Archives For Federal Trade Commission

A White House administration typically announces major new antitrust initiatives in the fall and spring, and this year is no exception. Senior Biden administration officials kicked off the fall season at Fordham Law School (more on that below) by shedding additional light on their plans to expand the accepted scope of antitrust enforcement.

Their aggressive enforcement statements draw headlines, but will the administration’s neo-Brandeisians actually notch enforcement successes? The prospects are cloudy, to say the least.

The U.S. Justice Department (DOJ) has lost some cartel cases in court this year (what was the last time that happened?) and, on Sept. 19, a federal judge rejected the DOJ’s attempt to enjoin United Health’s $13.8 billion bid for Change Healthcare. The Federal Trade Commission (FTC) recently lost two merger challenges before its in-house administrative law judge. It now faces a challenge to its administrative-enforcement processes before the U.S. Supreme Court (the Axon case, to be argued in November).

(Incidentally, on the other side of the Atlantic, the European Commission has faced some obstacles itself. Despite its recent Google victory, the Commission has effectively lost two abuse of dominance cases this year—the Intel and Qualcomm matters—before the European General Court.)

So, are the U.S. antitrust agencies chastened? Will they now go back to basics? Far from it. They enthusiastically are announcing plans to charge ahead, asserting theories of antitrust violations that have not been taken seriously for decades, if ever. Whether this turns out to be wise enforcement policy remains to be seen, but color me highly skeptical. Let’s take a quick look at some of the big enforcement-policy ideas that are being floated.

Fordham Law’s Antitrust Conference

Admiral David Farragut’s order “Damn the torpedoes, full speed ahead!” was key to the Union Navy’s August 1864 victory in the Battle of Mobile Bay, a decisive Civil War clash. Perhaps inspired by this display of risk-taking, the heads of the two federal antitrust agencies—DOJ Assistant Attorney General (AAG) Jonathan Kanter and FTC Chair Lina Khan—took a “damn the economics, full speed ahead” attitude in remarks at the Sept. 16 session of Fordham Law School’s 49th Annual Conference on International Antitrust Law and Policy. Special Assistant to the President Tim Wu was also on hand and emphasized the “all of government” approach to competition policy adopted by the Biden administration.

In his remarks, AAG Kanter seemed to be endorsing a “monopoly broth” argument in decrying the current “Whac-a-Mole” approach to monopolization cases. The intent may be to lessen the burden of proof of anticompetitive effects, or to bring together a string of actions taken jointly as evidence of a Section 2 violation. In taking such an approach, however, there is a serious risk that efficiency-seeking actions may be mistaken for exclusionary tactics and incorrectly included in the broth. (Notably, the U.S. Court of Appeals for the D.C. Circuit’s 2001 Microsoft opinion avoided the monopoly-broth problem by separately discussing specific company actions and weighing them on their individual merits, not as part of a general course of conduct.)

Kanter also recommended going beyond “our horizontal and vertical framework” in merger assessments, despite the fact that vertical mergers (involving complements) are far less likely to be anticompetitive than horizontal mergers (involving substitutes).

Finally, and perhaps most problematically, Kanter endorsed the American Innovative and Choice Online Act (AICOA), citing the protection it would afford “would-be competitors” (but what about consumers?). In so doing, the AAG ignored the fact that AICOA would prohibit welfare-enhancing business conduct and could be harmfully construed to ban mere harm to rivals (see, for example, Stanford professor Doug Melamed’s trenchant critique).

Chair Khan’s presentation, which called for a far-reaching “course correction” in U.S. antitrust, was even more bold and alarming. She announced plans for a new FTC Act Section 5 “unfair methods of competition” (UMC) policy statement centered on bringing “standalone” cases not reachable under the antitrust laws. Such cases would not consider any potential efficiencies and would not be subject to the rule of reason. Endorsing that approach amounts to an admission that economic analysis will not play a serious role in future FTC UMC assessments (a posture that likely will cause FTC filings to be viewed skeptically by federal judges).

In noting the imminent release of new joint DOJ-FTC merger guidelines, Khan implied that they would be animated by an anti-merger philosophy. She cited “[l]awmakers’ skepticism of mergers” and congressional rejection “of economic debits and credits” in merger law. Khan thus asserted that prior agency merger guidance had departed from the law. I doubt, however, that many courts will be swayed by this “economics free” anti-merger revisionism.

Tim Wu’s remarks closing the Fordham conference had a “big picture” orientation. In an interview with GW Law’s Bill Kovacic, Wu briefly described the Biden administration’s “whole of government” approach, embodied in President Joe Biden’s July 2021 Executive Order on Promoting Competition in the American Economy. While the order’s notion of breaking down existing barriers to competition across the American economy is eminently sound, many of those barriers are caused by government restrictions (not business practices) that are not even alluded to in the order.

Moreover, in many respects, the order seeks to reregulate industries, misdiagnosing many phenomena as business abuses that actually represent efficient free-market practices (as explained by Howard Beales and Mark Jamison in a Sept. 12 Mercatus Center webinar that I moderated). In reality, the order may prove to be on net harmful, rather than beneficial, to competition.

Conclusion

What is one to make of the enforcement officials’ bold interventionist screeds? What seems to be missing in their presentations is a dose of humility and pragmatism, as well as appreciation for consumer welfare (scarcely mentioned in the agency heads’ presentations). It is beyond strange to see agencies that are having problems winning cases under conventional legal theories floating novel far-reaching initiatives that lack a sound economics foundation.

It is also amazing to observe the downplaying of consumer welfare by agency heads, given that, since 1979 (in Reiter v. Sonotone), the U.S. Supreme Court has described antitrust as a “consumer welfare prescription.” Unless there is fundamental change in the makeup of the federal judiciary (and, in particular, the Supreme Court) in the very near future, the new unconventional theories are likely to fail—and fail badly—when tested in court. 

Bringing new sorts of cases to test enforcement boundaries is, of course, an entirely defensible role for U.S. antitrust leadership. But can the same thing be said for bringing “non-boundary” cases based on theories that would have been deemed far beyond the pale by both Republican and Democratic officials just a few years ago? Buckle up: it looks as if we are going to find out. 

[This post from Jonathan M. Barnett, the Torrey H. Webb Professor of Law at the University of Southern California’s Gould School of Law, is an entry in Truth on the Market’s FTC UMC Rulemaking symposium. You can find other posts at the symposium page here. Truth on the Market also invites academics, practitioners, and other antitrust/regulation commentators to send us 1,500-4,000 word responses for potential inclusion in the symposium.]

In its Advance Notice for Proposed Rulemaking (ANPR) on Commercial Surveillance and Data Security, the Federal Trade Commission (FTC) has requested public comment on an unprecedented initiative to promulgate and implement wide-ranging rules concerning the gathering and use of consumer data in digital markets. In this contribution, I will assume, for the sake of argument, that the commission has the legal authority to exercise its purported rulemaking powers for this purpose without a specific legislative mandate (a question as to which I recognize there is great uncertainty, which is further heightened by the fact that Congress is concurrently considered legislation in the same policy area).

In considering whether to use these powers for the purposes of adopting and implementing privacy-related regulations in digital markets, the commission would be required to undertake a rigorous assessment of the expected costs and benefits of any such regulation. Any such cost-benefit analysis must comprise at least two critical elements that are omitted from, or addressed in highly incomplete form in, the ANPR.

The Hippocratic Oath of Regulatory Intervention

There is a longstanding consensus that regulatory intervention is warranted only if a market failure can be identified with reasonable confidence. This principle is especially relevant in the case of the FTC, which is entrusted with preserving competitive markets and, therefore, should be hesitant about intervening in market transactions without a compelling evidentiary basis. As a corollary to this proposition, it is also widely agreed that implementing any intervention to correct a market failure would only be warranted to the extent that such intervention would be reasonably expected to correct any such failure at a net social gain.

This prudent approach tracks the “economic effect” analysis that the commission must apply in the rulemaking process contemplated under the Federal Trade Commission Act and the analysis of “projected benefits and … adverse economic effects” of proposed and final rules contemplated by the commission’s rules of practice. Consistent with these requirements, the commission has exhibited a longstanding commitment to thorough cost-benefit analysis. As observed by former Commissioner Julie Brill in 2016, “the FTC conducts its rulemakings with the same level of attention to costs and benefits that is required of other agencies.” Former Commissioner Brill also observed that the “FTC combines our broad mandate to protect consumers with a rigorous, empirical approach to enforcement matters.”

This demanding, fact-based protocol enhances the likelihood that regulatory interventions result in a net improvement relative to the status quo, an uncontroversial goal of any rational public policy. Unfortunately, the ANPR does not make clear that the commission remains committed to this methodology.

Assessing Market Failure in the Use of Consumer Data

To even “get off the ground,” any proposed privacy regulation would be required to identify a market failure arising from a particular use of consumer data. This requires a rigorous and comprehensive assessment of the full range of social costs and benefits that can be reasonably attributed to any such practice.

The ANPR’s Oversights

In contrast to the approach described by former Commissioner Brill, several elements of the ANPR raise significant doubts concerning the current commission’s willingness to assess evidence relevant to the potential necessity of privacy-related regulations in a balanced, rigorous, and comprehensive manner.

First, while the ANPR identifies a plethora of social harms attributable to data-collection practices, it merely acknowledges the possibility that consumers enjoy benefits from such practices “in theory.” This skewed perspective is not empirically serious. Focusing almost entirely on the costs of data collection and dismissing as conjecture any possible gains defies market realities, especially given the fact that (as discussed below) those gains are clearly significant and, in some cases, transformative.

Second, the ANPR’s choice of the normatively charged term “data surveillance” to encompass all uses of consumer data conveys the impression that all data collection through digital services is surreptitious or coerced, whereas (as discussed below) some users may knowingly provide such data to enable certain data-reliant functionalities.

Third, there is no mention in the ANPR that online providers widely provide users with notices concerning certain uses of consumer data and often require users to select among different levels of data collection.

Fourth, the ANPR unusually relies substantially on news websites and non-peer-reviewed publications in the style of policy briefs or advocacy papers, rather than the empirical social-science research on which the commission has historically made policy determinations.

This apparent indifference to analytical balance is particularly exhibited in the ANPR’s failure to address the economic gains generated through the use of consumer data in online markets. As was recognized in a 2014 White House report, many valuable digital services could not function effectively without engaging in some significant level of data collection. The examples are numerous and diverse, including traffic-navigation services that rely on data concerning a user’s geographic location (as well as other users’ geographic location); personalized ad delivery, which relies on data concerning a user’s search history and other disclosed characteristics; and search services, which rely on the ability to use user data to offer search services at no charge while offering targeted advertisements to paying advertisers.

There are equally clear gains on the “supply” side of the market. Data-collection practices can expand market access by enabling smaller vendors to leverage digital intermediaries to attract consumers that are most likely to purchase those vendors’ goods or services. The commission has recognized this point in the past, observing in a 2014 report:

Data brokers provide the information they compile to clients, who can use it to benefit consumers … [C]onsumers may benefit from increased and innovative product offerings fueled by increased competition from small businesses that are able to connect with consumers that they may not have otherwise been able to reach.

Given the commission’s statutory mission under the FTC Act to protect consumers’ interests and preserve competitive markets, these observations should be of special relevance.

Data Protection v. Data-Reliant Functionality

Data-reliant services yield social gains by substantially lowering transaction costs and, in the process, enabling services that would not otherwise be feasible, with favorable effects for consumers and vendors. This observation does not exclude the possibility that specific uses of consumer data may constitute a potential market failure that merits regulatory scrutiny and possible intervention (assuming there is sufficient legal authority for the relevant agency to undertake any such intervention). That depends on whether the social costs reasonably attributable to a particular use of consumer data exceed the social gains reasonably attributable to that use. This basic principle seems to be recognized by the ANPR, which states that the commission can only deem a practice “unfair” under the FTC Act if “it causes or is likely to cause substantial injury” and “the injury is not outweighed by benefits to consumers or competition.”

In implementing this principle, it is important to keep in mind that a market failure could only arise if the costs attributable to any particular use of consumer data are not internalized by the parties to the relevant transaction. This requires showing either that a particular use of consumer data imposes harms on third parties (a plausible scenario in circumstances implicating risks to data security) or consumers are not aware of, or do not adequately assess or foresee, the costs they incur as a result of such use (a plausible scenario in circumstances implicating risks to consumer data). For the sake of brevity, I will focus on the latter scenario.

Many scholars have taken the view that consumers do not meaningfully read privacy notices or consider privacy risks, although the academic literature has also recognized efforts by private entities to develop notice methodologies that can improve consumers’ ability to do so. Even accepting this view, however, it does not necessarily follow (as the ANPR appears to assume) that a more thorough assessment of privacy risks would inevitably lead consumers to elect higher levels of data privacy even where that would degrade functionality or require paying a positive price for certain services. That is a tradeoff that will vary across consumers. It is therefore difficult to predict and easy to get wrong.

As the ANPR indirectly acknowledges in questions 26 and 40, interventions that bar certain uses of consumer data may therefore harm consumers by compelling the modification, positive pricing, or removal from the market of popular data-reliant services. For this reason, some scholars and commentators have favored the informed-consent approach that provides users with the option to bar or limit certain uses of their data. This approach minimizes error costs since it avoids overestimating consumer preferences for privacy. Unlike a flat prohibition of certain uses of consumer data, it also can reflect differences in those preferences across consumers. The ANPR appears to dismiss this concern, asking in question 75 whether certain practices should be made illegal “irrespective of whether consumers consent to them” (my emphasis added).

Addressing the still-uncertain body of evidence concerning the tradeoff between privacy protections on the one hand and data-reliant functionalities on the other (as well as the still-unresolved extent to which users can meaningfully make that tradeoff) lies outside the scope of this discussion. However, the critical observation is that any determination of market failure concerning any particular use of consumer data must identify the costs (and specifically, identify non-internalized costs) attributable to any such use and then offset those costs against the gains attributable to that use.

This balancing analysis is critical. As the commission recognized in a 2015 report, it is essential to strike a balance between safeguarding consumer privacy without suppressing the economic gains that arise from data-reliant services that can benefit consumers and vendors alike. This even-handed approach is largely absent from the ANPR—which, as noted above, focuses almost entirely on costs while largely overlooking the gains associated with the uses of consumer data in online markets. This suggests a one-sided approach to privacy regulation that is incompatible with the cost-benefit analysis that the commission recognizes it must follow in the rulemaking process.

