Regular Article
Place-based preferential tax policy and industrial development: Evidence from India’s program on industrially backward districts

https://doi.org/10.1016/j.jdeveco.2020.102621Get rights and content

Highlights

  • The backward district program led to a 60% increase in number of firms and employment in light manufacturing industries of the better-off backward districts over four years since it was introduced.

  • The effects represented gains in real economic activity in the treated areas as no significant effect on firm formalization was detected.

  • The program have generated negative spatial spillovers to the non-backward districts geographically and economically close to the better-off backward districts.

  • The program’s effects did not persist six years after the program cut its tax benefits by three fourths and one year after the program ended.

Abstract

We evaluate a tax-exemption program initiated by the Indian government in 1994 to promote manufacturing in districts designated as industrially backward on the basis of a continuous gradation score that reflected district characteristics in early 1990s. Employing a regression discontinuity design, we find that the program led to a significant increase in firm entry and employment, especially in light manufacturing industries of the better-off backward districts in the short run. However, this was partly driven by spatial displacement of economic activity from neighboring districts that narrowly missed qualifying for the program. ​Further, we do not find the effects of the program to persist after it ended.

Introduction

Place-based policies have been popular in both developed and developing countries. Examples of large scale place-based policies include the federal Empowerment Zone Program in the United States established in 1993, European Union (EU)’s various initiatives supported under its Structural Funds targeted at disadvantaged areas and countries within EU, and China’s Special Economic Zones starting in late 1970s, just to name a few (Neumark and Simpson, 2015).

A common goal of place-based policies is to create jobs and spur economic activity by attracting new firms to and/or promoting growth of existing firms in selected areas. The policies usually take one or a combination of the following forms: tax exemptions and subsidies, discretionary grants, special economic zones or industrial parks, and infrastructural support. Place-based policies are often targeted at certain types of industries. For instance, some industrial parks are established to host firms in high-technology manufacturing only.

This paper studies the backward district program of India which was initiated in 1994 and offered new firms a 5-year tax exemption on profits and another 5-year tax reduction afterwards. The program is a typical place-based policy as it was only applicable to 123 industrially backward districts identified out of 360 districts from 14 major states of India. The program was focused only on firms engaged in manufacturing activities.

We attempt to examine both the short-term and long-term impacts of the program with two rounds of economic census data administered in 1998 and 2005, respectively. A key challenge in evaluating the impact of place-based policies lies in the fact that treated areas are often not randomly selected. For a program supporting underdeveloped areas, not only do the treated and untreated areas differ in initial conditions, their development trajectories could also be distinct in the absence of the program. In such cases, untreated areas do not constitute good counterfactuals for quantifying program impacts.

The approach used by the Indian government to identify backward districts, however, offers us a unique opportunity to credibly evaluate the program’s impacts. In short, the central government calculated a composite gradation score for each district across 14 major states based on several indicators such as the urbanization rate and per capita gross value added in the registered manufacturing sector in the early 1990s. Districts with scores below a specific cutoff point were designated as backward districts qualifying for the tax exemption scheme. This setting allows us to use a (sharp) regression discontinuity (RD) design in identifying the causal effects of the program. Upon checking the distribution of the gradation scores, we show that there is no bunching of districts near the cutoff point. Moreover, non-backward districts with scores right above the cutoff point are statistically indifferent from backward districts right below the cutoff in terms of pre-treatment covariates such as population, labor participation rate, and number of manufacturing workers. Hence, RD design, especially applied to districts near the cutoff point should yield plausible estimates of the program’s effects.

Our main findings are as follows. First, the program led to remarkable growth in the manufacturing sector in the better-off backward districts, which account for about one third of all backward districts with gradation scores nearest to the cutoff, over four years after the policy was introduced. The effects largely accrued to light manufacturing industries (with about a 60% increase in both number of firms and employment), while some results suggest a modest increase (about 20%) in heavy manufacturing industries. The results prove to be highly robust when a variety of alternative sample treatments, estimation methods, and geographically overlapping policies are considered. A falsification test also lends credibility to the results.

