id s ra m P y to e s a k n c Pa City Form to Promote Sustainable Growth Somik Lall, Mathilde Lebrand, Hogeun Park, Daniel Sturm and Anthony Venables Wichtiger HINWEIS ! Innerhalb der Schutzzone (hellblauer Rahmen) darf kein anderes Element platziert werden! Ebenso darf der Abstand zu Format- resp. Papierrand die Schutzzone nicht verletzen! Hellblauen Rahmen der Schutzzone nie drucken! Siehe auch Handbuch „Corporate Design der Schweizerischen Bundesverwaltung“ Kapitel „Grundlagen“, 1.5 / Schutzzone www. cdbund.admin.ch Supported by: In partnership with: © 2021 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved. 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 other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creativecommons. org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Lall, S. V., M. Lebrand, H. Park, D. Sturm, and A. J. Venables. 2021. Pancakes to Pyramids: City Form to Promote Sustainable Growth. Washington, DC: World Bank. License: Creative Commons Attribution CC BY 3.0 IGO. Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org. Design: Zephyr www.wearezephyr.com Foreword What drives the shape of cities, and what Building tall is not a matter of creating a actions can policy makers take to guide their distinctive skyline with notable skyscrapers. growth? Much more than a vanity project, enabling the construction of taller buildings—say, The authors of Pancakes to Pyramids set out 5 to 10 stories high—is a matter of livability; to find out. I am pleased to say that they have vertical layering creates enough floor space succeeded in increasing our understanding to accommodate growing populations without of the economic variables that drive urban packing people into smaller and smaller expansion, while challenging conventional spaces. It’s the difference between crowding wisdom about sprawl. Most importantly, they versus livable density, slums and sweatshops have opened up a field of inquiry that will be versus more humane housing and office central to the World Bank’s mission of poverty conditions. reduction and sustainable and inclusive development in the years ahead as leaders The Covid-19 pandemic has highlighted strive to create green, resilient, and inclusive the life-and-death implications of crowded cities that attract people and businesses. neighborhoods that are ill equipped to curb the spread of disease. As countries slowly As low- and middle-income countries urbanize extricate themselves from the pandemic, in the decades ahead, this report provides planning for a better urban future requires new evidence for city leaders interested in understanding the forces that have shaped the managing spatial growth. It also provides a cities we inhabit today. theoretical model to test assumptions about compactness and public transport that will be My hope is that this report helps to start a crucial to rein in commuting time, fuel use, and conversation about urban growth and policy greenhouse gas emissions. choices at all country income levels, to afford people everywhere the opportunity to live in Drawing on ground-breaking data covering decent housing, apply for competitive jobs, almost 10,000 cities between 1990 through access affordable services, and thrive. 2015, this report shows a dynamic two-way relationship between a city’s economy and the height of its buildings. The extent to which firms benefit from concentrating their workers in one place is one of the factors that determines how compact a city is likely to be. Juergen Voegele Other factors include the total size of a city, Vice President the ease with which people can move around Sustainable Development Practice Group the city, and the ability of developers to build World Bank tall—an ability that may encounter regulatory, technical, or financial constraints. Pancakes to Pyramids | City Form to Promote Sustainable Growth i Contents Foreword by Juergen Voegele i Acknowledgments vii About the authors viii Abbreviations ix Why this report? x Overview 1 Pancakes to pyramids: Physical manifestations of city development 3 Sustainable densities 6 Five stylized facts 8 — Stylized fact 1: The growth of urban built-up area worldwide is not as 8 large as conventional wisdom suggests — Stylized fact 2: In developing country cities, spatial expansion is 9 happening mostly through horizontal spread — Stylized fact 3: Increasing incomes are the one indispensable driver 9 of vertical layering, because building tall is capital intensive — Stylized fact 4: Dysfunctional urban land markets, along with zoning and restrictive building regulations, are factors that can militate 10 against taller structures, economic density, and pyramidal growth — Stylized fact 5: Improved transport technology enables economic concentration in urban cores, supports cities’ economic and spatial 11 growth, and boosts demand for livable residential floor space Policy implications 12 I How are cities accommodating families and firms—and how fast is global 1.  16 built-up area expanding? The spread of urban built-up area is slower than has been thought 24 Who is contributing the most to urban built-up expansion? 25 Spotlight 1 Stages of urban economic development can be seen in cities’ 28 average building heights—and in their skylines ii Pancakes to Pyramids | City Form to Promote Sustainable Growth II What has driven urban spatial evolution since 1990? 2.  The answers are complicated—but incomes are key 32 How cities grow 33 How income and population growth drive overall urban built-up expansion 33 How income and population growth drive vertical layering 37 How cities supply floor space by building upward and expanding outward 40 How population density responds to income growth 43 Spotlight 2 Disruptive technologies and the future of cities: 45 A policy and analytic agenda III 3.  What drives one city to grow differently from another? 46 Using a structural urban model to clarify cities’ spatial development 47 What determines the shape of the city? 52 — The absence—or presence—of agglomeration economies as a key 52 determinant of city shape — Construction costs as a constraint on floor space growth and 55 building height — City population size as a driver of downtown rent, building height, and 56 employment density — Commuting costs as drivers of rents, job locations, and residential 58 densities across the city Projecting the effects of policy on cities’ spatial growth: The new urban model 60 in action Using the new urban model to estimate the impacts of reduced travel times 62 and increased housing supply in metropolitan Dhaka Spotlight 3 Urban form and greenhouse gas emissions 69 IV 4.  From pancakes to pyramids: What city leaders need to know 72 Policy making for economic growth and productivity—strengthening 75 institutional foundations — Reform land markets and clarify land and property rights 76 — Strengthen urban land use planning 79 — Manage land valuation and prices 81 Infrastructure planning for economic density, livability, and sustainability— 82 scaling up and evaluating investments — Understand path dependence 82 — Account for the interdependence of infrastructure and physical 83 structures — Coordinate infrastructure investments, land use plans, and zoning 83 regulations Pancakes to Pyramids | City Form to Promote Sustainable Growth iii — Make infrastructure investments as early as possible—while 84 coordinating them with urban plans Durable financing for capital investment costs and recurring expenses— 87 mobilizing urban revenue sources — Tax the value of urban land and property 87 — Consider public–private partnerships (PPPs) for service delivery 88 — Borrow on capital markets—if the city’s creditworthiness allows 88 — Fix the fundamentals—fiscal, economic, and institutional—to mobilize 89 urban finance Annex 1 Income elasticity of population density 90 Annex 2 Data and methodological details 91 Global Human Settlement Built-Up grid (GHS-BUILT) 91 — Comparative advantages of GHS-BUILT 91 — Concern with GHS-BUILT 92 Degree of Urbanization 92 — Comparative advantages of Degree of Urbanization 92 — Concerns in use of Degree of Urbanization and GHS-UCDB 93 World Settlement Footprint (WSF) 3D 96 Annex 3 Spatial development of the five largest cities in each region of the 98 world Works cited 134 Figures Figure 1 The growth of a city along three margins—horizontal spread, infill 3 development, and vertical layering Figure 2 The virtuous cycle of urban economic development and spatial 5 transformation Figure 3 Two forms of urban residential density: Crowding in Dharavi, tall 6 buildings in Manhattan Figure 4 Weak zoning and land use plans in cities of low- and lower-middle- 11 income countries Figure 5 Flexible rules enabled Seoul’s infill redevelopment 13 Figure 1.1 Coverage of cities studied in this report 17 Figure 1.2 The growth of a city along three margins—horizontal spread, infill 22 development, and vertical layering Figure 1.3 Urban built-up area expansion by national per capita income level, 25 1990–2015 Figure 1.4 Urban built-up area expansion by region, 1990–2015 26 Figure 1.5 Urban built-up area expansion in China and India, 1990–2015 27 Figure S1.1 Average building height in Accra (low income) 29 Figure S1.2 Average building height in Lagos (lower middle income) 29 Figure S1.3 Average building height in Mexico City (upper middle income) 29 Figure S1.4 Average building height in Paris (high income) 29 Figure S1.5 Average building height in Houston (high income) 30 iv Pancakes to Pyramids | City Form to Promote Sustainable Growth Figure S1.6 Average building height in Sydney (high income) 30 Figure 2.1 Richer cities are more pyramid shaped, with peaked skylines 37 Figure 2.2 How income and population are associated with cities’ vertical 39 structure Figure 2.3 The elasticities of a city’s built-up area, average building height, 42 and total floor space to its population and income Figure 2.4 Predicted density over built-up area 44 Figure 3.1 Some patterns predicted by the model for a hypothetical city—built 50 on a “featureless plain” Figure 3.2 Spatial distribution of residents and employment (tradables and non-tradables sectors) and heights of buildings in a city lacking 53 agglomeration economies Figure 3.3 Effects of agglomeration economies on other city characteristics in 54 the new urban model Figure 3.4 Effects of construction costs on other city characteristics in the 56 new urban model Figure 3.5 Effects of city population size on other city characteristics in the 57 new urban model Figure 3.6 Effects of commuting costs on other city characteristics in the 58 new urban model (fixed population) Figure 3.7 Population and employment density in Dhaka 63 Figure 3.8 Height of buildings and volume of floor space in the model and the 65 data for Dhaka Figure 3.9 Estimated land rent in Dhaka 66 Figure 3.10 Estimated productivity in Dhaka 66 Figure 3.11 Estimated impact of a 25 percent increase in the housing supply 67 elasticity in Dhaka Figure 3.12 Estimated impact of a new north–south radial road in Dhaka 68 Figure S3.1 While GHG emissions per capita decline with urban density in high- and upper-middle-income countries, low- and lower-middle-income 70 cities buck the trend Figure 4.1 Land values of new development projects are higher than values in 85 other neighborhoods of Dar es Salaam, including rich ones Figure 4.2 Differential impacts of new development projects and upgrading 86 projects in Dar es Salaam Figure A2.1 Different urban boundaries in Beijing, China 93 Figure A2.2 Estimated population within the urban area in the top 50 growing 95 cities, 1990 and 2015 Figure A2.3 Schematic process of WSF 3D development 96 Figure A2.4 Average building height in Chicago, Illinois 97 Maps of built-up area for 35 cities 99–133 Tables Table S1.1 Average building height moving away from the central business 31 district Pancakes to Pyramids | City Form to Promote Sustainable Growth v Table A1.1 Determinants of naïve population density (population over built-up 90 area) Table A2.1 Comparison between GHS-UCDB and World Urbanization 94 Prospects Table A2.2 Built-up area comparison between GHS-UCDB and FUA 94 Table A2.3 Built-up area comparison in high-income land-rich countries 95 Table A2.4 Selected cities for WSF-3D data 97 Boxes Box 1 Will the Covid-19 pandemic reshape urbanization? 2 Box 2 Cultural and natural amenities are important over the course of a city’s 14 growth and development Box 1.1 What is a city? 18 Box 1.2 Height data on cities everywhere: The World Settlement Footprint 3D 19 Box 1.3 Seeking livable urban densities around the world: From population 20 density to per capita floor space Box 1.4 The benefits—and costs—of urban density 23 Box 1.5 Sustainable Development Goal 11 and urban built-up area expansion 24 Box 2.1 Econometric issues in estimating the drivers of built-up area 34 expansion Box 2.2 Two concurrent processes: Horizontal spread and infill development 36 Box 2.3 Technical specifications of city height estimates 38 Box 2.4 Technical specifications of floor space estimates 41 Box 3.1 What is the new urban model, and how can its analytic framework 47 help city leaders? Box 3.2 Tradables and non-tradables in the urban economy 49 Box 3.3 Agglomeration economies 51 Box 3.4 How transport innovations have shaped cities 61 Box 3.5 Density-enhancing land value capture schemes can amplify the 62 economic benefits from Bogotá’s Transmilenio bus rapid transit system Box 3.6 Transit infrastructure lowers informality rates in Mexico City 62 Box 4.1 Data requirements for the new urban model 74 Box 4.2 Urbanization in low-income and lower-middle-income countries: How 76 will it drive global demand for urban floor space in the coming decades? Box 4.3 Urban slums—a lagging indicator of inefficient land markets 77 Box 4.4 Using the new urban model to scrutinize received assumptions about 83 the economic consequences of urban policies and investments Box 4.5 The global urban infrastructure financing gap: A challenge for 88 developing country cities Box 4.6 Taxing land and property values to finance urban infrastructure 89 vi Pancakes to Pyramids | City Form to Promote Sustainable Growth Acknowledgments This report, a flagship report of the World The report was produced under the Bank’s Sustainable Development Practice supervision of Sameh Wahba, Global Director, Group, was prepared by a team led by Somik V. Urban, Disaster Risk Management, Resilience, Lall and comprising Mathilde Lebrand, Hogeun and Land Global Practice at the World Bank. Park, Daniel Sturm, and Anthony J. Venables. Generous funding was provided by the Swiss It benefited from the contributions of Gaurav State Secretariat for Economic Affairs (SECO) Bhardwaj, Josep Casas, Maria Edisa Soppelsa, through the World Bank Umbrella Multi-Donor Benjamin Stewart, and Kohei Takeda. Thomas Trust Fund Sustainable Urban and Regional Esch and Mattia Marconcini from the German Development (SURGE), and the Global Aerospace Center were key research partners. Environment Facility (GEF) through the Global Lewis Dijkstra, Indermit Gill, Forhad Shilpi, and Platform for Sustainable Cities (GPSC). Nick Sameh Wahba were key advisors to the report Moschovakis and Bruce Ross-Larson, with a team. team at Communications Development, edited the report, which Zephyr designed. The report draws on three background papers: The Evolution of City Form: Evidence from Satellite Data. World Bank Policy Research Working Paper 9618. Somik Lall, Mathilde Lebrand, and Maria Edisa Soppelsa. Exploring the Urban Model: Employment, Housing, and Infrastructure. Daniel M. Sturm, Kohei Takeda, and Anthony J. Venables. Applying the Quantitative Urban Model to Cities in Developing Countries. Daniel M. Sturm, Kohei Takeda, and Anthony J. Venables. The team received valuable comments from Arturo Ardila Gomez, Judy Baker, Louise Cord, Richard Damania, Georges Darido, Lewis Dijkstra, Peter Ellis, Matthew Kahn, Paul Kriss, William Maloney, Nancy Lozano Gracia, Mark Roberts, and Forhad Shilpi. It benefited from discussions with Maitreyi Das, Olivia Daoust, Lincoln Lewis, Sally Murray, Ethel Sennhauser, Anastasia Touati, Xueman Wang, and Roland White. And it appreciated the opportunity to discuss the analytic and policy framework and findings at various forums, including those at the North American Meetings of the Urban Economics Association and the SECO-World Bank annual roundtable. Pancakes to Pyramids | City Form to Promote Sustainable Growth vii About the authors Somik V. Lall is the World Bank’s global lead Daniel Sturm is a professor of economics in on territorial development solutions, a lead the Department of Economics at the London economist for urban development, and the School of Economics (LSE) and a research team leader of this report. He is a recognized fellow of the Centre for Economic Policy expert on development policy related to urban Research (CEPR). He received his PhD in and territorial competitiveness, agglomeration Economics from the LSE in 2001. Prior to joining and clusters, and infrastructure, with more the LSE in 2006, he was an assistant professor than 20 years’ global experience. He was the of economics at the University of Munich. lead author of the World Bank’s flagship report His current research focuses on modeling on urbanization: Planning, Connecting, and the agglomeration and dispersion forces that Financing Cities—Now. Somik heads a World shape the distribution of economic activity Bank global research program on urbanization both within and across cities. His work has and spatial development and founded the been published in leading academic journals, Bank’s Urbanization Reviews program. He has including the American Economic Review, written more than 40 publications in peer- Econometrica, Quarterly Journal of Economics, reviewed journals, edited volumes, and working Review of Economic Studies, and Review of papers. Economics and Statistics. He won the 2018 Frisch Medal for the paper “The Economics of Mathilde Lebrand is an economist in the Density: Evidence from the Berlin Wall,” with World Bank’s Chief Economist Office for Gabriel Ahlfeldt, Steve Redding, and Nikolaus Infrastructure. She previously worked in Wolf, and his research is currently supported the Transport Global Practice and Europe by a five-year Advanced Grant from the and Central Asia Chief Economist Office at European Research Council (ERC). the World Bank, the University of Montreal, and the World Trade Organization. She is a Anthony Venables is professor of economics Research Fellow at the Center for Economic and research director of the Productivity Studies ifo Institute (CESifo) and holds a PhD Institute, University of Manchester, and senior in economics from the European University research fellow at the University of Oxford. Institute. Mathilde’s interests are in economic He is a fellow of the Econometric Society, geography, urban studies, transport, and the Regional Science Association, and the international trade. British Academy. Former positions include professor of economics at Oxford University, Hogeun Park is a junior professional officer chief economist at the UK Department for with the World Bank’s Urban, Disaster Risk International Development, and professor Management, Resilience, and Land Global at the London School of Economics. He has Practice for the Middle East and North Africa. published in the International Trade Journal, His work provides advanced analytics for Journal of Economic Geography, and Natural complex urban challenges through a spatial Resources Journal. His publications include lens. Hogeun was Big Pixel postdoctoral fellow The Spatial Economy: Cities, Regions and at University of California San Diego. He has International Trade, written with M. Fujita and P. published widely in leading peer-reviewed Krugman (MIT Press, 1999). journals on urban development, spatial analytics, and remote sensing. viii Pancakes to Pyramids | City Form to Promote Sustainable Growth Abbreviations AMM Alonso-Muth-Mills model BRT Bus rapid transit CBD Central business district CEPAC Certificate of Potential Additional Construction CCFLA Cities Climate Finance Leadership Alliance CIESIN Center for International Earth Science Information Network DLR German Aerospace Center DP Displaced person EAP East Asia and Pacific FAR Floor-area ratio FUA Functional urban area GHS-BUILT Global Human Settlement Built-up Area Grid (dataset) GHS-UCDB Global Human Settlement Urban Center Database (dataset) GDP Gross domestic product GGMCF Global Gridded Model of Carbon Footprints (dataset) GHG Greenhouse gas GLTN Global Land Tool Network GPR Gross plot ratio GUF Global Urban Footprint (dataset) JRC European Commission Joint Research Centre KRIHS Korea Research Institute for Human Settlements LVC Land value capture MENA Middle East and North Africa MODIS Moderate Resolution Imaging Spectroradiometer NASA National Aeronautics and Space Administration N-sector Non-tradable sector OECD Organisation for Economic Co-operation and Development PPP Public–private partnership SDG Sustainable Development Goal SSA Sub-Saharan Africa TDR Transferable development rights T-sector Tradable sector UN United Nations UNDESA United Nations Department of Economic and Social Affairs UNESCO United Nations Educational, Scientific and Cultural Organization UN-Habitat United Nations Human Settlement Programme WSF-3D World Settlement Footprint 3D (dataset) WUP World Urbanization Prospects Pancakes to Pyramids | City Form to Promote Sustainable Growth ix Why this report? Policy makers have often misjudged the potency of market forces. Many policy makers perceive cities as constructs of the state—to be managed and manipulated to serve some social objective. In reality, cities and towns, just like firms and farms, are creatures of the market. WORLD DEVELOPMENT REPORT 2009: RESHAPING ECONOMIC GEOGRAPHY Towns and cities are economic and social central districts to increase economic and microcosms in which large numbers of people residential densities. Both types of density and firms interact. These interactions largely can help cities overcome the challenges shape how a city looks, how it functions, that come with population growth, and and how it grows. But how exactly does this most urgently, evolving from a pancake into many-sided relationship work? What are a pyramid, creating a platform with more the specific drivers of urban economic and options for controlling greenhouse gas spatial development? emissions. Pancakes to Pyramids brings us closer to The report draws on new evidence, answering these questions, beginning with econometric analysis, and predictive an idealized contrast between two patterns modeling to relate the economic growth of of urban spatial growth. Pancakes are cities cities to their past spatial evolution—and that grow outward and remain relatively to the possibility and conditions for future low-built. Pyramids are cities that grow partly pyramidal growth. Urban spatial expansion outward, but also partly inward and upward, is examined across 9,500 cities worldwide, filling vacant parcels and adding height to making novel use of satellite imaging data x Pancakes to Pyramids | City Form to Promote Sustainable Growth from the Global Human Settlement Urban City leaders who understand these dynamics Centre Database (GHS-UCDB). Econometric are best able to anticipate future needs— analysis is used to distinguish multiple choosing the right urban plans, policies, and drivers of urban spatial growth. Based on this investments, in the near term, that will create analysis, a canonical economic model—the enabling conditions for the city’s sustainable new urban model—is described and then growth at later development stages. applied to counterfactual experiments that predict how cities may evolve under different plans and policy scenarios. The stakes of these spatial evolutionary processes for today’s rapidly growing lower income cities are high. In South Asia, Sub- Saharan Africa, and elsewhere, wherever urban horizontal expansion is not managed well, cities will become unlivable (Ellis and Roberts 2016; Lall et al. 2017). The harmful effects of uncontrolled pancake expansion are overwhelmingly likely to appear in cities that get big while remaining poor. But they can also occur in cities that grow at higher income levels, if the spatial expansion of these cities is unplanned—or if their planners make poorly founded assumptions about how cities grow. How can today’s low- and lower-middle- income cities plan for growth and development tomorrow? What should leaders do now to put these cities on a livable, sustainable path, averting a future of sprawl and smog? Pancakes to Pyramids describes how economic and spatial development processes go hand in hand. The dynamics of urban spatial evolution can be complex, but their relation to economic growth and development is increasingly well understood. Pancakes to Pyramids | City Form to Promote Sustainable Growth xi Overview There is no logic that can be superimposed on the city; people make it, and it is to them, not buildings, that we must fit our plans. Jane Jacobs 1958 Spatial expansion of cities requires land, but the final product of urbanization is floor space. Alain Bertaud 2018 xii Pancakes to Pyramids | City Form to Promote Sustainable Growth Overview The main report consists of four chapters: • Chapter 1 describes new global Today more than half the world’s population evidence—derived from satellite data— (55 percent) resides in urban areas. Three for rates and patterns of urban spatial decades hence, this share will likely surpass development since 1990. It examines two-thirds (68 percent by 2050 in a recent spatial development across 9,500 cities projection; United Nations 2018c). Such rapid worldwide, using satellite imagery from the growth in cities creates new opportunities—but Global Human Settlement–Urban Centre it is also putting new pressures on cities and Database 2015 (GHS-UCDB). Developed countries. Much of the new urbanization will by the European Commission’s Joint take place in Sub-Saharan African and South Research Centre (JRC), the GHS-UCDB Asian countries with low incomes and weak estimates city incomes and populations institutional and fiscal capabilities. By one for 1990, 2000, and 2015. To analyze estimate of the Organisation for Economic what has changed in a geographical Co-operation and Development, growth and space and to reliably measure how cities development globally will require infrastructure have evolved, it uses the urban extent investments of $6.3 trillion a year over 2016– corresponding to the most recent data, 30 (OECD 2017). those from 2015. A novel feature is the report’s definition of cities using the How will city leaders and national governments “Degree of Urbanization” approach: a keep up with urbanization, enabling cities new, consistent basis for international to adapt and thrive—sustainably—as their comparisons of urban growth developed populations continue to grow rapidly? by the European Commission and Recent approaches to this question in urban endorsed in March 2020 by the United development policy have tended to focus on Nations Statistical Commission. The urban density as key to livability and efficient methodology is explained in chapter 1 and resource use, inspiring plans that seek to annex 2 of the main report, as well as in recreate features of downtown districts in Lall, Lebrand, and Soppelsa (2021). rich countries. While Pancakes to Pyramids • Chapter 2 combines this global evidence also looks at density, it does so differently, with economic data and econometric reframing the question as an empirical one analysis to identify the drivers of recent about how cities evolve. The most useful urban spatial growth along three margins: guidance for planners starts not with emulating horizontal spread (outward extension), efficient cities, but with understanding infill development (inward additions in the processes that affect all cities due to urban gaps left between earlier structures), and spatial evolution. vertical layering (upward construction). The end product of this growth is floor What drives changes in a city’s shape, in the space, the amount and distribution of pattern of its spatial expansion, and in its which are central to understanding how a related distribution of housing and of economic city becomes livable and sustainable. activity? Which economic and institutional forces do most to determine the evolution of • Chapter 3 provides an analytical urban form and function? How do these various framework—based on the analysis in drivers interact? Combining new data with chapter 2—to explain why one city grows new analysis, Pancakes to Pyramids advances differently from another. An important policy makers’ insight into these questions— contribution of chapter 3 is to incorporate and helps city and country decision makers the vertical structure of cities, a nascent seek evidence-based, actionable answers to field in urban economics. Most previous the challenges facing cities now. urban economics literature on city Pancakes to Pyramids | City Form to Promote Sustainable Growth 1 structure has focused on horizontal development (Ahlfeldt and Barr 2020). • Chapter 4 guides leaders in using the analytical framework to assess and quantify the likely impacts of particular plans and policies on urban spatial growth. For example, city leaders can use the framework and its underlying predictive model to better project the impacts of investments such as transport infrastructure and of changes to zoning laws, building codes, and land use regulations on the city’s productivity and its future physical form and human geography. The report thus draws on recent advances Box 1 Will the Covid-19 pandemic reshape in two areas—satellite imagery and urban urbanization? economics—to give leaders new tools for investigating the combined effects of There is much current discussion about how the economic drivers and policy choices on a city’s Covid-19 pandemic will affect the future shape of development path. Governing the drivers of cities. Although it is impossible to say with certainty urban spatial form and function are complex what the lasting impacts of Covid-19 will be on institutional and decision-making processes urbanization, a number of reasons suggest that the that combine with fundamental economic long-run impact may not be very large. Historically, outbreaks of the plague often killed large shares of the forces to affect the size and shape of a city’s urban population, but the terrible death tolls did little to built-up area, the heights of its structures, stop the continuous march toward urbanization. the contours of its skyline, the distribution of Four reasons suggest why the impact of Covid-19 its population densities, and its floor space on modern cities might be largely temporary. First, per person. By illuminating how economic only a relatively small share of jobs can be done from productivity shapes location decisions by home. Dingel and Neiman (2020) estimate that about households and firms over time, and how the 37 percent of jobs in the United States can be done from home, and a substantially lower share, estimated quantity and spatial distribution of urban floor at less than 10 percent, in less developed countries space respond to these changes in demand, (Gottlieb et al. 2021). Second, although many jobs Pancakes to Pyramids can help decision can in principle be performed from home, face-to-face makers identify the planning and regulatory interactions in an office are likely to offer substantial productivity benefits, as argued by Storper and approaches most likely to promote prosperity Venables (2004). and sustainability. As countries and cities grapple with the challenges brought about Third, cities are places not just to work but also to consume. To consume urban amenities such as by Covid-19, the fundamental forces shaping restaurants, theaters, or museums requires travel, urbanization are likely to be central for a which advantages central locations in cities. The resilient recovery (box 1). analysis in this report shows that such consumption access can generate concentrations of service sector jobs in central locations, even without agglomeration forces in the production of tradable goods (chapter 3). Fourth, in a world of hybrid working—with some days in the office and some at home—long commutes from the fringes of cities to well-paying jobs in urban centers could actually become more attractive rather than less. 2 Pancakes to Pyramids | City Form to Promote Sustainable Growth Pancakes to pyramids: Physical manifestations of city development Cities are economic and social microcosms in which thousands—or millions—of people and firms interact. These interactions shape a city’s growth along three margins (figure 1): • Horizontal spread—extending beyond the city’s previously built-up area. • Infill development—closing gaps between existing structures. • Vertical layering—raising the skyline of the existing built-up area. Figure 1 The growth of a city along three margins—horizontal spread, infill development, and vertical layering Built-up area Defined by the presence of buildings Urban area Cities grow along three margins Horizontal spread Infill development Vertical layering Source: Authors’ depiction. Pancakes to Pyramids | City Form to Promote Sustainable Growth 3 As cities grow in productivity and in population, While pancake expansion and pyramidal they add floor space by expanding outward, expansion are physical manifestations of city inward, upward, or—more usually—along all development, the drivers of that development three margins to varying degrees. The report are largely economic and institutional. What uses the terms pancakes and pyramids as decides whether a pancake city evolves into a shorthand for two broadly different tendencies pyramid? in the physical manifestation of city growth: Much of the answer—though not all—lies in • Cities with low productivity and income economic growth, productivity, and trade. A levels generally grow as pancakes—flat pancake city’s chance for pyramidal expansion and spreading slowly.1 Low economic hinges on its success at nurturing highly demand for land and floor space productive economic activities that benefit keeps land prices low and structures from urban scale and agglomeration potential. close to the ground, especially at the These activities are likely to be in tradable urban edge. Given slow expansion, sectors: concentrating in cities, firms produce growth in population density is often goods and services to supply to buyers accommodated by crowding, starkly outside the city and possibly internationally.2 visible in the slums of developing country cities. The other piece of the answer is found in laws, • Cities with higher productivity may institutions, and capacity. Pyramids are more evolve from pancakes into pyramids— likely to evolve in countries and municipalities their horizontal expansion persists, yet where property rights are clear, land values it is accompanied by infill development are transparent, land use and zoning are and vertical layering. A rising demand for compatible with local preferences, and the floor space in economically productive enabling environment encourages durable cities (especially near downtown centers), investment in infrastructure—especially early combined with a related rise in housing investment, informed by forward-thinking investment and consumption, leads urban plans. developers to fill vacant or underused land at and within the city edge with new A canonical framework developed for this structures. These pockets of close-in land report (Sturm, Takeda, and Venables 2021a) become dense with office and residential outlines the key institutional and economic space. The same demand for floor space drivers: drives expansion not just horizontally in • Economic structure—the extent to which two dimensions, but also in the third— tradable sector firms can benefit from the vertical. Structures are built taller on agglomeration economies. average, and at the urban core, they are built much taller, forming sharply peaked • Population size—the ability of larger skylines. cities to have higher productivity, because agglomeration economies reflect scale along with density. Large urban populations also create opportunities for scale economies in providing network infrastructure and local public goods. 1 Pancakes typically range from 4 to 10 inches in diameter and are 1/3 inch thick (https://en.wikipedia.org/wiki/Pancake). They don’t hold well if they are made any larger. Also known traditionally in the United States as griddle cakes, pancakes are found under different monikers around the world: injira (Ethiopia), lahoh (Somalia and Yemen), crepe (Belgium, France), blini (Russia), dosa (India), panekuk (Indonesia), tiganites (Greece). 2 A tradable sector supplies goods and services that are exported to other regions or countries. These include  manufacturing and various services, from business, legal, and financial services to media and education. 4 Pancakes to Pyramids | City Form to Promote Sustainable Growth • Ability of developers to build tall —an While pyramidal development with peaked ability that may encounter regulatory, skylines may be the outcome of strong technical, or financial constraints. It agglomeration forces and complementary requires: institutions, steepening a city’s skyline should not be the focus of policy, nor the key metric — Institutions, governance, and urban by which urban success is judged. Improving plans and policies. A city’s shape living standards and sustainability should be. reflects the institutional structure of the land market, building and land use Urban economic development and spatial regulations, taxes, and the investment transformation form a virtuous cycle that in and placement of public assets—in enhances livability and promotes sustainability. particular, the transportation network. Pyramidal growth is preferable to pancake These factors—grounded in law, growth because pyramidal growth keeps this governance, and policy—are explored virtuous cycle going (figure 2). in some of the experiments conducted for this report. — Building technologies—innovations that make increases in building height less costly. • Transport and mobility —the ease with which people can move around the city, including but not limited to public transport investments (such as metro rail systems). Figure 2 The virtuous cycle of urban economic development and spatial transformation Better & more tradable sectors Greater, stronger agglomerated urban economy Better & more human resources Pyramidal urban development Better & more sustainable living conditions Source: Authors’ depiction. Pancakes to Pyramids | City Form to Promote Sustainable Growth 5 Sustainable densities Density can help cities become productive, to livability when they enhance sociability and livable, and sustainable, though it does not access to amenities. Finally, higher densities inherently make them so. Population density promote sustainability if they enable urban alone does not guarantee that a city will residents to use resources less intensively, realize its potential for economic efficiency especially by reducing carbon emissions. and productivity. And not all densely settled neighborhoods provide residents with decent Residential density in cities can take two forms. housing and amenities—or with affordable One less livable form of density is crowding— transport options to connect workers to job people are packed into small amounts of floor opportunities. space per person. The other, more livable form is vertical layering, as defined above—elevating Urban plans and policies thus do not just seek a city’s skyline with upper stories that amply high residential density, or a large number accommodate large numbers of residents and of residents per unit of land area. Rather, firms, while leaving room among buildings for planners seek density of a particular kind: that green space. The contrast appears in figure 3, which promotes productivity and livability, which juxtaposes Dharavi, a slum neighborhood along with sustainability. Higher densities of Mumbai, with midtown Manhattan, New contribute to productivity if they make it York. Both have similar density—around 60,000 easier for households—as consumers and as people per km2—but midtown Manhattan has workers—to connect with firms that serve and about 8 times more floor space per person employ them. Higher densities also contribute than Dharavi. Figure 3 Two forms of urban residential density: Crowding in Dharavi, tall buildings in Manhattan Dharavi, Mumbai Average household income Approximately $ 812 62,000 people per km2 Total floor space 96,000 m2 Midtown East, New York Median household income $ 137,130 58,000 people per km2 Total floor space 760,000 m2 Source: Upper image from A. Savin (https://w.wiki/p3), distributed under a copyleft license; lower image from Google, “Streetview,” digital images, Google Maps (http://maps.google.com), photograph of 18 Sutton Place South, New York, taken May 2019. Total floor space was calculated by authors based on WSF-3D data. Note: The total floor space numbers refer to the floor space built on the inset square section of 90,000 m2 (300 m x 300 m). 