Private-Ordering Approaches to Consumer-Data Regulation

Suppose that a rigorous and balanced cost-benefit analysis determines that a particular use of consumer data would likely yield social costs that exceed social gains. It would still remain to be determined whether and howa regulator should intervene to yield a net social gain. As regulators make this determination, it is critical that they consider the full range of possible mechanisms to address a particular market failure in the use of consumer data.

Consistent with this approach, the FTC Act specifically requires that the commission specify in an ANPR “possible regulatory alternatives under consideration,” a requirement that is replicated at each subsequent stage of the rulemaking process, as provided in the rules of practice. The range of alternatives should include the possibility of taking no action, if no feasible intervention can be identified that would likely yield a net gain.

In selecting among those alternatives, it is imperative that the commission consider the possibility of unnecessary or overly burdensome rules that could impede the efficient development and supply of data-reliant services, either degrading the quality or raising the price of those services. In the past, the commission has emphasized this concern, stating in 2011 that “[t]he FTC actively looks for means to reduce burdens while preserving the effectiveness of a rule.”

This consideration (which appears to be acknowledged in question 24 of the ANPR) is of special importance to privacy-related regulation, given that the estimated annual costs to the U.S. economy (as calculated by the Information Technology and Innovation Foundation) of compliance with the most extensive proposed forms of privacy-related regulations would exceed $100 billion dollars. Those costs would be especially burdensome for smaller entities, effectively raising entry barriers and reducing competition in online markets (a concern that appears to be acknowledged in question 27 of the ANPR).

Given the exceptional breadth of the rules that the ANPR appears to contemplate—cover an ambitious range of activities that would typically be the subject of a landmark piece of federal legislation, rather than administrative rulemaking—it is not clear that the commission has seriously considered this vital point of concern.

In the event that the FTC does move forward with any of these proposed rulemakings (which would be required to rest on a factually supported finding of market failure), it would confront a range of possible interventions in markets for consumer data. That range is typically viewed as being bounded, on the least-interventionist side, by notice and consent requirements to facilitate informed user choice, and on the most interventionist side, by prohibitions that specifically bar certain uses of consumer data.

This is well-traveled ground within the academic and policy literature and the relative advantages and disadvantages of each regulatory approach are well-known (and differ depending on the type of consumer data and other factors). Within the scope of this contribution, I wish to address an alternative regulatory approach that lies outside this conventional range of policy options.

Bottom-Up v. Top-Down Regulation

Any cost-benefit analysis concerning potential interventions to modify or bar a particular use of consumer data, or to mandate notice-and-consent requirements in connection with any such use, must contemplate not only government-implemented solutions but also market-implemented solutions, including hybrid mechanisms in which government action facilitates or complements market-implemented solutions.

This is not a merely theoretical proposal (and is referenced indirectly in questions 36, 51, and 87 of the ANPR). As I have discussed in previously published research, the U.S. economy has a long-established record of having adopted, largely without government intervention, collective solutions to the information asymmetries that can threaten the efficient operation of consumer goods and services markets.

Examples abound: Underwriters Laboratories (UL), which establishes product-safety standards in hundreds of markets; large accounting firms, which confirm compliance with Generally Accepted Accounting Principles (GAAP), which are in turn established and updated by the Financial Accounting Standards Board, a private entity subject to oversight by the Securities and Exchange Commission; and intermediaries in other markets, such as consumer credit, business credit, insurance carriers, bond issuers, and content ratings in the entertainment and gaming industries. Collectively, these markets encompass thousands of providers, hundreds of millions of customers, and billions of dollars in value.

A collective solution is often necessary to resolve information asymmetries efficiently because the benefits from establishing an industrywide standard of product or service quality, together with a trusted mechanism for showing compliance with that standard, generates gains that cannot be fully internalized by any single provider.

Jurisdictions outside the United States have tended to address this collective-action problem through the top-down imposition of standards by government mandate and enforcement by regulatory agencies, as illustrated by the jurisdictions referenced by the ANPR that have imposed restrictions on the use of consumer data through direct regulatory intervention. By contrast, the U.S. economy has tended to favor the bottom-up development of voluntary standards, accompanied by certification and audit services, all accomplished by a mix of industry groups and third-party intermediaries. In certain markets, this may be a preferred model to address the information asymmetries between vendors and customers that are the key sources of potential market failure in the use of consumer data.

Privately organized initiatives to set quality standards and monitor compliance benefit the market by supplying a reliable standard that reduces information asymmetries and transaction costs between consumers and vendors. This, in turn, yields economic gains in the form of increased output, since consumers have reduced uncertainty concerning product quality. These quality standards are generally implemented through certification marks (for example, the “UL” certification mark) or ranking mechanisms (for example, consumer-credit or business-credit scores), which induce adoption and compliance through the opportunity to accrue reputational goodwill that, in turn, translates into economic gains.

These market-implemented voluntary mechanisms are a far less costly means to reduce information asymmetries in consumer-goods markets than regulatory interventions, which require significant investments of public funds in rulemaking, detection, investigation, enforcement, and adjudication activities.

Hybrid Policy Approaches

Private-ordering solutions to collective-action failures in markets that suffer from information asymmetries can sometimes benefit from targeted regulatory action, resulting in a hybrid policy approach. In particular, regulators can sometimes play two supplemental functions in this context.

First, regulators can require that providers in certain markets comply with (or can provide a liability safe harbor for providers that comply with) the quality standards developed by private intermediaries that have developed track records of efficiently establishing those standards and reliably confirming compliance. This mechanism is anticipated by the ANPR, which asks in question 51 whether the commission should “require firms to certify that their commercial surveillance practices meet clear standards concerning collection, use, retention, transfer, or monetization of consumer data” and further asks whether those standards should be set by “the Commission, a third-party organization, or some other entity.”

Other regulatory agencies already follow this model. For example, federal and state regulatory agencies in the fields of health care and education rely on accreditation by designated private entities for purposes of assessing compliance with applicable licensing requirements.

Second, regulators can supervise and review the quality standards implemented, adjusted, and enforced by private intermediaries. This is illustrated by the example of securities markets, in which the major exchanges institute and enforce certain governance, disclosure, and reporting requirements for listed companies but are subject to regulatory oversight by the SEC, which must approve all exchange rules and amendments. Similarly, major accounting firms monitor compliance by public companies with GAAP but must register with, and are subject to oversight by, the Public Company Accounting Oversight Board (PCAOB), a nonprofit entity subject to SEC oversight.

These types of hybrid mechanisms shift to private intermediaries most of the costs involved in developing, updating, and enforcing quality standards (in this context, standards for the use of consumer data) and harness private intermediaries’ expertise, capacities, and incentives to execute these functions efficiently and rapidly, while using targeted forms of regulatory oversight as a complementary policy tool.

Conclusion

Certain uses of consumer data in digital markets may impose net social harms that can be mitigated through appropriately crafted regulation. Assuming, for the sake of argument, that the commission has the legal power to enact regulation to address such harms (again, a point as to which there is great doubt), any specific steps must be grounded in rigorous and balanced cost-benefit analysis.

As a matter of law and sound public policy, it is imperative that the commission meaningfully consider the full range of reliable evidence to identify any potential market failures in the use of consumer data and how to formulate rules to rectify or mitigate such failures at a net social gain. Given the extent to which business models in digital environments rely on the use of consumer data, and the substantial value those business models confer on consumers and businesses, the potential “error costs” of regulatory overreach are high. It is therefore critical to engage in a thorough balancing of costs and gains concerning any such use.

Privacy regulation is a complex and economically consequential policy area that demands careful diagnosis and targeted remedies grounded in analysis and evidence, rather than sweeping interventions accompanied by rhetoric and anecdote.

The Federal Trade Commission (FTC) wants to review in advance all future acquisitions by Facebook parent Meta Platforms. According to a Sept. 2 Bloomberg report, in connection with its challenge to Meta’s acquisition of fitness-app maker Within Unlimited,  the commission “has asked its in-house court to force both Meta and [Meta CEO Mark] Zuckerberg to seek approval from the FTC before engaging in any future deals.”

This latest FTC decision is inherently hyper-regulatory, anti-free market, and contrary to the rule of law. It also is profoundly anti-consumer.

Like other large digital-platform companies, Meta has conferred enormous benefits on consumers (net of payments to platforms) that are not reflected in gross domestic product statistics. In a December 2019 Harvard Business Review article, Erik Brynjolfsson and Avinash Collis reported research finding that Facebook:

…generates a median consumer surplus of about $500 per person annually in the United States, and at least that much for users in Europe. … [I]ncluding the consumer surplus value of just one digital good—Facebook—in GDP would have added an average of 0.11 percentage points a year to U.S. GDP growth from 2004 through 2017.

The acquisition of complementary digital assets—like the popular fitness app produced by Within—enables Meta to continually enhance the quality of its offerings to consumers and thereby expand consumer surplus. It reflects the benefits of economic specialization, as specialized assets are made available to enhance the quality of Meta’s offerings. Requiring Meta to develop complementary assets in-house, when that is less efficient than a targeted acquisition, denies these benefits.

Furthermore, in a recent editorial lambasting the FTC’s challenge to a Meta-Within merger as lacking a principled basis, the Wall Street Journal pointed out that the challenge also removes incentive for venture-capital investments in promising startups, a result at odds with free markets and innovation:

Venture capitalists often fund startups on the hope that they will be bought by larger companies. [FTC Chair Lina] Khan is setting down the marker that the FTC can block acquisitions merely to prevent big companies from getting bigger, even if they don’t reduce competition or harm consumers. This will chill investment and innovation, and it deserves a burial in court.

This is bad enough. But the commission’s proposal to require blanket preapprovals of all future Meta mergers (including tiny acquisitions well under regulatory pre-merger reporting thresholds) greatly compounds the harm from its latest ill-advised merger challenge. Indeed, it poses a blatant challenge to free-market principles and the rule of law, in at least three ways.

  1. It substitutes heavy-handed ex ante regulatory approval for a reliance on competition, with antitrust stepping in only in those limited instances where the hard facts indicate a transaction will be anticompetitive. Indeed, in one key sense, it is worse than traditional economic regulation. Empowering FTC staff to carry out case-by-case reviews of all proposed acquisitions inevitably will generate arbitrary decision-making, perhaps based on a variety of factors unrelated to traditional consumer-welfare-based antitrust. FTC leadership has abandoned sole reliance on consumer welfare as the touchstone of antitrust analysis, paving the wave for potentially abusive and arbitrary enforcement decisions. By contrast, statutorily based economic regulation, whatever its flaws, at least imposes specific standards that staff must apply when rendering regulatory determinations.
  2. By abandoning sole reliance on consumer-welfare analysis, FTC reviews of proposed Meta acquisitions may be expected to undermine the major welfare benefits that Meta has previously bestowed upon consumers. Given the untrammeled nature of these reviews, Meta may be expected to be more cautious in proposing transactions that could enhance consumer offerings. What’s more, the general anti-merger bias by current FTC leadership would undoubtedly prompt them to reject some, if not many, procompetitive transactions that would confer new benefits on consumers.
  3. Instituting a system of case-by-case assessment and approval of transactions is antithetical to the normal American reliance on free markets, featuring limited government intervention in market transactions based on specific statutory guidance. The proposed review system for Meta lacks statutory warrant and (as noted above) could promote arbitrary decision-making. As such, it seriously flouts the rule of law and threatens substantial economic harm (sadly consistent with other ill-considered initiatives by FTC Chair Khan, see here and here).

In sum, internet-based industries, and the big digital platforms, have thrived under a system of American technological freedom characterized as “permissionless innovation.” Under this system, the American people—consumers and producers—have been the winners.

The FTC’s efforts to micromanage future business decision-making by Meta, prompted by the challenge to a routine merger, would seriously harm welfare. To the extent that the FTC views such novel interventionism as a bureaucratic template applicable to other disfavored large companies, the American public would be the big-time loser.

[This post is an entry in Truth on the Market’s FTC UMC Rulemaking symposium. You can find other posts at the symposium page here. Truth on the Market also invites academics, practitioners, and other antitrust/regulation commentators to send us 1,500-4,000 word responses for potential inclusion in the symposium.]

The Federal Trade Commission’s (FTC) Aug. 22 Advance Notice of Proposed Rulemaking on Commercial Surveillance and Data Security (ANPRM) is breathtaking in its scope. For an overview summary, see this Aug. 11 FTC press release.

In their dissenting statements opposing ANPRM’s release, Commissioners Noah Phillips and Christine Wilson expertly lay bare the notice’s serious deficiencies. Phillips’ dissent stresses that the ANPRM illegitimately arrogates to the FTC legislative power that properly belongs to Congress:

[The [A]NPRM] recast[s] the Commission as a legislature, with virtually limitless rulemaking authority where personal data are concerned. It contemplates banning or regulating conduct the Commission has never once identified as unfair or deceptive. At the same time, the ANPR virtually ignores the privacy and security concerns that have animated our [FTC] enforcement regime for decades. … [As such, the ANPRM] is the first step in a plan to go beyond the Commission’s remit and outside its experience to issue rules that fundamentally alter the internet economy without a clear congressional mandate. That’s not “democratizing” the FTC or using all “the tools in the FTC’s toolbox.” It’s a naked power grab.

Wilson’s complementary dissent critically notes that the 2021 changes to FTC rules of practice governing consumer-protection rulemaking decrease opportunities for public input and vest significant authority solely with the FTC chair. She also echoed Phillips’ overarching concern with FTC overreach (footnote citations omitted):

Many practices discussed in this ANPRM are presented as clearly deceptive or unfair despite the fact that they stretch far beyond practices with which we are familiar, given our extensive law enforcement experience. Indeed, the ANPRM wanders far afield of areas for which we have clear evidence of a widespread pattern of unfair or deceptive practices. … [R]egulatory and enforcement overreach increasingly has drawn sharp criticism from courts. Recent Supreme Court decisions indicate FTC rulemaking overreach likely will not fare well when subjected to judicial review.

