American antitrust regulators have recently taken aim at noncompete clauses. They argue that noncompetes suppress labor bargaining power and thus wages. The Italian labor market differs from its American counterpart in its rigid protections for labor, but the use of noncompetes in Italy occur at about the same rate as in the United States and shows a correlation with lower wages for workers whose noncompete clauses are unjustified because their jobs require little training and do not grant access to trade secrets. The evidence from Italy suggests that better regulation of noncompetes and informing workers of their rights is justified on the whole.
The “antitrust and labor” debate recently peaked in the United States when the U.S. Federal Trade Commission proposed a new rule in January that would ban employers from imposing noncompete clauses on their workers. A noncompete is an agreement or clause included in an employment contract by which an employee agrees to refrain from competing with her employer (such as working for a business competitor or starting a competing business) for a given period after the end of the employment relationship. In most countries, noncompetes are lawful (under certain conditions) and justified by the need to protect legitimate business interests.
However, noncompetes can also reduce competition in the product market by restraining the ability of competitors to hire workers or deterring departing employees from creating a new competing company. They also reduce competition in the labor market by, for example, limiting workers’ outside options and, consequently, their bargaining power. Evidence available in the U.S. shows that noncompetes are widespread and cover low-wage and low-skilled workers, which suggests that employers use them beyond the need to protect trade secrets and specific investment in the employment relationship, such as training, as low-skilled workers generally do not have access to trade secrets and often require little training.
The arguments for and against noncompetes in rigid European labor markets
There is no comprehensive evidence on the diffusion of noncompetes and their impact on workers outside the U.S. In Europe, even among law professionals, many consider the phenomenon an essentially American one, typical of a fluid labor market characterized by employment-at-will and high workers’ mobility. Several European countries are characterized by strict employment protection legislation and a strong reach of labor unions. Strict regulation reduces workers’ mobility across firms by affecting not only layoffs but also hiring rates. Moreover, workers with a permanent contract have to provide advanced notice within a certain period when they intend to quit, which is usually regulated by the law or by industry-wide collective agreements. Meanwhile, employees with a temporary contract cannot quit (and, conversely, cannot be fired) before the foreseen end date of their contract (if they do so, they may be subject to the payment of a hefty penalty. Such strict regulation of hiring, firing, and quitting reduces the need of noncompetes, as employers run a lower risk of losing employees to competitors in a rigid labor market.
At the same time, under strict employment protection, firms changing production methods and lines are often compelled to retrain the existing workforce rather than upgrade workers’ skills by adjusting along the extensive margin (i.e. hiring other workers with the newly required skill set). This retraining involves the typical hold-up problems of investment under uncertainty: an employer may refrain from offering training to workers because of concerns that human capital investment may give them increased bargaining power, and thus an incentive to leave, due to the better outside opportunities they may have after the training. Under these circumstances, noncompetes may contribute to reducing the holdup problem as they are a powerful commitment device. The employer deciding upon investing in training is reassured that the worker will not leave, or at least will not leave immediately, after the costly investment in their human capital is undertaken.
In addition, collective bargaining agreements negotiated by labor unions could also affect the spread of noncompetes as industry-level negotiations may in principle internalize the costs of noncompetes in terms of misallocation of workers. Moreover, collective bargaining is generally associated with stronger worker bargaining power, reducing the possibility that employers impose noncompetes and push wages below the value of marginal productivity. Conversely, collective bargaining often induces coordination among employers sitting at the same bargaining table, thus potentially enhancing mutual support in imposing noncompetes.
While it is perhaps not surprising that companies make (good and bad) use of noncompetes in a fluid labor market like that of the U.S., in a rigid labor market there are arguments for both a greater and a lower scope for noncompetes.
