Research

About 45 percent of U.S. union members work in management, professional, and related occupations, yet most evidence on what unions do comes from blue-collar jobs and occupations paid on fixed salary schedules. Faculty unionization in U.S. higher education offers a setting to study bargaining where pay is set individually and tied to productivity. We combine the staggered timing of faculty unionization between 1993 and 2023 with institution-level salaries and headcounts from IPEDS and restricted-use Survey of Doctorate Recipients (SDR) microdata, which allow us to follow doctorate-holding faculty over time.

Unionization does not raise average salaries at any tenure-line rank: under parallel trends, we rule out gains of more than about $900 a year for assistant professors and $1,400 for full professors. Unionization does change employment. Total faculty employment is unchanged, but the median unionizing institution employs four to eight fewer assistant professors than comparable institutions, following an expansion in the years before unionization. Faculty who stay at unionizing institutions see faster salary growth, largest for non-tenure-track faculty and, among tenure-line faculty, concentrated below full professor. Exits and movement along the tenure track do not change. The findings broaden the margin of adjustment to collective bargaining from pay alone to the composition and careers of the academic workforce.

We examine whether legalizing local alcohol sales affects traffic safety, exploiting close local option elections in Texas as a quasi-random source of variation. Over 300 Texas cities voted to legalize alcohol sales between 2003 and 2019. Legalization generated large increases in off-premise alcohol licenses (liquor stores and grocery stores) with little change in on-premise establishments, and shifted alcohol sales from bars and restaurants toward retail outlets for home consumption. Despite these large changes in access, we find no meaningful increase in total crashes, DUI crashes, or fatal crashes, and no evidence of spillover effects on neighboring cities. This null result reflects which type of access was expanded: on-premise licenses are positively associated with crash rates while off-premise licenses are not, consistent with bars and restaurants (not retail stores) being the channel through which alcohol access raises traffic risk.

We will estimate the causal impact of legalized sports betting on the academic success of college students. Specifically, we employ a quasi-experimental research design to investigate how sports betting affects students’ academic performance and engagement, including course grades, classroom participation, and academic progress at Purdue University. Because betting activity and betting-related attention can respond sharply to the timing and salience of live sporting events, we leverage quasi-random variation in students’ exposure to such events that arises from the interaction of individual fandom and course-specific academic schedules. This design compares students with similar interests in sports who, due to differences in course schedules, are more or less exposed to games on the days of classes and exams, generating plausibly exogenous variation in engagement and wagering behavior. Our findings will inform university administrators and state lawmakers as they fine-tune policies surrounding this newly legalized practice, particularly with respect to its potential impacts on college students’ academic progress.