Tag: chapter-19

Author’s Note*

Introduction

There has been increasing interest in regulating firms’ use of non-compete agreements (NCA).[1] NCAs are contractual provisions in employment agreements that prohibit an employee from working for a competing employer for a period of time post-employment. Currently, the legality and enforceability of NCAs are determined by state law, with the majority of states allowing the use and enforcement of reasonable NCAs.[2] However, three states have passed statutes that ban enforcement of NCAs as a matter of public policy.[3] Other states have challenged the use of NCAs or passed statutes or that prohibit the use of NCAs in specific settings, such as when applied to low wage workers.[4] And even in those states that allow the use and enforcement of such terms, the courts and legislatures in those states have limited enforcement of NCAs when their terms are deemed to be unreasonable.[5] For example, state courts decline to enforce NCAs that lack a legitimate business interest or contain prohibitions that are unreasonably broad or lengthy, and state statutes restrict the scope, duration, and circumstances under which an NCA can be included in an employment contract and enforced post-employment.[6]

Much of the recent attention on the effects of NCAs has focused upon their widespread use in contracts with low-wage workers. For example, recent legislative actions have adopted statutory prohibitions on NCA use that targets low-wage workers,[7] and several states have brought actions against franchised fast food chains to prevent the unlawful imposition and use of NCAs on at-will low wage employees.[8] An economic reason for this focus is that many of the traditional efficiency based reasons for the use of non-competes seem less plausible in many low-wage settings. For example, use of NCAs to protect trade secrets or other valuable business information seems implausible in this setting, as low-wage employees are unlikely to possess such information. Moreover, use of NCAs in response to concerns over employee holdup of employer investments in training and other forms of employee development, while plausible in some cases involving low-wage workers (e.g., client centered services such as hairdressers or financial analysts), seem less compelling in other low-wage settings where NCAs are observed.[9] In addition, survey evidence suggests NCAs are common in contracts with low- wage workers, even in states that preclude their use or enforcement, and that such workers are likely to be unaware of the existence of NCAs in their employment contracts.[10] Moreover, there is evidence based on causal designs that worker mobility and wages increased when a state restricted the use of NCAs, suggesting that use of NCAs in the settings studied cause decreases in labor mobility and wages.[11]

While large digital companies are not immune from criticism for using broad NCAs in the context of low wage workers,[12] many of the applications of NCAs in digital markets will involve higher-wage settings where the use of NCAs is generally more defensible when viewed from and economic and public policy standpoint. In higher-wage settings, many of the traditional reasons for the imposition of NCAs based on the legitimate business reasons mentioned above are more plausible. Under these circumstances, the proper analytic focus is whether any net costs borne by workers are offset by the benefits generated to market participants. These benefits include enhancement of firms’ ability to protect their interests in intellectual property and other valuable business information. Consumers can also benefit from lower prices, higher quality, and product improvements generated by a greater rate of innovation.

Moreover, the same national survey by Starr et al. that found a lack of awareness and consideration for low-wage workers found the opposite in high-wage settings.[13] In particular, for surveyed employees who are aware of the NCA prior to accepting the job, Starr et al. report that these workers have 9.7% higher earnings, are 4.3% more likely to have information shared with them, are 5.5% more likely to have received training in the last year, and are 4.5% more likely to report satisfaction with their job when compared to employees that did not have NCAs.[14] If informed high-wage workers benefit from signing NCAs, then allowing enforcement of reasonable NCAs for higher wage workers to protect legitimate business interests is a sensible approach that avoids interfering with efficient uses of NCAs. More generally, most states enforce NCAs in a manner consistent with this analytic approach. In particular, enforcement of NCAs in many states is conditioned, either through statute or via court enforcement, on the existence of adequate disclosure, limits on the duration of any agreement, and a narrow focus tailored to protect legitimate business interests.[15] Such interests include trade secrets, confidential information, or the employer’s goodwill. Some states also protect investments in skills training and client lists.[16]

The view that it is defensible to use NCAs in employment contracts with higher income employees is not universally shared.[17] Indeed, criticism of the use of NCAs by digital firms has been extended to its application to higher-paid workers. Some would use the federal antitrust laws to ban the use of NCAs irrespective of the context in which they are used.[18] However, it is not clear that the solution to the potential harms generated by the use of NCAs is best addressed federally, or through the antitrust laws. In this short chapter, I address three primary points with respect to calls for a broad federal antitrust based prohibition of NCAs. First, evidence of losses in both wages and worker mobility is not sufficient to support a broad condemnation of NCAs under the antitrust laws unless it is accompanied by systematic evidence that NCAs also negatively affect consumer or total welfare. And unless one adopts the view that harm to an input supplier is sufficient to conclude that a firm using an NCA in its employment contract has violated the antitrust laws,[19] NCAs that may simultaneously reduce mobility or wages but decrease prices and increase output, quality, and innovation should not be condemned as per se illegal under the antitrust laws.

