83905 JANUARY 2014 • Number 132 What Makes Cities More Competitive? Lessons from India Ejaz Ghani, William Kerr, and Stephen D. O’Connell Policy makers in both developed and developing countries want to accelerate spatial development, make cities more competitive, attract new entrepreneurs, boost economic growth, and promote job creation. These are commendable goals given that city populations in developing countries are expected to double from 2 billion to 4 billion people between 2000 and 2030. So what makes some cities more competitive than others? This note examines city competitiveness in India through the lens of spatial location choices of new and young entrepreneurs using plant-level data from the manufactur- ing and services sectors, including formal and informal operations. Findings show that the two most consistent factors that predict overall entrepreneurship for a district are its population's level of education and the quality of local physical infrastructure; these patterns are true for manufacturing and services. Agglomeration economies are much stronger in India than in the United States, but there is much greater variation in spatial outcomes in India than in the United States. Micro evidence for India also suggests that while strict labor regulations discourage formal sector entry, better household banking environments encourage entry into the informal sectors. Informal sectors conform much more closely to the overall contours of India’s economic geography than formal sectors. Policy makers looking to promote competitive- ness in their local areas have several policy levers to exploit. Many policy makers want to encourage entrepreneurship at mies in advanced countries, but the relevance of these pat- the local level, given its central role in economic growth and terns in developing economies has not been consistently development. Multiple studies have examined this question established. of how to promote entrepreneurship in advanced economies. A recent working paper by Ghani, Kerr, and O’Connell However, similar work for developing countries is still at an (2012), key elements of which are featured in a forthcoming early stage. Why do some cities in developing countries attract article in Regional Studies (Ghani, Kerr, and O’Connell forth- more entrepreneurs and become engines of growth? Why are coming), examines these questions for manufacturing and other cities short of entrepreneurs? services in India for 630 districts spread across 35 states/ Finding and understanding the answers to these ques- union territories. Within these two industry groups, analysis tions from a developing-country perspective can help these also compares the formal and the informal sectors. Results countries jump start economic growth in their cities. The show that certain factors and traits of districts and industries roles that education or infrastructure play, for example, re- systematically predict stronger entry of new firms. Ghani, garding entry into an advanced economy like the United Kerr, and O’Connell (2012) seek to determine the general de- States may be quite different than their impact in a setting gree to which the economic geography of India can be ex- like India, where lack of human capital and lack of roads plained with a parsimonious set of specifications, and to com- continue to hamper development. Likewise, there is exten- pare the specific factors identified as important relative to sive evidence on the importance of agglomeration econo- those in the United States. 1 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise Drivers of Entrepreneurship and District- the degree to which industries interact (following Glaeser and Level Competitiveness Kerr [2009]). The study examines the roles of demographic traits (age pro- Empirical Findings files, population, and population density), structural traits Table 1 presents the key findings by sector, with more detailed (education of the local labor force, quality of local physical coefficients provided in referenced studies. The first column infrastructure, travel time to major cities, stringency of labor laws, and household banking conditions), and agglomeration considers the organized manufacturing sector. An initial un- economies and their impacts on entrepreneurship and dis- reported analysis first considers the predictive power for en- trict-level competitiveness. While these traits do not consti- trepreneurship of a parsimonious regression that includes tute an exhaustive list of local conditions, they are motivated district populations, district-industry employments, and in- by the literature on India’s development. dustry-fixed effects as explanatory variables. Not surprisingly, The analysis develops metrics that unite the incumbent existing district-industry employment strongly shapes the industrial structures of districts with the extent to which in- spatial location of entry: a 10 percent increase in incumbent dustries interact through three traditional agglomeration employment raises entry employment by around 2 percent. channels that have been discussed since Alfred Marshall. The In addition, a district’s population increases entry rates with first agglomeration channel is proximity to customers and an elasticity of 0.5. suppliers, which reduces transportation costs and thereby in- Table 1 also includes district traits and agglomeration creases productivity. The second channel is the Chinitz ef- economies, of