VIBRANT CITIES ON THE BEDROCK OF STABILITY, PROSPERITY, AND SUSTAINABILITY Somik V. Lall Jon Kher Kaw Forhad J. Shilpi Sally Beth Murray VIBRANT CITIES ON THE BEDROCK OF STABILITY, PROSPERITY, AND SUSTAINABILITY Somik V. Lall Jon Kher Kaw Forhad J. Shilpi Sally Beth Murray © 2023 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; internet: www.worldbank.org ISBN 978-0-9765655-8-1 Some rights reserved This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, 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 any other information shown on any map in this work do not imply, on the part of the World Bank Group, any judgment on the legal status of any territory, or any 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://creativecom- mons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, trans- mit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Lall, Somik V.; Kaw, Jon Kher; Shilpi, Forhad J.; Murray, Sally Beth. 2023. Vibrant Cities—On the Bedrock of Stability, Prosperity, and Sustainability. Washington, DC: World Bank. 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 responsi- bility 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 reuse a component of the work, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Examples of compo- nents can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; email: pubrights@worldbank.org. This report received support from the Sustainable Urban and Regional Development program (SURGE), a World Bank Umbrella Multi-Donor Trust Fund financed by the Swiss State Secretariat for Economic Affairs (SECO), and NDC-­ Support Facility via funding from Germany’s Federal Ministry for Economic Cooperation and Development (BMZ). Follow @WBG_Cities or @WorldBankMENA on Twitter for more information. #WBG_VibrantCities What’s new? The report provides new evidence, analysis, and • City-specific assessments show how urban policy insights to advance Green, Resilient, and Inclu- congestion­ —­i n transport, housing, and sive Urban Development. In doing so, it operational- pollution­—­erode productivity-enhancing ag- izes the World Bank’s GRID approach and lays the glomeration economies in developing country foundations to shore up our technical assistance cities, and what that means for jobs and eco- and policy engagements for urban development nomic dynamism. in MENA­ —­the Middle East and North Africa­ —­ and • Those models and assessments increase under- elsewhere. standing of the roles of economic development, public investment, and market/government in- NEW EVIDENCE surance in building resilience. • The first cost-benefit analysis of the impacts of • How cities can be resilient in the face of natural urban investments in mass transit and urban hazards and the Covid-19 pandemic. density on mitigating CO2 emissions­—­using • How cities are managing CO2 emissions and the valuations of carbon savings­ —­strengthens the drivers of city-specific emissions. appraisal of infrastructure projects. • What are the patterns of inequality within cities and intergenerational mobility in urban slums SHARP POLICY INSIGHTS and informal settlements. • Building state capacity as a pre-requisite for NOVEL ANALYSIS Green, Resilient and Inclusive Urban Develop- ment. Using a political economy lens­ ­ rawing —d • City-specific models examine the comple­ on the latest thinking in public economics­ has —­ mentarities between land use and transport major implications for our Program-for-Results policies in expanding jobs, raising productivity, (P4R) financing operations. and reorienting urban development. • Policy framework­ Inform, support, protect. —­ iii i v VI B R A N T C I T I E S • Inform­—­ to set expectations and motivate assistance­—­t o help more marginalized private action, including for climate adap- groups navigate adverse shocks and open tation. Well-functioning markets­—­for land, escape routes from inequality traps. labor, capital, and products­ —­can provide • The policy framework can help with World credible information to enable people, rich Bank operations and engagements, such as the and poor, to become active participants in design of prior actions for Development Policy building resilience, move up the urban lad- Operations, as in the new Egypt DPO. It can der, and become productive workers and inform Investment Policy Financing for service entrepreneurs. delivery and infrastructure provision, as with • Support­ —­firms and households with broad new BRT investment in Amman, Jordan. And investments in connectivity and services, for it can open opportunities for new programs, as efficiency, inclusion, and climate resilience. with expanding opportunities through housing • Protect­ —­ tailor provision to different needs and school vouchers. and extend public safety nets and targeted Contents What’s new? iii —­ Inform, support, and protect­ boosting low-carbon Preface viii urban development through urban planning and Acknowledgments x coordinated public investment 58 Overview 1 Annex: Impact of underground subway systems on CO2 concentrations: An econometric analysis 61 Chapter 1 Notes 62 Making cities resilient 21 References 62 Floods and urban resilience 21 New evidence on economic development, private Chapter 3: Making cities more inclusive­ —­ action, and public interventions in urban resilience 25 for economic growth, stability, and prosperity 67 The Covid-19 pandemic and city resilience 29 Productive cities tend to be more unequal 67 Enhancing resilience through self-protection, public Inequality in developing country cities 68 investment, and market insurance 34 Are cities escalators out of poverty? 69 Inform, support, protect 35 Inform, support, and protect to make cities more What decisionmakers need first: economic growth inclusive 80 and a market-conscious policy lens 37 References 87 Annex: Does flood protection infrastructure investment Notes 86 enhance resilience? An econometric analysis 38 Notes 40 Chapter 4: Making cities more References 41 productive—connecting markets, upgrading skills, raising expectations 91 Chapter 2: Decarbonizing cities—data, Muted agglomeration benefits in developing and density, and delivery 47 emerging market cities 92 Recent evidence on the magnitude and sources of CO2 Four foundations for productive cities 97 emissions for cities worldwide 48 Modeling the spatial dynamics of city productivity 112 Sustainable population density: a debate complicated Making city growth resilient and inclusive 120 by disparate patterns at different incomes 51 Notes 124 Public infrastructure for urban sustainability: subway References 125 systems mitigating GHG emissions 53 v vi VI B R A N T C I T I E S Chapter 5: Taking the next steps for 4.8 Tackling central land congestion supports a tomorrow’s vibrant cities—Governance shift toward spatial specialization and tradables reforms to inform, support, and protect 129 productivity in Dhaka 107 Fiscal capacity 129 4.9 Bogota’s BRT—accounting for private responses City governance 131 to transport investments, not only travel time Inform market actors­ setting expectations and —­ saved, and allowing land redevelopment 108 motivating private action 136 4.10 MENA’s ghost cities versus Tangier’s more Support private action­ —­investing in connection and promising model 111 service provision 136 4.11 New urban models capture urban form and Protecting the less fortunate­—­when needed, on a productivity 113 timely, targeted, and temporary basis 138 4.12 Calibration of the new urban model for Cairo 115 References 139 4.13 Cairo has been dense since the 11th century 117 Notes 139 4.14 Addressing gendered barriers to rail commuting in Cairo 122 Boxes 5.1 Legitimacy and trust 132 1 MENA’s cities: Today’s underperformers, 5.2 How governance through political accountability tomorrow’s powder kegs­ —­or drivers of prosperity? 2 improved health services in Ceara, Brazil 133 2 Sluggish urban economies: The perils of sterile 5.3 Casablanca­ —­Delegated public services and agglomeration in MENA countries 4 public sector companies 135 3 Active adaptation for MENA cities 6 5.4 Amman­ Municipal-owned companies for land —­ 4 Reconnecting Cairo: Investments to get the development and public transport 137 urban economy moving 14 5 High costs undermine labor productivity 16 Figures 1.1 Gender disparities in disaster risk reduction 24 1 Inform, support, and protect­ —­ for vibrant and 3.1 Why should cities care about inclusion and equity? 70 sustainable cities 5 3.2 Why is children’s socioeconomic status 2 The 100 top CO2 emitters 9 positively correlated with that of parents? 71 3 Intracity inequality is high in large cities of 3.3 Does city growth improve female labor force developing countries, and highest in Latin America 11 participation? 75 1.1 Percentage of population exposed to 1, 2, or 3 or 3.4 Access to credit as a tool for neighborhood more climate change hazards 23 segregation? 77 1.2 Accessibility to social infrastructure with areas at 3.5 Refugees and low-quality housing in Amman 80 risk of flooding in Fez 24 3.6 The brunt of climate-related hazards 1.3 Flooding in Accra leads to heavy asset loss for a and the COVID-19 pandemic has fallen large share of poor households 25 disproportionately on the poor, women, elderly, 1.4 Three ways to adapt to climate change: an and other vulnerable groups in three MENA Ehrlich and Becker framework 26 countries 81 1.5 Night lights before and after floods in Wuhan, 2016 27 3.7 Are housing vouchers enough to induce families 1.6 Night lights before and after floods in Chennai, to move to high upward-mobility neighborhoods? 84 October 2015 to January 2016 30 4.1 Like many MENA countries, Egypt has struggled 1.7 Covid-19’s impact is more severe in developing to transition out of government employment 94 countries 31 4.2 Population density with and without capital density 97 1.8 Cities in MENA suffered a major economic loss 4.3 Human capital deficits and uncontestable labor following the onset of the pandemic 32 markets undermine labor productivity in MENA 99 1.9 Population density and city income, 2015, 400 4.4 Cairo’s congested transport network and cities worldwide 33 fragmented urban form—a burden for commuters 101 1.10 Covid-19 contagion hotspots were throughout 4.5 Transport investments help overcome Kinshasa in 2020 33 informality in Mexico City 103 1.11 Compounded shocks from the overlapping of 4.6 Past studies construed price levels for flood risk and COVID-19 transmission risk 34 productivity and agglomeration economies as 2.1 The 100 top CO2 emitters 49 high in developing countries 105 2.2 Average annual CO2 emissions, 2014–2021 49 4.7 High costs of living may help explain 2.3 Econometrically estimated CO2 emissions uncompetitive wages in Sub-­ Saharan African cities 105 sources for sample cities 50 2.4 City emissions performance by region 50 Contents vii 2.5 Emission performance for selected cities in Egypt 51 4.4 Steam rail enabled the spatial separation of work 2.6 CO2 concentration and population density at and residences in Greater London 109 different income levels 52 4.5 Government spending by spatial category in 2.7 Subway system development, Tokyo, Moscow, several MENA countries differs substantially and Madrid, 1935–2010 56 from that of international comparators 112 2.8 Qualifying functional urban areas and net 4.6 Location of new cities, and other major land present value ratios: varying social costs of types, in Greater Cairo 114 carbon and UCs 59 4.7 Ongoing transit investments in Greater Cairo 115 3.1 Large cities in developing countries are more 4.8 East, west, north, south, and central Cairo 116 unequal 68 4.9 Change in population density 118 3.2 Large Latin American cities are the most unequal 69 4.10 Change in employment density 118 3.3 Intergenerational mobility­—­absolute (better 4.11 Baseline spatial variation in labor productivity off than parents) and relative (moving up from across Greater Cairo 119 parents’ position) 71 4.12 Baseline spatial variation in residential amenity 3.4 Absolute mobility in schooling climbed in across Greater Cairo 119 developing countries as a group 72 5.1 Decentralized spending and own-source 3.5 Absolute mobility in schooling picked up in most revenues are low in MENA 130 developing regions 73 5.2 Egypt has seen a persistent decline in 3.6 Relative mobility in schooling over time 73 subnational revenue and local taxes as a 3.7 Relative mobility in schooling by regions 74 percentage of national revenue 131 3.8 Commuting distance and neighborhood 5.3 Trust in government and the civil service in characteristics in Amman 76 MENA countries 133 3.9 East Amman has more low-income households 5.4 Cairo, Amman, and Casablanca: three points on than west Amman 77 a spectrum of urban government accountability 134 3.10 Southwestern Mexico City has more vulnerable household clusters than the northeast 77 Tables 3.11 Ulaanbaatar’s ger districts have poor access to A1.1 Dams and resilience 38 sanitation 78 A1.2 Recovery dynamics for floods and extreme rain 39 3.12 Long times to reach schools, health facilities, and 2.1 Urban growth model 53 green spaces 79 2.2 Urban development and GHG emissions 4.1 Cities in MENA display a lower share of urban modeling scenarios for Amman 54 tradable employment than other world regions 93 2.3 CO2 emissions impact of operating subway systems 57 4.2 World map of production cities and A2.1 Determinants of CO2 concentration anomalies 61 consumption cities, c. 2000 96 4.1 Impact of transit investments in Greater Cairo 4.3 Transport costs are often prohibitive for the (new urban model results) 117 urban poor 101 4.2 Maximum heights under new, 2021, building regulations in Cairo 119 vi i i VI B R A N T C I T I E S Preface The motivation for this report started with a focus process, and when is it a systemic reflection of on the MENA region­ —­a region at the epicentre of unfairness linked to a person’s origin and family? multiple crises. Beyond recent shocks associated with Covid-19 and turbulent oil prices, the region To address related knowledge gaps with new has long been suffering from anemic growth and analytical findings, the report extends empirical poor labor market outcomes. MENA also faces a analysis of the shape and growth of cities around jobs and youth unemployment crisis, and urban the world and findings from Pancakes to Pyramids: centers are on the frontlines of climate change. City Form to Promote Sustainable Growth2 that the And it faces enormous demand for reforms in the most successful urban areas are those that connect wake of the Arab Spring, as policymakers across physical growth to economic demand and support the region try to find pragmatic policy reforms to it with good plans, policies, and investments to change their countries’ growth and job paths. avoid uncontrolled sprawl. The report addresses the link between city development, productivity, The report builds on the recent World Bank report, jobs, and GHG emissions. It also brings in new Convergence: Five Critical Steps Toward Integrating perspectives on private action and choices and Lagging and Leading Areas in the Middle East and public investment in building urban resilience, North Africa.1 It explains why MENA cities are not extending the discourse in Place Productivity, and benefiting from agglomeration, migration, and Prosperity: Revisiting Spatially-Targeted Policies for specialization. It identifies the underlying causes of Regional Development.3 spatial fragmentation, limited economic mobility of its people, and walled-off national economies The report’s analysis of resilience, inclusion and from regional and global markets. It provides fur- growth complements those in recent flagship ther evidence on inequality within cities and inter- reports. The recently launched Sustainable Devel- generational mobility in urban slums and informal opment Network’s Flagship, Thriving: Making settlements. And it also tackles the questions: Cities Green, Resilient, and Inclusive in a Changing When is inequality an outcome of the urbanization Climate, also deals with green and resilient urban Preface ix development. While this report focuses more on NOTES deeper analysis of flooding and Covid-19 shocks, the SD flagship provides a more comprehensive 1. World Bank 2020. account of urban resilience in response to differ- 2. Lall et al. 2021. ent types of climate shocks. This report provides 3. Grover, Lall, and Maloney 2022. a benchmark for city level GHG emissions com- parable across the globe. The SD flagship covers REFERENCES issues of urban livability beyond GHS emissions, considering such issues as air and water pollution. Grover, A., S. Lall, and W. Maloney. 2022. Place Productiv- The inclusion aspects are covered in more detail in ity, and Prosperity: Revisiting Spatially-Targeted Policies this report – bringing in complementary analysis as for Regional Development. Washington, DC: World well as dynamics of economic and social mobility. Bank. In addition, this report devotes a chapter to urban Lall, S., M. Lebrand, H. Park, D. Sturm, and A. J. Ven- growth, which is not been covered in the SD flag- ables. 2021. Pancakes to Pyramids: City Form to Promote ship. Together these two reports­ —­complementing Sustainable Growth. Washington, DC: World Bank. each other­ —­provide a comprehensive analysis of World Bank. 2020. Convergence: Five Critical Steps Toward urban development within the Green Resilient Integrating Lagging and Leading Areas in the Middle East Inclusive Development (GRID) framework. and North Africa. Washington, DC: World Bank. x VI B R A N T C I T I E S Acknowledgments This report was prepared by a team led by Jon Kher management, resilience, and land); Ayat Soliman Kaw (senior urban development specialist) and (director, strategy and operations, Latin America Somik V. Lall (economic advisor and staff director and the Caribbean); Meskerem Brhane (regional to the World Bank Chief Economist). The core director for sustainable development, Middle East team also included Forhad J. Shilpi (senior econ- and North Africa); Sameh Wahba (regional director omist), Sally Beth Murray (economist), Hogeun for sustainable development, Europe and Central Park (urban specialist), and Ban Edilbi (young Asia); Maitreyi B. Das (director for trust funds and professional). partner relations); and Catherine Tovey, Angelica Nunez del Campo, Marianne Grosclaude, and Jaa- The following conducted research and prepared far Sadok Friaa (practice managers). working papers that contributed to the final report: Nancy Lozano Gracia (lead economist); Mark Rob- The team was fortunate to receive excellent erts (lead urban economist); Sohaib Athar (senior advice and guidance from the following peer urban specialist); Stuti Khemani (senior econo- reviewers at various points in the report prepara- mist); Olivia D’aoust (senior urban economist); tion process: Jesko Hentschel (country director Amjad Khan and Tatjana Kleineberg (economists); for the Maghreb countries and Malta); Roberta Roman David Zarate Vasquez (research econo- Gatti and Richard Damania (chief economists); mist); and Myriam Ababsa, Olga Kaganova, Hala Ming Zhang (practice manager); Thomas Farole, Mujally, Giuseppe Rossitti, and Sanji Wibisana Ruth Hill, Judy Baker and Jaime de Piniés (lead (consultants). economists); Sateh Chafic El-­Arnaout (lead urban development specialist); Joanna Masic (lead urban The work was conducted under the general guid- specialist); Mathilde Lebrand and Harris Selod ance of Juergen Voegele (vice president, sustain- (senior economists); Wael Zakout (senior techni- able development); Ferid Belhaj (regional vice cal adviser, land and geospatial); Victoria Stanley president, Middle East and North Africa); Bernice (senior land administration specialist); Syed Adeel Van Bronkhorst (global director, urban, disaster risk Abbas (senior climate change specialist); Dilnara Preface xi Isamiddinova and Mayu Sakota (senior operations Facility via funding from Germany’s Federal Min- officers); Jonna Lundwall (senior social scien- istry for Economic Cooperation and Development tist); and Chaymae Belouali (urban development (BMZ), as well as from the United Kingdom For- specialist). eign, Commonwealth and Development Office (FCDO) via the Growth Research Team, Research Mr. Abderrahim Kassou, Dr. Randa A. Mahmoud, and Evidence Division. and Magd Zahran made invaluable contributions to this paper by collecting data and information ABOUT THE AUTHORS on Casablanca and Cairo, respectively. The spatial analysis for Amman, Jordan, was made possible by Somik Lall is staff director of the World Bank’s 2024 Microsoft’s AI for Humanitarian Action program World Development Report, which examines the and was carried out by the collaboration with challenges of economic growth in middle-­ income Microsoft AI for Good Research Lab, including countries. He is also an economic adviser in the Rahul Dodhia, Juan M. Lavista Ferres, Anthony Office of the World Bank Group Chief Economist. Ortiz, Caleb Robinson, and Tina Sederholm. The Previously, he headed the climate economics and findings were presented at the 2022 Conference on policy team in the World Bank’s Equitable Growth, Computer Vision and Pattern Recognition. Finance and Institutions vice presidency, where he developed and supervised high-­ quality research The authors would like to thank Yousra Mohamed programs on resilience and economic develop- Abdelrahman, Adelaide Barra, Victoria Ahlonkoba ment, the macro-­ criticality of climate change, Bruce-­ Goga, Priyantha Jayasuriya Arachchi, and and innovations for the low-­ carbon transition. His Thao Phuong Tuong for their unfailing adminis- other roles include global lead for territorial and trative support. spatial development and lead economist for the World Bank’s Urban, Disaster Risk, Resilience, and A team at Communications Development Land Global Practice. Dr. Lall also teaches at Johns Incorporated­ —­l ed by Bruce Ross-­ L arson and Hopkins University and has been a Visiting Pro- including Joe Caponio, Mike Crumplar, Debra fessor at the National Institute of Public Finance Naylor (Naylor Design), Chris Trott, and Elaine and Policy. Wilson­ —­e dited, designed, and produced the report. Report and chapter covers were provided Jon Kher Kaw is a senior urban development spe- by Belinda Tato and Lily Liebes of Ecosistema cialist with the World Bank’s Urban, Disaster Risk Urbano. Laura Ivers and Susan Pleming (senior Management, Resilience and Land Global Practice. external affairs officers) and Melina Fleury Franco Since 2013, he has led numerous city engagements, (communications consultant) provided publicity, investment operations, and advisory and analyti- social media, and dissemination support. cal work in the South Asia, Middle East and North Africa regions. Prior to joining the World Bank, he The authors are also deeply thankful to the gov- held various key positions at the Urban Redevelop­ ernment officials and private-­ sector experts in ment Authority, the national planning agency of Amman, Cairo, and Casablanca who shared infor- Singapore, where he oversaw technical and policy mation and their professional insight. work on urban planning and design, property and land markets, and urban resilience. He also under- This report received support from the Sustain- took academic research and private sector work in able Urban and Regional Development program Singapore and London. He holds a master’s degree (SURGE), a World Bank Umbrella Multi-­ Donor with a specialization in urbanization and real estate Trust Fund financed by the Swiss State Secretariat from Harvard University, and a master’s degree in for Economic Affairs (SECO) and NDC-­ Support architecture from Columbia University. x i i VI B R A N T C I T I E S Forhad J. Shilpi is a senior economist in the topics of urban development, spatial analytics, and Sustainability and Infrastructure Team of the remote sensing. He completed his PhD in Planning Development Research Group. Her current from Michigan State University with a fellowship research focuses on the impacts of infrastructure from the Korean government. and communication on rural-­ urban transforma- tion, the role of domestic market institutions in Ban Edilbi is a young professional working in the the transmission of international price signals, and World Bank’s Urban, Disaster Risk Management, intergenerational mobility in developing countries. Resilience and Land Global Practice. Her areas of Her research has been published in leading devel- expertise include urban design, city resilience and opment and economics journals such as Economic cultural and natural heritage. Prior to joining the Journal, Journal of Development Economics, Journal World Bank, she worked at UNESCO’s World Heri- of Human Resources, Canadian Journal of Economics, tage Centre to increase awareness of the impacts of Oxford Bulletin of Economics and Statistics and Eco- climate change, and with UN-­ Habitat’s Urban Lab nomic Development and Cultural Change. She holds in Nairobi, Kenya on addressing sustainable urban a PhD degree in economics from Johns Hopkins development challenges such as urban sprawl, University, where she taught several economics basic services provision, resilience, climate change courses. Prior to joining the World Bank, she and migration. She has a Master in Architecture worked as a research associate at the Bangladesh and Urban Design from Columbia University in the Institute of Development Studies. City of New York, and a Bachelor in Architecture from the University of Bath, United Kingdom. Sally Beth Murray is an economist in the World Bank’s Urban, Disaster Risk Management, Resil- Stuti Khemani is a senior economist in the Develop- ience and Land Global Practice. Her current work ment Research Group of the World Bank. Her area focuses on climate change, cities, and urban and of research is the political economy of public policy spatial economics. Before joining the World Bank choices, and the role of institutions in economic in 2018, Sally was a senior country economist for development. She led the World Bank’s Policy the International Growth Centre (IGC) in Rwanda, Research Report on Governance: Making Politics where she led policy research on cities, taxation, Work for Development: Harnessing Transparency and and public sector performance. Sally holds a Citizen Engagement. Her work is published in leading MSc in Development Studies (Research) from the economics and political science journals, such as London School of Economics and a Bachelors in the American Economic Journal, Journal of Development Philosophy, Politics, and Economics from the Uni- Economics and American Political Science Review. She versity of Oxford. is currently examining how the role of government has resurged in the 21st century to address problems Hogeun Park is an urban specialist in the World such as climate change, water scarcity, public health, Bank’s Urban, Disaster Risk Management, Resil- conflict, and (lack of) fairness in economic systems ience and Land Global Practice. His work focuses which appears to fuel social unrest. She applies on producing advanced analytics for complex economic theory of public sector institutions to urban challenges through a spatial lens. He has led develop policy ideas for how to strengthen state the spatial analytics for various knowledge prod- capacity to address these problems and build trust ucts at the Bank, including the SD flagship report and legitimacy in society. Her research and advisory “Pancakes to Pyramids” and the MENA flagship work span a diverse range of countries and regions, report “Land Matters,” and multiple analytics to including Benin, China, India, the Philippines, the design and deliver operational projects with empir- Middle East and North Africa (MENA), Nigeria, Tan- ical and rigorous evidence. He has published exten- zania and Uganda. She holds a PhD in Economics sively in leading peer-­ reviewed journals on the from the Massachusetts Institute of Technology. Preface xiii Olivia D’aoust is a senior urban economist with Warsaw. In the US, she advised the State of Cal- extensive experience in issues related to the eco- ifornia and the US Government Accountability nomics of urbanization, regional development, Office, a watchdog agency of the federal govern- and conflict and fragility. She has led and contrib- ment. She has published extensively, including the uted to several analytical products focusing on books Government Asset Management: International the drivers of and impediments to cities’ produc- Experiences (Urban Institute Press, 2006) and tivity and livability in Sub-­ Saharan Africa, Latin Guidebook on Capital Investment Planning for Local America and the Caribbean, and the Middle East Governments (World Bank, 2011; 2018); numerous and North Africa. Her work expanded to inform papers, including How American Cities Can Benefit spatial development strategies globally, identify from “Land Value Capture” Instruments in the Time of priorities for action in lagging regions, and explore 19 and Beyond (2020) and Central Government Covid-­ avenues for operations to leverage spatial synergies Property Asset Management: A Review of International and address missing complements. Olivia holds a Changes (2020). Dr. Kaganova served as an advi- doctorate in economics from the European Center sor for the Canadian National Executive Forum for Advanced Research in Economics and Statistics on Public Property. She holds a PhD in Applied at the Solvay Brussels School of Economics and Mathematics from the Institute of Physics, Acad- Management in Brussels, and a master’s degree in emy of Science, Russia. demography from the Center for Demographic Research at UCLouvain. Myriam Ababsa is a social geographer affiliated with the French Institute for the Near East (Ifpo Tatjana Kleineberg is an economist in the Macro- Amman) and an urban planner. Her work focuses economics and Growth team of the World Bank’s on decentralization, land taxation and housing pol- Research Department. Her research is at the icies in the Middle East (Jordan, Syria, Palestine). intersection of macroeconomics, economic devel- She is the author of Raqqa : territoires et pratiques opment, and urban economics. In particular, she sociales d’une ville syrienne (Ifpo, 2009). She has studies how labor market frictions and spatial coedited Popular Housing and Urban Land Tenure in frictions affect countries’ growth and development the Middle East (American University in Cairo Press, as well as countries’ regional, sectoral, and occu- 2012) and edited the Atlas of Jordan (Ifpo, 2013). A pational structure. She holds a Bachelor from Sci- former student at ENS Fontenay and Sorbonne ences Po Paris and a PhD in Economics from Yale. Paris 1 (1993), she holds a doctorate from the Uni- versity of Tours, France. She received the Best Doc- Olga Kaganova, PhD is an internationally rec- toral Dissertation Prize, Special Mention, by the Syrian ognized expert on government property assets Studies Association in 2006. and has worked in more than 35 countries. She is currently an independent consultant, after a Giuseppe Rossitti is an economist specialised in successful career at senior research positions at urban economics and political economy. He estab- the Urban Institute (Washington, DC) and NORC matter expert lished himself as a regional subject-­ at the University of Chicago. For the past 20 on Europe and Central Asia through his consulting years she has been on a forefront of USAID and work at the World Bank on numerous urbanisation World Bank efforts to integrate asset manage- projects, assessing cities’ evolution, productivity ment into public management and sustainability and their role as drivers of economic develop- agendas. She advised the governments of Chile, ment. He has also advised on projects promoting Colombia, Egypt, Ethiopia, Kosovo, Kuwait, Kyr- sustainable development and resilience across gyzstan, Macedonia, Mexico, and Saudi Arabia, different geographies, such as Sub-­Saharan Africa, and hundreds of cities, including Bishkek, Cape Latin America and the Caribbean, and the Middle Town, Jeddah, Moscow, Riyadh, Ulaanbaatar, and East and North Africa. Giuseppe holds a PhD in x i v VI B R A N T C I T I E S Economics from the London School of Economics of Government at Harvard University, where he was and Political Science. a Fulbright Scholar, and a Bachelor of Science in Accounting and Finance from the Lahore University Sohaib Athar is a senior urban development specialist of Management Sciences. in the World Bank’s Urban, Disaster Risk Manage- ment, Resilience and Land Global Practice. His Amjad Khan is an economist working in sustainable areas of expertise include urban governance and development, providing evidence-­ based insights to financing, decentralization and intergovernmental inform the design of policy and investment proj- financing for local and subnational governments ects, with a focus on evaluating the impacts asso- and entities, municipal service delivery and infra- ciated with regional development, human capital, structure, and local economic development. He has conflict and natural resource management. His led and contributed to several World Bank financing work specializes in building analytical narratives programs and technical assistance and analytical that engage with contextual nuances through large activities on these topics in various regions. Prior to amounts of quantitative and spatial data and eco- joining the World Bank, he worked as an economist nomic theory. Amjad is a native of Pakistan, holds at the International Growth Centre where he helped a PhD in Economics from the George Washington managed the Pakistan country program. Sohaib has University, and has been working with the World a Master in Public Policy from the Kennedy School Bank since 2016. Overview How will the world’s developing cities become equitable provision of basic services and opportu- vibrant­—­capable of meeting the climate, social, and nities. To do this, tomorrow’s vibrant cities will be: economic challenges of tomorrow? Like cities that • Resilient and low carbon­—L­ imiting greenhouse developed in the past, those that grow and thrive in gas emissions, reducing vulnerability to climate the coming decades will need endowments, institu- related hazards, and rebounding from disasters tions, and infrastructure to boost the expectations and pandemics. for incomes and welfare. That is how cities attract • Inclusive­—­Meeting basic needs for all residents, capital and foster productive livelihoods. But while enabling all to aspire realistically to a bet- tomorrow’s vibrant cities also urgently need basic ter life through investment in skills and through survival plans, both to mitigate climate change equitable access to job opportunities. and to reduce exposure to disaster risks while • Productive­—­Driving economic growth, creating protecting those most exposed. And with popu- jobs, boosting incomes, and financing critical lations in developing country cities still growing social and infrastructure investments. rapidly, inclusive opportunities in housing, health, education, and employment are not luxuries­ —­ they Drawing on current economic research and are necessities. One of the key expectations under- analysis, the report charts a course for today’s lying urban prosperity is that poor people hope to developing cities to become tomorrow’s vibrant escape poverty. Without this hope, precious talent cities. While spotlighting the Middle East and is wasted­—­or turned to destructive use. North Africa region (MENA), it offers general insights for city and country leaders around the Vibrant cities offer firms and households high world. Compared with cities globally, MENA cities expectations for good returns on investments, for are delivering little for most of their residents today a sustainable and resilient future, and for dynamic (box 1). To suggest what vibrant cities could look and inclusive growth. Cities thrive not only by like and what forces have historically prevented increasing incomes and wealth for a select few their emergence, the report focuses on Amman, but by improving common welfare through the Cairo, and Casablanca. 1 2 VI B R A N T C I T I E S BOX 1 MENA’S CITIES: TODAY’S UNDERPERFORMERS, TOMORROW’S POWDER —O KEGS­ ­ R DRIVERS OF PROSPERITY? Cities ought to be vibrant centers of economic productivity and growth­ —b­ ut MENA’s cities are far from realizing this potential, and their underperformance poses risks to the region’s future stability. Despite high urbanization shares, MENA countries have notably lower firm-entry rates than countries elsewhere (box figure 1). With this lack of economic dynamism comes a lack of opportunity. In 2018, MENA had a youth unemployment rate of 26 percent, the highest globally. And in the next two decades, the region’s labor force will grow by 300 million people. Where will all those young people look for employment and income? BOX FIGURE 1 MENA has significantly lower firm-entry rates for its level of urbanization New business density (new registrations per 1,000 people ages 15–64) 30 MENA All countries 20 10 0 0 20 40 60 80 100 Urbanization rate (percent) Source: World Bank Enterprise Surveys. Shocks including Covid-19, oil price volatility, and growing risks and impacts from climate change are putting ever greater pressure on MENA cities’ productivity. This pressure will prevent them from becoming inclusive and resilient, unless major commitments are made now to change their economic dynamics. Cities in MENA require concerted institutional reforms to enable more efficient urban land and housing markets, to provide those markets with more reliable information on land prices and land use, and to foster private investment. Supporting these reforms, public investments and interventions will —i also be needed for connectivity, for inclusive public services, and for resilience­ ­nvestments that must be financed through rising private sector growth and productivity. To make cities more resilient, urban policymakers making cities livable and for building their resil- can make cities better at preventing shocks­—­or cur- ience to shocks. Infrastructure maintenance and tailing exposure­ ­ o reduce vulnerability; better —t good basic service provision reduce the impact of at preparing for shocks to facilitate speedy adjust- shocks and prevent indirect damage. Investments ment; and better at responding to shocks to mitigate in enabling infrastructure are key complements to adverse impacts. Building resilience requires, first these efforts, and climate resilient infrastructure and foremost, good general urban management. is especially relevant in regions prone to natural Land use planning and zoning are critical for disasters. To enhance the resilience of individuals, Overview 3 a core task for city authorities is to provide accu- The first need of developing cities­ —­and a vital pre- rate information to enable land, housing, and condition for vibrant cities­—­is effective urban gov- insurance markets to operate efficiently. ernance and management, with efficient financing and evidence-based participatory policymaking. If To make cities low carbon, globally comparable, pub- this primary need is addressed, city and country licly disclosed metrics on city level carbon emissions leaders can act to: are needed to motivate people and businesses, as • Inform private firms and households about land well as national and city governments, to inform use, land values, and disaster risks­ —m­ otivating and calibrate their approaches to decarbonizing them to act on market signals and on credible development. City and country leaders can coordi- knowledge of government policies and plans. In nate three kinds of support for low carbon devel- well-functioning markets for land, labor, capi- opment: pricing instruments, urban planning, and tal, housing, and insurance, prices signal true public infrastructure investment. A combination of values, and the rules for market participation policies including urban planning, energy demand are fair, reliable, and transparent. Information side management, and building code enforcement about land use, land values, and disaster risks can result in significantly greater carbon abatement is readily available. And city plans communi- than any of these levers by itself. Policies that price cate the future placement of infrastructure and carbon emissions can be more effective when com- amenities. plemented by urban plans that support densification • Support firms and households with investments and by investments in mass transit. in broad-based infrastructure and service provision­ —­m aking private endeavors more To make cities more inclusive, universal basic productive. Infrastructure should aim chiefly at service provision and human capital development creating efficient connections between work- are not just moral imperatives, but links in a vir- ers and job locations and between firms and tuous cycle with productivity­—­at once reflecting regional and global markets. In some cases, in- and promoting economic growth. Developing frastructure may efficiently increase resilience cities cannot improve welfare solely by mandat- to floods and other climate-induced hazards. ing equity. Incomes must also rise, and for this to And services are essential for inclusive access happen, economic productivity must increase. Yet to health, education, and urban amenities. inclusion not only depends on prosperity­ —­it also • Protect poor and vulnerable people against promotes prosperity, notably through human cap- pandemics, severe weather events, and other ital accumulation. emergencies with transfers and safety nets that are timely, targeted, and temporary. There are To make cities more productive, coordinated pro- times that challenge even vibrant cities. And grams can overcome vicious cycles of low expec- where market incentives for the provision of tations, investment, and productivity, beginning basic services or private insurance are inade- with initiatives to connect urban markets, build quate, cities can provide emergency transfers skills, and foster land and real estate markets. For and safety nets, especially for women, school- cities in developing countries­ —i­ncluding those age children, and informal workers. in MENA­ —­the first challenge is to reform land and labor market conditions that impede private The vision here is fundamentally economic: investment and to align land use regulations and tomorrow’s vibrant cities will empower firms and incentives for market-driven growth. Only if these households to make sound decisions about where foundations are laid can city and country leaders to locate­—­and how to allocate­—­ their human and aspire to change sterile agglomerations into pro- financial capital. Such decisions include, for exam- ductive ones (box 2). ple, the choice to insure oneself against climate 4 VI B R A N T C I T I E S BOX 2 SLUGGISH URBAN ECONOMIES: THE PERILS OF STERILE AGGLOMERATION IN MENA COUNTRIES A fundamental economic challenge for MENA countries is that their cities are not fully benefiting from the three dimensions of urbanization and spatial transformation: agglomeration, migration, and spe- cialization. What economists call agglomeration economies are basic to all vibrant cities, driving rising productivity and growth through economic density and efficiency. But where the potential gains from agglomeration are eaten by congestion costs­ —­ where transport connections cannot move goods and people efficiently into, from, and around the urban area­ —c ­ ities continue to be sterile agglomerations. Rather than driving productivity, income, and inclusive economic opportunity, sterile agglomerations tend to display three symptoms of institutional inefficiency that heighten social tensions and risks. Until the ongoing institutional challenges are addressed, these symptoms will persist: • Cities are physically and economically fragmented. The three MENA cities most intensively stud- ied for this report­ Amman, Cairo, and Casablanca­ —­ —m ­ anifest horizontal spatial development, both through extension (additive sprawl), and through leapfrogging (developing parcels beyond the urban periphery at a distance from its built-up areas). Such fragmentation partly reflects the usual spatial evolution of cities in the modern era. But it also partly reflects land and housing markets, often constrained and distorted by ineffective incentives and initiatives. In addition to denying cities the benefits of agglomeration, spatially sprawling and fragmented urban forms can potentially exacerbate social inequities, divisions, and tensions.1 • People are spatially and economically stuck in place. In most parts of the world, people vote with their feet and move toward opportunity­ —­ but this happens less in MENA: just 14 percent of households on average have moved from their place of birth, compared with 28 percent across countries elsewhere. The culprits include institutional obstacles to within-country economic migration and to efficiently functioning land markets, along with a lack of education and tradable skills. These challenges pose interlocking barriers to physical labor mobility, to formal economic development, and to inclusive economic opportunity across the region. • National economies are walled off from regional and global economies. Governments in MENA countries have either created or perpetuated inefficient barriers to trade and to the production of tradable goods and services, which are essential for productive urban agglomerations.2 Cities instead specialize in nontradable services for local consumption, financed by extractives exports. For example, Egypt and Jordan have experienced negative structural change in recent decades­ —­ away from manufacturing, toward natural resources and nontradable urban services. Unless MENA’s urban economies can achieve greater efficiency, they risk continuing on a downward —­ spiral­ with rapid population growth, but without rising economic density and productivity­ —­adding to crowding, informal slum housing, and discontent. As in other, successful developing cities, the shift onto a more efficient development path will need to begin with structural economic transformation. Notes: 1. Lall et al. 2021. 2. World Bank 2020. risks or to protect oneself through adaptive reloca- informed about markets, adequately supported by tion; the choice to reside in a slum or formal hous- infrastructure and services, and protected when all ing; and the choice to invest in land near an urban else fails (figure 1). center or at the periphery (or beyond). Grounded in economic analysis, this policy approach The interactions among three dimensions of the envisions the vibrant city as having actors well urban economy­ ­ ts resilience in the face of climate —i Overview 5 FIGURE 1 —F INFORM, SUPPORT, AND PROTECT­ ­ OR VIBRANT AND SUSTAINABLE CITIES Sustainable Resilience and Inclusive adaptation + growth and low-carbon OUTCOMES urban development PUBLIC AND Public Motivate private action with credible interventions knowledge of plans and policies PRIVATE ACTIONS SUPPORT INFORM PROTECT broad-based infrastructure through policies and institutions that lay foundations for vulnerable with targeted POLICY and services efficient decision-making by individuals and firms policies APPROACH Source: Authors. and other shocks, its inclusivity in fostering social can avoid these risks entirely.1 Climate change has and economic cohesion, and its growth dynamism already made destructive storms and floods more supporting resilience and inclusion­ —­determine likely than before. And urban economic activity its vibrancy. Richer cities are better equipped to tends to concentrate in low-lying, flood-prone withstand the vagaries of nature. Growing cities areas. Among the cities most exposed to flooding can tackle social and economic inclusion better, by their geography and topography are many of the but growth generates higher emissions and can world’s fastest-growing developing cities.2 increase the frequency and potency of climate shocks. At the same time, the policies for tackling Because climate change is already affecting cities, climate change­ —­s uch as creating more urban a sound scientific understanding of urban eco- green space­ —­have potential adverse implications nomic resilience­ —­how people and places cope for affordable housing that is essential for inclu- with shocks­ is vital to formulating effective devel- —­ sive cities. This report pays close attention to these opment policy. Around the world, floods have a interactions to identify the policy priorities for significant negative impact on economic activity. vibrant cities. Most cities recover economically from flood disas- ters within two months. Yet the impact on cities MAKING CITIES RESILIENT­—­THROUGH is more severe in poor nations than in rich ones.3 PRIVATE ADAPTATION, SELF- How should the global development agenda reflect PROTECTION, PUBLIC INVESTMENT, the need for adaptation and protection in develop- AND MARKET INSURANCE ing country cities around the world? Climate change is a key urban development chal- To the recurrent human and economic toll of lenge. Around the world, rising greenhouse gas climate-related disasters, add the ongoing impacts (GHG) emissions are increasing climate-related of Covid-19, in all its successive waves and variants. risks and environmental risks to people’s lives The pandemic has ended millions of lives and cut and livelihoods (box 3). In cases of flooding and deeply into livelihoods worldwide. No real end is extreme rainfall, cities in low-income countries in sight, though the danger of this virus may even- suffer the greatest economic harm­ —­yet no cities tually wane. 6 VI B R A N T C I T I E S BOX 3 ACTIVE ADAPTATION FOR MENA CITIES The scale to which cities must adapt to climate change in MENA is large, and governments cannot do it alone. In Egypt, more than 80 percent of the population in 14 major cities is increasingly exposed to at least one major climate risk such as flooding, sea-level rise, heat stress, land subsidence, or air pollution. Most core economic activities (industries, tourism, services, and so on) are concentrated in cities and urban agglomerations.1 In Jordan, air pollution is serious in the largest cities, where the PM2.5 particle concentration is higher than the WHO standard of 10 μg/m3­ —­a significant health risk for the country’s large urban population. Urban heat in Jordan, especially on summer days when the surface tempera- ture can exceed 40 degrees Celsius (104 degrees Fahrenheit), will be a daunting enemy to urban health and productivity. In addition, some scenarios project higher pluvial flooding risk in major cities, as extreme precipitation rises with climate change. In Amman, more than 50 percent of the households at risk from floods are low income.2 And for the region, the economic impact of climate-change–induced water scarcity are significant and rising.3 Notes: 1. World Bank 2022. 2. Kaw et al. 2022. 