Noncompetes in the Italian labor market
To bring another perspective to the growing literature, we conducted a survey—similar to the one J.J. Prescott, Norman D. Bishara, and Evan Starr ran in the U.S.—in a country that is practically at the opposite side of the labor regulation spectrum: Italy. The Italian labor market is characterized by one of strictest employment protection legislation in OECD countries, with hiring, separation, and quit rates about half of those in the U.S. In Italy, collective agreements are negotiated by labor unions which cover (on paper, at least) all employees. Noncompetes themselves are also more regulated in Italy than in most U.S. states. In Italy, a noncompete must financially compensate an employee for their restricted future economic opportunities (such as with a fixed sum payment or as additional compensation during employment), as well as time, sectoral, and geographical limits (otherwise they are simply unenforceable). Judges can always assess and invalidate a noncompete agreement if they find it excessively and unjustifiably restricts a worker’s future employment opportunities.
Yet, the results of our survey show a stunning similarity with the U.S. Figure 1 shows that about 16% of Italian private sector employees are bound by a noncompete, compared to 18% in the U.S. Noncompetes are not the only legal tool to regulate post-employment activity in the U.S. Out of the total population of private sector employees in Italy, 55% are not covered by any clause, 23% are covered by one clause (typically a non-disclosure agreement), and 22% are covered by more than one clause (typically a nondisclosure agreement and another clause, such as a noncompete).
Figure 1: Share of Italian employees bound by clauses regulating post-employment activity
As in the U.S., noncompetes in Italy are not restricted to high-skilled professionals, managers, or workers with access to confidential information. They are relatively frequent among low-educated and low-earning workers employed in manual and unskilled occupations and without access to any type of confidential information.
When we look at the enforceability of noncompetes, most of them appear not to comply with the minimum requirements set by Italian law (i.e., specifying a compensation as well as time, sectoral and geographical limits). This suggests that most noncompetes are unenforceable or that workers, even those who are sure to have signed one and declare to have read it carefully, have incomplete information about their content. However, Figure 2 shows that workers’ beliefs about the likelihood of enforcement by the employer and adverse litigation outcomes do not vary depending on the enforceability of the noncompete. It is the same with respect to the intention to quit. In fact, as already shown by previous research, what matters for workers’ behavior is their beliefs about the likelihood of a resulting trial and court enforcement, not the actual enforceability. The in terrorem effect of an unenforceable noncompete is enough to discourage voluntary quits and curtail workers’ bargaining power.
Figure 2: In Italy, likely unenforceable clauses still appear able to dissuade workers to change job
Panel A: Perceived likelihood of being brought to court and being found liable, by enforceability of the noncompete
Panel B: Share of employees bound by a noncompetes dissuaded from quitting because of the noncompete
Although our paper does not provide any causal evidence, we do show that the correlation with Italian labor market outcomes markedly differs across noncompetes. Consistent with the research in the U.S., we find that unenforceable clauses, i.e., clauses that are more likely to be used to deter workers from moving and not to protect business interests, go together with lower wages compared to enforceable ones, and are not associated with higher or lower training opportunities, information sharing, or job satisfaction. Likewise, noncompetes signed during the employment relationship without any change in job tasks are associated with lower wages and lower training opportunities (in particular, training paid by firms) than clauses signed at the onset of the employment relationship. In contrast, noncompetes signed during the employment relationship and associated with a change in job tasks go together with higher wages. Finally, noncompetes without access to any type of confidential information are consistently associated with lower wages, less information sharing and lower satisfaction with pay than clauses associated with access to confidential information.
All in all, the evidence emerging from our study suggests that 1) noncompetes are not a feature specific to a fluid labor market and 2) even when they are regulated, a mix of abuse (by firms) and lack of (workers’) awareness may lead to a distortion of their original purpose with negative effects in terms of equity and efficiency.
These novel findings lend some support to the relatively radical FTC’s proposal to ban noncompetes in the U.S. Even where labor regulation is much more protective of the worker, where the law already foresees some limitations to the use of noncompete, and where labor unions exert a much tighter control (again, at least on paper), noncompetes go well beyond what we should expect if they were used to protect trade secrets or specific investments in the employment relationship by the employer. However, although providing a simpler and clearer rule governing noncompetes can be helpful, an outright ban does not tackle the pervasiveness of unenforceable noncompetes and the general lack of knowledge and awareness. Our study confirms that, regardless of the content, regulation is ineffective when information does not reach workers (their behavior and beliefs, consequently), enforcement levels are low, and incentives to comply with it are inadequate.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
