The Effects of Noncompetes on Workers and Employers: Evidence from a Large Field Experiment
Early Career Research Grant
Over the last 40 years, the startup rate has halved, job mobility has declined 22% (Molloy et al. 2016), and real wages for the middle quintile of earners has risen on average by just 0.09% per year (Shambaugh et al. 2017). While many factors underlie these patterns, one recent argument emphasizes employer monopsony power (Council of Economic Advisors 2016). Labor market imperfections, such as search and moving costs, may allow firms to keep wages and job-to-job mobility artificially low. One prominently cited employment practice in particular has come under increased scrutiny: Covenants not to compete (“noncompetes”) which directly curtail individual freedom to pursue better, higher-paying jobs. Prior work has shown that noncompetes are surprisingly common in the U.S., affecting around 40% of US workers at some point in their work-life (Starr et al. 2019a). Interest in reforming noncompetes has exploded in policy circles, with dozens of recent bills proposing reforms at the federal and state levels. The central concern of policymakers is whether noncompetes suppress wages for workers and limit their freedom to work where they choose. Theoretically, it is unclear why workers would be hurt by noncompetes: Why would workers voluntarily agree to provisions that make them worse off (Callahan 1985, Friedman 1991)? This tension—between voluntary contracting and prohibiting worker mobility—is at the heart of the noncompete debate. However, because the use of noncompetes is nonrandom, existing empirical work has yet to causally identify the effects of noncompetes —a fact that has given policymakers pause. Indeed, a recent review of the literature by the Federal Trade Commission concluded that “Further research … is necessary to establish the causal impact such agreements have on market participants” (McAdams 2019). By assuming the role of the employer in a large field experiment, our study will be able to provide the first causal estimates of the effects of noncompetes themselves on worker wages and job mobility.