Phillips and Wilson’s warnings are fully warranted. The ANPRM contemplates a possible Magnuson-Moss rulemaking pursuant to Section 18 of the FTC Act,[1] which authorizes the commission to promulgate rules dealing with “unfair or deceptive acts or practices.” The questions that the ANPRM highlights center primarily on concerns of unfairness.[2] Any unfairness-related rulemaking provisions eventually adopted by the commission will have to satisfy a strict statutory cost-benefit test that defines “unfair” acts, found in Section 5(n) of the FTC Act. As explained below, the FTC will be hard-pressed to justify addressing most of the ANPRM’s concerns in Section 5(n) cost-benefit terms.

Discussion

The requirements imposed by Section 5(n) cost-benefit analysis

Section 5(n) codifies the meaning of unfair practices, and thereby constrains the FTC’s application of rulemakings covering such practices. Section 5(n) states:

The Commission shall have no authority … to declare unlawful an act or practice on the grounds that such an act or practice is unfair unless the act or practice causes or is likely to cause substantial injury to consumers which is not reasonably avoidable by consumers themselves and not outweighed by countervailing benefits to consumers or to competition. In determining whether an act or practice is unfair, the Commission may consider established public policies as evidence to be considered with all other evidence. Such public policy considerations may not serve as a primary basis for such determination.

In other words, a practice may be condemned as unfair only if it causes or is likely to cause “(1) substantial injury to consumers (2) which is not reasonably avoidable by consumers themselves and (3) not outweighed by countervailing benefits to consumers or to competition.”

This is a demanding standard. (For scholarly analyses of the standard’s legal and economic implications authored by former top FTC officials, see here, here, and here.)

First, the FTC must demonstrate that a practice imposes a great deal of harm on consumers, which they could not readily have avoided. This requires detailed analysis of the actual effects of a particular practice, not mere theoretical musings about possible harms that may (or may not) flow from such practice. Actual effects analysis, of course, must be based on empiricism: consideration of hard facts.

Second, assuming that this formidable hurdle is overcome, the FTC must then acknowledge and weigh countervailing welfare benefits that might flow from such a practice. In addition to direct consumer-welfare benefits, other benefits include “benefits to competition.” Those may include business efficiencies that reduce a firm’s costs, because such efficiencies are a driver of vigorous competition and, thus, of long-term consumer welfare. As the Organisation for Economic Co-operation and Development has explained (see OECD Background Note on Efficiencies, 2012, at 14), dynamic and transactional business efficiencies are particularly important in driving welfare enhancement.

In sum, under Section 5(n), the FTC must show actual, fact-based, substantial harm to consumers that they could not have escaped, acting reasonably. The commission must also demonstrate that such harm is not outweighed by consumer and (procompetitive) business-efficiency benefits. What’s more, Section 5(n) makes clear that the FTC cannot “pull a rabbit out of a hat” and interject other “public policy” considerations as key factors in the rulemaking  calculus (“[s]uch [other] public policy considerations may not serve as a primary basis for … [a] determination [of unfairness]”).

It ineluctably follows as a matter of law that a Section 18 FTC rulemaking sounding in unfairness must be based on hard empirical cost-benefit assessments, which require data grubbing and detailed evidence-based economic analysis. Mere anecdotal stories of theoretical harm to some consumers that is alleged to have resulted from a practice in certain instances will not suffice.

As such, if an unfairness-based FTC rulemaking fails to adhere to the cost-benefit framework of Section 5(n), it inevitably will be struck down by the courts as beyond the FTC’s statutory authority. This conclusion is buttressed by the tenor of the Supreme Court’s unanimous 2021 opinion in AMG Capital v. FTC, which rejected the FTC’s claim that its statutory injunctive authority included the ability to obtain monetary relief for harmed consumers (see my discussion of this case here).

The ANPRM and Section 5(n)

Regrettably, the tone of the questions posed in the ANPRM indicates a lack of consideration for the constraints imposed by Section 5(n). Accordingly, any future rulemaking that sought to establish “remedies” for many of the theorized abuses found in the ANPRM would stand very little chance of being upheld in litigation.

The Aug. 11 FTC press release cited previously addresses several broad topical sources of harms: harms to consumers; harms to children; regulations; automated systems; discrimination; consumer consent; notice, transparency, and disclosure; remedies; and obsolescence. These categories are chock full of questions that imply the FTC may consider restrictions on business conduct that go far beyond the scope of the commission’s authority under Section 5(n). (The questions are notably silent about the potential consumer benefits and procompetitive efficiencies that may arise from the business practices called here into question.)

A few of the many questions set forth under just four of these topical listings (harms to consumers, harms to children, regulations, and discrimination) are highlighted below, to provide a flavor of the statutory overreach that categorizes all aspects of the ANPRM. Many other examples could be cited. (Phillips’ dissenting statement provides a cogent and critical evaluation of ANPRM questions that embody such overreach.) Furthermore, although there is a short discussion of “costs and benefits” in the ANPRM press release, it is wholly inadequate to the task.

Under the category “harms to consumers,” the ANPRM press release focuses on harm from “lax data security or surveillance practices.” It asks whether FTC enforcement has “adequately addressed indirect pecuniary harms, including potential physical harms, psychological harms, reputational injuries, and unwanted intrusions.” The press release suggests that a rule might consider addressing harms to “different kinds of consumers (e.g., young people, workers, franchisees, small businesses, women, victims of stalking or domestic violence, racial minorities, the elderly) in different sectors (e.g., health, finance, employment) or in different segments or ‘stacks’ of the internet economy.”

These laundry lists invite, at best, anecdotal public responses alleging examples of perceived “harm” falling into the specified categories. Little or no light is likely to be shed on the measurement of such harm, nor on the potential beneficial effects to some consumers from the practices complained of (for example, better targeted ads benefiting certain consumers). As such, a sound Section 5(n) assessment would be infeasible.

Under “harms to children,” the press release suggests possibly extending the limitations of the FTC-administered Children’s Online Privacy Protection Act (COPPA) to older teenagers, thereby in effect rewriting COPPA and usurping the role of Congress (a clear statutory overreach). The press release also asks “[s]hould new rules set out clear limits on personalized advertising to children and teenagers irrespective of parental consent?” It is hard (if not impossible) to understand how this form of overreach, which would displace the supervisory rights of parents (thereby imposing impossible-to-measure harms on them), could be shoe-horned into a defensible Section 5(n) cost-benefit assessment.

Under “regulations,” the press release asks whether “new rules [should] require businesses to implement administrative, technical, and physical data security measures, including encryption techniques, to protect against risks to the security, confidentiality, or integrity of covered data?” Such new regulatory strictures (whose benefits to some consumers appear speculative) would interfere significantly in internal business processes. Specifically, they could substantially diminish the efficiency of business-security measures, diminish business incentives to innovate (for example, in encryption), and reduce dynamic competition among businesses.

Consumers also would be harmed by a related slowdown in innovation. Those costs undoubtedly would be high but hard, if not impossible, to measure. The FTC also asks whether a rule should limit “companies’ collection, use, and retention of consumer data.” This requirement, which would seemingly bypass consumers’ decisions to make their data available, would interfere with companies’ ability to use such data to improve business offerings and thereby enhance consumers’ experiences. Justifying new requirements such as these under Section 5(n) would be well-nigh impossible.

The category “discrimination” is especially problematic. In addressing “algorithmic discrimination,” the ANPRM press release asks whether the FTC should “consider new trade regulation rules that bar or somehow limit the deployment of any system that produces discrimination, irrespective of the data or processes on which those outcomes are based.” In addition, the press release asks “if the Commission [should] consider harms to other underserved groups that current law does not recognize as protected from discrimination (e.g., unhoused people or residents of rural communities)?”

The FTC cites no statutory warrant for the authority to combat such forms of “discrimination.” It is not a civil-rights agency. It clearly is not authorized to issue anti-discrimination rules dealing with “groups that current law does not recognize as protected from discrimination.” Any such rules, if issued, would be summarily struck down in no uncertain terms by the judiciary, even without regard to Section 5(n).

In addition, given the fact that “economic discrimination” often is efficient (and procompetitive) and may be beneficial to consumer welfare (see, for example, here), more limited economic anti-discrimination rules almost certainly would not pass muster under the Section 5(n) cost-benefit framework.     

Finally, while the ANPRM press release does contain a very short section entitled “costs and benefits,” that section lacks any specific reference to the required Section 5(n) evaluation framework. Phillips’ dissent points out that the ANPRM:

…simply fail[s] to provide the detail necessary for commenters to prepare constructive responses” on cost-benefit analysis. He stresses that the broad nature of requests for commenters’ view on costs and benefits renders the inquiry “not conducive to stakeholders submitting data and analysis that can be compared and considered in the context of a specific rule. … Without specific questions about [the costs and benefits of] business practices and potential regulations, the Commission cannot hope for tailored responses providing a full picture of particular practices.

In other words, the ANPRM does not provide the guidance needed to prompt the sorts of responses that might assist the FTC in carrying out an adequate Section 5(n) cost-benefit analysis.

Conclusion

The FTC would face almost certain defeat in court if it promulgated a broad rule addressing many of the perceived unfairness-based “ills” alluded to in the ANPRM. Moreover, although its requirements would (I believe) not come into effect, such a rule nevertheless would impose major economic costs on society.

Prior to final judicial resolution of its status, the rule would disincentivize businesses from engaging in a variety of data-related practices that enhance business efficiency and benefit many consumers. Furthermore, the FTC resources devoted to developing and defending the rule would not be applied to alternative welfare-enhancing FTC activities—a substantial opportunity cost.

The FTC should take heed of these realities and opt not to carry out a rulemaking based on the ANPRM. It should instead devote its scarce consumer protection resources to prosecuting hard core consumer fraud and deception—and, perhaps, to launching empirical studies into the economic-welfare effects of data security and commercial surveillance practices. Such studies, if carried out, should focus on dispassionate economic analysis and avoid policy preconceptions. (For example, studies involving digital platforms should take note of the existing economic literature, such as a paper indicating that digital platforms have generated enormous consumer-welfare benefits not accounted for in gross domestic product.)

One can only hope that a majority of FTC commissioners will apply common sense and realize that far-flung rulemaking exercises lacking in statutory support are bad for the rule of law, bad for the commission’s reputation, bad for the economy, and bad for American consumers.


[1] The FTC states specifically that it “is issuing this ANPR[M] pursuant to Section 18 of the Federal Trade Commission Act”.

[2] Deceptive practices that might be addressed in a Section 18 trade regulation rule would be subject to the “FTC Policy Statement on Deception,” which states that “the Commission will find deception if there is a representation, omission or practice that is likely to mislead the consumer acting reasonably in the circumstances, to the consumer’s detriment.” A court reviewing an FTC Section 18 rule focused on “deceptive acts or practices” undoubtedly would consult this Statement, although it is not clear, in light of recent jurisprudential trends, that the court would defer to the Statement’s analysis in rendering an opinion. In any event, questions of deception, which focus on acts or practices that mislead consumers, would in all likelihood have little relevance to the evaluation of any rule that might be promulgated in light of the ANPRM.    

A recent viral video captures a prevailing sentiment in certain corners of social media, and among some competition scholars, about how mergers supposedly work in the real world: firms start competing on price, one firm loses out, that firm agrees to sell itself to the other firm and, finally, prices are jacked up.(Warning: Keep the video muted. The voice-over is painful.)

The story ends there. In this narrative, the combination offers no possible cost savings. The owner of the firm who sold doesn’t start a new firm and begin competing tomorrow, and nor does anyone else. The story ends with customers getting screwed.

And in this telling, it’s not just horizontal mergers that look like the one in the viral egg video. It is becoming a common theory of harm regarding nonhorizontal acquisitions that they are, in fact, horizontal acquisitions in disguise. The acquired party may possibly, potentially, with some probability, in the future, become a horizontal competitor. And of course, the story goes, all horizontal mergers are anticompetitive.

Therefore, we should have the same skepticism toward all mergers, regardless of whether they are horizontal or vertical. Steve Salop has argued that a problem with the Federal Trade Commission’s (FTC) 2020 vertical merger guidelines is that they failed to adopt anticompetitive presumptions.

This perspective is not just a meme on Twitter. The FTC and U.S. Justice Department (DOJ) are currently revising their guidelines for merger enforcement and have issued a request for information (RFI). The working presumption in the RFI (and we can guess this will show up in the final guidelines) is exactly the takeaway from the video: Mergers are bad. Full stop.

The RFI repeatedly requests information that would support the conclusion that the agencies should strengthen merger enforcement, rather than information that might point toward either stronger or weaker enforcement. For example, the RFI asks:

What changes in standards or approaches would appropriately strengthen enforcement against mergers that eliminate a potential competitor?

This framing presupposes that enforcement should be strengthened against mergers that eliminate a potential competitor.

Do Monopoly Profits Always Exceed Joint Duopoly Profits?

Should we assume enforcement, including vertical enforcement, needs to be strengthened? In a world with lots of uncertainty about which products and companies will succeed, why would an incumbent buy out every potential competitor? The basic idea is that, since profits are highest when there is only a single monopolist, that seller will always have an incentive to buy out any competitors.

The punchline for this anti-merger presumption is “monopoly profits exceed duopoly profits.” The argument is laid out most completely by Salop, although the argument is not unique to him. As Salop points out:

I do not think that any of the analysis in the article is new. I expect that all the points have been made elsewhere by others and myself.

Under the model that Salop puts forward, there should, in fact, be a presumption against any acquisition, not just horizontal acquisitions. He argues that:

Acquisitions of potential or nascent competitors by a dominant firm raise inherent anticompetitive concerns. By eliminating the procompetitive impact of the entry, an acquisition can allow the dominant firm to continue to exercise monopoly power and earn monopoly profits. The dominant firm also can neutralize the potential innovation competition that the entrant would provide.

We see a presumption against mergers in the recent FTC challenge of Meta’s purchase of Within. While Meta owns Oculus, a virtual-reality headset and Within owns virtual-reality fitness apps, the FTC challenged the acquisition on grounds that:

The Acquisition would cause anticompetitive effects by eliminating potential competition from Meta in the relevant market for VR dedicated fitness apps.

