In new research exploiting state-level changes in non-compete enforceability, Kate Reinmuth and Emma Rockall find that stronger non-competes have historically reduced innovation in the United States. These declines are driven by sharp drops in inventor mobility and knowledge spillovers, especially in young, high-growth sectors.
For decades, American workers have been signing away their right to switch jobs. Non-compete agreements—clauses that restrict where employees can work after leaving a firm—have spread far beyond the executive suite. Today, estimates suggest that up to 45 percent of U.S. private sector workers, from software engineers to sandwich makers, are bound by these provisions.
Proponents argue that non-competes are essential for protecting companies’ investments in innovation: if workers can’t take ideas to competitors, firms will be more willing to invest in research and development in the first place. But this logic rests on the assumption that tying workers to firms ultimately produces more innovation than allowing them to move freely.
Our research suggests otherwise. We ask a simple but consequential question: do restrictions on labor mobility actually increase innovation? Using state-level policy changes as natural experiments and comprehensive U.S. data on a variety of innovation-relevant outcomes, we test this assumption directly. Our main finding is that increases in non-compete enforceability have reduced innovation on average.
How changes in state law let us test the innovation story
States differ in how strongly they enforce non-competes. Some, like California, make them almost entirely unenforceable. Others permit them broadly. Critically for our purposes, state policy has not been static: courts and legislatures have repeatedly changed enforceability rules over time, sometimes strengthening non-competes and sometimes weakening them.
These shifts create an unusually rich policy laboratory. We match these policy changes to detailed data on patenting, inventors’ job moves, business formation, and firm productivity. This setup allows us to observe what happens after states make non-competes easier or harder to enforce. The data show a consistent pattern: when non-competes become more enforceable, innovation declines.
When non-competes get stronger, patenting falls
The rate of patenting is a widely used metric for innovation. If non-competes worked as their proponents suggest, stronger enforceability should increase innovation and thus patenting. Our findings instead show that the net effect of stronger non-competes on innovation is negative. After a typical increase in enforceability, we find that corporate patenting falls by about 14 percent within five years. The decline unfolds gradually and persists, suggesting that stronger non-competes do not merely delay invention but reduce it.
We find particularly large declines among the most innovative patents, which we define in several ways. First, when we weight patents by how often they are cited by later inventions—a standard proxy for technological impact—patenting still falls at about the same rate as in the broader, unweighted case. This suggests that the results are not driven by the disappearance of low-value patents. High-impact inventions are disappearing too.
Second, we examine novelty using backward citations. Patents that cite fewer prior inventions tend to reflect more original ideas. These are the patents that push the technological frontier. Here, the effects are even stronger: novelty-weighted patenting falls by about 21 percent.
Third, we look at technology field age. In younger, faster-moving areas (such as many modern information technology and software fields), patenting falls by 32 percent, compared to about 13 percent in older fields. Innovation in emerging industries, where experimentation and worker movement are especially important, seems to be disproportionately impacted.
We also look at venture-backed firms, which are widely recognized as hubs of high-value innovation. For firms that have ever received venture capital funding, patenting declines by about 23 percent.
Taken together, these results show that stronger non-competes disproportionately reduce high-value innovation, particularly in young technological fields and venture-backed firms.
Productivity also declines
Patents capture only part of innovative activity. Many productivity improvements arise from organizational changes, new processes, and tacit know-how that never appear in patent filings.
To capture these broader effects, we study firm productivity using Compustat data on publicly traded firms. When non-competes become more enforceable, average firm productivity declines by about six percent after five years. The timing is intuitive: patenting falls first, and productivity follows with a lag as the slowdown in innovation works its way into firm operations.
Non-competes limit knowledge flows
Proponents argue that non-competes induce incumbents to innovate more by strengthening their incentives. We find the opposite. Incumbent firms innovate and patent less, not more, when non-competes become stronger. Why does this happen?
One possibility is that new firms stop entering. However, while non-competes may create barriers to hiring and entrepreneurship, we find little evidence that overall business formation declines significantly following increases in enforceability. When we focus specifically on innovative entry—patents by newly founded firms—the estimated effect is negative but small, while the effect on incumbent patenting remains large. Thus, entry alone cannot explain the overall drop in innovation.
What does change dramatically is labor mobility. Following a typical increase in enforceability, the share of inventors changing employers falls by 31 percent after five years. Among inventors at venture-backed firms, mobility drops by 64 percent. These are not small frictions; they represent large changes in how talent circulates through the economy.
This matters because inventors carry more than just their formal skills. They carry tacit knowledge: how a system really works, which ideas have been tried and abandoned, what failed in the lab, and what might work better next time. When inventors move, that knowledge travels with them. When they do not, it stays locked inside firms. We can see this dynamic in citation patterns: after non-competes become more enforceable, firms cite their own prior patents more and cite other firms’ patents less. Firms become more inward-looking in their technological development, consistent with weaker knowledge spillovers when workers are less able to move.
Thus, stronger non-competes do not simply bind workers. They inhibit the flow of knowledge.
Alternative explanations
Two alternative explanations could undermine our results but do not fit the data.
First, firms might respond to stronger non-competes by shifting their intellectual property strategy away from patents and towards trade secrecy. This would generate a decrease in patenting without a true decrease in underlying innovation. But if this were the case, we would expect to see compositional shifts in the types of patents filed—for example, a relative decline in process patents (which are easier to keep secret) compared to product patents (which are harder to hide once on the market). We do not observe these shifts. Instead, the structure of patenting remains broadly stable, suggesting that what changes is the overall level of innovative output.
Second, inventors might simply move across state borders when local non-competes tighten, such that we observe a reallocation of innovation rather than a net loss. In practice, we find that non-competes reduce within-state mobility but do not generate meaningful increases in cross-state inventor migration.
A challenge to conventional wisdom
For decades, non-competes have been justified on the grounds that they promote innovation by safeguarding firms’ investments in knowledge. Our evidence plainly contradicts that view—at least for current levels of enforceability in the United States. This does not mean that non-competes never confer benefits. At some levels of enforceability, incentives for firms to invest in employee training or long-term, risky research projects may offset the downsides of reduced mobility. But our results indicate that stronger non-compete enforceability at current levels reduces overall innovation and productivity, particularly in the most dynamic parts of the economy.
The broader lesson is that innovation is not generated by isolated firms alone. It emerges from networks of people and ideas, recombining knowledge across organizations and sectors. When legal rules fracture those networks by restricting worker movement, innovation weakens.
Author Disclaimer: The views expressed herein are those of the authors and should not be attributed to the IMF, its Executive Board, or its management.
Authors’ Disclosures: The authors report no conflicts of interest. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
Subscribe here for ProMarket’s weekly newsletter, Special Interest, to stay up to date on ProMarket’s coverage of the political economy and other content from the Stigler Center.