Second, those advocating broad bans on NCAs, including preventing their use to protect legitimate business interests, argue that such interests can be adequately protected by an alternative and less restrictive means. But by assuming a theoretical and less restrictive alternative exists, such an approach fails to confront the differing costs and performance of alternative approaches, and ignores the operation of actual markets. Such an approach is contrary to the approach taken in most states that recognize that use of NCAs are necessary to protect legitimate business interests that cannot be adequately protected through an alternative restrictive covenant, including but not limited to a non-solicitation agreement or a non-disclosure or confidentiality agreement.[20] An approach that ignores the differing costs and performance of alternative approaches is unlikely to produce a rational approach to the complex tradeoffs presented by the use of NCAs in digital industries.

This last observation leads to the third point. Proponents of widespread bans have focused on a uniform federal prohibition of NCAs. But such an approach would likely interfere with the balanced and reasonable approaches to the use and enforceability of NCAs taken by the states, reflected in recent statutes and in court decisions. In addition to limiting enforcement in specific settings, state approaches to NCAs also include duration limits, requirements for consideration, garden leave that requires firms asserting enforcement of NCAs to compensate workers post-enforcement, and consumer protection remedies aimed at reducing informational asymmetries can serve to mitigate the wage effects of NCAs. In addition, in light of the critical need for further evidence on the effects of legal restrictions on NCAs, a uniform federal approach may be particularly costly at this time, as it will destroy the best chance to use the laboratory of the states to generate the economic evidence that would best guide regulation in a rational way going forward.

Finally, the chapter briefly examines the relationship between NCAs and no-poach agreements between horizontal firms. NCAs and no-poach agreement are substitutes, but the use of horizontal agreements between firms does not rely on the assent of the worker and thus is unlikely to result in the compensating worker benefits generated when informed workers agree to NCAs. Moreover, these agreements can facilitate the exercise of monopsony power, resulting in reduced output and higher prices in the output market. Thus, the use of no-poach agreements between firms is not consistent with the measured approach of the states to the enforcement of NCAs. The approach taken by the federal antitrust agencies to treat naked no-poach agreements between firms competing to hire employees as a per se and even criminal offense going forward should give firms strong incentives to seek alternative ways, including through the use of NCAs, to advance legitimate business interests.

I. Economic Evidence on the Effect of NCAs as a Basis for an Antitrust Rule

In addition to the legislative and public policy interest in NCAs, there has been a corresponding academic interest in studying the causes and effects of NCA use. In addition to the Starr et al. national survey, there has been a rising interest in examining the effects of NCAs on worker mobility and wages. As noted above, an event study using Oregon’s choice in 2008 to ban the use of NCAs for low-wage workers finds the law caused increases in hourly wages and within-industry mobility relative to control states that did not change their laws over this time period. Moreover, other studies find that there may be spillover effects to the market for labor generally, with broad NCA use affecting the equilibrium wages of workers that are not bound by an NCA.

Because there have been several excellent and recent analytical surveys of the growing literature on the economic effects of NCAs, this section assumes, arguendo, that widespread use of NCAs cause lower wages and worker mobility.[21] The question addressed in this section is whether or not this result, if shown to be a causal effect of widespread NCA use, is sufficient to support a rule banning the use of NCAs.[22]

To help frame the discussion in this section as well as the next, consider the recent paper by Gurun et al. that analyzes the effects of NCAs used by financial advisory firms to constrain the mobility of their employees.[23] In particular, their research design analyzes the voluntary adoption in 2004 of a Protocol for Broker Recruiting by three major firms.[24] This protocol established a set of rules for member firms governing employee departures to other financial advisory firms. These rules eliminated enforcement of important aspects of employee NCAs for firms in the protocol. In particular, it allowed an adviser to take client lists and contact information to their new employer without having to fear litigation over breaching post-employment restrictions contained in the employment contract with their former employer. This in turn allowed clients of a member firm to follow their advisers to other firms that were also members of the protocol. Protocol membership was not limited to the original three firms, and over time more than 2,000 firms have joined the protocol in a staggered fashion.

