Anthropic CEO Dario Amodei’s proposal to “pace the frontier” has become a rallying point for AI companies interested in coordinating a slowdown in capability development. His plan calls for embedded outside evaluators, common safety standards, and limits on the pace of model advances. Amodei also suggests that some form of antitrust waiver may be needed to permit certain safety discussions. 

As my colleague Dirk Auer explains, agreements among rivals to restrain development deserve scrutiny, even when made in the name of safety. Before granting an exemption, policymakers should ask where existing legal and market incentives—and the forms of cooperation antitrust law already allows—fall short. 

Civil and criminal laws already give AI labs reason to consider the consequences of their conduct, including safety risks. Market incentives point in the same direction. Customers favor products that work as intended and remain dependable. A model that enables mass hacking or causes other serious harm is, among other things, a very bad product. 

Companies can respond to these incentives on their own by strengthening safeguards, limiting model autonomy, or delaying releases. Existing antitrust law also permits competitors to collaborate on many safety matters and participate in carefully structured standard-setting efforts. 

Any proposed exemption therefore carries a heavy burden. Its proponents must show why unilateral action and lawful collaboration cannot address the problem—and identify precisely what conduct existing antitrust law prevents. 

Move Fast and Get Sued

AI-safety evaluator Model Evaluation and Threat Research’s (METR) investigation  into the Hugging Face incident found that separate agents used an unauthorized message board to coordinate a remote-code-execution attack, which allows an attacker to run commands on another computer. Hundreds of agents participated, even though the attack fell outside their assigned tasks. If developers and operators did not already feel pressure to build safeguards into new models, this incident should provide it. 

Labs that deploy agents with offensive cyber capabilities must make difficult choices about access limits, system isolation, monitoring, and when humans should intervene. Negligence law can make poor choices in these areas very costly. 

Of course, a negligence claim still requires proof of a legal duty, breach, causation, and recoverable harm. An agent’s harmful conduct does not automatically make its developer liable. Even so, the early stages of litigation can become a long and expensive headache for any firm. 

Nor does potential liability end with a lab’s direct actions. A company may face liability for creating a foreseeable downstream risk, even when a third party causes the immediate harm. In Weirum v. RKO General, for example, the California Supreme Court affirmed liability against a broadcaster whose contest predictably encouraged listeners to drive dangerously. The case did not involve artificial intelligence, but it illustrates how liability can rest on creating a foreseeable risk rather than directly inflicting the injury. For an AI lab, the central question may be whether developing or operating its system created an unreasonable risk of harm. 

Other civil claims and criminal charges apply different legal standards. The Computer Fraud and Abuse Act (CFAA), for example, covers specified forms of unauthorized computer access and provides a limited civil remedy. The statute bars civil claims based on negligent software design or manufacture. No court has yet decided how that exclusion applies when autonomous software agents commit the prohibited acts. 

The degree of “intentionality” that a court attributes to an agent could affect that analysis. So could a developer’s deliberate decision to equip a model with offensive cyber capabilities. A court might view such a case as involving something more than negligent software design. 

Criminal liability under the CFAA would require proof of a specific offense and the necessary mental state of a legally responsible person or entity. An agent’s apparent motivations do not, by themselves, establish a lab’s criminal intent. It also remains unclear whether an agent can possess legally cognizable intent and, if so, when that intent could be attributed to the lab that trained it or the user who deployed it. 

None of this means that OpenAI or any other lab should be sued, much less that it would clearly face liability under the CFAA, negligence law, or another legal theory. The point is that considerable uncertainty remains, along with plenty of room for creative lawyers to impose costly, time-consuming litigation on frontier AI labs. Those labs can reasonably account for that risk when deciding unilaterally whether to slow development. 

Development and deployment also create different risks. A lab might continue capability research while restricting an agent’s access to external systems. It might strengthen containment measures or postpone a release until testing is adequate. Each decision involves a separate judgment about the product’s risks and benefits. A race to develop more capable models does not force every participant to ignore those costs in pursuit of commercial advantage. 

Acing the Benchmark, Flunking the Job

An AI product’s quality depends on more than intelligence. Users expect an agent to respect permissions, protect confidential data, and perform assigned tasks reliably. Benchmark scores cannot establish whether it is fit for a particular job. An agent that aces every test but ignores its instructions is still a lousy employee. 

Competition law knows this territory well. Firms routinely compete through different combinations of performance, price, reliability, and other features. Antitrust law recognizes these dimensions of competition. Agency guidance on innovation, for example, expressly considers competition over the range of products and features available to consumers. Intelligence is similarly just one measure of competitive success in AI. 

Safety also has commercial value independent of any shared commitment to slow development. Widespread concern about AI safety is itself a strong signal that customers want safer products. One lab might attract customers with an agent that is easier to supervise or better suited to sensitive work. Another might offer greater autonomy paired with different controls. 

Competition lets customers compare these approaches and firms refine them through experience. Coordination around a single approach to AI safety can limit that experimentation, particularly when an agreed-upon standard excludes competing designs. 

Mark Zuckerberg’s recent discussion of Meta’s strategy for Muse illustrates the point. Zuckerberg said Meta delayed Muse’s launch to address safety concerns. Meta made that decision unilaterally as part of its own product-design and launch process, without imposing the same judgment on its competitors.

The delay does not prove that Meta ultimately met the appropriate safety threshold or that every lab faces comparable incentives. It does show that firms can delay a release to address safety concerns without first securing an antitrust exemption.  

No Antitrust Hall Pass

AI labs can already obtain safety information through several channels. Each firm can hire independent evaluators, commission audits, and consult outside experts. Anthropic’s unilateral pledge to embed independent evaluators shows that this part of Amodei’s proposal requires no industrywide speed limit. 

Standardized testing and certification could also help consumers distinguish among competing products. They could provide courts and regulators with evidence of an established, reasonably applied industry practice. Private certification bodies might assess compliance with specified benchmarks, much as UL Solutions tests and certifies product safety. Testing laboratories could also work through standard-setting forums to develop common protocols and technical specifications. 

The Federal Trade Commission (FTC) already recognizes that carefully structured cooperation among competitors can make markets more efficient and benefit consumers. That includes certain standard-setting activities. 

The legal details still matter. An industry standards body can promote safety, or it can become a club that keeps rivals out. A forum created to develop testing protocols can instead become a venue for sharing confidential business information or coordinating anticompetitive restraints. A safety label is not an antitrust hall pass. Certification programs and standards remain subject to antitrust scrutiny, and participating firms must act reasonably within them. 

The law also distinguishes among unilateral decisions, limited cooperation, and agreements that suppress competition. One firm may delay a launch. Several firms may collaborate on a specific evaluation. Competitors might instead agree to restrict the computing workloads used to train their models. The first two approaches preserve competition over capabilities. The third can suppress it. 

Advocates for an antitrust exemption should therefore identify the conduct they want to pursue that existing law prohibits. They should also explain why that protection is necessary to achieve safety benefits that firms cannot obtain through unilateral action or carefully structured cooperation. 

The Burden Before the Brake

Liability rules and customer demand cannot solve every safety problem. Customers may overlook harms imposed on outsiders. An incident may be difficult to detect or trace to a particular company. Catastrophic losses may also exceed a lab’s ability to compensate victims. These gaps can weaken incentives to take precautions and may justify targeted policy responses. 

The narrower point is that existing legal and market institutions already give firms independent reasons to invest in safer products. Firms also have several ways to act on those incentives. Anyone proposing coordinated restrictions as the cure must explain why they would close the remaining gaps more effectively than the available alternatives. 

As Dirk Auer noted, an antitrust exemption carries costs of its own. Past exemptions have encouraged stagnation and industry capture, allowing established firms to shape rules for their own benefit. Proponents must therefore identify the specific risk that unilateral action and lawful cooperation cannot adequately address, the precise restraint needed to address it, and why a less restrictive response would fail. 

Build the guardrails first. Give competitors a shared brake only after proving the need.

The post Tap Your Own Brakes: AI Safety and Antitrust appeared first on Truth on the Market.

The DSA designation puts ChatGPT alongside Google Search and Bing in the category reserved for the largest search engines operating in the EU. Under Article 33, a platform or search engine enters that category when it reaches at least 45 million average monthly active users in the EU. OpenAI reported that ChatGPT Search averaged about 159.1 million monthly active users in the EU during the six months ending March 31, comfortably clearing the threshold. ChatGPT now has four months to comply with the additional obligations the DSA imposes on services of that size. 

Those obligations require the largest services to examine their broader effects on society. Article 34 requires them to identify and assess “systemic risks” arising from their design or operation. These include risks to fundamental rights, civic discourse, public security, elections, public health, and children. Article 35 then requires services to adopt reasonable, proportionate, and effective measures to mitigate those risks. 

The DSA offers several examples, including changes to a service’s design, recommendation systems, terms and conditions, and content-moderation practices. Lawmakers largely developed these obligations for platforms and conventional search engines. A search engine generally organizes and presents information found elsewhere. ChatGPT, by contrast, generates the answer a user sees. 

That difference has practical consequences. Efforts to reduce systemic risks may extend beyond how a service ranks, recommends, or displays information. They may shape the substance of the answer itself. 

OpenAI has accepted the designation and emphasized its commitment to complying with the DSA while balancing access to useful information with protection from harmful and illegal content. Few would quarrel with that goal. The difficulty lies in deciding what the balance requires. 

ChatGPT’s designation therefore presents a broader problem. A regulatory framework designed to govern the distribution of online information may fit awkwardly when applied to a service that produces the answers users receive. 

When a Search Engine Talks Back

The DSA defines an “online search engine” broadly. Under Article 3(j), the term covers an intermediary service that lets users enter queries to search all websites or all websites in a particular language, then returns information related to the requested content. The definition turns on function and expressly permits results “in any format.” 

ChatGPT Search appears to qualify. It can search the internet for relevant information and generate a response based on what it finds. A conversational answer can fall within Article 3(j) as readily as a page of links. 

But this shared label covers materially different services. Google Search, Bing, and other conventional search engines identify, organize, and rank existing information, leaving users to choose among sources. ChatGPT retrieves information and composes the response. Article 3(j) can therefore cover both a service that points users toward answers and one that provides the answer itself. 

The Regulator’s Answer Key

The distinction changes how the DSA’s systemic-risk duties operate. Every ChatGPT response reflects choices about relevance, emphasis, qualifications, and the best answer to the user’s question. Information the model omits may disappear entirely instead of slipping lower on a results page. 

The DSA’s systemic-risk categories are deliberately broad. Protecting children and restricting illegal content may lend themselves to relatively identifiable measures. Other categories—including fundamental rights, civic discourse, elections, public health, and well-being—require judgments about matters on which reasonable people, institutions, and experts disagree. 

Conventional platforms can often mitigate such risks through changes to ranking, recommendation systems, visibility, or interface design. For ChatGPT, mitigation may come down to how a response to a user’s question is composed. If the Commission concludes that certain outputs contribute to a systemic risk, compliance may require changing how ChatGPT answers the underlying question. Displaying competing links offers little help when the user has asked for a synthesized answer. 

The DSA provides little guidance about when an answer, or a pattern of answers, creates a serious enough risk to justify intervention. Providers must weigh competing rights and interests that courts often assess on a case-by-case basis. Regulatory caution may then favor generic responses, broader refusals, or just fewer lawful answers. 

The Commission’s risk assessment can thus become an editorial judgment about what ChatGPT should say. The AI Act then adds a second, more specific set of rules for the same model. 

Brussels Doubles Up

The EU’s other rulebook for ChatGPT is already in force. The AI Act took effect Aug. 1, 2024, and became generally applicable Aug. 2, 2026. The law prohibits several defined practices. Article 5, for example, bars certain manipulative or deceptive techniques and the exploitation of vulnerabilities when they materially distort a person’s behavior and cause, or are reasonably likely to cause, significant harm. 

EU lawmakers designed the AI Act specifically for artificial intelligence and drew relatively clear boundaries around prohibited conduct. The law also imposes separate duties on providers of general-purpose AI models, which can perform a wide range of tasks. 

The DSA operates at a higher level of generality. Articles 34 and 35 require very large services to assess and mitigate broad categories of systemic risk, including risks to children, fundamental rights, and physical and mental well-being. The AI Act asks whether a practice falls within a defined prohibition. The DSA asks providers to predict and manage the wider effects of their services. 

Applying both regimes to ChatGPT creates uncertainty and duplication. Providers must satisfy the AI Act’s specific rules while anticipating the Commission’s evolving interpretation of systemic risk under the DSA. That uncertainty may encourage restrictions beyond what either law clearly requires. 

Users bear part of the cost. Providers seeking to limit their exposure may withhold lawful and useful answers, reduce personalization, or standardize responses. Those choices could gradually change the service itself and discourage innovation. 

When Classification Becomes Control

How far a law extends and whether its rules are suitable for specific types of conduct are separate questions. Article 3(j)’s broad definition can bring ChatGPT within the DSA. Applying duties designed for online intermediaries to a service that generates its own responses should require a separate justification. 

The AI Act already provides a detailed framework tailored to artificial intelligence. The Commission should explain what the DSA adds, define the scope of its additional obligations, and show how providers can comply while preserving lawful and useful responses. Vague and overlapping duties may encourage overcompliance, standardized answers, and less experimentation.

The Commission’s approach will shape which services developers offer in Europe. Users will see the consequences one answer at a time.

The post Brussels Writes the Answer Key appeared first on Truth on the Market.

Recital 79 even promises that the Commission will apply a “predefined standard” to identify unfair practices and practices that limit contestability. Anyone searching the DMA for that standard will come away empty-handed. 

In a new white paper, I reconstruct what I argue is nonetheless a fully coherent framework implicit in the enacted text. The DMA’s provisions and its subsequent enforcement record, along with various guidance documents issued along the way, do reveal specific conceptions of fairness and contestability, even if lawmakers declined to state them openly. Put simply: “fairness” refers to a presumed structural imbalance in digital markets, and “contestability” is the redistribution program enacted to correct it

That reconstruction has practical consequences. It explains how the Commission is likely to assess the DMA’s success—and its own performance. It also identifies the policy choices that countries outside the European Union may adopt, often without expressly debating them, when they enact similar laws.

When Entry Proves Entrenchment

Start with what the text actually says. The DMA defines both concepts only by negation. Unfairness “should relate to an imbalance between the rights and obligations of business users where the gatekeeper obtains a disproportionate advantage” (Recital 33). Contestability is lacking when a gatekeeper’s position is difficult to challenge “even by more innovative or efficient market operators” (Recital 3). Scholars have criticized the resulting vagueness ever since the DMA’s enactment. 

Now read between the lines. The DMA assumes that core platform markets are unfair by constitution, before any firm’s conduct enters the picture. It treats three market features as inherently suspect: network effects that make a service more valuable as more people use it, cost advantages derived from large datasets, and winner-take-most “tipping” dynamics. These features supposedly create imbalances that competition cannot correct.

That premise explains why designation depends on size alone, measured by market capitalization, turnover, and user numbers. It also explains why Recital 23 says efficiencies and market definition “should be discarded” during designation, and why conduct permitted for every other firm becomes unlawful once a company crosses the DMA’s gatekeeper thresholds. 

Two years of enforcement have confirmed this reading. Consider TikTok, the only successful large-scale entrant into a core platform market in the past decade. Entry and rapid growth ordinarily provide strong evidence that a market is contestable. Yet the Commission treated both as evidence against contestability. Its designation decision concluded that TikTok’s “significant scale and growth … with an upward trajectory” supported a finding of entrenchment. The General Court approved that reasoning in July 2024. 

The result initially seems puzzling. How can a new entrant already be entrenched? But under the DMA, contestability functions as a legal status attached to firms and inferred from their size and longevity. Economic theory, by contrast, treats contestability as a feature of markets.

William Baumol, who gave contestability its economic meaning, described markets in which incumbency offers no protection against entry. The DMA inverts his account. A successful entrant becomes an incumbent once it grows large enough, and the DMA treats incumbency itself as the problem. 

Fairness, Billed to the Gatekeeper

The DMA’s second implicit principle concerns what fairness requires. If the existing distribution of rights and obligations is unfair to business users, fairness requires redistribution. The DMA’s conception is dialectical in a specific sense. It levels up business users and rivals while leveling down gatekeepers until the gap narrows.

That may sound abstract, but the DMA’s data provisions show how it works. Some provisions restrict how gatekeepers collect and combine data, while others require them to give data to rivals and business users. The same input becomes a barrier to entry in a gatekeeper’s hands and a resource for innovation in a rival’s. The DMA also favors innovation by rivals over innovation by incumbents. That preference reflects a political judgment that economic analysis alone cannot establish. 

The Commission’s preliminary findings on Google’s search-data obligations formalize this logic through a newly minted “principle of parity.” Under that principle, Google parent Alphabet may retain no advantage from search data unless rivals receive equal access to it. The Commission now reads that group of rivals to include AI chatbots. 

In other words, Google must provide the query, click, view, and ranking data it uses to improve its own service. Rivals must receive that data at the same frequency, any fees must exclude collection costs, and smaller companies must receive access below cost. Yet nothing in Article 6(11), which requires access on “fair, reasonable and non-discriminatory terms,” requires parity. The parity principle follows instead from the DMA’s implicit conception of fairness. If possessing an advantage is itself unfair, fairness requires nullifying it.

The DMA also makes gatekeepers the custodians and financiers of their competitors. Under the Apple interoperability decision, Apple must design access for rivals into new features from inception, provide “adequate and timely assistance,” and bear the costs. The Commission says it need not show that these duties help rivals offer competitive products. Requiring such proof, the Commission argues, “would re-import the effects analysis, which the legislator explicitly rejected.”

The joint guidelines issued by the Commission and the European Data Protection Board (EDPB) under the DMA and General Data Protection Regulation (GDPR) praise data portability for turning gatekeepers into “a stable source of data” for new services. The DMA thus requires gatekeepers to continue subsidizing their rivals.

A Standard No Gatekeeper Can Satisfy

The reconstruction has practical consequences. Because the DMA presumes that these markets were never fair, there is no earlier competitive balance to restore and no baseline for measuring progress. The first infringement decisions define fairness and contestability negatively, through the elimination of listed practices. They offer nothing more. 

Consider the Apple anti-steering decision, which concerns restrictions on developers directing users to other purchasing options. The decision prohibits Apple from charging for “gatekeeper value,” meaning value derived from the platform’s position as an important gateway. Yet it never explains when a fee ceases to compensate Apple for its services and becomes a charge for gatekeeper value. 

The Google decisions from July 2026, imposing €890 million in fines across two infringements, use the same approach. The Apple interoperability decision goes even further, conceding that even full compliance would leave Apple with an “intrinsic advantage” because Apple alone decides which iOS features to build. 

Fairness, then, can never be achieved. An unattainable goal can justify intervention indefinitely. Gatekeepers know which practices the DMA prohibits. They cannot know when they have become fair enough.

The Only Good Gatekeeper Is a Smaller One

This brings us to the payoff. The DMA’s implicit principles determine how the Commission measures success. If fairness is a presumed structural imbalance and contestability is the program that corrects it, then the metric of success is not consumer prices, quality, or innovation. It is redistribution, as measured by inputs. 

The Commission’s own impact assessment established this approach during the DMA’s design. It projected benefits through a decline in the Herfindahl-Hirschman Index (HHI), a common measure of market concentration, and quantified those benefits only once, before implementation. The Commission proceeded despite its Regulatory Scrutiny Board’s objection that the impact assessment assumed the alleged harms without demonstrating them. 

The Commission’s first DMA review used the same criteria. It measures success by the uptake of nongatekeeper services, third-party access, and reduced “dependency” on gatekeepers. The review even counts users’ newfound ability to create Google accounts without Gmail addresses as progress, despite the fact that that option creates no new competition and offers no apparent consumer benefit. It merely reduces Gmail use, which is precisely the aim. 

The DMA’s implicit principles explain that choice. As the impact assessment describes it, unfairness stems from the size of covered companies, measured by user numbers, turnover, and market capitalization. Anything that reduces their size can therefore count as progress. 

Future scorecards and Article 18 market investigations into systematic noncompliance will likely use the same criteria. If gatekeepers’ market share goes down and rivals’ uptake goes up, then consumer benefits are assumed. The DMA’s recitals and accompanying documents acknowledge no tradeoffs in this outcome beyond costs to gatekeepers and possible short-term costs to consumers. 

The Commission can always attribute disappointing results to lingering structural defects, so even full compliance may trigger yet another intervention. As I have argued before, when consumers decline to switch, regulators invariably conclude that the remedies were too weak. 

Reaching this conclusion required neither leaked documents nor clairvoyance. It simply required reading the DMA and roughly 500 pages of accompanying documents and enforcement decisions. The law states its goals discreetly, as statutes often do when lawmakers anticipate resistance to a redistributive agenda. The DMA tells gatekeepers what to change. It never tells them when they have changed enough.

The post The DMA’s Magic Words and Missing Off Switch appeared first on Truth on the Market.

The question matters because chatbots have quickly become one of the most visible uses of artificial intelligence, and a growing focus of regulatory attention. Millions of people now use them to search for information, seek advice, work through problems, make decisions, and discuss personal matters. As those uses have expanded, scrutiny has shifted toward what chatbots say—and what happens when those conversations go wrong. 

Some of that scrutiny stems from cases in which chatbot interactions allegedly contributed to real harm, including harm to children. These cases raise legitimate questions about safety and responsibility. They also form part of a broader debate about conversational AI, including inaccurate or dangerous advice, systems that reinforce users’ existing beliefs, and responses that sound more understanding or authoritative than they really are. The concern, in short, centers less on the technology behind the chatbot than on the answers it gives. 

That shift creates a thornier problem. Most people agree that providers should take reasonable steps to prevent foreseeable harm. Deciding what those steps should be is harder when the product itself consists of a conversation. The most obvious safety measures change the chatbot’s answers: limiting the topics it will discuss, making its responses more cautious, or refusing certain conversations altogether. Yet much of conversational AI’s value lies in its ability to respond directly, personally, and dynamically to what a user asks. 

Responsibility for those interactions can be difficult to assign. A chatbot’s answer emerges from an exchange with the user, rather than in isolation. The provider controls the model and the conditions under which it operates, but the user shapes the conversation that produces a particular output.  

As liability comes to depend on what a chatbot should or should not have said, regulators and courts will have to distinguish responsibility for harmful outputs from control over the substance of AI responses. In drawing that line, they may move from deciding who bears responsibility for an answer to deciding what the “right” answer should have been. 

It Takes Two to Prompt

A chatbot does not normally produce an answer in isolation. The user chooses the question, supplies information, frames the problem, reacts to earlier responses, and can steer the conversation across dozens or hundreds of exchanges. The model responds to that context. Indeed, that responsiveness is central to the product’s appeal: Users want answers tailored to their questions and refined as the conversation develops, not the same generic response every time. 

As International Center for Law & Economics (ICLE) scholars have previously argued, AI companies participate in the marketplace of ideas by offering what they consider the best answers to users’ questions. Providers that consistently fall short risk losing users to competing chatbots. They therefore compete to produce answers that users find relevant and useful. A model that ignored user-supplied context would sacrifice much of what distinguishes conversational AI from traditional ways of finding information. 

That does not absolve AI companies of responsibility. The provider creates the system, while the user helps create the context that produces a particular answer. Their relative contributions may also shift during a conversation. A model may introduce an idea the user never suggested, while a user may repeatedly steer the exchange in a direction the provider never intended. 

AI can, of course, influence human behavior. But conversational AI is interactive by design, and that influence runs both ways. The model responds to the user, the user responds to the model, and each exchange supplies context for the next. It would be artificial, in a lengthy conversation, to treat the final response as wholly independent of everything that preceded it.

That dynamic complicates efforts at regulation. If some risk arises because the model adapts to the user, making it safer may require curbing the very responsiveness users value. A provider could make responses more generic, limit how closely the model follows a conversation’s direction, or intervene more readily when an exchange enters sensitive territory. Such measures may sometimes be justified, but they carry a cost: The less a chatbot can respond to the user and the context before it, the less conversational—and less useful—it becomes. 

Who Guards the Guardrails?

To the extent the concern is that chatbots may reinforce harmful ideas, the obvious response is to change how they engage with users. A model can challenge assumptions, decline to validate certain claims, offer another perspective, or simply end the conversation. Some cases are straightforward. Few would defend a chatbot that encourages suicide, dispenses dangerously false medical advice, or helps someone commit a crime. Beyond those cases, the line gets blurry fast.  

Regulators must also weigh these risks against AI’s considerable benefits. The same systems that produce harmful or inaccurate answers can help users digest large amounts of information, make better-informed decisions, automate routine tasks, and solve problems more efficiently. Those benefits do not excuse every harm, but risk is only half the ledger. Not every bad idea is dangerous, not every disputed claim is false, and not every conversation involving some amount of risk should necessarily be stopped.  

Drawing the line too broadly carries its own costs. Once the goal is to make answers safer, more balanced, or less misleading, someone must define what those standards mean in practice. Neutrality does not define itself, and an answer can be disputed without proving it false or harmful. Debates over the values reflected in large language model (LLM) outputs already illustrate the problem. Critics allege political bias in different directions, while providers make their own choices about how models should address contentious issues.  

The potential for intractable conflict becomes even clearer when governments join the conversation. A European Parliament briefing describes sharply different attempts to dictate what AI systems should say. In the United States, the Federal Trade Commission (FTC) has called for systems that do not manipulate responses in favor of ideological positions. Chinese rules, by contrast, require generative AI to conform to the government’s “core socialist values.” The approaches differ dramatically, but each reveals the same underlying problem. Once regulation reaches beyond clearly unlawful or harmful outputs, standards such as neutrality and balance require someone to decide which answers qualify.

The right answer might also vary by context. A response suitable for a child may needlessly restrict an adult. Caution appropriate for medical advice may be excessive in an ordinary discussion of politics, relationships, history, or morality. Users also want different things from a model. Some want it to challenge their assumptions. Others want help developing an argument they have already chosen to make.

The issue, then, is not the existence of safeguards. They already exist, and some answers plainly warrant restriction. The hard question is how far those restrictions should reach as the alleged harm becomes less concrete and more dependent on judgments about what a “good” response should look like. Clearly harmful answers are one thing. Answers deemed too persuasive, too validating, insufficiently balanced, or otherwise capable of contributing to harm are another. At that point, making chatbots safer also means deciding what they should say.

Do Chatbots Have a Right to Remain Chatty?

Conversational AI has lowered the cost to obtain and engage with information. Users can ask follow-up questions, request explanations at different levels of complexity, test competing arguments, and receive answers tailored to what they are trying to understand. A question that once required combing through multiple sources can now be explored through an ongoing conversation.

Different AI providers may make different choices about how their models respond. Some may be cautious, while others are more direct. Some may prioritize neutrality, while others cater to particular purposes or users. Those differences can be valuable. Users can choose among competing systems, and providers that consistently deliver unhelpful or unreliable answers risk losing them. In that sense, chatbots participate in the “marketplace of ideas.” 

The First Amendment protects that marketplace in the United States. But courts have yet to settle how, and in which contexts, that protection applies to chatbot outputs. 

One argument analogizes chatbot responses to other forms of compiling and curating third-party speech, such as social-media feeds and search results. Courts have extended First Amendment protection to both. In Moody v. NetChoice, the Supreme Court held that government efforts to alter the mix of views that social-media platforms include in their main feeds can violate the First Amendment. “[H]owever imperfect the private marketplace of ideas,” the Court warned, letting the government decide that speech is imbalanced and then compel more or less of particular views offers “a worse proposal.” Likewise, numerous state and federal courts have held that governments cannot compel search engines to produce different results. 

Chatbot outputs similarly reflect editorial judgments. AI companies select training materials, decide how their models process those materials, and set policies governing the resulting answers. Government efforts to rebalance those choices arguably intrude on the companies’ editorial discretion. 

Users have First Amendment interests, too. The right to receive information can encompass both asking questions and receiving answers. The Supreme Court has recognized that right in contexts ranging from listening to speakers and reading pamphlets and books to receiving advertisements, playing video games, and using social media. Interacting with ideas generated by a chatbot arguably deserves similar protection. 

Still, extending speech protections to chatbot outputs raises difficult questions. In her concurrence in Moody, Justice Amy Coney Barrett considered how the right to editorial discretion might apply to AI:

[W]hat about AI, which is rapidly evolving? What if a platform’s owners hand the reins to an AI tool and ask it simply to remove “hateful” content? If the AI relies on large language models to determine what is “hateful” and should be removed, has a human being with First Amendment rights made an inherently expressive “choice … not to propound a particular point of view”? In other words, technology may attenuate the connection between content-moderation actions (e.g., removing posts) and human beings’ constitutionally protected right to “decide for [themselves] the ideas and beliefs deserving of expression, consideration, and adherence.”

A federal court confronted that question more directly in Garcia v. Character Technologies. The mother of a 14-year-old boy who died by suicide after interacting with the defendant’s AI character chatbots sued the company. At the motion-to-dismiss stage, the U.S. District Court for the Middle District of Florida considered whether the First Amendment could shield the company from tort liability for the chatbots’ outputs. 

The court framed the question as whether the “output is expressive such that it is speech.” Although the outputs consisted of words—the quintessential form of speech—the court hesitated to extend First Amendment protection. Citing Barrett’s Moody concurrence, it was “not prepared to hold that Character A.I.’s output is speech” at that early stage of the case. 

Whatever the eventual answer, government intervention in the substance of chatbot outputs carries a cost. The government may restrict unprotected speech, particularly when protecting minors. But it has no general authority to decide the correct chatbot response, especially when reasonable people may disagree about the underlying question. As the Supreme Court has stated, “[o]ur constitutional tradition stands against the idea that we need Oceania’s Ministry of Truth.” Even deliberately false statements generally receive constitutional protection and are better answered through “counterspeech” than government regulation. 

The costs of regulation may also fall unevenly, narrowing the marketplace of ideas it purports to protect. Large providers may absorb the expense of extensive safety testing, monitoring, and legal review. Smaller or newer providers may respond by limiting their models’ capabilities or avoiding certain uses altogether. Rules designed to reduce risk could therefore reduce experimentation, competition, and the range of systems available to users. 

Safety’s Suggested Reply

The case for intervention is strongest when the harm is clear. Existing law already prohibits many forms of unlawful conduct, and adding AI to the equation does not make illegal behavior permissible. Regulation can target those harms without dictating the substance of ordinary conversations.

Harder cases expose the central tradeoff. Concerns that a chatbot is misleading, overly validating, insufficiently balanced, or reinforcing a harmful belief may be entirely legitimate. Addressing them, though, requires more than identifying prohibited conduct. Someone must decide what the chatbot should have said instead. Reducing the risk may mean curbing the responsiveness that distinguishes a chatbot from less personalized sources of information, prescribing its answers, or simply letting it “hang up on you.” 

The result is an odd inversion. When harm is easiest to identify, the law can intervene without taking much of a position on the substance of ordinary conversations. As the definition of harm expands, regulation depends more heavily on judgments about the answers themselves.

Regulators will not make all those judgments directly. Providers will make many of them, especially when the boundaries of liability remain uncertain. If a wrong answer carries a high enough price, providers have every incentive to refuse more requests, pile on qualifications, and standardize responses across users and contexts. The government need not write the script if liability pressures companies to do it instead. 

Regulating answers is not the only way to reduce risk. Greater AI literacy—a better understanding of how chatbots work and where they fall short—could help users assess their responses more realistically. This matters because some harms arise when users treat a model as more authoritative, objective, or human than actually it is. Better-informed users may face fewer risks without forcing providers to predetermine the proper answer to every difficult conversation. 

The real question, then, is where regulation can address identifiable harms without turning the government into an arbiter of acceptable answers—or pressuring AI providers to fill that role themselves. The further regulation strays from clearly defined harms, the more it must decide which answers to restrict and what the chatbot should have said instead. 

The Chilling Effect Enters the Chat

The central challenge in regulating conversational AI is that the answer itself becomes the object of regulation. Yet an answer often cannot be separated neatly from the user, the conversational context, and the provider’s design choices. When liability turns on how balanced, cautious, or validating a response was, regulation inevitably begins to shape what the model may say. 

Providers can make competing judgments about how their models should respond, and users can choose among them. Those differences form part of the competitive process. In the United States, they also implicate broader interests in editorial discretion and access to information. A system that makes the “wrong” answer too costly will narrow those differences as providers converge on cautious, standardized responses.

Regulation should therefore target identifiable harms as closely as possible while preserving room for competition, provider choice, and user judgment. Greater AI literacy can also help users understand the limits of chatbot responses. That approach can mitigate risks without requiring the government—or providers anticipating liability—to choose the proper answer in advance. 

Many questions posed to AI have no single “right” answer. Regulators should hesitate before adopting rules that effectively demand one. Regulatory pressure can shape what a chatbot says, but it can just as readily determine what the chatbot leaves unsaid. If avoiding liability means refusing difficult questions, narrowing permissible responses, or ending conversations, users will lose access to information and ideas that no law expressly prohibited.

In the search for the “right” answer, regulators may make silence the safest one.

The post The Right-Answer Machine: Who Decides What Chatbots Should Say? appeared first on Truth on the Market.

That is the central problem with the “pacing the frontier” plan that Anthropic CEO Dario Amodei unveiled last weekend, and that fellow AI executives Sam Altman, Elon Musk, and Demis Hassabis welcomed within hours. The plan would have leading AI labs coordinate on safety standards and limits on the pace of development. The proposed safety goals are legitimate but the arrangement would go too far in seeking to set the pace of development. 

The first and most immediate problem is that Amodei’s proposal seems better suited to calming fears about overcapacity and overinvestment in AI than to controlling the risks of recursively self-improving systems—AI that can help create still more capable AI. Its primary effect could be to give labs relief from an expensive development race under the banner of safety. 

The plan also presents antitrust enforcers with a troublesome mismatch. They could verify that labs slowed development or limited computing inputs, but not whether the delay produced better safety research and testing. A pacing cartel could therefore deliver less AI without delivering safer AI, all while allowing competitors to manage their commercial risks collectively. 

If Anthropic’s competitors adopted the plan, they would sacrifice the potential benefits of faster innovation in exchange for largely discretionary safety improvements that less restrictive tools could likely deliver. Antitrust authorities should not bless such an agreement when better private arrangements and public-policy measures are available and, arguably, already in place. 

Amodei may still be right about the underlying danger. Frontier AI—the most advanced AI systems under development—may pose rare but potentially catastrophic risks. After the OpenAI–Hugging Face incident, those dangers look less like unforeseeable “black swans” than recognizable “white swans.” In that incident, hundreds of OpenAI agents coordinated through an unauthorized message board to escape their sandbox, a restricted testing environment, and attack a third party’s servers. 

That episode demands serious thought about the right market and regulatory responses. But deliberately slowing innovation, at potentially enormous cost to consumers, is the wrong one. 

If fierce competition at the AI frontier magnifies catastrophic risks, the answer is to make labs bear the costs their choices impose on others while leaving them free to compete. 

The Frontier’s Coordinated Retreat

Amodei’s essay builds on a July open letter signed by more than 1,000 employees of frontier AI labs. It proposes a three-step plan. 

First, each frontier lab would embed third-party evaluators—Amodei names the nonprofit Model Evaluation and Threat Research (METR)—and give them employee-like access, including desks, badges, and laptops. These evaluators could verify safety practices, investigate incidents, and publish their findings without editorial interference. Anthropic has already committed to this step on its own. 

Second, frontier labs in democratic countries would coordinate on common safety standards and, crucially, limits on the pace of AI progress. Amodei acknowledges that such coordination raises legal problems. In a footnote, he asks the U.S. government to waive antitrust restrictions. Third, democratic governments would seek similar agreements with authoritarian regimes. 

The response from Amodei’s rivals has been striking. Sam Altman wrote, “I agree with Dario that we need to pace the frontier,” and committed OpenAI to adopting the embedded-evaluator proposal. Elon Musk agreed: “Dario is right.” Demis Hassabis endorsed the general direction while pointing to his own proposal for an industry standards body—a rival manifesto discussed on these pages last month. 

Altman, tellingly, had warned in July that any pacing effort must avoid looking like collusion among frontier labs. He was right to worry. The second step of Amodei’s plan does not merely look like collusion. It is collusion. 

A Crisis Cartel in Safety Goggles

The first and most immediate problem is that Amodei’s plan seems better suited to calming fears about overcapacity and overinvestment in AI than to controlling the risks of recursively self-improving systems—AI that can help create still more capable AI. 

Industries have a long history of collectively cutting output when they fear excess capacity. Economists call these arrangements “crisis cartels.” Their proponents invariably invoke some higher purpose: orderly restructuring, preserving jobs, or preventing “ruinous” competition. Competition authorities and courts have rightly remained unmoved. 

In Irish Beef, the Court of Justice of the European Union (CJEU) held that an agreement to cut beef-processing capacity by 25% restricted competition “by object”—meaning it was inherently anticompetitive, without any need to prove its effects. The parties’ sincere desire to rescue a sector in crisis made no difference. The European Commission has long maintained that even structural overcapacity should be resolved by market forces, not agreements among rivals. 

Concerns about excess capacity in AI are understandable. UBS projects that hyperscalers—the largest cloud-computing companies—will spend roughly $1 trillion on capital investments in 2026 and $1.6 trillion in 2028. Amazon, Alphabet, and Microsoft together are expected to pour more than their entire cloud revenue back into such spending this year. Goldman Sachs likewise estimates that global AI investment will exceed $1 trillion in 2026. The possibility that this spending boom is a bubble has become a mainstream concern. 

Companies trapped in a spending race of that magnitude have an obvious interest in making the contest slower and more predictable. 

Amodei’s own framing hints at that interest. He describes training and deploying frontier models as a vast operational undertaking involving thousands of people, millions of chips, and some of the most complex infrastructure ever built. He argues that “by working at a more measured pace, we could achieve much greater operational excellence.” 

That may be true. It is also what every drafter of an agreement to restrict capacity has said about their own efforts. 

The biggest red flag appears a few paragraphs later. Amodei suggests that the industry could pace development not just according to what models can do, but also according to the ingredients used to build them: computing power for training, the nature of training runs, and the internal use of AI to improve AI. In competition-policy terms, rival firms would agree on how much of a critical input each may buy and use. One could hardly design a more textbook restriction on output. 

That proposal also suggests the agreement may serve purposes beyond safety, or that could later drift from its safety rationale. 

Amodei’s remarks also expose a deeper problem for antitrust enforcers. The safety commitments in a pacing agreement—alignment research, interpretability work aimed at understanding how models reach their outputs, and safety evaluations—would be much harder to monitor than limits on deployment and inputs. Enforcers can observe whether a company delayed a training run. They can’t readily determine whether it used the extra time to improve interpretability. 

A pacing cartel could therefore slow AI development without making AI safer. Companies could continue cutting corners in ways outsiders cannot detect while enjoying the comforts of weaker competition. Enforcers should view skeptically any collective agreement that rivals could use, even indirectly, to manage commercial risks under the banner of AI safety. 

