Research

We study how peers affect in-group bias. Exploiting several umpiring reforms in international cricket matches—where two umpires make independent decisions in each other’s presence—we show that home-team umpires are less biased when working with a neutral colleague, that is, one who is neither a national of the home nor the foreign team. This temporary debiasing is driven by the social pressure umpires feel to be impartial in the presence of neutral peers. Performance evaluation by visually non-salient monitors does not reduce bias, suggesting that physical presence is an important component of debiasing and peer influence.

Policies that broaden political representation empower new leaders who may lack knowledge of how government works. We study whether peer networks among local politicians facilitate knowledge transfer and improve the quality of governance. We run an experiment in partnership with the Government of Bihar, India, where we organised peer groups for randomly selected village leaders. We examine whether these peer networks increase politicians’ knowledge of how to manage the development programs under their charge and improve the delivery of public services. We also test whether politicians from marginalised groups benefit more from peer networks. To understand mechanisms, we test if peer groups diffuse information about governance best practices and help politicians organise collective action. 

We seek to understand how poor mental health, through its effects on cognition and perception, affects behavior with political implications. We will implement a randomized controlled trial around the 2024 general elections in India in which the treatment group will be offered one-month of subsidized mental health care through an online app. We will use this intervention to study how improvements in mental health change the news that people want to consume and affect their political attitudes and behaviors. 

I study how political dynasties affect economic development, by compiling biographical data on all Indian legislators since 1952 and analysing three distinct natural experiments. I present three results. First, descendants of political dynasties worsen local economic development (descendant effects), since inheriting political support weakens performance incentives. Second, founders of political dynasties improve economic development, partly because the opportunity to establish a dynasty encourages politicians to perform well in office (founder effects). Third, dynastic politics has negative long-run impacts, as descendant effects outweigh founder effects. A political agency model where descendants inherit ability and political capital can explain these patterns.

We compare how politicians and bureaucrats govern.  We analyse a natural experiment in Karnataka, India, which generated quasi-random variation in whether a village was governed by politicians or bureaucrats during 2020-21.  Politicians are better at translating citizens' preferences into policy outcomes: public spending in politician-run villages is more aligned with citizen preferences, reflecting politicians' greater engagement with citizens.  Politicians are also more responsive to citizens: they deliver more social assistance during periods of economic distress and get front-line officials to exert greater effort.  On the other hand, bureaucrats are less susceptible to elite capture and perform better at specialised governance tasks that utilise their expertise.  Politicians and bureaucrats perform similarly in terms of controlling corruption, internalising externalities, managing crises, and promoting local economic growth.  Nevertheless, most citizens prefer to be governed by an elected leader.

Salaries are the largest expenditure item for most governments. We examine the effects of increasing bureaucrats’ pay, by analyzing a policy change in Telangana, India, where one group of frontline bureaucrats received a 91% pay increase while another group of bureaucrats performing the same job did not. Using a difference-in-differences design, we show that, on average, higher salaries (i) had no impact on performance, (ii) did not affect dishonest or corrupt behavior, (iii) reduced quit rates by 2.4 percentage points (16%) after 2 years but did not affect the average quality of bureaucrats in service. Higher salaries marginally increased effort among bureaucrats who were unhappy about their pay. Expert forecasters incorrectly predicted that higher salaries would improve both bureaucrat performance and selection.

We evaluate two major efforts to decarbonise energy-intensive cryptocurrency mining.  First, we analyse the impacts of China's cryptomining ban using a difference-in-differences design.  The ban decreased electricity consumption in mining-intensive Chinese provinces by 5\% and shifted electricity generation towards non-fossil sources.  While effective within China, the ban induced a global relocation of mining activity, with China exporting and the US importing more cryptomining equipment.  This relocation increased electricity consumption and fossil generation in mining-heavy US states, diluting the global impact of China's ban by 30-40\%.  Second, we examine Ethereum's transition to Proof-of-Stake, a cleaner production technology.  Using a regression discontinuity design, we find that Ethereum's switch reduced US electricity consumption.  However, miners redeployed their equipment to mine other cryptocurrencies, offsetting the impact of Ethereum's transition by 30\%.  Our findings illustrate the difficulty of decarbonising energy-intensive industries with mobile, redeployable production factors.

In 1992, the 73rd Amendment to the Indian Constitution created 250,000 village democracies (called Gram Panchayats) covering 800 million citizens. It mandated regular elections, deliberative spaces, and political reservations for women and disadvantaged castes. The unprecedented variation in democratic experience that emerged from this has resulted in a large body of research that provides insights into the intersection between democracy, governance, and development. This paper reviews this literature, showing that India’s democratic trajectory has been shaped by four broad forces: a 3,000 year tradition of debate and deliberation, colonial policies, the contrasting ideologies of central players in the formation of modern India—Gandhi and Ambedkar—and the 73rd Amendment. The paper distills key findings from the empirical literature on the effectiveness of local politicians and bureaucrats, political reservations, public finance, deliberative democracy, and service delivery. It concludes with a set of policy recommendations for improving the functioning of the Panchayats in India, emphasizing the need for greater devolution and improved local fiscal capacity. It also argues that urban governments in India would benefit from learning from the experience of Gram Panchayats.

This paper studies the long-run labour impacts of vocational secondary education using a natural experiment in Singapore and social security administrative data.  Students are assigned to tracks in secondary school based on a cutoff score in the nationwide Primary School Leaving Examination.  Using a fuzzy regression discontinuity design, we find that students assigned to the vocational track are more likely to complete secondary school and attain a post-secondary (usually vocational) qualification, but less likely to attain a university degree.  Attending the vocational track has no average effect on later-life employment, earnings, home ownership or wealth, but this masks substantial heterogeneity: vocational education has positive effects at lower income quantiles but negative effects at higher quantiles.  Even though some students appear to benefit from vocational education, nearly all marginal students opt for the academic track when given the choice.  These patterns are consistent with a simple theory where individuals have different learning styles and there is social stigma against the vocational track.