Gurun et al. use the staggered entry and exit of firms from the protocol to identify the effect of the protocol on the performance of advisers working in protocol firms relative to the performance of advisers working for firms that were not members of the protocol. Gurun et al. find that working for a protocol firm resulted in increases in adviser welfare—the study finds increases in both adviser compensation levels and adviser mobility. However, advisers working for protocol firms were less likely to be disciplined, and the clients of these firms experienced higher levels of adviser misconduct, and also paid increased fees.

To the extent that the voluntary adoption of the protocol mimics the effects of an antitrust rule that would ban the use and/or enforcement of NCAs, the results from Gurun et al. suggest that the effects of a ban of NCAs in this setting would be to help the adviser employees by increasing mobility and wages. On the other hand, client investors would face higher fees and receive lower quality advice resulting from higher rates of adviser misconduct.

Going back to the question of whether the observation of reductions in wages and employee mobility is sufficient to conclude that NCAs are anticompetitive, these results demonstrate that a change in welfare in an input market does not directly map onto a similar change in consumer welfare in the output market, and may be negatively correlated with both consumer and total welfare. Indeed, such a negative relationship will be common when NCAs are used by firms in a procompetitive way to lower costs and increase quality by reducing agency costs. The point is that the procompetitive use of NCAs can result in less mobility and lower wages relative to a setting in which use and/or enforcement of NCAs are prohibited. As a result, the observation that NCAs reduce worker mobility and wages by itself is unlikely to produce a reliable screen between welfare increasing uses of NCAs and welfare decreasing uses of NCAs that create monopsony power.[25] And without evidence on the broader effects of NCAs on product quality, innovation, and downstream consumers, such an observation does not provide a rational or sufficiently informed basis to declare use of NCAs to be an unfair method of competition.[26] Such an approach could not be easily reconciled with the FTC’s commitment to use its Section 5 UMC authority to target practices that harm competition and consumer welfare, to use the rule of reason when both pro and anticompetitive outcomes are possible, and not to pursue broader public policy goals.[27]

II. The Availability of Less Restrictive Means

As set out in the prior section, proposals to unconditionally ban NCAs would include settings where legitimate business interests are involved. However, those advocating broad bans on NCAs, including preventing their use to protect legitimate business interests, argue that broad bans would not impose high costs in such situations because such interests can be adequately protected by an alternative and less restrictive means.[28]

But approaches that assume a theoretical and less restrictive alternative exists fail to confront the differing costs and performance of alternative approaches, and ignore the operation of actual markets.[29] Take for example the argument that in the absence of the ability to use and enforce NCAs, trade secret law and non-disclosure requirements could be used to prevent former employees from using a firm’s valuable business information to compete with them. This argument ignores the fact that use of these instruments is often ineffective and costly to enforce. In particular, enforcement of trade secret law, including through non-disclosure agreements, “has important gaps” and is “costly and uncertain to apply.”[30] Thus, employers often must supplement ineffective and costly legal protection from trade secret law with more effective direct restrictions on employees, including active monitoring of current and former employees use of firm information, and the use of post-employment contractual provisions such as NCAs. These contracts protect against the improper disclosure of valuable firm information by reducing the ability of a departing employee to benefit from using this information where it would be the most valuable. Indeed, under many state laws, use and enforcement of NCAs is supported by the fact that a legitimate business interest cannot be adequately protected through an alternative restrictive covenant, including but not limited to a non-solicitation agreement or a non-disclosure or confidentiality agreement.[31] Thus, rather than rely on costly and ineffective legal enforcement and monitoring, use of NCAs provide a substitute mechanism based on changing employees’ incentives to misappropriate their former employer’s information.[32]

More generally, how a firm addresses the problem of simultaneously sharing valuable and confidential information with its employees and preventing the loss of this information when employees depart the firm is a complex and individualized problem that depends upon many factors unique to both the firm and its employees. These include information and transactions costs, including the costs of monitoring post-employment use, the costs of litigation involved in enforcing post-employment restrictions, the relative effectiveness of a particular restriction, and the compensating differential required to compensate employees for the loss of mobility imposed by post-employment restrictions. These transactions, information, and litigation costs are important determinants of the nature of employment contracts, and how firms are organized generally.[33] Indeed, the Broker Protocol discussed in the prior section was initially designed to address the frequent and high costs of litigation related to the enforcement of post-employment contractual restrictions that prevented departing advisers from taking client information to their new firms.[34] While litigation is costly, the subsequent departure of large firms from the protocol suggests that, for these firms, the costs of not enforcing these contractual restrictions were even higher.[35] Moreover, technological changes in the way advisers and clients communicate with each other has made it more difficult to enforce post-employment restrictions on using the firm’s client information.[36] This suggests that any approach that fails to incorporate the complex tradeoffs presented by the use of NCAs in digital industries is unlikely to produce a rational approach to the regulation of NCAs.

