Research
Social media bots influence political news coverage
Winner of best student paper in the Information Technology, & Politics section at American Political Science Association
My paper examines how fake social media accounts boost politicians’ online popularity and this phenomenon’s subsequent spillover on traditional news coverage. Using the ‘Botometer’ algorithm, I assessed the proportion of bot accounts engaging with tweets from 382 U.S. Congress members on Twitter. A policy change to Twitter’s API infrastructure in late November 2022 was an exogenous shock to the platform that significantly hampered bot functionality. My first-stage analysis demonstrated that this policy change only affected high-bot-engagement politicians, who saw a substantial decline in followers after November 2022. Placebo comparisons show this decline was not observed in comparable data from Facebook ‘likes’ or Instagram followers. My second-stage analysis found that after the policy change, high-bot-engagement politicians also saw a decline in popularity in online news articles and broadcast news from December 2022 to February 2024.
New Evidence on the Effects of Randomly Assigned Reservations for Women Leaders in Indian Local Government (with Donald Green and Gaurav Sood)
Influential early studies of reservations for women in India’s local government bodies, like the Gram Panchayat, showed them to have important policy consequences. Outcomes such as local budget allocations seemed to shift as women suddenly entered leadership roles. Specifically, Gram Panchayats, where the role of the Pradhan was reserved for women, tended to allocate more resources to areas that women are thought to be especially concerned about, such as drinking water, sanitation, and maternal care. This paper revisits the ongoing naturally occurring experiment of reservations for women in India’s Gram Panchayats, focusing on Rajasthan, where reservations have been assigned by lottery (in most districts) since the 1990s. We focus on the effects of reservations in two election cycles, 2005 and 2010. Our dataset consists of 5,191 Gram Panchayats, which is more than 50 times larger than the original study of Rajasthan. Using administrative data for outcomes, including some of the same public works spending measures used by the authors of the original study, we find no evidence of substantively meaningful effects on the scale or content of public works programs, even in domains such as drinking water, sanitation, and maternal care. Censuses of local health, infrastructure, and schooling also show no apparent effects. We find no evidence of effects on budget allocations or Census outcomes even when comparing Gram Panchayats that were reserved for women for four successive 5-year terms starting in 2005 to their counterparts that were never reserved for women during this period. Nor do we find effects when examining public works allocations in a different state, Uttar Predesh. We conclude by discussing possible explanations for this pattern of null findings.
Can music concerts boost nationalist identity? (with Dejan Kovac and Jacob Shapiro)
Music and large-scale public events have long been used as part of nationalist projects, but their impact is unclear as those who engage may do so to express preexisting beliefs. We study whether attending nationalist concerts heighten nationalist identity in Croatia, where consumer behavior, i.e., purchase of nationally branded goods labeled ‘Authentic Croatian’ or ‘Croatian Quality’, provides a high frequency, individual-level revealed preference measure of nationalist identity. We first show a robust positive correlation at the store level between such goods and votes for right-wing political parties, which emphasize Croatian national identity. We then use loyalty card tracking to identify individuals who likely traveled to attend nationalist or non-nationalist music concerts. Using a regression discontinuity in time (RDiT) design, we show that attending concerts by a highly-political nationalist artist leads to an approximately three percentage point increase in the share of nationally branded goods purchased, a 17\% increase from the mean, and a 0.19 standard deviation treatment effect. We find no similar impact on consumption after attending non-nationalist concerts or for attending concerts by a Herzegovinian singer who sometimes performs with Croatian nationalist artists.