O. Ogutu, S. C. Herring. 2013. “Explaining Extreme M. Y. Said, T. L. Saidimu, S. C. Kifugo, D. S. Events of 2012 from a Climate Perspective.” Kaelo, and R. S. Reid. 2011. “Mobility and Bull. Am. Meteorol. Soc. 94: S1–S74. Livestock Mortality in Communally Used Pielke, R., J. Gratz, C. Landsea, D. Collins, Pastoral Areas: The Impact of the 2005– M. Saunders, and R. Musulin. 2008. 2006 Drought on Livestock Mortality in “Normalized Hurricane Damage in the United Maasailand.” Pastoralism 1: 1–17. States: 1900–2005.” Nat. Hazards Rev. 9: 29–42. T hreat M u ltiplier : C limate C hange , D isasters , and P oor P eople   109 Planning Department. 2015. Existing Land Use J. L. Wilhelm. 1996. “Heat-Related Deaths Maps in Mumbai. Municipal Corporation of during the July 1995 Heat Wave in Chicago.” Greater Mumbai. N. Engl. J. Med. 335: 84–90. Prudhomme, C., I. Giuntoli, E. L. Robinson, Shenoy, A. 2013. “Risk and Economic Under- D. B. Clark, N. W. Arnell, R. Dankers, Specialization: Why the Pin-Maker Grows B. M. Fekete, W. Franssen, D. Gerten, and Cassava on the Side.” Working paper S. N. Gosling. 2014. “Hydrological Droughts in available at SSRN: http://dx.doi.org/10.2139​ the 21st Century, Hotspots and Uncertainties from /­ssrn.2211799. a Global Multimodel Ensemble Experiment.” Stott, P. A., D. A. Stone, and M. R. Allen.2004. Proc. Natl. Acad. Sci. 111: 3,262–67. “Human Contribution to the European Public Health Department. 2015. Data on Malaria Heatwave of 2003.” Nature 432: 610–14. Cases in Mumbai, 2001 to 2011. Municipal doi:10.1038/nature03089. Corporation of Greater Mumbai. Tesliuc, E., and K. Lindert. 2003. “Vulnerability: Rabbani, G., A. Rahman, and K. Mainuddin. A Quantitative and Qualitative Assessment.” 2013. “Salinity-Induced Loss and Damage to Guatemala Poverty Assessment Program, Farming Households in Coastal Bangladesh.” World Bank, Washington, DC. Int. J. Glob. Warm. 5: 400–15. The Economist. 2015. “Why India’s Heatwaves Ranger, N., S. Hallegatte, S. Bhattacharya, Are So Deadly.” May 27. M. Bachu, S. Priya, K. Dhore, F. Rafique, Thieken, A. H., H. Kreibich, M. Müller, and P. Mathur, N. Naville, F. Henriet, C. Herweijer, B. Merz. 2007. “Coping with Floods: S. Pohit, and J. Corfee-Morlot. 2011. “An Preparedness, Response and Recovery of Assessment of the Potential Impact of Climate Flood-Affected Residents in Germany in 2002.” Change on Flood Risk in Mumbai.” Clim. Hydrol. Sci. J. 52: 1,016–37. Change 104: 139–67. Tschakert, P. Forthcoming. “The Ability of the Ranson, M., C. Kousky, M. Ruth, L. Jantarasami, Poor to Cope.” Background paper prepared for A. Crimmins, and L. Tarquinio. 2014. this report. “Tropical and Extratropical Cyclone Damages Tschakert, P. 2007. “Views from the Vulnerable: under Climate Change.” Clim. Change 127: Understanding Climatic and Other Stressors in 227–41. the Sahel.” Glob. Environ. Change 17: 381–96. Rentschler, J. E. 2013. “Why Resilience U NISDR (U nite d Nations Inte r national Matters—The Poverty Impacts of Disasters.” Strategy for Disaster Risk Reduction). 2011. Policy Res. Work. Pap. 6699, World Bank, 2011 Global Assessment Report . Geneva, Washington, DC. Switzerland: United Nations International Rodriguez-Oreggia, E., A. De La Fuente, R. De Strategy for Disaster Risk Reduction. La Torre, and H. A. Moreno. 2013. “Natural van Gelder, J.-L. 2010. “Tenure Security and Disasters, Human Development and Poverty at Housing Improvement in Buenos Aires.” Land the Municipal Level in Mexico.” J. Dev. Stud. Lines 19:4. 49: 442–55. Vice News. 2015. “Poor People Are Most Rogers, D. P., and V. V. Tsirkunov. 2013. Affected as Hundreds Die in Blistering Indian Weather and Climate Resilience: Effective Heatwave.” May 25. Preparedness through National Meteorological Viguie, V., and S. Hallegatte. 2012. “Trade-Offs and Hydrological Services. Washington, DC: and Synergies in Urban Climate Policies.” Nat. World Bank. Clim. Change advance online publication. Schewe, J., J. Heinke, D. Gerten, I. Haddeland, Ward, P. J., B. Jongman, F. S. Weiland, N. W. Arnell, D. B. Clark, R. Dankers, S. Eisner, A. Bouwman, R. van Beek, M. F. P. Bierkens, B. M. Fekete, F. J. Colón-González, S. N. Gosling, W. Ligtvoet, and H. C. Winsemius. 2013. H. Kim, X. Liu, Y. Masaki, F. T. Portmann, “Assessing Flood Risk at the Global Scale: Y. Satoh, T. Stacke, Q. Tang, Y. Wada, D. Wisser, Model Setup, Results, and Sensitivity.” T. Albrecht, K. Frieler, F. Piontek, L. Warszawski, Environ. Res. Lett. 8: 044019. and P. Kabat. 2014. “Multimodel Assessment of Whitman, S., G. Good, E. R. Donoghue, Water Scarcity under Climate Change.” Proc. N. Benbow, W. Shou, and S. Mou. 1997. Natl. Acad. Sci. 111: 3,245–50. “Mortality in Chicago Attributed to the July Semenza, J. C., C. H. Rubin, K. H. Falter, J. D. 1995 Heat Wave.” Am. J. Public Health 87: Selanikio, W. D. Flanders, H. L. Howe, and 1,515–18. 1 1 0    SHOCK WAVES Williams, B. A. 2002. Fran, Floyd and Mitigation Assessment Report of the Intergovernmental Policy. Collinsville, IL: Berry A. Williams and Panel on Climate Change. Cambridge, U.K. and Associated, Inc. New York: Cambridge University Press. Winsemius, H., L. P. H. Van Beek, B. Jongman, P. J. World Bank. 2007. Bangladesh–Dhaka: Ward, and A. Bouwman. 2013. “A Framework Improving Living Conditions for the Urban for Global River Flood Risk Assessments.” Poor. Washington, DC: World Bank. Hydrol Earth Syst Sci 17: 1,871–92. ———. 2013. Risk and Opportunity: Managing Winsemius, H., B. Jongman, T. Veldkamp, Risk for Development, World Development S. Hallegatte, M. Bangalore, and P. J. Ward. Report 2014. Washington, DC: World Bank. Forthcoming. “Disaster Risk, Climate Change, ———. 2014. Turn Down the Heat: Climate and Poverty: Assessing the Global Exposure Extremes, Regional Impacts, and the Case for of Poor People to Floods and Droughts.” Resilience. Washington, DC: World Bank. Background paper prepared for this report. ———. 2015a. Turn Down the Heat. Washington, Wodon, Q., A. Liverani, G. Joseph, and DC: World Bank. N. Bougnoux. 2014. Climate Change and ———. 2015b. Climate Data (No. Climate 4 Migration: Evidence from the Middle East and Development). Washington, DC: World Bank. North Africa. Washington, DC: World Bank. ———. 2015c. Another Nargis Strikes Everyday: Wong, P. P., I. J. Losada, J.-P. Gattuso, J. Hinkel, Post-Nargis Social Impacts Monitoring Five A. Khattabi, K. L. McInnes, Y. Saito, and Years On. Washington, DC: World Bank. A. Sallenger. 2014. “Coastal Systems and Low- World Bank and Australian Agency for Lying Areas.” In Climate Change 2014: Impacts, International Development (AUS AID) Adaptation, and Vulnerability. Part A: Global 2014. “Where Are We During Flooding?” and Sectoral Aspects, edited by C. B. Field, A Qualitative Assessment of Poverty and Social V. R. Barros, D. J. Dokken, K. J. Mach, Impacts of Flooding in Selected Neighborhoods M. D. Mastrandrea, T. E. Bilir, M. Chatterjee, of HCMC. K. L. Ebi, Y. O. Estrada, R. C. Genova, B. Girma, Yamano, T., H. Alderman, and L. Christiaensen. E. S. Kissel, A. N. Levy, S. MacCracken, 2005. “Child Growth, Shocks, and Food Aid P. R. Mastrandrea, and L. L. White, 361–409. in Rural Ethiopia.” Am. J. Agric. Econ. 87: Contribution of Working Group II to the Fifth 273–88. Under the Weather: Climate Change, Health, and the 4 Intergenerational Transmission of Poverty Main Messages • Climate change will magnify some threats to • Development—notably better access to health, especially for poor and vulnerable health care and to services such as water and people such as children and the elderly, but sanitation infrastructure—has the potential large uncertainties remain in what is still an to reduce, but not eliminate, the risks cli- emerging research field. mate change poses for health. • Health shocks and poor health contribute • Universal health coverage would contrib- to poverty through loss of income, health ute greatly to climate change adaptation— expenses, and caring responsibilities, so that and monitoring and surveillance systems climate change impacts on health will repre- (both in the health and environmental sec- sent an additional obstacle to poverty reduc- tors) will be critical to deal with emerging tion and will increase inequality. health issues. Introduction work and reduces productivity. It can also One reason why people fall into poverty, or diminish financial assets through medical cannot leave it permanently, is that they are expenditures, especially in the presence of sometimes affected by health shocks and dis- high borrowing costs (Krishna 2006). eases, or by a death in their household. Poor people are more vulnerable to health Illness can reduce human capital through shocks because they have fewer resources permanent health consequences and disability, with which to maintain good health, have less which makes it difficult or impossible to access to improved water and health care, 111 1 1 2    SHOCK WAVES and are more likely to depend on labor-­ Our main message is that economic devel- intensive livelihoods that require good health opment—notably better access to health care such as agriculture or construction. As a and to services such as water and sanitation result, they incur a more severe burden of dis- infrastructure—has the potential to reduce, ease than nonpoor people, and this burden but not eliminate, the enhanced risks that hampers their ability to accumulate and retain ­ climate change poses for health. assets and improve living conditions. Chapters 2 and 3 explored two key chan- nels of poverty—agriculture and ecosystems, Disease and poor health and natural disasters—noting that climate change impacts on agriculture and ecosys- contribute to poverty tems and that natural disasters affect health The impact of poor health on low-income in many different direct and indirect ways, households is already large. In Sub-Saharan such as through undernutrition due to lower Africa and South Asia, the number of crop productivity, or via the spread of dis- deaths in children under five is much higher eases after a disaster. than in other regions, as shown in figure This chapter explores how health, the third 4.1. Diarrheal and respiratory diseases, and key channel of poverty in this report, will be malaria in Sub-Saharan Africa, contribute affected by climate change. A changing significantly to this gap. For adults, there is climate can reduce the quantity and quality of ­ a striking difference in the diseases that water resources, along with altering the sus- affect rich and poor countries. In particu- ceptibility to, and spatial distribution of, lar, communicable diseases—including climate-sensitive diseases. And climate change ­ HIV—­ represent almost half of the cause of is likely to amplify many of the diseases that mortality for adults in Sub-Saharan Africa, already threaten poor households—for while in richer regions non-communicable instance, by allowing malaria-bearing mos- diseases, including cancer and cardiovascu- quitoes to spread in new places, or accelerat- lar diseases, dominate (WHO 2015). ing the replication of pathogens in water. Over 40 percent of the global burden of Even climate change adaptation measures can disease attributed to environmental factors contribute to worsening health conditions. falls on children under five years of age, most Irrigation dams, water storage receptacles, of them living in developing countries. An and other land use and water management estimated 800,000 out of the annual 2 million practices can create suitable conditions for deaths among children under the age of five vectors and pathogens to reproduce, thereby caused by respiratory infections is due to worsening the incidence of disease (Asenso- indoor air pollution; another 760,000 Okyere et al. 2011; Keiser et al. 2005; children die as a result of diarrhea (WHO ­ Medlock and Vaux 2011). 2013a). These deaths could be prevented with These potential impacts of climate change minimal health care and better hygiene, shift- on health are important in and of themselves, ing toward cleaner fuels, and by better access as they could directly cause a massive reduc- to safe water and improved sanitation. tion in well-being. But they could also affect Diseases—and more generally poor households in economic terms, magnifying health—increase poverty for several reasons. the initial impact on welfare. This chapter Health expenditures can absorb a large share begins by reviewing the evidence on the of a household’s income. Diseases reduce pro- impact of health shocks and poor health on ductivity because of missed work and school poverty. It then summarizes what we know days, caregiving responsibilities, or reduced about the health impacts of climate change productivity. And in the long term, they can and concludes with options to minimize these impair children’s development and reduce impacts, including better health infrastructure their ability to learn, which can affect and universal health coverage. earnings. Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    113 FIGURE 4.1  Diarrheal diseases, respiratory diseases, and malaria contribute to record child mortality rates in Sub-Saharan Africa and South Asia (Mortality rates by cause and region for children under 5) Sub-Saharan Africa South Asia Middle East and North Africa Europe and Central Asia East Asia and Pacific Latin America and Caribbean High income 0 10 20 30 40 50 60 70 80 90 100 Deaths per 1,000 Perinatal conditions Respiratory diseases Malaria Diarrheal diseases Other Source: Based on WHO 2015 (data from 2013). Health care costs are regressive and (Krishna et al. 2004). Similar results are found have large impacts on poor households in several countries (Krishna 2007) (figure Health expenditures absorb a large share of 4.3). And even when households have the poor households’ budget, especially in devel- means to cover health care costs, illness has a oping countries. Whereas higher-income regressive cost burden on poorer patients and countries tend to have more sophisticated households (Asenso-Okyere et al. 2011). social insurance systems to support access to Looking at malaria, a review of multiple health care, financial risk protection is largely studies in Burkina Faso, Cameroon, Ghana, absent in lower-income countries, where Malawi, Nigeria, and Sri Lanka finds the about half of health costs borne by house- cost of treatment ranging from $0.41 to holds are out of pocket (figure 4.2). These $5.98 per month, per person,1 which can be poorer households typically rely on their own a significant burden for poor people relative funds, remittances, private health insurance, to their monthly income or expenditure. In or external resources (such as development Malawi, malaria treatment represents 2 per- assistance or support from nongovernmental cent of monthly income for the average organizations [NGOs]). Public funds for household, but 28 percent for the poor health care are rarely available to those who (Ettling et al. 1994). In Kenya, malaria need them the most, and in most low-income accounts for 7.2 percent of household expen- countries the bottom quintile receives less diture on average in wet seasons and 5.9 per- than its share of public outlays for health. cent in dry seasons—but for the bottom Catastrophic health expenditures often quintile the ratios increase to 11 percent in drag people into poverty. In western Kenya, wet seasons and 16.1 percent in dry seasons nearly 73 percent of households mention (Chuma, Thiede, and Molyneux 2006). For health expenses as a principal reason for their diarrhea, the cost of treatment can also be decline into poverty, and 32 percent mention significant for poor households, especially if the death of a major earner as a result of ill- the cost of transport to health care facilities ness as a contributing factor to their poverty is included (Hutton and Haller 2004). 1 1 4    SHOCK WAVES FIGURE 4.2  In poorer countries, half of all health expenditures are paid out of pocket, unlike in richer ones 100 Share of health care expenditure (% of total in 2011) 90 80 70 60 50 40 30 20 10 0 Low income Lower-middle Upper-middle High income Global income income Out-of-pocket expenses External resources Other private expenditure Private health insurance Other government expenditure Social security Source: Watts et al. 2015. Note: Data show the global health care expenditure profile from 2011. PPP = purchasing power parity. FIGURE 4.3  Health and funeral expenses are a major reason why In the extreme case where one household households fall into poverty member dies, the economic impact on the (Percentage of households citing health and funeral expenditures as a principal reason other members can be large, not only through for their descent into poverty) the loss of income but also through funeral expenses. Household surveys in India, Kenya, 100 Peru, and Uganda find that in some places funeral expenses represent a significant cause of poverty, sometimes comparable to health Percentage of households 75 expenditures (figure 4.3) (Krishna 2007). 50 Forgone wages, reduced productivity, and caregiving responsibilities increase 25 poverty Missed days of work as a result of illness— 0 and resulting wage forgone—can also have a a , ny rn di t, In des ra di an an ern d In jara Pe ar nd di h, Ke este Pr ndh Ug est l an In sth significant impact on income. Table 4.1 , a ru ca a jam o a Gu ja W ntra W A a Ca un a Ra da reviews studies on the number of days lost P Ce because of malaria episodes in various coun- Health-related expenses Funeral-related expenses tries, for the sick and for the caregiver (which is an important component because it is often Source: Based on Krishna 2007. children who are affected). Because people can go through many episodes per year, the total number of days lost to malaria can be large. In Oyo State in Nigeria, adults lost on Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    115 TABLE 4.1  Many days of work are lost because of malaria episodes Country Number of missed days of work per episode Burkina Faso 4 days for sick adult and 1.2 days for caregiver for child Ethiopia 18 days for sick adult and 2 days for caregiver for child Ghana 5 days Kenya 2–4 days for sick adult, 2 days of lower productivity for sick adult, and 2–4 days for caregiver Malawi 2.7 days for sick adult and 1.2 days for caregiver for child Nigeria 1–3 days for sick adult, 3 days of lower productivity for sick adult, and 1–3 days for caregiver Sri Lanka 4 complete days and 5.3 days partially lost Source: Based on Guiguemde et al. 1994 for Burkina Faso; Cropper et al. 2000 for Ethiopia; Asenso-Okyere and Dzator 1997 for Ghana; Leighton and Foster 1993 for Kenya and Nigeria; Ettling et al. 1994 for Malawi; and Attanayake, Fox-Rushby, and Mills 2000 for Sri Lanka. average 22 working days per year to the dis- This lack of access to health care for chil- ease (Ajani and Ashagidigbi 2008). For diar- dren matters greatly because illness has par- rhea, which can reach 3 to 7 days per episode, ticularly severe impacts on children. it is usually the caregiver who is missing Children—particularly those under the age of work, since it mostly affects children (Hutton five—are in critical periods of their develop- and Haller 2004). And missed days of work ment so that illness and malnutrition can can be even more detrimental if people are affect lifelong health, educational attainment, fined or even fired when they miss work and labor market outcomes. And when poor (World Bank and Australian AID 2014). families cannot protect their children from Disease can also lower productivity or result these effects, poverty can be transmitted from in the complete inability to work. In agrarian one generation to the next, depriving children households in Africa, repeated malaria illness of a fair chance to escape poverty. has led to a decline in farm output and income Illness can lead to irreversible effects on and contributed to greater incidence of poverty cognitive function, either because it affects the (ESPD 2005). In Ho Chi Minh City, ­ frequent supply of nutrients to the brain or because of floods were found to be a cause of chronic responses in the immune system that damage respiratory disease, rheumatism, and skin and the structure of the brain (Jukes 2005). This is intestinal diseases, especially for children under particularly true of illnesses that affect the five, rendering affected people and caregivers central nervous system (such as severe cere- unable to work (World Bank and Australian bral malaria), which lead to lifelong cognitive AID 2014; this report, chapter 3). impairment in survivors. In Kenya, children aged six to seven who had suffered cerebral malaria were found to be 4.5 times more Impacts on child development result in likely than their peers to suffer cognitive the intergenerational transmission of impairments (ranging from mild challenges to poverty severe learning difficulties) (Holding et al. Children, with less mature immune systems, 1999). Similar results were found in Senegal are more susceptible to illness—and tend to among children between the ages of 5 and 12 have less access to health care when they most who had suffered cerebral malaria with coma need it after a shock. In Côte d’Ivoire, the before the age of 5 (Boivin 2002). share of children taken to a health practitio- Chronic undernutrition is associated with ner fell from 50 percent to around 33 ­percent impairment in the development of cognitive in areas of extreme rainfall (Jensen 2000). In functions in young children, with subsequent Nicaragua, after Hurricane Mitch, children in effects on sociability and educational attain- affected communities were 30 ­ percent less ment (Grantham-McGregor 1995; Whaley likely to be taken to health care facilities et al. 1998). In Jamaica, children who ­suffered when ill (Baez and Santos 2007). from severe undernutrition between 6 and 24 1 1 6    SHOCK WAVES months of age lagged behind their adequately estimate that by the year 2030, climate nourished peers in overall IQ, vocabulary, change could be responsible for an addi- and education tests, even when accounting tional 38,000 annual deaths due to heat for differences in backgrounds (Grantham- exposure among elderly people, 48,000 due McGregor et al. 1994). to diarrhea, 60,000 due to malaria, and More generally, reduced access to educa- about 95,000 due to childhood undernutri- tion associated with health shocks and disas- tion. Morbidity (incidence or prevalence of a ters impact lifelong prospects. Children who disease) would also increase, with the conse- are withdrawn from school to earn income quences for poverty we described above and support their households are particularly (Hales et al. 2014). These estimates assume at risk for long-term effects on their earning that socioeconomic development will reduce potential (Asenso-Okyere et al. 2011). mortality rates, so numbers may be higher if Children exposed to extreme natural disasters development is slower or adaptation is less tend to spend fewer years in school and have efficient than expected. lower educational achievement, delayed This section explores five health issues— development, behavioral issues, and lower IQ (i) vectorborne diseases (malaria); (ii) water- (Caruso 2015; Currie 2009; del Ninno and borne diseases (diarrhea); (iii) stunting; Lundberg 2005; Victora et al. 2008). (iv) mental disorders; and (v) productivity But these impacts are not unavoidable—as loss due to high temperatures—that are likely evidenced by the fact that long-lasting impacts to be sensitive to climate effects and to lead to on children’s health are only observed in a large impacts on the well-being of the poor. small share of the population and chronic Other issues, like diseases related to air pollu- impacts are not manifest in more than 30 per- tion, are also likely to worsen with climate cent of those affected by a disaster (Bonanno change (box 4.1). et al. 2010). Many of these impacts can be avoided or managed by strategic prioritiza- Climate change threatens to reverse tion, sufficient allocation of resources, and progress made on vectorborne political will to manage transient shocks and diseases such as malaria increase the resilience of individuals and households. Policies that reduce the exposure The first major health issue is vectorborne and vulnerability of children to risks and facil- diseases, accounting for over 17 percent of itate recovery after shocks will be essential to all infectious diseases and causing more than manage the impacts of climate change, par- 1 million deaths annually. Vectorborne dis- ticularly in poor communities (see chapter 5). eases are caused by an infectious microbe transmitted to people mainly by bloodsuck- ing insects. Their spread is determined by a Climate change magnifies combination of environmental and social threats to health, especially for factors. In recent years, globalization of travel and trade, urbanization, and environ- poor people mental challenges have had a significant So health matters for poverty, and the impact on transmission. Climate change also ­ evidence is growing that climate change mat- partly explains changes in the spatial distri- ters for health. We know that higher temper- bution of these diseases (Beugnet and atures, varied rainfall patterns, and more Chalvet-Monfray 2013). frequent droughts and floods will affect Malaria is the most prevalent vectorborne health in many ways—through heat expo- disease in the tropics and subtropics and prob- sure, undernutrition, natural disasters, and ably the most important for developing coun- increased proliferation and transmission of tries where insufficient health infrastructure, illnesses that affect poor households (such as favorable climate, drug resistance, and poverty malaria and diarrhea). Hales et al. (2014) have made it difficult to control. It occurs all Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    117 BOX 4.1  Getting harder to breathe Air pollution is already a challenge for human the contribution of outdoor air ­ p ollution to pre- health—exposure to pollution can lead to heart and mature mortality could double by 2050 (Lelieveld lung disease as well as increased hospital visits and et al. 2015). One study that looks at the impact mortality (Peel et al. 2005; Peel et al. 2007; Hoyt of climate change–induced changes on p ­ remature and Gerhart 2004; Moore et al. 2006). Globally, mortality estimates that there could be an increase about 3.3 million premature deaths occur every year of 100,000 premature deaths associated with small because of outdoor air pollution, predominantly in particulate matter exposure and 6,300 premature Asia (Lelieveld et al. 2015). Those who spend a lot of deaths associated with ozone exposure annually time outdoors and engage in physical activities (like (Fang et al. 2013). The incidence of cardiovascular outdoor workers, children, and athletes) and those and respiratory illnesses is also expected to increase who already suffer from respiratory diseases are the as a result of climate change (Takaro, Knowlton, most vulnerable. and Balmes 2013; D’Amato et al. 2014). Will climate change make the situation worse? Possible solutions lie in air surveillance systems We know that higher temperatures and lower rain- and information campaigns to encourage adaptive fall are likely to lead to a worsening of air quality behavior. In addition, certain technologies can be in some areas. It may increase exposure to ground used to reduce emissions of many pollutants at the level ozone, small particulate matter, and air con- combustion source (like air filters or fuel switching). taminants such as allergens and spores (Hogrefe Some health cobenefits can also be expected from et al. 2004). emissions-reduction policies (see chapter 6). Model projections indicate that, without green- house gas (GHG) emissions-reduction policies, Source: Based on Kinney 2008; Ebi and McGregor 2008; Tibbetts 2015. over the world, with an estimated 3.3 billion malaria could be associated with a 0.3 per- people exposed to the risk of infection. In 2013 cent increase in annual growth (Gallup and alone, around 198 ­ million cases occurred, Sachs 2001). Conversely, a study in Africa leading to 584,000 deaths, mostly among chil- found that a 1 percent increase in malaria dren. The challenge is greatest in Sub-Saharan morbidity leads to a reduction in real GDP Africa, which accounts for over 90 percent of growth of 0.4 percent in Ghana (Okorosobo malaria-related deaths—78 percent of which et al. 2011). In Kenya, 2–6 percent of produc- occur in children under the age of five (WHO tion loss could be attributed to malaria inci- 2014). Four out of every ten people who die of dence (Leighton and Foster 1993). malaria live in the Democratic Republic of Fortunately, since 2000, there has been Congo and Nigeria (WHO 2014). A large strong progress in reducing the incidence of share of the deaths occurs among poor and malaria (map 4.1). Moreover, since the launch vulnerable communities living in rural areas, of the World Health Organization’s (WHO’s) with limited access to health facilities. global malaria eradication program in the At the household level, malaria is a burden 1950s, 79 countries have eliminated malaria, on incomes, especially because of health care although mostly in temperate climates. With costs. These household-level impacts add up the shift in focus to malaria control (as to a significant impact on the growth and opposed to eradication) in the tropics, the development of countries—with particularly incidence has decreased by 33 percent in dire effects on low-income countries. At the Africa and 17 percent globally (Caminade country level, a 10 percent reduction in et al. 2014). 1 1 8    SHOCK WAVES MAP 4.1  Most countries on track for significant declines in the incidence of malaria (Projected change in malaria incidence in 2015 compared to 2000) Source: Based on WHO 2014. Climate change can favor vectorborne dis- Higher temperatures could have a major eases such as malaria. Climate variability and effect on malaria transmission. At the global change influence the epidemiology and geogra- level, increases of 2°C or 3°C could raise the phy of vectorborne diseases for several rea- number of people at risk for malaria by up to sons. Rising temperatures boost the odds that 5 percent—affecting more than 150 million climate-sensitive infectious diseases will emerge people (WHO 2003). In Africa, malaria could in new areas—as is already being observed in increase by 5 to 7 percent among populations the densely populated highlands of Colombia at risk in higher altitudes due to rising tem- and Ethiopia (Siraj et al. 2014). And they perature, possibly increasing the number of increase the likelihood of longer seasonal cases by up to 28 percent (Small, Goetz, and transmission and higher incidence in areas Hay 2003). with high current burdens. Variability in tem- Moreover, if adaptation measures to cope perature and precipitation affects the survival with other consequences of climate change and reproduction of vectors that carry disease are not designed carefully, they can also pathogens, their biting rate, and the incubation increase malaria prevalence. In Kumasi, rate of pathogens within the ­ vectors—either Ghana, a study found that irrigated urban raising or lowering ­ transmission. Changing agriculture led to higher densities of anophe- precipitation patterns also affect the quantity line mosquitoes (those responsible for trans- and quality of breeding sites for vectors like mitting malaria) in peri-urban and urban mosquitoes, and shelter and food availability locations and a subsequent higher reported for disease-spreading rodents, affecting the incidence of malaria than in nonagricultural odds of outbreaks (WHO 2003). parts of the city (Afrane et al. 2004). Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    119 On the positive side, malaria is likely to The highland of east Africa is particularly decrease in areas where warming brings very likely to experience significant impacts with high temperatures and less precipitation—the an increase in population at risk projected by expected case for Central America and the all models and a potential reach of over Amazon. Above 34°C, transmission is 200 million additional people at risk of reduced because it becomes more difficult for malaria by 2080 (Caminade et al. 2014). vectors and parasites to survive (Smith et al. Thus, greater efforts will still be needed if 2014; WHO 2003). Already in Senegal, less current gains in the malaria fight are to be precipitation and drought have led to the vir- maintained in spite of climate change. tual disappearance of some mosquito species Increased transmission is most alarming in and reduced malaria prevalence by over areas of unstable or seasonal transmission, as 60 percent (Githeko et al. 2000). these populations have no immunity and Development and specific policies can health systems are less equipped, leading to eradicate—or at least control—malaria . much higher mortality rates (Chima, Despite the expected change in climate con- Goodman, and Mills 2003). ditions, the future of malaria will be largely determined by socioeconomic conditions, such as access to resources and the efficient MAP 4.2  By 2050, socioeconomic development should reduce use of existing prevention and treatment malaria incidence, even with climate change mechanisms (Gething et al. 2010; Hales et al. 2014). One way to sort out the different influ- ences is to examine the separate impacts of climate change and economic growth on malaria, and their combined impacts by 2050 (Béguin et al. 2011). As shown in map 4.2, panel a, an expansion of malaria is expected globally when only considering climate change impacts. When only considering eco- nomic growth, as map 4.2, panel b, shows, malaria is expected to contract in most places with the notable exception of large areas of Sub-Saharan Africa. And what would hap- pen if we combined the two factors? We still expect to see a contraction of malaria by 2050, as map 4.2, panel c, shows. But cli- mate change—even if it does not reverse the positive impacts of development—could still significantly slow down the contraction of malaria, especially in Sub-Saharan Africa and India. By 2030, the same study finds that if both economic growth and climate change are present, an estimated 3.6 billion people could be at risk of malaria, including 100 ­million because of climate change. This number is about 30 percent lower than the estimate for the population at risk if eco- Source: Based on Béguin et al. 2011. Note: Map of projected areas of malaria presence for 2050. Areas where the malaria status changes nomic growth were absent, which is around between the baseline and the scenario period are shown in color. Absent means the model predicts 5.2 billion. no malaria transmission by 2050 in any scenario. 1 2 0    SHOCK WAVES BOX 4.2  Dengue’s future hinges on whether development or climate change prevails Over the past 50 years, there has been a 30-fold drought conditions when households store water increase in global cases of dengue fever, with an annual in containers suitable for mosquito breeding. This incidence of 390 million infections, mostly manifested means that if these extreme events occur more often, in the Asia-Pacific region (Smith et al. 2014). it may extend the suitability of the areas for dengue. Economic development and better infrastructure are As a result, the dengue risk is likely to become more expected to reduce dengue in the future. But will significant in parts of Europe (Bouzid et al. 2014) climate change undercut some of that progress? And and Africa. will it affect dengue’s spread to other regions? Predominantly an urban disease, dengue will be At this point, there is still a lot of uncertainty a greater risk in urbanized areas with poorly man- about how climate change will affect the distribution aged water and solid waste systems. Though climate of dengue. This uncertainty is related partly to the change may extend the suitability of areas for den- fact that there are other important factors—notably gue, economic development—including better access urbanization and migration—besides climate change to piped water, strong vector control programs, and (Hales et al. 2002; Wilder-Smith and Gubler 2008). air-conditioning—can counter this effect. But in However, several studies show that its incidence Africa, climate change could undermine the prog- is correlated with the weather, even though the ress that has been made thanks to higher economic effects are delayed for weeks or months (Hii et al. growth, making it more challenging to eradicate 2009; Johansson et al. 2009). Dengue fever seems dengue and reduce its impacts on welfare and pov- to proliferate under both heavy precipitation and erty (Åström et al. 2012). Malaria prevalence can be reduced by Valley fever, the plague, and chikungunya measures such as mosquito control, improved fever (Bouley and Planté 2014; Smith et al. access to bed nets and malaria treatments, 2014). Less is known about the relationship and better buildings with air-conditioning. In between climate change and these diseases, Oman, in 2000, malaria was pervasive except even though they could have significant and for high-altitude and desert areas (Gallup and increasing impacts on well-being and Sachs 2001), but in 2014, thanks to strategic poverty. intervention and significant resources, there are no indigenous cases of malaria (WHO Diarrhea and other waterborne diseases 2014). are also expected to increase because of Successful elimination requires organiza- climate change tion, resources, and strategies. Exposure is usually highest in remote and rural areas Poor water and food quality continue to where vector control (removal of larvae pose a major threat to human health, espe- breeding sites and residual indoor spraying) is cially for the poor. Soil-transmitted helmin- difficult or impossible, making developing thes and schistosomes (parasitic worms) are countries—like those in Africa where urban- some of the most prevalent chronic diseases ization and connectivity are quite low—more in poor regions, with serious and insidious vulnerable (WHO 2003). Also, medicine effects on health and nutrition (Stephenson, (such as chloroquine) has become ineffective Latham, and Ottesen 2000). Diarrhea alone in many areas, and new drugs are often unaf- is responsible for 1.5 million deaths every fordable for poor populations. year (WHO 2013a). Climate change is also expected to impact Diarrheal outbreaks can occur after drink- other vectorborne diseases—like dengue ing water becomes contaminated and are (box 4.2), encephalitis, Lyme disease, Rift often reported after flooding and related Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    121 displacement (Watson, Gayer, and Connolly transmission. In 18 Pacific islands, diarrhea 2007). In Bangladesh, after the 2004 floods, cases increased with reduced water availabil- more than 17,000 cases of diarrhea were reg- ity (Singh et al. 2001). Globally, low rainfall istered (Qadri et al. 2005), and the cholera locations are strongly associated with higher epidemic in West Bengal, India in 1998 was diarrhea disease prevalence among children attributed to preceding floods (Sur et al. (Lloyd, Kovats, and Armstrong 2007). 2000). In Pakistan, incidence of infectious dis- Hygiene is the main protection against diar- ease and diarrhea increased as a result of the rhea, but studies suggest that water scarcity impact of the 2010 floods on the quality of makes it difficult for households to pursue it. water. Ongoing efforts to eradicate polio were In Peru, it was a lack of water that prevented also interrupted, further setting back this a high awareness of the benefits from hand- agenda (Warraich, Zaidi, and Patel 2011). washing being translated into changed prac- Diarrhea and other waterborne diseases tices (Gilman et al. 1993). affect households’ well-being and prospects Overall, climate impacts could increase the because they are so widespread. The cost of burden of diarrhea by up to 10 percent by one episode of diarrhea is important for poor 2030 in some regions (Kolstad and Johansson households, with treatment costs of $2–4 and 2010; WHO 2003, 2002). An estimated the loss of a few days of work. But diarrhea 48,000 additional deaths annually among can have even larger impacts, because it can children under the age of 15 resulting from provoke undernutrition by making children diarrhea illness are projected by then (Hales unable to absorb nutrients, even if they con- et al. 2014). sume enough food. The total number of The combined effects of temperature deaths caused directly or indirectly by under- fluctuation, coastal salinity, humidity, nutrition induced by unsafe water, inadequate heavy rainfall, flooding, and drought are sanitation, and insufficient hygiene is esti- likely to contribute to outbreaks of other mated at 860,000 per year among children waterborne diseases such as cholera and under five (Vir 2011; this report, chapter 1). schistosomiasis (Bandyopadhyay, Kanji, Climate change can increase the risk of and Wang 2012; Cann et al. 2013; Delpla diarrhea through its impacts on temperature et al. 2009; Stephenson, Lathan, and and water scarcity. Diarrhea is highly sea- Ottesen 2000). Schistosomiasis is second to sonal and higher rates of diarrhea have been malaria as the most devastating i ­llness in associated with higher temperatures, the tropics, causing a debilitating illness although which specific pathogens are that not only damages the internal organs responsible for the association is unclear of its patients but also has lasting impact on (Kolstad and Johansson 2010; Paz 2009). In the growth and cognitive development of Lima, Peru, a 4 percent increase in hospital children (Asenso-Okyere et al. 2011). admissions for diarrhea occurred for each °C The risk of diarrhea and other waterborne increase in temperature during warmer diseases is reduced by better infrastructure, months and a 12 percent increase for every education, and hygiene. To eradicate diar- degree centigrade increase in cooler months rhea, both infrastructure improvement and over six years of observation (Checkley et al. education are needed. The risk for diarrheal 2000). disease outbreaks is higher in developing Greater water stress will further challenge countries than in industrialized countries countries’ ability to provide access to high- (Ahern et al. 2005; Noji 2000)—and a signifi- quality water and push people to use lower- cant number of these diseases could be pre- quality sources, increasing the risk from vented in developing countries through better contaminated water. Lower water quantity access to safe water supply, adequate sanita- also reduces dilution, degrades water quality, tion facilities, and better hygiene practices and can change how people use water, in (Bartram and Cairncross 2010). Indeed, diar- ways that can increase infectious disease rhea is an important risk for poor households 1 2 2    SHOCK WAVES because of unsatisfactory hygiene conditions episodes among children under five, but only if that are related to a lack of infrastructure combined with other behavioral interventions (WHO 2008). The Global Monitoring Report to promote good hygiene (Jalan and Ravallion 2014 (World Bank 2015a) shows large differ- 2003). Similar results are found elsewhere with ences in access across groups even within low- an emphasis on the importance of systemic income countries, with access to improved effects: for a family, gains from access to sani- sanitation for the poorest 40 percent of the tation are relatively small; most of the benefits population everywhere much worse than for arise when the entire community gains access. the richer 60 percent (figure 4.4). Within poor communities, children living in a The prevalence of diarrhea decreases with household with access to improved sanitation income, across and within countries, but there in a village with complete coverage manifest is a large variance at low-income levels 47 percent fewer cases of diarrhea than chil- figure 4.5), suggesting that even poor countries (­ dren living in a household without access to can do much to reduce the prevalence of the improved sanitation in a village without sani- disease among poor people. In rural India, tation coverage. One-fourth of this benefit can access to piped water can significantly reduce be attributed to household effects and the rest the frequency and duration of diarrhea is due to community gains (Andres et al. 2014). FIGURE 4.4  For poorer countries, access to better sanitation for the bottom 40 percent is much worse than for the top 60 percent 100 90 Percentage of population with access to basic sanitation 80 70 60 50 40 30 20 10 0 M C ep. L a Lib ia -B o Ni o o, pu n ya n an he an lad e Ha h Ug nya M da M Eth nea Ca biq a Sie mb ue nz e in ia Co can Be ria m lic Bu om au Ga Bu ar in T ar r Af bia, di ali Ke iti Ba bab al Gu wi ag ad rra odi ge am pi ea og as ng w Ta eon ng Re ni M ista es an rk al nm c De b m p m un ist an gh T e iss .R M ala oz io aF as ad h i Zi Ne jik r S Ta Gu fri lA ra nt Ce Bottom 40 percent Top 60 percent Source: World Bank 2015a. Note: Most recent data between 2005 and 2012 are used. Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    123 Climate change can exacerbate stunting FIGURE 4.5  As incomes rise, the prevalence of diarrhea for children under five falls Despite international efforts in the last 15 years—and large progress in Latin America, some parts of Asia, and northern Africa— 40 Prevalence of diarrhea, children under 5 severe undernutrition remains a problem in Sub-Saharan Africa and Southern Asia. Chronic undernutrition, or stunting, is 30 defined as a very low weight for height (below −3z scores of the median WHO growth standards), whereby children are 20 smaller and shorter but appear normal. Stunting can start before birth and is caused by poor maternal nutrition, poor feed- 10 ing practices, poor food or water quality, and frequent infections. Around 25 percent of stunting among young children can be linked 0 0 0 0 0 to having had five or more episodes of diar- ,00 ,00 ,00 ,00 10 20 30 40 rhea before the age of two (Checkley et al. Average income earned by quintile (US$, PPP 2011) 2008). The consensus among scientists is that the damage to physical growth, brain devel- Source: Based on data from World Bank 2014, 2015b. opment, and human capital formation that Note: Each dot represents one income quintile in one country; the figure therefore shows ­differences across and within countries. PPP = purchasing power parity. occurs in the period before pregnancy to 24 months of age is largely irreversible (World Bank 2006; Black et al. 2008). undernutrition (Black et al. 2013). And 165 Stunting in childhood has been associated million children under five years of age are with a greater risk of noncommunicable dis- stunted, 85 percent of whom live in 20 coun- eases and lower economic productivity in tries, mainly in South Asia and Sub-Saharan adulthood. The medium- and long-term Africa (UNICEF 2013). effects of an increase in stunting among chil- Climate change will likely be a strong dren could significantly reduce their ability to obstacle to the eradication of stunting. Its cope with shocks. Moderate stunting impacts on food production and ecosystems, increases the risk of death by 1.6 times and and natural disasters (like droughts or floods) severe stunting by a staggering 4.1 times force poor and uninsured households to (Black et al. 2008). Even when not mortal, reduce their food intake and quality—which severe stunting brings a higher risk of morbid- in turn can lead to more frequent manifesta- ity and significantly reduces future education tions of undernutrition and stunting, particu- and earning potential (Victora et al. 2008; larly among children. Modeling studies Maccini and Yang 2009; Maluccio et al. suggest that unabated climate change could 2009). Stunting therefore contributes to pov- significantly challenge the increase in avail- erty and its intergenerational transmission. able calories per capita in Sub-Saharan Africa Over 800 million people are currently and South Asia. Such an impact would undernourished, according to the Food and directly affect food intake and health. Poorly Agriculture Organization (FAO 2015). designed land-based mitigation policies could Children are highly represented in these magnify these issues by increasing competi- numbers, with over a third of the burden of tion for land, and thus reducing food avail- disease in undernutrition attributable to chil- ability and contributing to undernutrition and dren under five (Black et al. 2008). In 2011, stunting (see chapter 2). around 45 percent of deaths among children Regardless of socioeconomic development, (3.1 million deaths) could be attributed to climate change will largely increase severe 1 2 4    SHOCK WAVES stunting among children (figure 4.6). An addi- than 2 billion people are affected.2 Climate tional 7.5 million children are expected to be change could reduce the nutritional quality of stunted by 2030, of whom 3.9 million would food and worsen this issue (Myers et al. be affected by severe stunting (a 4 percent 2014), but no quantified estimate of the increase). A WHO report estimates this num- impact is available. ber will rise in 2050 to about 10 million addi- In Sub-Saharan Africa, there is evidence tional children stunted, with an increase of that households provide lower-quality nutri- moderate stunting to about 6 million children tion to children in response to weather shocks (Hales et al. 2014). These estimates include (Alderman, Hoddinott, and Kinsey 2006; assumptions on improved public health due Dercon and Porter 2014; Hoddinott 2006; to technology and economic development, Yamano, Alderman, and Christiaensen 2005), but they do not include nonagricultural inter- which in turn increases the likelihood that ventions, like water and sanitation provision. they will suffer illness (Dercon and Porter Climate change could lead to an increase in 2014). These household behaviors and trends severe stunting of up to 23 percent in Sub- have significant and long-lived impacts on Saharan Africa and up to 62 percent in South physical health, particularly for younger chil- Asia, compared to scenarios without climate dren and women. In Ethiopia, as early as six change (Lloyd et al. 2011). These increases months following a disaster, households that correspond to an absolute increase in the reduced the nutritional quality of their food number of stunted children in some parts of intake displayed lowered growth among chil- Sub-Saharan Africa, with the negative effect dren under two years by 0.9 cm (Yamano, of climate change outweighing the positive Alderman, and Christiaensen 2005). ­ conomic growth. effect of e The quality of the diet is also essential. At Climate-related shocks and disruptions least half of children worldwide aged six can increase mental disorders and may months to five years suffer from one or more exacerbate the “cognitive tax” micronutrient deficiencies (iron, iodine, vita- min A, folate, and zinc), and globally more The evidence is growing that poverty is asso- ciated with mental disorders, even though the FIGURE 4.6  Stunting projections for 2030 and causality is unclear (Patel and Kleinman 2050 suggest that regardless of the socioeconomic 2003). In developed countries, rates of mental scenario, climate change will increase severe disorders are much higher among low-income stunting among children under 5 and homeless people than the rest of the pop- ulation (Bassuk et al. 1998; Fazel et al. 2008). Stunting in children under 5 attributable 12 In low- and middle-income countries, a 10 review of 115 studies reports a positive asso- to climate change (millions) 8 ciation between a range of poverty indicators 6 4 and mental disorders (Lund et al. 2010). 2 We know that natural disasters can cause 0 high levels of stress and mental disorders. –2 Anxiety, depression, and post-traumatic –4 stress disorder (PTSD) have been reported –6 L B H L B H in populations affected by flooding and dur- 2030 2050 ing slow-onset events such as droughts Severe Moderate (Ahern et al. 2005; Paranjothy et al. 2011). In Nicaragua, a study of adolescents half a year Source: Hales et al. 2014. Note: The bars show the additional number of children under 5 stunted after Hurricane Mitch found instances of because of climate change in 2030 and 2050 under low-growth (L), PTSD, stress, and depression, particularly base-case (B), and high-growth (H) socioeconomic scenarios. Moderate stunting decreases in 2050 in the low-growth scenario mostly because among those in most affected communities stunting becomes more severe, not because fewer children are stunted. who suffered the highest impact and those Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    125 who experienced a death in the household mental health effects after exposure to extreme (Goenjian et al. 2001). In Sri Lanka, children weather events. After Hurricane Katrina in the between the ages of 8 and 14 in areas affected United States, those with anxiety prior to the by the 2004 tsunami had rates of PTSD rang- storm were more likely to experience PTSD ing from 14 to 39 percent within a month of symptoms afterward, and younger children the event (Neuner et al. 2006). These trends had more symptoms (Kronenberg et al. 2010; can lead to chronic distress and increased Weems et al. 2007). In the longer term, flood- incidence of suicide (Berry, Bowen, and ing affects perceptions of security and safety Kjellstrom 2010; Hanigan et al. 2012; and can lead to depression, anxiety, PTSD, Keshavarz et al. 2013). To mitigate impacts, and other chronic and severe mental health psychosocial and psychological interventions disorders (Ahern et al. 2005; Berry, Bowen, must be incorporated into disaster response and Kjellstrom 2010; Fritze et al. 2008; and recovery management interventions. Paranjothy et al. 2011). Higher temperatures The worry is that climate change may exac- and extreme rainfall also raise concerns about erbate mental disorders and stress. It could do more frequent conflicts, which tend to impede so directly through greater exposure to trauma poverty reduction (box 4.3). (from floods and other disasters) but also indi- Risks and stresses also affect cognitive per- rectly through impacts on physical health, formance and decision making. Planning for household dynamics, and community well- contingencies (for example, because of a being. Poor households, already strained by shock), unpredictable income, and constant the pressures of poor living conditions and worry about the financial situation create scarce resources, could be more prone to larger stress and depression for the poor, which BOX 4.3  The uncertain triangle of climate change, conflict, and poverty Conflict, fragility, and lack of security and stability change will increase murders, assaults, rape, and are fundamental barriers to poverty reduction and the other violence (Ranson 2014). In developed coun- well-being of the poor (World Bank 2011). This is tries, studies typically cite psychological factors for illustrated by the increasing share of poor people liv- this correlation (such as people become more aggres- ing in conflict environments. In the 33 countries (rep- sive during heat waves). In less developed countries, resenting half a billion people) classified by the World like India, the trigger may be lower income from Bank as fragile and conflict-affected states, the poverty higher temperatures, which in turn can raise crime headcount is 51 percent. Similarly, within c ­ ountries, rates (Iyer and Topalova 2014). poor people are also more exposed to crime: in Cape Intergroup conflict: A large body of literature sug- Town, South Africa, 44 percent of all homicides occur gests a link between weather or climate and conflict, in three neighborhoods that are among the city’s especially in low-income areas. A meta-analysis sug- poorest (World Bank 2013, chapter 4). gests that intergroup conflicts increase by 11 percent What type of impact does the environment—and when temperatures increase one standard deviation potentially climate change—have on conflicts? This and by 3.5 percent when rainfall deviates one stan- is an area of active research (Burke, Hsiang, and dard deviation (Hsiang, Burke, and Miguel 2013; Miguel 2014), which can be split into studies cover- Burke, Hsiang, and Miguel 2014). However, a vig- ing conflicts from either the interpersonal level (like orous debate has emerged around the robustness of assaults and rape) or the intergroup (national) level these results (Buhaug et al. 2014; Hsiang, Burke, and (like civil conflicts, wars, and riots). Miguel 2014), with the IPCC noting that “­ collectively Interpersonal conflicts: There is a strong correlation the research does not conclude that there is a strong in the United States between crime and violence and positive relationship between warming and armed temperature, raising the question of whether climate conflict” (Adger et al. 2014). 1 2 6    SHOCK WAVES reduces focus, lowers productivity, and inter- benefits. Health effects from extreme heat feres with making long-term decisions exposure are expected to result from both (Banerjee and Duflo 2012). Poor people may higher average seasonal temperature and more have little time or energy to think about the frequent and intense extreme heat wave events future, as their day-to-day economic lives are (Huang et al. 2011; IPCC 2014). Chapter 3 more consuming of cognitive control than for discussed exposure to heat waves, showing that the rich (Banerjee and Duflo 2012). This poor individuals are more likely to be exposed effect has been referred to as the “cognitive to higher temperature, especially in hot coun- tax”: the high level of stress of poor people tries. It also showed that poor people are par- acts like a tax that reduces their productivity ticularly vulnerable to high temperatures, and earnings, contributing to their poverty. because of their living conditions, the poor Already today, natural risks are a major quality of their housing, and a lack of access to source of stress. In flood-prone wards of conditioning. Here, we explore the conse- air-­ Mumbai, 71 percent of the households sur- quences of this exposure to high temperature, veyed cited flooding as a critical stressor, sec- looking at direct health consequences and their ond only to “hectic life,” and more important impacts on performance and productivity. than stress from transportation or congestion Heat-related problems will not be limited (figure 4.7) (Patankar, forthcoming). Against to developing regions. It is projected that, the background of anticipated more frequent globally, by 2030, without accounting for natural hazards due to climate change, this adaptation, there could about 100,000 addi- cognitive tax for poor people may increase. tional deaths annually in 2030—and 250,000 annual deaths in 2050—among those aged High temperatures are a health hazard 65 and over (Hales et al. 2014). Of course, and affect labor productivity humans adapt, so adaptation may lower these estimates. But these estimates do not include ­ Though extreme cold-related deaths will morbidity and mortality in other age groups decrease in temperate regions, the negative and among vulnerable people, nor do they effects of heat waves will likely outweigh these factor in extreme heat wave events. Most heat wave–related deaths occur FIGURE 4.7  Poor households in Mumbai face multiple stresses, with a key one the risk from floods among poor elderly people and people who have existing illnesses (such as cardiovascular 80 or chronic respiratory diseases and mental illness) (WHO 2003). Urban dwellers are ­ 70 % of households reporting stress particularly at risk because of inefficient hous- ­ 60 ing and the heat island effect—where urban 50 environments with high thermal mass and low ventilation retain heat, thereby amplify- 40 ing the rise in temperature, especially at night. 30 Air-conditioning practices in urban areas 20 further amplify this effect as indoor heat is ­ transferred outdoors, which hurts the desti- 10 tute and homeless. In Taiwan, China air-­ 0 conditioning was found to have added 0.7°C n on n e ng n ks s s to the outdoor temperature (Liu, Ma, and ue ti e ti o lif ti o ti o or s ti di ic ss ni rta di llu w oo ge ct yi tu n et Li 2011). And as a greater portion of the pop- po po He co Fl n ar or ln Co s et r pp g an ci a Ai in ulation becomes elderly and urbanization on Tr fo us so M o Ho No increases, a bigger share of the population ck La will be vulnerable to heat stress. Source: Patankar, forthcoming. Note: Y-axis shows the percent of respondents who cited each stress as being important. Multiple One possible solution is climate-smart stresses could be reported. urban design and innovative architecture, Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    127 which can reduce this effect while taking FIGURE 4.8  If it gets too hot, productivity falls advantage of shade provision, solar heat man- significantly agement, and other measures that use thermal insulation to minimize energy consumption 1.00 (Masson et al. 2013; Masson et al. 2014; Relative performance 0.95 Stone, Hess, and Frumkin 2010). Additionally, well-designed early warning and surveillance 0.90 systems can help detect and respond to heat waves. These must be adapted to the levels of 0.85 risk so as to have a significant impact on reducing mortality (Ebi et al. 2004; Schmier 0.80 and Ebi 2009). 15 20 25 30 35 High temperatures reduce labor produc- Temperature (°C) tivity and can thus increase poverty. A well- established medical and task productivity Source: Seppänen, Fisk, and Lei 2006. literature has uncovered a systematic rela- tionship between temperature stress of the incentives faced by workers. And even though human body and reduced performance these studies focus on the developed world, a (Seppänen, Fisk, and Lei 2006). Lab experi- similar impact in developing countries can be ments have quantified this relationship by expected. randomly assigning subjects to rooms of In developing countries, temperatures varying temperatures and asking them to above 24–25°C are associated with poorer perform cognitive and physical tasks. They performance (Federspiel et al. 2004). Indian find that extreme temperature reduces manufacturing worker efficiency at the plant human performance on a wide range of level declines substantially on hotter days, tasks, including time estimation, vigilance, with a magnitude of roughly minus and higher cognitive functions (like mental 2.8 ­percent per °C, an effect that is driven pri- arithmetic and simulated flight) (Grether marily by on-the-job task productivity decline 1973). A review of the experimental litera- as opposed to increased missed days of work ture finds that in laboratory settings task or absenteeism (Adhvaryu, Chari, and productivity improves up to a temperature Sharma 2013; Sudarshan et al. 2015). threshold of around 20°C to 25°C, but after Granted, air-conditioning is typically a scarce that it declines significantly—with the aver- commodity in the developing world, but there age productivity loss on the order of does seem to be evidence that the same phe- 2 ­percent per °C for the various tasks sur- nomena occur even in developed economies veyed (figure 4.8). Similarly, a review of his- such as the United States (Deryugina and torical fluctuations in temperature within Hsiang 2014; Park, forthcoming). This result countries identifies that higher temperatures suggests that air-conditioning, while useful, reduce economic growth in poor countries may not be able to cancel out all of the (Dell, Jones, and Olken 2012). impacts from higher temperature. Responses by workers to temperature There is also evidence of a drop in labor shocks may take many forms (Heal and Park supply in response to heat stress (Zivin and 2013; Zivin and Neidell 2014). There may be Neidell 2014). In U.S. industries with a high declines in task productivity, labor supply exposure to climate, workers report less time (hours worked), labor effort, or all three. The spent at work and less time spent on outdoor emerging microeconometric literature finds leisure activities, on hot and cold days. At tem- evidence for at least the first two in particular, peratures over 38°C, labor supply in outdoor and likely reflects a combination of all three— industries drops by as much as one hour per the specific breakdown of which will depend day compared to those in the 24–27°C range. on labor market institutions and specific Using U.S. plant-level output data from 1 2 8    SHOCK WAVES 1994–2004 for the automobile sector, a study Since then, numerous cross-country analy- found that hot days are associated with lower ses have suggested that hotter countries have output across the board. At the extreme, a tended to grow more slowly on average. Many week with six or more days above 32°C have noted that hotter countries tend to have reduces that week’s production by about 8 per- lower income levels generally—with a gradient cent (Cachon, Gallino, and Olivares 2012). of roughly minus 8.5 percent per capita income Poorer households are more likely to be per °C hotter average temperatures (Dell, affected by the downsides of higher tempera- Jones, and Olken 2009; Horowitz 2009). tures because they are less likely to benefit However, it is likely that unobservable effects from air conditioning and more likely to play an important role (like institutions, levels work in sectors that are more sensitive to of human capital, and agricultural productiv- temperature stress: namely, manual labor– ity) (Acemoglu, Johnson, and Robinson 2000). intensive industries, and outdoor work–­ This negative relationship between temper- intensive sectors (like agriculture and ature and income seems to hold within coun- construction). It is also likely the case that tries, albeit to a milder extent. This suggests manual labor and outdoor work occupations that institutional factors are not wholly respon- pay lower wages on average. The U.S. Bureau sible for the temperature-productivity relation- of Labor Statistics reports that the average ship and there are limits to adaptation through construction laborer earns 25 percent less better buildings and air-conditioning. An than the median U.S. worker, and laborers in assessment of incomes at the municipality level the farming, fishing, and forestry sector earn for 12 U.S. counties finds that a 1°C rise in 48 percent less (BLS 2015). temperature is associated with 1.2–1.9 p ­ ercent It remains unclear how much can be lower per capita income (Acemoglu and Dell expected from adaptation and how large the 2009). Similar results are found analyzing a expected social costs of adaptation will be— larger set of counties, using U.S. income and whether in the form of physical capital invest- payroll data from 1986–2012 (Deryugina and ments, relocation costs, or the nonpecuniary Hsiang 2014; Park, forthcoming). costs of changing habit patterns and social norms. Also, the role of technological change is unclear; the same goes for possible public Health care systems and investment in research and development on development pathways play a these issues. Climate change impacts on individual critical role productivity could have an effect at the mac- The outcome of climate-induced health roeconomic level. As to the toll climate effects will be determined by institutional change might take on economic growth at structures and the combined effects of the global and national levels, researchers ­ parallel global changes—such as urbaniza- have long noted a relationship between tem- tion, population growth, and demographic perature and macroeconomic variables shifts. Also relevant are social norms and (namely income and growth). As far back as behavior, along with differences in the vul- Montesquieu in the 18th century (Huntington nerability of populations due to nonclimatic 1922; Montesquieu 1758), there has been a factors (Ebi and Semenza 2008; Patz et al. suggestion that extreme climate may reduce 2005; Sutherst 2004). For instance, in devel- economic growth. Using data from agricul- oping countries, around 43 percent of the tural and manufacturing occupations in reduction in the number of children under- North Carolina, a study showed that aggre- weight between 1970 and 1995 can be gate productivity was highest in moderate attributed to greater access to education for temperatures (fall and spring), and lower women, 26 percent to greater access to food, in more extreme temperatures (summer, and 19 percent to improved water and sani- winter) (Huntington 1922). tation (Smith and Haddad 2000). Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    129 Health infrastructure, access, and supply in low-income countries to the level quality of care need to be improved of the best middle-income countries today (Jamison et al. 2013). Given that low-income countries will be the Although beyond the scope of this report, most vulnerable because of limited public the health sector can also play a role in emis- health infrastructure, a top priority should sions reductions. In the United Kingdom, this be improving health care. Often, treatable sector is responsible for about 25 percent of illnesses are not addressed because of lack of all public sector emissions and, in the United access to adequate health care services. In States, about 8 percent of total emissions. It is rural areas, transportation may not be avail- thus crucial that investments contribute to able to transfer the ill to clinics. Further, greening the health infrastructure. many of these rural clinics do not have ade- The risk from emerging diseases or unex- quate equipment or trained health personnel, pected crises also increases the urgency to put and require payment up front. in place effective risk monitoring systems Today, the share of births attended by (Wesolowski et al. 2015; Semenza and Menne skilled health staff is close to 100 percent for 2009) and to share experience and informa- countries above GDP per capita of $20,000 tion (Ebi and Burton 2008)—as illustrated by but varies widely below this level, suggesting the emergence of chikungunya in France, progress can be made even at low-income Italy, and the Caribbean. This means an levels (figure 4.9, panel a). If skilled health urgent need for surveillance systems that rely staff are not available—for birth, injuries, or on all participants in the health care system diseases—people are more likely to suffer (especially private practice physicians), effec- from permanent consequences on health, tive communication of good behaviors income, and well-being. Improving health through general media, and international care systems (staff training, vaccination pro- cooperation and exchange of information. grams, information campaigns, and access to Such efforts pay off. Following the 2003 heat rapid diagnostic kits and drugs for treat- wave in France, the government introduced a ment) is therefore essential. With significant heat wave warning system and national investments over the next 20 years, it is pos- action plan. Health worker training and new sible to improve the level of health care infrastructure helped avoid an estimated FIGURE 4.9  A lot of room to improve the quality and cost of health care in poor countries a. Quality: Births attended by skilled health staff b. Cost: Out-of-pocket health expenditure 100 100 Share of births attended by skilled health expenditure (%) Share of out-of-pocket 80 80 health staff (%) 60 60 40 40 20 20 0 0 00 00 00 00 00 00 0 0 0 0 0 0 ,00 ,00 ,00 ,00 ,00 ,00 ,0 ,0 ,0 ,0 ,0 ,0 10 20 30 40 50 60 20 10 30 40 50 60 GDP per capita (US$, PPP 2011) GDP per capita (US$, PPP 2011) Source: Based on data from World Bank 2015b. Note: PPP = purchasing power parity. 1 3 0    SHOCK WAVES BOX 4.4  Universal health coverage: Kenya’s bottom-up strategy Launched in 2010, Kenya’s Health Sector Services better reporting and accounting for user fees Fund (HSSF) aims to expand the supply of health collected and stronger information and governance care and strengthen primary care by providing systems. In some facilities, these funds were used direct cash transfers to public health facilities— effectively to upgrade service delivery (such as particularly in rural areas where health centers providing electricity in maternity wards for 24-hour and dispensaries are a primary source of care and services) and better maintain facilities by hiring over 80 percent of the population resides. Jointly more local staff. A successful pilot exploring the supported by the government of Kenya, the Danish incorporation of performance-based financing in International Development Agency (DANIDA), this program shows further potential for improving and the World Bank, this initiative complements service quality (for example, with better prenatal better access and transparency of resources with care and more child immunization). sector reforms to improve the availability of The direct funding provided by HSSF has human resources and essential medicine for service helped ensure that resources reach the periphery delivery. The HSSF encourages transparency, of the health system at low levels of bureaucratic community participation in decision making, and interference without compromising transparency. accountability for resource allocation and results. A demand-based “pull system” approach that Local communities are tasked with managing requires facilities to order supplies and commodities the funds through representative management based on need, as opposed to centrally determined committees and work with the staff of health allocation, has significantly reduced wastage and facilities to improve delivery of services. expiry of drugs while increasing the reliable avail- So far, this initiative has reached 3,000 public ability of health commodities. primary health facilities and boosted the use of health facilities. Preliminary observations show Source: Ramana, Chepkoech, and Workie 2013. 4,400 deaths in a subsequent 2006 heat wave to $3 billion—out of the nearly $250 billion (Fouillet et al. 2008; Pascal et al. 2006). spent annually on health-related R&D And research and development efforts in (Jamison et al. 2013). the health sector should be intensified to bet- Social protection systems can also play a ter prevent, diagnose, and treat diseases that significant role, especially in helping avoid affect poor people, especially those that are irreversible losses from undernutrition—but expected to increase over time, including only if they can be scaled up quickly after because of climate change. This is especially shocks and targeted to reach the poorest and true for the so-called “neglected tropical dis- most vulnerable (Alderman 2010; Clarke and eases”—those diseases that thrive mainly Hill 2013; this report, chapter 5). among the world’s poorest populations.3 Of these, several, such as dengue, leishmaniasis, Universal health care coverage is an and chikungunya are sensitive to climate adaptation priority and likely to change in spatial distribution with climate change. Private research and Even if skilled health care is available, its development (R&D) alone is unlikely to affordability is not a given. Health shocks develop the needed solutions without public tend to bring households into poverty even intervention (Trouiller et al. 2001). Today, more where people have to borrow, often at annual R&D spending on “infectious dis- high interest rates, creating debts that they eases of particular concern to low-income may never be able to repay (Krishna 2006). and m­ iddle-income countries” amounts only The WHO estimates that about 100 million Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    131 people fall into poverty each year just to pay important to assess the consequences of for health care (WHO 2013b). A big problem reforms on the nonpoor; (iii) solutions are is that financial risk protection varies widely, best designed starting from population with people in low-income countries having needs—including the local epidemiological to bear very high and variable fractions of profile, major barriers to access to care, out-​of-pocket health expenditure (figure 4.9, unsatisfied demand, and major sources of panel b). financial hardship; and (iv) highly focused Thus, better health care coverage and interventions (such as on one barrier to lower out-of-pocket expenses would be effi- access or one disease) can be a useful initial cient ways to reduce the health impacts of step toward universal coverage (Giedion, ­ climate change vulnerability and reduce pov- Andrés Alfonso, and Díaz 2013). erty, especially by helping the poor to man- age catastrophic health expenditures (Jamison et al. 2013). Providing health cov- erage is possible at all income levels, but In conclusion context and implementation challenges will Health shocks and poor health bring and determine the optimal path for countries, as keep people in poverty, and can reduce the case in Kenya illustrates (box 4.4). ­ lifelong earning prospects when children are Rwanda invested in a universal health affected. Climate change is expected to ­ coverage system in 1994, and today over worsen many of these issues—although 80 percent of its population is insured. big uncertainties remain, such as the Employment-based social insurance is extent to which climate change will affect limited to the formal sector. But strategic ­ the nutritional quality of food. Moreover, policies that promote equitable and pro- the combined effects of multiple health poor financing mechanisms can accelerate stressors are largely unknown, in spite of the the p­ rocess toward universal health cover- importance of interactions among diseases. age. In Thailand, the government has For instance, undernourished children are expanded coverage to the informal sector known to be more vulnerable to malaria and with a minimal charge of $0.70 per visit, other vectorborne or waterborne diseases, drawing on general tax revenues. In but these interactions have not been investi- Colombia, through a ­ multilevel government gated yet in the context of climate change. scheme and cross-subsidization from con- The encouraging news is that economic tributory schemes, the poor are covered development, poverty reduction, and better against primary care and catastrophic event infrastructure and access to health care could costs—with coverage among the poorest compensate for many of the negative pro- quintile rising from 3 to 8 percent in 1993 jected climate-related trends. Indeed, if to 47 percent by 1997. In parts of Africa developing countries could achieve the pres- and Asia, an efficient tool is community- ent level of health care access in industrial- financed coverage schemes that pool expen- ized countries by 2030, they could avoid diture risks at lower administrative levels. many of the impacts that would worsen Strong community solidarity and adminis- health conditions. Child mortality could fall trative capacity is important for these by an ­ estimated 63 percent globally if cover- interventions (O’Donnell 2007). ­ age rates of effective prevention and treat- What can we learn from past efforts to ment mechanisms rose to 99 percent (Jones expand health coverage? Four insights stand et al. 2003). out: (i) affordability is important but not suf- A recent Lancet Commission on Investing ficient to achieve universal access, and mea- in Health concludes that by 2035, a “grand sures to ensure affordability should be convergence” in mortality and morbidity included within a broader strategy; (ii) tar- rates across the world is achievable, as is the geting the poor is necessary, but it is also global provision of universal health care 1 3 2    SHOCK WAVES (Jamison et al. 2013). The report contends References that enhanced investments to scale up health Acemoglu, D., and M. Dell. 2009. Productivity technologies and systems in developing coun- Differences between and within Countries . tries could bring down mortality rates in National Bureau of Economic Research. most low-income and middle-income coun- Acemoglu, D., S. Johnson, and J. A. Robinson. tries to those presently seen in the best-­ 2000. The Colonial Origins of Comparative performing middle-income countries. This Development: An Empirical Investigation . would mean the prevention of about 10 mil- National Bureau of Economic Research. lion deaths in 2035 in these poorer countries Adger, W. N., J. M. Pulhin, J. Barnett, relative to a scenario of stagnant investments G. D. Dabelko, G. K. Hovelsrud, M. Levy, and no improvements in health technology. S. Oswald, and C. H. Vogel. 2014. “Human The benefit-cost ratio of these investments is Security.” In Climate Change 2014: Impacts, estimated between 9 and 20, without Adaptation, and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working accounting for climate change. And, as this Group II to the Fifth Assessment Report of the chapter has shown, climate change only Intergovernmental Panel on Climate Change. amplifies the benefits of acting now to [Field, C.B., V.R. Barros, D.J. Dokken, improve health services. K.J. Mach, M.D. Mastrandrea, T.E. Bilir, Of course, a “grand convergence” would M. Chatterjee, K.L. Ebi, Y.O. Estrada, R.C. significantly reduce the impact of climate Genova, B. Girma, E.S. Kissel, A.N. Levy, change on poor people through the health S. MacCracken, P.R. Mastrandrea, and L.L. channel. But even with prevention and White (eds.)], 755–91. Cambridge, U.K. and hygiene, and even if health care and health New York: Cambridge University Press. coverage are available, diseases and accidents Adhvaryu, A., A. Chari, and S. Sharma. 2013. cannot be fully prevented—and they will con- “Firing Costs and Flexibility: Evidence from Firms’ Employment Responses to Shocks in tinue to lead to reduced income, lost days of India.” Rev. Econ. Stat. 95: 725–40. work, and higher expenses. To help poor Afrane, Y. A., E. Klinkenberg, P. Drechsel, households cope with these shocks and avoid K. Owusu-Daaku, R. Garms, and T. Kruppa. the irreversible impacts on children’s develop- 2004. “Does Irrigated Urban Agriculture ment and education, other tools can be mobi- Influence the Transmission of Malaria in the lized and developed, such as financial City of Kumasi, Ghana?” Acta Tropica 89 (2): products to save and borrow, social safety 125–34. nets (such as cash transfers and social insur- Ahern, M., R. S. Kovats, P. Wilkinson, R. Few, and ance), and remittances and support from fam- F. Matthies. 2005. “Global Health Impacts of ily and friends. The next chapter explores Floods: Epidemiologic Evidence.” Epidemiol. how these instruments can help, and whether Rev. 27: 36–46. Ajani, O. I. Y., and W. M. Ashagidigbi. 2008. they need to be designed differently, in a con- “Effect of Malaria on Rural Households’ Farm text of climate change. Income in Oyo State, Nigeria.” African Journal of Biomedical Research 11 (3). Alderman, H. 2010. “Safety Nets Can Help Notes Address the Risks to Nutrition from Increasing 1. Asenso-Okyere and Dzator 1997; Climate Variability.” J. Nutr. 140: 148S–152S. Attanayake, Fox-Rushby, and Mills 2000; Alderman, H., J. Hoddinott, and B. Kinsey. 2006. Desfontaine et al. 1989; Ettling et al. 1994; “Long-Term Consequences of Early Childhood Guiguemde et al. 1994; Konradsen et al. Malnutrition.” Oxf. Econ. Pap. 58: 450–74. 1997; Louis et al. 1992; Onwujekwe, Chima, Andres, L., B. Briceño, C. Chase, and and Okonkwo 2000 J. A. Echenique. 2014. “Sanitation and 2. http://www.cdc.gov/immpact/micronutrients/ Externalities: Evidence from Early Childhood 3. http://www.who.int/neglected_diseases​ Health in Rural India.” Policy Res. Work. Pap. /­diseases/en/ 6737, World Bank, Washington, DC. Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    133 Asenso-Okyere, K., C. Chiang, P. Thangata, and Black, R. E., L. H. Allen, Z. A. Bhutta, K. Andam. 2011. “Interactions between Health L. E. Caulfield, M. de Onis, M. Ezzati, and Farm-Labor Productivity.” Food Policy C. Mathers, and J. Rivera, Maternal and Child Report, International Food Policy Research Undernutrition Study Group. 2008. “Maternal Institute and Child Undernutrition: Global and Regional Asenso-Okyere, W. K., and J. A. Dzator. 1997. Exposures and Health Consequences.” The “Household Cost of Seeking Malaria Care. Lancet 371: 243–60. A Retrospective Study of Two Districts in Black, R. E., C. G. Victora, S. P. Walker, Ghana.” Soc. Sci. Med. 45: 659–67. Z. A. Bhutta, P. Christian, M. De Onis, Åström, C., J. Rocklöv, S. Hales, A. Béguin, M. Ezzati, S. Grantham-McGregor, J. Katz, V. Louis, and R. Sauerborn. 2012. “Potential and R. Martorell. 2013. “Maternal and Distribution of Dengue Fever under Scenarios of Child Undernutrition and Overweight in Climate Change and Economic Development.” Low-Income and Middle-Income Countries.” ­ EcoHealth 9: 448–54. The Lancet 382: 427–51. Attanayake, N., J. Fox-Rushby, and A. Mills. 2000. BLS (Bureau of Labor Statistics). 2015. Overview “Household Costs of ‘Malaria’ Morbidity: of BLS Wage Data by Area and Occupation. A Study in Matale District, Sri Lanka.” Trop. Bureau of Labor Statistics. Med. Int. Health 5: 595–606. Boivin, M. J. 2002. “Effects of Early Cerebral Baez, J. E., and I. V. Santos. 2007. “Children’s Malaria on Cognitive Ability in Senegalese Vulnerability to Weather Shocks: A Natural Children.” J. Dev. Behav. Pediatr. 23: 353–64. Disaster as a Natural Experiment.” Social Bonanno, G. A., C. R. Brewin, K. Kaniasty, and Science Research Network, New York. A. M. La Greca. 2010. “Weighing the Costs of Bandyopadhyay, S., S. Kanji, and L. Wang. 2012. Disaster Consequences, Risks, and Resilience “The Impact of Rainfall and Temperature in Individuals, Families, and Communities.” Variation on Diarrheal Prevalence in Sub- Psychol. Sci. Public Interest 11: 1–49. Saharan Africa.” Appl. Geogr. 33: 63–72. Bouley, T., and C. Planté. 2014. Reducing Climate- Banerjee, A., and E. Duflo. 2012. Poor Economics: Sensitive Disease Risks (No. 84956-Glb). A Radical Rethinking of the Way to Fight Washington, DC: World Bank. Global Poverty, Reprint edition. New York: Bouzid, M., F. J. Colón-González, T. Lung, PublicAffairs. I. R. Lake, and P. R. Hunter. 2014. “Climate Bartram, J., and S. Cairncross. 2010. “Hygiene, Change and the Emergence of Vector-Borne Sanitation, and Water: Forgotten Foundations Diseases in Europe: Case Study of Dengue of Health.” PLoS Med 7, e1000367. Fever.” BMC Public Health 14: 781. Bassuk, E. L., J. C. Buckner, J. N. Perloff, and Buhaug, H., J. Nordkvelle, T. Bernauer, T. Böhmelt, S. S. Bassuk. 1998. “Prevalence of Mental M. Brzoska, J. W. Busby, A. Ciccone, H. Fjelde, Health and Substance Use Disorders among E. Gartzke, N. P. Gleditsch, J. A. Goldstone, Homeless and Low-Income Housed Mothers.” H. Hegre, H. Holtermann, V. Koubi, J. S. A. Am. J. Psychiatry 155: 1,561–64. Link, P. M. Link, P. Lujala, J. O’Loughlin, Béguin, A., S. Hales, J. Rocklöv, C. Åström, C. Raleigh, J. Scheffran, J. Schilling, T. G. V. R. Louis, and R. Sauerborn. 2011. “The Smith, O. M. Theisen, R. S. J. Tol, H. Urdal, Opposing Effects of Climate Change and and N. von Uexkull. 2014. “One Effect to Socioeconomic Development on the Global Rule Them All? A Comment on Climate and Distribution of Malaria.” Glob. Environ. Conflict.” Clim. Change 127: 391–97. Change 21: 1,209–14. Burke, M., S. M. Hsiang, and E. Miguel. 2014. Berry, H. L., K. Bowen, and T. Kjellstrom. 2010. “Climate and Conflict.” Working Paper “Climate Change and Mental Health: A Causal No. 20598, National Bureau of Economic Pathways Framework.” Int. J. Public Health Research. 55: 123–32. Cachon, G., S. Gallino, and M. Olivares. 2012. Beugnet, F., and K. Chalvet-Monfray. 2013. “Severe Weather and Automobile Assembly “Impact of Climate Change in the Epidemiology Productivity.” Columbia Bus. Sch. Res. Pap. of Vector-Borne Diseases in Domestic Caminade, C., S. Kovats, J. Rocklov, A. M. Tompkins, Carnivores.” Comp. Immunol. Microbiol. A. P. Morse, F. J. Colón-González, H. Stenlund, Infect. Dis. 36: 559–66. P. Martens, and S. J. Lloyd. 2014. “Impact of 1 3 4    SHOCK WAVES Climate Change on Global Malaria Distribution.” del Ninno, C., and M. Lundberg. 2005. “Treading Proc. Natl. Acad. Sci. 111: 3,286–91. Water: The Long-Term Impact of the 1998 Cann, K. F., D. R. Thomas, R. L. Salmon, Flood on Nutrition in Bangladesh.” Econ. A. P. Wyn-Jones, and D. Kay. 2013. “Extreme Hum. Biol. 3: 67–96. Water-Related Weather Events and Waterborne Delpla, I., A.-V., Jung, E. Baures, M. Clement, Disease.” Epidemiol. Infect. 141: 671–86. and O. Thomas. 2009. “Impacts of Climate Caruso, G. D. 2015. The Legacy of Natural Change on Surface Water Quality in Relation Disasters: The Intergenerational Impact of 100 to Drinking Water Production.” Environ. Int. Years of Natural Disasters in Latin America. 35: 1,225–33. Available SSRN 2560891. Dercon, S., and C. Porter. 2014. “Live Aid Checkley, W., G. Buckley, R. H. Gilman, Revisited: Long-Term Impacts of the 1984 A. M. Assis, R. L. Guerrant, S. S. Morris, Ethiopian Famine on Children.” J. Eur. Econ. K. Mølbak, P. Valentiner-Branth, C. F. Lanata, Assoc. 12: 927–48. and R. E. Black. 2008. “Multi-country Analysis Deryugina, T., and S. M. Hsiang, 2014. Does the of the Effects of Diarrhoea on Childhood Environment Still Matter? Daily Temperature Stunting.” Int. J. Epidemiol. 37: 816–30. and Income in the United States . National Checkley, W., L. D. Epstein, R. H. Gilman, Bureau of Economic Research. D. Figueroa, R. I. Cama, J. A. Patz, and Desfontaine, M., H. Gelas, A. Goghomu, R. E. Black. 2000. “Effects of EI Niño and D. Kouka-Bemba, and P. Carnevale. 