Given the prevalence of this perspective, it is important to examine the basic model’s assumptions. In particular, is it always true that—since monopoly profits exceed duopoly profits—incumbents have an incentive to eliminate potential competition for anticompetitive reasons?

I will argue no. The notion that monopoly profits exceed joint-duopoly profits rests on two key assumptions that hinder the simple application of the “merge to monopoly” model to antitrust.

First, even in a simple model, it is not always true that monopolists have both the ability and incentive to eliminate any potential entrant, simply because monopoly profits exceed duopoly profits.

For the simplest complication, suppose there are two possible entrants, rather than the common assumption of just one entrant at a time. The monopolist must now pay each of the entrants enough to prevent entry. But how much? If the incumbent has already paid one potential entrant not to enter, the second could then enter the market as a duopolist, rather than as one of three oligopolists. Therefore, the incumbent must pay the second entrant an amount sufficient to compensate a duopolist, not their share of a three-firm oligopoly profit. The same is true for buying the first entrant. To remain a monopolist, the incumbent would have to pay each possible competitor duopoly profits.

Because monopoly profits exceed duopoly profits, it is profitable to pay a single entrant half of the duopoly profit to prevent entry. It is not, however, necessarily profitable for the incumbent to pay both potential entrants half of the duopoly profit to avoid entry by either. 

Now go back to the video. Suppose two passersby, who also happen to have chickens at home, notice that they can sell their eggs. The best part? They don’t have to sit around all day; the lady on the right will buy them. The next day, perhaps, two new egg sellers arrive.

For a simple example, consider a Cournot oligopoly model with an industry-inverse demand curve of P(Q)=1-Q and constant marginal costs that are normalized to zero. In a market with N symmetric sellers, each seller earns 1/((N+1)^2) in profits. A monopolist makes a profit of 1/4. A duopolist can expect to earn a profit of 1/9. If there are three potential entrants, plus the incumbent, the monopolist must pay each the duopoly profit of 3*1/9=1/3, which exceeds the monopoly profits of 1/4.

In the Nash/Cournot equilibrium, the incumbent will not acquire any of the competitors, since it is too costly to keep them all out. With enough potential entrants, the monopolist in any market will not want to buy any of them out. In that case, the outcome involves no acquisitions.

If we observe an acquisition in a market with many potential entrants, which any given market may or may not have, it cannot be that the merger is solely about obtaining monopoly profits, since the model above shows that the incumbent doesn’t have incentives to do that.

If our model captures the dynamics of the market (which it may or may not, depending on a given case’s circumstances) but we observe mergers, there must be another reason for that deal besides maintaining a monopoly. The presence of multiple potential entrants overturns the antitrust implications of the truism that monopoly profits exceed duopoly profits. The question turns instead to empirical analysis of the merger and market in question, as to whether it would be profitable to acquire all potential entrants.

The second simplifying assumption that restricts the applicability of Salop’s baseline model is that the incumbent has the lowest cost of production. He rules out the possibility of lower-cost entrants in Footnote 2:

Monopoly profits are not always higher. The entrant may have much lower costs or a better or highly differentiated product. But higher monopoly profits are more usually the case.

If one allows the possibility that an entrant may have lower costs (even if those lower costs won’t be achieved until the future, when the entrant gets to scale), it does not follow that monopoly profits (under the current higher-cost monopolist) necessarily exceed duopoly profits (with a lower-cost producer involved).

One cannot simply assume that all firms have the same costs or that the incumbent is always the lowest-cost producer. This is not just a modeling choice but has implications for how we think about mergers. As Geoffrey Manne, Sam Bowman, and Dirk Auer have argued:

Although it is convenient in theoretical modeling to assume that similarly situated firms have equivalent capacities to realize profits, in reality firms vary greatly in their capabilities, and their investment and other business decisions are dependent on the firm’s managers’ expectations about their idiosyncratic abilities to recognize profit opportunities and take advantage of them—in short, they rest on the firm managers’ ability to be entrepreneurial.

Given the assumptions that all firms have identical costs and there is only one potential entrant, Salop’s framework would find that all possible mergers are anticompetitive and that there are no possible efficiency gains from any merger. That’s the thrust of the video. We assume that the whole story is two identical-seeming women selling eggs. Since the acquired firm cannot, by assumption, have lower costs of production, it cannot improve on the incumbent’s costs of production.

Many Reasons for Mergers

But whether a merger is efficiency-reducing and bad for competition and consumers needs to be proven, not just assumed.

If we take the basic acquisition model literally, every industry would have just one firm. Every incumbent would acquire every possible competitor, no matter how small. After all, monopoly profits are higher than duopoly profits, and so the incumbent both wants to and can preserve its monopoly profits. The model does not give us a way to disentangle when mergers would stop without antitrust enforcement.

Mergers do not affect the production side of the economy, under this assumption, but exist solely to gain the market power to manipulate prices. Since the model finds no downsides for the incumbent to acquiring a competitor, it would naturally acquire every last potential competitor, no matter how small, unless prevented by law. 

Once we allow for the possibility that firms differ in productivity, however, it is no longer true that monopoly profits are greater than industry duopoly profits. We can see this most clearly in situations where there is “competition for the market” and the market is winner-take-all. If the entrant to such a market has lower costs, the profit under entry (when one firm wins the whole market) can be greater than the original monopoly profits. In such cases, monopoly maintenance alone cannot explain an entrant’s decision to sell.

An acquisition could therefore be both procompetitive and increase consumer welfare. For example, the acquisition could allow the lower-cost entrant to get to scale quicker. The acquisition of Instagram by Facebook, for example, brought the photo-editing technology that Instagram had developed to a much larger market of Facebook users and provided a powerful monetization mechanism that was otherwise unavailable to Instagram.

In short, the notion that incumbents can systematically and profitably maintain their market position by acquiring potential competitors rests on assumptions that, in practice, will regularly and consistently fail to materialize. It is thus improper to assume that most of these acquisitions reflect efforts by an incumbent to anticompetitively maintain its market position.

It’s been a busy summer, and promises to be a busier fall. So the UMC Roundup is on hiatus this week.

But because the news doesn’t stop even when we do, we’re using this week’s Roundup to announce a call for submissions relating to the FTC’s ANPR on Commercial Surveillance and Data Security. Submissions relating to various aspects of the ANPR will be considered for publication as part of our ongoing FTC UMC Symposium. We have already previously offered some discussion of the ANPR on Truth on the Market, here and here.

Posts should substantively engage with the ANPR and will generally be between 1,800-4,000 words. We are interested in all topics and perspectives. Given that this is the UMC symposium, we are particularly interested in submissions that explore the competition aspects of the ANPR, including the mysterious Footnote 47 and the procedural and substantive overlaps between the FTC’s UDAP and UMC authorities that run throughout the ANPR.

Submissions should be sent to Keith Fierro (kfierro@laweconcenter.org). To maximize the likelihood that we will publish your submission, we encourage potential authors to submit a brief explanation of the proposed topic prior to writing. Because selected submissions will be published as part of the ongoing UMC Symposium, we anticipate beginning to publish selected submissions immediately and on a rolling basis. For full consideration, contributions should be submitted prior to Sept. 8, 2022.

The FTC UMC Roundup, part of the Truth on the Market FTC UMC Symposium, is a weekly roundup of news relating to the Federal Trade Commission’s antitrust and Unfair Methods of Competition authority. If you would like to receive this and other posts relating to these topics, subscribe to the RSS feed here. If you have news items you would like to suggest for inclusion, please mail them to us at ghurwitz@laweconcenter.org and/or kfierro@laweconcenter.org.

[TOTM: This guest post from Svetlana S. Gans and Natalie Hausknecht of Gibson Dunn is part of the Truth on the Market FTC UMC Symposium. If you would like to receive this and other posts relating to these topics, subscribe to the RSS feed here. If you have news items you would like to suggest for inclusion, please mail them to us at ghurwitz@laweconcenter.org and/or kfierro@laweconcenter.org.]

The Federal Trade Commission (FTC) launched one of the most ambitious rulemakings in agency history Aug. 11, with its 3-2 vote to initiate Advance Notice of Proposed Rulemaking (ANPRM) on commercial surveillance and data security. The divided vote, which broke down on partisan lines, stands in stark contrast to recent bipartisan efforts on Capitol Hill, particularly on the comprehensive American Data Privacy and Protection Act (ADPPA).  

Although the rulemaking purports to pursue a new “privacy and data security” regime, it targets far more than consumer privacy. The ANPRM lays out a sweeping project to rethink the regulatory landscape governing nearly every facet of the U.S. internet economy, from advertising to anti-discrimination law, and even to labor relations. Any entity that uses the internet (even for internal purposes) is likely to be affected by this latest FTC action, and public participation in the proposed rulemaking will be important to ensure the agency gets it right.

Summary of the ANPRM  

The vague scope of the FTC’s latest ANPRM begins at its title: “Commercial Surveillance and Data Security” Rulemaking. The announcement states the FTC intends to explore rules “cracking down” on the “business of collecting, analyzing, and profiting from information about people.” The ANPRM then defines the scope of “commercial surveillance” to include virtually any data activity. For example, the ANPRM explains that it includes practices used “to set prices, curate newsfeeds, serve advertisements, and conduct research on people’s behavior, among other things.” The ANPRM also goes on to say that it is concerned about practices “outside of the retail consumer setting” that the agency traditionally regulates. Indeed, the ANPRM defines “consumer” to include “businesses and workers, not just individuals who buy or exchange data for retail goods and services.”

Unlike the bipartisan ADPPA, the ANPRM also takes aim at the “consent” model that the FTC has long advocated to ensure consumers make informed choices about their data online. It claims that “consumers may become resigned to” data practices and “have little to no actual control over what happens to their information.” It also suggests that consumers “do not generally understand” data practices, such that their permission could be “meaningful”—making express consumer consent to data practices “irrelevant.”

The ANPRM further lists a disparate set of additional FTC concerns, from “pernicious dark pattern practices” to “lax data security practices” to “sophisticated digital advertising systems” to “stalking apps,” “cyber bullying, cyberstalking, and the distribution of child sexual abuse material,” and the use of “social media” among “kids and teens.” It “finally” wraps up with a reference to “growing reliance on automated systems” that may create “new forms and mechanisms for discrimination” in areas like housing, employment, and healthcare. The issue the agency expresses about these automated systems is with apparent “disparate outcomes” “even when automated systems consider only unprotected consumer traits.”

Having set out these concerns, the ANPRM seeks to justify a new rulemaking via a list of what it describes as “decades” of “consumer data privacy and security” enforcement actions. The rulemaking then requests that the public answer 95 questions, covering many different legal and factual issues. For example, the agency requests the public weigh in on the practices “companies use to surveil consumers,” intangible and unmeasurable “harms” created by such practices, the most harmful practices affecting children and teens, techniques that “manipulate consumers into prolonging online activity,” how the commission should balance costs and benefits from any regulation, biometric data practices, algorithmic errors and disparate impacts, the viability of consumer consent, the opacity of “consumer surveillance practices,” and even potential remedies the agency should consider.  

Commissioner Statements in Support of the ANPR

Every Democratic commissioner issued a separate supporting statement. Chair Lina Khan’s statement justified the rulemaking grounds that the FTC is the “de facto law enforcer in this domain.” She also doubled-down on the decision to address not only consumer privacy, but issues affecting all “opportunities in our economy and society, as well as core civil liberties and civil rights” and described being “especially eager to build a record” related to: the limits of “notice and consent” frameworks, as opposed to withdrawing permission for data collection “in the first place”; how to navigate “information asymmetries” with companies; how to address certain “business models” “premised on” persistent tracking; discrimination in automated processes; and workplace surveillance.   

Commissioner Rebecca Kelly Slaughter’s longer statement more explicitly attacked the agency’s “notice-and-consent regime” as having “failed to protect users.” She expressed hope that the new rules would take on biometric or location tracking, algorithmic decision-making, and lax data security practices as “long overdue.” Commission Slaughter further brushed aside concerns that the rulemaking was inappropriate while Congress considered comprehensive privacy legislation, asserting that the magnitude of the rulemaking was a reason to do it—not shy away. She also expressed interest in data-minimization specifications, discriminatory algorithms, and kids and teens issues.

Commissioner Alvaro Bedoya’s short statement likewise expressed support for acting. However, he noted the public comment period would help the agency “discern whether and how to proceed.” Like his colleagues, he identified his particular interest in “emerging discrimination issues”: the mental health of kids and teens; the protection of non-English speaking communities; and biometric data. On the pending privacy legislation, he noted that:

[ADPPA] is the strongest privacy bill that has ever been this close to passing. I hope it does pass. I hope it passes soon…. This ANPRM will not interfere with that effort. I want to be clear: Should the ADPPA pass, I will not vote for any rule that overlaps with it.

Commissioner Statements Opposed to the ANPRM

Both Republican commissioners published dissents. Commissioner Christine S. Wilson’s urged deference to Congress as it considers a comprehensive privacy law. Yet she also expressed broader concern about the FTC’s recent changes to its Section 18 rulemaking process that “decrease opportunities for public input and vest significant authority for the rulemaking proceedings solely with the Chair” and the unjustified targeting of practices not subject to prior enforcement action. Notably, Commissioner Wilson also worried the rulemaking was unlikely to survive judicial scrutiny, indicating that Chair Khan’s statements give her “no basis to believe that she will seek to ensure that proposed rule provisions fit within the Congressionally circumscribed jurisdiction of the FTC.”  

Commissioner Noah Phillips’ dissent criticized the ANPRM for failing to provide “notice of anything” and thus stripping the public of its participation rights. He argued that the ANPRM’s “myriad” questions appear to be a “mechanism to fish for legal theories that might justify outlandish regulatory ambition outside our jurisdiction.” He further noted that the rulemaking positions the FTC as a legislature to regulate in areas outside of its expertise (e.g., labor law) with potentially disastrous economic costs that it is ill-equipped to understand.

Commissioner Phillips further argued the ANPRM attacks disparate practices based on an “amalgam of cases concerning very different business models and conduct” that cannot show the prevalence of misconduct required for Section 18 rulemaking. He also criticized the FTC for abandoning its own informed-consent model based on paternalistic musings about individuals’ ability to decide for themselves. And finally, he criticized the FTC’s apparent overreach in claiming the mantle of “civil rights enforcer” when it was never given that explicit authority by Congress to declare discrimination or disparate impacts unlawful in this space. 

