Gollin urbanization With and Without Industrialization

Transcript

1 Introduction Evidence Model Implications Conclusion with and without Urbanization Industrialization Douglas Gollin University of Oxford Remi Jedwab (GWU) Dietrich Vollrath (Houston) WB-GWU Conference on Urbanization and Poverty Reduction Douglas Gollin (Oxford) Urbanization with and without Industrialization

2 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Research Question I Urbanization associated with structural change: Green revolution, industrial revolution → industrialization. I In some parts of the world (Africa), urbanization without in- dustrialization. Where do these cities come from? Douglas Gollin (Oxford) Urbanization with and without Industrialization

3 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Where do these cities come from? Traffic jam in Lagos, Nigeria (Source: Engineering Consultancy (Nigeria) Ltd). 300,000 inh. in 1956, 11 millions today, empl. share of manufacturing: 3%. Douglas Gollin (Oxford) Urbanization with and without Industrialization

4 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Where do these cities come from? Kinshasa, DRC (Source: Jean-Jacques Frouein, Visoterra). 400,000 inh. in 1958, 9 millions today, empl. share of manufacturing < 10%. Douglas Gollin (Oxford) Urbanization with and without Industrialization

5 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Where do these cities come from? Khartoum, Sudan (Source: Marco di Lauro/Getty Images). 460,000 inh. in 1964, 5 millions today, empl. share of manufacturing: 7%. Douglas Gollin (Oxford) Urbanization with and without Industrialization

6 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Research Question I Urbanization associated with structural change: Green revolution, industrial revolution → industrialization. I In some parts of the world (Africa), urbanization without in- dustrialization. Where do these cities come from? What we do: 1. Motivating evidence on urbanization w/o industrialization. 2. Structural transformation model where resource exports drive urbanization: Surplus income → consumption cities , with mostly non-tradable services. 3. Implications of this type of urbanization on long-run growth. Douglas Gollin (Oxford) Urbanization with and without Industrialization

7 I Urbanization promotes self-sustained growth: I Returns-to-scale and agglomeration economies: I Developed countries (Glaeser et al. 1992, Duranton & Puga 2003, Rosenthal & Strange 2004, Henderson 2005) I Developing countries (Overman & Venables 2005, Duranton 2008, Henderson 2010, Felkner & Townsend 2012) Virtuous circle between development and urbanization: Europe/US, Asia Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Motivating Evidence I Urbanization associated with structural transformation: I Decline of agriculture, rise of manufacturing, tradable services: I (Schultz 1953, Matsuyama 1992, Gollin, Green revolution Push: Parente & Rogerson 2002, Michaels, Rauch & Redding 2011) I (Lewis 1954, Matsuyama 1992, Hansen Pull: Industrial revolution & Prescott 2002, Galor & Mountford 2008, Lucas 2009) Douglas Gollin (Oxford) Urbanization with and without Industrialization

8 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Motivating Evidence I Urbanization associated with structural transformation: I Decline of agriculture, rise of manufacturing, tradable services: I Green revolution (Schultz 1953, Matsuyama 1992, Gollin, Push: Parente & Rogerson 2002, Michaels, Rauch & Redding 2011) I Pull: Industrial revolution (Lewis 1954, Matsuyama 1992, Hansen & Prescott 2002, Galor & Mountford 2008, Lucas 2009) I Urbanization promotes self-sustained growth: I Returns-to-scale and agglomeration economies: I Developed countries (Glaeser et al. 1992, Duranton & Puga 2003, Rosenthal & Strange 2004, Henderson 2005) I Developing countries (Overman & Venables 2005, Duranton 2008, Henderson 2010, Felkner & Townsend 2012) Virtuous circle between development and urbanization: Europe/US, Asia Douglas Gollin (Oxford) Urbanization with and without Industrialization

9 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Urbanization and Industrialization for Developing Countries, 2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012. Data for 119 developing countries: Asia (30), Africa (46), LAC (26) and MENA (17). Douglas Gollin (Oxford) Urbanization with and without Industrialization

10 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Urbanization and Industrialization for Developing Countries, 2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012. Data for 119 developing countries: Asia (30), Africa (46), LAC (26) and MENA (17). NRX = natural resource exports in GDP (%) Douglas Gollin (Oxford) Urbanization with and without Industrialization

11 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Urbanization with Industrialization for Resource Poor Countries, 2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012. Data for 49 resource poor developing countries in 2000. NRX = natural resource exports in GDP (%) Douglas Gollin (Oxford) Urbanization with and without Industrialization

12 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Urbanization without Industrialization for Resource Rich Countries, 2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012. Data for 70 resource rich developing countries in 2000. NRX = natural resource exports in GDP (%) Douglas Gollin (Oxford) Urbanization with and without Industrialization