To understand how much of the growth effect arose from a net increase in new economic activity, we investigate the effects of the program on firm formalization and spatial spillovers generated by the program. The evidence suggests that the program did not have a sizable effect on the share of formal firms or their employment in backward districts as of 1998. We do find some evidence suggesting net negative spatial spillovers, although the spatial displacement was not large enough to offset the direct effects of the program on the better-off backward districts. In sum, the analyses support the notion that the program promoted manufacturing development in the treated areas in the short run and generated overall positive effects even after accounting for negative spillovers to the untreated areas.

We do not find the gains in manufacturing growth to persist. In 2005, six years after the program cut its tax benefits by three fourths and one year after the program ended, the estimated effects for manufacturing industries diminish and lose statistical significance. This indicates that the program failed to generate self-sustained agglomeration economies in the treated areas. Thus, the economic returns to the program seem to be rather limited.

This paper contributes to a vast literature on place-based policies in a few ways.1 First, studies on the effects of economic/enterprise zones in creating jobs yield mixed results. For instance, Neumark and Kolko (2010) find the California Enterprise Zones program to have no significant effects in generating employment, whereas Freedman (2013) finds positive effects for the Texas Enterprise Zones program. Hanson (2009) and Busso et al. (2013) also produce an opposite set of findings with respect to the Federal Empowerment Zones program. Our own findings suggest that the success of a place-based policy may depend on the time frame within which it is examined and the specific interventions it adopts. Mere time-bounded tax exemption may attract firms and contribute to increased employment in the short run, but it is less effective in empowering local communities to generate sustainable economic development in the longer term. Studies that find persistent effects of a temporary place-based policy stress the importance of accumulated capital or agglomeration economies resulting from the policy (Ehrlich and Seidel, 2018; Kline and Moretti, 2014), which may help explain the inefficacy of the program in longer term.

Second, work on the spatial spillovers of place-based policies is relatively sparse. Givord et al. (2013) found strong displacement effects of the French Zones Franches Urbaines (ZFUs) program on the nearby non-ZFU areas, which were of comparable magnitude to the positive effects the program generated inside the ZFUs. However, the results for the US programs are quite mixed (e.g. Ham et al., 2011; Hanson and Rohlin, 2013). In our case, we find net negative spatial spillovers for untreated areas which were geographically and economically close to treated areas, although the magnitude is not large enough to completely offset the positive effects of the program.

This paper also extends a growing literature on place-based policies in developing countries, which can be categorized into two types based on their primary goals. One involves programs to boost economic growth in leading areas such as China’s special economic zones, originally established in coastal areas. The other aims to reduce geographical disparities and targets policy interventions at underdeveloped areas. The latter, to which the backward district program belongs to, sometimes bear more political than economic significance.

In an evaluation of China’s experiment with special economic zones, Wang (2013) found that they increased the level of FDI and exports as well as average wages of workers in the hosting municipalities and generated a moderate displacement effect in adjacent municipalities. Shenoy (2018) shows that a comprehensive package including infrastructure investment, subsidy and tax exemption promoted economic development in the new state of Uttarakhand, India in early 2000s. Unlike these policies, the backward district program offered only tax exemption to new firms in selected areas.

More similar to the backward district program is India’s New Industrial Policy that Chaurey (2017) studies, which offered tax exemptions and capital subsidies for firms in two states, Uttarakhand and Himachal Pradesh, since 2003. Applying a difference-in-differences approach, the author showed that the policy resulted in large increases in outcomes such as employment, number of firms, and total output in treatment states relative to control states. In addition, there is no evidence for any spatial displacement of firms.

However, the two policies differ in a few important aspects. For instance, the backward district program had nationwide coverage with treatment implemented at district level and only new firms eligible for the program. These features enable us to implement a distinctive identification strategy and probably contribute to our different conclusions. In particular, while we obtain similar effects in the short run, these do not persist after the program ended. Moreover, we confirm that attention should be paid to spatial displacement in designing place-based policies in the future.

