Welcome to my website! I am a Ph.D. candidate in economics at the University of California, Berkeley, with a Designated Emphasis in Political Economy.
I work in political economy and development economics. My job market paper studies collusion and prices in public procurement in India. I have also worked on political reservations and conflict in India, democratic transitions and state repression, electoral cycles in deforestation, and healthy ageing.
Before my Ph.D., I worked as a predoc with Robin Burgess and Oriana Bandiera at the LSE, where I also did a masters in economics.
Working Papers
Job Market Paper
Measuring Bid Rigging at Scale: Detection and Prices in Indian Procurement
Draft coming soon.
Abstract [+] [–] Public procurement accounts for a third of government spending across the world. Collusion among firms bidding for government contracts is a central concern of competition authorities because it can drive up the prices governments pay, yet its prevalence and effects are largely unmeasured, particularly in settings with limited data and weak enforcement. Existing methods to detect collusion test whether priced bids are consistent with competitive bidding. I develop a new method that detects collusion from the behavioral markers that colluding pairs of firms leave when they coordinate. The excess mass of these markers in the observed data, relative to data generated without coordination, identifies both the prevalence of collusion and a ranking of pairs of bidders by how likely they are to collude. I apply the method to a novel dataset covering the near-universe of Indian electronic procurement: 42.3 million bids on 15.7 million tenders from 76 government portals since 2008. The markers follow the Competition Commission of India's guidance to procurement officers, such as bids submitted minutes apart or rotation of the winning bidder. Cartels are heterogeneous in the markers they adopt, so detecting collusion one marker at a time misses much of it. Collusion is widespread and diffuse: about half of regular bidders collude with at least one partner, one in six regularly co-bidding pairs collude, and colluding pairs enter together and win a quarter of tenders. These tenders settle 2.1–3.5 percent higher than comparable tenders. Incidence varies with procurement rules: it is higher under minimum-bidder requirements and lower under price floors with lottery tie-breaks. Over a decade, this overpayment amounts to roughly $18–30 billion across Indian public procurement.
Abstract [+] [–] This paper analyzes the effect of an important power-sharing institution — political reservations for historically disadvantaged minorities — on low-level subnational conflict in India. Previous research has identified positive impacts of reservations on public goods provision benefitting the targeted minorities. However, reservations may come to the detriment of other groups, leading to backlash and potentially higher conflict. I empirically investigate this question with panel data from Indian districts and assembly constituencies. Exploiting exogenous variation in reservations stemming from rules in the process of apportioning reserved seats, I show that reservations consistently lead to lower levels of conflict. The combination of district- and constituency-level analysis shows that this is not a displacement effect, and that potential backlash at the district level is outweighed by direct conflict reduction. I explore mechanisms behind this result and show that (1) conflict reductions have been particularly pronounced for killings of security personnel, (2) conflict reductions have been stronger in areas with active Naxalite insurgent activity, and (3) conflict reductions were accompanied by reductions in caste-related conflict events and crimes against lower castes.
The Journal of the Economics of Ageing, Volume 26 (2023), 100470[PDF]
Abstract [+] [–] Growing life expectancy and a rising proportion of older people make the issue of whether cohorts are ageing better a key individual, social and economic issue. Using data from the English Longitudinal Study of Ageing we characterise how frailty develops with age, how this differs across demographic groups, whether more recent cohorts are ageing better and what the key areas of focus for health policy should be. We find cohort effects such that frailty at each age has been decreasing over time but that this trend shows modest signs of slowing and is less pronounced for those with lower wealth. Improvements across cohorts reflect improvements in ADLs, cognitive function, and mobility but limited progress in reducing the incidence of diseases such as cancer, cardiovascular disease, etc. We find mobility and ADLs the main driver of average differences across regions but cross-regional differences are driven more by within than between group inequality.
International Studies Quarterly, September 2021, Vol. 65, Issue 3, 753–767[PDF]
Abstract [+] [–] This paper examines the transnational dimensions of low-level conflict and state repression. In this regard, special emphasis is placed on the role of political regimes. Drawing on a simple model, we argue that democracy has opposing effects on conflict intensity. On the one hand, democracy satisfies demand for political participation and thus reduces conflict potential. On the other hand, we highlight that domestic democracy may spur dissatisfaction and conflict abroad, which, in turn, may induce conflict spillovers. As a result, the net effect of democracy on low-level conflict and state repression is ambiguous and depends on the level of democracy in the neighborhood: We predict that democracy is more pacifying in democratic environments and may spur conflict in autocratic environments. By the symmetry of the model, we also predict that democratic environments are more pacifying for democratic countries and may spur conflict in autocracies. Empirical evidence using panel data on different types of low-level conflict and state repression for 160 countries in the period from 1950 to 2011 supports these hypotheses. Additionally, two case studies illustrate the mechanisms of our model.