6 Pancakes to Pyramids | City Form to Promote Sustainable Growth Each form of residential density—crowding and of cities today often favors a preference for vertical layering—comes at a cost: car ownership, it does not need to: providing lower-emission transport options can shape • Crowding is common where cities are mobility preferences (Mattauch, Hepburn, growing at low incomes, as most are now and Stern 2018; Weinberger and Goetzke doing in Africa and South Asia. In the 2010). Furthermore, the dense social and average Sub-Saharan African city, 60 economic interactions that occur in cities can percent of the population lives in slums—a amplify an emergent cultural preference for much larger share than the 34 percent sustainability, supporting and complementing average in cities around the world (Lall, institutional transitions to less carbon-intensive Henderson, and Venables 2017; United development (Nyborg et al. 2016). Nations 2015)—and the lack of residential floor space takes a severe toll on livability. While rising productivity and incomes are thus • Vertical layering, while creating more the necessary foundations of pyramids— floor space per unit of land area and thus thriving, livable cities that achieve density allowing more livable densities, is more through vertical layering and increased floor expensive than single-story construction. space—the benefits of pyramidal growth Because building tall requires better extend to an enhanced potential for city life technology, it depends on larger capital that respects planetary boundaries. Dense investments by land developers, who connections among people, in cities that require an expectation of high returns. abound in both economic and social capital, Such expectations depend on prior public can accelerate a “social tipping” toward investments in transport infrastructure sustainability values and commitments, as and public services, as well as on neighbors and peers mutually reinforce each the economic outlook and business other’s awareness of climate change and the environment. need to step up both privately and publicly. Vertical layering and infill development, like The trade-off for crowded cities and reduced retrofitting, can make cities more energy floor space is thus its cost in quality of life. The efficient and reduce overall resource trade-off for vertical layering and increased consumption. And, over time, the freedom of floor space is the quantity of private and public pyramidal cities to preserve green space even capital that is needed to build tall, and that will as densities rise, and to promote walking and not be present unless the city’s productivity biking as primary urban travel modalities, will and institutional capacity are sufficient to reduce total greenhouse gas emissions while create an enabling business environment. Only adding to livability. Where economic density cities that attract and nurture highly productive and population density allow, investing in firms—firms that benefit from agglomeration public space, sidewalks, and bicycle lanes economies—are able to attract the private (as opposed to car roads and parking lots or capital needed for vertical layering and structures) is a comparatively efficient use of pyramidal growth. public funds, at a time when municipal revenues are under severe pressure from the economic Pyramid cities, if managed well, can provide impacts of the Covid-19 crisis and its aftermath. not only the residential density but also the financial means and the political will to support sustainable urban development. The economic density and productivity of pyramids can generate enough capital for expensive public and private investments in shifts to low carbon technology. Although the built environment Pancakes to Pyramids | City Form to Promote Sustainable Growth 7 Five stylized facts If agglomeration dynamics are not present, regulations, are factors that can militate cities normally will not evolve from pancakes against taller structures, economic into pyramids, because their lower incomes and density, and pyramidal growth. productivity will keep them flat. The best-laid urban plans, the most thoughtful zoning and 5. Improved transport technology enables height regulations, even costly investments in economic concentration in urban cores, mass transit: none of these will create livable, supports cities’ economic and spatial sustainable densities without well-functioning growth, and increases the demand for markets, economic growth, and increasing livable residential floor space. demand for floor space near the urban core. To this extent, economy is destiny. STYLIZED FACT 1   The growth of urban built-up area worldwide is not as large as conventional Yet economic density, rising productivity, wisdom suggests and rising incomes are not sufficient. Over the quarter century between 1990 and Without appropriate and early infrastructure 2015, the urban built-up area worldwide grew investments, enabled by institutional capacity by 30 percent—or 66,000 km2, the size of and leadership, poor pancake cities can grow the island nation of Sri Lanka—through both richer and yet remain pancakes. Moreover, horizontal spread and infill. In developing even cities that become pyramids will never countries, total urban built-up area increased entirely escape the economic drivers of by 34 percent.3 Significant, to be sure. But not horizontal spread. Rich cities continue to grow quite the explosive and rapacious expansion out at the edges. How far a city shifts toward estimated in many recent studies. For example, pyramidal expansion as it becomes richer is the Atlas of Urban Expansion (Angel et al. 2016) a relative question: the answer is dictated by argued that in this period the total area of cities circumstances, some historical or accidental, in less developed countries grew by a factor of others a result of planning and coordination. 3.5. The authors added that if this (overstated) rate were to continue through 2050, the total From this report’s new empirical and analytic territory occupied by cities globally would be work on the drivers of urban spatial evolution, large enough to cover all of India. five stylized facts emerge: 1. The growth of urban built-up area Globally, between 1990 and 2015, an outsized worldwide is not as large as conventional share of total urban built-up area growth was wisdom suggests. concentrated in high-income and upper-middle- income countries. In 1990, cities in high-income 2. In developing country cities, spatial countries accounted for 48 percent of global expansion is happening mostly through urban built-up area. These same rich country horizontal spread. cities then contributed 29.5 percent of the world’s growth in built-up area. More striking is 3. Increasing incomes are the one the rapid expansion of urban built-up area in indispensable driver of vertical layering, upper-middle-income countries: these countries because building tall is capital intensive. contained one-third of the world’s urban built-up 4. Dysfunctional urban land markets, along area in 1990, but they contributed 44 percent with zoning and restrictive building of its expansion between 1990 and 2015. 3 Forecasts by Jones et al. (forthcoming), prepared for the European Commission using the Degree of Urbanization approach, predict that total urban land area will grow 29 percent between 2015 and 2050—in line with the spatial growth rates we measure here for 1990–2015. 8 Pancakes to Pyramids | City Form to Promote Sustainable Growth STYLIZED FACT 2  In developing country STYLIZED FACT 3   Increasing incomes are the cities, spatial expansion is happening mostly one indispensable driver of vertical layering, through horizontal spread because building tall is capital intensive Built-up area can grow at a city’s extensive A city that grows in population, but not margins as the city expands outward through productivity and incomes, will not generate what this report calls horizontal spread. enough economic demand for new floor space But built-up area can also expand inside for its spatial expansion to keep pace with the margins through what economists call population growth. For example, if population “intensive margin development”—here termed grows by 100 percent but incomes stay infill development. constant, the city’s total floor space increases by 60 percent. This 60 percent increase is too While most cities grow through a combination small to allow a newly doubled population the of horizontal spread and infill, the relative same amount of floor space per person as prominence of each type of expansion before: each inhabitant’s residential and work changes with successive stages of economic space will shrink, eventually making the city development. It also reflects changes in less livable. Our econometric results show that: construction technologies, preferences, and • If a city’s population doubles but incomes local government priorities. Cities in poorer stay constant, the city’s floor space per countries tend to expand more horizontally person declines by 40 percent. than cities in rich countries—an indication that rich country cities increase their total urban • If per capita income doubles but population floor space through vertical layering, as well as stays constant, the city’s total floor space along the other two margins. per person increases by 29 percent. In low-income and lower-middle-income Increasing incomes are a uniquely necessary countries, 90 percent of urban built-up area condition for a rise in floor space per person expansion occurs as horizontal spread. through vertical layering and pyramidal Nevertheless, there is a silver lining: in high- growth: the reason is that building tall is capital income and upper-middle-income country intensive. Whether developers are putting up cities, a larger share of new built-up area high-rise office buildings or substituting formal is provided through infill development. For apartment blocks for informal slum dwellings, example, a city in a high-income country that the needed investment is one that can be increases its built-up area by 100 m2 will add made only with adequate incomes, capital about 35 m2 through infill development and wealth, and financial institutions. Even if the 65 m2 through horizontal spread. But a similar levelized lifetime costs of different building city in a low-income country will add about 90 types proved similar—an unlikely assumption— m2 through horizontal spread and only 10 m2 meeting the upfront capital costs of load- from infill. These findings are consistent with bearing structures would still be more feasible our intuition that agglomeration economies, in cities with higher incomes and productivity. incomes, and supply capabilities all improve with a country’s transition to upper-middle- Higher incomes further enable vertical layering income status. by reducing the marginal cost of building tall, as opposed to building single-story structures, in cities with sufficient market demand for land and floor space. In a productive urban agglomeration, land is the scarcest of resources, and it becomes still more valuable as it is connected to infrastructure. Any structure built with substantial materials Pancakes to Pyramids | City Form to Promote Sustainable Growth 9 needs sound foundations and installed water, habitation to higher-density apartments sewerage, and electricity. If the fixed costs of or to build new commercial structures in initial investments in land, foundations, and economically dense clusters. Formal land infrastructure account for a large share of a transactions are long, costly, and complicated building’s overall cost, then adding stories will (World Bank 2015c). These land market reduce the unit cost of floor space: building constraints reduce the collateral value of upper stories multiplies floor space by a larger structures and give developers little incentive factor than it increases overall costs. to build tall, while they also tempt all parties to do business informally (Lall, Henderson, and Vertical layering and infill development slow Venables 2017). horizontal spread, but they do not stop it. Even as richer and more productive cities add floor Land management is essential, especially space by reaching upward, they continue to for emerging and lower income cities: their expand outward. Workers are more likely in potential for future pyramidal growth hinges rich cities than in poor cities to live in a less largely on their ability to plan for productive residentially dense suburb while commuting to economic agglomerations at early development an economically dense center. Pyramid cities stages, both at the city’s edge and in areas thus require economic growth, not only to of new expansion. Building regulations, supply the capital needed for vertical layering, such as floor-area restrictions, can be but also to fund local coffers and make the economically counterproductive by limiting large public transport investments that support density (residential and commercial) and by concentrated hubs of economic activity and lengthening commutes. Zoning restrictions longer residence–workplace commutes. can be damaging if they lock in patterns of land use that become inefficient as a city STYLIZED FACT 4   Dysfunctional urban land develops. Appropriate zoning, however, can markets, along with zoning and restrictive reduce negative externalities that affect building regulations, are factors that can households and can encourage commercial militate against taller structures, economic concentrations that promote positive spillovers density, and pyramidal growth and externalities among firms. While low incomes may constrain the ability to In developing-country cities, current urban develop urban land, they are not the only such plans and planning institutions are often constraint: a city’s spatial evolution can also ineffective—neither coordinating market-driven reflect numerous impediments related to policies investment in structures nor managing the and weak institutions. Formal institutions for spatial form of cities for efficiency, livability, titling and for property transfer tend to function and sustainability (figure 4). One challenge is more smoothly and predictably in countries with the inappropriate adoption of building codes higher levels of human capital and government and planning models through inheritance from effectiveness (La Porta et al. 1999; Glaeser et former colonial regimes or importation from al. 2004). In developing country cities, a lack of richer countries (Goodfellow 2013). Another legal clarity in land tenure often deters investors challenge is that plans lack credible accounts from sinking capital into formal structures and of finance, market dynamics, and distributional contributes to the persistence of slum areas. impacts. Take minimum lot sizes: though Many of these cities struggle with overlapping they may be intended as pro-poor land use and sometimes conflicting property rights regulations, in practice they limit households’ systems—formal, customary, and informal—that investment choices. In Brazil, they appear together pose barriers to urban land access, associated with slum formation (Lall, Wang, and to plot consolidation, and to evolution in land Da Mata 2007). Still other challenges arise from use. Developers cannot readily buy downtown capacity and resource constraints. land, whether to convert it from low-density 10 Pancakes to Pyramids | City Form to Promote Sustainable Growth Weak zoning and land use plans in cities of low and lower-middle-income countries Figure 4 Weak zoning and land use plans in cities of low- and lower-middle-income countries Yes Partial No/not enforced 100 80 Share of cities (%) 60 40 20 0 Low income Lower middle Upper middle High income income income Source: Authors’ analysis, based on Regulatory Survey data by Angel et al. (2016) retrieved from http://atlasofurbanexpansion.org/. STYLIZED FACT 5   Improved transport workplace from residence—promoting economic technology enables economic concentration density in the center and supporting the city’s in urban cores, supports cities’ economic and spatial expansion (Heblich, Redding, and Sturm spatial growth, and boosts demand for livable 2020). Later, the mass production of cars residential floor space dramatically reshaped urban density and living space. Future investments in transport policies Many developing country cities today struggle and infrastructure will similarly affect a city’s with the high road congestion and commuting ability to grow, lower its pollution levels, and costs—in time and money—that result from reduce its carbon footprint. poor transport infrastructure and limited public transit options. Such congestion impedes Which transport technologies are best? the separation of residence from workplace, The answer depends on circumstances. limiting cities’ spatial expansion along with their Underground rail (metro or subway) economic growth and productivity. According construction is capital intensive, and its to the 2019 Tom-tom Index, the most extreme success will reflect a city’s ability to achieve urban congestion occurs largely in developing economic and population density. Recently, countries: those countries represented on bus rapid transit (BRT) systems have rapidly the list of the world’s 10 most congested gained prominence as an alternative to cities, including Colómbia, India, Peru, and the subways in developing country cities—yet Philippines. Transportation in such cities is BRT systems have also encountered often informal and chaotic. Few have been able challenges. While early transport infrastructure to afford large mass transit interventions. investments can be critical for a city’s potential evolution into a dense pyramid, the Past transport technologies have shaped cities’ identification of appropriate modalities and the economic and spatial growth. For example, planning of networks requires extensive and in London in the 19th century, the revolution evidence-based study of a city’s particular in steam railways slashed travel times and geographic, economic, and social contexts. allowed the first large-scale separation of Pancakes to Pyramids | City Form to Promote Sustainable Growth 11 Policy implications incomes—at earlier stages of economic development—to develop physically as Governing both the economic and the pancakes. As incomes grow, cities start institutional drivers of urban spatial form are growing horizontally, building low and policies and decision-making processes, which extending outward. Only at later stages, with combine with fundamental forces to affect a higher incomes and productivity, can they city’s pattern of spatial development—and, in hope to evolve into pyramids. The spatial particular, its potential to evolve from earlier development of cities is thus closely tied to pancake growth to later pyramidal growth. their economic transformation: a city’s shape Urban plans and policies influence the tends to track its per capita GDP, from low- to pace of new construction to meet demand, lower-middle- to upper-middle-income. To the height of the resulting structures, the enable these transformations, there is urgent contours of the city’s skyline, the clustering need to coordinate land management with of its firms into hubs of productive activity, infrastructure, natural resources, and hazard the pattern of its residential densities, the risk to reflect market need and societal distance from residences to workplaces, and preferences. the extent to which commutes are costly, time consuming, and carbon-intensive. The An important consideration is that urban city’s livability and sustainability reflect the development decisions are long-lived. Because effects of policy on land transfer, on the ease the economic system reorganizes itself around of formal development, on patterns of formal infrastructure and urban plans, and because and informal settlement, on pollution, on so much of current growth is in cities, this transportation modalities, on the availability of inertia can extend over centuries. A delay in amenities, and on floor space per person. greening city investments may therefore prove extremely costly if it locks in technologies that From the five stylized facts that emerge turn out to no longer be appropriate (because from this report’s analysis, four main policy of their excessive carbon, land, or water implications can be drawn—implications that intensity) or settlement patterns that prove especially concern leaders in developing vulnerable to changing climatic conditions. countries with emerging cities: Second, city shape is driven by economic First, setting the stage for pyramidal growth fundamentals. The forces that raise pyramids requires planning for expansion along all with peaked skylines, signaling high demand margins—not only the vertical layering and for floor space in the central business district, infill development that will be enabled by are those of economic agglomeration. They future productivity, but also the horizontal are the same forces that drive tradables spread that will occur at the city’s edge production and productive job creation. and that requires urban planning to be The hope of “leapfrogging” urban spatial strengthened and land to be laid out for development stages—turning pancakes into development. City leaders need to plan for economically dense pyramids despite weak spatial development along all three margins. demand for downtown space, and in the The analysis in this report shows that context of factor markets that are currently economic drivers induce cities with lower inefficient or sluggish—is likely quixotic. 12 Pancakes to Pyramids | City Form to Promote Sustainable Growth Third, allow for redevelopment in the future. development and neighborhood regeneration As cities develop economically, their (box 2). management and financing capabilities improve. Regulations need to be adaptable Fourth, Infrastructure matters. A principal to changing demand and supply conditions, factor in setting a city’s growth path is its plan anticipating the emergence of new uses for for networked infrastructure, including roads scarce urban land. When Seoul changed its and transit arteries between the center and density regulations to increase development periphery. Some capital investments are best potential and density inside its greenbelt, the to plan for—or even finance and execute—in change spurred a sharp rise in redevelopment the earlier spatial development stages, when and revitalized the city (figure 5). Also cities stand to benefit disproportionately important here is to conserve irreplaceable over the life of the investment. If a smaller, cultural and natural amenities—these have low-income city spreads quickly into a not only intangible permanent value, but also larger pancake with no effective plan, its a unique potential to create economic value opportunities to reap future returns from large and attract investment at later stages of urban infrastructure investments may fade. Figure 5 Flexible rules enabled Seoul’s infill redevelopment Sejong-Ro, 1962 Sejong-Ro, 2015 Source: Left image from Korea Research Institute for Human Settlements (KRIHS); right image from Seoul Research Data Service, http://data. si.re.kr/seoulphoto. Pancakes to Pyramids | City Form to Promote Sustainable Growth 13 These policy implications point to three 2. Evaluate infrastructure investments. To recommendations for city leaders and decision support the market forces that drive urban makers. economic agglomeration, productive job creation, and income growth, governments 1. Strengthen institutional foundations. can give priority to policies and Although lower-income, less productive investments that coordinate infrastructure cities cannot be molded into pyramids investment with land management today, decision makers can set the under forward-looking plans. In addition, stage for future transformative growth governments can provide public goods and by establishing the right institutional amenities that directly enhance livability. environment. Because economic and The new urban model used in this report physical evolution go hand in hand, cities and its annexes (Sturm et al. 2021a) need integrated legal and regulatory offers city leaders economic intelligence, reforms and frameworks that will enable helping them to assess the likely impacts both economic and spatial development. of large infrastructure investments as well Especially important are steps to as regulatory approaches. The new urban strengthen land markets and urban model also allows leaders to pinpoint planning institutions. challenges and risks that call for careful assessment and management. Box 2 Cultural and natural amenities are important over the course of a city’s growth and development Cultural and natural amenities • Heritage investment. In regeneration project used make cities vibrant and attractive Ireland, Dublin’s Talent Hub PPPs to engage multiple places, enhancing their quality leverages cultural assets stakeholders in its strategic of life (UNESCO 2016; Tweed and to promote knowledge plan, formed by government Sutherland 2007). As developing industries. By undertaking agencies working with private countries face rapid urbanization, major urban regeneration and developers. they must not assume that growth historic preservation projects • Financing through taxation comes at the cost of these unique in central Dublin over the past coupled to incentives. assets (World Bank 2012). Far from 30 years, the government Emerging cities often lack an inevitable trade-off, cultural and helped the city attract the infrastructure and public natural amenities make cities and young and creative classes, services needed for urban regions competitive and innovative boosting the local economy. regeneration. The Akaretler (Alberti and Giusti 2012). Cultural • Public–private partnerships Row Houses project used and natural assets also matter for (PPPs). In developing land value capture (LVC) to sustainability in urban development, countries, the public sector recover the cost of urban according to the new Habitat III by itself is likely to lack redevelopment in Istanbul. urban agenda (UN 2016). both the capacity and the Through LVC, the municipality Caring for cultural and natural funds for urban regeneration collected locally generated amenities thus needs to be projects that preserve tax revenue from project complementary to cities’ economic cultural and natural assets. areas to fund infrastructure development, with a central place in Moreover, such projects improvements—creating urban plans. Here infill development generally depend on private a positive feedback loop. and redevelopment are especially sector stakeholders— The project succeeded in relevant. To balance cultural and among others—to identify maintaining sites’ cultural natural conservation with urban economically viable new values while promoting local economic development, instruments uses for existing areas and businesses and creating include (World Bank 2012): structures. In Istanbul, Turkey, employment. the Akaretler Row Houses 14 Pancakes to Pyramids | City Form to Promote Sustainable Growth 3. Mobilize finance for durable urban transactions, and they should streamline investment. By a conservative estimate, intergovernmental fiscal and institutional worldwide urban infrastructure financing frameworks. In addition, leaders should needs amount to $4.1–4.5 trillion annually keep in mind that while cheap financing (CCFLA 2015). Climate mitigation costs can raise money for investment, repaying are likely to add another $0.4–1.0 trillion it requires predictable revenue. For this annually—and adaptation costs add a purpose, strong economic fundamentals further $120 billion annually, which could based on agglomeration economies will become much greater if the average be vitally important. global temperature increases by more than 2°C above preindustrial levels. Pyramidal growth requires intensive investments in both structures and infrastructure. The main difficulty for city leaders is the lumpiness of investment, as the initial capital layout for any large investment—in transport, water provision, solid waste management, or sewage removal and treatment—is likely to far exceed the budget of any city government. Leaders must therefore identify financing solutions that enable the city to anticipate and meet its future needs. City leaders should clarify regulatory frameworks for municipal borrowing, public–private partnerships (PPPs), and land value capture Photo: iStock, Akaretler Row, Istanbul, Turkey Pancakes to Pyramids | City Form to Promote Sustainable Growth 15 Photo: World Bank/Flickr, Bogota, Colombia I How are cities accommodating families and firms—and how fast is global built-up area expanding? The spread of urban built-up area is slower than has been thought – Page 24 Who is contributing the most to urban built-up expansion? – Page 25 SPOTLIGHT 1 Stages of urban economic development can be seen in cities’ average building heights—and in their skylines – Page 28 16 Pancakes to Pyramids | City Form to Promote Sustainable Growth How are cities accommodating families and firms—and how fast is global built-up area expanding? Chapter 1 provides empirical evidence on Beginning with data showing that global the pace and patterns of urban spatial urban built-up area has been expanding more development, drawing on new observational slowly than is sometimes claimed, chapter data for city growth worldwide since 1990 1 continues with an account of how various (figure 1.1, box 1.1, box 1.2). From these data, income groups, regions, and countries have new facts emerge on: contributed to this global expansion of cities. • The extent to which growing cities expand their built-up area and their height. • The implications of this spatial evolution for the amount of floor space available (box 1.3). • The drivers of changing residential densities over time—both the ratio of population to built-up area, and the ratio of population to floor space. Figure 1.1 Coverage of cities studied in this report Population Less than 100,000 Less than 250,000 Less than 500,000 Less than 1 million Less than 5 million More than 5 million Source: GHS–Urban Centre Database. Pancakes to Pyramids | City Form to Promote Sustainable Growth 17 Box 1.1 What is a city? To overcome the vagaries in and settlements globally (Dijkstra limited (with relatively low shares of how countries define towns and et al. 2020). In contrast, other global impervious surface). cities, this report uses a recently urban mapping exercises have limited The report defines an urban center devised methodology—the Degree city samples (for example, Angel et as “a population of at least 50,000 of Urbanization—which allows al. 2011). Because the Degree of inhabitants in contiguous dense grid for international comparisons of Urbanization measure is based not cells with population densities greater urban growth. Endorsed by the only on built-up areas but also on than 1,500 inhabitants per square UN Statistical Commission for this population density, it may be better kilometer” (Florczyk et al., 2019). Box purpose, the Degree of Urbanization suited to capture rapidly transitioning table 1 reports the distribution of cities was developed by the European areas in developing country cities. across world regions and city sizes. Commission in collaboration with five Other exercises often fail to identify international organizations. A strength these transitional areas because, of this measure is that it can be despite a huge population influx, the Does the density threshold of 1,500 consistently applied to identify cities areas’ built environments may remain people per square km underestimate Box table 1 Distribution of cities analyzed in this report, by city population and region Less than 70,000– 100,000– 200,000– 500,000– 1 1 million– More than 10 Region 70,000 100,000 200,000 500,000 million 10 million million East Asia & Pacific 612 609 819 476 131 116 11 Europe & Central Asia 345 297 324 225 77 56 2 Latin America & Caribbean 267 228 255 158 60 51 3 Middle East & North Africa 221 137 199 113 43 40 2 North America 77 86 92 60 22 33 2 South Asia 312 362 588 390 122 84 7 Sub-Saharan Africa 300 290 439 217 55 52 1 Source: Authors’ analysis, based on GHS–Urban Centre Database. Box figure 1 Population estimates with additional buffered urban areas are similar to the Box figure 1 population within the urban centers boundary 50 40 Population (million) 30 20 10 0 -5km -4km -3km -2km -1km Urban +1km +2km +3km +4km +5km boundary Source: Authors’ analysis, based on GHS–Urban Centre Database. 18 Pancakes to Pyramids | City Form to Promote Sustainable Growth the size of cities? No—this density 50 fastest-growing cities from 1990 (3,205 square km) while increasing threshold is robust. Because the to 2015 were selected to assess its population by only 4.6 percent. arbitrary selection of the density population variations across different This report also uses data from the threshold (1,500 people per square sizes of urban areas. Multiple ring German Aerospace Center (DLR) km) could introduce bias in estimating buffers (from −5 km to +5 km from on cities’ height based on its World the size of cities, two additional the edge of the urban boundary) Settlement Footprint 3D product correction methods were used to were added to the urban boundary. (WSF-3D; see box 1.2). Height data supplement the density threshold As shown in box figure 1, except are available for 400 cities—and the and reduce bias: adding pixels with for two outliers—Guangzhou, China sample is chosen to be representative built-up coverage over 50 percent to and Jakarta, Indonesia—the median of the Global Human Settlement urban centers, and gap-filling based on values of population estimates with Layer database, which covers all neighboring pixels (Dijkstra et al. 2020). additional buffered urban areas are cities in the world. The data on city Based on these combined methods, similar to the population within the heights add to our understanding the cities defined by Degree of urban boundary. For instance, the of complexity and layering in city Urbanization are well aligned (91.9 additional 5 km buffer increases development. Though these data are percent) with nationally defined Mexico City’s urban area by 74 not the final word, they are a good cities having more than 300,000 percent (1,556 square km) while step toward understanding urban population globally (Dijkstra et al. increasing its population by only spatial evolution. Annex 2 provides 2020). In a further sensitivity test of 3.3 percent, and it increases Los details on the data and measurement. different urban boundary sizes, the Angeles’s urban area by 57 percent Source: Authors. Box 1.2 Height data on cities everywhere: The World Settlement Footprint 3D (WSF-3D) The World Settlement Footprint 3D mission. in Marconcini et al. (2020) and Esch (WSF 3D), developed by the German • 10 m multispectral Sentinel-2 et al. (2020, 2021). Aerospace Center (DLR), is a three- imagery. For the present study, the 30 m ALOS dimensional model of the built • The human settlement mask World 3D elevation model (AW3D30) environment worldwide. The WSF of the World Settlement Foot- was used as the standard input to 3D estimates building area, height, print 2015 (WSF 2015). estimate built-up height, because and density at an aggregated 90 m • Where available, vector data German data distribution regulations spatial resolution. It uses extensive for building location—for exam- (SatDSiG) still require a specific machine learning to jointly analyze ple, data from the Open Street clearance for the use of TanDEM-X data of four kinds: Map initiative. height information. • 12 m digital elevation and The WSF 2015 and WSF 3D radar intensity data from the methodologies are described in detail German TanDEM-X satellite Source: WSF-3D data of “Building Fraction” and “Building Height” derived for the city of Santiago, Chile. Pancakes to Pyramids | City Form to Promote Sustainable Growth 19 Box 1.3 Seeking livable urban densities around the world: From population density to per capita floor space A novel feature of Pancakes to use this conventional measure The reason for this seeming Pyramids is its use of global data of population density advisedly, anomaly in box figure 1 is that on building heights. These data are because density as measured Hong Kong’s total built-up area is produced by German Aerospace by built-up area does not in itself limited by its island geography, Center (DLR), using their World make cities livable. To get at pushing its population density— Settlement Footprint 3D product livability, we add a measure of per defined as the ratio of people to (WSF-3D; see box 1.2). capita floor space. built-up area—extremely high. More than most high-income cities, Hong Why seek to measure and To appreciate the practical Kong has had to build upward and understand cities’ height profiles? difference between measuring a inward rather than outward. And Because these profiles indicate city’s population density—relative yet, as box figure 2 shows, Hong vertical layering. And layering to built-up area—and its per capita Kong’s per capita floor space is in is key to providing floor space floor space, compare box figures 1 line with that of a high-income city together with urban density— and 2. Each figure plots population such as London or Tokyo. the combination that reshapes per square kilometer on the vertical pancakes as pyramids, enabling axis, per capita income on the Kinshasa and Karachi lack as much cities to be not only dense with horizontal axis. But in box figure 1, per capita floor space as Hong people, but also thriving, livable, the unit of urban area used for the Kong because their population and sustainable. vertical axis is built-up area. In box densities (relative to built-up area) figure 2, it is floor space. are achieved through crowding— Without accurate estimates not vertical layering. If it is critical of height data, we could not Both figures reveal similar general to avoid confusing the economic confidently compare floor space patterns: cities in richer countries geography of crowded pancakes across cities. Nor could we track its tend to be less dense relative to like Kinshasa and Karachi with distribution within the geography of built-up area, and also to allow that of an efficient pyramid like a single city. Following convention, more per capita floor space. But in Hong Kong, measuring cities’ floor this report defines urban box figure 1, Hong Kong seems at space–adjusted density makes population density as the ratio of first glance to be an exception to more sense. And that can be done population to horizontal built-up this pattern. Its ratio of population only by measuring height. area—not floor space (which can to built-up area resembles that of also be added vertically). But we Kinshasa or Karachi. 