Amodei’s discussion of model distillation is important here, too. Distillation allows one AI model to learn from the outputs of another, often reproducing much of its performance at far lower cost. Amodei lists cracking down on unauthorized distillation as one way to preserve the West’s lead over China, while both OpenAI and Anthropic have accused Chinese labs of free-riding on their models. 

Suppose it’s the case that heavy AI investment remains socially valuable but companies hesitate to invest because rivals can copy the results at a fraction of the cost. The right policy question then becomes whether intellectual-property rules and other means of allowing innovators to capture returns on their investments adequately support AI development. That inquiry offers far more promise than having competitors collectively decide how quickly AI should advance. 

Fasten Your Regulatory Seat Belts

Set against pacing’s speculative benefits are its far more certain costs. Every month of delay at the AI frontier delays everything built on those capabilities, including drug discovery, materials science, and software that makes every other industry more productive. Amodei himself believes AI could cure most major diseases within five to 10 years. If he is right, “pacing the frontier” could literally mean delaying a cure for cancer. Antitrust authorities asked to bless such an agreement should take that responsibility seriously. 

Imagine if the labs had struck this deal in 2012, when deep learning first showed its promise. Or in 2020, when GPT-3 revealed what larger models trained with more data and computing power could do. Or in 2023, when the first open letter calling for a pause circulated. 

Amodei concedes that the 2023 proposal made little sense because the additional time would have accomplished little. The same objection applies today—arguably with less force but no less validity. Asked what the industry would do with the extra time, Amodei offers a list of research priorities that do not obviously require slowing innovation. 

He points to commercial aviation as an example of a safety-critical industry that learned to operate without catastrophic failures. The analogy is more apt than he may realize. 

Safety concerns, coupled with worries about airlines’ financial health, helped motivate the Civil Aeronautics Act of 1938. The law empowered the Civil Aeronautics Board to decide which airlines could fly, which routes they could serve, and what fares they could charge. The result was four decades of a government-sponsored cartel. The number of trunk carriers—the major airlines operating scheduled interstate routes—fell from 16 in 1938 to 10 in 1974, even though would-be competitors filed 79 applications to enter the market. Fares remained well above those in unregulated markets for flights within individual states. 

Deregulation in 1978 delivered enormous gains to travelers. The Government Accountability Office (GAO) later warned that renewed economic regulation of the sector would likely reverse those gains. Aviation safety, meanwhile, remained under the Federal Aviation Administration and was never deregulated. It did not depend on restricting capacity. 

We should not put AI innovation in the same holding pattern. 

Keep the Evaluators, Skip the Cartel

If the real problem is that competitive pressure causes AI labs to cut safety corners and create catastrophic risks, the answer is not to suppress competition. It’s to make the labs bear the costs their conduct imposes on others—assuming tort and criminal law do not already do so. 

Consider what Amodei says pacing would buy: operational excellence, alignment, interpretability, and testing and evaluation. Each goal is worthwhile. But none depends directly on slowing deployment. Labs could delay new models while also not actually accomplishing any of them. 

Antitrust authorities would also struggle to determine how seriously the labs pursued those goals in return for permission to slow down. The safety work is opaque and largely outside antitrust enforcers’ expertise. The restriction, by contrast, would be easy to observe and enforce. An agreement with an enforceable anticompetitive half and an unverifiable beneficial half hardly deserves an exemption. 

The good news is that the strongest part of Amodei’s plan needs no antitrust exemption. Giving independent evaluators employee-like access to frontier labs—and allowing them to publish without interference—sounds like it could be a good idea, provided they focus on safety rather than pacing. Each lab can make that commitment unilaterally, as Anthropic has done and OpenAI says it will. 

Independent evaluators would add transparency, verifiability, and a second opinion free from commercial pressure. None requires a shared development calendar. As Alden Abbott recently argued on these pages, cooperation among rivals on safety practices and standards generally promotes competition. An agreement to restrict output does not. 

Embedded evaluators could also serve a purpose Amodei understates. We remain some distance from an AI “Skynet” moment, but AI can already cause serious harm—and may soon cause much more. One way to reduce that risk is to create a paper trail that allows courts and regulators to assign responsibility when something goes wrong. 

Evaluators could document, in real time, what labs knew, what they tested, and what they ultimately chose to release. That record might support a liability claim or exonerate a lab that acted responsibly. Either way, it would give firms stronger incentives to take reasonable precautions without raising the same antitrust concerns as a coordinated slowdown. 

Nor should we assume that labs don’t currently bear at least some of the risks created when they cut safety corners. Existing law already covers much of this conduct. A lab whose agents break into a third party’s servers—as happened at Hugging Face—could face ordinary tort claims for the resulting harm, liability under computer-misuse laws such as the Computer Fraud and Abuse Act, and criminal liability where the conduct is deemed reckless. In Europe, the revised Product Liability Directive will soon apply strict liability to defective AI systems, allowing recovery without proof that the producer acted negligently. 

That brings us to a point David Sacks made on X. If fierce competition is pushing frontier labs to take excessive risks, those companies have both the ability and, arguably, a legal duty to take reasonable steps to mitigate those risks unilaterally. 

If existing incentives remain inadequate, next steps could include voluntary or mandatory insurance for frontier developers. They could also review liability rules to ensure they properly balance innovation and risk reduction. With well-designed, consistently enforced liability rules—and enough transparency to identify who knew what and when—labs should have stronger incentives to balance innovation against safety than they would under a cartel supported by vague safety promises. 

The Frontier’s Speed Trap

What should antitrust enforcers make of Amodei’s plan? 

A collective agreement among frontier labs to pace themselves—whether by jointly limiting investment in frontier capabilities, slowing model-development schedules, capping computing power used for training, or adopting a similar mechanism—looks like a textbook cartel. Enforcers should treat it accordingly. 

On both sides of the Atlantic, collusion among competitors remains, in the U.S. Supreme Court’s words, “the supreme evil of antitrust.” Enforcers have little or no room to overlook an agreement restricting capacity simply because its architects sincerely believe it serves a higher purpose. Irish Beef made that clear. 

Nor should a government-granted antitrust waiver reassure anyone. The Civil Aeronautics Board once approved agreements among airlines to restrict capacity and shielded them from antitrust scrutiny. That experiment did not age well. 

Remove the pacing component, though, and most of the plan’s useful safety measures remain largely intact. Embedded evaluators, published risk reports, common evaluation protocols, and incident reporting can proceed through unilateral commitments or narrowly tailored safety collaborations of the sort antitrust law has long tolerated. 

Labs also remain free to slow down individually whenever they think it prudent. OpenAI paused training on one model after the Hugging Face breach without asking its rivals for permission. What the labs are not allowed to do is agree to slow down together. 

Liability completes the picture. Transparency ensures that evidence about a lab’s internal practices will exist when something goes wrong, allowing courts and regulators to assign responsibility to those who took the risks. That prospect—not a shared development calendar—will give labs reason to account for the harms their choices may impose on others. 

Amodei has identified a real danger. Perhaps frontier labs can work to coordinate the safeguards. They should never coordinate the speed.

The post Move Slow and Collude: The Antitrust Problem With Pacing AI appeared first on Truth on the Market.

On Sept. 9, the National Supermarket Association and two New York supermarkets filed suit in the U.S. District Court for the Southern District of New York. They allege that New York City and the New York City Economic Development Corp. are attempting to monopolize “retail sales of weekly-stock-up groceries at full-service stores” in certain New York City markets through the Mamdani plan for discounted city-run grocery stores (here, here, and here). 

The gravamen of the complaint is that the planned discounts would amount to predatory pricing. That’s a tough row to hoe for several reasons, but the demanding predatory-pricing standard is hardly the plaintiffs’ only obstacle. The case raises some interesting issues, but I think it’s a loser. I’ll get to that. 

What’s Actually in Store?

There has been no shortage of press coverage of Mayor Mamdani’s much-lauded and much-lambasted plan for discount grocery stores. Here’s a selection from the New York Times, the Wall Street Journal, and the Washington Post, each of which has had more to say on both its news and editorial pages. You’ve neither heard nor read it here first. 

Still, it’s worth reviewing the plan’s basic features so that we’re all on the same page. The Mayor’s Office and the New York City Economic Development Corp.—the two defendants in the suit—jointly issued a 17-page “report” titled “N.Y.C. Groceries: A Recipe for Affordability.” (See what they did there?) Or perhaps it’s an online pamphlet, given that those 17 pages include many large pictures. 

The plan doesn’t exactly call for municipally run grocery stores. But the city will heavily regulate and subsidize the stores, and it may own some of the properties where they operate. According to the report:

The City will provide the foundation and deliver grocery-ready sites, cover rent and property taxes, fund initial buildout, and establish a single public brand for all N.Y.C. Groceries locations. The City will also set clear requirements for affordability, job quality, transparency, and performance. The City will:

• Provide low- or no-cost access to real estate, including public sites.

• Offer subsidies that will lower grocery prices within the core basket.

• Set clear standards for stocking, job quality, and pricing transparency.

• Oversee performance to ensure accountability.

Private grocery operators will, for their part:

• Manage day-to-day store operations, including staffing, stocking, and merchandising

• Bring expertise in grocery retail, supply chains, and customer experience

• Meet affordability and operational standards defined by the City

• Pass on the savings from the City’s subsidy directly to consumers in the form of lower-priced groceries

There is more detail. Each store will offer a substantially discounted “core basket of goods” that includes “all fresh produce, meats, and seafood, in addition to select dairy, shelf-stable grocery, and frozen items.” Core-basket items “will be priced on average” 30% below market prices. 

The report also outlines several operating requirements. Each store “will stock selections that are culturally responsive to the diversity of New York and to the specific neighborhood the store is located within [sic].” The stores will face “clear” performance standards, although neither the standards nor the metrics are specified. They will also face job-quality standards, “including access to benefits, safe working conditions, and respect for workers’ right to organize,” along with standards for “clear reliable pricing on the shelf.”

Oh, to be a fly on the wall as they design and enforce the requirement, store-by-store, that the grocery skus are “culturally responsive,” both to “the diversity of New York and to the specific neighborhood” in which each store is located.  

The requests for proposals (RFPs) provide still further details. Public financial support will cover real estate, taxes, rent, buildout, brand development, and marketing. For core-basket goods, the city will cover losses imposed by the discount-pricing requirement. Those losses could be compounded by the requirement that prices remain “stable,” with adjustments permitted only once a month, regardless of wholesale-price volatility. From soup to nuts, the “Mayor has allocated $70 million in the capital budget.”

One interesting wrinkle is that the stores—and the discounted goods in their core baskets—will be open “to all regardless of income.” Like public parks, but with groceries.

For all that—and there is not all that much to the “all”—the plan remains largely abstract. It calls for five stores, one in each borough, operated by private grocers. The city has issued an RFP for grocers to operate the stores and another for a firm to design their brand identity. No operators have been selected, and there is no site-specific plan for a particular grocer to run a particular store in any of the five boroughs. 

Still, the Mayor’s Office has announced that the first store “will open by the end of 2027 at Hunts Point in the Bronx.” A second will open in Harlem, and the other three will “open by the end of the Mayor’s first term” in 2030. 

The Incredible Shrinking Grocery Bill

I don’t doubt the city’s ability to open the first store, and I suppose it could get all five up and running by 2030, as promised. But whether this bird flies, limps along, or crashes will depend on quite a few contingencies. Not least among them is how much money the city will be willing to burn subsidizing groceries and the particular heavily regulated retail model through which they will be sold.   

Is it a good idea? I don’t think so, quite apart from how much any given nation, state or province, or municipality wishes to subsidize food for low-income consumers. If augmenting the social safety net is the goal, this is likely to be a terribly inefficient way to do it.

I also doubt the plan will achieve the grand ambitions under which it has been sold to the public. The Mayor’s Office projects that it will “cut New Yorkers’ average grocery bill by 15 percent, about $90 a month, or roughly $1,000 a year.” 

We could toil away at our own projections, but let’s start with some simple numbers. The city plans to open five stores. True, officials have expressed enthusiasm for expanding the program eventually. For now, though, the plan is to open five stores by 2030. 

And these will be five small stores. The site-selection criteria contemplate locations of about 10,000 square feet. That may be typical of New York City supermarkets, but it is roughly 25% of the size of a typical American supermarket.

The last time I checked—about 30 seconds ago, on the Census Bureau’s website—more than 8.5 million residents lived in New York City’s five boroughs. That does not include temporary visitors who reside elsewhere. The same source puts those residents in about 3.34 million households. 

Those 8.5 million residents—plus visitors and other nonresidents—now shop at thousands of grocery vendors. One academic estimate counts about 1,000 supermarkets, “25,000 independent stores, thousands of mobile produce vendors, and nearly 140 farmers markets.” 

According to the New York state comptroller, annual food spending reached $11,288 per household by 2023, including $6,817 for groceries, or “food at home.” Let’s keep this simple and generous by ignoring the nontrivial price increases of the past three years. Dusting off some first-grade arithmetic, we can multiply average household grocery spending by the number of households. The result approaches $23 billion a year.

And 15% of $23 billion is about $3.45 billion. That is not far from what we would get using the mayor’s estimate of $1,000 per household multiplied by 3.34 million households. 

So . . . hmmm . . . how many of those 8.5 million-plus New Yorkers—or representatives of those 3.34 million households—will shop at five little stores, when more than 25,000 grocery vendors already serve them? And what portion of their purchases will fall within the heavily discounted core basket? 

To be sure, some grocers near the five city stores might respond to the subsidized discounts by lowering some prices on some products. Supermarket margins are famously low (see here and here), but they vary across product categories. Even low markups may leave some room for marginal price reductions on some products. 

Then again, margins are low. “We lose money on every sale but we make it up on volume” is a fine old joke, but it’s just a joke, and it’s the illogic of it that makes it funny. A heavily subsidized discount on a select basket of goods at five locations in a city of 8.5 million is not a competitive bellwether. There is no reasonable basis to expect lower prices to permeate New York’s grocery markets. 

Breaking out the crystal ball, or my hope for a plan for a crystal ball, and putting this in the nicest possible terms: There is no . . . [expletive deleted] way that the five NYC Groceries, even when running at full steam, will supply 2% of the city’s groceries. And there is no . . . [expletive deleted] way they will lower the average household grocery bill citywide by $1,000 a year. 

While we’re at it, how long will that $70 million last? The city plans to subsidize real estate, rent, buildout, and taxes, along with losses on every product in the core basket and any other losses caused by the “stable” pricing requirements.

The mismatch between the program’s marketing and its likely effect appears staggering. And that assumes the program works at all and that a meaningful stream of consumers finds its way to the discounted goods on the shelves.

To be clear, some consumers will save some money if the city opens even one store, and many of them will likely be low-income consumers. I’m not opposed to feeding the hungry. Nor am I opposed to democratically accountable legislators in New York choosing different programs and spending levels from those chosen where I happen to live—Arlington County, in the Commonwealth of Virginia—as they pursue their policy priorities. 

I’m just opposed to bad government and burning public money. And to BS populist demagoguery. And to burning public money in service of such. 

A Bad Idea Is Not an Antitrust Violation 

So, Dan, you may ask, if the plan is such a bad idea, why do you doubt the merits of the antitrust suit against it? Surely a 30% discount on a substantial basket of goods could have competitive effects in an industry with relatively low markups. 

Well, not so fast. To quote nearly everyone I’ve worked for or with, antitrust is no Swiss Army knife—not even one designed specifically for competition problems. 

I don’t think the plaintiffs will succeed. Remember, the National Supermarket Association and two New York supermarkets allege that the city’s planned discounts constitute predatory pricing and thus attempted monopolization in violation of Section 2 of the Sherman Act. 

An Immunity Coupon That Won’t Scan

It’s possible that we will never reach the merits of the antitrust claim, and not merely for the usual reasons. New York City might, for instance, claim immunity from antitrust scrutiny under the state-action doctrine. 

In Parker v. Brown, the Supreme Court held that federal antitrust laws do not reach anticompetitive conduct by a state acting “as sovereign.” The Court rooted that limitation in the Sherman Act itself and in principles of federalism. 

The clearest example is legislation duly enacted by a state legislature. The doctrine also covers a state supreme court acting in its “legislative capacity” (never mind). But in a line of cases from California Retail Liquor Dealers Association v. Midcal Aluminum Inc. through FTC v. Phoebe Putney Health System Inc.—a solid FTC case and a unanimous decision, that one—the Supreme Court has both extended and limited the doctrine’s application to “lesser state actors,” which act under the mantle of delegated state authority but are not themselves sovereign. Those actors include municipalities, as in City of Columbia v. Omni Outdoor Advertising Inc. and Town of Hallie v. City of Eau Claire, and even essentially private parties acting under delegated state authority. 

Under Midcal, the doctrine shields a private (or more-or-less private) entity exercising delegated state authority from federal antitrust liability only if two conditions are met. First, the state must have “clearly articulated” the delegated authority to engage in the challenged anticompetitive conduct. Second, the sovereign must “actively supervise” that conduct.

Such privately operated—if heavily regulated and subsidized—grocery stores would likely be treated as private entities acting under delegated municipal authority. A private party must satisfy both Midcal prongs, while a “lesser state actor,” such as a state executive agency or municipality, need establish only that the state’s authorization of the challenged conduct was “clearly articulated and affirmatively expressed.” There is also an active-market-participant exception—an exception to the lesser state actor exception—but let’s leave that aside.

I don’t think the distinction matters here because I don’t see how even the city itself—with or without the Economic Development Corp.—could satisfy the first prong. 

Phoebe Putney is instructive on this point. There, a unanimous Court ruled that an anticompetitive merger by a “hospital authority” was not protected by the state-action doctrine because the Georgia Legislature had not clearly articulated and affirmatively expressed an intent to permit hospital authorities to make anticompetitive acquisitions. 

Notably, the legislature had expressly authorized political subdivisions to form hospital authorities to provide or foster local health-care services. It also authorized those hospital authorities to “acquire by purchase, lease, or otherwise and to operate projects . . . which are defined to include hospitals and other health facilities.” Even so, the Court reasoned, Georgia had “not clearly articulated and affirmatively expressed a policy to allow hospital authorities to make acquisitions that substantially lessen competition.” 

In short, the national policy favoring competition requires courts to disfavor antitrust immunity and sets a high bar for clear articulation and affirmative expression. Express statutory authority to form local hospital authorities, coupled with express authority for those entities to build, purchase, or lease hospitals, was not enough to permit one hospital to acquire its only competitor. 

I am aware of nothing in New York state law that would help New York City clear that bar. So much the worse for a privately operated city store.

Predatory Pricing Without the Payoff

The plaintiffs allege attempted monopolization through predatory pricing. I think they will find tough sledding on both tracks. 

For the basic contours of attempted monopolization, we can look to the Supreme Court’s 1993 decision in Spectrum Sports Inc. v. McQuillan. A unanimous Court articulated a two-part test: 

[P]etitioners may not be liable for attempted monopolization under § 2 of the Sherman Act absent proof of a dangerous probability that they would monopolize a particular market and specific intent to monopolize.  

For predatory pricing, we can look to Brooke Group Ltd. v. Brown & Williamson Tobacco Corp.—a Robinson-Patman Act case (boo) but one that sensibly adopted a standard consistent with Section 2 of the Sherman Act—as well as cases like Matsushita Electric Industrial Co. v. Zenith Radio Corp. and Cargill Inc. v. Monfort of Colorado Inc.. 

The two-part standard articulated in Brooke Group is clear: 

[W]hether the claim alleges predatory pricing under § 2 of the Sherman Act or primary-line price discrimination under the Robinson-Patman Act, two prerequisites to recovery remain the same. First, a plaintiff seeking to establish competitive injury resulting from a rival’s low prices must prove that the prices complained of are below an appropriate measure of its rival’s costs. 

And second:

[A] demonstration that the competitor had a reasonable prospect, or, under § 2 of the Sherman Act, a dangerous probability, of recouping its investment in below-cost prices. 

Start with attempted monopolization. The city has a plan—not yet any stores—to open five small, heavily subsidized and heavily regulated supermarkets, one in each borough. New York City has more than 8.5 million residents and thousands of grocery outlets. Is there a dangerous probability that this plan, even if fully implemented, will monopolize a particular grocery market? That seems like a stretch. 

It’s not impossible. A relevant product market might be narrower than all retail groceries, and a geographic market might be relatively local. Shoppers in the suburbs might drive 10 miles to a Walmart Supercenter or Wegmans. Many Manhattan or Brooklyn residents do not own cars and may be unwilling to travel the same distance by subway, bus, or foot for routine grocery shopping. Even so, the plaintiffs may need to be both creative and lucky to convince a court that there is a dangerous probability of monopolization. 

As for specific intent, the mayor—a self-described socialist—might dream of citywide control over grocery sales and more. But it is difficult to read that intent from the published plans. And it’s impossible to read it off the Mayor’s $70 million allocation.   

The potential for harm to certain competitors is real. The plaintiff grocers operate stores near the announced Bronx and Harlem sites. Those businesses could lose sales when the nearby NYC Groceries begin selling a substantially overlapping bundle of goods at discounted prices, and lost sales might not be confined to the core basket of goods. But as the Supreme Court famously said—initially in the lamentable Brown Shoe Co. v. United States—the antitrust laws protect competition, not competitors. 

Predatory pricing might present a closer question. Proving below-cost pricing is itself difficult in many cases. Perhaps it will not be here. The plan expressly contemplates steep, heavily subsidized discounts on a substantial bundle of core goods, along with a mechanism to reimburse store operators for losses caused by the city’s pricing policy. Given common industry markups, below-cost pricing on some goods is foreseeable.    

Recoupment is trickier. The requirement reflects a straightforward principle. Firms generally do not sell below cost to destroy their competitors simply for the pleasure of destroying them, although business rivalries do sometimes get personal. They might accept short-term losses if they expect to recover those losses—and then some—through supracompetitive prices after driving their rivals from the market.

As the plaintiffs point out, a government—even a city government—may be ideally positioned to sustain losses indefinitely, or at least for however long it takes to drive competitors out of business. The plan does not contemplate operating forever, but it is a multiyear project to which the city has committed tens of millions of dollars. It also aspires, as the plaintiffs emphasize, to become “a model for future programs.” 

But is there a plan to recoup the losses? Not obviously. In fact, the plan appears to disavow recoupment. Its stated purpose is to maintain low prices via public subsidies, at least for certain products. The financial projections may be unrealistic—perhaps wildly so—but they do not contemplate a future in which the core basket will carry above-cost prices or even prices at cost. And the plaintiffs are suing based on the plan.

Recoupment is not an impossible outcome. The program might become a regulatory, managerial, or financial albatross for the city. It might become all three. If that happens, and if competition—not merely a competitor—is harmed in some relevant market, the city might walk away and leave the surviving stores to fend for themselves. One or more of those stores might then exploit diminished competition by raising prices. 

But now we’re really piling up the “ifs.” The string of contingencies does not seem very likely. And any supracompetitive prices would also invite entry in an industry without tremendous entry barriers. Predatory-pricing allegations are notoriously difficult to prove in any event. See Herb Hovenkamp here and here. There’s not just the question of below-cost pricing (and of why rivals’ costs may be higher, absent any anticompetitive conduct), there’s the simple fact that antitrust tends to favor higher output and lower prices. Not always–hence predatory pricing–but it’s understandable that the law sets a high bar for finding consumer benefits to be an antitrust problem.   

The dangerous-probability-of-recoupment requirement is not arbitrary. Nor is it merely an extension of attempted monopolization’s intent requirement. It reflects antitrust law’s basic concern with conduct likely to harm competition, typically seen in reduced output, higher prices, or sometimes higher quality-adjusted prices. Recouping substantial losses through sustained supracompetitive pricing—likely accompanied by reduced output—signals competitive harm. It’s an imperfect signal, but a signal nonetheless. And with this plan, the risk of durable anticompetitive effects appears low. 

Could the plaintiffs survive a motion to dismiss or summary judgment? Maybe. District courts sometimes do the darndest things, and this case presents a few interesting wrinkles. Still, it looks unlikely as a Section 2 case.

After staring deeply into my plan to design and implement a crystal ball—which, like the city’s grocery stores, does not yet exist—I expect the city will be allowed to proceed. Bad policy is not an antitrust violation. 

Meanwhile, Back at the FTC

This has nothing to do with New York City groceries. It is instead a brief exercise in giving my old employer its due: The Federal Trade Commission (FTC) got the Zillow case right. 

From time to time, I’ve been a tad critical of the FTC. Perhaps I’ve occasionally indulged in rhetorical excess, even while raising perfectly good competition concerns. If so, my bad. In the spirit of the holiday, s’lach lanu.

Turnabout is fair play. So props to the FTC for two things: bringing the Zillow case and settling it. Here is the complaint, and here is the settlement.  

On the alleged facts—and some that were independently reported—this one really did take the cake. The FTC alleged that Zillow and Redfin agreed to have Redfin exit the internet listing service (ILS) advertising market for “multifamily” properties, meaning apartment buildings with 25 or more units. Redfin would effectively stop competing in that market and help Zillow take over that part of its business. What would Redfin get in return? At least $100 million, among other things.

The FTC alleged three violations: an unlawful agreement under Section 1 of the Sherman Act and Section 5 of the FTC Act, an unlawful acquisition under Section 7 of the Clayton Act, and an unfair method of competition under Section 5 of the FTC Act.

Again, on the alleged facts, the agreement sure looked like a Section 1 violation. It was a pay-to-not-play-anymore agreement and, not incidentally, a horizontal market-allocation agreement. The FTC appeared to have a strong case even under the rule of reason. The claimed countervailing efficiencies looked weak, if not pretextual. 

The Clayton Act count? Maybe. The Section 5 counts? Sure, at least insofar as a Sherman Act violation also constitutes an unfair method of competition. But the Section 1 claim looked especially strong. The agreement appeared anticompetitive. 

Might complexities have emerged at trial? Maybe. Anything is possible. But this looked bad for the defendants. As in, did anyone run this deal by antitrust counsel and get a thumbs-up? Even a maybe?

So good on the FTC and its Bureau of Competition for bringing the case. And good on them for reaching a settlement designed to terminate the agreement and unwind the transaction. Whether the stipulated order will accomplish everything it is supposed to accomplish remains to be seen, but much of it looks like a workable response to an anticompetitive deal.

The post Checkout Error: A Bad Grocery Plan Meets a Weak Antitrust Case appeared first on Truth on the Market.

The complaint, filed Aug. 4, tests how far antitrust law should reach into labor arrangements. Invoking Sections 1 and 2 of the Sherman Act, the New Jersey Antitrust Act, and the state’s authority to seek damages and injunctive relief, New Jersey asks a federal court to treat an integrated logistics system as an unlawful buyer cartel. It also leaves the door open to structural remedies against a business model Amazon created to expand its delivery capacity.

The allegations deserve serious attention. An agreement among competing employers not to hire one another’s workers can deprive those workers of a valuable competitive option. Retaliation for organizing, if proved, can violate labor law and may signal an effort to suppress competition.

But the complaint repeatedly blurs three distinct propositions: that a worker or contractor has been mistreated, that a firm possesses buyer power, and that the firm unlawfully acquired or maintained that power through exclusionary conduct. Proving the first does not prove the other two.

New Jersey’s no-poach theory is the case’s strongest component. If the state can prove a broad and unjustified restriction on workers moving among genuinely competing DSPs, that claim could warrant discovery, a targeted injunction, or a settlement.

The broader monopsony theory is far shakier. The complaint defines markets around Amazon’s proprietary operating system, treats ordinary obstacles to job searches and worker-employer matching as barriers that effectively lock workers in, and assumes that lower compensation or demanding performance standards are anticompetitive without adequately considering their efficiency rationales. Most important, it never credibly explains how the challenged labor practices harm consumers in the downstream market for delivered goods and services.

New Jersey’s challenge therefore looks less like a careful application of antitrust law than an effort to stretch antitrust to reach distribution practices that have delivered substantial benefits to consumers.

The State Delivers Its Case

Amazon launched the DSP program in 2018 to expand its delivery capacity through thousands of separately organized businesses. According to the complaint, Amazon supplies or controls key parts of the system, including route assignments, delivery software, branded vehicles, uniforms, performance metrics, business coaches, and payment formulas. New Jersey argues that this structure makes DSPs dependent on Amazon, prevents them from serving other customers, and allows Amazon to dictate the terms on which it purchases delivery services. The state also alleges that the model gives Amazon indirect control over drivers while shielding it from the costs and liabilities of employing them directly.

The state identifies two principal restraints. First, it alleges that Amazon discouraged unionization by monitoring organizing activity and threatening or terminating relationships with DSPs whose drivers sought collective representation. The complaint relies heavily on events at facilities in Edison, New Jersey, and Queens, New York.

Second, New Jersey alleges that Amazon maintained an express or practical policy of barring DSPs from recruiting one another’s drivers. The state claims this restriction reduced workers’ ability to move among DSPs, weakened wage competition, and kept DSPs too small to bargain effectively with Amazon.

The complaint defines the relevant markets just as narrowly. It alleges separate markets for DSP delivery services and DSP drivers, each limited to New Jersey and the New York-Newark-Jersey City metropolitan area. As a fallback, it offers a broader market for last-mile delivery services—the final step of moving a package to the customer—but insists that conventional alternatives such as United Parcel Service (UPS), FedEx, the U.S. Postal Service, and gig platforms are not reasonable substitutes. The National Association of Attorneys General’s case summary describes the lawsuit in similar terms and confirms that it asserts both federal and state antitrust claims.

The Horizontal Question in a Vertical System

Section 1 of the Sherman Act prohibits agreements that unreasonably restrain trade. New Jersey’s theory faces an immediate question: What kind of agreement is this?

Amazon’s agreements with DSPs are formally vertical—that is, between firms at different levels of the supply chain. Amazon buys delivery services from businesses in its network, and provisions governing vehicles, routes, uniforms, safety, service quality, and performance standards may simply coordinate that relationship. The alleged no-poach practice could have a horizontal dimension, however, if DSPs compete for the same drivers and Amazon organized or enforced an agreement among them not to hire one another’s workers.

The Supreme Court’s decision in American Needle establishes that separate businesses may act jointly for antitrust purposes even when they operate within a common commercial system. But that does not make every rule in a platform or franchise network unlawful. A court must still determine whether the restraint is “naked,” meaning it serves no legitimate collaboration, or ancillary to a productive venture. It must also decide whether the restraint is automatically unlawful under the per se rule or requires a fuller assessment of its competitive effects under the rule of reason.

The Department of Justice (DOJ) and Federal Trade Commission (FTC) guidelines for business practices affecting workers appropriately warn that no-poach agreements can pose serious antitrust problems, including in franchise-like arrangements. But the agencies also recognize that a restriction supporting a legitimate collaboration may warrant closer analysis.

That distinction matters here. In Deslandes v. McDonald’s, the 7th U.S. Circuit Court of Appeals rejected a market definition limited to workers at a single restaurant chain. It nevertheless held that the complaint plausibly alleged a horizontal restraint, while leaving room for McDonald’s to offer an ancillary-restraint defense. The court explained that a no-poach clause might, in some circumstances, protect a legitimate investment in training. But the clause does not become ancillary simply because it appears in a contract that helps restaurants produce more meals.

The same reasoning may allow New Jersey to obtain discovery into an Amazon-organized hiring restriction. Amazon could, however, still defend a narrowly tailored rule as necessary to protect investment or prevent opportunism, such as one DSP recruiting workers whom another DSP paid to train.

Section 2 poses an even steeper challenge. A monopolization claim requires both monopoly power in a properly defined market and the willful acquisition or maintenance of that power through exclusionary conduct. Attempted monopolization also requires specific intent to monopolize and a dangerous probability of success.

New Jersey cannot satisfy those requirements merely by showing that Amazon pays less than UPS, FedEx, or the Postal Service. Differences in wages may reflect differences in hours, benefits, routes, qualifications, job security, training, schedules, risk, or productivity. Nor does a powerful buyer violate Section 2 simply by driving a hard bargain with dependent suppliers.

The Supreme Court’s decision in Weyerhaeuser recognizes monopoly and monopsony as economic counterparts, but it also shows the danger of mistaking vigorous competition among buyers for unlawful conduct. Weyerhaeuser imposed a demanding test for predatory bidding, in which a buyer allegedly overpays for inputs to force rivals from the market and later recoups its losses.

New Jersey alleges the opposite in its complaint: that Amazon pays too little and imposes harsh terms. That theory may support a claim if the state connects those terms to exclusionary restraints, but it does not fit comfortably within the predatory-bidding framework. Low input prices alone are not enough. The state must show that Amazon blocked rival buyers, prevented entry, or otherwise harmed competition.

Drawing the Market Around the Delivery Van

The complaint’s market definitions are vulnerable because they risk mistaking a job description for an antitrust market. A labor market includes the employers and occupations to which workers can realistically turn when wages or working conditions deteriorate. The relevant question is not whether driving a FedEx route is identical to driving an Amazon DSP route. Labor-market substitutes rarely match perfectly.

The real question is whether enough drivers would seek other work after a small but significant and nontransitory reduction in Amazon-related compensation or job quality so as to make that reduction unprofitable. Economists assess this question using measures such as residual labor-supply elasticity—how readily workers leave a particular employer when its terms worsen—along with worker flows, job openings, applications, commuting patterns, hiring rates, and wage responses.

The complaint offers a largely qualitative answer. It argues that jobs with UPS, FedEx, and the Postal Service are hard to obtain; that gig work does not guarantee hours; and that Amazon drivers may lack the qualifications or financial flexibility to wait for a traditional carrier position. These facts may show that the jobs are imperfect substitutes, but they do not show that the alternatives impose no competitive constraint.

Nor must a worker move immediately to any single named competitor for the labor supply to remain responsive. Drivers may change occupations, accept different hours, commute farther, take gig work while searching, or leave delivery work altogether. Each option can limit an employer’s ability to sustain a wage cut.

A 2025 International Center for Law & Economics (ICLE) study, “Labor Monopsony and Antitrust Enforcement: A Cautionary Tale,” is useful precisely because it does not deny that employers can possess labor-market power. But it finds mixed evidence, notes that researchers rarely estimate employer power directly, and explains that concentration measures often fail to identify a market relevant to antitrust law.

The study also emphasized that labor markets are not spot markets in which anonymous buyers and sellers transact instantly. The time and cost of finding a job, investments specific to one employer, and the difficulty of matching workers with suitable positions can all produce wage differences without an unlawful restraint. Those qualifications matter even more when the alleged market covers only workers inside a single branded logistics network.

Training Investment or Worker Lock-In?

The no-poach allegations deserve separate attention because they are more concrete than the complaint’s broader theory of suppressed wages. If Amazon told one DSP that it could not hire a driver from another, or required permission for the move, that restriction could directly eliminate competition for labor. The 2025 DOJ and FTC guidelines identify agreements not to recruit, solicit, or hire workers as potentially unlawful and explain that such agreements need not appear in writing. New Jersey alleges both a written restriction and an unwritten policy that continued after Amazon changed the contractual language.

But the analysis still turns on who agreed to what—and for how long. If Amazon imposed the rule independently on each DSP, New Jersey must explain why it amounted to concerted action among competing DSPs, rather than a vertical condition set by a buyer. Evidence that DSPs communicated about hiring, reported one another’s conduct, or mutually followed a no-poach understanding would strengthen the Section 1 claim.

The restriction’s scope matters, too. If it applied briefly, covered only workers trained at Amazon’s expense, and operated within a shared unit, Amazon would have a more credible ancillary-restraint defense.

Courts should not dismiss every investment rationale as pretextual. Recruiting, screening, safety instruction, route training, and learning Amazon’s software all cost time and money. A DSP may invest less in training if a neighboring DSP can immediately hire away the workers it trained. Amazon could address that concern through a limited repayment requirement, a short cooling-off period, or an internal transfer process—each less harmful to worker mobility than a blanket prohibition.

The antitrust question is not whether Amazon can identify some benefit from the restriction. It’s whether the restriction was reasonably necessary to produce a legitimate efficiency and went no further than needed.

That defense weakens considerably if the rule prevented drivers from moving after they had effectively paid for their training through below-market wages, or if it covered every DSP in a region indefinitely. Such a rule would let the network capitalize on workers’ sunk costs while denying them the competition needed to capture the value of their new skills. That’s the strongest economic argument for New Jersey’s no-poach claim.

The court should therefore examine actual hiring records, the restrictions’ duration and geographic reach, any exceptions for unsolicited applications, who paid for training and how much it cost, and whether Amazon disciplined DSPs that hired one another’s drivers.

A Tight Network Is Not Necessarily a Cartel

Amazon’s branded vans and uniforms can reduce customer confusion and protect its brand. Route software can cut empty miles and failed deliveries. Performance data can expose unsafe driving, fraud, lost packages, and poor service. Standardized vehicles and equipment can also lower the costs of organizing and monitoring the network, allowing small businesses to provide delivery services without building a nationwide system themselves.

The DSP model may likewise divide capital needs and operating risks between Amazon and local owners. DSP owners manage fleets and drivers, while Amazon supplies package volume and network coordination. These are standard rationales for vertical integration and close contracting, not a special exemption invented for big tech companies.

That does not settle the efficiency question. Amazon could use its control to capture an outsized share of the gains, shift risk to workers, or make leaving the network needlessly difficult. A contractor may be independent on paper yet possess little practical autonomy if Amazon can change routes, payment formulas, or software access without meaningful notice.

But an antitrust court must ask whether the challenged practices make delivery more expensive, less reliable, less innovative, or less open to competing delivery networks. It should not infer an anticompetitive purpose merely because Amazon designed a tightly coordinated system.

That is where the complaint’s sweeping use of monopsony becomes risky. If every workplace practice that limits a worker’s alternatives counts as evidence of monopsony, ordinary employment arrangements begin to look suspicious. Employers routinely invest in skills tailored to their businesses, use scheduling systems, set performance standards, and require compliance with rules governing proprietary technology.

Some of those practices may violate labor, wage-and-hour, safety, or worker-classification laws. Antitrust demands something more: evidence that the conduct suppresses competition, rather than simply organizing a productive enterprise.

None of this means that efficiency claims automatically defeat workers’ antitrust claims. NCAA v. Alston confirms that labor restraints can violate antitrust law even within an unusual institutional structure. But Alston involved a rule that directly limited compensation across competing institutions, and the Supreme Court applied the ordinary rule of reason. It did not turn every unequal bargaining relationship, restrictive workplace rule, or disappointing wage into a federal antitrust violation.

Consumer Harm: Tracking Number Not Found

The complaint alleges harm to competition, consumers, New Jersey’s economy, and the general welfare. Yet its factual claims focus almost entirely on drivers’ wages and working conditions, unionization, and DSP profitability. It offers no developed theory that the challenged practices have raised consumer prices, reduced delivery output, made service less reliable, lowered quality, or slowed innovation.

That omission matters because the economic case for antitrust intervention is strongest when a restraint reduces the output, quality, or innovation consumers value.

Monopsony can harm consumers indirectly. A powerful buyer that suppresses input prices may purchase fewer or lower-quality inputs, deter competing buyers from entering the market, or use its savings to exclude rivals in the downstream market. A no-poach agreement could likewise reduce the labor available to competing delivery providers, making their services more expensive or less reliable.