1989. Ambient Temperature on Hospital Admissions “Evaluation of Practices and Costs of for Diarrhoeal Diseases in Peruvian Children.” Antivectorial Control at the Family Level in The Lancet 355: 442–50. Central Africa, I. Yaoundé City.” Bull. Société Chima, R. I., C. A. Goodman, and A. Mills. 2003. Pathol. Exot. Ses Fil. 82: 558–65. “The Economic Impact of Malaria in Africa: Ebi, K. L., and I. Burton. 2008. “Identifying A Critical Review of the Evidence.” Health Practical Adaptation Options: An Approach Policy 63: 17–36. to Address Climate Change-Related Health Chuma, J. M., M. Thiede, and C. S. Molyneux. Risks.” Environ. Sci. Policy 11: 359–69. 2006. “Rethinking the Economic Costs of Ebi, K. L., and G. McGregor. 2008. “Climate Malaria at the Household Level: Evidence Change, Tropospheric Ozone and Particulate from Applying a New Analytical Framework in Matter, and Health Impacts.” Environmental Rural Kenya.” Malar. J. 5: 76. Health Perspectives 116 (11): 1449–55. Clarke, D. J., and R. V. Hill. 2013. “Cost-Benefit Ebi, K. L., and J. C. Semenza. 2008. “Community- Analysis of the African Risk Capacity Facility.” Based Adaptation to the Health Impacts of IFPRI Discuss. Pap. 01292. International Food Climate Change.” Am. J. Prev. Med . 35: Policy Research Institute. 501–07. Cropper, M., M. Haile, J. A. Lampietti, C. Poulos, Ebi, K. L., T. J. Teisberg, L. S. Kalkstein, and D. Whittington. 2000. “The Value of L. Robinson, and R. F. Weiher. 2004. Preventing Malaria in Tembien, Ethiopia.” Policy “Heat Watch/Warning Systems Save Lives: Res. Work. Paper, World Bank, Washington, DC. Estimated Costs and Benefits for Philadelphia Currie, J. 2009. “Healthy, Wealthy, and Wise: 1995–98.” Bull. Am. Meteorol. Soc. 85: Socioeconomic Status, Poor Health in 1,067–73. Childhood, and Human Capital Development.” ESPD (Economic and Social Policy Division). J. Econ. Lit. 47: 87–122. 2005. “Enhancing Health Systems: Malaria’s D’Amato, G., L. Cecchi, M. D’Amato, and Negative Impact in Africa.” ESPD Poverty Soc. I. Annesi-Maesano. 2014. “Climate Change Policy Team Geneva 40. and Respiratory Diseases.” European Ettling, M., D. McFarland, L. Schultz, and Respiratory Review 23 (132): 161–69. L. Chitsulo 1994. “Economic Impact of Dell, M., B. F. Jones, and B. A. Olken. 2009. Malaria in Malawian Households: A Nation- “Temperature and Income: Reconciling New Wide Malaria Knowledge, Attitudes and Cross-Sectional and Panel Estimates.” Am. Practices Survey in Malawi.” Trop. Med. Econ. Rev. 99: 198–204. Parasitol. 45: 74–79. ———. 2012. “Temperature Shocks and Economic Fang, Y., D. L. Mauzerall, J. Liu, A. M. Fiore, Growth: Evidence from the Last Half Century.” and L. W. Horowitz. 2013. “Impacts of Am. Econ. J. Macroecon. 4: 66–95. 21st Century Climate Change on Global Air Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    135 Pollution-Related Premature Mortality.” and Depressive Reactions among Nicaraguan Climatic Change 121 (2): 239–53. Adolescents after Hurricane Mitch.” Am. FAO (Food and Agriculture Organization of the J. Psychiatry. 158 (5): 788–94. United Nations). 2015. The State of Food Grantham-McGregor, S. 1995. “A Review of Insecurity in the World 2015 . Rome, Italy: Studies of the Effect of Severe Malnutrition Food and Agriculture Organization. on Mental Development.” J. Nutr . 125: Fazel, S., V. Khosla, H. Doll, and J. Geddes. 2008. 2,233S–38S. “The Prevalence of Mental Disorders among Grantham-McGregor, S., C. Powell, S. Walker, the Homeless in Western Countries: Systematic S. Chang, and P. Fletcher. 1994. “The Long- Review and Meta-regression Analysis.” PLoS Term Follow-up of Severely Malnourished Med 5: e225. Children Who Participated in an Intervention Federspiel, C. C., W. J. Fisk, P. N. Price, G. Liu, Program.” Child Dev. 65, 428–39. D. Faulkner, D. L. DiBartolomeo, D. P. Sullivan, Grether, W. 1973. “Human Performance at and M. Lahiff. 2004. “Worker Performance Elevated Environmental Temperatures.” and Ventilation in a Call Center: Analyses of Aerosp. Med. 44: 747–55. Work Performance Data for Registered Guiguemde, T. R., F. Dao, V. Curtis, A. Traore, Nurses.” Indoor Air 14: 41–50. B. Sondo, J. Testa, and J. B. Ouedraogo. Fouillet, A., G. Rey, V. Wagner, K. Laaidi, 1994. “Household Expenditure on Malaria P. Empereur-Bissonnet, A. L. Tertre, P. Frayssinet, Prevention and Treatment for Families in the P. Bessemoulin, F. Laurent, P. D. Crouy-Chanel, Town of Bobo-Dioulasso, Burkina Faso.” E. Jougla, and D. Hémon. 2008. “Has the Impact Trans. R. Soc. Trop. Med. Hyg. 88: 285–87. of Heat Waves on Mortality Changed in France Hales, S., N. De Wet, J. Maindonald, and since the European Heat Wave of Summer A. Woodward. 2002. “Potential Effect of 2003? A Study of the 2006 Heat Wave.” Int. Population and Climate Changes on Global J. Epidemiol. 37: 309–17. Distribution of Dengue Fever: An Empirical Fritze, J. G., G. A. Blashki, S. Burke, and J. Wiseman. Model.” The Lancet 360: 830–34. 2008. “Hope, Despair and Transformation: Hales, S., S. Kovats, S. Lloyd, and D. Campbell- Climate Change and the Promotion of Mental Lendrum, 2014. Quantitative Risk Assessment Health and Wellbeing.” International Journal of of the Effects of Climate Change on Selected Mental Health Systems 2: 13. Causes of Death, 2030s and 2050s. Geneva, Gallup, J. L., and J. D. Sachs. 2001. “The Switzerland: World Health Organization. Economic Burden of Malaria.” Am. J. Trop. Hanigan, I. C., C. D. Butler, P. N. Kokic, and Med. Hyg. 64: 85–96. M. F. Hutchinson. 2012. “Suicide and Drought Gething, P. W., D. L. Smith, A. P. Patil, A. J. Tatem, in New South Wales, Australia, 1970–2007.” R. W. Snow, and S. I. Hay. 2010. “Climate Proc. Natl. Acad. Sci. 109: 13,950–55. Change and the Global Malaria Recession.” Heal, G., and J. Park. 2013. “Feeling the Heat: Nature 465: 342–45. Temperature, Physiology & the Wealth of Giedion, U., E. Andrés Alfonso, and Y. Díaz. Nations.” Working Paper No. 19725, National 2013. The Impact of Universal Coverage Bureau of Economic Research. Schemes in the Developing World: A Review Hii, Y. L., J. Rocklöv, N. Ng, C. S. Tang, of the Existing Evidence. Washington, DC: F. Y. Pang, and R. Sauerborn. 2009. “Climate World Bank. Variability and Increase in Intensity and Gilman, R. H., G. S. Marquis, G. Ventura, Magnitude of Dengue Incidence in Singapore.” M. Campos, W. Spira, and F. Diaz. 1993. Glob. Health Action 2. “Water Cost and Availability: Key Determinants Hoddinott, J. 2006. “Shocks and Their of Family Hygiene in a Peruvian Shantytown.” Consequences across and within Households Am. J. Public Health 83: 1,554–58. in Rural Zimbabwe.” J. Dev. Stud . 42: Githeko, A. K., S. W. Lindsay, U. E. Confalonieri, 301–21. and J. A. Patz. 2000. “Climate Change and Hogrefe, C., B. Lynn, K. Civerolo, J.-Y. Ku, Vector-Borne Diseases: A Regional Analysis.” J. Rosenthal, C. Rosenzweig, R. Goldberg, Bull. World Health Organ. 78: 1,136–47. S. Gaffin, K. Knowlton, and P. L. Kinney. Goenjian, A. K., L. Molina, A. M. Steinberg, 2004. “Simulating Changes in Regional Air L. A. Fairbanks, M. L. Alvarez, H. A. Goenjian, Pollution over the Eastern United States due to and R. S. Pynoos. 2001. “Posttraumatic Stress Changes in Global and Regional Climate and 1 3 6    SHOCK WAVES Emissions.” Journal of Geophysical Research: Jamison, D. T., L. H. Summers, G. Alleyne, Atmospheres 109 (D22): D22301. K. J. Arrow, S. Berkley, A. Binagwaho, F. Bustreo, Holding, P., J. Stevenson, N. Peshu, and K. Marsh. D. Evans, R. G. A. Feachem, J. Frenk, G. Ghosh, 1999. “Cognitive Sequelae of Severe Malaria S. J. Goldie, Y. Guo, S. Gupta, R. Horton, with Impaired Consciousness.” Trans. R. Soc. M. E. Kruk, A. Mahmoud, L. K. Mohohlo, Trop. Med. Hyg. 93: 529–34. M. Ncube, A. Pablos-Mendez, K. S. Reddy, Horowitz, J. K. 2009. “The Income–Temperature H. Saxenian, A. Soucat, K. H. Ulltviet-Moe, and Relationship in a Cross-Section of Countries G. Yamey. 2013. “Global Health 2035: A World and Its Implications for Predicting the Effects of Converging within a Generation.” The Lancet Global Warming.” Environ. Resour. Econ. 44: 382 (9908): 1,898–1,955. 475–93. Jensen, R. 2000. “Agricultural Volatility and Hoyt, K. S., and A. E. Gerhart. 2004. “The San Investments in Children.” Am. Econ. Rev. 90: Diego County Wildfires: Perspectives of 399–404. Healthcare Providers [corrected].” Disaster Johansson, M. A., D. A. Cummings, and G. E. Glass. Management & Response: DMR: An 2009. “Multiyear Climate Variability and Official Publication of the Emergency Nurses Dengue—El Nino Southern Oscillation, Association 2 (2): 46–52. Weather, and Dengue Incidence in Puerto Rico, Hsiang, S. M., M. Burke, and E. Miguel. 2013. Mexico, and Thailand: A Longitudinal Data “Quantifying the Influence of Climate on Analysis.” PLoS Med. 6: 1,262. Human Conflict.” Science 341: 1235367. Jones, G., R. W. Steketee, R. E. Black, ———. 2014. “Reconciling Climate-Conflict Meta- Z. A. Bhutta, and S. S. Morris, Bellagio Child analyses: Reply to Buhaug et al.” Clim. Change Survival Study Group. 2003. “How Many 127: 399–405. Child Deaths Can We Prevent this Year?” The Huang, C., A. G. Barnett, X. Wang, P. Vaneckova, Lancet 362: 65–71. G. FitzGerald, and S. Tong. 2011. “Projecting Jukes, M. 2005. “The Long-Term Impact of Future Heat-Related Mortality under Climate Preschool Health and Nutrition on Education.” Change Scenarios: A Systematic Review.” Food Nutr. Bull. 26: 193–201. Environ. Health Perspect. Keiser, J., M. C. de Castro, M. F. Maltese, R. Bos, Huntington, E. 1922. Civilization and Climate. M. Tanner, B. H. Singer, and J. Utzinger. New Haven, CT: Yale University Press. 2005. “Effect of Irrigation and Large Dams Hutton, G., and L. Haller. 2004. Evaluation of on the Burden of Malaria on a Global and the Costs and Benefits of Water and Sanitation Regional Scale.” Am. J. Trop. Med. Hyg. 72: Improvements at the Global Level. Geneva, 392–406. Switzerland: World Health Organization. Keshavarz, H., D. Fitzpatrick-Lewis, D. L. Streiner, IPCC (Intergovernmental Panel on Climate R. Maureen, U. Ali, H. S. Shannon, and Change). 2014. Climate Change 2014: P. Raina. 2013. “Screening for Depression: A Impacts, Adaptation, and Vulnerability. Part Systematic Review and Meta-analysis.” Can. A: Global and Sectoral Aspects. Contribution Med. Assoc. Open Access J. 1: E159–E167. of Working Group II to the Fifth Assessment Kinney, P. L. 2008. “Climate Change, Air Quality, Report of the Intergovernmental Panel on and Human Health.” American Journal of Climate Change. [Field, C.B., V.R. Barros, Preventive Medicine 35 (5): 459–67. D.J. Dokken, K.J. Mach, M.D. Mastrandrea, Kolstad, E. W., and K. A. Johansson. 2010. T.E. Bilir, M. Chatterjee, K.L. Ebi, Y.O. Estrada, “Uncertainties Associated with Quantifying R.C. Genova, B. Girma, E.S. Kissel, A.N. Levy, Climate Change Impacts on Human Health: S. MacCracken, P.R. Mastrandrea, and A Case Study for Diarrhea.” Environ. Health L.L. White (eds.)]. Cambridge, U.K. and New Perspect. 119: 299–305. York: Cambridge University Press. K o n r a d s e n , F. , W. v a n d e r H o e k , Iyer, L., and P. B. Topalova. 2014. “Poverty and P. H. Amerasinghe, F. P. Amerasinghe, and Crime: Evidence from Rainfall and Trade K. T. Fonseka. 1997. “Household Responses to Shocks in India.” SSRN Scholarly Paper No. Malaria and Their Costs: A Study from Rural ID 2419522, Social Science Research Network, Sri Lanka.” Trans. R. Soc. Trop. Med. Hyg. 91: Rochester, NY. 127–30. Jalan, J., and M. Ravallion. 2003. “Does Piped Krishna, A. 2006. “Pathways out of and into Water Reduce Diarrhea for Children in Rural Poverty in 36 Villages of Andhra Pradesh, India?” J. Econom. 112: 153–73. India.” World Dev., Part Special Issue: 324–404. Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    137 Corruption and Development: Analysis and Nutrition During Early Childhood on Measurement 34: 271–88. Education among Guatemalan Adults.” The ——— . 2 0 0 7 . “ T h e S t a g e s o f P r o g r e s s Economic Journal 119 (537): 734–63. Methodology and Results from Five Masson, V., Y. Lion, A. Peter, G. Pigeon, Countries.” Reducing Glob. Poverty Case J. Buyck, and E. Brun. 2013. “Grand Paris: Asset Accumul. 62–79. Regional Landscape Change to Adapt City to Krishna, A., P. Kristjanson, M. Radeny, and Climate Warming.” Climatic Change 117 (4): W. Nindo. 2004. “Escaping Poverty and 769–82. Becoming Poor in 20 Kenyan Villages.” Masson, V., C. Marchadier, L. Adolphe, J. Hum. Dev. 5: 211–26. R. Aguejdad, P. Avner, M. Bonhomme, Kronenberg, M. E., T. C. Hansel, A. M. Brennan, G. Bretagne, X. Briottet, B. Bueno, and C. de H. J. Osofsky, J. D. Osofsky, and B. Lawrason. Munck. 2014. “Adapting Cities to Climate 2010. “Children of Katrina: Lessons Learned Change: A Systemic Modeling Approach.” about Postdisaster Symptoms and Recovery Urban Climate 10: 407–29. Patterns.” Child Dev. 81: 1,241–59. Medlock, J. M., and A. G. Vaux. 2011. “Assessing Leighton, C., and R. Foster. 1993. Economic the Possible Implications of Wetland Expansion Impacts of Malaria in Kenya and Nigeria . and Management on Mosquitoes in Britain.” Bethesda, MD: Abt Associates. Eur. Mosq. Bull. 29: 38–65. Lelieveld, J., J. S. Evans, M. Fnais, D. Giannadaki, Moore, D., R. Copes, R. Fisk, R. Joy, K. Chan, and A. Pozzer. 2015. “The Contribution of and M. Brauer. 2006. “Population Health Outdoor Air Pollution Sources to Premature Effects of Air Quality Changes due to Forest Mortality on a Global Scale.” Nature 525 Fires in British Columbia in 2003: Estimates (7569): 367–71. from Physician-Visit Billing Data.” Canadian Liu, J., F. Ma, and Y. Li. 2011. “The Effect Journal of Public Health [Revue Canadienne of Anthropogenic Heat on Local Heat De Santé Publique] 97 (2): 105–08. Island Intensity and the Performance of Air Montesquieu, C. L. d. S. 1758. The Spirit of Laws. Conditioning Systems.” Adv. Mater. Res . Myers, S. S., A. Zanobetti, I. Kloog, P. Huybers, 250-53: 2,975–78. A. D. Leakey, A. J. Bloom, E. Carlisle, Lloyd, S. J., R. S. Kovats, and B. G. Armstrong. L. H. Dietterich, G. Fitzgerald, and T. Hasegawa. 2007. “Global Diarrhoea Morbidity, Weather 2014. “Increasing CO 2 Threatens Human and Climate.” Clim. Res. 34: 119. Nutrition.” Nature 510 (7503): 139–42. Lloyd, S. J., R. S. Kovats, and Z. Chalabi. Neuner, F., E. Schauer, C. Catani, M. Ruf, and 2011. “Climate Change, Crop Yields, and T. Elbert. 2006. “Post-tsunami Stress: A Study Undernutrition: Development of a Model to of Posttraumatic Stress Disorder in Children Quantify the Impact of Climate Scenarios Living in Three Severely Affected Regions in Sri on Child Undernutrition.” Environ. Health Lanka.” J. Trauma. Stress 19: 339–47. Perspect. 119: 1,817–23. Noji, E. K. 2000. “The Public Health Consequences Louis, J. P., A. Trebucq, H. Gelas, E. Fondjo, of Disasters.” Prehospital Disaster Med. 15: L. Manga, J. C. Toto, and P. Carnevale. 1992. 147–57. “Malaria in Yaounde (Cameroon): Cost and O’Donnell, O. 2007. “Access to Health Care Antivectorial Control at the Family Level.” in Developing Countries: Breaking Down Bull. Société Pathol. Exot. 1990 85: 26–30. Demand Side Barriers.” Cad. Saúde Pública 23: Lund, C., A. Breen, A. J. Flisher, R. Kakuma, 2,820–34. J. Corrigall, J. A. Joska, L. Swartz, and V. Patel. Okorosobo, T., F. Okorosobo, G. Mwabu, 2010. “Poverty and Common Mental Disorders J. N. Orem, and J. M. Kirigia. 2011. “Economic in Low and Middle Income Countries: Burden of Malaria in Six Countries of Africa.” A Systematic Review.” Soc. Sci. Med. 71: Eur. J. Bus. Manag. 3: 42–62. 517–28. Onwujekwe, O., R. Chima, and P. Okonkwo. Maccini, S., and D. Yang. 2009. “Under the 2000. “Economic Burden of Malaria Illness Weather: Health, Schooling, and Economic on Households versus That of All Other Illness Consequences of Early-Life Rainfall.” Am. Episodes: A Study in Five Malaria Holo- Econ. Rev. 99: 1,006–26. endemic Nigerian Communities.” Health Policy Maluccio, J. A., J. Hoddinott, J. R. Behrman, 54: 143–59. R. Martorell, A. R. Quisumbing, and A. D. Paranjothy, S., J. Gallacher, R. Amlôt, G. J. Rubin, Stein. 2009. “The Impact of Improving L. Page, T. Baxter, J. Wight, D. Kirrage, 1 3 8    SHOCK WAVES R. McNaught, and S. R. Palmer. 2011. Health Coverage (UNICO) Studies, No.5. “Psychosocial Impact of the Summer 2007 Washington, DC: World Bank. Floods in England.” BMC Public Health 11: 145. Ranson, M. 2014. “Crime, Weather, and Climate Park, J. Forthcoming. “Will We Adapt? Change.” J. Environ. Econ. Manag . 67: Temperature Shocks, Labor Productivity, and 274–302. Adaptation to Climate Change in the United Schmier, J. K., and K. L. Ebi. 2009. The Impact States 1986–2012.” of Climate Change and Aeroallergens on Park, J., S. Hallegatte, M. Bangalore, and Children’s Health. 229–237. Presented at the E. Sandhoefner. Forthcoming. “Households Allergy and Asthma Proceedings. OceanSide and Heat Stress: Estimating the Distributional Publications, Inc. Consequences of Climate Change.” Background Semenza, J. C., and B. Menne. 2009. “Climate paper prepared for this report. Change and Infectious Diseases in Europe.” Pascal, M., K. Laaidi, M. Ledrans, E. Baffert, Lancet Infect. Dis. 9: 365–75. C. Caserio-Schönemann, A. Le Tertre, Seppänen, O., W. J. Fisk, and Q. Lei. 2006. Effect J. Manach, S. Medina, J. Rudant, and of Temperature on Task Performance in Offfice P. Empereur-Bissonnet. 2006. “France’s Environment . Lawrence Berkeley National Heat Health Watch Warning System.” Int. Laboratory. J. Biometeorol. 50: 144–53. Singh, R. B., S. Hales, N. de Wet, R. Raj, Patankar, A. Forthcoming. “The Exposure, M. Hearnden, and P. Weinstein. 2001. “The Vulnerability, and Ability to Respond of Poor Influence of Climate Variation and Change Households to Recurrent Floods in Mumbai.” on Diarrheal Disease in the Pacific Islands.” Background paper prepared for this report. Environ. Health Perspect. 109: 155. Patel, V., and A. Kleinman. 2003. “Poverty and Siraj, A., M. Santos-Vega, M. Bouma, D. Yadeta, Common Mental Disorders in Developing D. R. Carrascal, and M. Pascual. 2014. Countries.” Bull. World Health Organ. 81: “Altitudinal Changes in Malaria Incidence in 609–15. Highlands of Ethiopia and Colombia.” Science Patz, J. A., D. Campbell-Lendrum, T. Holloway, 343: 1,154–58. and J. A. Foley. 2005. “Impact of Regional Small, J., S. J. Goetz, and S. I. Hay. 2003. Climate Change on Human Health.” Nature “Climatic Suitability for Malaria Transmission 438: 310–17. in Africa, 1911–1995.” Proc. Natl. Acad. Sci. Paz, S. 2009. “Impact of Temperature Variability 100: 15,341–45. on Cholera Incidence in Southeastern Africa, Smith, K. R., A. Woodward, D. Campbell- 1971–2006.” Ecohealth 6: 340–45. Lendrum, D. D. Chadee, Y. Honda, Q. Liu, Peel, J. L., K. B. Metzger, M. Klein, W. D. Flanders, J. M. Olwoch, B. Revich, and R. Sauerborn. J. A. Mulholland, and P. E. Tolbert. 2007. 2014. “Human Health: Impacts, Adaptation, “Ambient Air Pollution and Cardiovascular and Co-benefits.” In Climate Change 2014: Emergency Department Visits in Potentially Impacts, Adaptation, and Vulnerability. Part Sensitive Groups.” American Journal of A: Global and Sectoral Aspects. Contribution Epidemiology 165 (6): 625–33. of Working Group II to the Fifth Assessment Peel, J. L., P. E. Tolbert, M. Klein, K. B. Metzger, Report of the Intergovernmental Panel on W. D. Flanders, K. Todd, J. A. Mulholland, Climate Change. [Field, C.B., V.R. Barros, P. B. Ryan, and H. Frumkin. 2005. “Ambient D.J. Dokken, K.J. Mach, M.D. Mastrandrea, Air Pollution and Respiratory Emergency T. E .   B i l i r, M . C h a t t e r j e e , K . L . E b i , Department Visits.” Epidemiology (Cambridge, Y.O. Estrada, R.C. Genova, B. Girma, E.S. Mass.) 16 (2): 164–74. K i s s e l , A . N . L e v y, S . M a c C r a c k e n , Qadri, F., A. I. Khan, A. S. Faruque, Y. A. Begum, P.R. Mastrandrea, and L.L. White (eds.)]. F. Chowdhury, G. B. Nair, M. A. Salam, Cambridge, U.K. and New York: Cambridge D. A. Sack, and A. M. Svennerholm. 2005. University Press. “Enterotoxigenic Escherichia Coli and Vibrio Smith, L. C., and L. J. Haddad. 2000. Explaining Cholerae Diarrhea, Bangladesh, 2004.” Emerg Child Malnutrition in Developing Countries: Infect Dis. 11 (7): 1104–07. A Cross-Country Analysis. International Food Ramana, G. N., R. Chepkoech, and N. W. Policy Research Institute. Workie. 2013. Kenya - Improving Universal Stephenson, L. S., M. C. Latham, and E. Ottesen. Primary Health Care by Kenya: A Case Study 2000. “Malnutrition and Parasitic Helminth of the Health Sector Services Fund, Universal Infections.” Parasitology 121: S23–S38. Under the W eather : C limate C hange , H ealth , and the I ntergenerational T ransmission of P overt y    139 Stone, B., J. J. Hess, and H. Frumkin. 2010. M. Lott, R. Lowe, Y. Luo, G. Mace, M. Maslin, “Urban Form and Extreme Heat Events: Are M. Nilsson, T. Oreszczyn, S. Pye, T. Quinn, Sprawling Cities More Vulnerable to Climate M. Svensdotter, S. Venevsky, K. Warner, Change than Compact Cities?” Environ. Health B. Xu, J. Yang, Y. Yin, C. Yu, Q. Zhang, Perspect. 118: 1,425–28. P. Gong, H. Montgomery, and A. Costello. 2015. Sudarshan, A., E. Somanathan, R. Somanathan, “Health and Climate Change: Policy Responses and M. Tewari. 2015. “The Impact of to Protect Public Health.” The Lancet. Temperature on Productivity and Labor Weems, C. F., A. A. Pina, N. M. Costa, S. E. Watts, L. Supply: Evidence from Indian Manufacturing.” K. Taylor, and M. F. Cannon. 2007. “Predisaster Work. Pap. No 244, Centre for Development Trait Anxiety and Negative Affect Predict Economics, Delhi School of Economics. Posttraumatic Stress in Youths after Hurricane Sur, D., P. Dutta, G. B. Nair, and S. K. Bhattacharya. Katrina.” J. Consult. Clin. Psychol. 75: 154. 2000. “Severe Cholera Outbreak Following Wesolowski, A., T. Qureshi, M. F. Boni, Floods in a Northern District of West Bengal.” P. R. Sundsøy, M. A. Johansson, S. B. Rasheed, Indian J. Med. Res. 112: 178–82. K. Engø-Monsen, and C. O. Buckee. 2015. Sutherst, R. W. 2004. “Global Change and Human “Impact of Human Mobility on the Emergence Vulnerability to Vector-Borne Diseases.” Clin. of Dengue Epidemics in Pakistan.” Proceedings Microbiol. Rev. 17: 136–73. of the National Academy of Sciences , Takaro, T. K., K. Knowlton, and J. R. Balmes. September. 201504964. 2013. “Climate Change and Respiratory Whaley, S. E., M. Sigman, M. P. Espinosa, and Health: Current Evidence and Knowledge C. G. Neumann. 1998. “Infant Predictors of Gaps.” Expert Review of Respiratory Medicine Cognitive Development in an Undernourished 7 (4): 349–61. Kenyan Population.” J. Dev. Behav. Pediatr. Tibbetts, J. 2015. “Air Quality and Climate 19: 169–77. Change: A Delicate Balance.” Environmental WHO (World Health Organization). 2002. The Health Perspectives 123 (6): A148. World Health Report. Geneva, Switzerland: Trouiller, P., E. Torreele, P. Olliaro, N. White, World Health Organization. S. Foster, D. Wirth, and B. Pécoul. 2001. ———. 2003. Summary Booklet: Climate Change “Drugs for Neglected Diseases: A Failure of the and Human Health – Risks and Responses. Market and a Public Health Failure?” Trop. Geneva, Switzerland: World Health Organization. Med. Int. Health 6, 945–951. ———. 2008. Global Burden of Disease: 2004 UNICEF (United Nations Children’s Fund). 2013. Update. Geneva, Switzerland: World Health Improving Child Nutrition: The Achievable Organization. Imperative for Global Progress . New York: ———. 2013a. Diarrhoeal Disease: Fact United Nations Children’s Fund. Sheet . Geneva, Switzerland: World Health Victora, C. G., L. Adair, C. Fall, P. C. Hallal, Organization. R. Martorell, L. Richter, and H. S. Sachdev. ———. 2013b. Universal Health Coverage: Report 2008. “Maternal and Child Undernutrition: by the Secretariat. Geneva, Switzerland: World Consequences for Adult Health and Human Health Organization. Capital.” The Lancet 371: 340–57. ———. 2014. World Malaria Report 2014 . Vir, S. C. 2011. Public Health and Nutrition in G e n e v a ,   S w i t z e r l a n d : Wo r l d H e a l t h Developing Countries (Part I and II). Boca Organization. Raton, FL: CRC Press. ———. 2015. Global Health Observatory Data Warraich, H., A. K. Zaidi, and K. Patel. 2011. Repository. World Health Organization. http:// “Floods in Pakistan: A Public Health Crisis.” apps.who.int/gho/data/node.main. Bull. World Health Organ. 89: 236–37. Wilder-Smith, A., and D. J. Gubler. 2008. Watson, J. T., M. Gayer, and M. A. Connolly. “Geographic Expansion of Dengue: The Impact 2007. “Epidemics after Natural Disasters.” of International Travel.” Med. Clin. North Am. Emerg. Infect. Dis. 13: 1–5. 92: 1,377–90. Watts, N., W. N. Adger, P. Agnolucci, J. Blackstock, World Bank. 2006. Repositioning Nutrition as P. Byass, W. Cai, S. Chaytor, T. Colbourn, Central to Development: A Strategy for Large- M. Collins, A. Cooper, P. M. Cox, J. Depledge, Scale Action. Washington, DC: World Bank. P. Drummond, P. Ekins, V. Galaz, D. Grace, ———. 2011. Conflict, Security, and Development, H. Graham, M. Grubb, A. Haines, I. Hamilton, World Development Report 2011. Washington, A. Hunter, X. Jiang, M. Li, I. Kelman, L. Liang, DC: World Bank. 1 4 0    SHOCK WAVES ———. 2013. Risk and Opportunity: Managing World Bank and Australian Agency for International Risk for Development, World Development Development (AUS AID). 2014. “Where Are We Report 2014. Washington, DC: World Bank. During Flooding?” A Qualitative Assessment ———. 2014. Health Nutrition and Population of Poverty and Social Impacts of Flooding in Statistics by Wealth Quintile. Washington, DC: Selected Neighborhoods of HCMC. World Bank. Yamano, T., H. Alderman, and L. Christiaensen. ———. 2015a. Global Monitoring Report 2005. “Child Growth, Shocks, and Food Aid in 2014/2015: Ending Poverty and Sharing Rural Ethiopia.” Am. J. Agric. Econ. 87: 273–88. Prosperity. Washington, DC: World Bank. Zivin, J. G., and M. Neidell. 2014. “Temperature ———. 2015b. World Development Indicators. and the Allocation of Time: Implications for Washington, DC: World Bank. Climate Change.” J. Labor Econ. 32: 1–26. Lend a Hand: Poor People, Support Systems, Safety Nets, 5 and Inclusion Main Messages • Poor people struggle more than others to • Given the limits to how much protection the cope with and adapt to climate change and financial system can offer, especially to the natural hazards: not only are they more poorest, social safety nets are needed to pro- exposed and vulnerable to shocks but the vide effective protection to poor households. support they receive from families, commu- To be effective, safety nets must be rapidly nities, financial system, and government is scalable, even if speed of delivery may come also weaker, and they are often not granted a at the cost of targeting. voice in decision-making processes. • An adaptive social protection system creates • Financial inclusion, insurance, social safety a formal liability for the government, which nets, and remittances complement each other may need to draw on instruments such as in protecting different populations against reserve funds, contingent finance, reinsur- different types of shocks. ance products, or even international aid. Introduction a shock or a change in economic conditions (like higher food or energy prices), they may Poor people are particularly exposed and vul- adapt to reduce the losses or even benefit nerable to the physical impacts of climate from the changes. Hence, the overall impact change, such as reduced crop yields, more on welfare and quality of life also depends on intense floods, or lower productivity due to how well people cope and adapt. extreme temperature, making climate change One reason why poor people struggle to and disasters a magnifier of existing inequali- adapt to changes in environmental and eco- ties. However, these direct impacts tell only nomic conditions is limited resources—for part of the story. When people are affected by instance, limited financial resources can push 141 1 4 2    SHOCK WAVES poor people to live in the most flood-prone change and shocks arises not just from the areas of cities, even if they are aware of the fact that they are more exposed and vulnera- risk (chapter 3). In addition, as discussed in ble but also because they receive less support the World Bank’s World Development from financial instruments, social protection Report 2014: Risk and Opportunity, the schemes, and private remittances. For ability to manage risk also depends on instance, in response to flooding and the “support systems” available to them: the ­ landslides in communities in Nepal in 2011, household, the community, the enterprise only 6 percent of the very poor sought gov- and financial sectors, and the state. ernment support, compared to almost These support systems are also critical to 90 ­ p ercent of the well-off (Gentle et al. help individuals and firms adapt to the 2014). Besides suffering from larger immedi- effects of climate change, cope with the ate shocks than the wealthier, poor people impacts that cannot be avoided, and deal also tend to be more alone in the struggle to with the potential adverse side effects from cope and recover. emissions-reduction policies. The key message of this chapter is the This chapter investigates how these sys- need for a holistic and flexible risk man- tems support people in the face of natural agement strategy—with a range of policy hazards, environmental changes, and eco- instruments appropriate for different disas- nomic and policy transitions. It assesses the ters and affected populations (figure 5.1). obstacles that prevent effective risk manage- Revenue diversification and basic social ment and adaptation to environmental and protection, where it exists, can help house- economic changes—and then suggests avail- holds at all income levels cope with small able policies to help poor people adapt to shocks. But, when a shock is larger, these climate change and cope with its conse- instruments will not be sufficient, and addi- quences. It follows by reviewing the role of tional tools are needed. For relatively financial instruments, social protection sys- wealthier households, savings will help; tems and safety nets, and migrations and and market insurance can provide them remittances, closing with thoughts on gover- with efficient protection for larger losses. nance and the poor. However, for the poorest households, sav- The chapter finds that poor people’s ings are often not an option; and high d isproportionate suffering from climate ­ transaction costs make private insurance unattainable. For the poorest households—and to FIGURE 5.1  Poorer households need different types of solutions cover the largest shocks—well-targeted and easily scalable social safety nets are needed. More International aid intense These systems need to be designed so as to events Social insurance maintain incentives to invest in long-term and scaled-up adaptation to economic and environmental social safety nets Market insurance changes. Such an adaptive social protection Government system creates a liability for the govern- insurance and contingent finance ment, which may need to rely on financial Savings, credit, and instruments such as reserve funds (for Government scaled-up remittances small-scale events), contingent finance, rein- reserve funds surance products, or even international aid Smaller Basic social protection, remittances, if its capacity is exhausted. Social protec- events and revenue diversification tion systems and financial protection have always been needed to help people cope Poorer Richer with individual and systemic shocks, includ- households households ing disease and natural disasters, but this L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   143 need has been intensified and made more Poor people lack access to savings urgent by climate change. and credit, but policies can enhance financial inclusion Saving, borrowing, and Unfortunately poor people often lack access to formal financial instruments—possibly insurance help people adapt to because of the cost of bank accounts, large changes and cope with shocks, distance and time to access a financial agent, but are not always accessible for or lack of documentation and mistrust in banks. Some people also prefer to stay in the poor people informal sector, or are not fully aware of the People use financial instruments—notably benefits of using financial tools for risk man- their savings and assets—to smooth consump- agement (Allen et al. 2012). For instance, in tion and limit the effects of income shocks. Indonesia, more than 20 percent of house- While evidence suggests that this smoothing is holds in the top 60 percent of the income significant, it is also far from complete (Kinnan distribution saved at a financial institution in and Townsend 2012; Morduch 1995). In 2011, compared to less than 8 percent from extreme cases, households have even been the bottom 40 percent. Poor people also found to reduce consumption to protect their have less access to credit when affected assets, thus exacerbating consumption shocks by a shock or worsening environmental (Zimmerman and Carter 2003). In general conditions. however, financial instruments (such as bank What can governments do? The 2014 accounts and insurance contracts) can com- World Development Report (World Bank plement informal risk-sharing mechanisms 2013a, chapter 6) discusses the actions that and play a critical role in helping households governments can take to help households at and firms adapt to climate change, prepare for all income levels gain access to financial natural shocks, and recover when affected instruments for risk management. Two cate- (World Bank 2013a, chapters 4 and 6). gories emerge. We know that savings at financial institu- Financial infrastructure. To reduce costs tions tend to be less vulnerable to natural and improve trust in the banking system, gov- disasters than in-kind savings (chapter 3). ernments can strengthen the financial infra- Livestock and housing can be washed away by structure (including payment and security a flood, while savings accounts are far less settlement systems and public credit regis- likely to be affected. Financial inclusion makes tries). Physical access to financial instruments it possible for households to protect their sav- can be improved by using the postal network, ings and lower the vulnerability of their asset and improving infrastructure for hosting portfolio, so they can cope with income losses financial agents and facilitating transport while maintaining consumption and avoiding (roads and public transit). Wide coverage by radical coping measures, such as reduced food mobile phone networks can make financial intake. instruments accessible virtually, by using cel- Financial inclusion also enables people and lular banking and electronic payment tech- firms to reduce risk in the first place. For nologies (Gupta 2013). instance, if changes in rainfall patterns call for Competition, protection, and flexibility. adapting farming activities, farmers will need Governments can help keep service costs low to invest in new machines and seeds, or pos- by ensuring fair competition and consumer sibly learn new techniques. Without access to protection, or by requiring the introduction credit, these measures may be unaffordable, of low-cost bank accounts for vulnerable thus locking them into activities with declin- populations. Moreover, ensuring convenience ing productivity and income. and flexibility is critical for accommodating 1 4 4    SHOCK WAVES the relatively high frequency at which poor The development of insurance markets people tend to make deposits and withdraw- in low-income economies faces many als, or take out and repay loans. For instance, obstacles poor households benefit greatly from flexible To cope with large shocks that affect many loan schedules that can be readily renegoti- people, savings or borrowing may not be ated or forborne in “hungry months,” or paid adequate. Instead, insurance products can in advance when the household enjoys extra provide protection at a lower cost. However, liquidity. Experience from microcredit also insurance markets are complex, and behav- demonstrates the benefits from schemes that iors often deviate from what theory suggests, create self-discipline—like using planned sav- making it challenging to provide appropriate ing schedules (Banerjee and Duflo 2012). insurance products to poor households or Poor people are also less able to protect them- small firms in developing countries, which selves against fraud and abuse, underscoring are often exposed to many risks (Kunreuther, the need for adequate and accessible con- Pauly, and McMorrow 2013). sumer protection schemes. In Mexico and The classical indemnity insurance products South Africa, the government has established are commonplace in high-income countries financial ombudsmen to resolve disputes in and are based on the observation of losses, consumer finance (Brix and McKee 2010). with insurance payments triggered once losses While borrowing can help maintain con- occur. Classical indemnity insurance requires sumption in the short term, it can also create that robust data be available for the insurer a debt trap from which poor households have to assess risks ex ante—something that is trouble escaping. For instance, health shocks often lacking in developing countries (Rogers are more likely to push people into poverty in and Tsirkunov 2013). And loss assessment the presence of high borrowing costs, may be costly if it requires that an expert visit ­ precisely because of such trap effects (Krishna every victim. 2006). Stringent borrowing conditions paired One problem with indemnity insurance with postdisaster destitution mean that poor is that asymmetric information results households affected by disasters are likely to in  adverse selection and moral hazard . quickly incur high levels of debt. Adverse selection refers to the fact that In Bangladesh after the 1998 floods, if the price of insurance cannot be adjusted ­ borrowing was by far the most common cop- to the level of risk that clients face— ing mechanism chosen by a sample of 757 because the information is unavailable or households in a postdisaster survey (del Ninno too costly to collect—then those clients fac- et al. 2001). Almost 60 percent of these house- ing more risk will demand more insurance, holds were in debt in the months immediately threatening the sustainability of the insur- following the floods, with average debt rising ance scheme. But this problem can be to almost 1.5 months of average consump- solved by making insurance mandatory, tion. Furthermore, 57 percent of flood- although it will be tough to do in countries affected households in the bottom three with little capacity or where premiums quintiles resorted to purchasing food on would be particularly high. credit. This borrowing mitigated the shock, Moral hazard refers to the fact that people but higher prices meant that poor flood- protected against the negative impact of a affected households consumed less. The shock could choose to do less to prevent the ­ financial cost of borrowing was found to vary shock or reduce associated losses. In the case widely, with interest rates from zero (for of insurance against natural hazards, moral many loans from family and neighbors) to an hazard is mitigated by the fact that house- average of 50 percent for loans from banks holds suffer from significant nonmonetary and cooperatives. Better access to finance ex uninsured losses (such as the hassle of reloca- ante and lower interest rates could have tion and loss of personal property if one’s reduced the debt trap and improved recovery. L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   145 house is flooded). In most contracts, it is miti- government, the case of France’s Cat-Nat gated further by deductibles, which ensure storm and drought insurance (Paudel 2012). that the insured face a portion of the losses in However, subsidizing insurance can be pro- case of a shock. hibitively expensive, and the penetration of Even in developed countries, penetration indemnity insurance is very low in developing of indemnity insurance against natural haz- countries. ards remains low, albeit with some excep- These problems have led to the develop- tions: (i) when insurance is subsidized, as with ment of index-based insurance products, in floods in the United States with the National which insurance payments are not made Flood Insurance Program; (ii) when insurance based on observed losses, but when a phys- is mandatory and backed by the g ­ overnment— ical variable—such as a rainfall deficit or such as the Turkish insurance against earth- wind speed—exceeds a predetermined quakes and fires, which is an excellent threshold (regardless of the existence of example of how insurance access can be losses). For instance, a farmer will receive increased in middle-income countries a predefined insurance payment if rainfall (box 5.1); or (iii) when insurance is manda- is below a minimum threshold over a one- tory, cross-subsidized, and backed by the month period. Index-based insurance BOX 5.1  Developing catastrophe insurance in Turkey through public-private partnerships Few countries in the world are more exposed to Insurance Pool (TCIP)—the first national catastro- earthquakes than Turkey. Around 70 percent of its phe insurance pool in World Bank partner countries population and 75 percent of its industrial facilities that provides a stand-alone earthquake insurance are exposed to large-scale earthquakes. Since 1984, coverage to homeowners and small and medium direct property and infrastructure losses due to earth- enterprises. The TCIP provides mandatory property quake episodes in Turkey have frequently exceeded earthquake insurance for owners of private dwell- $5 billion (in current US$ terms). The last major ings built legally on registered land. Premium rates earthquake in the Marmara region in 1999 resulted are actuarially sound, not subsidized, and vary with in the loss of 15,000 lives and placed an enormous construction type and property location. Covered financial burden on the economy and the government. risks include earthquakes and fire. Before 1999, earthquake insurance uptake The catastrophe risk financing strategy of the had traditionally been low in Turkey (at around TCIP relies on both risk retention and reinsurance. 3 ­p ercent of residential buildings) because house- The TCIP absorbs the first $80 million of losses holds traditionally relied on the government to through its reserves (initially complemented by a finance the reconstruction of private property after $100 million World Bank contingent loan facility) major natural disasters. This presented massive and transfers excess losses to the international rein- challenges to government budgets. But, in the after- surance markets. The government covers losses that math of the Marmara earthquake, the government would exceed the overall claims-paying capacity of decided to develop a catastrophe risk insurance the TCIP, which is estimated to be able to withstand mechanism to reduce its fiscal exposure to natural a 1-in-350 year earthquake. Economies of scale are hazards—­ a rising from publicly funded reconstruc- obtained through countrywide pooling of the risk tion of private property. In 2000, it created a com- and transaction costs, which results in more afford- pulsory earthquake insurance system for all residen- able premium rates. tial buildings on registered land in urban areas. The World Bank provided financial and techni- Source: Global Facility for Disaster Reduction and Recovery (GFDRR) policy note cal assistance for creating the Turkish Catastrophe on the TCIP. More information can be found in Gurenko 2006. 