20 Pancakes to Pyramids | City Form to Promote Sustainable Growth Box figure 1 Urban population density, 2015 Box 1.3 Fig 1 Log of density (population/km2 of built-up space) 1,000,000 100,000 Mumbai Kinshasa Karachi Hong Kong Dhaka Lagos Beijing 10,000 Shanghai Tokyo London Dallas 1,000 100 1,000 10,000 100,000 Log of city income, 2015 ($) Source: Authors’ analysis, based on GHS–Urban Centre Database. Box figure 2 Floor space–adjusted density, 2015 Box 1.3 Fig 2 100,000 Log of density (population/km2 of floor space) Karachi Kinshasa Mumbai 10,000 Dhaka London Lagos Hong Kong Beijing Tokyo Shanghai 1,000 100 1,000 10,000 100,000 Log of city income, 2015 ($) Source: Authors’ analysis, based on WSF-3D data. Pancakes to Pyramids | City Form to Promote Sustainable Growth 21 Motivating this and the following chapters is a market demand for land and floor space keeps basic contrast between two idealized physical land prices in pancake cities low and keeps city types resulting from opposed patterns structures close to the ground, especially at of spatial growth: some cities are pancakes, the city’s edge. The slow rate of spatial growth and others are pyramids. Generally, cities at may lead the city to accommodate population low productivity and income levels grow as growth through crowding—a process starkly pancakes: flat and expanding slowly, chiefly visible in the slums of developing country through horizontal spread (figure 1.2). Low cities. Figure 1.2 The growth of a city along three margins—horizontal spread, infill development, and vertical layering Built-up area Defined by the presence of buildings Urban area Cities grow along three margins Horizontal spread Infill development Vertical layering Source: Authors’ depiction. In contrast to cities at low levels of city edge, and vertical layering, or the building productivity, which have no choice but to of taller structures around centers of economic grow as pancakes, cities that attain higher activity (see figure 1.2). As pancakes turn into productivity may evolve from pancakes into pyramids, their skylines grow more sharply pyramids. Their pancake expansion goes on, peaked—and they accommodate population yet it is accompanied by growth along two growth, not by crowding, but by adding total other margins: infill development, or the filling of floor space throughout the city and providing vacant and underused land at and within the ample floor space per person (box 1.4). 22 Pancakes to Pyramids | City Form to Promote Sustainable Growth While pancake expansion and pyramidal used to model and predict the effects of plans expansion are physical manifestations of city and policies on urban spatial growth, density, development, the drivers of that development productivity, livability, and sustainability. The are largely economic and institutional. Later sections in this chapter set the scene by chapters in this report identify and explain describing key patterns that emerge from the these drivers, indicating how they can be global satellite data. Box 1.4 The benefits—and costs—of urban density Economists and planners working and Xiong 2017). This is true of is about 0.1. Measuring the elasticity on issues of city development certain amenities in rich countries, of house prices with respect to often showcase urban population too. Across lower and higher income density—accomplished simply by density—that is, a high number of countries, both the share of people conditioning out the city’s total residents per land area unit—and with access to the Internet and land area—results in a figure of 0.3. focus on pursuing its advantages. the share with bank accounts are With housing constituting a third of Two compelling cases for urban highest in cities and decline at lower household expenditure in large cities, population density appear in Edward densities. the cost of living in a city that is 10 Glaeser’s Triumph of the City and percent denser is about 1 percent Yet urban population density— Serge Salat’s Cities and Forms. A higher. But these costs appear to considered in itself as the ratio of recent survey of empirical work become manageable if cities both population to built-up area, without by Gabriel Ahlfedt and Elisabetta expand horizontally and add floor reference to floor space consumption Pietrostefani shows that urban space (Duranton and Puga 2020). or other measures of livability—can population density is associated also have large drawbacks. We have Developing country cities are where with many benefits: doubling urban seen that a city with a growing the demons of density are most density is associated with premiums population but stagnant incomes apparent. The relatively flat and in wages (4 percent), patent activity will become increasingly crowded, crowded centers of these cities (19 percent), consumption variety as floor space fails to keep up with often afford only tiny amounts value (12 percent), the preservation population growth. And even as more of floor space per person. The of green spaces (23 percent), people compete for the same floor dynamics of richer cities are different: and the use of non-car transport space, land and housing prices go vertical layering adds floor space modes (7 percent), as well as with up. Other things, too, become more on buildings’ upper stories, even as reductions in average vehicle mileage expensive: with rising population and continuing horizontal spread also (–8.5 percent), energy consumption residential density come rapidly rising adds to the average floor space (–7 percent), pollution concentration urban transport costs. And then available for each resident. (–8 percent), crime (–8.5 percent), there is traffic congestion. Unlike the and unit costs of providing local The lesson is that population density benefits of density, these costs or public services (–14.4 percent) (the ratio of population to built-up “demons of density” (Glaeser 2011) (Ahlfeldt and Pietrostefani 2019). area) is not an end in itself—certainly have received comparatively little it is not a sufficient measure of either In developing countries, from empirical research attention. livability or sustainability. Cities India and Vietnam to Brazil and Examining the costs of population should also afford decent quantities Colombia, larger cities with denser density, Duranton and Puga (2020) of per capita floor space, as well as populations mostly have better living showed that a typical elasticity of amenities, efficient transit systems, standards than smaller and less the price of housing at the center of and controls on harmful externalities densely populated ones (Glaeser a city with respect to city population of urban growth. Pancakes to Pyramids | City Form to Promote Sustainable Growth 23 The spread of urban built-up area is slower than has been thought How are cities evolving in space—horizontally and vertically—around the world? The satellite Box 1.5 Sustainable Development Goal 11 and data show that between 1990 and 2015 the urban built-up area expansion world’s total urban built-up area increased by 66,000 square km, or the size of Sri Lanka. Sustainable Development Goal (SDG) indicator 11.3.1— Urban built-up area increased from 226,768 the sustainable urbanization rate—is defined as the ratio square km to 294,550, a growth of 30 percent. of the land consumption rate to the population growth rate. The SDG tracker reports that data are unavailable for In developing countries, urban built-up area this indicator.1 Between 1990 and 2015, the global urban population increased by 35 percent, and the developing increased over 1990–2015 from 117,977 country urban population increased by 36 percent. In low- square km to 166,231. That is a 34 percent income countries, the urban population grew by as much as rate of built-up expansion, an annual growth of 72 percent.2 If we measure the land consumption rate as total 1.36 percent. This growth is significant, to be built-up area growth, the sustainable urbanization rate can be benchmarked as follows (box figure 1): sure, but not quite the explosive and rapacious spread estimated by many recent studies. For • Global: 0.75. 2 example, the Atlas of Urban Expansion argued • Low-income countries: 0.47 that during this period the area occupied by • Lower-middle-income countries: 0.86 cities in less developed countries increased • Upper-middle-incomes countries: 1.05 by a factor of 3.5—and that if this rate were to • High-income countries: 0.88 continue, the total amount of land taken over by urban land use would be equivalent to the entire Box figure 1 Sustainable urbanization rate: Ratio country of India by 2050 (Angel et al. 2016). of urban land consumption growth rate to urban population growth rate Box Figure 1_SUrb-rate Others have predicted recently that developing countries’ urban population will have doubled between 2000 and 2030 and built-up area will 0.75 Global have tripled. Urban land cover in Sub-Saharan Africa is expanding at the fastest rate—more Low income 0.47 than 12-fold, it is said, between 2000 and 2050 (Angel et al. 2016). Again, urban land areas Lower middle income 0.86 have been projected to expand by 0.6–1.3 million square km between 2015 and 2050, an 1.05 increase of between 78 and 171 percent over Upper middle income the urban footprint in 2015 (Seto et al. 2014). High income 0.88 This report employs urban built-up area estimates from satellite imagery at high Source: Authors’ analysis, based on GHS–Urban Centre Database. spatial resolutions to examine urban spatial Note: Box figure 1 shows that the relative growth of city built-up development (see box 1.1). These granular data area—while inherent in early urban development stages—is far suggest that horizontal expansion in growing slower in low- and lower-middle-income countries than was recently thought. cities is considerably slower than population 1. https://sdg-tracker.org/. increase—a finding with considerable bearing  sing the urban centers thresholds described in this report. Total 2. U on Sustainable Development Goal 11, “Make urban population increased from 2.2 billion to 3.1 billion from 1990 cities inclusive, safe, resilient and sustainable” to 2015; developing countries’ urban population increased from 1.7 billion to 2.5 billion. (box 1.5).  rban built-up area increase of 33 percent alongside population 3. U growth of 72 percent. 24 Pancakes to Pyramids | City Form to Promote Sustainable Growth These findings cast doubt on an often repeated commercial, and industrial structures. The account of unfettered urban spatial expansion. demand for these structures tends to rise They should not come as a surprise, however. only with economic development, and when Land area is built up mainly for residential, supported by a conducive policy environment. Who is contributing the most to urban built-up expansion? Globally, the expansion of urban built-up accounted for 48 percent of global urban built- area over 1990–2015 was disproportionately up area (108,726 square km). The same rich concentrated in high-income and upper- country cities contributed 29.5 percent of the middle-income countries. As figure 1.3 shows, world’s growth in built-up area between 1990 in 1990, cities in high-income countries and 2015. Figure 1.3 – p9 Figure 1.3 Urban built-up area expansion by national per capita income level, 1990–2015 Built-up in 1990 Built-up expansion, 1990-2015 140,000 Share of expansion: 29.5% 120,000 Share of expansion: 44.3% 100,000 Built-up area (km2) 80,000 Share of expansion: 60,000 23.4% 40,000 20,000 Share of expansion: 2.8% 0 Low income Lower middle Upper middle High income income income Source: Authors’ analysis, based on GHS–Urban Centre Database. Even more striking in figure 1.3 is the rapid As is also shown in figure 1.3, lower-middle- expansion of urban built-up area in upper- income countries contributed 23 percent to middle-income countries. Containing one-third urban built-up area expansion over 1990– of the world’s urban built-up area in 1990, 2015, while low-income countries contributed these countries contributed 44 percent to its just 2.8 percent. Nevertheless, the 2.8 percent expansion between 1990 and 2015: their rate contribution represents a 32 percent rate of of urban built-up expansion was about 1.5 expansion from the low initial amount of urban times as high as that of high-income countries. built-up area in low-income countries. Pancakes to Pyramids | City Form to Promote Sustainable Growth 25 The distribution of urban built-up area The smallest total contributions were those of expansion across regions appears in figure 1.4. Middle East and North Africa (MENA) and Sub- Over 1990–2015, East Asia and the Pacific Saharan Africa (SSA). Urban built-up area in (EAP) made a much higher total contribution EAP grew seven times more than in MENA—and urban1.4 toFigure than area Urban built-up expansion expansion any other by region, 1990–2015 region. 1.7 times more than in cities in North America. Figure 1.4 Urban built-up area expansion by region, 1990–2015 Built-up in 1990 Built-up expansion, 1990-2015 Share of 100,000 expansion: 36.3% Share of 80,000 expansion: 20.0% Built-up area (km2) Share of 60,000 expansion: 9.5% Share of 40,000 expansion: Share of Share of 10.5% Share of expansion: expansion: expansion: 10.7% 20,000 7.7% 2.8% 0 Middle East Sub- South Latin Europe & North East & North Saharan Asia America & Central America Asia & Africa Africa Caribbean Asia Pacific Source: Authors’ analysis, based on GHS–Urban Centre Database. Two countries, China and India, led globally India have been pushed to the outskirts of in expanding urban built-up area over 1990– large metropolitan areas by rising economic 2015. China alone contributed 60 percent of demand for cities—following India’s economic all urban built-up area expansion by upper- liberalization in the early 1990—in combination middle-income countries during this period. In with urban land use regulations and density absolute terms, China added more than 17,000 caps affecting metropolitan cores (World Bank square km of such area (figure 1.5). India’s 2013b; Ellis and Roberts 2016). urban built-up expansion was also impressive: starting from 9,463 square km in 1990, it had added 5,421 square km by 2015, contributing 8.2 percent of the world’s total urban built-up area expansion. While smaller in absolute terms than China’s urban expansion, India’s was more nearly comparable in its rate of growth: urban built-up area expanded by 45 percent in India over 1990–2015. People and firms in 26 Pancakes to Pyramids | City Form to Promote Sustainable Growth Figure 1.5 Urban built-up area expansion in China and India (1990-2015) Figure 1.5 Urban built-up area expansion in China and India, 1990–2015 Built-up in 1990 Built-up expansion, 1990-2000 Built-up expansion, 2000-2015 50,000 8,243 km2 40,000 8,960 km2 Built-up area (km2) 30,000 20,000 31,882 km2 1,892 km2 10,000 3,520 km2 9,463 km2 0 China India Source: Authors’ analysis, based on GHS–Urban Centre Database. Photo; Li Yang/Unsplash Pancakes to Pyramids | City Form to Promote Sustainable Growth 27 Photo: World Bank/Flickr, Accra, Ghana Photo: Eart/Unsplash, Paris, France SPOTLIGHT 1 Stages of urban economic development can be seen in cities’ average building heights—and in their skylines Richer cities are taller. This simple, intuitive truth about the impact of economic development on building height can be broadly confirmed by comparing six cities at different income levels and urban structures. The two poorer cities—Accra (low income) and Lagos (lower middle income)— are built very low on average, with few buildings taller than 10 m, and almost none above 20 m (figures S1.1 and S1.2). These poorer cities also lack central clusters of higher structures to make their hubs visually identifiable. In contrast, the two richer cities—Mexico City (upper middle income) and Paris (high income)—are built at far greater average heights, with cores towering over surrounding areas (figures S1.3 and S1.4). 28 Pancakes to Pyramids | City Form to Promote Sustainable Growth Figure S1.1 Average building height in Accra Figure S1.2 Average building height in Lagos income) (lowAverage building height in Accra (lower middle income) Average building height in Lagos (lower middle income) 50 50 40 40 Average Height (m) Average Height (m) 30 30 20 20 10 10 0 0 0 1 2 3 4 5 0 1 2 3 4 5 Distance from central business district (km) Distance from central business district (km) Figure S1.3 Average building height in Figure S1.4 Average building height in Paris Mexico City Average (upper building middle height in Mexico income) City (upper middle income) (high income) Average building height in Paris (high income) 50 50 40 40 Average Height (m) Average Height (m) 30 30 20 20 10 10 0 0 0 1 2 3 4 5 0 1 2 3 4 5 Distance from central business district (km) Distance from central business district (km) Source: Authors’ construction, based on WSF-3D data. The four cities at different income levels in Are cities in land-rich countries, such as the figures S1.1–S1.4 suggest a general pattern: United States and Australia, too sprawling to a pyramidal height distribution sloping down display pyramidal growth? No: the data show from the city center. While the degree of that, to some extent, even hugely expansive concentration differs across cities’ income and high-income cities are pyramid cities. In structure, the central area tends to be taller Houston (figure S1.5) and Sydney (figure S1.6), than other areas. Yet this general tendency building heights are highest at the center, does not mean that all city centers benefit declining continuously with distance. from economic concentration. In low-income cities with high residential concentrations and lower building height, the more central neighborhoods are likely to feel more crowded— compared with surrounding areas. Pancakes to Pyramids | City Form to Promote Sustainable Growth 29 Figure S1.5 Average building height in Houston (high income) Average building height in Houston (high income) 50 40 Average Height (m) 30 20 10 0 0 1 2 3 4 5 Distance from central business district (km) Source: Authors’ construction, based on WSF-3D data. Source: Houston, Texas. March 14, 2018. Image ©2018 Planet Labs, Inc. cc-by-sa 4.0. Figure S1.6 Average building height in Sydney (high income) Average building height in Sydney (high income) 50 40 Average Height (m) 30 20 10 0 0 1 2 3 4 5 Distance from central business district (km) Source: Authors’ construction, based on WSF-3D data. This report also documents a less familiar, perhaps less obvious, fact: richer cities’ skylines peak more sharply. In cities in upper-middle- and high-income countries, building heights decline more steeply from the inner urban core (table S1.1). Cities in these higher income countries have more centrally concentrated floor area than those in lower income countries—and the richer city centers stand out more visibly from less central neighborhoods. 30 Pancakes to Pyramids | City Form to Promote Sustainable Growth Table S1.1 Average building height moving away from the central business district Average building height (m) Average distance (km) Accra Lagos Mexico City Paris Houston Sydney 0–1 2.93 4.70 8.74 14.87 13.83 15.51 1–3 2.12 3.26 5.61 5.67 3.69 6.38 3–5 1.82 3.04 4.42 4.59 3.00 3.88 Source: Authors’ analysis, based on WSF-3D data.. The pyramidal height distribution patterns picture with their many outlying clusters of seen in satellite data on Houston (see figure taller buildings. These outliers likely reflect S1.5) and Sydney (see figure S1.6) show that edge-city development enabled by networks of even these sprawling cities conform to the infrastructure that encourage agglomerations general pattern associating higher productivity of tradable sectors along main arteries. What is and income with higher urban structures and clear is that, as a general rule, building height is steeper skylines. Yet on a closer look, Houston associated with economic activity. and Sydney may complicate this general Pancakes to Pyramids | City Form to Promote Sustainable Growth 31 II What has driven urban spatial evolution since 1990? The answers are complicated—but incomes are key How cities grow – Page 33 How income and population growth drive overall urban built-up expansion – Page 33 How income and population growth drive vertical layering – Page 37 Photo: World Bank/Flickr, Bogota, Colombia How cities supply floor space by building upward and expanding outward – Page 40 How population density responds to income growth – Page 43 SPOTLIGHT 2 Disruptive technologies and the future of cities: A policy and analytic agenda – Page 45 32 Pancakes to Pyramids | City Form to Promote Sustainable Growth What has driven urban spatial evolution since 1990? The answers are complicated—but incomes are key The econometric analysis in chapter 2 aims Residential densification, or a rise in population to isolate and measure the drivers of urban density—defined as a rise in population spatial evolution in its main dimensions, using relative to built-up area—is derived from data on economic and population growth population growth combined with the three (1990–2015). The analysis yields useful margins: horizontal spread, infill development, insights into the complex etiology of cities’ and vertical layering. Critically, an increase in spatial form. It also informs the analytic population density need not cause a decline in framework and predictive application of the floor space per person. Both may coincide, as new urban model in chapter 3. when developing country cities accommodate growing populations in crowded neighborhoods How cities grow and slums (because only limited income is available for housing investment and How does a city adjust to accommodate consumption). But this is not inevitable: as increasing numbers of people—other than by a productive city grows richer, with people simply becoming crowded and congested? and firms generating greater demand for land, Basically, it can accommodate its growing housing, and centrally located office space, population by expansion along three margins. it can add to its population density while still These margins are: expanding (or only slightly shrinking) its floor space per person. • Horizontal spread. At or near the city edge, the city can add land area through extensive growth. How income and population growth drive overall urban built-up • Infill development. Inside the city edge, the city can add built cover by filling in expansion green or “brown” land with structures— Both a city’s population size and its income or (less frequently) it can replace old level drive expansion in its built-up area—but structures with new ones. of these two drivers, population is by far the • Vertical layering. The city can increase stronger. Specifically: its floor-area ratio—available floor space • The elasticity of built-up area to per unit of land area—by building taller population is 0.35. If the city’s population structures, thus fitting more total floor increases by 10 percent —while incomes space onto the same amount of land. remain constant—its built-up area increases by 3.5 percent. This 3.5 percent Each of these three margins reflects a specific figure represents a striking response to combination of demand signals. Generally, population growth. if productivity and incomes are sufficient to generate demand for more built-up area, then • The elasticity of built-up area to income supply will respond to that demand, driving is 0.1. If the city’s average annual the city to expand and increase its total floor income (per capita GDP) increases by space. The extent to which this happens in 10 percent—while population remains practice will change—as the population grows, constant—its built-up area increases and as productivity and incomes rise through by 1 percent. Rising incomes thus drive successive stages of a city’s economic expansion, but they do so far more slowly development. than rising population. Pancakes to Pyramids | City Form to Promote Sustainable Growth 33 Box 2.1 Econometric issues in estimating the drivers of built-up area expansion At least two econometric issues and the explanatory variables (city its key predictors. In each case, a arise when estimating the impacts population size and annual per capita simple ordinary least squares (OLS) of growing populations and rising income). Second is the omission of estimator will be biased. The panel incomes on urban built-up area. First variables: certain unobserved factors, regression analysis is based on the is endogeneity: unobserved factors such as geographic and weather following specification: may simultaneously affect both the amenities, may independently affect dependent variable (built-up surface) built-up area or even be among For city c, at time t, where BUc,t is built-up area for city c, at time t, and FE city and year fixed-effects. Box figure 1 shows the results for the previous specification for all countries first and then for cities per income group. Box figure 1 How is a 10 percent increase in population and income associated with built-up expansion? Population Per capita income 3.5 1.2 3.5% 3.4% 3.0 1.2% 3.1% 1.0 Change in built-up space (%) Change in built-up space (%) 1.0% 1.0% 2.8% 2.5 0.9% 2.5% 0.8 2.0 0.6 1.5 0.5% 0.4 1.0 0.5 0.2 0.0 0.0 All countries High income Upper middle Lower middle Low income All countries High income Upper middle Lower middle Low income income income income income Source: Econometric analysis by report team; details in Lall, Lebrand, and Soppelsa (2021). The figure shows the results of the and within estimator, adding both specification uses functional specification using city and year year and city fixed effects. The panel urban areas (FUAs) to measure fixed effects, for all countries and by data allow for city fixed effects to urban boundaries. Our preferred income group. More specifications control for the large heterogeneity of specification is given by the between are reported in Lall, Lebrand, and unobserved city characteristics. and within estimator, which is Soppelsa (2021). A first specification reported in the figure. Income and Two additional specifications report reports the results of the between population elasticities can be inferred the results using an instrument estimator, which includes year fixed from the regression results. Estimates strategy for both population effects and city controls. Year fixed are similar when using instrumental and income to correct potential effects control for aggregate changes variables, nighttime lights, and FUAs. endogeneity issues. The first that have affected all cities at each The exception is when we use FUAs instrument is the lag of population period. The built-up area of the for land-rich countries (Australia, and income. However, it reduces the city in 1975 is added as a control Canada, New Zealand, and the sample of observations by a third. for unobserved variables such as United States)—the elasticity of built- The second instrument used for geographic characteristics that are up area to population is 0.21, and the income is the country per capita GDP. a main determinant of city growth elasticity of built-up area to income over time. A second specification Another specification uses nighttime is 0.38. reports the results for the between lights as a proxy for income. A final 34 Pancakes to Pyramids | City Form to Promote Sustainable Growth In box 2.1, a figure provides income and In poorer countries, population growth is population elasticities of built-up area associated with faster built-up area expansion expansion using a panel of observations from than in richer countries. As the urban the Global Human Settlement Urban Centre population increases by 10 percent, a city’s Database (GHS-UCDB) for 1990, 2000, and built-up area increases by 3.4 percent in low- 2015.1 City income and population estimates income countries. The corresponding figure is are also from the GHS-UCDB (similar estimates 3.1 percent for lower-middle-income countries, are obtained when using country income). 2.5 percent for upper-middle-income countries, and 2.8 percent for high-income countries.2 In Cities grow differently at different stages of contrast, rising incomes are associated with economic development. Box figure 1 in box slower built-up expansion in poorer countries 2.1 reports the results of the elasticities of than in richer ones. As income increases by built-up area to population and incomes at 10 percent, the city’s built-up area increases various stages of economic development (as by 0.5 percent in low-income countries, 0.9 defined by World Bank classifications). These percent in lower-middle-income countries, 1.1 estimates are based on the between and percent in upper-middle-income countries, within estimator for city built-up area. and 1 percent in high-income countries. These results are consistent across cities of different As box figure 1 in box 2.1 shows, as a city’s sizes. However, the income elasticity of built- population increases by 10 percent, its up area is smaller for cities with more than built-up area increases by 3.5 percent when 500,000 persons: for these cities, as income keeping income constant and controlling for increases by 10 percent, the city’s built-up unobserved city and year characteristics. area increases by 0.7 percent.3 This number is far lower than previous estimates reported in the literature (Angel Cities in poorer countries thus expand more et al. 2016). It shows that cities have not horizontally than cities in richer countries—a expanded their built-up area as much as finding that suggests richer countries are their population growth may have seemed increasing their total urban floor space through to suggest. The figure also shows that as vertical layering, as well as along the other a city’s per capita income increases by two margins (horizontal expansion and infill 10 percent, its built-up area increases development; box 2.2). by 1 percent when keeping population constant and controlling for unobserved city and year characteristics. Higher income increases the demand for residential and commercial space and is associated with a city’s growing ability to build higher and acquire more efficient public services, such as subways (metro rail systems). 1 While built-up area for all cities with more than 50,000 inhabitants globally is available for four years—1975, 1990, 2000, and 2015—the 1975 data are dropped due to measurement concerns. The various Landsat missions used in the GHS- UCDB—Landsat 1–3 for 1975, Landsat 4–5 for 1990, Landsat 7 for 2000, and Landsat 8 for 2015—were designed for compatibility (in spectral bands and wavelength, for example). Still, the 1975 built-up layer may underestimate built-up area, as Landsat 1–3’s spatial resolution (60 meters) is coarser than that of Landsat 4–8 (30 meters). Details are in annex 2. 2 There is no contradiction in the fact that the estimated coefficients per income group are all lower than the estimated coefficient when including all cities. 3 The corresponding increase in built-up area for smaller cities is 0.1 percent (associated with a 10 percent increase in incomes). Pancakes to Pyramids | City Form to Promote Sustainable Growth 35 Box 2.2 Two concurrent processes: Horizontal spread and infill development Built-up area can grow at a city’s technologies, preferences, and local high-income country that increases extensive margins, as the city political choices. its built-up area by 100 square m expands outward through what will add about 35 square m through Horizontal spread dominates the this report calls horizontal spread. infill development and 65 square m growth of built-up area—and its share But built-up area can also expand through horizontal spread. However, was even greater in 2000–2015 within the city’s boundaries through a similar city in a low-income than in the previous decade. In what economists call intensive country will add about 90 square m low-income countries, 90 percent margin development—here termed through horizontal spread, with only of built-up area is provided through infill development. While most 10 square m from infill. These findings horizontal spread. Nevertheless, cities grow through a combination are consistent with the underlying there is a silver lining: in high-income of horizontal spread and infill intuition that agglomeration and upper-middle-income country development, the proportions of economies, incomes, and supply cities, a larger share of new built- each are likely to change at different capabilities all improve with a up area is provided through infill stages of economic development— country’s transition to upper-middle- development. For example, a city in a and with changes in construction income status. Box figure 1 The contribution of horizontal spread and infill development to total additions to built-up area, 2000–15 The spatial pattern of urban growth differs by income level Growth period 2000–2015 LOW INCOME In low-income countries, development nearly all built-up area growth is provided through horizontal 9% spread outside the existing urban extent. HIGH INCOME 35% Horizontal spread 91% By comparison, one-third of built-up area growth is through infill construction 65% in high-income countries. Source: Authors’ analysis, based on GHS-BUILT data. Source: Authors’ analysis, based on GHS-BUILT data. Note: These estimates of infill development are likely to be biased downward—the satellite imagery cannot identify changes in land use and the intensity of land development in places that have already been built up. 36 Pancakes to Pyramids | City Form to Promote Sustainable Growth How income and population growth drive vertical layering Cities can add floor space through vertical representative of the larger Global Human layering as well as horizontal spread and infill— Settlement Layer sample—to examine how by building upward, as well as by expanding heights vary with economic development. outward. Vertical layering plays a key role in These data were described in box 1.2. A visual adding downtown floor space, which alleviates inspection of the data shows that richer cities crowdedness in urban centers and enables are physically higher and more structurally cities to achieve livable population densities. compact near the center: they are more pyramid shaped, with peaked skylines and a To isolate and measure the effects of income higher concentration of downtown floor area and population growth on vertical layering, for residents and businesses (figure 2.1). we use a subsample of almost 400 cities— Figure 2.1 Richer citiesare cities Richer more pyramid aremore shaped, with peaked skylines pyramid shaped, with peaked skylines 30 Average building height (meters) High income 20 Upper middle income Lower middle income 10 Low income 0 10 7.5 5.0 2.5 0 2.5 5.0 7.5 10 Distance from city center (km) Source: Authors’ Source: Authors’ analysis, analysis, based based data.data.. on WSF-3D on WSF-3D Note: The graph reports the smoothed line of average pixel heights per city. Each pixel contains information on both its average height and the calculated distance to the highest pixel of the city. First, in each city, pixel heights are averaged for each interval of distance to the highest point. Second the lowess function is used to report the smoothing lines of average heights for cities per income group. The figure above reports the lowess smooth of average height per distance to the highest city point per income group. The patterns displayed in figure 2.1 reflect Richer cities thus have taller structures, economic processes. Highly productive with more concentrated commercial and agglomeration economies drive rising demand residential space near the center, than poorer for commercial floor space and for housing, ones—yet rich cities also tend to expand both of which are normal goods (demand more. Even as richer cities add floor space by increases with income). Higher income cities building upward, they still expand outward: meet part of this demand through developers their workers are more likely than workers in who increase building height in response poor cities to live in a less residentially dense to rising land prices, especially downtown: suburb while commuting to an economically pyramids form through infill development, dense center. utilizing all available land area near the center. Pancakes to Pyramids | City Form to Promote Sustainable Growth 37 While low-income cities are built much closer height of urban structures is somewhat larger to the ground, they are also less outwardly than that of income growth (figure 2.2): expansive. Low economic demand for • The elasticity of height to population is suburban residential space keeps people in 0.25. If the city’s population increases poor cities crowded together. Most workers in by 10 percent (while incomes remain poor cities have lower incomes and few transit constant), its average building height options, so they cannot afford the time or increases by 2.5 percent. expense of a long commute and must live near their jobs. Crowding into the urban core, they • The elasticity of height to income is increase its population density—importantly 0.185. If the city’s income increases by defined here as the ratio of population to built- 10 percent (while population remains up area, rather than floor space—to levels that constant), its average building height are unlikely to be seen at later stages if the city increases by 1.9 percent. becomes richer. Box 2.3 provides the technical specifications of the drivers of building heights. The independent effect of population growth on the average Box 2.3 Technical specifications of city height estimates This chapter provides results from Height is measured as the average average over the whole urban extent econometric analysis to assess the height of the city’s buildings within as defined by GHS-UCDB, or using income and population elasticities of 5 km of the highest point in the city.2 A quantiles over the whole distribution. city height.1 similar analysis is conducted using the Using the cross-section data for almost 400 cities, the following specification is estimated for all cities: with the average height of buildings within 5 km of the highest point in city c. Several specifications are estimated with region dummies as additional controls and replacing city income with its equivalent at the country level.3 1 The subsample of 400 cities was built to be as representative as possible at the global level of the Global Human Settlement Layer sample of cities. As a first check, the same regression is run using both the universe and the subsample. The within estimator of the income and population elasticities of built-up area was obtained for each sample and was shown to be very similar. See Lall, Lebrand, and Soppelsa (2021) for details. 