Noncompetes in the U.S. Labor Force

Abstract

Using nationally representative survey data on 11,505 labor force participants, we examine the use, implementation, and effects of noncompete agreements. Nearly 1 in 5 labor force participants were bound by noncompetes in 2014, and nearly 40% had signed at least one noncompete in the past. Noncompetes are more likely to be found in high-skill, high-paying jobs, but they are also surprisingly common in low-skill, low-paying jobs as well as in states that do not enforce them. We document that less than 10% of employees negotiate over their noncompetes and that roughly one-third of noncompetes are signed after applicants have already accepted their job offers (but not with a raise or promotion). Differences in the timing of the noncompete are associated with starkly different outcomes: those presented with a noncompete before they accepted the job offer earn 9.7% higher wages, receive 11% more training, and are 6.6% more satisfied in their job than those not bound by noncompetes. However, those asked to sign after accepting a job offer are 12.5% less satisfied in their job and experience no wage and training benefits. We conclude that noncompetes can be efficient contracting tools, but also serve as intertemporal conduits of monopsony power.

Keywords: covenants not to compete, monopsony power, employment law, contracts, wages, training

JEL Classification: J4, J6, K31, L41, M5