On May 7, 2024, the Federal Trade Commission published its final rule related to non-compete agreements, which essentially bans such agreements in all new employment contracts (“Final Rule”). The issuance of the 570 page rulemaking for the Final Rule followed a comment period in which the FTC received over 26,000 public comments on its Notice of Proposed Rulemaking (NPRM) issued in January 2023. This article reviews the FTC’s rulemaking related to economic research and other evidence the FTC considered in evaluating the impact of non-competes, and how the FTC took this evidence into account in finalizing its rule.
As the rule undergoes challenges about whether the FTC has statutory authority to promulgate such a rule, and on other grounds, at least one court found the FTC did not do an adequate job supporting the rule’s across-the-board ban on virtually all non-compete agreements. This article describes and evaluates the FTC’s support for its rule, and in particular evaluates the strength of the evidence the FTC relied upon to justify prohibiting non-competes not just for low-skilled workers, but for high-skilled workers as well.
The FTC’s Final Rule has been challenged in multiple jurisdictions since its issuance. These challenges include one in the Northern District of Texas, where the judge issued a nationwide injunction blocking enforcement of the Final Rule, and the FTC has indicated it will consider appeal. In that case, among other factors the judge faulted the FTC for not adequately considering alternatives to a broad “one-size-fits-all approach.” The Final Rule is also being challenged in other lawsuits, such as in the Eastern District of Pennsylvania, where the court declined to issue a preliminary injunction barring enforcement of the rule, finding that the plaintiff had failed to show it was reasonably likely to prevail on the merits in its argument that the FTC lacked authority to issue the Final Rule. In the Middle District of Florida, the rule was blocked for enforcement against only the complainant in the case. The FTC has since filed notice of appeal. As a result, given the differing court decisions and appeals, it appears that it will be some time before a final decision on the FTC’s ability to enforce the Final Rule will be determined. As such, a review of the process that the FTC followed in its rulemaking and the evidence the FTC relied on can be useful in evaluating the Final Rule’s economic merits and whether the rulemaking reliably supported the Final Rule.
In June 2023, we performed a literature review of the economic evidence the FTC provided to support its NPRM, analyzing the degree to which the literature supported the FTC’s rulemaking. We found that the NPRM showed a serious effort to incorporate a review of economic literature. The FTC’s Final Rule, while largely preserving the provisions of the NPRM, contains a limited number of substantive changes, as well as the citation of additional economic literature on the effects of non-compete agreements. In this article, we examine the changes to and analyze the additional support the FTC provided for the Final Rule.
The FTC’s Final Rule relies on many of the same articles that it cited in the NPRM, providing more in-depth evaluation of the articles that it cited earlier. The Final Rule also discusses additional recent research. It also provides an explanation for how heavily it weighs different types of studies, though the FTC’s evaluation of empirical studies might be considered “holistic”. Specifically, the FTC states it gives more weight to studies:
The FTC’s literature review for the Final Rule was expanded to include articles released after the NPRM was issued. For example, Johnson, Lipsitz, and Pei (2023) study the impact of non-competes on firm investment, and Jeffers (2024) studies the impact of non-competes on departure rates. The FTC cites these studies in claiming that non-compete enforceability is associated with reduced high-tech firm entry and reduced worker departure rates, respectively. Given the FTC’s detailed discussion and apparent reliance on at least some of the research discussed for the first time in the NPRM, the FTC somewhat surprisingly states that they use the studies beyond those cited in the NPRM “only where they check or confirm analyses discussed in the NPRM, or where the Commission is responding to comments raising them. The Commission’s findings do not rest on these studies, however, and they are not necessary to support its findings.” Below, we review in more detail the literature, comments, and reasoning described for the substantive changes to the Final Rule.
The substantive changes in the Final Rule, compared to the NPRM, include:
The FTC largely cites public comments rather than new economic literature in justifying its changes. In addition, the FTC updated its estimate of the financial benefits of the rule. Each of these changes are discussed in more detail below.
Some commenters, such as the American Bar Association’s Antitrust Law Section, had suggested eliminating non-competes for only low wage workers. Our previous article showed that the FTC’s cited literature suggested that non-competes can result in higher wages for high-wage workers, though such effects were empirically difficult to isolate from the effects of other arrangements like non-solicitation agreements. However, the FTC Rule still eliminated employers’ ability to create new non-competes going forward, regardless of wage level.
The FTC appears to acknowledge the differences between senior executives and other types of workers, stating that “[t]he final rule allows existing non-competes with senior executives to remain in force because this subset of workers is less likely to be subject to the kind of acute, ongoing harms currently being suffered by other workers subject to existing non-competes.” Instead, the FTC finds that the harm caused by non-competes for senior executives largely involves reduced competition in product and services markets, as well as in labor markets. This reduced competition seen by the FTC is based on the ability of senior executives, absent being subject to non-competes, to potentially start or spinoff companies, increasing the number of competitors in the market, fostering innovation, and employing more workers in doing so. Thus, while the FTC does not find the same harms for senior executives as other workers, the Final Rule imposes the same categorical ban eliminating the ability to create any new non-competes with senior executives going forward.
Interestingly, the FTC finds these potential harms from non-competes powerful enough to eliminate them for senior executives in the future, but apparently not powerful enough that existing agreements should also be rendered unenforceable. The FTC states:
[b]ecause the harm of these non-competes is principally that they tend to negatively affect competitive conditions (rather than exploiting or coercing the executives themselves), and due to practical concerns with extinguishing existing non-competes for such executives, the final rule prohibits employers only from entering into or enforcing new non-competes with senior executives.
The FTC’s reasoning for allowing existing non-competes to remain for senior executives revolves around the complex negotiations and logistical issues that would result from their recission. This logic is primarily drawn from commenters’ concern that the ban on existing non-competes would erode existing financial agreements between employers and senior executives, leading to a loss in investment for the firm, and/or a loss in agreed-upon benefits (e.g. bonuses, severance pay) for the worker. It is not clear why the Final Rule’s ban on future non-competes would not similarly deter investments.
The FTC’s reasoning behind allowing notice of non-competes’ unenforceability, rather than a requirement to formally rescind them, appears to be largely practical. The FTC claims that the notice requirement will help workers be “meaningfully informed” of the change, assist with compliance, and eliminate an unnecessary administrative burden on employers. The FTC cites literature showing that workers with even unenforceable non-competes may believe they are bound by them, suggesting that notices rather than recission of non-competes may be effective because workers should be made aware that their non-competes are no longer being enforced. This is the only economic literature cited by the FTC related to the issue of notice rather than recission. The FTC relies largely on comments that raise concerns about the potential challenges and implications associated with putting the recission requirement in place. These include compliance costs related to rescinding and revising employment contracts, as well as potential legal obstacles related to contract severability.
The FTC notes how commenters argued that a 25% equity stake threshold is not appropriate because owners with less than 25% equity could still have significant goodwill in the business that is not reflected in their ownership percent. The FTC also added that non-competes are only valid for a bona-fide sale to “address commenters’ concerns that employers will use sham transactions, stock-transfer schemes or other mechanisms designed to evade the rule.”
The FTC’s adjustment to remove the 25% equity stake threshold and instead allow non-competes for any bona-fide business interest sale was not based on any particular literature, but rather consideration of the received comments. In fact, the FTC declines to eliminate non-competes for such circumstances due to a lack of relevant literature support. The FTC states:
[T]he Commission acknowledges that some non-competes between the seller and buyer of a business may be exploitative and coercive due to an imbalance in bargaining power and/or may tend to harm competitive conditions. However, commenters did not present empirical research on the prevalence of non-competes between the seller and the buyer of a business or on the aggregate economic effects of applying additional legal restrictions to non-competes between the seller and buyer of a business. The Commission’s decision to adopt § 910.3(a) reflects the view of the Commission and most commenters that, compared to non-competes arising solely out of an employment relationship, non-competes between the sellers and buyers of businesses may implicate unique interests and have unique effects that this rulemaking record does not address.
For example, the FTC cited comments that the presence of non-compete agreements within a particular industry may imply that buyers would be less willing to pay for a business. This is because they know they are not protected from future competition by the seller, which could lead to a slowdown in M&A transactions and stifled innovation.
Interestingly, while the FTC does make this exception to its Final Rule for business entity sales, it devotes significant discussion to the role that non-competes play in stifling spinoffs and business formation. For example, the FTC says:
Non-competes, however, tend to negatively affect competitive conditions in product and service markets by inhibiting new business formation in two ways. First, since many new businesses are formed by workers who leave their jobs to start firms in the same industry, non-competes reduce the number of new businesses that are formed in the first place. Second, non-competes deter potential entrepreneurs from starting or spinning off new businesses—and firms from expanding their businesses—by locking up talented workers. Non-competes thus create substantial barriers to potential new entrants into markets and also stymie competitors’ ability to grow by making it difficult for those entrants to find skilled workers. [footnotes omitted]
The FTC states that it is not finding that non-competes in business entity sales are procompetitive, rather that because it has “seen fewer changes [related to non-competes and business entity sales] recently than the laws applicable to non-competes that arise solely out of employment, there have not been natural experiments allowing researchers to assess this type of non-compete’s effect on competition.” It seems that the FTC distinguishes application of the literature between the situations where part of an existing business is spun off into a new entity (which creates a new competitor) versus when part of an existing entity is sold (which does not). The FTC makes this distinction despite the fact that permitting non-competes for sellers of business entities could prevent them from then starting new competitors of their own.
As described in our previous article, the NPRM quantifies the benefits of eliminating non-competes by, in part, multiplying a 3.3% wage effect from one study by the entire annual wage earnings in the US from private employers, resulting in an annual increase in wages of $250-296 billion. Our previous research found the approach to be an oversimplification of the overall effects of eliminating non-competes since it did not account for how the effects of the rule would vary for different types of workers. It also did not account for the fact that a portion of calculated benefits would likely represent transfers rather than net benefits.
In the Final Rule, the FTC applies a similar approach to compute the earnings effect, although the figures they use to compute the “% increase in earnings caused by the change in enforceability of non-competes” vary. In the NPRM, the FTC relies on a result from a study conducted by Johnson, Lavetti, and Lipsitz (2020) that found earnings would increase, on average, by 3.3-13.9% in the presence of the ban, which was updated to a range of 3.2-14.2% in a December 2023 version of the paper. In the Final Rule, the FTC computes the percent increase in worker earnings by combining results from Johnson et al. (December 2023) and Johnson et al. (July 2023), both of which were published after the NPRM was released. Specifically, the FTC utilizes Johnson et al. (July 2023)’s results measuring the annual change in non-compete enforceability, and incorporates Johnson et al. (December 2023)’s measure of the average effect of non-compete enforceability on worker earnings. Combining the results from these studies, the FTC arrives at a wage effect of 0.86% resulting from the Final Rule, which corresponds with a greatly reduced annual earnings effect of $53 billion.
While the smaller earnings effect calculated by the FTC is based on findings in the literature, the effect of a complete nationwide elimination of non-competes might be larger than those suggested in the literature. The FTC includes a full and partial sensitivity analysis in the Final Rule to account for the more pronounced earnings effect that the FTC expects to be realized from the ban—as opposed to the smaller changes in non-compete enforceability cited in the literature that is used in the primary analysis. The purpose of the sensitivity analyses is an attempt to estimate the full economic impact of the Final Rule by extrapolating the impact of earnings on (1) a full drop in enforceability of non-competes in all states and (2) a partial drop in enforceability of non-competes in all states to the minimum level of existing enforceability—which corresponds to the level of 53% enforceability in New York. Under the assumption of a full drop in enforceability, the FTC finds an effect on workers’ earnings equal to 3.2%, which results in an annual earnings effect of $234 billion. Under a reduced “enforceability score,” the FTC finds an annual earnings effect of $161 billion.
The updated analysis of the Final Rule also attempts to take into account any lost benefits from the non-competes, such as the potential for reduced training and investment found in some of the research. The FTC finds significant uncertainty in whether the Final Rule will result in an economic benefit or cost of the Final Rule related to investments in human capital, providing a range of a benefit of $32 billion to a cost of $41 billion over ten years. The FTC cites two sources of uncertainty in its estimates: (1) without non-competes, firms may be able to more readily obtain new workers with core skills (resulting in the Final Rule providing benefits from lower search costs), but may instead lead to less advanced training of skilled workers (representing offsetting costs of the Final Rule due to lost productivity), and (2) a lack of estimates in the literature specifically analyzing the training effect of a ban across the entire workforce.
The upper bound of the training-related benefit from the Final Rule of $32 billion over ten years assumes that none of the training foregone due to the Final Rule is “advanced” training of skilled workers, and is instead training foregone because firms would need to provide less unnecessary “core” training to workers. However, at least some of the foregone training is likely to be advanced training. For example, the FTC cites a study showing that “knowledge-intensive firms invest substantially less in capital equipment following decreases in the enforceability of non-competes…” Such knowledge-intensive firms are likely to have highly trained workers. Indeed, the FTC acknowledges “[t]here is some empirical evidence that non-competes increase investment in human capital of workers, capital investment, and R&D investment.” The estimated cost of the Final Rule of $41 billion assumes that all of the training foregone due to the Final Rule is advanced training that increases productivity. The FTC’s estimates of a potential loss in output due to reduced investment in training are only between $0.7 and $1.5 billion lost in the first year of the Final Rule. To a large degree these costs are driven by one economic study that estimates the decrease in trained workers due to reduced training would only be 0.4 percent.
Section IX of the FTC’s Rulemaking discusses a number of alternative rules, and why the FTC did not adopt them. For example, Section IX.A.2 states the FTC considered “. . . Alternative #2 (categorical ban below some threshold, no requirements above) . . .” The FTC concluded about this alternative that “non-competes tend to negatively affect competitive conditions in labor markets and in product and service markets, and that non-competes are also exploitative and coercive for workers other than senior executives. For these reasons, the Commission declines to adopt Alternative #2, which contemplated imposing no requirements on workers above a certain wage or other threshold.” This statement appears to be an acknowledgement that non-competes are not “exploitative and coercive” for senior executives, which would undercut the decision not to adopt Alternative #2. Moreover, the FTC does not present the costs and benefits of this or the other alternative rules.
The FTC’s estimate of the effect of the Final Rule on wages is more detailed than that in the NPRM. Nevertheless, it still oversimplifies the effect of the ban, which in all likelihood would vary for different regions and different types of workers. It is far from clear that a more targeted rule limited to lower-skill workers would have resulted in higher wages for those workers, without discouraging investments in training high-skilled workers and subsequently reducing productivity.
The Final Rule adopted by the FTC is quite similar to the proposal in the NRPM, despite tens of thousands of comments and some new relevant literature. The Final Rule relies primarily on the FTC’s reading of the literature discussed in the NPRM, with changes addressing practical considerations for implementation raised by commenters. Interestingly, the Commission faulted some commenters for not presenting empirical research on topics such as the prevalence of non-competes between the seller and the buyer of a business.
The FTC presents detailed estimates of the costs and benefits of the Final Rule, but does not do so for the alternative rules it considered. For example, a more targeted rule that did not eliminate non-compete agreements for high-skilled workers may have resulted in higher wages for low-skilled workers without the potential losses in productivity from reduced incentives to train and invest in high-skilled workers. Depending on the various assumptions needed to calculate costs and benefits, an alternative rule might result in higher net benefits than the Final Rule. One challenge for analyzing such an alternate rule would be the appropriate definition of high-skilled workers.
Regarding the costs and benefits from the Final Rule, the Commission lowered its estimates somewhat of the Final Rule’s impact on wages based on recent research, and recognized there may be a cost to the Final Rule due to lost productivity from reduced training of more highly skilled workers. One may disagree with the Commission’s reading of the economic research (or lack of research) and the large number of comments, but they appear to have influenced and provided a basis for both the NRPM and the Final Rule.