1 4 6    SHOCK WAVES schemes have major advantages compared Overall, evidence suggests that the take-up with traditional contracts: (i) transaction of index-based insurance requires large costs are reduced because losses do not ­ s ubsidies, although, as with indemnity need to be measured; (ii) individuals are insurance, subsidies can make the schemes still encouraged to take preventive mea- unsustainable (Brown, Zelenska, and sures since the payout does not depend on Mobarak 2013; Cole et al. 2012; Cole the losses or the actions taken to reduce et al. 2013). risks (in other words, there is no moral Some of these obstacles can be removed hazard with index-based insurance); and by improving technology, policy design, (iii) the payment decision is simple and and adopting best practices—for example, objective, making it easier to enforce modernizing observation systems and contracts. improving index designs may reduce the However, index-based insurance suffers basis risk and strengthen index-based instru- from basis risk (that is, the difference ments (Barnett, Barrett, and Skees 2008; between the payment received by contract Rogers and Tsirkunov 2013). holders and the actual losses they suffer). If the index is well correlated with actual losses, contract holders will receive an adequate insurance payment when (and Social protection schemes are only when) they have losses. But, in prac- critical for helping people adapt tice, the correlation between losses and and cope with shocks, but must payout can be low, because of wide varia- tions in impacts from natural hazards and be flexible and easily scalable limitations of hydrometeorological obser- Against this backdrop of the poor being at vation systems. This means that people a major disadvantage in terms of financial may receive a payment in the absence of resources, it is critical that governments losses, or receive nothing even in the pres- also provide social protection—that is any ence of large losses—which would be cata- government program concerned with pre- strophic for those close to the subsistence venting, managing, and overcoming situa- level. t i o n s t h a t a d v e r s e l y a f f e c t p e o p l e ’s Despite its advantages, the take-up of well-being. Social protection schemes can index-based insurance is low, with several act as a crucial complement to formal risk reasons being proffered (Brown, Zelenska, management tools provided by markets. and Mobarak 2013; Cole et al. 2012; Cole They also complement informal support et al. 2013). One is that basis risk plays a from communities and informal insurance, key role, because a low correlation between which tend to be insufficient in the face of losses and payout undercuts the product’s large or systemic shocks, and too often benefits (Karlan et al. 2012; Mobarak exclude the most vulnerable (World Bank and Rosenzweig 2013). Another is that 2013a). index insurance typically covers only one The three main types of social protection type of risk, while producers may be are (i) social safety nets (also known as social exposed to many (like price risk or supply assistance), which include conditional and chain risk). Other reasons include a general unconditional cash transfers, public work distrust in the insurance policy, limited programs, subsidies, and food stamps; financial ­l iteracy, and insufficient under- (ii)  social insurance, which consists of con- standing of the product. The decision to tributory pensions and contributory health purchase an insurance contract may hinge insurance; and (iii) labor market measures, on whether the individual has had prior which include instruments such as unemploy- experience with it (especially having ment compensation (table 5.1) (World Bank received a p ­ ayout) (Karlan et al. 2012). 2012, 2015a).1 L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   147 TABLE 5.1  Social protection includes safety nets, social insurance, and labor market policies Category Examples Social safety nets Conditional and unconditional cash transfers, including noncontributory pensions and disability, (or social assistance) birth and death allowances Food stamps, rations, emergency food distribution, school feeding and food subsidies Cash or food-for-work programs Free or subsidized health services Housing and utility subsidies Scholarships and fee waivers Social insurance Old age, survivor, and disability contributory pensions Occupational injury benefits, sick or maternity leave Health insurance Labor market policies Unemployment, severance, and early retirement compensation Training, job sharing, and labor market services Wage subsidies and other employment incentives, including for disabled people Source: ASPIRE documentation, World Bank 2015a. Social safety nets can reduce the response, the incomes of those in the bottom poverty impact of disasters and 20th percentile of the distribution would have economic shocks fallen by over 8 percent; by expanding the cash transfer program this fall was reduced to A growing body of evidence shows that social 5 percent (Grosh, Bussolo, and Freije 2014). insurance and social safety nets are efficient Similar instruments can help poor people cope tools to support poor people affected by with increases in food or energy prices (due to disasters or environmental and economic droughts or ambitious climate policies). shocks. In Kenya, the Hunger Safety Net But social protection may be less effective Program prevented a 5 percent increase in at protecting against prolonged adverse poverty among beneficiaries following the trends, such as sea level rise. A background 2011 drought (Merttens et al. 2013). In paper for this report explores how including Bangladesh, the Chars Livelihood Program nonpoor but vulnerable households in social protected 95 percent of recipients from losing protection can prevent them from falling into their assets after the 2012 floods (Kenward, poverty. In the long term, this reduces the Cordier, and Islam 2012). In Mexico, benefi- number of people in poverty, and thus allows ciaries of Prospera, the national cash transfer spending more to support each poor person. program (previously known as Oportunitades (Carter and Janzen, forthcoming). The same or Progresa), are less likely to withdraw their paper suggests that there is a limit: if shocks children from the classrooms following a become too frequent and intense, social safety shock (de Janvry et al. 2006; Fiszbein, Schady, nets become inefficient and livelihood changes and Ferreira 2009; Gertler 2004). are needed. In the short term, social protection helps mitigate adverse effects on livelihoods during economic crises (Akresh, De Walque, and Social protection can support long-term Kazianga 2013; Handa et al. 2015) and disas- transformations toward more adaptive ter shocks (World Bank 2012). In Latin and resilient societies if it does not lock America, social safety nets played a critical people into unsustainable locations or role in helping poor people cope with the activities food, fuel, and financial crisis in 2008. In Mexico, the expansion of the Progresa pro- Poor people, with fewer resources, tend to gram significantly mitigated the impacts of the invest less in preventing and mitigating crisis for the poor: without the policy adverse effects of natural hazards and 1 4 8    SHOCK WAVES environmental changes. In China, Indonesia, Social protection can also improve the Philippines, Thailand, and Vietnam, e ­ ducation and health levels, improving poor wealthier households are more likely to take people’s ability to escape poverty and adapt proactive ex ante adaptation measures, while to environmental and economic changes poorer households mostly react to shocks ex (Adger et al. 2014). post (Francisco et al. 2011). In addition, • In Burkina Faso, cash transfers (condi- poorer individuals, lacking resources for tional and unconditional) helped increase long-term investments and proactive risk enrollment rates of primary and second- management, often rely on short planning ary children by 18 percent compared to a horizons (Lawrance 1991). However, wealth control group (families not receiving a is not the only determinant: policies favoring transfer) and, in Chile, by 8 percent training in disaster preparedness and higher (Akresh, De Walque, and Kazianga education can help both rich and poor 2013; Martorano and Sanfilippo 2012). households (Francisco et al. 2011). Similar positive outcomes are consis- Social protection and safety nets can sup- tently found for health, nutrition, and port long-term adaptation to changing risks food security status of participants (FAO or environmental and economic situations. In 2015). Nicaragua, the Red de Protección Social cash • In Peru, women of childbearing age transfer scheme greatly helped beneficiary enrolled in the Juntos cash transfer pro- households cope in the aftermath of the “cof- gram were 91 percent more likely to have fee crisis” (coffee price decline); it also helped a doctor-assisted delivery compared to coffee laborers intensify alternative agricul- those not participating in the program tural activities even before the crisis (Maluccio (Perova and Vakis 2012). 2005). In the Philippines (see annex 5A for • In Ecuador, a supplementary feeding pro- case study) and many other cases (Arnold gram more than halved child mortality in et al. 2014), community-driven development households exposed to the program for projects can align the response to a shock, at least 8 months (Meller and Litschig building longer-term resilience and empower- 2014). ing the poorest. Further, countries with strong social pro- One potential drawback of strong social tection can provide better support for work- safety nets is that they can lower incentives ers transitioning from declining to growing for people to adapt and change occupation sectors. The United States has done this with or activity as early as possible, when the first trade liberalization—typically through unem- effects of climate change appear (Chambwera ployment insurance for laid-off workers and et al. 2014). If poorly designed, safety nets wage subsidies in sectors that benefit to help can even lock them into locations or activi- them absorb workers from declining sectors. ties that will become more dangerous or less Studies show that these measures can mitigate productive. But then this challenge is not most of the losses at a very small overall cost new or specific to climate change. We discuss (Porto and Lederman 2014; Trebilcock in the next section the role of migration in 2014). poverty reduction, and the efforts made to Social protection can also be used directly make social protection a facilitator of—and to facilitate long-term economic transforma- not an obstacle to—long-term change and tion. In Ethiopia, the Productive Safety Net adaptation (Brown, Zelenska, and Mobarak Program (PSNP) contributes to increased 2013; Bryan, Chowdhury, and Mobarak resilience and climate change adaptation by 2014). In terms of design, the need to sup- investing in the creation of community assets port long-term change favors the portability to reverse the severe degradation of water- of benefits if the recipient decides to move to sheds and by providing a more reliable water capture better opportunities (Gentilini, supply under different climatic conditions. forthcoming). L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   149 Poor people often lack coverage, employment, whereas most poor ­ people work or amounts are too small to make a in the informal economy. Also, poor people in difference remote rural areas can be difficult to reach. And conditional and unconditional cash Unfortunately, poorer households often have transfer programs that have revolutionized limited access to social protection and safety social protection over the last decade are nets. One reason is limited coverage. Social much easier to deploy in rural than in urban assistance consistently reaches more poor areas, given the challenge of targeting the than nonpoor people (figure 5.2, left panel)— poor in cities, where they often live next door conditional and unconditional cash transfers to the wealthier (Gentilini, forthcoming). As a specifically target poor households and are result, even social assistance shows a large increasingly associated with good coverage range of coverage for poor people: in many among households in the bottom quintile countries, coverage does not exceed 50 per- (World Bank 2015b). But the two other types cent, meaning that half of poor people within of transfers (social insurance and labor mar- a country do not receive any social assistance, ket policies) reach poor and nonpoor house- and even below 10 percent in many low- holds in about the same proportion. This income countries (­figure 5.3). does not necessarily mean that those schemes Even when poor households are covered by are poorly designed; some programs, such as social protection schemes, amounts received contributory pensions, are designed for those are often too small to enable better coping who can afford to contribute. strategies. According to the World Bank’s H o w e v e r, p o o r p e o p l e a r e o f t e n ASPIRE database (World Bank 2015a), excluded from programs they should bene- within countries the average per capita trans- fit from. Some programs are tied to formal fer received by households in the bottom FIGURE 5.2  Coverage of poor people is often under 50 percent, and they often receive lower transfer amounts (Coverage and average transfer received, by household income category) a. Coverage (% of population) b. Total public transfers received ($/year) 2,000 80 1,500 60 Nonpoor Nonpoor 1,000 40 500 20 0 0 0 20 40 60 80 0 500 1,000 1,500 2,000 Poor Poor Labor market policies Social assistance Social insurance Source: ASPIRE database, World Bank 2015a. Note: Each dot represents a country and shows the coverage of poor people (horizontal axis) and nonpoor people (vertical axis) in panel a, and the average transfer toward poor and nonpoor people in panel b. 1 5 0    SHOCK WAVES FIGURE 5.3  Poor people in the poorest countries barely covered by social safety nets (Coverage of the poorest quintile, by social safety net program and country income group) 60 50 Percent of poor people covered 40 30 20 10 0 g ks In g T Sc e w CT fe s d cw N he g g SN In ks e w nd ic s e w ks s SN T ol ind e w CT he s SN fe d T ic T he T ic T s ol er din bl er er Ot iver rk UC in Ot -kin UC bl CC Ot UC CC bl CC Ot din in bli SS ol in or r Fe wor Fe C Fe U Fe -ki ho aiv rS rS rS wo wo Pu aiv aiv ed ed ho In-k ho In-k Pu her fee e a fe ol Pu Pu ho Sc Sc Sc Low-income (19) Lower-middle-income Upper-middle-income High-income (8) (41) (36) Income level/type of transfer Source: World Bank 2015b. Note: Coverage rates refer to the percent of poor receiving any social safety net transfer. The number of countries with available survey data for at least one program category in each income group is indicated in parentheses. CCT = conditional cash transfer; SSN = social safety net; UCT = unconditional cash transfer. quintile from social protection is lower than the flood borrowed about six to eight times the transfer received by the four other quin- more compared to the level of government tiles (right panel of figure 5.2). In Malawi, the transfers. poorest quintile receives on average 0.5 cents per day, while the richest 20 percent receives Social protection schemes can be made more than 17 cents. In Vietnam, transfers are more responsive respectively 9 cents and $1.6; in Colombia, the poorest receive 23 cents per day and the To help the population cope with shocks, richest more than $4.6. disasters, and environmental and economic After a disaster, amounts can also be insuf- change, social protection programs must be ficient when examining ad hoc schemes to designed for scalability and flexibility, espe- support affected people. In Bangladesh fol- cially for coverage. Moreover, they need to lowing the 1998 Great Flood, 66 percent of do so while encouraging adaptation and households in the bottom quintile received asset accumulation, without locking benefi- transfers, compared with 33 percent in the ciaries into unsustainable activities and loca- top quintile; and 53 percent of the flood- tions. Thus, the choice of the right exposed households received transfers, com- instrument is context specific. Cash transfers pared with 34 percent of non-flood-exposed cannot ensure short-term food security if households (del Ninno et al. 2001). While tar- food supply is limited, making the case for geting was relatively good, however, transfer dedicated measures for food provision in amounts were small: they represented only some emergencies (box 5.2). Similarly, it is 4 percent of total household monthly expen- futile to attempt to implement theoretically diture for poor households, and 2 percent for optimal policies if institutional capacity is all households. Household borrowing high- weak in practice—using simpler policy tools lights this limit: poor households affected by may be more realistic, even if they are less L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   151 BOX 5.2  Food provision and school feeding schemes are commonplace and effective The distribution of food is a common measure in most common social protection and safety net sys- situations of humanitarian emergency, following tem in the world (figure B5.2.1), even though they disasters, severe economic crises, or conflicts, even reach less than 15 percent of the poor, on average, if it can distort local markets and reduce local pro- in each country (World Bank 2015b). School feed- duction. During the food, fuel, and financial crisis of ing programs are efficient in times of crisis because 2007–2008, Benin, Burkina Faso, Mali, and Niger they rely on existing infrastructure and human introduced emergency food distribution and used resources—schools themselves, as well as the teach- cereal banks to sell food at reduced prices (World ers and parents who are part of school systems Bank 2015b). (Bundy et al. 2009). Moreover, they have the advan- Food distribution is also widely used beyond cri- tage of discouraging parents from taking children sis situations. School feeding programs remain the out of school in times of crises (FAO 2013). FIGURE B5.2.1  School feeding programs are the most prevalent type of social safety net 140 120 Number of countries 100 80 60 40 20 0 rs s rs rs s g rk er in fe fe fe wo aiv ed ns ns ns fe ew tra tra tra ic bl ol sh d sh Fe Pu ho in ca ca -k Sc al al in on on al on iti iti nd nd iti nd co Co Un co Un Type of transfer Source: World Bank 2015b. efficient or less well targeted (World Bank approaches stand out: (i) increasing the 2013a). amount transferred by an existing program to While designing effective social protection its beneficiaries or relaxing rules and condi- can be a challenge, recent experience from tionality such that the transfers increase; social protection systems globally offers (ii) extending the coverage of an existing pro- encouraging and valuable lessons. It suggests gram to include new beneficiaries; and that countries at all income levels can set up (iii) introducing extraordinary payments or systems that increase resilience to natural haz- creating an entirely new program (Bastagli ards. But, to do so, the systems need to be rap- 2014). idly scalable in case of crisis and feature Increasing the amount or value of transfer. targeting mechanisms flexible enough to This works best when beneficiaries of existing adjust quickly to new situations. Three key social protection programs are those who are 1 5 2    SHOCK WAVES affected the most by the crisis, the shock 4Ps cash transfers was relaxed in response to affects primarily the poorest, and there is Typhoon Yolanda, allowing the government already at least one large-scale social protec- to quickly release a total of P550.5 million tion program in place with efficient delivery (US$12.5 million) between November 2013 systems for disaster response. An example of and February 2014 in temporarily uncondi- such a program with built-in mechanisms for tional transfers (see annex 5A). rapid scale-up in response to a shock is Expanding the coverage. In case of severe Mexico’s Temporary Employment Public shocks and those with heterogeneous impacts Works Program (PET). Similarly, after (such as a flood), even relatively well-off Typhoon Yolanda hit the Philippines in 2013, households may lose enough to be pushed external actors such as the World Food into poverty—possibly becoming poorer than Program and the United Nations Children’s existing beneficiaries. To provide adequate Fund (UNICEF) used the preexisting Pantawid support to such at-risk households, the Pamilyang Pilipino Program (4Ps) conditional p rogram must be expanded to include ­ cash transfer program to deliver their support people affected by the shock. In 2008, the the ­ to affected 4Ps beneficiaries—in effect, increas- Mexican government expanded the coverage ing the value of the transfer (see annex 5A). of the national Oportunidades cash transfer For some shocks, such as changes in food scheme by 1 million recipients to mitigate the prices, indexing of social transfers provides a food and fuel crisis. The total number of method for automatically adjusting the Mexicans assisted by the program reached amount of transfers to a changing situation, 5 million households (one out of four fami- without a discretionary decision (box 5.3). lies) (Demeke, Pangrazio, and Maetz 2009). It is also possible to increase transfers by In Ethiopia, the Productive Safety Net relaxing program rules and conditionality. Program incorporates innovative features to Disasters may make existing program rules scale up automatically and enroll additional unpractical or inappropriate: if a disaster beneficiaries when there is poor rainfall (see destroys schools in a region, attendance is no annex 5A). longer an applicable condition for disbursing Creating a new program. In the absence of conditional cash transfers. In Colombia, the an appropriate program that can be used or cash transfer scheme Familias en Acción sus- extended to respond to the crisis, it is possible pended conditionality temporarily in 2008 to to introduce new programs or initiatives— accommodate the shortfalls in service provi- sometimes, a disaster or a crisis even creates sion as a result of damaged infrastructure. In the opportunity to strengthen or reform the the Philippines, all conditionality linked to the social protection systems. In certain cases, BOX 5.3  Indexing as an automatic scale-up mechanism Price changes pose a major threat to the smooth and services. In Malawi, two schemes—the Food working of social protection schemes. Take, for and Cash Transfers and Dowa Emergency Cash instance, what happened in Kenya when food ­Transfers —adjust the transfers before each monthly prices rose in 2007–2008: the Hunger Safety Net disbursement based on observed prices (Sabates- Programme’s cash transfer scheme lost more than Wheeler and Devereux 2010). Experience has shown half of its value over 18 months (Devereux 2015). that indexing works best when there is a contin- To avoid such risks, many countries index the gency fund to absorb changes in the program’s cost. benefits in their social protection systems by using inflation data or the price of a basket of goods Source: Based on Bastagli (2014). L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   153 countries have used extraordinary payments. balance the urgency of postdisaster support In Chile, the government paid a one-time and the need to carefully target the larger bonus (Ch$40,000 or about US$66) in March transfers supporting reconstruction (see 2009 to 1.7 million poor families to cope with annex 5A). the effects of the ongoing financial crisis.2 A similar measure was introduced in March Postcrisis responses need to balance 2010 following a major earthquake. In other timeliness with targeting accuracy cases, new durable programs have been intro- duced. The 1990 Honduran Programa de In the aftermath of a crisis or a disaster, it Asignación Familiar and the 2001 Colombian can be tough to identify those affected and at cash transfer scheme Familias en Acción were risk of being pushed into poverty. There are launched during recessions and macroeco- several approaches to targeting beneficiaries, nomic adjustment periods—and transformed all of which face challenges (table 5.2). into permanent programs, part of the national Economic shocks or disaster consequences safety net system. In Guatemala, the food and are often heterogeneous, making geographic fuel crisis in 2008 prompted the introduction or demographic targeting approaches diffi- of a new program, Mi Familia Progresa. cult (Alderman and Haque 2006; Grosh But the challenge is larger when respond- et al. 2008). Registries with socioeconomic ing to a disaster or a crisis with immediate information and precise location are seldom and urgent needs. Creating and rolling out a available—and, as in Nepal, there may not new program takes time—this is why coun- even be a reliable street address system. tries with existing scalable programs are more Usual targeting methods (like proxy-mean resilient and better placed to respond to crises testing) are based on slowly changing house- and disasters. hold characteristics (like assets) and are slow To extend support to new beneficiaries— and expensive to implement—meaning that whether through an existing or a new they cannot capture sudden changes in program—it is necessary to be able to iden- ­ income and consumption. And affected pop- tify them rapidly. A challenge is to strike a ulations are often displaced in camps or with balance between providing rapid support family or friends, and thus hard to reach. when needed and targeting precisely the Because these approaches will always have most in need. Case studies suggest that the inclusion and exclusion errors, grievance cost of a drought to households can increase appeal mechanisms are critical. In Pakistan, from zero to about $50 per household if the grievance redress system in the second support is delayed by four months, and to phase of the Citizen’s Damage Compensation about $1,300 if support is delayed by six to Program cut exclusion errors from an initial nine months (Clarke and Hill 2013). This 61 percent to 32 percent (see annex 5A). rapid increase is due to irreversible impacts Options to manage this challenge include on children and distress sales of assets (espe- developing—before a crisis occurs—large and cially livestock). flexible social registries that include both Thus, most postdisaster responses have potential and existing beneficiaries, the use of multiple stages, with initial (survival-related) self-targeting methods, and the use of support delivered quickly even at the subsidies. expense of targeting and accuracy, and Social registries. These are crucial because reconstruction support provided later with they facilitate quickly identifying households more effort to target support appropriately that are vulnerable to being pushed into pov- (de Nicola 2015). In Pakistan after the 2010 erty by a disaster. Social registries should floods, the government implemented the include demographic, socioeconomic, and Citizen’s Damage Compensation Program location information on households that (CDCP), a rapid response cash grant pro- can potentially be supported by a social gram that included two phases to better program. 1 5 4    SHOCK WAVES TABLE 5.2  Methods for targeting beneficiaries with social safety nets are more or less appropriate during a crisis or after a disaster Principle Advantage Limitation Means testing Benefits are allocated High level of accuracy and Costly (thus only appropriate for large and damage conditional on income therefore appropriate in benefits) and requires high administrative assessments and the presence or heterogeneous situations capacity magnitude of losses Can be adjusted after a shock, It takes a long time to collect the data, so taking into account the it cannot be applied for rapid response specificity of the shock and the varying impacts at the household level Proxy-mean Income and losses Verifiable and objective, cheap Focuses on slow-changing household testing and from the shock are because based on available characteristics (for example, assets) that damage proxied by quantifiable data may ignore income shocks assessments and easily measurable Able to capture heterogeneous Ignores large shocks that may change the characteristics, such as situations, if proxies are well statistical relationship between poverty ownership of a house selected and the selected proxies made of bricks, or visible Captures asset losses from It takes a long time to collect the data, so damages disasters data cannot be applied for rapid response Community Communities eligible Can be quickly adjusted in Works only in sufficiently cohesive targeting for support are selected, response to a shock or disaster, communities, and may exacerbate social and distribution based on low-resolution exclusion and affect authority of local of benefits is then estimates of the impact actors delegated to the head (leaving local decision makers Requires an estimate of community- of a formal or informal to identify small-scale needs) level needs, often based on proxy-mean community Makes best use of local testing at the community level (leading to knowledge on needs and the same issues as above) priorities Can include nonmonetary dimensions of poverty and nonmonetary impacts of the shock Demographic Benefits are given based Simple and cheap to administer, Requires good demographic data targeting on characteristics such and usually popular Inaccurate when impacts are imperfectly as age or gender Appropriate for supporting correlated with demographics highly vulnerable groups such as children Geographical The program covers only Simple and cheap to administer Appropriate only for large-scale shocks targeting inhabitants of specific Can be quickly adjusted in with relatively homogeneous impacts regions response to a shock or disaster, Unable to account for household-level based on low-resolution vulnerability and heterogeneous impacts estimates of the impact Performs poorly where poverty is not concentrated (for example, in urban areas) Can be politically controversial and limit migration and its benefits Self-targeting Mainly cash and food- Simple and cheap to administer, Cannot be used to deliver large benefits for-work programs as no registry of beneficiaries is and may stigmatize the affected needed population Can be quickly adjusted in Inaccurate or inadequate if the demand response to a shock or disaster for work exceeds supply, as the poorest Can be used to reduce risks (for and most in need are typically excluded example, public work programs Requires the availability of good projects, that improve drainage) or to appropriate for the beneficiaries’ skill set reconstruct after a disaster Source: Based on Gentilini, forthcoming. L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   155 In Brazil, the Cadastro Unico registry Self-targeting. This approach, which does includes households with a per capita not necessitate much institutional capacity, income below half the national minimum can be done via work programs—which pro- wage, a threshold that is higher than the vide jobs and income by putting in place income eligibility threshold of existing social public projects (like road construction, main- programs. As a result, the registry includes tenance, irrigation infrastructure, reforesta- households that are not currently beneficia- tion, soil conservation) or, especially in ries of social protection but are considered postdisaster situations, reconstruction tasks. to be vulnerable to economic shocks or It usually works by offering a below-market disasters. Moreover, individuals can register wage; people join only if alternative income at any time based on self-reported income, sources are lacking (Cazes, Verick, and thereby reducing transaction costs (Bastagli Heuer 2009). In Côte d’Ivoire, the Highly 2009). Such a design ensures that the Bolsa Labour Intensive Works Program was cre- Familia cash transfer scheme can be rapidly ated to support and rehabilitate 35,000 for- adjusted when shocks occur, thus acting as mer combatants via road building and an insurance facility for vulnerable reconstruction work. The key drawback is households. that works programs fail to reach those who Large social registries make it possible to face constraints that prevent them from introduce dynamic targeting, in which poten- working (like those facing disabilities, sick- tial beneficiaries are segmented—before a ness, and exclusion) and who are often the disaster or a crisis—into multiple categories, poorest (McCord 2013). based on their income, assets, location, or The use of works programs as a social occupation (like farmers and fishermen). ­ protection measure in postdisaster situations Then, the different categories receive a vary- requires that cost-effective and socially benefi- ing level of support depending on the situa- cial projects be readily identified before a cri- tion. For instance, potential beneficiaries can sis strikes. In practice, however, extreme be ranked starting from the poorest, and the natural events, such as storms or floods, are number of people provided with support typically associated with obvious and signifi- (“how far you go down the list”) can depend cant labor needs. Reconstruction of public on the situation, for instance to provide more infrastructure and buildings and the clearing people with support during a drought. The of rubble are examples of needs that can be level of support in each category can even be met by works programs, which can benefit based on an objective rule or a weather index affected poor and vulnerable people (even (like using cumulative rainfall or a trigger those with low skills), as well as the wider based on wind speed). community. When social registries are not available, Subsidies. These are widely used to help an alternative is combining geographical tar- poor people, especially in the absence of other geting (to concentrate resources in the most social protection programs, and not least affected municipalities or communities) with because they can be simple and quick to community targeting (to use local knowl- implement. The Egyptian food subsidy pro- edge to concentrate resources on the most gram was expanded in 2008 to include 15 affected households). Pakistan used this million additional beneficiaries (Jones et al. approach in the first phase of the CDCP 2009)—thereby avoiding an increase in the after the 2010 floods, when timeliness was a poverty rate from 22 percent to 31 percent priority and there was no reliable data on due to food price increases. Indonesia used a the distribution of losses. The second system of generalized subsidies as a safety net phase—less urgent but with larger trans- during the 1997 financial crisis. fers—put a stronger emphasis on targeting, But the drawbacks of subsidies are many. using housing damages as a proxy for liveli- They can lead to waste and corruption. For hood losses (see annex 5A). instance, analyses of India’s Public Food 1 5 6    SHOCK WAVES Distribution Program , which provides Building solid social protection systems subsidized food and fuel, found a number of ­ requires resources but is affordable operational challenges—including underpro- even for the poorest countries vided entitlements as a result of “leakages” Governments in poor countries face many of food through the supply chain, commodi- competing needs and have limited resources, ties being diverted, food getting under- so the development of social protection and weighted, beneficiaries being overcharged, safety net programs needs to be justified shops being closed, and food falsely being carefully. Overall, however, costs are moder- declared out of stock (Drèze and Khera ate, even in low-income countries—and 2015; Government of India 2011; World instruments are available to help govern- Bank 2011). ments face the liability created by social pro- In addition, subsidies are often difficult tection programs. to remove when the crisis is over, and they Experience suggests the cost of social pro- are an expensive and inefficient tool for tection can be managed. For instance, a recent supporting poor people because in many study assesses how much social protection cases a large fraction of the funds go to would be needed to support vulnerable peo- those who do not need them the most. ple in the Horn of Africa and the Sahel in Fossil fuel subsidies, for instance, are typi- 2030 (accounting for population growth and cally implemented and publicly justified socioeconomic and climatic change) (del with the rationale of helping poor people Ninno and Coll-Black 2015). It finds that— gain access to energy and energy services. assuming that vulnerable people can be pro- But while low energy prices indeed reduce tected against the worst effects of droughts poverty by reducing the cost of energy ser- with an annual social protection package of vices, they do so in an extremely inefficient $300 per capita (the typical size of such sup- way, since energy is overwhelmingly con- port systems in the region)—1 percent of the sumed by the wealthier. region’s GDP would be sufficient to cover this Thus, it is strongly in the poor’s interest to population, although more is needed in some reallocate resources used to subsidize basic countries (figure 5.4). goods in order to implement better-targeted This would be a moderate cost compared and more efficient support measures instead. with the cost of some short-term coping strat- Ghana’s 2005 fossil fuel subsidy reform egies, such as reduced food intake or suspen- increased the price of transport fuels by sion of schooling, which can have irreversible, 50 percent, but also included in-kind benefits life-long effects, especially on children. While for the poor—an e ­ xpansion of primary health such a social protection package can by no care and electrification in poor and rural means prevent all negative impacts of ­ istribution of efficient light areas, large-scale d droughts, it reduces the need for expensive bulbs, public transport improvements, and humanitarian relief. In fact, the total cost of immediate elimination of school fees at gov- providing this protection to disaster victims in ernment-run primary and secondary schools Africa during the period from 2010 to 2013 is (IMF 2013; Vagliasindi 2012). Indonesia has lower than what was spent on humanitarian introduced programs to mitigate the effects of relief measures (del Ninno and Coll-Black higher energy prices through subsidized rice, 2015). free health care, cash assistance to poor stu- The government’s ability to provide social dents, and a one-year conditional cash trans- protection to poor households will be greater fer scheme targeting poor households with if the middle class has access to instruments pregnant women or school-age children to manage risks, such as private insurance (Perdana 2014). Iran implemented a quasi- (box 5.4). Otherwise, the middle class is often universal cash transfer (about $45 per month better able to demand and obtain support per capita) when it reformed its energy subsi- from governments, at the expense of the dies (IMF 2013). L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   157 FIGURE 5.4  Providing safety nets in the Horn of Africa and Sahel is affordable, but the cost is very volatile (Cost of providing safety net coverage) 6 5 4 Percent of GDP 3 2 1 0 a ia a da ia a in in o te ad m ire e a a au ali ia r ria l go l ga ta ge ric ny ric an ne Th as op an an n M iss an Cô Ch ge to Ga ’lvo To ne Be Ni aF Af Af Ke Gh Gu Gui a, nz rit hi -B Ug Ni d Se bi d rn rn Et au an Ta ea rk ste te M Gr in Bu es Ea W Severe Moderate Mild Regular Source: del Ninno and Coll-Black 2015. Note: Cost of providing safety net coverage to the fraction of Sahel’s population affected by drought in a regular year (25 percent of vulnerable population is affected), and in years characterized by mild, moderate, and severe drought (with 35, 50, and 65 percent of the vulnerable population affected, respec- tively). For the entire region, the cost would increase from 0.3 percent to 0.5 percent, 0.76 percent, and 1 percent of GDP, respectively. BOX 5.4  Private insurance and social protection schemes are complements, not substitutes The different instruments for managing risk and cop- mechanisms: one cannot insure against regularly ing with crises—from private insurance to uncondi- recurring events. Similarly, self-insurance and risk tional cash transfers—are not substitutes: they are retention will be preferred for some risks, while con- part of the toolbox available to governments, individ- tingent finance, private insurance, and risk sharing uals, and firms, and can be used together, depending will be preferred for rarer events. As a result, the opti- on the context and the considered risks. For instance, mal risk management strategy for a government or some tools can be more appropriate and efficient an individual typically consists of a series of tools, than others, depending on the probability with which combined within a consistent and holistic strategy. a natural hazard turns into disaster. It is widely Moreover, the adequacy of different tools depends accepted that risk-reduction investments—such as not only on the risk itself but also the varying char- physical protection against floods—are more effi- acteristics of individuals and firms. While the middle cient at dealing with frequent events than risk-sharing class and formal firms can have access to private box continues next page 1 5 8    SHOCK WAVES BOX 5.4 (continued) insurance markets, providing insurance products direct access: if higher-income households are cov- adapted to poor people’s needs can prove extremely ered by private insurance schemes, public funds for challenging—for them, social protection schemes postdisaster relief can be better focused on supporting may be a more effective alternative to private insur- poorer households. In Turkey, the fact that all dwell- ance. Thus, protecting a population against shocks ers in urban areas are covered by an insurance prod- may require both developing a stable and competitive uct (the TCIP) makes it easier for the government to insurance market and providing social safety nets. focus public resources on rural areas, which are poorer Well-established private insurance markets can and where market insurance would be challenging to be beneficial for poor populations, even if they lack introduce. poorer if resources are scarce. In two case Managing such increases in social expen- studies on Thailand, it was found that the ditures can be a challenge for governments majority of post-flooding government support who often face reduced tax revenues follow- was benefiting the well off, with 500 Bahts ing a disaster (Noy and Nualsri 2011; per capita (about US$14) going to the richest Ouattara and Strobl 2013). To cover these quartile, compared to 200 Bahts for the poor- liabilities created by natural hazards and est quartile (Noy and Patel 2014). other environmental risks, different instru- ments have been developed and imple- mented (Mahul and Ghesquiere 2007; Scalable social protection creates a Ghesquiere and Mahul 2010; Hochrainer- formal liability for the government, but Stigler et al. 2014; Cardenas et al. 2007). contingent finance and insurance are The optimal choice of instruments is coun- available to manage it try specific and depends on both costs and Just how costly is social protection? Certainly, timeliness (Clarke and Poulter 2014). These the cost of providing coverage to vulnerable instruments include: people affected by natural hazards changes Reserve funds. The Risk Financing from year to year. In the case of the Horn of Mechanism in Ethiopia is a fund dedicated to Africa and Sahel regions, protection costs can scaling up social protection, which allows the increase by a factor of four in a year of severe PSNP to disperse additional transfers to exist- drought, compared to an average year (figure ing recipients or temporarily expand its cov- 5.4). For other (rarer) disasters or shocks, the erage to reach beneficiaries not enrolled in the impact on social protection costs can be even regular PSNP program, but who are affected larger. In Bangladesh in 1998, households by a shock (see annex 5A). In the Philippines, had to borrow at high costs to cope with the National Disaster Risk Reduction and floods, with long-term impacts on welfare Management Fund finances a range of and poverty (del Ninno et al. 2001). Avoiding ­ disaster-related expenditures but is not able such emergency borrowing would have to disburse rapidly in the case of crisis; this is required a transfer of approximately Tk why the government created the Quick 5,000 (approximately US$90 at the time) for Response Fund, which focuses on emergency each of the indebted households. For the gov- response (see annex 5A). Mexico’s Natural ernment, total costs would have amounted to Disasters Fund (FONDEN) was created as a more than $1.5 billion (3.5 percent of GDP), budgetary tool to rapidly allocate federal a significant impact on public finances and funds for rehabilitation of public infrastruc- social program budgets. ture affected by disasters. L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   159 However, reserve funds have limited offer quick-disbursing index-based coverage c ­ apacities and cannot be designed to cope against tropical cyclones, earthquakes, or with the more rare and extreme events. In the droughts. In response to Cyclone Pam in Philippines, Typhoon Yolanda raised ques- March 2015, the PCRAFI provided Vanuatu tions as to the adequacy of the Quick with a rapid $1.9 million payment to support Response Fund volume and the process to immediate postdisaster needs. This payout is replenish it if it gets emptied by a major event limited compared with total losses and recon- (or a series of smaller disasters). Thus, addi- struction needs—estimated at $184 million— tional instruments have been developed to but was still 8 times the annual emergency protect public finances. relief provision held by the government, and 7 Insurance and catastrophe bonds. The times higher than the annual insurance pre- contingent fund FONDEN