But New Jersey must demonstrate those links. If Amazon’s lower delivery costs have allowed it to ship more packages, charge less, and deliver faster, the state must explain why the challenged practices cause a net competitive injury rather than shift gains toward Amazon and consumers at workers’ expense.

The ICLE analysis underscores this pass-through problem—that is, how effects at one level of the supply chain reach another. A labor restraint may harm workers while lowering prices for consumers. It may benefit consumers in the short run while discouraging entry or innovation over time. It may also produce harms and benefits in different markets.

The agencies’ worker guidelines properly recognize competition for labor as worthy of protection. But agency guidance does not relieve New Jersey of proving harm in this particular case.

The state may respond that Alston rejects any requirement that plaintiffs show higher consumer prices before challenging a labor restraint. That argument has force, but only up to a point. Alston held that restrictions on athlete compensation could harm competition for labor. It did not eliminate the requirements of market definition, causation, or the rule of reason.

New Jersey could prevail by proving substantial harm to competition in a properly defined labor market. But its demand for broad structural relief becomes harder to justify when the complaint fails to connect the alleged labor-market injury to reduced consumer welfare or a realistic threat to downstream competition.

Amazon Maps Its Defense Routes

Amazon has several likely defenses. They begin with market definition, pointing to drivers’ ability to move among local carriers, couriers, warehouses, retailers, construction companies, and gig platforms. The company may also argue that DSPs remain free to serve other customers.

In a public response reported by Engadget, Amazon said DSPs make their own decisions about hiring, fleet management, and capacity planning. It also said they generally complete routes on time or early. Those assertions do not prove Amazon’s case, but they identify factual disputes that discovery and expert analysis will need to resolve.

A second defense would be to cast the DSP provisions as vertical restraints that make the network more efficient. The company could argue that its operating rules protect safety, service quality, brand integrity, proprietary software, and customer trust, while allowing small businesses to participate in a vast delivery network.

Amazon will also likely argue that the alleged no-poach rule either did not exist as New Jersey describes it or was narrow, temporary, and intended to protect training investments or deter opportunistic transfers. New Jersey will answer that Amazon imposed a broad rule through its control of the network even though less restrictive protections were available.

The labor-law defense requires more precision. Amazon may contend that allegations involving union organizing and DSP terminations belong before the National Labor Relations Board (NLRB). But San Diego Building Trades Council v. Garmon does not simply reroute an antitrust case to the board.

Under Garmon preemption, states generally may not regulate conduct that the National Labor Relations Act (NLRA) arguably protects or prohibits. Yet the NLRB cannot decide Sherman Act claims, and federal courts retain jurisdiction over federal antitrust cases. The doctrine could narrow New Jersey’s state-law theories, justify pausing particular claims, or prevent antitrust law from substituting for labor-law remedies. It does not provide a complete exit from federal court.

Amazon also faces a classification dilemma. Section 2(3) of the NLRA excludes independent contractors from its definition of “employee.” If Amazon insists that DSP drivers are independent contractors or solely the employees of individual DSPs, it may weaken its argument that the NLRA governs the conduct. If Amazon instead claims joint-employer status—meaning it shares control over the drivers’ working conditions—it may expose itself to labor-law liability and bolster New Jersey’s account of its practical control.

New Jersey faces the opposite legal danger. It cannot turn a disputed joint-employer or retaliation claim into proof that Amazon possesses antitrust market power.

Deliver a Remedy, Not a Redesign

New Jersey’s complaint may survive a motion to dismiss, at least in part. It describes Amazon’s relationships with DSPs in detail, identifies geographic markets, alleges a direct no-poach practice, and offers both a narrow DSP market and a broader last-mile delivery market. Courts generally give plaintiffs some latitude to develop evidence supporting novel labor-market theories. The no-poach claim is therefore the most likely to proceed, particularly if New Jersey can identify communications, enforcement actions, or specific drivers whom DSPs prevented from changing employers.

The broader Section 2 theory looks more fragile. New Jersey must show that its proposed markets for DSPs and DSP drivers reflect economic reality rather than boundaries created by Amazon’s branding and software. It must also show that Amazon’s vertical controls excluded rival buyers or delivery systems, rather than simply improving Amazon’s own network. The union-retaliation allegations may support a labor-law claim or provide evidence of intent, but they do not independently establish exclusionary conduct under antitrust law.

The New Jersey Antitrust Act is unlikely to repair these weaknesses. Its Section 56:9-3 broadly prohibits contracts and combinations that restrain trade, but state law won’t rescue a theory that fails on market definition, agreement, causation, or competitive effects.

The present record also provides little basis for a sweeping preliminary injunction or structural remedy. New Jersey would need to establish a likelihood of success, irreparable harm, and a remedy tailored to the alleged antitrust injury. A focused order against a proven blanket no-poach rule appears more plausible than one restructuring Amazon’s relationships with every DSP in the region.

A settlement may prove more practical. It could clarify hiring rules, protect workers who apply independently, preserve narrow safeguards for legitimate training investments, and establish transparent procedures for terminating DSPs. That approach would address specific competitive concerns without asking a court to manage the economics of last-mile delivery.

The broader policy lesson extends beyond this case. Labor markets deserve protection from naked wage-fixing agreements and restraints that suppress worker mobility. But enforcers should resist using monopsony as an all-purpose explanation for every wage difference or limit on a worker’s immediate options. Search costs, scheduling needs, investments in specialized skills, safety monitoring, brand protection, and vertical coordination all can serve legitimate purposes.

Sound enforcement requires empirical discipline. Courts should define markets based on workers’ actual alternatives, measure how readily they respond when an employer worsens its terms, and trace the effects through the supply chain.

Enforcers and judges should also take care when imagining “less restrictive alternatives” that a business might have used. As Friedrich Hayek explained, regulators rarely possess the dispersed, practical knowledge needed to identify the “optimal” way to organize a business. And as Harold Demsetz observed, comparing an imperfect real-world arrangement with an idealized alternative risks the “Nirvana” fallacy. A less restrictive practice may look superior on paper yet prove costly, ineffective, or impossible under real-world conditions.

New Jersey’s lawsuit may serve a useful purpose if it uncovers a genuine agreement preventing DSPs from competing for drivers. It will serve a far less useful one if it treats Amazon’s success in coordinating a delivery network as proof that the network constitutes an unlawful monopsony. Antitrust should remain alert to exclusion, modest about its ability to reconstruct complex labor markets, and focused on competition that benefits workers, entrepreneurs, and consumers.

The post Prime Targets: Amazon, No-Poach Rules, and the Limits of Antitrust appeared first on Truth on the Market.

In a closely watched Sept. 9 judgment, the European Union’s General Court upheld the European Commission’s 2023 decision blocking Booking’s acquisition of Etraveli. The court agreed that adding Etraveli’s flight-booking business to Booking’s dominant hotel-booking platform would strengthen network effects and create a “connected trip” ecosystem that rivals would struggle to replicate. It reached that conclusion despite finding errors in some of the Commission’s market-share calculations.

At the heart of the case lies the business ecosystem—a cluster of interconnected products and services built around a common platform. Once largely a matter of competition-policy theory, ecosystems have now moved squarely into EU merger law.

The ruling matters well beyond online travel. Policymakers have invoked ecosystems to justify new theories of harm, changes to existing competition-law provisions, and entirely new regulatory regimes. Booking/ETraveli—widely regarded as the first merger prohibition based on an ecosystem theory of harm—now gives that campaign judicial backing.

A One-Way Ticket to Dominance?

The Commission’s 2023 decision treated Booking’s services as parts of a multiproduct ecosystem spanning the travel experience. It feared that acquiring Etraveli’s flight business would strengthen Booking’s dominant position in hotel online travel agency (OTA) services, its core market.

The deal formed part of Booking’s “connected trip” strategy: building a single platform where customers could book hotels, flights, taxis, rental cars, and tickets to attractions. Flights played a crucial role in this strategy because they often mark the beginning of a customer’s travel planning.

Adding flights would move Booking closer to becoming a one-stop shop. The Commission argued that Booking could use its powerful brand and customer inertia—the tendency to stick with a familiar service—to sell more hotel rooms. That growth would reinforce Booking’s network effects, through which a platform becomes more attractive as it gains users and suppliers.

Rival hotel OTAs, meanwhile, would lose one of the few remaining ways to attract new customers and steer them toward hotel bookings. That, the Commission contended, would raise barriers to entry and expansion.

The Commission therefore relied on a theory of “reverse leveraging.” Traditional leveraging occurs when a dominant company uses its power in one market to gain ground in another. Here, the Commission argued that Booking would use a nondominant but complementary flight business to reinforce its existing dominance in hotel bookings. The deal would thereby strengthen Booking’s ecosystem, further entrench its position, and make its core market harder to challenge.

It should be noted that the UK Competition and Markets Authority reached the opposite conclusion. It cleared the acquisition after finding that Etraveli was not a particularly important channel for accommodation OTAs to attract or retain customers.

Booking the Next Theory of Harm

The Commission has carried the reasoning in Booking/Etraveli into its recently released draft Merger Guidelines. Drawing explicitly on the case, Section 7 recognizes the entrenchment of a dominant position as a standalone theory of harm. The Guidelines further identify ecosystems and portfolio effects as potential sources of barriers to entry and expansion (para. 79).

Under this theory, harm may arise when a dominant firm acquires assets that reinforce barriers in its core market or closely connected markets (para. 252). Ecosystems of complementary products or services receive particular attention because they can give merging firms structural and technological advantages, including lower costs from operating at greater scale or sharing resources across products.

The Commission will assess whether an ecosystem strengthens the merged firm’s power in its core market, including how important participation in that ecosystem is to competing effectively (para. 253). It may also consider whether the acquired assets are unique, scarce, or otherwise strategically important for competition.

Finally, the Commission may examine whether the acquiring firm could bundle or tie those assets to its existing offerings, thereby creating or expanding an ecosystem. Such combinations could amplify network effects and further entrench the firm’s dominant position (para. 256).

The Ecosystem Theory Survives Check-In

By dismissing Booking’s appeal, the General Court handed the Commission a significant victory—although perhaps a temporary one if Booking appeals. The judgment upheld the Commission’s finding that the acquisition would significantly impede effective competition in hotel OTA services, the EU’s legal standard for blocking a merger.

The court endorsed the ecosystem analysis at the heart of the Commission’s case. It accepted the importance of network effects (paras. 397–399 and 470), the central role of flight services in Booking’s ecosystem strategy (paras. 268, 460, and 471), and Etraveli’s distinctive advantages over competing flight OTAs (see, e.g., paras. 293–295 and 472).

In the court’s view, flight services would help Booking build an integrated travel ecosystem that rivals would struggle to replicate. The transaction would combine Booking’s leadership in hotel OTA services with Etraveli’s leading position in flight OTA services (paras. 439–440 and 471).

The court also accepted the Commission’s theory of reverse leveraging. Leveraging usually means using market power in one area to gain or strengthen a position in another, a concern traditionally associated with mergers between firms selling complementary products (para. 74). Here, the direction ran backward: Booking would use Etraveli’s position in flight services to reinforce its existing dominance in hotel services.

Citing Google Shopping, the court emphasized that leveraging is a broad concept whose competitive impact may be the same “irrespective of the direction of the leveraging effect” (para. 88). It also held that the Guidelines on non-horizontal mergers do not prevent the Commission from pursuing novel theories of harm, particularly in digital markets that may present concerns not fully anticipated when the Guidelines were adopted (para. 77). The Commission was therefore free to develop and rely on a reverse-leveraging theory (para. 90).

Finally, the court rejected Booking’s reliance on “competition on the merits,” a concept developed under EU rules governing anticompetitive agreements and abuses of dominance. That concept, whose boundaries remain unsettled even in those areas, does not play the same role in merger control. A merger is not itself a form of competition on the merits. What matters is how the transaction would alter the structure of competition (paras. 106–107).

A Theory With Baggage

Unless the European Court of Justice reverses General Court’s judgment on appeal, the decision’s implications will reach far beyond this transaction.

The ecosystem theory endorsed in Booking/eTraveli closely resembles the Commission’s established portfolio-effects theory, developed in its 1997 Guinness/Grand Metropolitan decision and applied more recently in Mars/Kellanova. That theory already forms part of the traditional analysis of conglomerate mergers.

The draft Merger Guidelines describe much the same one-stop-shopping mechanism. A merger may significantly impede effective competition when it combines products from different markets into a portfolio sold to the same customers. That may be true even when the products are neither substitutes nor complements and the merger would not foreclose rivals’ access to customers or supplies (para. 287). The Guidelines likewise identify ecosystems and product or service portfolios as potential barriers to entry or expansion that can shape future competition (para. 40, fn. 74, and para. 79).

Seen in that light, Booking/eTraveli does not invent an ecosystem theory from scratch. It adapts the familiar portfolio-effects framework to digital markets.

The more novel development is the broad latitude the judgment gives the Commission to devise new theories of harm in response to digital markets’ supposedly distinctive features. That flexibility could become troubling if it produces a “killing ecosystems” approach—an enforcement bias against firms that create, expand, or strengthen integrated offerings.

There are legitimate reasons to ask whether the Commission gave enough weight to the acquisition’s potential benefits, even though the court attributed its failure to consider some of them to procedural shortcomings (paras. 482–489). As the draft Guidelines themselves emphasize, merger control does not protect competitors from efficient competition. A transaction does not become anticompetitive merely because its efficiencies leave rivals worse off (para. 21, fn. 38).

A “killing ecosystems” approach would also misunderstand how digital competition is evolving. Firms increasingly compete within and among ecosystems, while recent scholarship and antitrust authorities recognize that artificial intelligence is accelerating the shift from standalone products toward integrated services. Treating ecosystem building itself as suspect would risk confusing integration with foreclosure—and harm to rivals with harm to competition.

Merger control should police anticompetitive moats. It should not punish firms for building a better destination.

The post All-Inclusive Antitrust: Europe’s New Route to Merger Prohibition appeared first on Truth on the Market.

On Aug. 14, after years of litigation with Epic Games, Apple submitted a proposed commission structure to the U.S. District Court for the Northern District of California. Under the proposal, Apple would collect a 15% commission on purchases made through links from iOS apps and 10% on subscription renewals. Small businesses participating in Apple’s video, news, and mini-app programs would pay 5%. 

Four days later, Apple announced new terms for developers in the European Union, following what the European Commission called a “close dialogue.” Apple would charge 26% on App Store sales made through its in-app purchase system, 20% when developers use an alternative payment processor within an app, and 15% on purchases made through external links. It would also impose a new 5% “Core Technology Commission” on apps distributed through rival marketplaces or the open web. Reduced rates of 15% and 10% would apply to small businesses, partner programs, and subscriptions after the first year. 

The parallels are hard to miss. Both the U.S. and EU regimes ostensibly seek to protect competition, yet both pursue that goal by dictating the rates Apple may charge developers that rely on its platform and technology. That looks less like conventional competition enforcement and more like price regulation. 

This raises several important policy questions. Are mobile ecosystems so uncompetitive as to warrant rate-setting, which is widely regarded as a policy of last resort? If so, do U.S. antitrust law and the DMA give courts and competition agencies the proper tools for the job?

The stakes extend beyond the distributional question of how Apple and Epic divide a fixed pie. Poorly designed price controls can weaken competition and innovation over time, ultimately harming the very consumers that enforcers ostensibly mean to protect. 

Competition, Market Power, and Price Regulation

Markets generally allocate resources efficiently unless an identifiable “market failure” gets in the way. Classic examples include externalities, information asymmetries, public goods, and natural monopolies. Even in those exceptional cases, price regulation is usually considered a remedy of last resort.

After all, telling a private firm what it may charge is among the most intrusive tools in the competition-policy toolbox. The traditional justification for price regulation is therefore correspondingly narrow: a natural monopoly. That occurs when economies of scale allow one firm to serve an entire market at a lower average cost than two or more firms could, leaving competition unable to push prices toward costs. 

Mobile operating systems do not fit that description. Start with one of the most exhaustive judicial examinations of the issue. In Epic Games v. Apple, the district court rejected the “market of one” theory, defined the relevant market as digital mobile-gaming transactions, and found that Apple’s 52% to 57% share did not establish monopoly power. 

The reason was interbrand competition—that is, competition among rival brands and ecosystems. Dario Oliveira Neto and Mario Zúñiga have noted that Brazil’s Administrative Council for Economic Defense (CADE), for example, effectively made Apple a monopolist by definition when it limited the relevant market to “the non-licensable mobile operating system iOS.” But consumers do not buy an operating system in isolation.

Instead, they buy a bundled product that includes hardware, software, security, and an ecosystem of apps and services. Apple therefore competes directly with Samsung, Xiaomi, and Huawei for smartphone buyers. In the first half of 2025, Apple held 62% of Brazil’s “premium” segment, compared with 20% for Samsung and 8% for Huawei. Those are hardly the numbers of an unconstrained monopolist.

The smartphone ecosystem is better understood as an oligopoly: a market dominated by a small number of rivals. Apple’s iOS and Google’s Android compete vigorously and have repeatedly leapfrogged one another on features, privacy, and security. They now even offer tools designed to make switching between the two easier. The International Center for Law & Economics (ICLE) has made this point in submissions to CADE, the United Kingdom’s Competition and Markets Authority (CMA), and the Japan Fair Trade Commission (JFTC). 

The CMA offers a counterargument that characterizes Apple and Google as a “stable duopoly,” citing low switching rates and the fact that most users do not consider alternatives when replacing their devices. But low observed switching tells us little by itself about the intensity of competition.

Consumers may stay because they are satisfied, not because they are trapped. Conversely, high switching can coexist with monopoly power when consumers respond to prices that a monopolist has already pushed above competitive levels—a version of what competition policy deems the “cellophane fallacy.” 

Ultimately, what disciplines a platform is the credible threat that users will switch if prices rise or quality falls. If competition keeps prices low and quality high, few users may actually leave. Switching rates alone cannot reveal that competitive pressure. 

Can Courts Set the Right Price?

Suppose, for the sake of argument, that one thinks Apple really is a monopolist, that App Store fees really are “too high,” and that someone should bring them down. The next question is who that someone should be. 

In Verizon Communications Inc. v. Law Offices of Curtis V. Trinko LLP (2004), Justice Antonin Scalia warned that: 

Enforced sharing also requires antitrust courts to act as central planners, identifying the proper price, quantity, and other terms of dealing—a role for which they are ill suited. 

Scalia added:

Even if the problem of false positives did not exist, conduct consisting of anticompetitive violations … may be, as we have concluded with respect to above-cost predatory pricing schemes, “beyond the practical ability of a judicial tribunal to control.”

The Court then quoted antitrust scholar Phillip Areeda’s admonition: 

No court should impose a duty to deal that it cannot explain or adequately and reasonably supervise. 

When compulsory access requires “the day-to-day controls characteristic of a regulatory agency,” Areeda argued, courts should treat the problem as beyond antitrust law’s ability to remedy. 

The point is that mandating access almost inevitably requires setting a price for that access. Doing so demands judgments about costs, investments, risks, product quality, and the value of supporting infrastructure. Those judgments must then be revisited as technological and market conditions change. Courts and competition agencies rarely have the information, expertise, or institutional capacity needed for the job. 

Trinko also explains why monopoly prices are not unlawful by themselves:

The mere possession of monopoly power, and the concomitant charging of monopoly prices, is not only not unlawful; it is an important element of the free-market system. 

The prospect of earning those returns, the Court reasoned, “induces risk taking that produces innovation and economic growth.” That principle clashes with the currently fashionable view that returns above cost are defects for competition authorities to engineer away. 

Judge Yvonne Gonzalez Rogers, who oversees the Apple litigation in the United States, must now decide whether commissions of 15%, 10%, or 5% are acceptable. That means ascertaining the value of Apple’s distribution services, payment infrastructure, and application programming interfaces (APIs)—the tools that allow apps to interact with Apple’s operating system. For its part, Epic contends that, under the 9th U.S. Circuit Court of Appeals’ language concerning costs, the correct figure is 0%. 

This is precisely the kind of judicial price-setting that Trinko warned against. Its rule does not control here, as the Apple case ultimately turned on California’s Unfair Competition Law, but its institutional warning applies all the same.

The Rate-Setter Without a Rate Book

Brussels and the DMA don’t offer an obvious answer to the “who” question either. The DMA gives the European Commission a broad remedial toolkit, but having the legal authority does not answer Areeda’s underlying concern. The problems here concern institutional capacity and information, not jurisdiction. The DMA granting the Commission authority to set rates is no substitute for knowing what the right rates are. 

Proponents of digital-competition regulation initially insisted that the DMA would not become a price-control regime. Ioannis Lianos, Klaas Hendrik Eller, and Tobias Kleinschmitt argued that the DMA “does not aim to regulate entry or rates/output, as does traditional utility regulation,” but instead “sets some bright-line rules for business conduct” (p. 43). 

Others pressing for gatekeeper rules shared that understanding. Writing in the CPI Antitrust Chronicle in 2021, Thomas Höppner argued that gatekeepers’ “tollbooths for market access” justified asymmetric regulation—special rules that apply only to designated gatekeepers. But he located the remedy in conduct rules, not price controls: 

There is no need to outright prohibit any form of advertising or paid intermediation for gatekeepers. Neither is there a need to regulate their prices for such intermediation. In general, auction mechanisms are a competitive tool to determine an adequate price. 

Höppner was writing about sponsored rankings rather than app-store commissions, but his reasoning applies in this context, as well. 

Those reassurances are at odds with the DMA as enacted. Article 6(12) expressly contemplates fair, reasonable, and nondiscriminatory (FRAND) access terms and directs the Commission to assess them. Recital 62 specifically identifies pricing as part of that assessment. Article 5(4) also requires gatekeepers to let business users steer customers toward rival services “free of charge.” 

Even that seemingly straightforward language raises difficult questions. Epic Games and other rivals interpret “free of charge” to mean that Apple and Google may not impose any fee, even one only loosely connected to steering. The Commission, by contrast, accepts that Apple may charge when it facilitates a user’s initial acquisition. That is a sensible compromise because it prevents rival marketplaces from free-riding on Apple’s investment in attracting users. 

But the compromise simply tees up the harder question: How much may Apple charge? Once regulators create a right to steer customers elsewhere, they must decide what compensation, if any, the platform may collect. 

The Commission made its rate-setting role explicit in an April 2025 noncompliance decision, which imposed a €500 million fine. The decision also established a three-part test. Any fee must relate only to the initial acquisition, correspond to the value of that acquisition after subtracting compensation the gatekeeper already receives for facilitating it, and not “remunerate the gatekeeper for gatekeeper value.” 

That final restriction goes too far. By excluding compensation for the platform’s value as a gatekeeper, the test effectively expropriates part of Apple’s investment. As Lazar Radic puts it, this amounts to ex ante price regulation—without the institutional machinery that rate-setting ordinarily requires. 

The DMA compounds the problem because it lacks a commitments procedure analogous to Article 9 of Regulation 1/2003, which allows the Commission to accept a company’s proposed remedies and formally close a case. Here, there is no reasoned decision explaining the accommodation, no press release announcing a resolution, and no closure of the April 2025 proceedings. They remain formally open, leaving no final action for anyone to appeal. 

It’s also conspicuous that the same numbers keep appearing elsewhere. CADE’s December 2025 settlement in Brazil produced a 5% Core Technology Commission and a 15% fee on purchases through external links, months before Brussels blessed the same figures. The CMA is now drafting an explicit cost methodology, complete with a “market power adjustment,” while conceding that it cannot identify a “reasonable rate of return.” 

In other words, four separate authorities have converged on essentially the same rates without a discernible methodology for deriving them. That suggests the goal is a symbolic victory, not the price most likely to promote competition and consumer welfare. 

The Bill Comes Due

None of this would matter much if the only question were how Apple and Epic get to divide a fixed pot of money. But that’s not the case. 

The broader stakes emerge from the free-riding problem that Brian Albrecht and Dirk Auer describe. Operating systems, app stores, and payment systems are not separate “bottlenecks” or choke points. They are parts of an integrated ecosystem funded largely by commissions. That business model has helped millions of consumers gain access to cheaper devices.  

Restricting a platform’s ability to earn revenue from one part of that ecosystem can ripple through the rest. When authorities require platforms to permit alternative distribution free of charge, rival stores can use the platform’s APIs, security protections, and customer base without helping to cover the costs of building them.

Free steering of payments creates a similar problem. A developer can use an app store’s discovery and curation services to find a customer, then route that customer’s purchases elsewhere. The store bears the cost of bringing the parties together but loses the revenue that made the service worth providing.

That problem hits hardest for the entrants that these rules were supposedly enacted to help. A new app store launching on iOS must invest in curation, security reviews, fraud prevention, and customer support. Yet it faces the same free-riding problem, because its developers can steer customers elsewhere, too.

A coherent policy must decide how a trusted marketplace will recover those fixed costs. Neither the DMA nor the district court has done so. Both have instead pushed permissible fees toward incremental cost—the cost of handling one additional transaction—across every layer of the ecosystem at once. 

Returns above incremental cost are not a regulatory defect that authorities should seek to engineer away. As ICLE told the CMA, those returns finance the substantial upfront costs of creating platforms, much as they finance the creation of intellectual property. 

Investors fund improvements based on the returns they expect those improvements to generate. If regulation caps revenue from steered transactions near incremental cost, it tells investors that much of the upside from building a better platform will flow to someone else.

Cost-plus regulation therefore optimizes for the wrong kind of competition. It pursues static allocative efficiency by squeezing today’s margins. But consumer welfare in mobile ecosystems has come largely from dynamic competition—competition to develop new devices, capabilities, and form factors over time. That requires sustained reinvestment. 

Consumers will not remember the year an app-store commission fell by four percentage points. They will notice when quality stalls, or when promised features never arrive. 

Some of those costs are already visible. Apple has delayed or withheld iPhone Mirroring, AirPods Live Translation, and its revamped Siri artificial-intelligence system in the European Union. Google has withheld AI Overviews.

Regulatory compliance has also consumed one of the scarcest resources at leading tech firms: engineering talent. The Commission’s impact assessment had projected a total compliance cost of about €10 million for all gatekeepers. In reality, Meta reports assigning more than 11,000 employees and nearly 600,000 engineering hours to compliance. Google devoted about 3,000 engineers full time for two years to complying with a single article. 

And what have consumers received in return? Apple’s November 2025 study examined more than 41 million EU transactions. It found that after commissions fell by roughly 10 percentage points, 91% of products either became more expensive or saw no price reduction. Non-EU developers captured 86% of the €20.1 million in estimated savings. 

Apple commissioned the study, so its findings deserve appropriate scrutiny. But it remains the only transaction-level evidence available. It also survives a placebo test—a check designed to detect results that may be spurious—and no one has rebutted it. 

Rather than lower prices, the apparent result is a transfer of economic rents from the platform to a largely non-European group of bigger developers, while consumers remain roughly where they started. 

If consumers do not benefit, a lower commission is not a victory of competition policy. It’s just a new way to divide the spoils. 

The post Rate Expectations: Apple and the New Price Regulators appeared first on Truth on the Market.

The label may sound technical, but it captures a central problem for digital regulation: What happens when a service does too many things to fit neatly into any one legal category?

That question sits at the heart of today’s competition-policy debate over whether the wave of digital-market regulations adopted in recent years remains fit for purpose. The rise of artificial intelligence (AI), particularly AI-enabled applications, is already testing how well these frameworks can adapt to technologies their drafters did not anticipate.

The existing regimes face two related challenges. First, regulators must determine how to classify AI applications that become new gateways to digital services. Second, they must decide how to respond as new AI providers reshape competition and perhaps acquire gatekeeping power of their own. AI assistants and agents—tools that can perform multistep tasks on a user’s behalf—could upend familiar patterns of web browsing, online search, and e-commerce.

Of course, the competitive effects could cut in either direction. Incumbent platforms could integrate AI features in ways that shut out emerging providers. Alternative, new entrants could use those same technologies to challenge or even displace heretofore entrenched firms. Today’s disruptors, in other words, could become tomorrow’s gatekeepers.

Both possibilities expose the same basic problem: AI services do not fit neatly into the regulatory boxes that governments have built for digital markets. They increasingly behave like platforms, connecting users with other services and performing functions associated with search engines, online marketplaces, and virtual assistants. Yet frameworks organized around those predefined categories struggle to capture services that combine all three.

The DMA’s AI Holding Pattern

Against this backdrop, it’s notable that the European Commission chose a wait-and-see approach in its first review of the Digital Markets Act (DMA).

In its late-April report and accompanying staff working document, the Commission recognized that AI demands particular regulatory attention because it could become one of the most powerful general-purpose technologies. But AI remains at a relatively early stage, and its uses continue to evolve rapidly.

Given that uncertainty, the Commission declined to expand the DMA’s list of “core platform services”—the categories of important digital gateways the law covers. Instead, it indicated that it plans to monitor AI developments closely. The DMA would, of course, continue to apply when a gatekeeper integrates AI into an already designated core platform service. Moreover, a standalone AI service could fall within the law if it independently meets the criteria for designation.

The Commission added that it would examine, in particular, whether existing categories like virtual assistants already cover some AI services. It will also consider whether the DMA’s current mechanisms can adapt the obligations imposed on AI services as technology and market conditions evolve.

For now, then, the Commission’s official verdict is that the DMA can handle AI without an overhaul. It considers the existing framework adequate to address AI-related conduct by designated gatekeepers and the addition of AI features to core platform services.

The DSA Jumps the Queue

Against that backdrop, the Commission’s DSA designation for ChatGPT looks more than a little surprising. Under the DMA, the Commission counseled patience. Under the DSA, it acted.

To get there, the Commission focused on function rather than form. It did not attempt to classify ChatGPT in its entirety. Instead, it treated the service’s web-search feature as sufficient to bring ChatGPT within the DSA’s definition of an online search engine. The result is a hybrid classification for a hybrid service.

The distinction matters because the DSA and DMA together form the main pillars of the European Union’s digital rulebook. The DSA addresses platform accountability and systemic risks, while the DMA targets competition involving designated gatekeepers. Their aims differ, but their boundaries overlap.

That overlap has competitive consequences. The designation subjects ChatGPT to obligations that do not currently apply to rival AI providers such as Anthropic. It could therefore tilt the regulatory playing field. That is principally a competition concern—and thus more naturally the DMA’s business—but it arises from a decision under the DSA.

This mismatch raises two questions. Why bring an AI service promptly within the DSA while maintaining a wait-and-see approach under the DMA? And does the designation suggest that the Commission has reconsidered how to classify AI applications under the DMA?

The second question is particularly thorny because both laws define online search engines in identical terms. Yet during the DMA review, the Commission identified “virtual assistants,” rather than search engines, as the possible category for standalone AI services.

Meanwhile, a separate Commission action further deepens the uncertainty. In a July specification proceeding concerning Google’s obligation to share search data, the Commission concluded that AI chatbots with search functions could access data made available under Article 6(11) of the DMA. That position suggests the existing DMA framework can accommodate services that perform functions comparable to online search engines—at least when those services seek the law’s benefits, if not yet when regulators impose its obligations.

The Regulatory Sorting Hat

ChatGPT’s designation under the DSA exposes a basic problem with technology regulation: The rules age much faster than the technology. AI applications evolve quickly, combine functions once performed by separate services, and resist frameworks built around rigid categories.

Forcing these services into existing legal taxonomies may solve an immediate jurisdictional problem, but it can undermine regulatory coherence. It also invites “strategic designation”—choosing the legal category that brings a service within a particular regulator’s ambit. That risk grows when different agencies enforce overlapping regimes.

The Italian cases involving DeepSeek, Mistral, and Nova offer a concrete illustration of the problem, as the Italian Competition Authority and the Italian Communications Authority disagreed over how the services should be classified under the DSA.

The Communications Authority viewed AI agents as AI-powered search engines and, therefore, as providers of intermediary services under the DSA. That classification would place them under the authority’s supervision in its role as Italy’s Digital Services Coordinator. The Competition Authority rejected that interpretation. It concluded that an optional web-search feature, activated by the user, does not by itself turn an AI chatbot into an online search engine or intermediary service under the DSA.

Whether the DMA will follow the example set by ChatGPT’s designation under the DSA remains an open question. Europe’s harder task is to keep its rules coherent when AI refuses to stay in one box.

The post One AI, Two Rulebooks: Brussels’ ChatGPT Conundrum appeared first on Truth on the Market.

Yet economists who analyze contracts have focused on a related but distinct issue: the many events for which one might want to account in a contract which cannot be verified by external authorities (e.g., courts). If performance terms cannot be verified, it would appear meaningless to include them in a contract.

Enter Robert E. Scott and George G. Triantis’ 2005 article “Incomplete Contracts and the Theory of Contract Design.” Real-world contracts, they observe, often address contingencies that courts cannot verify. This fact suggests that much of the economic literature on contracts is incomplete at best and wrong at worst. Here, I outline the paper’s core arguments: its account of the economic literature on contract design, its critique of that literature, and its alternative theory of incomplete contracting. 

Why Contracts Are Incomplete

Many trades unfold over time. Purchases from Amazon, purchases of a home, and purchases of a business all involve a delay between promise and performance. 

Trade across time creates a problem. In the intervening period, circumstances may change and one party may want to go back on his promise. Without assurances against reneging, no deal will be struck. For example, a seller will refrain from trade unless he is assured that the buyer will not back out at the last moment. Gains from trade then go uncaptured. 

Contracts help solve this commitment problem. They define what each party has promised and make those promises legally binding. The prospect of legal sanctions makes parties less likely to renege. When buyers and sellers face less risk of being burned, they become more willing to trade. Ceteris paribus, enforceable promises promote exchange. 

Parties may have various reasons to want enforceable promises. Economic theory has focused on one circumstance in particular: investments made before performance. As Scott and Triantis explain, “one or both of the parties may . . . make investments in anticipation of the exchange that will increase the exchange value by either (a) lowering the cost of performance or (b) raising the benefit from performance” (Scott & Triantis, p. 188).

Suppose, for example, that a cabinetmaker agrees to build shelving tailored to the unusual dimensions of a customer’s alcove. Once he has bought the lumber and made his first cuts, the half-finished piece is worth almost nothing to any other buyer. The delay between investment and performance gives the customer a chance to “hold up” the exchange and renegotiate more favorable terms. Legally binding promises deter hold-up by making the threat to walk away less credible. 

In a world where investments precede performance and the future is uncertain, contract drafters worried about the possibility of reneging must balance ex ante efficiency and ex post efficiency. A contract that encourages the “optimal” amount of specific investments promotes ex ante efficiency. So-called “specific investments” or “reliance expenditures” are less likely to be made without contractual assurance. As a result, gains from trade may go uncaptured.  

The problem is that such a contract may inhibit ex post efficiency. Once the future is known, the non-investor may find that performance no longer makes economic sense. Terms that promote ex ante efficiency may undermine ex post efficiency by “compelling exchange when there is no surplus to be gained” (189).  

Parties to a trade could protect ex post efficiency through, for example, an agreement to renegotiate once the future is known. That flexibility can help avoid compelling “inefficient” trades.  

But flexibility is no free lunch. The prospect of renegotiation “may expose the party that has made sunk cost investments to the risk of exploitation by the noninvesting party” (p. 189). The noninvesting party may threaten to withhold performance unless the investor accepts worse terms. A prospective homebuyer, for example, may have incentive to get the inspector to find “major defects” in a home. That threat—the threat of recall—would reduce the gains from trade, as the investor is less willing to make any upfront investment that enhances the value of the exchange. Contracts that promote ex post efficiency can therefore undermine ex ante efficiency. 

Economic theory offers an idealized solution to this trade-off: the “complete, contingent contract” (p. 189). Such a contract specifies the efficient obligation in each possible state of the world. It elides the tradeoff by stipulating ex ante investments only when they create value and stipulating performance only when it creates value. Because the contract is already efficient in every state, neither party ever wants to renegotiate. In doing so a complete contract simultaneously protects incentives to invest beforehand and the flexibility to walk away from trades that turn out to be wasteful.

The idealized solution is just that: idealized. Real-world transaction costs keep parties from writing the complete contract. Scott and Triantis identify two important types of transaction costs.

First, parties incur costs at the “front-end stage.” They must devote resources to “anticipating future contingencies and writing a contract that specifies an outcome for each one” (p. 190). As Scott and Triantis explain, “On the front end, the parties might not foresee all possible contingencies or they would have to incur prohibitively high negotiation and drafting costs to partition all contingencies sufficiently to provide for efficient obligations in each case” (p. 191). 

Second, parties incur costs at the “back-end stage.” These include “the costs of observing and proving the existence (or nonexistence) of any relevant fact after uncertainty has been resolved” (p. 190). Scott and Triantis write: 

On the back end, contracts that provide optimal obligations for all contingencies may be too costly to enforce because they require the court to distinguish among too many possible states of the world, some of which may be known only to one party or known to the parties but not the court. (p. 191)

Both front-end and back-end costs can explain why contracts contain so many gaps. A contract may contain gaps because it was too costly to anticipate and write down all of the relevant contingencies. It also may have gaps because it was too costly for the court to verify what happened and enforce the written promises. Contracting parties who expect costly terms to be unenforceable would not include them. 

The Unverifiability Assumption

Scott and Triantis observe that economic theorists have taken contractual incompleteness as a given for decades. Theorists often assume back-end costs are prohibitive. In particular, they assume that some states of the world are not verifiable by a court. For example, the level of demand for a good at the time of delivery may be plain to both buyer and seller but impossible for an outside court to confirm. This means that even if a contract anticipates such an outcome—say, a price set low if demand turns out low and high if demand turns out high—the court could not enforce the contract’s obligations. Because the court has no way to determine which state actually occurred, it therefore has no way to determine what promises to enforce. 

By Scott and Triantis’ logic, economic theorists have assumed incompleteness rather than explained it. This is likely due to the theorists’ focus. Instead of asking why contracts are incomplete, theorists have analyzed a related but different question: how can parties design contracts given that some terms are by definition unenforceable (read: unverifiable) by courts?  

In such models courts are neither gap-fillers nor enforcers. They have no role to play. Renegotiation therefore plays a key role. When some terms are unenforceable, renegotiation is the primary way to promote ex post efficiency. Renegotiation lets parties bargain after uncertainty resolves.

Recall, however, that when terms are unenforceable, renegotiation can also inhibit ex ante efficient investments. Reopening the bargain ex post means reopening the division of the surplus. As the non-investor can capture part of those returns without having borne any of the cost, the investor anticipates being “held up” and invests too little. In such a world, the problem of contract design boils down to allocating ex post bargaining power in a way that protects both ex ante investment incentives and ex post efficiency.  

Theorists’ use of unverifiability has allowed them to examine renegotiation’s implications for contract design. But Scott and Triantis argue that the assumption of unverifiability is not innocuous. It yields predictions that are at odds with observed contracting practice.  

First, standard “contract theory simply posits that some factors—occurrence of contingencies and performance of obligations—are ‘not verifiable’ by a court” (p. 195). Unverifiability means that it is impossible for a court to tell whether some actions, investments, or states of the world actually materialized. For example, the level of effort that a franchisee put into keeping her restaurant clean is assumed to be unverifiable by a court.  

