Research

Working Papers

"Who Watches the Watchmen? Local News and Police Behavior in the United States"

with Nicola Mastrorocco (Trinity College Dublin). April 2020.

How do the media influence local institutions? We explore the question by looking at how acquisitions of local TV stations by a large broadcast group, which are likely to decrease local coverage in favor of national news, affect U.S. municipal police departments. To capture variation in local media, we implement a triple differences-in-differences design which exploits the staggered acquisition of local TV stations by the Sinclair Broadcast Group 2010-2017, together with cross-sectional variation in whether municipalities tend to be covered by local news at baseline. First, using a newly collected dataset of transcripts of local newscasts, we document that when a station is acquired by Sinclair, covered municipality experience a decline in the probability of appearing in the news with a crime story relative to non-covered municipalities. Second, we find that after Sinclair enters a media market, covered municipalities have lower violent crime clearance rates with respect to non-covered municipalities. Instead, we find no differential effect of Sinclair acquisitions on property crime clearance rates, which are minimally covered in local news. Increases in crime rates or resources allocated to police departments do not explain the result, reinforcing the interpretation that police effort allocation across different activities is changing as a result of changes in local news content.

"Subsidies and the Dynamics of Selection: Experimental Evidence from Indonesia's National Health Insurance" (submitted)

with Abhijit Banerjee (MIT), Amy Finkelstein (MIT), Rema Hanna (Harvard University), Ben Olken (MIT) and Sudarno Sumarto (TNP2K and SMERU). April 2020.

Previous version: NBER Working Paper #26204; VoxDev article; Online Appendix.

How can developing countries increase health insurance? We experimentally assessed three approaches that simple theory suggests could increase coverage and potentially reduce adverse selection: temporary price subsidies, registration assistance, and information. Temporary subsidies attracted lower-cost enrollees, in part by reducing strategic coverage timing. While subsidies were active, coverage increased more than eightfold, at no higher unit cost to the government; after subsidies ended, coverage remained twice as high, again at no higher cost. However, subsidies are not sufficient to achieve universal coverage: the most intensive intervention – a full one-year subsidy combined with registration assistance – resulted in only 30 percent enrollment.

"Stereotypes in High-Stakes Decisions: Evidence from U.S. Circuit Courts" (submitted)

with Elliott Ash (ETH Zurich) and Daniel L. Chen (Toulouse School of Economics). March 2020.

Stereotypes are thought to be an important determinant of decision making but are hard to systematically measure, especially for individuals in policy-making roles. In this paper, we propose and implement a novel language-based measure of gender stereotypes for the high-stakes context of U.S. Appellate Courts. We construct a judge-specific measure of gender-stereotyped language use – gender slant – by looking at the linguistic association of words identifying gender (male versus female) and words identifying gender stereotypes (career versus family) in the judge's authored opinions. Exploiting quasi-random assignment of judges to cases and conditioning on detailed biographical characteristics of judges, we study how gender stereotypes influence judicial behavior. We find that judges with higher slant vote more conservatively on women's rights’ issues (e.g. reproductive rights, sexual harassment, and gender discrimination). These more slanted judges also influence workplace outcomes for female colleagues: they are less likely to assign opinions to female judges, they are more likely to reverse lower-court decisions if the lower-court judge is a woman, and they cite fewer female-authored opinions.

"Closing Time: the Local Amenities Effect of Prohibition" (submitted)

with Greg Howard (University of Illinois Urbana-Champaign). January 2020.

How do amenities affect local land values, production, and sorting? We study the question exploiting a large historical policy change: U.S. Alcohol Prohibition in the early 20th century. Comparing same-state early and late adopters of county dry laws in a difference-in-differences design, we find that early Prohibition adoption increased population and farm real estate values. Moreover, we find strong effects on farm productivity consistent with increased investment due to a land price channel. In equilibrium, the amenity change disproportionately attracted immigrants and African-Americans.

Does reducing politicians’ control over public employees’ hiring and firing improve bureaucratic performance? I answer this question exploiting population-based mandates for U.S. municipal police department merit systems in a regression discontinuity design. Merit system mandates improve performance: crime rates are lower in departments operating under a merit system than in departments under a spoils system. Changes in resources or police officers’ characteristics do not drive the effect, but I provide suggestive evidence that the limitations to politicians’ ability to influence police officers are instead important.