Implications for Regulated Entities and Others Concerned with Potential Agency Overreach

The sheer breadth of the ANPRM demands the avid attention of potentially regulated entities or those concerned with the FTC’s aggressive rulemaking agenda. The public should seek to meaningfully participate in the rulemaking process to ensure the FTC considers a broad array of viewpoints and has the facts before it necessary to properly define the scope of its own authority and the consequences of any proposed privacy regulation. For example, the FTC may issue a notice of proposed rulemaking defining acts or practices as unfair or deceptive “only where it has reason to believe that the unfair or deceptive acts or practices which are the subject of the proposed rulemaking are prevalent.”(emphasis added).

15 U.S. Code § 57a also states that the FTC may make a determination that unfair or deceptive acts or practices are prevalent only if:  “(A) it has issued cease and desist orders regarding such acts or practices, or (B) any other information available to the Commission indicates a widespread pattern of unfair or deceptive acts or practices.” That means that, under the Magnuson-Moss Section 18 rulemaking that the FTC must use here, the agency must show (1) the prevalence of the practices (2) how they are unfair or deceptive, and (3) the economic effect of the rule, including on small businesses and consumers. Any final regulatory analysis also must assess the rule’s costs and benefits and why it was chosen over alternatives. On each count, effective advocacy supported by empirical and sound economic analysis by the public may prove dispositive.

The FTC may have a particularly difficult time meeting this burden of proof with many of the innocuous (and currently permitted) practices identified in the ANPRM. For example, modern online commerce like automated decision-making is a part of the engine that has powered a decade of innovation, lowered logistical and opportunity costs, and opened up amazing new possibilities for small businesses seeking to serve local consumers and their communities. Commissioner Wilson makes this point well:

Many practices discussed in this ANPRM are presented as clearly deceptive or unfair despite the fact that they stretch far beyond practices with which we are familiar, given our extensive law enforcement experience. Indeed, the ANPRM wanders far afield of areas for which we have clear evidence of a widespread pattern of unfair or deceptive practices. 

The FTC also may be setting itself on an imminent collision course with the “major questions” doctrine, in particular. On the last day of its term this year, the Supreme Court handed down West Virginia v. Environmental Protection Agency, which applied the “major questions doctrine” to rule that the EPA can’t base its controversial Clean Power Plan on a novel interpretation of a relatively obscure provision of the Clean Air Act. An agency rule of such vast “economic and political significance,” Chief Justice John Roberts wrote, requires “clear congressional authorization.” (See “The FTC Heads for Legal Trouble” by Svetlana Gans and Eugene Scalia.) Parties are likely to argue the same holds true here with regard to the FTC’s potential regulatory extension into areas like anti-discrimination and labor law. If the FTC remains on this aggressive course, any final privacy rulemaking could also be a tempting target for a reinvigorated nondelegation doctrine.  

Some members of Congress also may question the wisdom of the ANPRM venturing into the privacy realm at all right now, a point advanced by several of the commissioners. Shortly after the FTC’s announcement, House Energy and Commerce Committee Chairman Frank Pallone Jr. (D-N.J.) stated:

I appreciate the FTC’s effort to use the tools it has to protect consumers, but Congress has a responsibility to pass comprehensive federal privacy legislation to better equip the agency, and others, to protect consumers to the greatest extent.

Sen. Roger Wicker (R-Miss.), the ranking member on the Senate Commerce Committee and a leading GOP supporter of the bipartisan legislation, likewise said that the FTC’s move helps “underscore the urgency for the House to bring [ADPPA]  to the floor and for the Senate Commerce Committee to advance it through committee.”  

The FTC’s ANPRM will likely have broad implications for the U.S. economy. Stakeholders can participate in the rulemaking in several ways, including registering by Aug. 31 to speak at the FTC’s Sept. 8 public forum. Stakeholders should also consider submitting public comments and empirical evidence within 60-days of the ANPRM’s publication in the Federal Register, and insist that the FTC hold informal hearings as required under the Magnuson-Moss Act.

While the FTC is rightfully the nation’s top consumer cop, an advanced notice of this scope demands active public awareness and participation to ensure the agency gets it right.  

 

In a recent op-ed for the Wall Street Journal, Svetlana Gans and Eugene Scalia look at three potential traps the Federal Trade Commission (FTC) could trigger if it pursues the aggressive rulemaking agenda many have long been expecting. From their opening:

FTC Chairman Lina Khan has Rooseveltian ambitions for the agency. … Within weeks the FTC is expected to begin a blizzard of rule-makings that will include restrictions on employment noncompete agreements and the practices of technology companies.

If Ms. Khan succeeds, she will transform the FTC’s regulation of American business. But there’s a strong chance this regulatory blitz will fail. The FTC is a textbook case for how federal agencies could be affected by the re-examination of administrative law under way at the Supreme Court.

The first pitfall into which the FTC might fall, Gans and Scalia argue, is the “major questions” doctrine. Recently illuminated in the Supreme Court’s opinion in West Virginia v. EPA decision, the doctrine holds that federal agencies cannot enact regulations of vast economic and political significance without clear congressional authorization. The sorts of rules the FTC appears to be contemplating “would run headlong into” major questions, Gans and Scalia write, a position shared by several contributors to Truth on the Market‘s recent symposium on the potential for FTC rulemakings on unfair methods of competition (UMC).

The second trap the authors expect might trip up an ambitious FTC is the major questions doctrine’s close cousin: the nondelegation doctrine. The nondelegation doctrine holds that there are limits to how much authority Congress can delegate to a federal agency, even if it does so clearly.

Curiously, as Gans and Scalia note, the last time the Supreme Court invoked the nondelegation doctrine involved regulations to implement “codes of fair competition”—nearly identical, on their face, to the commission’s current interest in rules to prohibit unfair methods of competition. That last case, Schechter Poultry Corp. v. United States, is more than 80 years old. The doctrine has since lain dormant for multiple generations. But in recent years, several justice have signaled their openness to reinvigorating the doctrine. As Gans and Scalia note, “[a]n aggressive FTC competition rule could be a tempting target” for them.

Finally, the authors anticipate an overly aggressive FTC may find itself entangled in yet a thorny web wrapped around the very heart of the administrative state: the constitutionality of so-called independent agencies. Again, the relevant constitutional doctrine giving rise to these agencies results from another 1935 case involving the FTC itself: Humphrey’s Executor v. United States. While the Court in that opinion upheld the notion that Congress can create agencies led by officials who operate independently of direct presidential control, conservative justices have long questioned the doctrine’s legitimacy and the Roberts court, in particularly, has trimmed its outer limits. An overly aggressive FTC might present an opportunity to further check the independence of these agencies.

While it remains unclear the precise rules the FTC seek try to develop using its UMC authority, the clearest signs are that it will focus first on labor issues, such as emerging research around labor monopsony and firms’ use of noncompete clauses. Indeed, Eric Posner, who joined the U.S. Justice Department Antitrust Division earlier this year as counsel on these issues, recently acknowledged that: “There is this very close and complicated relationship between labor law and antitrust law that has to be maintained.”

If the FTC were to upset this relationship, such as by using its UMC authority either to circumvent the National Labor Relations Board in addressing competition concerns or to assist the NLRB in exceeding its own statutory authority, it would be unsurprising for the courts to exercise their constitutional role as a check on a rogue agency.

[TOTM: The following is part of a digital symposium by TOTM guests and authors on Antitrust’s Uncertain Future: Visions of Competition in the New Regulatory Landscape. Information on the authors and the entire series of posts is available here.]

When I was a kid, I trailed behind my mother in the grocery store with a notepad and a pencil adding up the cost of each item she added to our cart. This was partly my mother’s attempt to keep my math skills sharp, but it was also a necessity. As a low-income family, there was no slack in the budget for superfluous spending. The Hostess cupcakes I longed for were a luxury item that only appeared in our cart if there was an unexpected windfall. If the antitrust populists who castigate all forms of market power succeed in their crusade to radically deconcentrate the economy, life will be much harder for low-income families like the one I grew up in.

Antitrust populists like Biden White House official Tim Wu and author Matt Stoller decry the political influence of large firms. But instead of advocating for policies that tackle this political influence directly, they seek reforms to antitrust enforcement that aim to limit the economic advantages of these firms, believing that will translate into political enfeeblement. The economic advantages arising from scale benefit consumers, particularly low-income consumers, often at the expense of smaller economic rivals. But because the protection of small businesses is so paramount to their worldview, antitrust populists blithely ignore the harm that advancing their objectives would cause to low-income families.

This desire to protect small businesses, without acknowledging the economic consequences for low-income families, is plainly obvious in calls for reinvigorated Robinson-Patman Act enforcement (a law from the 1930s for which independent businesses advocated to limit the rise of chain stores) and in plans to revise the antitrust enforcement agencies’ merger guidelines. The U.S. Justice Department (DOJ) and the Federal Trade Commission (FTC) recently held a series of listening sessions to demonstrate the need for new guidelines. During the listening session on food and agriculture, independent grocer Anthony Pena described the difficulty he has competing with larger competitors like Walmart. He stated that:

Just months ago, I was buying a 59-ounce orange juice just north of $4 a unit, where we couldn’t get the supplier to sell it to us … Meanwhile, I go to the bigger box like a Walmart or a club store. Not only do they have it fully stocked, but they have it about half the price that I would buy it for at cost.

Half the price. Anthony Pena is complaining that competitors such as Walmart are selling the same product at half the price. To protect independent grocers like Anthony Pena, antitrust populists would have consumers, including low-income families, pay twice as much for groceries.

Walmart is an important food retailer for low-income families. Nearly a fifth of all spending through the Supplemental Nutrition Assistance Program (SNAP), the program formerly known as food stamps, takes place at Walmart. After housing and transportation, food is the largest expense for low-income families. The share of expenditures going toward food for low-income families (i.e., families in the lowest 20% of the income distribution) is 34% higher than for high-income families (i.e., families in the highest 20% of the income distribution). This means that higher grocery prices disproportionately burden low-income families.

In 2019, the U.S. Department of Agriculture (USDA) launched the SNAP Online Purchasing Pilot, which allows SNAP recipients to use their benefits at online food retailers. The pandemic led to an explosion in the number of SNAP recipients using their benefits online—increasing from just 35,000 households in March 2020 to nearly 770,000 households just three months later. While the pilot originally only included Walmart and Amazon, the number of eligible retailers has expanded rapidly. In order to make grocery delivery more accessible to low-income families, an important service during the pandemic, Amazon reduced its Prime membership fee (which helps pay for free delivery) by 50% for SNAP recipients.

The antitrust populists are not only targeting the advantages of large brick-and-mortar retailers, such as Walmart, but also of large online retailers like Amazon. Again, these advantages largely flow to consumers—particularly low-income ones.

The proposed American Innovation and Choice Online Act (AICOA), which was voted out of the Senate Judiciary Committee in February and may make an appearance on the Senate floor this summer, threatens those consumer benefits. AICOA would prohibit so-called “self-preferencing” by Amazon and other large technology platforms.

Should a ban on self-preferencing come to fruition, Amazon would not be able to prominently show its own products in any capacity—even when its products are a good match for a consumer’s search. In search results, Amazon will not be able to promote its private-label products, including Amazon Basics and 365 by Whole Foods, or products for which it is a first-party seller (i.e., a reseller of another company’s product). Amazon may also have to downgrade the ranking of popular products it sells, making them harder for consumers to find. Forcing Amazon to present offers that do not correspond to products consumers want to buy or are not a good value inflicts harm on all consumers but is particularly problematic for low-income consumers. All else equal, most consumers, especially low-income ones, obviously prefer cheaper products. It is important not to take that choice away from them.

Consider the case of orange juice, the product causing so much consternation for Mr. Pena. In a recent search on Amazon for a 59-ounce orange juice, as seen in the image below, the first four “organic” search results are SNAP-eligible, first-party, or private-label products sold by Amazon and ranging in price from $3.55 to $3.79. The next two results are from third-party sellers offering two 59-ounce bottles of orange juice at $38.99 and $84.54—more than five times the unit price offered by Amazon. By prohibiting self-preferencing, Amazon would be forced to promote products to consumers that are significantly more expensive and that are not SNAP-eligible. This increases costs directly for consumers who purchase more expensive products when cheaper alternatives are available but not presented. But it also increases costs indirectly by forcing consumers to search longer for better prices and SNAP-eligible products or by discouraging them from considering timesaving, online shopping altogether. Low-income families are least able to afford these increased costs.

The upshot is that antitrust populists are choosing to support (often well-off) small-business owners at the expense of vulnerable working people. Congress should not allow them to put the squeeze on low-income families. These families are already suffering due to record-high inflation—particularly for items that constitute the largest share of their expenditures, such as transportation and food. Proposed antitrust reforms such as AICOA and reinvigorated Robinson-Patman Act enforcement will only make it harder for low-income families to make ends meet.

[TOTM: The following is part of a digital symposium by TOTM guests and authors on Antitrust’s Uncertain Future: Visions of Competition in the New Regulatory Landscape. Information on the authors and the entire series of posts is available here.]

If S.2992—the American Innovation and Choice Online Act or AICOA—were to become law, it would be, at the very least, an incomplete law. By design—and not for good reason, but for political expediency—AICOA is riddled with intentional uncertainty. In theory, the law’s glaring definitional deficiencies are meant to be rectified by “expert” agencies (i.e., the DOJ and FTC) after passage. But in actuality, no such certainty would ever emerge, and the law would stand as a testament to the crass political machinations and absence of rigor that undergird it. Among many other troubling outcomes, this is what the future under AICOA would hold.

Two months ago, the American Bar Association’s (ABA) Antitrust Section published a searing critique of AICOA in which it denounced the bill for being poorly written, vague, and departing from established antitrust-law principles. As Lazar Radic and I discussed in a previous post, what made the ABA’s letter to Congress so eye-opening was that it was penned by a typically staid group with a reputation for independence, professionalism, and ideational heterogeneity.

One of the main issues the ABA flagged in its letter is that the introduction of vague new concepts—like “materially harm competition,” which does not exist anywhere in current antitrust law—into the antitrust mainstream will risk substantial legal uncertainty and produce swathes of unintended consequences.