in Mexico now mium paid by the government of Vanuatu. leverages private sector financing as part of Contingent credit: The Cat-DDO. In 2007, a strategy combining risk retention and risk the World Bank introduced Catastrophe transfers. In 2006, FONDEN issued a $160 Deferred Drawdown Options (Cat-DDOs), million catastrophe bond to transfer a new financing instrument allowing ­ countries Mexico’s earthquake risk to the interna- to access budget support in the immediate tional capital markets—the first parametric aftermath of a disaster. A contingent loan can catastrophe bond issued by a national gov- be rapidly disbursed if a state of emergency is ernment. However, studies suggest that declared, and thus help governments finance using reinsurance or international capital the upscaling of social protection (see markets for financial protection can be annex 5A). Cat-DDOs can also be used to more expensive than building additional back up an existing insurance pool, as is the reserves (that is, the opportunity cost of case in Turkey (box 5.1). public funds) (Cardenas et al. 2007). Further, Cat-DDOs do not only aim to Nevertheless, insurance products offer ben- provide immediate liquidity. They also efits in the form of fiscal discipline and function as a mechanism to incentivize pro- timeliness of budget allocation. In emer- active actions toward risk reduction. To be gency situations, these financial schemes are eligible for a Cat-DDO, governments are able to disburse funds rapidly—more rap- required to develop ex ante capacity to idly than would be possible with public manage natural risks. As such, it is the first budgets. And by predefining payout rules instrument linking immediate disaster for allocating postdisaster support, formal response funding with proactive engage- insurance and financial products can reduce ment in risk reduction. Other institutions, political economy biases (Clarke and such as the Inter-American Development Poulter 2014). Bank and the Japan International Regional risk-sharing facilities. Regional Cooperation Agency have since introduced mechanisms are also popular solutions. The similar instruments. Caribbean Catastrophic Risk Insurance Cat-DDOs have proven to be an effective Facility (CCRIF) currently pools disaster risk instrument for implementing disaster risk across 16 countries. It was the world’s first management strategies. However, experi- regional catastrophe insurance facility, using ence shows that, facing a finite financing parametric insurance to provide participating envelope, governments tend to favor cash in governments quick, short-term liquidity for hand at the expense of contingent instru- financing responses and early recovery from ments. As a result—and despite strong inter- major earthquakes or hurricanes. The Pacific est from client countries—the uptake of Catastrophe Risk Assessment and Financing Cat-DDOs has been limited. One option to Initiative (PCRAFI) and African Risk improve access to contingent finance and Capacity are other, more recent examples of build the resilience of developing countries donor-supported regional mechanisms that would be to remove this trade-off between 1 6 0    SHOCK WAVES cash in hand and contingent finance by Migration and remittances play ­ separating the budget allocated to contin- gent instruments from the budget allocated an increasingly important role to traditional lending. and need to be supported by International aid. When a country’s policies capacity to cope with a disaster is exceeded, ­ international aid and humanitarian ­ emergency Migration plays a key role in the ability of measures can be critical. Foreign aid includes poor households to escape poverty by open- essential in-kind support (including emer- ing opportunities for better jobs, higher pay, gency equipment such as emergency water and better access to services and education. treatment stations, reconstruction material, Migrants typically benefit from relocating, as equipment and machinery, and emergency do their families and areas of origin, through relief goods like food, blankets, and clothes), remittances, enhanced social networks, and as well as financial aid for social protection access to information (Adger et al. 2002; and reconstruction costs. Bryan, Chowdhury, and Mobarak 2014; However, in the past, increases in foreign Moser and Felton 2007). aid have been low, averaging only a small per- cent of total economic losses stemming from Migrations help households adapt to the disaster (Becerra, Cavallo, and Noy and cope with shocks 2013). Generally, studies have found that increases in financial aid are larger for more Migration can be an important way of adapt- severe disasters and for particularly poor ing to extreme weather events and climate countries with limited disaster management change impacts, and thus of reducing impacts capacities. This suggests that these resources that lower welfare (Adger et al. 2014; Black are relatively well targeted and not politically et al. 2011b; Jülich 2011). Particularly in biased (Becerra, Cavallo, and Noy 2013). areas where in situ adaptation is difficult or Nevertheless, increases in foreign aid in extremely costly (such as in low-density response to disasters remain sensitive to coastal areas or remote areas with low pro- media coverage, are hardly predictable, and ductivity), migration can be critical. By can be slow to arrive—all of which make it migrating, individuals and households can ever more difficult to prepare contingency reduce their exposure to natural hazards and plans based on available resources. Foreign increase the set of available opportunities, aid should thus be regarded as a resource of thus improving well-being and livelihood last resort. prospects. However, the poorest house- To improve the timeliness, transparency, holds have a lower capacity to migrate and and predictability of postdisaster or crisis may therefore be unable to use this option international aid, and provide additional (Black et al. 2011a). This can also be the financing, a special Crisis Response Window case for households in conflict and fragile (CRW) was created as part of the International areas, or those being socially excluded or Development Association, the World Bank marginalized. Group’s fund for the poorest countries, in Climate change can affect migration deci- 2011. Its primary objective is to (i) provide sions, but migration is usually driven by a poor countries with extra resources in a timely variety of pull and push drivers, both environ- manner; (ii) help them respond to severe eco- mental and socioeconomic (Adger et al. 2014; nomic crises, price shocks, and major natural Black et al. 2011a). In the past, direct environ- disasters; and (iii) return to their long-term mental factors have generally played a minor development paths. In Malawi, the CRW pro- role (Black et al. 2011b), except in extreme vided $40 million of postdisaster support circumstances such as large disasters. The after the large floods that affected the country most important factor remains the socioeco- in January 2015. nomic context in the origin and destination L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   161 areas (Wodon et al. 2014). In the future, as the is adopted, the provision of information is effects of climate change intensify, environ- critical—both on the procedure of how to mentally induced migration is expected to remain in the program and on the location of increase although no robust global estimates registration offices. are available (Adger et al. 2014). A study Modern technologies can help simplify investigating this question in five countries access to benefits, at least when the program in the Middle East and North Africa region involves cash transfers rather than vouchers concludes that a significant deterioration of and in-kind transfers. The Bolsa Familia pro- climatic conditions would lead to an increase gram in Brazil provides beneficiary cards that of about one-tenth to one-fifth of current can be redeemed at outlets across many urban migration levels (Wodon et al. 2014). centers. A similar system exists in Ecuador, Climate change and climate policies can but with mobile vendors that visit beneficia- also impede migration—whether through ries. While effective and secure, such systems constraints on urban development and higher can be costly to implement at a large scale. housing costs linked to natural risks (higher Mobile money programs can be an efficient construction costs due to stricter building alternative, as they are typically low-cost and norms or restrictive flood zoning) or through have a wide reach (Aker et al. 2014; Vincent increased conflict and exclusion (crime and and Cull 2011). violence or civil unrest). In that case, climate change would diminish the opportunities that Domestic and global remittances are individuals and households can capture. In key to increase recipients’ resilience addition, the ability to migrate depends on household assets (including land tenure), the Remittances—that is, the private transfer of ability to sell assets, information and social money by a foreign worker to individuals in capital, financial resources, and human his or her home country—are estimated at capital. $584 billion in 2014. They are a vital Given the importance of mobility as an resource for developing countries and sig- instrument for poverty reduction, climate nificantly exceed official development assis- change adds to the rationale for portable tance and foreign direct investment social protection benefits (Holzmann, Koettl, everywhere except China (Ratha et al. and Chernetsky 2005; Kuriakose et al. 2013): 2015). Fragile and conflict-affected coun- safety nets that are linked to specific locations tries, in particular, have large diaspora sav- could tie poor people to places that may no ings as a share of GDP: some 81 percent in longer support livelihoods. Safety net pro- Somalia, and 53 percent in Haiti (Ratha grams must thus consider the portability of et al. 2015). In addition, domestic remit- benefits—that is, help households or individu- tances can play an important role, especially als to remain engaged in programs and to in rural areas. In India, the domestic remit- maintain their benefits even as they move tance market was estimated to be $10 billion (Gentilini, forthcoming). in 2007–08, with 80 percent of that being For portability of participation, in the directed toward rural households for whom absence of a central registry, programs require this represents a large fraction of total con- systematic tracking of beneficiaries. Some sumption (Tumbe 2011). programs place the responsibility on International remittance flows are a stable beneficiaries to inform programs of their ­ source of finance that are generally not corre- migration. In the Philippines, beneficiaries lated with capital flows and that can help must declare a change in residency and notify hedge against shocks (Bugamelli and Paterno the program six months in advance of a move. 2009; Chami, Hakura, and Montiel 2009; Alternatively, many countries have program World Bank 2006, 2015c). After natural, offices in major urban centers where migrants ­ economic, financial, and political shocks, can register upon arrival. Whichever strategy these flows have been found to either remain 1 6 2    SHOCK WAVES stable or even increase (Clarke and Wallsten FIGURE 5.5  Within a country, remittances tend 2004; Fagen 2006; World Bank 2006). to be higher for the wealthier (Remittance transfers for poor and nonpoor people) Unsurprisingly, countries with a larger stock of emigrants as a share of the home population tend to experience a greater surge in remit- Private tranfers received ($US 2005 PPP/year) 300 tances following natural disasters (Mohapatra, Joseph, and Ratha 2009). 250 Remittances can help smooth consumption and finance recovery and reconstruction. After the 1998 flood in Bangladesh, consump- 200 tion was higher in remittance-receiving house- Nonpoor holds (Mohapatra, Joseph, and Ratha 2009). 150 In the Philippines, it was estimated that remit- tances compensated for nearly 65 percent of 100 lost income after rainfall shocks (Yang and Choi 2007). Despite disruptions in transfer 50 channels and financial services, remittances remained relatively stable after disasters hit 0 Pakistan and Indonesia, and they were an 0 50 100 150 200 250 300 important factor in recovery and reconstruc- Poor tion (Suleri and Savage 2006; Wu 2006). In Indonesia, households that received remit- Source: ASPIRE database, World Bank 2015a. Note: Each dot represents a country for which adequate data exist. tances in the Aceh region recovered faster from the 2004 tsunami, despite disruptions in financial services and informal transfer chan- nels (Wu 2006). However, international and domestic Sustainable Development has proposed remittances tend to benefit the wealthier reducing remittance costs to 3 percent, which within a country (figure 5.5). They also have would translate into savings of over sometimes been shown to lower government $20 ­billion annually for migrants. Commonly spending through a substitution effect available technologies (like instant money between private insurance provided by remit- transfers through cell phones) could play a tances and public insurance provided through key role in streamlining processes and reduc- government expenditures (Kapur and Singer ing transaction costs. 2006). But policies that encourage and facili- tate the use of remittances for investments can also promote microsaving and microinsur- Voice and governance ance, and lead to cobenefits such as enhanced When it comes to adaptation and coping, financial integration. the affected populations must have access To support the positive impacts of remit- to and some control over the country’s tances, adequate financial and banking infra- economic, social, and institutional structure and frameworks are essential. resources. Because poor communities lack Globally, the burden of transfer costs stood human and social capital (like social net- at 7.7 percent of overall transfers in 2014— works and influence on policies and strate- and they tend to be the highest in Sub- gies that impact well-being), they are Saharan Africa, where they average typically excluded from accessing such 11.5 percent (Ratha et al. 2015), partly reflect- resources. A background paper for this ing limited competition among service pro- report argues that only with inclusive and viders. The UN Open Working Group on participatory decision-­m aking processes L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   163 can policies be designed to protect the poor A narrow focus on poor people’s and vulnerable effectively (Tschakert, v ­ ulnerabilities to climate change can lead to forthcoming). When such ­ processes fail, as the justification of top-down interventions in conflict-affected states, ­g overnments that undermine the role of communities may be unwilling or unable to support (Tschakert, forthcoming). Adaptation strate- those affected, with poor people being the gies that fail to account for the needs and cir- first to suffer. cumstance of marginalized groups (including women or ethnic minorities) can exacerbate risk dynamics (Vincent et al. 2014). In con- Vulnerability to climate change trast, well-designed adaptation projects can interplays with many other promote equity, as in India’s Karnataka vulnerabilities watershed project—which increased income, Efforts to improve the ability of poor peo- employment, and agricultural productivity ple to cope with climate shocks can be among the poorest participants (Olsson et al. more effective if they address the broader 2014). issues related to power relations within When social dimensions are disregarded, societies, instead of narrowly focusing on climate change and climate policies can sys- one particular shock (such as disasters or tematically reinforce and exacerbate inequali- changes in agricultural yields). It may be ties along with other drivers that amplify more effective to boost income-generating vulnerability. Gender impacts can be rein- activities for the poor to enable them to forced leading to differential access to social afford living in safe areas, rather than to and environmental adaptation resources and implement strict land use regulations that exclusion from decision making and planning prevent destitute people from settling in (Vincent et al. 2014). In Ghana, observations flood-prone areas. Assessments of vulnera- show that some husbands seek to maintain bility need to go beyond analyzing physical their power positions by preventing women assets and location, and on to exploring from cultivating their own plots, even though the structural drivers of poverty (like social it could compensate for yield losses due to capital, institutional arrangements, and shifting precipitation patterns (Carr 2008). governance). Overlooking local marginalized groups has Poor people are often confronted with major implications for the efficacy of climate multiple dimensions of inequality (including policies and the ability of these communities gender, age, race, caste, ethnicity, and disabil- to cope and adapt. Natural r­ esource-dependent ity), with implications for their capacities and communities—including indigenous groups— opportunities to cope and participate in adap- often have extensive knowledge and long-­ tive decision making (Tschakert, forthcom- lasting traditions for adapting to changes in ing). These inequalities can marginalize the climate and environment, but are over- specific groups and further aggravate their looked in policy-making and planning pro- vulnerability. In Benin, while progress in the cesses. These cultural practices are at risk of multidimensional poverty index can be disappearing because of the effects of climate observed for most ethnic groups, no reduc- change on the livelihoods and identities of tion in poverty was observed among the Peul, these communities. Other factors exacerbate the poorest ethnic group (Alkire, Roche, and this risk further, sometimes because of the Vaz 2014). These types of inequalities not dependencies created by policies, inadequate only reinforce systematic constraints to mar- entitlements and rights, obstacles for intergen- ginalized groups’ access to opportunities, but erational knowledge transmission, and the they can also be a source of conflict that fur- lack of inclusion in formal decision-making ther amplifies the stress caused by climate processes (Adger et al. 2014; Tschakert, impacts (Stewart 2010). forthcoming). 1 6 4    SHOCK WAVES Poor people need a voice in decision- of different vulnerabilities, rather than just making processes—community-based their consequences. development and strong institutions One study, which analyzes how gover- can help nance structures affect the variety of coping strategies for hurricanes in the Mexican When poor people are excluded from gover- Caribbean, finds that regions with less hege- nance and have no say in the d ­ ecision-making monic political structures have developed process, the policy options discussed in pre- more diversified coping strategies, compared vious sections are unlikely to be implemented to regions with strong top-down management in a timely and adequate manner. This issue that discourages participation (Manuel- is closely linked to the values and criteria Navarrete, Pelling, and Redclift 2011). that determine which segment of the poor Another study shows that community-driven population is considered “deserving” of sup- development projects have in various cases port and the openness, inclusiveness, and succeeded in helping communities deal with fairness of the decision-making processes. disaster and climate risks (Arnold et al. 2014). Decision-making processes matter. A cost-benefit analysis of adaptation invest- ments would favor policies that protect h i g h e r- i n c o m e a s s e t s r a t h e r t h a n In conclusion less-productive assets. Without an explicit ­ This chapter provides evidence that poor focus on the poor and vulnerable, such an people suffer more from economic and envi- efficiency criterion may fail to help poor ronmental shocks than does the rest of the communities and instead concentrate sup- population—and not only because they are port and resources on the wealthier. more exposed or vulnerable to the impacts A recent study assessing impacts of sea level discussed in chapters 2 (agriculture and eco- rise on U.S. coastal communities found that systems), 3 (natural disasters), and 4 (health). 99 percent of the most vulnerable popula- Another crucial difference between poor tions in the gulf region of the United States and nonpoor people is the strength and scope live in areas where protection from inunda- of “support systems” available to them: fam- tion (such as sea wall construction) is not ily, friends, communities, financial institu- cost-effective, compared to only 8 percent tions, and the government. These systems of the most resilient segments of the popu- provide tools for people to manage risks or lation (Martinich et al. 2013). Protecting cope with disasters and shocks (such as social only areas where the benefits from avoided safety nets, market insurance, savings, access property loss exceed the costs of protective to credit, and remittances). They are crucial measures is a sure way of directing protec- for reducing vulnerability and adapting to tive investments toward rich areas. Explicit changing environmental and economic condi- choices to support or compensate poor tions (World Bank 2013a). And they are criti- communities are thus necessary to ensure cal for complementing other climate change that adaptation policies support communi- adaptation measures—and preventing (at ties with the least adaptive capacities. least partially) the associated harmful impacts And yet for poor people the ability to influ- on poverty. ence such decisions is often limited, thus con- The chapter calls for a comprehensive tributing to their vulnerability to climate strategy, combining multiple tools that can change and climate mitigation policies— protect against a variety of events (like small which, in turn, aggravates preexisting poverty and frequent shocks, rare extreme disasters, (Lawson and Elwood 2014). That is why par- and adverse long-term trends in ticipatory decision-making processes can ­ precipitation)—and remain suitable and acces- improve the diversification of coping strate- sible for various income and demographic gies for disasters and help address the causes groups. For instance, developing market L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   165 insurance for the middle class to protect agricultural assets) and a Direct Support against relatively ­frequent events helps gov- component (provides social assistance to ernments concentrate resources on the poorest the poorest). and vulnerable people and on hedging against • Household Asset Building Program exceptional shocks that exceed the capacity of (HABP): This is designed to empower private insurers. While the private provision r u ral households to increase thei r of insurance for the very poor is associated incomes, food production, and assets, by with large challenges, governments can supporting livelihood activities, extension p rovide quasi-insurance via tax-financed ­ services, and access to financial services. social protection schemes, and in turn reinsure against the resulting liability. But, for such Households that receive support from protection strategies to be effective, they must these programs are expected to “graduate” rely on a good understanding of the benefits from chronic food insecurity to food of different tools for the poor and facilitate ­ self-reliance. The FSP is implemented by the participation in the decision-making process. Ministry of Agriculture (MoA) and is largely Now that this report has underscored the funded by international donors and financial vulnerability and exposure of poor people to institutions. the adverse effects of climate change—and the PSNP targets the poorest and fights various channels through which this can chronic food insecurity.The PSNP uses a mix occur—we can ask what the overall impact of geographic and community-based target- on poverty will be by 2030. The next chapter ing to identify chronically food-insecure tackles this question, along with weighing the households in chronically food-insecure implications of these results for climate woredas (or districts) in rural areas. The change mitigation and its role in contributing greater use of c­ ommunity-based targeting has to poverty reduction. led to more participatory and accountable targeting processes. As a result, there is a gen- eral consensus among the highlands commu- nities that the PSNP targets the poorest Annex 5A. Case studies of social households. In fact, it is considered better tar- protection and risk management geted than any other African safety net pro- in Ethiopia, the Philippines, and gram (Coady, Grosh, and Hoddinott 2004; Coll-Black et al. 2013). Pakistan The PSNP provides cash or food transfers, Ethiopia: Moving from crisis and typically for six months each year, coinciding humanitarian response to resilience with the lean season (between June and building and proactive risk August). The transfer value from the program management3 launch in 2005 to June 2015 equaled 15 kg of cereals per household member per month, or Given the persistence of food insecurity in its cash equivalent. Households with able- Ethiopia, the government of Ethiopia bodied adults contribute to developing their launched the Food Security Program (FSP) in communities through public works activities 2005. The FSP was designed to institute a (such as soil and water conservation mea- movement away from ad hoc responses to sures, school room construction or rehabilita- food insecurity—as characterized by a major tion, water point development, and road drought in 2002—to a planned and system- r ehabilitation). Households with no ­ atic approach. It includes two programs: ­ able-bodied adults are considered “­ permanent • Productive Safety Net Program (PSNP): direct support” households and are not This is the primary program, which required to undertake public works activities. includes a Public Works component The PSNP can also scale up in a crisis, by (builds community infrastructure and drawing on contingency budgets. 1 6 6    SHOCK WAVES Scaling up to ensure prompt, proactive PSNP smooths consumption and protects response to crisis and humanitarian needs. assets in times of crisis. How is the PSNP far- The PSNP has various components that can ing? Biannual PSNP Impact Evaluations have scale up in response to shocks to support concluded that the program is succeeding in transitory needs. A key one is the Risk smoothing consumption and protecting Financing Mechanism (RFM), which facili- assets—even during times of crisis (see tates additional transfers to existing recipi- Box 5A.1). The transfers provided to PSNP ents or a temporary expansion of coverage to households are the equivalent of 45 percent of reach nonrecipients who are affected by a annual food needs for public work beneficia- shock. The RFM acts as an intermediate pol- ries, and 90 percent for “permanent direct icy response between the PSNP (addressing support” beneficiaries. Moreover, while much chronic food insecurity) and emergency of the cash transfer is spent on consumption, operations. The process of scaling up has around 25 percent of funds are invested in three stages: productive assets. Further ­ evidence suggests that the PSNP helps beneficiaries boost agri- 1. Early warning triggered: The early warn- cultural investments, ­ leading to higher rates ing system routinely collects and analyzes of agricultural productivity. early warning data. When the early warn- Although emergency operations are not ing system triggers a response, a request necessarily less efficient than PSNP risk- for the release of funds is prepared and the financed transfers at delivering food, the RFM Management Committee determines potential for greater cost-effectiveness lies in the number of beneficiaries and length of (i) the typically lower cost (though not always support required. Through contingency greater cost-efficiency) of cash as opposed to plans developed at the woreda level, bot- food delivery, alongside a range of other less tom-up needs are reconciled with available quantifiable benefits of cash; (ii) the use of resources and funds are released for existing capacity to deliver; and (iii) the pre- distribution. vention of asset erosion and weakening of 2. Resource transfer: Funds are released by future household livelihoods with timely risk- the RFM Management Committee either financed transfers, rather than lengthy for transfer to the regions and onward to humanitarian appeals. woredas or for food to be purchased, which PSNP explicitly supports long-term adap- is then dispatched directly to the woredas. tation and resilience. As for concerns that 3. I m pl e m e n t i ng c o n t i nge n c y pl a n s : safety net programs such as the PSNP can, Although the contingency plans are to varying extents, be “maladaptive” in the woreda plans, implementation of most face of climate change, the public works activities such as public works will be component of the PSNP and the HABP/live- carried out at the ward level with tech- lihoods component are designed to explic- nical support from woreda and sectoral itly encourage adaptation. Factors like soil experts. Normal public works procedures erosion and water scarcity are being actively will be used. Notably, however, the RFM countered to make locales and livelihoods Guidelines provide an option to waive sustainable through regeneration. In fact, the public works requirements in severe 60 percent of the PSNP’s public works sub- situations. projects target soil and water conservation, The advantage of this approach is that it is strengthening both livelihoods and resilience early and preventive, rather than late and to the impacts of variable rainfall. Together, reactive (IDL Group 2009). The newest phase soil erosion– and water conservation– of the program, PSNP 4, involves measures focused PSNP projects have led to signifi- that support a better emergency response by cant and visible increases in wood and integrating with the humanitarian system herbaceous vegetation cover and a broader more directly. diversity of plant species. L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   167 BOX 5A.1  How the PSNP helped households cope with Ethiopia’s 2011 food crisis The risk financing procedure was set in motion fol- intervals built into public works transfers and to pre- lowing several successive inadequate seasonal rains vent households from neglecting agricultural activi- and local crop failures—made worse by a sharp rise ties during this period. in food prices during the main PSNP transfer period. Over 80 percent of 11.1 million people identified Initially, the cash value of the transfer was raised as experiencing transitory food insecurity in 2011 from 10 to 15 birr (US$0.50 to $0.75) for a wage were brought into the PSNP, leaving only 20 per- day doing public works (and the same increase for cent to be covered by emergency operations. Of the direct support beneficiaries). Next, a detailed assess- total 9.26 million people to be assisted from con- ment was done for the size and distribution of a tingency funds, 6.5 million (70 percent) were PSNP contingent risk financing operation (GFDRE 2011). clients and 2.8 million (30 percent) were humanitar- It found that 9.26 million PSNP and nonPSNP cli- ian recipients. Since a total of 11.1 million people ents in hotspot woredas would require risk financing were determined to need transitory assistance in support to meet consumption needs for up to three Ethiopia as a whole, this highlights the size of the months between September and November 2011. It potential humanitarian caseload shifted from emer- also found that risk finance payments should be pro- gency operations to contingency funding thanks to vided as unconditional transfers to avoid payment risk financing. Philippines: A postdisaster response Development (DSWD) is the lead agency for to Typhoon Yolanda based on existing “disaster response” within the government’s social protection institutions and National Disaster Risk Reduction and instruments4 Management Plan (NDRRMP). It also has responsibilities across the national prevention Typhoon Yolanda (internationally referred and mitigation, preparedness, recovery, and to as Typhoon Haiyan) struck the Philippines rehabilitation pillars of the NDRRMP. And it on November 8, 2013, costing the country is the lead agency of four coordinating clus- an estimated P571.2 billion (US$12.9 billion) ters of the UN cluster system—food security, in damages, with over a million homes dam- shelter, camp coordination and camp man- aged or destroyed. Nearly 6,300 people died. agement, and protection. As a result of this A further 4.1 million people were displaced. linkage of social protection and disaster risk It affected some of the country’s poorest management, SP programs are well posi- regions and was estimated to increase tioned to respond to disasters. national poverty incidence by 1.9 percent, From the emergency to long-term recon- with estimates of up to an additional struction. In response to Yolanda, DSWD million people falling into poverty. Given 1 ­ implemented a variety of SP and social welfare such immense human and economic conse- programs: distribution of in-kind relief items, quences, Typhoon Yolanda provides an cash transfers (unconditional and conditional), interesting case study on how existing social shelter, and community-driven development. protection (SP) systems can be used to coor- These programs are mapped out in dinate and implement postdisaster support. figure 5A.1, illustrating the postdisaster phases ­ The Philippines has one of the most in which each program was implemented. advanced SP systems—backed with advanced Initially, the emphasis was on food and information and delivery systems—in the East nonfood items (like mats, blankets, tarpau- Asia Pacific region, designed to help poor lins, hygiene kits, and clothing) to meet the households manage risk and shocks. The immediate and urgent survival needs, plus D e p a r t m e n t o f S o c i a l We l f a r e a n d temporary shelter assistance for displaced 1 6 8    SHOCK WAVES households. By end-November 2013, 375,000 The  National Community-Driven food packs had been distributed, rising to 5.1 Development (NCDD) program of DSWD million by end-December. was set up in 2002 to alleviate rural poverty. After immediate survival needs were A contingent component of the NCDD was addressed, DSWD delivered a number of designed to adjust and simplify procedures cash-based response programs, such as Cash in the case of disasters, triggered by the dec- for Work , Cash for Building Livelihood laration of a state of calamity. Under this Assets , and cash for shelter (“ Emergency program, i ­nfrastructure selected by commu- Shelter Assistance”)—then transformed into nities is being constructed (or reconstructed) the Core Shelter Assistance Program to in Yolanda-affected areas. rebuild permanent housing. DSWD also tem- Integration of postdisaster support with porarily removed all conditionality of existing social protection systems. In the Pantawid Pamiliya Pilipino Program (4Ps), a event of a disaster, the 4Ps regular condi- usually conditional cash transfer program. In tional cash transfer can be leveraged to addition, at least 45 international humanitar- deliver its cash transfers unconditionally, ian agencies implemented cash transfers through the removal of grant conditions for (unconditional and conditional), partly deliv- beneficiaries in affected areas for a defined ered through the 4Ps infrastructure. Four period of time. After a natural disaster, it is million, agencies alone distributed around $34 ­ unrealistic to assume that beneficiaries can benefiting 1.4 million disaster-affected meet the conditions of the conditional cash people. transfer, especially in instances where the Over the longer term, the Yolanda experi- supply side may be down, with schools and ence has demonstrated the important health centers destroyed or being used for role that community-driven development relief operations. In such cases, the DSWD ­ programs can play in the recovery of poor Regional Director of an affected area sub- and vulnerable people from disasters. mits a request to deem all beneficiaries in the FIGURE 5A.1  Multiple programs answer different needs in postdisaster contexts in the Philippines Community-driven National community-driven development Core shelter assistance program Emergency shelter assistance Shelter Temporary shelter assistance Programs Cash for building livelihood assets Cash transfer Temporarily unconditional 4Ps cash transfer program Cash for work Nonfood items Relief Food for work Family food packs Relief Early recovery Recovery Reconstruction “Response” “Recovery and rehabilitation” Time Source: Bowen, forthcoming. L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   169 affected area as compliant to the 4Ps’ beneficiaries, whether they are vulnerable National Project Management Office. If (like older persons or lactating mothers), approved at the level of the Secretary, bene- and their estimated monthly income. ficiaries will be deemed compliant for a Humanitarian agencies were encouraged maximum of three consecutive compliance by the government to adopt the DAFAC verification periods (three months). This is as their primary tool for monitoring their not an unprecedented procedure, with the programs to increase coordination Bolsa Familia program in Brazil having between the government agencies and undertaken similar steps to deliver uncondi- humanitarian programs. tional payments to beneficiaries affected by • The Listahanan is a national registry flooding in 2011. Using the preexisting 4Ps used to target beneficiaries of the national system in this way, DSWD was able to conditional cash transfer program and quickly release a total of P550.5 million other government social programs. In (US$12.5 million) in unconditional cash 2011, it contained information on 11 mil- transfers to Yolanda-affected 4Ps beneficia- lion households, but it is only updated ries between November 2013 and February every four years. A process was created 2014 (DSWD 2014). for humanitarian agencies to coordinate Considerable effort was also made to with the Listahanan, thereby helping the coordinate the large number of cash trans- DSWD to meet its mandate to increase fer providers with the government cash- the 4Ps coverage by 20,000 poor house- based response programs—with some holds affected by Typhoon Yolanda. positive results. In the case of Yolanda, • The 4Ps Beneficiary Update List (BUL) both the World Food Program (WFP) and is a subset of the Listahanan. Updated UNICEF used the 4Ps system to deliver every two months, it contains informa- additional cash transfers to affected house- tion on poor households (as identified holds. The WFP and UNICEF “topped up” using the Listahanan and a Proxy Means the amount delivered by DSWD to house- Test targeting method) that are also ben- holds in affected areas, effectively scaling eficiaries of the national 4Ps program. up the conditional cash transfer grant A number of humanitarian cash transfer amount during a time of increased need for programs directed assistance directly to affected beneficiaries. This was an innova- 4Ps households using this up-to-date ben- tive and replicable practice illustrating the eficiary information. potential efficiency gains of delivering post- • The 4Ps Grievance Redress System (GRS) disaster grants through a national cash is a component of the 4Ps that allows any- transfer program to pretargeted poor and one to direct queries, clarifications, com- vulnerable households. Finally, DSWD and plaints, grievances, and appeals to the the UN Office for the Coordination of appropriate 4Ps committees at the com- Humanitarian Affairs (UNOCHA) also munity, provincial, regional, and national encouraged the humanitarian cash transfers levels (UNOCHA 2014). to leverage existing social protection and other systems to increase coordination in The extent to which these recommenda- their delivery, including: tions were adopted by the interagency humanitarian response remains unclear, and • Disaster Affected Family Assistance Card details on lessons learned are yet to emerge. (DAFAC) is the primary tool for monitor- Nevertheless, the potential value of a more ing the receipt of all benefits by affected unified overall response facilitated by preex- households, including cash. It records isting SP information systems is clear, and basic information pertaining to whether further research into the streamlining of these the beneficiaries’ houses were damaged information systems into future interagency (partially or totally), whether they are 4Ps response would add much value. Ultimately, 1 7 0    SHOCK WAVES however, post-Yolanda response documenta- including specific contingent financing or tion and reviews reveal that (i) there has been linking to the broader government risk no rigorous evaluation of the impact of the financing strategy (like a Catastrophe overall response and (ii) there nevertheless Deferred Drawdown Option (CAT-DDO)). appear to have been issues in coordinating In this way, the Department of Finance cash transfer programs, leading to coverage would be able to utilize a proven delivery gaps and duplication. mechanism to better cover household level As currently designed, the postdisaster risk, delivering additional assistance directly cash transfers are only provided to 4Ps benefi- to those most in need. ciaries and not to other affected households that may be equally or more poor and as or Pakistan: A two-phase window to more affected by the disaster. A solution balance urgency with targeting5 could be to implement a complementary but separate emergency cash transfer program, In July and August 2010, during the mon- based on the 4Ps information and payment soon season, Pakistan experienced the systems, that would be able to reach these worst floods in its history. The floods cov- households. This would create a more equi- ered all four provinces of the country table and efficient postdisaster cash-based (Sindh, Punjab, Khyber Pakhtunkwa, and intervention while preserving the integrity of Baluchistan), as well as the autonomous the existing 4P conditional cash transfer and territories of Gilgit-Baltistan and Azad its long-term human capital accumulation Jammu and Kashmir (AJK). More than 20 objectives. million people were affected, with over Scaling up of social protection requires 1,980 reported deaths. About 1.6 million funding—while reserve funds were avail- homes were destroyed, 2.4 million hectares able, contingent finance mechanisms could of crops damaged, and both farm and non- also help. Accessing finance to fund relief farm livelihoods were severely affected programs through preestablished budget (United Nations 2010). lines in a timely fashion is a necessity postdi- Pakistan’s main response was the cre- saster, when the affected may be slipping ation of the federal government’s Citizen’s into destitution by the hour. The Philippines Damage Compensation Program (CDCP), a recognized that response activities of the sort rapid-response cash grant program—rather carried out by DSWD require immediate than using an existing social safety net liquidity to be delivered rapidly, and that mechanism. Drawing on positive prior this could not be achieved quickly enough experience from the 2009 civil crisis, it through the regular calamity fund—the decided to deliver the cash transfers through National Disaster Risk Reduction and commercial banks, working closely with Management Fund (NDRRMF). The Quick provincial governments and the National Response Fund (QRF) was created to Database Registration Authority (NADRA). address this issue, with dedicated reserves Selected program beneficiaries were issued for response activities. Nevertheless, the spe- Visa direct debit cards, called Watan cards, cific case of Yolanda highlights problems in for collecting grants from ATM machines sourcing financing from even the QRF in a or designated points of sale. timely fashion. It also raises questions about Phase I of the CDCP focused on immedi- the adequacy of the QRF amount for a fiscal ate support. In Phase I (September 2010 to year’s worth of disasters because, although June 2011), the goal was to provide quick the QRF may be replenished once emptied, assistance to families who lost their homes or that is a lengthy process. faced a serious threat to their well-being. The There are a number of risk financing program was funded by the government, mechanisms that could allow DSWD to which provided almost $400 million in cash fund the grant top-ups in times of disaster, grants to more than 1.62 million families L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   171 (World Bank 2013b). Eligible households were provided with cash payments that could were given a one-off cash grant in the be used to meet any recovery needs (like amount of PRs 20,000 (about US$213), reconstructing houses, restoring livelihoods, based on funds available to cope with the or repaying accumulated debt). The size of urgent needs of a very large flood-affected the grant to eligible households was doubled target population. to PRs 40,000 (around US$426), a more suit- The provincial and regional governments able amount to support recovery, provided in identified CDCP beneficiaries in two ways. A two installments of PRs 20,000 each. geographical targeting system was used in Measures were taken from the outset to Punjab, Sindh, and Balochistan. Entire com- address the targeting issues found in Phase I. munities were identified as calamity affected These changes meant that not all Phase I ben- through notification by each province of the eficiaries were eligible for Phase II support, flood-affected areas (determined by visually and some people excluded from Phase I were calculating that at least 50 percent of houses included in Phase II. Housing damage was or crops were lost). A Rapid Housing Survey adopted as a proxy indicator for livelihood was used in Khyber Pakhtunkwa (KP) prov- losses nationwide, rather than the geographic ince and the autonomous territories of Gilgit targeting method previously used in Baltistan and AJK, with families living in Balochistan, Punjab, and Sindh. This meant flood-affected housing units, rather than that the existing Rapid Housing Surveys communities, being identified as flood could be used for targeting in KP and the affected. autonomous regions, whereas new surveys The findings from the Phase I evaluation needed to be conducted in the other three showed that the funds helped households provinces. cover their needs at a crucial time (World The eligibility criteria were further refined Bank 2013b), with the grants mostly used to filter out the wealthier and to include par- for food, health needs, housing repair, and ticularly vulnerable households by adding debt repayment (Hunt et al. 2011). However, two new eligibility criteria: the amount was insufficient for the flood- • Well-off households are excluded from affected households to recapitalize their receiving the Phase II transfer. Wealth is damaged or lost assets. The evaluation also measured by a combination of proxies— suggested that for every 100 potentially eli- such as those having bank accounts in gible family heads, only 43 had received the international banks, frequent interna- Watan card. Inclusion and exclusion errors tional travel activities, and executive explain these results. Geographical targeting jobs. missed households that suffered from dam- • All legitimate vulnerable beneficiaries ages but lived in a weakly affected commu- (defined as female- and disabled-headed nity; and the Rapid Household Surveys families in the NA DR A's database) missed some damages and were conducted included in Phase I but not captured as by local notables, not by experts able to head of household through the housing detect all damages. Finally, the beneficiary damage survey will, de facto, become selection and verification process proved to Phase II beneficiaries. be lengthy and cumbersome, particularly for those who had lost the documentation neces- The vulnerability characteristics of flood- sary for verification either prior to, or affected families or households were profiled ­ during, the floods. by analyzing a random sample from Phase II of the CDCP distributed larger NADRA’s flood registration database and amounts and was better targeted. In Phase II linking this with information on gender, dis- (June 2011 to June 2013), with total resources ability, and educational levels in the civil reg- of around $600 million, flood-affected house- istration database (World Bank 2013b). holds, including many of those from Phase I, Additionally, the outstanding legitimate 1 7 2    SHOCK WAVES grievance claims from Phase I were settled processes reduced exclusion errors in and considerable resources used to strengthen Balochistan to 30 percent, compared to a the government’s communications, grievance baseline of 73 percent. redress, and policy and implementation A model of beneficiary registration and capacities at different levels. payment. The CDCP offers a model of how The new targeting mechanism was cum- to e­ stablish an efficient decentralized benefi- bersome for good reason—with fairly large ciary ­ r egistration system for a very large transfers, the focus had to be on reducing number of clients over a widespread geo- inclusion errors and then addressing the graphic area. By the end of Phase I, more exclusion challenge through a rigorously than 1.6 million families had been enrolled, applied grievance appeals system that came to and RPs 33 ­ b illion (US$374 million) was play a critical role in determining legitimate distributed (World Bank 2013b). A further beneficiaries and drawing up beneficiary lists. 