which three factors stand out as discouraging fect, which descends from the work of Benjamin Chinitz on entrepreneurship in organized manufacturing: high popula- how small-scale suppliers can provide specific aid to new tion density, strict labor regulations, and greater distance firms. The third channel is labor pooling, which picks up on from 1 of India’s 10 biggest cities. The first pattern has been the themes of specialized workers tightly clustered together. observed in many settings—the traded nature of manufactur- These metrics unite the industrial structures of cities with ing products allows more rural settings for firms, and manu- Table 1. Summary of District-Industry Entrepreneurship Estimations Organized Unorganized Organized Unorganized manufacturing manufacturing services services sector sector sector sector Log of incumbent employment in district-industry 0 0 -- + Log of district population +++ +++ +++ +++ District traits: Log of district population density --- 0 0 - Share of population with graduate education +++ 0 +++ +++ Demographic dividend for district (age profiles) 0 +++ + ++ Index of infrastructure quality for district 0 +++ +++ +++ Strength of household banking environment 0 +++ ++ +++ Stringency of labor laws in district's state --- 0 - --- Log travel time to closest large city --- 0 0 0 Local industrial conditions by incumbent firms: Labor market strength for district-industry 0 +++ Inputs/supplier strength for district-industry +++ +++ Outputs/customer strength for district-industry +++ +++ Chinitz strength for district-industry 0 Industry-fixed effects Yes Yes Yes Yes Observations 4843 6451 3340 6552 Adjusted R-squared 0.218 0.264 0.252 0.536 Source: Authors' compilation. Notes: Estimations quantify the relationship between district-industry employment in new establishments and local conditions. 2 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise facturers often seek cheaper environments than the wages Among district traits, education and infrastructure mat- and rents associated with high-density areas. The second pat- ter the most. Overall, education is found to be generally im- tern connects with the earlier studies of India that argue strict portant, and particularly relevant in the organized sectors of labor laws reduce economic growth. Entrepreneurship in la- manufacturing and services. Physical infrastructure is also bor-intensive industries is disproportionately reduced by important, particularly in the unorganized sectors. The strict labor laws. The final factor highlights that while manu- strength of the household banking sector is also very impor- facturers avoid the high costs of urban areas, they also avoid tant in the unorganized sectors. the most remote areas of India in favor of settings that are rela- Agglomeration economies operate as strongly for en- tively near large population centers, likely to access customers trants in India as they do in advanced economies. The impor- directly or to connect to shipping routes. tance of the Chinitz effect is concentrated among small en- The education of a district’s workforce is strongly linked trants. The importance of overall output markets and labor to higher entry rates. The elasticity is, in fact, stronger in eco- spillovers grows with entrant size. It appears that input cost nomic magnitude than that evident in comparable studies of factors are more influential in the location choices of small advanced economies. Looking at agglomeration economies, start-ups in India, while output conditions and labor markets the qualities of input and output markets are exceptionally are more important for large entrants. strong, with 0.4–0.5 elasticities. Labor market and Chinitz Conclusions measures have positive coefficients, but are not statistically significant. The decline in the main effect of incumbent em- India’s city competitiveness and industrial landscape is still ployment suggests that these four new metrics capture the taking shape. A comparison of India with United States positive channels of agglomeration on entry. shows that existing city population levels, city-industry em- Column 2 repeats this approach for the informal/unorga- ployment, and industry-fixed effects can explain 80 percent nized manufacturing sector: several distinct differences exist. of the spatial variation in entry rates in the United States. First, local population plays a much greater role with unit elas- The comparable explanatory power for India is just 29 per- ticity evident, much stronger than for organized manufactur- cent for manufacturing and 33 percent for services. A large ing. This greater connection of entry to the overall size of local portion of this gap is because India is in a much earlier stage markets almost certainly reflects unorganized entry being of development. India’s industrial landscape is also adjusting proportionate to market size and servicing local needs. Unor- after the deregulations of the 1980s and 1990s. At such an ganized manufacturing clearly conforms much more closely early point, and with industrial structures not yet en- to the overall contours of India’s economic geography than trenched, local policies and traits can have profound and organized manufacturing. lasting impacts by shaping where industries plant their The other two district traits that are associated with roots. strong entry rates are the strength of local, within-district, Agglomeration economies are