3. World Bank 2017. The impact of both climate and pandemic shocks consider the interactions among private invest- often falls heavily on vulnerable groups, such as ments in prevention and self-protection, public women, school-age children, and informal and investments in infrastructure, and market insurance. unskilled workers. Resilience includes not only the ability of economies to recover in the aggregate, Generally, economies can become more resilient but also their effectiveness at protecting more to climate change risks through self-protection. A vulnerable groups. In Morocco, economically and city’s exposure to storm and flood risks is deter- socially vulnerable groups­ including women, the —­ mined largely by its geography and topography.4 In elderly, persons with disabilities, and others­—a­ re many circumstances, migration is an efficient way disproportionately exposed to climate-related for families and firms to mitigate their exposure disasters, especially flooding. In Bangladesh, poor to these risks. But development discourse has yet people report losing 70 percent of their incomes to adequately consider migration, and policies to following a disaster, and nonpoor people 45 per- enable migration, as effective adaptation strategies. cent. Flooding in Accra results in significant asset An alternative is to make adaptive investments in loss for a larger share of poor than nonpoor house- land and structural improvements­ —­such as elevat- holds, despite their having a similar probability of ing structures above expected flood lines. being affected. Local resilience can be built through public invest- People in developing cities do not passively experi- ment to reduce risk in particular places. Dams and —­ ence the effects of climate change­ they take action levees, for example, may protect urban districts to buffer themselves and their livelihoods against from flooding. But decisions about which places emerging risks. Urban residents are not mere vic- to protect can be fraught with tension. Should the tims: they inform themselves about circumstances, government protect richer areas because they are and they act on that information. To take this active the most productive, or poorer areas because their dimension of adaptation into account, a realistic inhabitants stand to suffer the most? No city can framework for development policymaking should build floodwalls everywhere. Overview 7 Resilience can also be built through market municipal institutions must be structured and insurance­ —­r isks of damage from storms and enabled to provide markets accurate informa- flooding can be pooled, to compensate for harm tion on risks. Market-conscious adaptation pol- to individual firms and households. The growth of icies require accessible, up-to-date information. global insurance companies creates new opportu- Affected individuals, firms, and governments must nities for deep-pocketed insurers to expand in the have both the incentive and the capacity to identify developing world: while offering insurance pol- and assess adaptation costs and benefits. As satel- icies, they can diversify their risk exposure using lite images and big data improve climate science instruments such as catastrophe bonds. The rise and expand the scope for its use in private and of global financial markets can thus protect more public decisionmaking, governments and other individuals from the losses associated with place- development actors should focus on seeking out based climate shocks. synergies among self-protection, public invest- ment, and market insurance. These efforts will be Self-protection, public investment, and market crucial in determining the medium-term costs of insurance depend on economic growth in differ- climate change. ent ways. When people, households, and firms can invest more to reduce risk exposure through Fortunately, urban prosperity increases people’s private decisions and group insurance, they will and firms’ ability to adapt to climate change by do so. And when governments are better funded protecting themselves and pooling risk through by tax revenues­ —­whether national, provincial, or insurance­ while it can also help countries and —­ municipal­ —­they will have more and better options cities finance needed public investments in mitiga- for climate risk mitigation through infrastructure tion and resilience. In the longer term, balancing investment. National governments may also have urban prosperity with necessary reductions in a greater capacity for an effective centralized GHG emissions will depend on a shift toward envi- disaster response. These readily understood rela- ronmentally sustainable middle-class preferences, tionships do much to explain why death tolls from encouraged and complemented by policy reforms disasters are generally lower in more economically (such as carbon pricing) along with investments in developed countries. low-carbon technologies for energy, transport, and construction. For now, policies must harness the For city and country decisionmakers, any debate power of markets to adapt and become resilient. on urban adaptation measures should begin with a sound analysis of how households and firms will Governments can set expectations for urban resil- likely respond­ —­through self-protection, public ience by informing market actors, by supporting interventions, and market insurance­ —­because firms and households with investments in broad- each of these options will influence investment and based infrastructure and service provision, and by location decisions. A public investment proposal protecting poor and vulnerable people. Cities and should be assessed for its impacts on market incen- countries can: tives and behavior: Will it encourage or discourage • Inform market actors by publicizing risks to motivate private adaptation and access to private insurance? adaptation­—­while ensuring that markets work well Similarly, access to private insurance should be enough to support such efforts. To promote active valued in part according to its implications for adaptation, a country’s national, provincial, self-protection: Will policy buyers be more or less and municipal institutions must be structured efficient in taking measures to adapt to risk? to provide land, housing, and insurance markets with accurate information on risks; they must To support active adaptation by households also enable these markets to operate efficiently. and firms, a country’s national, provincial, and Consider land markets. Most parts of the world 8 VI B R A N T C I T I E S today do not inform households or firms ade- that generate the largest shares of GHG emissions. quately to motivate private adaptation. But to Staying below the 1.5° C threshold means mas- align incentives and encourage self-protection, sively decarbonizing cities, in part through carbon cities and countries must allow investors to ac- pricing, in part through reforms to product and cess land with clear property rights and must factor markets, and in part through investments establish the institutional conditions for a in low-carbon energy and transport systems and well-functioning land market. programs to reduce urban sprawl. • Support firms and households with investments in broad-based infrastructure and service provision. As cities rise in population and incomes, they tend Regular maintenance of public services is key; to generate increasing GHG emissions, with the so is smart land use management. Floods in de- urban middle class growing and consuming higher veloping country cities often reflect inadequate amounts of energy intensive goods and services: maintenance of drainage and other systems: in durables, transport, housing, and other struc- South Asia, monsoon rains overflow drainage tures. Despite covering less than 3 percent of the ditches that are used as garbage dumps because world’s land area, urban areas currently account regular refuse collection is insufficient, so settle- for roughly two-thirds of global energy consump- ments are inundated by runoff. Land use plan- tion.5 The top 100 urban areas in CO2 emissions ning, a core task of city governments, shapes emit 5.1 gigatons of carbon annually, accounting hazard risk; for risk reduction, the chief aim for 14 percent of global emissions and 23 percent should be to prevent development on hazard-­ of urban emissions (figure 2).6 The top emitters are prone land. Tokyo, New York, Chicago, Guangzhou, Dortmund, • Protect poor and vulnerable people against pandem- and Beijing, in that order­ —­each emitting more ics, severe weather events, and other emergencies than 100 million tons a year. Finding strategies to with transfers and safety nets that are timely, tar- control GHG emissions in all urban areas, rich and geted, and temporary. When market incentives poor, is a vital task for policymakers globally. But for basic services or insurance are inadequate, so is building economic resilience in developing cities can provide emergency transfers and cities, which are generally the most exposed to safety nets, especially for women, school-age climate change risk. children, and informal workers. Preemptive so- cial transfers and safety nets assist vulnerable The Glasgow Pact, an outcome of the COP26 meet- households to take self-protection and self-in- ings in November 2021, emphasizes much deeper surance measures. emissions cuts, noting that without major decar- bonization efforts the world risks catastrophic MAKING CITIES LOW CARBON­ impacts on people and ecosystems. The COP26 FOR LOCAL BENEFITS AND —­ conclusions present an opportunity to reflect on GLOBAL SUSTAINABILITY the role of cities in contributing to greenhouse gas emissions (GHGs). Identifying the sources of these Low carbon urban development will be key to emissions, along with ways to accurately measure addressing climate change. Cities account for over and monitor them, is a first step toward meaning- 70 percent of global CO2 emissions­ —­most of it fully reducing them. from industrial and motorized transport systems that use huge quantities of fossil fuels and that rely Assessing decarbonization progress requires on spatially extensive infrastructure constructed new metrics, because not all cities are the same. with carbon-intensive materials. The World Bank’s Rapid decarbonization will need to be promoted Climate Change Action Plan (2021–2025) thus or rewarded by emission-based performance rat- recognizes city systems as one of five key systems ings that make urban governments and businesses Overview 9 FIGURE 2 THE 100 TOP CO2 EMITTERS Top 100 cities, by total CO2 emissions (million tons) Less than 30 30–34 35–49 50–64 65 or more Source: Dasgupta, Lall, and Wheeler 2022. more accountable to their stakeholders. However, finds that these systems cut their cities’ current performance measures cannot ignore the effects of GHG emissions by about half (on average), col- differing demographic, economic, and geographic lectively accounting for an 11 percent reduction conditions on actual CO2 levels in cities. in global emissions.7 But financing transport infrastructure is not easy­ ­east of all for cities in —l New metrics will require more accurate and developing countries­ —­ and many pitfalls can attend objective CO2 measures. Assessing city progress a large infrastructure investment, such as an under- on emission reductions has been hindered by ground subway (metro) system. Early planning the near-total absence of directly measured CO2 decisions may not be reversible at later stages in a emissions data. Consistently measured emissions city’s spatial development. From the outset, then, estimates are available only for about 80 cities, with urban plans must integrate scientific assessments only half in developing countries. What’s more, of GHG mitigation strategies, based in part on hardly any estimates are based on actual emis- economic models that compare costs and benefits sions. Most rely on parameters from engineering from alternative spatial development paths. studies applied to survey-based activity measures for transport, energy production, and manufac- Policymakers can motivate cities to shift onto low turing. These estimates are especially suspect for carbon development paths by informing stakehold- developing countries, since many of the measures ers about local and global urban GHG emissions, are calibrated using databases and models from by supporting economically credible planning for high-income economies. urban spatial efficiency, and by protecting vulner- able households from adverse impacts on high Cities, and urban development policies, can do carbon industries that employ the poor. Leaders much to mitigate GHG emissions­ if the capacity —­ should act to: and financing are available for early infrastruc- • Inform individuals and city governments to set ex- ture investments based on rigorous sustainability pectations and motivate private action. Globally analyses and sound economic modeling. Urban comparable, publicly disclosed metrics on transport infrastructure, for example, can vastly city level CO2 emissions will incentivize peo- reduce GHG emissions when done right. A new ple and businesses­ —­ as well as city and coun- study of 192 urban underground transit systems try governments­ —­t o make more informed 10 VI B R A N T C I T I E S and calibrated approaches to low carbon contrast, higher intracity economic inequality can development. create “inequality traps,” stark income differences • Support private investment with credible plans to that persist and prevent the rise of virtuous cycles coordinate land use management, demand manage- of consumption, production, and investment. ment, and building standards, along with sustainable connections. From the outset, urban plans should Large cities in developing countries exhibit higher integrate rigorous assessments of GHG mitiga- intracity inequality than smaller ones. Globally, the tion strategies, including economic modeling relationship between city population and the GINI of all proposed policies and investments. Such index of inequality displays a U-shaped curve (figure modeling allows comparisons of costs and ben- 3). Large cities in upper middle-income countries­ —­ efits from alternative spatial development paths. such as Bogota, Cape Town, Johannesburg, Mexico Investments in mass transit­—­subways­—­have set City, and Rio de Janeiro­ —­display particularly high cities on a course for lower GHG emissions over intracity inequality. At the regional level, Latin time while also integrating urban labor markets. American cities are on average characterized by • In the short term, protect vulnerable groups from the the highest intracity inequality and European and adverse effects of a transition to low-carbon develop- Central Asian cities by the lowest (see figure 3). ment. Poor people are most at risk in this tran- sition because they are most likely to earn in- Urban areas are generally expected to show higher come from informal, agricultural, and “brown” inequality than rural areas as a result of agglom- activities. It is estimated that about 40 percent eration externalities­ —­but when this inequality is of Egypt’s employment share is concentrated in exacerbated by a socially unfair distribution of brown sectors (such as construction and trans- opportunity, policymakers should raise the alarm. port) and 58 percent in yellow sectors (crops Intracity income inequality need not cause grave such as rice, for example). To capture the large concern when it is driven by entrepreneurship, potential for reduced emissions from these sec- innovation, and stronger efforts by individual tors while protecting the poor, cities and coun- workers. For example, higher intracity inequality tries can adapt and expand safety nets while typically appears in productive cities that attract investing in reskilling and supporting labor clusters of skilled workers with high earning power. mobility. Such inequality may be economically benign. More frequently, however, the outcomes for individuals MAKING CITIES MORE INCLUSIVE­ in a city depend on their family background, eth- FOR STABILITY, EFFICIENCY, —­ nicity, gender, race, birthplace, and so on. When AND ECONOMIC GROWTH a child born to a wealthy family pulls ahead sim- ply because of birth circumstances­ —­not through Inclusive economic development­ —­the wider diffu- —­ individual effort or talent­ this inequality is unfair. sion of gains from rising productivity and incomes­ Such unfair inequality is rooted in an unequal dis- —­promotes social stability while driving growth, tribution of opportunity, and it impedes cities’ and productivity, and prosperity. Although developing countries’ economic development. countries generally lack reliable granular data on urban growth, evidence from developed countries An unfair distribution of opportunity traps chil- suggests that cities grow more rapidly when their dren into a social and economic situation reflect- intracity income inequality is lower (that is, when ing that of their parents, leading to high relative income differences within the urban area are intergenerational persistence­ —­a measure defined smaller). Thus, in the United States, lower intracity as the correlation between children and parents’ inequality is significantly correlated with higher outcomes. Relative intergenerational persistence growth in urban populations and incomes.8 By provides a versatile measure of inequality of Overview 11 FIGURE 3 INTRACITY INEQUALITY IS HIGH IN LARGE CITIES OF DEVELOPING COUNTRIES, AND HIGHEST IN LATIN AMERICA Average Gini Average population inequality index from GMD surveys Gini index (unweighted) (millions) 0.6 0.5 5 High income Upper middle income Lower middle income Low income 0.4 4 0.5 Amman Sana’a 0.3 3 Casablanca 0.4 Tehran 0.2 2 Djibouti Qom Tunis 0.3 0.1 1 Baghdad 0.2 0.0 0 11 12 13 14 15 16 17 Europe & Middle Sub- East Latin Central East & Saharan Asia & America & Population (millions, log) Asia North Africa Pacific Caribbean Africa Source: Park et al. 2022. opportunity. The inverse measure, relative inter- In cities that lack basic institutions­ —e­ specially for generational mobility, registers the extent to which land ownership and tenure­ —­these discriminatory (in more inclusive societies) a talented child from patterns will inevitably continue to exclude people a disadvantaged family can climb the social ladder from opportunities for relative upward mobility, as efficiently as a similarly talented child from a imposing huge economic costs. Consider the social richer and higher status family. In an economically residential segregation in many developing cities, inclusive community, the poor but more talented with certain subgroups spatially concentrated in child will sometimes overtake a less talented child crowded slum districts and largely deprived of pub- with far greater inherited advantages. lic services. Lacking formal rights to the land they occupy, potentially subject to eviction from one day Among the many factors in relative intergenera- to the next, these residents have little or no economic tional persistence­ —­unfair inequality passed down incentive to invest in local infrastructure, services, from one generation to the next­ —­a re various and amenities. However, even after land ownership social, political, and market forces that interact to and tenure are clarified by laws and regulations­ —­a constrain opportunity for specific households and top priority for all developing cities, including those individuals. Unfair inequality often reflects social in MENA­ —­ market forces still need to be comple- prejudices that punish individuals for their race, mented with public policies for inclusion. Social and gender, ethnicity, religion, place of birth, family political identities can be sticky; left to themselves, background, or the like. Often tightly linked to they will often continue to skew even functioning these social prejudices are political interests. But markets in fundamentally unfair ways. Unfairness­ whether discrimination arises from social behav- —­not inequality­—­ is more likely the reason for many ior or from political patronage (or both), it results developing cities to become powder kegs of turmoil. in large distortions of urban factor markets, as investments and opportunities are preferentially The transmission of unfair inequality across gen- channeled to particular groups without regard for a erations hinders economically efficient urban city’s overall economic efficiency and productivity. resource allocation and accumulation. How much 12 VI B R A N T C I T I E S human capital does a city, or a country, lose in each from one generation to the next, the average per- generation to racism, to sexism, to poor children’s son by definition will experience upward absolute inability to benefit from school? The economic intergenerational mobility. Such upward mobility costs of racism and sexism are staggering. One means that economic growth lifted all boats at study finds that if women received compulsory sec- least a little. It does not mean that opportunities ondary education, global human capital wealth­ —­ are generally available for children to climb to a defined as the present value of the future earnings higher rung on the income distribution than that of of the labor force­ would have been US$15–30 tril- —­ their parents, even those who start at the bottom. lion higher.9 Closing the racial gap between whites Relative intergenerational mobility is this upward and blacks could add 0.35 percentage points to mobility from a lower rung to a higher rung of the annual GDP growth in United States.10 Bad inequal- income distribution: it directly measures the inclu- ity and wasted potential exact a toll on economic sive growth seen in vibrant cities. growth, as well as on social and political cohesion, in poor and rich countries alike. Relative intergenerational mobility is higher in some cities than others, suggesting higher degrees of social Despite high urban income inequality of both kinds and economic inclusion. Wide differences in rela- (fair and unfair), developing country cities continue tive mobility are often observed among cities in the to attract migrants­ —­simply because cities, overall, same country.12 Mobility tends to be higher in cities provide greater social and economic opportuni- with less segregation, better schools, higher social ties than rural areas. Urbanization tends to widen capital, lower income inequality, and more stable income gaps: in developing countries, inequality is family units.13 Among developing countries, relative higher within cities than in rural areas, and intracity mobility is higher in countries that spend more on inequality increases with population size.11 Despite education, have better child health indicators and this higher inequality within cities, urban areas higher school quality, and display less segregation.14 score better than rural areas on intergenerational mobility. That is, developing country cities offer The main reason developing country cities display residents greater opportunities (on average) to higher intracity inequality than developed country surpass their parents’ income and education out- cities, along with lower absolute and relative inter- comes than rural areas can provide. One reason generational mobility, is less equitable access­ —­to developing country cities have lower relative inter- jobs, services, and land, credit, and housing mar- generational persistence, and thus higher relative kets. Where basic infrastructure and services are intergenerational mobility, is that cities tend to not accessible to residents in all of a city’s districts, offer better opportunities to people in socially dis- opportunities are profoundly unequal. A new advantaged groups. These groups typically include study of 27 African countries shows that school women and may include members of marginalized outcomes improve significantly after a move to a castes, ethnicities, and religious communities. better neighborhood, especially at an early age, and that about a third of the improvement is due to For developing cities to become more inclusive, the neighborhood’s independent effect (not to the increases in relative intergenerational mobility are sorting of households in better neighborhoods).15 needed­ —­ it is not enough to demonstrate absolute Developed country cities can be expected to show intergenerational mobility (children’s outcomes less inequality, in part, because of smaller dispari- improving overall yet still co-varying closely with ties in service provision from one neighborhood to parents’ outcomes). Absolute intergenerational the next. Barriers to market participation by low- mobility measures the likelihood that a child will, er-income people and members of disadvantaged in adult life, surpass a parent’s income or education groups­ for example, in credit provision­ —­ —­ are also outcomes. If average incomes rise across the board lower in developed than in developing countries. Overview 13 Inclusive urban growth is an even higher prior- and credit markets function more efficiently, ity for developing than for developed countries. with fewer market failures and distortions from Given the evidence that intracity inequality is patronage. bad in developed country cities, it is likely to be • Support broad-based infrastructure for efficient con- still worse in developing country cities. In devel- nection and service provision, not just to the wealth- oping countries, markets for fundamental services iest and most influential urban residents, but to all. such as credit and insurance are often missing Along with digital connectivity and efficient or distorted, property rights are often enforced transport connections­ —­both vital for expand- selectively, and public service provision is skewed ing livelihood opportunities and for efficiently to favor the wealthy. The despair that arises from matching jobs and skills (chapter 3)­ —­ inclusion persistent intracity inequality periodically leads requires basic sanitation, health services, and ac- to social unrest and destruction. For developing cess to a high-quality education. Making service countries to realize and maintain stable economic providers accountable to citizens is also key. growth and prosperity, their urban economies • Protect poor urban residents with augmented safety must become more inclusive. nets as a short to medium term buffer against des- titution and adverse shocks. Such shocks include Higher intergenerational mobility, both absolute job losses, health crises, and disasters (chapter and relative, makes resource use more efficient and 1). Consider innovative investments in child boosts urban economic growth: an ironclad eco- health and education: for example, housing nomic case exists for building urban escalators out vouchers to help disadvantaged parents move of poverty while also more efficiently identifying to neighborhoods with better services and and developing human capital throughout cities. amenities, and incentives for developers to To further these ends requires both city growth and build affordable housing in relatively affluent policy interventions. City and country leaders can neighborhoods. begin with steps to reform land and labor markets; to provide better services and meet basic needs for MAKING CITIES MORE PRODUCTIVE­ all, especially through greater citizen engagement; —­WITH BETTER CONNECTIONS, and to improve governance and accountability, UPGRADED SKILLS, AND enabling cities to better raise and manage funds HIGHER EXPECTATIONS for services and safety nets. Increasingly productive cities can not only create Developing cities can become more inclusive by more and better jobs, but also expand inclusive supporting all urban residents with equitable access opportunities and finance more comprehensive to infrastructure and services; by protecting the dis- resilience measures. Today’s rapid urbanization in advantaged and vulnerable from disproportionate low- and middle-income countries could be har- shock impacts and from unfair disparities in hous- nessed to drive rising productivity, incomes, and ing and education opportunities; and by informing investor expectations. Instead, too many growing market actors from all segments of society, provid- cities are crowded with people who lack support ing all with access to the same data and subjecting from productivity-enhancing infrastructure, all to the same rules. Cities and countries can: services, markets, and human capital. The result is • Inform to spur private actions that are more inclusive high-cost urbanization with sterile agglomeration in their effects. Markets can be reformed to signal economies, leaving cities unable to compete in scarcity and returns with greater accuracy. And growth-enhancing and labor-intensive global trad- all urban residents can be provided with access ables. In MENA cities, labor productivity is often to reliable data and consistent rules for transac- undermined by fragmented and distorted labor tions. Such information can help labor, housing, markets and a shortage of marketable skills.16 And 14 VI B R A N T C I T I E S Saharan Africa, wages for urban tradables in Sub-­ cities. In contrast, nontradable services­ —­and much workers often are not internationally competitive­ of the informal sector­ —­cater only to a limited local —­driven up, in part, by high urban congestion market, resulting in diminishing returns to scale and costs.17 and in lower price levels. Unfortunately, many of today’s rapidly growing developing cities focus on As long as developing urban economies keep nontradables, limiting their chances for productiv- specializing in goods and services for local con- ity growth through agglomeration (box 4). sumption, they are likely to remain sterile agglom- erations, unable to leverage population density for To compete in global tradables, cities need interna- economic density. Tradable sectors are most able to tionally competitive productivity and prices­ —­with benefit from the agglomeration economies of large labor productivity attracting labor-intensive and BOX 4 RECONNECTING CAIRO: INVESTMENTS TO GET THE URBAN ECONOMY MOVING Despite the vast population and high residential densities of Cairo, Egypt, people who live there are not as productively employed as they could be. Greater Cairo is home to more than 20 million people, with an average of about 12,000 inhabitants per km2. But even as the city’s population keeps growing, local jobs are becoming increasingly informal, small-scale, and precarious. For example, in 2018, 89 percent of construction jobs and 95 percent of transport jobs were provided “outside of establishments”­ —­up from 85 percent for construction and 86 percent for transport in 2004.1 In recent years, a large number of Cairo residents have exited the labor market altogether. Cairo’s economy is not structured to benefit from urban agglomeration economies. Historically, countries transitioning from low to high income have transformed their economies structurally across sectors, shifting from a predominantly agricultural economy toward tradable urban services and manu- facturing. These countries have also boosted their productivity within sectors, especially in global trad- ables. In contrast, private sector employment in Cairo is heavily concentrated in nontradable sectors. And despite recent initiatives to promote private investment, the government continues to provide a large share of urban employment. An innovative analysis of productivity and spatial development in Cairo finds that better connecting Cairo’s fragmented and congested urban fabric may help shift employment toward globally tradable goods and services­ —­substantially boosting output and welfare. Using a tailored version of the new urban model developed in the past several years,2 analysts estimated the impact of ambitious new investments (which are already under way) to connect Cairo’s outlying “New Cities” with older neighborhoods. The analysis concluded that the new transit links­ —­ a large expansion of the central metro system, plus a new monorail and a new e-train­ —­ will increase output in Greater Cairo by almost 6 percent, with the largest gains in tradable services. Also increased will be the shares of labor in manufacturing (+4 percentage points) and in tradable services (+2 percentage points), while the share of labor in nontradables will decline (–2 percentage points). The combined transit investments are expected to drive striking welfare improvements. And although the model assumes that total population and employment remain unchanged, such large welfare and productivity gains will­ —­ —­ in reality­ likely entice more working-age residents to join or rejoin Cairo’s labor market. In addition, improved welfare will likely attract higher in-migration to the city. Notes: 1. World Bank 2020. 2. Lall et al. 2021. Overview 15 job-creating investment. Cities with high produc- labor markets, more efficient spatial development, tion costs risk being out-competed in tradables­ —­ and more flexible land use; by supporting labor mar- both by imports at home, and by other exporters kets through new human capital and infrastructure internationally. This is the lesson of several studies investments to lower urban costs; and by protect- finding that, in manufacturing, nominal labor pro- ing labor market access for people who may face ductivity and price levels are similar (and converge social or financial barriers to commuting. Cities over time) around the world.18 The established pres- and countries can: ence of large agglomeration economies in tradables • Inform markets to motivate private action, setting makes it even harder for new locations to become high expectations with fair labor markets and land internationally competitive through efficiency: use policies that complement plans for connective underdeveloped locations lack the agglomeration infrastructure­—­and that anticipate indirect effects economies needed for tradables productivity, of future spatial evolution. Zoning and other reg- creating a vicious cycle of low agglomeration, low ulations must signal to private investors that the productivity, and low investment.19 use of scarce urban land can respond to market forces, evolving as the city evolves. To break into global tradables, cities need to take • Support productivity growth by investing in human coordinated action on several fronts to raise pro- capital and connective infrastructure to overcome ductivity and private sector expectations. Cities labor market fragmentation and lower urban costs. should deliver internationally competitive labor Connective investment should be tailored to local productivity, not through a race to the bottom on density, complemented by land value capture for labor standards, but through strategic investments sustainable financing, and complemented by and policies to lower living costs, improve amenities, flexible land use policies that allow private in- and increase the efficiency of urban labor markets­ —­ vestors to respond to the opportunities created. because today’s high nominal wages in developing • Protect labor market participation opportunities country cities reflect congestion and inefficiency for people who face commuting challenges­ —­women, (box 5). Beyond population density, developing —­ low-income residents, the elderly, and the disabled­ country cities will require high physical connectiv- through targeted investments to help them reach job ity, efficient land use, investments to raise human locations. For example, tailor transport modes to capital, and fair, efficient, contestable markets. the needs of particular populations. To ensure that reforms can overcome cycles of low NEXT STEPS FOR TOMORROW’S expectations, strategic multidimensional programs VIBRANT CITIES: INFORM, to raise urban productivity should be comple- SUPPORT, PROTECT mented by strong engagement with local and inter- national investors whose job-creating potential is Only with functional urban governance institutions­ demonstrably high. Unlike traditional policy tools —­especially for formal land rights and land market that allow the analysis of only one reform dimen- transactions­ can developing country cities posi- —­ sion, New Urban Models can help city leaders tion themselves for a vibrant future. Unless clear appraise multidimensional reform programs­ both —­ land ownership and tenure records are consis- to analyze the potential impacts on productivity, tently maintained and publicly available, nothing jobs, and inclusion, and also to plan for second-or- contemplated in this report will happen. The other der and unintended consequences. main challenges for developing country cities are to raise revenue and provide high-quality services. Developing city and country leaders can make These tasks, too, require governments to function their urban economies more productive by inform- in a professionally credible manner. City govern- ing investors about policies for more competitive ments must be accessible, expected to interact with 16 VI B R A N T C I T I E S BOX 5 HIGH COSTS UNDERMINE LABOR PRODUCTIVITY High costs and high nominal wages make developing cities less able to compete in global tradables. A recent meta-analysis of 1,200 estimates of urban productivity­ —­from 70 studies, covering 33 countries, over 1973 to 2020­ —­shows that while the elasticity of nominal labor productivity to density is similar in develop- ing and developed country cities, this seeming resemblance conceals dissimilar dynamics. In developing country cities, high nominal wages are often needed to overcome high urban costs from congestion, crime, and pollution (box figure 1). Rather than reflecting productivity, these wages trap developing country cities into nontradable vocations­ —­ often fueled by natural resource rents or a large government sector. BOX FIGURE 1 Urban costs are higher in developing countries than in advanced economies a. Pollution disamenity elasticity b. Congestion disamenity elasticity c. Crime agglomeration elasticity PM2.5 pollution (log) Hours lost to travel time (log) Homicide rate (log) 5 6 6 4 4 5 3 2 4 2 0 1 3 –2 7 8 9 10 11 7 8 9 10 11 8 9 10 11 Population density (log) Population density (log) Population density (log) High income Not high income Note: In panel a, city level pollution reflects particulate matter (PM2.5) data for 2014 for 298 cities (78 in non–high income countries). In panel b, congestion reflects the annual additional hours spent driving in rush hours, measured in 2018, and contains data for 337 cities (69 in non–high income countries). In panel c, homicide rate data reflect 124 cities (63 in non– high income countries) in 2015. In all three panels, population density is calculated from built-up area per person in 2015. The top and bottom 1 percent of population density observations are dropped for clarity. Source: Grover, Lall, and Timmis 2021. the residents whom they serve and to be account- • Trust consists in the belief that others are behav- able to them in clearly specified ways. ing cooperatively in a particular “game” of life, —­ society, or politics­ motivating similar coopera- For city leaders to implement reforms and raise tive behavior from the subject in question.20 revenue, they need to earn legitimacy and trust. Commonly recognized as the two elements of a By building trust and legitimacy at all layers social contract, legitimacy and trust are related in of governance, a city can move from a low-­ principle yet distinct in practice: performance–low-accountability equilibrium to a • Legitimacy is the ability of leaders to win com- high-performance–high-accountability equilibrium. pliance with new laws or public orders because people share a widespread belief that others are In relationships between city leaders and citizens­ —­ also complying. —­ that is, households, firms, and society at large­ trust Overview 17 and legitimacy can be built through local political and households through investments in connec- contestation and through two-way communication. tivity and service provision­—f­ or resilient and low In MENA cities, local elections with high participa- carbon urban development, for inclusion, and for tion strengthened political leaders’ motivation to productivity. For resilience and for a low carbon enact policies for the public good.21 In Ceara, Brazil, future, public investment may be needed­ but it —­ municipal leaders used communication campaigns, should be based on weighing costs and benefits, along with the accountability mechanism of local not just for sustainability but also for productiv- elections, to cement and reward political commit- ity. For inclusion, ensuring universal access to ments to healthcare reform and to motivate a new decent basic services is indispensable, though, cadre of public health workers; the result was a by itself, it is not enough. For productivity, it is large improvement in health outcomes.22 essential to efficiently connect fragmented urban labor markets, to let land use evolve with demand, In the relationships between city leaders and city and to harness the gains for urban agglomeration workers, trust and legitimacy can be built by judi- economies. cious use of monetary incentives, motivations, and norms. To effectively manage city resources and Policymakers, finally, can protect poor and vul- service delivery, the varied personnel in municipal nerable urban residents by extending public government departments need appropriate moti- safety nets and targeted assistance­ to help the —­ vation, an understanding of merit, and norms for unfortunate weather adverse shocks, and to open professional conduct­ —a­ ll elements of staff capac- escape routes from inequality traps. Getting ity. Instilling intrinsic motivation and professional markets right will unlock urban productivity and norms in city officials starts with recruitment: this growth, but the results may not be fair to every- must be based on merit and motivation, empha- one. Too many children are trapped in persistent sizing qualifications and commitment to public intergenerational poverty, not just by low neigh- services. Unfortunately, in many developing borhood school quality, but by weak social net- countries­—­including some in MENA­—­public sec- works and a lack of positive role models and peer tor jobs typically are obtained through connections interactions­ —­and, perhaps even more important, rather than through qualifications.23 by limited economic growth and job creation. The vicious cycles that create these inequality traps Once a city has begun to meet some of the basic are difficult to break even in developed countries, governance requirements for rising productivity especially when reinforced by social discrimi- and economic growth, policymakers can aspire to nation and divisions based on identity. Creative set higher expectations for residents and investors solutions are needed. Policies are also needed to by informing them about the city’s vibrant future. fill gaps in insurance markets, protecting people The challenge for city governments is to show and firms from risks when private action is not and persuade citizens that markets are governed feasible. by common rules, including price signals­ not by —­ privilege and influence. Where to reside or locate To recap: a business? How to invest resources wisely? What • First, inform market actors to motivate the pri- to do to protect and insure against risks? These are vate decisions and investments that make cities the questions that people and firms should be free vibrant. to answer­—­for the most part­ —­for themselves, given • Next, support markets through rigorous and what they bring to a city and what they find there. economically defensible public investments in connectivity and service provision. When cities are seen as credible informants by • Then, protect urban populations as needed in a market actors, policymakers can support firms targeted, timely, and timebound fashion. 18 VI B R A N T C I T I E S NOTES Dasgupta, S., S. Lall, and D. Wheeler. 2022. “Subways and CO2 Emissions: A Global Analysis with Satellite Data.” 1. Kocornik-Mina et al. 2020. Duarte, M., and D. Restuccia. 2010. “The Role of the 2. Gandhi et al. 2022. Structural Transformation in Aggregate Productivity.” 3. Gandhi et al. 2022. The Quarterly Journal of Economics 125 (1): 129–173. 4. Gandhi et al. 2022. Gandhi, S., M. E. Kahn, R. Kochhar, S. Lall, Y. Peng, and 5. UN Habitat 2020. V. Tandel. 2022. “Measuring the Short-Run Economic 6. Dasgupta, Lall, and Wheeler 2022. Impact from Urban Disasters in Rich and Poor 7. Dasgupta, Lall, and Wheeler 2022. Nations.” 8. Glaeser, Resseger, and Tobio 2009. Gelb, A., C. Meyer, V. Ramachandran, and D. Wadhwa. 9. Wodon et al. 2018. 2017. “Can Africa Be a Manufacturing Destination? 10. Peterson and Mann 2020. Labor Costs in Comparative Perspective.” Working 11. Park et al. 2022. Paper 466, Center for Global Development, Washing- 12. Chetty et al. 2014. ton, DC. 13. Chetty et al. 2014. Glaeser, E. L., M. Resseger, and K. Tobio. 2009. “Inequality 14. Van der Weide et al. 2021. in Cities.” Journal of Regional Science 49 (4): 617–646. 15. Alesina et al. 2021. Grover, A., S. Lall, and W. F. Maloney. 2022. Place, Pro- 16. Jones 2016; Gelb et al. 2017. ductivity, and Prosperity: Revisiting Spatially-Targeted 17. Nakamura, Sigurdsson, and Steinsson 2016; Policies for Regional Development. Washington, DC: Nakamura et al. 2019. World Bank. 18. Duarte and Restuccia 2010; Mano and Castillo 2015. Grover, A., S. Lall, and J. Timmis 2021. “Agglomer- 19. Venables 2021. ation Economies in Developing Countries: A 20. Alesina and Giuliano 2015; Algan and Cahuc 2014. Meta-Analysis.” Policy Research Working Paper 9730, 21. World Bank 2016. World Bank, Washington, DC. 22. Athar and Khemani 2022. Jones, P. 2016. “Closed for Business: High Urban Costs.” 23. World Values Survey 2018. Working Paper, Department of Economics, Oxford University. REFERENCES Kaw, J., M. Morad, Y. Li., W. Tammaa, M. Alam, and N. Daito. 2022. “Pathways for Decarbonizing Growing Alesina, A., and P. Giuliano. 2015. “Culture and Institu- Cities across the Urban-Transport-Energy Nexus in tions.” Journal of Economic Literature 53 (4): 898–944. Jordan.” Alesina, A., S. Hohmann, S. Michalopoulos, and E. Kocornik-Mina, A., T. McDermott, G. Michaels, and F. Papaioannou. 2021. “Intergenerational Mobility in Rauch. 2020. “Flooded Cities.” American Economic Africa.” Econometrica 89 (1): 1–35. Journal: Applied Economics 12 (2): 35–66. Algan, Y., and P. Cahuc. 2014. “Trust, Growth, and Mano, R., and M. Castillo. 2015. “The Level of Productivity Well-Being: New Evidence and Policy Implications.” in Traded and Non-Traded Sectors for a Large Panel In Handbook of Economic Growth, Volume 2, 49–120. of Countries.” Working Paper WP/15/48, International Amsterdam: Elsevier. Monetary Fund, Washington, DC. Athar, S., and S. Khemani. 2022. “Unlocking Institutions Lall, S. V., M. S. M. Lebrand, H. Park, D. M. Sturm, and of Urban Governance to Transform Cities.” Back- A. J. Venables. 2021. Pancakes to Pyramids: City Form ground paper for the Vibrant Cities report, World to Promote Sustainable Growth. Washington, DC: World Bank, Washington, DC. Bank. Chetty, R., N. Hendren, P. Kline, and E. Saez. 2014. Nakamura, S., R. Harati, S. Lall, Y. Dikhanov, N. Hama- “Where Is the Land of Opportunity? The Geography deh, W. Vigil Oliver, M. Rissanen, and M. Yamanaka. of Intergenerational Mobility in the United States.” 2019. “Is Living in African Cities Expensive?” Applied The Quarterly Journal of Economics 129 (4): 1553–1623. Economics Letters 26 (12). Overview 19 Nakamura, E., J. Sigurdsson, and J. Steinsson. 2016. “The Wodon, Q., C. Montenegro, H. Nguyen, and A. Ona- Gift of Moving: Intergenerational Consequences of a goruwa. 2018. Missed Opportunities: The High Cost of Not Mobility Shock.” Technical Report, National Bureau of Educating Girls. Washington, DC: World Bank. Economic Research, Cambridge, MA. World Bank. 2016. “Population living in slums (% of urban Park, H., N. Lozano-Gracia, G. Rossitti, and O. D’Aoust. population) – Cote d’Ivoire.” DataBank. Washington, 2022. “A Global Review of Intra-Urban Spatial DC: World Bank. Inequality.” World Bank: Washington, DC. World Bank. 2017. Beyond Scarcity: Water Security in the Peterson, D., and C. Mann. 2020. Closing the Racial Inequal- Middle East and North Africa. MENA Development ity Gaps: The Economic Cost of Black Inequality in the US. Series. Washington, DC: World Bank. Citi GPS: Global Perspectives & Solutions. World Bank. 2020. Convergence: Five Critical Steps Toward UN-Habitat (United Nations Human Settlements Integrating Lagging and Leading Areas in the Middle East Programme). 2020. World Cities Report 2020: The Value of and North Africa. Washington, DC: World Bank. Sustainable Urbanization. Nairobi, Kenya: UN-Habitat. World Bank. 2022. Egypt: Country Climate and Devel- Van der Weide, R., C. Lakner, D. Mahler, A. Narayan, and opment Report (English). Washington, DC: World R. Ramasubbaiah. 2021. “Intergenerational Mobility Bank. http://documents.worldbank.org/curated/en​ around the World.” Policy Research Working Paper /099510011012235419/P17729200725ff0170ba05031a8 9707, World Bank, Washington, DC. d4ac26d7. Venables, A. 2021. “Winners and Losers in the Urban Zimmerman, S. 2019. “Elite Colleges and Upward Mobil- System.” In: Urban Empire: Cities as Global Rulers in the ity to Top Jobs and Top Incomes.” American Economic New World. Routledge. Review 109 (1): 1–47. MAKING CITIES RESILIENT reducing vulnerability to climate-related disasters. rebounding from disasters and pandemics. 1 Making cities resilient Economic resilience is an economy’s capacity Sustainable Development Group provides a to prepare for, mitigate, and adapt to adverse comprehensive treatment of urban resilience in shocks that affect people’s lives and livelihoods.1 response to different types of shocks.) The shocks covered in this chapter—linked to climate-related weather events and the Covid-19 The Covid-19 pandemic, in all of its successive pandemic—harm incomes, assets, and living con- waves and variants, has ended millions of lives and ditions. These shocks typify two broad categories cut deeply into livelihoods worldwide. At the time of socioeconomic and natural shocks that have of writing (mid-2022), no real end is in sight. The dominated policy and popular discourse since the danger of this virus may wane eventually. pandemic’s onset. And because different types of shocks can hit simultaneously, compounding their This chapter provides an economic framework effects, it is important to consider multiple aspects for policymaking to enhance urban resilience. It of resilience. This chapter focuses on individual combines fundamental insights from economic urban areas as well as a country’s system of cities. theory with new sources of empirical evidence on economic development, private action, and public Among climate-related shocks, flooding has interventions and with new econometric work, the taken the most devastating human toll. In 2019, framework absorbs the experiences of recurrent floods accounted for 44 percent of deaths, fol- flooding around the world as well as various cities’ lowed by extreme temperatures (25 percent), and efforts to cope with Covid-19 and the economic storms (22 percent). Floods forced over 13 million contractions induced by pandemic policies. people into poverty, followed by drought (8 mil- lion).2 Flooding along with droughts also represent FLOODS AND URBAN RESILIENCE the top climate-related shocks in Middle East and North Africa (MENA) countries, the geographic Today, more than 4 billion people around the “deep dive” in this report. (A recent flagship report world—over half the global population—live in cit- on climate change and cities by the World Bank’s ies; by 2050, with the urban population more than 21 22 VI B R A N T C I T I E S doubling, nearly 7 of 10 people will live in cities. a medium sea-level rise scenario (RCP4.5 SSP2) As cities increase in population and physical size, of 1 percent of GDP annually by 2030, growing to more people and assets will be exposed to shocks. about 3 percent in by 2100. Coastal flooding sce- The interplay of economic and physical geography narios conducted for the Egypt Country Climate is one major reason for high hazard risks in urban and Development Report (CCDR) estimate that in areas. Many cities have historically emerged at Alexandria, rising sea levels will also lead to saltwa- locations with good accessibility or favorable nat- ter intrusion, inundation, and erosion, which will ural endowments, such as a river crossing, a coastal amplify climate change impacts on water available location, or fertile volcanic soils. Those geographic for agriculture and affect the quality and availabil- settings are often associated with a higher proba- ity of drinking water.5 Further, Egypt’s urbanization bility of hazard events including floods, cyclones, is centered on 14 medium-to-large cities and urban and volcanic eruptions. agglomerations, which account for 72 percent of the country’s urban population. Greater Cairo Many of the world’s largest cities are on the coast, alone contributes about 50 percent of Egypt’s GDP, which once offered ease of trading and plentiful and coastal governorates around 25 percent. Many food supply, but now presents increasing threats people in these 14 cities are exposed to at least one from sea-level rise and flooding. While it may major climate risk including flooding, heat stress, take another 100 years for the sea level to rise by air pollution, desertification, and, for coastal areas, 1.5 meters, this slow but accelerating onset event sea-level rise (figure 1.1). This concentration can is already causing more frequent flooding, with further exacerbate urban population vulnerabili- higher intensity and duration. In the summer of ties given that informal settlements covered around 2021, a damaging flood hit Germany, and urban 38 percent of the built-up area and housed around flooding submerged China’s Zhengzhou with eight 5.2 percent of the urban population in 2019.6 inches of water in just one hour. India and Belgium also faced devastating floods over the past year. Across the MENA region, pluvial and fluvial flood- Globally, from 2012 to 2018, an annual average of ing will severely undermine urban livability and 2,600 floods inflicted large economic losses.3 hurt vulnerable populations. Resilience of MENA cities against even low-intensity rainfall is gener- Econometric analysis for this report shows that ally weak due to gaps in and inadequacy of urban floods impede economic activity.4 The effect, drainage infrastructure; poor planning, operation, unsurprisingly, is higher in cities in low-income and maintenance; and a lack of early warning countries, with night lights (a proxy for economic systems. In Cairo, the built-up area exposed to a activity) falling by 8.3 percent, against 1.4 percent pluvial flood hazard up to a 1-in-1,000-year return in high-income countries. The findings are robust period rose from about 18 km2 in 1985 to 40 km2 in to using extreme precipitation events as a measure 2015, and in Alexandria, from 9 km2 to 24 km2 over of floods. In addition, the “recovery dynamic” the same period. Coastal cities such as Alexandria, results indicate that economic activity is restored Ismailia, Port Said, and Suez face the highest level to pre-disaster levels within one month in cities in of climate and economic risks.7 high-income countries, but two months in those in low-income countries. In Jordan’s capital Amman, more than half of the at-risk households are low income with low One thing is clear: mother nature will throw even capacity to cope with floods.8 About 66,900 of harder punches in the years ahead. Consider Egypt, its households are potentially exposed to pluvial which ranks fifth in the world on the potential eco- flooding, including 34,300 low-income house- nomic impact of sea-level rise’s potential economic holds. Although the risks faced by low-income impact on urban areas, with damage costs under households are not disproportional—about the Chapter 1  Making cities resilient 23 FIGURE 1.1 PERCENTAGE OF POPULATION EXPOSED TO 1, 2, OR 3 OR MORE CLIMATE CHANGE HAZARDS 3 or more hazards 2 hazards 1 hazard Zagazig Tanta Suez Port Said Mansoura Luxor Kafr El Dawwar Ismailia Faiyum El Mahalla El Kubra Cairo Asyut Aswan Alexandria 0 20 40 60 80 100 Percent Source: World Bank 2022a. same percentage of households from each income vulnerability of these areas and their populations, tier is subject to potential pluvial flooding—the suggesting unevenly distributed socioeconomic exposure is spatially uneven. The city center has impacts of climate-related disaster events. Neigh- a higher concentration of at-risk low-income borhoods like Zouagha in Fez are also projected to households, which doubtless are less equipped to experience the highest population growth, perhaps deal with flood events than high-income families. tripling by 2045—a trend likely to further increase Amman’s settlements are projected to further the neighborhood’s vulnerability to flooding.11 expand into places facing current flood risks. Under both SSP1 and SSP5, the area of at-risk set- Beyond MENA, poor people in developing coun- tlements will reach 26 km2 by 2030 and 31 km2 by try cities take a harder hit to their living standards 2050. Climate change and urban development may than non-poor people. In Mumbai, India, poor increase further the exposure of land and built-up people report losing 80 percent of their assets after areas to flooding.9 a disaster, and nonpoor people 40 percent.12 In Bangladesh, poor people report losing 70 percent In Morocco, economically and socially vulnerable of their incomes/assets after a disaster, non-poor groups—including women, the elderly, persons people 45 percent.13 And in Accra, flooding causes with disabilities, and others—are disproportion- heavy asset loss for a large share of poor house- ately exposed to climate-related disaster hazards, holds, despite having a similar probability of hitting especially flooding (box 1.1). Urban resilience nonpoor households (figure 1.3). diagnostics and mapping supported by the World Bank in the cities of Fez and Mohammedia suggest Floods also trigger internal displacement, and that poorer areas, often in peripheries with poor “climate migrants” are likely to resettle in riskier access to services and mobility, are disproportion- urban areas. In Morocco, according to the Inter- ately exposed to floods.10 Moreover, these areas are national Displacement Monitoring Centre, floods usually less likely to be close to health centers and caused more than 17,000 internally displaced peo- fire stations, have poorly connected transportation ple between 2000 and 2020. In addition, up to 1.9 networks, and have only limited access to the elec- million climate migrants in that country (due to trical grid (figure 1.2). These factors increase the increasing water scarcity in the decades to come) 24 VI B R A N T C I T I E S BOX 1.1 GENDER DISPARITIES IN DISASTER RISK REDUCTION Floods, droughts, and earthquakes may be gender-neutral, but their impacts are not. Due to gen- der disparities in Morocco, women suffer from higher exposure and vulnerability and more limited preparedness and coping capacity than men in disaster and climate-related events. These barriers emerge from Moroccan women’s limited access to disaster risk information, early warning systems, and post-disaster interventions. Moroccan women have higher illiteracy rates, especially in rural areas (61 percent against 35 percent for men). They also have restricted mobility due to social and cultural norms that prevent them from accessing disaster risk reduction services. Additional factors also limit their preparedness and coping capacity. They have lower labor force participation (22 percent against 71 percent for men). They have a higher unemployment rate (18 per- cent against 11 percent for men). They have a persistent wage gap (at similar academic and professional levels as men, they earn about 17 percent less than their male counterparts). And they have less access to and control over assets (only 7 percent of Moroccan women have access to immovable property, a figure that drops to 1 percent in rural areas), depriving them of financial security and economic independence. Source: Ennaji 2018; World Bank 2015; Morocco High Commission for Planning 2019, 2020; World Bank 2019; Bordat and Kouzzi 2018. FIGURE 1.2 ACCESSIBILITY TO SOCIAL INFRASTRUCTURE WITH AREAS AT RISK OF FLOODING IN FEZ Source: World Bank 2021. Chapter 1  Making cities resilient 25 FIGURE 1.3 FLOODING IN ACCRA LEADS TO HEAVY ASSET LOSS FOR A LARGE SHARE OF POOR HOUSEHOLDS Percent of households Affected by flood Lost at least 10 percent of assets 50 40 30 20 10 0 Poorest 2nd quartile 3rd quartile Richest Source: Erman et al. 2018. are likely to resettle in vulnerable urban neighbor- People can adapt to climate change in three hoods. According to UN-Habitat, close to 9 percent ways: private investments in prevention and self- of the country’s urban population already lives in protection; public investments in infrastructure slums.14 Moreover, some of the main climate immi- and social protection; and insurance. They can: gration hotpots identified by the modeling exercise • Make private investments to protect homes and live- in the Groundswell report15 were identified as lihoods from risk exposure. Private self-protection “extremely flood sensitive” areas.16 That could fur- often amounts to migrating from hazard-prone ther reinforce this adverse feedback loop between areas. An alternative is to make adaptive invest- water scarcity and exposure to natural hazards for ments in land and structural improvements—for the poor and vulnerable. example, by raising homes on stilts to put them above expected floodlines. Private adaptive re- NEW EVIDENCE ON ECONOMIC sponse can simply include pre-cautionary sav- DEVELOPMENT, PRIVATE ACTION, ings for rainy days as well. AND PUBLIC INTERVENTIONS • Rely on public investments to protect localities and IN URBAN RESILIENCE safety-nets to protect the vulnerable. Sea walls and levees are among the investments most dis- Households and firms are not, however, passive cussed today. And social insurance and safety victims. They can—and where possible do —act to nets in urban areas gained traction during the reduce their exposure to natural disasters, and to Covid-19 pandemic. buffer against losses from disasters and pandem- • Make private investments in market insurance to ics alike. Most urban residents exposed to natural reduce economic losses from disasters to guarantee hazards can take steps to reduce risks to their lives, that they receive a payment if a disaster hits. This livelihoods, and property. The options available option requires the presence of a robust insur- to the poor and vulnerable urban residents may ance market which is not commonly the case in on the other hand be limited. Given that damage most developing countries. will inevitably occur somewhere in a given year or decade, a resilient society will have insurance These insights (figure 1.4) stem from seminal work mechanisms and safety-nets in place to help those done 50 years ago by Isaac Ehrlich and Nobel affected rebuild their lives. Laureate Gary Becker, to study interactions 26 VI B R A N T C I T I E S FIGURE 1.4 THREE WAYS TO ADAPT TO CLIMATE CHANGE: AN EHRLICH AND BECKER FRAMEWORK PRIVATE investments for self-protection Does improved risk Do public interventions (households) pricing through market strengthen or discourage insurance enhance private actors’ adaptive self-protection—or do policy investments? buyers invest less to protect themselves? PUBLIC PRIVATE investments interventions Does public investment in market crowd in or crowd out insurance private insurance? Source: Adapted from Kahn and Lall 2021. and tradeoffs among private investment for The advent of digital technology to minimize self-protection, public investment, and market adverse selections and moral hazards and of global insurance.17 insurance companies to overcome the danger of correlated shocks present opportunities to develop Private markets to insure against climate shocks and expand insurance markets to deal with climate have been limited in developing countries. The shocks in developing countries. The policies to take drought and flood shocks to crop production advantage of these opportunities should be a prior- have been a part of life for farmers around the ity. But they fall in the jurisdiction of national gov- world. Despite repeated efforts, the costs of ernment. Keeping this in mind, this chapter focuses such insurance remain unusually high. The high on the roles of private investment in self-protection costs are the result of familiar adverse selection, and of public investment in infrastructure in com- where farmers most exposed to these shocks bating climate shocks—mainly flooding. self-select into insurance program—and moral hazard, where farmers with insurance exert less This section provides new evidence on the roles effort to protect crops. The index-based insur- of economic development, private investments ance uses a village level index to decide payouts for self-protection, and public investment in infra- and thus avoids the adverse selection and moral structure—as well as strong urban management—as hazard. But take-up of index-based insurance has forces for shaping urban climate change resilience. not been widespread, partly due to the inability Some places can cope better through private adap- of these products to overcome ambiguity aver- tive action. Other places may have governments sion by poorer farmers, who prefer the inferior that develop and implement strong plans and but known outcomes over uncertain outcomes invest more in infrastructure. Still other places promised by the insurance. The uncertainty over may have governments that act faster, both before insurance payouts arises because climate shocks a disaster hits and after (including investments in are correlated local shocks, raising doubts about resilience). To explore these intuitions, research for the solvency of insurance companies, particu- this chapter examines three questions: larly when coverage by an insurance company is • Are people getting out of harm’s way (private localized. self-protection)? Chapter 1  Making cities resilient 27 • Does flood protection infrastructure investment self-protect and cope. Natural hazards such as (place-based public intervention, specifically floods can lead to power failures, disrupt essen- dams) enhance resilience? tial services, damage property, temporarily close • Does economic development—including in- offices and factories, and in all cases affect the come as a proxy for government quality—en- normal functioning of economic life. hance resilience? 