Science Policy Research Report: Employee Non-compete Agreements

Massachusetts Institute Of Technology, Cambridge MA

Investigators

Abstract

Post-employment covenants (a.k.a. "non-compete agreements") have been used for more than 600 years but have attracted considerable attention from U.S. policymakers in the past quarter-century. U.S. states including Florida, Idaho, Georgia, Utah, Hawaii, New York, Oregon, New Hampshire, and Illinois have adopted new legislation regarding non-competes. Supreme Court rulings in Texas, Louisiana, Wisconsin, Illinois, Montana, Colorado, and South Carolina have materially affected the enforceability of non-competes. Moreover, reform efforts are underway in Massachusetts and elsewhere. At the national level, legislation aimed at curbing the use of non-competes for low-wage workers has been considered. The optimal policy prescription for non-compete agreements is hardly straightforward, however, as evidenced by the opposing reforms adopted by various states, some tightening non-compete enforceability whereas others have loosed enforceability. This ambiguity exists in part given that there are multiple stakeholders (workers, firms, and regions) for whom the theoretical impact of non-competes is non-obvious; moreover, the interests of these parties may be opposed. Given both the importance of the policy issue and its theoretical ambiguity, empirical evidence is essential to help guide legislative initiatives. The project focuses on published empirical articles regarding non-compete agreements in the context of the United States. Levels of analysis include implications for workers, firms, and regions. At the individual level, outcomes analyzed include intra- and inter-state mobility, wages, training, motivation, and specialization. At the firm level, outcomes analyzed include financial performance, innovation, employee retention, and merger & acquisition activity. At the regional and national level, outcomes analyzed include productivity, entrepreneurship, job creation, and knowledge diffusion. Causal evidence across these outcomes at various levels is critiqued by reviewing the identification strategies to produce the findings as well as the datasets used to execute those strategies. These identification strategies include assessing the changes in state-level non-compete enforceability, comparing firms or workers more vs. less likely to be affected by non-competes, using instrumental variables that account for the endogeneity of actions that might be influenced by differential non-compete policies across time and states, and randomized controlled experiments. In addition to causal evidence, stylized facts regarding the use of non-competes from surveys and other fieldwork are assessed.