13 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Urbanization without Industrialization for Resource Rich Countries, 2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012. Data for 70 resource rich developing countries in 2000. NRX = natural resource exports in GDP (%) Douglas Gollin (Oxford) Urbanization with and without Industrialization

14 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline This Paper’s Approach I We take a theoretical approach and adapt a standard model of structural change to explain how natural resource exports: 1. Drive urbanization. 2. Give rise to a “different” type of urbanization (sectoral compo- sition), which may have implications for long-run growth. I Empirical analysis using a sample of 119 developing countries: 1. Additional motivating evidence at the continental level. 2. Cross-sectional multivariate analysis (not intended to show cau- sation). Douglas Gollin (Oxford) Urbanization with and without Industrialization

15 Introduction Evidence Model Implications Conclusion Question Motivation Approach Outline Outline I Additional motivating evidence I Conceptual Framework I Implications for Growth I Conclusion Douglas Gollin (Oxford) Urbanization with and without Industrialization

16 Introduction Evidence Model Implications Conclusion Additional Motivating Evidence Multivariate Analysis Outline I Additional motivating evidence I Conceptual Framework I Implications for Growth I Conclusion Douglas Gollin (Oxford) Urbanization with and without Industrialization

17 Introduction Evidence Model Implications Conclusion Additional Motivating Evidence Multivariate Analysis Urbanization Rate for Four Groups of Countries, 1950-2010 Sources: WUP 2011. Data for 119 developing countries: Asia (30), Africa (46), LAC (26) and MENA (17). Averages are estimated using the population weights for the same year. Douglas Gollin (Oxford) Urbanization with and without Industrialization

18 Introduction Evidence Model Implications Conclusion Additional Motivating Evidence Multivariate Analysis Urbanization with Industrialization in Asia, 2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012. Data for 29 Asian countries. Douglas Gollin (Oxford) Urbanization with and without Industrialization

19 I Is African urbanization different? I Not in line with standard structural change models: I Low food yields (Restuccia, Yang & Zhu 2008, Gollin 2010) I Small and/or unproductive manufacturing and service sectors (Caselli 2005, McMillan & Rodrik 2011) I The most urbanized African countries export natural resources: I Significant resource endowments at independence I Fuels, mining products or cash crops (cocoa, coffee, etc.) I High mark-up rates (inelastic international supply) I Jedwab 2012, Henderson, Roberts & Storeygard 2012 Additional Motivating Evidence Multivariate Analysis Introduction Evidence Model Implications Conclusion Additional Motivating Evidence I Africa has recently experienced dramatic urban growth: I Urbanization rate: ≈ Asia) 10% in 1950 ( Asia) 40% in 2010 ( ≈ 750,000): 40% of urban pop ( > Large cities ( ≈ Asia) Douglas Gollin (Oxford) Urbanization with and without Industrialization

20 Introduction Evidence Model Implications Conclusion Additional Motivating Evidence Multivariate Analysis Additional Motivating Evidence I Africa has recently experienced dramatic urban growth: I Urbanization rate: 10% in 1950 ( ≈ Asia) 40% in 2010 ( Asia) ≈ Large cities ( > 750,000): 40% of urban pop ( ≈ Asia) I Is African urbanization different? I Not in line with standard structural change models: I Low food yields (Restuccia, Yang & Zhu 2008, Gollin 2010) I Small and/or unproductive manufacturing and service sectors (Caselli 2005, McMillan & Rodrik 2011) I The most urbanized African countries export natural resources: I Significant resource endowments at independence I Fuels, mining products or cash crops (cocoa, coffee, etc.) I High mark-up rates (inelastic international supply) I Jedwab 2012, Henderson, Roberts & Storeygard 2012 Douglas Gollin (Oxford) Urbanization with and without Industrialization

21 Introduction Evidence Model Implications Conclusion Additional Motivating Evidence Multivariate Analysis Urbanization without Industrialization in Africa, 2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012. Data for 46 African countries. Douglas Gollin (Oxford) Urbanization with and without Industrialization

22 Introduction Evidence Model Implications Conclusion Additional Motivating Evidence Multivariate Analysis Urbanization without Industrialization in Africa, 2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012. Data for 46 African countries. Douglas Gollin (Oxford) Urbanization with and without Industrialization

23 Introduction Evidence Model Implications Conclusion Additional Motivating Evidence Multivariate Analysis Urbanization without Industrialization in Africa, 1960-2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012. Data for 46 African countries. Douglas Gollin (Oxford) Urbanization with and without Industrialization

24 Introduction Evidence Model Implications Conclusion Additional Motivating Evidence Multivariate Analysis Urbanization with Industrialization in LAC, 2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012. Data for 26 LAC countries. Correlation robust to dropping Guyana (GUY), Trinidad (TTO), Haiti (HTI) and Venezuela (VEN). Douglas Gollin (Oxford) Urbanization with and without Industrialization

25 Introduction Evidence Model Implications Conclusion Additional Motivating Evidence Multivariate Analysis Urbanization without Industrialization in MENA, 2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012. Data for 17 MENA countries. Correla- tion robust to dropping Lebanon (LBN), Jordan (JOR) and Yemen (YEM). Douglas Gollin (Oxford) Urbanization with and without Industrialization

26 Introduction Evidence Model Implications Conclusion Additional Motivating Evidence Multivariate Analysis Cross-Sectional Multivariate Analysis I c for the year 2000: Cross-sectional regression for countries . rate = α Urb Mfgserv β + Nrx γ + u + , c ,2000 c ,2000 c c ,2000 t Mfgserv : Share of manufacturing and services in GDP (%) c ,2000 Nrx : Share of natural resource exports in GDP (%) c ,2000 I Population-weighted, area FE (Asia, Africa, LAC, MENA), ctrls: I Urban definition: type (admin, threshold, threshold + admin, threshold + condition) and threshold I Physical geography: area, small island, landlocked, droughts I Economic geography: population, rural density (rural pop per sq km of arable area), pop growth rate in 1960-2000 I Institutions: autocracy (av. combined polity score < -5), dummy interstate or civil conflict since independence Douglas Gollin (Oxford) Urbanization with and without Industrialization

27 Introduction Evidence Model Implications Conclusion Additional Motivating Evidence Multivariate Analysis TABLE 1: MULTIVARIATE CROSS-SECTIONAL ANALYSIS Dependent Variable: Urbanization Rate in 2000 (%) (3) (5) (6) (1) (4) (2) Asia Africa 0.89*** 0.68*** 0.50*** -0.10 0.55*** Natural Resource Exports 1.28*** [0.19] (% of GDP, 2000) [0.13] [0.11] [0.47] [0.10] [0.17] Manufacturing & Services 0.91*** 0.57*** 0.33*** 1.26*** 0.40*** 1.29*** [0.21] [0.23] [0.10] [0.10] [0.28] [0.19] (% of GDP, 2000) Area Fixed Effects (4) N Y Y Y Y Y Controls N Y Y Y Y N Ctrl Initial Conditions 1960 N N Y N Y N Observations 119 119 119 114 30 46 R-squared 0.68 0.82 0.90 0.81 0.71 0.53 Notes: Robust SEs in parentheses; * p < 0.10, ** p < 0.05, *** p < 0.01. The sample consists of 119 developing countries across 4 areas: Asia (30), SSA (46), LAC (26) and MENA (17). Controls: urban definition, area, population, rural density, population growth, small island, landlocked, droughts, autocracy, conflict. Initial conditions: urbanization rate and resource exports in 1960. Douglas Gollin (Oxford) Urbanization with and without Industrialization

28 Introduction Evidence Model Implications Conclusion Intro Outline I Additional motivating evidence I Conceptual Framework I Implications for Growth I Conclusion Douglas Gollin (Oxford) Urbanization with and without Industrialization

29 Introduction Evidence Model Implications Conclusion Intro Conceptual Framework (Model in the paper) I Structural transformation: decline in agriculture, rise of manu- facturing (and/or tradable services). I Standard models of structural change: I Labor push: a rise in food productivity ( ) in- green revolution creases food availability and release labor for the modern sector. I Labor pull: a rise in mfg productivity ( industrial revolution ) attracts underemployed labor from agri. into the modern sector. I → mfg Trade: comparative advantage in mfg (tradable services) X and food M → acceleration of structural change. I Production cities: trade → mfg exporting cities that arise be- cause of a specialization effect (in urban-based sectors). I Urbanization without green revolution and industrial revolution? Douglas Gollin (Oxford) Urbanization with and without Industrialization

30 I Resource production has consumption linkages: I Massive surplus generated for producing countries (10x more than aid and remittances) I Relaxes the food constraint through food imports I Surplus income spent on urban goods and services (labor pull) I Urbanization through the rise of consumption cities I Urban tradables and/or urban non-tradables? Introduction Evidence Model Implications Conclusion Intro Conceptual Framework (Model in the paper) I Resource production does not contribute to urban growth: I Mineral production capital-intensive (e.g, oil 60% of GDP in Angola, only employs 10,000 nationals) I Cash crops produced in fields (rural employment) I urban tradables Model with four sectors: food , (tradable mfg, urban non-tradables (non-tradable mfg, serv), resources serv), Douglas Gollin (Oxford) Urbanization with and without Industrialization

31 Introduction Evidence Model Implications Conclusion Intro Conceptual Framework (Model in the paper) I Resource production does not contribute to urban growth: I Mineral production capital-intensive (e.g, oil 60% of GDP in Angola, only employs 10,000 nationals) I Cash crops produced in fields (rural employment) I Model with four sectors: food , urban tradables (tradable mfg, serv), urban non-tradables (non-tradable mfg, serv), resources I Resource production has consumption linkages: I Massive surplus generated for producing countries (10x more than aid and remittances) I Relaxes the food constraint through food imports I Surplus income spent on urban goods and services (labor pull) I Urbanization through the rise of consumption cities I Urban tradables and/or urban non-tradables? Douglas Gollin (Oxford) Urbanization with and without Industrialization

32 Introduction Evidence Model Implications Conclusion Intro Conceptual Framework (Model in the paper) I A subsistence economy urbanize without industrializing (unlike the Dutch Disease model of deindustrialization). I Stronger effects for mining/fuels vs cash crops. I If no resource sector, the country industrialize and urbanize through the rise of production cities. I Extension: Non-homothetic preferences accelerate urbanization if the distribution of the rent is unequal. I Extension: Government taxation accelerates urbanization (gvt’s Engel curve: civil servants, gvt region: primacy). I Extension: No capital accumulation in the model. Consump- tion cities cannot evolve into production cities. Douglas Gollin (Oxford) Urbanization with and without Industrialization

33 Introduction Evidence Model Implications Conclusion Other Stylized Facts Discussion Outline I Additional motivating evidence I Conceptual Framework I Implications for Growth I Conclusion Douglas Gollin (Oxford) Urbanization with and without Industrialization

34 Introduction Evidence Model Implications Conclusion Other Stylized Facts Discussion Other Stylized Facts I Stylized fact: urbanization associated with resource exports. I In the model, the resource rich countries: I urbanize because they are wealthier, I import more food and tradables, I their consumption cities have mostly non-tradable workers. I The numerical simulations match the patterns found in the data. I For example, cities in resource rich countries associated with less employment in urban tradables? Outcome: employment share of manufacturing and finance, insurance, real estate and business services for the urban areas in 2000. Douglas Gollin (Oxford) Urbanization with and without Industrialization

35 Introduction Evidence Model Implications Conclusion Other Stylized Facts Discussion Resource Exports and the Sectoral Composition of the Urban Areas 2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012, IPUMS 2012, labor force and household surveys, censuses. Data for 65 countries. Mfg: manufacturing. Fire: finance, insurance, real estate and business services. Douglas Gollin (Oxford) Urbanization with and without Industrialization

36 Introduction Evidence Model Implications Conclusion Other Stylized Facts Discussion Resource Exports and the Sectoral Composition of the Largest City 2000 Sources: WUP 2011, FAO 2012, USGS 2012, WDI 2012, IPUMS 2012, labor force and household surveys, censuses. Data for 66 countries. Mfg: manufacturing. Fire: finance, insurance, real estate and business services. Douglas Gollin (Oxford) Urbanization with and without Industrialization

37 Introduction Evidence Model Implications Conclusion Other Stylized Facts Discussion Discussion: Implications for Growth? I production cities : structural change → cities with Literature on → agglomeration economies → growth. tradable workers I Consumption cities → cities with non-tradable workers → second best type of urbanization (welfare-increasing). I Productivity growth is slower in services, while unconditional convergence in manufacturing. I Potential mechanisms? (i) lack of competition in non-tradable sectors (no global competitors), (ii) incentives for human K ac- cumulation, (iii) agglomeration economies, (iv) reforms (urban middle-class vs. urban elite). I In the long run, consumption cities could evolve into production cities. Constraints? Trade, institutions? Douglas Gollin (Oxford) Urbanization with and without Industrialization

38 Introduction Evidence Model Implications Conclusion Conclusion Outline I Additional motivating evidence I Conceptual Framework I Implications for Growth I Conclusion Douglas Gollin (Oxford) Urbanization with and without Industrialization

39 Introduction Evidence Model Implications Conclusion Conclusion Conclusion I Most developing countries have experienced urbanization with industrialization (e.g, Asia, LAC). Urbanization in production cities (tradable and non-tradable workers). I Most resource rich countries have experienced urbanization without industrialization (e.g., Africa, MENA). Urbanization in consumption cities (mostly non-tradable services). I This may have implications for long-run growth. I Research agenda: I Panel data analysis for 119 countries in 1960-2000. I Can consumption cities evolve into production cities? Yes: US, Dubai, South Africa, etc. Constraints? Trade, institutions? Douglas Gollin (Oxford) Urbanization with and without Industrialization

Related documents