Peer-Reviewed Research

Descriptive Representation and Conflict Reduction: Evidence from India’s Maoist Rebellion. (with Aidan Milliff). Accepted at Journal of Peace Research. Replication Data.


Can greater inclusion in democracy for historically-disadvantaged groups reduce rebel violence? Democracy-building is a common tool in counterinsurgencies and post-conflict states, yet existing scholarship has faced obstacles in measuring the independent effect of democratic reforms. We evaluate whether quotas for Scheduled Tribes in local councils reduced rebel violence in Chhattisgarh, an Indian state featuring high-intensity Maoist insurgent activity. We employ a geographic regression discontinuity design to study the effects of quotas implemented in Chhattisgarh, finding that reservations reduced Maoist violence in the state. Exploratory analyses of mechanisms suggest that reservations reduced violence by bringing local elected officials closer to state security forces, providing a windfall of valuable information to counterinsurgents. Our study shows that institutional engineering and inclusive representative democracy, in particular, can shape the trajectory of insurgent violence.

New Data on Indian Internal Security Force Fatalities and Demographics. (with Paul Staniland). India Review. 2018. Replication Data.


National and state-level security forces across India operate against insurgents, criminals, and external threats. These operations are politically consequential, yet these forces tend to be quite opaque. This article provides new data on the fatalities that these forces have suffered in order to explore the location and nature of political violence in India. We create several new datasets of security force fatalities extracted from commemorative security force “martyrs” documents and online databases published by Indian state-level police organizations, the Ministry of Home Affairs (MHA), and the Ministry of Defence, as well as semi-official sources. The data vary wildly in quality and detail, and there are serious limits to their use. Nevertheless, they allow us to – with caveats – measure the location and incidence of violence, as well as the demographic underpinnings of the Indian Army, the two largest MHA paramilitaries, and several state police forces. Caveats aside, we anticipate that subsets of these data are sufficiently high in quality, facilitating future rigorous quantitative analysis on political violence in India. The entire dataset will be made publicly available.

Selected Research in Progress

Civil Wartime Political Parties: Evidence from Four Rebellions in India. Under Review.


Elections frequently occur during civil wars. Yet, existing research does not adequately conceptualize the political parties contesting them nor, in particular, the party organizations which are ideologically aligned with insurgents but opposed to overthrowing the state. This study conceptualizes such non-revisionist parties as *aligned status quo* party organizations and situates them in a novel typology of civil wartime political parties. Regression discontinuity and qualitative analyses of four civil wars in India find that in localities where aligned status quo parties won elections, overall voter turnout significantly declined in subsequent election cycles and can be explained by rebels and state security forces intensifying their operations in these areas. The results illustrate that aligned status quo parties uniquely face counterinsurgent and insurgent coercion, reinforce the typology's utility, and contribute to our understanding of political parties and civil wartime democratic institutions.

On the reliability of published findings using the regression discontinuity design in political science. (with P. M. Aronow and Fredrik Sävje). Under Review.


The regression discontinuity (RD) design offers identification of causal effects under weak assumptions, earning it the position as a standard method in modern political science research. But identification does not necessarily imply that the causal effects can be estimated accurately with limited data. In this paper, we highlight that estimation is particularly challenging with the RD design and investigate how these challenges manifest themselves in the empirical literature. We collect all RD-based findings published in top political science journals from 2009--2018. The findings exhibit pathological features; estimates tend to bunch just above the conventional level of statistical significance. A reanalysis of all studies with available data suggests that researcher's discretion is not a major driver of these pathological features, but researchers tend to use inappropriate methods for inference, rendering standard errors artificially small. A retrospective power analysis reveals that most of these studies were underpowered to detect all but large effects. The issues we uncover, combined with well-documented selection pressures in academic publishing, cause concern that many published findings using the RD design are exaggerated, if not entirely spurious.

Partisan alignment and rebel violence: Evidence from India.


Elections to multiple levels of government often occur during civil wars, generating uncertainty about whether elected officials in insurgency-affected areas exhibit partisan alignment with the central government. Prior scholarship is ambiguous about whether this alignment affects rebel violence. This study implements a regression discontinuity design using evidence from six civil war settings in India to evaluate how electing a state-level representative aligned with the central government affects insurgent violence. The main RD results indicate that while alignment does not affect rebel violence shortly after elections, insurgent attacks increase two years later and intensify thereafter for the next three years. Complementary analyses demonstrate that this result is best explained by the central government favoring localities represented by copartisans with higher levels of natural resource extraction as well as industrial and agricultural development, which jointly pose an increased threat to rebels and incite their violence. The findings in this study shed light on the importance of democratic institutions and economic activity during civil war.

Who Should Fight? Experimental Evidence on Policy Corrections to the Unequal Costs of U.S. Wars. (with Daniel Goldstein). Under Review.


The physical costs of war --- who fights and experiences casualties --- are borne unequally in the United States. Yet, little is known regarding how informing individuals of this disparity affects preferences over how to address it. We introduce a framework of `policy corrections,' which differentially allocates to socioeconomic groups the costs associated with public good provision. A survey experiment demonstrates how informing Americans that low-income communities disproportionately bear the physical costs of U. S. wars impacts their support for specific policy corrections. We find enhanced support for greater military recruitment on the richest half of Americans (a direct correction) but unaltered preferences for increasing taxes on this group (an indirect correction). Effects are consistent across income, partisanship, and race. Our results suggest that war casualties transcend socioeconomic in-group calculus and, moreover, even individuals who benefit from present policies support redressing the unequal costs associated with the provision of defense.