According to some, however, the bill’s inherent uncertainty is a feature, not a bug. It leaves enough space for specialist agencies to define the precise meaning of key terms without unduly narrowing the scope of the bill ex ante.

In particular, supporters of the bill have pointed to the prospect of agency guidelines under the law to rescue it from the starkest of the fundamental issues identified by the ABA. Section 4 of AICOA requires the DOJ and FTC to issue “agency enforcement guidelines” no later than 270 days after the date of enactment:

outlining policies and practices relating to conduct that may materially harm competition under section 3(a), agency interpretations of the affirmative defenses under section 3(b), and policies for determining the appropriate amount of a civil penalty to be sought under section 3(c).

In pointing to the prospect of guidelines, however, supporters are inadvertently admitting defeat—and proving the ABA’s point: AICOA is not ready for prime time.

This thinking is misguided for at least three reasons:

Guidelines are not rules

As section 4(d) of AICOA recognizes, guidelines are emphatically nonbinding:

The joint guidelines issued under this section do not … operate to bind the Commission, Department of Justice, or any person, State, or locality to the approach recommended in the guidelines.

As such, the value of guidelines in dispelling legal uncertainty is modest, at best.

This is even more so in today’s highly politicized atmosphere, where guidelines can be withdrawn at the tip of the ballot (we’ve just seen the FTC rescind the Vertical Merger Guidelines it put in place less than a year ago). Given how politicized the issuing agencies themselves have become, it’s a virtual certainty that the guidelines produced in response to AICOA would be steeped in partisan politics and immediately changed with a change in administration, thus providing no more lasting legal certainty than speculation by a member of Congress.

Guidelines are not the appropriate tool to define novel concepts

Regardless of this political reality, however, the mixture of vagueness and novelty inherent in the key concepts that underpin the infringements and affirmative defenses under AICOA—such as “fairness,” “preferencing,” “materiality”, or the “intrinsic” value of a product—undermine the usefulness (and legitimacy) of guidelines.

Indeed, while laws are sometimes purposefully vague—operating as standards rather than prescriptive rules—to allow for more flexibility, the concepts introduced by AICOA don’t even offer any cognizable standards suitable for fine-tuning.

The operative terms of AICOA don’t have definitive meanings under antitrust law, either because they are wholly foreign to accepted antitrust law (as in the case of “self-preferencing”) or because the courts have never agreed on an accepted definition (as in the case of “fairness”). Nor are they technical standards, which are better left to specialized agencies rather than to legislators to define, such as in the case of, e.g., pollution (by contrast: what is the technical standard for “fairness”?).

Indeed, as Elyse Dorsey has noted, the only certainty that would emerge from this state of affairs is the certainty of pervasive rent-seeking by non-altruistic players seeking to define the rules in their favor.

As we’ve pointed out elsewhere, the purpose of guidelines is to reflect the state of the art in a certain area of antitrust law and not to push the accepted scope of knowledge and practice in a new direction. This not only overreaches the FTC’s and DOJ’s powers, but also risks galvanizing opposition from the courts, thereby undermining the utility of adopting guidelines in the first place.

Guidelines can’t fix a fundamentally flawed law

Expecting guidelines to provide sensible, administrable content for the bill sets the bar overly high for guidelines, and unduly low for AICOA.

The alleged harms at the heart of AICOA are foreign to antitrust law, and even to the economic underpinnings of competition policy more broadly. Indeed, as Sean Sullivan has pointed out, the law doesn’t even purport to define “harms,” but only serves to make specific conduct illegal:

Even if the conduct has no effect, it’s made illegal, unless an affirmative defense is raised. And the affirmative defense requires that it doesn’t ‘harm competition.’ But ‘harm competition’ is undefined…. You have to prove that harm doesn’t result, but it’s not really ever made clear what the harm is in the first place.”

“Self-preferencing” is not a competitive defect, and simply declaring it to be so does not make it one. As I’ve noted elsewhere:

The notion that platform entry into competition with edge providers is harmful to innovation is entirely speculative. Moreover, it is flatly contrary to a range of studies showing that the opposite is likely true…. The theory of vertical discrimination harm is at odds not only with this platform-specific empirical evidence, it is also contrary to the long-standing evidence on the welfare effects of vertical restraints more broadly …

… [M]andating openness is not without costs, most importantly in terms of the effective operation of the platform and its own incentives for innovation.

Asking agencies with an expertise in competition policy to enact economically sensible guidelines to direct enforcement against such conduct is a fool’s errand. It is a recipe for purely political legislation adopted by competition agencies that does nothing to further their competition missions.

AICOA’s Catch-22 Is Its Own Doing, and Will Be Its Downfall

AICOA’s Catch-22 is that, by making the law so vague that it needs enforcement guidelines to flesh it out, AICOA renders both itself and those same guidelines irrelevant and misses the point of both legal instruments.

Ultimately, guidelines cannot resolve the fundamental rule-of-law issues raised by the bill and highlighted by the ABA in its letter. To the contrary, they confirm the ABA’s concerns that AICOA is a poorly written and indeterminate bill. Further, the contentious elements of the bill that need clarification are inherently legislative ones that—paradoxically—shouldn’t be left to competition-agency guidelines to elucidate.

The upshot is that any future under AICOA will be one marked by endless uncertainty and the extreme politicization of both competition policy and the agencies that enforce it.

[TOTM: The following is part of a digital symposium by TOTM guests and authors on Antitrust’s Uncertain Future: Visions of Competition in the New Regulatory Landscape. Information on the authors and the entire series of posts is available here.]

Earlier this month, Professors Fiona Scott Morton, Steve Salop, and David Dinielli penned a letter expressing their “strong support” for the proposed American Innovation and Choice Online Act (AICOA). In the letter, the professors address criticisms of AICOA and urge its approval, despite possible imperfections.

“Perhaps this bill could be made better if we lived in a perfect world,” the professors write, “[b]ut we believe the perfect should not be the enemy of the good, especially when change is so urgently needed.”

The problem is that the professors and other supporters of AICOA have shown neither that “change is so urgently needed” nor that the proposed law is, in fact, “good.”

Is Change ‘Urgently Needed’?

With respect to the purported urgency that warrants passage of a concededly imperfect bill, the letter authors assert two points. First, they claim that AICOA’s targets—Google, Apple, Facebook, Amazon, and Microsoft (collectively, GAFAM)—“serve as the essential gatekeepers of economic, social, and political activity on the internet.” It is thus appropriate, they say, to amend the antitrust laws to do something they have never before done: saddle a handful of identified firms with special regulatory duties.

But is this oft-repeated claim about “gatekeeper” status true? The label conjures up the old Terminal Railroad case, where a group of firms controlled the only bridges over the Mississippi River at St. Louis. Freighters had no choice but to utilize their services. Do the GAFAM firms really play a similar role with respect to “economic, social, and political activity on the internet”? Hardly.

With respect to economic activity, Amazon may be a huge player, but it still accounts for only 39.5% of U.S. ecommerce sales—and far less of retail sales overall. Consumers have gobs of other ecommerce options, and so do third-party merchants, which may sell their wares using Shopify, Ebay, Walmart, Etsy, numerous other ecommerce platforms, or their own websites.

For social activity on the internet, consumers need not rely on Facebook and Instagram. They can connect with others via Snapchat, Reddit, Pinterest, TikTok, Twitter, and scores of other sites. To be sure, all these services have different niches, but the letter authors’ claim that the GAFAM firms are “essential gatekeepers” of “social… activity on the internet” is spurious.

Nor are the firms singled out by AICOA essential gatekeepers of “political activity on the internet.” The proposed law touches neither Twitter, the primary hub of political activity on the internet, nor TikTok, which is increasingly used for political messaging.

The second argument the letter authors assert in support of their claim of urgency is that “[t]he decline of antitrust enforcement in the U.S. is well known, pervasive, and has left our jurisprudence unable to protect and maintain competitive markets.” In other words, contemporary antitrust standards are anemic and have led to a lack of market competition in the United States.

The evidence for this claim, which is increasingly parroted in the press and among the punditry, is weak. Proponents primarily point to studies showing:

  1. increasing industrial concentration;
  2. higher markups on goods and services since 1980;
  3. a declining share of surplus going to labor, which could indicate monopsony power in labor markets; and
  4. a reduction in startup activity, suggesting diminished innovation. 

Examined closely, however, those studies fail to establish a domestic market power crisis.

Industrial concentration has little to do with market power in actual markets. Indeed, research suggests that, while industries may be consolidating at the national level, competition at the market (local) level is increasing, as more efficient national firms open more competitive outlets in local markets. As Geoff Manne sums up this research:

Most recently, several working papers looking at the data on concentration in detail and attempting to identify the likely cause for the observed data, show precisely the opposite relationship. The reason for increased concentration appears to be technological, not anticompetitive. And, as might be expected from that cause, its effects are beneficial. Indeed, the story is both intuitive and positive.

What’s more, while national concentration does appear to be increasing in some sectors of the economy, it’s not actually so clear that the same is true for local concentration — which is often the relevant antitrust market.

With respect to the evidence on markups, the claim of a significant increase in the price-cost margin depends crucially on the measure of cost. The studies suggesting an increase in margins since 1980 use the “cost of goods sold” (COGS) metric, which excludes a firm’s management and marketing costs—both of which have become an increasingly significant portion of firms’ costs. Measuring costs using the “operating expenses” (OPEX) metric, which includes management and marketing costs, reveals that public-company markups increased only modestly since the 1980s and that the increase was within historical variation. (It is also likely that increased markups since 1980 reflect firms’ more extensive use of technology and their greater regulatory burdens, both of which raise fixed costs and require higher markups over marginal cost.)

As for the declining labor share, that dynamic is occurring globally. Indeed, the decline in the labor share in the United States has been less severe than in Japan, Canada, Italy, France, Germany, China, Mexico, and Poland, suggesting that anemic U.S. antitrust enforcement is not to blame. (A reduction in the relative productivity of labor is a more likely culprit.)

Finally, the claim of reduced startup activity is unfounded. In its report on competition in digital markets, the U.S. House Judiciary Committee asserted that, since the advent of the major digital platforms:

  1. “[t]he number of new technology firms in the digital economy has declined”;
  2. “the entrepreneurship rate—the share of startups and young firms in the [high technology] industry as a whole—has also fallen significantly”; and
  3. “[u]nsurprisingly, there has also been a sharp reduction in early-stage funding for technology startups.” (pp. 46-47.)

Those claims, however, are based on cherry-picked evidence.

In support of the first two, the Judiciary Committee report cited a study based on data ending in 2011. As Benedict Evans has observed, “standard industry data shows that startup investment rounds have actually risen at least 4x since then.”

In support of the third claim, the report cited statistics from an article noting that the number and aggregate size of the very smallest venture capital deals—those under $1 million—fell between 2014 and 2018 (after growing substantially from 2008 to 2014). The Judiciary Committee report failed to note, however, the cited article’s observation that small venture deals ($1 million to $5 million) had not dropped and that larger venture deals (greater than $5 million) had grown substantially during the same time period. Nor did the report acknowledge that venture-capital funding has continued to increase since 2018.

Finally, there is also reason to think that AICOA’s passage would harm, not help, the startup environment:

AICOA doesn’t directly restrict startup acquisitions, but the activities it would restrict most certainly do dramatically affect the incentives that drive many startup acquisitions. If a platform is prohibited from engaging in cross-platform integration of acquired technologies, or if it can’t monetize its purchase by prioritizing its own technology, it may lose the motivation to make a purchase in the first place.

Despite the letter authors’ claims, neither a paucity of avenues for “economic, social, and political activity on the internet” nor the general state of market competition in the United States establishes an “urgent need” to re-write the antitrust laws to saddle a small group of firms with unprecedented legal obligations.

Is the Vagueness of AICOA’s Primary Legal Standard a Feature?

AICOA bars covered platforms from engaging in three broad classes of conduct (self-preferencing, discrimination among business users, and limiting business users’ ability to compete) where the behavior at issue would “materially harm competition.” It then forbids several specific business practices, but allows the defendant to avoid liability by proving that their use of the practice would not cause a “material harm to competition.”

Critics have argued that “material harm to competition”—a standard that is not used elsewhere in the antitrust laws—is too indeterminate to provide business planners and adjudicators with adequate guidance. The authors of the pro-AICOA letter, however, maintain that this “different language is a feature, not a bug.”

That is so, the letter authors say, because the language effectively signals to courts and policymakers that antitrust should prohibit more conduct. They explain:

To clarify to courts and policymakers that Congress wants something different (and stronger), new terminology is required. The bill’s language would open up a new space and move beyond the standards imposed by the Sherman Act, which has not effectively policed digital platforms.

Putting aside the weakness of the letter authors’ premise (i.e., that Sherman Act standards have proven ineffective), the legislative strategy they advocate—obliquely signal that you want “change” without saying what it should consist of—is irresponsible and risky.

The letter authors assert two reasons Congress should not worry about enacting a liability standard that has no settled meaning. One is that:

[t]he same judges who are called upon to render decisions under the existing, insufficient, antitrust regime, will also be called upon to render decisions under the new law. They will be the same people with the same worldview.

It is thus unlikely that “outcomes under the new law would veer drastically away from past understandings of core concepts….”

But this claim undermines the argument that a new standard is needed to get the courts to do “something different” and “move beyond the standards imposed by the Sherman Act.” If we don’t need to worry about an adverse outcome from a novel, ill-defined standard because courts are just going to continue applying the standard they’re familiar with, then what’s the point of changing the standard?

A second reason not to worry about the lack of clarity on AICOA’s key liability standard, the letter authors say, is that federal enforcers will define it:

The new law would mandate that the [Federal Trade Commission and the Antitrust Division of the U.S. Department of Justice], the two expert agencies in the area of competition, together create guidelines to help courts interpret the law. Any uncertainty about the meaning of words like ‘competition’ will be resolved in those guidelines and over time with the development of caselaw.

This is no doubt music to the ears of members of Congress, who love to get credit for “doing something” legislatively, while leaving the details to an agency so that they can avoid accountability if things turn out poorly. Indeed, the letter authors explicitly play upon legislators’ unwholesome desire for credit-sans-accountability. They emphasize that “[t]he agencies must [create and] update the guidelines periodically. Congress doesn’t have to do much of anything very specific other than approve budgets; it certainly has no obligation to enact any new laws, let alone amend them.”