This paper is also connected to the literature on the location and growth of firms in response to local taxation. The available evidence is somewhat mixed. Some studies, for example, Rathelot and Sillard (2008) find a weak response of firms’ location choice to higher taxes, while others, for example, Bartik (1991) and Guimaraes et al. (2004), find a negative relationship. After correcting for potential endogeneity issues, Duranton et al. (2011) find that local taxation has a negative impact on firm employment but no effect on firm entry in the United Kingdom. Our work complements this literature by examining the entry and formalization decisions of industrial firms in response to a large-scale preferential tax policy.

Section snippets

Policy background

The spatial pattern of economic activity in India has been characterized by a concentration of industrial development around large cities and development skewed towards a few states. This concentration has grown with time as shown by studies such as that of Ghosh et al. (1998) which reported a rise in the coefficient of variation of per-capita state domestic product from 25% in 1950–51 to 35% in 1993–94. This has occurred despite balanced regional development being a key policy objective of the

Data

We use establishment-level data from India’s Economic Census of 1998 and 2005 (EC 1998 and EC 2005 thereafter), and district-level data from the Primary Census Abstract of 1991 (PCA 1991), in conjunction with the districts’ gradation scores published in the “All India Gradation List” of the Income Tax Act, to evaluate the impacts of the backward districts program on local industrial development.6 The EC, administered by the

Empirical strategy

The major identification challenge in evaluating the impact of the backward district program is that backward districts are likely to be distinct from the non-backward districts in both pre-program conditions and development trends without the program. However, the way the government identifies backward districts allows us to apply a regression discontinuity design to credibly estimate the causal effects of the program.

As described earlier, each of the 360 districts from 14 states considered

Economic activity

We first examine how the program affected local economic activity as measured by the number of firms and employment. Before reporting the RD estimation results, we show plots of the log transformed data by industry category in Fig. 3. Each dot represents mean counts of firms or employment averaged across 2-digit industries within each category (i.e. light manufacturing, heavy manufacturing and ineligible industries) and districts falling in a bin of size of 40 gradation scores. The solid line

Long-term impacts of the program

While the backward districts program lasted only 10 years, it is important to see whether the program generated persistent effects on treated areas after it was phased out. As has been recognized in the literature (for example, Krugman, 1991), temporary policy shocks can have persistent effects. For example, an initial spurt in economic activities due to the program might generate self-sustaining agglomeration externalities.

We examine the long-term impacts of the program using Economic Census

Conclusion

Place-based policies have been popular in both developed and developing countries. We evaluate the backward district program initiated by the Indian government in 1994, which aimed to promote manufacturing in underdeveloped areas across the country. The way the program selected backward districts allows us to reliably estimate the program’s impacts on industrial development, at least for a subset of treated districts.

We find that over the short run the program led to faster growth in light

References (36)

  • A. Shenoy

    Regional development through place-based policies: evidence from a spatial discontinuity

    J. Dev. Econ.

    (2018)
  • J. Wang

    The economic impact of special economic zones: evidence from Chinese municipalities

    J. Dev. Econ.

    (2013)
  • T. Bartik

    Who Benefits from State and Local Economic Development Policies, Kalamazoo, MR

    (1991)
  • M. Bruhn et al.

    Entry regulation and the formalization of microenterprises in developing countries

    World Bank Res. Obs.

    (2014)
  • M. Busso et al.

    Assessing the incidence and efficiency of a prominent place based policy

    Am. Econ. Rev.

    (2013)
  • S. Calonico et al.

    Robust nonparametric confidence intervals for regression-discontinuity designs

    Econometrica

    (2014)
  • S. Calonico et al.

    Optimal bandwidth choice for robust bias corrected inference in regression discontinuity designs

    Econom. J.

    (2020)
  • S. Calonico et al.

    Rdrobust: software for regression discontinuity designs

    STATA J.

    (2017)
  • Cited by (17)

    View all citing articles on Scopus

    The authors are grateful to Gilles Duranton for insightful guidance and three referees and one co-editor for helpful comments. Seminar participants at ADB, Centre for Policy Research (New Delhi), University of Tokyo, and the Urban Economic Association Meeting 2018 also provided useful suggestions. The views expressed in this publication are those of the authors and do not necessarily reflect the views and policies of ADB or its Board of Governors or the governments they represent. ADB does not guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use.

    View full text