18 A key variable for measuring both resilience These three questions combine attention to the indicators is satellite-observed night lights data. first two adaptation strategies in the modified Such data—collected by satellites at a uniform and Ehrlich and Becker framework above (see figure disaggregated spatial scale for the whole world— 1.4)—private self-protection and public interven- allow comparisons of economic activity on a fine tions. Keep in mind the key role of prosperity in spatial scale and circumvents the issue of poorly enhancing resilience by households, firms, and measured or missing estimates of GDP at a local governments alike. (Market insurance will be level. Such night lights data have a spatial resolu- picked up in future work.) tion of 465 meters × 465 meters (grids) and have been provided monthly since April 2012. For these The research uses a global city dataset of 9,468 reasons, multiple studies use night lights as a proxy cities, drawing on the Global Human Settlement for urban economic activity, population density, Urban Center Database (GHS-UCDB).19 The urban and built-up area.21 extent of cities includes city centers and suburbs. A similar approach was used in the World Bank Using the global dataset of cities, an economet- report From Pancakes to Pyramids: City Form for Sus- ric model is estimated to examine the impact of tainable Growth.20 floods on economic activity.22 For example, Visible Infrared Imaging Radiometer Suite (VIIRS) night The impact of a flood on urban economic activity lights show Wuhan, China before and after floods depends on the resilience of the city’s core infra- in 2016, with the lights dimmer for May, after the structure and the ability of firms and families to floods.23 FIGURE 1.5 NIGHT LIGHTS BEFORE AND AFTER FLOODS IN WUHAN, 2016 Note: Average monthly night light intensity in Wuhan of Hubei province, China. Wuhan suffered from major floods between May and July 2016, causing economic losses estimated to be greater than US$350 million. Source: Gandhi et al. 2022. 28 VI B R A N T C I T I E S Are people getting out of harm’s way the number of extreme precipitation events that (private self-protection)? struck a city, the authors construct a distribution of monthly rainfall for 1958–2015 for each city, and Natural disaster–prone areas are likely to be hit then add up the total number of 90th percentile time and again. So it is therefore surprising that events in 1970–2012. The assumption is that the recent empirical work on disasters finds that extreme precipitation events led to flooding, and incumbents tend to remain in a shocked area therefore, a higher number of events in 1970–2012 unless their housing is destroyed.24 But when peo- would indicate recurrent shocks, which they ple do get out of harm’s way, they may find a silver hypothesize should enable a city to adapt. The lining. According to some empirical studies, people results: Cities that had flood events in the past see displaced from their origin by a natural disaster a lower impact of floods than cities with no such enjoy improved material living standards in later experience. years.25 Does flood protection infrastructure Yet, residents of cities in developing countries— investment (place-based public especially the poor—are likely to be stuck in harm’s intervention, specifically dams) enhance way by various economic disincentives to migra- resilience? tion. Cities experiencing a higher frequency of extreme events between 2000 and 2015 saw lower Local resilience can be built through infrastruc- population growth in the same period,26 suggesting ture investments that reduce the vulnerability of that cities with recurrent floods lose some of their particular areas. Dams and levees, for example, attraction for current and potential residents. This may protect some urban districts from flooding. slowing effect on population growth is, however, But decisions about which places to protect can statistically significant only for cities in high- be fraught with tension: should the government income countries. protect richer areas because they are the most productive? Or poorer areas because their inhab- So, what explains people’s choice to stay put itants stand to suffer the most from flooding? No despite high exposure to climate-related risks? government can build levees everywhere; and sea- Two likely factors are high migration costs27 and walls, the most ambitious solution, are not always built-up local social capital.28 Further, deaths from feasible or affordable. Reliance on place-based urban floods are lower (1.1 percent) in risky cities, protection can also create a moral hazard that may but only in high-income countries. For vulnerable be counterproductive in the longer term. cities in low-income countries, the death rate is in fact 2.7 percentage points higher. And if poor peo- Evidence to inform these decisions is scant, how- ple are least likely to vote with their feet by moving ever, and this report provides new empirical evi- to higher ground, climate change will likely cause dence on dams in China, India, Mexico, and the the poor to remain more exposed than others. United States. The countries contain some 3,600 cities of the total dataset of cities. China (a rapidly People and public authorities learn from expe- growing economy) and the United States (a rich rience and prepare better against disasters. Do economy) have 71 percent and 67 percent of their public authorities and citizens with experience of river cities protected by a dam. Surprisingly India, extreme weather events know what to expect and a lower-middle-income country, has the exact same can plan better for future events. An alternative percentage of cities with dams as the United States hypothesis is that places that face repeat flooding (71). Only in Mexico does this number fall, to 58 experience disinvestment as the repeat events act percent. Annex table A1.1 presents the key findings as a “tax” on capital investment.29 To determine from the econometric analysis. Chapter 1  Making cities resilient 29 Cities with dams face more floods, a result of flooding (panels b and c). By January 2016, the driven by high-income countries. This result is intensity of lights started recovering (panel d). unsurprising given that dams are more likely to be placed in areas at greater risk. An econometric analysis was conducted to assess the length of time it takes cities to recover from Controlling for whether a city is at risk from floods. In addition to “all” flooding events, the floods, cities protected by dams are found to have analysis also considered the impact of “extreme” higher population growth, though some heteroge- precipitation—that is, in the 95th percentile rela- neity is observed among cities protected by dams, tive to a city’s 60-year history. The focus on extreme depending on whether they are exposed to flood events is motivated by broader discussions of “tip- risks. Those at risk from floods see lower popu- ping” points when climate-related shocks exceed lation growth than those not at risk from floods. particular thresholds (annex table A1.2). On average, cities protected by dams have a sig- nificant 1.5 percentage point lower population The key finding is that the negative impact of floods growth rate. This result too is also being driven persists for at least one month after the disaster mostly by cities in high-income countries. But this and is more severe for low-income than high- masks considerable heterogeneity between high- income countries. The striking findings for extreme and low-risk cities. Low-risk cities with dams precipitation are the larger impact overall and for actually experience a higher population growth cities in high-income countries. Considering that rate, and the risky cities with dams experience infrastructure in developed countries is abundant lower population growth. and of higher quality than in developing countries, this finding has multiple implications for measures Cities protected by dams see a smaller decline in to enhance infrastructure resilience. night lights during floods. The effect of floods is negative and significant, with night lights falling by Still, the econometric analysis highlights that cities 4.2 percentage points after a flood. However, the in richer countries experience a lower impact from effect is attenuated by 1.7 percentage points in cit- floods. On average, cities that suffer from floods ies with dams. Surprisingly, no mitigating effect of see economic activity decline by around 3 percent, dams is seen during extreme precipitation events. but those in low-income countries see it decline by It is likely that dams are beneficial in mitigating the 5–8 percent. effects of riverine flooding that may be caused by factors upstream. This is not the case with extreme THE COVID-19 PANDEMIC precipitation, however, which falls directly on the AND CITY RESILIENCE city, and may cause waterlogging in low-lying areas. This last result presents a puzzle that merits future The sudden, global onset of the Covid-19 pan- research. demic greatly disrupted human activity globally. Governments almost everywhere adopted counter- Does economic development—including measures to contain its spread and the related eco- income as a proxy for government nomic damage. The pandemic hit urban economic quality—enhance resilience? activity particularly hard, given the central place of cities in global transport networks and their After a flood, how many months does it take for greater susceptibility to contagion, given their high the city to recover? Consider Chennai, the capital densities. of Tamil Nadu state, which suffered from heavy floods in November–December 2015 (figure 1.6). Whether the concentration of economic activity The light intensity declined during the months in large cities would survive the pandemic was 30 VI B R A N T C I T I E S FIGURE 1.6 NIGHT LIGHTS BEFORE AND AFTER FLOODS IN CHENNAI, OCTOBER 2015 TO JANUARY 2016 Note: Average monthly night light intensity in Chennai, India. The floods in November–December 2015 caused economic losses estimated at US$1 billion. Source: Gandhi et al. 2022. one of the issues most debated at the start of determinant of epidemic diffusion, because den- the pandemic. The role of urban density during sity is generally proportional to the number of such a crisis is ex ante ambiguous as there are contacts of an infected agent, and is especially true two opposing forces at play: “health-preserving in dense, developing country cities where density behavior” and a “resilience effect.” First, denser is often associated with crowding in slums, and cities are more prone to contagion risks during where social distancing is difficult.32 an outbreak,30 which nudges economic agents to reduce economic activity more in denser areas Second, the strong agglomerations that underly to avoid health risks (that is, health-preserving denser cities create more resilience to large neg- behavior, including both state-sanctioned as well ative shocks—city resilience. Long-run persistence as voluntary reductions of activities that require in spatial distribution of activity can arise because movement outside the home and interactions with of, say, locational fundamentals (that is, natural others). Bisin and Moro (2022) show that incorpo- advantages), sunk investments (man-made advan- rating a spatial dimension into the now well-known tages), agglomeration effects (the direct effect “SIR”31 model reveals that city density is a crucial of population scale), or institutions.33 Cities can Chapter 1  Making cities resilient 31 also benefit from greater industrial diversifica- contagion risk are based on satellite imagery along tion and higher human capital. In the developing with granular data on building heights. world, denser cities also tend to have more access to WASH and health services—and higher local Economic impacts of Covid-19 are worse capacities to deal with shocks—than less dense in developing country cities locations. If forward-looking economic agents believe that denser cities are more resilient, this Covid-19 pandemic impacts are amplified in devel- may result in an increased concentration of eco- oping country cities. While both the disease and nomic activity and investments in these locations the countermeasures hit urban economic activity, in the face of a large “symmetric” shock that affects the impact on city economies has been much all cities. For these reasons, density may also have more severe in developing countries than devel- a positive role in limiting the adverse economic oped countries (figure 1.7). Econometric analysis impacts of Covid-19. suggests that cities in developing country cities have experienced deeper economic contraction This section provides new evidence on the differ- (measured by the loss in night lights) and slower ential economic impact of Covid-19 on economic recovery than developed-country cities. The large activity in developed-country and developing share of jobs in nontradable sectors, along with country cities—and on the vulnerability and con- underdeveloped public services and infrastruc- tagion risk for people living in crowded informal ture, have increased the vulnerability of cities in settlements within developing country cities. The developing countries. analysis combines spatial data to measure aggre- gate economic activity in cities, the progress of For MENA, cities suffered a massive fall in the Covid-19, and the policy measures implemented intensity of economic activity with the onset of in response. The primary data are nightlights (as the pandemic and the launch of strict lockdowns in the previous section) using the VIIRS Day-Night and other nonpharmaceutical interventions (NPIs) Band (DNB) monthly composites obtained from (figure 1.8). Correcting for “normal” monthly vari- the Earth Observation Group (EOG) at the Col- ation for each city, the median city’s lights were orado School of Mines. The data cover the study about 10 percentage points less bright in March period April 2012 through March 2021. The data on 2020 than their pre-Covid levels. For a sample of FIGURE 1.7 COVID-19’S IMPACT IS MORE SEVERE IN DEVELOPING COUNTRIES Median change in ln(SOL) from trend paths 1.0 0.5 High-income countries 0.0 –0.5 Lower-income countries –1.0 Dec. Dec. Dec. Mar. Dec. 2017 2018 2019 2020 2020 Source: Khan et al. 2022. 32 VI B R A N T C I T I E S FIGURE 1.8 CITIES IN MENA SUFFERED A MAJOR ECONOMIC LOSS FOLLOWING THE ONSET OF THE PANDEMIC Percentage point change since December 2019 5 0 –5 –10 Initial impact: 10 percentage point Apparent “double dip” decrease in nighttime light intensity in light intensity relative to pre-Covid-19 trend –15 Jan. Feb. Mar. Apr. May June July Aug. Sep. Oct. Nov. Dec. 2020 Note: Based on analysis of a sample of 140 cities whose population in 2015 was at least 200,000. Source: Khan et al. 2022. 140 cities spanning MENA, monthly composites Differentiating between crowded and livable den- of cloud-free satellite images and changes in night sities has been important for developing country lights (a proxy for changes in economic activity) cities in managing the pandemic, given their very assessed the role of lockdowns and other NPIs. limited resources. In the first year of the pandemic, As cities began to ease their NPIs, the intensity of prevention through containment and social dis- lights began to recover slightly in April, and more tancing, as well as frequent handwashing, appeared rapidly in May. But cities felt a double dip in eco- to be the only strategy. To slow transmission to nomic activity as new strains of the virus emerged avoid overwhelming their health system, especially and countries adopted policies to curb the spread. beds in intensive care unit and ventilators, would have led to a collapse of the health system if large Vulnerable groups in slums and informal numbers of people had been infected at the same settlements have borne the burden of time. Figure 1.9 shows a scatterplot of population Covid-19 density and city incomes, for 400 cities around the world. Population density is people per built-up Cities in low- and lower-middle-income cities tend area of the city. to be crowded, but not as economically dense as in more developed parts of the world.34 The crowd- Two striking features emerge. First, population edness is most apparent in their slums, with more densities decline as incomes rise. Kinshasa is a than 35 percent of urban residents in developing crowded and fast-growing city of more than 14 country cities in slums. In Sub-Saharan Africa, million inhabitants that concentrates the highest the share is 60 percent.35 In Dar es Salaam, 28 number of people living in poverty in the Dem- percent of residents live at least three to a room.36 ocratic Republic of Congo. Mumbai and Karachi In Abidjan, the equivalent figure is 50 percent.37 are four to five times more densely populated than Most residents in these locales live in substandard Shanghai, Tokyo, or London. Second, naïve mea- housing, lack open space, and have inadequate sures of population density can conflate crowding infrastructure. In South Africa, only 44 percent of with livable densities. Kinshasa, Mumbai, and people have access to water inside their house, and Hong Kong SAR have similar population densities only 61 percent have access to a flush toilet.38 but very different per capita incomes. Chapter 1  Making cities resilient 33 FIGURE 1.9 POPULATION DENSITY AND CITY INCOME, 2015, 400 CITIES WORLDWIDE Density (population per square kilometer of built-up space, log scale) 150,000 100,000 Mumbai 50,000 Kinshasa Hong Kong Karachi Dhaka Lagos Beijing Shanghai 10,000 Tokyo London 5,000 2,000 Dallas 500 1,000 5,000 10,000 40,000 Annual income per capita, 2015 (US$, log scale) Source: Lall et al. 2021. Unlivable, crowded cities and neighborhoods in Kinshasa’s high COVID-19 risk comes from both low- and middle-income countries were hit hard- the expanse and intensity of its hotspots, which are est by the pandemic and economic contraction. seen throughout the city (figure 1.10). Most people simply did not have the capacity to socially distance themselves, leading to “contagion Finally, compounding shocks can increase vulner- hotspots” in cities—areas where people cannot ability. A particular challenge in many developing keep the required social distance of 2 meters. country cities is the coincidence of natural hazard FIGURE 1.10 COVID-19 CONTAGION HOTSPOTS WERE THROUGHOUT KINSHASA IN 2020 Note: The color represents the intensity of the hotspot from low intensity (yellow) to high intensity (red), but all hotspots are already high-risk locations. Source: Lall and Wahba 2021. 34 VI B R A N T C I T I E S and health risk. In Dhaka, more than 2.5 million protection—reinforcing one another—policies people live in Covid-19 contagion hotspots with to enhance both will, together, add up to more high flood risk (figure 1.11). In Dar es Salaam and than the sum of their parts. A property owner Nairobi, more than half a million people face such who puts her house on stilts in a levee-protected compounded shocks. area is better adapted to sea-level rise thanks to the synergies between the investments. Progress ENHANCING RESILIENCE over time in civil engineering means that such THROUGH SELF-PROTECTION, new ideas can diffuse across the world. PUBLIC INVESTMENT, AND • Private investments for self-protection made less ur- MARKET INSURANCE gent by private market insurance. If global insurers offer insurance buyers a good premium price on For policymakers, what matters is the interaction coverage, policy buyers will not need to invest among three strategies for coping with risk. The as much in self-protection. The growth of global adoption of one can affect—positively or nega- insurance giants creates new opportunities for tively—uptake of the others, leading to indirect insurers seeking to expand in the developing effects and potentially inefficient outcomes. The world by offering policies and using their “deep framework in figure 1.4 offers a heuristic for exam- pockets” to diversify their risk exposure using in- ining every resilience policy proposal in light of its struments such as catastrophe bonds. While the strategic interactions with the others. In some cir- rise of global financial markets helps to protect cumstances, well intended policies could actually more individuals from the ex-post losses asso- discourage private adaptation efforts.39 Consider ciated with place-based climate-related shocks, four interactions: it may also reduce private adaptive investment. • Private investments for self-protection comple- • Private incentives for self-protection discouraged by mented by public infrastructure investments. If public infrastructure investments. If an at-risk lo- private investment and public infrastructure cation invests in spatial protection, this could investments provide complementary forms of attract more people to move to the area. This FIGURE 1.11 COMPOUNDED SHOCKS FROM THE OVERLAPPING OF FLOOD RISK AND COVID-19 TRANSMISSION RISK Percent of population Millions of people 12 3.0 10 2.5 8 2.0 6 1.5 4 1.0 2 0.5 0 0.0 g M a di ang a am th ne ee u N n M obi a La s M re Za to i om ra Ka sa la Sa r r Ka r ch es wa sa a go k ub ab uj d w on pa c ib ho u ba ha u an to ra as Ab Ac la r ap s La Bh Ab P Ad al nz ai ag m D ane D ak m itt s Fr M Ch Ka ar Source: Lall and Wahba 2021. Chapter 1  Making cities resilient 35 dynamic can increase the population’s overall nation builds infrastructure to protect its produc- risk exposure, while also possibly enhancing ag- tive hubs, the economic growth that this creates glomeration economies. If the national govern- can be taxed and used for redistribution. Poor ment subsidizes disaster insurance, this creates people in other areas can be urged to move to the an incentive for individuals to invest less in their productive places. Climate change will pose deep own self-protection. issues of how to protect poor people who seek to • Private investments in market insurance discouraged remain in less productive and increasingly at-risk by public infrastructure investments. Public invest- places. ment to protect locales from climate change impacts can crowd out efforts by private sector Maximizing synergies among the resilience insurers to invest in the research to price risk channels­ —­of self-protection, self-insurance, pub- and incentivize asset owners to invest in their lic infrastructure, government safety nets, and own self-protection. Given the risks in design- market insurance—will be crucial in determining ing new products and the required upfront fixed the medium-term costs of climate change. The costs, for-profit firms will be less likely to incur market approach to adaptation emphasizes having these costs if they are competing with the gov- up-to-date information on emerging risks and pro- ernment selling a subsidized version of the same viding incentives and a menu of strategies to help product.40 affected individuals, firms, and governments cope with these risks. Improvements in climate science, These interactions illuminate how economic thanks to satellite deployment and big data pro- growth, and policies to promote economic growth, cessing, will help. make people more resilient. Poor people are less likely to afford self-protection strategies ranging INFORM, SUPPORT, PROTECT from food and medicines to air conditioning and high-quality housing that help to maintain one’s Urban economic productivity and growth are fun- health and well-being. Such person-based strate- damental for building resilience. Rising incomes gies provide the resources people need to face their make firms and households better able to adapt climate challenges. As poor people grow richer, to climate change hazards. Firms can relocate to they have the resources to finance migrating to less exposed places. People can migrate to protect geographic areas that offer them a higher quality themselves. Both can pool risk through insurance. of life. Person-based resilience strategies also Prosperity can also help finance needed public have implications for development, as individuals investments in sustainability and interventions to choose where they want to live. protect particular places from risk. Place-based public interventions for resilience Governments can set expectations for urban resil- raise questions of spatial resource allocation, ience in three ways: informing private actions, sup- leading to familiar efficiency-equity tradeoffs. Any porting firms and households with investments in government investment in infrastructure targeted broad-based infrastructure and service provision, to reduce risk exposure will be place based. A and protecting poor and vulnerable people. levee is built in a specific location. Even rich gov- ernments do not have the resources to build levees Inform: publicizing risks to help firms and everywhere. Should governments use their scarce individuals take adaptive actions funds to protect productive, populated places or target their poor population? This question shows To support active adaptation by households and that climate change will exacerbate the equity– firms, a country’s national, provincial, and munic- efficiency tradeoff that nations often confront. If a ipal institutions must be structured and enabled 36 VI B R A N T C I T I E S to provide markets with accurate information on improving their housing situation.41 People living in risks—and to have land, housing, and insurance these settlements manage risk by playing too safe: markets to operate efficiently. Market-conscious they invest less than optimally in inputs that would adaptation policies require accessible, up-to- enhance their living standards. date information. Providing such information is increasingly feasible as satellites and big data Without perceived tenure security, households improve climate science and expand the scope will not invest in self-protection—or in any for its use in private and public decisionmaking. improvements to their neighborhoods or places of Data on hazard probabilities and vulnerability of residence—beyond the absolute minimum repairs structures and people feed into comprehensive needed. Nearly 1 billion people globally live in risk assessments. These should be considered a slums and informal settlements, where they lack public good, accessible to all. Such information security of tenure and live in substandard housing then allows residents to make informed location with poor or missing infrastructure, especially choices, enables markets to price hazard risk for water, sanitation, and stormwater drainage.42 appropriately, provides the basis for the emer- Unsurprisingly, these are the places that correlate gence of private insurance markets, and serves as most with the Covid-19 hotspots seen earlier. The a sound basis for transparent zoning decisions and same places are often most exposed to climate- other land use restrictions. related disasters. Government will also typically not extend infrastructure and services to infor- To estimate adaptation costs and benefits, firms, mally or illegally developed settlements. Without households, and governments need not only the either private or public investment in shelter and capacity and resources, but also the economic infrastructure, slums and informal settlements will incentives. These incentives can be fully apparent remain overcrowded and unlivable places, at high only when markets provide credible information risk of disease contagion and climate vulnerability. to all parties on an investment’s future value. While market actors require information from scientific Support firms and households with institutions about climate conditions and risks, investments in broad-based infrastructure markets themselves provide a lot of information to and service provision coordinate transactions and influence investment decisions, when they are enabled to function well. Maintenance of public services. Natural disasters And if land and other factor markets work well, are the man-made consequences of geo-physical information on the location of natural hazards is hazard events, whether large or small. But smaller priced into rent or home prices. disasters can be more easily avoided. Good rou- tine urban management already reduces hazard Land is a case in point. In most parts of the world risk considerably. By mainstreaming hazard risk today, land markets are far from adequately infor- reduction in everyday urban planning and manage- mative to motivate private adaptation efforts. To ment, damage can be avoided early. For instance, align incentives and encourage self-protection by floods in developing country cities are often households and firms, a city or country must allow the consequence of insufficient maintenance of investors to access land with clear property rights. drainage systems. In South Asia, drainage ditches Another prerequisite is a well-functioning land are often used as garbage dumps, because regular market. Globally, however, some 70 percent of all refuse collection is insufficient, so drains lose their land lacks secure tenure. This informality of land capacity to transport monsoon runoff from settle- tenure leaves households at risk of eviction and ments. For example, Mumbai spends about 1 billion unable to benefit from their property, whether by rupees ($25 million) a year preparing for monsoon selling it, using it as collateral to access a loan, or rains. Yet, the 2005 monsoon caused 300 deaths. Chapter 1  Making cities resilient 37 Unchecked urban development that leaves too be ex-post. The ex-ante social protection programs little porous green space further increases runoff provide vulnerable populations with the means to and flood risk. take informed action to self-protect themselves in anticipation of climate shocks. For instance, one Smart land use management. Land use planning such social protection program—Mexico’s indexed is a core task of city government that shapes haz- disaster fund (Fonden)—accelerates economic ard risk. For risk reduction, the main objective is recovery after a disaster considerably particularly to prevent development of hazard-prone land. In in less resilient municipalities.43 Similarly an antic- fast-growing cities, land for new development is ipatory cash transfer provided to households fore- scarce. Poor people often cannot afford transport cast to experience extreme floods in Bangladesh charges and need to locate close to city centers helps households maintain food consumption, to have access to labor markets. They end up on earning potentials and overall well-being.44 the least desirable land, such as flood plains or steep and unstable slopes. To reduce settlement WHAT DECISIONMAKERS NEED of these areas, cities need to combine regulation FIRST: ECONOMIC GROWTH AND A and incentives. Zoning enforcement must attempt MARKET-CONSCIOUS POLICY LENS to prevent settlement of the riskiest areas, but it is not easy because informal settlements can spring Information and markets for adaptive action, up overnight and once established are difficult to investments for sustainable connection, and place- relocate. To absorb a growing population while based public interventions all depend, in different excluding risk-prone areas, cities need to ensure a ways, on economic growth. When people, house- supply of suitable land for new development. Since holds, and firms can invest more to reduce risk these areas will be farther from economic opportu- exposure through private decisions and through nities, land development must be accompanied by group insurance, they will do so. And when govern- affordable transport services. ments are better funded by tax revenues—whether national, provincial, or municipal— they will have Protect poor and vulnerable people more and better options for building resilience through infrastructure investment. National Protect poor and vulnerable people against pan- governments with deeper pockets may also have demics, severe weather events, and other emergen- a higher capacity for an effective, centralized cies with transfers and safety nets that are timely, response to disasters. targeted, and temporary. When market incentives for basic services or insurance are inadequate, Finally, governments and other urban develop- cities can provide emergency transfers and safety ment actors need to seek the synergies among nets, especially for women, school-age children, self-protection and market insurance, and public and informal workers. These transfers need not investments in infrastructure. 38 VI B R A N T C I T I E S ANNEX: DOES FLOOD PROTECTION INFRASTRUCTURE INVESTMENT ENHANCE RESILIENCE? AN ECONOMETRIC ANALYSIS TABLE A1.1 DAMS AND RESILIENCE Floodcjmy CHANGE IN Populationcj ln(Night Lights)cjmy 1 2 3 4 5 Damscj 0.003** –0.015** 0.095** (0.001) (0.007) (0.037) Riskycj –0.002 (0.002) Riskycj × Damscj –0.006*** (0.002) Floodcjmy –0.042*** (0.006) Extreme Raincjmy –0.071*** (0.005) Floodcjmy × Damscj 0.017** (0.008) Extreme Raincjmy × Damscj –0.008 (0.006) Fixed effects City ✔ ✔ Country ✔ ✔ ✔ Month × Year ✔ Country × Month × Year ✔ ✔ Number of observations 292,459 3,820 3,820 277,179 277,179 Adjusted R2 0.107 0.123 0.132 0.919 0.919 Note: Clustered robust standard errors in parentheses. *** p < 0.01; ** p < 0.05; * p < 0.1. The dependent variable in column 1 is Floodcjmy, a dummy indicating whether city e in country j was hit by a flood in month m of year y.  The regression includes log values of GDP per capita, builtup area per km2, and population pertaining to the year 2015 as controls. The dependent variable in columns 2 and 3, change in population, refers to the population growth between 2000 and 2015 in city c in country j. Riskycj, is a continuous variable that measures the number of extreme precipitation events in city c in country j between 2000 and 2015, where extreme precipitation is a dummy indicating whether the precipitation in the month m and year y in city c in country j fell in the 90th percentile of the distribution of rainfall in the said city. The distribution was created using 1958–2015 precipitation data. Both regressions include log values of GDP per captta, builtup area per km2, and population pertaining to the year 2015 as controls. The dependent vartable in columns 5 and 6, ln(Night Lights)cjmy, is the natural log of mean light intensity in city c in country j in month m of year y. Both regressions include the controls Stormcjmy and Landslidecjmy, dummies indicating whether city c in country j was hit by a storm or landslide, respectively, in month m of year y. Two-month lead and lags for all three disaster types have also been included as controls. Observations include all city-country-month-year observations which had a nonzero value of night lights. Each observation in the two regressions was weighted by the mean of the cloud free coverage for the city-country-month-year observation. Standard errors are clustered at the city level. Column 1 presents the correlation between cities with dams and flood events. Columns 2 and 3 report population growth regressions;  the difference in column 3 is that the variable for dams is interacted with a proxy for vulnerability of the city (Riskycj). The coefficient on dams in column 2 tells us about population growth in cities with dams, while the coefficient on the interaction term in column 3 indicates whether vulnerable cities with dams have a different population growth rate from low-risk cities with dams. Columns 4 and 5 focus on resilience, in exploring the extent to which dams mitigate the impacts of floods and extreme rain. The coefficient on the interaction term signifies the magnitude by which the effect of the floods on economy activity is mitigated by the presence of a dam within a 100-kilometer radius. The table also shows the impact of extreme rain. Source: Gandhi et al. 2022. Chapter 1  Making cities resilient 39 TABLE A1.2 RECOVERY DYNAMICS FOR FLOODS AND EXTREME RAIN DEPENDENT VARIABLE: ln(Night Lights)cjmy ALL HIGH INCOME LOW INCOME Panel A: Recovery dynamics for floods Floodcj{m}y –0.033*** –0.016** –0.079*** (0.009) (0.007) (0.021) Floodcj{m–1}y –0.017 –0.002 –0.049* (0.011) (0.009) (0.025) Floodcj{m–2}y 0.010 0.007 0.016 (0.011) (0.009) (0.028) Floodcj{m–3}y 0.020 0.013* 0.029 (0.013) (0.008) (0.030) Panel B: Extreme precipitation Extreme Raincj{m}y –0.049*** –0.047*** –0.052*** (0.011) (0.009) (0.018) Extreme Raincj{m–1}y –0.022*** –0.014** –0.034** (0.007) (0.006) (0.014) Extreme Raincj{m–2}y 0.003 0.001 0.005 (0.008) (0.007) (0.015) Extreme Raincj{m–3}y 0.011 0.015** 0.007 (0.009) (0.006) (0.017) Fixed effects City ✔ ✔ ✔ Country × Month × Year ✔ ✔ ✔ Number of observations 606,353 311,287 295,066 Adjusted R2 0.953 0.938 0.932 Note: Two-way clustered robust standard errors in parentheses. *** p < 0.01; ** p < 0.05; * p < 0.1. The dependent variable in all regressions, ln(Night Lights)cjmy, is the natural log of mean light intensity in city c in country j in month m  of year y. Floodcj{m–t}y (Extreme Raincj{m–t}y ) indicates whether city c in country j was hit by a flood (extreme precipitation) event i months prior. Extreme Raincjmy is a dummy indicating whether the precipitation in the month m and year y in city c in country j was greater than the 95th percentile of the city-specific distribution of precipitation, which was created using 1958–2018 data. Observations include all city-country-month-year observations which had a non-zero value of night lights. All regressions in Panel A include the controls Stormcjmy and Landslidecjmy, dummies indicating whether city c in country j was hit by a storm or landslide, respectively, in month m of year y. The controls in Panel B include 7 lags and 3 leads for each of the three disaster types. Controls in Panel B include 7 lags and 3 leads for extreme precipitation events. Each observation was weighted by the mean of the cloud-free coverage for the city-country-month-year observation. Standard errors are clustered at the city and month-year level. Source: Gandhi et al. 2022. 40 VI B R A N T C I T I E S NOTES Day-Night Band on the Suomi satellite. The DMSP night lights data has been shown to have some flaws. 1. World Bank 2021. The values are top coded, leading to saturation in 2. Hallegatte, Vogt-Schilb, and Rozenberg, 2017. core cities (Hsu et al. 2015), and the data do not cor- 3. Elliott et al. 2015; Hsiang and Jina 2014; Kocornik- relate well with output in less dense areas (Chen and Mina et al. 2020. Nordhaus 2011). While some of the issues were rec- 4. Gandhi et al. 2022. tified in an updated Radiance Calibrated Nighttime 5. World Bank 2022a. Light dataset, the data are only available annually 6. World Bank 2022a. and only for seven years through 2010. 7. World Bank 2022a. 22. The econometric specification is described in 8. Jordan CCDR background paper. Gandhi et al. (2022). 9. Kaw et al. 2022. 23. The city-level aggregate measures used here should 10. This result is consistent with the results of a solid account for cases in which economic activity does empirical literature showing that the impacts of not cease but simply moves around the urban area, disasters tend to be much larger for the poorer seg- if people remain within the geographic boundary. ments of the population in developing countries. See, (If one geographic neighborhood is evacuated after for instance, Hallegatte, Vogt-Schilb, and Rozenberg a flood, this displacement effect may increase eco- (2017). nomic activity in that part of the city to which dis- 11. World Bank 2022b. placed residents move.) 12. Patnakar and Patwardhan 2016. 24. Boustan et al. 2012; Kocornik-Mina et al. 2020. 13. Rabbani et al. 2013. 25. Deryugina et al. 2018; Nakamura et al. 2016. 14. UN-Habitat n.d. 26. Gandhi et al. 2022. Extreme events are defined 15. Rigaud et al. 2021. as rainfall in the 90th or 95th percentile of a city’s 16. Sterzel et al. 2020. monthly rainfall distribution between 1960 and 2015. 17. Ehrlich and Becker 1972. 27. Grover, Lall, and Maloney 2022. 18. The research for this report assumes that the qual- 28. Glaeser et al. 2002. ity of government is directly related to the nation’s 29. Gandhi et al. 2022. degree of economic development. 30. The higher risk of spread of contagious diseases in 19. The database identifies urban centers and their more densely populated cities has been cited by boundaries, defining urban areas as “contiguous grid Glaeser as one of the “demons of density”—to quote, cells with a density of at least 1,500 inhabitants per km2 “If two people are close enough to exchange an idea of permanent land or with a built-up surface share on face to face they are also close enough to exchange a permanent land greater than 0.5, and has at least 50,000 contagious disease…” (Glaeser 2016). inhabitants in the cluster with smoothed boundaries” 31. Susceptible, infected, recovered, or removed. (Florczyk et al. 2019, p. 3). Further details can be found 32. Bhardwaj et al. 2020; Jedwab et al. 2021. in the background paper—Gandhi et al. (2022), Adapt- 33. Davis and Weinstein 2002; Bleakley and Lin 2012; ing to Flood Risk: Evidence from a Panel of Global Cities. Henderson et al. 2018; Dell and Olken 2019. 20. Lall et al. 2021. 34. Lall and Wahba 2021. 21. For example, Henderson et al. 2011; Donaldson and 35. United Nations 2015. Storeygard 2016; Henderson et al. 2018. This report 36. World Bank 2015. draws on VIIRS night lights data (Elvidge et al. 2017), 37. World Bank 2016. which has been shown to be accurate and reliable 38. Lall, Henderson, and Venables 2017. (Gibson et al. 2021). Research studies make use of two 39. Kousky et al. 2006. major sources of night lights data: the U.S. Defense 40. Kahn et al. 2017. Meteorological Satellite Program (DMSP) and 41. Global Land Tool Network: https://gltn.net/access-to​ Visible Infrared Imaging Radiometer Suite (VIIRS) -land-and-tenure-security/ Chapter 1  Making cities resilient 41 42. https://unstats.un.org/sdgs/report/2019/goal-11/ (last Ehrlich, I., and G. S. Becker. 1972. “Market Insurance, accessed February 20, 2021). Self-Insurance, and Self-Protection.” Journal of Political 43. de Valle, de Janvry, and Sadoulet 2020. Economy 80 (4): 623–648. 44. Pople et al. 2021. Elliot, R., E. Strobl, and P. Sun. 2015. “The Local Impact of Typhoons on Economic Activity in China: A View REFERENCES from Outer Space.” Journal of Urban Economics 88: 50–66. Bhardwaj, G., T. Esch, S. Lall, M. Marconcini, M. Sop- Elvidge, C., K. Baugh, M. Zhizhin, F. Hsu, and T. Ghosh. pelsa, and S. Wahba. 2020. Cities, Crowding, and the 2017. “VIIRS Night-time Lights.” International Journal of Coronavirus: Predicting Contagion Risk Hotspots. Wash- Remote Sensing 38: 5860–5879. ington, DC: World Bank. Ennaji, M. 2018. “Morocco’s Experience with Gender Gap Bisin, A., and A. Moro. 2022. “Spatial-SIR with Network Reduction in Education.” Structure and Behavior: Lockdown Rules and the Emran, M., W. Greene, and F. Shilpi. 2018. “When Mea- Lucas Critique.” Working Paper 28932, National sure Matters: Coresidency, Truncation Bias, and Bureau of Economic Research, Cambridge, MA. Intergenerational Mobility in Developing Countries.” Bleakley, H., and J. Lin. 2012. “Portage and Path Depen- Journal of Human Resources 53 (3): 589–607. dence.” The Quarterly Journal of Economics 127 (2): Florczyk, A., M. Melchiorri, C. Corbane, M. Schiavina, 587–644. M. Maffeni, M. Pesaresi, P. Politis, S. Sabo, S. Freire, Bordat, S., and S. Kouzzi. 2018. “USAID/Morocco Gender D. Ehrlich, T. Kemper, P. Tommasi, D. Airaghi, and Analysis.” Prepared by Banyan Global. L. Zanchetta. 2019. “Description of the GHS Urban Boustan, L., M. Kahn, and P. Rhode. 2012. “Moving to Centre Database, 2015.” Public Release 2019, Version Higher Ground: Migration Response to Natural 1.0. Luxembourg: Publications Office of the European Disasters in the Early Twentieth Century.” American Union. Economic Review 102: 238–244. Gandhi, S., M. Kahn, R. Kochhar, S. Lall, and V. Tandel. Chen, X., and W. Nordhaus. 2011. “Using Luminosity Data 2022. “Adapting to Flood Risk: Evidence from a Panel as a Proxy for Economic Statistics.” PNAS 108 (21): of Global Cities.” Working Paper 30137, National 8589–8594. Bureau of Economic Research, Cambridge, MA. Davis, D., and D. Weinstein. 2002. “Bones, Bombs, and Gibson, J., S. Olivia, G. Boe-Gibson, and C. Li. 2021. Break Points: The Geography of Economic Activity.” “Which Night Lights Data Should We Use in Econom- American Economic Review 92 (5): 1269–1289. ics, and Where?” Journal of Development Economics 149: del Valle, A., A. de Janvry, and E. Sadoulet. 2020. “Rules 102602. for Recovery: Impact of Indexed Disaster Funds on Glaeser, E. 2016. “Poor World Cities: A Conversation with Shock Coping in Mexico.” American Economic Journal: Edward Glaeser.” Council on Foreign Relations, Octo- Applied Economics 12 (4): 164–95. ber 26. https://www.cfr.org/event/poor-world-cities​ Dell, M., and B. Olken. 2019. “The Development Effects -conversation-edward-glaeser. of the Extractive Colonial Economy: The Dutch Cul- Glaeser, E., D. Laibson, and B. Sacerdote. 2002. “An tivation System in Java.” The Review of Economic Studies Economic Approach to Social Capital.” The Economic 87 (1): 164–203. Journal 112: F437–F458. Deryugina, T., L. Kawano, and S. Levitt. 2018. “The Eco- Grover, A., S. Lall, and W. F. Maloney. 2022. Place, Pro- nomic Impact of Hurricane Katrina on its Victims: ductivity, and Prosperity: Revisiting Spatially-Targeted Evidence from Individual Tax Returns.” American Eco- Policies for Regional Development. Washington, DC: nomic Journal: Applied Economics 10: 202–233. World Bank. Donaldson, D., and A. Storeygard. 2016. “The View Hallegatte, S., A. Vogt-Schilb, M. Bangalore, and J. Rozen- from Above: Applications of Satellite Data in berg, 2017. Unbreakable: Building the Resilience of the Poor Economics.” Journal of Economic Perspectives 30 (4): in the Face of Natural Disasters. Climate Change and 171–198. Development;. Washington, DC: World Bank. 42 VI B R A N T C I T I E S Henderson, J., T. Squires, A. Storeygard, and D. Weil. /210662​/gender-wage-gap-moroccan-women-make​ 2018. “The Global Distribution of Economic Activity: -17-less-men/. Nature, History, and the Role of Trade.” Quarterly Jour- Lall, S., J. V. Henderson, and A. J. Venables. 2017. Africa’s nal of Economics 133: 357–406. Cities: Opening Doors to the World. Washington, DC: Henderson, J., A. Storeygard, and D. Weil. 2011. “A Bright World Bank. Idea for Measuring Economic Growth.” American Eco- Lall, S., and S. Wahba. 2021. “Crowded Cities: New Meth- nomic Review 101: 194–199. odology in Covid-19 Risk Assessment.” Sustainability Hsiang, S., and A. Jina. 2014. “The Causal Effect of Envi- 2021 (13): 7167. ronmental Catastrophe on Long-Run Economic Lall, S. V., M. S. M. Lebrand, H. Park, D. M. Sturm, and Growth: Evidence from 6,700 Cyclones.” Working A. J. Venables. 2021. Pancakes to Pyramids: City Form Paper 20352, National Bureau of Economic Research, to Promote Sustainable Growth. Washington, DC: World Cambridge, MA. Bank. Hsu, F.-C., K. Baugh, T. Ghosh, M. Zhizhin, and C. Elvidge. Nakamura, E., J. Sigurdsson, and J. Steinsson. 2016. “The 2015. “DMSP-OLS Radiance Calibrated Nighttime Gift of Moving: Intergenerational Consequences of a Lights Time Series with Intercalibration.” Remote Mobility Shock.” Technical Report, National Bureau of Sensing 7 (2): 1855–1876. Economic Research, Cambridge, MA. Jedwab, R., A. Khan, J. Russ, and E. Zaveri. 2021. “Epi- Patnakar, A., and A. Patwardhan. 