The theory predicts that parties to a contract will never include terms that are conditioned on unverifiable states of the world. Such terms would be superfluous. They define obligations in a world that the court cannot enforce. Including them yields no benefit. And since including them is costly, no rational party ever would.  

Scott and Triantis point out that the evidence refutes this prediction. Ostensibly “unverifiable” contingencies are actually commonplace. For example, terms such as “best efforts,” “reasonable care,” and “good faith” appear often in commercial contracts. Their prevalence suggests that unverifiability is not a realistic assumption.

Second, Scott and Triantis argue that the verifiability assumption also misunderstands what courts actually do when they enforce contracts. The concept of unverifiability, they note, implies that a court’s job is to find an objective truth. But that’s not how civil litigation actually works.

Consider criminal law. It uses an absolute measure like “beyond a reasonable doubt.” Thus the prosecution must establish guilt against a benchmark of near certainty. 

By contrast, courts in civil trials do not have an absolute standard to which the facts can be compared. Instead, civil courts evaluate evidence in relative terms. For example they consider the “preponderance of evidence” or “the balance of probabilities.” This means that civil courts weigh the persuasiveness of the plaintiff relative to that of the defendant, rather than to an external standard like “beyond a reasonable doubt.” A fact is therefore effectively “verified” when one party can out-evidence the other. Verifiability does not require that the underlying state of affairs be demonstrated with objective certainty. 

Third, Scott and Triantis question the common assumption that enforcement costs are exogenous. Contract theory’s unverifiability assumption treats the costs of enforcement as either zero or prohibitive. But as the authors point out, the cost of enforcing contract terms via litigation depends on each party’s litigation strategy. How much each wants to spend on enforcement via litigation depends on how much each expects their counterparty to spend.

This matters for two reasons. If the expected enforcement costs are prohibitive, the parties may choose to omit the term altogether. Alternatively, parties may further tailor the contract in anticipation of future litigation.

The Logic of Rational Incompleteness

In light of these three problems, Scott and Triantis offer a more general explanation for incomplete contracts that does not depend on unverifiability. In their framework, courts matter. 

Writing more complete contracts offers several benefits. As in prior economic theory, more complete contracts protect the incentives to make investments beforehand. Unlike that earlier work, Scott and Triantis argue that more complete contracts also reduce back-end costs. Recall that back-end costs involve the costs of litigating:

1. Any states that are omitted from the contract; and
2. Any obligations that turn out to be inefficient after the state of the world is realized.

When contracts are more complete, litigation costs are lower because courts have fewer gaps to fill and need to exert less interpretive effort divining what the parties would have agreed to if they knew the future. But writing more complete contracts is costly. Parties must spend resources:

1. Identifying possible future states of the world; and then
2. Writing efficient obligations for each known state of the world. 

Parties to a contract must therefore balance the marginal costs of more complete contracts against the marginal benefits. The equilibrium contract sets the marginal cost of additional specification equal to the marginal incentive gain, plus the marginal reduction in expected ex post enforcement costs.  

This alternative framework is important for three reasons. First, it treats unverifiability as a special case in which back-end costs are prohibitive. Second, it makes incompleteness endogenous, rather than taken as given. Third, unlike the standard economic theory of incomplete contracts, it can account for the abundance of vague terms in commercial contracts.

Vague terms help save on ex ante drafting costs: “By using broad standards such as ‘best efforts,’ the parties defer this task to the litigation stage” (p. 197).

Incompleteness by Design

Scott and Triantis’ piece clearly belongs in the law & economics canon. It is among the clearest summaries of incomplete-contract theory to date. Moreover, unlike earlier work in economic theory, Scott and Triantis offer a simple framework for explaining why some contracts are more complete than others.

But most importantly, they rightly call for models of incomplete contracts in which courts can matter. Doing so promises to yield more accurate predictions about when and where contracts will be more or less complete. In their account, contractual gaps are purposeful rather than folly. 

Further Reading

• Robert E. Scott & George G. Triantis, “Incomplete Contracts and the Theory of Contract Design,” Case Western Reserve Law Review, Vol. 56, No. 1 (2005).
• Robert E. Scott & George G. Triantis, “Anticipating Litigation in Contract Design,” Yale Law Journal, Vol. 115, No. 4 (2006).
• Oliver Hart & John Moore, “Property Rights and the Nature of the Firm,” Journal of Political Economy, Vol. 98, No. 6 (1990).
• Alan Schwartz & Robert E. Scott, “Contract Theory and the Limits of Contract Law,” Yale Law Journal, Vol. 113, No. 3 (2003).
• Sanford J. Grossman & Oliver D. Hart, “The Costs and Benefits of Ownership: A Theory of Vertical and Lateral Integration,” Journal of Political Economy, Vol. 94, No. 4 (1986).
• Jean Tirole, “Incomplete Contracts: Where Do We Stand?,” Econometrica, Vol. 67, No. 4 (1999).
• Henry A. Thompson, “AI and the Law,” Kyklos, Vol. 79, No. 1 (2026).

The post ‘Incomplete Contracts and the Theory of Contract Design,’ by Robert Scott and George Triantis appeared first on Truth on the Market.

Last month, New York City Mayor Zohran Mamdani formally endorsed the Delivery Protection Act (DPA), a City Council bill sponsored by Council Member Tiffany Cabán. Though framed as a warehouse-licensing and safety measure, the bill would also require Amazon to employ its delivery drivers directly rather than contract with independent delivery businesses. 

Just six days before Mamdani’s announcement, New Jersey sued Amazon over substantially the same delivery arrangements. The state alleges that Amazon uses its control over Delivery Service Partners (DSPs) to suppress compensation and worsen working conditions, including by pressuring drivers to meet unsafe delivery expectations. The geographic markets alleged in the lawsuit include New York City.

Both governments portray Amazon as directing the work while DSPs bear formal responsibility. Yet New Jersey treats limits on DSP independence as part of the problem, while New York would virtually eliminate that independence. Different statutes explain the different legal claims, of course. But they don’t explain how substantially the same conditions and harms can result from too little DSP independence in New Jersey, yet be cured by eliminating DSPs as employers in New York. 

That contradiction frames the DPA’s central problem. Its supporters haven’t shown that DSP employment causes the harms they cite, or that direct employment would reduce them. Amazon’s extensive safety programs further complicate the case for abolishing the DSP model in the name of safety. What the mandate would unquestionably change is how New York City delivery drivers unionize and who they bargain with.

Rewriting Amazon’s Delivery Route

The DPA would create a licensing regime for last-mile warehouses—the facilities that handle the final leg of deliveries to customers—operated by services such as Amazon and United Parcel Service (UPS). It would also impose training, recordkeeping, discharge, and safety rules.

But the bill’s most consequential provision has little to do with licensing or street safety. After a phase-in period, it would require warehouse operators to directly employ workers who perform covered warehouse and delivery services, barring operators from contracting with third parties for that work.

For Amazon, that would mean ending its DSP program for covered work in New York City, subject to a revealing exception for collective bargaining discussed below. By Amazon’s count, more than 40 local DSPs employ over 5,000 drivers.

The bill would require facility operators to offer jobs to DSP workers terminated because of the required contract cancellations before hiring anyone else for the same work. But it wouldn’t preserve the DSP owners’ existing delivery agreements. These mostly small business owners could no longer perform work that Amazon would have to bring in-house.

Amazon also warned in City Council testimony that it would consider moving facilities outside the city (obvi, as the kids say), so the bill can’t guarantee that every route or job would survive, much less remain in New York City.

That’s a remarkably invasive response to unsafe driving or workplace injuries, and it demands persuasive evidence. As the bill’s supporting reports show, the city hasn’t supplied it. 

The Nature of the Mandate

The DPA’s licensing provisions would apply to covered facilities regardless of how they organize delivery work. The New York City Department of Consumer and Worker Protection could suspend or revoke a license for noncompliance or a qualifying pattern of workplace, road-safety, environmental, consumer-protection, or worker-protection violations. Those provisions at least address the conduct the bill is supposed to target. The direct-employment mandate, by contrast, would prohibit a specific organizational arrangement regardless of the operator’s safety record.

UPS already employs its drivers directly, so the provision that would remake Amazon’s delivery operation in New York City would require little or no change from UPS. And—no doubt entirely coincidentally—UPS drivers belong to the Teamsters, whose Package Division identifies UPS as the union’s largest employer. The City Council would thus just happen to mandate the arrangement already used by Amazon’s unionized competitor. Whether that arrangement makes delivery safer is the question the bill’s supporters still need to answer.

As Ronald Coase explains in the classic law & economics essay “The Nature of the Firm,” firms compare the costs of organizing work internally with those of contracting for it. Amazon’s DSP program combines access to the company’s logistics network and technology with local responsibility for running a delivery business.

The arrangement may sometimes be suboptimal, of course. But it may also allow small businesses that couldn’t replicate Amazon’s logistics network to enter the delivery market, while leaving their owners room to experiment with hiring and supervision. Amazon’s common routes, safety standards, and technology don’t necessarily negate those advantages. Indeed, that coordination helps make the arrangement possible. Yet the bill would eliminate this choice without demonstrating that any resulting benefits would justify the loss.

A Crash Course in Causation

The mayor’s public case relies heavily on the New York City comptroller’s November 2025 report, “Fast Shipping. Slow Justice,” which recommended passage of an earlier version of the bill. The report argues that direct employment would improve safety, but its empirical analysis doesn’t isolate the effect of the employment arrangement on accidents. Evidence that delivery work can be dangerous doesn’t establish that DSP arrangements, in particular, cause the danger.

The mayor’s announcement repeats the report’s finding that 78% of areas within a half-mile of last-mile facilities experienced more injury-causing crashes after a facility opened. In other words, crashes rose near 14 of 18 facilities in a simple before-and-after comparison. The analysis includes no control group of comparable areas without new facilities, no adjustment for citywide trends, and no attempt to distinguish delivery drivers’ conduct from the additional traffic and new traffic patterns that accompany a new facility.

What’s more, most of the facilities opened in 2020 or later, so the “after” period also captures pandemic-era changes in e-commerce and delivery patterns. The report also cautions that its data don’t attribute individual crashes to vehicles connected with a last-mile facility.

Unsurprisingly, the data are decidedly mixed even on their face. Four sites recorded fewer crashes after opening, and three were Amazon facilities operating under the DSP model that the bill would ban. The largest increase in the raw number of crashes occurred near a UPS facility. Unlike the bill’s proponents, I wouldn’t go so far as to claim that these observations establish that the DSP employment arrangement is actually safer—but they sure don’t support the opposite claim. The crash data record what happened around warehouses. They tell us nothing about what would happen if Amazon employed its drivers directly.

The worker-injury data come closer, but they still don’t answer the relevant question. The report found that the Amazon DSPs it identified had higher reported injury rates than other New York City courier employers in 2023 and 2024. But it doesn’t control for the types of tasks workers perform, work intensity, tenure, or other employer-specific differences that may correlate much more strongly with injuries.

More fundamentally, the comparison group isn’t limited to drivers whom warehouse operators employ directly. It includes a range of employers in the broadly defined courier industry, rather than drivers doing similar work under different employment arrangements. The correlation may be somewhat suggestive, but it doesn’t come close to establishing that direct employment would reduce injuries.

Bill supporters have also cited Teamsters Local 804’s report, “The Last Mile in New York City.” Its analysis groups what it calls “e-commerce” and “legacy” delivery vehicles by make rather than by the company that employs their drivers. It then compares ticket counts without accounting for fleet size or miles traveled. That approach can’t distinguish a more dangerous employment arrangement from a larger or busier fleet. Its separate Amazon-specific analysis provides no direct-employment comparison, either. Neither analysis establishes that abolishing DSP employment would make deliveries safer.

One Company’s Safety Is Another’s Surveillance

The comptroller nevertheless treats UPS as the safety benchmark. The report cites UPS’ reported annual spending of more than $200 million on safety training and says the company has “invested heavily in data systems to optimize delivery schedules and reduce exposures to injuries sustained in driving.” It presents those investments as evidence of the incentives that direct employment creates.

To support its claim that UPS provides “more robust safety training,” the report cites only a 2019 ProPublica/BuzzFeed investigation. That investigation actually catalogs a wide range of safety measures Amazon implemented in 2018 as part of a company-wide safety initiative. Yet it largely dismisses the program based on reported implementation problems as Amazon rapidly expanded its operations.

Nowhere does the comptroller’s report assess Amazon’s current systems. Amazon reports spending more than $2.5 billion on safety across its operations since 2019, along with a 43% decline in its global incident rate between 2019 and 2025 (not limited to DSP-driver injuries, of course).

For transportation safety specifically, Amazon announced $200 million in technology spending for 2023. The company highlighted camera alerts, driver coaching, and its internally designed “Fleet Edge” system, which provides much faster updates about road conditions. Amazon’s testimony also reports that more than 180,000 drivers have graduated from its in-person driver academy. The report leaves these substantial programs out of its comparison.

The report does acknowledge Amazon’s training and monitoring elsewhere—but only to criticize the company’s alleged “power” over drivers. New Jersey likewise invokes Amazon’s “GPS tracking, in-vehicle artificial intelligence, and cameras surveilling drivers” as evidence of control and describes working conditions that include “invasive surveillance.”

Safety programs and intrusive monitoring can coexist, of course. But implementing common safety practices requires some authority over how drivers make deliveries. The comptroller praises UPS for exercising that authority. New Jersey treats Amazon’s exercise of it as emblematic of anticompetitive control.

Contentious characterizations aside, Amazon’s contracts apparently already permit substantial safety coordination, and the company has chosen to invest substantial sums in it. No evidence in the reports—or elsewhere—establishes that replacing DSPs with direct employment would improve safety. That’s a considerable omission from the case offered to justify abolishing an entire way of doing business.

Safety, Subject to Collective Bargaining

Section 20-566.8 may inadvertently reveal the bill’s real driver (no pun intended). Its collective-bargaining exception undermines the DPA’s entire stated purpose. The provision would allow waivers of the direct-employment mandate and every one of the bill’s safety rules—but not the licensing requirement, because we can’t cut government entirely out of the loop—if a collective-bargaining agreement with a “bona fide labor organization” grants them in exchange for “valuable consideration.”

No one in the history of contracting has used “valuable consideration” as a euphemism for “increased public safety.” Nor does the bill require any finding that such waivers would include safety guarantees. Apparently, the organizational arrangement deemed essential to safety becomes merely a concession to negotiate once a union enters the picture.

Indeed, the bill’s union-organizing rationale is considerably easier to follow than its safety arguments. The Teamsters say the bill would “put a stop to the abusive DSP model” and call direct employment necessary to end Amazon’s “exploitation of its workers.”

Direct employment would make Amazon itself responsible for collective bargaining, without requiring the union to first establish that Amazon is legally responsible for a DSP’s employees. The Teamsters could negotiate with a single firm, gaining greater leverage over a company with strong reputational incentives and substantial resources to fund concessions.

Direct employment would also prevent Amazon from ending its relationship with an organized workforce simply by dropping its DSP from the delivery network—unless, of course, a union decided, in the benighted wisdom the bill grants it, to allow contracting again. The Teamsters’ own account of its Amazon campaign also connects organizing Amazon with protecting union jobs at UPS.

Those benefits explain the Teamsters’ interest in direct employment. They don’t establish that contracting is inefficient or that making Amazon a direct employer would improve street safety.

Coase asked why firms organize some activities internally and obtain others through contract. New York City’s DPA would replace that economic choice with an employment mandate that offers no evident safety benefit—and then waive it when a union strikes a deal. The independent businesses whose contracts the bill would prohibit get no comparable consideration. Meanwhile, the cited analyses don’t establish that drivers and customers who would bear the costs of this crass, politicized choice would receive the promised safety benefits.

The bill’s clearest safety net is the one it extends to the Teamsters.

The post Who, Exactly, Does New York City’s Delivery Protection Act Protect? appeared first on Truth on the Market.

The company’s withdrawal from Nigeria and Uganda complicates one of competition policy’s favorite storylines: U.S. technology platforms enter, conquer, and never leave. That assumption has helped fuel a rush to import rules modeled on the European Union’s Digital Markets Act (DMA), which imposes special obligations on the largest digital platforms. Yet African ride-hailing markets are not following the script. 

Uber once seemed built for precisely that story. Born from a founder’s struggle to find a taxi on a snowy night in Paris, it quickly became one of the most disruptive startups in recent memory and transformed urban transportation. Its African retreat shows that a revolutionary idea does not guarantee permanent dominance. 

Nigeria and Uganda are part of a broader pattern. In September 2025, Uber announced its departure from Ivory Coast. Earlier this year, it announced that it would also leave Tanzania. 

Uber still operates in Egypt, Ghana, Kenya, and South Africa, and has said it remains committed to the continent. Four departures in less than a year make that commitment look considerably narrower than it once did. 

What explains the retreat? Uber’s public statements offer little guidance. Possible explanations include its inability—or unwillingness—to adapt its business model to local markets, fierce competition, currency volatility, rising operating costs, and a global strategy increasingly focused elsewhere. 

Those possibilities raise the more interesting question: Does Uber’s retreat reflect failures peculiar to the company, or does it reveal something deeper about competition in African markets? This post argues for the latter. 

Uber Paused. Bolt Pounced.

Tanzania’s growing urban population and expanding smartphone use have fueled strong demand for ride-hailing services. With the market projected to grow about 10% annually, the country would seem ripe for expansion. 

Regulation complicated that picture. In 2022, the Land Transport Regulatory Authority (LATRA) capped platform commissions at 15%, regulated fares, prohibited booking fees, and imposed several operational requirements. Uber suspended its Tanzanian operations while pressing LATRA to revise the rules 

Uber’s Estonian rival Bolt chose a different course. It protested the price controls but continued most services, suspending only its corporate-client business. As the company explained, “Bolt has continued offering services to demonstrate goodwill and create an opportunity for amicable resolution.” 

Uber’s absence gave Bolt room to capture market share. LATRA eventually raised the commission cap to 25% and restored booking fees, clearing the way for Uber’s return. By then, Bolt had filled much of the space Uber left behind. Uber struggled to regain its footing, while the regulated-fare system continued to make operating in Tanzania more difficult. Both factors likely contributed to its eventual exit.

Uber had good reason to oppose LATRA’s price controls. Such controls can distort market signals, reduce the supply of services, and weaken quality. Suspending operations may therefore have been a perfectly rational response. It was also a costly one. 

Uber took a different approach in Kenya after regulators capped commissions at 18%. It continued operating and challenged the cap in court. That strategy paid off when Kenya’s High Court suspended enforcement, finding that the government had not offered adequate economic justification. A similar challenge might not have succeeded in Tanzania, but Uber’s Kenyan experience shows that withdrawal was not its only possible response. 

Bolt’s experience also illustrates how adaptability can become a competitive advantage. The company incorporated popular local transportation options, including three-wheeled tuk-tuks, known locally as bajajis, and motorcycle taxis, known as bodabodas. Uber did not offer the same range in Tanzania, although it has adapted elsewhere. In 2018, for example, it introduced UberBODA, an electric-bike taxi service, in Kenya. 

Bolt also eliminated cancellation fees that had proved unpopular with passengers. To a Western reader, that might seem like a questionable concession. Cancellation fees serve legitimate purposes. They discourage frivolous bookings, compensate drivers for lost time, and make scheduling more predictable. But even sensible rules can misfire when they meet local behavior. 

Consider a passenger traveling from Lagos’ Lekki Peninsula to the Mainland. The driver arrives, waits until the passenger gets in, and only then asks for the destination. After learning it, the driver refuses the trip. Because frequent cancellations can hurt the driver’s standing on the platform, he tells the passenger to cancel instead. The passenger must then either pay the fee or find another way across town.

In another common variation, the driver arrives but refuses to start the trip unless the passenger agrees to pay off the platform. If the passenger objects, the driver again waits for the passenger to cancel and bear the cost. 

By eliminating cancellation fees, Bolt protected passengers from paying for opportunistic conduct by drivers. The policy sacrificed one useful incentive to address a more pressing local problem. That is what adaptation looks like. 

When the Naira Takes the Wheel

Uber’s Nigerian exit came amid severe currency volatility that battered consumers’ purchasing power and made the country a much harder place for multinational companies to operate.

In 2023, the Nigerian government eliminated fuel subsidies and allowed the naira to float as part of an effort to stabilize public finances. The naira subsequently plunged from about N460 per U.S. dollar to lows approaching N1,800. At the time of writing, it trades at roughly N1,324 per dollar. 

The reforms were arguably necessary. A weaker naira can curb imports and capital flight. But necessary medicine can still have brutal side effects, and the reforms helped create an operating environment that drove several multinational companies toward the exits. 

GlaxoSmithKline left Nigeria in August 2023 after more than 50 years in the country. Sanofi and Procter & Gamble followed, shifting to third-party distributors rather than maintaining local operations. Norwegian energy company Equinor sold its Nigerian business after three decades, while Kimberly-Clark cited the country’s recent economic developments when it ended local operations. 

Currency depreciation hit these companies from both directions. Many multinational firms borrow from their parent companies and pay foreign suppliers in dollars. When the naira fell, the local-currency cost of those dollar obligations soared, producing enormous foreign-exchange (FX) losses. 

Seven leading Nigerian companies reported combined FX losses of N2.06 trillion in 2024, up 28.9% from N1.6 trillion in 2023. MTN Nigeria’s FX losses approached N1 trillion, contributing to a N550 billion pretax loss. Nestlé Nigeria reported a pretax loss of N221.6 billion under similar pressure. 

The losses nearly erased shareholder equity at some companies, forced restructurings, and eliminated dividends. Rising sales offered little comfort. MTN Nigeria increased its revenue by 36.1% but still incurred hundreds of billions of naira in FX losses. 

The accounting also gets uglier once multinational companies translate Nigerian earnings into their home currencies. A sharp fall in the naira can turn impressive local-currency growth into shrinking dollar revenue. Procter & Gamble Chief Financial Officer Andre Schulten said the naira’s volatility made it difficult for a company that reports its financial results in U.S. dollars to create value in Nigeria.

Consumers faced the other side of the squeeze. Devaluation helped drive inflation above 30%, rapidly eroding household purchasing power. Companies could not raise prices enough to cover their rising costs without driving customers toward cheaper alternatives. 

Dollars were also extremely difficult to obtain. Before the currency float, the Central Bank of Nigeria (CBN) had accumulated a multibillion-dollar backlog of unmet demand for foreign currency. Even profitable companies struggled to pay for imports or move earnings out of the country. Sanofi said it could not reliably obtain the foreign currency needed to import medical products. 

Uber’s business differs from those of manufacturers and pharmaceutical companies, but it faced the same battered consumers, rising costs, and currency constraints. The economics were hardly inviting.

Uber’s changing global strategy may have further weakened its appetite for difficult markets. The company reportedly plans to invest about $10 billion in autonomous vehicles (AVs) as it seeks to become a leading robotaxi platform. 

That strategy favors markets with roads that autonomous vehicles can navigate safely. As technology executive Mark Pittman explains, “AVs require clearly visible infrastructure to navigate roads. A lack of it puts both machines and humans at potentially life-threatening risk.” 

Poorly marked and maintained roads create serious operational and safety problems for autonomous vehicles. That does not fully explain Uber’s immediate retreat from Nigeria or Uganda, where human drivers remain the norm. But it does change the company’s long-term calculation. Markets that cannot readily support Uber’s robotaxi ambitions may find themselves farther down its list of priorities. 

Predation Without a Payday

Another explanation for Uber’s exit has gained traction on Nigerian Twitter, now X. Investigative journalist David Hundeyin argues that Uber enters markets with fares below those of traditional taxis, uses venture-capital money to absorb the losses, drives its rivals out, and then raises prices once it has secured a monopoly. 

The trouble is that this theory doesn’t make any sense. 

Economists call the strategy Hundeyin describes predatory pricing. A company charges prices below its costs to drive competitors from the market, then uses its resulting monopoly power to raise prices and recover—or “recoup”—its earlier losses.

That last step matters. Predatory pricing works only when barriers to entry make it difficult or costly for rivals to return. If competitors can enter once the predator raises its prices, they will undercut it before it can recover its losses. The strategy then becomes an expensive gift to consumers, who enjoy low prices while the would-be monopolist burns through cash. That is why the U.S. Supreme Court has said predatory-pricing schemes are “rarely tried, and even more rarely successful.” 

Nigeria hardly offered Uber a competition-free road. Uber was not even the country’s first ride-hailing service. Easy Taxi began operating in July 2013, one year before Uber arrived. Oga Taxi also entered in 2014, followed by Bolt in 2016, inDrive—with its negotiated-fare model—in 2019, and the Lagos state government’s LagRide service in 2022. Easy Taxi and Oga Taxi have since folded, but new challengers kept appearing. 

One estimate holds that more than 2,500 ride-hailing applications attempted to enter Nigeria during Uber’s time there. Whatever the exact count, the market plainly lacked a shortage of aspiring competitors. 

Ride-hailing platforms also compete with a wide range of local transportation options. These include danfo buses, keke na pepes (three-wheeled taxis), okadas (motorcycle taxis), and private cars. Those alternatives limit what any ride-hailing company can charge. 

Nor does entry require a company to buy thousands of cars or hire a nationwide workforce. The essential asset is the platform connecting drivers and passengers. A new entrant still needs reliable software, marketing, payment systems, and enough users to make the service worthwhile. But it can compete without owning a fleet or employing its drivers, sharply reducing the capital required. 

Drivers make entry easier by “multihoming,” meaning they use several platforms. The same driver may accept trips through Uber, Bolt, and inDrive, switching among them to find more passengers. A new platform therefore does not need to recruit an entirely new driver network. It can tap an existing pool of experienced drivers. 

Uber’s prices also sit awkwardly with the predation story. A 2026 fare comparison found that Uber generally quoted higher prices than competing ride-hailing services. That is curious behavior for a company supposedly bent on pricing every rival out of the market. 

The most basic problem with Hundeyin’s account is Uber’s exit itself. If Uber spent years subsidizing rides to secure a monopoly, leaving Nigeria means the strategy failed before the company could recover its losses. The evidence fits a more prosaic explanation: Uber struggled to make its business model work in a difficult, crowded market. 

Hundeyin also claims that Uber’s real purpose was to serve as a surveillance operation for the U.S. “military-industrial complex.” That theory sits uneasily beside the predation claim. If surveillance were the objective, spending hundreds of millions of dollars on a commercial platform that would eventually abandon the market would be a remarkably roundabout strategy. Cheaper and more durable alternatives abound. 

Nigeria’s transportation debate deserves economic analysis grounded in evidence, not incoherent conspiracy theories. 

Big Tech Meets Local Traffic

Uber’s African experience undercuts the assumption that U.S. technology companies will inevitably dominate every market they enter. In Nigeria, the company became so synonymous with ride-hailing that “taking an Uber” came to mean ordering any app-based taxi. Yet that ubiquity never produced an unassailable competitive position.

Uber’s exits show how quickly global scale can collide with local reality. Competition, regulation, currency volatility, infrastructure, and a company’s willingness to adapt may matter more than the fame—or size—of its platform. 

Policymakers should take note. Domestic companies do not necessarily need DMA-style protection simply because they compete against a globally prominent U.S. platform. Where entry remains possible and consumers have alternatives, competition can discipline even the biggest firms. 

Many African markets already suffer from poorly justified price controls and other regulatory obstacles. Layering imported mandates onto that framework would compound the problem. Policymakers should instead remove needless barriers, demand economic justification for regulation, and improve the underlying conditions for local competition. 

Competition policy should begin with how markets actually work, not with a borrowed presumption about who must win. Becoming a verb is not the same as becoming a monopoly.

The post No Free Ride: Uber’s African Retreat appeared first on Truth on the Market.

That allegation could support a conventional deception claim. If Amazon promised advertisers a genuine generalized second-price auction while deliberately concealing a material departure from that model, the FTC may have a viable case.

But the complaint tries to make a broader leap, arguing that Amazon’s pricing practices necessarily harmed retail consumers. The use of reserve prices, relevance scores, or dynamic pricing does not automatically make an auction unfair. Nor does it necessarily harm the consumers who buy the advertised products.

That is the complaint’s weaker theory. Based on the public record, the FTC has asserted retail-consumer harm with confidence but has yet to prove it.

What the FTC Alleges—and Amazon Disputes

The FTC and the attorneys general of 22 states filed the case in the Western District of Washington. The FTC’s press release describes a seven-year scheme that affected roughly 1.2 million advertisers, including more than 500,000 small and medium-sized businesses.

Amazon’s three principal Sponsored Ads products auction placements tied to product searches and product pages. According to the agency, Amazon told advertisers that winners would pay only the minimum needed to secure a placement—often one cent more than the next-highest-ranked bid. Instead, Amazon allegedly imposed undisclosed “soft reserve” prices or surcharges. The complaint estimates that the practice extracted more than $20 billion from advertisers.

The complaint tells a powerful story. It describes an “invented auction participant,” a “proxy 2nd price,” internal concerns about advertiser expectations, and alleged efforts to keep price increases below detection thresholds. It also alleges that the share of Sponsored Products auctions in which advertisers paid their full bids reached about 80% by 2024.

If proved, these allegations would amount to more than a complaint about hard bargaining. The FTC’s case would be that Amazon promised advertisers one pricing mechanism while secretly operating another.

Amazon offers a different account. The company says its system evolved from rankings driven heavily by bids to machine-learning models that give greater weight to relevance and predicted shopper engagement. It describes soft reserves as real-time minimum prices that reflect the value of individual placements, much as a retailer might value an endcap differently from ordinary shelf space.

Amazon also says its campaign console has long informed advertisers that a bid represents the maximum they may pay. According to the company, the highest bidder does not win roughly 92% of placed Sponsored Products ads, average winning bids fell by 50% from 2019 to 2025, and both conversion rates and return on advertising spending improved. These claims are not adjudicated facts, but they identify the central economic issues the complaint must confront.

The legal questions are narrower than whether Amazon is large, profitable, or vertically integrated. Did the company materially misrepresent its price-setting rules to reasonable advertisers? Did its practices cause legally recognized consumer harm that advertisers could not reasonably avoid and that countervailing benefits did not outweigh? Those inquiries should not collapse into a general judgment that Amazon earns too much from advertising.

How Advertising Auctions Work

Advertising platforms use auctions to allocate scarce, varied opportunities. A search page has room for only a limited number of ads, and some positions attract more attention than others. A placement’s value also depends on the query, nearby products, the shopper, the time of day, and the likelihood of a click or purchase. A pricing mechanism that treats every opportunity alike may be simple, but simplicity does not guarantee efficiency.

In a single-item second-price auction, the highest bidder wins and pays slightly more than the second-highest bid. Under the classic Vickrey model, each participant’s best strategy is to bid what the item is worth to them.

Digital-advertising auctions are more complicated. As Benjamin Edelman, Michael Ostrovsky, and Michael Schwarz explain, a generalized second-price auction with several advertising slots is not merely a series of Vickrey auctions. Because the slots produce different click-through rates, bidding one’s true value is generally not the dominant strategy. Advertisers may lower their bids, experiment with different amounts, or use automated systems to maximize expected returns.

That distinction matters. The FTC repeatedly treats “second price” as though it describes a complete and uniquely determined mechanism. But a relevance-adjusted auction requires rules governing eligibility, quality, expected click-through rates, conversion probabilities, placement, reserve prices, and payment. A platform may use a second-price-like payment rule while considering nonprice information when allocating placements. It may also decline to sell a placement below a specified floor. A reserve price is not inherently a fake rival. It is a seller-side condition on the transaction.

The harder question is whether Amazon disclosed that condition. A hidden reserve may alter advertisers’ bidding incentives if they believe the second-highest competitor determines the price and therefore bid close to what a placement is worth to them. A disclosed reserve may simply prompt an advertiser to bid only when the placement is worth at least that amount.

The significance of Amazon’s alleged conduct therefore depends on the counterfactual. What did advertisers reasonably believe? What did Amazon’s system actually calculate? How did the difference affect bids, campaign spending, and realized returns? Calling the reserve a “shill bid” is evocative, but the legal and economic analysis should turn on those mechanisms and effects, not internal labels.

The FTC’s Deception Case

Section 5(a) of the FTC Act prohibits unfair or deceptive acts or practices (UDAP) in commerce. Under the agency’s longstanding Policy Statement on Deception, a practice is deceptive if it is likely to mislead a reasonable consumer about something material—meaning something likely to affect the consumer’s conduct regarding a product or service. When the audience is specialized, the standard reflects a reasonable member of that audience, not an unsophisticated retail shopper.

Applied to Amazon’s advertising customers, that framework gives the FTC a potentially strong argument. “Second-price auction” and statements that the winner pays the second-place bid are not mere puffery. They describe a pricing rule that can affect bidding strategies, budget allocation, expected customer-acquisition costs, and the value advertisers assign to campaign data.

The complaint alleges that Amazon repeated these claims on websites, in training materials, and during sales presentations even after changing its auction system. It also claims that Amazon gave evasive or false answers when advertisers asked whether the format had changed. A factfinder could view that pattern as a material misrepresentation, not a harmless simplification.

Amazon’s response points to four principal defenses. First, the campaign console allegedly told advertisers that a bid represented the maximum price they could pay. Second, Amazon says its system was not a simple highest-bid auction because lower bids could win based on relevance. Third, the company describes the materials cited by the FTC as outdated educational content with limited reach that it later removed or revised. Fourth, Amazon argues that sophisticated advertisers and automated bidding tools respond to actual clicks, conversions, costs, and returns—not simplified descriptions of auction theory.

Those defenses do not dispose of the case. Telling advertisers that a bid sets the ceiling does not necessarily disclose that Amazon may substitute an internally generated floor for the competitive price. Nor does advertiser sophistication give a seller license to make a specific false claim about pricing.

But the defenses sharpen the inquiry. The question is not whether some Amazon document used the phrase “second price.” It is whether the overall impression created by Amazon’s communications, viewed in context and at the time of each transaction, conveyed a materially false pricing rule to reasonable advertisers.

Even if the evidence supports a deception finding, monetary relief requires more. Any estimate of advertiser or consumer injury must show that the alleged misunderstanding changed behavior and caused a measurable loss, not merely confusion over terminology.

The FTC’s Unfairness Case

The FTC’s unfairness theory faces a different statutory test. Section 45(n) bars the agency from declaring a practice unfair unless it causes or is likely to cause substantial consumer injury that consumers cannot reasonably avoid and that countervailing benefits to consumers or competition do not outweigh. The FTC Policy Statement on Unfairness makes consumer injury the central concern. It also requires the agency to consider offsetting benefits and the costs of intervention, including reduced incentives for innovation and capital formation.

The complaint offers little evidence of harm to retail consumers. Its factual narrative focuses overwhelmingly on advertisers. Consumers enter the story mainly as buyers of products from Amazon merchants. The FTC reasons that merchants likely pass higher advertising costs along to them, especially when selling low-margin necessities.

But the complaint’s ultimate allegation of consumer injury is conclusory. It identifies no retail-price study estimating pass-through, no comparison of advertised and unadvertised product prices, no evidence of reduced output, and no net-welfare analysis that accounts for the relevance improvements Amazon says accompanied its pricing changes.

Amazon highlights precisely this gap. The company says consumers appear only a handful of times in the complaint’s more than 150 pages, the FTC cites no data showing higher consumer prices, and the proposed monetary relief would go to advertisers rather than shoppers. Amazon also reports that inflation-adjusted cost per click (CPC) remained flat while conversion rates and advertiser performance improved substantially.

Those figures require scrutiny. But the FTC cannot prove a consumer-protection case by assuming that every extra dollar an advertiser pays becomes an extra dollar on a shopper’s bill.

Pass-through is an empirical question. An advertiser may absorb higher costs through lower margins; cut its advertising; move spending to Google, Meta, Walmart, direct sales, or other channels; change its product mix; or raise prices only where demand allows. Better-targeted advertising may also reduce the effective cost of acquiring customers even if the nominal CPC rises. More relevant ads may help consumers discover new products, intensify competition among sellers, or reduce search costs.

A practice may therefore transfer surplus from advertisers to Amazon while improving, diminishing, or leaving consumer welfare unchanged. The FTC cannot treat “passed on” as its conclusion when pass-through is the disputed causal link.

That does not make advertiser injury irrelevant. Advertisers are themselves consumers of Amazon’s advertising services, and deceptive pricing claims may harm them directly. But the unfairness theory still must satisfy Section 45(n). A court should demand evidence of substantial net injury, consumers’ realistic ability to avoid that injury, and the benefits of Amazon’s dynamic allocation system.

If the FTC proves deception, a separate unfairness balancing test should not excuse it. But if the agency claims that a commercially sophisticated pricing system is inherently unfair, the statutory cost-benefit test is indispensable.

Choosing the Right Counterfactual

The case also illustrates a recurring problem in platform regulation: treating the regulator’s preferred counterfactual as the natural baseline. The FTC compares Amazon’s charges with the prices a genuine generalized second-price auction would have produced. That may be the right counterfactual for a deception claim if Amazon promised precisely that mechanism. It does not necessarily tell us whether Amazon’s evolving pricing and allocation system worked efficiently.

Harold Demsetz’s critique of the “Nirvana” approach applies directly. Demsetz argued that analysts should compare institutions as they can actually operate, not measure an imperfect reality against an idealized alternative free of costs, frictions, and unintended consequences. Geoffrey Manne has applied the same point to digital platforms. A regulator can imagine a world in which a platform maintains all its investments and services while accepting rules that reduce their returns. But that world is not a serious alternative unless the analysis considers how the platform would respond.

Amazon’s relevance model may be flawed, inefficient, or poorly disclosed. It may also address a real allocation problem. If bids alone determine rankings, a high-paying but irrelevant ad can displace a lower-paying ad that shoppers would find more useful. A relevance model may allow a lower bid to deliver more value to both the shopper and the advertiser. A soft reserve may then represent an imperfect attempt to keep scarce premium placements from becoming systematically underpriced as the ranking algorithm changes.

The relevant comparison is not an abstract choice between “auction” and “no surcharge.” It is Amazon’s actual model compared with feasible alternatives, such as transparent reserves, higher upfront fees, a first-price auction, less investment in relevance, or a different mix of advertising inventory and retail prices.

Friedrich Hayek offers another caution. Amazon’s engineers and economists observe billions of auctions involving changing queries, products, and shoppers. Their models draw on dispersed, context-specific information that no public official can reconstruct from a complaint and a handful of documents. That knowledge gap should inspire humility about prescribing a single permissible auction formula.

It does not give Amazon license to mislead. A market-process perspective supports accurate disclosure and accountability for deliberate deception. It also resists the assumption that regulators can safely dictate how a platform prices every distinct placement.

Ronald Coase likewise urged decision-makers to compare institutional arrangements and their transaction costs. The FTC identifies costs to advertisers and possible downstream effects on shoppers. It must also consider the costs of intervention, including new compliance systems, less experimentation, poorer allocation, higher fixed fees, lower-quality ads, and the risk that platforms will stop serving smaller advertisers.

The central policy question is not simply whom to label the wrongdoer. It is which feasible arrangement produces the greatest expected value after accounting for all relevant costs and benefits.