AICOA does not, however, confer rulemaking authority on the agencies; it merely directs them to create and periodically update “agency enforcement guidelines” and “agency interpretations” of certain affirmative defenses. Those guidelines and interpretations would not bind courts, which would be free to interpret AICOA’s new standard differently. The letter authors presume that courts would defer to the agencies’ interpretation of the vague standard, and they probably would. But that raises other problems.

For one thing, it reduces certainty, which is likely to chill innovation. Giving the enforcement agencies de facto power to determine and redetermine what behaviors “would materially harm competition” means that the rules are never settled. Administrations differ markedly in their views about what the antitrust laws should forbid, so business planners could never be certain that a product feature or revenue model that is legal today will not be deemed to “materially harm competition” by a future administration with greater solicitude for small rivals and upstarts. Such uncertainty will hinder investment in novel products, services, and business models.

Consider, for example, Google’s investment in the Android mobile operating system. Google makes money from Android—which it licenses to device manufacturers for free—by ensuring that Google’s revenue-generating services (e.g., its search engine and browser) are strongly preferenced on Android products. One administration might believe that this is a procompetitive arrangement, as it creates a different revenue model for mobile operating systems (as opposed to Apple’s generation of revenue from hardware sales), resulting in both increased choice and lower prices for consumers. A subsequent administration might conclude that the arrangement materially harms competition by making it harder for rival search engines and web browsers to gain market share. It would make scant sense for a covered platform to make an investment like Google did with Android if its underlying business model could be upended by a new administration with de facto power to rewrite the law.

A second problem with having the enforcement agencies determine and redetermine what covered platforms may do is that it effectively transforms the agencies from law enforcers into sectoral regulators. Indeed, the letter authors agree that “the ability of expert agencies to incorporate additional protections in the guidelines” means that “the bill is not a pure antitrust law but also safeguards other benefits to consumers.” They tout that “the complementarity between consumer protection and competition can be addressed in the guidelines.”

Of course, to the extent that the enforcement guidelines address concerns besides competition, they will be less useful for interpreting AICOA’s “material harm to competition” standard; they might deem a practice suspect on non-competition grounds. Moreover, it is questionable whether creating a sectoral regulator for five widely diverse firms is a good idea. The history of sectoral regulation is littered with examples of agency capture, rent-seeking, and other public-choice concerns. At a minimum, Congress should carefully examine the potential downsides of sectoral regulation, install protections to mitigate those downsides, and explicitly establish the sectoral regulator.

Will AICOA Break Popular Products and Services?

Many popular offerings by the platforms covered by AICOA involve self-preferencing, discrimination among business users, or one of the other behaviors the bill presumptively bans. Pre-installation of iPhone apps and services like Siri, for example, involves self-preferencing or discrimination among business users of Apple’s iOS platform. But iPhone consumers value having a mobile device that offers extensive services right out of the box. Consumers love that Google’s search result for an establishment offers directions to the place, which involves the preferencing of Google Maps. And consumers positively adore Amazon Prime, which can provide free expedited delivery because Amazon conditions Prime designation on a third-party seller’s use of Amazon’s efficient, reliable “Fulfillment by Amazon” service—something Amazon could not do under AICOA.

The authors of the pro-AICOA letter insist that the law will not ban attractive product features like these. AICOA, they say:

provides a powerful defense that forecloses any thoughtful concern of this sort: conduct otherwise banned under the bill is permitted if it would ‘maintain or substantially enhance the core functionality of the covered platform.’

But the authors’ confidence that this affirmative defense will adequately protect popular offerings is misplaced. The defense is narrow and difficult to mount.

First, it immunizes only those behaviors that maintain or substantially enhance the “core” functionality of the covered platform. Courts would rightly interpret AICOA to give effect to that otherwise unnecessary word, which dictionaries define as “the central or most important part of something.” Accordingly, any self-preferencing, discrimination, or other presumptively illicit behavior that enhances a covered platform’s service but not its “central or most important” functions is not even a candidate for the defense.

Even if a covered platform could establish that a challenged practice would maintain or substantially enhance the platform’s core functionality, it would also have to prove that the conduct was “narrowly tailored” and “reasonably necessary” to achieve the desired end, and, for many behaviors, the “le[ast] discriminatory means” of doing so. That is a remarkably heavy burden, and it beggars belief to suppose that business planners considering novel offerings involving self-preferencing, discrimination, or some other presumptively illicit conduct would feel confident that they could make the required showing. It is likely, then, that AICOA would break existing products and services and discourage future innovation.

Of course, Congress could mitigate this concern by specifying that AICOA does not preclude certain things, such as pre-installed apps or consumer-friendly search results. But the legislation would then lose the support of the many interest groups who want the law to preclude various popular offerings that its text would now forbid. Unlike consumers, who are widely dispersed and difficult to organize, the groups and competitors that would benefit from things like stripped-down smartphones, map-free search results, and Prime-less Amazon are effective lobbyists.

Should the US Follow Europe?

Having responded to criticisms of AICOA, the authors of the pro-AICOA letter go on offense. They assert that enactment of the bill is needed to ensure that the United States doesn’t lose ground to Europe, both in regulatory leadership and in innovation. Observing that the European Union’s Digital Markets Act (DMA) has just become law, the authors write that:

[w]ithout [AICOA], the role of protecting competition and innovation in the digital sector outside China will be left primarily to the European Union, abrogating U.S. leadership in this sector.

Moreover, if Europe implements its DMA and the United States does not adopt AICOA, the authors claim:

the center of gravity for innovation and entrepreneurship [could] shift from the U.S. to Europe, where the DMA would offer greater protections to start ups and app developers, and even makers and artisans, against exclusionary conduct by the gatekeeper platforms.

Implicit in the argument that AICOA is needed to maintain America’s regulatory leadership is the assumption that to lead in regulatory policy is to have the most restrictive rules. The most restrictive regulator will necessarily be the “leader” in the sense that it will be the one with the most control over regulated firms. But leading in the sense of optimizing outcomes and thereby serving as a model for other jurisdictions entails crafting the best policies—those that minimize the aggregate social losses from wrongly permitting bad behavior, wrongly condemning good behavior, and determining whether conduct is allowed or forbidden (i.e., those that “minimize the sum of error and decision costs”). Rarely is the most restrictive regulatory regime the one that optimizes outcomes, and as I have elsewhere explained, the rules set forth in the DMA hardly seem calibrated to do so.

As for “innovation and entrepreneurship” in the technological arena, it would be a seismic shift indeed if the center of gravity were to migrate to Europe, which is currently home to zero of the top 20 global tech companies. (The United States hosts 12; China, eight.)

It seems implausible, though, that imposing a bunch of restrictions on large tech companies that have significant resources for innovation and are scrambling to enter each other’s markets will enhance, rather than retard, innovation. The self-preferencing bans in AICOA and DMA, for example, would prevent Apple from developing its own search engine to compete with Google, as it has apparently contemplated. Why would Apple develop its own search engine if it couldn’t preference it on iPhones and iPads? And why would Google have started its shopping service to compete with Amazon if it couldn’t preference Google Shopping in search results? And why would any platform continually improve to gain more users as it neared the thresholds for enhanced duties under DMA or AICOA? It seems more likely that the DMA/AICOA approach will hinder, rather than spur, innovation.

At the very least, wouldn’t it be prudent to wait and see whether DMA leads to a flourishing of innovation and entrepreneurship in Europe before jumping on the European bandwagon? After all, technological innovations that occur in Europe won’t be available only to Europeans. Just as Europeans benefit from innovation by U.S. firms, American consumers will be able to reap the benefits of any DMA-inspired innovation occurring in Europe. Moreover, if DMA indeed furthers innovation by making it easier for entrants to gain footing, even American technology firms could benefit from the law by launching their products in Europe. There’s no reason for the tech sector to move to Europe to take advantage of a small-business-protective European law.

In fact, the optimal outcome might be to have one jurisdiction in which major tech platforms are free to innovate, enter each other’s markets via self-preferencing, etc. (the United States, under current law) and another that is more protective of upstart businesses that use the platforms (Europe under DMA). The former jurisdiction would create favorable conditions for platform innovation and inter-platform competition; the latter might enhance innovation among businesses that rely on the platforms. Consumers in each jurisdiction, however, would benefit from innovation facilitated by the other.

It makes little sense, then, for the United States to rush to adopt European-style regulation. DMA is a radical experiment. Regulatory history suggests that the sort of restrictiveness it imposes retards, rather than furthers, innovation. But in the unlikely event that things turn out differently this time, little harm would result from waiting to see DMA’s benefits before implementing its restrictive approach. 

Does AICOA Threaten Platforms’ Ability to Moderate Content and Police Disinformation?

The authors of the pro-AICOA letter conclude by addressing the concern that AICOA “will inadvertently make content moderation difficult because some of the prohibitions could be read… to cover and therefore prohibit some varieties of content moderation” by covered platforms.

The letter authors say that a reading of AICOA to prohibit content moderation is “strained.” They maintain that the act’s requirement of “competitive harm” would prevent imposition of liability based on content moderation and that the act is “plainly not intended to cover” instances of “purported censorship.” They further contend that the risk of judicial misconstrual exists with all proposed laws and therefore should not be a sufficient reason to oppose AICOA.

Each of these points is weak. Section 3(a)(3) of AICOA makes it unlawful for a covered platform to “discriminate in the application or enforcement of the terms of service of the covered platform among similarly situated business users in a manner that would materially harm competition.” It is hardly “strained” to reason that this provision is violated when, say, Google’s YouTube selectively demonetizes a business user for content that Google deems harmful or misleading. Or when Apple removes Parler, but not every other violator of service terms, from its App Store. Such conduct could “materially harm competition” by impeding the de-platformed business’ ability to compete with its rivals.

And it is hard to say that AICOA is “plainly not intended” to forbid these acts when a key supporting senator touted the bill as a means of policing content moderation and observed during markup that it would “make some positive improvement on the problem of censorship” (i.e., content moderation) because “it would provide protections to content providers, to businesses that are discriminated against because of the content of what they produce.”

At a minimum, we should expect some state attorneys general to try to use the law to police content moderation they disfavor, and the mere prospect of such legal action could chill anti-disinformation efforts and other forms of content moderation.

Of course, there’s a simple way for Congress to eliminate the risk of what the letter authors deem judicial misconstrual: It could clarify that AICOA’s prohibitions do not cover good-faith efforts to moderate content or police disinformation. Such clarification, however, would kill the bill, as several Republican legislators are supporting the act because it restricts content moderation.

The risk of judicial misconstrual with AICOA, then, is not the sort that exists with “any law, new or old,” as the letter authors contend. “Normal” misconstrual risk exists when legislators try to be clear about their intentions but, because language has its limits, some vagueness or ambiguity persists. AICOA’s architects have deliberately obscured their intentions in order to cobble together enough supporters to get the bill across the finish line.

The one thing that all AICOA supporters can agree on is that they deserve credit for “doing something” about Big Tech. If the law is construed in a way they disfavor, they can always act shocked and blame rogue courts. That’s shoddy, cynical lawmaking.

Conclusion

So, I respectfully disagree with Professors Scott Morton, Salop, and Dinielli on AICOA. There is no urgent need to pass the bill right now, especially as we are on the cusp of seeing an AICOA-like regime put to the test. The bill’s central liability standard is overly vague, and its plain terms would break popular products and services and thwart future innovation. The United States should equate regulatory leadership with the best, not the most restrictive, policies. And Congress should thoroughly debate and clarify its intentions on content moderation before enacting legislation that could upend the status quo on that important matter.

For all these reasons, Congress should reject AICOA. And for the same reasons, a future in which AICOA is adopted is extremely unlikely to resemble the Utopian world that Professors Scott Morton, Salop, and Dinielli imagine.

[On Monday, June 27, Concurrences hosted a conference on the Rulemaking Authority of the Federal Trade Commission. This conference featured the work of contributors to a new book on the subject edited by Professor Dan Crane. Several of these authors have previously contributed to the Truth on the Market FTC UMC Symposium. We are pleased to be able to share with you excerpts or condensed versions of chapters from this book prepared by authors of of those chapters. Our thanks and compliments to Dan and Concurrences for bringing together an outstanding event and set of contributors and for supporting our sharing them with you here.]

[The post below was authored by former Federal Trade Commission Acting Chair Maureen K. Ohlhausen and former FTC Senior Attorney Ben Rossen.]

Introduction

The Federal Trade Commission (FTC) has long steered the direction of competition law by engaging in case-by-case enforcement of the FTC Act’s prohibition on unfair methods of competition (UMC). Recently, some have argued that the FTC’s exclusive reliance on case-by-case adjudication is too long and arduous a route and have urged the commission to take a shortcut by invoking its purported authority to promulgate UMC rules under Section 6(g) of the Federal Trade Commission Act.

Proponents of UMC rulemaking rely on National Petroleum Refiners Association v. FTC, a 1973 decision by the U.S. Court of Appeals for the D.C. Circuit that upheld the commission’s authority to issue broad legislative rules under the FTC Act. They argue that the case provides a clear path to UMC rules and that Congress effectively ratified the D.C. Circuit’s decision when it enacted detailed rulemaking procedures governing unfair or deceptive acts or practices (UDAP) in the Magnuson Moss Warranty-Federal Trade Commission Improvement Act of 1975 (Magnuson-Moss).

The premise of this argument is fundamentally incorrect, because modern courts reject the type of permissive statutory analysis applied in National Petroleum Refiners. Moreover, contemporaneous congressional reaction to National Petroleum Refiners was not to embrace broad FTC rulemaking, but rather to put in strong guardrails on FTC UDAP rulemaking. Further, the congressional history of the particular FTC rule at issue—the Octane Ratings Rule—also points in the direction of a lack of broad UMC rulemaking, as Congress eventually adopted the rule solely as a UDAP provision, with heightened restrictions on FTC rulemaking.

Thus, the road to UMC rulemaking, which the agency wisely never tried to travel down in the almost 50 years since National Petroleum Refiners, is essentially a dead end. If the agency tries to go that route, it will be an unfortunate detour from its clear statutory direction to engage in case-by-case enforcement of Section 5.