2016. “Estimating the demics, Pandemics, and Social Conflict: Lessons from Uninsured Losses Due to Extreme Weather Events the Past and Possible Scenarios for COVID-19.” World and Implications for Informal Sector Vulnerability: A Development 147: 105629. Case Study of Mumbai, India.” Natural Hazards: Journal Kahn, M., and S. Lall. 2021. “Enhancing Climate Change of the International Society for the Prevention and Mitiga- Resilience through Self-Protection, Public Invest- tion of Natural Hazards 80 (1): 285–310. ment, and Market Insurance.” Brookings, October 15. Pople, A., R. Hill, S. Dercon, and B. Brunckhorst. 2021. Khan, A., H. Park, M. Roberts, and P. Wibisana. 2022. “Anticipatory Cash Transfers in Climate Disaster “When the Lights Go Out: New Empirical Evidence Response.” CSAE Working Paper Series 2021-07, Cen- of the Economic Impacts of the COVID-19 Crisis for tre for the Study of African Economies, University of a Global Sample of Cities.” Background paper for the Oxford. Vibrant Cities report, World Bank, Washington, DC. Rabbani, G., Rahman, S. H., and Faulkner, L. 2013. Kahn, M. E., B. Casey, and N. Jones. 2017. “How the “Impacts of Climatic Hazards on the Small Wetland Insurance Industry Can Push Us to Prepare for Cli- Ecosystems (ponds): Evidence from Some Selected mate Change.” Harvard Business Review. https://hbr. Areas of Coastal Bangladesh.” Sustainability 5: org/2017/08/how-the-insurance-industry-can-push​ 1510–1521. -us-to-prepare-for-climate-change. Rigaud, K., A. de Sherbinin, B. Jones, J. Bergmann, V. Kaw, J., M. Morad, Y. Li., W. Tammaa, M. Alam, and N. Clement, K. Ober, J. Schewe, S. Adamo, B. McCusker, Daito. 2022. “Pathways for Decarbonizing Growing S. Heuser, and A. Midgley. 2021. Groundswell: Preparing Cities across the Urban-Transport-Energy Nexus in for Internal Climate Migration. Washington, DC: World Jordan.” Bank. Kocornik-Mina, A., T. McDermott, G. Michaels, and F. Sterzel, T., B. M. K. B. Lüdeke, C. Walther, M. T. Kok, D. Rauch. 2020. “Flooded Cities.” American Economic Sietz, and P. L. Lucas. 2020. Typology of coastal urban Journal: Applied Economics 12 (2): 35–66. vulnerability under rapid urbanization. PLoS ONE Kousky, C., E. Luttmer, and R. Zeckhauser. 2006. “Private 15 (1). https://doi.org/10.1371/journal.pone.0220936. Investment and Government Protection.” Journal of United Nations. 2015. “Millenium Development Goals Risk and Uncertainty 33: 73–100. Indicators. Indicator 7.10 Proportion of Urban Pop- Lahsini, C. 2017. “Gender Wage Gap: Moroccan Women ulation Living in Slums.” New York: United Nations. Make 17% Less Than Men.” Morocco World News, March http://mdgs.un.org/unsd/mdg/seriesdetail.aspx?srid​ 11. https://www.moroccoworldnews.com/2017/03​​ =710. Chapter 1  Making cities resilient 43 UN-Habitat (United Nations Human Settlements World Bank. 2019. “Country Partnership Framework for Programme). n.d. “Population living in slums (% the Kingdom of Morocco for the Period FY19–FY24.” of urban population) - Morocco.” Nairobi, Kenya: Washington, DC: World Bank. UN-Habitat. World Bank. 2021. Green, Resilient, and Inclusive Develop- World Bank. 2015. Morocco—Mind the Gap: Empowering ment (GRID). Washington DC: The World Bank. Women for a More Open, Inclusive and Prosperous Society. World Bank. 2022a. “Egypt Country Climate Develop- Washington, DC: World Bank. ment Report.” Washington, DC: World Bank. World Bank. 2016. Côte d’Ivoire Urbanization Review: Diver- World Bank. 2022b. “Morocco Country Climate Develop- sified Urbanization. Washington, DC: World Bank. ment Report.” Washington, DC: World Bank. MAKING CITIES SUSTAINABLE mass transit investments. lower carbon emissions. reduced GHG footprint. energy management. 2 Decarbonizing cities— data, density, and delivery Low-carbon urban development will be key to that are harmful to green urban growth is provided in addressing climate change. Cities account for more a separate World Bank report, Thriving: Making Cities than 70 percent of global CO2 emissions, most from Green, Resilient, and Inclusive in a Changing Climate. industrial and motorized transport systems that use huge quantities of fossil fuels and rely on far-flung The Glasgow Pact, a result of the 26th UN Cli- infrastructure constructed with carbon-intensive mate Change Conference of the Parties (COP26) materials. To that end, the World Bank’s Climate in Glasgow, held 31 October–12 November 2021, Change Action Plan (2021–2025) recognizes city emphasizes the need for much deeper emission systems as one of five key systems that generate cuts, noting that without major decarbonization the most greenhouse gas (GHG) emissions.1 Staying efforts, the world risks catastrophic impacts on below the 1.5°C threshold means massive decarbon- people and ecosystems. The COP26 conclusions ization of cities, which will require carbon pricing, present an opportunity to reflect on cities’ contri- reforms to product and factor markets, investments butions to GHGs. Identifying the sources of these in low-carbon energy and transport systems, and emissions, along with ways to accurately measure programs to reduce urban sprawl. and monitor them, is a first step toward meaning- fully reducing them. This chapter focuses on CO2 emission because of its overwhelming importance in the global effort to Assessing decarbonization progress requires new mitigate climate change. In addition to CO2’s role in metrics, because not all cities are the same. Rapid global warming, air pollution is inimical to health and decarbonization will need to be promoted or productivity in cities. Cutting CO2 emissions in cities rewarded by emission-based performance ratings will fortunately reduce PM2.5, a harmful air pollutant, that make urban governments and firms more which is implicated in a host of respiratory diseases, accountable to their stakeholders. But effective reduced labor productivity, impaired cognitive func- performance measures cannot ignore the effects of tion, and stunting.2 More comprehensive treatment differing demographic, economic, and geographic of factors (air pollution, availability of green space) conditions on actual CO2 levels in cities. 45 46 VI B R A N T C I T I E S New metrics will require more accurate and Report, Thriving: Making Cities Green, Resilient, and objective CO2 measures. Assessing city progress Inclusive in a Changing Climate, covers these issues. on emissions reductions has been hindered by the near-total absence of directly measured CO2 RECENT EVIDENCE ON THE emissions data. Consistently measured emissions MAGNITUDE AND SOURCES OF CO2 estimates are only available for about 80 cities, only EMISSIONS FOR CITIES WORLDWIDE half in developing countries. What’s more, hardly any estimates are based on actual emissions. Most An extensive empirical literature has explored estimates rely on parameters from engineering the determinants of CO 2 emission growth. It studies applied to survey-based activity measures largely relies on estimates from inventories of CO2 for transport, energy production, and manufac- emissions that apply engineering parameters to turing. They are particularly suspect for develop- measures of activity in industry, transport, land ing countries, because many of the measures are clearing, residential heating, and other sectors. calibrated using databases and models from high- Most attention has focused on drivers of CO 2 income economies. emissions from fossil-fuel combustion and cement production.5 Fires associated with agriculture Coordinating and harmonizing pricing instru- and land-use change also produces heavy CO 2 ments, urban planning, and public infrastructure emissions, estimated more precisely by recent investment will support low-carbon development. research.6 However, emission drivers in this sector Economists generally support emissions reduction have received less attention than work on indus- via Pigouvian policies such as emission taxation or trial determinants.7 Worldwide, consistently mea- permit trading.3 Many policy analysts who support sured GHG estimates are available for only about such pricing also argue for a non-Pigouvian supple- 80 urban areas, with more than half in developed ment: coordinated public investment in low-car- countries.8 bon land development, energy, and transport that will accelerate the transition to low-carbon econ- A new study done for this report takes a com- omies.4 And reduced motor vehicle emissions and pletely different approach, employing direct energy efficiencies associated with higher-density CO2 observations from satellites. The advent of development are frequently cited as carbon-saving satellite-based CO 2 measurement has greatly advantages of mass transit. expanded the potential for empirical assessment. High-resolution observations of atmospheric GHG This chapter provides recent evidence on the concentrations are now available from several magnitude and sources of CO2 emissions for cit- platforms, including NASA’s OCO-2 and OCO-3 ies worldwide. It combines econometric analysis instruments, 9 the European Space Agency’s and urban modeling; a recent survey of empirical METOP-A and ­ TROPOMI (Sentinel-5P) platforms, work highlighting the benefits of urban popula- TANSAT, and the Japan Space Exploration China’s ­ tion density; a valuation of benefits from public Agency’s G ­ OSAT and GOSAT‑2. Detailed techni- investments in subway systems mitigating GHG cal assessments of measures from these platforms emissions; and recommendations for supporting verify that they provide useful and comprehensive low-carbon urban development through infor- information for global carbon emissions analysis.10 mation, urban planning, and coordinated public Dasgupta, Lall, and Wheeler (2021) use NASA’s investment. It does not, however, attempt to pro- OCO-2 because it offers open access (JPL/NASA vide a comprehensive treatment of broader issues 2021); a long panel of consistently measured daily of sustainable urban development, including local observations (September 6, 2014–present); and amenities and congestion, or air and water pol- the highest spatial resolution among the sources lution. A new Sustainable Development Flagship (1.29 × 2.25 km). Chapter 2  Decarbonizing cities 47 FIGURE 2.1 THE 100 TOP CO2 EMITTERS Top 100 cities, by total CO2 emissions (million tons) Less than 30 30–34 35–49 50–64 65 or more Source: Dasgupta, Lall, and Wheeler 2022. Using this satellite data, analysts can rank cities every location around the world using satellite data globally on their CO2 emissions. The top 100 urban highlights the roles of cities in the United States, areas for CO2 emissions emit 5.1 gigatonnes (Gt) Europe, and China, along with the forested regions of carbon annually, accounting for 14 percent of of South America, Central Africa and Southeast global emissions and 23 percent of urban emissions Asia (figure 2.2). (figure 2.1). The top emitters are Tokyo, New York, Chicago, Guangzhou, Dortmund, and Beijing­ —­ Thanks to a new econometric model, estimates are each emitting more than 100 million tonnes a year. now available for key components of observation- A global snapshot of annual observed emissions for based CO2 emissions in any urban area around FIGURE 2.2 AVERAGE ANNUAL CO2 EMISSIONS, 2014–2021 Thousand tonnes per year 0–143 144–473 474–1,056 1,057–1,999 2,000–3,350 3,351–5,269 5,270–8,160 8,161–13,761 13,762–27,569 27,570–53,728 Source: Dasgupta, Lall, and Wheeler 2022. 48 VI B R A N T C I T I E S the world. Policymakers can use the model to city’s mean annual observed emissions with the gauge a city’s deviation from the predicted annual emissions it predicted for that city (figure 2.4). emissions for a place with its characteristics. City performance tends to exceed expectations The model estimates the contributions of vari- (emissions below predictions) in India, Western ous sources to CO2 emissions in each city, based Europe, and the former Comecon countries (the on its demographic, economic, and geographic Soviet Union and Eastern Europe). In contrast, characteristics (figure 2.3). It also compares each cities tend to underperform (emissions greater FIGURE 2.3 ECONOMETRICALLY ESTIMATED CO2 EMISSIONS SOURCES FOR SAMPLE CITIES Industry Fires Income Population Climate Source: Dasgupta, Lall, and Wheeler 2021. FIGURE 2.4 CITY EMISSIONS PERFORMANCE BY REGION City residual (observed – predicted) CO2 concentration (ppm) 3 2 1 0 –1 –2 –3 China Sub- East Asia Middle Latin South North India Western Former Saharan & Pacific East & America & Asia America Europe Soviet Union Africa (excluding North Caribbean (excluding & Eastern China) Africa India) Europe Source: Dasgupta, Lall, and Wheeler 2022. Chapter 2  Decarbonizing cities 49 than predicted) in China, the rest of East Asia and individual cars, decarbonizing may weigh on their Pacific, the Middle East and North Africa, and economic attractiveness. Sub-­Saharan Africa. SUSTAINABLE POPULATION DENSITY: Closer analysis in Egypt and Jordan provide A DEBATE COMPLICATED BY DISPARATE insights on how some cities are managing carbon PATTERNS AT DIFFERENT INCOMES emissions. In Egypt, the concentration of CO2 emissions varies from that predicted, indicating A recent survey of empirical work, largely focusing differential performance of cities. While the accu- on developed countries, highlights the benefits mulated CO2 concentration observed in a few cit- of urban density. A doubling of urban density is ies is in the range of 402.39–406.75 parts per million associated with premiums in wages (4 percent), (ppm) due to their socioeconomic and geographic patent activity (19 percent), consumption variety characteristics, cities such as Port Said and Giza value (12 percent), the preservation of green spaces fall in the 90th percentile of cities in the MENA (23 percent), and the use of non-car transport modes region. The cities on the right of figure 2.5 emit (7 percent).11 It is also associated with reductions more CO2 than expected for their characteristics. in average vehicle mileage (–8.5 percent), energy For comparison, human activity generates about consumption (–7 percent), pollution concentration 36 Gt of CO2 emissions each year, increasing the (–8 percent), crime (–8.5 percent), and unit costs atmospheric CO2 concentration by about 2 ppm. of providing local public services (–14.4 percent). In Jordan, Amman and Irbid­ —­despite having con- Although the cost of living in a city increases as its centrations of CO2 emissions on par with other population grows­ —­reflecting demand for urban cities in MENA­ exhibit higher than expected CO2 —­ housing, transport, and services­—­developed-country emissions, and an emissions profile driven mainly cities more than compensate for this pattern through by industrial activity and income effects. The the rising productivity and incomes made possible ability to decarbonize will be one factor in cities’ by economically efficient agglomerations. competition to attract people and investment. For cities that have developed in carbon-intensive ways Developing countries are not, however, realizing with inefficient buildings and high dependence on these theoretical benefits from urban population FIGURE 2.5 EMISSION PERFORMANCE FOR SELECTED CITIES IN EGYPT Residual (observed – predicted) CO2 concentration (ppm) Observed CO2 concentration (ppm) 1.00 407 Higher than predicted Lower than predicted 0.75 406 0.50 405 0.25 404 0.00 403 –0.25 402 –0.50 401 6th of October Cairo Giza Shubra Alexandria Halwan Al Mahallah Port Said city al Khaymah al Kubra Source: World Bank 2022. 50 VI B R A N T C I T I E S density. Especially in South Asia and Sub-­Saharan because small reductions in sparse mechanized Africa, cities are not fully reaping the potential transport do not offset increased CO2 emissions economic productivity gains (see chapter 4). And from factors like more concentrated household urban costs grow even faster in these countries than cooking and heating. And in many low-income cit- in developed countries, largely due to inadequate Saharan Africa, households spend about ies of Sub-­ housing and strained transport and infrastructure 50 percent of their income on food, and among the systems.12 More broadly, little empirical evidence poorest urban households, even 60 percent.14 exists on the links between urban density and urban sustainability as measured by GHG emis- An urban growth model complements the econo- sions. The effects of urban form on sustainability metric analysis a nd compares a city’s future condi- thus remain subject to debate. tions in urban form, transport, and energy sector characteristics, as well as co-benefits for the follow- To address this knowledge gap, the econometric ing policy levers (table 2.1). model can be used to examine the population den- sity–CO2 relationship at different income levels. The urban growth model scenarios are applied to The relationship is strongly positive at US$400 per Amman, Jordan. The country’s economy, popula- capita and mildly so at US$800.13 It then switches tion, and GHG emissions are spatially concentrated sign and becomes progressively more negative in the Amman agglomeration. The municipality for US$1,200, US$2,000, and US$5,000 per capita has committed to reaching near net zero carbon (figure 2.6). emissions by 2050. The model scenarios suggest that Amman can benefit from deeper integration The wide variation in density–CO2 relationships and coordination across sectors, and more efficient at different incomes has various explanations. In use of urban land supported by targeted public- high-income areas that rely heavily on mechanized and private-sector interventions. Four scenarios transport, increased population density may lower for 2050 were modeled: aggregate emissions by reducing travel require- • Business as usual without any climate action ments. But higher population density in low-income policies and urban expansion based on histor- cities may translate into higher CO2 emissions ical patterns. FIGURE 2.6 CO2 CONCENTRATION AND POPULATION DENSITY AT DIFFERENT INCOME LEVELS CO2 concentration (ppm) 406.1 $400 406.0 405.9 $800 405.8 $1,200 405.7 $2,000 405.6 $5,000 405.5 0 10,000 20,000 30,000 40,000 50,000 Population density (people per km2) Source: Dasgupta, Lall, and Wheeler 2021. Chapter 2  Decarbonizing cities 51 TABLE 2.1 URBAN GROWTH MODEL POLICY LEVER PROCESS IN WHICH THE MODEL: A. Urban planning Defines the pattern in which the future population will settle, either in defined expansion areas or unoccupied land within the urban area. B. Renewable energy contribution Analyzes solar photovoltaic, wind, biogas, and hydroelectric generation. C. Demand-side management Analyzes energy measures related to the curtailing of energy demand by numerous strategies, including behavioral changes and financial benefits. D. Green building codes Analyzes the potential contribution of implementing green building measures and guidelines as described in the Jordanian green code. Electricity consumption in E.  Analyzes the required electricity for street lighting and the type of technology public lighting in operation (incandescent, LED, smart street lighting, solar, among others). F. Electrification of transport Analyzes the transformation from gasoline and diesel vehicles to electric technology. G. Mass public transit Analyzes an expansion of the public transport routes and an increase in capacity from buses to bus rapid transit. Source: Kaw et al. 2022. • A plan, taking into account Amman’s climate instance, a combination of urban planning plus action and spatial plans. energy demand-side management and enforce- • Ambitious, based on the plan scenario but scal- ment of building codes could potentially unlock ing up with more aggressive targets and com- almost seven times the GHG abatement potential pact urban growth. of individual policy levers alone. Amman could do • Net-zero, which includes investing heavily in more to reach its net-zero emissions goals by 2050­ renewable energy. —­by controlling urban expansion and investing in green infrastructure for the long run. However, Managing urban expansion and supporting dense this would mean more costly investments upfront urban development is a key policy lever in the in renewable energy penetration and aggressive modeling scenario. It defines where the future energy efficiency measures (such as public infra- population will settle, either in defined expansion structure and green buildings), and fully utilizing areas or unoccupied land within the urban area. infill land and controlling city expansion (table 2.2). The compact growth and densification for the ambitious scenario suggests that the municipality PUBLIC INFRASTRUCTURE FOR URBAN should encourage occupying vacant land within SUSTAINABILITY: SUBWAY SYSTEMS the urban area’s current boundaries, including inte- MITIGATING GHG EMISSIONS grating suburban areas within a buffer zone of 2.5 kilometers around the 2020 city boundaries. The Do mass transit systems help cities reduce CO2 net-zero strategy suggests that the municipality will emissions while integrating labor markets? The limit urban expansion and occupy the vacant land analytical framework outlined in the World in the existing urban footprint. Bank report Pancakes to Pyramids highlighted that improved transport technology enables economic The results suggest that Amman should adopt more concentration in urban cores, supports cities’ eco- aggressive targets and integrated measures. Such nomic and spatial growth, and boosts demand for an approach would generate greater synergies than livable residential floor space.15 The underlying just standalone policy and achieve mitigation and assumption is that cities benefiting from mass tran- adaptation outcomes more cost-effectively. For sit will benefit from agglomeration economies in 52 VI B R A N T C I T I E S TABLE 2.2 URBAN DEVELOPMENT AND GHG EMISSIONS MODELING SCENARIOS FOR AMMAN INSIGHTS FROM INTEGRATED SPATIAL SCENARIOS FOR GREATER AMMAN FOR 2050 Amman can reduce its energy consumption and Energy use by 2050 Greenhouse gas emissions GHG emissions by 27 percent and 41 percent, (megajoules by 2050 (kilograms of respectively, through more aggressive measures per capita per year) CO2 equivalent per year) beyond GAM’s current plans. Reaching net-zero 15,000 3,000 2,826 carbon emissions by 2050 requires additional 2,479 measures. 10,000 2,000 1,452 5,000 1,000 Commuting Water 17 Public lighting Buildings 0 0 BAU Plan Ambitious Net zero Wastewater Solid waste (do nothing) (current) (scale up) (do more) This scenario would entail a more spatially Urban land consumption and natural land loss by 2050 (km2) efficient urban form, through better planning, 1,000 910 controlled urban expansion, and protection of 784 natural and agricultural areas. The ambitious 615 570 scenario anticipates a 22 percent reduction in urban 500 expansion from current plans. 0 –4 0 –128 –500 –467 Land consumption forecast (2050) Natural land loss –1,000 BAU Plan Ambitious Net zero Urban footprint (2020) Agricultural land loss (do nothing) (current) (scale up) (do more) (continued) tradable manufacturing and services sectors, such their installation (annex table A2.1).17 An examina- that economic fundamentals favor a concentration tion of subways worldwide draws subway data from of economic activity and lead workers to seek jobs two sources.18 The first is a global subway survey by in these dense urban cores. For this chapter, the Turner and Gonzalez-Navarro (2018), encompass- key question is whether mass transit also consti- ing 137 systems installed before 2011, and provid- tutes an investment in urban sustainability, and if ing digital subway maps at five-year intervals from so, in what context and to what extent. 1930 to 2010. The second is the authors’ survey of 55 subway systems installed since 2010.19 Subways An econometric study of underground transit numbered 26 before 1948, 39 in 1948–1980, 72 in systems (subways)16 suggests that the systems have 1981–2010, and 55 in 2011–2021. Figure 2.7 draws on significantly reduced CO2 emissions in the cities these maps to illustrate the historical progression where they operate. The impacts affected by both of subway development in Tokyo, Moscow, and the scale of the systems and the time elapsed since Madrid. The estimated effects of subway systems is Chapter 2  Decarbonizing cities 53 TABLE 2.2 continued URBAN DEVELOPMENT AND GHG EMISSIONS MODELING SCENARIOS FOR AMMAN INSIGHTS FROM INTEGRATED SPATIAL SCENARIOS FOR GREATER AMMAN FOR 2050 Amman can realize cost-efficiencies Municipal and private expenditure by 2050 (US$ billions) and other co-benefits by investing in complementary combinations of policy levers 80 to achieve synergies. A shift toward smarter and complementary policies across the urban, 60 transport, and energy sectors can potentially reduce total infrastructure investment costs by 25 percent by 2050 under the ambitious scenario. 40 Reaching net-zero emissions will require a doubling of capital investment from the ambitious scenario 20 due to costs of renewable energy and related infrastructure. 0 –20 Municipal services New infrastructure and policies BAU Plan Ambitious Net zero Private sector Carbon market (do nothing) (current) (scale up) (do more) Total capital expenditure for new infrastructure and local policies by 2050 (US$ billions) 14 11.94 12 9.68 10 7.48 8 5.59 6 Renewable energy Electrification of public buses Green infrastructure Public lighting 4 Green spaces and parks Enforcement of building codes Water management Demand-side management 2 Waste management Urban planning Controlled parking Grey infrastructure for floods Pedestrianization Grey infrastructure 0 BAU Plan Ambitious Net zero Mass public transit (do nothing) (current) (scale up) (do more) Note: Model based on CAPSUS Suitability and Urban Performance tools. Source: Kaw et al. 2022. negative in terms of CO2 emissions, is highly signifi- The difference between present and counterfactual cant, and is strongly consistent with the hypotheses emissions provides an estimate of the impact of that the impacts of subways accumulate through subway systems to date (table 2.3). As column 2 time, with system scale also playing an important shows, estimated global population-related CO2 role (see annex table A2.1). emissions were 23.3 Gt in 2021. Emissions from “functional urban areas” (FUAs) with subways were The overall effect of subway systems on global 3.0 Gt­—­19.1 percent of emissions from all FUAs CO2 emissions is estimated with a counterfactual and 12.8  percent of global population-related exercise. First, emissions are predicted for the 192 emissions. Column 1 displays the counterfactual urban areas, given the ages and track lengths of results. For the 192 FUAs that currently operate their systems in 2021. Then emissions are predicted subways, estimated emissions without subways are after setting all subway lengths and ages to zero. —­ 6.1 Gt­ 32.7 percent of emissions from all FUAs 54 VI B R A N T C I T I E S FIGURE 2.7 SUBWAY SYSTEM DEVELOPMENT, TOKYO, MOSCOW, AND MADRID, 1935–2010 Source: Dasgupta, Lall, and Wheeler 2022. Chapter 2  Decarbonizing cities 55 TABLE 2.3 CO2 EMISSIONS IMPACT OF OPERATING SUBWAY SYSTEMS ESTIMATED POPULATION- RELATED CO2 EMISSIONS, 2021 (Gt) (1) (2) (3) (3) WITHOUT WITH EMISSION PERCENT GROUP SUBWAYS SUBWAYS CHANGE (Gt) CHANGE Subway functional urban areas 6.1 3.0 –3.1 –51.2 Other functional urban areas 12.6 12.6 Other areas 7.8 7.8 Total 26.4 23.3 –3.1 –11.8 Subway functional urban areas as a share of all functional urban areas (percent) 32.7 19.1 Subway functional urban areas as a share of total (percent) 23.1 12.8 Source: Dasgupta et al. 2021. and 23.1  percent of global population-related commuting times and vehicular pollution, which emissions. Column 3 summarizes the changes have provided the traditional rationale for subway attributable to subways: a reduction of 3.1 Gt in installation. CO2 emissions, which translates to reductions of 51.2 percent for subway FUAs and 11.8 percent for Until recently, empirical studies of transit project all areas. To the extent that subway construction benefits focused on reductions in traffic congestion has been done in conjunction with land use plans and local air pollution. Significant impacts on traf- to keep cities compact pyramids, the expected fic congestion have been found for Beijing,20 Los reductions in CO2 emissions from subways are Angeles,21 and Rotterdam.22 Significant effects on likely to be even larger, making them economically local air pollution have been found in city-specific viable in more cities. studies for Barcelona,23 Taipei,24 and Delhi,25 as well as national studies for Germany26 and China.27 The welfare economics of future subway investments Recent exploratory empirical studies have also con- sidered co-benefits from carbon emission reduc- The installation costs of currently operating sub- tion by mass transit for individual cities: They cover ways are sunk, but the effects of future systems on Delhi,28 Beijing,29 Nanjing,30 and Malmo, Sweden.31 CO2 emissions can also be assessed from a social Additional studies at the national level have been benefit-cost perspective. Carbon benefit values for conducted for China.32 But until now, no systematic future subways depend on their emissions-reduc- attempts have tried to quantify climate co-benefits ing impacts and the social cost of carbon, while in proportion to the financial magnitudes of sub- their costs reflect installation investment and way investments. Our model and results provide an operation and maintenance (O&M) costs. Subways opportunity to perform this analysis. can be assessed for their carbon-reducing benefits alone, or they can be viewed as a potentially im- For the financial component, this report draws portant source of co-benefits along with reduced on a recent study by Levy (2019) that estimates 56 VI B R A N T C I T I E S installation costs per kilometer (ICs) for 207 on investment. In principle, this can vary from a subway projects in 47 countries. From the study large negative number (for cases where annual CO2 sample, we draw ICs for percentiles 33, 50, and reduction benefits are overwhelmed by the initial 67­—­US$140  million, 200 million, and 280 mil- investment and O&M costs) to a large positive lion per kilometer, respectively. Annual subway number (for the converse cases). NPVRs that are O&M costs are estimated at 2 percent of the initial greater than 0 and less than 1 represent cases where investment.33 The social cost of carbon (SCC) for subway installation provides a positive co-benefit an emitted unit of CO2 is the discounted value of along with reductions in commuting time and the present and future damage caused by its pres- vehicular pollution, but cannot be justified on ence in the atmosphere. Damage estimation and climate grounds alone. NPVRs between 1.0 and 1.8 the appropriate discount rate remain objects of indicate positive paybacks but social rates of return controversy, although the consensus estimate of (SRRs) below 2 percent (which is equivalent to an the SCC has clearly been rising. A recent paper NPVR of 1.8 for our 30-year exercise). NPVRs above criticizes the SCC employed by the Obama admin- 1.8 represent cases where subway systems meet the istration in the United States (US$50/ton of CO2) consensus SRR standard and can be justified on and argue that appropriately valuing the welfare of climate grounds alone. future generations may warrant an SCC of US$100/ ton by 2030.34 For this exercise, we use three SCCs: Figure 2.8 provides a robust sense of the overall US$50, 100, and 150 per ton of CO2. implications with three cases that are “optimistic” for subway installation (SCC $150/ton; unit cost Using the projected values for population, income, $140 million/km); “midrange” (SCC $100/ton; unit heating degree days, and cooling degree days, cost $200 million/km; and “pessimistic” (SCC $50/ population-related CO2 emissions are estimated ton; unit cost $280 million/km). The horizontal axis for functional urban areas from 2020 to 2050, with presents NPVRs (in log scaling) from 0.10 to the max- and without subways (see annex table A2.1). The imum observed value, and the vertical axis measures difference in emissions is calculated in each year the number of candidate functional urban areas and is valued at SCCs of US$50, 100, and 150 per (FUAs) that have at least the corresponding NPVR ton of CO2. For standard 16.5 km track lengths, on the horizontal axis. Of the 1,214 candidate FUAs, we calculate subway installation costs at US$140, positive co-benefits (NPVR ≥ 0.1) are obtained for 200, and 280 million per km. We compute annual 294, 465, and 760 FUAs in the pessimistic, midrange, O&M costs at 2 percent of installation costs. To and optimistic cases, respectively. Full payback keep the exercise tractable, we adopt the extremely (NPVR = 1) is achieved by 79, 312, and 434 FUAs in conservative assumption that all O&M costs are the three cases, and an SRR of at least 2 percent is attributable to carbon reduction alone, with no achieved by 15, 228, and 370 FUAs. components for reduced commuting times or vehicular emissions.35 INFORM, SUPPORT, AND PROTECT­ —­BOOSTING LOW-CARBON URBAN Annual net benefits are computed for 2020–2050 DEVELOPMENT THROUGH URBAN from the with/without subway emissions differ- PLANNING AND COORDINATED ence, valued at the SCC, minus the annual O&M PUBLIC INVESTMENT cost. To obtain the present value, we draw on a sur- vey of 200 economists by Drupp et al. (2015) that Inform­—­to set expectations and motivate yields a median value of 2 percent for the social private action rate of discount. Dividing discounted net benefits by installation cost, we obtain a net present value To keep the global temperature below a 1.5°C ratio (NPVR) that can be used to assess returns increase relative to the preindustrial era requires Chapter 2  Decarbonizing cities 57 FIGURE 2.8 QUALIFYING FUNCTIONAL URBAN AREAS AND NET PRESENT VALUE RATIOS: VARYING SOCIAL COSTS OF CARBON AND UNIT COSTS Source: Dasgupta et al. 2021. huge and continued cuts in GHGs. Nations around plans that reorient the spatial organization of the world have pledged steep emission reductions development and the quality of structures. Coordi- as part of their nationally determined contribu- nating urban plans with investments in mass transit tions, but against the background of a near-total infrastructure can set cities on a course for lower absence of directly measured local and regional carbon emissions while also integrating urban GHG data for emissions performance assessment labor markets. and program design. The research for this report and encapsulated in this chapter provides a sum- —­ Protect­ aiding vulnerable groups during mary of globally comparable and publicly accessi- the transition to low-carbon development ble metrics on city-level carbon emissions that can incentivize people and firms as well as national The poor are often at increased risk in the transi- and city governments to make more informed and tion to a low-carbon economy, because they are calibrated approaches to low-carbon development. more engaged in informal, agricultural, and “brown” activities. It is estimated that about 40 percent of Support­—­ private action by developing Egypt’s employment is concentrated in brown sec- credible plans for coordinating action and tors, and 58 per percent in yellow sectors.36 Hence investing in sustainable connections the need to design policies ensuring that poor and vulnerable people benefit from the transition. Marginal policies such as setting the carbon price equal to the social cost of carbon have had limited While not the focus of this report, ongoing analyt- traction globally, have faced political resistance, ical work at the World Bank including the Sustain- and need complementary policies in the presence able Development flagship suggests that ex-ante of strategic complementarities in demand and sup- policy reform analysis, appropriate compensatory ply of renewable energies as well as lock-ins from measures, and targeted investments are three preferences and spatial organization. While carbon essential ingredients for a distribution-­ sensitive pricing can be a sharp instrument in influencing reform package. behavior of individual firms and households, their • Assessing the distributional impact of climate actions actions can be strengthened with credible urban using ex-ante analysis. This identifies negative 58 VI B R A N T C I T I E S impacts of climate action. Identifying the win- policies that make needed changes in carbon ners and losers of different actions is a key prices. For example, in the Russian Federa- part of understanding the political economy of tion, welfare analysis on the removal of energy reform. It also allows win-win policy choices­ subsidies showed that the poorest 40 percent —­climate actions that benefit the poor. One of households could be shielded from welfare such example is the Poverty and Social Impact losses in the short term by redistributing one- Analysis conducted in Mexico highlighting that third of the additional revenues collected from the proposed changes to agricultural subsidies higher utility prices through lump-sum trans- would benefit the poor. Ex-ante analysis of the fers. The experience of ESMAP in removing en- impact of climate actions on poverty and in- ergy subsidies shows the need for compensatory equality is being undertaken in more than 30 measures plus good communication. countries to contribute to the Bank’s Climate • Investing in skills and addressing labor market Change Development Reports. rigidities that limit the ability of poorer workers. • Compensatory measures targeted at poor households Low hydrocarbon growth is expected to shift when climate actions are costly for poor people. Some the composition of jobs and occupations. Poor climate actions such as policies to increase car- households often find it hardest to make these bon prices are progressive but make the poor transitions, because of lower skills and barriers poorer, as prices for energy and energy-inten- to migration or sectoral transitions. Work has sive goods increase. Reducing these impacts and started to identify skill mismatches and labor minimizing political pushback requires compen- market rigidities using standardized approaches satory measures to be implemented alongside and standardized multicountry survey data. Chapter 2  Decarbonizing cities 59 ANNEX: IMPACT OF UNDERGROUND SUBWAY SYSTEMS ON CO2 CONCENTRATIONS: AN ECONOMETRIC ANALYSIS TABLE A2.1 DETERMINANTS OF CO2 CONCENTRATION ANOMALIES Dependent variable: XCO2 Anomaly (parts per billion) NL NL (ROBUST) NL (CLUSTER)a Industry CO2 emissions (thousand tons) 0.277*** 0.277*** 0.277*** (26.19) (18.85) (9.88) Industry CO2 wind-displaced emissions 0.275*** 0.275*** 0.275*** (thousand tons, weighted) (44.59) (31.60) (13.82) Fires CO2 emissions (thousand tons) 0.359*** 0.359* 0.359* (24.79) (2.35) (2.03) Fires CO2 wind-displaced emissions 0.573*** 0.573*** 0.573*** (thousand tons, weighted) (61.85) (8.64) (4.46) Population (thousands) × Heating degree days 5.362*** 5.362*** 5.362*** (69.53) (17.25) (6.18) Population (thousands) × Cooling degree days 0.398*** 0.398*** 0.398 (5.48) (5.29) (1.32) Population (thousands) × Income per capita 13.54*** 13.54*** 13.54*** (US$ thousands) (24.45) (15.94) (3.75) IHSTb (subway scale [track length in km]) + –0.193*** –0.193*** –0.193*** IHSTb (subway age) (–37.48) (–22.29) (–10.51) Constant –193.4*** –193.4*** –193.4*** (–192.01) (–153.66) (–10.32) Observations 1,961,754 1,961,754 1,961,754 Note: t-statistics in parentheses. *** p < 0.001; ** p < 0.01; * p < 0.05. a. GADM (2021) Level 1 administrative divisions (states, provinces). b. IHST is inverse hyperbolic sine transformation. 60 VI B R A N T C I T I E S NOTES 27. Zheng et al. 2017. 28. Doll and Balaban 2013. 1. World Bank 2021. 29. Guo et al. 2021. 2. World Bank 2021. 30. Zheng et al. 2021. 3. Stiglitz and Stern 2021; Jacobs and van der Ploeg 31. Forsberg and Krook-Riekkola 2021. 2019; King et al. 2019; Klenert et al. 2018. 32. Mao et al. 2012; He and Qiu 2016. 4. van der Ploeg and Venables 2020. 33. KPMG International, cited in Kohli, Linn and Zucker 5. Raupach et al. 2007; Jotzo et al. 2012. (2019). 6. Gasser et al. 2020; Winkler et al. 2021. 34. Stern and Stiglitz 2021. 7. Sanchez and Stern 2016. 35. Taking the other approach would require a con- 8. C40 2021. voluted iteration with successive calculations for 9. Orbiting Carbon Observatory. alternative apportionment of O&M costs across 10. Pan et al. 2021; Wu et al. 2020; Labzovskii et al. 2019. reductions in CO2 emissions, commuting times, and 11. Ahlfeldt and Pietrostefani 2019. vehicular pollution. By implication, our estimates of 12. Lall et al. 2021; Grover, Lall, and Maloney 2022. net benefits should be regarded as lower bounds. 13. Dasgupta, Lall, and Wheeler 2021. 36. Sectors like construction and transport are brown 14. Hommann and Lall 2020. sectors, and crops yellow sectors. Both have large 15. Lall et al. 2021. potential to reduce their carbon emissions. 16. Gendron-Carrier, Gonzalez-Navarro et al. (2020) provide the following definition: “These data define REFERENCES a ‘subway’ as an electric-powered urban rail system isolated from interactions with automobile traffic Adler, M., and J. van Ommeren. 2017. “Does Public Tran- and pedestrians. This excludes most streetcars sit Reduce Car Travel Externalities? Quasi-natural because they interact with vehicle and pedestrian Experiments’ Evidence from Transit Strikes.” ERSA traffic at stoplights and crossings, but underground Conference Papers, European Regional Science streetcar segments are counted as subways. The data Association. do not distinguish between surface, underground, Ahlfeldt, G., and E. Pietrostefani. 2019. “The Economic or aboveground subway lines as long as the exclu- Effects of Density: A Synthesis.” Journal of Urban Eco- sive right of way condition is satisfied. To focus on nomics 111 (C): 93–107. intraurban subway transportation systems, the data Anderson, M. 2014 “Subways, Strikes, and Slowdowns: exclude heavy rail commuter lines (which tend not The Impacts of Public Transit on Traffic Congestion.” to be electric powered). For the most part, these data American Economic Review 104 (9): 2763–96. describe public transit systems that would ordinarily Basagaña, X., M. Triguero-Mas, D. Agis, N. Pérez, C. be described as ‘subways’, e.g., the Paris metro and Reche, A. Alastuey, and X. Querol. 2018. “Effect of the New York city subway, and only such systems.” Public Transport Strikes on Air Pollution Levels in 17. Cities are defined using “functional urban areas” Barcelona (Spain).” Science of the Total Environment (FUAs) developed by Schiavina et al. (2019). 610–611. 18. Dasgupta, Lall, and Wheeler 2022. C40. 2021. “C40 Greenhouse Gas Emissions Dashboard 19. Dasgupta, Lall, and Wheeler 2022. Data, March.” 20. Yang et al. 2018. Chen, Y., and A. Whalley. 2012. “Green Infrastructure: The 21. Anderson 2014. Effects of Urban Rail Transit on Air Quality.” American 22. Adler and van Ommeren 2017. Economic Journal: Economic Policy 4 (1): 58–97. 23. Basagana et al. 2018. Dasgupta, S., S. Lall, and D. Wheeler. 2021. “Urban CO2 24. Chen and Whalley 2012. Emissions: A Global Analysis with New Satellite Data.” 25. Goel and Gupta 2017. Policy Research Working Paper 9845, World Bank, 26. Lalive et al. 2018. Washington, DC. Chapter 2  Decarbonizing cities 61 Dasgupta, S., S. Lall, and D. Wheeler. 2022. “Subways and Emissions Rebound.” Nature Climate Change 2 (4): CO2 Emissions: A Global Analysis with Satellite Data.” 213–214. Doll, C., and O. Balaban 2013. “A Methodology for Eval- JPL/NASA. 2021. OCO-2 Data Set. Jet Propulsion Labo- uating Environmental Co-benefits in the Transport ratory. California Institute of Technology. https://co2. Sector: Application to the Delhi Metro.” Journal of jpl.nasa.gov/?mission=oco-2. Cleaner Production 58: 61–73. Kaw, J., M. Morad, and Y. Li. 2022. “Pathways for Decar- Drupp, M., M. Freeman, B. Groom, and F. Nesje. 2018. bonizing Growing Cities across the Urban-Trans- “Discounting Disentangled.” American Economic Jour- port-Energy Nexus in Jordan.” nal: Economic Policy 10 (4): 109–34. King, M., B. Tarbush, and A. Teytelboym. 2019. “Target Forsberg, J., and A. Krook-Riekkola. 2021. “Recou- Carbon Tax Reforms.” European Economic Review 119. pling Climate Change and Air Quality: Exploring Klenert, D., L. Mattauch, E. Combet, O. Edenhofer, C. Low-Emission Options in Urban Transportation Hepburn, R. Rafaty and N. Stern. 2018. “Making Car- Using the TIMES-City Model.” Energies 14 (11): 1–26. bon Pricing Work for Citizens.” Nature Climate Change Gasser, T., L. Crepin, Y. Quilcaille, R. Houghton, P. Ciais, 8: 669–677. and M. Obersteiner. 2020. “Historical CO2 Emissions Kohli, H. S., J. F. Linn, and L. M. Zucker. 2019. China’s Belt from Land Use and Land Cover Change and Their and Road Initiative: Potential Transformation of Central Uncertainty.” Biogeosciences 17: 4075–4101. Asia and the South Caucasus. Gendron-Carrier, N., M. Gonzalez-Navarro, S. Polloni, Labzovskii, L., S. Jeong, and N. Parazoo N. 2019. “Work- and M. A. Turner. 2020. “Subways and Urban Air ing towards Confident Spaceborne Monitoring of Pollution.” Working Paper 24183, National Bureau of Carbon Emissions from Cities using Orbiting Carbon Economic Research, Cambridge, MA. Observatory-2.” Remote Sensing of Environment 233: Goel. D., and S. Gupta. 2017. “The Effect of Metro Expan- 111359. sions on Air Pollution in Delhi.” World Bank Economic Lalive, R., S. Leuchinger, and A. Schmutzler. 2018. Does Review 31 (1) 271–294. https://doi.org/10.1093/wber/ expanding regional train service reduce air pollution? lhv056. Journal of Environmental Economics and Management 92: Grover, A., S. Lall, and W. Maloney. 2022. Place, Productiv- 744–764. https://doi.org/10.1016​/j.jeem.2017.09.003 ity, and Prosperity: Revisiting Spatially Targeted Policies for Lall, S. V., M. S. M. Lebrand, H. Park, D. M. Sturm, and A. Regional Development. Washington, DC: World Bank. J. Venables. 2021. Pancakes to Pyramids: City Form to Pro- Guo, Q., Z. Wang, Z. He, X. Li, J. Meng, Z. Hou, and J. mote Sustainable Growth. Washington, DC: World Bank. Yang, 2021. “Changes in Air Quality from the COVID to Levy, A. “What is the cost of building a subway?” 2019. Tianjin-­ the Post-COVID Era in the Beijing-­ Tangshan https://marroninstitute.nyu.edu/uploads/files​ Region in China.” Aerosol Air Quality Research 21: /construction_costs_2.xls. 210–270. https://doi.org/10.4209/aaqr.210270. Mao, X., S. Yang, Q. Liu, J. Tu, and M. Jaccard. 2012. He, L.-Y., and L.-Y. Qiu. 2016. “Transport demand, “Achieving CO 2 emission reduction and the co-­ harmful emissions, environment and health co-ben- benefits of local air pollution abatement in the efits in China.” Energy Policy 97: 267–275. https://doi​ transportation sector of China.” Environmental .org/10.1016/j.enpol.2016.07.037 Science and Policy 21: 1–13. https://doi.org/10.1016​ Hommann, K., and S. Lall. 2020. Which Way to Livable and /j.envsci.2012.03.010. Productive Cities? A Road Map for Sub-Saharan Africa. Pan, G., X. Yuan, and M. Jieqi. 2021. “The Potential of CO2 Washington, DC: World Bank. Satellite Monitoring for Climate Governance: A Review.” Jacobs, B., and F. van der Ploeg. 2019. “Redistribution Journal of Environmental Management 277: 111423. and Pollution Taxes with Non-linear Engel Curves.” Raupach, M., G. Marland, P. Ciais, L. Le Quere, J. A. Journal of Environmental Economics and Management 95 Canadell, G. Klepper, and C. B. Field. 2007. “Global (C): 198–226. and Regional Drivers of Accelerating CO2 Emissions.” Jotzo, F., P. Burke, P. Wood, A. Macintosh, and D. Stern. Proceedings of the National Academy of Sciences 104 (24): 2012. “Decomposing the 2010 Global Carbon Dioxide 10288–10293. 62 VI B R A N T C I T I E S Sanchez, L., and D. Stern. 2016. “Drivers of Industrial and World Bank. 2021. World Bank Group Climate Change Action Non-industrial Greenhouse Gas Emissions.” Ecological Plan 2021–2025: Supporting Green, Resilient and Inclusive Economics 124 (C): 17–24. Development. Washington, DC: World Bank. Schiavina, M., A. Moreno-Monroy, L. Maffenini, and P. ment World Bank. 2022. “Egypt Country Climate Develop­ Veneri. 2019. “GHS-FUA R2019A: GHS Functional Report.” Washington, DC: World Bank. Urban Areas, Derived from GHS-UCDB R2019A Wu, D., J. Lin, T. Oda, and E. Kort. 2020. “Space-based (2015).” European Commission, Joint Research Cen- Quantification of per Capita CO2 Emissions from Cit- tre Dataset. http://data​.europa​.eu​/89h​/347f0337​-f2da​ ies.” Environmental Research Letters 15: 035004. -4592​-87b3​-e25975ec2c95. Yang, J., S. Chen, P. Qin, F. Lu, and A. A. Liu. “The effect Stiglitz, J., and N. Stern. 2021. “The Social Cost of Carbon, of subway expansions on vehicle congestion: Evidence Risk, Distribution, Market Failures: An Alternative from Beijing.” Journal of Environmental Economics and Approach.” Working Paper 28472, National Bureau of Management 88: 114–133. Economic Research, Cambridge, MA. Zheng, B., G. Geng, P. Ciais, S. J. Davis, R. V. Martin, Turner, M., and M. Gonzalez-Navarro. 2018. “Subways J. Meng, N. Wu, F. Chevallier, G. Broquet, F. Boersma, and Urban Growth: Evidence from Earth.” Working R. Van der A, J. Lin, D. Guan, Y. Lei, K. He, and Paper 24996, National Bureau of Economic Research, Q. Zhang. “Satellite-based estimates of decline and Cambridge, MA. rebound in China’s CO2 emissions during COVID-19 van der Ploeg, F., and A. Venables. 2020. “Natural pandemic.” Science Advances 6 (2020), 49. https://doi​ Resource Wealth: The Challenge of Managing a Wind- .org/10.1126/sciadv.abd4998 fall.” Annual Review of Economics 4: 313–337. Zheng, S., X. Zhang, W. Sun, and J. Wang. “The effect of Winkler, K., R. Fuchs, M. Rounsevell, and M. Herold. 2021. a new subway line on local air quality: a case study “Global Land Use Changes Are Four Times Greater than in Changsha.” Transportation Research Part D: Transport Previously Estimated.” Nature Communications 12: 2501. and Environment 68 (2017), 26–38. MAKING CITIES INCLUSIVE equitable access to opportunities. investment in skills. aspire to a better life. 3 —­ Making cities more inclusive­ for economic growth, stability, and prosperity PRODUCTIVE CITIES TEND against 12.2 percent in San Francisco.2 As many TO BE MORE UNEQUAL scholars put it, inequality is like cholesterol, which can be good or bad. If income inequality is due to The massive intracity inequality in developing the city’s rewarding skill, ingenuity, talent, effort, and countries is visible even to casual observers. Take grit, it is fair and conducive to city growth in both Mumbai­ —­home to the world’s most expensive pri- the short and longer terms. But if it is due to returns vate residence (Antilia) and to Asia’s largest slum to one’s circumstances of birth in place, gender, (Dharavi). Or Lagos­ —­with about 70 percent of the ethnicity, religion, race, and family background, it is population in slums, and the richest person in Afri- unfair to individuals, unsettling to society, and costly ca.1 Such striking inequality raises the question of to the economy in the longer term.3 This distinction why the benefits of city growth in developing coun- between fair and unfair is particularly important in tries do not reach the mass of citizens. This issue finding out what developing country cities should has grown in recent years as globalization widened do to become more inclusive and productive.4 inequality across the world, and social and political unrest ensued. Inequality in outcomes, such as education or income, is influenced by preferences, efforts, tal- The reason: Productive cities attract skilled and ents, and sometimes pure luck, and thus is not the innovative workers who create jobs for unskilled right measure for policymakers. Instead, inequality workers­ —a­ nd such cities tend inherently to be of opportunity should be the benchmark for ascer- more unequal. But social, racial, and religious divi- taining the extent of inequality that is socially and sions and active segregation policies can also create economically inefficient. Equality of opportunity unequal cities regardless of their productivity. Miami means that life’s outcomes, in its many dimensions, and San Francisco are similar in intracity inequality should reflect mainly effort and talent, not back- (with a Gini coefficient of 0.51), yet a child born in ground. Gender, race, birthplace, family origins, the bottom fifth of the income distribution has a and the like should not determine economic and 7.3 percent chance of reaching the top fifth in Miami social success.5 65 66 VI B R A N T C I T I E S This chapter approaches inequality of opportu- cities the “land of opportunity” for all of their citi- nity using two metrics: whether an individual’s zens? The Sustainable Development Flagship report outcomes depend on parental and neighborhood Thriving: Making Cities Green, Resilient, and Inclusive background, and whether access to markets, basic in a Changing Climate also considers inclusion in infrastructure, and services in a city varies by urban areas, particularly as it pertains to climate neighborhood characteristics. It uses the recently shocks. The analysis here goes beyond cross-sec- available methodology of using correlations tional poverty analysis and focuses more on inter- between parents and children­ —­known as inter- generational mobility. generational correlation­ as a measure of inequal- —­ ity of opportunity. The measure has two advantages INEQUALITY IN DEVELOPING over other measures of inequality of opportunity.6 COUNTRY CITIES First, for children’s outcomes, parental outcomes can be considered predetermined, providing a Large cities in developing countries exhibit higher more direct estimate of inequality of opportunity intracity inequality (figure 3.1). But one of the main instead of a lower bound estimate. Second, the gaps in the economic literature on inequality is measure can be also used to ascertain the extent to the lack of a global characterization of intracity which neighborhood quality matters for children’s inequality due to the unavailability of an inter- outcomes. Recent advances in this literature allow nationally consistent measure. The solution here for estimates of intergenerational mobility compa- exploits survey responses from the Global Monitor rable across geographic areas at increasingly fine Database (GMD) surveys, a multitopic World Bank resolutions and over time.7 collection of income and expenditure surveys, typically collected by the country statistical office The evidence on intracity inequality in developing but later harmonized and compiled in a uniform countries is quite limited. The chapter starts with format. By matching administrative areas and the this evidence using a new dataset, then deals with Global Human Settlement-Urban Centre Database two policy-relevant questions: Are cities escalators (GHS-UCDB), 180 observations from 96 cities in out of poverty over time and generations? And are 60 countries are compiled. Multiple surveys are FIGURE 3.1 LARGE CITIES IN DEVELOPING COUNTRIES ARE MORE UNEQUAL Gini index 0.6 High income Upper middle income Lower middle income 0.5 Low income Casablanca 0.4 Sana’a Tehran Djibouti Amman Tunis Qom 0.3 Baghdad 0.2 11 12 13 14 15 16 17 Population (millions, log) Note: Global intracity inequality across countries is measured by the Gini index. Source: Park et al. 2023. Chapter 3  Making cities more inclusive 67 available for 42 cities, with a global average of 2.6 cities are indeed escalators out of poverty, this observations per city (3.3 excluding single-survey chapter relies on data on intergenerational mobil- cities). ity and access to jobs and health, education, and other services by different groups as indicators of This consolidated dataset makes it possible to unfair inequality. compute different measures of inequality based on the relevant measure of welfare in the survey Intergenerational mobility as an indicator (income or expenditure).8 Globally, the relation- of unfair inequality ship between city population and the Gini index displays a U-shaped curve. Large cities in upper- If all children living in a city enjoy the same oppor- middle-income countries (such as Bogota, Cape tunities regardless of their parents’ status and Town, Johannesburg, Mexico City, and Rio de neighborhoods’ characteristics, their outcomes Janeiro) display particularly high intracity inequal- should mainly reflect their own efforts, talents, ity. Latin American cities on average have the high- and work ethic. There may still be some positive est intracity inequality, and Europe and Central correlation between children’s and parents’ out- Asian cities the lowest (figure 3.2). comes because of genetics, but these associations should be weak. Larger correlations reflect greater ARE CITIES ESCALATORS inequality of opportunity and vice versa. OUT OF POVERTY? Intergenerational mobility in developing Larger cities in developing countries are home to countries a disproportionate share of skilled workers, which explains part of their higher intracity inequality. The literature on intergenerational mobility is But the extent of unfair inequality that results quite sparse, mainly due to sparse data. Income from inequality of opportunity and impedes city data for both parents’ and children’s generations growth is not known in developing countries are nonexistent in developing countries. So, inter- because data are lacking. To find out whether generational mobility estimates for developing FIGURE 3.2 LARGE LATIN AMERICAN CITIES ARE THE MOST UNEQUAL Average Gini inequality index (unweighted) Average population from GMD surveys (millions) 0.5 5 0.4 4 0.3 3 0.2 2 0.1 1 0.0 0 Europe & Middle East & Sub-Saharan East Asia Latin America Central Asia North Africa Africa & Pacific & Caribbean Note: Global intracity inequality across countries is measured by the Gini index. Source: Park et al. 2023. 