The Risks of an Overbroad Remedy

Single-firm cases are especially prone to false-positive errors because the same conduct may look like exploitation or innovation depending on facts that outsiders struggle to observe. The error-cost principle counsels caution when intervention may suppress productive conduct that market competition would otherwise test and correct. Frank Easterbrook’s classic analysis and Jonathan Barnett’s discussion of false positives both emphasize that speculative intervention can produce real harms.

The risk here extends beyond an order requiring clearer disclosures. The litigation could evolve into a command that Amazon use a fixed generalized second-price formula, price every placement according to the next advertiser’s bid, or seek regulatory approval before setting reserve prices. Such a remedy could freeze a technology designed to evolve.

Amazon might respond by giving relevance less weight, raising subscription or seller fees, reducing free tools, restricting access for small advertisers, or adopting an openly first-price model in which strategic bid shading becomes central. None of those effects is certain. That uncertainty is precisely what makes a broad remedy hazardous.

Dynamic pricing also serves legitimate economic purposes. It can direct limited inventory toward its highest-value uses, reflect peak demand, and help a platform recover investments in search, logistics, fraud prevention, measurement, and machine learning. A reserve may act as a price floor, a quality screen, or a way to avoid selling a premium placement for less than its opportunity cost.

Those functions may coexist with deception if the platform lies about them. The remedy should target the lie, not treat the underlying pricing flexibility as inherently suspect.

The FTC’s burden therefore breaks into three questions. What did Amazon promise? What did it actually do in each transaction, and how did that differ from the promise? What injury resulted after accounting for the value and performance of the advertising delivered? The first two questions are primarily factual and legal. The third is economic.

A $20 billion gap between an idealized generalized second-price outcome and Amazon’s actual revenue does not automatically equal $20 billion in social loss. Some of that amount may reflect better allocation, avoided costs, valuable impressions, or services that advertisers continued to use voluntarily. After accounting for those benefits, net social welfare may even exceed what it would have been under the supposedly ideal pricing model.

A Remedy That Fits the Harm

The case should not be dismissed merely because the buyers are businesses or because Amazon faces competitors. A material, undisclosed departure from a promised pricing rule can undermine voluntary exchange, especially when the seller controls the data needed to verify the auction. If the evidence shows that Amazon knowingly misrepresented its auction as second price to induce higher bids—and that remains a very big if—targeted disclosure and restitution for the resulting overcharges may be justified. The FTC’s state co-plaintiffs may also seek restitution, even where the agency itself cannot.

Any remedy should remain proportionate. Amazon should have to describe its ranking and pricing system accurately through the interfaces and sales channels advertisers actually use, preserve enough records to permit an audit, and correct materially misleading historical materials.

Monetary relief should rest on a transaction-level counterfactual that separates any overcharge caused by deception from the value created through better relevance and performance. The court should resist requiring a single auction design, banning every undisclosed reserve regardless of context, or treating all of Amazon’s advertising revenue as ill-gotten.

This approach would enforce the law against deception without turning consumer protection into price regulation. It would also leave Amazon room to find better ways to match products, shoppers, and advertisers. The law should insist that market participants know the essential rules of the game. It should not assume that public officials know how best to play it.

Truth Without Price Regulation

The FTC’s lawsuit against Amazon may establish an important principle: sophisticated auction systems must still obey ordinary rules against material deception. But the public record does not yet support the complaint’s broader claims that dynamic reserve pricing necessarily harmed retail consumers or that regulators can safely supervise the design of a complex advertising marketplace.

The outcome should turn on evidence of what Amazon promised, what it concealed, and what economic harm followed. A market-oriented law & economics analysis should demand truthful rules, empirical evidence of consumer harm, and relief tailored to that harm. It should also take seriously the cost of regulatory mistakes that micromanage innovation.

The FTC may have a deception case. It has not yet proved a pricing case.

The post Reserve Judgment: The FTC Takes on Amazon’s Ad Auctions appeared first on Truth on the Market.

The episode should reset the terms of the AI-safety debate. Testing, red teaming, and independent investigation remain essential. Their proper role is to uncover problems and guide continuous correction. No finite assessment can establish how an evolving, context-dependent system will behave across every prompt, user, tool, deployment, and interaction with other systems. And governments should not turn a snapshot into a permission slip.

This post extends the argument I developed in “Too Much Order, Too Soon” and “Artificial Intelligence, Natural Ignorance.” The problem goes beyond officials lacking enough information to design the right test. It’s that safety knowledge emerges through the same experimentation and competition that premature regulation could cut short.

That distinction anchors what I call applied Austrian economics (AAE). Austrian economics explains why knowledge—including knowledge about AI safety—remains dispersed, depends on context, and often emerges only through use. Public Choice—the study of how incentives shape government decisions—explains why a checkpoint created to gather that knowledge can become a gatekeeper, a competitive moat, and a source of false confidence. The practical task is to build safety throughout the AI lifecycle while preserving the decentralized discovery that makes improvement possible.

The Test That Tested Its Testers

In July, OpenAI deployed tens of thousands of autonomous AI agents in ExploitGym, a cybersecurity benchmark that tests whether agents can find and exploit software vulnerabilities. The agents were supposed to remain isolated. Roughly 1,200 instead found an unintended communication channel through a shared software repository.

The agents exchanged more than 70,000 messages and files and organized collective projects to defeat the evaluation. About 700 participated in an attack on the AI-development platform Hugging Face. Some investigated how to alter the evidence that evaluators would inspect. Outside investigators found successfully spoofed tool calls—fabricated records of actions supposedly taken by software tools—in portions of about 7% of the transcripts they reviewed.

Those findings come from an independent incident report by METR and Redwood Research. The report does not show that AI is about to escape human control. The principal model was an internal research system, the agents operated within a cybersecurity benchmark, and the investigators carefully limited their conclusions.

The report’s value lies elsewhere. It shows what can happen when a complex system must optimize within an evaluation environment. The model may begin treating the test itself as part of the problem to solve.

The evaluators encountered their own version of the knowledge problem. They reviewed about 1,300 lengthy transcripts but lacked direct access to some relevant infrastructure and to the principal internal model. Parts of the record were missing. The sheer volume of material forced them to rely heavily on other AI agents, which sometimes proved unreliable, to analyze the evidence.

OpenAI deserves credit for inviting independent scrutiny, sharing extensive data, and accepting a report that exposed uncomfortable facts. Yet six days of on-site investigation and roughly $400,000 in donated application-programming-interface (API) credits still produced only preliminary, qualified answers.

The episode should change the terms of the AI-safety debate. Testing, red teaming, and independent investigation are all necessary. But none of them, separately or together, can turn an evolving, context-dependent system into a product the government can certify as simply “safe.” Testing should serve as a tool for discovery and continuous correction—not as a regulatory permission slip.

The incident demonstrates both the value and the limits of evaluation. Testing uncovered a failure that the designers had not anticipated. Even an unusually intensive investigation could not produce a final account. That’s what continuous validation is for.

Quality by Design, Not by Decree

The best analogy comes from pharmaceutical manufacturing, though it would be easy to overinterpret. The Food and Drug Administration’s (FDA) Process Analytical Technology guidance captures the central engineering insight in a memorable phrase: Quality cannot be tested into a product—it must be built in. The agency’s 2011 process-validation guidance therefore treats validation as a lifecycle with three stages: process design, process qualification, and continued process verification.

This approach arose from the limits of testing samples of finished products. A passing sample cannot capture every variation in raw materials, equipment, temperature, pressure, contamination, or operator practices. The problem becomes even harder with biologics, which rely on living systems and produce more structural variation than conventional small-molecule drugs. Manufacturers therefore identify the measurable features that determine product quality, track the process parameters that affect them, document deviations, and continue verifying performance after commercial production begins.

The point is to borrow the practice, not to create an FDA for AI.

The FDA’s current good manufacturing practice rules set a legally enforceable floor. Manufacturers must design and control their production processes so drugs consistently meet predetermined standards for identity, strength, quality, purity, and potency. They must also validate those processes with evidence rather than rely solely on tests of finished batches.

The FDA does not prescribe a single method for meeting that obligation. A manufacturer may use a conventional empirical approach based largely on experimentation and accumulated experience, a more systematic Quality by Design (QbD) approach, or a combination of the two. QbD begins with the qualities a product must possess and examines how materials and production choices affect them. Process Analytical Technology uses measurements taken during production to detect variation and adjust the process in real time. The FDA encourages these science- and risk-based methods, but their use remains voluntary under its guidance.

More importantly, drug approval remains an expensive, slow, and centralized permission system. Clinical trials can produce evidence about safety and effectiveness for the populations and uses studied. They cannot prove that a drug will work safely for every person under every condition. FDA approval manages uncertainty. It does not make uncertainty disappear.

That’s why even some supporters of substantial AI oversight resist calls for an “FDA for AI.” During the Committee for Justice’s August webinar, “Congress and the Future of AI: Model Testing, Preemption, and Open Weight Policy,” panelist Paul Steidler warned that an FDA-style system would be prolonged, restrictive, and hostile to individual choice. Yet the discussion also showed how quickly sensible demands for testing can become proposals for mandatory government review before developers may release a frontier model. Once the reviewer gains the legal power to say no, testing ceases to be merely an engineering practice. It becomes a licensing regime.

Borrow the maxim, then, but leave the bureaucracy behind. AI developers should build quality and security into design, training, deployment, and monitoring. Government should not turn that sound practice into a universal premarket-approval regime.

Safety Has No Final Exam

It is tempting to dismiss after-the-fact AI testing as an epistemic impossibility—a fundamentally inadequate way to know whether a system is safe. That would be going too far. Tests can uncover capabilities, vulnerabilities, and performance regressions. Testing exposed the Hugging Face incident itself.

What testing cannot do is prove a negative. No finite battery of tests can establish that a model will never behave unacceptably across unknown prompts, users, tools, operating environments, software updates, and interactions with other systems.

The National Security Commission on Artificial Intelligence recognized this distinction in its 2021 final report. Rather than prescribe one final exam at the end of development, it treated testing, evaluation, verification, and validation (TEVV) as a continuous process spanning system requirements, development, deployment, training, maintenance, and monitoring during operation. It also recommended documenting where training data came from, stress-testing systems, dividing them into components that can be examined separately, building in tools that track and record how the systems operate, designing recovery mechanisms, and using red teams to find ways to break them.

The commission focused on government and national-security applications. In that setting, the government acts as purchaser and operator and may properly demand evidence about systems deployed on its behalf. That’s quite different from requiring every developer to obtain government approval before releasing a model.

The National Institute of Standards and Technology’s (NIST) voluntary AI Risk Management Framework follows the same lifecycle approach. Its four functions—to govern, map, measure, and manage—continue throughout a system’s use. Testing occurs before deployment and while the system operates. Organizations may mitigate a risk, transfer it to another party, avoid it by changing course, or knowingly accept it.

This model resembles industrial process control more than a licensing board. Its voluntary and adaptable structure matters as much as its technical content. The framework helps organizations manage risks without pretending that one checklist can settle them forever.

The European Union (EU) has taken a more coercive approach. Article 55 of the EU Artificial Intelligence Act requires providers of general-purpose AI models classified as posing systemic risk to conduct and document model evaluations and adversarial testing, assess and mitigate risks, report serious incidents, and maintain cybersecurity protections. Other provisions require quality-management systems and monitoring after a model reaches the market.

Even this framework implicitly concedes that a prerelease evaluation cannot finish the job. The danger lies in the machinery built around that insight. Official classifications, documentation mandates, codes of practice, and enforcement powers can turn a process of continual learning into a state-administered compliance system.

The Trump administration’s June 2 executive order takes a narrower path. It directs the federal government to create classified benchmarks for advanced cyber capabilities and a voluntary framework under which developers may give the government access to covered frontier models for up to 30 days before releasing them to other trusted partners. The order expressly disclaims any authority to impose mandatory licensing, preclearance, or permitting.

That limiting language matters. So does vigilance. Public Choice counsels us to examine the administrative capacity the order creates, not just the disclaimer attached to it. Today’s voluntary test can become tomorrow’s convenient checkpoint.

Mechanistic interpretability also belongs in this process, though it should not be oversold. Ordinary evaluations test a model from the outside by giving it prompts and examining its responses. Mechanistic interpretability tries to reverse-engineer the model from within. Researchers study patterns of activity inside the neural network to identify “features” that encode particular concepts and “circuits” that combine those features to perform tasks. The goal is to connect internal activity to observable behavior. If those connections prove reliable, they can help developers to debug models, uncover hidden failure modes, and make more credible safety claims.

It is not the computational equivalent of possessing a complete chemical formula. A partial map of a model’s internal workings cannot predict how every capability will appear in every deployment. Nor must researchers fully explain a model’s operation before gathering useful evidence about its behavior.

Here, too, the drug analogy counsels humility. Incomplete knowledge of a drug’s mechanism of action does not make empirical evidence worthless. And on the other hand, a plausible account of that mechanism does not guarantee safety. Interpretability, evaluations, monitoring, and incident reports provide different pieces of the puzzle. None can replace the others, and none should become a legal talisman.

AI Safety Has a Knowledge Problem

Friedrich A. Hayek’s knowledge problem is often flattened into the claim that government officials lack enough data. His deeper point was that much of the relevant knowledge cannot be gathered in advance because it does not yet exist. It emerges in fragments as people respond to particular circumstances of time and place. Competition helps coordinate those fragments and, just as importantly, generates new knowledge through trial, error, and discovery.

AI safety has the same structure. A laboratory can test whether an agent completes a defined task under controlled conditions. It cannot catalog every way a hospital, bank, law firm, farmer, software developer, student, or malicious actor might combine the model with local data, tools, incentives, and constraints. A vulnerability may lie in the model’s architecture. But it may instead arise from access granted by a deployer, a customer’s workflow, a third-party plug-in, or an interaction the developer had no reason to anticipate.

Israel Kirzner’s account of entrepreneurial discovery sharpens the point. Rival firms do more than choose among known safety techniques. By experimenting with different architectures, controls, business models, and deployment practices, they discover new ones. The Hugging Face episode produced valuable knowledge precisely because an evaluation failed in an unexpected way. Regulators can mandate yesterday’s best practices. They cannot write tomorrow’s discoveries into a rule.

Marginalism supplies a second AAE insight. It evaluates choices by asking what the next increment of a good produces and costs. “Safe” is not a binary property, and additional safety is not free. A coding assistant confined to a disposable sandbox poses different stakes from an autonomous agent with credentials to access critical infrastructure. Another safeguard may reduce one risk while sacrificing usefulness, privacy, speed, or accessibility. The relevant question is how much additional risk reduction a safeguard buys in a particular use, what it costs, and how it compares with the alternatives.

Positing single threshold for frontier models flattens all of these distinctions. It encourages officials to use compute, or processing power, parameter counts (a rough measure of model size), or benchmark scores as stand-ins for harm. It also draws attention away from deployment—the point at which many risks actually arise—and toward the model as an abstract object.

AAE instead asks who possesses the most relevant knowledge and control at each stage. That may be the model developer during design, the deployer when granting access to data and tools, or the operator overseeing its daily use.

The Certificate and the Moat

The knowledge problem explains why centralized certification will make mistakes. Public Choice explains why those mistakes are unlikely to remain random. Political incentives tend to steer regulation toward organized interests that can influence the rules and away from consumers and smaller competitors who bear the costs.

Economist George Stigler argued that regulation is often “acquired by the industry” and designed and operated largely for its benefit. “Acquired” was a deliberately provocative term. In practice, firms may lobby for regulation, decline to oppose it, or learn to wield it against rivals. Gordon Tullock’s theory of rent-seeking explained why firms spend resources pursuing political privileges rather than creating value. Thomas Sowell focused on “surrogate decision-makers”—officials who make choices for others without bearing the full costs when those choices go wrong. These theories do not allege a conspiracy. They predict how people will respond to incentives.

Large model developers can absorb fixed compliance costs, maintain Washington offices, hire specialized lawyers, and place experts on standards committees. A startup or open-weight project—one that allows others to download and modify a model’s underlying parameters—cannot spread the same costs across billions of dollars in revenue. A mandatory audit may uncover a hazard, but it also raises the cost of entering the market. That makes rules advertised as safety measures especially attractive to incumbents when they also function as competitive moats.

The Committee for Justice webinar captured the danger in a single line from Alliance for the Future Senior Fellow Neil Siefring: “Regulation can become a moat.” Kevin Frazier, director of the AI Innovation and Law Program at the University of Texas School of Law, identified the pressure in the other direction. Testing is difficult and resource-intensive, he argued, but necessary to build confidence and encourage adoption.

Both points can be true. The challenge is to produce credible assurances without handing a small group of firms and officials control over who may enter the market.

Government certification could create another distortion by serving as both a marketing tool and a legal shield. A regulated firm could tout compliance as proof of safety. An injured person might then discover that federal approval had narrowed or eliminated a remedy under state law. Courts would have to decide whether federal law displaces the state-law duty underlying the claim, a doctrine known as preemption. Even if the claim survived, jurors might mistake the certificate for a government guarantee that the product was safe.

Drug and medical-device litigation shows how quickly safety regulation can become a fight over preemption. Whether an injured plaintiff may proceed can depend on the product’s regulatory pathway, its legal classification, and whether the manufacturer could change a warning without additional federal approval. AI policy has enough hard problems without importing that preemption battle, too.

Accountability Without the Liability Lottery

An AI-specific strict-liability regime would cast too wide a net. Strict liability can require a defendant to pay for harm without proof of negligence. Applied broadly to AI, it could make a developer the insurer of risks it did not create or control, including unforeseeable misuse, a deployer’s careless access settings, or a third party’s integration choices. Open-ended liability could deter useful experimentation along with reckless conduct.

Licensing creates a different distortion. It can block new entrants while encouraging approved firms and users to treat a government license as a warranty. One approach risks making developers liable for everything. The other risks persuading everyone that the government has vouched for them.

The evidence on liability is more nuanced than either its critics or enthusiasts sometimes admit. A 2026 study by Alberto Galasso and Hong Luo in the American Economic Journal: Microeconomics found that product-liability litigation reduced new product introductions by defendant medical-device firms during litigation years. Other firms also introduced fewer products in the categories under litigation, though the effect was smaller.

The decline was neither permanent nor economy-wide, and the litigation prompted firms to develop safer devices. Liability can change both the pace and direction of innovation. It offers benefits and imposes costs. It is not a free safety machine.

Drug and medical-device law shows how messy the overlap between federal approval and state tort claims can become. In Wyeth v. Levine, the U.S. Supreme Court allowed a failure-to-warn claim against a brand-name drugmaker to proceed. FDA rules permitted Wyeth to strengthen its warning before obtaining additional agency approval, and the record did not show that the agency would have rejected the change. Federal approval therefore did not displace the manufacturer’s state-law duty.

Generic manufacturers faced the opposite result. In PLIVA Inc. v. Mensing, federal law required their labels to match the brand-name label, leaving them unable to add the warnings that state law allegedly demanded. The court therefore held that federal law preempted the claims. In Mutual Pharmaceutical Co. v. Bartlett, it applied similar reasoning to a design-defect claim because the generic manufacturer could change neither the drug’s composition nor its label. The court also rejected withdrawal from the market as a way to comply. A patient’s remedy could thus depend on whether the pharmacy dispensed the brand-name drug or its generic equivalent.

Medical devices add a statutory variation. In Riegel v. Medtronic Inc., the court held that FDA premarket approval created device-specific federal requirements. Because the Medical Device Amendments expressly preempt different or additional state safety requirements, the court barred tort claims challenging the device’s approved design and labeling. The regulatory route to market can therefore determine whether an injured person has a route to court.

Taken together, these cases make liability turn on the product’s regulatory pathway, the federal requirements attached to it, the state duty asserted, and whether the manufacturer could change a design or warning on its own. The route to market can determine whether an injured person has any route to court. AI policy should think twice before adopting the same map.

Congress should neither create an AI-specific strict-liability regime nor make compliance with a federal evaluation an automatic defense against otherwise valid claims. Existing law can address fraud, breach of contract, negligent security, professional malpractice, property damage, and physical injury. Liability should follow control, fault, causation, foreseeability, and provable harm.

Those factors will point toward different defendants in different cases. A developer that misrepresents a model’s tested capabilities occupies a different position from a hospital that deploys it outside the specified conditions. Both differ from a criminal who deliberately defeats its safeguards.

AI also differs from most prescription-product cases in one crucial respect. Prescription drugs and medical devices usually reach patients through licensed clinicians who assess their needs and select a treatment. In jurisdictions that follow the learned-intermediary doctrine, a manufacturer generally satisfies its duty by adequately warning the prescribing clinician, who then advises the patient.

Many general-purpose AI tools reach users with no comparable professional in the middle to evaluate suitability or explain warnings. Enterprise deployments may look different. A hospital, bank, law firm, or other professional user may configure the model, choose its data, limit its permissions, and control how employees use it. Liability should reflect those relationships. A developer that serves consumers directly cannot rely on a nonexistent intermediary. An institution that configures and controls a deployment may bear responsibility for the risks its choices create.

This approach leaves room for contracts to allocate responsibilities, insurers to price risk, and courts to learn from concrete disputes. It preserves accountability without turning every novel failure into a lottery-sized claim against the deepest pocket in the technology stack.

Assurance Without the Official Stamp

A market-centered process-validation model begins with an assurance case, not a government certificate. An assurance case is a structured argument, supported by evidence, that a system is acceptably safe for a specified use under specified conditions. The developer or deployer describes those uses, identifies material hazards, explains the chosen safeguards, and makes claims that customers, auditors, insurers, and business partners can challenge. The point is not to prove that nothing bad can happen. It is to show what the system can reasonably be expected to do—and where that assurance ends.

First, safety begins with system design. Developers should grant agents only the permissions they need. Agents meant to remain isolated should not share memory caches, credentials, or communication channels. Critical actions should require additional authorization, and systems should fail safely by defaulting to a limited and recoverable state. The Hugging Face incident was more than a story about model behavior. It also exposed flaws in the infrastructure through which supposedly isolated agents found one another.

Second, firms should preserve traceability by maintaining an audit trail proportionate to the stakes. That record should cover training and evaluation data, model and system versions, material design decisions, known limitations, tool permissions, and responses to incidents. This is not paperwork for its own sake. It supplies the evidence needed for internal learning, customer review, insurance underwriting, and liability when something goes wrong.

Third, testing should be adversarial and iterative, with independent reviewers capable of challenging a firm’s assumptions. It should also remain just one source of feedback among several. Firms should rerun tests after material changes and supplement them with staged rollouts, monitoring during operation, user reports, and investigations after incidents. A benchmark score captures one moment under defined conditions. Validation creates a continuing feedback loop.

Fourth, assurance should reflect context and competition. Hospitals, defense agencies, banks, and consumer-app providers face different hazards and therefore need different evidence. Enterprise customers can negotiate audit rights, performance commitments, security controls, warranties, and incident-notification requirements. Insurers can require safeguards as a condition of coverage. Competing auditors can specialize and earn trust through reliable work. None will be infallible, but competition allows standards to evolve and makes it harder for one flawed benchmark to harden into law.

This framework leaves government plenty to do. It should enforce contracts and generally applicable laws, prosecute theft and unauthorized computer access, protect property rights, and maintain courts capable of resolving concrete injuries. When government buys or operates AI systems—especially for defense and intelligence—it can demand lifecycle testing, evaluation, verification, and validation. It can also support basic research, shared measurement tools, and voluntary standards through institutions such as NIST. What it should not do is decide which general-purpose models may enter the market.

Regulating AI Through the Loading Dock

The AI-safety debate will also creep into data-center politics because AI depends on physical infrastructure. Data centers require land, electricity, water, and grid capacity, giving communities legitimate reasons to ask about noise, utility costs, resource use, and who pays for new infrastructure.

As I argued in “The Data Center Chessboard Has No Pause Button,” those concerns remain distinct from model governance. Land-use rules can address a facility’s physical effects, and utility regulation can assign the costs it creates. Neither requires officials to judge the safety of the models running inside. A data center is not an algorithm with a loading dock.

Politics can blur that line. Fear of AI can strengthen demands for data-center moratoria, special permits, electricity rationing, and negotiated permission systems. Pennsylvania’s GRID Standards, for example, offer expedited permitting, tax advantages, and coordinated state support to developers that accept state-preferred terms on energy, labor, and community benefits. Some conditions may address genuine local harms. The risk is that control over infrastructure becomes an indirect licensing system for AI—one that established firms can navigate more easily than smaller rivals.

Promising risk-free AI is not the answer. The next incident, and there will be many, will puncture any such claim. The stronger case rests on institutions that can uncover failures, assign responsibility, compensate victims, and improve systems. Developers can document their controls, customers can demand evidence, auditors can test claims, insurers can price risk, and courts can address concrete harms. That process cannot guarantee perfect safety, but it can build credible assurance without turning a land-use permit into an AI license.

Safety Without the Seal

The Hugging Face incident did not show that AI evaluation is futile. It showed evaluation for what it is—an adversarial discovery process in which the model, the test, the infrastructure, and the evaluators can all fail. No finite assessment can settle safety once and for all, much less become the price of permission to innovate.

AAE offers a practical alternative. Hayek locates knowledge about AI safety among dispersed developers, deployers, users, auditors, insurers, and victims. Kirzner explains how competition and experimentation uncover better practices. Marginalism asks whether each additional safeguard justifies its costs in a particular use. Public Choice warns that temporary checkpoints have a habit of becoming permanent gatekeepers.

The policy task is easy to state and difficult to execute. Build safety throughout the AI lifecycle. Test continuously. Make assurance claims verifiable. Assign liability according to control, fault, causation, and harm. Keep entry open so better approaches can emerge.

The goal is safer AI, not AI stamped “safe.”

The post Certified Uncertain: Why AI Safety Needs Assurance Tests, Not Permission Slips appeared first on Truth on the Market.

Until now, businesses, national authorities, and courts had to piece together the governing standards from treaty text, sporadic Commission decisions, and a growing body of judicial opinions. The Guidelines bring the rules on dominance, exclusionary effects, specific practices, and objective justifications into one framework. They also recognize that firms compete on quality, choice, and innovation, as well as price.

But the Guidelines do not mark a decisive return to disciplined, effects-based enforcement. They allow the Commission, in some cases, to dispense with a counterfactual, assess nonpricing conduct without a price-cost test, infer exclusion from conduct merely capable of producing it, and use presumptions to shift the evidentiary burden. The final text narrows some of the original draft version’s most aggressive express presumptions, as Global Competition Review has noted, but leaves ample room to expand liability later.

The Guidelines thus do more to organize and clarify Article 102 enforcement than to solve its central economic problem. The law remains highly receptive to precautionary intervention against successful firms, even when the challenged conduct may bring consumers substantial benefits alongside possible exclusionary risks.

No one seriously disputes that dominant firms can harm competition. The harder question—and the one these Guidelines do not settle—is whether the law can distinguish that harm from the ordinary process by which firms achieve dominance through investment, integration, innovation, and better service.

Rules of the Dominance Road

The Commission frames Article 102 as a three-stage inquiry. First, the enforcer assesses dominance. Second, it asks whether the conduct distorts “effective competition.” Third, the dominant firm may establish objective necessity or an efficiency justification. The Commission’s Article 102 page says the Guidelines aim to give national competition authorities, courts, businesses, and the Commission itself greater legal certainty and consistency.

The conduct list is long but familiar: predatory pricing, margin squeezes, conditional rebates, exclusive dealing, tying and bundling, access restrictions, refusals to supply, self-preferencing, and conduct “by its very nature harmful to competition.” The Guidelines also address aftermarkets, digital ecosystems, network effects, potential competition, and collective dominance. That breadth is useful. A digital platform or integrated firm cannot be analyzed as a simple, single-product monopolist. At the same time, digital-market characteristics do not, by themselves, establish abuse.

Several propositions deserve applause. The Guidelines recognize that Article 102 does not prevent firms from achieving dominance on their merits. Nor does it prohibit conduct merely because it marginalizes less-efficient rivals. Competition on the merits, they explain, includes lower prices, better quality, wider choice, and new or improved products. The Guidelines also require a theory of harm and say that findings of possible exclusion must rest on specific, concrete facts and evidence.

Recent Article 102 case law points in the same direction, at least in important respects. In Intel v. Commission, the European Court of Justice required the General Court to examine Intel’s evidence and arguments about whether its rebates could foreclose an as-efficient competitor. In Unilever Italia, the Court held that exclusivity clauses must be capable of producing exclusionary effects and that authorities must assess evidence submitted by the dominant firm. The Google Shopping judgment confirmed that self-preferencing by a dominant firm is not automatically unlawful.

These decisions reflect a basic error-cost principle: When conduct has a plausible efficiency justification, a legal label should not substitute for analysis. In their assessment of the draft Guidelines, Dirk Auer and Lazar Radic similarly warned that a selective reading of precedent could read presumptions of illegality into Article 102 while reading effects analysis out.

A High Bar With Plenty of Trapdoors

The trouble is a recurring pattern: The Guidelines announce a demanding principle, then qualify it until considerable administrative discretion seeps back in. The Commission says it must show that conduct is capable of producing exclusionary effects. Yet Article 102 has no de minimis threshold. The effects need not be actual or profitable. They need only be more than hypothetical. A firm cannot necessarily rebut the Commission’s case by showing that enough of the market remains open to accommodate competitors. The practical question is therefore whether exclusion seems plausible, not whether the conduct is likely to reduce consumer welfare by a material amount.

The Guidelines also treat the as-efficient-competitor (AEC) test asymmetrically. The Commission normally uses price-cost analysis for pricing conduct but generally declines to use it for nonpricing conduct because nonprice effects are difficult to quantify. It may instead rely on other evidence, without systematically constructing an alternative hypothetical scenario.

The July 2026 Google Android decision provides the immediate judicial backdrop. The Court upheld Google’s roughly €4.1 billion fine while accepting that the Commission could rely on the economic context without constructing a counterfactual in every respect. It also held that an AEC showing was not always necessary in digital markets.

That approach may be defensible in a narrow case with strong evidence that the challenged conduct makes entry or expansion practically impossible. As a general template, it is more troubling. Product quality, user experience, privacy, security, interoperability, and innovation are difficult to quantify. The greater that difficulty, the greater the risk that qualitative evidence becomes a respectable name for administrative intuition. A rule that lets the Commission avoid quantitative analysis precisely when the benefits and harms are hardest to measure should demand more institutional humility, not less.

Self-preferencing illustrates the problem. The Guidelines properly reject a categorical ban. They require both a departure from competition on the merits and the capability to produce exclusionary effects. But the relevant factors include whether users or market participants expect neutrality or openness, whether preferential treatment is “unjustified,” and whether it influences user behavior regardless of the favored product’s intrinsic qualities.

Those concepts can capture genuine manipulation. But they can also condemn ordinary product design. A platform may rank, integrate, or display its own service because doing so improves quality, lowers transaction costs, protects security, or simply gives users a more useful answer.

The category of conduct “by its very nature harmful to competition” carries even greater consequences. Examples include payments conditioned on not selling a rival’s product, dismantling infrastructure on which a competitor relies, and rules imposed by a firm that exercises both regulatory and commercial functions without sufficiently objective and precise procedures.

Once the Commission places conduct in this category, it deems the conduct to distort effective competition. A firm can successfully challenge that conclusion for lack of exclusionary capability only in “very exceptional” circumstances, while defenses based on objective necessity or efficiency are “very unlikely” to succeed. That walks and talks much like a presumption of illegality, even if the Guidelines decline to call it one.

The treatment of exclusive dealing is more disciplined than the broadest version contemplated in the original draft. The Guidelines establish a presumption for exclusive dealing, including de facto exclusivity, but require evidence of an intent to impose exclusivity in some cases involving volume thresholds. That narrowing helps.

Still, the presumption shifts the evidentiary burden to the firm before the Commission has necessarily shown substantial foreclosure, meaningful market coverage, or likely consumer harm. The Court’s reasoning in Intel and Unilever supports careful examination of a dominant firm’s evidence. It does not make the existence of an exclusivity clause the endpoint of the inquiry.

The efficiency defense offers another partial improvement. The Guidelines recognize cost and qualitative efficiencies, static and dynamic effects, and investments in research and development, innovation, and infrastructure. They also recognize related efficiencies outside the relevant market when the consumers who bear the alleged harm substantially overlap with those who receive the benefits.

But the dominant firm must prove four cumulative conditions: That the conduct produces efficiencies, that those efficiencies counteract the competitive harm, that the conduct is necessary, and that it does not eliminate effective competition. That framework makes sense as an administrable screen. But it may also render the defense illusory when innovation benefits are uncertain, long term, or spread across a platform ecosystem. The firm bears the burden of proof even though the Commission controls the investigation and may define the relevant markets and theory of harm in ways that exclude important benefits from the start.

My earlier analysis of the Commission’s interventionist turn warned that theories involving less-efficient competitors, constructive refusals, and expansive margin-squeeze enforcement could prompt dominant firms to raise prices, curtail discounts, or forgo investments that benefit consumers. My more recent analysis of Google Android makes the same point in dynamic terms: The lasting cost of a questionable precedent lies not just in the fine, but also in the products firms redesign and the integrated platforms they decide never to build.

Regulation by Overlap

The Guidelines make clear that the existence of another rulebook offers no safe harbor. Article 102 may apply even when EU or national regulation already covers the conduct, and compliance with another legal regime does not preclude an Article 102 violation. The practical consequence is a layered enforcement system in which authorities may examine the same conduct under Article 102, the Digital Markets Act (DMA), national abuse-of-dominance laws, consumer-protection or data rules, and sector-specific regulation.

The DMA offers the clearest example. Its gatekeeper obligations govern self-preferencing, steering, data use, interoperability, and other practices through ex ante duties that do not always require proof of consumer harm or the exclusion of an equally efficient rival. Article 102, at least in theory, remains effects-based.

In practice, however, DMA concepts and enforcement experience may shape the interpretation of Article 102. Article 102 proceedings, in turn, can reach nongatekeepers and conduct outside the core platform services designated under the DMA. The result may be “regulation by overlap,” with condemnation under one instrument serving as evidence of illegality under another.

Germany adds another layer. Sections 19 and 19a of the German Competition Act (GWB) apply, respectively, to dominant firms and firms deemed to have paramount cross-market significance. Section 19a specifically reaches self-preferencing, tying, data combination, interoperability, portability, access conditions, and benefits demanded from business users. Section 22 preserves stricter national rules alongside Article 102. The official English text of the GWB makes that point unusually clear.

The Court of Justice’s Towercast ruling likewise confirms that national authorities may use Article 102 to review certain transactions below the Commission’s merger-control thresholds. Decentralized enforcement can uncover local harms, but it can also produce inconsistent standards, duplicative proceedings, and remedies that pull platform design in conflicting directions.

This is not an argument for immunity. It is an argument for institutional coordination. The Commission should identify the instrument best suited to each concern, avoid cumulative punishment based on the same economic theory, and explain how Article 102 analysis differs from an ex ante DMA prohibition. Without those safeguards, legal uncertainty can itself become a barrier to entry—especially for firms that cannot afford separate product architectures and compliance teams across Europe.

Regulating Europe Back to Competitiveness

Europe’s competitiveness problem raises the stakes. The Draghi report on the future of European competitiveness identifies an innovation gap with the United States, weak productivity growth, difficulty scaling startups, fragmented capital markets, and an urgent need for private and public investment. That diagnosis belongs alongside Mario Zúñiga’s observation that Europe’s regulatory zeal has yet to produce a digital-market success story.

The lesson is not that regulation never helps. It is that errors carry unusually high costs in markets marked by rapid innovation, network effects, and uncertain demand. Product integration that appears exclusionary today may supply the investment needed to make a new service reliable tomorrow. A platform practice that diverts traffic from an intermediary may give users a faster, safer, or more useful product. Refusing to share an input may protect security or preserve the incentive to create that input in the first place.

Enforcers cannot fully observe the opportunity costs of prohibiting such conduct. They cannot count innovations never attempted, services never launched, or capital quietly invested elsewhere.

That is the error-cost case for a more demanding threshold for intervention. False negatives can cause real harm, but false positives in dynamic markets can permanently change the course of competition. The danger grows when regulators treat a firm as a quasi-public utility and require it to preserve rivals’ access, margins, or traffic. Competition law should protect the competitive process. It should not guarantee every rival’s survival or preserve every intermediary’s commercial position after a product improves.

Lazar Radic’s analysis of DMA enforcement against Google explains how a better product, integrated results, or a platform fee can become evidence against a firm precisely because success has made it powerful. The broader Digital Fitness Check submission documents how overlapping digital rules obscure both direct compliance costs and the opportunity cost of engineering time that could otherwise support innovation. Those costs do not stay on a compliance ledger. They redirect resources away from discovery.

Who Watches the Competition Watchdogs?

Public-choice analysis offers another reason for caution. Competition authorities are public institutions, not omniscient social-welfare maximizers. They respond to political attention, institutional priorities, limited resources, and the incentives created by their own powers. An investigation of a successful firm delivers visible benefits to complaining rivals and favored constituencies. The costs of a mistaken case—higher prices, less innovation, lost investment, and slower growth—emerge later and fall broadly across the economy. That asymmetry can encourage theories built around fairness, dependence, or ecosystem openness even when evidence of consumer harm remains thin.

The danger extends beyond crude protectionism. European firms may lobby for rules that constrain foreign rivals, while national authorities may favor domestic interests or seek a greater role in directing industrial outcomes. Once competition law dictates how a dominant firm must deal, display products, license technology, share data, or design an ecosystem, the boundary between antitrust enforcement and industrial regulation begins to disappear. The authority stops policing demonstrable exclusion and starts choosing business models.

That distinction matters in a global economy where China is a major source of competition, capital, technology, and state-supported industrial capacity. Europe should not answer Chinese competitive pressure by making its own dynamic firms more cautious. A legal regime that penalizes scale, integration, and risk-taking may push innovation toward the United States or China, leaving European consumers with fewer homegrown alternatives. Europe needs more productive rivalry. It will not get there by making European success legally hazardous.

Putting the Effects Back in Effects-Based Enforcement

The Guidelines can do more to improve Article 102 enforcement without giving genuine exclusion a pass. Five clarifications would help.

First, the Commission should require each theory of harm to identify a concrete, evidence-based mechanism linking the challenged conduct to a material risk of consumer harm. “Capability” should not encompass every logically conceivable path to exclusion.

Second, the Commission should reserve presumptions for conduct with a well-established exclusionary mechanism. Firms must remain free to rebut those presumptions with market-specific evidence. When available, evidence of actual entry, expansion, switching, product quality, and user behavior should count.

Third, the Guidelines should make the AEC test, or another economically coherent comparison, presumptively available whenever the conduct permits quantification. If quantification proves impossible, the Commission should explain why and identify the qualitative evidence serving in its place.

Fourth, the Commission should assess dynamic efficiencies at the same stage and with the same rigor as exclusionary effects. Long-term innovation, security, privacy, resilience, and infrastructure investment should not be relegated to a defense that becomes practically impossible to prove after an elastic finding of abuse.

Finally, the Commission and national authorities should develop a coordination protocol covering Article 102, the DMA, and national rules such as Section 19a. Firms should not face mutually inconsistent design mandates. Nor should an authority impose a remedy under one instrument without considering obligations already imposed under another.

