Home Address:
  79 Avon St.
  New Haven, CT 06511

Telephone: (203) 606-4721 (cell)
                   (203) 436-4341 (office)

Development Economics
Microeconomic Theory

Development Economics
Microeconomic Theory
Applied Econometrics

2003 (Oral): Development Economics, Microeconomic Theory
2002 (Written): Microeconomic and Macroeconomic Theory

Factor Immobility and Regional Inequality in India

Professor Rohini Pande
Professor Christopher Udry
Professor Mark Rosenzweig

May 2006

M.Phil., Economics, Yale University, 2004
M.A., Economics, Yale University, 2003
M.A., Economics, Delhi School of Economics (Delhi, India), 2001
B.A. (Hons.) in Economics, St. Stephens College, Delhi University (Delhi, India), 1999

Social Science Research Council- PAE Summer Research Grant, 2004–2005
Yale Center for International and Area Studies Dissertation Research Grant, 2004
Ryoichi Sasakawa Fellowship, 2003–2004
Yale University Fellowship, 2001–2005
Scholarship in Economics, Delhi School of Economics, 1999-2000
Scholarship, National Council of Educational Research and Training (India), 1994–1999

Teaching Assistant, Economic Development in Africa (undergraduate), Yale University, Spring 2005
Teaching Assistant, Introductory Macroeconomics, Yale University Summer School, Summer 2005

Research Assistant, Professor Rohini Pande, Yale University, 2003–2004
   Project: "Local Government in India"
Research Assistant, Professor Leonard Wantchekon, Political Science, Yale University, 2002
   Pursued research on the theory and the empirical implications of power-sharing agreements as political
   risk-sharing contracts
Summer Research Position, Economic Research Group, ICICI Bank, Mumbai (India), 2000
   Worked on the project "Core Inflation in India"; estimated and studied a core inflation series for India

Northeast Universities Development Consortium (NEUDC) Conference, September 2000

Professor Rohini Pande
Department of Economics
Yale University
PO Box 208269
New Haven, CT 06520-8269
Fax: (203) 432-3898
E-mail: rohini.pande@yale.edu

Professor Mark Rosenzweig
Department of Economics
Yale University
PO Box 208269
New Haven, CT 06520-8269
Fax: (203) 432-3898
E-mail: mark.rosenzweig@yale.edu

This dissertation studies how factor immobility has affected the pattern of regional growth in India.  In particular, it examines whether, as a result of poor mobility of capital across regions, production decisions are constrained by the level of regional wealth.

The geographic mobility of labor is markedly low in India. In the absence of migration, the geographic spread of manufacturing can help reduce regional imbalances in growth. However, this mechanism requires a well functiong credit market.  There is considerable evidence in the literature that producers in developing countries face financing constraints. It is also well known that localized informal credit markets coexist with the formal credit market in developing countries. My dissertation explores a combined implication of these facts — namely, that producers are more credit constrained in poorer regions and this can contribute to continued divergence in incomes.

The dissertation has two parts. The first part examines factor immobility as a potential explanation for regional inequality in industrialization in India.  It exploits a policy-induced shock in the supply of credit to manufacturers to identify how credit constraints on factories vary across regions.  The second part of the dissertation examines if there is a similar regional pattern in the financing constraints that farmers face by exploiting a nationwide scheme to deliver bank credit to farmers.

Part 1: Factor Immobility and Regional Inequality:  Evidence from a Credit Shock in India

There is a significant disparity in the level of industrialization across districts in India.  Furthermore, looking across districts from 13 major states of India, I find that districts that became wealthy as a result of initial gains from the Green Revolution in agriculture have more industry.  In the literature that studies industrial growth in developing countries, regional inequality is commonly attributted to differences in the allocation of relatively immobile factors, such as natural resources and geographical location.  In this paper, I hypothesize that these disparities are related to labor and capital immobility across regions within a country.  Specifically, I agrue that during the 1990's, the uneven spread of industry within India can be partly explained by credit constraints on regional growth that are rooted in the spatial immobility of factors.

With factor immobility, disparities in wealth would lead to differences in factor prices across districts.  This paper tests this hypothesis by comparing growth in new factories in response to a credit shock to industry in both rich and poor districts.   The basis of the identification strategy is that if factor prices differ across districts, then there will be a difference across districts in the response of industrial growth to a credit shock. 

In 1998, a nationwide policy change in India resulted in an increase in the availablity of bank credit to a specific category of factories.  This credit shock was the outcome of a redefinition of "Small Scale Industry" (SSI).  The definition change in 1998 meant that some factories that were not eligible for directed credit to the SSI sector in 1997 became so in 1998.  The new credit was made available at the same interest rate in different districts.

I examine how growth in the number of factories in a district responded to this credit shock by using a triple difference approach.  The response to the credit shock is identified as follows:  within each district, growth in the set of factories that became newly eligible is compared to growth in the set of factories that were already classified as SSI.  Then, in every district, this relative growth is compared before and after the policy change.  Finally, this change in the relative growth in the newly eligible factories is compared across low and high wealth districts.   Furthermore, since it is possible that district wealth in the early 1990s may itself reflect trends in industrial growth, I instrument for wealth by the districts level of adoption of High Yield Varieties (HYVs) of seeds in the late 1960s, the initial years of the Green Revolution in agriculture.  Since the early seeds were not suited to all districts to the same extent, some districts gained more in the early stages of introduction of the HYVs.  Over time, HYV adoption became more widespread, but this initial disparity resulted in persistent wealth differentials.

Using initial HYV adoption as an instrument for wealth, I find that it is factory growth in low wealth districts that responds more to the credit shock in 1998. This differential response is limited to the duration of the policy shock, which shows that it does not reflect a diverging trend by district wealth in the relative growth of the newly eligible factories.  This finding has two implications.  First, since wealthier districts are more industrialized, the results in the paper show that in part, the concentration of industries in rich districts can be explained by factor immobility.   Second, they show that making capital more mobile will not only make the geographic allocation of capital more efficient, it will also lower the inequality in industrial development.

Part II: Financing Constraints in Agriculture (in progress)

This part of the dissertation looks at credit costs in agriculture. In an analysis that complements Part I of the dissertation, it examines if there are regional variations in these costs. It identifies credit costs by exploiting a scheme-induced variation in the supply of formal credit to farmers, and takes advantage of the fact that the scheme spans a large cross-section of India.

If production credit is fungible, then the possibility of a buffer stock motive means that a farmer could be credit constrained even when he is visibly underutilizing his credit limit. Conceptually, a clean way to test for the existence of such a credit constraint is to see if a farmer increases his borrowing when there is an exogenous increase in his credit limit. The Kisan Credit Card scheme (KCC), launched nationwide in India in 1999, makes it possible to apply this concept in a wholly agricultural context. The scheme issues a credit card to farmers with a well-specified annual credit limit. This scheme is being implemented by all large national banks, and these banks have different rules for fixing credit limits. These rules are the same across all the branches of a given bank. Because banking rules delineate exclusive zones for rural bank branches, similar farmers in the same district get different credit limits simply because they happen to be served by different banks.

This within-district variation in the credit limit allocated to farmers can be used to measure the extent of credit constraint. The response to this "treatment’’ – the within-district variation in limit to similar farmers — can be studied in terms of several outcomes, including actual borrowing and expenditure on farm inputs. Further, because the data spans several districts from across 11 states, it is possible to compare how this response varies across districts.