68 VI B R A N T C I T I E S BOX 3.1 WHY SHOULD CITIES CARE ABOUT INCLUSION AND EQUITY? The prospect of higher rewards often motivates workers to acquire human capital and exert more effort, firms to invest in technology embodied in human and physical capital, researchers and scientists to innovate and discover new efficiency-­ improving technology, and the wealthy to finance many research and development activities. Inequality would be instrumental in ensuring growth and prosperity­ —­ making equity and efficiency substitutes for each other. Should cities worry about inclusion given this equity–efficiency trade-off? Society may care for equity on moral grounds even if inequality is a substitute for efficiency, but there is a case for equity on economic grounds as well. As the 2006 World Development Report pointed out, equity and effi- ciency could complement each other for two important reasons. First, when markets, other economic institutions, and political and social arrangements systematically favor the interests of the wealthy, the middle class and the poor may end up with unexploited talent, reducing efficiency. Second, unequal opportunities could perpetuate inequality over time and generations, creating inequality traps, and making inequality bad for city growth in the longer term. Ultimately, the question whether equity and efficiency are substitutes or complements in the context of cities is empirical, and empirical evidence­ —­ mainly from the United States­ —­provides considerable support for the complementarity hypothesis. How does intracity inequality affect city growth? Developing countries generally lack reliable gran- ular data on urban growth and inequality to answer this question. In the United States, the raw correla- tion between intracity income inequality and city growth (population and income) is positive, as both high- and low-skilled workers concentrate in larger cities.1 But after controlling for skill composition, the effect of inequality on city growth is significant and negative, supporting complementarity. More diverse cities increase the productivity of both skilled and unskilled workers and nurture more start-ups and young enterprises.2 City outcomes show strong persistence over time and generations. Pollution during the industrial revolution sorted the poor to the east side of most cities in the United Kingdom, and this past pollution still explained 20 percent of income segregation in 2011.3 The path dependence in neighborhood effects arises because higher housing costs force the poor to sort into less desirable neighborhoods.4 These less desirable neighborhoods have significantly higher crime rates and poor public services and amenities­ —­ and are vulnerable to pollution and environmental hazards. The evidence from developed countries suggests that intracity inequality is bad for city growth and can create inequality traps. The adverse effects of inequality are expected to be much more pro- nounced in cities in developing countries where markets are missing or distorted (as for credit and insurance), property rights are selectively enforced, and the distribution of public services favors the wealthy. The despair from persistent inequality periodically leads to social unrest and destruction. In short, cities in developing countries should care about inclusion for their own growth, stability, and prosperity. Notes: 1. Glaeser, Resseger, and Tobio 2009. 2. Eeckhout, Pinheiro, and Schmidheiny 2014; Duranton and Puga 2001. 3. Heblich, Trew, and Zylberberg 2021. 4. Guerrieri, Hartley and Hurst 2013; Couture et al. 2019. countries are available only for education. Fortu- equality of opportunity.9 Estimates of absolute and nately, education mobility estimates are strongly relative mobility for urban areas are not available and positively correlated with income mobility separately for most developing countries.10 This and thus provide reasonable approximations of chapter relies on national estimates to detect Chapter 3  Making cities more inclusive 69 FIGURE 3.3 INTERGENERATIONAL MOBILITY­ ABSOLUTE (BETTER OFF THAN PARENTS) AND —­ RELATIVE (MOVING UP FROM PARENTS’ POSITION) Parents Offspring Absolute upward intergenerational mobility Offspring are better off than their parents Relative intergenerational mobility Offspring of parents who are relatively poor can become middle class or upper class among their generation Source: Narayan et al. 2018. BOX 3.2 WHY IS CHILDREN’S SOCIOECONOMIC STATUS POSITIVELY CORRELATED WITH THAT OF PARENTS? Most parents invest in their children’s education, a decision guided by children’s expected returns. These returns are equated to the opportunity cost of investment to determine the optimal investment. They vary inversely with parents’ status if children from poorer families and different racial, ethnic, and religious backgrounds face barriers to better jobs due to low access to hiring networks and physical isolation from city centers where those jobs are. Similarly, the opportunity costs of funds required for investment are typically higher for disadvantaged groups due to unequal access to credit and insurance markets. Parents with disadvantaged backgrounds may also not be efficient decisionmakers due to a lack of information and experience in navigating schooling institutions. Thus, public investment in schooling can relieve the credit constraint and reduce the correlation between children and parents, but it is not enough to ensure equality of opportunity due to unequal access to labor markets and information. trends and regional differences and draw lessons Absolute mobility in education attainment is much for urban areas from the literature. Much of the lower in developing countries, though it has been analysis here is made possible by the recent avail- improving over time, narrowing the gap with devel- ability of these estimates from the Global Database oped countries. Absolute mobility here is the share on Intergenerational Mobility (GDIM) dataset.11 of respondents with higher education attainment 70 VI B R A N T C I T I E S than both parents.12 It displayed an upward trend Having been higher, relative mobility has been for developing countries between 1950 and 1980 lower in developing than developed countries (figure 3.4), when the trend in high-income econ- since around 1970. Based on the intergenerational omies was downward. But much of the developed correlation between children’s and parents’ educa- world had already experienced significant improve- tion from the GDIM dataset,18 relative mobility was ments in absolute mobility before the 1950s, with lower in developing countries before the 1970s than a much sharper decline in recent decades.13 In in developed countries (figure 3.6). But persistence developing countries, the gain in absolute mobility in developing countries increased substantially for was quite broad-based, as all regions except Europe those born in the 1980s, exceeding that of developed and Central Asia experienced substantial gains countries, with regional variations (figure 3.7). Rela- (figure 3.5). With increased public investment in tive mobility has been lowest in LAC countries, but it education and stronger economic growth, it is safe improved over time. For all other developing regions, to conclude that this trend has continued since the relative mobility worsened for recent birth cohorts. 1990s. Though intracity inequality is higher in larger cit- Urban areas in developing countries fared much ies, intergenerational relative mobility tends to be better than rural areas on absolute mobility. higher in cities. Although several recent articles Among 27 African countries, absolute mobility followed the lead of Chetty et al. (2014) in analyz- in urban areas is on average significantly lower ing the determinants of absolute intergenerational than in rural areas, and the rural–urban gap is the mobility, similar analyses of relative intergenera- highest in countries with low mobility and litera- tional mobility are still rare. Using Latinobarome- cy.14 In addition, proximity to larger cities, and to ter data for 18 LAC countries, a recent study finds colonial-era railroads and mission schools, is the that from the 1950s, relative mobility in large cities strongest determinant of mobility.15 These results has been higher than that in smaller cities, and that are largely confirmed for India16 and for 24 Latin over time, large cities gained more mobility.19 For American and Caribbean (LAC) countries,17 where other developing regions, relative mobility patterns the rural–urban gap appears to be narrowing. in urban areas are unavailable. In India, however, FIGURE 3.4 ABSOLUTE MOBILITY IN SCHOOLING CLIMBED IN DEVELOPING COUNTRIES AS A GROUP Percent of adults 70 High-income economies 60 Developing economies 50 40 30 1950 1960 1970 1980 Cohort (decade people are born in) Source: GDIM 2020. Chapter 3  Making cities more inclusive 71 FIGURE 3.5 ABSOLUTE MOBILITY IN SCHOOLING PICKED UP IN MOST DEVELOPING REGIONS Percent of adults 1950 1960 1970 1980 80 60 40 20 0 East Asia Europe & High Latin America Middle East & South Sub-Saharan & Pacific Central Asia income & Caribbean North Africa Asia Africa Source: GDIM 2020. FIGURE 3.6 RELATIVE MOBILITY IN SCHOOLING OVER TIME 1 – intergenerational correlation 0.62 High-income economies 0.60 0.58 Developing economies 0.56 1950 1960 1970 1980 Cohort (decade people are born in) Note: Relative mobility is 1 – intergenerational correlation (IGC), since IGC captures the intergenerational persistence. Source: GDIM 2020. relative mobility in urban areas tends to be slightly lower absolute and relative mobility than men and lower than in rural areas, though urban areas wit- forward-caste (higher social status) groups.21 Gen- nessed a larger improvement in mobility in the der gaps in absolute mobility are not significant in 1980s and 1990s.20 Indonesia shows almost no dif- African and Latin American countries. But even in ference in mobility between rural and urban areas. India, women and low-caste groups experienced higher gains in absolute and relative mobility in Recent gains in absolute and relative mobility have urban areas after economic liberalization, as eco- been higher among women and disadvantaged nomic growth following liberalization, along with groups in urban areas than in rural areas. In India, affirmative action for women and disadvantaged women and disadvantaged groups tend to have groups, spurred the gains (box 3.3). 72 VI B R A N T C I T I E S FIGURE 3.7 RELATIVE MOBILITY IN SCHOOLING BY REGIONS 1 – intergenerational correlation 1950 1960 1970 1980 0.80 0.60 0.40 0.20 0.00 East Asia Europe & High Latin America Middle East & South Sub-Saharan & Pacific Central Asia income & Caribbean North Africa Asia Africa Note: Relative mobility is 1 – intergenerational correlation (IGC), since IGC captures the intergenerational persistence. Source: GDIM 2020. Neighborhood and regional quality matters for Are cities the land of opportunity for all intergenerational mobility in developing coun- citizens? Access to jobs, housing, and tries. Children growing up in the same neighbor- services hood face common neighborhood factors such as school quality, crime, security, and peer and This section addresses the question in two ways. role model effects. Some of the effects are due While service provision tends to be better in cities to families with similar attributes sorting into than in rural areas, intracity differences are large the same neighborhood. But neighborhoods can in developing countries (as seen). But compre- have a direct causal effect beyond sorting. Liter- hensive treatment of how access to basic services ature from developed countries shows that being varies within developing country cities is again stuck in a poor neighborhood during childhood impossible, given the lack of data. So, this section leads to lower future earnings, a lower likelihood draws on case studies. Urban slums house perhaps of going to college, and a higher chance of being the poorest residents of a city. Slums lack access incarcerated or having a teenage pregnancy.22 to basic services and have unclear property rights. Social interactions in the neighborhood help Mobility in slums should provide the clearest evi- with job search and human capital accumula- dence of what poor access to markets (land, labor, tion. 23 In 27 African countries, moves during and credit), shaky and unclear property rights, and childhood to a better neighborhood, particularly adverse neighborhood environments mean for its at an early age, improve children’s educational residents’ ability to move up the welfare ladder outcomes significantly, and about one-third of over time and generations. It then draws on the the exposure effect is due to the neighborhood’s sociology literature to reflect on intergenerational independent effect, not due to the sorting of mobility within slums. households to better neighborhoods. This effect (region in this case) is unsurprising in developing Access to workplaces and commuting countries, because their richer regions provide higher-quality infrastructure and more and bet- High housing costs make living in areas closer to ter services. employment centers unaffordable for poorer city Chapter 3  Making cities more inclusive 73 BOX 3.3 DOES CITY GROWTH IMPROVE FEMALE LABOR FORCE PARTICIPATION? In many MENA and South Asian countries, female labor force participation in cities is dismal, and is much lower than that in rural areas, and is even trending downward in South Asia (box figure 1). BOX FIGURE 1 Female labor force participation is low in MENA and South Asia Female labor force participation rate in urban areas, by decade (percent) 2000 2010 2020 80 60 40 20 0 East Asia Europe & Latin Middle East North South Sub-Saharan & Pacific Central Asia America & & North America Asia Africa Caribbean Africa Source: Based on ILO stats. While social, cultural, and religious norms and economic circumstances play an outsized role in determining female labor force participation, expansion of tradable activities requiring low-skilled workers and white-collar jobs can melt some of these barriers in cities. An example is the expansion of employment in ready-made garments in Bangladesh. The first garment factory was established in Bangladesh in the late 1970s, and the industry’s robust growth generated more than 5 million jobs­ —a ­ quarter of industrial employment and three-quarters of the country’s export earnings in recent years. More important, about 60 percent of garment workers are women. Employment in garments enabled migrant women from rural areas to earn cash income and had far-reaching positive effects on the welfare and empowerment of women in rural and urban areas. The expansion of nontraditional white-collar jobs improved schooling and employment of low-caste women relative to similar men in Mumbai as well.1 In contrast, expansion of activities requiring “brawn”, such as mining, benefits men more. Note: 1. Munshi and Rosenzweig 2006. residents, so they sort into neighborhoods that usu- People in poorer neighborhoods face longer ally are not well connected to transport networks. commuting times24 in Belo Horizonte (Brazil) and Residents deal with this constraint either by work- Cape Town (South Africa).25 In Amman, Jordan, ing near where they live or by commuting longer. origin–destination matrices were computed for commuting patterns using cell phone data (fig- The poor commute to work over longer distances ure 3.8).26 The average commuting distance to the and spend a higher share of their income on trans- —­ workplace in Amman is 2.95 kilometers­ from poor port than residents from affluent neighborhoods. neighborhoods (bottom 10 percent of the income 74 VI B R A N T C I T I E S FIGURE 3.8 COMMUTING DISTANCE AND NEIGHBORHOOD CHARACTERISTICS IN AMMAN Average commuting distance (km) 12 9 6 3 0 0 20 40 60 80 100 Percent of low-income households Source: Park et al. 2023. distribution) about 4 kilometers but from affluent vulnerability index using data for demographics, neighborhoods (top 10 percent) only 2 kilometers. housing, labor, education, and health, based on And in Ulaanbaatar, Mongolia, due to inefficient and the 2010 and 2020 Mexican population censuses limited public transit, low-income commuters at the (figure 3.10). The highly vulnerable areas (clusters urban periphery spend as much as 24–36 percent of 4 to 6) are mainly toward the southwestern part of their monthly expenses on bus fare.27,28 Such long the city, and the less vulnerable (most of clusters 1 commuting times and high transport costs extract to 3) are toward the northeast. Such spatial sorting time and money that could be used for other activi- is observed in nearly all world cities. ties. They may also increase the risk of being a crime victim at night; the commuting time and cost; and Developing country cities are growing rapidly due the disconnections for the poor­ —­ especially women­ to natural increase and migration from rural areas in —­from job opportunities, thus increasing their the face of climate shocks and, in many cases, to ref- unemployment rate.29 So, women choose jobs near ugees and internally displaced refugees. Case stud- their place of residence. Residential networks have ies highlight different aspects of population growth a strong effect on hiring and on worker–employee and what they mean for housing. In Ulaanbaatar, the matches, but workers in isolated neighborhoods rapid expansion of the population has dramatically face exclusion from these networks.30 increased the use of portable felt tents as housing.31 Such ger districts now host about 60 percent of the Access to housing and services total in Ulaanbaatar. In Jordan, the vast majority of ­refugees­—­81 percent­—­live in cities instead of camps. Unaffordable housing in developing country cities In Metro Manila, Philippines, about 600,000 house- creates income segregation as the poor and vulner- holds with about 3 million people live in informal able congregate in poorer neighborhoods. House- settlements, severely challenging the provision of holds sort into neighborhoods they can afford, basic infrastructure and public services. leading to segregated residential patterns (box 3.4). This pattern is clear in Amman and Mexico City. Informal settlements are vulnerable to natural haz- Amman is divided into east and west, with the west ards and have inadequate services and amenities. having fewer low-income households than the east Ulaanbaatar’s sprawling ger districts bring huge (figure 3.9). Duque et al. (2021) developed a spatial infrastructure challenges for sanitation, water, Chapter 3  Making cities more inclusive 75 BOX 3.4 ACCESS TO CREDIT AS A TOOL FOR NEIGHBORHOOD SEGREGATION? In 1933, the Home Owners Loan Corporation (HOLC) was created to provide property appraisals to home buyers in the United States. The HOLC drew maps for more than 200 cities as part of its City Survey Program to document the relative riskiness of lending in neighborhoods. Neighborhoods were classified based on detailed risk-based characteristics of the house as well as non-housing attributes such as race, ethnicity, and immigration status. The lowest-rated neighborhoods were drawn in red and often had a vast majority of African-American residents. Borrowers from these redlined neighborhoods were denied access to credit due to the demographic composition of their neighborhoods. For more than two decades, the redlining in effect barred African-Americans from buying homes in attractive neighborhoods, even when they could afford them. This discriminatory practice was banned by the Fair Housing Act in 1968, yet more than 50 years later, the formerly redlined neighborhoods fare worse on housing price, home ownership, racial com- position, and pollution exposure. In other words, the redlining practices ended up creating inequality traps in poorer neighborhoods. Source: Aaronson, Hartley, and Mazumder 2021; Lane et al. 2022. FIGURE 3.9 FIGURE 3.10 EAST AMMAN HAS MORE LOW-INCOME SOUTHWESTERN MEXICO CITY HAS HOUSEHOLDS THAN WEST AMMAN MORE VULNERABLE HOUSEHOLD CLUSTERS THAN THE NORTHEAST Intraurban clusters, 2020 123456 Low-income households (%) 0 100 Source: Park et al. 2023; Housing Census 2015 (Department of Statistics 2015). health, and roads. On the outskirts of Mongolia’s cities, more than 70 percent of households in ger districts lack access to sanitation facilities (figure 3.11). Coal-burning stoves used in ger districts directly generate as much as 80  percent of air pollution, with severe health implications espe- cially for children and the elderly prone to respi- Note: Green indicates high vulnerability. ratory disease.32 Over the past decade, respiratory Source: World Bank datalibweb. 76 VI B R A N T C I T I E S FIGURE 3.11 and building heights, unregulated construction, ULAANBAATAR’S GER DISTRICTS HAVE and unclear property rights and their enforcement. POOR ACCESS TO SANITATION Intergenerational mobility in urban slums Sükhbaatar and informal settlements in developing Chingeltei countries Songino Khairkhan Slums have precarious legality and almost nonex- istent services.35 An estimated 55 percent of the Bayangol Bayanzurkh urban population in Sub-­ Saharan Africa, 30 per- cent in South Asia, and 20 percent in Latin Amer- ica live in slums.36 Living conditions range from Khan Uul homes erected with four poles and cardboard and tarpaulins, with no access to municipal services, to multistoried concrete structures connected to most services.37 Slums’ residents generally lack property title, work in informal jobs, and remain Khoroo District disconnected from jobs, infrastructure, and social Probability that people lack access to sanitation Less than 30% 30–70% More than 70% services. Source: Singh 2017. Educational mobility in urban slums is high but occupational mobility is low. In 14 slums in Ban- infections increased by 2.7 times, and children in galore, India, investing in children’s education was Ulaanbaatar have 40 percent lower lung function the top priority for households, and slum children than children in rural areas. are more educated than their parents.38 They also attend higher-quality schools. Most families Access to education and health facilities and to in Bangalore gain income and assets over time, green space differs across neighborhoods, reflect- except longer-term (fourth generation) residents ing income segregation (figure 3.12). In Amman, stuck in the slums. Jakarta slums have higher inter- only 49 percent of students reside within a 20 generational absolute and relative mobility among minute walk to kindergarten or school. More than younger children than older children.39 The gains in 10 percent of residents lack accessible medical educational mobility in these slums echo broader services, particularly on the city’s outskirts. Elderly improvements in schooling attainment in Indo- people reside farther from healthcare facilities nesia, alleviating concerns that slum residents are than other age groups.33 And a quarter of Amman trapped in a low, human capital equilibrium. But residents are far from any public space and parks, educational mobility does not translate into higher taking more than an hour to walk there; another occupational mobility in either India or Indonesia. tenth of residents take more than two hours to Most slum residents­—­particularly women­—­work in reach parks and public spaces.34 slums and cannot secure more formal, high-paying jobs due to the lack of access to job networks and Housing and land markets and inadequate urban to isolation from city centers. Slum residents also planning are spreading cities and proliferating face substantial downward mobility in the event of informal settlements. Most developing country health and other negative shocks.40 cities pay little if any attention to accommodating an unexpected rise in population. Serious issues in Economic growth alone is not enough to trans- urban land management include restricted zoning form slums into neighborhoods well connected Chapter 3  Making cities more inclusive 77 FIGURE 3.12 LONG TIMES TO REACH SCHOOLS, HEALTH FACILITIES, AND GREEN SPACES Educational facilities (K–12) and Health care facilities and elderly Parks and public spaces and total school-aged population population (65 and older) population 49% 40% 35% 34% 30% 23% 19% 16% 17% 15% 10% 10% Less than Less than Less than More than Less than Less than Less than More than Less than Less than Less than More than 20 minutes 40 minutes 60 minutes 60 minutes 20 minutes 40 minutes 60 minutes 60 minutes 20 minutes 40 minutes 60 minutes 60 minutes Source: Park et al. 2023. to jobs and services and to lift slum dwellers up the early 1900s.42 This period also saw wide expan- the economic ladder. Slums provide affordable sion of social safety nets, solid public provision of housing for unskilled migrants and native work- education and health care, protection for workers ers, keeping wages in cities low and contributing in the labor market, and reforms of property rights to city growth. With the intergenerational upward in the land markets and housing sector.43 mobility in slums, city growth is expected to allow residents to move to better neighborhoods or spur Climate change­—­how it affects the poor the development of slums. There is little evidence and vulnerable disproportionately of slums developing on their own in developing countries, though housing tends to improve over The poor and vulnerable are doubly disadvan- time, as reflected in roofing material from satellite taged in dealing with both slow- and sudden-on- images.41 set climate events. The poor are exposed more frequently to climate shocks because of informal In developed countries, the rebuilding of slums housing. Once exposed, they are also less equipped and the movement of residents to better neighbor- to deal with the shocks because of lack of access to hoods were facilitated by robust urban growth in services. The disproportionate burden of climate 78 VI B R A N T C I T I E S BOX 3.5 REFUGEES AND LOW-QUALITY HOUSING IN AMMAN Over the past decade, most new buildings in Amman were constructed at the fringe of or in poor neigh- borhoods. With the vast majority of refugees living in cities instead of camps, the incoming population has caused severe challenges in maintaining and providing basic infrastructure and public services. Limited planning resources and statistics hamper understanding of how Amman has spatially evolved recently. But by combining machine learning and very high-resolution imagery, building footprint layers for 2010 and 2020 were developed to identify growth at an urban scale. The entire study had 138,500 buildings in 2010 and 150,500 buildings in 2020, for an increase of 12,000 (box figure 1, left image). Of 150,500 buildings in Amman, 5 percent of them are low quality, mainly in the eastern part of the city (right image). BOX FIGURE 1 12,000 new buildings from 2010 to 2020 Number of new buildings in 2020 Predicted regular buildings Predicted low-quality buildings 0 200 Refugee camps Source: Robinson et al. 2022. change for the poor, women, children, and elderly poor and impair their forward-looking decision- are observed in three MENA countries (box 3.6). making. Second, the inequality of opportunity resulting from inequitable public investment in INFORM, SUPPORT, AND PROTECT basic services is due to the poor’s lack of voice TO MAKE CITIES MORE INCLUSIVE and agency. Yet cities can be made more inclusive in three ways: improving access to housing, jobs, Unfair intergenerational inequality that creates finance and services, protecting through targeted inequality traps can be traced to two fundamental safety nets and incentives, and ensuring that tje factors. First, markets and institutions disfavor the voices of the poor are heard. Chapter 3  Making cities more inclusive 79 BOX 3.6 THE BRUNT OF CLIMATE-RELATED HAZARDS AND THE COVID-19 PANDEMIC HAS FALLEN DISPROPORTIONATELY ON THE POOR, WOMEN, ELDERLY, AND OTHER VULNERABLE GROUPS IN THREE MENA COUNTRIES Jordan. Unemployment rates among women and youth are very high even in normal times (box fig- ure 1). Official figures show that unemployment rose sharply during the pandemic, reaching 25 percent at its highest point in 2021 Q1. Unemployment rates for women and youth have been hit the hardest by the pandemic, when youth unemployment jumped to unprecedented rates­ —­50 percent in 2020 Q4 among 15- to 24-year-olds. Job creation has been focused on low-skilled, low-wage jobs. BOX FIGURE 1 Unemployment rates in Jordan Unemployment rate (percent) 50 Total 40 Female 30 20 Youth (age 20–24) 10 Structural break 0 2000 2005 2010 2015 2020 Egypt. In metropolitan cities, the poor are more likely to locate in more hazardous sites and slums, and have higher exposure to air and water pollution and to congestion, and lack green amenities. Further, poor households do not own many assets nor have the right skills to allow them to cope with shocks, self-insure, or adapt. Women are also more vulnerable to climate change due to their low access to income opportunities, limited productive assets and access to services, and limited social capital and voice in decisionmaking. The poor are also at increased risk from the transition to a low-carbon econ- omy because they are more engaged in informal and “brown” activities such as construction, which emits far more carbon than the world’s median emission level. Morocco. Climate-related hazards and their cascading and compounding effects are increasingly con- tributing to adverse health outcomes, including communicable and noncommunicable diseases and chronic and severe mental health disorders, leading to increased mortality and morbidity. Extreme heat exposure is a growing risk to the population’s health, particularly for aging populations, pregnant women, and other vulnerable groups. The incidence of communicable diseases such dengue fever, malaria, and schistosomiasis is also projected to increase. These health outcomes are often unevenly distributed, and notably worse for vulnerable groups. Source: Egypt, Jordan, and Morocco Country Climate and Development Reports. 80 VI B R A N T C I T I E S Improve access to housing, jobs, and Cities should reform zoning regulations to allow finance mixed-use neighborhoods, building codes to facilitate construction of low-end apartments, As people act to protect themselves from a tsunami and laws on land use to contain the “not in my by moving to higher ground, so they strive to better backyard” phenomenon from existing property their lives by taking action to overcome the myriad owners. Cities and countries should also ban constraints imposed by market and information discriminatory practices where landowners can failures. All households­—­rich or poor­—­decide draft covenants that make it almost impossible where to work, where to live, and what to invest in for owners to sell houses to racial, ethnic, and on the basis of information about labor, housing, religious minorities. and credit markets; about commuting; and about • Provide information on credit and labor markets. school quality and other neighborhood amenities. The poor have limited access to formal credit Markets for housing, credit, and labor influence markets because they lack collateral, and even these decisions directly, and failures in these mar- when they have access, they are assessed as kets often cause inequitable access to work, hous- high-risk borrowers. This designation is a pow- ing, and investment. Having these markets work erful tool to restrict the mobility of the poor with clear rules rather than patronage is perhaps to better neighborhoods and has been used to the first step in making cities inclusive. enforce racial segregation within cities in the • Revamp land and housing markets. Slums and United States. There is evidence of an implicit informal settlements signal inadequate sup- bias in hiring as well. In India, low-caste appli- plies of affordable housing stemming from the cants need to send 20 percent more resumes inefficiency of land and housing markets. The than high-caste applicants to get the same calls inefficiency arises from a lack of clear property back.45,46 Discriminatory practices in credit and and tenancy rights, as well as regulations that labor markets, whether conscious or uncon- restrict housing supply. Clarifying property scious, should be addressed through legal pro- rights fosters investment from developers and visions at the national level. Cities can do their homeowners alike, and tenancy rights foster in- part by banning neighborhood-level demo- vestment by tenants. These reforms increase the graphic characteristics as criteria for risk assess- supply of formal housing and improve neigh- ment. Cities can also confront labor market dis- borhood quality. Reforming property rights to crimination by being vigilant in their own hiring allow women to own and use land can also help and procurement, and by encouraging the re- improve the participation of women in eco- cruitment of women and minority candidates. nomic activities. • Use regulations to promote inclusive housing de- Improve access to infrastructure and velopment. Yet increased formal housing can service provision have negative spillovers to surrounding low- cost informal settlements.44 In the 2000s, about Along with reforms to deal with market failures 15 percent of land in Mumbai became avail- and enable poorer households to make better deci- able for redevelopment after the closing of sions, public investments are needed to connect defunct textile factories. As developers built them to opportunities. high-rise and high-end housing in these areas, • Invest in the provision of basic services to build the gentrification of slums in nearby locations human capital. All children regardless of circum­ forced poorer households out. So, policies to stances of birth should have access to qual- increase housing supply have to be combined ity education and health services. A defining with policies to maintain affordable low-cost characteristic of urban slums and informal set- housing in relatively affluent central locations. tlements is the lack of access to basic services Chapter 3  Making cities more inclusive 81 such as hygiene, sanitation, and safe drinking conditional cash transfers, unconditional cash water. About one-third of slum populations are transfers, social assistance, pensions, and pub- children, whose lack of access to these services lic works is 10 percentage points higher in rural has serious implications for their health­ —­even areas than in urban areas.47 In contrast to social survival­ and human capital. The health shocks —­ assistance programs, social insurance programs hold down the labor productivity of adult work- such as social security and unemployment ers. The poorer households in slums are also benefits, which are not needs-based but con- least equipped to absorb these shocks and often tribution-based, are more prevalent in urban get stuck in slums for generations. areas, though the gap between rural and urban • Plan and invest in urban mobility to connect the poor areas is not as large as for social assistance. In to employment opportunities. Urban planning has response to massive job loss among informal not kept pace with urban population growth, workers during the Covid-19 lockdowns, many particularly given the influx of refugees and in- developing countries had to expand their social ternally displaced persons. These marginalized safety-net programs in urban areas on an emer- populations­—­often women­—­are forced into in- gency basis. City leaders should draw on the les- formal settlements at the edge of cities with poor sons from expanding urban safety nets during connectivity to centers of employment. Con- the pandemic to provide systematic protection necting poorer neighborhood through transport to its poorest citizens so that they can escape investment is, however, a double-edged sword. destitution, weather the shocks, and keep their It can lead to gentrification, which will force the children in school. poor to a different and poorly connected cor- • Adopt incentives to break the curse of poor neigh- ner of the city, or the relatively rich can leave borhood quality: Because of path dependence, poor the neighborhood, making neighborhood envi- neighborhoods in the past stayed poor even in the ronments even bleaker. Transport infrastructure longer term, as with the redlined neighborhoods (see plans should be combined with housing devel- box 3.5). Many developed countries designed opment plans to guard against these possibili- programs such as the U.S. Moving to Opportu- ties. In contrast to transport, digital connectivity nity program to introduce more diversity in rel- does not involve such reallocation and might be atively affluent neighborhoods, while allowing easier to implement and equally beneficial. households to move out of low-quality neigh- borhoods. These programs provide housing Protect by providing targeted safety nets vouchers to disadvantaged groups so they can and incentives move to better neighborhoods (box 3.7). House- holds are given school vouchers to enable their Informing to act and investing to connect are not children to attend better schools outside the always enough to protect the poor from sliding neighborhood. These programs are found to be down the poverty ladder in response to job loss or effective in improving households’ outcomes climate or health shocks. National and city leaders as well as intergenerational mobility. Another should consider direct interventions to protect the way to increase diversity in neighborhoods is to most vulnerable citizens through safety nets and provide incentives to developers and builders to incentives. construct affordable housing in relatively afflu- • Invest in urban safety-net programs to protect ent neighborhoods. against temporary adverse shocks and chronic des- titution: The safety-net programs that provide cash Give agency to the vulnerable support in times of need are quite widespread, though the Covid-19 pandemic uncovered a lingering gap in The agency of the poor and vulnerable can be pro- social safety nets in urban areas. The coverage of moted in two ways: providing relevant information 82 VI B R A N T C I T I E S BOX 3.7 ARE HOUSING VOUCHERS ENOUGH TO INDUCE FAMILIES TO MOVE TO HIGH UPWARD-MOBILITY NEIGHBORHOODS? The U.S. Moving to Opportunity program offers housing project residents vouchers to move to low- er-poverty neighborhoods. This program has high uptake as most recipients use their vouchers to rent an apartment somewhere­ —­though not always in the high-quality neighborhoods. A group of researchers ran an experiment in Seattle and King County to see whether customized housing assis- tance designed to reduce barriers to moving to high-upward-mobility neighborhoods indeed leads to higher rentals in those areas by the recipients of housing vouchers, and what impacts such moves have on household outcomes, including income mobility. For the experiment, families who applied for housing vouchers were divided into treatment and control groups randomly. Families from the treatment group were offered a supplementary program called Creating Moves To Opportunity (CMTO), which provided additional services such as customized search assistance and landlord engagement. All families in the treatment group had the option to use their housing voucher in any neighborhood within the housing authorities’ jurisdictions (although CMTO services were provided only in high-opportunity areas). The CMTO treatment increased the share of families who leased units in high-opportunity neighborhoods by 38 percentage points. The CMTO changed where families moved, not whether they moved at all with a Housing Choice Voucher: in both the treatment and control groups, approximately 87 percent of families leased a unit somewhere using their housing vouchers. The CMTO intervention increased upward income mobility by 1.6 percentage points. The CMTO also created persistence in new neighborhoods: Among families who leased up to at least one year earlier, 60 percent of families in the treatment group live in high-opportunity areas, compared with 19 percent in the control group. More stan- dardized policies such as increasing voucher payment standards in high-opportunity areas or just informational interventions have much smaller impacts. So, most low-income families do not have a strong preference to stay in low-opportunity areas; instead, barriers in the housing search pro- cess are a central driver of residential segregation by income. Combining customized assistance in housing search with housing vouchers could reduce residential segregation and increase upward mobility substantially. Source: Bergman et al. 2019. to inform their decisionmaking; and engaging them Engage the vulnerable to community decisionmaking: into community decisionmaking. Often the vulnerable not only lack access to services but also have little influence on the very Provide information: Reforming markets and services that influence their livelihood. With digital institutions­—­by laying the foundation for creating portals, cities can encourage citizens’ engagement —­ inclusive cities­ is not enough to break the inequal- in monitoring and improving service provision. ity trap if not well communicated to the relevant Various social protection measures can also be population. The poor often lack the information tailored to engage the vulnerable in community and experience to make informed decisions. A job- decisionmaking. seeker living in a slum may harbor distorted beliefs about what is feasible in employment opportuni- All these actions are complementary. Property ties.48 With the advent of digital portals, cities can right reforms will induce households to make provide information on jobs and services. the last-mile investment in connecting the house Chapter 3  Making cities more inclusive 83 —­ to water and sanitation lines­ made available by neighborhoods but perhaps also financial assis- public investment. Improvements in water and tance from social assistance programs as well as sanitation will improve children’s health and active engagement of the community and the induce parents to invest more in their education vulnerable in decisionmaking. These actions by sending them to better schools in better neigh- correspond to the inform, support, and protect borhoods. These actions may require not just framework of this report. Any break in the chain better information about those high-opportunity could make other actions moot. 84 VI B R A N T C I T I E S NOTES 24. Kilroy 2009. 25. Rospabe and Selod 2006; Haddad 2020. 1. Krishna and Rains 2021. 26. Mobile phones have proven good instruments to 2. Chetty et al. 2014. measure human behavior (Toole et al. 2012). These 3. World Bank 2006. patterns can be used to track people’s activities. Par- 4. Productivity is discussed in the next chapter. ticularly, by using mobile phone records and their 5. World Bank 2006. patterns, we identified Home and Work locations 6. Brunori, Ferreira, and Peragine 2013; Ferreira and and developed Origin-Destination (OD) tables. Gignoux 2011. 27. Baker and Gadgil 2017. 7. Chetty et al. 2014. 28. Considering the relatively higher price of public 8. The algorithm relied on the World Bank datalibweb transit (0.25–0.65 JD, which is equivalent to US$0.35– (https://github.com/worldbank/GLAD/blob/master​ 0.92), this may reduce poor people’s economic /00_documentation/Datalibweb_Guidelines.md) data opportunities. It is important to note that USAID system. (2014) reported that about 23 percent of young peo- 9. Van der Weide et al. 2021. ple’s salary is allocated to cover transport costs. 10. This is due to data limitations as most household 29. Gobillon, Selod, and Zenou 2007. surveys lack enough sample size to allow nationally 30. Hellerstein, McInerney, and Neumark 2011; Heller- representative estimates of mobility parameters sep- stein, Kutzbach, and Neumark 2019. arately for urban areas. 31. A portable small felt tent used by nomads in Mongolia. 11. GDIM 2020; Van der Weide et al. 2021. 32. UNICEF 2018. 12. Because education is top-coded, these estimates con- 33. Low-income elderly people are, on average, more dition on neither parent having tertiary education. frequent users of hospital care (Billings et al. 1993; 13. Salvanes forthcoming. Evashwick et al. 1984). Loo and Lam (2011) reported 14. Alesina et al. 2021. that about 20 percent of their elderly survey respon- 15. Mobility is better in rural areas than urban areas in dents’ mode of transportation is walking. the United States and high mobility areas have (1) less 34. Computed by synthesizing open-sourced Point of residential segregation, (2) less income inequality, (3) Interests data (such as Open Street Map), gridded better primary schools, (4) greater social capital, and travel time (Weiss et al. 2018), and gridded popula- (5) greater family stability (Chetty et al. 2014). tion (such as WorldPoP) in cloud computing (such as 16. Asher, Novosad, and Rafkin 2021. Google Earth Engine). 17. Munoz 2021. 35. UN-Habitat 2016. 18. Intergenerational correlation estimates have the 36. Rains and Krishna 2021. advantage that they are less biased when there is 37. Krishna, Sriram, and Prakash 2014. co-residency issue in the sample and less subject to 38. Krishna 2013. shifts in children and parents’ marginal distributions 39. Wong 2019. (Salvanes forthcoming; Emran, Greene, and Shilpi 40. Perlman (2006) for Brazil, Moser (2009) for Ecuador, 2018) Zulu et al. (2011) for Kenya, and Turok and Borel-­ 19. Neidhofer, Ciaschi, and Gasparini 2021. Saladin (2018) for South Africa. 20. Emran and Shilpi 2015; Emran, Jiang, and Shilpi 2021. 41. Rains and Krishna 2019. 21. Asher, Novosad, and Rafkin 2021; Emran and Shilpi 42. Glaeser 2011. 2015; Emran, Jiang, and Shilpi 2021. 43. See Rains and Krishna (2021) for a survey of these 22. Chetty, Hendren, and Katz 2016; Chetty and Hendren policies in Europe and the United States. 2018a, 2018b. 44. Gechter and Tsivanidis 2018. 23. Topa and Zenou 2015; Hellerstein, McInerney, and 45. Siddique (2009) for India study. Neumark 2011; Hellerstein, Kutzbach, and Neumark 46. A study by Bertrand and Mullainathan (2004) sent 2019. out resumes with traditionally white-sounding Chapter 3  Making cities more inclusive 85 names like Emily and Greg and also resumes with on Hospital Use in New York City.” Health Affairs 12 (1): Black-sounding names like Lakisha and Jamal and 162–173. found a white applicant was 49 percent more likely Bocquier, P., D. Beguy, E. Zulu, K. Muindi, A. Konseiga, than their Black counterpart to receive a call back and Y. Yé. 2011. “Do Migrant Children Face Greater in Chicago and 50 percent more likely in Boston. A Health Hazards in Slum Settlements? Evidence from 2015 experiment involving baseball card auctions on Nairobi, Kenya.” Journal of Urban Health 88 (2): 266–281. eBay found that baseball cards held by dark-skinned/ Brunori, P., F. Ferreira, and V. Peragine. 2013. “Inequality of African American hands sold for approximately Opportunity, Income Inequality, and Economic Mobil- 20 percent less than cards held by light-skinned/ ity: Some International Comparisons.” In Getting Devel- Caucasian hands, despite the cards held by the Afri- opment Right, 85–115. New York: Palgrave Macmillan. can American hands being more valuable on average. Chetty, R., and N. Hendren. 2018a. “The Impacts of Neigh- 47. Gentilini, Khosla, and Almenfi 2021. borhoods on Intergenerational Mobility I: Childhood 48. Abebe et al. 2020. Exposure Effects.” The Quarterly Journal of Economics 133 (3): 1107–1162. REFERENCES Chetty, R., and N. Hendren. 2018b. “The Impacts of Neighborhoods on Intergenerational Mobility II: Aaronson, D., D. Hartley, and B. Mazumder. 2021. “The County-level Estimates.” The Quarterly Journal of Eco- Effects of the 1930s HOLC ‘Redlining’ Maps.” American nomics 133 (3): 1163–1228. Economic Journal: Economic Policy 13 (4): 355–92. Chetty, R., N. Hendren, and L. Katz. 2016. “The Effects of Abebe, G., S. Caria, M. Fafchamps, P. Falco, S. Franklin, Exposure to Better Neighborhoods on Children: New S. Quinn, and F. Shilpi. 2020. “Marching Frictions and Evidence from the Moving to Opportunity Experi- Distorted Beliefs: Evidence from a Job Fail Experi- ment.” American Economic Review 106 (4): 855–902. ment.” Unpublished working paper, Department of Chetty, R., N. Hendren, P. Kline, and E. Saez. 2014. Economics, Oxford University. “Where Is the Land of Opportunity? The Geography Alesina, A., S. Hohmann, S. Michalopoulos, and E. Papa- of Intergenerational Mobility in the United States.” ioannou. 2021. “Intergenerational Mobility in Africa.” The Quarterly Journal of Economics 129 (4): 1553–1623. Econometrica 89 (1): 1–35. Couture, V., C. Gaubert, J. Handbury, and E. Hurst. 2019. Asher, S., P. Novosad, and C. Rafkin. 2021. “Intergenera- “Income Growth and the Distributional Effects of tional Mobility in India: New Methods and Estimates Urban Spatial Sorting. Working Paper 26142, National Across Time, Space, and Communities.” Bureau of Economic Research, Cambridge, MA. Ayres, I., M. Banaji, and C. Jolls. 2015. “Race Effects on Carneiro, P., I. García, K. Salvanes, and E. Tominey. 2021. eBay.” The RAND Journal of Economics 46 (4): 891–917. “Intergenerational Mobility and the Timing of Parental Baker, J., and G. Gadgil (eds.). 2017. East Asia and Pacific Income.” Journal of Political Economy 129 (3): 757–788. Cities: Expanding Opportunities for the Urban Poor. Wash- Duque, J. C., N. Lozano-Gracia, J. E. Patino, and P. ington, DC: World Bank. Restrepo. 2021. “Urban Form and Productivity: What Bergman, P., R. Chetty, S. DeLuca, N. Hendren, L. F. Katz, Shapes Are Latin American Cities?” Environment and and C. Palmer. 2019. “Creating Moves to Opportunity: Planning B: Urban Analytics and City Science 49 (4). Experimental Evidence on Barriers to Neighborhood Duranton, G., and D. Puga. 2001. “Nursery Cities: Urban Choice.” Working Paper 26164, National Bureau of Diversity, Process Innovation, and the Life Cycle of Economic Research, Cambridge, MA. Products.” American Economic Review 91 (5): 1454–1477. Bertrand, M., and S. Mullainathan. 2004. “Are Emily and Eeckhout, J., R. Pinheiro, and K. Schmidheiny. 2014. “Spa- Greg More Employable than Lakisha and Jamal? A tial Sorting.” Journal of Political Economy 122 (3): 554–620. Field Experiment on Labor Market Discrimination.” Emran, M., W. Greene, and F. Shilpi. 2018. “When Mea- American Economic Review 94 (4): 991–1013. sure Matters: Coresidency, Truncation Bias, and Billings, J., L. Zeitel, J. Lukomnik, T. Carey, A. Blank, and Intergenerational Mobility in Developing Countries.” L. Newman. 1993. “Impact of Socioeconomic Status Journal of Human Resources 53 (3): 589–607. 86 VI B R A N T C I T I E S Emran, M., H. Jiang, and F. Shilpi. 2021. “Is Gender Des- Hellerstein, J., M. Kutzbach, and D. Neumark. 2019. tiny? Gender Bias and Intergenerational Educational “Labor Market Networks and Recovery from Mass Mobility in India.” March 22. Layoffs: Evidence from the Great Recession Period.” Emran, M., and F. Shilpi. 2015. “Gender, Geography, and Journal of Urban Economics 113: 103192. Generations: Intergenerational Educational Mobility Hellerstein, J., M. McInerney, and D. Neumark. 2011. in Post-reform India.” World Development 72: 362–380. “Neighbors and Coworkers: The Importance of Res- Evashwick, C., G. Rowe, P. Diehr, and L. Branch. 1984. idential Labor Market Networks.” Journal of Labor “Factors Explaining the Use of Health Care Services Economics 29 (4): 659–695. by the Elderly.” Health Services Research 19 (3): 357. Iversen, V., A. Krishna, and K. Sen (eds.). 2021. Social Ferreira, F., and J. Gignoux. 2011. “The Measurement of Mobility in Developing Countries: Concepts, Methods, and Inequality of Opportunity: Theory and an Application Determinants. Oxford University Press. to Latin America.” Review of Income and Wealth 57 (4): Kilroy, A. 2009. “Intra-Urban Spatial Inequality: Cities as 622–657. ‘Urban Regions.’” Washington, DC: World Bank. GDIM (Global Database on Intergenerational Mobility). Krishna, A. 2013. “Stuck in Place: Investigating Social 2020. Global Database on Intergenerational Mobility, Mobility in 14 Bangalore Slums.” The Journal of Devel- Development Research Group, World Bank. Washing- opment Studies 49 (7): 1010–1028. ton, DC: World Bank. Krishna, A., and E. Rains. 