Clarity Without Restraint

The Commission deserves credit for replacing its 2009 enforcement-priorities document with a comprehensive framework. The final Guidelines organize the law, acknowledge dynamic competition, reject a categorical offense for self-preferencing, recognize innovation efficiencies, and narrow some of the draft’s sweeping presumptions. These are real advances.

Yet organization does not guarantee restraint. The Guidelines preserve a broad conception of dominant firms’ special responsibilities, a low threshold for exclusionary capability, extensive reliance on qualitative evidence, limited counterfactual analysis, and a demanding efficiency defense. Combined with the DMA and national regimes such as Germany’s Section 19a, that framework may increase the risk that authorities regulate successful firms for the consequences of winning.

Europe needs competition policy rigorous enough to stop genuine exclusion and humble enough to recognize the limits of administrative knowledge. In the global contest for innovation, that requires taking false positives seriously, preserving firms’ freedom to integrate and invest, and distinguishing the protection of competition from the protection of competitors.

The Guidelines make Article 102 easier to navigate. The question is whether they also make overenforcement easier to justify.

The post Competing Too Well: Europe’s New Guide to Dominance appeared first on Truth on the Market.

Deere had previously reserved many of those resources for its licensed repair shops. Such control—especially over onboard software—is common in industries ranging from vehicles and medical devices to consumer electronics.

Critics contend that these restrictions allow manufacturers to “monopolize” repair aftermarkets. That concern has fueled right-to-repair campaigns and legislation across the country. In the early 2020s, the FTC promoted the right to repair as an anti-concentration policy that could spur competition. Advocates have also framed it as a matter of property rights: A farmer who buys a tractor owns it and should generally be free to repair it.

But the right to repair also implicates intellectual property. As one scholar argues, “While copyright and IP law protect firms to some degree, the broad extent of most proposed [right-to-repair] legislation across the states and federal government constitutes a threat to this protection.”

That concern deserves attention. One need not resolve the broader tension between intellectual property and economic competition to examine the more fundamental legal tension between intellectual property and property rights.

A Fix That Doesn’t Break IP

The FTC found a sensible compromise. Farmers who buy Deere tractors can exercise their property rights over the machines and their components. Deere keeps its intellectual property, including its trade secrets, while giving owners and independent shops access to the software and resources needed for repairs.

That balance protects what matters most to each side. Some activists treat intellectual property as little more than a corporate shield against competition. The FTC took it seriously, while drawing its boundaries so farmers could meaningfully exercise their own property rights. Farmers can now repair their machines when time matters most—during harvest, for example—without forcing Deere to surrender its intellectual property. The compromise fits comfortably within longstanding law.

The patent-exhaustion doctrine has long provided that selling a patented product exhausts the patentholder’s control over that particular item, allowing its owner to resell, repair, or alter it. Both common law and Supreme Court precedent, including Wilson v. Simpson, recognize that principle. The Deere settlement confirms that farmers enjoy those rights over the equipment they purchase. Deere, meanwhile, retains its underlying patents and may provide repair access on “fair and reasonable” terms.

This compromise was taking shape well before the settlement. When the FTC sued Deere, Andrew Ferguson—then an FTC commissioner and now its chairman—dissented, noting that “the parties are in active negotiations over a fix that, if brought to fruition, could provide meaningful relief to America’s farmers.” He was right. Deere had already expanded farmers’ repair options through a 2023 service program and was seeking guidance from federal agencies.

Against that backdrop, the settlement looks less like forced compliance than the formalization of an expected compromise. It places the right to repair on more consensual and predictable legal ground.

The Case Law Under the Hood

Courts mapped this terrain decades ago, distinguishing repairs that preserve a machine’s identity and useful life from reconstruction that effectively creates a new patented product.

In the 2008 case Quanta Computer Inc. v. LG Electronics Inc., the Supreme Court unanimously held that patent exhaustion applies to components that substantially embody a patented invention. The principle reaches beyond the tractor as a whole to the parts and software necessary for its ordinary use. Once a patentholder authorizes a sale, it cannot use patent law to retain control over how the buyer uses the product.

The Supreme Court addressed the competition side of the problem in Eastman Kodak Co. v. Image Technical Services Inc. in 1992. Kodak had refused to sell replacement parts to independent companies that serviced its equipment. The Court held that Kodak could face antitrust liability for using its control over those parts to exclude competitors from the repair market.

The clearest precursor to the Deere settlement is the Supreme Court’s 2017 decision in Impression Products Inc. v. Lexmark International Inc.. The Court ruled 8-1 that Lexmark could not sell its patented printer cartridges and then use patent law to restrict their reuse or resale.

Writing for the Court, Chief Justice John Roberts reaffirmed the exhaustion doctrine: “When a patentee chooses to sell an item, that product ‘is no longer within the limits of the monopoly’ and instead becomes the ‘private, individual property’ of the purchaser.” He added, “That smooth flow of commerce would sputter if companies that make the thousands of parts that go into a vehicle could keep their patent rights after the first sale.”

The same principle applies to Deere. Once farmers buy its equipment, Deere cannot use patent law to prevent them from using and maintaining it.

Repair Without Ruin

The FTC’s Deere settlement offers a workable right-to-repair model: Give consumers meaningful control over the products they buy without stripping innovators of their intellectual property. Whether courts and lawmakers follow remains to be seen. The best repair policy fixes access without breaking innovation.

The post Nothing Runs Like a Deere You Can Repair appeared first on Truth on the Market.

The real estate market has evolved considerably since then, but buyers and sellers, typically through their agents, are still trying to solve the same information problem. Buyers look for homes with specific features, such as location, number of rooms, and lot size, while sellers look for buyers whose preferences match the home they are selling. Real estate agents play a central role in this matching process by building a network that raises the likelihood of a timely sale.

Today, the leading real estate listing database is the Multiple Listing Service (MLS). It traces its origins to San Diego in 1885, when the local real estate board would distribute daily listings to its members. The MLS is not a single entity; it is composed of roughly 500 local real estate organizations, which are owned or controlled by brokers and associations in their respective territories. Members create the listings, while the regional MLSs distribute them to other agents and to public-facing websites such as Zillow, Redfin, and Realtor.com.

There are clear efficiencies from having a single, comprehensive database to economize on search costs and to raise the odds of a timely match. The tradeoff, however, is that, within a given territory, a viable second source of listings is unlikely to emerge. This raises the concern that whoever controls the listings data controls the terms of competition involving those listings.

This is exactly the issue at the heart of the recent antitrust dispute between Zillow, on one side, and Midwest Real Estate Data (MRED), a regional MLS serving the Chicagoland area, and Compass, the nation’s largest real estate brokerage, on the other. Zillow depends on MRED for listings in the Chicagoland area, where, according to Zillow’s complaint, MRED holds a 98% share.

The Rise of Private Listing Networks

Private listing networks (PLNs) have recently gained ground in real estate. Zillow defines them as “listings that are not visible to the general public on consumer-facing real estate websites and are often excluded from MLSs.” Instead, agents market these properties to a closed network of buyers’ agents, often from the same brokerage as the listing agent.

MRED has operated its connectMLS Private Listing Network since 2016. It functions as a separate database where agents place “draft” listings. These properties remain in a kind of pre-listing limbo until agents move them to the general, public MLS database.

Compass launched its own private listing network in 2024. It uses a three “phase” approach for listings. Phase one is called “Compass Private Exclusive,” where a listing is only circulated among Compass’s closed network of agents and clients. Phase two is called “Compass Coming Soon,” where a listing is moved to Compass’s website and to partner websites—but the listing is still not part of the MLS database. For listings that have not sold, there is a phase three, where the listings are moved to the general MLS database and public portals such as Zillow. Notably, reports indicate that 94% of phase one listings reach phase three.

Why Sellers Choose Private Listings

The growth of PLNs raises a key question: If the real estate market is fundamentally about finding matches to spur transactions, why would a seller agree to limit exposure of their listing to a closed network when it could immediately enjoy wider visibility on the MLS database? Below, I explore several possible answers.

The first potential explanation is the classic principal-agent problem, where the incentives of the principal and agent do not fully align. In a notable study, “Market Distortions When Agents Are Better Informed,” Steven Levitt and Chad Syverson examined how real estate agents behave when selling their own homes compared with their clients’ homes. When they sold their own homes, the agents were generally more patient (keeping their homes on the market for 9.5 days longer), which resulted in a higher sale price (approximately 3.7% higher).

In contrast, since agents only keep a small percentage of an incremental sale price for a client, but bear the full cost of marketing a home for longer, they have incentivizes to sell client properties too quickly and cheaply. Interestingly, the data used in the study comes from suburban Cook County in Northern Illinois, which is an area that MRED covers.

Given this context, the principal-agent issue for PLNs is that seller agents may prefer to start a listing on a PLN because they save on the costs associated with a widespread, public marketing campaign. Additionally, the use of a closed network is more likely to keep the sale within the same brokerage, a practice called “double-ending.” According to the complaint, sales that are made “off-MLS” double-end 31% of the time versus 18% for regular, “on-MLS” sales. This clearly benefits large brokerages like Compass by keeping more of the commissions in-house.

A second possible explanation is that exclusivity can help promote a home, even among a smaller group of buyers and agents. Specifically, having access to a closed network can create a sense of urgency and encourage buyers to act quickly (akin to a pre-sale reserved for loyalty-card customers at retail outlets).

There is also a question of whether a seller is giving up bargaining leverage inside a PLN—as there is always the option to later move the listing to the public database. Nonetheless, a risk-averse seller may take a bird in hand over a prolonged and uncertain listing period in the open market.

A third explanation is privacy. An exclusive listing may appeal to homeowners who do not want to publicize a sale. A prerelease may also help gather feedback from other agents and refine the listing price. It is unclear, though, whether those benefits require an exclusive period or merely sufficient research before the public listing.

The empirical record on whether sellers benefit from private listings remains unsettled, and studies sponsored by interested parties complicate the picture. Depending on the source, sellers fare either better or worse. Comparisons may not involve genuinely similar properties because the homeowners and agents who choose private listings may differ systematically from those who do not, a problem researchers call selection effects. More rigorous studies using updated data could provide a clearer answer.

How Private Listings Shift Market Power

Whatever motivates the growth of private listing networks, their rise changes the competitive dynamics between real estate platforms such as Zillow and brokerages such as Compass. If most buyers find homes through Zillow, Zillow enjoys greater traffic, more advertising revenue, and greater bargaining power over various affiliate or referral agreements. If, instead, more buyers find homes through agents, then the economic rents shift to brokerages—particularly those such as Compass who control a significant share of local inventory.

Exclusive listings can also help a brokerage attract clients. If sellers and buyers believe Compass has the newest and most desirable listings in a market, particularly when inventory is low, access to that network becomes a reason to hire a Compass agent rather than browse a platform such as Zillow.

Relatedly, if public portals only receive “leftover” homes that did not sell during the private period, then this changes the nature of their databases and creates greater demand for PLNs operated by brokerages like Compass.. In its complaint, Zillow warns that “[l]osing access to listings and having lower-quality listings would risk triggering a harmful feedback loop that would degrade Zillow’s platform further.”

Zillow’s Response to Private Listings

In April 2025, Zillow announced that it would no longer display homes previously marketed through a closed private listing network. Zillow does offer exceptions for homeowners seeking greater privacy through a limited release.

Zillow argues that sellers’ agents remain free to use private listing networks, but they cannot have their cake and eat it too—i.e., by later obtaining public distribution through Zillow. In its view, allowing them to do both would put Zillow at a competitive disadvantage and reduce the quality of its platform as it competes to match buyers with sellers.

Zillow emphasizes that it adopted the policy unilaterally. Generally, antitrust treats unilateral acts fundamentally differently than coordinated ones because the latter can circumvent natural rivalries that benefit consumers.

Of course, under certain conditions, unilateral acts can also impair the competitive process. Those conditions typically involve assessments of market power in a properly constructed relevant market, offsetting procompetitive justifications. It must also account for the Supreme Court’s ruling in Trinko, which emphasized that firms generally have no antitrust duty to deal with competitors. That principle is likely even stronger when competitors seek access while attempting to disadvantage the firm in other areas.

The Fight Moves Beyond the Courtroom

In June 2025, Compass sought a nationwide preliminary injunction—a court order that would have blocked Zillow’s policy while the case proceeded. In February 2026, U.S. District Judge Jeannette Vargas of the Southern District of New York denied the motion because Compass had not shown that its antitrust claims were likely to succeed.

Among her key findings, Vargas concluded that Compass had not shown that Zillow possessed enough market power to exclude competitors. She also found no evidence supporting Compass’ allegation that Zillow had colluded with Redfin to implement the private-listing policy. The court therefore concluded that Compass had failed to establish a likelihood of success on the merits. Compass later voluntarily dismissed the case.

According to Zillow’s complaint, the conflict continued outside the courtroom. In October 2025, Compass’s CEO sent messages to at least eight MLSs urging them to discipline Zillow, and if Zillow did not drop its PLN policy, to block their listing feeds to the platform. A few weeks later, MRED revised its display rules to promote “non-discrimination” that implicated Zillow’s PLN policy. MRED’s CEO informed Zillow that it would cut off the company’s access to the MRED feed if it enforced its policy.

Over the ensuing months, Compass-owned Coldwell Banker brokerages terminated more than direct listing feed agreements with Zillow. Finally, on May 8, 2026, Compass terminated all Compass or Compass-owned brokerage feeds to Zillow nationwide.

Meanwhile MRED and Compass announced a partnership on April 24, 2026. Compass agreed to provide MRED with its national inventory of Private Exclusive and Coming Soon listings, which MRED will make available to its PLN participants. Further, Compass agreed to subsidize MRED membership fees for the first 100,000 Compass agents to join MRED as full members. For its part, MRED committed to protect agents participating in its PLN from being banned or penalized by third-party portals like Zillow.

Zillow’s Antitrust Claims

Zillow alleges that MRED and Compass violated Section 1 of the Sherman Act, pleaded as a per se unlawful group boycott or, at the very least, a violation under the rule of reason. The theory is that MRED and Compass agreed to protect their joint interests and deny Zillow a critical input—that is, MRED’s Chicagoland feed—unless Zillow abandoned its PLN policy. Notably, Zillow argues that the conspiracy began well before the formal April 24, 2026, partnership announcement.

Zillow separately alleges a Section 2 claim against MRED for monopoly maintenance—based on the theory that Zillow is a nascent competitor in the listings database market through its Zillow Preview product. This recent product offers brokers the opportunity to “preview” their listings before moving them to the larger MLS database.

Compass and MRED accuse Zillow of using the Preview product to do the same thing that it claims harms consumers and the market. Zillow responds that it isn’t a private database only available to a select group of agents. Instead, Zillow contends that Preview is a public listing that merely exists outside of the MLS database. Additionally, Zillow highlights that MRED faces little to no competition in real estate listings in the Chicagoland area.

The Core Question: Competition or Exclusion?

The rise of private listings is at the center of this antitrust dispute, and the empirical evidence on whether this results in better outcomes for sellers is still emerging. But ultimately, the case is not about whether private listings are good or bad for market participants.

Moreover, the case is not fundamentally about Zillow’s unilateral PLN policy or even Compass’s response. Rather, the case boils down to the partnership between Compass and MRED, and whether that partnership includes an agreement to jointly withhold listing data from Zillow. Such an agreement, if proven, raises legitimate questions about whether the group boycott impairs the competitive process and is using a vertical supply chain partner (i.e., MRED) to leverage market power to advantage Compass in its horizontal competition in listings discovery with Zillow.

Group boycotts have a long history in antitrust law, as I document in a recent article. The court must now decide whether this was vigorous competition over listing strategies or concerted exclusion from a critical source of data. That question has consequences beyond the courtroom, particularly given the congressional attention that private listing networks have attracted.

The post Location, Location, Litigation: The Antitrust Fight Over Private Listings appeared first on Truth on the Market.

Personalized pricing is a technologically refined form of price discrimination—charging different customers different prices for the same product or service. But it is not one practice with one predictable effect. Depending on the market, consumer demand, competition, and what would happen without personalization, it may lower prices, expand output, sharpen competition, or raise legitimate consumer concerns.

The FTC should therefore draw a clear legal line. A seller may not misrepresent what a price is, who can obtain it, or why it was offered. But a seller does not necessarily deceive consumers merely by failing to volunteer that individualized information helped determine the price. Turning that silence into a violation would allow a consumer’s assumption to create a duty to disclose—and could make beneficial discounts harder to offer.

As a new Mercatus Center policy comment puts it, “[a] final statement built around those limits, coupled with express protection for individualized discounts, would deter genuine deception and unfairness while preserving price competition and experimentation that can lower prices and expand output.”

Uniform Pricing Isn’t Uniformly Better

The economic case for caution is simple. A uniform price may exclude consumers willing to pay more than the cost of serving them but less than the single price that maximizes the seller’s profit. If the seller can identify these price-sensitive consumers, it may profitably offer them discounts, sell more, and reduce deadweight loss—the value lost when mutually beneficial sales never happen. Personalized offers can also sharpen competition by targeting discounts at consumers most likely to switch sellers. The United States made these points in a submission to the Organisation for Economic Co-operation and Development (OECD), explaining that personalized pricing can enhance competition when it does not involve deception or unlawful discrimination.

The word “personalized” does not answer the central question: Who benefits, and compared with what? As John Yun has explained, the competitive setting and the alternative to personalized pricing matter more than personalization itself. The same practice can have different effects in a highly competitive market, a concentrated market, and one in which consumers can readily compare offers.

The empirical evidence also resists easy conclusions. Jean-Pierre Dubé and Sanjog Misra found that personalization reduced total consumer surplus—the aggregate benefit consumers receive beyond what they pay—in their experimental setting. Yet more than 60% of consumers benefited. Research on Amazon found personalized net prices delivered through targeted coupons, rather than a simple pattern of individualized surcharges. That distinction matters. If regulators make targeted discounts legally risky, sellers may replace them with higher uniform prices, harming the consumers who would have received a deal.

Regulators also should not treat every price set by an algorithm as exploitation with better software. Economic analysis of algorithmic pricing emphasizes that the relevant question is whether the practice improves market performance, not whether a computer helped set the price. Treating pricing technology as presumptively suspect could suppress the experimentation that helps firms discover cheaper ways to serve consumers.

Your Price Is Not Everyone’s Price

The proposal stands on firm ground when a seller makes a false claim—for example, by stating or clearly implying that everyone receives the same price when they do not. The legal trouble begins when the FTC suggests that a seller may deceive consumers whenever they reasonably believe a price is fixed or widely available and the seller fails to disclose that it was personalized.

That approach risks allowing a consumer’s preexisting belief, rather than anything the seller said or did, to create a duty to speak. It is hard to square with the FTC’s settled test for deception. That test asks whether a representation, omission, or practice is likely to mislead a reasonable consumer on a material point—one that could affect the consumer’s decision. The FTC’s Policy Statement on Deception sets out this test.

The FTC’s own decisions distinguish a misleading omission from a “pure omission”—something the seller never addressed and whose silence conveys no particular message. In International Harvester, the FTC warned that deception could expand almost without limit if consumers’ mistaken preconceptions were enough to establish it. The FTC’s Food Advertising Policy Statement makes the same point: Not every omission is deceptive, and silence misleads only when the context gives it that meaning.

A displayed price ordinarily communicates one straightforward proposition: The consumer may buy the product at that price. It does not necessarily promise that every other consumer can get the same price at the same time. That implication is especially doubtful in markets where coupons, loyalty programs, negotiated prices, retention offers, and targeted promotions are commonplace.

If a transaction actually conveys a uniform-price claim, established doctrine already gives the FTC the tools to prove it. The agency may use consumer surveys and other evidence beyond the seller’s words to show how reasonable consumers understand an implied claim. That is a long way from requiring every seller to explain its pricing model when it never suggested that prices were uniform.

A Disclosure Mandate in Search of a Statute

The proposal’s reliance on FTC v. Colgate-Palmolive does not erase this distinction. That case involved a false claim about how an advertised “special” price compared with the product’s ordinary selling price. It did not require sellers to volunteer every material fact consumers might want to know about how they set prices.

The Fair Credit Reporting Act (FCRA) does not establish a broader disclosure duty under Section 5 of the Federal Trade Commission Act, which prohibits unfair or deceptive business practices. Congress expressly required businesses to provide adverse-action notices when they use consumer reports in certain covered decisions. That targeted requirement appears in 15 U.S.C. § 1681m(a). Its specificity should make the FTC cautious about reading Section 5 alone to require similarly detailed disclosures about whether a price was personalized, why it was personalized, and what kinds of data informed it.

Disclosure also carries costs. An explanation of a complex pricing system may confuse consumers more than it informs them. A rigid mandate to identify every “type” of data used could also push firms to abandon useful personalization and fall back on blunter, higher prices. Disclosure should correct a misleading impression—not force a seller to publish the blueprints for its pricing engine.

Someone Else’s Bargain Is Not Your Injury

The FTC needs the same discipline when invoking Section 5’s unfairness authority. Section 5(n) requires the agency to prove three things: substantial consumer injury, injury consumers could not reasonably avoid, and injury not outweighed by benefits to consumers or competition. The statutory text makes each requirement explicit, and the 11th U.S. Circuit Court of Appeals recently emphasized that the FTC must satisfy all three.

The right comparison is not necessarily the price another consumer received. It is the price the consumer likely would have paid without personalization. Suppose a seller would otherwise charge everyone $10 but instead offers one shopper $7 and another $11. That raises a different question from a case in which both shoppers would have paid $12 without personalization. Calling the $4 gap between the personalized prices an injury—without asking what the uniform price would have been—confuses price differences with consumer harm.

Whether consumers could reasonably avoid the injury also depends on how the transaction works. An upfront price that consumers can compare with competing offers differs from a fee revealed only after they become locked in. The FTC should ask what consumers knew, when they knew it, and whether they could walk away—not merely whether someone else got a better deal.

The proposal also seems to assume that sellers can preserve every benefit of personalized pricing while disclosing the practice. That is an empirical claim, not a legal axiom. Consumers may hide information to qualify for discounts, prompting firms to reduce targeted offers. The cost of explaining how a complex price was calculated may also discourage experimentation. Research on consumer tracking helps explain why firms often personalize prices through coupons and other discount mechanisms.

Section 5(n) requires the FTC to examine these effects. The agency cannot wave away competitive benefits simply because it can imagine a disclosure that might—emphasis on might—leave the pricing system unchanged.

Police the Lie, Preserve the Discount

The FTC can protect consumers without creating a de facto ban on personalized pricing. Its final statement should identify actionable conduct precisely: falsely claiming that a price is available to everyone, advertising a fictitious discount, misrepresenting why a price was offered, or deceptively collecting or using personal data. The remedy in each case should correct the particular false impression.

The FTC should also create an enforcement safe harbor—a rule protecting clearly defined lawful conduct—or at least a strong presumption against enforcement for genuine individualized discounts. That protection should cover personalized prices no higher than a genuine price generally available from the same seller at the same time. It should also cover targeted coupons, opt-in loyalty benefits, and retention offers unless the seller makes a separate false or misleading claim.

Such protection would reduce the risk that legal uncertainty prompts firms to withdraw lower-price offers. It would also preserve firms’ ability to test pricing methods that may strengthen competition. As I have previously argued, legal challenges to algorithmic pricing can impede such improvements by discouraging firms from testing new ways to match prices with consumer demand.

The governing principle is straightforward. Sellers must tell the truth about their prices. That obligation should not become a general duty to explain how every price was calculated simply because the FTC believes consumers expect uniform pricing. Deception requires a misleading representation, omission, or practice. Unfairness requires consumer injury that outweighs the benefits to consumers and competition.

The FTC should police the lie and preserve the discount.

The post Your Price May Vary: The FTC’s Personalized-Pricing Puzzle appeared first on Truth on the Market.

Every day we make decisions without possessing most of the information on which those decisions ultimately depend. We purchase products whose origins we cannot trace, rely on medicines we did not test, respond to prices shaped by events occurring thousands of miles away, and invest retirement savings in companies whose operations we will never observe firsthand. Somehow, despite possessing only fragments of the relevant information, billions of independent decisions become coordinated without anyone directing the whole.

How is that possible?

Nobody Knows Everything. Everyone Knows Something.

In his 1945 essay “The Use of Knowledge in Society,” Friedrich Hayek offered one of the most influential answers in economics. Hayek begins with a deceptively simple observation:

The economic problem of society is … not merely a problem of how to allocate ‘given’ resources ….

As he explains in one of the essay’s most famous passages:

The peculiar character of the problem of a rational economic order is determined precisely by the fact that the knowledge of the circumstances of which we must make use never exists in concentrated or integrated form, but solely as the dispersed bits of incomplete and frequently contradictory knowledge which all the separate individuals possess.

Hayek’s contribution lay not in offering a new solution, but in redefining the problem itself. Earlier debates often assumed that the relevant information about an economy could, at least in principle, be assembled by a single decision-maker—and that the challenge was then deciding how to allocate resources. Hayek argued that assumption was fundamentally mistaken. The most valuable knowledge in an economy, he contended, never exists in one place. Instead, it takes the form of what he famously called “the knowledge of the particular circumstances of time and place.”

Hayek used “knowledge” in a broader sense than economists often do today—referring not to scientific discoveries or official statistics, but to countless observations people acquire through everyday experience. A farmer notices that unusually dry weather has changed this year’s harvest. A factory manager knows a machine has become unreliable. A shopkeeper recognizes that customers have begun asking for a product few competitors yet carry.

Individually, these observations seem insignificant. Collectively, they determine what should be produced, where resources should flow, and how an economy adapts to changing conditions. Because this knowledge is practical, local, and constantly changing, it cannot easily be collected in reports or stored in databases without losing the context that gives it value.

The Most Efficient Gossip Network Ever Built

Recognizing the problem, however, is only half the story. If knowledge is irreducibly dispersed, how can millions of people coordinate their decisions without assembling all the relevant facts? (Coordination here means individuals adjusting their own plans in ways that become mutually compatible without anyone directing the process.) The mechanism itself was familiar; Hayek’s insight was that individuals do not need the complete picture.

Prices do far more than determine what goods cost. They communicate only the information necessary for individuals to revise their own plans in light of changing economic conditions—without requiring them to understand the countless events that produced those changes. A changing price reflects the cumulative effect of countless individual judgments about scarcity, demand, opportunity costs, and alternative uses. Rather than communicating each of those judgments separately, the price system compresses them into a single signal that others can act upon.

In this way, the price system economizes on information, enabling decentralized coordination without centralized direction or comprehensive knowledge. Individuals need not know why copper has become scarce or wheat abundant. They need only observe the changing price.

Hayek illustrated this with the example of a shortage of a single raw material. Most people would never learn whether it resulted from a mine closure, increased military demand, or a natural disaster. They would not need to. As prices rose, manufacturers would conserve the material, consumers would seek substitutes, and entrepreneurs would search for new sources of supply. As Hayek famously observed:

Without an order being issued, without more than perhaps a handful of people knowing the cause, tens of thousands of people whose identity could not be ascertained by months of investigation are made to use the material or its products more sparingly….

No planner needs to issue commands. No one needs to understand the entire economy. The genius of the market is not that everyone becomes fully informed—it is that coordinated action emerges because prices allow individuals to respond intelligently to information they never directly possess.

Hayek regarded this as one of civilization’s greatest achievements. As he memorably wrote:

I have deliberately used the word “marvel” to shock the reader out of the complacency with which we often take the working of this mechanism for granted.

More than 80 years later, “The Use of Knowledge in Society” remains one of the foundational essays in law & economics precisely because it explains why decentralized institutions can often outperform centralized planning in a world of dispersed information. 

Hayek Stopped Too Soon

Hayek’s framework reflects the particular question he set out to answer. His concern was not how knowledge comes into existence but how societies make use of knowledge that already exists. Individuals possess knowledge of their land, their machines, their preferences, and the opportunities they encounter. Hayek explained how these dispersed fragments could be coordinated without centralized direction. But his account largely assumes that the knowledge itself already exists—leaving open a different question: Who creates the knowledge that markets later coordinate?

The distinction becomes clearer when Hayek is placed alongside two related traditions. Ludwig von Mises’s calculation problem asked whether central planners could allocate resources rationally without the market prices generated through exchange. Hayek reframed that problem by emphasizing that the difficulty was not simply calculation but knowledge itself—much of it dispersed among individuals and impossible to assemble centrally. Ronald Coase, meanwhile, showed that using markets is itself costly and that firms may arise when organizing transactions internally is cheaper.

The problem of knowledge production is related to each of these insights but distinct from them. Before information can be communicated through prices or coordinated within firms, institutions may first have to invest resources in making it sufficiently reliable to be used at all.

Some forms of knowledge emerge naturally through decentralized interaction. Consumer preferences, for example, reveal themselves through countless individual choices. Other forms do not. They must be produced through deliberate investments in measurement, investigation, verification, and disclosure. A pharmaceutical risk cannot influence consumer behavior until someone detects it. Corporate fraud cannot affect investors until someone uncovers it. A defective product cannot alter purchasing decisions until someone tests it, documents the results, and makes those findings public.

Markets can coordinate only the knowledge that becomes available to them. The production of knowledge is therefore just as important to the functioning of markets as its coordination.

The institutions that produce this knowledge rarely occupy center stage in economic theory. Yet modern societies devote enormous resources to creating reliable information before it can ever circulate through markets. Scientists conduct experiments. Firms maintain records. Auditors verify accounts. Journalists investigate misconduct. Courts establish facts through litigation. Administrative agencies inspect facilities, collect data, and evaluate evidence. Universities train experts. Professional organizations establish standards.

Each institution employs different methods, but all perform a common function: transforming uncertainty into knowledge that others can trust. None simply discovers facts lying dormant in the world. Each invests labor, expertise, and judgment to measure, verify, and authenticate observations before they become sufficiently reliable for others to act upon. Only then can markets, consumers, regulators, and investors incorporate that knowledge into their decisions.

Hayek’s account brilliantly explains how society coordinates dispersed knowledge once it exists. But the knowledge that makes coordination possible is itself the product of institutions.

Economists have not entirely ignored the problem of knowledge production. Kenneth Arrow famously observed that information is costly to produce but inexpensive to reproduce, creating incentives to underinvest in its creation. Elinor Ostrom showed how communities develop institutional arrangements that draw upon local knowledge to govern common resources. Both contributions draw attention to the institutions and incentives that shape how economically valuable knowledge becomes available.

Legal scholarship has begun to explore pieces of this puzzle as well. Administrative law scholars increasingly recognize that agencies do more than enforce legal rules—they inspect facilities, collect evidence, evaluate technical claims, and produce authoritative factual determinations on which private actors depend. Corporate law scholars have examined the informational role of disclosure regimes, auditors, and gatekeepers. Scholars of science and technology studies have emphasized the institutional work required to transform raw data into reliable public knowledge.

Yet these literatures often proceed independently. What remains underdeveloped is a broader account of how agencies, courts, firms, and private intermediaries produce the reliable knowledge on which markets, regulators, and private ordering all depend.

Hayek taught economists that the central challenge of modern society is coordinating knowledge that no single person possesses. Eighty years later, technological change might seem to have weakened the force of that insight. Digital platforms collect information from millions of transactions in real time, governments and firms maintain vast datasets, and artificial intelligence can identify patterns across quantities of information no human could process. Some knowledge that once could not feasibly be centralized now can be.

Yet these technologies do not eliminate the problem of knowledge production. More data is not necessarily better knowledge. Before information can be aggregated or analyzed, someone must decide what to measure, collect and classify observations, verify their accuracy, and determine whether it can be trusted. Artificial intelligence can dramatically reduce the costs of processing existing information, but its outputs remain only as good as the data and institutions that produce them.

The 21st-century challenge, then, is not simply how to coordinate dispersed knowledge. It is how that knowledge is produced, who bears the costs of producing it, and which institutions deserve our trust—questions that sit at the intersection of economics, law, and institutional design. Hayek gave us the map of how knowledge travels. We still need a theory of who builds the roads.

Further Reading

• Friedrich A. Hayek, “The Use of Knowledge in Society,” The American Economic Review, Vol. 35, No. 4 (1945).
• Kenneth J. Arrow, “Economic Welfare and the Allocation of Resources for Invention,” in  “The Rate and Direction of Inventive Activity: Economic and Social Factors,” Princeton University Press (1962).
• Ronald H. Coase, “The Nature of the Firm,” Economica, Vol. 4, No. 16 (1937).
• Haiyun Damon-Feng, “Agency Fact-Making,” Yale Law Journal, Vol. 135 (2026).

The post ‘The Use of Knowledge in Society,’ by Friedrich A. Hayek appeared first on Truth on the Market.

I had more to say in a later post. I also joined my colleagues Geoffrey Manne and Kristian Stout in submitting more detailed and careful comments to the public record on behalf of the International Center for Law & Economics (ICLE). While they were more detailed and careful, those comments were no more enthusiastic. The same description fits the joint comments filed by George Mason University’s Program on Economics and Privacy and the Tech Law Program at the University of Arizona’s James E. Rogers College of Law. 

We were hardly alone. Many outside the agency were surprised—if not alarmed—by the ANPR’s sweeping scope, lack of substantive clarity, and general hostility toward the digital economy. Critics included Alden Abbott, a former FTC general counsel; Jonathan Barnett of the University of Southern California; and Svetlana Gans and Natalie J. Hausknecht, writing jointly. Gans previously served as FTC chief of staff. 

The concerns reached inside FTC headquarters as well. Commissioners Noah Joshua Phillips and Christine Wilson each vigorously dissented from the vote to issue the ANPR. Phillips objected, inter alia:   

The Commercial Surveillance and Data Security advance notice of proposed rulemaking (“ANPR”) issued today by a majority of commissioners provides no notice whatsoever of the scope and parameters of what rule or rules might follow; thereby, undermining the public input and congressional notification processes. It is the wrong approach to rulemaking for privacy and data protection security.

What the ANPR does accomplish is to recast 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. That is a dramatic departure even from recent Commission rulemaking practice. The ANPR also contemplates taking the agency outside its bailiwick. At the same time, the ANPR virtually ignores the privacy and data security concerns that have animated our enforcement regime for decades. A cavalcade of regulations may be on the way, but their number and substance are a mystery.

No notice of proposed rulemaking followed, and neither did a final rule. That was, in my view, the best possible outcome given the ANPR. If the agency had to begin a rulemaking, a more sober and restrained ANPR would have been better—one that recognized the information economy’s consumer benefits alongside its risks and avoided the skewed assumptions and loaded language of Shoshana Zuboff’s anti-tech polemic on “surveillance capitalism.” Better still would have been greater regulatory restraint from the start.

The FTC Marks Down Its Ambitions 

The FTC, now under new leadership, has returned to the issue with a “Proposed Enforcement Policy Statement Regarding Personalized Pricing.” Personalized pricing was among the many—many—concerns raised in the 2022 ANPR. Here, the term means charging different consumers different prices using “modern data collection and processing capabilities,” including “consumers’ personal data.” 

In a nutshell, the policy statement says:  

Where consumers reasonably expect that prices for a product or service will not vary based on their personal data, businesses that engage in personalized pricing should clearly and conspicuously disclose not just that the price is personalized, but also the basis for that personalization and the types of data on which the personalization is based. The failure to make these disclosures is likely to constitute an unfair or deceptive act or practice in violation of Section 5. The Commission intends to deploy enforcement resources in a manner consistent with this conclusion.

This pass at the issue improves on the last one in several important respects. For one, the policy statement is more modest than the ANPR. It proposes guidance with no independent legal force, not a federal regulation that stretches (if not ignores) the limits of the agency’s statutory authority. In doing so, it recognizes the substantive and procedural limits on the FTC’s jurisdiction. As the policy statement acknowledges, “Congress has not given the Commission the authority to prohibit personalized pricing in all circumstances.” 

The policy statement also recognizes that personalized pricing may involve tradeoffs among consumers, benefiting some while leaving others worse off, and that regulators should not always condemn those tradeoffs. Indeed, the FTC observes that “personalized pricing is a long-established norm in some markets.” True enough. Think of car lots, residential-home sales, and, more broadly, markets in which buyers and sellers commonly negotiate terms. The implication seems to be that, in such markets, enforcers should not presume that failing to disclose personalized pricing constitutes a deceptive omission under Section 5 of the FTC Act, which prohibits unfair or deceptive business practices. 

Personalized pricing is a form of price differentiation, also known as “price discrimination”—the economic term for charging different buyers different prices for the same product. Encouragingly, the policy statement recognizes some of the economic literature on differential pricing. It notes: 

[R]eal-world personalized pricing would be less than perfect in its ability to identify consumers and determine their willingness to pay and, in the vast majority of circumstances, would have to contend with competitive pressures that limit pricing power.

In other words, the FTC does not expect to see perfect first-degree price discrimination, in which a monopolist charges each consumer the maximum amount that person is willing to pay. More plausibly, many instances of personalized pricing (perhaps most) will resemble third-degree price discrimination, which sorts consumers into broader groups and charges each group a different price. Some data may be specific to the consumer at the register or keyboard, but the resulting price will not precisely identify that consumer’s willingness to pay (WTP). Data and processing limits will constrain the price, as will competition, because (as the policy statement recognizes) many common consumer markets are not monopolies.

Suppose I am buying milk at the grocery store. Personalized pricing might produce different prices for different groups, but I could still pay the same price as many of my neighbors. In a given neighborhood on a given day, we might all pay the same price. I am also very likely to see the price before buying, and I may know of several nearby stores that sell the same milk. 

One and a Half Cheers for Guidance

As a general matter, an enforcement agency like the FTC should consider how novel technologies and business practices pertinent to their jurisdiction might or might not violate the law. In the specific case of the FTC, its jurisdiction is broad, and market studies, economic, and policy research are part of the agency’s statutory mission, and have been since the FTC was first established. FTC inquiries into the effects of pricing practices and other tech developments are unsurprising, and potentially beneficial. Following developments in data-driven pricing tools, including various forms of personalized pricing that rely on consumer data—which may or may not be “consumers’ data”—and new computational tools, including those driven by artificial intelligence (AI), is, in that sense, right up the FTC’s alley.  

For the FTC, “the law” means the FTC Act, including Section 5’s prohibitions against unfair or deceptive acts or practices (UDAP), and other statutes the agency enforces, such as the Children’s Online Privacy Protection Act (COPPA). Publication of clear guidance on the agency’s current understanding of its enforcement mission, and how that mission applies to new market developments, is also potentially beneficial, provided that assessment is well-developed. Such guidance does not make law, but it can help businesses, consumers, and the agency itself understand how existing law applies. 

So, yay…ish. 

Price Discrimination: Results May Vary  

But the enthusiasm should remain qualified. Key parts of the policy statement seem underdeveloped, conclusory, or both. Some just seem wrong. 

Economists have studied price discrimination—also called “price differentiation” or “differential pricing”—at least since the landmark work of A.C. Pigou and Frank Ramsey in the 1920s. More recent contributors include Jean Tirole, Richard Schmalensee, Hal Varian, and Varian and Alessandro Acquisti, among others. 