2 Given that most surface is expected to be built around the highest point of the city, the measured height per pixel in this 5 km radius will correspond to both the average over the buildings and the average for the total area of the pixel. DLR data provide height averages over the area of the pixel. 3 Population and income from 2000 are used for two reasons. First, height data have been collected between 2007 and 2011, with the year of collection reported in the data. Second, adding some lag in the explanatory variable can be thought of as an instrumental strategy to deal with the endogeneity issues associated with the initial specification. 38 Pancakes to Pyramids | City Form to Promote Sustainable Growth Figure 2.2 How income and population are associated with cities’ vertical structure Figure 2.2 How income and population are associated with cities’ vertical structure 10% increases in population and income associated with… 3.0 8,243 km2 2.5 2.5 Change in city height (%) 2.0 1.9 1.5 1.0 0.5 0.0 Population Income Note: The figure shows the population and income elasticities of average height results from the regression reported in Lall et al. (2021). Data on average height have been collected between 2006 and 2011 and correspond to only one point in time for each of the 397 cities. Several specifications, OLS, with nighttime lights and region dummies, are tested in the paper by Lall et al. (2021). The figure reports the results of the OLS estimation. Combined with the elasticities of built-up area to and from higher densities of local economic income and population described above, these activity. These incentives boost the demand elasticities of height to income and population for downtown floor space and thus for higher fully capture the association between each commercial buildings in urban centers. On the driver and total floor area growth across cities. supply side, higher land prices and the density Notably, however, the elasticity of height to of demand create incentives for developers to population is lower than that of built-up area supply more floor space in a given land unit area. to population—while the elasticity of height to The need for higher structures also improves income exceeds that of built-up area to income. capabilities in the construction sector, which requires advanced engineering technologies to Why are rising incomes a stronger driver of build upward, as well as in the real estate sector vertical layering than of horizontal spread? (Ahlfeldt and Barr 2020). Intuitively, these findings suggest that households—which can afford higher rents Across geographic regions, cities display a as their incomes rise—prefer to pay more for large heterogeneity in their patterns of vertical housing closer to job centers. Living near the expansion (Lall, Lebrand, and Soppelsa 2021). center reduces commuting time, and it allows Compared with cities in the East Asia and access to downtown amenities. Pacific region, cities in Latin America and the Caribbean have shorter buildings. But the widest Richer cities also tend to produce more divergences from global averages are seen in tradable services—which rely more heavily South Asia and in Sub-Saharan Africa. Cities than non‑tradables on knowledge workers, and in these regions have the lowest buildings which gain the most from agglomeration effects (controlling for income and population). Pancakes to Pyramids | City Form to Promote Sustainable Growth 39 Vertical layering, in addition to horizontal 1.5 percent annual population growth—housing spread, is essential for cities to increase demand will rise by 3 percent per year. If that their floor space and remain livable as their rate holds for the next 20 years, the result populations grow. Duranton and Puga (2020) will be a cumulative 80 percent increase in make back-of-the-envelope calculations of demand, or a near doubling in the amount of the floor space that Latin American cities will floor space needed to accommodate firms need over the next 20 years. Assuming that and families. Achieving this doubling of total 2 percent income growth will increase the floor space through horizontal spread alone is demand for housing by 1.5 percent per year unlikely—and undesirable. (with housing prices constant)—and assuming How cities supply floor space by building upward and expanding outward Floor space is critical for cities to building both upward and outward. The effects accommodate families and firms. It can be on height and built-up area are: added through horizontal spread, vertical • The elasticity of height to the combination layering, or infill development. This section of population and income is 0.44. If uses the previous horizontal and vertical the city’s population and income both growth estimates to measure the elasticities increase by 10 percent, its average of floor space to income and population, both building height increases by 4.4 percent. individually and in combination (specifications appear in box 2.4). • The elasticity of built-up area to the combination of population and income is Figure 2.3 shows these elasticities and 0.45. If the city’s population and income confirms that as cities grow in population, in both increase by 10 percent, its built-up incomes, or in both, they add floor space by area increases by 4.5 percent. 40 Pancakes to Pyramids | City Form to Promote Sustainable Growth Box 2.4 Technical specifications of floor space estimates Income and population elasticities of total floor space are defined as the following:1 The elasticities of per capita floor space to income and to population are then expressed as:2 1. These expressions are derived using the log properties. The elasticities for total floor area with respect to income or population are then derived. 2. The elasticities for total floor area per person with respect to income or population are then derived. Pancakes to Pyramids | City Form to Promote Sustainable Growth 41 Figure 2.3 The elasticities of a city’s built-up area, average building height, and total floor Fig. 2.3 space to its population and income POPULATION INCOME POPULATION AND INCOME A 10% increase A 10% increase 10% increases in in population is in income is population and income are associated with… associated with… associated with… 7% 3.5 4 Change in total floor area per person (%) 3 6% 3.0 2.9 2.5 6.0 1.9 2.9 2 5% 2.5 1 Percentage Percentage 4% 2.0 0 -4 3.5 -1 3% 1.5 -2 2% 1.0 1.0 -3 1% 0.5 -4 0% 0.0 -5 Built-up area Height Total floor space Built-up area Height Total floor space Population Income Source: Authors’ analysis, based on GHS–Urban Centre Database and WSF-3D data. Note: Because population is kept constant, the income elasticities for total floor space and total floor space per person (per capita floor space) are the same (0.29). These estimates have profound implications for • The elasticity of total floor space to a growing city’s supply of floor space, and also population is 0.60. If the city’s population for its tendency to sprawl outward—especially increases by 10 percent (holding income in the absence of rising productivity and constant), its total floor space increases incomes. If incomes do not grow but remain by 6 percent through a combination of constant, a 10 percent increase in the urban built-up area expansion (3.5 percent) and population increases total floor space by only vertical layering (2.5 percent). 6 percent. Furthermore, the 6 percent increase in total floor space is achieved more through horizontal spread (built-up area expansion) than through vertical layering (upward construction). 42 Pancakes to Pyramids | City Form to Promote Sustainable Growth The 6 percent increase in total floor space, as a intuitively plausible explanation for the income result of increasing the city’s population by 10 gains not reflected in added floor space is that percent, falls short of the 10 percent increase some are dissipated in higher housing prices. in the number of people needing space. This is Also, rising incomes are likely to increase suggestive evidence that floor space does not demand for housing services, which include increase in proportion to population and that the quality and variety (not just the quantity) residents of larger cities consume substantially of floor space. But unless per capita income less floor space per person. increases by at least 14 percent, if population increases by 10 percent then floor space per Cities with rising productivity and incomes person will fall. fare better. If a city’s income increases by 10 percent, holding population constant, total One of the key stylized facts is that increasing floor space—and thus per capita floor space— incomes are the one indispensable driver increases by almost 3 percent. of vertical layering, because building tall is capital intensive. This is true for office blocks, • Elasticity of total floor space to income: and also for the move from informal to formal 0.29. If the city’s income increases by 10 settlement. It therefore requires income levels, percent (holding population constant), capital wealth, and financial institutions that its total floor space increases by 2.9 enable these investments to be made. Even if percent through a combination of built-up the levelized lifetime costs of different building area expansion (1 percent) and vertical types were similar, meeting the upfront capital layering (1.9 percent). costs of load-bearing structures would still be more feasible in cities with higher incomes and Because the floor space available per person productivity. rises with income—though not proportionally so—the city becomes more livable. An How population density responds to income growth Finally, how do cities’ population densities— of population densities to incomes.) Annex 1 defined as their ratios of population to built-up provides these estimates. As shown previously, area—change in response to income growth? as cities become richer, their demand for floor And how do these densities evolve over time? space rises and their supply of built-up area To answer these questions, a panel regression increases in response, even with population (along the lines of equation 1) was estimated held constant. So as cities expand in built-up with density measures for cities around the area, they become less densely populated. world. (One modification to the estimation was The income elasticity of density is −0.10 using the introduction of a quadratic term to pick the between and within estimator; introducing out potential nonlinearities in the response instruments for incomes increases it to −0.15. Pancakes to Pyramids | City Form to Promote Sustainable Growth 43 Figure 2.4 Predicted density over built-up area Income elasticities of density Cities with more than 500,000 inhabitants 17,000 16,000 Predicted density 15,000 14,000 13,000 12,000 0 20,000 40,000 60,000 GDP per capita ($) Source: Authors’ analysis, based on GHS–Urban Centre Database. Note: Density is measured with respect to built-up area, not floor area. Figure 2.4 shows the predicted densities at various income levels. As cities become richer, they expand in response to a higher demand for floor space, making themselves less densely populated. However, a nonlinear specification for cities larger than 500,000 inhabitants reveals that population density stops declining and more or less stabilizes when cities reach $20,000 per capita GDP: if anything, density after that point begins a gradual increase. As the populations of large cities keep growing, some of these cities will add more space through vertical layering—adding both to livability and to economic agglomeration gains. 44 Pancakes to Pyramids | City Form to Promote Sustainable Growth SPOTLIGHT 2 Disruptive technologies and the future of cities: A policy and analytic agenda Cities across the globe are entering a realm of What lessons from classic urban economics and rapid disruption brought about by a wave of economic geography get overturned, and which transformational technological progress. The hold up well under rapid technological change? new informational, transactional, and operational Will new digital technologies lead to a rethinking technologies—popularly termed Industry 4.0—are of common economic wisdom, perhaps in the likely to change the shape of cities and could well way that the new trade theory provides a new require a rethinking of how we define density. perspective on globalization? The most profound disruptions are likely to come How are the impacts of the current technological/ from operational technologies that combine data digital revolution on cities different from the with automation. Examples include hyperloop, impacts of other general technology revolutions? robotics, machine learning, 3-D printing, New digital innovations have spread faster autonomous vehicles, artificial intelligence, and around the globe, but that does not mean that off-grid energy systems. The fundamental driver the benefits in each country will be universally is the falling cost of routine functions. These shared. Just as in previous episodes, there disruptors are at the frontier of innovation and will be winners and losers. Yet, the new are likely to fundamentally reshape notions of digital technologies, because of their rapidly density and economic geography. falling costs and the extensive scope of their applications, could yield higher returns more The city of the future is also likely to have a quickly and for more people around the world. different economic base. The future of work will be one based more on the: What time frames should we be thinking of? The fast pace of digital technological innovation • Gig economy—based on flexible, temporary, clashes with the long time scales inherent in or freelance jobs, often involving connecting urban development. Technologies tend to be with clients or customers through an online outdated within five years. Major infrastructure platform. and zoning choices, in contrast, are policy • Sharing economy—involving short-term decisions with consequences through peer-to-peer transactions, often through generations. some type of online platform that connects buyers and sellers. Are there important interactions with other major global trends? Digital technologies and With such technology come changes in urbanization are not the only mega trends connectivity and proximity. These types of work shaping economic development. Climate change have a different infrastructure base. For example, (and environmental degradation more generally), they may not need networked services but will demographic trends, the possible reversal of instead be on micro-grids and will be more home- globalization, or a gradual shift in global economic based. and political power could co-determine how cities will fare. Do possible interactions affect To help city leaders make more informed choices how we should view digital technology change in in anticipating and planning for the future spatial cities? organization of cities, a forward-looking policy and analytic program are urgently needed to addresses the following issues: Pancakes to Pyramids | City Form to Promote Sustainable Growth 45 III What drives one city to grow differently from another? Using a structural urban model to clarify cities’ spatial development – Page 47 What determines the shape of the city? – Page 42 Projecting the effects of policy on cities’ spatial growth: The new urban model in action – Page 60 Using the new urban model to estimate the impacts Photo: World Bank/Flickr, Dhaka, Bangladesh of reduced travel times and increased housing supply in metropolitan Dhaka – Page 62 SPOTLIGHT 3 Urban form and greenhouse gas emissions – Page 69 46 Pancakes to Pyramids | City Form to Promote Sustainable Growth What drives one city to grow differently from another? Chapters 1 and 2 of this report have examined derived from what urban economists know as the three main margins of urban spatial the new urban model (box 3.1). The resulting expansion observed around the world from analytic framework can provide insight into 1990–2015: horizontal spread, vertical layering, the likely productivity and density effects of and infill development. To clarify what drives various policies and policy changes, including each of these margins, chapter 3 now digs large investments such as transit systems and more deeply into the determinants of a city’s other infrastructure. spatial form, applying a structural urban model Using a structural urban model to clarify cities’ spatial development This section presents a version of the new linkages, and general equilibrium effects may urban model that is tailored to help policy be important but are hard to discern using analysts estimate the aggregate quantitative reduced form approaches. The framework can impacts of economic policies—and compare also help decision makers predict the gains and the impacts of alternative policies using distributional impacts from future investments counterfactual simulations. Such a framework and policy changes. is especially useful when large spillovers, Box 3.1 What is the new urban model, and how can its analytic framework help city leaders? The structural approach to urban Brueckner and Sridhar (2012), showed households, this in turn changing the economics that informs Pancakes to how central area height restrictions form of clusters of employment and Pyramids is based on a new urban could encourage urban sprawl and hence of productivity. model, which builds on the earlier reduce real incomes in the city. The new urban model also fills out urban tradition of the Alonso-Muth- The new urban model—first developed further aspects of the city in greater Mills (AMM) model (Alonso 1964, Mills by Ahfeldt et al. (2015) who build detail. There is rich modeling of 1972, Muth 1969), and does so in a on earlier work by Lucas and construction technologies, infill within way that is both more general and Rossi-Hansberg (2002) and Fujita the city boundary as well as extensive can capture heterogeneity of cities. and Ogawa (1982)—differs in two growth at the city margin, and multi- The new urban model can be applied main respects. First, agglomeration sector approaches with different to help city leaders understand how forces are made explicit, and worker economic activities having different economic productivity and various productivity is assumed to depend locational priorities. Dynamic aspects aspects of urban form mutually on the scale and density of economic of the model make it possible to look interact. activity in the city, in line with the at the process of development and The AMM model generally assumed findings of empirical research on redevelopment of a growing city. that all employment takes place in agglomeration. Second, firms, as Importantly, new sources of data the city center and that all workers well as households, choose where provide information on building height commute to this central district (or to locate in the city, responding to (lidar), on traffic and commuting point). There is an implicit assumption costs of commuting, wages and flows on the city’s transport network that agglomeration forces create this floor space prices in different parts (mobile phone data), and hence on the central clustering, but firm productivity of the city. These changes give a possibility of linking people’s places is taken to be fixed. This basic model better description of the city as of work, home, and recreation. These yields important insights, particularly some employment may be dispersed, developments enable the new urban about the determination of land rents and some residences will be in or model to be calibrated to particular and the varying density of settlement close to the center. They also give cities, and to serve as a tool for policy at different distances from the center. a much richer set of responses to analysis. Policy applications of the model to investment and policy changes. For particular cities, such as those of example, height restrictions may Bertaud and Brueckner (2005) and change the location of firms as well as Pancakes to Pyramids | City Form to Promote Sustainable Growth 47 Informing the structural urban model are report’s structural urban model broadly several basic intuitions about the decisions distinguishes two production sectors: that face firms, workers, developers, and tradables and non-tradables (box 3.2). The two landowners: sectors have different priorities in choosing urban locations: • Firms must choose their locations. Production can, in principle, occur • Many non-tradables producers prefer anywhere in the city, but in practice a to be near the final consumer, so they firm’s choice will reflect characteristics are less centralized than tradables such as productivity, access to labor, and producers. Non-tradables firms will be the supply of commercial floor space. somewhat dispersed throughout the city. In contrast, tradables firms—those that • Workers must choose where to reside trade externally—are less constrained by and where to work across all locations a need to be near local consumers and so in a city. Both choices will depend on how may be more centralized. attractive locations are as places to live in and to work in, as well as on the cost • Firms in certain specific sectors of commuting between residence and are especially likely to benefit from workplace. agglomeration economies, so they will cluster together. Business services, • Developers create housing and finance, and many creative sectors tend commercial space using the capital to cluster and are likely to locate in a and land available in each location. central business district (CBD). Demand for floor space will determine the profitability of their investments. • Employers require floor space to operate, and their varying demands for • Landowners determine land use. Their space—and for specific building types decisions reflect tradeoffs between to supply it—are a further crucial factor estimated returns on residential or in firms’ location decisions. Office-based commercial use, along with any regulatory employers can build tall, enabling high land use restrictions. density in the CBD. In contrast, most manufacturing firms require more space All these individual decisions are related per employee, sometimes in structures through general equilibrium market clearing that are necessarily low-built: such conditions, which equate the demand and manufacturers are more likely to occupy supply for each factor in each location and pin the city edge. down prices, wages, and rents. Just as important as firms’ choices about Because all employers need access to where to locate are workers’ choices about workers, urban firms on average tend to favor where to live—along with the related decisions central locations. Neighborhoods near the of developers and landowners about where to center are closer to a larger share of the city’s provide residential floor space. The decisions total land area than are areas near the urban by developers and owners that produce this edge. Furthermore, edge areas may not be stock reflect both demand and supply factors: favored by radial transport networks. • Housing floor space demand comes from For many firms, however, access to workers households that seek greater access to is not the only consideration—and firms in jobs, services, and amenities—and that some sectors have other overriding reasons act on a variety of innate preferences to be located away from the center. To better for particular locations, needs for floor understand firms’ location decisions, this space, and degrees of affordability. These 48 Pancakes to Pyramids | City Form to Promote Sustainable Growth choices are shaped substantially by All of the aspects described above interact: available and accessible transport modes the places where people work and live (and and mobility technology. commute from), the choices made by firms about location, production, and hiring, and the • Housing supply is shaped by the construction decisions of land developers. A availability and price of land and by key mechanism that coordinates these choices available construction techniques. Multi- is the land market and the price—or rental story apartment blocks will be built where rate—of land in different neighborhoods. The land is expensive. One- and two-story structural urban model used here captures housing will be built farther from the center, these interactions in a consistent manner. where land is cheaper. Construction types vary widely in durability and capital Later sections of this chapter apply this model cost, and informal settlements arise to a particular city—Dhaka. General insights because of people’s inability to afford the are obtained by using the model to analyze capital cost of durable structures—along the interactions of firms, people, and other with institutional barriers to efficient elements in a hypothetical city located on a land use. Importantly, decisions about “featureless plain.” This illustrative hypothetical durable structures must largely reflect city will be monocentric, with rotational expectations about future demand and symmetry and a perfectly circular boundary. how the city is likely to grow (Henderson, Regan, and Venables 2020). Key variables for this hypothetical city are illustrated in figure 3.1. Plotted against the Overarching the spatial distribution of firms and horizontal axis of each panel (a, b, c, and d) residents are three further factors: is distance from the city center: as we look • Natural geography. Geographic features from left to right, we are thus traveling along a influence a city’s transport options radial slice of the city, from the CBD to a point and accessibility. For this report, the beyond the city edge. most stylized model ignores geographic differences and assumes that every city Box 3.2 Tradables and non-tradables in the is on a “featureless plain.” But in applying the framework to real cities, geography urban economy matters. This includes elevation, coastal location, and local amenities, including What does a city produce—what goods and services do its firms hire people to provide? The larger share of parks and open spaces. its labor force is probably employed in non-tradables: activities that supply goods and services to the local • Institutions, governance, and urban plans urban population. Such non-tradables include personal and policies. A city’s shape reflects the services, retail, hospitality, construction, building institutional structure of the land market, maintenance, transport, and public services. building and land use regulations, taxes, A smaller share of workers are likely to be employed in and the investment in and placement tradables, or goods and services that are exported to of public assets—in particular the other regions or countries (as well as being consumed transportation network. These factors— locally). Tradables include manufacturing and varied services, from business, legal, and financial services to grounded in law, governance, and policy— media and education. are explored in some of the experiments A city must export something in exchange for any food, conducted for this report. fuel, or other goods and services that are not locally produced. But the size of its tradables sector varies • Population size. The model employed here with many factors. The more the city’s imports can be reflects city size through a variable for financed by funds transferred from outside the city urban growth. (government transfers, resource rents, or development aid), the smaller its tradables sector will be (Gollin, Jedwab, Vollrath 2016). Pancakes to Pyramids | City Form to Promote Sustainable Growth 49 Figure 3.1 Some patterns predicted by the model for a hypothetical city—built on a “featureless plain” a. Land rent b. Height of buildings Land rent (outside = 1) Residential Commercial 8 12 10 6 8 Land rent Height 4 6 4 2 2 0 0 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 Distance from central business district Distance from central business district c. Employment and residential density d. Productivity Resident Employment-T Employment-N Productivity-T 2.0 1.3 1.5 1.2 Productivity Population 1.0 1.1 0.5 0.0 1.0 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 Distance from central business district Distance from central business district Source: Authors’ analysis. Note: The model is simulated on a two-dimensional plain and for now considers a fixed city population. The figure reports a one- dimensional slice from center to edge. The city edge is slightly jagged, because discrete finite size cells are used in simulation. Units used for all four charts, whether for distance or for other variables, are abstracted model units—they are not directly interpretable as kilometers, people, or dollars. But because each model unit is consistent wherever it appears, the charts allow comparisons across cases. 50 Pancakes to Pyramids | City Form to Promote Sustainable Growth Plotted against the vertical axis of figure 3.1a The presence of agglomeration economies is land rent, which is costly at the center and in the model used for figure 3.1—along with declines with distance, until reaching the level developers’ ability to build tall in the CBD, and of land rent outside the city—this point defining the consequent clustering of T-sector firms the edge of the city. On the vertical axis of near the center—gives the city the benefit of figure 3.1b is building height, which responds high productivity. As indicated in figure 3.1d, to rent: as developers seek to economize on productivity is highest for the firms nearest the land area relative to other factors, they build core: that is what justifies the high rents they taller structures in the center with more stories are willing to pay. than at the edge. Note that the model example for figure 3.1 allows residential and commercial structures to use different technologies, so figure 3.1b reflects how commercial firms—to the extent that they are office-based—will consistently build taller than residential developers. Patterns of employment and residential density appear in figure 3.1c. Plotted against the vertical axis are three distinct aspects of density, defined as people per unit of land area: the density of workers in the tradables sector Box 3.3 Agglomeration economies (T-sector), the density of workers in the non- tradables sector (N-sector), and the density Firms and workers in some types of economic activity of residents. Readers will note that the plotted have higher productivity when they are located in a heights measure density, not the numbers of large and dense cluster of economic activity. Numerous people working or living at each distance from factors create this effect. In a cluster, firms have lots the center since a belt far from the CBD is far of suppliers and customers nearby, not only saving on transport costs but also allowing knowledge transfer larger in area than a belt of the same width and the use of just-in-time technologies. The cluster may closer in. be large enough for highly specialized firms to develop, with very high productivity in a particular activity—an Why is T-sector density so strikingly skewed example of Adam Smith’s scale and specialization. toward the center, in contrast to the other two Pools of highly skilled labor may develop, with skills being passed from worker to worker—or generation densities shown in figure 3.1c? Because, in the to generation in some cases. These agglomeration model used for figure 3.1, firms in the tradables economies may operate across a wide range of sector benefit from agglomeration economies activities, or in particular sectors such as financial (box 3.3). Accordingly, firms in this sector services, high-technology clusters, and film and media cluster in the center: a preference that does centers. much to drive the high land rents in figure 3.1a, A good deal of research has quantified these effects. and building heights in figure 3.1b. In contrast, This suggests that, in high-income countries, the agglomeration scale elasticity is around 3–5 percent, the N-sector in figure 3.1c is highly dispersed, implying that a city with a population of 8 million has spreading outward through residential areas at productivity 15 percent higher than one with 250,000 increasing distances from the center. Finally, people. In upper-middle-income countries the effect residential density is highest at some distance has been estimated to be several times larger again. from the CBD: workers want access to the Less research on this subject has been done in low- income countries, but such as there is suggests a cluster of central employment, but commercial much weaker effect, perhaps because the sectors in development prices most of them out of the which agglomeration economies are most powerful are innermost core—the downtown CBD itself. largely absent from low-income cities (see Grover, Lall, and Maloney 2021 for a review of evidence and further discussion). Pancakes to Pyramids | City Form to Promote Sustainable Growth 51 What determines the shape of the city? Underlying the benchmark city of figure • The city’s total population size. 3.1 are the economic, technical, and policy relationships that factor into a city’s economic • Transport and commuting costs, affecting and spatial development. Four of these factors the ease with which people can move are especially important: around the city. • Agglomeration economies, or the extent Cities vary in each of these dimensions, along to which the city’s economic structure with others—and each is critical to determining enables T-sector firms to benefit from a city’s shape, as well as its prosperity. urban agglomerations. • Construction costs, reflecting constraints—whether financial, technological, or regulatory—on developers’ ability to build tall. The absence—or presence—of agglomeration economies as a key determinant of city shape Agglomeration economies are a key driver of (N-sector) employment density highest in the the centripetal forces and density of the CBD center. Residential density is also now higher in in the benchmark city. But many cities lack the the center than anywhere else in the city. The types of productive sectors that are able to reason is that non-tradables firms—including benefit from agglomeration economies. What is retail and personal services—gain from easy the shape of such cities? access to large numbers of customers, while residents gain from easy access to these Figure 3.2 describes a city, the same as the firms both as employees and customers. benchmark in all respects except that the When the T-sector has no incentive to cluster, T-sector does not benefit from agglomeration the densities of N-sector employment and economies, having instead constant returns of residents can drive each other ever higher to scale. The differences are striking. In near the center without heavy competition figure 3.2—compared with figure 3.1c—the from the T-sector. Land rent in the city center employment structure of the city is inverted, is much lower, this implying less tall buildings with tradable (T-sector) employment density (figure 3.2b). highest at the city edge, and non-tradable 52 Pancakes to Pyramids | City Form to Promote Sustainable Growth Figure 3.2 Spatial distribution of residents and employment (tradables and non-tradables sectors) and heights of buildings in a city lacking agglomeration economies a.Employment and residential density b. Height of buildings Resident Employment–tradable Employment–non-tradable Residential Commercial 0.4 12 10 0.3 8 Population Height 0.2 6 4 0.1 2 0.0 0 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 Distance from central business district Distance from central business district Source: Authors’ analysis. The scale and further implications of these benchmark city of figure 3.1) to a higher value. agglomeration effects appear in figure All variables measured on the vertical axis are 3.3. On the horizontal axis is a measure of expressed relative to the benchmark city (all agglomeration economies, ranging from 0 (the take value 1 at point 0.1). city described in figure 3.2) through 0.1 (the Pancakes to Pyramids | City Form to Promote Sustainable Growth 53 Figure 3.3 Effects of agglomeration economies on other city characteristics in the new urban model Figure 3.3a Figure 3.3b a. City shape b. Economic productivity Average residential density Maximum land rent Real wages (real income) Average land rent per person Average commute distance Peak productivity 2.0 1.2 1.8 1.1 1.6 1.0 1.4 1.2 0.9 1.0 0.8 0.8 0.6 0.7 0.4 0.6 0.2 0.0 0.5 0 0.02 0.04 0.06 0.08 0.1 0.12 0 0.02 0.04 0.06 0.08 0.1 0.12 Agglomeration economies Agglomeration economies Source: Authors’ analysis. The city shape captured in figure 3.3a gives without. This is largely a direct consequence average residential density, maximum land rent, of switching on these economies—but it is and average commute distance. Maximum land amplified by the fact that, the larger these rent—a reflection of demand for space in the scale economies are, the more firms and CBD—increases sharply with agglomeration employment cluster. The additional income economies, varying ninefold across the range generated by higher productivity is divided provided. Average residential density and between worker households and landowners. commute distance increase slightly: while We measure the worker household benefit the locational patterns of residence and by average utility (the average wage net employment vary substantially, there is little of commuting costs and housing costs).1 change in the overall land area of the city. The figure indicates that the benefits of agglomeration-driven productivity go Figure 3.3b gives real income measures. Peak disproportionately to land rents (which more productivity is more than 30 percentage points than double across the range of the figure) higher with agglomeration economies than 1  he same amount of floor space costs different amounts at different places in the city. The utility measure deflates T nominal wages by a location-specific price index to control for this variation. 