Do non-compete agreements help or hurt workers? Evidence from the NLSY97

While non-compete agreements (NCs) are prevalent, the incentives driving their use and their causal effects on workers remain poorly understood. We develop a model with asymmetric information to show that NCs shift the nature of allocative inefficiency by reducing inefficient quits and increasing inefficient retention, while mitigating the canonical hold-up problem. The model predicts that NCs are more likely to be used in industries with high returns on industry-specific investments, and that signers have longer job tenures, higher wages, and receive more firm-provided investment than similar workers without such agreements. To test these predictions, we use panel data from the NLSY97 and a difference-in-differences research design to estimate the causal impact of signing an NC. We find that NCs raise job tenures by 6% and lead to an immediate wage increase of 10%. Six years after signing, the wage premium falls to 5%. There is also substantial heterogeneity across worker demographics, with non-White, non-college and lower-wage workers experiencing lower wage-growth after signing an NC. While the theory links NCs to firm investment, we find no evidence of increased investment in formal training, suggesting investments prompted by the agreement are likely informal. Our findings caution against blanket bans on NC usage, favoring a more targeted approach focusing on lower-wage workers.