Research
Publications
Polity size and local government performance: Evidence from India with Jeff Weaver
American Economic Review (2024), 114(11): 3385-3426. (lead article)
Developing countries have increasingly decentralized power to local governments. This paper studies the implications of a central element of decentralization – polity size – using population-based discontinuities that determine local government boundaries for over 100,000 Indian villages. Over the short and long-run, individuals allocated into local governments with smaller populations have better access to public goods. We provide suggestive evidence that these results are related to heightened civic engagement and stronger political incentives, without any observable evidence of other mechanisms such as elite capture.
Working Papers
The Limits of Political Representation: Evidence from India with Jeff Weaver
Conditionally Accepted at American Economic Review: Insights
Political inclusion is widely believed to improve governance, which has motivated the creation of elected representatives for highly localized constituencies. This paper studies 1.2 million “hyperlocal” representatives across 150,000 local governments in rural India. Exploiting discontinuities that determine the number and identity of these representatives, we assess how expanded representation affects governance outcomes. We find precisely estimated null effects on core functions, including public project management, intermediation in access to welfare programs, oversight of public finances, and the equity of benefit allocation. These findings highlight the limits of expanding political representation if representative capacity is weak.
The Road Not Taken (Anymore): Checkposts and Local Economic Development with Barthélémy Bonadio
Slides (New draft coming soon)
We examine the impact of a 2017 reform in India that eliminated interstate checkpoints and digitized the process of transporting goods across state borders. The reform generated a significant reduction in inter-state trade barriers and administrative burdens, resulting in a decline in the overall concentration of economic activity near state borders. More specifically, we show that this exogenous shock to location-based forces, generated an overall decline in local economic patterns, and we provide evidence on what characteristics enabled a market to sustain itself despite the removal of these traditional forces of attraction. This quasi-experimental approach provides insights into the dynamics of economic activity in response to policy changes affecting trade and transport.
Work in Progress
Rural Development, Migration & Aggregate Growth: Evidence from India
(New draft coming soon)
This paper examines the effects of rural development policies on economic growth and rural-urban migration in developing countries, using a 1992 reform in India. Despite traditional economic models emphasizing urban migration, rural populations remain high in low-income countries. Using a regression discontinuity design for Uttar Pradesh, I analyze the impact of independent versus shared village councils on economic activity, infrastructure, and population growth. I find that villages with independeny councils witness improved economic activity, enhanced infrastructure, and heightened population growth. While rural-rural migration primarily drives this population surge, districts with more independent councils surprisingly show increased rural-urban migration. This suggests that local-level enhancements reduce migration barriers, thus boosting rural-urban shifts, and ultimately benefiting district-level GDP.
Aggregate Regression Discontinuity Design with Jordan Norris and Zheng Wang
We generalize the regression discontinuity design (RDD) to identify the effect of a binary treatment on aggregated outcomes. The treatment is at the unit level, and units are partitioned into groups. We assume a partial interference structure, where arbitrary interference among units within a group is allowed, but not interference across groups. In this environment, we show that the standard RDD assumption of imperfect control at the cutoff is sufficient for the identification of the effect of the unit- level treatment on the group-level aggregated outcomes. In addition, we derive estimators to estimate the non-linear effect of the number of treated units on the group-level outcomes. We demonstrate our methodology empirically to study the effect of a more uniform partisan composition across US regions on spatial tax competition.