874,000 Watan cards have been issued Data suggest that Phase II has been reason- since then and nearly RPs 31.9 billion ably successful in targeting the most severely (US$337.6 million) disbursed during Phase affected and the most vulnerable households II up to June 2012 (World Bank 2013b). (like the poorest and least educated) in the This is an impressive logistical and adminis- four provinces (except Balochistan). trative achievement. However, beneficiaries experienced payment Over the course of Phases I and II, NADRA delays, with just 63 percent of households established 101 CDCP local offices, named having receiving both tranches and 13 per- Watan Card Facilitation Centers (WCFCs), cent having received neither tranche—more covering all of the flood-affected districts. The than a year after the inauguration of Phase II WCFCs serve as a “one-stop shop,” where and almost three years after the 2010 the beneficiaries are enrolled, register com- flooding. plaints or grievances, and often receive Grievance appeals are standard features in ­ p ayments via a Point of Sale machine. safety net systems, and can be particularly Biometric screening is used to verify the ben- important in emergency-type interventions, eficiary identities to prevent fraudulent claims. where often simple and straightforward Beneficiaries are then registered and issued a selection procedures are applied within rela- Watan card, which can be used at the Point tively short periods of time. They made for Of Sale desk or any ATM. In certain districts, substantially improved targeting. Whereas the placement of a cash desk at the WCFC initial beneficiary identification (the baseline) (like on-site cash storage) was deemed a secu- had resulted in the estimated exclusion of rity risk and payments have been processed at 61 ­ p ercent of potentially eligible flood- a local bank branch, usually one or two kilo- affected households, with modified pro- meters away from the WCFC. cesses, and especially a fully operational grievance-redress system, errors of exclusion were reduced to 32 ­ percent. Where exclusion Notes did occur, it was driven by difficulties in cap- 1. Grosh et al. (2008) provide an extensive turing people without identity cards (who review of existing social protection schemes. would be automatically excluded), those 2. http://www.bcn.cl/de-que-se-habla/bono​ without a ­ permanent address, very isolated -­solidario-marzo-2009. households, and those living in insecure 3. Source: Background paper for this report by areas. Enrollment difficulties were encoun- Johnson and Bowen (forthcoming). tered in particular in Balochistan and Gilgit- 4. Source: Background paper for this report by Balistan, where low administrative capacity Bowen (forthcoming). 5. Source: GFDRR case study on Pakistan in both areas, and stone-built, less flood-­ (World Bank 2013b) and Implementation sensitive houses in the latter, resulted in few completion and results report ICR00003119 households being verified. Still, improved (World Bank 2014). L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   173 References 2045-2102. The Volatility of Overseas Aid 36: 1,766–85. Adger, W. N., P. M. Kelly, A. Winkels, L. Q. Huy, Bastagli, F. 2009. “From Social Safety Net to and C. Locke. 2002. “Migration, Remittances, Social Policy? The Role of Conditional Cash Livelihood Trajectories, and Social Resilience.” Transfers in Welfare State Development in AMBIO J. Hum. Environ. 31: 358–66. Latin America.” Working Paper, International Adger, W. N., J. M. Pulhin, J. Barnett, Policy Centre for Inclusive Growth. G. D. Dabelko, G.K. Hovelsrud, M. Levy, ———. 2014. Responding to a Crisis: The Design S. Oswald, and C. H. Vogel. 2014. “Human and Delivery of Social Protection. London, Security.” In Climate Change 2014: Impacts, U.K.: Overseas Development Institute. Adaptation, and Vulnerability. Part A: Global Becerra, O., E. Cavallo, and I. Noy. 2013. “Where and Sectoral Aspects, edited by C. B. Field, Is the Money? Post-disaster Foreign Aid V. R. Barros, D. J. Dokken, K. J. Mach, Flows.” Environ. Dev. Econ. 1–26. M. D. Mastrandrea, T. E. Bilir, M. Chatterjee, Black, R., W. N. Adger, N. W. Arnell, S. Dercon, K. L. Ebi, Y. O. Estrada, R. C. Genova, B. Girma, A. Geddes, and D. Thomas. 2011a. “The Effect E. S. Kissel, A. N. Levy, S. MacCracken, of Environmental Change on Human P. R. Mastrandrea, and L. L. White, 755–791. Migration.” Glob. Environ. Change 21: S3–S11. Contribution of Working Group II to the Fifth Black, R., S. R. G. Bennett, S. M.Thomas, and Assessment Report of the Intergovernmental J. R. Beddington. 2011b. “Climate Change: Panel on Climate Change. Cambridge, U.K. and Migration as Adaptation.” Nature 478: New York: Cambridge University Press. 447–49. Aker, J. C., R. Boumnijel, A. McClelland, and Bowen, T. Forthcoming. “Social Protection and N. Tierney. 2014. “Payment Mechanisms and Disaster Risk Management in the Philippines: Anti-Poverty Programs: Evidence from a The Case of Typhoon Yolanda (Haiyan).” Mobile Money Cash Transfer Experiment in Background paper prepared for this report. Niger.” Unpubl. Work. Pap. Brix, L., and K. McKee. 2010. “Consumer Akresh, R., D. De Walque, and H. Kazianga. Protection Regulation in Low-Access 2013. “Cash Transfers and Child Schooling: Environments: Opportunities to Promote Evidence from a Randomized Evaluation of the Responsible Finance.” Focus Note 60. Role of Conditionality.” Policy Res. Work. Brown, J. K., T. V. Zelenska, and A. M. Mobarak. Paper, World Bank, Washington, DC. 2013. Barriers to Adoption of Products and A l d e r m a n , H . , a n d T. H a q u e . 2 0 0 6 . Technologies That Aid Risk Management in “Countercyclical Safety Nets for the Poor and Developing Countries. Washington, DC: World Vulnerable.” Food Policy 31: 372–83. Bank. Alkire, S., J. M. Roche, and A. Vaz. 2014. Bryan, G., S. Chowdhury, and A. M. Mobarak. “Multidimensional Poverty Dynamics: 2014. “Underinvestment in a Profitable Methodology and Results for 34 Countries.” Technology: The Case of Seasonal Migration in OPHI Res. Prog. 3. Bangladesh.” Econometrica 82: 1,671–1,748. Allen, F., A. Demirgüç-Kunt, L. F. Klapper, and Bugamelli, M., and F. Paterno. 2009. “Do M. S. Martinez Peria. 2012. “The Foundations Workers’ Remittances Reduce the Probability of Financial Inclusion: Understanding of Current Account Reversals?” World Dev. Ownership and Use of Formal Accounts.” 37: 1821–38. Policy Res. Work. Paper, World Bank, Bundy, D., C. Burbano, M. Grosh, A. Gelli, Washington, DC. M. Jukes, and L. Drake. 2009. Rethinking Arnold, M., R. Mearns, K. Oshima, and V. Prasad. School Feeding. Washington, DC: World Bank. 2014. Climate and Disaster Resilience: The Cardenas, V., S. Hochrainer, R. Mechler, G. Pflug, Role for Community-Driven Development . and J. Linnerooth-Bayer. 2007. “Sovereign Washington, DC: World Bank. Financial Disaster Risk Management: The Case Banerjee, A., and E. Duflo. 2012. Poor Economics: of Mexico.” Environ. Hazards 7: 40–53. A Radical Rethinking of the Way to Fight Carr, E. R. 2008. “Between Structure and Global Poverty. Reprint edition. New York: Agency: Livelihoods and Adaptation in PublicAffairs. Ghana’s Central Region.” Glob. Environ. Barnett, B. J., C. B. Barrett, and J. R. Skees. 2008. Change 18: 689–99. “Poverty Traps and Index-Based Risk Transfer Carter, M. R., and S. A. Janzen. Forthcoming. Products.” World Dev. , Special Section: “Social Protection in the Face of Climate 1 7 4    SHOCK WAVES Change: Targeting Principles and Financing Household Risk Management: Evidence from Mechanisms.” Background paper prepared for India.” Am. Econ. J. Appl. Econ. 5: 104. this report. Coll-Black, S., D. O. Gilligan, J. Hoddinott, Cazes, S., S. Verick, and C. Heuer. 2009. “Labor N. Kumar, A. S. Taffesse, and W. Wiseman. Market Policies in Times of Crisis.” 2013. Targeting Food Security Interventions: Employment Sector Employment Working The Case of Ethiopia’s Productive Safety Net Paper No. 35, International Labour Office, Programme. Washington, DC: International Geneva. Food Policy Research Institute. Chambwera, M., G. Heal, C. Dubeux, de Janvry, A., F. Finan, E. Sadoulet, and R. Vakis. S. Hallegatte, L. Leclerc, A. Markandya, 2006. “Can Conditional Cash Transfer B. A. McCarl, R. Mechler, and J. E. Neumann. Programs Serve as Safety Nets in Keeping 2014. “Economics of Adaptation.” In Climate Children at School and from Working When Change 2014: Impacts, Adaptation, and Exposed to Shocks?” J. Dev. Econ. 79: 349–73. Vulnerability. Part A: Global and Sectoral del Ninno, C., and S. Coll-Black. 2015. Protecting Aspects, edited by C. B. Field, V. R. Barros, the Vulnerable in the Drylands: The Role of D. J. Dokken, K. J. Mach, M. D. Mastrandrea, Social Protection . Washington, DC: World T. E. Bilir, M. Chatterjee, K. L. Ebi, Bank. Y. O. Estrada, R. C. Genova, B. Girma, del Ninno, C., P. A. Dorosh, L. C. Smith, and E. S. Kissel, A. N. Levy, S. MacCracken, D. K. Roy. 2001. The 1998 Floods in P. R. Mastrandrea, and L. L. White, 945–977. Bangladesh Disaster Impacts, Household Contribution of Working Group II to the Fifth Coping Strategies, and Response (Research Assessment Report of the Intergovernmental Report No. 122). International Food Policy Panel on Climate Change. Cambridge, U.K. Research Institute. and New York: Cambridge University Press. Demeke, M., G. Pangrazio, and M. Maetz. 2009. Chami, R., D. Hakura, and P. J. Montiel. 2009. Country Responses to the Food Security Crisis: “Remittances: An Automatic Output Nature and Preliminary Implications of the Stabilizer?” IMF Work. Pap. 1–31, International Policies Pursued . Rome, Italy: Food and Monetary Fund. Agricluture Organization. Clarke, D. J., and R. V. Hill. 2013. “Cost-Benefit de Nicola, F. 2015. “Expanding or Increasing: Analysis of the African Risk Capacity Facility.” Index-Based Scalable Social Protection in IFPRI Discuss. Pap. 01292, International Food Niger.” Presentation at Disaster Risk Financing Policy Research Institute. and Insurance (DRFI) Conference, FERDI, Clarke, D., and R. Poulter. 2014. “A Methodology Clermont-Ferrand, France, June. for Calculating the Opportunity Cost of Devereux, S. 2015. “Social Protection for Layered Sovereign DRFI Strategies.” Impact Enhanced Food Security in Sub-Saharan Appraisal for Sovereign Disaster Risk Financing Africa.” Food Policy. and Insurance Project: Phase 1. SDRFI Impact Drèze, J., and R. Khera. 2015. “Understanding Appraisal Project. Leakages in the Public Distribution System.” Clarke, G., and S. Wallsten. 2004. Do Remittances Econ. Polit. Wkly. 50: 39. Protect Households in Developing Countries DSWD (Department of Social Welfare and against Shocks? Evidence from a Natural Development). 2014. Response and Early Disaster in Jamaica. Washington, DC: World Recovery Updates . Department of Social Bank. Welfare and Development, Government of the Coady, D., M. E. Grosh, and J. Hoddinott. 2004. Philippines. Targeting of Transfers in Developing Countries: Fagen, P. W. 2006. “Remittances in Conflict and Review of Lessons and Experience. Crises: How Remittances Sustain Livelihoods in Washington, DC: World Bank. War, Crises and Transitions to Peace.” Policy Cole, S., G. Bastian, S. Vyas, C. Wendel, and Paper, International Peace Academy. D. Stein. 2012. The Effectiveness of Index- FAO (Food and Agriculture Organization of the Based Micro-insurance in Helping Smallholders United Nations). 2013. A Panorama of School Manage Weather-Related Risks. London: EPPI- Feeding and the Possibilities for Direct Centre, Social Science Research Unit, Institute Purchases from Family Farming—Case Studies of Education, University of London. in Eight Countries . Rome, Italy: Food and Cole, S., X. Giné, J. Tobacman, R. Townsend, Agriculture Organization of the United P. Topalova, and J. Vickery. 2013. “Barriers to Nations, Rome, Italy. L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   175 ———. 2015. Qualitative Research and Analyses Handa, S., A. Peterman, C. Huang, C. Halpern, of the Economic Impacts of Cash Transfer A. Pettifor, and H. Thirumurthy. 2015. “Impact Programmes in Sub-Saharan Africa: Synthesis of the Kenya Cash Transfer for Orphans and Report. Rome, Italy: Food and Agriculture Vulnerable Children on Early Pregnancy and Organization of the United Nations. Marriage of Adolescent Girls.” Soc. Sci. Med. Fiszbein, A., N. R. Schady, and F. H. Ferreira. 141: 36–45. 2009. Conditional Cash Transfers: Reducing Hochrainer-Stigler, S., R. Mechler, G. Pflug, and Present and Future Poverty. Washington, DC: K. Williges. 2014. “Funding Public Adaptation World Bank. to Climate-Related Disasters. Estimates for a Francisco, H. A., C. D. Predo, Global Fund.” Glob. Environ. Change 25: A .   M a n a s b o o n p h e m p o o l , P.   Tr a n , 87–96. R. Jarungrattanapong, L. M. Penalba, N. Tuyen, Holzmann, R., J. Koettl, and T. Chernetsky. 2005. T. H. Tuan, D. D. Elazegui, and Y. Shen. 2011. Portability Regimes of Pension and Health Determinants of Household Decisions on Care Benefits for International Migrants: An Adaptation to Extreme Climate Events in Analysis of Issues and Good Practices. Global Southeast Asia. Economy and Environment Commission on International Migration. Program for Southeast Asia (EEPSEA). Hunt, S., S. O’Leary, T. N. Jones, and Z. Ali. Gentilini, U. Forthcoming. Safety Nets in Urban 2011. Evaluating Implementation of Pakistan’s Areas: Emerging Issues, Evidence and Practices. Citizens Damage Compensation Program Washington, DC: World Bank. (Phase 1) Final Report. Oxford, U.K.: Oxford Gentle, P., R. Thwaites, D. Race, and K. Alexander. Policy Management. 2014. “Differential Impacts of Climate Change IDL Group. 2009. Guidelines for the PSNP Risk on Communities in the Middle Hills Region of Financing Mechanism in Ethiopia: Final Nepal.” Nat. Hazards 74: 815–36. Guidelines. Gertler, P. 2004. “Do Conditional Cash Transfers IMF (International Monetary Fund). 2013. Energy Improve Child Health? Evidence from Subsidy Reform: Lessons and Implications. P R O G R E S A’s C o n t r o l R a n d o m i z e d Washington, DC: International Monetary Experiment.” Am. Econ. Rev. 336–41. Fund. GFDRE (Government of the Federal Democratic Johnson, K., and T. Bowen. Forthcoming. Republic of Ethiopia) 2011. Productive Safety “Ethiopia’s Food Security Program: Social Net Program—2011 Public Works Impact Protection, Disaster Risk Management and Assessment (No. Phase 1, Final Report). Addis Climate Change Adaptation.” Background Ababa, Ethiopia: Government of Ethiopia. paper prepared for this report. Ghesquiere, F., and O. Mahul. 2010. Financial Jones, N., C. Harper, S. Pantuliano, S. Pavanello, Protection of the State against Natural Disasters: K. Kyunghoon, S. Mitra, and K. Chalcraft. A Primer. Washington, DC: World Bank. 2009. Impact of the Economic Crisis and Food Government of India. 2011. Report of the and Fuel Price Volatility on Children and Working Group on Urban Poverty, Slums and Women in the MENA Region. London, U.K. Service Delivery System . New Delhi: and New York: Overseas Development Institute Government of India. and UNICEF. Grosh, M., M. Bussolo, and S. Freije. 2014. Jülich, S. 2011. “Drought Triggered Temporary Understanding the Poverty Impact of the Migration in an East Indian Village.” Int. Migr. Global Financial Crisis in Latin America and 49: e189–e199. the Caribbean. Washington, DC: World Bank. Kapur, D., and D. A. Singer. 2006. “Remittances, Grosh, M., Carlo del Ninno, E. Tesliuc, and Government Spending and the Global A. Ouerghi. 2008. For Protection and Economy.” Presented at the Annual Meeting of Promotion: The Design and Implementation of the American Economic Association, San Effective Safety Nets . Washington, DC: Diego, CA. World Bank. Karlan, D. S., R. Osei, I. Osei-Akoto, and C. Udry. Gupta, S. 2013. “The Mobile Banking and 2012. Agricultural Decisions after Relaxing Payment Revolution.” Eur. Financ. Rev. 2. Credit and Risk Constraints. National Bureau Gurenko, E. 2006. Earthquake Insurance in of Economic Research. Turkey: History of the Turkish Catastrophe Kenward, S., L. Cordier, and R. Islam. 2012. I n s u r a n c e P o o l . Wa s h i n g t o n , D C : Chars Livelihoods Programme: A Study to World Bank. Assess the Performance of CLP Raised Plinths, 1 7 6    SHOCK WAVES Low Cost Latrines and Access to Clean Water McCord, A. 2013. “The Public Pursuit of Secure During the July 2012 Flood . Dhaka, Welfare: Background Paper on International Bangladesh: Chars Livelihoods Programme and Development Institutions, Social Protection & Maxwell Stamp PLC. Developing Countries.” European Commission, Kinnan, C., and R. Townsend. 2012. “Kinship and Brussels. Financial Networks, Formal Financial Access, Meller, M., and S. Litschig. 2014. “Saving Lives and Risk Reduction.” Am. Econ. Rev. 102: Evidence from a Conditional Food 289–93. Supplementation Program.” J. Hum. Resour. Krishna, A. 2006. “Pathways out of and into 49: 1,014–52. Poverty in 36 Villages of Andhra Pradesh, Merttens, F., A. Hurrell, M. Marzi, R. Attah, India.” World Dev. , Part Special Issue: M. Farhat, A. Kardan, and I. MacAuslan. 324–404. Corruption and Development: 2013. Kenya Hunger Safety Net Programme Analysis and Measurement 34: 271–88. Monitoring and Evaluation Component. Kunreuther, H. C., M. V. Pauly, and S. McMorrow. Impact Evaluation Final Report: 2009 2013. Insurance and Behavioral Economics: to 2012 . Oxford, U.K.: Oxford Policy Improving Decisions in the Most Management. Misunderstood Industry. Cambridge, U.K. and Mobarak, A. M., and M. R. Rosenzweig. 2013. New York: Cambridge University Press. “Informal Risk Sharing, Index Insurance, and Kuriakose, A. T., R. Heltberg, W. Wiseman, Risk Taking in Developing Countries.” Am. C. Costella, R. Cipryk, and S. Cornelius. 2013. Econ. Rev. 103: 375–80. “Climate-Responsive Social Protection.” Dev. Mohapatra, S., G. Joseph, and D. Ratha. 2009. Policy Rev. 31: o19–o34. “Remittances and Natural Disasters: Ex Post Lawrance, E. C. 1991. “Poverty and the Rate of Response and Contribution to Ex Ante Time Preference: Evidence from Panel Data.” Preparedness.” Policy Res. Work. Pap. 4972, J. Polit. Econ. 54–77. World Bank, Washington, DC. Lawson, V., and S. Elwood. 2014. “Encountering Morduch, J. 1995. “Income Smoothing and Poverty: Space, Class, and Poverty Politics.” Consumption Smoothing.” J. Econ. Perspect. Antipode 46: 209–28. 103–14. Mahul, O., and F. Ghesquiere. 2007. “Sovereign Moser, C., and A. Felton. 2007. “Intergenerational Natural Disaster Insurance for Developing Asset Accumulation and Poverty Reduction in Countries: A Paradigm Shift in Catastrophe Guayaquil, Ecuador, 1978–2004.” Reducing Risk Financing.” Policy Res. Work. Pap. 4345, Glob. Poverty Case Asset Accumul. 15–50. World Bank, Washington, DC. Noy, I., and A. Nualsri. 2011. “Fiscal Storms: Maluccio, J. A. 2005. Coping with the “Coffee Public Spending and Revenues in the Aftermath Crisis” in Central America: The Role of the of Natural Disasters.” Environ. Dev. Econ. 16: Nicaraguan Red de Protección Social . 113–28. Washington, DC: International Food Policy Noy, I., and P. Patel. 2014. “Floods and Spillovers: Research Institute. Households after the 2011 Great Flood in Manuel-Navarrete, D., M. Pelling, and Thailand.” Victoria University of Wellington M. Redclift. 2011. “Critical Adaptation to SEF Working Paper 3609. Hurricanes in the Mexican Caribbean: O l s s o n , L . , M . O p o n d o , P. Ts c h a k e r t , Development Visions, Governance Structures, A. Agrawal, S. H. Eriksen, S. Ma, L. N. Perch, and Coping Strategies.” Glob. Environ. and S. A. Zakieldeen. 2014. “Livelihoods and Change 21: 249–58. Poverty.” In Climate Change 2014: Impacts, Martinich, J., J. Neumann, L. Ludwig, and Adaptation, and Vulnerability. Part A: Global L. Jantarasami. 2013. “Risks of Sea Level Rise and Sectoral Aspects, edited by C. B. Field, to Disadvantaged Communities in the United V. R. Barros, D. J. Dokken, K. J. Mach, States.” Mitig. Adapt. Strateg. Glob. Change M. D. Mastrandrea, T. E. Bilir, M. Chatterjee, 18: 169–85. K. L. Ebi, Y. O. Estrada, R. C. Genova, Martorano, B., and M. Sanfilippo. 2012. B .   G i r m a , E . S . K i s s e l , A . N . L e v y, “Innovative Features In Poverty Reduction S. MacCracken, P. R. Mastrandrea, and L. L. Programmes: An Impact Evaluation Of Chile White, 793–832. Contribution of Working Solidario On Households And Children.” J. Int. Group II to the Fifth Assessment Report of the Dev. 24: 1,030–41. Intergovernmental Panel on Climate Change. L end a H and : P oor P eople , S u pport S y stems , S afet y N ets , and   I ncl u sion   177 Cambridge, U.K. and New York: Cambridge UNOCHA (United Nations Office for the University Press. Coordination of Humanitarian Affairs). 2014. Ouattara, B., and E. Strobl. 2013. “The Fiscal Cash Coordination Update: Response to Implications of Hurricane Strikes in the Typhoon Yolanda (Yolanda). New York: United Caribbean.” Ecol. Econ. 85: 105–15. Nations. Paudel, Y. 2012. “A Comparative Study of Public- Vagliasindi, M. 2012. Implementing Energy Private Catastrophe Insurance Systems: Lessons Subsidy Reforms: Evidence from Developing from Current Practices.” Geneva Pap. Risk Countries. Washington, DC: World Bank. Insur.-Issues Pract. 37: 257–85. Vincent, K., and T. Cull. 2011. “Cell Phones, Perdana, A. A. 2014. “The Future of Social Electronic Delivery Systems and Social Cash Welfare Programs in Indonesia: From Fossil- Transfers: Recent Evidence and Experiences Fuel Subsidies to Better Social Protection,” from Africa.” Int. Soc. Secur. Rev . 64: Briefing Note, The International Institute for 37–51. Sustainable Development. Vincent, K. E., P. Tschakert, J. Barnett, Perova, E., and R. Vakis. 2012. “5 Years in Juntos: M. G. Rivera-Ferre, and A. Woodward. 2014. New Evidence on the Program’s Short and “Cross-Chapter Box on Gender and Climate Long-Term Impacts.” Economía 35: 53–82. Change.” In Climate Change 2014: Impacts, Porto, G., and D. Lederman. 2014. “The Price Is Adaptation, and Vulnerability. Part A: Global Not Always Right: On the Impacts of and Sectoral Aspects, edited by C. B. Field, (Commodity) Prices on Households (and V. R. Barros, D. J. Dokken, K. J. Mach, Countries).” Policy Research Working Paper M. D. Mastrandrea, T. E. Bilir, M. Chatterjee, S6858, World Bank, Washington, DC. K. L. Ebi, Y. O. Estrada, R. C. Genova, Ratha, D., S. De, E. Dervisevic, S. Plaza, B.   Gir ma, E. S. K is s e l, A . N. Le vy, K. Schuettler, W. Shaw, H. Wyss, S. Yi, and S. R. S. MacCracken, P. R. Mastrandrea, and Yousefi. 2015. Migration and Development L. L. White, 105–107. Contribution of Working Brief, No. 24. Washington, DC: World Bank. Group II to the Fifth Assessment Report of the Rogers, D. P., and V. V. Tsirkunov. 2013. Weather Intergovernmental Panel on Climate Change. and Climate Resilience: Effective Preparedness Cambridge, U.K. and New York: Cambridge through National Meteorological and University Press. Hydrological Services. Washington, DC: World Wodon, Q., A. Liverani, G. Joseph, and N. Bank. Bougnoux, N. 2014. Climate Change and Sabates-Wheeler, R., and S. Devereux. 2010. “Cash Migration : Evidence from the Middle East Transfers and High Food Prices: Explaining and North Africa . Washington, DC: World Outcomes on Ethiopia’s Productive Safety Net Bank. Programme.” Food Policy 35: 274–85. World Bank. 2006. Economic Implications of Stewart, F. 2010. “Horizontal Inequalities in Remittances and Migration (Global Economic Kenya and the Political Disturbances of 2008: Prospects). Washington, DC: World Bank. Some Implications for Aid Policy.” Confl. Secur. ———. 2011. Social Protection for a Changing Dev. 10: 133–59. India: Main Report. Washington, DC: World Suleri, A. Q., and K. Savage. 2006. Remittances in Bank. Crises: A Case Study from Pakistan. Sustainable ———. 2012. The World Bank’s Social Protection Development Policy Institute. and Labor Strategy: Resilience, Equity, and Trebilcock, M. J. 2014. Dealing with Losers: The Opportunity. Washington, DC: World Bank. Political Economy of Policy Transitions . ———. 2013a. Risk and Opportunity: Managing Oxford University Press. Risk for Development, World Development Tschakert, P. Forthcoming. “The Ability of the Report. Washington, DC: World Bank. Poor to Cope.” Background paper prepared for ———. 2013b. Pakistan’s Citizens Damage this report. Compensation Program (CDCP): Case Study. Tumbe, C. 2011. “Remittances in India: Facts & Washington, DC: World Bank. Issues.” SSRN Scholarly Paper 2122689, Social ———. 2014. Pakistan—Flood Emergency Cash Science Research Network, Rochester, NY. Transfer Project . Washington, DC: World United Nations. 2010. Multi-Cluster Initial Rapid Bank. Humanitarian Needs Assessment. New York: ———. 2015a. ASPIRE Database. Washington, United Nations. DC: World Bank. 1 7 8    SHOCK WAVES ———. 2015b. The State of Social Safety Nets. Yang, D., and H. Choi. 2007. “Are Remittances Washington, DC: World Bank. Insurance? Evidence from Rainfall Shocks in ———. 2015c. “Can Remittances Help Promote the Philippines.” World Bank Econ. Rev. 21: Consumption Stability?” In Global 219–48. Economic Prospects . Washington, DC: Zimmerman, F. J., and M. R. Carter. 2003. “Asset World Bank. Smoothing, Consumption Smoothing and the Wu, T. 2006. The Role of Remittances in Crisis: Reproduction of Inequality under Risk and An Aceh Research Study . London, U.K.: Subsistence Constraints.” J. Dev. Econ. 71: Overseas Development Institute. 233–60. A Window of Opportunity: Climate-Informed Development 6 and Pro-Poor Climate Policies Main Messages • In a pessimistic development scenario, climate • To ensure that emissions-reduction policies change could drag more than 100 million peo- do not slow down poverty reduction, coun- ple into poverty by 2030. This number can be tries need to focus on options that yield local reduced to fewer than 20 million, if rapid, (health or economic) cobenefits and protect inclusive, and climate-informed development poor people from the negative consequences is combined with targeted adaptation actions. of mitigation policies. • The impacts of climate change on poverty by • In poor countries where domestic resources 2030 mostly depend on development policy are insufficient to protect poor people, sup- choices. port from the international community is • Immediate emissions-reduction policies are essential. This is particularly true for invest- needed to reduce the longer-term threat of ments with high upfront costs that are criti- climate change to poverty and avoid the cal to prevent lock-ins into carbon-intensive post-2030 impacts on poverty that develop- patterns (such as for urban transport, energy ment policy alone cannot manage. infrastructure, or deforestation). Introduction framework—that the uncertainty around the So far, this report has presented plenty of potential impacts on poverty, development, ­ evidence that unmitigated climate change and welfare are extremely large. could become a significant obstacle to Our analysis has been centered on the ­ development and poverty eradication and— three channels through which climate-related maybe even more important in a risk-based people events already affect the ability of poor ­ 179 1 8 0    SHOCK WAVES to escape poverty: (i) agricultural production, to give a sense of the magnitude of the impact ecosystems, and food security; (ii) natural climate of development on the vulnerability to ­ disasters; and (iii) health. change impacts. But this report has also shown that there are limits to adaptation. The • Chapter 2 shows that losses in the agricul- evidence is strong that climate change mitiga- tural sector and spikes in food prices tion action is needed now to prevent much pushed people into poverty in the past, more severe impacts later this century (Fay as exemplified by the 2010–2011 episode et al. 2015; IPCC 2014; NCE 2014; OECD that increased poverty by 44 million 2015). This chapter provides thoughts on p eople—and that climate change will ­ how mitigation policies can stabilize the cli- likely magnify this threat. mate and the risks it creates, and do so with- • Chapter 3 discusses how natural disasters out slowing down poverty reduction. are already preventing many households from escaping poverty, with the poor more vulnerable to these events—and how climate change is likely to worsen By 2030, climate change will the situation. increase; but rapid, inclusive, • Chapter 4 shows that climate change will and climate-informed magnify the type of health shocks that are already a serious burden for poor peo- development can minimize ple and poor countries—such as malaria, its impact on poverty diarrhea, and stunting. Between now and 2030, climate policies can These chapters also identify many options to do little to reduce the amount of global warm- reduce the impact of climate change, especially ing already under way because of the long lag on the poor and vulnerable—​ from dikes and between the introduction of mitigation poli- irrigation systems to early warning systems, cies, their impact on emissions, and the effect better connection to markets, and universal of emissions reductions on the ­climate system. health coverage. In addition, chapter 5 pres- Only changes in short-lived climate pollutants ents instruments to make the population (like black carbon and methane) could have a more resilient to shocks at an affordable cost rapid impact, especially at local scale, but for public finances—for example, financial their potential at global scale remains rela- inclusion and social safety nets that are scal- tively limited (Rogelj et al. 2014). This scien- able and can target people hit by a health tific certainty means that the only way to shock or a flood. reduce climate change impacts by 2030 is by The message that emerges from these sec- lowering socioeconomic vulnerability to cli- toral and thematic chapters is that poverty is mate change—through both climate-informed one of the key markers of vulnerability and development and targeted adaptation efforts. that much of what is recommended as mea- We investigate here the potential efficacy of sures to make people and societies more resil- these development policies to reduce the pov- ient is simply good development policy. And, erty consequences of climate change (full given that greenhouse gas (GHG) emissions- results and technical details of the analysis reduction policies have limited impacts on presented below can be found in a background ­ climate change between now and 2030, it is paper for this report, Rozenberg and development progress that will be the key Hallegatte, forthcoming). To do so, we first determinant of the impacts of climate change look at what the future could look like with- on poverty in the short to medium term. out climate change. We then use the likely How much of a difference could good impacts of climate change on the poor identi- development make to the resilience of indi- fied in chapters 2–4 to examine how the viduals and societies? This chapter tries to aggregate impact of climate change on poverty answer this question—and for the first time, is affected by overall development progress. A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   181 The results are unequivocal: the impact of of socioeconomic scenarios and exploring climate change on poverty is conditioned by ­ how climate change would affect develop- overall development progress. ment in each of these scenarios. These sce- narios do not correspond to particularly likely futures (box 6.1). Rather, they are pos- Even without climate change, very sible and internally consistent futures, cho- different futures can be imagined for sen to cover a broad range of possible futures poverty and development to facilitate assessing possible orders of mag- It is impossible to forecast future socioeco- nitude of future climate change impacts. nomic development. Experience suggests we People sometimes refer to these scenarios as are simply not able to anticipate structural “what-if” scenarios because they can help shifts, technical breakthroughs, and geopo- answer questions such as: “What would the litical changes (Kalra et al. 2014). Here, we climate change impact be if socioeconomic do not predict future socioeconomic change development followed a given trend?” Our and we do not predict the impact of climate goal is to better understand how the impact change on poverty. of climate change on poverty depends on Instead, we follow a scenario-based socioeconomic development, estimate the approach that is the basis of all potential impacts in “bad” scenarios, and Intergovernmental Panel on Climate Change explore possible policy options to minimize (IPCC) reports. It consists of analyzing a set the risk that such a bad scenario occurs. BOX 6.1  It is possible to inform decision making, even in a context of deep uncertainty We do not attribute probabilities or likelihood to the future of technologies and most political and our scenarios. These scenarios thus cannot be used socioeconomic trends. In such a context, exploring as forecasts or predictions of the future of poverty scenarios without attributing probabilities to them is or as inputs into a probabilistic cost-benefit a ­ nalysis. commonplace. Since the 1990s, the IPCC and cli- That said, they can still be an important input into mate community have used such long-term socioeco- decision making. Indeed, decisions often are not nomic scenarios—the Special Report on Emissions based on average or expected values or on the most Scenarios (SRES) and now the Shared Socioeco- likely outputs, but instead on the consequences of nomic Pathways (SSPs)—to link policy decisions to relatively low-probability outcomes. For instance, their possible outcomes (Edenhofer and Minx 2014). insurers and reinsurers are often regulated on the Similarly, the U.K. government performs national basis of the 200-year losses (that is, the losses that risk assessments using “reasonable worst case sce- have a 0.5 percent chance of occurring every year). narios” (for example, regarding pandemics, natural And we buy insurance to protect ourselves against disasters, technological accidents, or terrorism), low-probability events that could have a large which are considered plausible enough to deserve impact on our well-being. attention, even though their probability is unknown Moreover, in a situation of deep uncertainty, it is (World Bank 2013, chapter 2). often impossible to attribute probabilities to possible While these scenarios cannot be used to perform outcomes (Kalra et al. 2014). For example, we know a full cost-benefit analysis, they make it possible to that conflicts, such as those in North Africa and the elicit trade-offs and to support decision making. Middle East, could continue over decades, slowing For instance, they help identify dangerous vulnera- down growth and poverty reduction. But they also bilities that can be removed through short-term could subside, allowing for rapid progress. While interventions (Kalra et al. 2014). In our case here, these two scenarios are obviously possible, it is our two scenarios help us explore and quantify how impossible to attribute probabilities to them in any poverty reduction can reduce the vulnerability to reliable way. The same deep uncertainty surrounds ­climate change. 1 8 2    SHOCK WAVES To explore how climate change could Prosperity scenario. This scenario is affect poverty reduction, we create two sce- o ­ ptimistic in that it assumes that the World narios for the future of poverty by 2030, Bank’s twin goals of extreme poverty eradica- in the absence of climate change (figure 6.1; tion and shared prosperity are met by 2030— box 6.2). with less than 3 percent of the world BOX 6.2  Building two scenarios to explore the large uncertainty on the future of poverty To build representative scenarios that are sufficiently L ­ empert et al. 2006; Rozenberg et al. 2014), in which differentiated, we first identify the potential driv- all uncertain parameters are varied systematically ers of future poverty—like demography, structural across the full range of possible values. This enables change, technological change and productivity, and us to generate hundreds of scenarios for the future redistribution—and explore the range of uncertainty socioeconomic development of each of the 92 coun- for each of these drivers. We combine them to create tries. Then, we select two representative scenarios per hundreds of socioeconomic scenarios for 92 coun- country—one optimistic and one pessimistic in terms tries. This analysis combines homogenized household of poverty reduction and changes in inequality—​ surveys (from the I2D2 database) and microsimula- and we aggregate them into two global scenarios tion techniques (Bourguignon, Ferreira, and Lustig labeled “prosperity” and “poverty” (table B6.2.1). 