54 Pancakes to Pyramids | City Form to Promote Sustainable Growth relative to worker household utility, which agglomeration economies, others do not. increases by around 25 percent. Still others may be prone to agglomeration economies but also be very land-intensive— The message of figure 3.3 is that the shape car factories are built sprawling, not tall—so of the city varies greatly with its economic clusters develop on the city edge, while the function. Some T-sectors have strong center attracts N-sector activities. Construction costs as a constraint on floor space growth and building height Land is the city’s scarcest factor. Development agglomeration spillovers between firms. Zoning and construction technology transform land can be done in ways that are damaging, into floor space. Numerous constraints affect locking in land-use patterns that become this transformation, each with effects on the inappropriate as a city develops. But zoning city’s shape and growth. can also yield benefits by reducing negative externalities that affect households—and also, Building technology varies, from the mud possibly, by encouraging a concentration of floors and iron roofs of informal settlements commercial activities that promotes positive to the steel and glass of modern high-rise spillovers and externalities between firms. districts. Costs vary—especially the capital costs of building durable, load-bearing, and To see the implications of these factors for the tall structures. The capital intensity of these shape of the city, figure 3.4 varies a parameter structures means that they are relatively more of the model that captures construction costs expensive in capital scarce countries than and the obstacles to building tall. The horizontal in capital abundant ones: as incomes rise, axis is the construction cost parameter, relative building types evolve. The associated capital to its value in the benchmark. Increasing this requirements appear in the data (Dasgupta, parameter by 10 percent implies that at a given Lall, and Lozano-Gracia 2014) and can be land rent, buildings will be about one-third as captured in modeling. Henderson, Regan, and tall as in the benchmark case. Associated with Venables (2020) model the evolution of slum this, the city sprawls, with lower density and and formal settlements theoretically and in an longer commutes. application to Nairobi, incorporating the loss of amenities due to high slum density (see also Bird and Venables 2019). Whereas formal structures can deliver density with height, informal ones can deliver it through crowding (little floor space per person, little green space per house). Building and development choices are also shaped by institutional and regulatory factors. A lack of clarity in land tenure creates a bias against sinking capital into formal structures—a factor in the perpetuation of slum areas. Building regulations, such as floor area restrictions, often restrict the density of both commercial and residential development: such restrictions lengthen commutes and weaken Pancakes to Pyramids | City Form to Promote Sustainable Growth 55 Figure 3.4 Effects of construction costs on other city characteristics in the new urban model Figure 3.4a Figure 3.4b a. City shape b. Economic productivity Average residential density Maximum land rent Real wages (real income) Average land rent per person Average commute distance Peak productivity 1.2 1.03 1.02 1.1 1.01 1.00 1.0 0.99 0.9 0.98 0.97 0.8 0.96 0.95 0.7 0.94 0.6 0.93 0.95 1.00 1.05 1.10 1.15 0.96 0.98 1.00 1.02 1.04 1.06 1.08 1.10 1.12 1.14 Construction costs Construction costs Source: Authors’ analysis. Note: Population size is fixed. Income effects are shown in figure 3.4b. As lower rents in the central area are offset by the central cluster becomes less dense, somewhat higher rents in outlying areas. productivity is about 3 percent lower, reducing worker household utility by 5 percent. Average land rents per capita also fall, but by less— City population size as a driver of downtown rent, building height, and employment density Large cities tend more than small ones to form the benchmark city: a nearly fourfold variation pyramids. Varying the model city’s population in city population size is indicated in the figure size can quantify this generalization. In figure (50–180 percent of the benchmark). 3.5, the horizontal axis is population relative to 56 Pancakes to Pyramids | City Form to Promote Sustainable Growth Figure 3.5 Effects of city population size on other city characteristics in the new urban model Figure 3.5a Figure 3.5b a. City shape b. Economic productivity Average residential density Maximum land rent Real wages (real income) Average land rent per person Average commute distance Peak productivity 1.3 1.04 1.03 1.2 1.02 1.1 1.01 1.0 1.00 0.99 0.9 0.98 0.8 0.97 0.7 0.96 0.5 1.0 1.5 2.0 0.5 1.00 1.5 2.0 Population Population Source: Authors’ analysis. Figure 3.5a indicates that, as expected, cities As the city grows more rapidly in population with larger populations have higher rent in the size than in productivity, worker household center—a sign of taller central buildings and utility (real income) bears the penalty of higher higher employment density. Yet residential commuting costs and higher average rent per density varies little, indicating that the city’s person—even though nominal wages may built-up area grows almost in proportion to remain broadly in line with productivity. From its population size. Along with this horizontal left to right in figure 3.5b, as the population spread, average commute distance rises grows, average rent per person rises by 6 sharply (by more than 50 percent). Figure 3.5b percent, while real wages fall by 4 percent. Any shows the productivity and wage effects of productivity gains that firms may be realizing population size variations. From left to right, in this city—which is likely to be increasingly a rise in population results in somewhat crowded—are either dissipated in commuting higher productivity: agglomeration economies costs or transferred to landowners. reflect scale as well as density. Yet this rise in productivity is small relative to the nearly fourfold increase in population size as the additional employment in the city spreads over a larger land area. Pancakes to Pyramids | City Form to Promote Sustainable Growth 57 Commuting costs as drivers of rents, job locations, and residential densities across the city Reducing commuting costs has the dual effect levels? Figure 3.6 shows the effects of varying of enabling both employment concentration commuting costs, relative to the benchmark and residential dispersion: forces that appear case. Thus, a value of 0.9 on the horizontal axis to pull in opposite directions. What is the net indicates that all the travel costs in the city effect on the city’s expansion, distribution are cut by 10 percent from their benchmark of densities, and income and productivity values. Figure 3.6 Effects of commuting costs on other city characteristics in the new urban model (fixed population) Figure 3.6a Figure 3.6b a. City shape b. Economic productivity Average residential density Maximum land rent Real wages (real income) Rent per person Average commute distance Peak productivity 1.10 1.03 1.08 1.02 1.01 1.06 1.00 1.04 0.99 1.02 0.98 1.00 0.97 0.98 0.96 0.96 0.95 0.94 0.94 0.8 0.85 0.9 0.95 1.0 1.05 1.1 0.80 0.85 0.90 0.95 1.00 1.05 1.10 Travel costs Travel costs Source: Authors’ analysis. 58 Pancakes to Pyramids | City Form to Promote Sustainable Growth Figure 3.6a captures both the employment Transport improvements, however, generally are concentration and the residential dispersion not citywide—instead, they entail developing that result from lowering commuting costs. particular road or rail networks that benefit From right to left, as the city grows in area some locations more than others. Most cities relative to its population (a spread that is have some sort of hub and spoke transport implied by the decline in commuting costs), network, with arteries radiating out from the residential density declines. But because the center to outlying districts and beyond the average commute grows longer in about the city. What is the effect of adding such a radial same proportion as the decline in density, transport network to the benchmark city? the concentration of jobs around the CBD In a dramatic experiment in the model, six (reflected in the maximum value of land new roads lead radially out from the CBD, on rent) does not vary much. A slightly greater which journeys are 50 percent faster (shorter employment density near the center has a distance) than elsewhere.3 Adding this network small positive effect on maximum productivity, creates powerful centripetal effects: the shown in figure 3.6b. Worker household utility new roads advantage the central area, along (real income) now receives a double benefit with other locations along the new travel from the commuting cost reduction—a user corridors. At the center, employment density benefit, or direct effect from the improvement, and maximum rent nearly double, as the road plus wider benefits that result from the city’s system enables residents to move away additional productivity. Average land rents from the CBD and downtown land becomes per person fall substantially, however: as available for businesses and jobs—with lower commuting costs increase the city’s consequent agglomeration benefits. physical size, the effective supply of land for development increases. The real income effects of the radial road network are substantial, comprising three Because reduced commuting costs imply an components. The first and smallest component increase in worker household utility, a further is the network’s direct user benefit—the value predictable effect of lower commuting costs of time and cost savings, holding firm and is migration: drawn by higher urban wages, household locations unchanged. This direct migrants add to the city’s population. This benefit accounts for a 3 percent increase in-migration response can be captured in the in worker household utility. In addition, model by supposing that there is a supply productivity gains arising from the city’s new curve of labor to the city. If the elasticity of this spatial pattern lead to wider, indirect benefits supply curve is set, for illustrative purposes, that boost real wages: these indirect benefits at 2, then the 3 percent increase in real wages add a further 4 percent to worker household illustrated in figure 3.6b will lead to a 6 percent utility. And if all these effects are accompanied increase in city population. When combined, by in-migration, real wages rise still further. All the insights from figures 3.5 and 3.6 predict told, the combined impact of direct and indirect dual effects from transport improvement: benefits, including effects from in-migration, the combination of travel costs and induced increases worker household utility by 9 in-migration leads to lower residential density, percent compared with the city that lacks the longer commutes, and a higher concentration radial road network. of employment in the center.2 2 The maximum rent schedule of figure 3.5a becomes U-shaped, increasing at low travel costs as greater population leads to a denser central area. 3 Travel choices are then made efficiently, on the shortest path between each node pair in the city. Pancakes to Pyramids | City Form to Promote Sustainable Growth 59 The key messages from this experiment are (the combined effect of firm and worker two: relocation). • Transport networks are key to creating hubs of economic activity in city centers. The modeling approaches outlined here can be used to forecast the possible effects • Transport networks’ direct income of transport policies on city growth and benefits (reductions in travel time and development. Because these effects are cost) may only account for a small part highly sensitive to place and context (box of their total benefits, which also include 3.4), scenarios require the detailed calibration indirect benefits driven by agglomeration that we describe in the next section (see also economies and increased firm productivity Sturm, Takeda, and Venables 2021b). Projecting the effects of policy on cities’ spatial growth: The new urban model in action Whereas the previous section provided general The new urban model can be used to insights into urban growth for a hypothetical understand the trade-offs and unintended city, the new urban model can also be applied consequences of local land use rules to real cities. and initiatives, which are likely to have complicated impacts throughout a city. For Modeling the evolution of a city such as example, one model allows for both formal Bogotá, Dhaka, or Mexico City enables city and informal construction, enabling policy leaders to consider granular predictions about makers to quantify the costs of formalizing the direct and indirect impacts of potential slums (Henderson, Regan, and Venables policy interventions. Unlike reduced form 2020). Another model allows planners to methods—which can estimate only a part of characterize optimal zoning across residential these policy impacts—structural urban models and commercial use, taking into account the estimate the broader impacts and understand fact that households and firms do not fully the mechanisms behind them. They offer internalize the consequences for others of their economic intelligence on the effects of policy location choices (Allen, Arkolakis, and Li 2016). packages as opposed to standalone efforts, In Bird and Venables (2019), a similar model is and on the advisability of complementary applied to evaluate the impact of tenure reform policies to manage the downside risks of in Kampala, while also accounting for various specific investments. types of municipal housing policy: public housing projects, slum upgrading programs, Structural urban models can thus be pragmatic and land change. tools. One practical use of the new urban model is to look at the wide-ranging effects of Gechter and Tsivanidis (2020) quantify the urban land use policies in combination with impacts of redeveloping Mumbai’s 58 textile transit systems and other large improvements. mills during the 2000s—finding that while the Most transportation investments are amount of formal housing in the city center expensive, especially in developing cities, increased, poor residents of nearby slums were and have broad efficiency and distributional also displaced as housing prices rose across implications. Recent modeling studies have the area. Anticipating such effects would allow predicted the impacts of such investments— for complementary and compensating policies along with complementary policies—in Bogotá to manage these downside risks. (box 3.5) and Mexico City (box 3.6). 60 Pancakes to Pyramids | City Form to Promote Sustainable Growth Box 3.4 How transport innovations have shaped cities Historically, the main effect of on the example of London from 1801 Subway construction, however, is transport technology—steam engines, to 1921). capital intensive. Furthermore, the subways, cars, and so on—has success of subway investments In the 20th century, the mass been to enable urban expansion and depends on a city’s density. These production of cars dramatically increase the living space consumed requirements partly explain why not affected both density and living space. by households (Glaeser 2020). The all cities have subways. Among cities In the United States especially, cars transport innovations that most that do, a majority began providing reduced transport costs and strongly affected cities’ spatial evolution in the subway services only after they drove the horizontal spread of built- 19th century were large public transit reached $5,000 per capita GDP or up area—making urban sprawl, with systems, such as street cars and higher (box figure 1). its associated externalities, a central subways. By contrast, the innovation challenge for city planners today Today, in the poorest countries, car that most affected cities in the 20th (Burchfield et al. 2006). ownership remains very low, and century was mass-produced cars. many workers still walk to work or Generally, subways lead residential In the 19th century, while streetcars to access services. As incomes rise, locations to disperse centrifugally—but and subways initially enabled ownership of private vehicles is likely they do so less than highways have urban sprawl, they also induced to follow, pushing cities outward and been shown to do (Gonzales-­ Navarro urban economic density through widening the spatial separation of and Turner 2018). Subways are thus agglomeration. Enabling residences residences from workplaces. But less conducive to sprawl. According to to exist farther from workplaces new transportation trends—such as analyses done for this report, cities at supported spatial expansion but also autonomous vehicles—and broader given population and income levels are attracted jobs to city centers (see technological changes (notably more residentially dense when they Heblich, Redding, and Sturm [2020] increased telecommuting) will also have subways. affect city shapes (Glaeser 2020). Box 3.3_p39 Box figure 1 Most cities provide subway services only after reaching at least $5,000 per capita GDP GDP at subway starting year Average GDP for all cities Average GDP for cities without a subway Dubayy (Dubai) San Juan Las Vegas Seattle Jacksonville København (Copenhagen) Detroit Los Angeles 30,000 Washington, DC Atlanta Hiroshima Torino (Turin) Vancouver Toulouse Daejon San Francisco Rotterdam Bruxelles Singapore Kyoto Montréal München (Munich) Valencia Daegu Incheon Makkah (Mecca) Sapporo Yekaterinburg Kazan Toronto GDP per capita ($) Milano (Milan) Valparaíso Caracas Monterrey Santo Domingo Ciudad de México Kuala Lumpur Busan Ankara Istanbul (Mexico City) Medellín Tehran Chengdu Roma (Rome) Rio de Janeiro BrasÍlia Kaohsiung Lima Porto Alegre Krung Thep (Bangkok) Nagoya São Paulo Nanjing, Jiangsu Dnipropetrovsk Santiago Shenzhen Lisboa (Lisbon) Dalian 3,000 Manila Guangzhou, Guangdong Al-Qahirah (Cairo) Taipei Delhi Shanghai Kolkata (Calcutta) Beijing Tianjin Hong Kong 300 1940 1950 1960 1970 1980 1990 2000 2010 2020 Source: Authors’ calculations using data from Gonzalez-Navarro and Turner (2018). Pancakes to Pyramids | City Form to Promote Sustainable Growth 61 Other questions remain open for which regulations are changed to improve aggregate a structural model is useful to capture efficiency in resource allocation?” If efficiency all following impacts and unintended enhancement is the main objective of reforms, consequences. For example, while building the outcome would be Pareto optimal if, in higher in the city center might seem a good principle, there would be net gains after idea, it also attracts more people to the city displaced people are compensated. In practice, center and generates additional negative however, it is not clear that mechanisms externalities that are not internalized by exist to compensate those who are worse off the households or firms moving there. following changes in housing and land markets. Spotlight 3 discusses this question in the trade-off of urban form and greenhouse gas The absence of compensating mechanisms emissions. Another important question relates means that the displacement of poor to unintended gentrification and spatial households or other vulnerable groups inequalities that could accompany policies that reduces their living standards. Although the enhance the overall economic efficiency of the model presented here can highlight potential city. The main normative question here is, “How displacement and gentrification following do we compensate and protect the welfare efficiency-enhancing policies, further work of the poor and vulnerable communities as is needed to understand how to design and implement complementary compensatory policies. Box 3.5 Density-enhancing land value capture schemes can amplify the economic Using the new urban model to benefits from Bogotá’s Transmilenio bus estimate the impacts of reduced rapid transit system travel times and increased housing supply in metropolitan Dhaka Quantitative models in the spirit of the new urban model can identify policies to complement expensive This section illustrates how the new urban infrastructure and maximize returns on investments. One model can be used to analyze and model such analysis recently examined the bus rapid transit (BRT) system in Bogotá, Colombia (Tsivanidis 2019). various impacts on a real urban agglomeration: Using a model that allowed for multiple skill groups of workers with non-homothetic preferences over various transport modes, Tsivanidis quantified the BRT system’s Box 3.6 Transit infrastructure lowers impact on aggregate performance not only through informality rates in Mexico City reduced time losses from travel, but also through improved allocation between workers and places of employment and residence. After accounting for Informality presents policy challenges for most policy reallocation and general equilibrium effects, the analysis makers in developing cities. Can transit infrastructure suggested that welfare gains were 20–40 percent larger increase allocative efficiency by reducing informality? than otherwise estimated. A new study explores this question by analyzing the effects of metro system construction in Mexico City The quantitative model generated other striking findings. (Zarate 2020). For example, it showed that Bogotá’s feeder bus system (which partly solves the last-mile problem of getting Assuming that high urban transit costs may prevent residents between poor, dense peripheral neighborhoods workers from commuting and thus limit their access and the BRT) improves welfare more than any single to formal employment, and finding that informality trunk line. Also, Tsivanidis ran a counterfactual exercise declines by 4 percentage points in areas near the suggesting that if the government had adopted a land new metro stations, Zarate uses a model to estimate value capture scheme—increasing zoning densities the efficiency gains that result. He determines that near BRT stations, and auctioning building permits to workers’ reallocation to the formal sector explains developers—welfare gains could have been 18 percent approximately 17–25 percent of the total gains from higher. Further, government revenues could cover the metro system—and that for every dollar spent on between 8 and 40 percent of the systems’ construction its construction, average real income increases by 20 costs. percent relative to a perfectly efficient economy. 62 Pancakes to Pyramids | City Form to Promote Sustainable Growth the city of Dhaka, Bangladesh. One of the and informal employment) and residential major cities of South Asia, Dhaka is home to a population (from the 2010 population dynamic ready-made garments industry that census)—both of these across the 266 wards connects the city’s economy with global value of Dhaka’s metropolitan area. Figure 3.7 shows chains. Having experienced rapid population the enormous variation in population and growth since the partition of India in 1947, employment density across those wards. Dhaka continues to attract new residents at a Central parts of Dhaka have vastly higher rapid pace. Today the metropolitan area covers employment densities and much higher about 1,465 square km—comparable to the population densities. Employment at workplace Greater London Authority. We first outline the density is just under 3 workers per hectare at steps involved in applying the model to Dhaka, the 5th percentile, while it is 485 workers per and then the results of policy experiments, hectare at the 95th percentile, and the median including transport improvements and lower- is 41 workers per hectare. Similarly, population cost building technologies. density is just over 3 people per hectare at the 5th percentile, while it is 232 people per Application of the model involves several steps. hectare at the 95th percentile, and the median First is obtaining basic data, including counts is 53 people per hectare. of employment at workplaces (from the 2013 employment census, capturing both formal Figure 3.7 Population and employment density in Dhaka Source: Sturm, Takeda, and Venables 2021b. Second is establishing information on the They estimate that the elasticity of commuting key relationships in the model. Travel times flows with respect to travel time is −2.5: between and within wards can be obtained that is, a 1 percent increase in travel times from city sources, but a full description of reduces commuting flow by 2.5 percent on commuting flows is less easy to obtain. average. Kreindler and Miyauchi also report a Kreindler and Miyauchi (2020) use mobile decomposition of this overall effect into the phone data from Dhaka to estimate the flows. Pancakes to Pyramids | City Form to Promote Sustainable Growth 63 separate contributions of commuting costs of the United States. Two reasons may lie and of preference heterogeneity. behind this difference: one is the higher costs of building tall structures in Dhaka compared A further relationship is that between house with US cities, while the other is more stringent prices and the amount of floor space supplied regulatory constraints in Dhaka. in various neighborhoods of the city—the elasticity of housing supply. As is the case in most developing country cities, there is no systematic data on the price of a square meter of residential or commercial floor space in different parts of Dhaka. To overcome this problem, we use newly available data from the German Aerospace Centre (DLR), in which satellite pictures allow measurements of built- up area and building height for cities around the world. Using these data, we compare the heights predicted by the model to the heights observed in the data for different values of housing supply elasticity in the model. Figure 3.8a shows the variation in the height of buildings across the wards of Dhaka in the DLR data and the heights predicted by the model for the best fit value of housing supply elasticity. While the correlation between the DLR height data and the model-predicted heights is not perfect, the two are clearly strongly correlated (the correlation coefficient is 0.49) and lie close to the 45 degree line.4 Figure 3.8b compares the volume of buildings in the data and the model, where we feed the observed built-up area into the model as data. Comparing the model-predicted heights to the observed heights in the data, the best fit value of the housing supply elasticity is 1.45: that is, a 10 percent increase in the price of a square meter of floor space in a location triggers a 14.5 percent increase in the supply of floor space. This value for the housing supply elasticity is slightly below the average housing supply elasticity estimated by Saiz (2010) across different metropolitan areas 4 T  he DLR height data come with their own measurement error, which is—on a percentage basis—likely to be especially pronounced in lower-density peripheral locations of Dhaka. In these areas, the time lag between the census and height data probably also leads to further substantial measurement error. We take these errors into account when we estimate land prices in each location of Dhaka, as discussed in Sturm, Takeda, and Venables (2021b). 64 Pancakes to Pyramids | City Form to Promote Sustainable Growth Figure 3.8 Height of buildings and volume 3.8 floor space in the model and the data for Dhaka Fig of a. Height of buildings 20 Average height of buildings in the model (m) 15 10 5 0 0 5 10 15 20 Average height of buildings in the DLR data (m) b. Volume of floor space 60 Volume of buildings in the model (millions of m3) 50 40 30 20 10 0 0 10 20 30 40 50 60 Volume of buildings in the DLR data (millions of m3) Source: Sturm, Takeda, and Venables 2021b. Another relationship is the agglomeration estimate the strength of this agglomeration parameter—the elasticity of productivity with force in cities. Relying on this literature we set respect to employment density. An extensive this elasticity to 0.05, which is close to the literature has used different strategies to average value of this parameter reported in Pancakes to Pyramids | City Form to Promote Sustainable Growth 65 the meta study of Melo, Graham, and Noland sector can bid for land in a particular location, (2009). Increasing the density of employment given its production technology and the in a ward by 10 percent therefore increase demand for floor space in that location. The productivity in this location by 0.5 percent. land prices predicted by the model vary by a factor of roughly 18 between the 5th and Using the basic data and these key 95th percentile of the land price distribution—a relationships, the model can be calibrated to range similar to the variation of land prices in the observed distribution of employment and cities where we have good measures of the population that was illustrated in figure 3.7. value of land in different locations of the city. The model then provides estimates of other key variables that are not directly observed. Figure 3.10 combines the estimated variation Intuitively, the model-estimated wages ensure in wages and floor space prices across the that workers decide to commute so that the wards of Dhaka to estimate productivity in inflows of commuters to a destination match different locations of the city. Intuitively, the the observed employment at workplaces. model predicts that firms in locations with high Land rents are such that the observed floor space prices and wages must have higher built structures and population densities levels of productivity, as they otherwise would are consistent. And spatial variations in not be able to break even in a competitive productivity are consistent with the observed market. The model predicts considerable pattern of employment. variation in productivity across different locations, with locations at the 95th percentile Figure 3.9 shows the distribution of land having roughly 51 percent higher productivity prices estimated by the model. The land price than locations at the 5th percentile. is the maximum price that the construction Figure 3.9 Estimated land rent in Dhaka Figure 3.10 Estimated productivity in Dhaka Source: Sturm, Takeda, and Venables 2021b. Having estimated the key parameters of the a change in the density of development and model and fitted it to Dhaka, we use it to the building of a north–south road through evaluate the impact of two policy interventions: the city. The results presented here are for a 66 Pancakes to Pyramids | City Form to Promote Sustainable Growth variant of the model in which all employment either an improvement of the technology of the is assumed to be in the tradables sector construction sector or a relaxation of planning (T-sector), and the costs of construction for constraints. Our model estimates that such a commercial and residential buildings are the change would increase worker welfare by 5.4 same. percent, while the income of landowners would fall by nearly 3 percent and the total population Figure 3.11 examines the impact of a 25 of the city would expand by 11 percent. percent increase in the housing supply elasticity. To achieve this change would require Figure 3.11 Estimated impact of a 25 percent increase in the housing supply elasticity in Dhaka Source: Sturm, Takeda, and Venables 2021b. Figure 3.11 shows that while we assume that actually increases marginally by just under the housing supply elasticity changes by the 1 percent. This at first sight counterintuitive same amount in all wards of Dhaka, the impact result is driven by two opposing developments. of this change is highly uneven. The heights of Those living in the high-density core of the buildings in the very center of Dhaka increase city experience a fall in average commuting by over 50 percent, while building heights times. However, the remaining residents in in the periphery of Dhaka are essentially the peripheral wards experience a reduction unchanged. This asymmetry is also mirrored in local jobs and are more likely to commute by employment and population density, which to the dense urban center as a result. These both increase in Dhaka’s central areas—while results show how important it is to take both decline in its peripheral wards, even as general equilibrium forces into account when the city grows in population by 11 percent. assessing the impact of urban interventions. Intuitively, relaxing the housing constraint allows employment to further agglomerate Figure 3.12 examines the impact of in the highly productive city center, and constructing a new north–south road through residents follow this trend to be close to the Dhaka that would cut travel times on this route concentration of jobs. by 25 percent relative to current travel times. As the figure illustrates, the road acts as a A further striking insight from this magnet for economic activity in Dhaka with counterfactual is that while the average height increases in both employment, population, of buildings in Dhaka increases by nearly and building heights along the road. The road 30 percent, the average commuting time would increase aggregate worker welfare by Pancakes to Pyramids | City Form to Promote Sustainable Growth 67 just under 0.5 percent, while the income of the policy changes that can be explored in the landowners would increase by approximately model. Other experiments include the general 1 percent. This illustrates that improvements equilibrium impacts of much more targeted in commuting speeds not only benefit interventions, such as an increase in permitted commuters, but are also capitalized into land building heights in particular locations of a city prices. These increases in land values are or improvements in travel speeds on particular highly unequal, with landowners close to the travel corridors. In each of these cases the new road benefiting while land values in other model illuminates the general equilibrium parts of Dhaka decline marginally. linkages that need to be taken into account in evaluating urban policies. Transport improvements—either across the city or in particular places—are just a few of Figure 3.12 Estimated impact of a new north–south radial road in Dhaka Source: Sturm, Takeda, and Venables 2021b. The starting point of this report is that cities uncertainties will remain. Furthermore, adding are about interaction, and that effective detail to a model does not necessarily add policy requires that the interactions are insight. The new urban model should be viewed understood and that their implications are as one part of a toolkit for evaluating urban factored into decision making. The modeling development and policy change, while being approach outlined here is a tool for deepening far from encompassing all the consequences this understanding and for seeing the fuller of such changes. consequences of planned or likely changes in aspects of the city. As with all modeling exercises, many important features of the city and the urban system are omitted. For example, the data allow us to infer the productivity that a place offers to firms and the amenity that it offers to households, but is the amenity due to access to jobs; proximity to a park, a train station, or a school; or the social capital of the neighborhood? The research literature provides results on some of these causal channels, but others will remain hard or impossible to quantify. In other dimensions, too, such as dynamics and the full costs of adjusting to change, progress will be made, but 68 Pancakes to Pyramids | City Form to Promote Sustainable Growth Photo: Fran/Flickr, Shanghai, China SPOTLIGHT 3 Urban form and greenhouse gas emissions This report provides new evidence on Manhattan had a bunch of bungalows. Careful the three dimensions along which a city city-specific analytics can help untangle grows as well as an analytic framework the granular implications of policy efforts. to examine the institutional and economic forces that shape how a city is built. The jury is still out on the impact of urban form on greenhouse gas emissions. For example, The analytic framework developed in chapter Xu et al. (2019) [correct capitalization] show 3 can provide city-specific insights on the that population density and the overall physical implications of density-enhancing policies and compactness of urban land patches have transportation investments on greenhouse gas opposing influences on energy-related per capita (GHG) emissions. The policy counterfactuals— GHG emissions across EU member countries. changes in density by increasing housing High population density, mixed-use urban supply and the changes in transport speeds—as development with a lower degree of physical described in the chapter, show that implications compactness is advisable in terms of reducing of these improvements are not straightforward. energy footprints and mitigating GHG emissions. For example, increasing the density of a location Ma, Liu, and Chai (2015) find that compact urban makes a place more attractive and brings in development leads to low-carbon travel behavior. commuters from distant areas. Manhattan The evidence on the links between urban density has a lot of skyscrapers so a lot of people and greenhouse gas emissions also varies across can live close to work, but because it’s such a countries at different stages of development, great place people commute into Manhattan like the relationships of building heights and from distant places. They wouldn’t do that if built-up expansion examined in chapter 2. Pancakes to Pyramids | City Form to Promote Sustainable Growth 69 Box 1.5 Fig 1 Figure S3.1 While GHG emissions per capita decline with urban density in high- and upper- middle-income countries, low- and lower-middle-income cities buck the trend High and upper middle income Low and lower middle income 30 30 Detroit 25 25 Carbon footprint per capita (metric tons CO2) Carbon footprint per capita (metric tons CO2) Dubai 20 20 15 15 10 10 Hong Kong Ulaanbaatar 5 5 Kyiv 0 0 Islamabad 0 10,000 20,000 30,000 40,000 50,000 0 10,000 20,000 30,000 40,000 50,000 Population density (population/km of built-up area) 2 Source: Authors’ analysis, based on GHS–Urban Centre Database and Global Gridded Model of Carbon Footprints (GGMCF) retrieved from Moran et al. (2018). Figure S3.1 shows that while GHG emissions Finally, research is needed to understand per person decline with urban densities for the relative contribution to GHG emissions high- and upper-middle-income countries, the of urban structures and industrial emissions association does not seem to hold for low- that are co-located in cities. These include and lower-middle-income countries. Another power plants, steel plants, and cement insight here is that cities in low- and lower- plants. Such a decomposition is part of middle-income countries are already far denser future work being done at the World Bank. than cities in high- and upper-middle-income countries. As chapters 1 and 2 showed, cities in low- and lower-middle-income countries are likely to see greater urban expansion and reduction in densities as they develop. 