2021. “Will Urbanization Raise Gechter, M., and N. Tsivanidis, 2018. “Spatial Spillovers Social Mobility in the South, Replicating the Eco- from Urban Redevelopments: Evidence from Mum- nomic History of the West?” In V. Iversen, A. Krishna, bai’s Textile Mills.” Unpublished working paper. and K. Sen (eds.), Social Mobility in Developing Countries: Gentilini, U., S. Khosla, and M. Almenfi. 2021. “Cash in Concepts, Methods, and Determinants. Oxford University the City: Emerging Lessons from Implementing Cash Press. Transfers in Urban Africa.” Social Protection and Jobs Krishna, A., M. Sriram, and P. Prakash. 2014. “Slum Types Discussion Paper 2101, World Bank, Washington, DC. and Adaptation Strategies: Identifying Policy-relevant Gheorghe, A., B. Ankhbayar, H. van Nieuwenhuyzen, and Differences in Bangalore.” Environment and Urbaniza- R. de Sa. 2018. “Mongolia’s Air Pollution Crisis: A Call to tion 26 (2): 568–585. Action to Protect Children’s Health.” Ulaanbaatar, Mon- Lane, H., R. Morello-Frosch, J. Marshall, and J. Apte. golia: National Center for Public Health and UNICEF. 2022. “Historical Redlining Is Associated with Pres- Glaeser, E. 2011. Triumph of the City: How Urban Spaces Make ent-Day Air Pollution Disparities in US Cities.” Envi- Us Human. Macmillan. ronmental Science & Technology Letters. Glaeser, E., M. Resseger, and K. Tobio. 2009. “Inequality in Loo, B., and W. Lam. 2011. “Geographic Accessibility Cities.” Journal of Regional Science 49 (4): 617–646. around Health Care Facilities for Elderly Residents Gobillon, L., H. Selod, and Y. Zenou. 2007. “The Mechanisms in Hong Kong: A Microscale Walkability Assessment.” of Spatial Mismatch.” Urban Studies 44 (12): 2401–2427. Environment and Planning B: Urban Analytics and City Gourevitch, R., S. Greene, and R. Pendall. 2018. Place and Science 39 (4): 629–646. Opportunity: Using Federal Fair Housing Data to Examine Moser, C. O. 2010. Ordinary Families, Extraordinary Lives: Opportunity across US Regions and Populations. Washing- Assets and Poverty Reduction in Guayaquil, 1978–2004. ton, DC: Urban Institute. Brookings Institution Press. Guerrieri, V., D. Hartley, and E. Hurst. 2013. “Endogenous Munshi, K., and M. Rosenzweig. 2006. “Traditional Insti- Gentrification and Housing Price Dynamics.” Journal tutions Meet the Modern World: Caste, Gender, and of Public Economics 100: 45–60. Schooling Choice in a Globalizing Economy.” Ameri- Haddad, M. 2020. “Residential Income Segregation and can Economic Review 96 (4): 1225–1252. Commuting in a Latin American City.” Applied Geog- Munoz, E. 2021. The Geography of Intergenerational Mobility raphy 117: 102186. in Latin America and the Caribbean. Washington, DC: Heblich, S., A. Trew, and Y. Zylberberg. 2021. “East-side World Bank. Story: Historical Pollution and Persistent Neighborhood Narayan, A., R. Van der Weide, A. Cojocaru, C. Lakner, Sorting.” Journal of Political Economy 129 (5): 1508–1552. S. Redaelli, D. Mahler, R. Ramasubbaiah, and S. Chapter 3  Making cities more inclusive 87 Thewissen. 2018. Fair Progress? Economic Mobility across Singh, G. 2017. Urban Poverty in Ulaanbaatar: Understanding Generations around the World. Washington, DC: World the Dimensions and Addressing the Challenges. Washing- Bank. ton, DC: World Bank. Neidhöfer, G., M. Ciaschi, and L. Gasparini. 2021. “Inter- Toole, J., M. Ulm, M. González, and D. Bauer. 2012. generational Mobility in Education in Latin America.” “Inferring Land Use from Mobile Phone Activity.” In Working Paper 2021/14, CAF–Development Bank of Proceedings of the ACM SIGKDD International Workshop Latin America, Caracas. on Urban Computing, 1–8. Neupert, R. 1999. “Population, Nomadic Pastoralism and Topa, G., and Y. Zenou. 2015. “Neighborhood and Net- the Environment in the Mongolian Plateau.” Popula- work Effects.” In Handbook of Regional and Urban Eco- tion and Environment 20 (5): 413–441. nomics, Volume 5, 561–624. Elsevier. Park, H., P. Fan, R. John, and J. Chen. 2017. “Urbanization Turok, I., and J. Borel-Saladin. 2018. “The Theory and on the Mongolian Plateau after Economic Reform: Reality of Urban Slums: Pathways-out-of-poverty or Changes and Causes.” Applied Geography 86: 118–127. Cul-de-sacs?” Urban Studies 55 (4): 767–789. Park, H., N. Lozano-Gracia, G. Rossitti, and O. D’Aoust. UN-Habitat (United Nations Human Settlement Forthcoming. “A Global Review of Intra-Urban Spatial Programme). 2016. Slum Almanac 2015–2016: Track- Inequality.” World Bank: Washington, DC. ing Improvement in the Lives of Slum Dwellers. Nairobi, Perlman, J. 2006. “The Metamorphosis of Marginality: Kenya: UN-Habitat. Four Generations in the Favelas of Rio de Janeiro.” UNICEF (United Nations Children’s Fund). 2018. Mongo- The Annals of the American Academy of Political and Social lia’s Air Pollution Crisis: A Call to Action to Protect Chil- Science 606 (1): 154–177. dren’s Health. New York: UNICEF. Rains, E., and A. Krishna. 2019. “Will Urbanization Raise USAID (United States Agency for International Devel- Social Mobility in the South, Replicating the Eco- opment). 2014. Transport Services and their Impact on nomic History of the West?” WIDER Working Paper Youth Employment and Development in Jordan. Amman, 2019/102, United Nations University–World Institute Jordan: International Youth Foundation Jordan for Development Economics Research, Helsinki. Office. Rains, E., and A. Krishna. 2021. “Informalities, Volatility, Van der Weide, R., C. Lakner, D. Mahler, A. Narayan, and and Precarious Social Mobility in Urban Slums.” Social R. Ramasubbaiah. 2021. “Intergenerational Mobility Mobility in Developing Countries. around the World.” Policy Research Working Paper Robinson, C., A. Ortiz, H. Park, N. L. Gracia, N., J. K. 9707, World Bank, Washington, DC. Kaw, T. Sederholm, R. Dodhia, and J. M. Lavista Weiss, D., A. Nelson, H. Gibson, W. Temperley, S. Peedell, Ferres. 2022. “Fast Building Segmentation from Sat- A. Lieber, M. Hancher, E. Poyart, S. Belchior, N. Full- ellite Imagery and Few Local Labels.” In 2022 IEEE/ man, B. Mappin, U. Dalrymple, J. Rozier, T. Lucas, R. CVF Conference on Computer Vision and Pattern Howes, L. Tusting, S. Kang, E. Cameron, D. Bisanzio, Recognition Workshops (CVPRW), New Orleans, LA. K. Battle, S. Bhatt, and P. Gething. 2018. “A Global https://doi.org/10.48550/arXiv.2206.05377. Map of Travel Time to Cities to Assess Inequalities in Rospabe, S., and H. Selod. 2006. “Does City Structure Accessibility in 2015. Nature 553 (7688): 333–336. Cause Unemployment? The Case of Cape Town.” Pov- Wong, M. 2019. “Intergenerational Mobility in Slums: erty and Policy in Post-Apartheid South Africa, 262–287. Evidence from a Field Survey in Jakarta.” Asian Devel- HSRC Press. opment Review 36 (1): 1–19. Salvanes, K. G. Forthcoming. “What are the Drivers Inter- World Bank. 2006. World Development Report 2006. Wash- generational Mobility? The Role of Family, Neighbor- ington, DC: World Bank. hood, Education, and Social Class: A Review of Bukodi Zulu, E. M., D. Beguy, A. C. Ezeh, P. Bocquier, N. J. and Goldthorpe’s Social Mobility and Education in Madise, J. Cleland, and J. Falkingham. 2011. “Overview Britain.” Journal of Economic Literature (forthcoming). of migration, poverty and health dynamics in Nairobi Siddique, Z. 2009. “Caste-based Discrimination: Evidence City’s slum settlements.” Journal of Urban Health 88 and Policy.” Discussion Paper 3737, Institute for the .org/10.1007​ (Supplement 2): 185–199. https://doi​ / Study of Labor, Bonn, Germany. s11524-011-9595-0. MAKING CITIES PRODUCTIVE driving economic growth. creating jobs. boosting incomes. financing critical investments. 4 Making cities more productive— connecting markets, upgrading skills, raising expectations The rapid growth of cities in developing countries of the economy from agriculture into tradable is a potential springboard for productivity and manufacturing and services. These sectors are income growth. In 2020, 3.4 billion people in low- more able to benefit from the agglomeration econ- and middle-income countries (LMICs) lived in omies in large, vibrant cities. In most countries cities. By 2050, this number is projected to increase that transitioned from low- to high-income status, by more than 2 billion to 5.4 billion.1 The Middle urbanization was complemented by a dual pro- East and North Africa (MENA) region is highly cess of within-sector and across-sector productivity urbanized: 13 of the region’s 20 countries are more growth. Productivity improved within sectors­ —­ than 75 percent urbanized, and only Egypt and particularly, within internationally tradable manu- Yemen had urban shares below 50 percent in 2020. facturing and/or services sectors, often supported Sixty-six percent of the region’s total population by a “Green Revolution” that released labor from lived in cities in 2020­ 10 percent above the global —­ agriculture and lowered urban food prices.3 This average. within-sector channel is consistent with a Solow growth model of accumulated savings financing The continuing urbanization of LMICs creates efficiency improvements and innovation.4 But a substantial opportunity to boost productivity these countries also enjoyed a shift across sectors, and incomes, as people escape the productivity as labor and value addition were drawn out of constraints in sparse rural areas, and benefit from (less-productive) rural agriculture and into more the productivity boost offered by urban agglom- (and increasingly) productive urban services and eration. Urbanization is typically associated with manufacturing. large labor productivity gains, and no country has reached high income without big shares of the pop- Cities can support both within-sector and ulation and economy in large vibrant cities.2 across-sector productivity growth­ —­if properly managed. Urbanization is the spatial corollary of Cities drive growth most effectively when urbaniza- the across-sector structural transformation out of tion is accompanied by a structural transformation agriculture (mainly rural) and into services and 89 90 VI B R A N T C I T I E S manufacturing (mainly urban). Cities also support to employ more skilled workers, who are more able within-sector productivity growth, by concentrating to boost their productivity from agglomeration and connecting large and diverse markets of peo- economies in large cities, due to greater capacity ple, firms, products, and ideas to generate agglom- to learn from others and spread knowledge and eration economies. Duranton and Puga (2004) ideas (learning), and to benefit from the matching codify the agglomeration benefits that this gen- of worker skills with jobs.10 erates as “sharing, matching, and learning”: fixed costs and risks are shared between more people and Because tradables benefit the most from agglomer- firms, allowing more efficient and higher-quality ation economies, cities with fast income growth in infrastructure and services, and better insurance recent decades have typically been led by expand- against risk.5 With a larger and more diverse mar- ing tradable manufactures and services. A study ket, individual firms and workers can also raise of the world’s largest 750 cities found that in the their efficiency by specializing. Specialization fastest-growing cities (the top 10 percent in GDP and diversity allow firms and households to meet per capita growth from 2005 to 2012), tradable sec- more specialized, niche needs at the right price tor employment growth exceeded nontraded sec- (matching), whether their need is a skilled worker, tor employment growth by 2.5 percentage points business services, raw material, machinery parts, annually on average. These cities enjoyed 13.5 per- specialized medical care, or a job that uses their cent per capita income growth and 9.2 percent skills.6 And interactions among firms and workers jobs growth annually on average­ compared with —­ in the city facilitate the generation and diffusion 4.7 percent and 1.9 percent on average in other of ideas (knowledge spillovers), supporting human cities.11 capital accumulation and productive innovation (learning).7 MUTED AGGLOMERATION BENEFITS IN DEVELOPING AND Not all sectors benefit equally from urban sharing, EMERGING MARKET CITIES matching, and learning­ —­with “tradable” sectors tending to benefit most. Nontradables are defined Despite the potential agglomeration benefits as sectors that can cater only to a local market (a offered by cities, rapid urbanization in the Mid- neighborhood, city, or country),8 because, for var- dle East and North Africa (MENA), Sub-­ Saharan ious reasons (such as perishable items, or the need Africa (SSA), and Latin America and the Carib- for face-to-face interaction), they cannot be easily bean (LAC) in the past half century has rarely traded over long distances. Their consumer base been accompanied by strong productivity gains. is therefore limited by the size of the local market, Unfortunately, many migrants from rural to urban meaning returns to scale begin to diminish­ both —­ labor markets in these countries have moved into for technological reasons, and because prices are urban nontradable services and informal manufac- set locally and decline as supply increases.9 This turing, limiting the returns available from urban contrasts with tradables, which are defined as agglomeration. products (and services) that can be traded over long distances. Tradables can tap into compar- MENA is a highly urbanized region, with one of atively inexhaustible international demand to the world’s lowest specializations in tradables. exploit large economies of scale (and to simply sell At 66 percent, MENA is 10 percent more urban- more product). Producers of tradables in develop- ized than the global average. More than half (11) ing countries may also benefit from (often higher) of MENA’s countries were more than 80 percent international income and price levels, while non- urbanized in 2020. Despite MENA’s much higher tradables producers are limited by (typically lower) income and skill levels and high urbanization rates, local incomes and price levels. Tradables also tend the average tradable share in MENA cities is lower Chapter 4  Making cities more productive 91 FIGURE 4.1 CITIES IN MENA DISPLAY A LOWER SHARE OF URBAN TRADABLE EMPLOYMENT THAN OTHER WORLD REGIONS Share of tradable sector in urban employment (percent) 60 50 40 30 20 10 0 South Europe East Asia Latin Sub-Saharan Middle East Mashreq GCC Maghreb Asia & Central & Pacific America & Africa & North Asia Caribbean Africa Source: World Bank 2020a. than that in SSA (figure 4.1).12 MENA’s exports are heard elsewhere in the region. The region has the dominated by extractives and other raw materials highest unemployment rate (10.5 percent in 2021) (constituting one third of the region’s exports), and youth unemployment rate (26 percent in 2019) while MENA’s shares of global intermediate, con- in the world, and one in three MENA youth are sumer, and capital goods exports are considerably (15–24) not in education, employment, or training below its share of global GDP.13 MENA cities have (NEET) (higher than the one in five in SSA). Female typically not grown due to positive productivity labor force participation is also the lowest in the shocks in urban tradables, but as sites for the dis- world, at 20 percent. Those employed in MENA’s tribution of natural resource rents (for example, cities overwhelmingly rely on government or infor- through government employment, political con- mal employment.14 nections, and demand for local nontradables) or due to inflows of displaced people and refugees. Many SSA countries have enjoyed substantial Two countries of focus for this report, Egypt and structural change out of agriculture and into Jordan, have experienced negative structural services and manufacturing as they urbanized in change alongside urbanization in recent decades, recent decades. However, newly urbanized labor with their economies transitioning away from has tended to move into nontradable urban services tradable manufacturing and toward natural and small-scale, informal manufacturing­ with lim- —­ resource exports and nontradable urban services. ited potential to grow through exports, within-firm scale economies, and urban sharing, matching, This concentration in nontradables has contrib- and learning economies. There has been limited uted to low economic dynamism in MENA’s cities. within-sector productivity growth accompanying MENA’s private sector has struggled to grow and urbanization in most SSA countries.15 Although create jobs, resulting in some of the worst employ- unemployment rates are lower than in MENA, ment outcomes globally in the region and its cit- these tend to mask high underemployment, with ies. Firm entry and firm growth are low relative many urban workers employed for few and irreg- to urbanization rates, and labor productivity is in ular hours for low pay, often in self-employment, decline in many countries. Echoes of Egypt’s recent household employment, day labor, urban agricul- challenges of urban job creation (box 4.1) can be ture, and so on. 92 VI B R A N T C I T I E S BOX 4.1 LIKE MANY MENA COUNTRIES, EGYPT HAS STRUGGLED TO TRANSITION OUT OF GOVERNMENT EMPLOYMENT In the past two decades, Egypt has implemented wide-ranging reforms to wean the economy off public sector jobs. Before these reforms, the economy was dominated by the public sector, which accounted for 33 percent of employment and 35 percent of GDP in 2003. From 2003 to 2018, with a raft of reforms, the private sector gradually assumed a larger role, and by 2019, the public sector share of employment had fallen by more than 10 percentage points. But the public share of GDP fell by just 4 percentage points, reflecting a shortage of good income opportunities in the private sector, as it took on greater responsibility for employment creation. As public employment declined, Egypt’s private sector struggled to create high-quality, productive jobs for a growing workforce. In FY2004–2018, Egyptian workers tended to transition into sectors with higher, but declining, labor productivity, often accompanied by rising informality. Sectors that gained in their share of employment despite falling or stagnating productivity include private manufacturing, hotels and restaurants, private social services, utilities, and (perhaps surprisingly) finance. Meanwhile, sectors in which labor productivity was growing tended to shed workers, or exhibit insignificant employ- ment growth (including many public sectors, as well as private agriculture). BOX FIGURE 1 Sector employment BOX FIGURE 2 Increasing share of growth and sector labor productivity are employment “outside establishments” negatively correlated (FY 2004–18) (an indicator for informality) Change in GDP per worker in sector as a share of contribution to overall productivity growth Employment outside establishments (percentage points) as a share of total employment (percent) 12 100 FY 2004 FY 2010 Social services (public) Manufacturing (public) FY 2016 FY 2018 95 6 Finance (public) Wholesale and retail (private) ICT (private) Construction Agriculture (private) (private) 90 Construction (public) Utilities (public) 0 Agriculture (public) Transport (private) Restaurants and hotels (private) 85 Manufacturing (private) Social services (private) –6 80 –10 –5 0 5 10 Private sector Private sector Change in employment share (percentage points) construction transport Note: Size of bubbles corresponds to employment share in FY 2018. ICT is information and communication technology. Importantly, the period saw a substantial shift of labor toward nontradable urban services, where opportunities for scale, specialization, and exporting are limited, and informality is high. Employment grew substantially in the private construction, transport, and wholesale and retail sectors. With their focus on highly localized markets, these sectors offer limited opportunities for productivity-boosting scale and specialization. Nontradable sector employment is often highly informal: 89 percent of con- struction and 95 percent of transport jobs were outside establishments in 2018 (up from 85 percent and 86 percent respectively in 2004). Whereas government jobs offer full-time work, health insurance, pensions, predictability, stability, and paid parental and sick leave, Egypt’s private sector is dominated by informal jobs that lack basic protections. In response, increasing numbers of working-age Egyptians are leaving the labor force. (continued) Chapter 4  Making cities more productive 93 BOX 4.1 continued LIKE MANY MENA COUNTRIES, EGYPT HAS STRUGGLED TO TRANSITION OUT OF GOVERNMENT EMPLOYMENT Labor force participation BOX FIGURE 3 BOX FIGURE 4 Low and declining “quality and employment are declining or stagnant of work” indicators Percent of working-age population Percent of wage workers (age 15 and over) with… 50 80 FY 2016 FY 2020 Labor force participation rate 45 60 40 40 Employment rate 35 20 30 0 2016 2017 2018 2019 2020 2021 Social Health Work Permanent insurance insurance contract work Source: World Bank 2020b. The largest LAC cities have experienced negative of its service sector. The pattern is similar in indus- structural change, and within-sector productivity try, with a 1 percent increase in industry employ- growth has been stifled, in recent decades. Many ment share associated with a 2.2 percent labor were relatively industrialized in the 1970s, with productivity gain in the rest of the world, and only high shares of tradable manufacturing employment 0.9 percent in LAC. supported by import-substitution industrializa- tion policies. With the removal of these policies, These trends in MENA, SSA, and LAC contrast LAC cities became increasingly deindustrialized with successful growth stories in Asia, where posi- “consumption cities,” focused on nontradable tive productivity shocks in urban tradable services services. Manufacturing employment shares fell, and manufacturing drew in labor from agriculture, and employment transitioned to local nontradable creating both across-sector and within-sector pro- services.16 A study of nine major LAC countries ductivity growth.18 The specialization of cities in shows that while the share of services employ- tradables production is most pronounced in cities ment (nationally) went up by 32 percentage points in coastal China, but is also seen in certain cities in between 1960 and 2009, services’ share of value South Asia such as Bangalore, Kanpur, and Dhaka added increased by only 5 percentage points in the (figure 4.2). period, suggesting (akin to Egypt’s case) a declining quality of service sector employment opportuni- This first raises the question­ why have many —­ ties.17 In other world regions, a 1 percent increase developing countries become, or are becoming, in the services share of employment is linked to highly urbanized, if their cities are not deliver- a 3.8 percent increase in overall labor productiv- ing productive agglomeration benefits? Part of ity, but in LAC this increase was just 0.7 percent, the answer may be that cities offer better basic underscoring the depressed productivity potential services than­—­even more underserviced­—­rural 94 VI B R A N T C I T I E S FIGURE 4.2 WORLD MAP OF PRODUCTION CITIES AND CONSUMPTION CITIES, C. 2000 Consumption High Mid Low Production Low Mid High Neutral City population 10 million 5 million 1 million 500,000 Note: In the figure, consumption cities (red to orange) are those specialized in nontradables, and production cities (blue) in tradables. Colors indicate the extent of the respective specialization, from low to high. Source: Jedwab, Ianchovichina, and Haslop forthcoming. areas. This both attracts in-migrants and leads to cities wishing to compete in tradables. To compete natural population growth through higher fertility in global tradables, cities must offer production and lower death rates. In MENA, some cities (such costs (including wages and skills) that are compet- as Amman) have also grown as refuges from rural itive internationally­ or risk being out-competed —­ and cross-border fragility shocks. Cities have also by imports on the domestic market, and by other grown as sites for the distribution of rents from countries’ exports on international markets. Goods natural resources (extractives and cash crops), for- or services that can easily be traded internationally eign aid, and/or government kickbacks.19 Natural will tend to be purchased from the most efficient resources can create large rents, which tend to be global producer, wherever they are located.21 For distributed in cities, through high demand for local this reason, nominal labor productivity and price nontradable services, as well as opportunities for levels in manufacturing are similar around the government employment, contracts, and so on­ —­a world: their easy tradability means that producers common pattern in the MENA region. Urban- who can’t produce at internationally competitive ization that represents a mere crowding around costs tend not to survive.22 This contrasts with (diminishing) returns from natural resource rents nontradable services, for which price levels and can result in cities with high local costs that pre- (nominal) labor productivity vary widely across clude competitiveness in globally tradable manu- countries, and have converged less over time, facturing or tradable services.20 because competition is only local, and prices are thus set locally. If high urban costs drive local price Once density and scale are established, why have levels and production costs too high, this can price cities struggled to translate these into productive cities out of competitiveness in global tradables, agglomeration economies supporting tradables while upholding the (inefficient) local nontradables and income growth? The bar is particularly high for sector.23 Chapter 4  Making cities more productive 95 For strong urban job creation, it is particularly im- and MENA reached 40 percent urbanization at portant that urban labor costs be internationally half the per capita income of EAP when it reached competitive. Gelb et al. (2017)24 highlight that the 40 percent urbanization, and SSA at one-third internationally mobile, foreign direct investment the per capita income,26 depriving cities of capi- (FDI)-led tradable industries that kickstarted large- tal necessary to finance productive investments. scale job creation and industrialization in countries Weak governance and markets compounded cap- like China, Bangladesh, and Vietnam tended to be ital accumulation challenges: when markets are highly labor intensive, making nominal labor costs missing, inefficient, or uncontestable, or when a a key determinant of investment location.25 Rather large government sector introduces distortions than engage in a “race to the bottom” on unskilled or crowds out the private sector, investment in real wages and labor standards to drive down costs, productive, efficient, urban capital is further cities can address weaknesses in human capital, depressed. Frictions and distortions in land and and poor connectivity and land use efficiency real estate markets, for example, prevent the pri- undermining the efficiency of labor markets­ —­to vate sector converting scarce urban land to its achieve competitive nominal wages that allow for most productive uses, depressing the density, effi- decent living standards locally. ciency, and specialization of land, to in turn reduce agglomeration economies and raise the urban costs The large, dense urban populations in many LAC, of living and producing. MENA, and SSA cities have not been supported by a complementary density of capital appropriate FOUR FOUNDATIONS FOR to raising their productivity. This includes human PRODUCTIVE CITIES capital, adequate connective infrastructure, built density, and private investment (box 4.2). LAC, Cities with low productivity and tradable special- MENA, and SSA urbanized at much lower income ization risk becoming stuck in vicious cycles­ —­of levels than East Asia and the Pacific (EAP). LAC low productivity, informing low private expecta- tions, low investment, and hence low agglomer- BOX 4.2 ation, which further cements low productivity. Leaders need to break free of these vicious cycles POPULATION DENSITY WITH through coordinated and strategic policy choices AND WITHOUT CAPITAL DENSITY to raise productivity, investment, and positive agglomeration economies. A range of national Kinshasa, the capital of the Democratic macroeconomic, trade, and business environ- Republic of Congo, and Hong Kong SAR, ment drivers (including ‘natural resource curses’) China, have similar population densities but contribute to cities’ productivity challenges, but are at two extremes of low and high invest- ment in complementary capital. In Hong are beyond the main scope of this report.27 The Kong SAR, China, strong human capital, con- focus here is on four factors particularly relevant nective transport and other public infrastruc- for urban policy and strengthening agglomera- ture, and vertical stacking of the population tion economies­ —­human capital accumulation, in tall buildings support a livable, productive transportation, land use efficiency, and private density. In Kinshasa, the density of people expectations. without human capital, basic services, or ver- tical creation of floorspace results in crowded Skilled urban labor in efficient markets and congested slums with low welfare and little scope for economic transformation. Through an economic lens, vibrant cities can be Source: Lall, Lebratnd, and Soppelsa 2021. seen as labor markets that efficiently match job seekers to employers who can use their skills.28 In 96 VI B R A N T C I T I E S developing countries, the ability to pool labor­ —­not • Through learning, skilled workers accelerate input suppliers­ —­appears to be the greatest attrac- their human capital accumulation as they live tion of firms to large cities.29 Cities scarce in human and work in large cities, learning from others in capital, and whose labor markets are spatially their vicinity. Skilled workers also create human fragmented or uncontestable, cannot perform this capital spillovers that raise the productivity of matchmaker role efficiently.30 others. La Roca and Puga (2017) find that, after the initial income boost of moving to a big city, Cities’ function of pooling and matching workers workers subsequently also accumulate human and jobs is most transformative in labor markets capital faster, as they “learn by working” in large with high scale, skills, and diversity. The literature cities; the authors also find that this learning is distinguishes quantity effects­—­the scale of the urban most powerful for more skilled workers. Duran- labor markets­—­from quality effects­—­the tendency ton (2015) estimates that this learning within of large cities to attract higher-quality, more skilled the city may add 50–125 percent to a worker’s workers. This split is helpfully complemented by a income over their working life. Bacolod, Blum, third component­—­diversity.31 Vibrant labor mar- and Strange (2009) demonstrate that this abil- kets offer diverse, complementary skills to enable ity to learn in cities depends not just on cogni- a strong ‘matching’ of workers to specialized needs. tive skills, but also softer “people skills,”34 while Skills will be increasingly important as workers’ Charlot and Duranton (2004) show that urban income aspirations rise and as technological advan- workers enjoy a wage premium to the extent tages allow more routine tasks to be automated. that they communicate with workers from outside their own firm. Jaffe, Trajtenberg, and Quality and quantity effects are in fact complemen- Henderson (1993) show that proximity matters tary: large, thick, urban labor markets (quantity) to knowledge diffusion, finding that innovators are powerful catalysts for the conversion of skills are more likely to cite patents from other inno- (quality) into productivity. More skilled workers vators located nearby than those far away. not only enjoy a larger productivity boost from working in large cities than unskilled workers­ —­ they Supporting urban workers with productive human also contribute more to other residents’ productiv- capital requires an education system fit to deliver ity by generating human capital externalities.32 The marketable skills, as well as well-functioning labor key catalysts for this process are the agglomeration markets that efficiently match workers with jobs economies of matching and learning. that utilize their skills. Yet human capital develop- • Through matching, large cities host a diversity ment is often undermined by the low or unequal of specialized firms and jobs, making them quality of education. Existing human capital can better able to match high-skill workers to roles also be underutilized, when the efficiency of that can utilize their skills. The tradable sectors worker–firm matching is eroded by spatial frag- that thrive in large cities also typically require mentation of the urban labor market (due to poor higher and more diverse skills. In Spanish cities, transport connectivity and land markets), or by skilled workers receive an initial income boost inefficient, stagnant, or discriminatory labor mar- when moving to a city from a rural area, which kets. These may erect barriers to the employment persists over time; this benefit is greater in large of minority groups, women, or people without cities than small cities, and in small cities than political connections, or burden employers with rural areas.33 By contrast, in occupations with excessive costs when taking on new workers, such low skill requirements (such as informal non- as high employment taxes or severance require- tradables), worker skills are more homogenous ments. Fragmented or distorted markets make it and replaceable, reducing the need for careful hard for workers to identify, secure, and travel to matching of workers to jobs. jobs that make the best use of their skills. Chapter 4  Making cities more productive 97 Cities can in fact help accelerate social change Connected urban markets toward more equitable participation in labor mar- kets, by increasing people’s exposure to disadvan- Many developing country cities have become both taged groups in socially valued roles, and raising congested and spatially fragmented­ —­ undercutting the opportunity costs of the non-participation of the connectivity needed for sharing, matching, disadvantaged groups in labor market.35 In many and learning. High transport costs (in time and/ MENA countries, this mechanism is hampered by or money) and low-density, inefficient land use exclusionary employment restrictions, unequal fragment what should be a connected urban fab- remuneration, transport options that are poorly ric, impeding the flow of ideas, labor, and products tailored for women or minorities, and arduous reg- around a city. When urban scale is not matched by ulations and employment taxes that disincentivize connectivity, what should be large, diverse urban firms from taking on new workers. Such bottle- labor and product markets begin to instead frag- necks are undermining urban labor productivity ment and operate as multiple smaller, more ring- in many MENA cities (box 4.3), despite relatively fenced, markets, with reduced opportunities for high formal education levels. sharing, matching, and learning.36 BOX 4.3 HUMAN CAPITAL DEFICITS AND UNCONTESTABLE LABOR MARKETS UNDERMINE LABOR PRODUCTIVITY IN MENA In the MENA region, urban labor productivity is depressed by low human capital relative to income levels. MENA’s workforce suffers a deficit in marketable skills, which are not commensurate with the region’s income level or formal education credentials. Acquisition of core skills for each year of school- ing is comparably low in MENA, and students regularly underperform in tests of core skills like the Trends in International Mathematics and Science Study (TIMSS) and Progress in International Reading Literacy (PIRLS). This is reflected in low private sector demand for formal education credentials, and little to no correlation between education and private sector wages, private sector employment, and the likelihood of migration to a city (box figure 1). Indeed, unlike in other regions, where educated people tend to migrate to large urban areas where they can enjoy a substantial wage premium thanks to matching and learning agglomeration benefits, a MENA citizen with a university degree is no more likely to migrate to a city than a worker with no education at all. And while education in MENA is not transferring skills valued by the private sector, education credentials are important to secure public sector jobs, which generally offer higher pay, benefits, and job security. MENA’s labor productivity is also undermined by an “unfair” playing field in the labor market and by regulations that make it burdensome for firms to expand or renew the workforce.1 Formal firms bear a relatively high burden when taking on workers in MENA, with labor taxes constituting a large share of the overall tax burden, and high severance pay required by law, alongside limited public unemployment benefits. Most formal firms have expanded employment by only 1 percent annually in recent years, versus 5 percent in income peers in other regions. The costs of firing workers have made it challenging to take on younger workers with higher skills, contributing to the concentration of unemployment among youth (and the highest youth unemployment among world regions, at 26 percent). Restrictions on women working in specific industries, restricted working hours for women, unequal pay relative to men’s, and the need for women to obtain spousal permission to get certain jobs also leave an unequal playing field for women, whose labor force participation is the lowest among global (continued) 98 VI B R A N T C I T I E S BOX 4.3 continued HUMAN CAPITAL DEFICITS AND UNCONTESTABLE LABOR MARKETS UNDERMINE LABOR PRODUCTIVITY IN MENA regions, at just 20 percent. Cronyism also undermines labor market contestability, when firms are required to take on employees in return for political favors, given a high government presence in many markets. While beyond the scope of this report, Islam, Moosa, and Saliola (2022) also demonstrate how the presence of state-owned enterprises and favoritism toward politically connected firms further limit market contestability, to the detriment of labor and broader firm productivity in MENA. BOX FIGURE 1 Education credentials do not affect chances of private employment in MENA Effect of education on probability of working Effect of education on probability of working in the public sector outside the public sector Marginal effect of education Marginal effect of education 0.4 0.4 0.3 0.3 0.2 0.2 0.1 0.1 0.0 0.0 Primary Secondary Tertiary Primary Secondary Tertiary Educational attainment Educational attainment Source: World Bank 2020a. Note: 1. Islam, Moosa, and Saliola 2022. Weak connectivity separates workers and firms, distances), limiting workers’ access to jobs and impeding the worker–firm matching that would firms’ access to workers. In Uganda in 2010, 70 per- otherwise elevate labor productivity. In Cape Town, cent of urban commuters relied only on walking, the average resident can access only 6.5 percent of resulting in an average of just 19 percent of the city’s the city’s jobs in one hour, using the fastest existing jobs reachable within one hour.38 Nairobi, Kenya’s transport infrastructure and services. The figure intense congestion and weak transport services is under 25 percent for Dar es Salaam, Kigali, and mean that walking accounts for 41  percent of Harare, and under 35 percent in Nairobi, Conakry, commutes in the city.39 Informal transport options and Kampala.37 In many SSA cities, consuming two are also often inefficient and limit access to jobs: motorized transport trips per day would absorb Nairobi commuters relying on informal minibuses over a third of total household income for urban (matatu) can access only 4 percent of jobs in the city residents in the bottom income quintile (figure 4.3). within 30 minutes.40 Workers in Cairo, one of the Faced with costly or nonexistent transport services, world’s most congested and spatially fragmented many workers simply forgo transportation (or cities, also forgo job opportunities due to costly, limit commutes to shorter, for example, walkable, slow, or unsafe commutes (box 4.4). Chapter 4  Making cities more productive 99 FIGURE 4.3 TRANSPORT COSTS ARE OFTEN PROHIBITIVE FOR THE URBAN POOR Percent of bottom quintile household budget Spent on transportation Needed for two trips a day 100 80 60 40 20 0 Lagos Dar es Kigali Abidjan Kampala Ougadougou Nairobi Kinshasa Douala Addis Dakar Salaam Ababa Source: Kumar and Barrett 2008. BOX 4.4 CAIRO’S CONGESTED TRANSPORT NETWORK AND FRAGMENTED URBAN FORM—A BURDEN FOR COMMUTERS Cairo’s transport system puts the city in the top decile of congested cities worldwide. The share of Cairo’s land dedicated to roads has halved in the past 30 years, and major corridors now carry 3,000–7,000 vehicles per lane per hour, with average speeds falling to at least half (15–40 km/h) the normally expected speeds on major corridors (60–80 km/h). Speeds grind to as low as 6 km/h in rush hour on some local roads in central Cairo. Cairo’s metro was one of the most crowded in the world, with ridership rates (riders per km of metro per day) six times that of London, and a fifth higher than even Tokyo’s infamously crowded metro. Formal bus services per capita were much below benchmarks from comparable large, dense, emerging market cities, and there was no official information on bus routes and schedules. Cairo’s congested transport is exacerbated by a fragmented and sprawling urban form, which fur- ther raises transport and commuting costs, as well as poor tailoring of transport services to the needs of women and marginal groups—for safety and accessibility. There is evidence of high commuting costs and safety concerns forcing workers to turn down jobs, and even leave the labor market altogether. A 2016 survey found that 83 percent of women in Cairo did not feel safe in the street, 87 percent did not feel safe on public transport, 99 percent had experienced sexual harassment, and almost 70 per cent were dissuaded from using public trains to commute to work because of security concerns.1 For others, time and monetary costs of commuting are enough to deter employment: Barsoum2 reports the case of Ahmed Saleh, a 28-year-old male “who currently does not work. He left his position at a building firm after being transferred from an office job to a construction site. His entire monthly pay would have gone toward the extra transportation cost in getting to the new location.” Like many Cairo residents, he hopes to rely on personal connections to find employment that offers higher benefits and job security. Notes: 1. EBRD 2016. 2. Barsoum 2013. 100 VI B R A N T C I T I E S Women and marginalized groups may be particu- tightly in a few central locations where they can larly cut off from employment opportunities when interact with other high-skilled firms and more transport is not tailored to their needs. As many as efficiently reach skilled workers through trans- 65 percent of women in a selection of major MENA port networks across the city. cities, including Cairo and Amman, reported not • Manufacturing also relies on large pools of labor using existing public transport due to fears and/or due to strong internal scale economies, but has experiences of violence or harassment on public less need for skills, and higher land intensity transport. In Jordan, 40 percent of women were of production, creating some centripetal force found to have turned down job opportunities due to the urban periphery. Heavy equipment, as- to lack of viable transport to work.41 sembly lines, scale production, and so on drive up manufacturing’s land intensity versus other The commuting choices of lower-income and sectors. These dynamics create some cen- unskilled workers are particularly sensitive to tripetal force away from the most central and transport costs. Zaraté (2020, box 4.5) and Tsivan- dense urban areas (where land is most costly), idis (2020, see box 4.9) find that, in Mexico City and toward urban peripheries. Even so, manufac- Bogota, respectively, unskilled workers have the turing firms need access to a large and afford- highest sensitivity to commuting times and costs, able pool of labor in order to operate at scale, tending to work close to home rather than pay high as well as strong backward and forward linkages fares for employment across the city. This pattern with other firms (such as business services, re- reflects the low purchasing power of the poor and pairs, or parts), and high-quality, affordable the more limited benefits from matching econo- infrastructure and utilities. Access to a pool of mies for low-skill workers, who typically cannot cheap, low-skill labor can be undermined by find a substantial wage premium by commuting high transport costs; in these cases, firms must to work across town, compared with the wages in either compensate workers’ commutes through their local nontradable sector. By contrast, high- higher nominal wages (reducing labor produc- skilled and formal sector workers tend to endure tivity) or risk losing cheap lower-skilled workers longer and costlier commutes, and exhibit less to the more local nontradable sector. commuting elasticity to times or costs, because • Informal and nontradable services tend to their wage premia depend on their ability to work be scattered across the city, and rely little on in a few high-productivity locations, and/or spe- skilled labor. Nontradables exhibit limited in- cialized firms, that can use their skills (matching). ternal scale economies or specialization, and Although commuting costs fall directly on work- typically employ lower-skilled workers, result- ers, they also affect firms, which may need to offer ing in less need for access to a large or diverse higher wages to attract workers to overcome com- pool of labor. Nontradable firms gain their muting costs, or bear higher land costs to locate main competitive edge from physical proximity somewhere with better access to workers. to consumers, with benefits declining steeply with distance. For some services, this gradient The spatial fragmentation of urban labor markets­ —­ is on the order of meters rather than kilometers through inefficient land use and transport­—i­ s more (think, for example, of a neighborhood grocery harmful to tradable sectors than nontradables. store or café).42 For these reasons, nontradable • Formal tradable services tend to require high services tend to be spread more evenly through- human capital­ and hence to rely on connectiv- —­ out the city, roughly tracking local population ity to skilled workers, and benefit from learning or employment density. Lower-skilled workers agglomeration economies when skilled workers often turn to nontradables work closer to home interact and spread skills, knowledge, and ideas. when the costs to travel to alternative low-skill Tradable services thus tend to cluster more jobs across the city are high.43 Chapter 4  Making cities more productive 101 BOX 4.5 TRANSPORT INVESTMENTS HELP OVERCOME INFORMALITY IN MEXICO CITY Zaraté’s (2020) new urban model of Mexico City illustrates that investments to address urban con- gestion can be cost-effective counterparts to more traditional business environment investments, to raise formality and productivity. Informality is estimated to impose a high cost on productivity in Mexico. Busso et al. (2012) estimate that Mexico’s total factor productivity (TFP) could be 50 percent higher if informality were eradicated.1 Zaraté illustrates how urban transport improvements may cost-effectively support formalization, complementing more traditional approaches such as tax reform, surveillance, or digitization. He shows that workers and firms in Mexico City’s informal sector are more sensitive to transport costs than their formal counterparts.2 High congestion therefore keeps informal sector workers from accessing jobs in formal firms—which generally cluster in the city’s center and west, far from informal neighborhoods on the periphery and east. Zaraté models a new transit line linking remote eastern neighborhoods specialized in informal worker housing to central neighborhoods specialized in formal employment (box figure 1). By reducing the transport costs to which the informal sector is so sensitive, the transit line is expected to reduce worker informality in the city by 2–4 percentage points, and firms’ informality by 1–3 percentage points, with an overall welfare gain of around US$2 per dollar spent on the infrastructure, thanks to reduced distortions and higher tax revenues. Achieving the same reduction in informality through more tradi- tional business-environment measures was estimated to be far more costly, requiring costs of entry into the formal sector to be reduced by more than 25 percent. BOX FIGURE 1 Heatmap to show the change in commuter market access to formal sector jobs due to the construction of Transit Line B Line B Other lines Change in commuter market access to formal sector jobs Increased access Notes: 1. The informal sector’s non-payment of taxes gives it a competitive edge over tax-paying firms, leading it to become inflated in size, and the formal sector undersized, relative to socially optimal outcomes. 2. For example, informal workers spend 40 percent less time commuting on average, and are 10 percentage points more likely to work in the same municipality in which they reside. 102 VI B R A N T C I T I E S As well as fragmenting markets, inefficient trans- location choices of firms and households. These port and land markets directly raise the costs of second-order shifts and general equilibrium effects living and doing business in the city, which may can drive productivity benefits far exceeding those contribute to shutting cities out of tradables. Trans- from simple static cost savings, as illustrated by an port and land are direct components of the cost of array of urban modeling exercises.48 For this rea- living and producing in a city­ —­such as rents, trade, son, it is important to complement physical invest- and commuting costs. More spatially fragmented ments in connectivity with flexible and contestable cities have higher price levels: a 1 percent reduction land markets­ —­and a broader enabling business in spatial fragmentation (measured by the Puga10 environment­ —­that enable firms and households to index)44 is associated with a 12 percent reduction respond to the opportunities, as discussed below. in urban costs, controlling for income levels and city population.45 Nonmonetary transport costs Efficient land markets can also be large. In a meta-analysis covering most world regions (with some focus on LAC), Grover, Land is the defining scarce factor of production Lall, and Timmis (2021) estimate that commuters —­ in a city­ using it efficiently is important for an in developing country cities spend 23–43 percent efficient, competitive city. By fragmenting land, more time in congested traffic than commuters transport and real estate frictions prevent an in high-income cities. While urban costs of living efficient specialization of land use to drive urban directly affect households, firms may also have to productivity. Inefficient land and property markets pay higher (nominal) wages to attract workers or also prevent land being put to its most productive subsidize commuting costs in cities with high costs or welfare-enhancing uses, damping returns to and/or low amenity. Some previous studies that public investment. More enabling land markets found high agglomeration benefits in developing can allow livable, productive density to develop country cities in fact equated high urban costs in high-demand neighborhoods, and allow each (which drive up nominal wages) with high urban neighborhood to develop according to its changing productivity. Studies that instead control for urban comparative advantage. costs (for example, using real, rather than nominal, wages, or TFP) find much lower agglomeration The second-order response of land to transport benefits in developing country cities, and a much investments can drive higher productivity gains smaller gap between developing country and rich than the first-order travel time saving. When country city agglomeration elasticities (box 4.6).46 neighborhoods are poorly connected, people can High urban costs from inefficient land and trans- access jobs only by living in the neighborhood in port markets can leave cities unable to compete in which they work. By contrast, transport networks tradables, prices for which are set by global price allow households to separate spatially from firms. levels (box 4.7).47 The large productivity benefits of this mechanism allow neighborhoods to specialize according to Despite the importance of direct cost savings, the their comparative advantages (for example, in resi- largest benefits of transport and land reforms often dence or production, or production in a particular come from second-order effects­ —­ such as land real- sector), while also lowering the costs of land for location and investment, new commuting patterns, firms and households.49 and changes in productive activities. Improving the connectivity of urban neighborhoods and Efficient transport and land markets incentivize markets leads to shifts in the comparative advan- not only the specialization but also the densifica- tage of urban locations and firms­ precipitating —­ tion of land. By raising the productivity of neigh- shifts in the optimal productive activities, trade borhoods and of the city as a whole, transport and commuting activities, investment levels, and infrastructure can raise demand for land around Chapter 4  Making cities more productive 103 BOX 4.6 PAST STUDIES CONSTRUED PRICE LEVELS FOR PRODUCTIVITY AND AGGLOMERATION ECONOMIES AS HIGH IN DEVELOPING COUNTRIES In light of their focus on nontradables, and high urban congestion and costs, it is surprising that several studies found higher agglomeration benefits in developing country than developed country cities. Studies in the United States, France, and Spain showed a 0.3–0.5 percent increase in productivity for a 10 percent increase in density. In several developing countries and regions, this productivity increase (for 10 percent higher density) was found to be much higher: 1.9 percent in China, 1.2 percent in India, and 1.7 percent in Africa.1 The high agglomeration elasticity in Africa is surprising, considering that many African cities exhibit relatively low connectivity and skills, and focus on nontradables that can benefit less from agglomeration economies. A recent meta-analysis complements the picture, suggesting that earlier studies construed high urban costs for urban productivity, exaggerating agglomeration benefits.2 Studies finding higher agglomeration elasticities in developing countries tended to rely on nominal wages as a proxy for productivity. This is misleading, however, conflating high urban costs and price levels with true produc- tivity. When studies control for urban price levels, agglomeration benefits in developing country cities appear much smaller. For example, using TFP or real wages, instead of nominal wages, as the proxy for productivity, agglomeration elasticities are much lower overall (by 4.2 percentage points), and exhibit less significant difference between developed-country and developing country cities (0.1 percent for high-income countries and 1 percent for non-high-income countries). Notes: 1. Grover, Lall, and Maloney 2021. 2. Grover, Lall, and Timmis 2021. Source: Grover, Lall, and Timmis 2021. BOX 4.7 HIGH COSTS OF LIVING MAY HELP EXPLAIN UNCOMPETITIVE WAGES IN SAHARAN AFRICAN CITIES SUB-­ Many SSA cities exhibit high costs of living, high nominal wage costs, and low competitiveness in global tradables. Controlling for real national income and population sizes, the cost of living in cities in SSA is on average one-fourth higher than in other regions, with transport costs 42 percent higher, rents 55 percent higher, and housing services (like water and electricity) more than 100 percent higher (box figure 1).1 African cities also exhibit a higher urban wage premium (elevated urban versus rural wages).2 There is some evidence of firms paying higher nominal wages to overcome higher costs of living. Man- ufacturing firms in SSA cities pay higher nominal wages (15 percent more on average) than urban firms in other countries at similar stages of development.3 Manufacturing labor in Ethiopia, Kenya, Senegal, and Tanzania is much more costly than that in Bangladesh, despite comparable national income and World Economic Forum competitiveness ratings (proxies for labor and business environment quality, respectively), suggesting uncompetitive labor costs (box figure 2).4 Africa’s urban firms hire on average 20 percent fewer employees than comparable firms in other regions, tending to instead be more capital intensive.5 (continued) 104 VI B R A N T C I T I E S BOX 4.7 continued HIGH COSTS OF LIVING MAY HELP EXPLAIN UNCOMPETITIVE WAGES IN SAHARAN AFRICAN CITIES SUB-­ BOX FIGURE 1 Adjusted price level index of household consumption: SSA and other regions Adjusted price level index 200 Sub-Saharan Africa Other economies 150 100 50 6 8 10 12 Log of GDP per capita (2011 PPP$) Source: Nakamura et al. 2019. BOX TABLE 1 Comparing manufacturing firms’ labor costs in Bangladesh and four SSA countries WORLD ECONOMIC FORUM LABOR COST PER CAPITAL COST GDP PER COMPETITIVENESS WORKER ($) PER WORKER ($) CAPITA ($) RANKING Bangladesh 835.31 1,069.84 853.02 106 Kenya 2,118.01 9,775.45 1,116.69 96 Tanzania 1,776.65 5,740.99 1,094.95 116 Senegal 1,561.64 2,421.98 775.45 112 Ethiopia 909.28 6,137.98 471.19 109 Source: Gelb et al. 2017. Notes: 1. Nakamura et al. 2019; Nakamura et al. 2021. 2 Gollin, Lagakos, and Waugh 2014; Gollin, Kirchberger, and Lagakos 2021. 3. Jones 2016. 4. Gelb et al. 2017. 5. Iacovone, Ramachandran, and Schmidt 2014. the network and in the city overall. That contrib- In many cases, reorganizing and densifying land in utes to densification and scale (where it prompts response to transport boosts productivity by more in-migration, or labor market participation of than the direct cost savings from fares and travel existing residents), with attendant second-order times. Lall et al. (2021) elucidate this econometri- agglomeration benefits.50 cally, using a stylized city model to show that adding Chapter 4  Making cities more productive 105 a radial road network provides a 3 percent income than a simple travel-time cost saving calculation boost through direct cost savings, but a 4 percent (box 4.9). Heblich, Redding, and Sturm (2020) use boost through land reorganization and a further a quantitative urban model to show how the intro- 2 percent boost from in-migration. Bird and Ven- duction of steam railways in Greater London from ables (2019) also find large productivity benefits 1836 to 1921 enabled much wider commuting, and from making central land available for develop- an attendant spatial specialization of production in ment, and allowing unconstrained firm relocation, the City of London (central London) and residences in Dhaka (box 4.8). Tsivanidis (2020) presents a sim- in the wider commuting area (Greater London), ilar finding from an empirical analysis of the Trans- contributing to a rapid increase in central jobs and Milenio Bus Rapid Transit (BRT) in Bogota, finding productivity, and accounting for an approximately that gains are 74 percent higher when land realloca- 20 percent increase in population and land and tion and agglomeration benefits are accounted for building values of Greater London (figure 4.4). BOX 4.8 TACKLING CENTRAL LAND CONGESTION SUPPORTS A SHIFT TOWARD SPATIAL SPECIALIZATION AND TRADABLES PRODUCTIVITY IN DHAKA In Dhaka, a new urban model was applied to analyze the productivity impacts of making a huge tract of central urban land available for development. A key insight was that firms relocated around the city in pursuit of cheaper land and strong agglomeration benefits, to form more specialized hubs. This drove a large increase in tradables employment, and in city value added. Bird and Venables (2019) model the impacts of flood protections that make a large undeveloped area of Dhaka (“East Dhaka”) available for development. The area is over 100 square kilometers—7 percent of the total land area of Dhaka—and directly adjacent to the dense and overcrowded city center. But it is underdeveloped today due to its high risk of flooding (by contrast, West Dhaka is protected from flooding by a dyke). The authors model the prospective impacts of a dyke that protects East Dhaka from flooding, encouraging firms and development in this area.1 The dykes enhance productivity in Dhaka by supporting specialized tradable services and man- ufacturing hubs, and lowering congestion pressures on land. The investment(s) raise the supply of land in Dhaka by around 12 percent in areas well connected to the economic heart of the city. As the most land-intensive sector, manufacturing initially undergoes the largest shift, bringing 915,000 new manufacturing jobs to East Dhaka, generally from elsewhere in the city, making East Dhaka a new hub of manufacturing. This out-migration of manufacturing from other central areas helps address the intense congestion pressures elsewhere, lowering land and building rents in the central city, to the benefit of tradable services firms that flourish in these more central locations. Thus, by relieving central congestion, the dykes support the emergence of stronger, specialized tradable services hubs, increasing tradable services employment by 279,000 jobs, predominantly in central areas. The model estimated that the eastern dyke would raise city GVA (a local measure of GDP) by around 10 percent, per capita utility/income by around 5 percent, and the total population of Dhaka by around 6 percent. When combined with transport investments to better link East Dhaka to the central city, GVA rises even more—by 22 percent—and per capita income by around 10 percent. With these adjustments, nontradable services also benefit, primarily due to the increase in the city’s population. Note: 1. The authors assume this raises the amenity and productivity of land in East Dhaka to the (median) values seen in the central city. They posit that this may underestimate benefits, because East Dhaka benefits from large greenfield tracts of land for larger developments and proper servicing, while the central city includes a wide variety of neighborhood types, from the modern formal to highly informal, low-income, underserviced neighborhoods. 106 VI B R A N T C I T I E S BOX 4.9 BOGOTA’S BRT—ACCOUNTING FOR PRIVATE RESPONSES TO TRANSPORT INVESTMENTS, NOT ONLY TRAVEL TIME SAVED, AND ALLOWING LAND REDEVELOPMENT New urban models illustrate that transport investments drive productivity gains well beyond those implied by traditional, “travel time saved” metrics. Agglomeration economies, and the ability of firms and households to relocate around the city, have profound effects on the impacts of transport invest- ments. Transport benefits were traditionally assessed according to the direct economic “value of travel time saved” (VTTS) for commuters. New urban models show, by contrast, that the greatest benefits come from how firms and households respond to transport investments by relocating and developing land. Transport links make longer commutes less costly, allowing employment to concentrate more in central well-connected areas, and residents to disperse to more peripheral locations with cheaper land. This helps create or strengthen specialized production hubs in central areas, building their productivity through stronger agglomeration economies and improved access to workers. Lall et al. (2021) esti- BOX FIGURE 1 Change in residents’ mate that these second-order effects account for commuter market access with TransMilenio about two-thirds of the benefits of a transport investment in a stylized urban model, whereas the direct value of time saved accounts for only one-third. Tsivanidis (2022) uses a new urban model to estimate the impacts of Bogota’s TransMilenio BRT system on output and welfare in the city. The study estimates that the BRT raised welfare and output in Bogota by 2.2 percent and 3.0 percent respectively. These gains are significant nationally, considering that Bogota accounts for more than a quarter of Colombia’s GDP, more than half of its large and international firms, and 7.2 million residents. The analysis highlights that traditional VTTS analyses can substantially mischaracterize and understate the economic and distributional impacts of urban investments. VTTS analyses treat firm and household locations and com- mutes as fixed, then calculate the cost or time savings enabled by the transport as a proxy for its economic benefits. By contrast, urban models can simulate the second-order responses of firms, households, and developers to improved con- nectivity: this may include relocating within the city, changing their employment sector or loca- tion, switching commuting modalities, investing Note: Warmer colors (red, orange) indicate a larger in buildings, and so on. increase in commuter market access. Black lines show the TransMilenio routes as of 2013. Chapter 4  Making cities more productive 107 BOX 4.9 continued —A BOGOTA’S BRT­ ­ CCOUNTING FOR PRIVATE RESPONSES TO TRANSPORT INVESTMENTS, NOT ONLY TRAVEL TIME SAVED, AND ALLOWING LAND REDEVELOPMENT Accounting for just some of these second-order effects (allowing workers to change their residential location, workplace, and commuting modality), the welfare gains from TransMilenio are 74 percent larger than implied by a pure VTTS analysis; that is, VTTS accounts for only 57.5 percent of the welfare gain. The direction of benefits for low- versus high-skilled workers is also reversed. Although the BRT directly targets low-skilled commuters, high-skilled workers benefit more, because their skills are less easily substituted by other workers as the labor supply increases, and because high-skilled employment opportunities are more concentrated in locations connected by the BRT. The analysis also underscores the importance of anticipating and enabling a strong private sector response to infrastructure investments. The BRT raised the productivity of and demand for land close to the BRT, which would normally incentivize private developers to invest in building more vertical density, to optimize the use of scarce land. Unfortunately, strict zoning regulations in Bogota kept low building heights low along the line. Tsivanidis estimates that if these regulations were relaxed, allowing the government to auction off development rights for private vertical construction near the BRT (a form of land value capture), the BRT’s welfare gains would be 19 percent higher. This finding builds on those of the World Bank (2006), that about two-thirds of the average country’s nongovernmental capital stock is buildings,1 and urban construction represents a large and rising share of national investment (a key component of GDP). Note: 1. World Bank 2006. Source: Tsivanidis 2022. FIGURE 4.4 STEAM RAIL ENABLED THE SPATIAL SEPARATION OF WORK AND RESIDENCES IN GREATER LONDON a. Day and night population over time b. Overground and underground rail network across (City of London) Greater London, by 1921 Population (thousands) 400 Day population 300 200 100 Night population 0 Greater London London City Council 1800 1820 1840 1860 1880 1900 1920 City of London Administrative Overground rail boundaries Underground rail River Thames Source: Heblich, Redding, and Sturm 2020. 108 VI B R A N T C I T I E S Inappropriate land-use regulations can, however, and 70 percent of the population living in informal impede the private sector’s productive response to housing, in 2012, due to modernist regulations that changes in the comparative advantage and produc- priced out normal residents and speculative buy- tivity of land. Land is the binding scarce factor of ing.52 A 2018 study in Jordan estimated that zoning production in cities, so it is important to put it to its and building regulations made formal housing most productive use, guided by market signals. This unaffordable for all but the top 30 percent of the includes meeting market demand for floorspace income distribution, while more than 70 percent through taller structures in high-demand areas, of new housing was developed without a building and allowing repurposing land as the compara- permit.53 tive advantages of neighborhoods shift. Inflexible zoning, regulations, and land markets can greatly Private expectations depress the productivity gains attendant on public investments in cities, by preventing the private sec- The presence of agglomeration economies in cit- tor from responding appropriately to the opportu- ies in general, and in the urban tradables sector in nities created (see box 4.9), resulting in lower and particular, creates a first mover challenge for cities less efficient private investment, and sometimes a trying to break into tradables. The very agglomera- resort to informality to bypass codes. While build- tion economies that generate productivity benefits ing taller and more formally are partly a question from large cities or tradable sectors also create of finances, cities in LMICs often have inefficient large barriers to entry for new locations, due to land markets, unclear or insecure property rights, an incumbent advantage: established locations and excessively inflexible zoning regulations­ —­ all of offer stronger initial agglomeration benefits, driv- which raise costs and hinder efficient concentra- ing cycles of productivity, investment, and further tion and densification.51 agglomeration. New locations without an initial density of tradables face the reverse: a vicious cycle Vacant land, underoccupied formal buildings, of lower agglomeration benefits, contributing to and illegal construction in prime urban locations lower productivity, which in turn depresses inves- are common symptoms of dysfunctional land tor demand and density.54 This creates a first-mover markets and regulations. Many LMIC cities have problem, with no firm wanting to enter an undevel- considerable underdeveloped or vacant land and oped location or tradable sector alone, even if other buildings in well-connected central areas, while enabling conditions are in place. Further entrench- many firms and households are forced to locate ing low investment and scale are discontinuities in in disconnected peripheral areas or informal the impact of public investments and incentives, dwellings. The patterns often reflect building and and a strong incumbent advantage as new locations zoning regulations that are ill-attuned to local face a vicious cycle of low scale engendering low market signals and incomes (for example, forbid- agglomeration benefits and productivity.55 ding mixed uses and enforcing inefficient building parameters); dysfunctional or captured land and Expectations are also crucial in the building sec- property markets through which rights cannot be tor. Once built, formal buildings can be difficult secured; and in some cases large government land to modify, and may stay in place for more than holdings, developments, and speculative buying 150 years, with long-run impacts on the density (seen in the modernist new cities in the GCC and a neighborhood can host.56 Buildings are con- Egyptian desert (box 4.10). In Harare, Zimbabwe, structed at given densities based on expectations and Maputo, Mozambique, more than 30 percent of future demand for and productivity of a loca- of land within 5 kilometers of the central business tion. But the construction sector’s expectations in district remains unbuilt. In Cairo, 20–30 percent this regard can be partially self-fulfilling: building of formal housing units were found to be vacant, lower vertical density suppresses the possibility Chapter 4  Making cities more productive 109 BOX 4.10 MENA’S GHOST CITIES VERSUS TANGIER’S MORE PROMISING MODEL Both Egypt and Saudi Arabia constructed costly new cities in greenfield, often remote, desert loca- tions. In both countries, these new cities attracted populations much below targets, while absorbing large shares of public finance. In 2017, Saudi Arabia invested about 30 percent of its national budget to develop eight new planned cities, despite many planned cities remaining virtually unoccupied. In Egypt, the 23 new cities created from 1979 to 2000 attracted fewer than 800,000 people by 2012, versus the targeted 20 million. In 2015–16, Egypt invested 20–30 percent of the national built environment budget in these new cities and zones that hosted only 2 percent of the urban population—almost the same share (29 percent) as that allocated to all pre-existing cities and towns, which hosted 98 percent of the urban population. Tangier offers a more promising model—exploiting pre-existing advantages in market access, com- bined with a raft of well-thought-out complements. In 2002, Tangier was a lagging region in Morocco. However, the government of Morocco recognized that it could better leverage Tangier’s strategic location at a trade crossroads between the Middle East, Africa, and Europe. It invested €1 billion to develop and expand Tangier’s port and develop the surrounding city and region through transport infrastructure in roads and rail; industrial and logistics free-trade zones; training and education of the local workforce; and collaboration across government agencies and private sector stakeholders to spur international investment. The area was governed under an alternative legal framework that addressed core institutional and business environment challenges elsewhere in the country. The region enjoyed 28 percent employment growth in just two years (2002–04), and investment grew 13.2 percent annually. The local and regional economy became more diversified, and saw production increases in tradable manufacturing-intensive sectors such as the automotive industry. As a result, the region evolved from one of Morocco’s most lagging in 2000 to one of the most leading today. Source: World Bank 2020a. for future agglomeration economies long into efficiency, human capital, and private expectations the future, lowering demand for the location.57 At and market contestability and efficiency in a coor- these lower levels of demand and agglomeration, dinated and concerted effort. the lower-density building is in fact optimal. But building greater vertical density initially would These theories have gone awry, however, when have made stronger agglomeration benefits possi- governments have acted as property developers, ble later, resulting in higher demand. With higher rather than focusing on skills, connectivity, and demand and agglomeration, the higher-density institutions. When governments and state-owned building may be optimal. enterprises take the initiative to build high-density, expensive buildings in low-demand locations, the The presence of agglomeration economies in cit- result has often been costly “ghost cities” that absorb ies also means that progress along different policy huge shares of government finance for low eco- and investment dimensions is complementary, nomic returns and occupancy (see box 4.10). High and entails discontinuities in impacts rather than spending on “spatial bets” in MENA, for example, steady incremental progress. For this reason, large has had a large opportunity cost, including reduced developing country cities may need large and coor- fiscal space to invest in existing urban centers (which dinated advances along several fronts to overcome become increasingly congested, polluted, and infor- vicious cycles of low expectations and productivity, mal) and in broad-based fundamentals like educa- and break into tradables: raising land and transport tion, social protection, and institutions (figure 4.5). 110 VI B R A N T C I T I E S FIGURE 4.5 GOVERNMENT SPENDING BY SPATIAL CATEGORY IN SEVERAL MENA COUNTRIES DIFFERS SUBSTANTIALLY FROM THAT OF INTERNATIONAL COMPARATORS Percent of expenditure 100 80 60 40 20 0 Moroccoa Egypt, Jordanc Tunisiad Austria Spain South Hungary Poland Arab Rep.b Africa Institution-based interventions Place-based interventions 1. Broad-based governance and 5. Physical infrastructure for connectivity and to institutional reforms support local production People-based interventions 6. Subsidies and other incentives to capital 2. Basic health, education, and related human capital 7. Growth poles, industrial districts, other location improvements subsidies 3. Provision of basic public services 8. Public sector industrialization and industrial 4. Skill development, worker training, wage subsidies location regulations Source: World Bank 2020a. MODELING THE SPATIAL DYNAMICS complementarities (or the reverse) between poli- OF CITY PRODUCTIVITY cies, as well as second-order, general equilibrium effects. They not only allow a better assessment of To break the cycle of low expectations, low trad- the magnitude and distributional impacts of urban ables agglomeration, and low productivity, coor- policies, but can also help identify complementary dinated action is needed along the above reform measures that may be needed to achieve targeted dimensions. But most tools for policy or invest- desired benefits and manage second-order or unin- ment analysis look at one measure or dimension tended consequences (box 4.11). in isolation, struggling to account for interactions with other enabling (or undermining) reforms or Modeling illustrates how transport conditions. The presence of agglomeration econ- investments can boost productivity in Cairo omies in cities also entails sizable discontinuities in impacts, and the likelihood of second-order or The advances of new urban models are particularly general equilibrium effects not captured by static valuable for understanding how planned and on- analyses, which may be vastly different in both going investments to connect Greater Cairo may magnitude and direction. influence productivity and spatial development in the city. One of the defining urban policies in Cairo A generation of new urban models allow policy- in the past two decades has been the construction makers to analyze packages of investments and of new cities on the urban periphery and farther reforms. These spatial Computerized Gen- into the desert around Cairo, intended to become eral Equilibrium (SCGE) models account for centers of both employment and population. Some Chapter 4  Making cities more productive 111 BOX 4.11 NEW URBAN MODELS CAPTURE URBAN FORM AND PRODUCTIVITY The impact of urban reforms and investments is also highly sensitive to specifics of a city’s geography, economy, and policy environment, which are better captured in new urban models. Connecting a neighborhood through transport infrastructure can lead firms to either crowd into the neighborhood, or to vacate the neighborhood—depending on factors like transport speeds, costs, and productivity benefits in competing locations along the transport route.1 Skilled and unskilled workers, in tradable and nontradable sectors, also respond differently to transport improvements. The nature of transport infrastructure and services (speeds, affordability, reliability, gender sensitivity, quality of competing options), the prevalence of informality, the flexibility of real estate policies and markets, are among the many factors that affect how a city, and its composite neighborhoods, respond to urban policies aiming to boost productivity. A previous generation of urban models (Alonso-Muth-Mills models—AMM)2 offered insights into the second-order impacts of such investments as transport corridors and zoning, but with limitations for practical purposes. AMM assumed a monocentric urban form, with all firms located in the city center and all households on the periphery (commuting to the center). In these models, firms could not relocate in response to changing market forces, and firm productivity was fixed, so could not respond to changes in density, connectivity, or agglomeration.3 In reality, cities are continuously shaped and reshaped by the decentralized decisions of private firms, developers, and households, including decisions about where to live, build, or produce to maximize productivity or welfare.4 Further, the presence and nuances of agglomeration economies are central to many of these decisions, but were not explicitly considered in AMM models. A new generation of urban models, developed initially by Ahlfeldt et al. (2015), better captures the dynamics of urban spatial form and productivity. The model is a spatial computerized general equilib- rium model (SCGE), in which firms and households can chose where to locate (and relocate) in the city, according to the changing comparative advantages of each location. Developers can provide building floorspace in response to changes in the value of land (demand for locations), costs of construction, and land use regulations across the city. The new urban models allow firm location decisions and productivity to reflect agglomeration economies.5 That is, worker productivity is affected by the scale and density of employment in their work location, and these agglomeration impacts can vary by sector, worker type, for- mality, and so on. New urban models simulate multiple types of policies, both independently and simul- taneously, to explore their interdependencies and interactions, in areas such as transport infrastructure and services, land use regulation, firm incentives, human capital, urban amenities and services, and more. Notes: 1. Duranton and Venables 2018. 2. Alonso 1964, Mills 1972, Muth 1969. 3. Lall et al. 2021. 4. Bertaud 2018. 5. Firms can benefit from many forms of agglomeration economies: sharing goods or services that are indivisible or have high fixed costs (like infrastructure); better matching of workers to firms, thanks to the variety and density of the labor markets; and learning from the ideas and practices of proximate workers and firms, to improve innovation and efficiency. new cities are as far as 90 kilometers from central remained close to the city center, commuting out Cairo. And while new cities were intended to be to scattered employment locations.58 self-contained towns, hosting firms and their work- ers, Cairo has instead exhibited an unusual urban The new urban model developed for Cairo exam- pattern, with certain jobs dispersing to new cities, ines the interaction of congestion, land policies, while workers (households) have overwhelmingly and productivity in one of the world’s largest and 112 VI B R A N T C I T I E S FIGURE 4.6 densest cities. Greater Cairo’s population of more ONGOING TRANSIT INVESTMENTS IN than 20 million (in 2020) is on a par with Beijing, GREATER CAIRO Mumbai, or New York, and equal to more than a fifth of Egypt’s total population. Central Cairo also harbors some of the highest population densities in the world, reaching 30,000 per km2 in some neigh- borhoods, and averaging 12,000 per km2 across the wide area of Greater Cairo. Employment is even more concentrated, with almost half of Greater Cairo’s private sector jobs located in just 20 per- cent of land around the city center. Although this density could be an enabler for efficient mass transit systems, land-based financing, and tran- sit-oriented development, Cairo’s transit network has instead become among the most congested and underserviced globally, while central Cairo suffers high levels of informal, crowded, and even hazard- ous housing. Metro Monorail Cairo has also struggled to translate its density E-train into productive employment for its population. Sheyakha Cairo’s GDP is estimated at a quarter that of Beijing, despite its similar population size.59 Private employment has increasingly focused on informal nontradables, and formal jobs are concentrated in Source: Authors. the government sector. As government employ- ment opportunities have been cut in recent years, large numbers of residents have left the labor mar- investments, focusing on two questions: (1) Can ket altogether.60 these investments achieve the government’s aim of attracting populations and firms to new cities, Today, large transport investments are under way and protecting agricultural land in Greater Cairo’s to address Cairo’s central congestion, revive stag- north and south? (2) Can the investments improve nant new cities, and improve the performance of employment and productivity outcomes in Cairo, the urban economy. The government is investing by boosting tradables employment and overall out- to revive connectivity in the central city, including put and welfare? large expansion of the central metro (figure 4.6). In addition, whereas previously the government tried The model results suggest that the large ongoing to insulate new cities from the congested central transit investments may substantially boost output city, today, ambitious new investments to connect and welfare in Cairo, and shift employment toward new cities to the more historic urban fabric across tradables. Output may increase by almost 6 per- Greater Cairo are underway­ —­including transit lines cent, with the largest gains in tradable services. (a monorail and e-train) connecting the central city As a result of the combined transit investments, to more peripheral new cities in East and West. labor is expected to move away from nontradables (–2 percent) to manufacturing (+4 percent) and We use the new urban model developed for tradable services (+2 percent). Welfare is expected Cairo (box 4.12) to simulate the impacts of these to increase even more dramatically, by 16 percent. Chapter 4  Making cities more productive 113 BOX 4.12 the historic central city, deter expansion onto CALIBRATION OF THE NEW northern/­ southern agricultural land, and mobilize URBAN MODEL FOR CAIRO sunk-cost investments in new cities (table 4.1). It is most common for intra-urban transport to Data inputs: For each sheyakha (local gov- incentivize households to disperse to the urban ernment unit) in Greater Cairo, we used periphery (to consume more, cheaper land, now data on the: total population by sheyakha; better connected to the urban core), while firms employment by ISIC sector (QISM-level data was disaggregated to sheyakhas based take the more central locations thus vacated, to on built density); building volume (floor- enjoy the agglomeration benefits of interfirm link- space × heights); complete road network ages in dense urban cores. In Cairo, however, the with approximate speeds per type of road; result is somewhat reversed. Households indeed planned extension to the transit network disperse from the highest-density areas of historic (with speeds of 25 km/hour for the metro, Cairo to suburban areas served by the expanded and 45 km/hour for the e-train); average metro (figure 4.7). But firms do not crowd into the building regulation in each sheyakha (maxi- central land so vacated. Instead, they move even mum height and ground cover, per Decem- farther toward the urban periphery than house- ber 2021 regulations). holds: new cities in the East and West of Greater Cairo experience a more than doubling of jobs Data not acquired: Land values, incomes, (from 1.6 million at baseline to 3.7 million after and transport flows: these variables are estimated based on other input data, by inverting the model. FIGURE 4.7 Closure: Greater Cairo’s total population EAST, WEST, NORTH, SOUTH, AND and total employment are fixed. CENTRAL CAIRO Exogenous input parameters of interest: Agglomeration elasticities were assumed to be twice as high for tradables than non­tradables (1/3 vs 1/6). Building on Bird and Venables (2017), manufacturing was assumed to have a higher land intensity (14 percent land share in the production function, against 7  percent for tradable services, and 5.5 percent for nontradables). Source: For further details of the model inputs and calibration, see Zaraté (2022). (Total population and employment are fixed in the Center model.) In practice, such large gains in welfare are North likely to entice more working-age residents to rejoin East South the urban labor market, and may also attract higher West in-migration from areas beyond Greater Cairo. New cities The ongoing transit investments should sup- port the government’s spatial aims to decongest Source: Authors’ demarcation. 114 VI B R A N T C I T I E S TABLE 4.1 IMPACT OF TRANSIT INVESTMENTS IN GREATER CAIRO (NEW URBAN MODEL RESULTS) Change in population and jobs by sector, by region of Greater Cairo CHANGE IN JOBS (PERCENT OF BASELINE JOBS IN REGION) POPULATION TRADABLE NONTRADABLE REGION CHANGE MANUFACTURING SERVICES SERVICES TOTAL Center 1% –20% –9% –10% –12% North –7% –23% –18% –17% –19% South –5% –24% –20% –19% –20% West –1% –19% –17% –5% –9% New cities –1% 116% 150% 140% 130% Total 0% 4% 1% –2% 0% CHANGE IN JOBS (NUMBER) POPULATION TRADABLE NONTRADABLE REGION CHANGE MANUFACTURING SERVICES SERVICES TOTAL JOBS Center 184,998 (626,025) (237,579) (925,459) (789,063) North (117,651) (38,523) (17,468) (98,331) (154,323) South (51,610) (25,741) (10,178) (67,125) (103,044) West (4,700) (16,149) (6,984) (11,822) (34,955) New (11,037) 862,513 304,771 914,101 2,081,385 Total 0 156,074 32,561 (188,636) (0) Source: Authors. the investments), most of which are diverted from suggesting new cities were relatively unattractive Central Cairo (which sees the out-migration of locations for housing.61 1.8 million jobs, including 20 percent of its man- ufacturing, 9 percent of tradable services, and The specialization of peripheral new cities in 10 percent of nontradable jobs). Northern and employment, rather than housing (and resulting southern areas lose 20 percent of their jobs, mainly dispersal of jobs, rather than households, to these —­ to new cities­ supporting the government’s aim of areas), could reflect three important dynamics in protecting agricultural land in these locations, and Cairo. strengthening the agglomeration economies avail- able as firms shift to new cities. First, Cairo’s economy is focused on nontradables production, for which the gains from proximity to The dispersal of firms to the urban periphery is other firms are relatively low:62 consistent with some baseline characteristics of • 60 percent of private employment is in non- Greater Cairo­ and a specialization of neighbor- —­ tradables (nontradable services, plus agriculture hoods according to their comparative advantages. and mining).63 At baseline, new cities have higher labor productiv- • Only 17 percent of private employment is in ity and wages, but a lower share of the city’s jobs. tradable services­ —­the sector benefiting most This suggests substantial productivity gains from from external agglomeration economies. overcoming bottlenecks to expanding employ- • 23 percent of private employment is in man- ment in new cities. By contrast, amenity values ufacturing, where agglomeration forces are for households in new cities are relatively low, strong, but compete with manufacturing’s Chapter 4  Making cities more productive 115 FIGURE 4.8 greater need for cheap land, which creates CHANGE IN POPULATION DENSITY forces of dispersion toward the cheaper urban periphery. The result is that many of Cairo’s firms benefit more from reducing land costs (moving toward the urban periphery) than paying for the benefits from central proximity to other firms. Second, central Cairo is unusually dense and under- serviced, suffering major crowding disamenities. Central Cairo harbors some of the highest popu- lation densities in the world and has since the 11th century (box 4.13 and figure 4.8). Employment is even more centrally concentrated, with almost half Change in of Greater Cairo’s private sector jobs in just 20 per- residents per km2 +5,000 cent of land around the city center (figure 4.9). This density is often accommodated by crowding in tall and mid-height buildings, including considerable 0 –1,000 informal, hazardous construction and tenancy (such as Cairo’s characteristic “rooftop slums”). Cairo’s transport system has been ranked (by TomTom) in the top decile of congested cities worldwide.64 Cen- Source: Authors. tral congestion disamenities may exceed agglomera- tion benefits for the firms and households that leave by the government, while housing in new cities has central Cairo in response to transit connections. In been unaffordable for most Egyptians. The govern- addition, congested roads reduced willingness to ment has subsidized firms’ land and invested heav- commute to new cities in the baseline, despite the ily in productive infrastructure in new cities, which higher wages there; new transit connections (and may help explain their relatively high labor pro- the second-order effect of households relocating to ductivity (figure 4.10) and tradables specialization, suburbs closer to new cities) help make these com- despite their dearth of agglomeration benefits and mutes worthwhile for more workers. disconnection from the main city­ —­government investments acting similarly to place-based sub- Third, firms’ land and infrastructure in new cities sidies. By contrast, government involvement in have often been directly or indirectly subsidized housing development in the new cities has led to BOX 4.13 CAIRO HAS BEEN DENSE SINCE THE 11TH CENTURY “Fatimid Cairo was an urban marvel, and in time became the greatest city in the Islamic world. It was a triumph of architecture innovation and neck-craning verticality in which high-rise living became com- pletely normal almost 1,000 years before skyscrapers first sprouted in Manhattan”. This urban density included “seven- to fourteen- story communal houses.” Source: Marozzi 2019. 116 VI B R A N T C I T I E S FIGURE 4.9 FIGURE 4.10 CHANGE IN EMPLOYMENT DENSITY BASELINE SPATIAL VARIATION IN LABOR PRODUCTIVITY ACROSS GREATER CAIRO Change in jobs per km2 +2,000 0 1.798, 12.500 –6,000 0.004, 0.066 Source: Authors. a predominance of units that are high end and too Note: Darker areas denote higher labor productivity. expensive for most residents, with low uptake and Source: Authors. population density, lowering residential amenity in new cities (figure 4.11). A 2012 World Bank study especially as more migrants are attracted to the found that 20–30 percent of formal housing units city. in Cairo were vacant­ —­even as 70 percent of the city’s population occupies informal housing­ —d­ ue The government has been highly concerned to modern standards, high prices, and speculative with central density, particularly informal con- buying. struction of units that it sees as hazardous and unmanaged. 2020 saw on an outright ban on new Although the ongoing transit investments may be construction for six months, with the subsequent well calibrated to mobilize large sunk cost invest- introduction of highly restrictive building-height ments in new cities, there are important caveats. regulations (table 4.2).The model results suggest First, data are lacking on the costs of the transport that these December 2021 building regulations projects to translate results into a cost-benefit were non-binding, which reflects the fact that a lot analysis of the transit infrastructure. Second, hun- of new development is expected to take place in dreds of billions of dollars have been invested in new cities where the regulations do not apply­ and —­ Greater Cairo’s new cities in the last four decades that Greater Cairo’s total population is exogenously to bring these to their present levels of productiv- fixed in the model. ity, but complete data are lacking on these costs. Third, new building-height regulations could also Greater Cairo is likely to attract considerable result in greater encroachment of agricultural land, in-migration. Private developers need to be able Chapter 4  Making cities more productive 117 FIGURE 4.11 of vertical floorspace by a building sector that is BASELINE SPATIAL VARIATION IN well regulated for safety, but does not impose RESIDENTIAL AMENITY ACROSS modernist standards that render formal structures GREATER CAIRO unaffordable.65 Summary of insights from the new urban model for Cairo Attempts to develop new centers of agglomeration, and for cities to break into international tradables, suffer from a first mover problem.66 Developing new cities in the desert around Cairo has likely absorbed hundreds of billions of dollars of public investment over four decades. Despite these heavy investments, and although these new cities are rea- sonably close to one of the world’s largest and most overcrowded cities, the new cities have struggled to develop as viable agglomerations. Most housing 1.257, 8.610 in new cities was prohibitively expensive, while commutes to new cities were heavily congested and underserviced by public transport, resulting 0.000, 0.038 in most households and employers staying in the congested central city. Note: Darker areas denote higher amenity. The ongoing investments in connectivity across Source: Authors. Greater Cairo may be appropriate to overcome the first-mover challenge. They could unlock sunk TABLE 4.2 costs invested in new cities, and help develop them as viable agglomerations, including some special- MAXIMUM HEIGHTS UNDER NEW, 2021, ization in tradables. Holding Greater Cairo’s total BUILDING REGULATIONS IN CAIRO population and employment constant, the ongoing MAXIMUM MAXIMUM metro, monorail, and e-train investments could STREET WIDTH HEIGHT NUMBER OF attract 2 million jobs to the new cities’ locations, (METERS) (METERS) STORIES including 1.2 million tradable sector jobs, while 6–8 10 Ground +2 boosting Greater Cairo’s overall output by 6 per- 8–10 13 Ground +3 cent, concentrated in the tradable sector. 10–12 16 Ground +4 MAKING CITY GROWTH Source: https://invest-gate.me/news/new-building-regulations​ -are-issued/. RESILIENT AND INCLUSIVE to accommodate this influx on existing urban To translate population density into economic land, without crowding or spilling over on to the density, governments need to make strategic agricultural land that the government is trying to investments to support inclusive human capital protect. This means that the private sector must development and connectivity; ensure that all be allowed to build positive, formal density on swathes of the workforce are encompassed; inform existing urban land, allowing the construction firms and households with accurate and credible 118 VI B R A N T C I T I E S information; and ensure efficient, fair, and con- to equal pay, protection against sexual harassment, testable land and labor markets that can translate and job security after maternity leave). And they government investments into productive private should reduce the burden on firms of taking on action. new or young workers (for instance, reducing statutory severance pay requirements in favor of Upskill and strengthen urban labor markets government social protection, and shifting the tax burden from labor taxes). People in large cities need to be supported by strong, marketable skills. Skilled workers can Connect markets through transportation benefit most from the agglomeration economies of large cities, which help them find jobs to make A large, skilled labor force needs to be physically productive use of their unique and higher skills, connected, to benefit from and contribute to urban and to keep growing their human capital through agglomeration economies. Connecting fragmented urban knowledge spillovers. Underinvestment in markets through policies for efficient land use and marketable skills has left many urban residents intracity transport reduces costs of living and unable to benefit from and contribute to urban producing, improves the scale of agglomeration agglomeration economies, as they lack the skills economies in the city, and has second-order effects for high-quality employment in the tradable sector, through enabling greater specialization and densi- so as to reap the benefits of firm–worker match- fication of land. Especially in developing countries, ing and contribute to virtuous circles of urban cities will not overcome congestion by simply innovation and learning. Arguably the first urban building more roads. Not only is this very costly, policy for vibrant cities is thus in fact education­ but private vehicle use simply increases as more —­investing in an education system that delivers roads are built, keeping congestion high. Most globally competitive cognitive and interpersonal developing countries have more-transformative, skills to the workforce, while nurturing a diversity low-hanging fruit: build a foundation of effective of talents, technical skills, and specialties. traffic management, road and intersection design, parking regulation, and so on; ensure that existing To translate stronger human capital and connec- mass transport services are operated efficiently tivity into labor productivity, labor markets should and inclusively (for example, with enforced sched- be equal playing fields for all. This means over- uling, clear communication of routes and sched- coming physical, institutional, and social hurdles ules, interoperability and fare integration across that disconnect certain urban residents from job providers and modes, and accessibility for women opportunities. Cities can accelerate social change and other groups); favor public and non-motorized toward more equal participation in labor markets, transport (for example, concentrating investment increasing exposure to women or other disadvan- in bus lanes, cycle lanes, street lighting, and high- taged groups in socially valued roles, and raising quality sidewalks, while applying fees, taxes, and the opportunity costs of people’s non-participa- regulations on private vehicles, fuel, and parking). tion.67 But this process is hampered in many MENA More ambitious investments in public mass transit countries by exclusionary employment restrictions, should be calibrated to local density and incomes, unequal remuneration, poorly tailored transport to avoid bankruptcy and under-utilization. New options, and arduous regulations and employment urban models can help cities calibrate transport taxes that disincentivize firms from taking on new investments and reforms to meet goals for produc- workers. Labor regulations should ensure an equal tivity, inclusion, and financial sustainability. playing field for women, youth, and other disad- vantaged groups, abolishing any gendered restric- Heavy infrastructure must be strategically selected tions, providing protections needed (such as rights and prioritized to tap into a city’s latent potential Chapter 4  Making cities more productive 119 and meet its most pressing needs. Despite the to raise worker and firm productivity should be tai- productive benefits of connective infrastructure, lored to the needs of such subgroups. For example, examples abound of underused, expensive transit transport investments built with only able-bodied, investments that failed to deliver, while exerting high-income, urban men in mind may not connect a large drain on public resources­ —­as in bus rapid women, older or disabled workers, or unskilled transit systems in several African cities.68 This waste workers, to productive job opportunities. Less often follows from a mismatch between infrastruc- skilled workers are typically more sensitive to the ture and population/firm density, poor tailoring of price of transport, meaning that services must infrastructure to incomes and other needs, and a balance efficiency and affordability to integrate failure to first provide foundations of good trans- the urban poor. Women (especially in MENA) port management. Where large and costly trans- often report feeling unsafe, or may face cultural port investments are needed to relieve congestion barriers, when using public transport, walking, and overcome first mover problems, new urban cycling, or using motorbikes, and so require tai- models can complement other approaches to help lored solutions. Measures that have improved policymakers understand the risks of investments women’s transport access in developing country failing, appropriate tailoring to local density and cities around the world include a stronger security incomes, etc., and where such complementary presence on buses and trains, better street light- approaches may help turn unproductive invest- ing, women-only seats and carriages, and private ments, productive. “last mile” connections from home to bulk infra- structure (such as rickshaws). The right measures Broad-based investments risk excluding large for each city and subgroup should be informed swathes of the workforce, unless tailored to the by strong and regular stakeholder consultations needs of subgroups like women, people with dis- or surveys, feedback channels, and accountability abilities, and informal workers. Any broad support mechanisms (box 4.14). BOX 4.14 ADDRESSING GENDERED BARRIERS TO RAIL COMMUTING IN CAIRO Egyptian National Railways (ENR) studied, and subsequently reduced, gender imbalances in transport accessibility in Greater Cairo. After a 2013 UN Women survey found that 99 percent of Egyptian women had experienced sexual harassment, and 86 percent of women in Cairo did not feel safe or secure using public transport. ENR­ —­ with the European Bank for Reconstruction and Development (EBRD)­ —­assessed its services and operations from a gender perspective. Over 18 months, it surveyed 2,000 railway users in Cairo’s main railway station, conducted focus groups, and held validation workshops with ENR staff. The study confirmed that most women were deterred from using public transport because of safety and harass- ment concerns­ —­and identified key risk areas for harassment, such as platforms, ticket queues, and second- and third-class carriages. The summary report noted, “ENR has already started implementing some recommendations, such as providing a customer telephone hotline, installing surveillance cameras, improving lighting, and placing trained security personnel in the most crowded stations.” It also noted that, “participation of ENR staff throughout the process has unlocked internal dialogue within the company around possible gender-related improvements to transport services­ —­ something unheard of at the beginning of the EBRD’s involvement.” Source: EBRD 2016. 120 VI B R A N T C I T I E S Complement transport with land reform goods, leading to high urban costs and low produc- and servicing tivity.69 Public investments in transport often drive up nearby land and property values, handing gains For vibrant cities, land and real estate markets must for private landowners who have not contributed allow the private sector to make the most produc- to the investment. City governments should recap- tive use of land. Land rights must be transparent ture this publicly created value to finance further and secure, and land markets efficient and con- public spending, kickstarting a virtuous circle of testable, with prices accurately signaling scarcity, public investment, productivity, and tax revenue. demand, and productivity. Zoning and building Land value capture comes in many forms, includ- regulations should allow private investors to meet ing land or property taxes linked to market values, demand for floorspace (including market-appro- betterment fees, and auctioning building rights or priate vertical density, and efficient building foot- public land proximate to new infrastructure. New prints) in safe locations. And as the productivity urban models can help explore how alternative and comparative advantages of neighborhoods land value capture schemes can improve the pro- shift with public investments and rising skills, they ductivity and sustainability of investments. should allow private developers to repurpose land use­—­say, from residential to commercial. New Coordinate expectations urban models can help analyze where land use pol- icies may restrict productive private responses to The very agglomeration economies that generate public investment or where tighter land regulation productivity benefits from large cities or tradable may be needed to prevent spillovers into protected sectors also create a first mover problem, with no or vulnerable areas. firm wanting to enter an under-developed location or tradable sector alone, even if other enabling As effective transport reforms change demand conditions are in place. Further entrenching low for land, anticipate and provide for in-migration investment and scale are discontinuities in the to certain neighborhoods. Households or firms impact of public investments and incentives, and a relocating to a neighborhood due to improved strong incumbent advantage as new locations face a connectivity will also need adequate basic services vicious cycle of low scale engendering low agglom- like utilities, sanitation, healthcare, and policing, eration benefits and productivity.70 To attract trad- instead of informal, underserviced crowding. Some ables investment in new locations, it is necessary to neighborhoods may, however, see out-migration credibly set expectations that other complementary in response to new infrastructure­ —­ as residents firms (and workers) will also invest.71 leave a crowded city center to commute from higher-amenity suburbs. New urban models can Credible planning and communication can address help understand which neighborhoods may see the first mover problem to kickstart urban trad- in-migration, and which may see out-migration, ables. Well-designed transport infrastructure and stemming from transport investments, so that land use policies can help governments credibly complementary services can be properly coordi- set and coordinate expectations of future scale nated and avoid the appearance or expansion of and agglomeration, to overcome the first mover slums, or overprovision in declining areas. problem. Governments can provide transport infra- structure or services that better connect a targeted Cities should build positive fiscal feedback loops location to international markets (such as through through land value capture, to enable sustained port upgrading) or a large local labor market (such investments in connectivity and skills. Developing as through urban bus lanes, BRTs, or metros). country cities can become stuck in vicious cycle They can put in place flexible land use regulations of low tax revenues and low spending on public and efficient land markets to facilitate private Chapter 4  Making cities more productive 121 investment. These efforts are most powerful when directly. Government investments should focus on supported by broader reforms to address binding market failures and public goods, such as transport, constraints in the business environment and raise basic services and utilities (including education), the marketable skills of the labor force (see box 4.3). public spaces, and institutions, allowing the private New urban models can help stakeholder consulta- sector to supply buildings. tions and investor dialogue to credibly set private expectations and future productivity and density, New urban models can help governments select and analyze proposals submitted by stakeholders. optimal public investments to boost productivity and job creation. Selecting the right public invest- Investments in urban locations must be based on ments and locations can be challenging. Invest- accurate market expectations of future demand, ments for each neighborhood should be tailored rather than aspirational government plans, and to local comparative advantages and needs, and public investment should focus on public goods. accompanied by the right complements (such as Buildings sector activities have significant, long-run enabling land rights and a business environment). impacts on urban form and productivity, but costly New urban models are one tool for policymakers failed experiments around the world show that gov- to explore the appropriate combination of loca- ernments must focus on informing private develop- tion, investments, and complements to spur pro- ers’ expectations, rather than acting as a developer ductivity, tradables, and job creation. 122 VI B R A N T C I T I E S NOTES 26. Lall et al. 2017. 27. See, for example, Lall et al. (2017) on Sub-Saharan 1. UN Population Projections. Africa and World Bank (2020a) or Islam et al. (2022) 2. World Bank 2009; Annez and Buckley 2009. on Middle East and North Africa. 3. Gollin, Jedwab, and Vollrath 2016. 28. Bertaud 2018. 4. Solow 1956. 29. Grover, Lall, and Maloney. 2021. 5. Foster and Briceno-Garmendia 2010; Overman and 30. Lall, Henderson, and Venables 2017. Puga 2010. 31. Grover, Lall, and Maloney 2021. 6. Holmes (1999), for example, shows that in larger cit- 32. Duranton 2015; Moretti 2004. ies, firms are more likely to outsource to other firms, 33. La Roca and Puga 2017. consistent with higher specialization and input- 34. By contrast, cities offer relatively low returns to output linkages. motor skills and physical strength (Bacolod, Blum, 7. La Roca and Puga 2017; Duranton 2015; Charlot and and Strange Duranton 2004; Jaffe, Trajtenberg, and Henderson 2009; Andersson, Klaesson, and Larsson 2014). 1993. 35. Evans 2018; Evans forthcoming. 8. International trade in nontradables does in fact 36. Bertaud 2018; Grover, Lall, and Maloney 2020. take place in several settings with populations living 37. Peralta-Quiros, Kerzhner, and Avner 2019. either side of highly permeable land borders, such as 38. Bernard 2016. the border of Ireland and Northern Ireland, between 39. Lall, Henderson, and Venables 2017. Goma (Democratic Republic of Congo) and Rubavu 40. Avner and Lall 2016. (Rwanda), or many other cases in the European 41. Um and Effah 2020. Union, the East African Community, near poorly 42. Grover, Lall, and Maloney 2021a. policed rural borders, and so on. For simplicity, we 43. Zaraté 2020; Tsivanidis 2019. ignore these cases. 44. The Puga10 index reflects the number of people 9. Lall, Henderson, and Venables 2017. within 10 km of the average worker. 10. Grover, Lall, and Maloney 2021. 45. Lall, Henderson, and Venables 2017. 11. World Bank 2015. 46. Grover, Lall, and Timmis 2021. 12. World Bank 2020a. 47. Venables 2017. 13. World Bank 2020a. 48. Canning and Bennathan 1999; Tsivanidis 2020. 14. Islam, Moosa, and Saliola 2022. 49. Heblich, Redding, and Sturm 2020; Lall et al. 2021. 15. McMillan and Rodrik 2011; McMillan and Zeufack 50. Lall et al. 2021. 2022; Diao et al. 2021. 51. Lall, Lebrand, and Soppelsa 2021; see Fujita and 16. Jedwab, Ianchovichina, and Haslop, forthcoming. Ogawa (1982) and Heblich et al. (2018). Employment shares in tradable “FIRE” services have 52. Lall, Henderson, and Venables 2017. increased over time, but insufficiently to stem the 53. World Bank 2018b. overall transition toward nontradables. 54. Venables 2021. 17. World Bank 2018a. 55. Venables 2021. 18. Jedwab, Ianchovichina, and Haslop, forthcoming. 56. Hallegatte 2009. 19. Gollin, Jedwab, and Vollrath 2016; Ades and Glaeser 57. Venables 2021. 1995 58. Stewart 1996. 20. Gollin, Jedwab, and Vollrath 2016. 59. McKinsey 2011. 21. Venables 2017 60. After decades of government-led growth and job 22. Duarte and Restuccia 2010; Mano and Castillo 2015. creation, Egypt is attempting to transition to a more 23. Venables 2017. sustainable private sector–led model, but private 24. Gelb et al. 2017. firms have struggled to create the formal, high- 25. Gelb et al. 2017. quality jobs to which Egyptians are accustomed. Chapter 4  Making cities more productive 123 Government jobs offered full-time work, health Amiti, M., and L. Cameron. 2007. “Economic Geography insurance, pensions, predictability and stability, and and Wages.” The Review of Economics and Statistics 89 (1): paid parental and sick leave. By contrast, in recent 15–29. years, most employment growth was in nontradable Andersson, M., J. Klaesson, and J. P. Larsson. 2014. “The services, including private construction, transport, Sources of the Urban Wage Premium by Worker and wholesale and retail sectors. Nontradable sector Skills: Spatial Sorting or Agglomeration Economies?” employment in Egypt is typically highly informal, Papers in Regional Science 93 (4): 727–47. with 89 percent of construction and 95 percent of Spence, M., P. Annez, and R. Buckley. Urbanization and transport jobs “outside establishments” in 2018 (up Growth. Commission on Growth and Development. from 85 percent and 86 percent respectively in 2004). Washington, DC: World Bank. For more on Egypt’s stalled transformation and jobs Avner, P., and S. V. Lall. 2016. “Matchmaking in Nairobi: challenge, see World Bank (2020b). The Role of Land Use.” Policy Research Working 61. Amenity captures the attractiveness of a neighbor- Paper 7904, World Bank, Washington, DC hood not explained by its income opportunities, cap- Bacolod, M., B. S. Blum, and W. C. Strange. 2009. “Skills turing features like the quality of local public spaces in the City.” Journal of Urban Economics 65 (2) :136–53 and (non-transport) services, and affordability of Barsoum, G. 2013. “No Jobs or Bad Jobs.” The Cairo Review land and property. of Global Affairs. https://www.thecairoreview.com/ 62. In our urban model, nontradable agglomeration ben- essays/no-jobs-and-bad-jobs/. efits are assumed to be half those of manufacturing Bernard, L. 2016. “Job Access in Kampala.” University of or tradable services. Density, for the purpose of cal- Oxford, Oxford, UK. culating agglomeration benefits, is measured as the Bertaud, A. 2018. Order Without Design. Cambridge, MA: number of jobs in the sheyakha. MIT Press. 63. Employment in public administration is not included Bird, J., and A. Venables. 2019. “Growing a Developing in our data; if captured, this nontradables share City: A Computable Spatial General Equilibrium would be even higher. Model Applied to Dhaka.” Policy Research Working 64. https://www.worldbank.org/content/dam/World- Paper 8762, World Bank, Washington, DC bank/TWB-Executive-Note-Eng.pdf. Busso, M., M. V. Fazio, and S. Levy Algazi. 2012. “(In)For- 65. Lall et al. 2021. mal and (Un)Productive: The Productivity Costs of 66. Duranton and Venables 2018. Excessive Informality in Mexico.” IDB Working Paper 67. See Evans (2018) and Evans (forthcoming). Series No. IDB-WP-341, Inter-American Development 68. Lall, Henderson, and Venables 2017. Bank, Washington, DC. 69. Venables 2017. Canning, D., and E. Bennathan. 1999. “The Social Rate 70. Duranton and Venables 2018; Venables 2021. of Return on Infrastructure Investments.” Policy 71. Venables 2017. Research Paper, World Bank, Washington, DC. Charlot, S., and G. Duranton. 2004. “Communication REFERENCES Externalities in Cities.” Journal of Urban Economics 56 (3): 581–613. Ades, A. F., and E. L. Glaeser. 1995. “Trade and Circuses: Diao, X., M. Ellis, M. McMillan, and D. Rodrik. 2021. “Afri- Explaining Urban Giants.” Quarterly Journal of Econom- ca’s Manufacturing Puzzle: Evidence from Tanzanian ics 110: 195–227. and Ethiopian Firms.” Working Paper 28344, National Ahlfeldt, G. M., S. J. Redding, D. M. Sturm, and N. Wolf. Bureau of Economic Research, Cambridge, MA. 2015. “The Economics of Density: Evidence from the Duarte, M., and D. Restuccia 2010. “The Role of Structural Berlin Wall.” Econometrica 83: 2127–2189. Transformation in Aggregate Productivity.” The Quar- Alonso, W. 1964. Location and Land Use: Toward a General terly Journal of Economics 125 (1): 129–173. Theory of Land Rent. Cambridge, MA: Harvard Univer- Duranton, G. 2015. “Growing through Cities in Develop- sity Press. ing Countries.” The World Bank Research Observer 30 (1). 124 VI B R A N T C I T I E S Duranton, G., and D. Puga. 2004. “Micro-Foundations Grover, A., S. Lall, and J. Timmis. 2021. “Agglom- of Urban Agglomeration Economies.” Chapter 48 eration Economies in Developing Countries: A in Handbook of Regional and Urban Economics, Vol. 4, Meta-Analysis.” Policy Research Working Paper 9730, edited by V. Henderson and J.-F. Thisse, 2063–117. World Bank, Washington, DC. Amsterdam: North-Holland. Hallegatte, S. 2009. “Strategies to Adapt to an Uncertain Duranton, G., and A. Venables. 2018. “Place-based Policies Climate Change.” Global Environmental Change 19 (2): for Development.” Working Paper 24562, National 240–247. Bureau of Economic Research, Cambridge, MA. Heblich, S., S. J. Redding, and D. M. Sturm. 2020. “The Mak- EBRD (European Bank for Reconstruction and Develop- ing of the Modern Metropolis: Evidence from London.” ment). 2016. “Safe Transport for All: Issues and Oper- The Quarterly Journal of Economics 135(4): 2059–2133. ational Lessons from the Egyptian National Railways.” Holmes, T. J. 1999. “Localization of industry and vertical London: EBRD. disintegration.” Review of Economics and Statistics 81 (2), Ellison, G., E. Glaeser, and W. Kerr. 2010. “What Causes 314–332. Industry Agglomeration? Evidence from Coagglom- Iacovone, L., V. Ramachandran, and M. Schmidt. 2014. eration Patterns.” American Economic Review 100 (3): “Stunted Growth: Why Don’t African Firms Create 1195–1213. More Jobs?” Policy Research Working Paper 6727, Evans, A. 2018. “Cities as Catalysts of Social Change.” World Bank, Washington, DC. Annals of the American Association of Geographers 108 Islam, A., D. Moosa, and F. Saliola. 2022. Jobs Undone: (4): 1096–114. Reshaping the Role of Governments towards Markets and Evans, A. Forthcoming. “How Cities Erode Gender Workers in the Middle East and North Africa. Washington, Inequality: A New Theory and Evidence from Cam- DC: World Bank. bodia.” Gender and Society. Jaffe, A., M. Trajtenberg, and R. Henderson. 1993. “Geo- Foster, V., and C. Briceno-Garmendia. 2010. Africa’s Infra- graphic Localization of Knowledge Spillovers as structure: A Time for Transformation. Washington, DC: Evidenced by Patent Citations.” Quarterly Journal of World Bank. Economics 108 (3). Fujita, M., and H. Ogawa. 1982. “Multiple Equilibria and Jedwab, R., E. Ianchovichina, and F. Haslop. Forthcom- Structural Transition of Non-monocentric Urban ing. “Consumption Cities: Novel Considerations and Configurations.” Regional Science and Urban Economics Evidence.” 12 (2): 161–196. Jones, P. 2016. “Closed for Business: High Urban Costs.” Gelb, A., C. Meyer, V. Ramachandran, and D. Wadhwa. 2017. Department of Economics, Oxford University, “Can Africa Be a Manufacturing Destination? Labor Oxford, UK. Costs in Comparative Perspective.” Working Paper 466, Kumar, A., and F. Barrett. 2008. “Stuck in Traffic: Urban Center for Global Development, Washington, DC. Transport in Africa.” AICD Background Paper 1, Gollin, D., R. Jedwab, and D. Vollrath. 2016. “Urbaniza- Africa Infrastructure Country Diagnostic. tion with and without Industrialization.” Journal of Lall, S., M. Lebrand, H. Park, D. Sturm, and A. Venables. Economic Growth 21 (1): 35–70. 2021. Pancakes to Pyramids: City Form to Promote Sustain- Gollin, D., M. Kirchberger, and D. Lagakos. 2021. “Do able Growth. Washington, DC: World Bank. Urban Wage Premia Reflect Lower Amenities? Evi- Lall, S., J. Henderson, and A. Venables. 2017. Africa’s Cit- dence from Africa.” Journal of Urban Economics 121: ies: Opening Doors to the World. Washington, DC: World 103301. Bank. Gollin, D., D. Lagakos, and M. Waugh. 2014. “The Agri- Lall, S., M. Lebrand, and M. Soppelsa. 2021. “The Evo- cultural Productivity Gap.” The Quarterly Journal of lution of City Form: Evidence from Satellite Data.” Economics 129 (2): 939–993. Policy Research Working Paper 9618, World Bank, Grover, A., S. Lall, and W. Maloney. 2021. Place, Productiv- Washington, DC. ity, and Prosperity: Revisiting Spatially-targeted Policies for La Roca, J., and D. Puga. 2017. “Learning by Working in Regional Development. Washington, DC: World Bank. Big Cities.” Review of Economic Studies 84: 106–142. Chapter 4  Making cities more productive 125 Mano, R., and M. Castillo. 2015. “The Level of Productivity Pritchett, L. 2013. “The Rebirth of Education: Schooling in Traded and Non-Traded Sectors for a Large Panel Ain’t Learning.” Washington, DC: Center for Global of Countries.” Working Paper WP/15/48, International Development. Monetary Fund, Washington, DC. Solow, R. 1956. “A Contribution to the Theory of Eco- Marozzi, J. 2019. Islamic Empires: Fifteen Cities that Define a nomic Growth.” Quarterly Journal of Economics 70 (1). Civilization. Penguin. Stewart, D. 1996. Greater Cairo: Mother of the World. Cairo: Mckinsey. 2011. “Urban World: Mapping the Economic American University in Cairo Press. Power of Cities.” McKinsey & Company. Tsivanidis, N. 2019. “Evaluating the Impact of Urban McMillan, M., and D. Rodrik 2011. “Globalization, Struc- Transit Infrastructure: Evidence from Bogotá’s Trans- tural Change, and Productivity Growth.” Working Milenio.” University of California, Berkeley. Paper 17143, National Bureau of Economic Research, Um, P., and D. Effah. 2020. “If Urban Transport in MENA Cambridge, MA. Was Safe for Women, Economies Would Prosper.” McMillan, M., and A. Zeufack 2022. “Labor Productivity Arab Voices blog, World Bank, October 29. https://blogs​ Growth and Industrialization in Africa.” Journal of Eco- .worldbank.org/arabvoices/mena-if-urban-transport​ nomic Perspectives 36 (1): 3–32. -was-safe-women-economies-would-prosper. Mills, E. S. 1967. “An aggregative model of resource allo- Venables, A. 2017. “Breaking into Tradables: Urban Form cation in a metropolitan area.” American Economic and Urban Function in a Developing City.” Journal of Review Papers and Proceedings 57 (2): 197–210. Urban Economics 98: 88–97. Mills, E. S. 1969. “The value of urban land.” In Perloff, H. S. Venables, A. 2021. “Winners and Losers in the Urban (ed.), The Quality of the Urban Environment. Baltimore, System.” In Urban Empire: Cities as Global Rulers in the MD: Resources for the Future. New World. Routledge. Mills, E. S. 1972. Studies in the Structure of the Urban Econ- World Bank. 2006. Where Is the Wealth of Nations? Measur- omy. Baltimore, MD: Johns Hopkins University Press. ing Capital for the 21st Century. Washington DC: World Moretti, E. 2004. “Human Capital Externalities in Cities.” Bank. In Handbook of Regional and Urban Economics, Vol. 4, World Bank. 2009. World Development Report 2009: Reshap- edited by V. Henderson and J.-F. Thisse. Amsterdam: ing Economic Geography. Washington, DC: World Bank. North-Holland. World Bank. 2014. “Cairo Traffic Congestion Study.” Muth, R. F. 1969. Cities and Housing. Chicago: University Washington, DC: World Bank. of Chicago Press. World Bank. 2015. Competitive Cities for Jobs and Growth. Nakamura, S., R. Harati, S. Lall, Y. Dikhanov, N. Hama- Washington, DC: World Bank. deh, W. Oliver, M. Rissanen, and M. Yamanaka. 2019. World Bank. 2018. Raising the Bar for Productive Cities in “Is Living in African Cities Expensive?” Applied Eco- Latin America. Washington, DC: World Bank. nomics Letters 26 (12): 1007–1012. World Bank. 2020a. Convergence: Five Critical Steps toward Nakamura, S., R. Harati, S. Lall, Y. Dikhanov, N. Hama- Integrating Lagging and Leading Areas in the Middle East deh, W. Oliver, M. Rissanen, and M. Yamanaka. 2021. and North Africa. Washington, DC: World Bank. “Comparing Costs of Living Across World Cities.” The World Bank. 2020b. “Egypt Economic Monitor.” Washing- World Bank Economic Review 34 (Supplement): S79–S88. ton, DC: World Bank. Overman, H., and D. Puga. 2010. Labor Pooling as a Source Zaraté, R. 2020. “Spatial Misallocation, Informality, and of Agglomeration: An Empirical Investigation. Chicago, Transit Improvements: Evidence from Mexico City.” IL: University of Chicago Press. Policy Research Working Paper 9990, World Bank, Peralta-Quiros, T., T. Kerzhner, and P. Avner. 2019. Washington, DC. “Exploring Accessibility to Employment Opportuni- ties in African Cities.” Washington, DC: World Bank. GOVERNANCE governing efficiently. resilient. sustainable. inclusive. productive. 5 Taking the next steps for tomorrow’s vibrant cities— Governance reforms to inform, support, and protect Many of this report’s policy priorities are far from people but lacking in productive economic density. new­ —­enabling formal land and labor markets, con- For credible governance, cities need to grow their necting neighborhoods efficiently, providing basic fiscal capacity and establish accountable social services more equitably, and so on. Some of these contracts. policies are technically hard or costly, but others are less so. Yet too many developing country cities fail FISCAL CAPACITY to make even basic reforms and investments. Why? Most investments contemplated in this report The traditional answer emphasizes vested inter- are predicated on municipal fiscal capacity. Such ests that would lose from reforms and thus orga- capacity comes to the fore when comparing the nize themselves to successfully oppose them. But recent economic experience of China and India. recent work on the political economy of reforms Case studies, reports, and analyses of municipal has moved beyond vested interests to consider budgets show that Chinese cities have outper- how public servants’ and citizens’ expectations formed Indian ones by raising own-source revenue for each other’s behavior (codified as “legitimacy” and long-term financing to invest in urban infra- and “trust”) (box 5.1 below) shape social contracts structure, which in turn attracted private firms and and lead to low- or high-performance equilibria. economic activity to cities.1 Related to the social contracts, cities also lack the fiscal capacity to fund their investment and reform Financing urban infrastructure and safety-net programs. programs is especially hard for cities in developing countries. A distinguishing feature of urban policy- Effective urban governance institutions are vital to making is the sheer scale of the financing for urban a vibrant future. Cities that continue to grow with- infrastructure, adding long-term operating costs to out credible governance­ —­in basics such as land initial capital outlays. Similarly, expanding safety rights and fiscal management­ will remain ineffi- —­ nets to tackle climate and other shocks requires ciently and insufficiently developed, crowded with dedicated funding. 127 128 VI B R A N T C I T I E S For cities to be vibrant, they need large-scale and Financing for urban public services is both inade- long-term financing to build the infrastructure for quate and highly centralized in the MENA region sustainable growth and job creation. The financing relative to OECD countries. Central government gap for urban infrastructure globally is estimated entities are primarily responsible for public invest- to be more than US$4 trillion per year, much of it ments in urban infrastructure and service delivery. in developing countries, and official development Even where city/local governments undertake pub- assistance is only a tiny fraction of this financing. lic investment in infrastructure and service provi- Ultimately, filling this gap will primarily rely on the sion, they largely rely on fiscal transfers from higher ability of national, subnational, local, or city gov- levels of government to do so­ —­as seen in very low ernments to access commercial financing, which shares of own-source revenue in their total revenue is possible only if they can provide credible returns streams relative to OECD countries (figure 5.1). Cen- on capital investments by improving their fiscal tral transfers are the main source of local finance, base and creditworthiness. and often lack predictability, transparency, and equity, and fail to provide adequate performance The clear and established pathway to improve credit- incentives for cities. The costs to perform assigned worthiness and have a stronger fiscal base to support municipal functions often far exceed the proceeds higher commercial financing is the ability of these of local revenues, meaning that municipalities run governments to raise higher volumes of stable and perennial operational deficits. With high centraliza- recurrent revenue streams through taxation of grow- tion, cities have little flexibility in adjusting rates and ing incomes and property values. But city govern- mechanisms for stronger revenue generation.2 ments in the developing world lack fiscal capacity. A wealth of empirical evidence in the literature on For example, there has been a persistent decline state capacity shows that the ability to consistently in subnational revenue and local tax collection as raise revenue (that is, having strong fiscal capacity) a share of total national revenue in Egypt over the is a fundamental driver of building state capacity. past three decades (figure 5.2). Even in Morocco­ —­ City governments can raise stable local revenue only one of the better-performing MENA countries­ —­it if they are well-functioning, credible organizations is estimated that cities need five times more capital that citizens and firms can interact with. investment than current levels­ —­ or 2.5 times more FIGURE 5.1 DECENTRALIZED SPENDING AND OWN-SOURCE REVENUES ARE LOW IN MENA Subnational government spending as a share of total government spending (percent) 100 MENA OECD countries 80 60 40 20 Morocco Egypt West Bank and Gaza Iraq Jordan Tunisia 0 Djibouti 0 20 40 60 80 100 Revenue from noncentral sources as a share of total subnational government revenue (percent) Source: World Bank 2020a. Chapter 5  Taking the next steps for tomorrow’s vibrant cities 129 FIGURE 5.2 EGYPT HAS SEEN A PERSISTENT DECLINE IN SUBNATIONAL REVENUE AND LOCAL TAXES AS A PERCENTAGE OF NATIONAL REVENUE Percent 5 4 3 Total local revenue 2 1 Local taxes 0 1992 1996 2000 2004 2008 2012 2016 2020 Source: Egypt Ministry of Finance data collected by M. Nada et al. (World Bank Egypt country team) as part of technical assistance on fiscal decentralization and own-source revenue strategy. own-source revenue­ to meet their estimated —­ vicious circles of low performance and low legit- financing needs for urban infrastructure investment imacy and trust, which frustrate further reform over the medium term due to rising urbanization.3 efforts. For city leaders to be able to implement reforms, they need a foundation of legitimacy and Municipalities’ own fiscal capacity will remain trust (box 5.1). Without legitimacy, compliance limited as long as social contracts between city with regulations is low and needs more costly­ —­ governments and citizens are weak. Three main often infeasible­ enforcement, rendering reforms —­ ways to secure adequate own-source financing ineffectual. Without trust, citizens are more likely for the provision of infrastructure and services to pursue actions that bring selfish benefits at a are taxing the value of land and property, forming cost for society as a whole­ —­ such as extracting private–public partnerships, and raising funds from bribes, shirking on their public duties, dumping the capital markets.4 The need for such reforms is waste illegally, encroaching on land, and so on. not unknown to city leaders. The main obstacle lies in governance­ specifically, the inability of author- —­ Trust in governments and the civil service appears ities to implement required measures due to a lack low across MENA countries. Fewer than half of all of legitimacy and trust. MENA citizens report “trust in government.” An even lower share report “trust in the civil service,” CITY GOVERNANCE which generally includes municipal and local gov- ernment agents (figure 5.3). Reforms to incentives, Only with functional urban governance institutions to intrinsic motivations, and to professional norms can developing country cities position themselves for are desperately needed to recover the credibil- a vibrant future. In addition to raising revenue, devel- ity of city governments now seen as corrupt and oping country cities need to provide high-quality inefficient. services. These tasks require city governments to function in a professionally credible manner. For a city government to build legitimacy and trust, it must make public officials accountable The lack of market-enabling reforms generally and accessible to households and firms, through reflects, not ignorance of their necessity, but mechanisms such as public fora and local elections. 130 VI B R A N T C I T I E S BOX 5.1 LEGITIMACY AND TRUST Legitimacy is the ability of leaders to win compliance with new laws or public orders because people share a widespread belief that everyone is complying. Recent work on law and economics re-examines the puzzle of why developing countries have laws on paper that are not effectively implemented.1 Instead of relying on explanations about weak governance, low capacity, and perverse political incen- tives, a conceptually clearer way of thinking about compliance with a new law is whether the new law changes beliefs about how others are behaving, and thus legitimizes compliance.2 Legitimacy­—­as defined above­ —­depends on beliefs about how others are behaving in political and bureaucratic institu- tions. For example, whether households comply with tariff increases of urban utilities, or urban planning regulations, depends on the legitimacy of those policies, viewed as beliefs about how others are likely to behave. This view provides new ideas for using transparency and local political contestation to shift beliefs toward greater legitimacy. Trust consists in beliefs that others are behaving cooperatively in a particular game of life, society, or politics­—­ motivating similar cooperative behavior from the subject in question.3 High levels of trust are equated with beliefs that others are playing the game in cooperative ways that enhance the payoffs to all, while low levels of trust are equated with beliefs that others are playing the game in noncooperative ways, with worse outcomes for all players. Within the large body of research on different types of trust, the segment most relevant for understanding urban governance issues is about trust, or lack thereof, in public institutions. Corruption or rent-seeking is one manifestation of a lack of trust. If people believe that others are likely to be extracting rents in the public sector, they are likely to behave in the same way, yielding an equilibrium of high corruption and low trust. Trust can also be examined within urban government departments as the strength (or weakness) of professional norms among the multiple principals and agents involved in the sector. If agents believe that others are likely to behave unprofessionally, by holding up decisions, or not performing their assigned tasks on time or effectively, then they are likely to behave the same way, yielding an equilibrium of weak professional norms and low trust. This view of trust provides new ideas for improving the performance of complex organizations of urban government through management reforms that use greater autonomy along with communica- tion to the public and among peers to build peer-to-peer monitoring and accountability, or professional norms. Notes: 1. World Bank 2017. 2. Basu 2018. 3. Alesina and Giuliano 2015; Algan and Cahuc 2014. Source: Athar and Khemani 2022. Local political contestation and (two-way) commu- reform, and motivate a new cadre of public health nication with citizens are key to building legitimacy workers, leading to much improved health out- and trust. In MENA cities, local elections with high comes (box 5.2). participation were found to strengthen political leaders’ motivation to enact policies for the pub- Accountability, performance incentives, and city lic good.5 In Ceara, Brazil, municipal leaders used governance outcomes vary considerably across communication campaigns, and the account- the MENA region. MENA cities can be ranked on ability mechanism of local elections, to cement accountability to local citizens (figure 5.4). At one and reward political commitments to healthcare end of the spectrum is Cairo: controlled by officials Chapter 5  Taking the next steps for tomorrow’s vibrant cities 131 FIGURE 5.3 TRUST IN GOVERNMENT AND THE CIVIL SERVICE IN MENA COUNTRIES Trust in government Trust in the civil service Percent 60 Qatar Kuwait Jordan 40 Morocco Egypt West Bank & Gaza 20 Algeria Lebanon Tunisia 0 0 20 40 60 80 100 a t n ria a n co t yp ai isi az no a rd oc ge w Percent n Eg G ba Ku Tu Jo or Al & Le M k an tB es W Source: World Values Survey 2014. BOX 5.2 HOW GOVERNANCE THROUGH POLITICAL ACCOUNTABILITY IMPROVED HEALTH SERVICES IN CEARA, BRAZIL In 1987–94, reform leaders in the Brazilian state of Ceara used competitive municipal elections to provide better health services through governance that transformed incentives, intrinsic motivations, and professional norms. Governors with reformist programs established a new cadre of public health workers, who greatly improved the state’s health outcomes in a few years.1 The model was scaled up nationally, improving health outcomes in poor areas countrywide.2 To accomplish these aims, Ceara’s governors created a new, professional cadre of health service providers who were held accountable, eschewing a prior system of patronage jobs and political hin- drances to service delivery. As the reformers recruited and trained the new cadre, they took steps to reduce the risk of its capture by political patronage. They flooded news media with information about the value of public health and the role of the new cadre, creating public expectations and demands of public officials. This helped cement (and reward) elected officials’ commitment to the reform, and instilled public health workers with a sense of professionalism driven by peer pressure to perform. The evidence from Ceara suggests two keys to the emergence of effective municipal governance: locally contested elections, and a transformation of local public sector incentives and professional norms through recruitment based on merit and commitment to public service. It also highlights com- munication to ensure accountability to citizens. Notes: 1. Tendler and Freedheim 1994. 2. Rocha and Soares 2010. Source: Athar and Khemani 2022. 132 VI B R A N T C I T I E S FIGURE 5.4 CAIRO, AMMAN, AND CASABLANCA: THREE POINTS ON A SPECTRUM OF URBAN GOVERNMENT ACCOUNTABILITY Spectrum of accountability, incentives, and institutional performance Cairo Amman Casablanca Very centralized, minimal Hybrid institution, medium Decentralized, strong political engagement political engagement political engagement Selection of • Top officials directly • Mayor and part of city • Locally elected mayor and officials appointed by central council appointed by PM; city council government rest of council elected • Transparent, competitive • Little transparency in • City has power to appoint process for heads of selection its top staff service providers Accountability • Upward reporting, • Mayor mostly accountable • High local political mechanisms with limited political to central government, engagement engagement for urban less to elected council • Citizens’ and businesses’ services • Limited political representation mandated engagement by law Incentives for • Incentives for officials to • Incentives for mayor to • Relatively strong and well- performance meet priorities of central meet priorities of central aligned formal incentives government government • Uncertain tenure for officials Mandates and • No “municipal” • Relatively strong • Strong mandate, authority governance government; three mandates, with financial and coordinating role for separate governorates independence municipal services • Extreme fragmentation of • Need for better clarity and mandates, with no formal coordination with central coordination government Source: Athar and Khemani 2022. with national government appointments, the city Even the best-intentioned reform leaders cannot has no municipal government as such, and so min- deliver without a professional cadre of urban imal accountability mechanisms for inhabitants. At public officials who are motivated, accountable, the other end is Casablanca, where the government and capable. To effectively manage city resources is strongly decentralized, and local accountability and service delivery, the myriad human personnel is real. Amman is in the middle: its government is a who staff the city’s government departments need political hybrid­—­part elected, part appointed­—­and appropriate motivation and professional norms. has some accountability to citizens.6 A leader with legitimacy can enact reforms to develop the right set of incentives, motivations, and Strong governance and accountability are becom- norms to move the city’s public service equilibrium ing increasingly relevant in MENA, with several cit- from low performance and low accountability to ies using new delegated models of service delivery high performance and high accountability. and municipal land development. Moroccan cities rely on delegated service providers and public sec- Monetary performance incentives for public offi- tor companies to manage municipal public services cials are likely to be less effective than reforms under city government contracts (box 5.3). Munici- that improve intrinsic motivation and professional pal companies in cities in Jordan and Saudi Arabia norms. In Pakistan’s Punjab province, perfor- have been set up to develop municipal lands for mance-pay incentives for frontline tax officials services and investment (box 5.4). —­ led to large increases in property tax revenue­ but Chapter 5  Taking the next steps for tomorrow’s vibrant cities 133 BOX 5.3 —­ CASABLANCA­ DELEGATED PUBLIC SERVICES AND PUBLIC SECTOR COMPANIES Casablanca possesses a level of sophistication in its institutional structure, but with substantial chal- lenges that have not been overcome. Like other cities and local governments in Morocco­ —C­ asablanca has a locally elected mayor and city council. The municipality of Casablanca constitutes the core of a large agglomeration called “Greater Casablanca,” comprising 18 municipalities covering a population of more than 4 million people. Each municipality is administered by a council whose members are elected by direct universal suffrage. There is a high degree of local political engagement and partici- pation in local elections. This implies a relatively higher degree of legitimacy of the city government as it represents the interests of the local citizenry. The city also has relatively strong mandates for local service delivery and provision of infrastructure. Like other cities of the kingdom, the municipality of Casablanca is responsible for the management of a wide set of municipal services­ —s­ uch as water supply; sanitation and wastewater; solid waste man- agement; electricity distribution; urban public transport, traffic management, and public spaces. One of the decentralization objectives of constitutional and legal reforms in recent decades has been to move from an administrative model in which the central government exercises close supervision over local authorities to a model in which it acts more as a partner and supporter. Since 2015, municipalities are primarily responsible for managing municipal services and local infrastructure. A key aspect of local service delivery in Moroccan cities­—b­ ut especially prevalent in Casablanca­—­is the use of delegated service providers and public sector companies as parties contracted by the city government to manage municipal public services. City governments can delegate their responsibilities to a special purpose vehicle under contract. These can be either city-owned public companies gov- erned by the law applicable to public limited companies created at the initiative of elected councils, or private sector providers contracted by the city. The relationship between the city and delegated service providers is governed by a law that aims, in particular, to promote transparency in the awarding of management agreements and to promote a balanced contractual relationship between the delega- tor and the delegatee. The top officials of these service providers are appointed through a relatively transparent and competitive process, which implies that intrinsically motivated officials are selected for these assignments. The use of delegated public services has enabled Casabalanca to increase investment in urban infrastructure and services. Delegated management has also been recognized to have contributed to modernizing management practices in this sector, professionalizing jobs, and introducing information systems through staff training. These factors that have improved the quality of municipal services. Going forward, the city still faces several challenges in its governance that could be improved. • First, clarity and coordination between the relative mandates of various tiers of government (cities, regions, and other deconcentrated departments of the central government) in the provi- sion of localized and coordinated services and infrastructure. The prevailing law grants regional administrations a degree of defacto pre-eminence over municipalities notably in investment planning, authorization, and execution. The institutional environment is also still marked by rigid and compartmental administrative structures which are difficult to adapt to rapid urbanization, despite the substantial reforms to promote decentralization and local empowerment. • Second, horizontal coordination at the metropolitan level across the various city governments. Despite its demographic weight and its important place in the country’s economy, Casablanca is governed by the same status and rules as those applied to all other municipalities, including the smallest ones. (continued) 134 VI B R A N T C I T I E S BOX 5.3 continued —­ CASABLANCA­ DELEGATED PUBLIC SERVICES AND PUBLIC SECTOR COMPANIES • Third, even with the gains achieved by delegated service providers and public companies, there is room to improve governance mechanisms and capacity. The city council generally has weak capacity in monitoring and overseeing the performance the companies delegated with services, and holding them adequately accountable to agreed service delivery provisions. There is also a multiplication of companies and service providers that would benefit from an overall strategic vision and coordination. They are increasingly perceived as a mechanism by which elected rep- resentatives of the city are deprived of their prerogatives. The central government-appointed local representatives chair the boards of directors of these companies. Source: World Bank 2021. also to higher kickback payments. In contrast, risks. The most fundamental task is to show citi- allowing Punjab province procurement officers the zens and investors that markets can be governed autonomy to make purchasing decisions reduced by common rules, not by privilege and influence. purchasing costs by almost 10 percent without Where to reside or locate a business? How to invest lowering the quality of goods procured: the offi- resources wisely? What to do to protect and insure cials’ increased autonomy may also have reduced against risks? These are the questions that workers bureaucratic delays.7 Instilling intrinsic motivation and firms should be free to answer­ —­for the most and professional norms in city officials starts with part­—­for themselves, given what they bring to a recruitment. Recruitment needs to be merit- and city and what they find there. But before they can motivation-based, emphasizing qualifications and act, they need accurate and credible information. commitment to public services. Unfortunately, That means cities must first earn people’s trust. in many developing countries, including some in the MENA region, public sector jobs are obtained Let people make choices about their own destinies through connections rather than qualifications.8 and priorities. Some may choose to invest more in children’s human capital, others to start productive Once a city has begun to meet some of the basic enterprises. Some may pool their resources for governance requirements for rising productivity market insurance against disaster risks, others may and economic growth, policymakers can aspire migrate or relocate their firms to places less exposed. to make its future vibrant. They can do this by Well-functioning markets­ for land, labor, capital, —­ informing for action, by supporting investments in and products­ —­can provide credible information connection and service provision, and by protecting to enable people, rich and poor alike, to become the less fortunate with interventions when needed. active participants in building resilience, moving up the urban ladder, and becoming entrepreneurs and INFORM MARKET ACTORS­ —­ productive workers. Strengthening these market SETTING EXPECTATIONS AND institutions should be the bedrock of policy efforts. MOTIVATING PRIVATE ACTION —­ SUPPORT PRIVATE ACTION­ To set higher expectations for market actors in a INVESTING IN CONNECTION developing country city, the government must AND SERVICE PROVISION inform them credibly about the city’s vibrant future­—­but also about risks to prosperity and When market actors see city governments as investment, including climate change–related credible informants, policymakers can effectively Chapter 5  Taking the next steps for tomorrow’s vibrant cities 135 BOX 5.4 AMMAN—MUNICIPAL-OWNED COMPANIES FOR LAND DEVELOPMENT AND PUBLIC TRANSPORT The Greater Amman Municipality (GAM) recently professionalized its public land management services to better respond to market demands thorough a municipally owned corporation, Amman Vision Investment and Development (AVID), launched in 2018. AVID also owns Amman Vision Transport, which aims to develop and improve the public transportation system in Amman, and more recently, GAM has also set up a municipal corporation to deliver solid waste management services. The mission of AVID is to spur economic development using municipal land and property assets to attract private sector investments and development. AVID is mandated to act as a “developer of municipality-owned lands, properties, and investment portfolios” and “offers investment partners access to several investment opportunities that are either market-driven, investor-driven or GAM-driven, to cover both market and community needs.”1 To leverage on these models of service and infrastructure delivery, a robust set of incentives and governance structure must be put in place to balance economic development, and service delivery needs. To do so, GAM and AVID can further enhance this model by: • Strengthening governance to clearly stipulate what the business model of its municipal compa- nies is, and how land-value capture is implemented. Because AVID has received control of about 1,100 of the most valuable public land parcels to allocate to private investors (about one-fourth of the parcels that Amman owned in late 2021), there needs to be robust policies or directives in place to use the market value of the entrusted land–such as how company revenue and profits will be spent or captured by the municipality. • Calibrating land development policies and company mandates towards positive spatial out- comes, and more optimal land use. AVID has great influence on how land use and city form will be shaped. A striking example is the allocation of the second-largest municipal parcel of land of 92 hectares by AVID, to a single investor to develop a Motor City to accommodate all dealerships in Amman–a one-stop location for selling, renting and servicing motor vehicles. While investments such as this has the potential of developing the local economy, place-based investments such as this needs to be considered carefully from a citywide perspective. Questions have been raised if investments such as this would change the land use dynamics of the city, and impact transportation infrastructure, as car owners in a 4.5-million-person city may need to travel to a single location 21 kilometers from the city center to access these services. • Coordinating closely with the municipality on service delivery mandates, and on efficient man- agement of municipal lands and assets. AVID and GAM will need a systematic way to balance the tradeoffs and/or synergies between using municipal land for investments and economic returns versus the provision of services to its populations. Furthermore, there needs to be in place a system to use land efficiently by governments or by private providers of social services when land is allocated anew. For example, AVID plans to allocate a large parcel of more than 19 hectares for a healthcare project, which could perhaps be developed with higher density on a smaller and/or mixed-used site by GAM. The governing structure between GAM and AVID–their institutional arrangement, incentives, and account- ability is critical. Special-purpose land development corporations for developing surplus public land— either a particular area or a whole portfolio of sites in multiple locations—are a well-known instrument. And all examples of international good practices have one thing in common: they are set up with a clear goal, enshrined in their governing documents and practices, to first enhance and then use the market value of public land entrusted them. They do this by maximizing commercial revenues from land and channeling profit to the government that created a corporation or reinvesting in building public infrastructure. Note: 1. https://ammanvision.jo/about-us/.  Source: Kaganova, Kaw, and Ababsa (forthcoming). 136 VI B R A N T C I T I E S invest in connection and service provision. For To mitigate the often unfair­ —a ­ nd ultimately sustainability and growth, it is essential to connect inefficient­ distributional effects of market-driven —­ fragmented urban markets, and allow changes urban development, countervailing policies are in land use in response, to harness the gains for needed. Too many children are trapped in per- urban agglomeration economies, specialization, sistent intergenerational poverty, not just by failures and inclusion. It is also essential to ensure that of service provision such as low school quality, but infrastructure plans are assessed not only for their also by a lack of social networks, positive role mod- direct impacts but also for second-order impacts els, and peer interactions. The vicious circles that on spatial form and agglomeration, as well as emis- cause such inequality traps are difficult to break, sions. For inclusion, investment in wider access to even in developed countries­ —e ­ specially when basic services is indispensable. reinforced by social discrimination and divisions based on identity. Creative solutions are needed, PROTECTING THE LESS FORTUNATE­ and programs to expand opportunity through —­WHEN NEEDED, ON A TIMELY, vouchers for housing and schooling are promising. TARGETED, AND TEMPORARY BASIS Policies are also needed to fill gaps in insurance Policymakers can protect by extending public markets, protecting people and firms from risks safety nets and targeted assistance­ —­to help the when private adaptation is infeasible. Public inter- less fortunate weather adverse shocks and to ventions may be needed for building protection open escape routes from inequality traps. Get- and resilience, chiefly in cases of emergencies and ting markets right will enable urban productivity insurance market failures­ remembering that inter- —­ and growth, but the results may not be fair to ventions (especially when place-based) create moral everyone. In developing countries, markets may hazard and inflict massive costs if done wrong. systematically favor the interests of the wealthy, as may social and political arrangements. Poor (and All policies, whether to inform, support, or protect, some middle-class) people may therefore not gain need to begin with a sound scientific assessment from market reforms alone. For example, when a that estimates direct and indirect economic effects. city enables formal land markets and invests in New urban models can help policymakers identify new services or transport infrastructure, poor expectation-setting strategies, connective invest- tenants are likely to be displaced from gentrifying ments, and complementary policies for vibrant areas, with little to show for the increased value cities, anticipating indirect effects and preventing of the land that they were forced to leave. Many unintended consequences. Certain reforms and will move out to informal settlements at the city’s investments will repay their cost many times over­ edge, where they may be disconnected from jobs —­the challenge is to learn which ones, and to make and services. them happen.9 Chapter 5  Taking the next steps for tomorrow’s vibrant cities 137 NOTES and S. David. “Making politics work for development: harnessing transparency and citizen engagement.” 1. Mookherjee 2021. Policy Research Reports, World Bank, Washing- 2. World Bank 2020a. ton, DC. http://documents.worldbank.org/curated​ 3. World Bank 2020b. /en/268021467831470443/Making-politics-work-for​ 4. Lall et al. 2021. -development-harnessing-transparency-and-citizen​ 5. World Bank 2016. -engagement. 6. Athar and Khemani 2022. Lall, S. V., M. S. M. Lebrand, H. Park, D. M. Sturm, and 7. Athar and Khemani (2022), citing Bandiera et al. A. J. Venables. 2021. Pancakes to Pyramids: City Form (2021). to Promote Sustainable Growth. Washington, DC: World 8. World Values Survey 2018. Bank. 9. Lall et al. 2021. Mookherjee, D. 2021. “The Tale of Two New Asian Tigers: Comparative Development of Selected Provinces of REFERENCES China and India since 1980.” Working Paper, Depart- ment of Economics, Boston University. Alesina, A., and P. Giuliano. 2015. “Culture and Institu- Rocha, R., and R. R. Soares. 2010. “Evaluating the Impact tions.” Journal of Economic Literature 53 (4): 898–944. of Community-based Health Interventions: Evidence Algan, Y., and P. Cahuc. 2014. “Trust, Growth, and from Brazil’s Family Health Program.” Health Econom- Well-Being: New Evidence and Policy Implications.” ics 19 (S1): 126–158. In Handbook of Economic Growth, Volume 2, 49–120. Tendler, J., and S. Freedheim. 1994. “Trust in a Rent-­ Amsterdam: Elsevier. seeking World: Health and Government Transformed Athar, S., and S. Khemani. 2022. “Unlocking Institutions in Northeast Brazil.” World Development 22 (12): of Urban Governance to Transform Cities.” Back- 1771–1791. ground paper for the Vibrant Cities report, World World Bank. 2017. World Development Report 2017: Gov- Bank, Washington, DC. ernance and the Law. Washington, DC. https://www​ Bandiera, O., M. C. Best, A. Q. Khan, and A. Prat. 2021. .worldbank.org/en/publication/wdr2017. “The Allocation of Authority in Organizations: A Field World Bank. 2020a. Convergence: Five Critical Steps toward Experiment with Bureaucrats.” The Quarterly Journal of Integrating Lagging and Leading Areas in the Middle East Economics 136 (4): 2195–2242. and North Africa. Washington, DC: World Bank. Basu, K. 2018. The Republic of Beliefs: A New Approach to Law World Bank. 2020b. “Unlocking Urban Development in and Economics. Princeton, NJ: Princeton University Morocco: Can Urbanization Pay for Itself, and How Press. to Make it Happen?” Report of PPIAF/World Bank Kaganova, O., J. Kaw, and M. Ababsa. Forthcoming. “The Technical Assistance to the Kingdom of Morocco on Need to Reform Land-Use Planning and Land Man- Financing Urbanization. Unpublished report. Wash- agement in MENA: The cases of Amman, Cairo, and ington, DC: World Bank. Casablanca.” World Bank. 2021. “Great Casablanca Governance Sys- Khemani, S., D. Bó, E. Ferraz, C. Finan, F. Shimizu, S. tem.” Unpublished report. Washington, DC: World Johnson, C. L. Odugbemi, A. M. Thapa, D. Abrahams, Bank. How will the world’s developing cities become vibrant—capable of meeting the social, climate, and economic challenges of tomorrow? Like cities that developed in the past, those that grow and thrive in the coming decades will need endowments, institu­ tions, and infrastructure to boost expec- tations for income and welfare. That is how cities attract capital and foster productive livelihoods. But tomorrow’s vibrant cities also urgently need basic survival plans to mitigate climate change and reduce exposure to disaster risks—while protecting those most exposed. Vibrant cities offer firms and households high expectations for good returns on investments, dynamic and inclusive growth, and a sustainable and resilient future. Cities thrive not by increasing incomes and wealth for a select few, but by improving common welfare through the equitable provision of basic services and opportu­ nities. To do this, tomorrow’s vibrant cities will be: Resilient and low carbon—Limiting greenhouse gas emissions, rebounding from disasters and pan- demics, and reducing vulnerability to climate-related hazards. ter Inclusive—Meeting basic needs for all residents, while enabling all to aspire realistically to a bet­ life, through investment in skills and equitable access to job opportunities. Productive—Driving economic growth, creating jobs, boosting incomes, and financing critical social and infrastructure investments. The report provides new evidence, analysis, and policy insights to advance green, resilient, and inclusive urban development. It draws on the latest thinking in spatial urban development and public economics. The report spotlights the Middle East and North Africa region (MENA) but offers general insights for city and country leaders around the world. It lays the foundation to shore up technical assistance and policy engagement for urban development in MENA and elsewhere through a new policy framework—inform, support, and protect.