2005; Bussolo, De Hoyos, and Medvedev 2008; To guide the selection of the “prosperity” and Olivieri et al. 2014). “poverty” scenarios in our large set of possible We start from a database of 1.4 million house- futures, we use socioeconomic scenarios developed holds (representing 1.2 billion households and by the scientific community to support climate 4.4 billion people in 92 countries). We transform the change research, the Shared SocioEconomic Path- population structure to account for demographic ways (SSPs). Our prosperity scenario is chosen such changes, and we modify the income of each house- that it is consistent with the 5th SSP (or SSP5; see hold to account for socioeconomic changes, by O’Neill et al. 2013) for population and GDP. SSP5 is 2030. We factor in assumptions on future demo- the scenario with the largest economic growth and a graphic changes (How will fertility or education small population. We also ensure that in this sce- change over time?); structural changes (How fast nario, extreme poverty is below 3 percent of the will developing countries grow their manufacturing global population in 2030. Similarly, we select the sector or shift to services?); technology, productiv- “poverty scenario” using the population and GDP ity, and economic growth (How fast will productiv- pathways of the SSP4, the most pessimistic in terms ity grow in each economic sector? What is the future of poverty and inequalities, and we minimize struc- of technologies and their productivity?); and policies tural change, so that 11 percent of the world popula- (How much redistribution will occur?). tion lives in extreme poverty in 2030. Because of Since evolutions are uncertain, we use a frame- constraints on microsimulations, our scenarios have work inspired by robust decision-making tech- a 2030 time horizon, and we cannot use this tool to niques (Groves and Lempert 2007; Kalra et al. 2014; explore the future after that point. TABLE B6.2.1  Our optimistic and pessimistic scenarios (Population, GDP, and extreme poverty in the 92 modeled countries, in 2030 and in the absence of climate change) Average income per Number of people below poverty line in 2030 in Population (billions) capita (US$ 2005 PPP) the absence of climate change (million people) Prosperity scenario 5.9 4,100 142 (2% of global population) Poverty scenario 6.2 3,700 900 (11% of global population) Source: Rozenberg and Hallegatte, forthcoming. Note: PPP = purchasing power parity. A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   183 FIGURE 6.1  Our model for estimating the number of people in poverty because of climate change (A schematic to represent the modeling undertaken to estimate the impact of climate change on extreme poverty in 2030 under different scenarios of future development, and thus in worlds with different levels of exposure and vulnerability) In the absence of climate change, we can imagine two different ways for the world to evolve Prosperity Poverty More optimistic on: Less optimistic on: • Economic growth • Economic growth • Poverty • Poverty • Inequality • Inequality • Basic services • Basic services With climate change, we can be more or less optimistic on the future magnitude of sectoral impacts Low impact High impact There are uncertainties on the impacts, in the short and the long run. By 2030, di erences in the physics (and biology) of climate change and sectoral adaptation to climate impacts may give us di erent outcomes (e.g., on local rainfall patterns and crop yields). By 2080, the level of emissions, and thus development patterns and climate mitigation polices, also matter. We introduce climate change impacts from the low-impact and high-impact scenarios into each scenario without climate change (Prosperity and Poverty). We model what poverty looks like in each scenario and then compare the difference. What development can achieve: Comparing the effect of low-impact climate change on poverty, in a world that would be more or less prosperous in the absence of climate change Uncertainty from the magnitude of climate change impacts What development can achieve: Comparing the effect of high-impact climate change on poverty, in a world that would be more or less prosperous in the absence of climate change 1 8 4    SHOCK WAVES population living in extreme poverty;1 that The effect of climate change on population growth is slow in developing poverty is a combination of many countries; that education levels and labor pro- sectoral impacts ductivity increase rapidly; and that the pro- In each country and for each of the two ductivity gap between developing and selected socioeconomic scenarios (prosperity developed countries decreases quickly. It also and poverty) we introduce climate change assumes fast globalization and technology impacts on food prices and production, nat- transfers between countries, allowing rapid ural disasters, and health, drawing on the structural changes in developing countries and results from chapters 2–4 (figure 6.1). In the the reduction of the share of unskilled jobs in projections of the 1.4 million households agriculture in favor of the industry and service modeled in our scenarios, we adjust the sectors. Governance is good, and fiscal sys- income and prices to reflect the impact of tems are efficient, allowing for high levels of climate change on their ability to consume, ­ redistribution. Even the most vulnerable pop- and thus derive the impact on poverty ulations have access to universal health cover- (box 6.2). The impacts are estimated using age, water and sanitation, and efficient safety sectoral models (such as crops and agricul- nets. And agricultural workers have enough tural trade models) and include adaptive market power to receive a large share of agri- behaviors (such as changing agricultural cultural price increases if price shocks occur. practices or trade patterns). Poverty scenario. This scenario is pessimis- With a 2030 horizon, impacts barely tic in that it assumes high population growth depend on emissions between 2015 and 2030 in developing countries and more particularly because these affect the magnitude of climate in Africa, low economic growth, and greater change only over the longer term, beyond inequalities between and within countries— 2050. Regardless of socioeconomic trends with 11 percent of the world population liv- and climate policies, the mean temperature ing in extreme poverty. The world is assumed increase between 2015 and 2035 is between to be fragmented, with few technology trans- 0.5 and 1.2°C—depending on the response of fers, low structural change, and in 2030 a sig- the climate system (IPCC 2013). The impacts nificant share of the global population still of such a change in climate are highly uncer- unskilled and working in agriculture. Many tain and depend on how global climate near-poor people remain vulnerable and risk change translates into local changes, on the falling back into poverty if a shock occurs ability of ecosystems to adapt, on the respon- because of low redistribution levels and inex- siveness of physical systems such as glaciers istent or inefficient safety nets. Health care and coastal zones, and on spontaneous adap- and water and sanitation are not accessible to tation in various sectors (such as adoption of all, making the eradication of vectorborne new agricultural practices or improved diseases more difficult. hygiene habits). These two scenarios are counterfactual To account for this uncertainty, we define a reference scenarios, which do not include ­ low-impact and a high-impact scenario that ­ climate change. In a second stage, we add cli- represent the uncertainty on the magnitude of mate change impacts into each of these sce- the physical and biological impacts of climate narios. We do not attribute probabilities or change. For agriculture, for instance, the dif- likelihood to our scenarios because we are not ference between the low-impact and the high- interested in forecasting the future of poverty. impact scenario comes from the uncertainty Instead, we want to explore how the impacts in the global climate system, crop responses, of climate change on poverty are different in and trade models that are used. For health, different development scenarios, with and one difference across low-impact and high- without climate change, to inform decision impact scenarios comes from the uncertainty making on poverty reduction and climate on the additional number of cases of dengue policies (box 6.1). ­ A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   185 and malaria due to climate change and on the effect depends on the balance between changes cost of treatment. in prices and quantities produced. There are several limits to our approach. Using the data from our analysis of food First, we follow a bottom-up approach and prices and production, we change the sum the sectoral impacts, assuming they do income of all workers in the agricultural sec- not interact. We do not focus on the macro- tor, according to the combination of changes economic impact of climate change and its in prices and in the quantities that are pro- effects on overall economic growth—and duced in a region (see Rozenberg and thus on the secondary impact on poverty Hallegatte, forthcoming, for details). We reduction, a limitation considering the evi- also rescale the (real) income of all house- dence that overall growth is a major driver holds, accounting for the change in food of poverty reduction (Dollar, Kleineberg, prices and the share of food in households’ and Kraay 2013; Dollar and Kraay 2002). budget. The impact of the agriculture chan- We do so because previous research suggests nel on poverty depends on the number of that the macroeconomic impact of climate farmers in each country, the income of these change is likely to remain limited by 2030, farmers, and the income of the entire popu- and because we hypothesize that the main lation (which affects the share of food in channel from climate change impacts to pov- consumption). erty is through the direct impacts, which are Our results show that in the high-impact largely invisible in macroeconomic models scenario, the number of people living below (chapter 1). Second, we consider only a sub- the extreme poverty line in 2030 increases by set of impacts, even within our three 67 million people in the poverty scenario sectors—for instance, we do not include the ­ because of climate change impacts on agricul- loss of ecosystem services and the nutritional ture, and by 6.3 million people in the prosper- quality of food. Third, we cannot assess the ity scenario. Thus, on average, the negative poverty impact everywhere. Our household impact of climate change on yields and prices database represents only 83 percent of the outweighs the potential positive impacts on population in the developing world. Some income that will come from higher food highly vulnerable countries (such as small prices. Those numbers are possibly an under- islands) cannot be included in the analysis estimation of actual impacts because both cli- because of data limitations, in spite of the mate scenarios (low and high impact) assume large effects that climate change could have that there is CO2 fertilization. The removal of on their poverty rates. the CO2 fertilization assumption could bring Food prices and food production. Impacts 3 million additional people into poverty in of climate change on agriculture affect poverty the prosperity scenario and 12 million in the in two ways, first through prices and con- poverty scenario. sumption, and second through farmers’ Note that we did not model the impact of incomes (chapter 2). Higher food prices climate change on ecosystem services—even reduce households’ available income—­ though those will likely have a strong impact especially for the poor, who spend a large on poverty—partly because the income share of their income on food products. In our derived from ecosystems represents a small scenarios, the impact depends on the ­ fraction part of the ecosystem’s role, but mostly of food expenditure in total expenditure, and because ecosystem impacts remain impossible this fraction decreases as households get to anticipate. richer. Food price changes also affect farmers’ Natural disasters. We estimate that the incomes. However, this channel is complex number of people who lose income as a since lower yields (which are expected in result of a natural disaster is on average 100 many areas because of climate change) mean million people per year (or 1.4 percent of the that higher food prices do not necessarily world population).2 To account for climate translate into higher farmer revenues: the net change by 2030, we assume that the fraction 1 8 6    SHOCK WAVES of the world population that will be annually between household income and stunting. affected by a disaster rises from an average We find that the prevalence of stunting drops today of around 1.4 percent to 2 percent in for families whose income is above $8,000 per the low-impact case and 3 percent in the year. We calculate the fraction of the stunted high-impact case. This is an increase of 40 to individuals in the families with income below 120 percent, which is in the range reported $8,000, so that stunting prevalence is consis- by Bouwer (2013) and the IPCC (2012 and tent with data for the current situation. Then, 2014) for the expected rise in economic we increase this fraction using projections losses. It means that between 0.6 percent and from Hales et al. (2014) to account for climate 1.6 percent of the world population would change. We assume that stunted individuals be affected by natural disasters because of have lifelong earnings reduced by 5 percent climate change, on top of the reference risk (low-impact scenario) and 15 ­ percent (high- without climate change. Ultimately, these impact scenario), regardless of employment numbers will depend on the effectiveness sector and skill level. and timeliness of adaptation to new climate For malaria, we increase the number of conditions. malaria cases in 2030 in each country fol- In the low-impact case, we assume that lowing Caminade et al. (2014). As with affected people lose 20 percent of their stunting, we calculate the fraction of people annual income if they are poor and 10 per- who are affected by malaria, based on cur- cent if they are nonpoor; in the high-impact rent prevalence, and we vary this fraction case, the losses would be 30 percent for the using estimates of future change due to cli- poor and 15 percent for the nonpoor. 3 mate change in various regions. Then, based These numbers are in line with postdisaster on the literature reviewed in chapter 4, we household surveys, even though much assume that these people are affected higher values are often observed (Patankar, between 0.1 and 2 times per year and lose forthcoming; Patankar and Patwardhan, income through the cost of treatment forthcoming; Noy and Patel 2014; Carter (between $0.7 and $6 per occurrence) and et al. 2007). We also assume that natural lost days of work (by the sick or caregivers, disasters affect income only during one between 1 and 5 days per occurrence). Note year, which is a conservative estimate that that we consider only the monetary expenses is valid for small disasters, but not for large- due to the disease and do not model non- scale events like Typhoon Yolanda in the monetary effects (like the cost of life or loss Philippines or Hurricane Katrina in the in well-being from being sick), which would United States.4 be important in a multidimensional analysis Our results show that for natural disasters, of poverty. in the high-impact scenario, the number of For diarrhea, we start from data on the poor people rises by 5.6 million people in the number of cases per country today, the cost of poverty scenario and by 1.5 million in the treatment (between $2 and $4 per episode), prosperity scenario. and the number of days out of work (between Health and high temperatures. We now 3 and 7 days for the sick and caregivers) include a set of additional impacts of climate (Hutton and Haller 2004). We assume that change on health (malaria, diarrhea, and the prevalence of diarrhea will increase by stunting), based on the literature reviewed in 10 percent by 2030 because of climate change chapter 4. (in all regions), using results from Kolstad For stunting, we include the additional and Johansson (2010). To account for devel- share of children estimated to be stunted opment, we use DHS data to explore the rela- because of climate change in 2030. To factor tionship between household income and in development, we use data from the exposure to diarrhea. We find a threshold at Demographic and Health Surveys (DHS) by $15,600 per year, and we assume that only wealth quintile to explore the relationship households with income below this level are A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   187 affected by impact of climate change on presented in chapter 4), we assume that, in diarrhea. hot countries, people working outside or Further, we assume that fast progress in without air-conditioning will lose between 1 access to water and sanitation in the pros- and 3 percent in labor productivity because perity scenario would halve the number of of this change of climate, compared with a cases, which is consistent with a recent baseline with no climate change. To assess assessment in India (Andres et al. 2014). Of the number of people affected, we estimate course, this assumes that the new water and the shares of people working outside or sanitation infrastructure can continue to without air-conditioning in the two socio- perform well in 2030 and beyond—in other economic scenarios. We find that with high words, that development has been climate climate change impacts, 19 million people informed. For that to occur, the uncertainty would be pushed into poverty in 2030 in the in climate projections would need to be poverty scenario, and 2.7 million people in accounted for in the design phase, as would the prosperity scenario because of the impact the extra funds needed to invest in more of temperature. resilient infrastructure (possibly factoring in Comparing sectoral influences. Which of safety margins and retrofit options) (Kalra these sectors has the greatest impact on pov- et al. 2014). erty in our simulations? As ­ figure 6.2 illus- Our results show that the health impacts trates, agricultural impacts are the chief of climate change are severe: in the high- culprit in all four scenarios (prosperity and impact case, 28 million people would be poverty , combined with high and low pushed into poverty in 2030 in the poverty impacts). Next come health impacts (diar- scenario and 4.1 million people in the pros- rhea, malaria, and stunting), and the labor perity scenario. The impact is smaller than productivity effects of high temperature with that of agriculture for both scenarios but a second-order but significant role. Disasters remains significant. play a limited role, but we have to be careful As for the impact of high temperatures on because only the direct impact of income labor productivity (also based on results losses was accounted for. FIGURE 6.2  Agriculture is the main sectoral factor explaining higher poverty due to climate change (Summary of climate change impacts on the number of people living below the extreme poverty threshold, by source) Prosperity scenario (high impact) Prosperity scenario (low impact) Poverty scenario (high impact) Poverty scenario (low impact) 0 20 40 60 80 100 120 140 Additional people (million) below the extreme poverty threshold by 2030 Agriculture Health Labor productivity Disasters Source: Rozenberg and Hallegatte, forthcoming. 1 8 8    SHOCK WAVES By 2030, climate change is not the • In the prosperity scenario, the increase dominant driver of global poverty in poverty due to a high-impact climate but can have a large impact if change scenario is “only” 16 million development is not rapid, inclusive, people, suggesting that development and and climate informed access to basic services (like water and sanitation) are effective in reducing poor So how do these sectoral results add up in people’s vulnerability to climate change. terms of climate change’s effect on future For the low-impact scenario, the increase poverty trends? We definitely find that a is 3 million people. large effect on poverty is possible, even though our analysis is partial and does not Note that the large range of estimates in include many other possible impacts (for our results—3 to 122 million—may incor- example through tourism, energy prices, for- rectly suggest that we cannot say anything eign direct investment, or remittances) and about the future impact of climate change on looks only at the short term (during which poverty. The reason for this rather wide there will be small changes in climate condi- range is not just scientific uncertainty on cli- tions compared with what unabated climate mate change and its impacts. Instead, it is change could bring over the long term). predominantly policy choices—particularly Indeed, our overall results show that between those concerning development patterns and 3 million (in the prosperity scenario with poverty reduction policies between now low impact) and 122 million (in the poverty and 2030. While emissions-reduction poli- scenario with high impact) additional people cies cannot do much regarding the climate would be in poverty because of climate change that will happen between now change (Table 6.1). and 2030 (since that is mostly the result of past emissions), development choices can • In the poverty scenario, the total number affect what the impact of that climate change of people living below the extreme pov- will be. erty line in 2030 is 1.02 billion people in In the prosperity scenario, the lower the high-impact climate change scenario; impact of climate change on poverty comes this represents an increase of 122 million from a reduced vulnerability of the develop- people compared to a scenario with no ing world to climate change compared to the climate change. For the low-impact sce- poverty scenario. This reduced vulnerability, nario, the increase is 35 million people. in turn, stems from several channels. TABLE 6.1  Climate change can have a large impact on extreme poverty, especially if socioeconomic trends and policies do not support poverty eradication (Poverty headcount in the four scenario types) Climate change scenario No climate change Low-impact scenario High-impact scenario Number of people in Additional number of people in extreme poverty Policy choices extreme poverty because of climate change Prosperity scenario 142 million +3 million +16 million Minimum Maximum Minimum Maximum +3 million +6 million +16 million +25 million Poverty scenario 900 million +35 million +122 million Minimum Maximum Minimum Maximum −25 million +97 million +33 million +165 million Source: Rozenberg and Hallegatte, forthcoming. Note: The main results use the two representative scenarios for prosperity and poverty. The ranges are based on 60 alternative poverty scenarios and 60 alternative prosperity scenarios. For full details, see Rozenberg and Hallegatte, forthcoming. A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   189 • People are richer and fewer households our model shows that almost 90 percent of live with a daily income close to the pov- the uncertainty on poverty impacts arises erty line. Wealthier people are less from the uncertainty on the local agriculture exposed to health shocks (such as stunt- impacts (like how crops respond to higher ing and diarrhea) and are less likely to be temperatures and resulting impact on yields), pushed into poverty when hit by a shock. which is due to the different climate models • The global population is smaller in the used in the agricultural analysis (chapter 2). prosperity scenario in 2030, by 2 ­percent This uncertainty prevents us from provid- globally, 4 percent in the developing ing a precise estimate of the future impacts of world, and 10 to 20 percent in most climate change on poverty, even for a given A frican countries. This difference in ­ socioeconomic development trend. And the population makes it easier for global present analysis underestimates this uncer- food production to meet demand, thereby tainty since many of the least-known impacts mitigating the impact of climate change have been disregarded—such as recent find- on global food prices. The prosperity ings of the impact of climate change on the scenario also assumes more technology nutritional quality of food (Myers et al. transfers to developing countries, which 2014), or the possibility of a more rapid rise further mitigates agricultural losses. in sea level than expected. • There is more structural change (involving Since most of the variation in our estimate shifts from unskilled agricultural jobs to of the climate change impact on poverty arises skilled manufacturing and service jobs), so from the socioeconomic trends and policies, fewer workers are vulnerable to the nega- we explore this variation further and use 60 tive impacts of climate change on yields. In alternative prosperity and 60 alternative pov- the prosperity scenario, a more balanced erty scenarios. These scenarios represent dif- economy and better governance mean ferent world evolutions that achieve similar that farmers capture a larger share of the progress to the two reference scenarios in income benefits from higher food prices. terms of economic growth and poverty reduc- tion. We assess the poverty impacts of climate Up to 2030, climate change remains a sec- change on all 120 scenarios. We find that the ondary driver of global poverty compared to range of possible impacts is extremely large, development: the difference across reference especially in the poverty scenario (table 6.1)— scenarios due to socioeconomic trends and which also features more uncertainty. In the policies (that is, the difference between the poverty scenario, some scenarios (12 out of poverty and prosperity scenarios in the 60) show a decrease in global poverty num- absence of climate change) is almost 800 mil- bers. These are scenarios where climate lion people. This does not mean that climate change impacts remain moderate (low- change impacts are secondary at the local impact), where a large share of the population scale: in some particularly vulnerable places still works in the agricultural sector, and (like small islands or in unlucky locations where farmers benefit the most from higher affected by large disasters), the local impact food prices (assuming a proportional pass- could be massive. through of higher revenues to their incomes). Note that although climate change impacts Our global results in the representative are secondary in our scenarios, they are also prosperity and poverty scenarios also hide highly uncertain. There is a big difference in higher impacts at a finer scale. At the country poverty outcomes when we consider climate and regional level, the hotspots for increased change in the low-impact or high-impact sce- poverty because of climate change are Sub- nario. This occurs because of the large uncer- Saharan Africa and—to a lesser extent—India tainty surrounding the future magnitude of and the rest of South Asia, especially in the physical impacts, largely in agriculture. In poverty scenario (map 6.1). Those countries, fact, a systematic sensitivity analysis based on in Africa in particular, bear a higher burden 1 9 0    SHOCK WAVES MAP 6.1  Sub-Saharan Africa and—to a lesser extent—India and the rest of South Asia are the most vulnerable (Increase in poverty rate due to climate change in the high-impact scenario) Source: Rozenberg and Hallegatte, forthcoming. A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   191 because they have the highest initial number development is rapid, inclusive, and climate of poor people and the steepest projected informed, than in a world where extreme food price increases. poverty would persist without climate change. In almost all countries, the additional Development policies thus appear to be good number of poor people due to climate change adaptation policies, in addition to the more is higher in the poverty scenario than in the targeted sectoral interventions described in prosperity scenario. Two exceptions are previous chapters. Liberia and the Democratic Republic of Congo, for which the number of poor people pushed into poverty because of climate Pro-poor mitigation policies are change is higher in the prosperity scenario needed to reduce the long-term than in the poverty scenario. This is because, in the poverty scenario, 70 percent of the threat of climate change population still lives below the extreme pov- So far, we have looked only at what occurs erty threshold in 2030 even without climate by 2030—a period during which emissions- change. There are fewer people at risk of fall- reduction policies have almost no impact on ing into poverty because most of the popula- the magnitude of climate change (IPCC tion is already poor—a reminder that the 2013). By this time, climate change impacts depth of poverty (not just the poverty head- also remain moderate compared with what count) also matters. is expected in 2050 and beyond. Indeed, the Moreover, our results show that it is not impacts of climate change will grow with its just the extreme poor who are affected. By magnitude, which will continue increasing as 2030, the income of the bottom 40 percent is long as net emissions of carbon dioxide reduced compared to the scenarios without cli- (CO2) are not reduced to zero. mate change by more than 4 percent in many While chapters 2 to 5 propose options to countries in the high-impact climate change reduce these impacts, they also point to the scenario. In most Sub-Saharan African coun- limits of these options. Land use planning tries and Pakistan, the income of the bottom faces difficult political economy obstacles, 40 percent decreases by more than 8 percent in financial constraints make it tough to invest the high-impact climate change scenario. in protection infrastructure, the provision of What messages should we take away from health care in rural areas remains challenging, all of these results? and targeting social assistance after a disaster First, the quantitative impacts of climate and in emergency conditions will always be change on poverty are uncertain, but could difficult. There are clear limits to what adap- be significant, even over the relatively short tation can achieve, and these limits will be term. It is true that our analysis does not tested by climate change. As summarized in cover all climate change impacts (like those IPCC (2014), “without additional mitigation on ecosystem services) or the entire develop- efforts beyond those in place today, and even ing country population (17 percent is left with adaptation, warming by the end of the out), yet we still find that more than 21st century will lead to high to very high risk 100 ­ million people may be pushed into pov- of severe, widespread and irreversible impacts erty because of climate change impacts. globally.” Second, most of the uncertainty surround- Moreover, the long-term impacts of cli- ing these impacts comes from development mate change are highly uncertain. How will choices made between now and 2030, and ecosystems react to rapid changes in temper- can therefore be actively reduced by imple- ature, rainfall, and ocean acidity? How fast menting the right policies. The quantitative will icecaps disappear, raising global sea impacts of climate change on poverty are ­ levels and threatening coastal settlements? much smaller in a world where socioeco- Could more pressure on natural resources nomic trends and policies ensure that trigger more conflicts? Importantly, this 1 9 2    SHOCK WAVES uncertainty is skewed toward catastrophic t ­ ransportation, and other productive activities outcomes: while climate change impacts that support development. At high-income might turn out to be moderate, they could levels (above about $10,000 per capita), econ- escalate to extremely high levels and in that omies and growth diversify away from manu- case—again—poor people would be the facturing, and energy consumption increases most affected (Pindyck 2013; Stern 2013; more slowly with income (Medlock III and Weitzman 2014). Soligo 2001; van Benthem 2015). Thus, uncertainty is not a reason to delay Looking ahead, can this relationship climate change mitigation action. On the con- change through energy leapfrogging? Thanks trary, the need for climate stabilization arises to technological progress, the energy efficiency from a risk management approach that takes of lighting, vehicles, appliances, and industrial into account threats created by long-term processes has improved considerably in the impacts and the fact that GHG emissions lock past decades. This means that, when countries us into irreversible warming. These risks— that are currently less developed reach the that remain impossible to quantify in terms of income per capita levels that today’s devel- consequences or probability—largely explain oped countries had, say in the 1960s, they will why the international community has com- have access to more energy-efficient technol- mitted to the goal of stabilizing global tem- ogy than was available for developed c ­ ountries perature (16th Conference of the Parties of at that time. Will this reduce energy consump- the United Nations Framework Convention tion associated with future development? So on Climate Change, UNFCCC 2010), and far, there is no evidence of leapfrogging: eco- thus to the full decarbonization of the global nomic development has not been less energy economy (Fay et al. 2015; G7 2015). intensive in follower, developing countries than past growth in leader, now-developed countries. Three factors can explain this result Climate stabilization requires (van Benthem 2015). immediate departure from current First, developing countries may not fully development trends adopt available efficient technology because What might a mitigation game plan look their regulations are less stringent, access to like? To begin with, there is agreement that technologies remains limited by trade barriers current development trends are incompatible and skill mismatch, and governments invest- with the internationally agreed climate tar- ing in infrastructure and firms investing in gets (IPCC 2014). Energy consumption, the productive capital face strong constraints in main driver of GHG emissions worldwide, is terms of access to capital and financial mar- expected to increase over time in a develop- kets (Fay et al. 2015; World Bank 2012). As a ment-as-usual scenario—reflecting the huge result, they typically favor technologies with income gaps among regions and countries. lower upfront capital costs, in effect, favoring Typically, very poor, agriculture-focused less energy-efficient technology. countries do not consume a lot of energy. In Second, globalization and outsourcing 2011, the 900 million persons (13 percent of mean that developing countries today are the population) living in the 50 poorest coun- manufacturing not just for themselves, as tries emitted only 0.8 percent of global CO2 developed countries did during their develop- emissions (figure 6.3). Indeed, below about ment, but also for the developed world. Their $5,000 GDP per capita, increases in income economy thus relies relatively more on manu- tend to result in only modest increases in facturing for exports, which tends to increase energy consumption (figure 6.4). But, beyond energy consumption. this threshold, a major factor in development Third, more efficient technology is off- has been industrialization, which comes set by increased use of such technology with a tighter link between GDP growth (Gertler et al. 2013; Gillingham et al. 2013). and energy consumption growth. And energy Developing countries may use more (or larger) is required to fuel hospitals, schools, cars and refrigerators than developed A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   193 c ­evels— ­ ountries in the past at similar income l 2014), or insufficient investment in zero-­ this earlier access of poor people to energy carbon electricity (Lecuyer and Vogt-Schilb services is one of the positive impacts of higher 2014). Without early retirement, the lifetime of energy efficiency, but it results in more energy energy infrastructure that is built now ranges consumption. Energy leapfrogging could occur in the FIGURE 6.3  One billion people living in the poorest countries emit future if (i) developing countries enact policies less than 1 percent of global emissions that favor the adoption of cleaner technolo- (Carbon emissions and population by country) gies (like performance standards on light 35 bulbs, appliances, buildings, or private vehi- cles); (ii) they correct other market or govern- 30 ment failures that prevent technology Cumulative emissions (GtCO2) adoption (like mandating labels that inform 25 on energy consumption, or removing energy subsidies); (iii) manufacturing patterns 20 become more balanced (for instance if devel- 15 oped countries increasingly use robots to manufacture locally); or (iv) technology adop- 10 tion in developing countries saturates at lower levels than in developed countries (car owner- 5 ship may end up being lower in developing 0 countries than in already developed countries 0 1 2 3 4 5 6 7 if developing countries build mass transit–­ Population (billions) oriented cities). But the evidence suggests that energy consumption and related emissions are Source: World Bank calculation based on World Development Indicators data for 2011. Note: This emission Lorenz curve shows cumulative population ranked by income (­horizontal unlikely to decrease by themselves. axis) and cumulative carbon emissions (vertical axis). Each rectangle represents a country. Policies are thus needed to make develop- GtCO2 = ­gigatons of carbon dioxide. ment and climate change stabilization com- patible. Energy consumption and related FIGURE 6.4  Energy consumption is low when GDP per capita is emissions are unlikely to decrease by them- below $5,000, but then increases fast until $10,000 selves, and maintaining global warming (Per capita final energy consumption and GDP, 1960–2006) below 2°C, or even below 3°C, will require reducing emissions to zero by 2100 Canada (­figure 6.5). Modern living standards will United thus need to be supported in a more efficient States Energy consumption per capita (GJ) and radically less carbon-intensive way, and Germany Finland residual emissions offset though natural car- 100 United bon sinks like forests (Fay et al. 2015). Kingdom With this goal in mind, it makes economic sense for all countries to account for the car- Taiwan, China bon constraint, especially in decisions with Malaysia long-term consequences, and to drive their Turkey China development toward efficient patterns (Fay Spain et al. 2015; Vogt-Schilb and Hallegatte 2014; Vietnam Portugal World Bank 2012). If the carbon constraint is Indonesia Korea, Rep. India not accounted for now, development will cre- 10 ate lock-ins into energy- and carbon-intensive 250 2,500 25,000 patterns—such as inefficient urban forms GDP per capita (US$2005 PPP) (Avner, Hallegatte, and Rentschler 2014), Source: van Benthem 2015. insufficient investment in public transport Note: Both axes have a logarithmic scale. GDP = gross domestic product; GJ = gigajoule; (Vogt-Schilb, Hallegatte, and de Gouvello PPP = purchasing power parity. 1 9 4    SHOCK WAVES FIGURE 6.5  Carbon neutrality is needed by 2100 to achieve environmental performance standards, climate goals ­ information labels, financing facilities, and (Global CO2 emissions in the scenarios analyzed by IPCC 2014) land use and urban planning (Fay et al. 2015; NCE 2014; OECD 2015). These packages 140 must be designed in a way that does not 120 threaten the objective of eradicating poverty by 2030. Annual CO2 emissions (GtCO2/yr) 100 80 Climate change mitigation need not slow down poverty alleviation, as 60 long as climate mitigation policies are done right 40 What would such ambitious mitigation 20 policies portend for poverty reduction? They ­ could reduce GDP growth, in turn slowing 0 down poverty reduction. Higher energy prices (due to more expensive low-carbon energy –20 technologies) could reduce poor p ­ eople’s 2000 2020 2040 2060 2080 2100 Year consumption, as would higher food prices ­ (due to land use for bioenergy or ­ carbon Baseline 3°C 2°C sequestration). Source: Adapted from IPCC 2014. However, reviews of modeling exercises Note: GtCO2 = ­gigatons of carbon dioxide. suggest that mitigation policies would not lead to large losses in this area, even without con- from 20 to 60 years, in effect creating an sidering benefits from lower climate change “­emission commitment” (Davis, Caldeira, and impacts and cobenefits. The IPCC (2014) esti- Matthews 2010; Davis and Socolow 2014; mates that mitigation policies would reduce Guivarch and Hallegatte 2011). This commit- global consumption by 1–4 ­ percent in 2030 ment is rapidly increasing today, especially and 3–11 percent in 2100 relative to an because much coal-related infrastructure con- expected consumption growth of more than tinues to be built across the world ­ (figure 6.6) 300 percent in all s­ cenarios. But these limited (Steckel, Edenhofer, and Jakob 2015). costs at the global scale remain uncertain and These carbon-intensive patterns would be heatedly debated; models still neglect many costly—or sometimes impossible—to reverse mechanisms that could magnify these losses, later on, which would impair an efficient such as imperfections in labor markets. More transition toward a zero-carbon economy and important, global estimates hide large impacts make it much more expensive and politically on certain countries or sectors. difficult (Rozenberg, Vogt-Schilb, and Even so, policy makers can design climate- Hallegatte 2014). Thus, it is urgent that all mitigation policies that do not threaten pov- countries—especially developing ones that are erty eradication. This can be done in three building their infrastructure stocks at present—​ ways: (i) building on no-regret options and take steps to redirect investment in long-lived focusing on local and immediate cobenefits; capital and infrastructure toward low- or (ii) protecting the poor and vulnerable popu- zero-emission alternatives. lations against adverse consequences of costly To achieve an efficient decarbonization of emissions reduction options; and (iii) in the the world economy, all countries must work poorest countries, using support from the on enacting comprehensive packages of miti- international community to offset potential gation policies (IPCC 2014)—ranging from trade-offs between poverty reduction and cli- carbon pricing and innovation support to mate change mitigation—especially for the A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   195 options that involve immediate costs but are FIGURE 6.6  Investments in coal-related infrastructure have urgently needed to prevent irreversibility and created large emission commitments (Total emissions committed by existing infrastructure, for oil, gas, and coal) lock-ins into carbon-intensive patterns (like those regarding deforestation or urban transport). 300 Climate change mitigation offers ­ cost-effective opportunities and cobenefits, 250 especially for poor people. Many climate miti- gation policies are consistent with development Total remaining commitments objectives and contribute to higher productiv- 200 ity and efficiency (Fay et al. 2015; World Bank 2012, 2014a). In other words, sometimes the (GtCO2) most effective development options also reduce 150 GHGs emissions (or increase them in a negli- gible manner; see box 6.3). 