70 Pancakes to Pyramids | City Form to Promote Sustainable Growth Photo: Nik Shuliahin/Unsplash, Kentucky, U.S. Pancakes to Pyramids | City Form to Promote Sustainable Growth 71 IV From pancakes to pyramids: What city leaders need to know Policy making for economic growth and productivity—strengthening institutional foundations – Page 75 Infrastructure planning for economic density, livability, and sustainability—scaling up and evaluating investments – Page 82 Durable financing for capital investment costs and recurring expenses—mobilizing urban revenue sources – Page 87 Photo: South Korea/Flickr 72 Pancakes to Pyramids | City Form to Promote Sustainable Growth From pancakes to pyramids: What city leaders need to know Pyramids are generally better than pancakes How can city leaders and decision makers at meeting three key urban planning objectives: act to shift urban expansion to a pyramidal driving prosperity, ensuring livability, and trajectory? First, they should never aspire to respecting planetary boundaries. Compared transform pancakes into pyramids where urban with a pancake city, a pyramid city will economic productivity is low, where demand drive more growth in urban productivity and for floor space is weak, or where inefficient incomes because it is more economically land markets impede formal investment and dense and efficient—its inward and vertical redevelopment. Plans that try to force a expansion reduce the distances between neighborhood’s vertical expansion through firms, jobs, and workers. A pyramid is also legal and regulatory incentives alone—without better at achieving livable urban population sufficient market demand for floor space, densities, accompanied not by crawling or without functioning factor markets and traffic and crowded slums, but by efficient adequate private investment—will yield only transport connections and decent formal ghost districts, their tall structures disused housing. And while a sprawling pancake is or underused. Such plans run afoul of basic likely to impose steep burdens on the climate principles of economic geography. through unmanaged vehicle emissions, a pyramid allows leaders to plan for the city’s Second, once leaders recognize why future population growth and spatial expansion mandated pyramidal growth plans cannot work in ways that will limit or reduce its carbon independently of the economic drivers of urban footprint. spatial evolution, they should investigate the more realistic options that remain. While these But not every pancake can become a pyramid. options differ from city to city, they can appear When a city with low productivity and low more readily through the new urban model incomes adds to its population, it cannot described in chapter 3 (data requirements are accommodate this growth through a costly summarized in box 4.1). Leaders can use this vertical layering of built-up area. Instead, such model to assess the likely impacts of particular a poor and economically inefficient city can plans and policies on future urban spatial absorb newcomers only by crowding them into growth—given a range of alternative scenarios low-built quarters and by spreading outward for economic transformation and rising where land is cheapest. Such a city will remain productivity—along with the predicted effects of a pancake—and it will continue to expand in these scenarios on firm and household location two dimensions, rather than three, as long as decisions over time. The new urban model its economy remains sluggish and its average can thus clarify how the city’s future physical resident household remains poor. form and human geography might respond to public investments—notably in transport As chapters 1, 2, and 3 have shown, pyramidal infrastructure—and to changes in zoning laws, expansion flows from economic transformation. building codes, and land use regulations. All of Only agglomeration economies, based on these policy choices will affect the quantity specialization and tradables production, can and spatial distribution of urban floor space, be counted on to set a city’s productivity and including patterns of residential density, as incomes on an upward path. And only a city developers respond to changes in demand. that is economically on the rise will generate increasing economic demand for floor space— After determining which urban plans, the prerequisite for land developers to invest regulations, and investments are best aligned in multistory construction around business to promote future urban prosperity and districts and elevate the urban skyline. sustainability, city leaders must put their Pancakes to Pyramids | City Form to Promote Sustainable Growth 73 strategies into practice through coordinated appropriate and not needlessly restrictive. action on three fronts—economic, spatial, and In the absence of a predictably functioning financial. Chapter 4 analyzes these closely and fluid land market, developers cannot linked challenges under the three section confidently expect high returns—and will headings that follow: never make the high capital investments needed for vertical layering, which would • Planning for economic growth and define the city’s evolution from a pancake productivity. Leaders must provide an to a pyramid. institutional environment that not only enables economic agglomerations for • Infrastructure planning for economic increased productivity and incomes, but density, livability, and sustainability. In also ensures demand-responsive urban evaluating infrastructure investments, planning regimes and allows for the leaders should focus on supporting future provision of floor space to meet the market forces that will drive urban rising demand. Land must be formally economic agglomeration and productive transferable and thus accessible to job creation within an environmentally development: ownership must be legally responsible vision for longer-term spatial clear, and land market transactions expansion. The central challenge is efficient. Land use and building height to coordinate investments with land regulations, where used, must be management, aligning connective infrastructure plans with land use regulations to shrink economic distances Box 4.1 Data requirements for the new between areas of potentially high urban model economic productivity and high residential density. Also vital are public goods and The basic data needed to compute the new urban amenities that will directly result in a more model—as described in chapter 3—are the following: livable and sustainably dense city. The • A GIS shapefile that partitions the city into a set new urban model can provide leaders of locations. with valuable economic intelligence to • Population data (or data on the economically ac- assess the likely impacts of infrastructure tive population) for each location in the city. investments and regulation—and can • Employment data (if possible, by sector) for each location in the city. point out risks that demand careful • Travel times between all pairs of locations in the management. city. • Durable financing for capital investment Note that in defining the city by area, it is important to include enough of the hinterland that everyone working costs and recurring expenses. The vital in the city can be plausibly assumed to live within the link between economic growth policies city perimeter. and infrastructure plans, durable financing In a developing country, two additional data types are is essential to pyramidal development. likely to be valuable: Investing in urban infrastructure is a capital • Mobile phone data to estimate commuting flows intensive proposition in the short, medium, and, if needed, employment and population in and long terms. To build and maintain each location of the city. a high quality and affordable transport • Satellite data on built-up area and building heights system—or any other urban public service to estimate the housing supply elasticity and floor or facility—means mobilizing revenue to space prices. cover both the initial investment and later Further data that could be used to inform or check operating costs. Accordingly, leaders model parameters include household surveys reporting household expenditure shares, firm surveys reporting must identify financing solutions for the firm input shares, commuting surveys, and real estate city to anticipate and meet its future price data. needs. City governments should clarify Source: World Bank. 74 Pancakes to Pyramids | City Form to Promote Sustainable Growth regulatory frameworks that pertain to Another general lesson is that urban economic municipal borrowing, public–private and spatial transformation form a virtuous partnerships (PPPs), and land value cycle: if rising productivity and incomes capture transactions. Also helpful is to in urban agglomerations can drive vertical streamline intergovernmental fiscal and layering and livability through the addition institutional frameworks. While cheap of floor space, this pyramidal expansion financing can raise money for investment, and increasing residential density can then repaying the principal requires a solid feed back into the economic density that funding base—and that will require strong drives agglomeration forces. Furthermore, economic fundamentals based on urban this economic and spatial virtuous cycle can agglomeration economies. spin off many options for sustainable urban development—from the revenues that finance Alongside this chapter’s specific practical low-carbon transit infrastructure, to the shifting guidance, city leaders ought also to keep in preferences that lead to greener commutes mind some broader lessons from Pancakes to and less carbon-intensive consumption in rich Pyramids. One is that every city, regardless cities. of its productivity and income, must plan for spatial development along all three margins: A final key point is that while urban pancakes not just inward and upward, but outward. While and pyramids are spatially opposed, the poorer cities tend to develop as pancakes fundamental drivers of urban development— for fundamental economic reasons, it is also and of the well-being of urban residents—are normal for richer ones—including pyramids—to not spatial. Instead, they are economic and continue their horizontal expansion. No city institutional. A rise in average population that keeps increasing its population will stop densities across a city and the rise of a extending outward altogether, unless it is peaked downtown skyline may reflect strong inhibited from doing so (and such inhibitions can agglomeration forces and complementary be costly to the city’s economy and to urban institutions. Nevertheless, achieving pyramidal residents). So sustainability initiatives should growth should not be the focus of urban policy not obstruct horizontal spread but should or the key metric by which success is judged. plan for it. Meanwhile, urban regulations and Improving a city’s livability and sustainability policy interventions, including investment, are must be. necessary to create a favorable environment for infill construction and vertical layering. Policy making for economic growth and productivity—strengthening institutional foundations While low-income and low-productivity cities important are steps to strengthen land markets cannot be transformed into pyramids today, and urban planning institutions. Low- and decision makers can set the stage for future lower-middle-income countries are expected transformative growth by establishing the right to dominate demographic urbanization in the institutional environment. Because economic coming decades (box 4.2)—and these, in and physical transformation go hand in hand, general, are the countries with the weakest cities need integrated legal and regulatory urban institutions and planning capacities reforms and frameworks that will enable both today. economic and spatial development. Especially Pancakes to Pyramids | City Form to Promote Sustainable Growth 75 City leaders and planners will need adaptable Reform land markets and clarify land strategies. Plans and regulations should allow and property rights the best use of land—but they must also allow changes in land use, and in users, as demand Land market inefficiencies appear widely in evolves further. Three key considerations will cities in the developing world, for reasons be how to: that range from unclear ownership rights to inefficient land allocation mechanisms and • Reform land markets and clarify land and development regulations. And wherever land property rights. markets do not function well, urban land tends to be underutilized. Prices for central locations are • Strengthen urban land use planning. likely to be too high. Land in those locations may go unused, or it may be underdeveloped with • Manage land valuation and prices. low floor space or low built-up area. To enable these cities’ future transformation into dense and livable urban areas, the first prerequisite is to improve access to land. Facilitating developers and households’ access Box 4.2 Urbanization in low-income and to buildable land requires unambiguous and lower-middle-income countries: How will it tradable land and property rights. Informal drive global demand for urban floor space in land markets are just not good enough for the coming decades? cities in developing countries, especially those experiencing rapid economic and population The United Nations projects that 2.3 billion people will growth. Urban land is a vital economic asset, be added to urban areas between 2020 and 2050, and asset transactions are viable only where with 90 percent of this increase taking place in Asia purchasers can rely on enduring documentation and Africa (UNDESA 2018). Urban population growth will be concentrated in a few countries. China, India, of ownership. A formal market both offers and Nigeria will account for 35 percent of the projected purchasers the state’s protection and—because growth. These are significant numbers, to be sure—but transactions are readily observable and proactive planning can help developing country cities recorded—generates the public good of accurate prepare for such growth. valuation. It is useful to place these projections in perspective. An additional 2.3 billion people over 30 years translates The first requirement of an effective and into 42 percent growth, or 1.4 percent annually. enforceable land and property registration Low-income countries will experience 65 percent system is tenure security. Such security growth (2.15 percent annually), and lower-middle- income countries 107 percent growth (3.57 percent enables land transactions to proceed annually)—considerably higher than richer parts of the unhindered by insecure land tenure world. arrangements—arrangements that limit trade, If the past foretells the future, the estimates in in practice, to parties who live locally or inside chapter 2 can help us understand how population small circles of trust. Globally, an estimated 70 growth will affect global demand for floor space in percent of all land lacks formal title.1 Alternative cities of low- and lower-middle-income countries forms of tenure security include certificates of globally. A conservative estimate, with no productivity occupancy and usufruct rights, along with other growth, would suggest that between 2020 and 2050 these cities’ built-up area will need to expand by 64 forms of partial tenure that protect the rights percent (low-income cities) and 36 percent (lower- to use and exploit land. Some forms of tenure middle-income cities). An optimistic estimate, with security, though not equivalent to formal titles, higher productivity and doubled incomes, would generate an additional 10 percentage points in demand for built-up area in each city category. Rising populations and incomes would also increase the 1 https://www.worldbank.org/en/news/feature/ demand for vertical layering. 2017/03/24/why-secure-land-rights-matter. 76 Pancakes to Pyramids | City Form to Promote Sustainable Growth enable land transactions or permit access to credit—giving assurance to households to invest in improving their housing conditions without fear of uncompensated eviction (GLTN n.d.). Perceptions of weak or nonexistent land Box 4.3 Urban slums—a lagging indicator of tenure security are widespread. According inefficient land markets to a recent global study on perceptions of property rights by Prindex (2019), one-fifth Land market inefficiencies often arise from a lack of persons globally expressed fears of being of formal title and from perceived tenure insecurity. forced out of their homes against their will Such inefficiencies inhibit cities’ formal development, in the coming five years. And because this making housing unaffordable within cities and study was conducted prior to the Covid-19 contributing to exurban sprawl. In addition, these market failures restrict central population densities outbreak, it likely underestimates current and thus price out poor residents from adequately perceptions of tenure insecurity. Countries located land. Poor people who need to live within reach that have invested in strengthening their land of job opportunities, yet who lack access to public administration systems—for example, Rwanda transportation infrastructure, often find themselves and Singapore—have only a limited share of with no option but to settle on marginal land, which the population (less than 10 percent) reporting may be at risk of flooding and landslides. Such settlements tend to become crowded and unlivable such perceptions. But this figure rises up to slums. about half the population in other countries Nearly a billion people globally live in these slums that have not invested sufficiently in land (including squatter and other informal settlements). registration, such as Burkina Faso and the Their land tenure is insecure, their housing is Philippines (Prindex 2019). substandard, and their infrastructure connections are typically poor, lacking water, sanitation, or stormwater Without secure property rights or perceived drainage.1 Not surprisingly, such places correlate with tenure security, households will not invest in Covid-19 hotspots (Lall and Wahba 2021). improving their shelter beyond minimal repairs. The good news is that many low- and lower-middle- Similarly, governments will rarely invest in income countries are taking steps to clarify land improving access to infrastructure and services rights and thus make land markets more efficient. Botswana took the bold step of regularizing customary for settlements that lack formal property rights. lands in 2008, partly because the Land Boards faced Such settlements often do not even appear in challenges to administering tribal land (Malope and official city maps. Without private investment Phirinyane 2016). Namibia recognizes traditional in shelter or public investment in infrastructure, leaders as part of the formal land system; they are slums and informal settlements remain designated by the president, and their details are published in the government gazette (United Nations unlivable (box 4.3). 2015). And Zambia passed a new planning bill in 2015, extending planning controls across state and Besides clear land and property rights, a customary land and designating all local authorities as government’s ability to assemble privately planning authorities (Wesseling 2016). owned land for urban expansion or Some countries and cities are also developing hybrid redevelopment also requires a coherent policy regimes to make formal and customary land rights and regulatory framework for land assembly: administration more compatible. For example, in one such framework is land readjustment Nigerian states with largely Muslim populations, the (Rabé 2010). The most commonly used tool for emir’s representatives subdivide and allocate land with the help of volunteer professionals from government: expanding urban boundaries on the periphery an example is the city of Rigasa, in the extreme west of of cities, land readjustment is also used for infill Kaduna (Igabi, Local Government Area, Nigeria; Lloyd- development. The government pools privately Jones et al. 2014). owned parcels in an area and prepares a land use plan, designating spaces for public 1 https://unstats.un.org/sdgs/report/2019/goal-11/ (last accessed February 20, 2021). Pancakes to Pyramids | City Form to Promote Sustainable Growth 77 infrastructure and services such as roads and disputed between landowners and 10,000 open spaces. It then implements the plan, slum dwellers—allowed the building of high providing trunk infrastructure, and distributes rises for existing residents, while releasing lots to landowners, proportional to the original other portions of the land for lucrative real parcels but smaller (for example, 50–60 estate development. In all seven cases, the percent). Because the new lot is serviced, it slum dwellers paid for part of the construction is worth more than the landowner’s original through a loan program. Generally, land sharing parcel. The government retains selected, can work both for squatter households, which strategic land parcels, which it auctions or gain the right to remain on the site (though in sells at market rate to recover the cost of new, multifamily, medium- to high-rise housing), infrastructure and service delivery (Lozano-­ and landowners, who recover and benefit from Gracia et al. 2013). part of their land (Rabé 2010). Land readjustment is useful for urban Another use of land readjustment was in regeneration where land ownership is Mumbai, where the World Bank supported divided among many private parties because the India Mumbai Urban Transport Project: such readjustment avoids the need for a vertical resettlement program for about the government to buy land outright. Yet 100,000 residents who formerly lived in urban it presupposes strong local institutions slums and shantytowns along aging roads and a sound legislative framework. Land and railway tracks. Through consultations readjustment has been used in Germany, with local displaced persons (DPs), the Japan, and the Republic of Korea to assemble Maharashtra State, Indian Railway Authorities, and plan privately owned land on the peri- and nongovernmental organizations, the urban fringe and develop it with infrastructure project resettled DPs into apartment buildings and services. In Japan, 40 percent of the close to their current locations in an effort to total annual supply of urban building plots preserve their social fabric. As an incentive from 1977 to 2000 was secured through land to apartment builders, tradable development readjustment. In the Republic of Korea, 95 rights— described in the next subsection percent of urban land delivery between 1962 below—enabled builders to acquire subsidized and 1981 occurred in the same manner (Povey floor area ratios (FARs) for constructing and Lloyd-Jones 2000). additional commercial space.2 Land readjustment has proved a successful An alternative to land readjustment is state instrument for urban redevelopment—notably expropriation of private land. Through powers in Bangkok in the 1970s and 1980s, when of eminent domain, governments can seize rapid economic growth drove up urban land land to pursue a public purpose—such as prices. Many slums were in accessible the provision of public infrastructure—while urban areas, which now became desirable following a due process, which includes to developers. To accommodate commercial compensation for the previous owners at development without displacing residents, the the market or replacement value of the land government brokered seven land-sharing deals expropriated. In instances where governments with slum dwellers. Existing development was control public land, an up-to-date inventory of to be densified, enabling the verticalization landholdings and a market-based allocative of low-rise or low-density residential uses system (auctions, sealed bids, or market and the opening of some of the land for valuations) can ensure that land is allocated to new development. The seven deals—struck its most efficient use, and affordable housing in cases where land rights had long been subsidies can be granted to eligible residents. 2 https://www.worldbank.org/en/country/india/brief/mumbai-urban-transport-project. 78 Pancakes to Pyramids | City Form to Promote Sustainable Growth Strengthen urban land use planning Land use planning. Well-functioning cities will provide public services, such as policing and Two government instruments are vital for health care, along with physical infrastructure— enhancing urban connectivity, productivity, roads, drainage, street lighting, electricity, water, and livability: sewerage, and waste disposal—in ways that • Land use planning —the distribution of benefit from complementarities and economies land uses across space together with of scale. Moreover, all these services and transportation and mobility infrastructure. infrastructure elements must be provided at once: addressing just one or two of them is of • Zoning regulations—the rules that govern little value if the others remain unresolved. Land population density, building heights, and use planning can help prevent these failures floor space. through foresight and strong implementation. What leaders do not always recognize is that That said, not all land use planning approaches both these instruments have fundamental lead to viable results. The preparation of a economic effects, and that their design will master plan is too often a static exercise, taking influence the market drivers of urban economic so much time that when the plan is adopted, development—for good or for ill. conditions have changed and the plan is no Economically, land use planning and zoning longer enforceable. Or the process may be a regulations are critical because their absence supply-driven exercise based on rigid planning will generate negative externalities and and engineering norms, such as the strict coordination failures. Unregulated markets are separation of land uses, the adoption of road unlikely to yield the most economically efficient hierarchies, and low built-up densities: norms quantity or density distribution of urban built-up inherited from colonial regimes or imported area, and they are unlikely to provide a city from other countries. These cases reflect a with its economically ideal form. The reason is disconnection between the planning process that, while firm productivity and job generation and the land uses that firms and households through density are positive externalities require—and, perhaps, a disconnection from the accruing freely to all, the investments needed to city’s socioeconomic and cultural reality. make higher population densities economically The more viable land use planning processes efficient—such as roads, buildings, and network are those that are more responsive to demand. utilities—are not fully internalized by firms and Cities including Seoul and Singapore have households. These market and coordination adopted demand-responsive planning through failures lead to suboptimal investment and, the integration of land use and transport ultimately, weaker productivity gains, slower job infrastructure, and through the adoption of high creation, and lower wages. FAR densities and mixed-use development. For In addition to land use planning and zoning instance, the FAR for commercial developments regulations, a supplementary instrument in downtown Singapore (called the Gross Plot considered in this subsection is transferable Ratio, or GPR) ranges around 10–15, which development rights. Described below, such allows for efficient land development and rights can introduce dynamism into land markets urbanization.3 In contrast, cities such as Mumbai and enable efficient development and density and São Paulo have traditionally had low FARs, distributions in line with market demands. around 1.5–2.5—a figure that is too low and does not allow for optimal land development.4 3 In fact, the Urban Redevelopment Authority (URA) of Singapore provides a map with the detailed land use and development control information at the parcel level for the city at: https://www.ura.gov.sg/maps/?service=mp. 4 São Paulo allows for the FAR to increase to 4 in certain specially designated areas for urban operations with the additional development rights—called certificates of potential additional construction or CEPACs—described in the subsection on “Tradable development rights” (World Bank 2013; Smolka 2018). Pancakes to Pyramids | City Form to Promote Sustainable Growth 79 Cities including Curitiba and Tokyo have Transferable development rights. In some also promoted transit-oriented development instances, a property’s location and land value approaches that create higher densities and would warrant higher density development, mixed uses around public transportation nodes, but the property is legally protected through resulting in a dynamic, dense, and livable built preservation—either as a historic building or environment. as an open space that provides a valuable amenity. Landowners can be compensated To be both viable and strong, urban land use for such government restrictions on their planning institutions must have a unique power development rights through a transfer to of enforcement. Empowered public authorities them of transferable development rights are essential to enforce private property (TDRs). Owners can sell permissible TDRs to rights. Because building the city depends developers in areas designated for additional on private rights over land and structures, densification potential.5 planning enforcement is fundamental to successful urbanization. For land registers In the United States, New York City used TDRs and mortgage collateral to perform their core to preserve its historic Grand Central Station functions—supporting a land parcel market, and its High Line (which was transformed providing finance for investment in structures— into a linear park), and Boston has used TDRs they need well-functioning on-the-ground to preserve historic neighborhoods such as enforcement. Back Bay, while cities have also used TDRs to control urban sprawl and preserve agricultural Zoning regulations. Zoning regulations similarly lands at the peri-urban fringe (Waldek require public enforcement. For the purpose 2018). In Brazil, São Paulo has pioneered of increasing market efficiency, zoning related market-based tools, such as the regulations have two functions: coordination auctioning of additional development rights and information. (CEPACs, or certificates of potential additional • Coordination—or the alignment of firms construction) beyond the maximum FAR in the to a common, publicly set standard— designated densification zones called “urban lowers development costs by providing operations.” The sale proceeds are used to standardized designs. finance infrastructure upgrades in the area. São Paulo city also has an instrument called • Information is provided by standards that Outorga Onerosa, which allows property govern structural features observable owners to build as much as 20 percent extra only during construction, such as onto their development by paying a fee into foundations. Such standards enable a general fund for infrastructure improvement property transactions by assuring later in the city. The city’s rationale for introducing purchasers that they have sufficient such instruments is to decouple land information about what they are buying. ownership from the development rights and to monetize the latter for revenue generation, By supporting property valuation, both the offsetting infrastructure improvement costs standardization and the information functions (Smolka 2018). of zoning regulations enhance structures’ collateral value. 5 Djankov et al. (2020) make an important distinction between possessory and transferable rights in the context of urban spatial structure. While weak possession rights reduce the incentives to build better housing or own more land, limited ability to transfer property makes it difficult to match workplace with home location and enable redevelopment of urban land. When possessory rights are limited, residents must either invest in self-protection or risk losing their property, and both the costs and the risk scale up with investment: thus, residents invest less, they consume lower quality housing, and urban density levels are higher. When transfer rights are limited, people are stuck in place even when economic circumstances change. 80 Pancakes to Pyramids | City Form to Promote Sustainable Growth Manage land valuation and prices The key to economically efficient urban land even have such rolls (World Bank 2015a). In use is the land market. Because a land parcel’s Malawi, only ratable areas are listed and valued initial use may become less efficient over time, for tax purposes, even though some nonratable efficiency requires land to shift among various areas have become indistinguishable from uses and thus among various owners: such ratable areas. As a result, Lilongwe City shifts are central to a city’s expansion and Council’s property valuation roll is estimated to infill development. Although land use generally list about 45 percent of the properties in the should be as free of restrictive regulation as city, and Blantyre’s lists about a third (World possible, public intervention is sometimes Bank 2016b). In Ghana, property valuations required to offset market failures (Henderson have not changed in the past 15–20 years and Wang 2007; World Bank 2013). For (World Bank 2015b). example, while land markets allocate land between urban and rural uses, governments Many developing countries thus lack the may create incentives to conserve farmland capacity to systematically record and manage and green space. Similarly, while markets information on land transactions—and where allocate land among various urban uses, transaction data exist, they may not reflect governments need to legislate and regulate to the true price of land, whether because of prevent neighborhoods from being underserved widespread informality in land transfers (to and to prevent disruptive land use. save on duties) or because of heavy public subsidies on housing and land use. With no The market’s efficient allocation of land use credible system to discover and disseminate requires not only legal mechanisms for land land values, the risk of land undervaluation is transfer—including clear titling and tenure substantial, as buyers may attempt to defraud arrangements, discussed in the subsection existing landowners. But even honest buyers above—but also ways for market actors to can face challenges in determining a fair price land parcels in accordance with their offer: too often, developing country cities lack fair value. In developed countries, markets ancillary data to indicate a parcel’s income determine land values and prices through an generation potential and its development examination of property attributes and market input costs. data from similar transactions. Governments manage existing data on land prices to Public initiatives to improve land valuation provide up-to-date and reliable information for systems can promote economic development. professional appraisers as well as the general During the 1970s, the Republic of Korea public. brought transparency to land valuations and made information on land values widely In developing countries, by contrast, market accessible by encouraging the development actors face many challenges in valuing and of a cadre of property appraisers. In previous pricing land. One common obstacle to land land acquisitions, local government officials valuation and pricing in developing countries had assessed market values and asset is a lack of basic institutions. If land registries replacement costs. In 1972, the government exist, they are likely to be archaic, lacking the introduced the Basic Land Prices system, dynamic functionality that allows them to be which mandated the assessment of land searched or updated quickly. For example, and buildings by certified private appraisers. Kenya’s valuation and rating system has Estimated property values from two appraisers not been updated since colonial times, and were averaged for a final value. If the two property rolls are outdated: Mombasa’s was appraisals differed by more than 10 percent, a last updated in 1992, Nairobi’s in 1981 (World third private appraiser was selected, and a new Bank 2016a). Some cities in Ethiopia do not average calculated (World Bank 2013). Pancakes to Pyramids | City Form to Promote Sustainable Growth 81 City leaders in several developing countries developed countries for various reasons, not are now making similar efforts to improve all of them dictated by economic efficiency land valuation systems. Between 2008 and or by the public good: in the United States, 2010, the city of Bogota updated its cadastral tax assessors’ calculated property values are database, revaluing the 2.1 million properties assumed to reflect around two-thirds of market it contains and generating a new revenue value.) stream of $171 million annually (Ruiz and Vallejo 2010). To complete the cadastral The dearth of publicly available data on land update, the government introduced information and property prices in developing countries technology, consulted with stakeholders, prevents analysis that is critical for appraisals, and began estimating property values not only by market actors seeking to value using spatially detailed information from GIS land, but by governments seeking to tax real systems (Uribe 2010). Because no property property and sales of land. The result is to transaction information was available, a team deprive cities of public revenues that could of expert appraisers collected price data be raised through local financing mechanisms using a combination of approaches to yield an involving real estate and infrastructure. The appraised value. But to keep property taxes integral role of land valuation in local public progressive—and to avoid resistance from revenue generation—whether through taxation property owners—the city imposed a cap on or through the sale or lease of public land—is property tax increases. (Such caps exist in discussed in a separate section below. Infrastructure planning for economic density, livability, and sustainability—scaling up and evaluating investments Linked to the functioning of a city’s land Understand path dependence markets is its demand for physical structures, infrastructure, housing, and amenities—and To attain pyramidal growth, urban development some parts of this demand are especially needs to follow a path that makes financing tricky to meet, because they need to be as feasible as possible. It is critical for leaders anticipated early. To set the stage for future to recognize that not all paths are equally economic growth and pyramidal development, affordable, and that cities are less likely to city leaders in developing countries not only reach the goal if they set out in the wrong need to plan for the present: they need to plan direction. New housing, infrastructure, and far ahead. industrial premises will vary widely in cost according to how they are sequenced. All three The emergence of thriving, livable, sustainable will be least expensive and most feasible if cities depends in large part on physical they are made in the following order: structure and infrastructure investments, 