Postemployment restraints and the displacement of human capital

Matt Marx

Harvard Business School

mmarx@hbs.edu

Scholars and practitioners alike have long maintained that intellectual property protection is a prerequisite for investing in technological innovation (Arrow 1962, Teece 1986).   Indeed, both patents and trade secret protection are widely implemented by technology companies whose most valuable assets “walk out the door every night.”  Recently, however, some have questioned whether such protections (Jaffe and Lerner 2003) may threaten the very process of innovation.  Moreover, work in this vein focuses on patents (Griliches 1990); trade secrets are a “neglected orphan” (Friedman et. al 1991).  

This project investigates the implications for technological innovation and entrepreneurship of contracts designed to enforce trade secret protection.  Although employees signing non-disclosure agreements (NDAs) promise not to divulge trade secrets, the difficulty of detecting violations leads firms to adopt post-employment covenants not to compete agreements (“noncompetes”) to prevent ex-employees from disclosing trade secrets to a competitor (Hyde 2003).  Despite the rapid growth of noncompete litigation (81% in the last ten years), current and pending legislative reforms, and an active discussion of the topic in the venture capital community, our understanding of the effect of noncompetes is limited to a few articles in the economic and sociological literature (Stuart and Sorenson 2003, Fallick et. al 2006). This project aims both to fill that theoretical gap as well as to inform individual, firm, and policy choices.

In a first paper (Marx, Strumsky, and Fleming, under review at Management Science) we exploit Michigan's 1985 inadvertent reversal of its noncompete enforcement policy as a natural experiment to assess whether noncompetes prevent knowledge workers from changing jobs.  Controlling for trends in the auto industry, we find that the intrastate job mobility of patenting inventors dropped 19.8% following Michigan's decision to enforce noncompetes.  Moreover, the mobility of highly-specialized inventors falls further than that of their peers with more generally-applicable skills.  A second paper examines whether similar interstate effects obtain.  Preliminary results suggest a “brain drain” from enforcing states to non-enforcing states, both for knowledge workers as well as knowledge spillovers.  

In a third, qualitative paper, I delve further into the impact of noncompetes on highly-specialized skills.  How do individuals respond when they are restrained from practicing their trade?  Might noncompetes create a “deadweight loss” of specialized human capital?  Given the few existing studies, I undertake a field study to order to build grounded theory. In an initial stage, I interviewed a dozen specialists constructed from a snowball sample of respondents to a Boston Globe story on the aforementioned article as well as others.  This pilot study yielded insights as to how specialists cope with the displacement of their skills, including involuntary “career detours” as well as isolating themselves from extraorganizational colleagues in order to continue to practice their trade without being detected.

In a second stage currently underway, I am refining the theory by interviewing a random sample of the 550 inventors who obtained at least two patents in the automatic speech recognition industry between 1970 and 2006.  (Speech recognition inventors are highly specialized, many of them having obtained a Ph.D. in the field.)  Following this second stage, I plan to conduct a larger-scale survey.  I look forward to feedback from the CCC attendees on my findings thus far as well as the next steps for the fieldwork.