100 For instance, using modern, e­ nergy-efficient technologies for lightning and transportation can help provide cheap energy services at a 50 low environmental and economic cost. The Global Fuel Economy Initiative’s (GFEI) goal of doubling the efficiency of the global 0 fleet of cars (from 8 to 4 liters per 100 km) 1950 1960 1970 1980 1990 2000 2012 would result in savings in annual oil import Year bills alone worth over $300 billion in 2025 Coal Gas Oil and $600 billion in 2050 (based on an oil Source: Davis and Socolow 2014. price of $100/barrel). According to the United Note: GtCO2 = gigatons of carbon dioxide. Nations Environment Programme (UNEP)/ Global Environment Facility (GEF) en.lighten many regions, especially before 2030 (Shindell initiative, eliminating inefficient lighting by et al. 2012; Shindell 2015; Thompson et al. 2030 would save about 1,000 ­ terawatt-hour 2014; West et al. 2013). A pathway leading to (TWh)/year in electricity consumption and a reduction in CO2 concentrations from 720 more than $100 billion in electricity bills. to 525 parts per million (ppm) in 2100 would Also, renewable energy can meet the avoid 0.5 million premature deaths annually needs of poor households at competitive in 2030, 1.3 million in 2050, and 2.2 million prices, especially in remote rural areas where in 2100, compared to a scenario with only the grid development and centralized production progress that can be expected from the histori- would be expensive (Deichmann et al. 2011). cally observed uptake of pollution-control U n d e r t h e Wo r l d B a n k – m a n a g e d technologies (­figure 6.7a). Community Development Carbon Fund In places where air pollution has reached Nepal Micro-Hydro Promotion project, 426 alarming levels in the past decade, health community-run micro-hydropower plants cobenefits can be particularly large (Matus were installed, benefitting 625,000 people et al. 2012). In China, air pollution is esti- and avoiding the emissions of about 66,000 mated to result in 7.4 times more premature tons of CO2 per year. deaths than in the European Union (EU) In addition, climate mitigation efforts can (Watts et al. 2015), and the estimated cost of lower local air pollution, thereby providing ambient air pollution in terms of morbidity massive health benefits and higher agricultural and mortality is around $1.9 trillion annually yields. Recent studies have found that the ben- in China and India alone (OECD 2014a). In efits from lower air pollution alone could East Asia, about 500,000 premature deaths more than offset the cost of mitigation in would be avoided annually in 2050 under 1 9 6    SHOCK WAVES FIGURE 6.7  Lower air pollution means lower mortality rates report reviews market and government fail- (Changes in mortality in baseline and mitigation scenarios, from small particulate ures that hamper their adoption—including matter (PM2.5) at the global scale and in East Asia) incorrect pricing, split incentives, poor enforcement of existing regulations, lack of a. Global b. East Asia information, behavioral failures, and limits to 500 1 the financing capacity of stakeholders. It also Mortality (millions per year) 0 Mortality (1,000s per year) 0 –1 –2 –500 proposes available solutions to overcome –3 –1,000 them, like information labels or performance –4 –5 –1,500 standards (Fay et al. 2015). –6 –7 –2,000 Capturing these opportunities should be a –8 –9 –2,500 priority for all countries at all income levels. –10 –3,000 The international community can help, by 00 20 40 60 80 00 00 20 40 60 80 00 providing a combination of technical assis- 20 20 20 20 20 21 20 20 20 20 20 21 Year Year tance and better access to green technologies. Baseline RCP4.5 For instance, the UNEP/GEF en.lighten ini- Source: Adapted from West et al. 2013. tiative supports countries to implement mea- Note: Baseline is a reference scenario leading to CO2 concentration around 720 parts per million sures to reduce inefficient lighting. The GFEI (ppm) in 2100, which assumes an uptake of pollution control technology consistent with historical ­experience (Smith, West, and Kyle 2011). RCP4.5 is a scenario with mitigation policies that would builds administrative and technical capacity lead to 525 ppm in 2100. (More ambitious decarbonization would tend to increase cobenefits). The in developing countries, with the final objec- ­difference between Baseline and RCP4.5 can be attributed uniquely to GHG emissions reductions. tive of helping them implement the policies that will double the energy efficiency of the climate mitigation (figure 6.7b). In India, if global private car fleet. health benefits from lower PM2.5 emissions The international community—and (a decrease of 50 percent by 2050 in tiny par- high-income countries—can also help fund ­ ticulate matter) were valued similarly to the innovation to come up with the solutions that approach used in the EU for air pollution, developing countries need (like improved they would offset the cost of emission reduc- building design and materials for tropical tions in full (Markandya et al. 2009; Watts et climates). al. 2015). Public transportation is an example Multilateral or bilateral development of a measure that can reduce local pollut- banks (MDBs) can provide advisory services ants—in addition to transport costs, conges- to help countries develop strong capital mar- tion, and GHG emissions. kets and channel official development assis- Another area where cobenefits can be gen- tance. This is particularly important for erated using mitigation policies is land-based infrastructure: although MDBs’ financial mitigation and policies and payment for eco- resources are small relative to the need— system services. These schemes require devel- MDB lending for infrastructure were about oping effective institutions (like land tenure) $90 billion in 2011, whereas $1 trillion per and enforcement capacity, and they need to year would be needed to close the infrastruc- be designed explicitly to support poverty ture gap in developing countries—they often reduction. If carbon-related payments were fund a substantial share of infrastructure fully developed and pro-poor participation investments in the poorest countries. MDBs conditions secured, an estimated 25–50 mil- can have a significant impact if they are lever- lion low-income households could benefit aged to make emissions-­ reduction invest- from them by 2030 (Milder, Scherr, and ments more attractive to the private sector Bracer 2010). And climate-friendly landscape (for instance, by derisking projects with guar- management can be more productive and antees and blended financial instruments). more resilient to climate shocks (chapter 2). Financial tools for the private sector are In most cases, governments need to enact also important. The International Finance policies to actively promote the adoption of Corporation recently provided a $30 million such no-regret options. A recent World Bank loan to the responsAbility Energy Access A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   197 Fund, which finances manufacturers and dis- Fortunately, studies suggest that countries tributors of affordable solar-powered devices. can reduce their GHG emissions while ensur- These devices give underserved people access ing universal access to modern cooking (for to LED-based lighting and power for charg- example, Pachauri et al. 2013). One approach ing cell phones or small appliances, thereby is to sequence fossil fuel subsidy removal, supporting economic activity and better removing subsidies on LPG later on. Similarly, livelihoods. carbon taxes can be combined with policies To protect poor and vulnerable people, that help the shift to modern energy, such as climate mitigation policies can be combined ­ low-cost financing for clean cookstove pur- with complementary policies, including chase or temporary subsidies for modern social protection. To stay on a pathway energy. If well targeted, policies that support ­ compatible with zero net emissions before LPG use would have a negligible impact on 2100, countries will have to do more than GHG emissions (Pachauri et al. 2013). More implement win-win options, potentially creat- generally, providing universal access to basic ing costs and trade-offs with poverty reduc- services would have no significant impact on tion. For instance, a key concern is that carbon global emissions—even using current technol- taxes or fossil fuel subsidy removals can jeop- ogies (box 6.3). ardize the switch from traditional biomass In Peru, the Fondo de Inclusion Social (which would not be impacted by higher Energetico mails LPG vouchers to poor energy or carbon prices) to modern cooking households with their electricity bill (targeting fuels, such as electricity or liquefied petroleum households who own an LPG cookstove and gas (LPG) (which would become more expen- consume less than a given threshold of elec- sive). This matters greatly because traditional tricity per month). Under India’s Direct cooking fuels are unhealthy and worsen gen- Benefits Transfer for LPG program, cash der imbalances and educational opportunities, transfers are credited directly to the bank given the time women and children must accounts of LPG consumers (this is done spend to collect them (WHO 2006). instead of reducing the market price of LPG BOX 6.3  Is there a trade-off between climate mitigation and reducing extreme poverty? Many recent studies support the idea that providing However, these studies rely on a very restric- those who are currently extremely poor with access to tive definition of access to basic services—one basic services would not jeopardize climate mitigation. that remains far below what is considered accept- • Above a human development index (HDI) of able in developed countries. For instance, for 0.8 (the UN threshold to be considered a devel- access to electricity, the IEA uses two threshold oped country), carbon emissions and the HDI are levels of consumption: 250 kilowatt-hours decoupled (Steinberger and Roberts 2010). (kWh) per year for rural households and 500 • The International Energy Agency (IEA) estimates kWh per year for urban households. In rural that universal access to basic energy services by areas, this is sufficient to use a floor fan, a mobile 2030 could be achieved by increasing electricity telephone, and two compact fluorescent light consumption by 2.5 percent, and fossil fuel con- bulbs for about five hours per day. In urban sumption by only 0.8 percent (IEA 2011). areas, it can include an efficient refrigerator, a • The World Development Report 2010 estimates that the additional emissions needed to provide second mobile telephone per household, and universal access to electricity in 2010 could be another appliance (such as a small television or a offset by a switch of the U.S. vehicle fleet to Euro- computer). Even middle-class living conditions pean standards (World Bank 2010). imply a much higher level of consumption box continues next page 1 9 8    SHOCK WAVES BOX 6.3 (continued) (ESMAP and SE4ALL 2015). And, historically, affluence and modern living standards with cur- there has been a strong relationship between rent development patterns and technologies energy consumption and GDP—even though would result in much higher energy consumption there are decreasing returns on how much energy and GHG emissions (Rao, Riahi, and Grubler consumption helps increase life expectancy and 2014). This is why immediate action is needed in basic needs (Figure B6.3.1). all countries to achieve affluence and shared So eradicating extreme poverty can be done at prosperity while decarbonizing the global econ- low energy consumption levels, but generalizing omy by the end of the century. FIGURE B6.3.1  Energy use keeps rising with GDP even though less energy might be enough for basic human needs 100 Basic needs access (%) 90 80 70 60 50 40 30 0 50 100 150 200 250 300 90 Life expectancy (years) 80 70 60 50 40 0 50 100 150 200 250 300 GDP per capita (2005 US$ PPP) 100,000 10,000 1,000 100 1 10 100 1000 Total final energy consumption (GJ/capita) Source: Lamb and Rao 2015. Note: GDP = gross domestic product; GJ = gigajoule; PPP = purchasing power parity. A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   199 cylinders, to reduce fraud). Sometimes the tax (or fossil fuel subsidy removal) coupled modern cook fuel is electricity, and govern- with targeted or untargeted cash transfers, ments want to phase out LPG (this will likely achieves two different objectives: reducing be increasingly important as governments GHGs and improving income distribution engage on the path to zero net GHG emis- (Klenert et al. 2015). sions). Ecuador is considering how to remove All the revenues from carbon prices or fos- LPG subsidies to reduce LPG imports and sil fuel subsidy reform cannot always be used GHG emissions without hurting poor house- for direct redistribution to households, but holds. The idea is to facilitate the switch to that does not necessarily threaten the positive electric cookstoves before LPG subsidies are distributive impact. In British Columbia, reve- removed by providing financing options when nues from the carbon tax are used to cut taxes buying the stoves and temporarily subsidizing on both labor and capital, and the scheme is electricity (which comes mainly from local still progressive overall (Beck et al. 2015). hydropower). Poverty benefits can be further increased if However, fuel subsidies introduce risks of revenues are used for more targeted instru- subsidy diversion, smuggling, and fraud ments that help poor people (like targeted (Barnwal 2014; Cunha, Trezzi, and Calvo- cash transfers), or for better social safety nets Gonzalez 2015). In Ghana, illegal diversion (like school feeding). Based on current CO2 of a heavily subsidized fuel for fishing, “pre- emissions and without any international mix,” has been problematic for decades. For transfer, a $30/tCO2 (ton of carbon dioxide) many years, the price of gasoline has been domestic carbon tax would raise resources double—and between 2011 and 2013 even amounting to more than 1.5 percent of local triple or quadruple—the price of premix fuel, GDP in half of the 87 countries (both devel- creating enormous scope for commercial mal- oped and developing) where data are avail- practice and illegal gains (Kojima 2013). This able (figure 6.9, panel a). Remember from is why more countries are now turning chapter 5 that, in Sahel countries, 1.5 percent toward cash transfers to compensate poor of GDP is more than the amount needed to people and protect them against higher protect households affected by severe energy prices. droughts. And in 60 out of the 87 countries, a Indeed, the best strategy may be to imple- $30/tCO2 domestic tax would provide the ment carbon prices or remove fossil fuel sub- resources to more than double current levels sidies, while recycling revenues through cash transfers or programs that help the poor (OECD 2014b; Vagliasindi 2012). When car- FIGURE 6.8  Recycling $100 from the global fossil bon revenues or savings from fossil fuel sub- fuel subsidy budget as a universal cash transfer sidy removal are recycled in lump-sum cash would benefit poor people transfers to the population, the overall impact is to improve equity (Bento et al. 2009; Callan et al. 2009; Cohen, Fullerton, and Topel $13 Variation in annual income 2013). That result directly follows from the $9 fact that poor households consume less $4 energy, in absolute amounts, than nonpoor households. Data from developing countries –$3 suggest that taking $100 away from fossil fuel subsidies and redistributing the money equally throughout the population would on –$23 average transfer $13 to the bottom quintile and take $23 away from the top quintile Bottom Second Third Fourth Top quintile quintile quintile quintile quintile (Arze del Granado, Coady, and Gillinghan 2012) (figure 6.8). In other words, a carbon Source: Based on Arze del Granado, Coady, and Gillingham 2012. 2 0 0    SHOCK WAVES of social assistance in the country (figure 6.9, they tend to own more houses or land where panel b). Even a low carbon tax at $10/tCO2 photovoltaic panels can be installed, and can would make it possible to significantly scale better afford the high upfront cost of install- up social assistance, or other investments that ing panels. In contrast, everyone pays higher benefit poor people, such as connection to electricity tariffs to finance the scheme. This sanitation and improved drinking water or problem applies to other subsidies to encour- access to modern energy. Brazil, the age low-carbon investment—such as hybrid Dominican Republic, Indonesia, and Mexico or electric vehicles, residential heating, or provide examples where well-functioning air-conditioning—​ that are more likely to be cash-transfer programs have been used to undertaken by wealthy households protect basic consumption by the poor from (Borenstein and Davis 2015). But solutions price increases resulting from subsidy have been proposed, such as financing subsi- ­ removals (Beaton and Lonton 2010; Di Bella dies with progressive income taxes or specifi- et al. 2015; Vagliasindi 2012). cally encouraging poor households to Other emissions-reduction policies can participate (CPUC 2013; Granqvist and also have significant distributional impacts Grover 2015; Macintosh and Wilkinson that need to be explored before policies are 2010). implemented (Fay et al. 2015). For instance, More generally, the distributional impacts it has been shown that feed-in tariffs for of climate mitigation policies can in principle renewable energy in the United Kingdom be corrected using independent policies spe- and Germany are slightly regressive (Grösche cifically designed to redistribute income in the and Schröder 2014; Grover 2013). Wealthy economy, such as using income or consump- households benefit from the scheme, because tion taxes to fund cash transfers or social FIGURE 6.9  Using the revenue from a carbon tax could boost social assistance (Potential carbon revenue as a fraction of GDP and compared to current social assistance benefits) a. Revenue as a fraction of GDP b. Revenue compared to social assistance benefits 14 14 12 12 Carbon revenue (% GDP) Carbon revenue (% GDP) 10 10 8 8 6 6 4 4 2 2 0 0 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 0 2 4 6 8 10 12 14 GDP per capita (US$, PPP 2011) Average social assistance received (% GDP) Source: World Bank calculations, using data from World Development Indicators and ASPIRE. Note: Panel a: revenue of a $30/tCO2 carbon price expressed as a fraction of GDP. Each dot represents a country. Panel b: How this revenue compares to current social assistance benefits in the countries. In 60 out of 87 countries for which data are available, a $30/tCO2 tax would provide the resources to more than double current social assistance transfers (dots above the diagonal line on the right panel). Calculations assume unchanged energy consumption. GDP = gross domestic product; tCO2 = tons of carbon dioxide. PPP = purchasing power parity. A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   201 safety net programs (Borenstein and Davis regarding urban transport, energy infrastruc- 2015; Gahvari and Mattos 2007; Lindert, ture, or deforestation). Skoufias, and Shapiro 2006). A study based The typical example is urban transit. While on World Bank household surveys reveals transit-oriented development may require that most countries where the GDP per capita higher upfront costs and investments than is above $4,000 (in purchasing power parity) road-based low-density urbanization, the can reliably redistribute poverty away using high urbanization rate in many developing their own internal resources (Ravallion 2010), countries and the lifetime of urban forms and and thus can protect poor people against the transit infrastructure means that there is a potential negative effects of climate mitiga- window of opportunity now to build efficient tion policies. This is important because transit-oriented cities. After a city is devel- around 70 percent of people in extreme pov- oped, it is practically impossible to modify its erty live in countries with a GDP per capita urban form. This makes it essential to provide above $4,000, where they could be protected developing countries with the resources and by redistribution from possible negative financial instruments that make it possible for effects of climate mitigation. them to drive urban development toward the The international community has a criti- efficient patterns that are needed to decarbon- cal role to play in helping reconcile immedi- ize the economy before the end of the ate poverty-reduction objectives and climate century. stabilization. In very poor countries, how- One source of international funding is pri- ever, it may be difficult for economic, politi- vate climate finance, for instance through cal, or institutional reasons to protect poor interconnected carbon markets (World Bank people against possible negative side effects of 2014b). But these flows are likely to focus on climate policies. In particular, countries with the cheapest emissions-reduction options a GDP per capita below $4,000 per capita (in available in developing countries (Narain and purchasing power parity) cannot always rely Veld 2008; Rose, Bulte, and Folmer 1999). on internal redistribution (Ravallion 2010). Indeed, carbon markets are designed to help In the poorest countries, even the “middle economic actors capture the lowest cost class” is poor, and there are simply not options to meet a short-term emissions-​ enough resources for redistribution: even tax- ­ reduction target, not necessarily to trigger ing 100 percent of the income of the “rich” investment in long-lived low-carbon equip- would not suffice to lift the poorest out of ment that avoid lock-ins into carbon intensive poverty. In these very poor countries, even if patterns (Vogt-Schilb and Hallegatte 2014). most of the cost of climate mitigation is paid For example, these flows alone are unlikely to by the upper quintiles of the population, cli- finance the upfront cost of more efficient cit- mate mitigation could still aggravate poverty, ies and land use planning, or any other mea- because the top quintiles are still in or close to sure that generates benefits only over the very poverty. long term. In countries where poor people cannot eas- Thus, additional resources are needed that ily be protected by domestic resources and focus on these long-term challenges. In par- policies, support from the international ticular, they can substantially increase the effi- c ommunity is needed to offset potential ­ ciency of the global decarbonization by trade-offs between poverty reduction and cli- ­ financing urgent measures that avoid carbon- mate change mitigation. This is especially the intensive lock-ins in low-income countries case for investments that involve high imme- (like public transportation infrastructure)— diate costs—and therefore large trade-offs even if these measures are more expensive with other investments—but are urgently (per abated ton of carbon) than alternative needed to prevent irreversibility and lock-ins short-term emissions reductions (Vogt-Schilb, into carbon-intensive patterns (like those Hallegatte, and de Gouvello 2014). 2 0 2    SHOCK WAVES In conclusion that countries need to act now to reduce their emissions, using two approaches: This report provides new quantification of how climate change will affect poor people • Focus on emissions-reduction options that and poverty through agricultural impacts create synergies with development or yield (chapter 2), natural disasters (chapter 3), and health or economic cobenefits—like using health shocks (chapter 4). In each of these renewable power and minigrids in remote chapters, it also identifies opportunities for rural areas, or switching to energy-­ better policies or specific interventions that efficient light bulbs and appliances. can reduce these impacts, sometimes even • Protect poor people—for instance by below their current levels in spite of climate strengthening social protection and cash change. Chapter 5 builds on these sectoral transfers, possibly financed with energy solutions by exploring cross-cutting options taxes or fossil fuel subsidy removal. to enhance resilience (like financial inclusion But the second approach will be particu- and social safety nets). It also identifies larly challenging for low-income countries, options to adapt to a context of changing cli- because they sometimes lack the capacity or mate with more frequent and intense shocks simply the resources to implement substantial and changing environmental conditions, like redistribution policies. The international com- permanently reduced rainfall. And it stresses munity should support costly emissions the need for a governance system that gives a reduction in these countries, especially invest- voice to poor people. ing in long-lived low-carbon infrastructure This report suggests that developing coun- (like urban public transit in cities), because tries have a window of opportunity to build waiting will only make low-carbon develop- resilience and reduce short-term climate ment more expensive over the long term. change impacts on poverty through develop- Bringing together the short- and long-run ment policies that are inclusive and climate view, this report overall emphasizes the nega- informed. For governments, two implications tive impact of climate change on poverty emerge: eradication, and the risk that unabated cli- • Greater urgency in reducing poverty and mate change creates for the objective of erad- providing poor people with opportunities, icating extreme poverty. In parallel, it also basic services, and well-designed social identifies many policy options that can be safety nets to reduce their vulnerability implemented and would make it possible to before climate change impacts become achieve our poverty objectives in spite of cli- much larger. mate change. Doing so implies a combination • The critical importance of ensuring that of (i) rapid, inclusive, and climate-informed investments and development patterns development and targeted adaptation inter- are not creating future vulnerabilities as ventions to cope with the short-term impacts environmental and climate conditions of climate change; and (ii) pro-poor mitiga- change. tion policies to limit long-term impacts and create an environment that allows for global In parallel, the international community prosperity and the sustainable eradication can do much to ensure that development is of poverty. rapid, inclusive, and climate informed. It can offer resources for climate risk analysis and project preparation; and it can ensure that Notes financing instruments and support are avail- 1. These simulations are performed using 2005 able to cover higher upfront costs. PPP exchange rate and the $1.25 extreme However, in the absence of mitigation poli- poverty line, but results are not expected to cies, risks for development and poverty eradi- change significantly under the $1.90 poverty cation will only grow over time. This means line and using 2011 PPP. A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   203 2. We cannot use data on the number of Distribution Dynamics in East Asia and Latin “affected persons” because the usual defini- America. Washington, DC: World Bank. tion of affected is much broader and includes Bouwer, L. M. 2013. “Projections of Future people who do not lose income because of Extreme Weather Losses under Changes in the disasters. Climate and Exposure.” Risk Anal . 33: 3. Because the analysis proposed in chapter 3 915–30. does not include all countries, we assume Bussolo, M., R. De Hoyos, D. Medvedev. 2008. that poor and nonpoor people are equally “Economic Growth and Income Distribution: exposed to natural disasters. Linking Macroeconomic Models with 4. Note that the impact of droughts on chil- Household Survey Data at the Global Level.” dren through stunting is accounted for in the In International Association for Research in health impact category. Income and Wealth (IARIW) 30th General Conference , Portoroz, Slovenia, August , 24–30. Callan, T., S. Lyons, S. Scott, R. S. J. Tol, and References S .   Ve r d e . 2 0 0 9 . “ T h e D i s t r i b u t i o n a l Andres, L., B. Briceño, C. Chase, and Implications of a Carbon Tax in Ireland.” J. A. Echenique. 2014. “Sanitation and Energy Policy 37: 407–12. Externalities: Evidence from Early Childhood C a m i n a d e , C . , S . K o v a t s , J . R o c k l o v, Health in Rural India.” Policy Res. Work. Pap. A. M. Tompkins, A. P. Morse, F. J. Colón- 6737, World Bank, Washington, DC. González, H. Stenlund, P. Martens, and Arze del Granado, F.J., D. Coady, and S. J. Lloyd. 2014. “Impact of Climate Change R. Gillingham. 2012. “The Unequal Benefits of on Global Malaria Distribution.” Proc. Natl. Fuel Subsidies: A Review of Evidence for Acad. Sci. 111: 3,286–91. Developing Countries.” World Dev . 40: Carter, M. R., P. D. Little, T. Mogues, and 2,234–48. W. Negatu. 2007. “Poverty Traps and Natural Avner, P., S. Hallegatte, and J. Rentschler. 2014. Disasters in Ethiopia and Honduras.” World “Carbon Price Efficiency: Lock-In and Path Dev. 35: 835–56. Dependence in Urban Forms and Transport Cohen, M. A., D. Fullerton, and R. H. Topel. Infrastructure.” Policy Res. Work. Pap. No 2013. Distributional Aspects of Energy and 6941, World Bank, Washington, DC. Climate Policies. Cheltenham, U.K.: Edward Barnwal, P. 2014. “Curbing Leakage in Public Elgar Pub. Programs with Biometric Identification Systems: CPUC (California Public Utilities Commission). Evidence from India’s Fuel Subsidies.” Working 2013. California Solar Initiative Program Paper. Handbook. San Francisco: California Public Beaton, C., and L. Lonton. 2010. “Lessons Utilities Commission. Learned from Indonesia’s Attempts to Reform Cunha, B., R. Trezzi, and O. Calvo-Gonzalez. Fossil-Fuel Subsidies.” SSRN Scholarly Paper 2015. “When Winners Feel Like Losers: 1702880, Social Science Research Network, Evidence from an Energy Subsidy Reform.” Rochester, NY. Policy Res. Work. Paper 7265, World Bank, Beck, M., N. Rivers, R. Wigle, and H. Yonezawa. Washington, DC. 2015. “Carbon Tax and Revenue Recycling: Davis, S. J., K. Caldeira, and H. D. Matthews. Impacts on Households in British Columbia.” 2010. “Future CO2 Emissions and Climate Resour. Energy Econ. 41. Change from Existing Energy Infrastructure.” Bento, A. M., L. H. Goulder, M. R. Jacobsen, and Science 329: 1,330–33. R. H. von Haefen. 2009. “Distributional and Davis, S. J., and R. H. Socolow. 2014. Efficiency Impacts of Increased U.S. Gasoline “Commitment Accounting of CO2 Emissions.” Taxes.” Am. Econ. Rev. 99: 667–99. Environ. Res. Lett. 9, 084018. Borenstein, S., and L. W. Davis. 2015. “The Deichmann, U., C. Meisner, S. Murray, and Distributional Effects of U.S. Clean Energy Tax D. Wheeler. 2011. “The Economics of Renewable Credits.” Working Paper 21437, National Energy Expansion in Rural Sub-Saharan Africa.” Bureau of Economic Research. Energy Policy 39: 215–27. Bourguignon, F., F. H. Ferreira, and N. Lustig. Di Bella, G., L. Norton, J. Ntamatungiro, 2005. The Microeconomics of Income S. Ogawa, I. Samaké, and M. Santoro. 2015. 2 0 4    SHOCK WAVES “Energy Subsidies in Latin America and the Guivarch, C., and S. Hallegatte. 2011. “Existing Caribbean: Stocktaking and Policy Challenges.” Infrastructure and the 2°C Target.” Clim. SSRN Working Paper, Social Science Research Change 109: 801–05. Network, Rochester, NY. Hales, S., S. Kovats, S. Lloyd, and D. Campbell- Dollar, D., T. Kleineberg, and A. Kraay. 2013. Lendrum. 2014. Quantitative Risk Assessment “Growth Still Is Good for the Poor.” Policy of the Effects of Climate Change on Selected Res. Work. Pap. 6568, World Bank, Causes of Death, 2030s and 2050s. World Washington, DC. Health Organization. Dollar, D., and A. Kraay. 2002. “Growth Is Good Hutton, G., and L. Haller. 2004. Evaluation of the for the Poor.” J. Econ. Growth 7: 195–225. Costs and Benefits of Water and Sanitation Edenhofer, O., and J. Minx. 2014. “Mapmakers Improvements at the Global Level. Water, and Navigators, Facts and Values.” Science Sanitation, and Health, Protection of the 345: 37–38. H u m a n E n v i r o n m e n t . Wo r l d H e a l t h ESMAP (Energy Sector Management Assistance Organization. Program) and SE4ALL (Sustainable Energy 4 IEA (International Energy Agency). 2011. World All). 2015. Beyond Connections—Energy Energy Outlook 2011: Energy for All - Access Redefined. Washington, DC and Vienna, Financing Access for the Poor. Paris, France: Austria: Energy Sector Management Assistance International Energy Agency. Program and Sustainable Energy 4 All. IPCC (Intergovernmental Panel on Climate Fay, M., S. Hallegatte, A. Vogt-Schilb, Change). 2012. Managing the Risks of Extreme J. Rozenberg, U. Narloch, and T. Kerr. 2015. Events and Disasters to Advance Climate Decarbonizing Development: Three Steps to a Change Adaptation: Special Report of the Zero-Carbon Future. Washington, DC: World Intergovernmental Panel on Climate Change. Bank. Cambridge, U.K. and New York: Cambridge G7. 2015. Leaders’ Declaration G7 Summit University Press. Germany. WWW Documentan. URL http:// ———. 2013. “Summary for Policymakers.” In www.international.gc.ca/g8/g7_germany​ Climate Change 2013: The Physical Science _­d eclaration-g7_allemagne_declaration​ Basis, edited by T. F. Stocker, D. Qin, G.-K. .aspx?lang=eng (accessed 6.14.15). Plattner, M. Tignor, S. K. Allen, J. Boschung, Gahvari, F., and E. Mattos. 2007. “Conditional A. Nauels, Y. Xia, V. Bex, and P. M. Midgley, Cash Transfers, Public Provision of Private 1–30. Contribution of Working Group I to the Goods, and Income Redistribution.” Am. Econ. Fifth Assessment Report of the Rev. 97: 491–502. Intergovernmental Panel on Climate Change. Gertler, P., O. Shelef, C. Wolfram, and A. Fuchs. Cambridge, U.K. and New York: Cambridge 2013. “How Pro-Poor Growth Affects the University Press. Demand for Energy.” Working Paper 19092, ———. 2014. Climate Change 2014: Synthesis National Bureau of Economic Research. Report. Contribution of Working Groups I, II Gillingham, K., M. J. Kotchen, D. S. Rapson, and and III to the Fifth Assessment Report of the G. Wagner. 2013. “Energy Policy: The Intergovernmental Panel on Climate Change Rebound Effect Is Overplayed.” Nature 493: [Core Writing Team, R.K. Pachauri and 475–76. L.A. Meyer (eds.)]. Geneva, Switzerland: Granqvist, H., and D. Grover. 2015. “Who Should IPCC. Pay for Clean Energy? Distributive Justice Kalra, N., S. Gill, S. Hallegatte, C. Brown, Perspectives.” Working Paper. A. Fozzard, R. Lempert, and A. Shah. 2014. Grösche, P., and C. Schröder. 2014. “On the “Agreeing on Robust Decisions: New Processes Redistributive Effects of Germany’s Feed-In for Decision Making under Deep Uncertainty.” Tariff.” Empir. Econ. 46: 1,339–83. Policy Res. Work. Pap. 6906, World Bank, Grover, D. 2013. The British Feed-in Tariff for Washington, DC. Small Renewable Energy Systems: Can It Be Klenert, D., G. Schwerhoff, O. Edenhofer, and Made Fairer? London, U.K.: Grantham L. Mattauch. 2015. “Carbon Taxation, Research Institute on Climate Change and the Inequality and Engel’s Law–The Double Environment. Dividend of Redistribution.” Mimeo. Groves, D. G., and R. J. Lempert. 2007. “A New Kojima, M. 2013. “Drawing a Roadmap for Oil Analytic Method for Finding Policy-Relevant Pricing Reform.” Policy Res. Work. Pap. 6450. Scenarios.” Glob. Environ. Change 17: 73–85. World Bank, Washington, DC. A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   205 Kolstad, E. W., and K. A. Johansson. 2010. Economy Report . Washington, DC: New “Uncertainties Associated with Quantifying Climate Economy. Climate Change Impacts on Human Health: A Noy, I., and P. Patel. 2014. “Floods and Spillovers: Case Study for Diarrhea.” Environ. Health Households after the 2011 Great Flood in Perspect. 119: 299–305. Thailand.” Victoria University of Wellington Lamb, W. F., and N. D. Rao. 2015. “Human SEF Working Paper 3609. Development in a Climate-Constrained World: OECD (Organisation for Economic Co-operation What the Past Says about the Future.” Glob. and Development). 2014a. The Cost of Air Environ. Change 33: 14–22. Pollution: Health Impacts of Road Transport. Lecuyer, O., and A. Vogt-Schilb. 2014. “Optimal Paris: Organisation for Economic Co-operation Transition from Coal to Gas and Renewable and Development. Power under Capacity Constraints and ———. 2014b. Energy Sector Reform and Its Adjustment Costs.” Policy Res. Work. Pap. Impact on Households (Issue Note), Addressing 6985. World Bank, Washington, DC. Social Implications of Green Growth. Paris: Lempert, R. J., D. G. Groves, S. W. Popper, and Organisation for Economic Co-operation and S. C. Bankes. 2006. “A General, Analytic Development. Method for Generating Robust Strategies and ———. 2015. Aligning Policies for a Low-Carbon Narrative Scenarios.” Manag. Sci. 514–28. Economy. Paris: Organisation for Economic Lindert, K., E. Skoufias, and J. Shapiro. 2006. Co-operation and Development. “Redistributing Income to the Poor and the Olivieri, S., S. Radyakin, S. Kolenikov, M. Lokshin, Rich: Public Transfers in Latin America and the A. Narayan, and C. Sanchez-Paramo. 2014. Caribbean.” Soc. Saf. Nets Primer Ser. Simulating Distributional Impacts of Macro- Macintosh, A., and D. Wilkinson. 2010. The dynamics: Theory and Practical Applications. Australian Government’s Solar PV Rebate Washington, DC: World Bank. Program: An Evaluation of Its Cost- O’Neill, B. C., E. Kriegler, K. Riahi, K. L. Ebi, Effectiveness and Fairness. Australia Institute. S. Hallegatte, T. R. Carter, R. Mathur, and Markandya, A., B. G. Armstrong, S. Hales, D. P. van Vuuren. 2013. “A New Scenario A. Chiabai, P. Criqui, S. Mima, C. Tonne, and Framework for Climate Change Research: The P. Wilkinson. 2009. “Public Health Benefits of Concept of Shared Socioeconomic Pathways.” Strategies to Reduce Greenhouse-Gas Clim. Change 122: 387–400. Emissions: Low-Carbon Electricity Pachauri, S., B. J. van Ruijven, Y. Nagai, K. Riahi, Generation.” The Lancet 374: 2,006–15. D. P. van Vuuren, A. Brew-Hammond, and Matus, K., K.-M. Nam, N. E. Selin, L. N. Lamsal, N. Nakicenovic. 2013. “Pathways to Achieve J. M. Reilly, and S. Paltsev. 2012. “Health Universal Household Access to Modern Energy Damages from Air Pollution in China.” Glob. by 2030.” Environ. Res. Lett. 8: 024015. Environ. Change 22: 55–66. Patankar, A. Forthcoming. “The Exposure, Medlock III, K. B., and R. Soligo. 2001. “Economic Vulnerability, and Ability to Respond of Poor Development and End-Use Energy Demand.” Households to Recurrent Floods in Mumbai.” Energy J. 77–105. Background paper prepared for this report. Milder, J. C., S. J. Scherr, and C. Bracer. 2010. Patankar, A., and A. Patwardhan. Forthcoming. “Trends and Future Potential of Payment for “Estimating the Uninsured Losses due to Ecosystem Services to Alleviate Rural Poverty Extreme Weather Events and Implications for in Developing Countries.” Ecol. Soc. 15: 4. Informal Sector Vulnerability: A Case Study of Myers, S. S., A. Zanobetti, I. Kloog, P. Huybers, Mumbai, India.” Forthcoming in Nat Hazards. A. D. Leakey, A. J. Bloom, E. Carlisle, Pindyck, R. S. 2013. “Climate Change Policy: L. H. Dietterich, G. Fitzgerald, and What Do the Models Tell Us?” J. Econ. Lit. 51: T. Hasegawa. 2014. “Increasing CO2 Threatens 860–72. Human Nutrition.” Nature 510: 139–42. Porter, J. R., L. Xie, A. J. Challinor, K. Cochrane, Narain, U., and K. Veld. 2008. “The Clean S. M. Howden, M. M. Iqbal, D. B. Lobell, and Development Mechanism’s Low-Hanging Fruit M. I. Travasso. 2014. “Food Security and Food Problem: When Might It Arise, and How Might Production Systems.” In Climate Change 2014: It Be Solved?” Environ. Resour. Econ. 40: Impacts, Adaptation, and Vulnerability. Part A: 445–65. Global and Sectoral Aspects, edited by NCE (New Climate Economy). 2014. Better C. B. Field, V. R. Barros, D. J. Dokken, Growth Better Climate: The New Climate K. J. Mach, M. D. Mastrandrea, T. E. Bilir, 2 0 6    SHOCK WAVES M. Chatterjee, K. L. Ebi, Y. O. Estrada, Shindell, D. T. 2015. “The Social Cost of R. C. Genova, B. Girma, E. S. Kissel, Atmospheric Release.” Clim. Change 1–14. A. N. Levy, S. MacCracken, P. R. Mastrandrea, Smith, S. J., J. J. West, and P. Kyle. 2011. and L. L. White, 485–533. Contribution of “Economically Consistent Long-Term Working Group II to the Fifth Assessment Scenarios for Air Pollutant Emissions.” Clim. Report of the Intergovernmental Panel on Change 108: 619–27. Climate Change. Cambridge, U.K. and New Steckel, J. C., O. Edenhofer, and M. Jakob. 2015. York: Cambridge University Press. “Drivers for the Renaissance of Coal.” Proc. Rao, N. D., K. Riahi, and A. Grubler. 2014. Natl. Acad. Sci. 112: E3,775–81. “Climate Impacts of Poverty Eradication.” Nat. Steinberger, J. K., and J. T. Roberts. 2010. Clim. Change 4: 749–51. “From Constraint to Sufficiency: The Ravallion, M. 2010. “Do Poorer Countries Have Decoupling of Energy and Carbon from Less Capacity for Redistribution?” J. Glob. Human Needs, 1975–2005.” Ecol. Econ., Dev. 1. Special Section: Ecological Distribution Rogelj, J., M. Schaeffer, M. Meinshausen, Conflicts 70: 425–33. D. T. Shindell, W. Hare, Z. Klimont, Stern, N. 2013. “The Structure of Economic G .   J .   M .   Ve l d e r s , M . A m a n n , a n d Modeling of the Potential Impacts of Climate H. J. Schellnhuber. 2014. “Disentangling the Change: Grafting Gross Underestimation of Effects of CO2 and Short-Lived Climate Forcer Risk onto Already Narrow Science Models.” Mitigation.” Proc. Natl. Acad . Sci. 111: J. Econ. Lit. 51: 838–59. 16,325–30. Thompson, T. M., S. Rausch, R. K. Saari, nd Rose, A., E. Bulte, and H. Folmer. 1999. “Long- N. E. Selin. 2014. “A Systems Approach to Run Implications for Developing Countries of Evaluating the Air Quality Co-benefits of U.S. Joint Implementation of Greenhouse Gas Carbon Policies.” Nat. Clim. Change 4: Mitigation.” Environ. Resour. Econ . 14: 917–23. 19–31. UNFCCC (United Nations Framework Convention Rozenberg, J., C. Guivarch, R. Lempert, and on Climate Change). 2010. 16th Conference of S. Hallegatte. 2014. “Building SSPs for Climate the Parties of the United Nations Framework Policy Analysis: A Scenario Elicitation Convention on Climate Change. Methodology to Map the Space of Possible Vagliasindi, M. 2012. Implementing Energy Future Challenges to Mitigation and Subsidy Reforms: Evidence from Developing Adaptation.” Climatic Change 122 (3): 509–22. Countries. Washington, DC: World Bank. Rozenberg, J., and S. Hallegatte. Forthcoming. van Benthem, A. 2015. “Energy Leapfrogging.” “The Impacts of Climate Change on Poverty in J. Assoc. Environ. Resour. Econ. 2: 93–132. 2030, and the Potential from Rapid, Inclusive Vogt-Schilb, A., and S. Hallegatte. 2014. and Climate-Informed Development.” “Marginal Abatement Cost Curves and the Background paper prepared for this report. Optimal Timing of Mitigation Measures.” Rozenberg, J., A. Vogt-Schilb, and S. Hallegatte. Energy Policy 66: 645–53. 2 0 1 4 . “ Tr a n s i t i o n t o C l e a n C a p i t a l , Vogt-Schilb, A., S. Hallegatte, and C. de Gouvello. Irreversible Investment and Stranded Assets.” 2014. “Marginal Abatement Cost Curves and Policy Res. Work. Pap. 6859, World Bank, Quality of Emission Reductions: A Case Study Washington, DC. on Brazil.” Clim. Policy. Shindell, D., J. C. I. Kuylenstierna, E. Vignati, Watts, N., W. N. Adger, P. Agnolucci, J. Blackstock, R. van Dingenen, M. Amann, Z. Klimont, P. Byass, W. Cai, S. Chaytor, T. Colbourn, S. C. Anenberg, N. Muller, G. Janssens- M. Collins, A. Cooper, P. M. Cox, J. Depledge, Maenhout, F. Raes, J. Schwartz, G. Faluvegi, P. Drummond, P. Ekins, V. Galaz, D. Grace, L. Pozzoli, K. Kupiainen, L. Höglund-Isaksson, H. Graham, M. Grubb, A. Haines, I. Hamilton, L. Emberson, D. Streets, V. Ramanathan, A. Hunter, X. Jiang, M. Li, I. Kelman, L. Liang, K. Hicks, N. T. K. Oanh, G. Milly, M. Williams, M. Lott, R. Lowe, Y. Luo, G. Mace, M. Maslin, V. D e m k i n e , a n d D . F o w l e r. 2 0 1 2 . M. Nilsson, T. Oreszczyn, S. Pye, T. Quinn, “Simultaneously Mitigating Near-Term M. Svensdotter, S. Venevsky, K. Warner, B. Xu, Climate Change and Improving Human Health J. Yang, Y. Yin, C. Yu, Q. Zhang, P. Gong, and Food Security.” Science 335: 183–89. H. Montgomery, and A. Costello. 2015. “Health A W indow of O pport u nit y : C limate - I nformed D evelopment and P ro - P oor C limate P olicies   207 and Climate Change: Policy Responses to Protect World Bank. 2010. World Development Report Public Health.” The Lancet. 2010: Development and Climate Change . Weitzman, M. L. 2014. “Fat Tails and the Social Washington, DC: World Bank. Cost of Carbon.” Am. Econ. Rev . 104: ———. 2012. Inclusive Green Growth: The Pathway 544–46. to Sustainable Development. Washington, DC: West, J. J., S. J. Smith, R. A. Silva, V. Naik, World Bank. Y. Zhang, Z. Adelman, M. M. Fry, S. Anenberg, ———. 2013. Risk and Opportunity: Managing L. W. Horowitz, and J.-F. Lamarque. 2013. Risk for Development, World Development “Co-benefits of Mitigating Global Greenhouse Report 2014. Washington, DC: World Bank. Gas Emissions for Future Air Quality and ———. 2014a. Climate-Smart Development: Human Health.” Nat. Clim. Change 3: Adding Up the Benefits of Actions That Help 885–89. Build Prosperity, End Poverty and Combat WHO (World Health Organization). 2006. Fuel Climate Change. Washington, DC: World Bank. for Life: Household Energy and Health. World ———. 2014b. State and Trends of Carbon Pricing Health Organization. 2014. Washington, DC: World Bank. ECO-AUDIT Environmental Benefits Statement The World Bank Group is committed to reducing its environmental footprint. In support of this commitment, the Publishing and Knowl- edge Division leverages electronic publishing options and print-on- demand technology, which is located in regional hubs worldwide. Together, these initiatives enable print runs to be lowered and shipping distances decreased, resulting in reduced paper consumption, chemical use, greenhouse gas emissions, and waste. The Publishing and Knowledge Division follows the recommended standards for paper use set by the Green Press Initiative. The majority of our books are printed on FSC certified paper, with nearly all containing 50–100 percent recycled content. The recycled fiber in our book paper is either unbleached or bleached using Totally Chlorine Free (TCF), Processed Chlorine Free (PCF), or Enhanced Elemental Chlorine Free (EECF) processes. More information about the Bank’s environmental philosophy can be found at http://crinfo.worldbank.org/wbcrinfo/node/4. C limate change threatens the objective of eradicating poverty. Poor people and poor countries are already vulnerable to all types of climate-related shocks—natural disasters that destroy assets and livelihoods; waterborne diseases and pests that become more prevalent during heat waves, floods, or droughts; crop failure from reduced rainfall; and spikes in food prices that follow extreme weather events. Such shocks can erase decades of hard work and leave people with irreversible human and physical losses. Changes in climate conditions caused by increasing concentrations of greenhouse gases in the atmosphere will worsen these shocks and slow down poverty reduction. The good news is that, at least until 2030, “good development” can prevent most of these impacts. By “good development,” we mean development that is rapid, inclusive, and climate informed; includes strong social safety nets and universal health coverage; and is complemented with targeted adaptation interventions such as heat-tolerant crops and early warning systems. Absent such good development, many people will still be living in or close to extreme poverty in 2030, with few resources to cope with climate shocks and adapt to long- term trends, and climate change could increase extreme poverty by more than 100 million people by 2030. In the longer run, beyond 2030, our ability to adapt to unabated climate change is limited. To keep the longer-term impacts on poverty in check, immediate emissions-reduction policies are needed that bring emissions to zero by the end of the 21st century. These policies need not threaten short-term progress on poverty reduction—provided they are well designed and international support is available for poor countries. Ending poverty and stabilizing climate change will be unprecedented global achievements. But neither can be attained without the other: they need to be designed and implemented as an integrated strategy. Shock Waves: Managing the Impacts of Climate Change on Poverty brings together those two objectives and explores how they can more easily be achieved if considered together. The book provides guidance on how to design climate policies so they contribute to poverty reduction, and on how to design poverty reduction policies so they contribute to climate change mitigation and resilience building. ISBN 978-1-4648-0673-5 90000 9 781464 806735 SKU 210673