AEA Papers and Proceedings, May 2021, Vol. 111, 415–419[PDF]
Abstract [+] [–] This paper examines the link between electoral incentives and environmental degradation by exploiting a satellite data set on 107,000 forest fires and 879 asynchronous district elections in Indonesia. Fires represent a cheap but illegal means of converting forested land to other uses, but risk burning out of control and creating substantial negative environmental externalities. We find a significant electoral cycle in forest fires. Ignitions and area burned decline during election years, but steeply increase in the year after. The results suggest that politicians may suppress this activity at times when it might particularly dent their electoral chances.
Abstract [+] [–] We study political contributions by firms in India, which has substantial variation in its campaign finance disclosure rules. Between 2019 and 2024, "Electoral Bonds" allowed firms to donate to political parties without public disclosure. This regime ran in parallel to a mandatory disclosure system, and the bond donations were themselves made public in 2024 following a ruling by India's Supreme Court. Matching firms that donated under both regimes, we find that firms use the anonymous channel to fund parties that do not appear in the portfolio they disclose. While these firms often support large national incumbents when their donations are disclosed, a substantial share of firms evades public disclosure when supporting regional incumbents.
Cognitive Enrichment and Human Capital: Experimental Evidence from Developing Countries
Abstract [+] [–] Children in low- and middle-income countries progress through school without acquiring foundational skills. Most education interventions invest in content, the material children are taught. Less is known about how to build the capacity that makes instruction productive: sustained attention, planning, and reasoning. This paper asks whether this capacity can be developed directly, cheaply, and at scale. We report two randomized controlled trials of a school-based chess program, a game that trains attention and planning but teaches no curriculum. In Kenya, 551 primary school students in six schools were randomized to a seven-week program. In Malawi, we expanded the program to 40 public schools and more than 2,100 students, adding curriculum-based tutoring of equal time as a benchmark. In intent-to-treat estimates, chess raised cognitive performance by 0.16 standard deviations (SD) in both countries, with the largest gains in attention in Kenya and planning in Malawi. Despite teaching no mathematics, chess raised numeracy by 0.11 SD, matching the gain from tutoring, and raised children's confidence in their academic ability by 0.14 SD. Tutoring raised cognition by an amount statistically indistinguishable from chess but had no effect on confidence. School examination scores rose by 0.07–0.09 SD. These findings suggest that cognitive skills remain malleable during primary school and that investing in children's capacity to learn, not only the content they are taught, offers a promising and scalable way to build human capital.
Abstract [+] [–] Although there is consensus that full democracies are less repressive than other regime types, the establishment of full democracy is not always followed by reduced state repression. Against that background, this paper examines heterogeneity in the relationship between democracy and government violations of human rights. Drawing on arguments from the civil war literature, we develop a simple model that highlights opposing effects of democracy on state repression. Consequently, the net effect of democracy is shown to be ambiguous. Furthermore, the model reveals that pacifying (adverse) effects of democracy are more likely to dominate in countries with higher (lower) income levels. These implications are tested empirically using different methodological approaches, including time-series cross-sectional regressions, event studies, and a recent generalization of the synthetic control method. Our analyses confirm that democracy is related to reduced repression in relatively rich countries, whereas we find no or even adverse effects in poor countries.
Grier and Grier report that Indian income per head ended 2023 about $1,000 below its synthetic counterfactual. Widening the pool of comparison countries, switching to the GDP series the Penn World Table recommends for growth comparisons, or fitting growth rates rather than levels each flips the sign.
The Interaction Between Covid-19 and an Ageing Society. With Andrew J. Scott. In: VoxEU / CEPR Policy Portal
(2020). Link to article
Sparprojekt oder Milchmädchenrechnung? Die finanziellen Effekte eines Bundeslandes „Mitteldeutschland“ [Fiscal aspects of a ‘Länder’ merger]. With Felix Rösel. In: ifo Dresden berichtet 25/02 (2018). 12–17. Link to pdf
Die Preisentwicklung für Kaufimmobilien in Sachsen – Ein Überblick über die aktuellen Daten [Recent data on housing prices in Saxony]. With Carolin Fritzsche. In: ifo Dresden berichtet 24/06 (2017). 20–24. Link to pdf
Steht sächsischen Schulen ein Qualitätsverlust bevor? Ein Überblick internationaler Studien zur Lehrqualität von Seiteneinsteigern [Review of studies on teaching quality of ‘career changers’]. With Julia Sonnenburg. In: ifo Dresden berichtet 24/06 (2017). 31–34. Link to pdf
Resources
India Census Explorer
A tool to search and download Indian Census tables (1951–2011): filter by state, year, table series, and topic, across all 40,214 entries in the archive. It runs entirely in your browser. Open the explorer →
This website is inspired by Gautam Rao and Lei Yang. You can find the source code on my GitHub repository. Please feel free to copy it for your own website.
↑