Broad UMC-Rulemaking Authority Contradicts the History and Evolution of the FTC’s Authority

The FTC Act grants the commission broad authority to investigate unfair methods of competition and unfair and deceptive acts or practices across much of the American economy. The FTC’s administrative adjudicative authority under “Part 3” is central to the FTC’s mission of preserving fair competition and protecting consumers, as reflected by the comprehensive adjudicative framework established in Section 5 of the FTC Act. Section 6, meanwhile, details the commission’s investigative powers to collect confidential business information and conduct industry studies.

The original FTC Act contained only one sentence describing the agency’s ability to make rules, buried inconspicuously among various other provisions. Section 6(g) provided that the FTC would have authority “[f]rom time to time [to] classify corporations and . . . to make rules and regulations for the purpose of carrying out the provisions of this [Act].”[1] Unlike the detailed administrative scheme in Section 5, the FTC Act fails to provide for any sanctions for violations of rules promulgated under Section 6 or to otherwise specify that such rules would carry the force of law. This minimal delegation of power arguably conferred the right to issue procedural but not substantive rules.

Consistent with the understanding that Congress did not authorize substantive rulemaking, the FTC made no attempt to promulgate rules with the force of law for nearly 50 years after it was created, and at various times indicated that it lacked the authority to do so.

In 1962, the agency for the first time began to promulgate consumer-protection trade-regulation rules (TRRs), citing its authority under Section 6(g). Although these early TRRs plainly addressed consumer-protection matters, the agency frequently described violations of the rule as both an unfair method of competition and an unfair or deceptive trade practice. As the commission itself has observed, “[n]early all of the rules that the Commission actually promulgated under Section 6(g) were consumer protection rules.”

In fact, in the more than 100 years of the FTC Act, the agency has only once issued a solely competition rule. In 1967, the commission promulgated the Men and Boys’ Tailored Clothing Rule pursuant to authority under the Clayton Act, which prohibited apparel suppliers from granting discriminatory-advertising allowances that limited small retailers’ ability to compete. However, the rule was never enforced or subject to challenge and was subsequently repealed.

Soon after, the FTC promulgated the octane-ratings rule at issue in National Petroleum Refiners. Proponents of UMC rulemaking, such as former FTC Commissioner Rohit Chopra and current Chair Lina Khan, point to the case as evidence that the commission retains the power to promulgate substantive competition rules, governed only by the Administrative Procedure Act (APA) and, with respect to interpretations of UMC, entitled to Chevron deference. They argue that UMC rulemaking would provide significant benefits by providing clear notice to market participants about what the law requires, relieving the steep expert costs and prolonged trials common to antitrust adjudications, and fostering a “transparent and participatory process” that would provide meaningful public participation.

With Khan at the helm of the FTC, the agency has already begun to pave the way for new UMC rulemakings. For example, President Joe Biden’s Executive Order on promoting competition called on the commission to promulgate UMC rules to address noncompete clauses and pay-for-delay settlements, among other issues. Further, as one of Khan’s first actions as chair, the commission rescinded—without replacing—its bipartisan Statement of Enforcement Principles Regarding “Unfair Methods of Competition” Under Section 5 of the FTC Act. More recently, the commission’s Statement of Regulatory Priorities stated that the FTC “will consider developing both unfair-methods-of-competition rulemakings as well as rulemakings to define with specificity unfair or deceptive acts or practices.” This foray into UMC rulemaking is likely to take the FTC down a dead-end road.

The Signs Are Clear: National Petroleum Refiners Does Not Comport with Modern Principles of Statutory Interpretation

The FTC’s authority to conduct rulemaking under Section 6(g) has been tested in court only once, in National Petroleum Refiners, where the D.C. Circuit upheld the commission’s authority to promulgate a UDAP and UMC rule requiring the disclosure of octane ratings on gasoline pumps. The court found that Section 6(g) “clearly states that the Commission ‘may’ make rules and regulations for the purpose of carrying out the provisions of Section 5” and liberally construed the term ‘rules and regulations’ based on the background and purpose of the FTC Act.” The court’s opinion rested, in part, on pragmatic concerns about the benefits that rulemaking provides to fulfilling the agency’s mission, emphasizing the “invaluable resource-saving flexibility” it provides and extolling the benefits of rulemaking over case-by-case adjudication when developing agency policy.

National Petroleum Refiners reads today like an anachronism. Few modern courts would agree that an ambiguous grant of rulemaking authority should be construed to give agencies the broadest possible powers so that they will have flexibility in determining how to effectuate their statutory mandates. The Supreme Court has never adopted this approach and recent decisions strongly suggest it would decline to do so if presented the opportunity.

The D.C. Circuit’s opinion is in clear tension with the “elephants-in-mouseholes” doctrine first described by the U.S. Supreme Court in Whitman v. Am. Trucking Ass’n, because it largely ignored the significance of the FTC Act’s detailed adjudicative framework. The D.C. Circuit’s reasoning—that Congress buried sweeping legislative-rulemaking authority in a vague, ancillary provision, alongside the ability to “classify corporations”—stands in direct conflict with the Supreme Court’s admonition in Whitman.

Modern courts would also look to interpret the structure of the FTC Act to produce a coherent enforcement scheme. For instance, in AMG Capital Management v. FTC, the Supreme Court struck down the FTC’s use of Section 13(b) to obtain equitable monetary relief, in part, because the FTC Act elsewhere imposes specific limitations on the commission’s authority to obtain monetary relief. Unlike National Petroleum Refiners, which lauded the benefits and efficiencies of rulemaking for the agency’s mission, the AMG court reasoned: “Our task here is not to decide whether [the FTC’s] substitution of § 13(b) for the administrative procedure contained in § 5 and the consumer redress available under § 19 is desirable. Rather, it is to answer a more purely legal question” of whether Congress granted authority or not. The same rationale applies to UMC rulemaking.

The unanimous AMG decision was no judicial detour, and the Supreme Court has routinely posted clear road signs that Congress is expected “to speak clearly when authorizing an agency to exercise powers of vast economic and political significance,” as UMC rulemaking would do. Since 2000, the Court has increasingly applied the “major questions doctrine” to limit the scope of congressional delegation to the administrative state in areas of major political or economic importance. For example, in FDA v. Brown & Williamson, the Supreme Court declined to grant Chevron deference to an FDA rule permitting the agency to regulate nicotine and cigarettes. Crucial to the Court’s analysis was that the FDA’s rule contradicted the agency’s own view of its authority dating back to 1914, while asserting jurisdiction over a significant portion of the American economy. In Utility Air Regulatory Group v. EPA, the Court invoked the major questions doctrine to strike down the Environmental Protection Agency’s greenhouse-gas emissions standards as an impermissible interpretation of the Clean Air Act, finding that “EPA’s interpretation is [] unreasonable because it would bring about an enormous and transformative expansion in [the] EPA’s regulatory authority without clear congressional authorization.”

Most recently, in West Virginia v. EPAthe Court relied on the major questions doctrine to strike down EPA emissions rules that would have imposed billions of dollars in compliance costs on power plants, concluding that Congress had not provided “clear congressional authorization” for the rules despite explicitly authorizing the agency to set emissions levels for existing plants.  Because broad UMC-rulemaking authority under Section 6(g) is similarly a question of potentially “vast economic and political significance,” and would also represent a significant departure from past agency precedent, the FTC’s efforts to promulgate such rules would likely be met by a flashing red light.

Finally, while National Petroleum Refiners lauded the benefits of rulemaking authority and emphasized its usefulness for carrying out the FTC’s mission, the Supreme Court has since clarified that “[h]owever sensible (or not)” an interpretation may be, “a reviewing court’s task is to apply the text of the statute, not to improve upon it.” Whatever benefits rulemaking authority may confer on the FTC, they cannot justify departure from the text of the FTC Act.

The Road Not Taken: Congress Did Not Ratify UMC-Rulemaking Authority and the FTC Did Not Assert It

Two years after National Petroleum Refiners, Congress enacted the Magnuson-Moss Warranty-Federal Trade Commission Improvement Act of 1975 (Magnuson-Moss). Section 202(a) of Magnuson-Moss amended the FTC Act to add a new Section 18 that, for the first time, gave the FTC express authority to issue UDAP rules, while imposing heightened procedural requirements for such rulemaking. Magnuson-Moss does not expressly address UMC rulemaking. Instead, it says only that Section 18 “shall not affect any authority of the Commission to prescribe rules (including interpretive rules), and general statements of policy, with respect to unfair methods of competition in or affecting commerce.” Section 6(g) currently authorizes the FTC “(except as provided in [section 18] of this title) to make rules and regulations for the purpose of carrying out the provisions of this subchapter.”

UMC-rulemaking proponents argue Magnuson-Moss effectively ratified National Petroleum Refiners and affirmed the commission’s authority with respect to substantive UMC rules. This revisionist interpretation is incorrect. The savings provision in Section 18(a)(2) that preserves “any authority” (as opposed to “the” authority) of the commission to prescribe UMC rules reflects, at most, an agnostic view on whether the FTC, in fact, possesses such authority. Rather, it suggests that whatever authority may exist for UMC rulemaking was unchanged by Section 18 and that Congress left the question open for the courts to resolve. The FTC itself appears to have recognized this uncertainty, as evidenced by the fact that it has never even attempted to promulgate a UMC rule in the nearly 50 years following the enactment of Magnuson-Moss.

Congressional silence on UMC hardly endorses the commission’s authority and is not likely to persuade an appellate court today. To rely on congressional acquiescence to a judicial interpretation, there must be “overwhelming evidence” that Congress considered and rejected the “precise issue” before the court. Although Congress considered adopting National Petroleum Refiners, it ultimately took no action on the FTC’s UMC-rulemaking authority. Hardly the “overwhelming evidence” required to read National Petroleum Refiners into the law.

The Forgotten Journey: The History of the Octane-Ratings Rule Reinforces the FTC’s Lack of UMC Rulemaking Authority

Those who argue that National Petroleum Refiners is still good law and that Congress silently endorsed UMC rulemaking have shown no interest in how the journey of the octane-ratings rule eventually ended. The FTC’s 1971 octane-ratings rule declared the failure to post octane disclosures on gasoline pumps both an unfair method of competition and an unfair or deceptive practice. But what has remained unexplored in the debate over FTC UMC rulemaking is what happened to the rule after the D.C. Circuit’s decision upheld rulemaking under Section 6(g), and what that tells us about congressional and agency views on UMC authority.

The octane-ratings rule upheld by the D.C. Circuit never took effect and was ultimately replaced when Congress enacted the Petroleum Marketing Practices Act (PMPA), Title II of which addressed octane-disclosure requirements and directed the FTC to issue new rules under the PMPA. But despite previous claims by the FTC that the rule drew on both UDAP and UMC authority, Congress declined to provide any authority beyond UDAP. While it is impossible to say whether Congress concluded that UMC rulemaking was unwise, illegal, or simply unnecessary, the PMPA—passed just two years after Magnuson-Moss—suggests that UMC rulemaking did not survive the enactment of Section 18. A brief summary of the rule’s meandering journey follows.

After the D.C. Circuit remanded National Petroleum Refiners, the district court ordered the FTC to complete an environmental-impact statement. While that analysis was pending, Congress began consideration of the PMPA. After its enactment, the commission understood Congress to have intended the requirements of Title II of the PMPA to replace those of the original octane-ratings rule. The FTC treated the enactment of the PMPA as effectively repealing the rule.

Section 203(a) of the PMPA gave the FTC rulemaking power to enforce compliance with Title II of the PMPA. Testimony in House subcommittee hearings centered on whether the legislation should direct the FTC to enact a TRR on octane ratings under expedited procedures that would be authorized by the legislation, or whether Congress should enact its own statutory requirements. Ultimately, Congress adopted a statutory definition of octane ratings (identical to the method adopted by the FTC in its 1971 rule) and granted the FTC rulemaking authority under the APA to update definitions and prescribe different procedures for determining fuel-octane ratings. Congress also specified that certain rules—such as those requiring manufacturers to display octane requirements on motor vehicles—would have heightened rulemaking procedures, such as rulemaking on the record after a hearing.

Notably, the PMPA specifically provides that violations of the statute, or any rule promulgated under the statute, “shall be an unfair or deceptive act or practice in or affecting commerce.” Although Section 203(d)(3) of the PMPA specifically exempts the FTC from the procedural requirements under Section 18, it does not simply revert to Section 6(g) or otherwise leave open a path for UMC rulemaking.

The record makes clear, however, that Congress was aware of FTC’s desire to claim UMC authority in connection with the octane-ratings rule, as FTC officials testified in legislative hearings that UMC authority was necessary to regulate octane ratings. After Magnuson-Moss was enacted, however, neither Congress nor the FTC tried to include UMC rulemaking in the PMPA. In a written statement reflecting the FTC’s views on the PMPA incorporated in the House report, the FTC described its original octane-ratings rule as UDAP only.[2] While not dispositive, the FTC’s apparent abandonment of its request for UMC authority after Magnuson-Moss, and Congress’ decision to limit the PMPA exclusively to UDAP, certainly suggests that UMC did not survive National Petroleum Refiners and that Congress did not endorse FTC UMC rulemaking.

Conclusion

The FTC appears poised to embark on a journey of broad, legislative-style competition rulemaking under Section 6(g) of the FTC Act. This would be a dead end. UMC rulemaking, rather than advancing clarity and certainty about what types of conduct constitute unfair methods of competition, would very likely be viewed by the courts as an illegal left turn. It would also be a detour for the agency from its core mission of case-by-case expert adjudication of the FTC Act—which, given limited agency resources, could result in a years-long escapade that significantly detracts from overall enforcement. The FTC should instead seek to build on the considerable success it has seen in recent years with administrative adjudications, both in terms of winning on appeal and in shaping the development of antitrust law overall by creating citable precedent in key areas.


[1]     H. Rep. No. 95-161, at 45, Appendix II, Federal Trade Commission—Agency Views, Statement of Federal Trade Commission by Christian S. White, Asst. Director for Special Statutes (Feb. 23, 1977).

[2]     38 Stat. 722 § 6(g), codified as amended at 15 U.S.C. §  46(g).