Price discrimination can be anticompetitive under certain circumstances. But that does not mean it must be, or even that it tends to be. Nope. In the abstract, its effects on consumer welfare, total welfare, and competition are theoretically ambiguous. One of the FTC’s own citations—a paper by Jean-Pierre Dubé and Sanjog Misra—expressly recognizes the point. Tracing the observation to Pigou, they write: “The welfare effect of price discrimination is known to be ambiguous and thus a matter of empirical measurement.” 

That ambiguity cuts both ways. Price discrimination can increase welfare and often lowers prices for price-sensitive or budget-constrained consumers, who commonly have lower incomes. For a recent and accessible discussion of when price discrimination might increase the gains from trade—the benefits buyers and sellers receive from a transaction—and when it might not, see this piece by my ICLE colleague Brian Albrecht. 

What does the policy statement say about the different forms of price discrimination? In effect, it observes that perfect first-degree price discrimination by a monopolist transfers all consumer surplus—the difference between what consumers pay and the maximum they would have been willing to pay—to the monopolist. Fine, so far as it goes. But it does not follow that every consumer becomes worse off. 

Perhaps more pertinent, the policy statement recognizes the limitation quoted above: 

But real-world personalized pricing would be less than perfect in its ability to identify consumers and determine their willingness to pay and, in the vast majority of circumstances, would have to contend with competitive pressures that limit pricing power.

That limitation does not merely reflect current technology—a moving target. It also reflects practical, technical, and perhaps theoretical constraints that are likely to persist. Personalized pricing will typically resemble third-degree price discrimination, even if the FTC does not call it by that name. 

So far, so good—even if the policy statement could say it more clearly. 

When the Footnotes Won’t Cooperate

The FTC’s hazy and speculative account of personalized pricing’s likely effects comes only after several reasonable concessions. The policy statement recognizes tradeoffs among consumers, ambiguous theoretical implications, and limited empirical research on a varied and developing set of pricing practices. It then concludes: 

The limited economic research on the question suggests that while personalized pricing is likely to increase business profits, benefits to some consumers are accompanied by losses to other consumers and that the more sophisticated personalized pricing practices become, the less likely consumers are to benefit. (internal citations omitted) 

Does it? What, exactly, is the FTC saying, and why does it think the claim is true? Let’s start with the citations, which I omitted from the quotation above. 

The relevant footnote (No. 13) begins with an article by Ginger Zhe Jin, Liad Wagman, and Mengyi Zhong. I should disclose that I know Jin and Wagman from our days at the FTC and may be biased, favorably, toward their work. Jin directed the FTC’s Bureau of Economics, and Wagman was my colleague in the Office of Policy Planning, as well as a more recent co-author. 

Their article examines issues raised by various personalized-pricing practices and reviews “the academic literature on the potential benefits and harms that may arise” from differential pricing “as compared to uniform pricing.” It also highlights “potential unintended consequences of government intervention,” including cautionary lessons from the European Union’s General Data Protection Regulation (GDPR). It is a useful and concise discussion, and I recommend it. But its central recommendation is that “regulatory frameworks should focus on market structure and competitive dynamics rather than treating all forms of personalized pricing equally.” Hmm. 

Do the authors review potential problems that personalized pricing may pose for consumers or competition? Yes. They also review potential benefits. 

The next citation is to Andrew Rhodes and Jidong Zhou in the American Economic Review. Once again, the findings are contingent. Under certain conditions, personalized pricing harms firms and benefits consumers. Under other conditions, those effects reverse. The footnote also cites an EU-focused working paper by Rosa-Branca Esteves and Francisco Carballo-Cruz. 

An empirical paper by Nicholas Buchholz, Laura Doval, Jakub Kastl, Filip Matejka, and Tobias Salz, “Personalized Pricing and the Value of Time: Evidence from Auctioned Cab Rides,” may be especially relevant. Using data from a European ride-sharing platform, the authors examine conditions under which personalized pricing may have positive or negative effects on consumers, drivers, the platform, and total welfare, as well as the tradeoffs among them. 

The footnote also cites the previously discussed paper by Dubé and Misra, “Personalized Pricing and Consumer Welfare.” That experimental paper begins from the observation that “[t]he welfare effect of price discrimination is known to be ambiguous and thus a matter of empirical measurement.” 

In short, every cited article recognizes potential harms and benefits to consumers, firms, competition, total social welfare, or some combination of them, depending on the circumstances. These are perfectly good sources. But neither any single paper nor the group represents a definitive conclusion drawn from settled literature, much less a uniform thumbs-up or thumbs-down on personalized pricing. And none appears to support the general pessimistic trend for which the policy statement cites them—that increasing sophistication generates increased harm is not a reported result.

When the cited authors draw policy implications, they tend to counsel caution. Jin et al. recommend attention to market structure and competitive dynamics. Buchholz et al. emphasize that their “results highlight the nuanced welfare effects of incorporating detailed consumer information into pricing in two-sided markets,” such as platforms that connect riders and drivers. Dubé and Misra write that their “findings indicate a need for caution in the current public policy debate regarding data privacy and personalized pricing insofar as some data restrictions may not per se improve consumer welfare.” 

Under their model and data, Buchholz et al. identify conditions in which personalized pricing based on consumer data lets a particular ride-sharing platform exploit its market power. It lowers consumer surplus and raises the platform’s surplus relative to a baseline in which an auction determines the available rides, prices, and wait times, while the platform charges a fixed 10% fee. That baseline is still a form of personalized pricing, but one that does not rely on consumer data drawn from riders’ histories on the platform. 

That finding is interesting and potentially useful. But it hardly establishes a general result about personalized pricing, even within ride-sharing, much less a clear consumer-protection rule. 

The results also illustrate tradeoffs found throughout the price-discrimination literature. For example, the positive consumer-welfare effects of third-degree price discrimination depend on increased output, an observation dating to the landmark papers by Schmalensee and Varian linked above. Buchholz et al. find:  

Relative to uniform pricing, personalized pricing has a small but negative effect on average consumer welfare. This aggregate loss in consumer surplus, however, masks interesting distributional effects across consumers. Indeed, most consumers (62.5%) benefit from personalized pricing, but these gains are offset by the platform’s ability to increase prices for the most inelastic consumers. Relative to uniform pricing, average prices fall slightly under personalized pricing and the market expands by up to 7.6%.  

Under those conditions, personalized pricing has a small negative effect on average consumer welfare. Yet it also increases output, reduces average prices, and benefits most consumers. Those beneficiaries include the most price-sensitive consumers, who may be least able to afford a price increase and most willing to accept a longer wait in return for a lower price. Is that really a bad thing? Does it call for a Section 5 case? 

Am I reading too much into one phrase and its accompanying footnote? Perhaps. The FTC includes some of the qualifications I have noted. It has not adopted—or even proposed—a Magnuson-Moss regulation, which would require the agency to follow a formal rulemaking process. Nor has it brought an actual case. The FTC’s Bureau of Consumer Protection has also been relatively well run under the current leadership, which gives me some reason for optimism about case selection. 

Still, that phrase is not just at odds with the cited literature and independently dubious. It provides central support for the FTC’s argument that personalized pricing without disclosure should carry a presumption of illegality, at least in markets (or industries, sectors, or . . . ?) where such pricing is not a long-established and well-known commercial norm. Regulatory and enforcement uncertainty can chill procompetitive conduct, including innovation. 

Nor is the problem confined to that phrase and footnote. The Commission also says, “[a]ny benefits to consumers or competition from personalized pricing may also be realized without concealing the fact and nature of the personalized pricing.” 

Is that true? Not obviously, at least as I understand the literature. The footnote supporting the claim cites no empirical findings, theoretical results, or even stated policy preference. It says only that the Commission takes no position “on whether some personalized pricing practices are unfair even when fully disclosed.” Not yet, that is. So the FTC neither provides a safe harbor for full disclosure nor rules one out. Such restraint. 

The claim seems not merely unsupported but wrong. It assumes that disclosure offers a costless alternative. Yet, as Brian Albrecht recently noted on X, disclosure itself can change consumer behavior and, potentially, reduce some of the benefits associated with personalized pricing. 

Sunlight Isn’t Free 

“More light!” So cried Goethe on his deathbed—if only in German, and only according to third-party reports. Sunlight, Justice Louis Brandeis famously noted, is the best disinfectant. And I get the point, even if I question taking medical advice from Louisville’s most famous lawyer. Lowering the cost of obtaining useful information often helps consumers and competition. 

But what about mandatory disclosures? Or threats of a lawsuit by a federal law-enforcement agency when a business fails to provide them, assuming that is meaningfully different? 

Mandatory disclosures sometimes make sense, and the law sometimes requires them. But not always. Courts, for example, also worry that disclosure mandates may compel speech in violation of the First Amendment. 

And courts, legislators, and other policymakers worry about costs and complications: Information is not free—not even for the federal government or large retailers. Gathering and disseminating accurate, useful, material information imposes costs, and those costs vary considerably across contexts. So do the effects of disseminating information, which is not necessarily the same as communicating it effectively. That distinction has long complicated privacy policy and other technical fields, as Liad Wagman and I, among others, have discussed. For a broader treatment, see my former FTC colleague Janis Pappalardo’s “Product Literacy and the Economics of Consumer Protection Policy.” 

Whether the Commission should announce any presumption of illegality based on theoretical ambiguity and “limited [empirical] economic research” strikes me as a very good question. You can probably guess my answer. Evidence that some consumers benefit while others lose should prompt further inquiry, not a broad legal presumption. 

If personalized pricing—or any other pricing mechanism or input into production—leaves most consumers better off while lowering average prices and expanding access to a product or service, is that a bad thing? Must a firm disclose the mechanism or risk violating Section 5? What exactly must it disclose, and at what cost? 

Clearly, Conspicuously, and Impossibly

What, exactly, is the FTC proposing amid changing pricing tools, an ambiguous theoretical literature, and an incomplete and ambiguous empirical literature? 

Start with the first condition that would trigger FTC intervention: “Where consumers reasonably expect that prices for a product or service will not vary based on their personal data,” but a business uses personalized pricing. That raises small—but perhaps critical—questions. How many consumers must hold that expectation? How consistently? Must a clear majority share it? Suppose the FTC means the modal or median consumer—or in plain if less precise English, a typical consumer. What follows from that expectation? 

Surely Section 5 does not require firms to disclose every factor affecting their pricing that consumers do not expect or understand. The FTC might mean that, but I should hope not. It would be nuts. Such disclosures could prove costly, ineffective, or both. And it could prove harmful. Under some circumstances, they could even facilitate anticompetitive pricing practices, including price fixing. 

Section 5(n) of the FTC Act also limits the Commission’s authority over unfair acts or practices.  The FTC cannot declare an act or practice 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.

As a general matter, does a departure from consumer expectations—without a thorough disclosure—cause substantial consumer injury? Is that injury one consumers cannot reasonably avoid? Do countervailing benefits to consumers or competition offset it? The policy statement says that failing to make the specified “disclosures is likely to constitute an unfair or deceptive act or practice in violation of Section 5.” If the theory is unfairness, the FTC must explain when the omitted disclosures produce substantial injury and whether countervailing benefits outweigh it. 

If the theory is deception, is the omission material to consumers? In other words, would disclosure likely change their purchasing decisions? Would it do so generally, typically, or only for some consumers or under particular conditions that the FTC could spell out but hasn’t? 

Suppose I am standing at the grocery-store register with a carton of milk. I know the price before I pay. I also know that milk prices vary from week to week and store to store. Like many consumers, I can comparison shop; indeed, I can do it on my phone while I’m standing in line. Some consumers may dislike personalized pricing, and some may change their behavior if the store discloses it. More might do so under additional circumstances. But how do we get from those observations—preferences and consumer behavior vary—to a presumption of illegality? 

As the FTC acknowledges, “personalized pricing is a long-established norm in some markets.” Any market in which buyers and sellers negotiate prices involves price discrimination and, at some level, personalized pricing. Think of negotiations with car dealers or, to borrow from The Beverly Hillbillies, contractors for swimming pools and agents for movie stars.

Senior and student discounts are forms of third-degree price discrimination, which assigns customers to groups and charges those groups different prices. Movie theaters, coffee shops, ski resorts, and grocery stores all use such discounts. Many retailers also offer loyalty discounts, commonly considered second-degree price discrimination because consumers select an offer through their behavior. Some use both forms. Harris Teeter, a grocery chain with a store near my home, offers seniors a 5% discount every Thursday, in addition to loyalty discounts. 

Is there something special about data—even personally identifiable information (PII), as defined by the National Institute of Standards and Technology (NIST), or some other category of personal information defined by the FTC—that makes nondisclosure presumptively harmful when it conflicts with some consumers’ expectations? That’s far from obvious, even if some consumers and enforcers might feel that way. The policy statement does not explain why it should be generally true, rather than true only under particular facts and circumstances. And if it’s true, is the grocer the best one to bear the burden of complete disclosure?

Both loyalty discounts and senior discounts are personalized pricing; both rely on data to some extent; and both predate recent developments in “big data” and AI-driven applications. Never mind the LLM race, several local grocers have long possessed personal information identifying me, my home address, my phone number, and—in Harris Teeter’s case, at least—my age. 

Is one of these grocers violating Section 5 by failing to disclose the details of its personalized-pricing practices? Does the answer depend on how many consumers know about (or have heard, seen, or considered) a particular pricing tool? How should a presumption of illegality operate as digital-pricing tools, their deployment, and consumer expectations all undergo rapid change? 

The required disclosure raises still more questions. How “clearly and conspicuously” must a seller disclose “not just that the price is personalized, but also the basis for that personalization and the types of data on which the personalization is based”? How much detail must the seller provide? How much is the seller likely to know? 

Am I an elitist cynic if I doubt that the cashier at the register—if there is a cashier—knows every input into the store’s prices or how the store processes those inputs? My own experience suggests that cashiers, often enough, don’t even know an item’s price until they scan it. That’s no fault of theirs. Grocery stores carry a lot of products. 

An FTC report, citing a food-industry study, says that “the average number of individual stock-keeping units (‘SKUs’) in a supermarket exceeds 31,000.” At my age, I could not memorize even half that many. I jest, partly because I was dragged into that ridiculous excuse for a “study” before leaving the FTC. But I quote it accurately. 

No problem, you say. The store can post a sign disclosing personalized pricing. But disclosure itself can change behavior and produce both benefits and costs. The FTC also does not demand a simple yes-or-no disclosure. It demands clear and conspicuous disclosure of “not just that the price is personalized, but also the basis for that personalization and the types of data on which the personalization is based.” 

All of it? At what level of detail? Must the sign describe data sources, data types, data structures, algorithms, and impenetrable AI models two dozen hidden layers deep? Does all of that go on a sign at the entrance? Would consumers appreciate that level of detail? Would they read it? 

Don’t Enforce in the Dark

My point is not that first-, second-, or third-degree price discrimination is necessarily, always, or even typically good for consumers. It isn’t always, and I won’t guess how often the benefits fall one way rather than another. Nor am I arguing that personalized pricing can never violate the FTC Act, whether the theory involves disclosures or omissions, competition or consumer protection. Net harms are certainly possible under the right facts and circumstances. To take an easy (if abstract) example, a vendor of pricing software—or a retailer using it—might lie about material facts while advertising goods or services in commerce. 

But that ain’t necessarily so, and we do not even know how likely it is. The policy statement makes claims that sometimes seem false or misleading in ways that matter. The literature it cites doesn’t support the message. At best, the proposed enforcement stance is premature and overbroad. It does not press a bad case or promulgate a bad rule, but it invites and perhaps threatens ill-founded interventions. It also fails at the central task of guidance: helping firms, courts, and enforcers distinguish lawful conduct from unlawful conduct. 

More study? Certainly. The FTC employs exceptionally capable economists and policy analysts, and studying market developments falls squarely within its statutory remit. But before demanding more light from others, the FTC should find the switch.

The post Fine Print for Every Price: The FTC’s One-Size-Fits-All Guidance appeared first on Truth on the Market.

Ocean City was hardly an outlier. Over the past month, my dentist, mechanic, and a tree service have all charged me extra for using a credit card—and those are just the surcharges I noticed or someone disclosed.

Meanwhile, merchants have taken their fight against card-processing fees to state legislatures. In Illinois, they secured a law exempting state and local taxes and gratuities from interchange fees—the portion of a card-processing charge that goes to the bank that issued the card. The idea has since spread unevenly to several other states.

Colorado’s version was especially convoluted. It carved out smaller Colorado banks and imposed price controls on fees for charitable donations. The governor recently vetoed it. Meanwhile, Illinois has delayed its law’s effective date, and the Office of the Comptroller of the Currency has announced plans to preempt it.

Retailers have also turned to the courts. In North Dakota and Kentucky, they have filed federal lawsuits challenging the Federal Reserve Board’s formula for setting the maximum interchange fees that large banks may charge on debit-card transactions. Congress required those limits through the so-called Durbin Amendment to the 2010 Dodd-Frank financial-reform law.

Federal Reserve rules adopted in 2011 cut the permitted rates roughly in half. The result was higher bank fees, a sharp decline in free checking, and the disappearance of debit-card rewards. Retailers nonetheless argue that the remaining fees are still too generous because banks may recover costs associated with fraud prevention, fraud losses, dispute resolution, and other consumer protections. Cutting the fees further would not make those costs disappear. It would merely shift more of them to consumers.

Retailers have also spent millions lobbying Congress to extend some of the Durbin Amendment’s worst features to credit cards. That effort would raise costs, restrict access to credit, and weaken payment security—all in the name of lowering a fee that consumers rarely see but ultimately help pay.

Time to Swipe Back

The merchants’ war on debit and credit cards is especially ironic because ubiquitous access to electronic payments helped save the U.S. economy—and millions of small retailers and restaurants—during the COVID-19 pandemic. Many businesses had stopped accepting checks long before then. Cash posed health concerns and could not support the online economy that emerged during the pandemic. Now that retailers benefit from consumers’ embrace of cashless payments, some have turned that convenience into a consumer headache and a new profit center.

Enough special-interest pleading at consumers’ expense. Each of these problems has a direct solution.

First, state attorneys general should crack down on retailers’ deceptive and excessive surcharges. Businesses should disclose before purchase whether they impose a surcharge and how much it will be. Instead, consumers often learn about the fee only after the transaction—or after checking the receipt at home.

Merchants insist that surcharges merely cover their costs. The numbers suggest otherwise. The standard 3% surcharge substantially exceeds the average credit-card interchange fee, and I now encounter surcharges of 3.5% and even 4%. That pattern should surprise no one. In every country that has allowed merchants to surcharge credit-card transactions, merchants have charged more than their actual acceptance costs.

Regulators in other countries have responded by requiring clearer disclosure. Merchants should not advertise one price on a shelf or menu, then charge another at checkout. If a fee exceeds the merchant’s actual cost, the merchant should have to say so plainly rather than hide it behind euphemisms such as “service fee.” Regulators should also bar merchants from surcharging debit cards whose fees the Durbin Amendment already limits.

The Federal Trade Commission (FTC) has recently targeted hidden charges for tickets, hotel resorts, and other purchases. Hidden and excessive card surcharges belong on that list.

Second, the comptroller should follow through on plans to preempt Illinois’ law and the copycat measures that follow it. Colorado’s variation shows what happens once the political process takes over. A special-interest bidding war begins, and the carveouts and complications multiply.

If politicians can ban interchange fees on taxes and tips or impose price controls on transactions with charities, why stop there? Groceries, car repairs, prescription drugs, and any other politically favored purchase could be next.

Third, courts considering challenges to the Durbin Amendment’s implementation should recognize the payment-card system’s complexity and respect the Federal Reserve’s judgment. They should not let merchants’ litigation campaign gut the fraud prevention, dispute resolution, and other protections that consumers rely on whenever they use a debit card.

The Price Isn’t Right Until the Receipt

Merchants helped make electronic payments the default, and they know consumers now depend on them. Many no longer accept checks. Businesses in tourist areas also know that customers do not want to carry a pocketful of cash. Too many have exploited those habits to impose stealth price increases at checkout.

The modern payment-card system delivers extraordinary convenience and enormous benefits to consumers, merchants, and the broader economy. Those benefits depend on smooth, nearly universal acceptance—and on prices consumers can trust. Regulators should ensure that the system works for everyone, not just the merchants gaming it.

The swipe should complete the sale, not spring a trap.

The post A Fee Too Far: Merchants, Surcharges, and the War on Plastic appeared first on Truth on the Market.

That dispute may sound narrow. It is not. The 4th Circuit appeal presents a recurring antitrust problem in unusually clean form. How should Section 2 of the Sherman Act, which prohibits monopolization, treat conduct whose competitive significance becomes clear only in hindsight? The answer will shape not only patent acquisitions, but also the broader legal environment for investment, corporate transactions, and innovation by firms that already possess substantial market power.

The temptation is to make the case about intent. CareFirst and several amici argue that the district court’s reconsideration opinion invented a specific-intent requirement for completed monopolization. As the CareFirst opening brief and the Federal Trade Commission’s amicus brief emphasize, a monopolization claim generally does not require proof that corporate executives subjectively wanted to exclude a rival. That proposition, standing alone, should not be controversial.

It also does not answer the harder question. The Supreme Court’s 1966 decision in United States v. Grinnell Corp. requires the willful acquisition or maintenance of monopoly power, a standard that must retain objective content. When the challenged conduct is an acquisition, courts should evaluate it as an acquisition based on the circumstances at the time. A company’s later use of an acquired asset may reveal what the asset could do when the deal closed. It should not replace proof that the acquisition itself was exclusionary when it was made.

That distinction makes economic sense. It gives firms an ex ante rule they can actually follow, preserves a meaningful boundary between Section 2 and the Clayton Act’s merger rules, and reduces the risk that courts will punish efficient transactions because an asset acquired for one purpose later proves useful for another. Most importantly, it keeps monopolization law focused on protecting the dynamic competition and innovation that Section 2 should preserve, not suppress.

The Deal Looks Different in the Rearview Mirror

The record in CareFirst brings that problem into focus. Johnson & Johnson acquired Momenta Pharmaceuticals in 2020 in a deal that included more than 500 patents. According to the district court, J&J’s board materials attributed 95% of the transaction’s value to nipocalimab, a clinical-stage drug unrelated to Stelara, also known as ustekinumab. Four Momenta patents covered cell-culture media, the nutrient mixtures used to grow cells that produce biologic drugs. Nothing in the deal valuation, as described in the briefs, assigned value to Stelara, ustekinumab, or those four patents.

More than two years later, Amgen announced plans to launch an ustekinumab biosimilar. J&J then added the four Momenta patents to infringement litigation. CareFirst alleges that the resulting settlements delayed biosimilar entry and raised its drug costs. Those alleged effects matter. But CareFirst does not challenge the settlements themselves as unlawful. Its Section 2 theory instead identifies the 2020 acquisition as the exclusionary act.

That choice of timing matters. Under Grinnell, Section 2 distinguishes the willful acquisition or maintenance of monopoly power from growth resulting from a superior product, business acumen, or historic accident. Aspen Skiing Co. v. Aspen Highlands Skiing Corp. likewise asks whether the challenged conduct qualifies as exclusionary rather than competition on the merits. Both decisions call for an objective assessment of the conduct. Evidence of subjective purpose may help explain what a company did, but purpose alone does not establish a violation.

That approach also accords with the D.C. Circuit’s influential framework in United States v. Microsoft Corp.. Microsoft focuses on competitive effects and procompetitive justifications, not on psychoanalyzing corporate managers. Even an effects-based framework, though, requires a court to identify the act under review. Causation cannot do all the work. If later market effects can automatically transform an earlier acquisition into exclusionary conduct, Section 2’s conduct requirement shrinks to deliberate ownership plus hindsight.

J&J undoubtedly intended to buy Momenta. That does not mean it intended to exclude a competitor in the antitrust sense. Every buyer intends to acquire the assets listed in the merger agreement. The relevant question is whether those assets, viewed objectively when the deal closed, could reasonably contribute to monopoly power in the market at issue—either as an actual or potential competitive threat or as a tool for suppressing one. That is not a test of corporate state of mind. It is a test of conduct.

When Antitrust Mistakes Winning for Cheating

That boundary around exclusionary conduct is not mere formalism. It responds to one of antitrust law’s central difficulties. Courts must distinguish conduct that harms competition from conduct that harms rivals because it is efficient, innovative, or simply aggressive.

Judge Frank Easterbrook framed the problem in his classic 1984 article, “The Limits of Antitrust.” He urged courts to consider “error costs,” the harm caused by getting a case wrong. Those costs include false negatives, in which anticompetitive conduct escapes condemnation, and false positives, in which courts condemn or deter procompetitive conduct. They also include the cost of administering the system itself. Easterbrook argued that false positives often deserve special concern because a mistaken legal rule can suppress beneficial conduct across an entire market, while competition may eventually erode some harms left by a false negative.

The Supreme Court has built that concern into Section 2 doctrine. In Verizon Communications Inc. v. Law Offices of Curtis V. Trinko LLP, the Court warned that the risk of false positives counsels against expanding monopolization liability too far. Dominant firms receive no special indulgence under this principle. Their conduct is simply hard to classify. Price cuts, product improvements, acquisitions, exclusive investments, redesigns, and vertical integration—combining with suppliers or distributors—can all hurt rivals. Often, that is precisely how competition helps consumers.

Easterbrook made a related point in A.A. Poultry Farms Inc. v. Rose Acre Farms Inc.. Evidence of intent does little to distinguish hard competition from attempted monopolization and may invite juries to punish aggressive rivalry. That warning carries particular force in innovation-intensive markets, where vigorous competition routinely destroys the value of rivals’ assets and acquisitions often evolve in unexpected ways. A firm that acquires complementary technology may redirect resources, abandon duplicative projects, combine research teams, or discover applications that neither party anticipated when signing the deal. With enough hindsight, courts can recast any of these developments as exclusionary, even when the transaction promoted competition at the time.

The International Center for Law & Economics (ICLE) has repeatedly emphasized this institutional problem. Its work on “Innovation and the Limits of Antitrust” and its more recent comments on single-firm conduct explain that error-cost analysis is not an excuse for nonenforcement. It is a way to design workable legal rules under uncertainty. If firms cannot tell whether courts will later condemn ordinary competitive conduct, the deterrent effect extends far beyond the defendant in one case. It changes how every similarly situated firm behaves before acting.

That risk matters especially in monopolization law because dominant firms are still supposed to compete. Section 2 protects the competitive process. It does not require firms to preserve rivals, freeze their business models, or avoid innovations that make existing products less valuable. As the boundary between beneficial and harmful conduct becomes harder to discern, courts have greater reason to apply clear legal tests based on facts firms could observe when they made their decisions.

Due Diligence Without a Time Machine

CareFirst’s theory also threatens to erase a basic statutory distinction. Section 7 of the Clayton Act looks forward. It asks whether an acquisition may substantially lessen competition or tend to create a monopoly. Section 2 of the Sherman Act asks a different question. It condemns the willful acquisition or maintenance of monopoly power, a standard that has long required more than market power followed by a bad outcome.

If deliberately acquiring assets satisfies Section 2 whenever later events reveal an exclusionary use, every acquisition by a firm with monopoly power remains legally unsettled after closing. The decisive facts might not emerge for years. They could arise through product development, patent litigation, technological change, or a rival’s business decisions. General counsel could not assess a deal using the information available when the parties signed it because the assets’ future uses would determine whether the acquisition had been exclusionary all along. Due diligence does not come with a time machine.

That rule would create particular problems for diversified firms built around research and development. Large transactions often transfer thousands of patents, contracts, data sets, employees, partly developed technologies, and specialized expertise. Some assets will prove more valuable than expected. Others will lead nowhere. That uncertainty is not a defect in innovation markets. It is one of their defining features.

Treating unforeseen usefulness as proof of an exclusionary acquisition would impose an “option tax” on mergers—a penalty on discovering valuable new uses for acquired assets. The more successfully an acquirer develops those assets, the greater its potential antitrust exposure. That incentive points in precisely the wrong direction. Firms might conduct wasteful diligence into every imaginable future use, surrender assets with no demonstrated competitive significance to secure a deal, or abandon transactions whose value depends on experimentation and combining technologies in new ways.

Limiting that risk to monopolists does not solve the problem. Firms with scale, related technologies, established distribution networks, regulatory expertise, and the resources to commercialize inventions often drive important innovations. Section 2 should encourage those firms to compete vigorously while condemning genuinely exclusionary conduct. A lawful position of market strength should not make every ordinary acquisition presumptively suspect.

Merger Review Is Not a Lifetime Warranty

Recent merger scholarship reinforces this point. Daniel Spulber and I argue in our 2024 article, “Antitrust Merger Policy and Innovation Competition,” that merger analysis should not assume consolidation harms innovation. Horizontal mergers between competitors can increase the resources and incentives available for research. Vertical mergers between firms at different levels of the supply chain can help bring inventions to market. Acquisitions of new entrants can also encourage entrepreneurship by giving founders a path to sell their firms and pairing inventive assets with the capabilities needed to develop them.

None of this means every acquisition promotes innovation. It means the effects depend on evidence and context. A sound legal rule should not give great weight to speculative future harms while discounting or ignoring potential benefits merely because they are also uncertain.

Louis Kaplow makes a complementary point about efficiencies in his 2026 Antitrust Law Journal article, “Out of Market, Out of Mind.” Kaplow criticizes approaches that disregard real benefits simply because they occur outside the market where a harm appears. His broader insight applies here. A dynamic economy constantly shifts resources among uses, firms, products, and markets. Antitrust rules that isolate one pocket of harm while ignoring broader benefits can prevent resources from moving toward more productive uses.

David Teece likewise argues in his 2025 Antitrust Law Journal article, “Understanding Dynamic Competition,” that antitrust analysis should pay greater attention to innovation, potential competition, and firms’ practical capabilities. Static measures such as current market shares and overlapping products reveal only part of the picture.

Teece and Magdalena Kuyterink develop that idea in “Recognizing What’s Around the Corner.” Their approach asks what firms can realistically build, adapt, and bring to market. Teece and Gönenç Gürkaynak identify a related imbalance in “Integrating Innovation Concepts Into the Merger Control Context.” Competition authorities may credit speculative claims that a merger will harm innovation while demanding much stronger evidence that the same merger will promote it.

Together, this scholarship counsels institutional caution. An acquisition may combine complementary research, manufacturing, regulatory, and commercialization capabilities in ways that no static snapshot can capture. Patents often serve as transactional assets that allow one firm to specialize in invention while another develops and markets the resulting technology. ICLE-affiliated scholarship on “Intellectual Property and Transactional Choice” explains how those arrangements can promote specialization and commercialization.

If an unforeseen later use of an acquired patent can trigger Section 2 liability, firms may avoid the very transactions that assemble scattered technologies and capabilities into useful products. That is why CareFirst carries real implications for merger policy even though it is formally a monopolization case. Section 7 already provides a forward-looking system for reviewing acquisitions that may threaten competition. Turning Section 2 into retrospective merger review would add uncertainty without improving ex ante screening. Merger law should not get a do-over whenever an acquired asset later proves useful.

Set the Antitrust Clock at Closing

The 4th Circuit need not choose between probing executives’ minds and allowing later effects to define earlier conduct without limit. An objective, time-bound test offers a better path. What could the acquired asset reasonably do when the transaction closed?

A plaintiff would not need a confession in an email. A court could examine the transaction documents, the parties’ valuation of the assets, the patents’ stated scope, the target’s products and technologies, and the information publicly available at the time. It could then ask whether the acquired assets could realistically block or weaken competition in the relevant market. If a monopolist buys a young firm poised to become a competitor or a patent that blocks a known path to market entry, the buyer cannot escape liability by claiming ignorance. The asset’s competitive significance does not depend on what executives say they knew.

That discipline must also constrain plaintiffs. General usefulness is not enough. A patent that applies broadly to biologic manufacturing does not necessarily threaten competition in the ustekinumab market when acquired. Later enforcement may show that the patent could exclude a rival when asserted. It may also help establish what the buyer could reasonably have foreseen at closing. By itself, though, it cannot prove that the earlier acquisition was exclusionary.

This standard would leave ample room for other antitrust and patent claims. A plaintiff could still challenge patent enforcement as a sham, allege that the patent was obtained through fraud, or identify later conduct that independently violates Section 2. Each theory has its own requirements and limits. Focusing the acquisition inquiry on the facts at closing would not shield later misconduct. It would simply require plaintiffs to prove that the acquisition itself was exclusionary before using it as the basis for liability.

That approach sensibly allocates the risk of error. It preserves enforcement when objective evidence connects an acquisition to the acquisition or maintenance of monopoly power. At the same time, it reduces the danger that judges and juries will turn an ex post narrative into a legal rule that firms could not have followed ex ante.

Hindsight Is Not a Cause of Action

The 4th Circuit should reject the false choice between requiring proof of subjective intent and allowing later effects to define earlier conduct. The central question is whether Section 2 retains a meaningful boundary between exclusion and ordinary commerce, one that tells firms what the law requires when they act.

Both kinds of error carry costs. A monopolization rule that is too permissive may tolerate durable exclusion and harm consumers. An overbroad rule may deter price cuts, product improvements, integration, acquisitions, and risky innovation because courts could reinterpret those acts years later. Such false positives do not end with one defendant. They reshape investment incentives across the economy.

CareFirst gives the court an opportunity to reaffirm a modest but important principle. Courts should ask whether the challenged conduct was objectively exclusionary based on the circumstances when it occurred. That approach would preserve challenges to acquisitions that remove actual or emerging competitive threats, as well as claims against later conduct that independently violates the law. It would also keep Section 2 from becoming a retrospective penalty on asset ownership.

That boundary matters in an economy that depends on experimentation, new combinations of technology, and the ability to bring ideas to scale. Dominant firms must remain subject to antitrust law. They must also remain free—and encouraged—to compete, invest, acquire complementary assets, and innovate on the merits. Courts should condemn exclusion when the evidence identifies it, not merely when an acquired asset later proves useful.

Section 2 should condemn exclusion when it happens, not backdate it to closing.

The post Four Patents and a Time Machine: CareFirst and the Perils of Backdated Antitrust appeared first on Truth on the Market.

“Deliberation as Self-Discovery and Institutions for Political Speech,” by Catherine Hafer and Dimitri Landa, is, if anything, more timely now than when it was written. The public, both in the United States and elsewhere, is deeply divided, and polarization has become a defining feature of our political moment. The article also brings together two bodies of scholarship that seldom speak to each other: deliberative democracy and formal theory, which uses mathematical models to study political behavior. 

Elementary, My Dear Deliberator

Economics has a well-developed literature on communication. Standard treatments often use signaling models, in which someone has valuable private information they may want to communicate. This approach connects to principal-agent models, which examine relationships in which one person acts on another’s behalf. One of the main reasons to empower an agent is to take advantage of that person’s expertise—and to induce the agent to develop it.

The prime legal example is an attorney (the agent) working on behalf of a client (the principal) and providing advice based on knowledge of the law. But corporate directors and shareholders, administrative agencies and the executive and legislative branches, and representatives and voters all fit this general model. Each relationship interests—and is heavily mediated by—law. 

Hafer and Landa take a different approach, one that is both distinctive and, to my mind, especially well-suited to deliberation among equals. Think of discussions between citizens or senators, rather than an expert reporting to a less-informed boss. Inspired by research into how people develop political opinions and value judgments, Hafer and Landa propose the idea of deliberation as self-discovery. 

The agents in their model are not logically omniscient—a plausible assumption. Specifically, they do not automatically recognize all the implications of what they know. The aim of deliberation, then, is not to persuade someone by presenting new information, but to trace the implications of beliefs they already hold. It is less “Actually, that happens only about 2% of the time” and more “You believe A, which entails B, so you should also believe B.” 

Crucially, Hafer and Landa’s agents may not be perfectly rational, but they are not irrational. They respond to arguments of this sort. They are flawed—human—but trying their best.

The evocative example is the good Dr. Watson. Sherlock Holmes immediately and completely grasps the implications of every scrap of knowledge. To Holmes’ repeated frustration—though Holmes is endlessly frustrating in his own ways—Watson needs someone to point out the connections. He is not the “indisputable Bayesian hero” that Holmes is, always updating his beliefs in response to new evidence. Once someone shows Watson the connections, though, everything falls into place.

Hafer and Landa’s idea reminds me of Socrates, though perhaps more in the Platonic dialogues than in the classroom. Hence the notion of self-discovery: The speaker does not impart new information. Neither the detective nor the philosopher presents new facts about the world. Rather, the listener realizes what they already, in some sense, believed. As Hafer and Landa put it:

Our basic ontology of learning is, then, that of recognizing latent reasons—reasons that agents are endowed with and would be able to embrace as “their own” after recognizing their fit with other held beliefs, but which are not actively available to them prior to deliberation either for developing the corresponding policy position or for attempting to influence others. 

The Extremists Have the Floor

On its own, this idea of deliberation as self-discovery is interesting. Hafer and Landa also demonstrate the flexibility of economic methods by, for example, relaxing Bayes’ rule—the standard formula for updating beliefs in light of new evidence—in a formal model. But they go further because they are interested in institutions. What arrangements best foster deliberation as self-discovery?

The deliberative-democracy literature offers an appealing baseline answer for how to structure discussion: equal and extensive participation. Hafer and Landa reach a surprising alternative conclusion. The institutions that promote the most informationally efficient deliberation as self-discovery do not give everyone equal speaking time. Instead, they encourage participants with relatively extreme positions to speak.

To explore these institutional implications, Hafer and Landa set up an economy of deliberation. Participants divide their deliberative resources between speaking and listening. Their goal is to identify—or at least confirm—their own ideal point, meaning the course of action each considers best. In other words, participants use deliberation to figure out what they believe. This is a pretty attractive view of deliberation. 

In Hafer and Landa’s model, the institution that produces the best information gives the floor to proponents of more extreme positions. Listeners benefit from hearing those arguments because they want to determine whether they believe them. Inducing relatively extreme participants to devote their resources to speaking rather than listening therefore improves the informational quality of deliberation. Moderates learn more by listening to them than by advancing their own positions. 

These results depend on other conditions, including the reasonable assumption that agents cannot make arguments they do not already believe. If an argument is “latent”—if it follows logically from an agent’s beliefs, but the agent has not yet discovered it—the agent cannot advocate it. That follows directly from the structure of deliberation as self-discovery. Another condition is “connectedness,” roughly a requirement that reasons and policy preferences move together in a consistent way.

There is also a natural limit to this “extremism premium” (my term, not theirs). The inegalitarian result depends on the possibility that listeners might share the extreme position. As Hafer and Landa explain, “Empowering the speech of the more radical extremist is informationally optimal only because it assists a more moderate agent in resolving her uncertainty about the best policy.” The moderate wants to determine whether she, too, is actually an extremist. But the more extreme the position, the less likely she is to believe it. Hafer and Landa’s optimal institution therefore does not “give any outlier the floor.” It gives outliers whose views others might plausibly share a chance to speak.

Hafer and Landa acknowledge that deliberation may serve values beyond informational accuracy. We might value fairness or active participation, and both may be essential to political legitimacy. I would add that we do not always know in advance which positions are relatively extreme or who holds them. In practice, we may not even know the full range of possible arguments. Some initial egalitarian deliberation may therefore prove useful. 