III. The Benefits of State Law

Proponents of widespread bans have focused on a uniform federal approach to prohibiting the use and enforcement of NCAs. But such an approach would likely interfere with the balanced and reasonable approaches to the use and enforceability of NCAs currently taken by the states, reflected in recent statutes and in court decisions.[37] As noted above, state laws currently limit enforcement of NCAs to those that contain reasonable terms. In addition to limiting enforcement in specific settings, state approaches to NCAs also include duration limits, requirements for consideration, garden leave that requires firms asserting enforcement of NCAs to compensate workers post-enforcement, and consumer protection remedies aimed at reducing informational asymmetries can serve to mitigate the wage effects of NCAs.[38]

This section makes two additional points regarding the state law approach. First, as noted above, the state law approach is not based on a mismatched and novel use of competition law. Rather it is a contract/consumer protection approach that focuses on ensuring that employees are informed about the inclusion of NCAs in their employment contract and that any restrictions are reasonable and narrowly tailored to further a legitimate business interest. A focus on adequate disclosure makes it more likely that employees agreeing to NCAs will be compensated for any loss in mobility. As noted above, the Starr et al. survey found that employees that are informed about the existence of NCAs prior to accepting the job reported that they experienced better outcomes than employees not subject to NCAs.[39] In addition, while Starr finds that increased non-compete enforceability is associated with higher levels of training but lower hourly wages, this latter result is driven by employees in states that do not have consideration laws.[40] These laws condition NCA enforceability on the employee receiving something of value in exchange for agreeing to the NCA, and are associated with employee wage gains.[41] In contrast to an approach that would prohibit use and enforcement of NCAs by making NCAs an unfair method of competition, the contract/consumer protection approaches directly attack the underlying informational imbalances that seem to be a primary cause of poor outcomes associated with NCA use and make it more likely that the agreement to include an NCA in an employment contract benefits both the firm and employee.[42]

In addition, in light of the critical need for further evidence on the effects of legal restrictions on NCAs, a uniform federal approach may be particularly costly at this time, as it will destroy the best chance to use states’ varying approaches to enforcement of NCAs as social “laboratories of democracy,” where laws and policies are created and tested at the state level of the democratic system, in a manner similar (in theory, at least) to the scientific method.^[43] Indeed, many of the recent empirical studies of the effect of laws regulating NCAs on employee outcomes have identification strategies that are based on variations in state policies.[44] Similarly, California’s longstanding approach in banning the use of NCAs has been hypothesized to be responsible for the relative success of Silicon Valley over the Route 128 tech corridor in Massachusetts, which allowed enforcement of NCAs during that time period.[45]

Despite the growing and robust empirical literature on the effect of NCAs, there are many outstanding questions that remain unresolved. As noted in the previous section, many of the studies examine the effect of state NCA laws on worker outcomes, but generally do not measure outcomes for firms and consumers. While there have been some initial attempts at measuring the effect of NCAs on growth and innovation, it is probably too early to make any strong inferences about systematic effects of NCAs on dynamic competition.[46] The use of the laboratory of the states to generate the economic evidence that could be used to study the causal effect of NDAs holds promise as the best way to generate useful information that would best guide regulation in a rational way going forward. Policy makers should be wary of imposing a uniform and uninformed federal regulatory policy that would prevent the generation of this evidence.

Finally, to the extent that the law of a state attempts to prevent the use of NCAs, several viable alternatives exist, especially for large digital firms that wish to protect valuable business interests through NCAs. Firms can choose a favorable choice of law and forum through incorporation in a state that enforces reasonable NCAs. For example, franchise firms and other firms incorporated in Delaware have used choice of law clauses to choose Delaware law, which favors enforcement of NCAs.[47] To the extent that such an option is foreclosed by the banning states,[48] such firms can also use their mobility to lessen the impact of laws that inefficiently restrict the use of welfare increasing reasonable NCAs.[49] Firms can, under these circumstances, move their principal place of business and many of their operations from regulating states to states with more favorable treatment of NCAs.[50]

IV. NCAs versus No-Poach Agreements Between Firms

Finally, this chapter briefly examines the antitrust treatment of no-poach agreements between horizontal firms. In a broad sense, NCAs and no-poach agreements are substitutes, as both can be used by firms to restrict the mobility of workers.[51] A primary difference between no-poach agreements and NCAs, however, is that the use of horizontal no-poach agreements between firms does not rely on the agreement or assent of the worker and are less likely to result in a compensating differential. Such agreements between firms also strip employees of any of the statutory or common law protections under state law that would apply when NCAs were used. Thus, the use of no-poach agreements between firms to restrict employee mobility is not consistent with the measured and balanced approach used by the states to enforce NCAs.

No-poach agreements have been the focus of increased scrutiny under the antitrust laws.[52] Under the Department of Justice Antitrust Division and Federal Trade Commission Antitrust Guidance for Human Resource Professionals, “naked wage fixing or no-poaching agreements among employers . . . are per se illegal under the antitrust laws.”[53] Moreover, the DOJ announced that going forward, it intends to proceed criminally against naked wage-fixing or no-poaching agreements.

In addition to addressing naked horizontal no-poaching agreements, the DOJ has recently filed statements of interest in three private no-poach cases involving no-poach agreements included in franchisor-franchisee agreements.[54] A primary legal issue in these cases is whether a franchisor and its franchisees are deemed to be a single entity under the Court’s Copperweld and American Needle precedents.[55] The DOJ argued that franchisors and franchisees were not necessarily entitled to single entity status under Copperweld. Whether the franchisor is capable of concerted action with the franchisees depends on whether it has “distinct” “economic interests” from the franchisees.[56] Under the DOJ’s stated approach, evaluation of vertical no-poach agreements between franchisors and franchisees found to be capable of concerted action would be treated similar to other vertical contracts and proceed under the rule of reason.

From the standpoint of antitrust law, the purpose of Copperweld immunity was to identify business arrangements that were unlikely to present threats to competition and reduce consumer welfare. Antitrust claims against such business arrangements could then be dismissed at an early stage as implausible, saving party and court resources and reducing the costs of type I errors.[57] Economic analyses criticize the Copperweld approach and its focus on the existence of a unity of interest between entities as misguided and inconsistent with the modern economic theory of the firm.[58] An approach consistent with the modern economic theory of the firm would focus instead on the allocation of control rights between entities, including franchisors and franchisees, and an evaluation of the economic function of the agreement or contract challenged as anticompetitive.[59]

Such an analysis, however, is likely to be a costly and complex fact-based inquiry, which would diminish the value of using a such an inquiry as part of a Copperweld type early stage screen.[60] Thus, the DOJ approach is consistent with an economic approach that: (i) recognizes the limited utility of a Copperweld type approach; (ii) relies on the Court’s approach in Bell Atlantic v. Twombly to engage in early stage screening;[61] and (iii) uses the rule of reason to evaluate cases that survive a motion to dismiss.[62]

Conclusion

The use of NCAs is currently regulated under state law. These laws embody diverse approaches to the regulation of NCAs which generally limit enforcement of NCAs to those with reasonable and narrowly tailored terms. State law also incorporates various approaches to mitigating the negative effects NCAs might have on workers, including requiring adequate disclosure, consideration, and banning their use in certain employment contracts. These diverse approaches can leave regulatory gaps in certain states and can result in overregulation in others. But this can also be a feature—diverse approaches to legal regulation of NCAs at the state level takes advantage of the laboratory of the states to generate information that will allow economic research to better evaluate the effects of these diverse legal approaches without exposing the entire nation to potential mistakes.

There has also been a growing and robust literature that has examined state variation in the regulation of NCAs and the effect of these laws on worker welfare. However, current evidence does not yet suggest a reliable and predictable link between the use of NCAs and the effect on employee welfare. Moreover, changes in employee welfare can be negatively correlated to changes in consumer welfare. Any antitrust rule that broadly prohibits the use and enforcement of NCAs would be based on incomplete evidence, would interfere with the current operation of the balanced and nuanced state by state approach to regulating NCAs, and would potentially be at odds with the antitrust laws’ focus on consumer welfare.