1. Infrastructure, planned to set high which come with particular challenges. One is expectations for future economic path dependence. Another is interdependence. development stages. These challenges imply a need for thorough coordination among infrastructure investments, 2. Housing, enabled by connective land use plans, and zoning regulations. And infrastructure, a functioning land market, whichever viable options are chosen, they and demand-responsive land use must begin with early infrastructure investment regulation. to set cities on a path toward density and 3. Industrial premises, attracted by efficient productivity. connections inside and outside the urban 82 Pancakes to Pyramids | City Form to Promote Sustainable Growth area, by agglomeration economies, and by structures will be built nearby. Finally, these a reliable labor supply in a prosperous and predictions are self-fulfilling. As expectations livable city. affect investments, so investments affect expectations. The first structures built in a Why invest in infrastructure first? Because neighborhood or city will dictate the options sewerage, drainage, electricity, clean water, and for further investments in the vicinity. This internet connectivity are cheaper if they are circularity makes the challenges of path provided all at once—at full scale—than if they dependence and interdependence are all the are added individually to houses and factories more pressing. in a piecemeal fashion over time (Collier 2016). Furthermore, urban structures share a “putty Coordinate infrastructure clay” quality: once constructed, they are investments, land use plans, and difficult to modify and can stay in place for more than 150 years (Hallegatte 2009). zoning regulations For all the reasons discussed above, Account for the interdependence effective coordination is vital to the success of infrastructure and physical of developing cities at nurturing economic agglomerations and making the urban structures environment livable and sustainable. To manage Another challenge is the interdependence of public infrastructure investments with private investments in physical structures. For firms, the productivity of premises depends Box 4.4 Using the new urban model to on proximity to infrastructure, workers, and scrutinize received assumptions about the customers—a proximity defined not just by economic consequences of urban policies physical distance but by transportation and investments options. For households, the utility of housing depends on firms’ investments in accessible Given a typical developing city’s large need for urban jobs—an accessibility that also reflects infrastructure and limited financing capacity, city leaders transport network planning. However, for must carefully examine the likely impacts of policies and the city that must finance transportation investments. Efforts need to be directed where they infrastructure and decide on public transit can have the greatest positive impact, and unintended investments, a rapid transit system is more consequences must be anticipated and prevented. viable where both economic and population The new urban model outlined in chapter 3 of this densities are higher. report can help leaders compare the implications of various options for a city’s productivity, livability, and sustainability. This economic intelligence can Alongside this interdependence of public augment other sources of planning intelligence (such transport investment choices and firm and as the CAPSUS and CALTHORPE models, among many household location decisions, any additional others)—especially in scrutinizing received assumptions social returns to infrastructure—such as about the consequences of various urban policies and interventions. sustainability—will also reflect the proximity of dense residential neighborhoods to hubs of For example, will greenbelt policies increase economic efficiency and aggregate welfare? Or will they constrain economic activity where firms are located. In economic growth and ultimately impede a city’s addition, much of a structure’s financial value pyramidal transformation? Again: will land market reforms is determined by complementarities with other help poor people? If so, under what conditions? It is structures in the neighborhood or city. Yet path plausible that increased aggregate urban efficiency may dependence makes these complementarities come at the cost of poorer residents’ displacement to less efficient neighborhoods (for example, with poorer a matter of prediction as well as observation: access to jobs and amenities). So, if the city does not investors need to anticipate what other already provide mechanisms to compensate these residents, critical complementary policies may be called for to protect them. Pancakes to Pyramids | City Form to Promote Sustainable Growth 83 the twin challenges of urban path dependence grew rapidly during a period when the and interdependence, city leaders must seek government was unable to make investments— synergies among infrastructure investments, the civil war decade of 1991–2002. Now that land use plans, and zoning regulations. A all urban spaces are settled, local opposition to government’s ability to make early, coordinated road construction is perpetuating a severe lack public investments will directly influence later of road infrastructure. decisions by firms weighing their own private investments in the urban economy. Only Sites and services projects: Lessons efficient infrastructure and service provision from World Bank experiments with early will generate economic density and improve infrastructure investment. During the 1970s livability, job market matching, and productivity. and 1980s, the World Bank used what it called “sites and services” projects to install Especially important is to avoid coordination infrastructure ahead of urban settlement failures in which single-sector interventions growth. Undertaken in many cases to hinder urban economic density. Inefficient prevent slum formation—or to set up durable structures can set back productivity growth foundations for upgrading slums to formal and spatial transformation for decades. The neighborhoods—the projects covered more new urban model can help leaders review than 20 urban neighborhoods in Brazil, El options that, though in use elsewhere, Salvador, Jamaica, Peru, Senegal, Tanzania, could warrant skepticism—or could require Thailand, and Zambia, including tens of complementary policies to be implemented thousands of households. Some sites and efficiently (box 4.4). services projects made new investments in undeveloped land, while others focused on Make infrastructure investments upgrading existing slums. Both project types comprised infrastructure investment in roads, as early as possible—while electricity, water, and public buildings such coordinating them with urban plans as schools, clinics, and community centers. Infrastructure will guide the course of a High costs led to the projects’ discontinuance city’s development. Besides determining during the late 1980s, despite anecdotal where structures with various uses can be evidence of beneficial long-term impacts. located, it is also a signal to investors about the future functions of areas around the city. A recent research study examines the Infrastructure is thus a coordinating device—an longer-term benefits of these sites and irreversible, and therefore credible, commitment services projects and is expected to find them that is highly visible and so shapes private financially efficient, while also identifying a investors’ expectations and decisions. possible distinction between the benefits of the earlier infrastructure investments If infrastructure is postponed until after represented by new development projects population settlement, the investment and the later interventions entailed by slum will prove far more costly, and it may also upgrades (Michaels et al. forthcoming). The encounter political obstacles. Services costs and benefits that researchers examined that must be placed underground are less are of two sorts: effects on land values in expensive to install at scale on clear sites, certain areas, and physical effects on the rather than retrofitted beneath (or over or urban landscape. The researchers believe around) existing structures on previously that, in the long run, the sites and services developed land. Furthermore, belated projects tended to increase property values infrastructure initiatives pose a prospect of and to raise a city’s tax base. Moreover, the disruption to private homes and can face new construction projects likely realized higher public resistance. In Sierra Leone, Freetown long-term benefits than the slum upgrades did. 84 Pancakes to Pyramids | City Form to Promote Sustainable Growth In Dar es Salaam, for example, sites with new of figure 4.2—a part of Dar es Salaam where development projects have higher land values roads are disorganized, plots are small and than land elsewhere in the city—including rich irregular, and land is not even valued for tax neighborhoods—partly because the sites and collections (the tax benefit would not justify services areas have a higher average ratio of the assessment cost). Although drawn in the building footprint to plot area (figure 4.1). 1970s, the sites and services plans for Dar es Salaam closely match the shape of the city’s The study shows further that plots are now road network today. This durability confirms larger where investments were made ahead that investment in infrastructure determines of settlement: an example is Sinza, the area of a city’s future course, while leading to higher Dar es Salaam in the upper left of figure 4.2. land values that are taxable and can finance The same plots have higher land values per future investments. square meter than projects in upgraded slum areas, such as Manzese in the bottom right Figure 4.1 Land values of new development projects are higher than values in other neighborhoods of Dar es Salaam, including rich ones 40 Median value per square meter Sites and services 35 (Tanzania shillings) 30 25 Other areas 20 15 10 0 Kijitonyama Mikocheni Sinza Mikocheni (old) Magomeni Msasani Mwananyamala Mazimuni Ndugumbi Source: Michaels et al. forthcoming. Another recent study examined sites and • Tiny plots compared with standard services projects in India (Owens, Gulyani, plots of the time. The smallest sites and Rizvi 2018). The study found that these and services project plot was 33 m2 in projects resulted in the creation of well- Chennai, 21 m2 in Mumbai. In contrast, planned, well-serviced neighborhoods that are other housing developments in these both mixed-income and mixed-use. In Chennai cities had minimum plot sizes of about and Mumbai, the success of the sites and 150–200 m2. The smaller plots were far services projects hinged on four key features more affordable and allowed lower-income (Gulyani 2016): households to enter the housing market. Pancakes to Pyramids | City Form to Promote Sustainable Growth 85 Figure 4.2 Differential impacts of new development projects and upgrading projects in Dar es Salaam Source: Michaels et al. forthcoming. Note: The upper left of the photograph shows Sinza, a new development project. The bottom right shows Manzese, an upgrading project. • Spatially efficient site planning norms. from 21 to100 m2. Today these are true These site plans lowered the unit mixed-income neighborhoods, with lower- costs of developed plots while further income families occupying smaller plots increasing urban density. For example, and middle- and high-income families only 34 percent of land was allocated occupying larger ones. to streets and open spaces, compared with the 50–60 percent often seen in • Design for mixed use. The sites and other developments in India at the time. services projects in India included Even so, average road density in these commercial areas (shops), amenities neighborhoods exceeds that of the (schools, clinics), and locations for light parent city as a whole. Smart planning industrial production. All these types of thus lowered the cost of infrastructure businesses, services, and amenities exist provision and individual housing plots in the neighborhoods today. Mixed use while creating compact, walkable, livable has resulted in vibrant streets. neighborhoods. This evidence from Chennai, Mumbai, and Dar • Varied plot sizes, affordable to different es Salaam suggests that sites and services income groups. In Chennai, the plot sizes projects are potent instruments for managing ranged from 33 to 223 m2, in Mumbai urban expansion and creating affordable 86 Pancakes to Pyramids | City Form to Promote Sustainable Growth housing. City governments can use the sites residents. In particular, access to potable and services approach and planning norms to water and safe sanitation is essential to limit shape future urban growth—moving beyond the incidence of waterborne diseases and “putting stakes in the ground,” and using early enable a healthy population. Also critical infrastructure investment to earmark future are transportation options to connect neighborhoods. Governments and the private residents to jobs around the city. All these sector can create more affordable housing by investments are expensive—but they are scaling up delivery of small housing plots, on most expensive when they require retrofitting which families can build incrementally. previously settled, yet underserviced, neighborhoods. They are less expensive in The need to ensure basic services for undeveloped areas where the government all. Whether urban expansion (horizontal has previously secured its rights-of-way and spread) or redevelopment (infill) is planned, does not need to disrupt residents’ lives. city leaders must ensure that infrastructure and basic services are provided for all Durable financing for capital investment costs and recurring expenses— mobilizing urban revenue sources Achieving economic density will require Tax the value of urban cities in developing countries to make huge infrastructure investments (box 4.5). How can land and property city leaders bridge this financing chasm? What Urban land values can offer a durable basis revenues can they tap? for urban infrastructure financing. In cities in developing countries, municipal own-source In most developing countries, urban public revenues—especially from property taxes— finance to date has relied on intergovernmental are a generally underdeveloped financing fiscal transfer systems, which account source. In Mumbai, for example, property for about 60 percent of all subnational taxes constitute just 12 percent of total local expenditures in developing and emerging government revenues (Farvacque-Vitkovic economies (Shah 2006). In Uganda, the and Kopanyi 2014). A common reason for share of local government revenues coming low or nonexistent land and property taxes from central government grants reaches as is the absence of clear land ownership and much as 91 percent (Farvacque-Vitkovic and tenure rights: as discussed above, city leaders Kopanyi 2014). Such fiscal transfers are mostly should make it a top priority to strengthen untransparent and unpredictable. In addition, titling institutions, with a cadaster system that they involve a complex political economy that creates incentives for citizens to keep property may not favor cities and is typically biased and tenure information up to date. Without toward rural areas and smaller settlements. clear land records and land rights, land and property taxes cannot be relied on as a source A better financing approach is to tax the value of municipal revenue. of land and property for initial infrastructure investments, for the recurring cost of Land and property taxes are especially fruitful infrastructure (operation and maintenance), and when cities experience rapid economic growth: for the provision of public goods and services. steeply rising productivity and incomes lead Other financing options include private sector to steeply rising land prices, which will yield investment in service delivery through public– substantial revenue if the right valuation and private partnerships (PPPs), and borrowing on taxation enforcement mechanisms are in place capital markets for creditworthy cities. (box 4.6). This process is circular, however. Pancakes to Pyramids | City Form to Promote Sustainable Growth 87 While potentially generating increased Borrow on capital markets—if the revenues, rapid urban economic growth city’s creditworthiness allows also creates new infrastructure needs and thus calls for greater investment resources. The bar for borrowing on capital markets is Infrastructure financing in France, Japan, and even higher than that for securing private the United States was based most heavily investment through PPPs. Among the 500 on land values during periods of rapid urban largest World Bank client cities in developing growth, when urban investment made rapid and emerging economies, only 90 (18 percent) leaps in scale. are deemed creditworthy on domestic and international credit markets—and just 32 Consider public–private partnerships (6 percent) have ever issued a municipal bond.6 The reasons for this general lack of (PPPs) for service delivery creditworthiness are manifold: the lack of A city’s ability to attract private investment in buoyant municipal revenues, the lack of infrastructure and service delivery will hinge expenditure rationalization, the presence on its ability to develop a robust regulatory of debt and contingent liabilities, the lack framework for public–private partnerships of asset management strategies, and the (PPPs). From the city’s perspective, the ability lack of a national framework governing to design PPPs and enforce contracts will creditworthiness are among the many be critical. From the investor’s perspective, contributing factors. a key requirement will be the availability of adequate information, with credit enhancement mechanisms to attract investment. 6 World Bank City Creditworthiness database. Box 4.5 The global urban infrastructure financing gap: A challenge for developing country cities The gap in urban infrastructure (World Bank 2016a), which means such global figures include both is huge. Today, 156 million urban that in a one-hour commute, they developed and developing countries inhabitants live without access to can barely access 14 percent of the and both urban and rural space, the improved water sources and 700 jobs in the city (Avner and Lall 2016). annual gap in urban infrastructure million without improved sanitation.1 financing for developing countries Urban infrastructure financing In particular, slums and informal may be roughly guessed to reach shortfalls in developing countries settlements often lack individual at least $1.5–2 trillion. That and emerging economies are not access to water and sanitation, conservative figure dwarfs all the rigorously or coherently estimated. and the use of public facilities official development assistance that For the cost of reaching the exposes them to contagion risk. Poor is available globally every year. Sustainable Development Goals residents of informal settlements globally, the Global Infrastructure While much of the infrastructure often resort to purchasing water from Outlook estimates the total annual deficit will be in developing countries, informal vendors at much higher unit infrastructure financing gap at existing infrastructure in the rates—often five times as much as $3.7 trillion (Global Infrastructure developed world suffers in many what other households pay when Hub 2017). In contrast, the State countries from deferred maintenance, purchasing water from the municipal of the City Climate Finance Report and thus requires important capital system (Klein 2012). Underinvestment estimates the global annual gap investment. The United States alone is in public transportation and the at $4.1–4.5 trillion—rising to $4.5– estimated to need nearly $0.5 trillion lack of affordability means that 5.4 trillion if one were to include a of infrastructure investment a year poor people often have no means 9–27 percent premium to enhance between today and 2040, and Europe of getting to work except walking. the resilience of the infrastructure another $0.6 trillion a year (Global For instance, 48 percent of the (United Nations 2015). Given that Infrastructure Hub 2020). population in Nairobi walks to work 1 https://www.unwater.org/water-facts/urbanization/. 88 Pancakes to Pyramids | City Form to Promote Sustainable Growth Fix the fundamentals—fiscal, When connective infrastructure investments succeed, they help cities improve their economic, and institutional—to economic fundamentals: efficient connections mobilize urban finance foster agglomeration economies, which To mobilize finance, municipal governments increase local revenue potential. and utilities need to demonstrate the ability and commitment to pay for that finance Whatever a city’s economic development with funding.7 And funding, in a municipal stage, and whatever its productivity and environment, comprises either local (own- income level, sound municipal finance can source) revenues, generated from taxes and exist only where property rights are clear service charges, or fiscal transfers, including and urban land markets function efficiently. aid grants. The greater the volume of private For developing country cities, effective titling finance, the greater the need for funding. and tenure reforms to clarify land rights Financing more means funding better. can underpin an effective municipal asset management strategy by supporting an up-to- Fixing the fiscal fundamentals of municipalities, date land and property cadaster at the base utilities, and other public services is thus of a tax system. Only when cities can thus a core part of mobilizing finance for urban secure funding, and can use funding to obtain development. New investments in high-quality financing, can they empower themselves to and affordable transport systems depend on achieve their full potential—as dense, efficient fiscal as well as financial sustainability, not economic agglomerations that are livable for only for the initial capital investment, but also people and sustainable for the world. to fund ongoing operations and maintenance.8 7 White and Wahba (2019) make an important point that financing and funding are two different things. Finance is raising money for investment. Funding is the payment for the investment, including the financing cost, over the long term. Finance thus does not obviate the need for funding. Indeed, because finance comes at a price (interest or return on equity), it aggravates the funding need. 8 Urban transport financing needs to be based on an appropriate mix of complementary financing instruments (Ardila-Gomez and Ortegon-Sanchez 2016). Several solutions exist to enable cities to better capture the value created through integrated land use and transport planning (Salat and Ollivier 2017; Ollivier et al. 2021). Box 4.6 Taxing land and property values to finance urban infrastructure Taxes on land values can fund constitutes an economic rent for a • Monetization of underused pub- infrastructure for dense urban scarce resource—not a return on lic land. development. They can also make any economic activity by the owner Creating a land and real estate tax land use more efficient: taxes based (there is no economic behavior to be system to support urban development on market valuations give property distorted). and densification will be a challenge owners an incentive to develop land to for most developing country cities. Higher revenues from land and real its most profitable use commensurate Strong institutions are essential to estate can be realized in three ways: with its market value. define property rights clearly, to ensure • Increased valuation of land and standardized and objective methods When valuable downtown locations properties after infrastructure is of land valuation, and to support and with higher land prices are taxed in planned, bringing assessed val- oversee land management, land sales, accordance with their value, they will ues closer to market values and and tax collection. In addition, planners attract greater investment in residential thus deepening the tax base. considering sole reliance on property and commercial structures, making the city center more residentially and • Improved compliance, causing taxes as an option should be aware economically dense. As a result, land more property owners to pay that property values respond only values will rise further. Yet the resulting land and property taxes and slowly to annual changes in economic rise in taxes is nondistortionary, thus broadening the tax base. activity. because appreciation in land values Pancakes to Pyramids | City Form to Promote Sustainable Growth 89 Annex 1 Income elasticity of population density The evolution of built-up area has implications for population density. We estimate income and population elasticities of density measured as population over built-up area. Using the Global Human Settlement Layer panel data, different specifications are estimated including nonlinear effects for the largest cities based on the following regression: As shown previously, as cities become richer, their demand for floor space rises and their supply of built-up area increases in response, even with population held constant. So as cities expand in built-up area, they become less densely populated. The income elasticity of density is −0.10 using the between and within estimator; introducing instruments for incomes increases it to −0.15. Table A1.1 provides estimates of the income elasticity of density using various specifications. Table A1.1 Determinants of naïve population density (population over built-up area) Dependent Between + Between + within Between + within For cities > Nonlinear variable: within estimator estimator with lag estimator with IV 500,000 only specification for for income cities > 500,000 Population density over built-up area Ln Income −0.105*** −0.149*** −0.0762*** −0.397*** (0.00484) (0.00439) (0.0103) (0.0946) Ln Population 0.649*** 0.627*** 0.784*** −0.446 (0.0177) (0.00856) (0.0277) (0.380) Ln lagged Income −0.181*** (0.00650) Ln lagged 0.478*** population (0.0347) Square Ln Income 0.0197*** (0.00584) Square Ln 0.0432** Population (0.0132) Constant 3.270*** 5.699*** 3.860*** -0.615 9.400*** (0.218) (0.424) (0.119) (0.402) (2.768) FE City + Year City + Year City + Year City + Year City + Year Controls No No No No No R-squared 0.415 0.328 0.466 0.477 N. of obs 28352 18856 27460 2495 2495 * p < 0.05, ** p < 0.01, *** p < 0.001. Note: Standard errors in parentheses. 90 Pancakes to Pyramids | City Form to Promote Sustainable Growth Annex 2 Data and methodological details This report examines spatial development (GHS-BUILT), a new global definition of cities, across 9,500 cities worldwide by synthesizing and socioeconomic characteristics such as Global Human Settlement Urban Center income and population. By combining such Database 2015 (GHS-UCDB) and World data, the GHS-UCDB are comparable across Settlement Footprint 3D product (WSF- time and countries. Due to the nature of remote 3D) retrieved satellite imagery. To apply a sensing-derived information; however, there are consistent measure of cities, this report concerns about measurement errors. Here, we uses a new methodology endorsed by the explore comparative advantages and potential UN statistical commission—the Degree of uncertainties in using GHS-BUILT and Degree Urbanization—that allows for international of Urbanization, and illustrate how we address comparisons of urban growth across time and several concerns data and measurement countries. concerns. The GHS-UCDB is a combined dataset based on the Global Human Settlement Built-Up grid Global Human Settlement Built-Up grid (GHS-BUILT) Comparative advantages resolution. While there are different global urban mapping products such as Moderate of GHS-BUILT Resolution Imaging Spectroradiometer (MODIS) GHS-BUILT is a satellite imagery-derived Land Cover Type (MCD12Q1) and Global Land dataset to measure built-up area. The dataset Cover (Globcover), the spatial resolutions constructed a series of built-up layers across of those products are 500 m and 300 m, four different periods (1975, 1990, 2000, and respectively. This coarse resolution may lead 2015). Each year of the dataset is based on to omission and commission errors. The recent the different collections of Landsat satellite development of the global land cover product— imagery. A total of 33,202 scenes were Copernicus Global Land Service offers a implemented to produce the multitemporal 100 m resolution built-up layer; however, this layers of built-up (Florczyk et al., 2019). has limited temporal coverage (2015-2019). Another Landsat satellite-based built-up layer, GHS-BUILT applies consistent measurement GlobeLand30 (30 m spatial resolution), is also for detecting the built-up layer, globally. A only available after 2000. building extraction algorithm was applied to global imagery to extract full information of GHS-BUILT is accurate across diverse urban built-up areas rather than sampled areas in landscapes. Diverse approaches were applied selected regions. This consistent measurement to test the accuracies of GHS-BUILT. Yang et allows us to explore the presence of built-up al. (2019) test the accuracy of different urban without consideration of the geographical land products, including GlobeLand30, Global coverage constraints. Since cities in developing Urban Footprint (GUF), and GHS-BUILT, across countries often grow extensively within a different urban settings. The overall accuracy relatively short period, it is critical to capture a levels are 0.87, 0.81, and 0.89, respectively. full picture of built-up. Given the limited temporal resolution of GlobeLand30 and GUF and their lower Higher spatial resolution compared to other accuracy, GHS-BUILT is a suitable product for global urban mapping products. The GHS-BUILT this report. Extensive sampling and validation constructed the built-up layers of 30 m spatial Pancakes to Pyramids | City Form to Promote Sustainable Growth 91 done by Liu et al. (2018) reported moderate coarse resolution of Landsat 1-2 Multispectral accuracies in 1990, 2000, and 2014 as 0.72, Scanner (60 m) compared to others (30 m), 0.72, and 0.71, respectively. Leyk et al. (2018) the 1975 layer potentially underestimates also confirmed the higher accuracy of GHS- the built-up areas. To alleviate measurement BUILT in rural-to-urban transition areas. concern, we dropped the 1975 layer for the analysis. Concern with GHS-BUILT Under- and over-estimation of built-up Multiple Landsat missions and different areas. While GHS-BUILT has higher accuracy satellite sensors may deliver inconsistent compared to other global products, one results. The Landsat imagery used for GHS- might still have concerns of under- and BUILT varies across periods. The 1975 layer over-estimation of built-up areas. Based on is based on 7,597 scenes from the Landsat the extensive validation based on 40 million 1-3 Multispectral Scanner. The 1990 layer individual building polygons from 277 different used 7,375 scenes from the Landsat 4-5 cities, Corbane et al. (2019) reported the Thematic Mapper-TM. The 2000 layer used reliable accuracies across different regions 8,788 scenes from the Landsat 7 Enhanced (Oceania: 0.82, Africa: 0.79, Europe: 0.78, Asia: Thematic Mapper+. The 2014 layer used 9,442 0.76, and America: 0.77). While the field of scenes from the Landsat 8 Operational Land remote sensing keeps evolving with the new Imager (Florczyk et al., 2019). Since the series sensor developments such as GHS-BUILT S2 of Landsat satellites were designed to maintain based on Sentinel-2 satellite (10 m spatial compatibility across different Landsat missions, resolution), at this point, the GHS-BUILT is the spectral bands’ wavelength and resolution only product fitted to the analytical goals of should be similar (Corbane et al., 2019). the report in terms of spatio-temporal scales. In theory, GHS-BUILT was not affected by different Landsat missions. However, due to the Degree of Urbanization The Degree of Urbanization is a method for built-up areas but also population density. In delineating cities, urban, and rural areas the case of cities in developing countries, a endorsed by the UN Statistical Commission relatively low share of impervious areas often in March 2020 (Dijkstra et al., 2020). An urban hampers identification of cities despite the center is classified based on the condition of cities having experienced a massive population contiguous grid cells (1 km2) that have at least influx. 50 percent built-up and 1,500 inhabitants, and the total urban center population should have Measurement with consistent geographical more than 50,000 persons. boundary. The GHS-UCDB is developed based on the urban areas of 2015. Since Comparative advantages of Degree the database uses a constant geographical boundary across multiple years (i.e., 1975, of Urbanization 1990, 2000, and 2015), the data maintain Consistent measurement for defining urban compatibility across multiple periods, areas, globally. A major strength of this method unlike other urban mapping products that is that it can be applied globally to identify use different geographical boundaries for cities and settlements (Dijkstra et al., 2020). urban areas by period by period. This is a GHS-UCDB provides data on more than 9,500 fundamental difference compared with the cities around the world. This method can better urban mapping work done by Angel et al. capture the cities in developing countries (2016). Since the urban boundary (see because the method is not only based on the Beijing below as an example) defined by 92 Pancakes to Pyramids | City Form to Promote Sustainable Growth Angel et al. has changed over time, this may growth in Beijing is 5.2 percent (from 668 km2 underestimate built-up areas in 1990 and in 1988 to 2,654 km2 in 2013); by contrast, we overestimate built-up areas in 2000 and 2015. calculate annual growth of 0.42 percent (from The omitted built-up areas at the outskirt 1,246 km2 in 1990 to 1,382 km2 in 2015; figure of the cities in 1990 could be automatically A2.1). As mentioned above, having a consistent included in 2000 and 2015. For example, the measurement of urban boundaries enables built-up area outside of Beijing’s 1990 urban us to maintain compatibility across multiple boundary cannot be identified as built-up periods. area. In Angel’s calculation, the annual built-up Figure A2.1 Different urban boundaries in Beijing, China Source: Authors’ depiction, based on Angel et al. (2016) for left image and GHS–Urban Centre Database for right image. Urban area does not only consist of impervious Concerns in use of Degree of areas. Many global urban mapping products estimate the urban area only based on the Urbanization and GHS–UCDB spectral characteristics of each pixel. That is, Arbitrary density threshold. To examine the urban areas often were classified based on density thresholds, three additional robustness their physical status, whether the pixels are checks were carried out: adding pixels having impervious areas (i.e., built-up) or not. However, built-up coverage over 50 percent added to the urban area includes not only built-up urban centers, gap-filling based on neighboring areas but also other amenities. By combining pixels, and excluding pixels having less than 3 the population density parameter and the percent of built-up (Dijkstra et al., 2020). Based gap-filling method, the Degree of Urbanization on these combined methods, the cities defined enables classifying impervious areas and other by Degree of Urbanization are well aligned amenities into urban areas. (91.9 percent), with nationally defined cities with more than 300,000 population globally (ibid.). The UN World Urbanization Prospects (WUP) 2018 also reported similar numbers of cities with at least 300,000 inhabitants across Pancakes to Pyramids | City Form to Promote Sustainable Growth 93 the world (table A2.1). While the numbers GHS‑UCDB, the other five regions are similar of cities in South Asia and North America across four different periods (Dijkstra et al. exhibit relatively large gaps between WUP and 2020). Table A2.1 Comparison between GHS-UCDB and World Urbanization Prospects 1990 2000 2015 Degree of Urbanization 1,167 1,410 1,768 World Urbanization Prospects 976 1,276 1,772 Source: Based on figure 15 in Dijkstra et al. (2020). As another set of robustness checks, we large urban extents and low population density. consider Functional Urban Area (FUA) Since the FUA boundary contains neighboring boundaries defined on the urban centers and areas, it is considerably larger compared neighboring areas belonging to commuting with the GHS-UCDB. The FUAs of high-, zones. This is particularly important for upper-middle-, lower-middle-, and low-income understanding the spatial development of countries are 6.1, 3.6, 2.5, and 2 times larger cities in land-rich countries that have relatively than the areas of GHS-UCDB, respectively. Table A2.2 Built-up area comparison between GHS-UCDB and FUA   Income group 1990 2000 2015 Annual growth rate (%) High-income 108,726 122,023 128,237 0.66 Upper-middle-income 74,982 92,021 104,289 1.32 GHS-UCDB Lower-middle-income 35,969 46,330 51,462 1.43 Low-income 4,716 6,009 6,596 1.34 High-income 168,115 200,791 226,130 1.19 Upper-middle-income 98,904 125,034 152,056 1.72 FUA Lower-middle-income 41,916 54,733 63,164 1.64 Low-income 5,018 6,473 7,366 1.54 However, the total built-up area difference the annual built-up growth rate differences between GHS-UCDB and FUA is relatively small between GHS-UCDB and FUA are small except (table A2.2). While total area of FUA is 4 times for high-income countries. In the case of the larger than total area of GHS-UCDB, the amount United States, the annual built-up growth rate of built-up area in FUA is only 1.5 times larger jumps from 0.85 percent using GHS-UCDB to than in GHS-UCDB. That is, the GHS-UCDB 1.55 percent by FUA. Therefore, other than does not miss a significant portion of urban a few exceptions—high-income land-rich built-up areas. The difference mostly comes countries including the United States, Canada, from high-income and land-rich countries. For and Australia—the GHS-UCDB adequately example, the United States itself accounts for captures urban built-up areas (table A2.3). 