very important for India’s physical infrastructure and the strength of local household entry patterns. The results of this analysis show strong evi- banking environments. This contrasts with organized man- dence of agglomeration economies in India’s manufacturing ufacturing entry, where education stood out. An intuitive sector. In a similar manner, there is extensive evidence that explanation, which is also reflected in the services estima- the incumbent compositions of local industries influence tions, is that these patterns and their differences reflect the new entry rates at the district-industry level within manufac- factors on which each sector depends most. Organized turing. This influence is through both traditional Marshallian manufacturing establishments, for example, may have economies, like a suitable labor force and proximity to cus- broader resources that reduce dependency on local infra- tomers, and through the Chinitz effect, which emphasizes structure and household finance. Likewise, the unorga- small suppliers. nized sector depends less on educated workers than the or- Education and physical infrastructure matter greatly. The ganized sector. two most consistent factors that predict overall competitive- Columns 3 and 4 present comparable estimations for the ness and entrepreneurship for a district are its level of educa- services sector. The patterns and their contrast to organized tion and the quality its local physical infrastructure. These manufacturing are again quite intriguing. First, overall dis- patterns are true for both manufacturing and services. These trict population is as important as it was for unorganized relationships are much stronger in India than those found in manufacturing, with its elasticity greater than 1. Similar also the United States. Higher levels of education in a local area to unorganized manufacturing, population density and travel increase the supply of entrepreneurs and increase the talent time to major cities are not important in the multivariate set- available to entrepreneurs for staffing their companies. Invest- ting, while the district’s age profile does contribute to higher ment in people is an easy call for policy makers. Likewise, cit- entry levels. ies must improve their infrastructure—electricity, roads, tele- 3 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise com, and water/sanitation facilities—to attract these coming globally powerful and growing in efficiency, and they entrepreneurs, retain the new businesses, and strengthen too will shape employment opportunities in the decades their competitiveness. ahead. However, the history of regional development shows Eliminating extreme poverty and promoting shared pros- that big firms are not enough: an entrepreneurial foundation perity. South Asia is currently the least urbanized region of that provides for local growth and regeneration is essential for the world, with only about 30 percent of its population in long-term success and prosperity. cities. It also has the highest concentration of the world’s About the Authors poor. India and South Asia are now at the forefront of a major population shift, moving from a largely rural to a rapidly ur- Ejaz Ghani is a Lead Economist in the PREM Economic Policy, banizing population. Nearly half a billion more people will Debt, and Trade Department. William Kerr is an Associate Pro- live in metropolitan areas in the region as the urban popula- fessor at Harvard Business School, Harvard University. Stephen tion is expected to more than double over the next 25 years. D. O’Connell is Chancellors Fellow at City University of New While this shift presents special challenges, there will also be York Graduate Center. unique opportunities to make a difference and have develop- ment impacts that reduce extreme poverty and promote References shared prosperity. Ghani, Ejaz, William Kerr, and Stephen O’Connell. 2012. “What More research on agglomeration economies and entre- Makes Cities More Competitive? Spatial Determinants of En- preneurship in developing countries is important for urban trepreneurship in India.” World Bank Policy Research Working and development economics going forward. Identifying con- Paper Series 6198, Washington, DC. ditions and factors that support entrepreneurship and ag- Ghani, Ejaz, William Kerr, and Stephen O’Connell. Forthcoming. “Spatial Determinants of Entrepreneurship in India.” Regional glomeration economies and acting upon them is essential to Studies. fostering economic growth in the cities of developing coun- Glaeser, Edward, and William Kerr. 2009. “Local Industrial Condi- tries. While focusing on entrepreneurship, policy makers also tions and Entrepreneurship: How Much of the Spatial Distribu- need to recognize that large firms play an important role in tion Can We Explain?” Journal of Economics & Management regional development. The giant firms of South Asia are be- Strategy 18 (3): 623–63. The Economic Premise note series is intended to summarize good practices and key policy findings on topics related to economic policy. They are produced by the Poverty Reduction and Economic Management (PREM) Network Vice-Presidency of the World Bank. The views expressed here are those of the authors and do not necessarily reflect those of the World Bank. The notes are available at: www.worldbank.org/economicpremise. 4 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK    www.worldbank.org/economicpremise