|
Article Excerpt
Introduction
Since Arrow's (1962, p. 615) observation that the "mobility of personnel among firms provides a way of spreading information," the implications of interorganizational mobility have received widespread attention. Scholars have examined the connection between mobility and knowledge spillovers (Stolpe 2002, Rosenkopf and Almeida 2003, Agrawal et al. 2006), research and development (R&D) investment (Kim and Marschke 2005, Singh 2007), and entry by spinoffs (Klepper and Sleeper 2005, Gompers et al. 2005). Generally speaking, this literature has treated mobility as exogenous, paying less attention to its antecedents than its implications. To be fair, other scholars have sought to understand antecedents of employee turnover. Psychologists have documented the influence of attitudinal differences (Porter and Steers 1973, Mobley et al. 1979); sociologists have studied the role of organizational demography (Wagner et al. 1984) and social capital (Granovetter 1973, Marsden and Hurlbert 1988); organizational researchers have differentiated between the ease and desirability of turnover (March and Simon 1958); and labor economists have examined contractual conditions favoring the retention of key scientists (Pakes and Nitzan 1982, Anton and Yao 1995).
Few of these studies, however, have taken into account the potential influence on mobility of post-employment covenants not to compete (hereafter, "non-competes"). This omission is particularly puzzling given the prevalence of such contracts among technology companies whose most valuable assets "walk out the door every night" (LaVan 2000). Recent research, mainly in regional policy and entrepreneur-ship, has begun to investigate non-competes. Gilson (1999) proposed that Silicon Valley's entrepreneurial growth can be attributed to California's proscription of non-competes. While mobility of California inventors does appear to be high (Almeida and Kogut 1999) and more startups appear in regions that do not enforce non-competes (Stuart and Sorenson 2003), causal evidence for these assertions remains thin (Fallick et al. 2006). Furthermore, we know little about which groups of knowledge workers are likely to be more affected by non-competes (but see Garmaise 2007 for evidence that executives are among those affected).
This paper explores the impact of non-competes on interorganizational mobility by exploiting Michigan's apparently inadvertent 1985 reversal of its enforcement policy as a natural experiment. In particular, it argues that the constraint of non-competes will fall more heavily on individuals who have firm-specific skills or who specialize in a narrow range of technologies. We find support for these arguments using several decades of patent data and by employing a differences-in-differences method that ameliorates some of the challenges inherent in tracking mobility of individuals. The job mobility of inventors in Michigan fell 8.1% following the policy reversal compared to inventors in other states that continued to proscribe non-competes, and these effects were amplified for those with particular characteristics. Michigan inventors with skills one standard deviation above the mean in their firm-specificity experienced a decrease in their job mobility of 15.4% following the policy reversal compared to similar inventors in other states. Likewise, having skills one standard deviation above the mean in technical specialization decreased mobility by 16.2%. By comparing the change in the mobility of Michigan inventors relative to inventors in states that did not change their non-compete laws, the paper offers a "research tool" that could help to establish deeper causal evidence on spillovers and other implications of mobility.
Non-Competes: History and Prior Research
Non-competes appear to be nearly universal in employment contracts (LaVan 2000, Kaplan and Stromberg 2001, Stuart and Sorenson 2003), yet the components of non-competition law have not changed materially for centuries. The earliest recorded case was settled in England in 1414, only a few decades after the Bubonic plague had decimated the European labor supply and subsequent to the Ordinance of Labourers that essentially outlawed unemployment in post-medieval England. Thus a plaintiff's request to enjoin one of his former clothes dyers from working in the same town for six months was met with disdain from the judge, who threatened the plaintiff with jail time for having sought to restrict a citizen from practicing his trade (Decker 1993). The principle of keeping skilled labor in the public domain was reinforced during the rise of the craft guilds through the 16th century; not until the decline of the guilds and inception of the Industrial Revolution did the court begin to enforce non-competes entered into voluntarily by employees. The courts typically stipulated a "reasonableness test," limiting the geographic scope and duration of the agreement.
Firms use non-competes to protect their interests: to prevent the disclosure of trade secrets, to honor customer confidentiality, and to guard against competitors appropriating the specialized skills and knowledge of its employees (Valiulis 1985). One might argue that trade secrets are already protected by the nondisclosure agreement (NDA) employees are generally required to sign, but violations of an NDA can be difficult to detect or prove (Hyde 2003). Preventing an ex-employee from joining a competitor reduces the likelihood that an employee will violate the corresponding NDA via so-called "inevitable disclosure" of confidential information at a new job (Whaley 1999).
Although the law of trade secrets is fairly similar across U.S. states (Hyde 2003), enforcement of non-competes varies significantly from state to state. For example, California's Business and Professions Code Section 16600 is reminiscent of early English law: "Except as provided in this chapter, every contract by which anyone is restrained from engaging in a lawful profession, trade, or business of any kind is to that extent void." (1) Gilson (1999) traces the lineage of California's statute back to its inception in 1872 as an "historical accident" of rapid law-making while California sought statehood. Section 16600 has been upheld by the courts and was reaffirmed in August 2008 by the California Supreme Court's ruling in Edwards v. Arthur Andersen. (2) Citing the attenuating impact of non-competes on employee mobility, Gilson proposed that this practice is in fact "the causal antecedent" of the high-velocity labor market as well as the unique culture Saxenian (1994) attributes to Silicon Valley. Gilson's hypothesis went untested until Stuart and Sorenson (2003) examined the effect of initial public offerings (IPOs) and acquisitions on founding, rates of biotech firms in regions that enforced non-competes compared with those that did not. That proportionally more biotech firms were founded in states that proscribe enforcement of non-competes is consistent with Gilson's hypothesis. However, as the Stuart and Sorenson analysis measures firm foundings, it does not directly track individual mobility.
An individual-level study of mobility was undertaken in the Fallick et al. (2006) examination of the computer industry in Silicon Valley. Using month-by-month data from the Current Population Survey in the top 20 metropolitan areas, they found an increase in intraregional employee mobility for the California computer industry versus other states. The authors caution, however, against interpreting their results as unequivocal evidence linking non-competes and mobility:
[We] have no direct evidence that the California effect on mobility
is due to the absence of enforceable non-compete agreements. As a
result we cannot assess the role that other factors (such as local
culture) may play in sustaining high rates of employee turnover.
(Fallick et al. 2006, p. 481)
Ideally, the impact of non-competes on mobility would be established through a quasi-experiment that randomly reversed the non-compete enforcement policy in one state and compared changes in intraregional mobility rates between that state and those that did not change their non-compete laws. In the next section, we describe why Michigan may afford such an experiment.
Michigan's Reversal of Non-Compete Enforcement
At the turn of the 20th century, the metropolitan area of Detroit, Michigan, in many ways resembled the Silicon Valley of the last few decades. Growth of the nascent auto industry was explosive, with 500 firms entering before 1915 (Klepper 2002). Ten years prior, the Michigan legislature in 1905 had passed statute 445.761 (bearing resemblance to California's prohibition): "All agreements and contracts by which any person ... agrees not to engage in any avocation or employment ... are hereby declared to be against public policy and illegal and void." This law governed non-compete enforcement until March 27, 1985, when the Michigan Antitrust Reform Act (MARA) repealed section 445 and with it the prohibition on enforcing non-compete agreements.
More than 20 pages of legislative analysis of MARA by both House and Senate subcommittees does not mention non-competes as a motivation for the bill (Bullard 1983a, b; 1985). This may be a consequence of MARA having been modeled on the Uniform State Antitrust Act (Lifland 1984), designed to "make uniform the law with respect to the subject of this act among those states that enact similar provisions." Given that the impetus for the change in law appears to have been general antitrust reform and not specifically altering non-compete enforcement, it appears that the 1905 statute prohibiting non-competes was inadvertently repealed as part of the antitrust reform. If so, then Michigan's change in enforcement would be an exogenous event rather than an example of the legislature simply "catching up" with the courts or responding to lobbying efforts. Even if it were the case that behind-the-scenes lobbying by powerful interests contributed to the legislature's move--and we did not uncover any evidence of this--such a change would still be exogenous to the inventors who are the subjects of this study, assuming that they would have been unaware of such efforts.
Additional evidence for the accidental, exogenous interpretation of Michigan's non-compete reversal is found following the enactment of MARA in March 1985. Multiple law review journals in 1985 (Alterman 1985, Levine 1985, Sikkel and Rabaut 1985) drew attention to the change. Given the rise of commercial advertising by law firms in the 1980s, it is likely that news of the change would have disseminated quickly through law firms, which would have then brought the news to their clients in hopes of generating new contractual work and prosecuting cases (Bagley 2006). Furthermore, less than two years later, the Michigan legislature passed MARA section 4(a), effective retroactive to the enactment of MARA. This bill established the "reasonableness" doctrine in Michigan--limiting the scope and duration of non-competes--that is common to many states that enforce non-competes (Decker 1993). Although we would not expect legislative analysis to report that the purpose of this bill was to provide guidance to the judiciary in the wake of an accidentally repealed statute, both House and Senate legislative analyses do state that a motivation for 4(a) was "to fill the statutory void" (Trim 1987a, b).
Interviews with two Michigan labor lawyers, the authors of a Michigan Bar Journal article on noncompetes that appeared in October 1985, support the interpretation of...
|