31 percent (49,075 km2) of the total built-up area difference (158,132 km2). Additionally, 94 Pancakes to Pyramids | City Form to Promote Sustainable Growth Table A2.3 Built-up area comparison in high-income land-rich countries Square kilometers   Country 1990 2000 2015 Annual growth rate (%) United States 48,353 56,508 60,030 0.87 GHS-UCDB Canada 4,291 4,833 5,313 0.85 Australia 3,701 3,945 4,095 0.40 United States 76,254 97,001 112,456 1.55 FUA Canada 5,917 6,974 8,208 1.31 Australia 5,139 5,703 6,271 0.80 Missing people due to the underestimated WorldPop (2018) was overlaid to the urban extent of cities. Mainly, for rapidly growing boundary layers to calculate the population. cities, one may be concerned if urban centers, defined by the degree of urbanization, are As shown in figure A2.2, except for two able to capture the proper population of the outliers—Guangzhou, China and Jakarta, region. To assess this concern of missing Indonesia—the median values of population population, we develop a sensitivity test for estimates with additional buffered urban the top 50 (spatially) growing cities from 1990 areas are similar to the population within the to 2015. Multiple ring buffers (from negative urban boundary. For instance, the additional 5 km to positive 5 km away from the edge of 5 km buffer increases Mexico City’s urban the urban boundary) were added to the urban area by 74 percent (1,556 km2) while boundary in order to assess the population increasing its population by only 3.3 percent, variations across different sizes of urban areas. and it increases Los Angeles’ urban area by The gridded population dataset retrieved from 57 percent (3,205 km2) while increasing its population by only 4.6 percent. Figure A2.2 Estimated population within the urban area in the top 50 growing cities, cities 2015 1990 and Top 50 growing from 1990 to 2015 50 40 Population (million) 30 20 10 0 -5km -4km -3km -2km -1km Urban +1km +2km +3km +4km +5km boundary Source: Authors’ analysis. Pancakes to Pyramids | City Form to Promote Sustainable Growth 95 World Settlement Footprint (WSF) 3D While the recent development in the field represent the cities’ vertical heights due to of remote sensing enables capturing global their nature: self-reported data in selected urbanization extensively, the 3D structure cities and restricted samples of buildings (i.e., (building height) of the city is rarely height above 80 m). incorporated. Lidar and active remote sensing techniques are often employed to represent To better understand the spatial structure three-dimensional urbanization; however, of cities with a constant measurement, this most of those studies are restricted to report used newly developed global 3D built sampled areas due to the high cost of data environment data. The World Settlement acquisition. For the global level studies, Mahtta Footprint 3D (WSF 3D) is developed by the et al. (2019) explored vertical and horizontal German Aerospace Center (DLR) to derive a urbanization by combining NASA’s QuickSCAT worldwide 3D model of the built environment SeaWinds and Global Human Settlement Layer in a fine resolution (90 m x 90 m) based on (GHSL). However, SeaWinds is not designed the data collected by TanDEM-X satellites. By to measure urban dynamics, so the sensor synthesizing TanDEM-X and the 30 m ALOS cannot properly capture the complex building World 3D elevation model (AW3D30) with the structures in cities. A recent study done by 12 m spatial resolution of Digital Terrain Model Jedwab et al. (2021) used a global building and multi-spectral Sentinel-2 Imagery, DLR database (Council on Tall Buildings and Urban successfully develops a global building height Habitat) to measure cities’ vertical height. database with a vertical resolution of 1 m Unfortunately, these data also cannot fully (figures A2.3, A2.4). Figure A2.3 Schematic process of WSF 3D development ALOS World 3D World Settlement Footprint ALOS World 3D WSF 3D (30 m) (WSF 2015) (10 m) (30 m) (90 m) Sentinel-2 (10 m) Open Street Map (Building) 96 Pancakes to Pyramids | City Form to Promote Sustainable Growth Figure A2.4 Average building height in Chicago, Illinois Source: Author’s construction, retrieved from WSF-3D data. In this report, we collected WSF-3D data for approximately 400 cities around the world. The data represent diverse geographical areas that enable us to explore different spatial structures of cities across regions (table A2.4). Table A2.4 Selected cities for WSF-3D data Regions Number of cities Europe & Central Asia 57 Middle East & North Africa 31 South Asia 66 East Asia & Pacific 124 Latin America & Caribbean 38 Sub-Saharan Africa 45 North America 12 Pancakes to Pyramids | City Form to Promote Sustainable Growth 97 Annex 3 Spatial development of the five largest cities in each region of the world Built-up Urban Built-up Population expansion No. Region City Country areas expansion growth 1990-2015 (km2) (% per year) (% per year) (km2) 1 Tokyo Japan 5,318 385 0.44 0.75 2 Guangzhou China 6,622 1,342 2.65 2.54 East Asia 3 Osaka Japan 3,158 123 0.24 0.09 & Pacific 4 Shanghai China 3,318 905 2.25 3.41 5 Jakarta Indonesia 5,009 418 1.00 2.30 6 London United Kingdom 1,864 65 0.21 1.03 7 Moscow Russia 1,882 164 0.58 1.29 Europe & 8 Paris France 1,638 84 0.30 0.50 Central Asia 9 Dortmund Germany 1,315 71 0.37 –0.23 10 Istanbul Turkey 1,340 129 0.79 2.30 11 São Paulo Brazil 2,005 146 0.44 1.04 12 Buenos Aires Argentina 1,967 223 0.72 1.19 Latin America 13 Mexico City Mexico 2,114 510 2.00 0.59 & Caribbean 14 Rio de Janeiro Brazil 1,367 81 0.43 0.89 15 Monterrey Mexico 667 161 1.86 1.05 16 Tehran Iran 1,382 102 0.65 2.09 17 Cairo Egypt 1,585 119 0.83 1.77 Middle East & 18 Riyadh Saudi Arabia 1,016 136 0.96 3.05 North Africa 19 Baghdad Iraq 787 72 0.72 1.64 20 Kuwait City Kuwait 476 22 0.29 2.56 21 Los Angeles 5,633 643 0.60 0.89 22 New York 5,384 481 0.56 0.26 North United States 23 Chicago 3,830 414 0.64 0.21 America of America 24 Dallas 3,699 788 1.44 1.77 25 Houston 3,418 804 1.58 2.10 26 Kolkata India 2,817 113 0.39 0.92 Delhi 27 India 2,474 443 1.82 2.25 (New Delhi) South Asia 28 Dhaka Bangladesh 3,248 195 1.02 3.48 29 Hyderabad India 874 168 2.48 1.65 30 Mumbai India 1,077 32 0.36 1.25 31 Johannesburg South Africa 1,638 126 0.61 3.07 32 Lagos Nigeria 1,196 165 0.91 2.60 Sub-Saharan 33 Accra Ghana 846 84 0.74 3.33 Africa 34 Cape Town South Africa 697 42 0.46 2.75 35 Durban South Africa 745 61 0.82 1.25 98 Pancakes to Pyramids | City Form to Promote Sustainable Growth East Asia & Pacific Tokyo, Japan Built-up in 1990 Built-up in 2015 0 10 20 40 km Annual built-up expansion 0.44% Total built-up expansion (1990–2015) 385 km2 Annual population growth 0.75% Built-up area per person (2015) 108 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 99 East Asia & Pacific Guangzhou, China Built-up in 1990 Built-up in 2015 0 10 20 40 km Annual built-up expansion 2.65% Total built-up expansion (1990–2015) 1,342 km2 Annual population growth 2.54% Built-up area per person (2015) 61 m2 100 Pancakes to Pyramids | City Form to Promote Sustainable Growth East Asia & Pacific Osaka, Japan Built-up in 1990 Built-up in 2015 0 10 20 40 km Annual built-up expansion 0.24% Total built-up expansion (1990–2015) 123 km2 Annual population growth 0.09% Built-up area per person (2015) 133 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 101 East Asia & Pacific Shanghai, China Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 2.25% Total built-up expansion (1990–2015) 905 km2 Annual population growth 3.41% Built-up area per person (2015) 64 m2 102 Pancakes to Pyramids | City Form to Promote Sustainable Growth East Asia & Pacific Jakarta, Indonesia Built-up in 1990 Built-up in 2015 0 10 20 40 km Annual built-up expansion 1.00% Total built-up expansion (1990–2015) 418 km2 Annual population growth 2.30% Built-up area per person (2015) 49 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 103 Europe & Central Asia London, United Kingdom Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.21% Total built-up expansion (1990–2015) 65 km2 Annual population growth 1.03% Built-up area per person (2015) 131 m2 104 Pancakes to Pyramids | City Form to Promote Sustainable Growth Europe & Central Asia Moscow, Russia Built-up in 1990 Built-up in 2015 0 10 20 40 km Annual built-up expansion 0.58% Total built-up expansion (1990–2015) 164 km2 Annual population growth 1.29% Built-up area per person (2015) 82 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 105 Europe & Central Asia Paris, France Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.30% Total built-up expansion (1990–2015) 84 km2 Annual population growth 0.50% Built-up area per person (2015) 117 m2 106 Pancakes to Pyramids | City Form to Promote Sustainable Growth Europe & Central Asia Dortmund, Germany Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.37% Total built-up expansion (1990–2015) 71 km2 Annual population growth −0.23% Built-up area per person (2015) 224 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 107 Europe & Central Asia Istanbul, Turkey Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.79% Total built-up expansion (1990–2015) 129 km2 Annual population growth 2.30% Built-up area per person (2015) 47 m2 108 Pancakes to Pyramids | City Form to Promote Sustainable Growth Latin America & Caribbean São Paulo, Brazil Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.44% Total built-up expansion (1990–2015) 146 km2 Annual population growth 1.04% Built-up area per person (2015) 71 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 109 Latin America & Caribbean Buenos Aires, Argentina Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.72% Total built-up expansion (1990–2015) 223 km2 Annual population growth 1.19% Built-up area per person (2015) 95 m2 110 Pancakes to Pyramids | City Form to Promote Sustainable Growth Latin America & Caribbean Mexico City, Mexico Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 2.00% Total built-up expansion (1990–2015) 510 km2 Annual population growth 0.59% Built-up area per person (2015) 55 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 111 Latin America & Caribbean Rio de Janeiro, Brazil Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.43% Total built-up expansion (1990–2015) 81 km2 Annual population growth 0.89% Built-up area per person (2015) 79 m2 112 Pancakes to Pyramids | City Form to Promote Sustainable Growth Latin America & Caribbean Monterrey, Mexico Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 1.86% Total built-up expansion (1990–2015) 161 km2 Annual population growth 1.05% Built-up area per person (2015) 102 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 113 Middle East & North Africa Tehran, Iran Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.65% Total built-up expansion (1990–2015) 102 km2 Annual population growth 2.09% Built-up area per person (2015) 52 m2 114 Pancakes to Pyramids | City Form to Promote Sustainable Growth Middle East & North Africa Cairo, Egypt Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.83% Total built-up expansion (1990–2015) 119 km2 Annual population growth 1.77% Built-up area per person (2015) 30 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 115 Middle East & North Africa Riyadh, Saudi Arabia Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.96% Total built-up expansion (1990–2015) 136 km2 Annual population growth 3.05% Built-up area per person (2015) 106 m2 116 Pancakes to Pyramids | City Form to Promote Sustainable Growth Middle East & North Africa Baghdad, Iraq Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.72% Total built-up expansion (1990–2015) 72 km2 Annual population growth 1.64% Built-up area per person (2015) 76 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 117 Middle East & North Africa Kuwait City, Kuwait Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.29% Total built-up expansion (1990–2015) 22 km2 Annual population growth 2.56% Built-up area per person (2015) 96 m2 118 Pancakes to Pyramids | City Form to Promote Sustainable Growth North America Los Angeles, California Built-up in 1990 Built-up in 2015 0 10 20 40 km Annual built-up expansion 0.60% Total built-up expansion (1990–2015) 643 km2 Annual population growth 0.89% Built-up area per person (2015) 315 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 119 North America New York, New York Built-up in 1990 Built-up in 2015 0 10 20 40 km Annual built-up expansion 0.56% Total built-up expansion (1990–2015) 481 km2 Annual population growth 0.26% Built-up area per person (2015) 220 m2 120 Pancakes to Pyramids | City Form to Promote Sustainable Growth North America Chicago, Illinois Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.64% Total built-up expansion (1990–2015) 414 km2 Annual population growth 0.21% Built-up area per person (2015) 396 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 121 North America Dallas, Texas Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 1.44% Total built-up expansion (1990–2015) 788 km2 Annual population growth 1.77% Built-up area per person (2015) 459 m2 122 Pancakes to Pyramids | City Form to Promote Sustainable Growth North America Houston, Texas Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 1.58% Total built-up expansion (1990–-2015) 804 km2 Annual population growth 2.10% Built-up area per person (2015) 437 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 123 South Asia Kolkata, India Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.39% Total built-up expansion (1990–-2015) 113 km2 Annual population growth 0.92% Built-up area per person (2015) 54 m2 124 Pancakes to Pyramids | City Form to Promote Sustainable Growth South Asia New Delhi, India Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 1.82% Total built-up expansion (1990–2015) 443 km2 Annual population growth 2.25% Built-up area per person (2015) 40 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 125 South Asia Dhaka, Bangladesh Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 1.02% Total built-up expansion (1990–2015) 195 km2 Annual population growth 3.48% Built-up area per person (2015) 33 m2 126 Pancakes to Pyramids | City Form to Promote Sustainable Growth South Asia Hyderabad, India Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 2.48% Total built-up expansion (1990–2015) 168 km2 Annual population growth 1.65% Built-up area per person (2015) 45 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 127 South Asia Colombo, Sri Lanka Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.36% Total built-up expansion (1990–2015) 32 km2 Annual population growth 1.25% Built-up area per person (2015) 17 m2 128 Pancakes to Pyramids | City Form to Promote Sustainable Growth Sub-Saharan Africa Johannesburg, South Africa Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.61% Total built-up expansion (1990–2015) 126 km2 Annual population growth 3.07% Built-up area per person (2015) 133 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 129 Sub-Saharan Africa Lagos, Nigeria Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.91% Total built-up expansion (1990–2015) 165 km2 Annual population growth 2.60% Built-up area per person (2015) 67 m2 130 Pancakes to Pyramids | City Form to Promote Sustainable Growth Sub-Saharan Africa Accra, Ghana Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.74% Total built-up expansion (1990–2015) 84 km2 Annual population growth 3.33% Built-up area per person (2015) 106 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 131 Sub-Saharan Africa Cape Town, South Africa Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.82% Total built-up expansion (1990-2015) 42 km2 Annual population growth 2.75% Built-up area per person (2015) 111 m2 132 Pancakes to Pyramids | City Form to Promote Sustainable Growth Sub-Saharan Africa Durban, South Africa Built-up in 1990 Built-up in 2015 0 5 10 20 km Annual built-up expansion 0.82% Total built-up expansion (1990–2015) 61 km2 Annual population growth 1.25% Built-up area per person (2015) 115 m2 Pancakes to Pyramids | City Form to Promote Sustainable Growth 133 Works cited Ahlfeldt, G. M., and J. Barr. 2020. “The Economics Bertaud, A., and J. K. Brueckner. 2005. “Analyzing of Skyscrapers: A Synthesis.” CEPR Discussion Building-height Restrictions: Predicted Impacts Paper DP14987, Center for Economic and Policy and Welfare Costs.” Regional Science and Urban Research, Washington, DC. https://ssrn.com/ Economics 35 (2): 109–125. abstract=3650100. Bird, J., and A. Venables. 2019. “Growing a Developing Ahlfeldt, G. M., and E. Pietrostefani. 2019. “The City: A Computable Spatial General Equilibrium Economic Effects of Density: A Synthesis.” Model Applied to Dhaka.” Policy Research Working International Trade and Regional Economics Paper 8762, World Bank, Washington, DC. Discussion Paper 13440, Centre for Economic Brueckner, J., and K. Sridhar. 2012. “Measuring Policy Research, London School of Economics and Welfare Gains from Relaxation of Land-use Political Science. Restrictions: The Case of India’s Building-height Ahlfeldt, G. M., S. J. Redding, D. M. Sturm, and N. Wolf. Limits.” Regional Science and Urban Economics 2015. “The Economics of Density: Evidence from 42 (6): 1061–1067. the Berlin Wall.” Econometrica 83: 2127–2189. Bryan, G., E. Glaeser, and N. Tsivanidis. 2019. “Cities Alberti, F. G., and J. D. Giusti. 2012. “Cultural Heritage, in the Developing World.” Annual Review of Tourism and Regional Competitiveness: The Motor Economics 12: 273–297. Valley Cluster.” City, Culture and Society 3 (4): Burchfield, M., H. G. Overman, D. Puga, and M. A. 261–273. Turner. 2006. “Causes of Sprawl: A Portrait from Allen, T., C. Arkolakis, and X. Li. 2016. “Optimal City Space.” The Quarterly Journal of Economics 121 Structure.” 2016 Meeting Paper 301, Society for (2): 587–633. Economic Dynamics. Byamugisha, F. 2013. Securing Africa’s Land for Alonso, W. 1964. Location and Land Use. Cambridge, Shared Prosperity. Washington, DC: World Bank. MA: Harvard University Press. CCFLA (Cities Climate Finance Leadership Alliance). Angel, S. 2011. Making Room for a Planet of Cities. 2015. The State of City Climate Finance 2015. Washington, DC: World Bank. New York: CCFLA. Angel, S., J. Parent, D. L. Civco, A. Blei, and D. Potere. CIESIN (Center for International Earth Science 2011. “The Dimensions of Global Urban Expansion: Information Network). 2011. Global Rural-Urban Estimates and Projections for All Countries, Mapping Project, Version 1 (GRUMPv1): Urban 2000–2050.” Progress in Planning 75 (2): 53–107. Extents Grid. Palisades, NY: NASA Socioeconomic Data and Applications Center (SEDAC). https://doi. Angel, S., A. M. Blei, J. Parent, P. Lamson-Hall, N. G. org/10.7927/H4GH9FVG. Sánchez, D. L. Civco, R. Q. Lei, and K. Thom. 2016. Atlas of Urban Expansion—2016 Edition. New Collier, P., and A. J. Venables. 2016. “Urban York: NYU Urban Expansion Program at New York Infrastructure for Development.” Oxford Review of University, UN-Habitat, and the Lincoln Institute of Economic Policy 32 (3): 391–409. Land Policy. Corbane, C., M. Pesaresi, T. Kemper, P. Politis, A. J. Ardila-Gomez, A., and A. Ortegon-Sanchez. Florczyk, V. Syrris, and P. Soille. 2019. “Automated 2016. Sustainable Urban Transport Financing from Global Delineation of Human Settlements from the Sidewalk to the Subway: Capital, Operations, 40 Years of Landsat Satellite Data Archives.” Big and Maintenance Financing. Washington, DC: Earth Data 3 (2): 140–169. World Bank. https://openknowledge.worldbank.org/ Dasgupta, B., S. V. Lall, and N. Lozano-Gracia. 2014. handle/10986/23521. “Urbanization and Household Investment.” Policy Avner, P., and S. Lall. 2016. “Matchmaking in Nairobi: Research Working Paper 7170, World Bank, The Role of Land Use.” Policy Research Working Washington, DC. Paper 7904, World Bank, Washington, DC. Dijkstra, L., A. J. Florczyk, S. Freire, T. Kemper, M. Bertaud, A. 2018. Order Without Design: How Markets Melchiorri, M. Pesaresi, and M. Schiavina. 2020. Shape Cities. Cambridge, MA: MIT Press. “Applying the Degree of Urbanisation to the Globe: 134 Pancakes to Pyramids | City Form to Promote Sustainable Growth A New Harmonised Definition Reveals a Different Gechter, M., and N. Tsivanidis. 2020. “Spatial Spillovers Picture of Global Urbanisation.” Journal of Urban from Urban Renewal: Evidence from the Mumbai Economics 103312. Mills Redevelopment.” Working Paper. http://www. personal.psu.edu/mdg5396/MumbaiMills.pdf. Dingel, J., and B. Neiman. 2020. “How Many Jobs Can Be Done at Home?” Journal of Public Economics Glaeser, E. 2011. Triumph of the City: How Our 189 (September). Greatest Invention Makes Us Richer, Smarter, Greener, Healthier, and Happier. Penguin Books. Djankov, S., E. L. Glaeser, V. Perotti, and A. Shleifer. 2020. “Measuring Property Rights Institutions.” Glaeser, E. 2020. “Infrastructure and Urban Form.” Working Paper 27839, National Bureau of Working Paper 28287, National Bureau of Economic Research, Cambridge, MA. Economic Research, Cambridge, MA. Duranton, G., and D. Puga. 2015. “Urban Land Use.” Glaeser, E., R. La Porta, F. Lopez-de-Silanes, and A. In Handbook of Regional and Urban Economics, Shleifer. 2004. “Do Institutions Cause Growth?” edited by G. Duranton, J. V. Henderson, and W. C. Journal of Economic Growth 9: 271–303. Strange, 467–560. Amsterdam: Elsevier. Glaeser, E., and W. Xiong. 2017. “Urban Productivity Duranton, G., and D. Puga. 2018. “Urban Growth in the Developing World.” Working Paper and Its Aggregate Implications.” Working Paper 23279, National Bureau of Economic Research, 26591, National Bureau of Economic Research, Cambridge, MA. Cambridge, MA. GLTN (Global Land Tool Network). n.d. “Access Duranton, G., and D. Puga. 2020. “The Economics of to Land and Tenure Security.” https://gltn.net/ Urban Density.” Journal of Economic Perspectives access-to-land-and-tenure-security/. 34 (3): 3–26. Global Infrastructure Hub. 2017. Global Infrastructure Ellis, P., and M. Roberts. 2016. Leveraging Outlook 2017. Oxford, UK: Oxford Economics. Urbanization in South Asia: Managing Spatial Global Infrastructure Hub. 2020. Global Infrastructure Transformation for Prosperity and Livability. Outlook 2020. Oxford, UK: Oxford Economics. Washington, DC: World Bank. Grover, A., S. Lall, and W. Maloney. 2021. Place, Esch, T., J. Zeidler, D. Palacios-Lopez, M. Marconcini, Productivity, and Prosperity: Revisiting Spacially- A. Roth, M. Mönks, B. Leutner, E. Brzoska, A. Metz-­ targeted Policies for Regional Development. Marconcini, F. Bachofer, S. Loekken, and S. Dech. Washington, DC: World Bank. 2020. “Towards a Large-scale 3D Modelling of the Built Environment: Joint Analysis of TanDEM-X, Gollin, D., R. Jedwab, and D. Vollrath. 2016. Sentinel-­ 2 and Open Street Map Data.” Remote “Urbanization with and without Industrialization.” Sensing 12 (2391). Journal of Economic Growth 21 (1): 35–70. Farvacque-Vitkovic, C., and M. Kopanyi. 2014. Gonzalez-Navarro, M., and M. A. Turner. 2018. Municipal Finance Handbook: A Handbook for “Subways and Urban Growth: Evidence from Local Governments. Washington, DC: World Bank. Earth.” Journal of Urban Economics 108 (C): 85–106. Florczyk, A., M. Melchiorri, C. Corbane, M. Schiavina, M. Pesaresi, P. Panagiotis, F. Sabo, S. M. C. Freire, Goodfellow, T. 2013. “Planning and Development D. Ehrlich, T. Kemper, P. Tommasi, D. Airaghi, and L. Regulation amid Rapid Urban Growth: Explaining Zanchetta. 2019. “Description of the GHS Urban Divergent Trajectories in Africa.” Geoforum 48: Centre Database 2015.” Brussels: Publication 83–93. Office of the European Union. Gottlieb, C., J. Grobovšek, M. Poschke, and F. Saltiel. Florczyk, A., C. Corbane, D. Ehrlich, S. Freire, T. 2021. “Working from Home in Developing Kemper, L. Maffenini, and F. Sabo. 2019. “GHSL Countries.” European Economic Review 133, Data Package 2019.” European Commission Joint 103679. Research Center, 29788. Gulyani, S. 2016. “Success When We Deemed It Fujita, M., and H. Ogawa. 1982. “Multiple Equilibria and Failure? Revisiting Sites and Services 20 Years Structural Transition of Non-monocentric Urban Later.” World Bank Sustainable Cities blog, June Configurations.” Regional Science and Urban 23. https://blogs.worldbank.org/sustainablecities/ Economics 12 (2): 161–196. Pancakes to Pyramids | City Form to Promote Sustainable Growth 135 success-when-we-deemed-it-failure-revisiting-sites Policy Research Working Paper 3915, World Bank, -and-services-20-years-later. Washington, DC. Hallegatte, S. 2009. “Strategies to Adapt to an Lall, S., and S. Wahba. 2020. “Cities, Crowding, Uncertain Climate Change.” Global Environmental and the Coronavirus: Predicting Contagion Risk Change 19 (2): 240–247. Hotspots.” Sustainable Cities (blog), April 23, Heblich, S., S. J. Redding, and D. M. Sturm. 2020. “The 2020. https://blogs.worldbank.org/sustainablecities/ cities-crowding-and-coronavirus-predicting-contagion Making of the Modern Metropolis: Evidence from -risk-hotspots. London.” The Quarterly Journal of Economics 135 (4): 2059–2133. Lall, S. V., and S. Wahba. 2021. “Crowded Cities.” Sustainability (forthcoming). Henderson, J. V., T. Regan, and A. J. Venables. 2020. “Building the City: From Slums to a Modern Lall, S., H. G. Wang, and D. Da Mata. 2007. “Do Urban Metropolis.” LSE Research Online Documents on Land Regulations Influence Slum Formation? Economics 105054, London School of Economics Evidence from Brazilian Cities.” Brazilian and Political Science, LSE Library. Association of Graduate Programs in Economics. Henderson, J. V., and H. G. Wang. 2007. “Urbanization Leyk, S., J. H. Uhl, D. Balk, and B. Jones. 2018. and City Growth: The Role of Institutions.” “Assessing the Accuracy of Multi-temporal Regional Science and Urban Economics 37 (3): Built-up Land Layers across Rural-urban 283–313. Trajectories in the United States.” Remote Sensing of Environment 204: 898–917. Ingram, G., Z. Liu, and K. Brandt. 2013. “Metropolitan Infrastructure and Capital Finance.” In Financing Liu, X., G. Hu, Y. Chen, X. Li, X. Xu, S. Li, and S. Wang. Metropolitan Governments in Developing 2018. “High-resolution Multi-temporal Mapping of Countries. Lincoln Institute of Land Policy. Global Urban Land Using Landsat Images Based on the Google Earth Engine Platform.” Remote Jones, B., D. Balk, H. Engin, S. Leyk, and M. Montgomery. Sensing of Environment 209: 227–239. Forthcoming. Projecting Global Population Grids to 2100. Luxembourg: Publications Office of the Lloyd-Jones, T., M. Theis, S. Adenekan, S. Gusah, European Union. and M. Mutter. 2014. “The Informal Housing Development Process in Nigeria: The Case of Klein, M. 2012. “Infrastructure Policy: Basic Design Kaduna.” London: MLC Press. Options.” Policy Research Working Paper 6274, World Bank, Washington, DC. Lozano-Gracia, N., C. Young, S. V. Lall, and T. Vishwanath. 2013. “Leveraging Land to Enable Kreindler, G., and Y. Miyauchi. 2020. “Measuring Urban Transformation. Lessons from Global Commuting and Economic Activity inside Cities Experience.” Policy Research Working Paper 6312, with Cell Phone Records.” Working Paper. https:// World Bank, Washington, DC. economics.mit.edu/files/14088. Lucas, R. E., and E. Rossi-Hansberg. 2002. “On the La Porta, R., F. Lopez-de-Silanes, A. Shleifer, Internal Structure of Cities.” Econometrica 70 (4): and R. Vishny. 1999. “The Quality of 1445–1476. Government.” Journal of Law, Economics, and Organization 15 (1): 222–79. Ma, J., Z. Liu, and Y. Chai. 2015. “The Impact of Urban Form on CO2 Emission from Work and Non- Lall, S., J. V. Henderson, and A. J. Venables. 2017. work Trips: The Case of Beijing, China.” Habitat Africa’s Cities: Opening Doors to the World. International 47: 1–10. Washington, DC: World Bank. Mahtta, R., A. Mahendra, and K. C. Seto. 2019. Lall, S., M. Lebrand, and M. Soppelsa. 2021. “The “Building up or Spreading out? Typologies of Evolution of City Form: Evidence from Satellite Urban Growth across 478 Cities of 1 Million+.” Data.” Policy Research Working Paper 9618, World Environmental Research Letters 14 (12): 124077. Bank, Washington, DC. Malope, P., and M. Phirinyane. 2016. “Enhancing Lall, S., H. Selod, and Z. Shalizi. 2006. “Rural–Urban Property Rights through Land Tenure Regularization Migration in Developing Countries: A Survey of in Botswana.” Paper prepared for presentation at Theoretical Predictions and Empirical Findings.” the 2016 World Bank Conference on Land and Poverty, Washington, DC, March 14–18. 136 Pancakes to Pyramids | City Form to Promote Sustainable Growth Marconcini, M., A. Metz-Marconcini, S. Üreyen, D. Owens, K. E., S. Gulyani, and A. Rizvi. 2018. “Success Palacios-Lopez, W. Hanke, F. Bachofer, J. Zeidler, When We Deemed It Failure? Revisiting Sites and T. Esch, N. Gorelick, A. Kakarla, and E. Strano. Services Projects in Mumbai and Chennai 20 2020. “Outlining Where Humans Live: The World Years Later.” World Development 106: 260–272. Settlements Footprint 2015.” Scientific Data 7 (242). Prindex. 2019. “Press Release: Almost 1 Billion People Mattauch, L., C. Hepburn, and N. Stern. 2018. “Making Worldwide Expect to Lose Their Homes against Carbon Pricing Work for Citizens.” Nature Climate Their Will, Prindex Global Study Reveals.” Press Change 8: 669–677. Release, March 26. https://www.prindex.net/news -and-stories/press-release-almost-1-billion-people Melo, P., D. Graham, and R. Noland 2009. “A Meta- -worldwide-expect-lose-their-homes-against-their analysis of Estimates of Urban Agglomeration -will-prindex-global-study-reveals/. Economies.” Regional Science and Urban Economics 39 (3): 332–342. Povey, M., and T. Lloyd-Jones. 2000. “Mixed Use Development: Mechanisms for Sustaining the Michaels, G., D. Nigmatulina, F. Rauch, T. Regan, Livelihoods and Social Capital of the Urban N. Baruah, and A. Dahlstrand. Forthcoming. Poor in Core Areas.” Paper presented at the “Planning Ahead for Better Neighborhoods: Long 17th Inter-Schools Conference on Sustainable Run Evidence from Tanzania.” Journal of Political Cities: Sustainable Development, Oxford Brookes Economy (forthcoming). University, Oxford, UK, April. Mills, E. S. 1972. Studies in the Structure of the Urban Pulido, D., G. Darido, R. Munoz-Raskin, and J. Economy. Baltimore, MD: Johns Hopkins Press. Moody. 2018. The Urban Rail Development Moran, D., K. Kanemoto, M. Jiborn, R. Wood, J. Többen, Handbook. Washington, DC: World Bank. https:// and K. C. Seto. 2018. “Carbon Footprints of 13,000 openknowledge.worldbank.org/handle/10986/30392. Cities.” Environmental Research Letters 13 (6): Rabé, P. E. 2010 “Land Sharing in Phnom Penh 064041. and Bangkok: Lessons from Four Decades of Muth, R. F. 1969. Cities and Housing: The Spatial Innovative Slum Redevelopment Projects in Two Pattern of Urban Residential Land Use. Chicago, IL: Southeast Asian ‘Boom Towns.’” Paper presented University of Chicago Press. at the “Places We Live: Slums and Urban Poverty in the Developing World” policy workshop, World Navarro, I. A., and G. K. Turnbull. 2014. “Property Bank, Washington, DC, April 30. Rights and Urban Development: Initial Title Quality Matters Even When It No Longer Matters.” The Ruiz, F., and G. Vallejo. 2010. “Using Land Registration Journal of Real Estate Finance and Economics 49 as a Tool to Generate Municipal Revenue: Lessons (1): 1–22. from Bogota.” Annual Conference on Land Policy and Administration, World Bank, Washington, DC, Nyborg, K., J. M. Anderies, A. Dannenberg, T. Lindahl, April 26–27. C. Schill, M. Schlüter, W. N. Adger, K. J. Arrow, S. Barrett, S. Carpenter, F. S. Chapin III, A.-S. Crépin, Ryberg-Webster, S., and K. L. Kinahan. 2014. G. Daily, P. Ehrlich, C. Folke, W. Jager, N. Kautsky, S. “Historic Preservation and Urban Revitalization A. Levin, O. J. Madsen, S. Polasky, M. Scheffer, B. in the Twenty-first Century.” Journal of Planning Walker, E. U. Weber, J. Wilen, A. Xepapadeas, and Literature 29 (2): 119–139. A. de Zeeuw. 2016. “Social Norms as Solutions.” Saiz, A. 2010. “The Geographic Determinants of Science 354 (6308): 42–43. Housing Supply.” Quarterly Journal of Economics OECD (Organisation for Economic Co-operation and 125 (3): 1253–1296. Development). 2017. “Technical Note on Estimates Salat, S. 2013. Cities and Forms: On Sustainable of Infrastructure Investment Needs.” Background Urbanism. Urban Morphology Institute. paper, Investing in Climate, Investing in Growth, OECD Publishing, Paris. Salat, S., and G. Ollivier. 2017. Transforming the Urban Space through Transit-Oriented Development: The Ollivier, G., A. Ghate, K. Bankim, and P. Mehta. 3V Approach. Washington, DC: World Bank https:// 2021. Transit-Oriented Development openknowledge.worldbank.org/handle/10986/26405. Implementation Resources and Tools, 2nd Edition. Washington, DC: World Bank. https:// Seto K. C., S. Dhakal, A. Bigio, H. Blanco, G. C. Delgado, openknowledge.worldbank.org/handle/10986/34870. D. Dewar, L. Huang, A. Inaba, A. Kansal, S. Lwasa, J. E. McMahon, D. B. Müller, J. Murakami, H. Nagendra, Pancakes to Pyramids | City Form to Promote Sustainable Growth 137 and A. Ramaswa. 2014. “Human Settlements, UNESCO (United Nations Educational, Scientific Infrastructure and Spatial Planning.” In Climate and Cultural Organization). 2016. Culture for Change 2014: Mitigation of Climate Change, Sustainable Urban Development. Paris: UNESCO. Contribution of Working Group III to the Fifth UN-Habitat (United Nations Human Settlements Assessment Report of the Intergovernmental Programme). 2008. Mozambique Urban Sector Panel on Climate Change. Cambridge, UK and New Profile. Nairobi: UN-Habitat. York: Cambridge University Press. Uribe, M. C. 2010. “Land Information Updating, a Shah, A. 2006. “A Practitioner’s Guide to Local De Facto Tax Reform: Bringing Up to Date the Government Fiscal Transfer.” Policy Research Cadastral Database of Bogota.” In Innovations Working Paper 4039, World Bank, Washington, DC. in Land Rights Recognition, Administration and Smolka, M. 2018. “Assessing the Monetary Governance, edited by K. Deininger. Washington, Relevance of Land Value Capture: The Case for DC: World Bank. Charges for Additional Building Rights in São Paulo, Walkdek, S. 2018. “Inside the Secret Life of New Brazil.” International Conference on Municipal York’s Grand Central Terminal.” Architectural Fiscal Health, Lincoln Institute of Land Policy, Digest, August 14. https://www.architecturaldigest. Detroit, MI, May 2018. https://www.lincolninst.edu/ com/story/grand-central-terminal-history. sites/default/files/research_-_martim_smolka.pdf. Weinberger, R., and F. Goetzke. 2010. “Unpacking Storper, M., and A. J. Venables. 2004. “Buzz: Face-to- Preference: How Previous Experience Affects Auto face Contact and the Urban Economy.” Journal of Ownership in the United States.” Urban Studies Economic Geography 4 (4): 351–370. 47 (10): 2111–2128. Sturm, D., K. Takeda, and A. J. Venables. 2021a. “The Wesseling, T. 2016. New Approaches to Physical Urban Model: Jobs, Housing, Infrastructure, and Planning in Zambia. Royal HaskoningDHV. Policy.” Forthcoming. White, R., and S. Wahba. 2019. “Addressing Sturm, D., K. Takeda, and A. J. Venables. 2021b. “How Constraints to Private Financing of Urban Useful Are Quantitative Urban Models for Cities (Climate) Infrastructure in Developing Countries.” in Developing Countries? Evidence from Dhaka.” International Journal of Urban Sustainable Forthcoming. Development 11 (3): 245–256. Toulmin, C. 2005. Securing Land and Property World Bank. 2012. The Economics of Uniqueness: Rights in Sub-Saharan Africa: The Role of Local Investing in Historic City Cores and Cultural Institutions. Geneva: World Economic Forum. Heritage Assets for Sustainable Development. Tsivanidis, N. 2019. “Evaluating the Impact of Urban Washington, DC: World Bank. https:// Transit Infrastructure: Evidence from Bogotá’s openknowledge TransMilenio.” University of California, Berkeley. .worldbank.org/handle/10986/12286. https://static1.squarespace.com/static/55bb98e2e4b World Bank. 2013a. Urbanization beyond Municipal 0ba843f39599e/t/5d039b14655edc0001d608a9/ Boundaries: Nurturing Metropolitan Economies 1560517406526/TsivanidisTransMilenio_6.2019.pdf. and Connecting Peri-Urban Areas in India. Turnbull, G. K. 2005. “The Investment Incentive Effects Directions in Development. Washington, DC: World of Land Use Regulations.” The Journal of Real Bank. Estate Finance and Economics 31 (4): 357–395. World Bank. 2013b. Planning, Connecting, and Tweed, C., and M. Sutherland. 2007. “Built Cultural Financing Cities Now: Priorities for City Leaders. Heritage and Sustainable Urban Development.” Washington, DC: World Bank. Landscape and Urban Planning 83 (1): 62–69. World Bank. 2014. Urban China: Toward Efficient, United Nations. 2015. The State of City Climate Inclusive, and Sustainable Urbanization. Finance 2015. New York: United Nations. Washington, DC: World Bank. United Nations. 2018. World Urbanization Prospects: World Bank. 2015a. Ethiopia Urbanization Review: The 2018 Revision. New York: United Nations. Urban Institutions for a Middle-Income Ethiopia. UNDESA (United Nations Department of Economic Washington, DC: World Bank. and Social Affairs). 2018. 2018 Revision of World Urbanization Prospects. New York: UNDESA. 138 Pancakes to Pyramids | City Form to Promote Sustainable Growth World Bank. 2015b. Ghana Urbanization Review WorldPop and CIESIN (Center for International Overview: Rising through Cities in Ghana. Earth Science Information Network). 2018. Washington, DC: World Bank. Global High Resolution Population Denominators Project. School of Geography and Environmental World Bank. 2015c. Stocktaking of the Housing Science, University of Southampton; Department Sector in Sub-Saharan Africa: Challenges and of Geography and Geosciences, University of Opportunities. World Bank, Washington, DC: Louisville; Departement de Geographie, Universite World Bank. https://openknowledge.worldbank.org/ de Namur; and Columbia University. https://dx.doi. handle/10986/23358. org/10.5258/SOTON/WP00647. World Bank. 2016a. Kenya Urbanization Review. Wu, W. 2013. “Public Private Partnership for Urban Washington, DC: World Bank. Infrastructure Financing and Management.” World Bank. 2016b. Malawi Urbanization Review: Background paper for the World Bank Strengthening Rural-Urban Linkages. Washington, Urbanization Project, World Bank, Washington, DC. DC: World Bank. Xu, C., D. Haase, M. Su, and Z. Yang. 2019. “The World Bank. 2016c. Regenerating Urban Land: Impact of Urban Compactness on Energy- A Practitioner’s Guide to Leveraging Private related Greenhouse Gas Emissions across EU Investment. Urban Development. Washington, DC: Member States: Population Density vs. Physical World Bank. Compactness.” Applied Energy 254. World Bank. 2020. Public–Private Partnerships in Yang, F., Z. Wang, X. Yang, Y. Liu, B. Liu, J. Wang, and J. Land Administration: Analytical and Operational Kang. 2019. “Using Multi-Sensor Satellite Images Frameworks. Washington, DC: World Bank. and Auxiliary Data in Updating and Assessing the Accuracies of Urban Land Products in Different World Bank and Development Research Center of Landscape Patterns.” Remote Sensing 11 (22): the State Council, People’s Republic of China. 2664. 2014. Urban China: Toward Efficient, Inclusive, and Sustainable Urbanization. Washington, DC: World Zarate, R. D.. 2020. “Spatial Misallocation, Informality Bank. and Transit Improvements: Evidence from Mexico City.” Unpublished working paper. Pancakes to Pyramids | City Form to Promote Sustainable Growth 139 I II How are cities accommodating What has driven urban spatial families and firms—and how fast is evolution since 1990? The answers are global built-up area expanding? complicated—but incomes are key III IV What drives one city to grow From pancakes to pyramids: differently from another? What city leaders need to know Wichtiger HINWEIS ! Innerhalb der Schutzzone (hellblauer Rahmen) darf kein anderes Element platziert werden! Ebenso darf der Abstand zu Format- resp. Papierrand die Schutzzone nicht verletzen! Hellblauen Rahmen der Schutzzone nie drucken! Siehe auch Handbuch „Corporate Design der Schweizerischen Bundesverwaltung“ Kapitel „Grundlagen“, 1.5 / Schutzzone www. cdbund.admin.ch Supported by: In partnership with: