Anatomy of Public Comments: An Empirical Analysis of Comments on FTC’s Proposed Ban of Employee Non-Competes

Table 1 Public Comments Received on Selected Competition-Related Proposed Rules, Regulations, Guidelines, and Policy Statements from the FTC and DOJ

Table Footnotes

¹ Department of Justice Antitrust Division, “Draft Policy Statement on Licensing Negotiations and Remedies for Standards-Essential Patents Subject to Voluntary F/RAND Commitments,” https://www.regulations.gov/docket/ATR-2021-0001

² Federal Trade Commission, “Request for Information on Merger Enforcement,” https://www.regulations.gov/docket/FTC-2022-0003/comments.

³ Federal Trade Commission, “Non-Compete Clause Rule (NPRM),” https://www.regulations.gov/docket/FTC-2023-0007.

⁴ Federal Trade Commission, “Premerger Notification: Reporting and Waiting Period Requirements,” https://www.regulations.gov/document/FTC-2023-0040-0001.

⁵ Federal Trade Commission, “Draft Merger Guidelines for Public Comment” https://www.regulations.gov/docket/FTC-2023-0043.

Despite the importance of public comments to the agencies’ rulemaking and other activities, the large volume of public comments presents significant challenges to the agencies’ review, given time and resource constraints. Moreover, although the agencies expend considerable efforts in the public comments process, it is unclear whether the submitted comments actually provide the informed and diverse points of view that the agencies seek. What types of individuals, businesses, and organizations are submitting comments to the agencies? What are their viewpoints? What information or evidence are they offering? Do the comments improve the quality of the agencies’ policymaking and regulation efforts? Answering these questions requires a deep dive into the submitters behind and the content of public comments—a task made difficult by the volume of comments that are typically submitted.

To answer some of these questions, we conducted an analysis of the public comments submitted to the FTC on the proposed ban on non-competes. We evaluated a random sample of over 900 comments on the proposed ban to gain insight into who submitted comments on the proposed ban and what they had to say. A team of economists manually reviewed the comments in this sample to gather information on the characteristics of the individuals and organizations that submitted public comments and the bases for the positions the commenters took.

Our study sheds light on the sentiments expressed in the public comments and provides insight into the types of individuals and organizations that weighed in on the proposed ban. The most striking feature of our findings is that the random sample of public comments we reviewed does not reflect a representative sample of U.S. employers and employees. The vast majority of public comments were submitted by employees (not employers), particularly in the healthcare field. Highly skilled professionals, particularly physicians, are overrepresented relative to low-wage employees. Our analysis also shows that fewer public comments were submitted from states where restrictions on the enforceability of non-competes are already in place. These results suggest that certain segments of the economy may be more aware of the proposed ban and have more incentives to submit public comments than others. The views of those who decided to submit comments on the proposed ban may not represent the views of the broader population of employers and employees that the ban would affect. If one were to conduct a carefully designed survey where a representative sample of employers and employees in the United States were selected to comment on the proposed ban, the survey results may yield significantly different results from those that can be gleaned from public comments submitted to the agencies.

Our analysis highlights potential pitfalls in drawing generalizable conclusions from the public comments on the proposed non-compete ban. Although there are no limits on who can submit comments, some groups may not be well represented in the comment process. This results in selection bias (that is, the process collects comments from submitters that for some reason are over-represented, compared with their incidence in the population that the proposed ban is likely to affect). Consequently, the proposed rule is likely to have wider and more varied implications for the U.S. economy than are captured by the public comments, and care must be taken in drawing general conclusions from the comments.

A Brief Overview of the Economics of Non-Compete Agreements

A non-compete clause is an agreement between an employer and an employee that limits the employee’s ability to work for a competitor or to start a competing business after leaving their current job. A typical non-compete clause includes terms that for the time period, geography, and industry scope of the agreement. For example, a non-­compete agreement may specify that an employee cannot work for a competing business that is within a 50-mile radius of their current employer for two years after leaving the firm. Some non-compete agreements restrict an employee only from performing certain types of work for a competitor. For example, an employee who is currently in a sales position might be prohibited from working in a sales position at a competing business but may be free to seek employment with a competitor in another capacity.

From an economic perspective, non-compete agreements can have both procompetitive and anticompetitive effects. One procompetitive rationale for non-compete agreements is that they provide incentives for employers to invest in their employees by providing costly training or sharing trade secrets, client lists, or other proprietary information with employees. Without a non-compete agreement, employees could freely take the benefits of training and information with them to a competitor, and the employer has less incentive to invest in employees in the first place. To the extent non-compete agreements incentivize employers to invest in their employees, non-compete agreements may lead to more productive employees and higher wages and may enhance the employer’s ability to develop more innovative products and services to the benefit of consumers. On the other hand, restrictions on an employee’s mobility in non-competes may be considered anticompetitive effects. In particular, a non-compete agreement may limit an employee’s outside employment options and, by limiting competition among employers for their services, may suppress their wages.

The extent to which non-compete agreements are on balance procompetitive or anticompetitive may differ on a case-to-case basis, which is one reason non-compete agreements have been the subject of intense interest and debate. The random sample of public comments that we review below reflects several features of this debate.

Insights from the Public Comments on the FTC’s Proposed Rule

Between January 5 and April 19, 2023, the FTC received a total of 26,813 comments about its proposed ban of non-compete agreements. Among those submissions, 21,116 were posted and available for public review at the time of our download of these comments in May 2023. We drew a random sample of 926 comments from all available public comments after removing duplicates and short comments that lacked substance. A team of reviewers manually reviewed this random sample of comments. Each reviewer was asked to collect and record information on whether the comment expressed support for or opposition to the proposed rule, and why. In addition, reviewers collected any identifying information on characteristics of the individuals or organizations that submitted comments, including the perspective (employer or employee), the industries or occupations, and the states.

Given the random sampling, findings from the analyses of this sample of comments can provide insight into the broader set of 18,515 unique comments that were available for public review and met a minimum threshold for substance.

Did commenters support or oppose the ban? Why?

Based on the manual review of a random sample of submissions, the reviewers categorized 923 comments into two groups based on whether the submission was in support of or opposition to the ban. The vast majority of public comments (95.4%) supported the proposed ban.

As summarized in Figure 1, the most common reasons expressed for supporting the proposed ban related to potential economic effects of non-compete clauses on employee mobility and compensation. For example, commenters noted that non-compete agreements “stifle career advancement” and “prevent the free movement of labor.” As one commenter stated, “[i]f we had the opportunity to move around the free market, my corporation would have to become wage-competitive rather than keeping talent for less than we are worth.” Another commenter noted the potential impact of non-compete agreements on entrepreneurship, stating that such agreements “discourage ambitious, would-be entrepreneurs from starting new businesses.”

Reasons for Supporting the Proposed Ban of Non-Competes

As shown in Figure 2, the majority of commenters who opposed the proposed rule cited protection of trade secrets or proprietary information and incentives to invest in training and research and development. Several commenters from new or smaller businesses highlighted the need to protect confidential information and investment in training from larger employers. For example, one commenter stated that “[b]ecause larger organizations have more resources to entice personnel and their insider knowledge of a smaller competitor away from them, noncompete and nondisclosure agreements are a necessity for many small, innovative businesses.” Another commenter noted that “we spend months of paid training to prepare our new hires for their new position which is very costly for a small business. . . . We implement non-competes to prevent training those trying to get training for free and then becoming our competition.”

Figure 2 Reasons for Opposing the Proposed Ban of Non-Competes

Other comments reflected a more nuanced view of non-compete clauses, recognizing that non-competes may be appropriate in certain circumstances but not in others. Some commenters voiced opposition to a “blanket ban” on non-competes, stating that “[t]his is not a mandate that can be made for every business in the entire nation” and that “[c]hallenges to such provisions should be examined on a case-by-case basis.” Other commenters were generally in favor of banning non-compete agreements but said that exceptions should be made for executives, such as “certain senior employees with a significant ownership stake” or for “employees that are provided significant financial incentives by the Company that would allow them to sit out for the period of their noncompete.” Another commenter pointed to the differential effect that certain provisions in non-competes can have on subgroups of employees, noting that they “understand protecting specific proprietary information, however, restricting physical distance from employer is ridiculous and unfair” for employees in rural areas.

Who submitted comments on the FTC proposed ban on non-competes?

Many comments provided information sufficient to determine whether the commenter was an employer, employee, or a third party with a vested interest in the final rule issued by a government agency, which provides important context for the opinions expressed in the submissions. As discussed above, differences between the set of individuals and organizations represented in the comments and in the general public can present challenges for the agencies’ review. The group of individuals, businesses, and organizations that are aware of and respond to solicitations for public comments may be those that closely follow regulatory developments and have strong incentives to submit public comments. The reviewing authority might therefore form a skewed impression of the public’s view if the agency fails to consider who is submitting public comments. Below, we describe some of the characteristics of the commenters in our random sample and several potential implications of our findings.

Employer/employee. As may be expected, public comments that supported the proposed ban of non-competes overwhelmingly came from employees, while comments opposing the ban tended to represent the perspectives of employers. We identified 431 comments that were clearly from either an employee or an employer point of view. Of the 406 comments submitted by employees or employee organizations, all but one expressed support for the blanket non-compete ban. 25 (or 5.8%) comments were submitted by employers, among which 18 (or 72%) raised concerns about a blanket ban on non-competes, especially for “individuals who have an extensive understanding of business practices and business relationships.” Employers also expressed concern that “[t]he idea that companies can have other means to protect trade secrets is shallow and misleading and generally will not apply to small businesses.” The remaining seven employer comments supported the proposal. One small business owner noted, “I would never have been able to start my company had I been bound by a non-compete and now that I am responsible for ten livelihoods, the idea that I would prevent them from going on to reach their professional potential is absurd.”

Industry and Occupation. In our random sample of submissions, 407 comments contained sufficient information for us to identify the commenter’s industry, and 306 comments had sufficient information to identify the commenter’s occupation. As shown in Figure 3, the vast majority (70%) of the 407 commenters with information on industry were associated with the healthcare and social assistance industry, all but three of whom expressed their support for the proposed ban. The second-most common industry represented in the comments (7.9%), although with a significant gap from the first, was the professional, scientific, and technical services industry. All 32 commenters from this industry supported the proposed ban.

Figure 3 Top Five Industries Associated with Comments

We observed a similar pattern in the distribution of occupations. Among the 306 commenters who provided information to identify their occupations, 231 (or 75.5%) worked in occupations related to healthcare, as illustrated by Figure 4, and 230 were in favor of banning non-compete agreements.

Figure 4 Top Five Occupations Associated with Comments

It is interesting that over 70% of the comments (with sufficient information to identify the industry and/or occupation) were submitted by individuals associated with healthcare, given the prevalence of non-competes in the healthcare industry and the existence of state laws exempting healthcare employees from non-competes. The share of comments associated with healthcare is more than seven times the sector’s employment share in the United States (9.3% in 2022). Although the FTC must be aware of this heavily skewed distribution, it is unclear how the agency weighs the feedback from healthcare professionals or addresses any potential biases in the inferences that could be drawn about the appropriateness of the proposed rule based on the volume and/or share of comments from healthcare professionals.

It may also be the case that certain individuals are more likely to become aware of proposed regulatory changes because of their relationship to special interest groups. For example, on February 22, 2023, the American Hospital Association (AHA) issued a statement requesting that any final rule must “exempt physicians and senior hospital executives” at a minimum. Approximately one-third (74 out of 231) of the commenters working in healthcare occupations submitted their comment in the subsequent 15 days, suggesting that the AHA’s statement may have increased healthcare practitioners’ awareness of the proposed ban and motivated them to weigh in.

Coordinated comments. Many organizations and interest groups informed their members about the FTC’s invitation for public comments and encouraged them to participate. Some organizations even provided a template for individual members to use when submitting public comments. It is unclear how the FTC or other regulatory agencies weigh these coordinated comments in their rulemaking. Based on our analysis, it may not be straightforward to identify coordinated comments without careful, detailed review. Many templates have placeholders for personal information, and as a result, there are often slight variations in coordinated comments.

Public comments submitted on behalf of organizations. Among the 923 comments that we reviewed, 14 comments were submitted on behalf of organizations representing the viewpoints of a varied set of constituents, including businesses, unions, and groups advocating on behalf of specialty occupations such as physicians and accountants. Unlike comments that were submitted by individuals, seven out of the 14 organizational submissions opposed the proposed ban on non-compete agreements, with some advocating for exceptions for certain groups of employees. The finding that comments submitted by organizations were more evenly divided between supporting and opposing the proposed rule may not be surprising: organizations advocating on behalf of employers or employees were likely to be informed of potential regulatory changes and of the process for expressing their viewpoints through the public comment process, and both sides had the incentives and resources to attempt to influence the commission’s views.

As can be expected, submissions on behalf of specific groups of employers or employees were generally consistent with the stance of the employers or employees they represented. Organizations that represent employees, such as the Illinois College of Emergency Physicians and AFL-CIO, tended to support the ban. In contrast, organizations representing employers and businesses—the Restaurant Law Center, National Restaurant Association, and Tucson Metro Chamber, for example—tended to oppose the proposed ban.

In contrast to the vast majority of public comments on the proposed rule, submissions by organizations tended to take the form of formal letters, were longer and more detailed, and often contained both economic and legal arguments in support of or opposition to the proposed ban. The organizations that submitted public comments often advocated on behalf of thousands, if not millions, of constituents. These differences between comments submitted by individuals and organizations raise questions about how the FTC and other federal agencies do or should weigh these different types of submissions when evaluating public comments.

Where were commenters located? Among the random sample of comments that the team manually reviewed, 477 submissions contained information that allowed us to identify the state in which the commenter was located., At the time of comment submission, three states—California, North Dakota, and Oklahoma—had a near-total ban of non-competes in place. A total of 28 comments (5.9%) were from these states, and all but one of them supported the proposed ban. The remaining 47 states, and Washington D.C., either permitted non-competes with or without exemptions, as of February 2023. For example, Florida allowed non-competes but included exceptions for mediators and physician specialists; Illinois included exceptions for broadcasters, government contractors, physicians, low-wage employees, and certain nurses and certified nurse aides.

A natural question to ask is whether employers and employees from states that permitted non-compete bans, and would therefore be affected by the proposed rule, submitted more public comments than individuals in states that already banned non-competes. Figure 5 presents the geographic distribution of the public comments. The orange circle in each state represents the total number of comments submitted in the state. To account for the fact that states with larger populations may generate more public comments, we calculated the total number of submissions for every 10 million residents in each state. The results suggest that more public comments came from states where non-competes were permitted than states with bans on non-compete agreements.

Figure 5 Distribution of Comments by State

Can the review be automated?

As shown in Table 1 above, the volume of public comments appears to be increasing as means of electronic submissions are made available and as competition agencies take an expansive view of the government’s role in preserving competition—introducing new and updated enforcement tools and strategies. Conducting a manual review and analysis of nearly 27,000 comments submitted in response to the FTC’s proposed rule is a lengthy and costly process.

To understand the feasibility and reliability of automating the review of comments, we developed a machine learning algorithm based on findings from our manual review to categorize or predict the stance of the approximately 17,500 comments that were not manually reviewed. Our algorithm predicted that 97.5% of the 18,515 comments would support the proposed ban. However, our review of the prediction results from the machine learning algorithm raised questions about the reliability of an automated approach. At a minimum, many additional rounds of iterative training guided by human input would be required for the algorithm to accurately predict the stance of commenters.

Indeed, as shown in our manual review of the random sample, many comments presented nuanced views on the proposed ban. Although the public comments disproportionately represented certain constituents, many of the comments also expressed the view that non-competes might be appropriate in some cases and harmful in others, depending on the circumstances. To develop an algorithm that would be able to accurately identify, interpret, and categorize these nuances would require considerable time and investment that would detract from the time- or cost-savings of an automated process. Moreover, even if it were feasible to develop a reliable algorithm to categorize the nuanced views expressed in the comments, it is unclear whether simple quantitative metrics on the comments can provide a meaningful summary of the qualitative review of the complex information provided by commenters. Our review of the public comments also suggests the critical importance of considering the skewed distribution of the comments’ sources before characterizing the comments (or any analysis of the comments) as meaningfully reflecting the view of the general public.

Conclusion

The FTC’s proposed rule banning non-competes is part of a broader wave of policy developments, rulemaking, and enforcement activities by U.S. antitrust regulators—and their increasing concerns about the state of competition in the labor market. The growing number of rules, guidelines, and tools proposed by competition agencies, coupled with an increasing volume of public comments submitted electronically, should motivate a careful examination of the public comments process and its role in informing and shaping federal agencies’ views.

To date, however, there has been only limited research on the contents and impact of public comments on federal agencies’ rulemaking. One study—an analysis of 1,444 interest group comments in reaction to forty federal agency rules from 1994 to 2001—found some evidence that comments from interest groups can affect the content of final government regulations. Although the federal agencies included in the study did not include competition agencies, the study’s findings suggest that, in certain contexts, public comments can have a meaningful impact on federal agencies’ rule- and policy-making. Our study suggests that public comments may not be representative of the groups and individuals affected by a proposed rule, which highlights the need for further research on the implications of the public comments process on policy outcomes. 

The authors thank Josephine Kantawiria, Upamanyu Pathare, and Ana Sanchez Chico for their excellent research assistance. Any opinions expressed herein are the authors’ alone and do not necessarily reflect those of Analysis Group or any of its clients.