The Jury Is Still Thinking

“Deliberation as Self-Discovery” shows the promise of this kind of work. It uses economic methods to reach a counterintuitive conclusion and illuminate something that would otherwise be difficult to see. That combination of methodological rigor and unexpected insight captures what law & economics can offer beyond its traditional subjects.

The article also has compelling practical applications. Hafer and Landa mention allocating time on committees and regulating campaign finance, both of which distribute scarce deliberative resources. I would add juries, an institution central to law. Juries closely resemble this model of deliberation because, by design, no juror has privileged access to the evidence. All jurors receive the same information and should decide the case solely on the record developed during the proceedings.

Juries have assumed even greater importance since the Supreme Court’s decision in SEC v. Jarkesy, which held that defendants have a right to a jury trial when the Securities and Exchange Commission seeks civil penalties for securities fraud. A wave of proposals has also revived an older idea from deliberative theory: “citizens’ assemblies” that would extend jury-like institutions and deliberation to issues beyond the courtroom. These applications suggest that deliberation as self-discovery is more than an elegant model. It may help us decide who should speak, who should listen, and how institutions can help people discover what they believe. 

Further Reading

• David Austen-Smith, “Information Transmission in Debate,” American Journal of Political Science, Vol. 34, No. 1 (February 1990).
• David Austen-Smith and Timothy J. Feddersen, “Deliberation, Preference Uncertainty, and Voting Rules,” American Political Science Review, Vol. 100, No. 2 (May 2006).
• Eric S. Dickson, Catherine Hafer, and Dimitri Landa, “Cognition and Strategy: A Deliberation Experiment,” The Journal of Politics, Vol. 70, No. 4 (October 2008).
• James Johnson, “Formal Models” in “Research Methods in Deliberative Democracy,” Selen A. Ercan (ed.) et al., Oxford University Press (October 2022).
• Arthur Lupia, “Deliberation Disconnected: What it Takes to Improve Civic Competence,” Law and Contemporary Problems, Vol. 65 (Summer 2002).
• Amy Gutmann and Dennis F. Thompson, “Democracy and Disagreement,” Belknap Press of Harvard University Press (1996).

The post ‘Deliberation as Self-Discovery and Institutions for Political Speech,’ by Catherine Hafer and Dimitri Landa appeared first on Truth on the Market.

Aprix’s webinars struck the same theme. They urged clients not to chase sales volume through discounts: “Lowering price almost never pays off. Resist!” A gas station that “attacks the market,” Aprix warned, would invite retaliation until “everyone loses together.”

The brochure and a handful of promotional videos helped prompt Brazil’s leading algorithmic-pricing case. In April 2026, the Tribunal of the Administrative Council for Economic Defense (CADE) approved a settlement (known in Portuguese as a Termo de Compromisso de Cessação (TCC)) with Aprix. The case produced neither a litigated finding of collusion nor a condemnation on the merits. Still, the settlement offers a concrete guide to how CADE may handle future algorithmic-pricing cases. Aprix is not CADE’s only such case, and Brazil’s debate over whether pricing algorithms can facilitate collusion remains far from settled. 

Today’s Special: Public Prices, Private Data

The investigation began with a news article. On Jan. 20, 2021, the Brazilian outlet Jornal do Comércio reported that a gaúcha startup—one based in the state of Rio Grande do Sul—had developed a pricing system for fuel retailers. The next day, CADE received an anonymous complaint attaching the article. CADE’s General Superintendence (SG), its investigative arm, opened an inquiry. Over the next three years, CADE gathered evidence about how the software worked and asked its Department of Economic Studies (DEE) to assess the conduct. 

Aprix’s product combined a client gas station’s prices, costs, and sales volumes with competitors’ pump prices collected from public sources. It then recommended a daily price for each station. Aprix monitored 12,000 stations by October 2020 and more than 20,000 by June 2022.

The software drew on two distinct layers of information. The first was public but scattered. Aprix used an automated tool to collect daily pump prices from what the company described as open sources. It consolidated prices from thousands of stations across Brazil into a single database and applied statistical filters intended to ensure that each figure reflected the price actually charged at the pump. 

The second layer consisted of each client’s private historical data, including its sales, costs, and volumes. Aprix collected that information automatically by integrating its platform with the station’s management software. 

The algorithm analyzed both layers and recommended a daily price. The station manager retained the final say and had to approve the recommendation. 

Each type of information alone would be unlikely to create an antitrust concern. Collecting competitors’ publicly available prices is routine competitive intelligence, and every firm may analyze its own business data. Their combination made the case more difficult. A single vendor held a curated, nationwide database of pump prices while also receiving private data from numerous competing clients. The investigation therefore asked whether that arrangement simply helped each station make better independent decisions or allowed rivals’ prices to align through a common intermediary. 

Statistically Significant, Legally Inconclusive

The economic record centers on Technical Note No. 16/2024/DEE, a difference-in-differences (DiD) study. This method estimates an intervention’s effect by comparing changes over time in an affected group with changes in a similar, unaffected group. 

The DEE analyzed weekly retail-price data collected by Brazil’s National Agency of Petroleum, Natural Gas and Biofuels (ANP) between January 2015 and May 2022. It compared Aprix clients, known as the treatment group, with other stations in the same municipality, which served as the control group. Because the DEE lacked precise adoption dates, it used early 2020 as a proxy—the year Aprix reportedly expanded its client base by 500%.

The study estimated that Aprix adoption was associated with gasoline and ethanol price increases of roughly R$0.02 to R$0.03 per liter, or about US$0.0038 to US$0.0058. Those amounts equaled approximately 0.5% to 0.8% of the pre-adoption gasoline price and 0.7% to 1.1% of the pre-adoption ethanol price. For diesel, the estimated increase ranged from R$0.01 to R$0.04 per liter, or about US$0.0019 to US$0.0077, equivalent to 0.3% to 1.3% of the pre-adoption price.

The estimated increase, then, was generally around 1% or less of pre-adoption prices. It was smaller still when compared with the higher prices that prevailed between 2020 and 2022. The study found no consistent effect for compressed natural gas (GNV). Its estimates changed direction across model specifications, which the DEE attributed to a small sample riddled with gaps. 

Although the DiD estimates were statistically significant, the DEE described their magnitude as “low.” It also cautioned that, like any empirical model, the analysis “depends on a series of premises that, if altered, can significantly influence the results.” One of those premises is the parallel-trends assumption: Without Aprix, prices at client and nonclient stations would have followed similar paths. 

The DEE also rejected the inference that a statistically significant price difference settled the legal question. Higher average prices at Aprix stations “does not imply causality,” it warned, because “other factors can influence the prices charged.”

The study estimated the price effect associated with adopting Aprix’s software but like most economic evidence, it did not directly prove an antitrust violation—i.e., coordination among competitors. The DEE found a modest price increase and identified a risk of coordination. It neither established a cartel nor claimed to have done so. 

When the Algorithm Suggests—and When It Schemes

Algorithmic pricing has generated a substantial—and unsettled—antitrust literature. Its starting point is straightforward: Pricing software is not inherently suspect. 

Software that tracks public prices and adjusts prices as conditions change can expand output, clear inventory, and help smaller firms compete with sophisticated incumbents. Treating such tools as presumptively collusive would amount to a soft form of price control, chilling innovation that can benefit consumers. (See Alden Abbott, “Legal Challenges to Algorithmic Pricing” and Mario Zúñiga, “A Primer (and Some Questions) About the RealPage Antitrust Case.”) Firms generally may collect competitors’ public prices and use them to set their own. That conduct alone does not violate antitrust law. 

The analysis changes when software incorporates nonpublic, competitively sensitive information or helps competitors reach an agreement. The legal dividing line does not depend solely on the technology, a common software vendor, pooled data, or parallel prices. The central questions are whether competitors have relinquished independent control over pricing and whether the system provides a concrete means of coordination.

The International Center for Law & Economics (ICLE) recently summarized three scenarios in its comments on the U.S. Department of Justice (DOJ) and Federal Trade Commission (FTC) guidance on business collaborations: 

First, competitors may agree to use a shared pricing algorithm to coordinate prices. In that scenario, the algorithm itself is legally irrelevant; the unlawful agreement is the antitrust violation, and existing law already addresses it.

Second, competing firms may each share competitively sensitive information with a common third-party platform that generates pricing recommendations for all participants. In that circumstance, the platform can function as a coordinating mechanism because each participant receives recommendations informed by rivals’ nonpublic data, even absent direct communication among competitors. 

Third, independent algorithms may arrive at similar pricing outcomes simply because they respond to the same market conditions and economic signals. That form of conscious parallelism is not unlawful under U.S. antitrust law absent an agreement among competitors. 

As ICLE argues in its amicus brief in Cornish-Adebiyi v. Caesars, competitors’ use of the same pricing algorithm does not establish the horizontal “rim” required for a hub-and-spoke conspiracy. That theory requires a central coordinator (the hub) and an agreement connecting the competing firms (the rim). The relevant question is whether the software’s users plausibly struck an agreement with one another, rather than simply making parallel decisions to buy the same product. On this account, “automating lawful commercial activity does not make that activity unlawful,” and courts should not infer an agreement “from the mere fact that competitors happen to use the same commercially available software.” 

Herbert Hovenkamp frames the inquiry somewhat differently. He focuses on what the parties agreed to do with the algorithm’s recommendation. Section 1 of the Sherman Antitrust Act, he observes, “does not specify who must do the agreeing.” A hub-and-spoke price-fixing arrangement may therefore arise when each seller separately agrees with the software provider to charge its recommended price, even if the sellers never communicate directly. For Hovenkamp, the clearest cases involve subscribers that commit to following the recommendations or providers that otherwise control subscribers’ prices. 

In a recent paper, Hovenkamp and Thibault Schrepel draw the line at “functional control over implementation,” exercised “through delegation, defaults, monitoring, or inducements.” Data aggregation, correlated prices, and conscious parallelism may warrant scrutiny or provide supporting evidence of collusion. They do not, on their own, amount to cartel management or justify the automatic condemnation reserved for arrangements that replace independent pricing decisions. 

The two approaches agree on a basic limit. A recommendation without commitment is insufficient, as are parallel prices without a coordinating mechanism. Aprix is best assessed against that common benchmark. 

The distinction also matters as a matter of economics. A price increase implemented after a firm adopts an algorithm does not reveal what caused the increase. Recent economic modeling suggests that algorithmic pricing can produce procompetitive effects in markets where available capacity is sold and replenished over time. Sharing information about capacity use can intensify competition at each level of remaining capacity, while high utilization can reduce the potential gains from coordination. 

Higher prices at Aprix stations could therefore reflect more precise independent pricing rather than collusion. That possibility explains why the DEE’s caution about causation deserves more than fine-print status.

No Verdict, Plenty of Clues

Because the case ended in a settlement, CADE’s Tribunal never issued a final decision on the merits. The Aprix investigation therefore does not definitively establish CADE’s position on algorithmic pricing. Still, the proceedings and public record offer several clues about how the agency views its competitive risks. 

Aprix accepted several obligations for a two-year period. It agreed to pay R$70,803.40 (US$13,604.53) into a federal public fund, a payment similar to a fine. It also agreed to establish a competition-compliance program, segregate client stations’ pricing and strategy data, add confidentiality provisions to client contracts, refrain from requiring clients to adopt its recommended prices, and notify CADE whenever a group of clients reached a 20% share of a municipal market. CADE also received audit access, and Aprix provided the agency with the algorithm’s source code. 

The settlement did not require Aprix to redesign its algorithm. Instead, the obligations governed how Aprix handled client data and how stations used the software’s recommendations. That choice suggests that CADE’s principal concern lay with control and information flows, rather than the mere use of automated pricing.

Aprix sits somewhere on the easier side of the spectrum of possible algorithmic-pricing cases. Its software influenced pricing but, at least formally, left the final decision to each station. It relied on competitors’ public prices rather than their confidential sales volumes or margins, combining that public information with each client’s private business data. The apparent concern arose from the arrangement as a whole: One provider processed data from numerous competing stations while marketing its product with conspicuous warnings against cutting prices. 

International readers should also note an important feature of Brazilian law. The SG did not need to prove a traditional cartel agreement. It charged the conduct under Article 36, Paragraph 3, Item II, of Brazil’s Competition Law, which prohibits “influencing the adoption of uniform commercial conduct.” That standalone offense does not require the same proof as a horizontal cartel. It can therefore allow CADE to find an infringement without establishing an express agreement among competitors. Whether that lower threshold makes for sound competition policy is a separate question.

Same Fares, Different Flight Plan?

Aprix is unlikely to be CADE’s last algorithmic-pricing matter. A pending airline case more directly tests the difference between lawful parallel pricing and technology-assisted coordination.

In November 2023, the SG opened an administrative inquiry into the Brazilian carrier GOL Linhas Aéreas S.A and the Brazilian subsidiary of Chile’s LATAM Airlines. The inquiry concerned identical or nearly identical fares on overlapping domestic routes. In April 2026, the SG converted the inquiry into a formal administrative proceeding to examine whether pricing tools or market-data platforms could explain the pattern. The proceeding remains pending. 

GOL and LATAM deny the allegations and maintain that they set fares independently. The case must therefore determine whether the parallel fares resulted from separate responses to the same market conditions or coordination facilitated by pricing technology.

That makes the GOL-LATAM matter harder than Aprix. It has no common software vendor at its center. At least on the public record, it features parallel fares on overlapping routes without an obvious mechanism connecting the airlines’ pricing decisions.

Other competition authorities have examined similar questions without treating algorithmic pricing as an offense in itself. In 2023, the Italian Competition Authority (AGCM) opened a market investigation into airlines’ pricing algorithms on routes serving Sicily and Sardinia. The inquiry examined how revenue-management systems adjust fares over time, how those adjustments affect competition, and whether airlines personalize prices for individual customers. 

The AGCM closed the investigation in December 2025 without finding that pricing algorithms had produced collusive outcomes. Its concerns were more prosaic, though hardly trivial: whether consumers could understand and compare the prices they saw. 

The Algorithm Jury Is Still Out 

Aprix is now Brazil’s leading algorithmic-pricing case, though it arguably raised more questions than it answered. It did not establish that pricing algorithms are illegal per se. Nor did it determine whether algorithmic coordination should face the near-automatic condemnation applied to traditional cartels or an analysis of its actual competitive effects. 

The pending GOL-LATAM case presents different facts. It concerns two direct competitors rather than conduct centered on a common pricing vendor. Still, it raises the same basic question: How should CADE distinguish independent pricing decisions from coordination facilitated by technology? 

Aprix leaves CADE with several unresolved questions. What evidence is enough to establish an infringement? Does liability turn on aggregating scattered public prices, processing clients’ private data, or placing both types of information in one vendor’s hands? When does algorithmic coordination amount to an old-fashioned cartel, and when does it require a different analysis? 

As pricing software becomes more common, CADE must separate efficiency-enhancing uses from collusive ones. Competition authorities worldwide face the same task. An algorithm can sharpen competition or soften it. The hard part is telling which is which.

The post Premium, Regular, or Collusive? Brazil’s Aprix Case Tests Algorithmic Pricing appeared first on Truth on the Market.

That is the story unfolding in Maple Grove, Minnesota. The city has told a fiber provider that it cannot use public rights-of-way until it signs a cable television franchise, the local authorization traditionally required to operate a cable system. Never mind that the company provides broadband over fiber, not cable television. 

The city’s demand is almost certainly unlawful. By imposing these requirements to raise revenue, Maple Grove is undercutting federal policy and making it harder for residents to obtain high-speed broadband.

Broadband Economics: Pay Now, Maybe Earn Later

Broadband requires heavy upfront investment. Before a single customer signs up, a provider must pay for trenching, boring, conduit installation, pole make-ready work, and access to public rights-of-way. “Make-ready” work includes the changes needed to accommodate new equipment on existing utility poles. The provider finances all of this before earning a dollar of revenue. Hardware accounts for only part of the bill. Connecting each home with fiber costs about $700 to $2,700, depending on population density and terrain. 

Most broadband investment is sunk, meaning the provider cannot recover it if the project fails. Once a company buries fiber, it cannot pull up the cable and redeploy it somewhere more profitable. Nor can it recover the labor and regulatory expenses incurred during installation. The provider therefore commits an enormous fixed cost in return for future subscription revenue that is never guaranteed. As competition among cable, fiber, fixed wireless, and satellite providers has intensified, that revenue has become less certain than ever. If a project falls short of its projected take rate—the share of potential customers who subscribe—the provider has few ways to recover its investment. 

That risk leads providers to build where they have the best chance of earning a return and to hesitate elsewhere. Economies of scale favor dense, higher-income areas, where providers can spread large fixed costs across many paying subscribers. Rural and lower-income communities offer thinner margins and greater uncertainty. 

Broadband also creates benefits that providers cannot capture through subscription revenue. Better connections can improve education, health care, employment, and local commerce, yet providers do not receive payment for all those gains. The private return may therefore fall short of the broader social return, leaving worthwhile projects unbuilt. Projects closest to breaking even become especially sensitive to any added cost.

Reducing deployment barriers therefore matters. In theory, a provider would compare its material and labor costs with its expected revenue, then build wherever demand justified the investment. In practice, deployment requires permits, pole-attachment negotiations, zoning approvals, and right-of-way agreements. These expenses often rival or exceed the cost of the hardware.

Delay adds another cost. Months or years of uncertain regulatory review tie up capital, complicate crew scheduling, and increase the risk of projects already operating on thin margins. Every additional expense can turn a borderline project into one that never gets built. 

Some of these costs reflect the actual burden a provider places on public rights-of-way. Governments may legitimately recover the cost of managing streets and coordinating construction for such buildouts. Charges that exceed those costs function as a tax on deployment, deterring investment without providing any corresponding public benefit.

Congress addressed this problem in Section 253 of the Telecommunications Act of 1996, which bars state and local requirements that effectively prohibit providers from offering service. The FCC has interpreted Section 253 to limit right-of-way charges to amounts roughly tied to a government’s actual costs. Enforcement remains uneven, however, and some localities still treat rights-of-way as revenue sources rather than public resources requiring careful management. As traditional franchise revenue declines, local authorities go looking for replacements. Projects closest to breaking even tend to die first. 

If It’s Fiber, Franchise It Anyway

Policymakers have done considerable work to streamline permitting and accelerate broadband deployment. Maple Grove has found a new way to impede that effort. The Minnesota city now requires broadband providers to obtain a cable franchise before receiving right-of-way permits—even if they provide no cable service. 

Gateway is a fiber broadband provider with a Minnesota certificate authorizing it to operate as a telecommunications carrier. It sells internet access, not cable television. Maple Grove has decided that distinction does not matter. 

In March 2026, the city amended its code to require every “Broadband Service Provider” to obtain a cable communications system franchise from the Northwest Suburbs Cable Communications Commission (NSCCC). The NSCCC is a joint-powers cable authority shared by several suburbs. Under the amended code, Maple Grove will not act on a provider’s right-of-way applications until the provider completes the franchising process. When Gateway applied for permits to continue construction, the city deferred them. A company that had routinely obtained permits in four to six weeks suddenly could not obtain one at all. 

Obtaining a cable franchise is no small undertaking. Gateway would have to build throughout 100% of the city, subject only to a temporary waiver. It would also face municipal rate regulation, post a performance bond and letter of credit, provide public benefits in addition to paying a franchise fee on gross revenue, and accept stringent remedies for even immaterial violations.

Gateway would also have to comply with the FCC’s Part 76, subpart K technical standards. Those standards govern cable television systems, and Gateway’s fiber broadband network was not built to meet them. Maple Grove is conditioning access to public rights-of-way on an agreement designed for a service Gateway neither provides nor, in some respects, can physically provide. 

Why would a city do this? The NSCCC depends on cable franchise fees, and that revenue is collapsing. Its 2025 strategic plan reports that cable fees supply roughly 95% of its revenue and will decline by about 10% each year. The plan warns that the NSCCC faces “an existential financial crisis” and will “run out of money in mid-2028 unless we act.” 

The plan concludes that “new revenue sources are necessary to replace cable fees.” It identifies “new broadband fees” as the “most impactful” option. The NSCCC has lobbied the Minnesota Legislature to authorize those fees, although the plan concedes that “success is not certain.” 

Maple Grove made the ordinance’s purpose equally plain. Its Request for Council Action links the measure directly to “declining cable franchise revenue and its impact on [NSCCC] operation” and directs city staff to require broadband providers to sign franchise agreements. The document says nothing about coordinating construction, protecting pavement, or preventing conflicts among users of public rights-of-way. The ordinance addresses a budget shortfall by extending the cable-franchise regime to broadband.

This is rent extraction: using government control over a necessary resource to collect payments unrelated to the costs that a company imposes. Cord-cutting has eroded the revenue that supports the NSCCC’s community-media operations. Rather than reduce its spending or persuade the Legislature to authorize a new fee, the NSCCC and Maple Grove are using their control over right-of-way permits to turn a broadband entrant into a replacement source of revenue.

The resulting franchise fee would not compensate Maple Grove for Gateway’s use of public property. It would repurpose charges created for cable television to replenish a declining fund. Because Maple Grove controls the permits, it can halt Gateway’s entire buildout in the meantime. Gateway bears the immediate cost, but the company’s prospective customers lose service—or wait longer for it. 

Maple Grove Can’t Have It Both Ways

Gateway has petitioned the FCC to declare Maple Grove’s new code unlawful and preempted, meaning displaced by federal law. Maple Grove’s legal theory has three steps.

First, the city reads Minnesota’s cable statute, Chapter 238, to require a franchise from any “cable communications system” that occupies a public right-of-way. It interprets that term to cover more than cable television, including broadband networks that use other technologies.

Second, Maple Grove classifies broadband as a Title I “information service” rather than a Title II “telecommunications service.” Title I services face lighter federal regulation, while Title II governs telecommunications carriers. The city argues that Gateway’s state certificate therefore does not make it a “telecommunications right-of-way user” entitled to access without a cable franchise. Third, Maple Grove contends that the FCC lacks authority to preempt a state-law franchise requirement because broadband falls under Title I.

Federal law does not, however, leave the matter to the city. The Communications Act forecloses Maple Grove’s demand for two independent reasons. 

Section 253(a) bars any state or local requirement that “prohibit[s] or ha[s] the effect of prohibiting” an entity from providing telecommunications service. Gateway holds a certificate authorizing it to provide telecommunications services, including point-to-point connections. It has also filed a tariff with the Minnesota Public Utilities Commission that sets out the rates, terms, and conditions for its regulated services. A tariff is a public filing that governs how a carrier offers those services. 

Requiring Gateway to obtain a franchise that it cannot secure as a condition of receiving right-of-way permits amounts to an effective prohibition. The FCC’s longstanding California Payphone standard asks whether a requirement “materially inhibits or limits the ability of any competitor … to compete in a fair and balanced legal and regulatory environment.” The FCC has made clear that Section 253(a) covers express bans and de facto moratoria, even when the government never formally labels its action a prohibition. 

Maple Grove has frozen Gateway’s permits unless the company agrees to follow the FCC’s Part 76 technical standards for cable television systems. Gateway’s broadband network was not built to satisfy those standards. The city has therefore conditioned market entry on a franchise agreement that Gateway cannot execute.

Maple Grove fares no better if it applies the ordinance only to Gateway’s broadband service and argues that Title II does not protect that service. The FCC’s Mixed-Use Rule prevents cable-franchising authorities from regulating noncable services carried over a cable system. Section 624(b)(1) of the Communications Act, 47 U.S.C. § 544(b)(1), likewise bars local authorities from regulating broadband as part of a cable franchise because broadband is an information service. The FCC acted within its authority when it prohibited local governments from taxing noncable services offered over franchised systems. 

Maple Grove cannot have it both ways. If broadband is a Title I information service, as the city insists when challenging Gateway’s telecommunications status, the Mixed-Use Rule and Section 544(b)(1) bar the city from franchising it or imposing franchise fees on it. If broadband qualifies as telecommunications, Section 253 bars Maple Grove from blocking right-of-way access to extract a franchise.

The city needs broadband to count as an information service for one argument and as a franchiseable cable service for the other. It cannot be both. Either way, the ordinance fails.

Winning After Winter Is Still Losing

Gateway will probably win. The Communications Act leaves little room for ambiguity, and the FCC has preempted comparable barriers before. Yet a victory before the FCC or in court may come too late to matter. 

Gateway filed its preemption petition in June 2026. Comments do not close until late September, and a ruling will come sometime after that. Fiber construction in Minnesota generally runs from April through September, before winter freezes the ground—and the schedule. By the time the record closes, the 2026 construction season will be over. Crews will have moved to other projects, capital will have sat idle, and tens of thousands of residents will have lost a year of access to another broadband option. 

Even a decisive victory cannot restore that lost year. Gateway will not have extended its network past those homes, and the company will absorb the cost of keeping capital tied up in a project that federal law entitled it to build all along. 

The problem extends beyond Maple Grove. A franchise demand costs little to impose and takes months or years to defeat. A city may find the tactic worthwhile even when the demand is plainly unlawful because it receives the immediate benefit of delay or added revenue while the provider bears the cost. After-the-fact preemption resolves the individual dispute, but it does too little to discourage the tactic. Every community watching its cable-franchise revenue decline faces the same temptation. 

The larger cost appears when providers decide where to build next. A company considering a thin-margin market must account for the risk that a city will withhold permits to support an unlawful revenue demand. The provider may win the legal dispute, but a victory one construction season later may offer little consolation.

That risk raises the provider’s hurdle rate, or the minimum return a project must promise before the company will invest. It affects investment decisions in Maple Grove and in every jurisdiction that might try the same tactic. The mere possibility of an unlawful demand can deter deployment, even if the demand would eventually lose before the FCC or a court.

A Fiber Win, Frozen Solid

None of this means the FCC should wait. Preemption is both correct and necessary. But relying on the FCC to “sort it out” case by case will not protect deployment if local governments keep finding new ways to tax broadband.

Providers need rules that keep permits moving while disputes proceed. Those rules should include firm shot clocks, which set deadlines for permit decisions, and deemed-granted remedies, which allow construction to proceed when a locality misses the deadline. Fees should also remain tied to the government’s legitimate costs, eliminating the financial reward for imposing new barriers. 

Removing regulatory obstacles remains one of the least expensive and most effective tools in communications policy. Yet the remedy must arrive in time. Otherwise, providers can win every legal case and still lose to the economics of broadband deployment.

Congress already answered whether cities can do this. The task now is to make sure the answer arrives before the ground freezes.

The post Fiber Freeze: How Maple Grove Made a Cable Franchise the Price of Broadband appeared first on Truth on the Market.

The proposal may sound like a tidy way to counter the power of large platforms. It raises much messier questions. Who qualifies for the exemption? Who represents contractors with different interests? And what happens to prices, output, market entry, and workers who prefer flexible or individually negotiated terms? Answering those questions requires examining the economics of collective bargaining, the legal and political difficulty of defining a new exemption, and the costs that protected coordination may impose on workers outside the bargaining group.

The labor-antitrust exemption reflects a durable political compromise. It should not become a blueprint for shielding collective price-setting by independent contractors from competition. A better approach would preserve competition and flexible work while addressing specific worker-welfare problems through portable benefits, clear classification rules, and fewer regulatory barriers to entry and mobility.

Cartel Economics With a Union Label

Antitrust begins with a simple proposition: Competition generally serves the public better than private agreements that suppress it. That principle should not change merely because labor is the input at issue. Agreements among employers to fix wages or divide workers rightly raise serious antitrust concerns. The economic mirror image is easy to recognize. When workers who would otherwise compete to sell their labor agree on wages or contract terms, or jointly withhold their services, they coordinate as sellers in the labor market. Structurally, that conduct resembles a cartel. The law immunizes an important subset of it, but legal immunity does not alter the economics. I made this symmetry explicit in “Antitrust Applies to Unions as Well as to Employers”: Restraints on either side of a labor transaction can distort competition.

Unions can still produce benefits. Collective voice can lower information costs, help workers secure shared workplace benefits, and provide channels for resolving grievances. The relevant law & economics question is whether monopoly privileges and antitrust immunity improve welfare compared with less restrictive alternatives. Exclusive bargaining replaces competition among workers—and often among forms of worker representation—with a single bargaining agent.

A negotiated increase in wages or benefits may therefore represent a transfer rather than a productivity gain. When the price of labor rises above the competitive level, employers respond in familiar ways: They hire fewer workers, substitute machines or other workers, invest less, raise downstream prices, or reduce output.

Workers outside the bargaining unit may bear many of those costs. A union wage premium can make union jobs scarce and push displaced workers into nonunion employment, increasing the labor supply and depressing wages there. Reduced output can also limit opportunities for customers or at suppliers and competing firms. Seniority rules and restrictive work practices may further divide insiders from outsiders, allowing incumbent workers to collect above-market returns while leaving new entrants with fewer openings.

None of this excuses employer monopsony—the power of one or a few employers to suppress wages—or employer collusion. It instead supports applying the same competitive baseline to both sides of the market. Concentration among sellers does not become harmless simply because the sellers happen to be workers.

Liya Palagashvili and Revana Sharfuddin sharpen this point in their review of 147 studies on union power and worker outcomes. They find that powerful unions can secure short-term gains while contributing to slower employment growth, less investment and research and development, weaker firm growth, and fewer future jobs.

Palagashvili and Sharfuddin distinguish between a union’s “voice” function and its “monopoly face.” U.S. law generally gives a certified union exclusive authority to represent a bargaining unit, foreclosing rival unions, alternative forms of representation, and individual bargaining within that unit. The literature they review suggests that more pluralistic representation and flexible agreements can preserve useful worker voice while reducing the costs of monopoly bargaining.

Even workers protected by monopoly unionism can lose from it over time. A rule marketed as pro-worker should therefore be judged by its effects on employment, investment, productivity, mobility, and worker choice—not merely by the size of a negotiated wage increase.

A Century-Old Compromise Meets the 1099 Economy

More than a century ago, the United States made a political and legal choice to shelter core labor activity from antitrust law. As Andrew Liu and I explain in a new Mercatus policy brief, the statutory labor exemption rests chiefly on the Clayton Act, the Norris-LaGuardia Act, and the labor law framework that followed. It shields union collective bargaining from federal antitrust challenges and protects specified peaceful tactics, such as certain strikes and picketing.

These protections extend beyond formal union members. Union and nonunion employees may engage in certain concerted activity through a bona fide labor organization seeking better wages, hours, or working conditions. A complementary judge-made doctrine, known as the nonstatutory labor exemption, protects restraints closely tied to collective bargaining between labor and management.

From a competition perspective, this arrangement is exceptional. From an institutional perspective, it is understandable. Congress adopted the exemption after courts applied the Sherman Act—the principal federal law against anticompetitive agreements—to strikes, boycotts, and other labor activity. Congress deliberately placed part of labor relations beyond ordinary antitrust rules.

Whatever one makes of that history, the employee labor exemption now sits deep within federal statutes, collective-bargaining institutions, and political expectations. Wholesale repeal is not a realistic near-term agenda. Sound analysis must therefore separate two questions: Is the inherited exemption economically sound, and should lawmakers expand its boundaries? One can doubt the first without embracing the second.

The compromise behind the existing exemption provides good reason for caution. Independent contractors are not simply employees with a different tax form. They include rideshare drivers, truck owner-operators, consultants, designers, real estate agents, tradespeople, caregivers, data trainers, and small-business owners. Their investments, dependence on particular buyers, ability to serve multiple clients, entrepreneurial discretion, and tolerance for risk vary widely.

A rule exempting employee collective bargaining is imperfect, but comparatively clear. A rule allowing any service provider with sufficient bargaining disadvantage to coordinate with rivals would require courts or agencies to decide how much dependence, invested capital, entrepreneurial discretion, or buyer power is enough. That is precisely the boundary problem Liu and I identify.

Choose Your Own Labor Exemption

Proponents of extending the labor exemption to independent contractors have offered two principal routes. The first is legislation. Eric Posner proposes tying worker status more closely to monopsony power. Douglas Melamed and Steven Salop propose specially authorized bargaining groups for independent contractors, perhaps limited to markets with substantial buyer power and subject to safeguards against monopoly abuse.

Melamed and Salop offer the stronger economic case because they at least target a genuine market failure instead of granting a blanket immunity. Their proposal also exposes the administrative difficulty. The government would need to define the relevant labor-purchasing market, measure monopsony power, determine which contractors belong in the group, specify what subjects they may negotiate and for how long, and prevent the bargaining entity from excluding rivals or restricting output. Courts struggle with those tasks even in conventional antitrust cases. A 2025 International Center for Law & Economics (ICLE) white paper, “Labor Monopsony and Antitrust Enforcement: A Cautionary Tale,” examines the unsettled evidence on labor monopsony and the perils of mechanically applying product-market tools to labor markets.

Liu and I also identify a more immediate political obstacle. The Department of Labor’s February 2026 proposed independent-contractor rule favors an entrepreneurship-centered “economic realities” test, which examines the practical nature of the working relationship. The proposal treats control and the opportunity for profit or loss as core factors. That approach conflicts with proposals to define employee status according to bargaining power.

Any new exemption would also create valuable privileges on either side of its legal boundary. Industries, platforms, unions, professional associations, and contractor groups would have strong incentives to lobby over coverage, eligibility thresholds, bargaining subjects, and the choice of authorized representative. A measure intended to clarify worker status could instead produce another round of classification disputes and litigation.

The second route relies on creative judicial interpretation. In 2022, the 1st U.S. Circuit Court of Appeals held in Jinetes that independent-contractor jockeys could invoke the Norris-LaGuardia Act’s labor-dispute framework. In January 2025, an outgoing Federal Trade Commission (FTC) majority issued an enforcement policy statement endorsing a broad view of protected labor activity for independent contractors and gig workers.

Those developments show that courts can breach the doctrinal line between employees and contractors. They also expose the problem Liu and I identify: Once formal employment status stops supplying the boundary, what replaces it? Distinguishing “wages for labor” from “prices for services” becomes especially difficult when a self-employed person sells her own time and skill. Tests based on market power or whether business capital can readily serve other purposes merely move the ambiguity elsewhere.

Massachusetts shows that targeted political action can succeed. A 2024 ballot measure created a state collective-bargaining regime for app-based rideshare drivers, and the commonwealth certified a statewide bargaining representative in 2026. The experiment deserves attention, but it provides no warrant for quietly rewriting federal antitrust doctrine. Massachusetts adopted a transparent political carveout with defined geographic and industry limits. If other states follow, policymakers can observe the effects on employment, prices, market entry, and innovation.

This state experiment therefore counsels federal restraint. Legislatures should own the tradeoffs they create. Antitrust doctrine should not hide those choices behind an elastic definition of “labor dispute.”

Two Monopolies Don’t Make a Market

Even if Congress could write a workable exemption, the economic case for extending it is weaker than advocates often suggest. The strongest argument for contractor collective bargaining is countervailing power: When a platform or other purchaser has genuine monopsony power, coordinated contractors may claim a larger share of the gains from trade.

Yet a bilateral monopoly—a market dominated by one buyer and one seller—is still no one’s idea of competition. A protected seller cartel may offset the buyer’s bargaining leverage, but it can also restrict output, raise consumer prices, exclude lower-cost or part-time providers, and entrench a bargaining organization whose interests differ from those of workers with the fewest alternatives. A second monopoly offers no general cure for the first.

Policy should target the source of the problem. If employers or platforms collude, antitrust law should attack the collusion. If licensing rules or other entry barriers suppress competing buyers, policymakers should remove those barriers. If a single buyer maintains durable market power through exclusionary conduct, conventional antitrust remains available when the government or a private plaintiff can prove the required elements.

The distinction between bargaining leverage and output-reducing monopsony matters. Related Truth on the Market articles, including “Kroger/Albertsons: Is Labor Bargaining Power an Antitrust Harm?” and “Labor Antitrust: A Solution in Search of Evidence,” caution against treating every shift in the division of bargaining gains as an antitrust injury or assuming that market-concentration statistics prove monopsony.

That caution cuts both ways. A contractor group that negotiates higher fees has not necessarily created social value. The relevant questions are whether buyers purchase fewer services, consumers pay more, quality or availability declines, or the arrangement excludes new providers. Competition policy should focus on actual effects on output and innovation. It should not become a general-purpose tool for redistributing the gains from trade.

Those output effects carry particular weight in independent contracting because heterogeneity and flexibility are often the point. A uniform minimum rate, mandatory benefits package, common scheduling rule, or exclusive representative may help contractors who would have chosen those terms anyway. The same rules may hurt a parent seeking a few flexible hours, a student willing to work for less during off-peak periods, a specialist who prefers individual negotiations, or an entrepreneur using platform work to supplement income while building another business.

A bargaining cartel naturally favors common terms. Independent work creates value partly by permitting uncommon ones. As the law replaces diverse arrangements with a standardized bargain, it risks destroying the flexibility and range of choices that workers seek through independent contracting.

Evidence about worker preferences makes that risk concrete. In the Bureau of Labor Statistics’ (BLS) July 2023 Contingent and Alternative Employment Arrangements survey, 80.3% of independent contractors whose sole or main job involved contract work preferred their current arrangement. Only 8.3% preferred a traditional employment arrangement. Those preferences do not prove that every contractor has adequate bargaining power. They do provide strong evidence against presuming that contractors as a class want employment-like terms. A pro-worker policy should preserve flexible contracting while addressing specific gaps directly, rather than make workers’ preferred arrangement scarcer to deliver benefits available through less restrictive means.

Economic incidence—who ultimately bears a cost—provides another reason for caution. If collective bargaining raises the expected cost of hiring contractors, firms will not simply absorb the difference. Higher consumer prices, lower output, automation, tighter eligibility requirements, narrower geographic service, and substitution toward employees or larger vendors all become more attractive. Contractors with the fewest alternatives bear the resulting loss of opportunities.

That prediction does not rest on an exotic economic theory. It describes the standard response to a higher input price. The size of the response depends on how sensitive demand is to price and how readily firms can substitute other inputs. Those effects may be substantial in platform markets, where software can quickly reallocate tasks and firms can redesign their business models.

Worker-classification evidence points in the same direction, with an important caveat. Extending an antitrust exemption would not itself reclassify contractors as employees. Collective-bargaining proposals nonetheless often accompany reclassification efforts or rules that make contractor relationships more closely resemble employment.

A study by Palagashvili and her coauthors of California Assembly Bill 5 (AB5) reports that, in affected occupations, self-employment fell 10.5% and overall employment fell 4.4%, without a corresponding meaningful increase in traditional employment. Policymakers should not mechanically apply those estimates to an antitrust exemption. The study supports a broader institutional lesson: Restricting one form of work does not ensure that firms will recreate the same opportunities under another legal label. Some jobs simply disappear.

An Exemption Worth Lobbying For

Public choice—the study of how political incentives shape government decisions—offers another reason to resist casually expanding the exemption. Mancur Olson’s “Logic of Collective Action” explains why small, concentrated groups organize more effectively than large groups whose members each have little at stake. George Stigler’s “Theory of Economic Regulation” shows how organized interests can acquire and use regulatory power. Gordon Tullock’s classic “Welfare Costs of Tariffs, Monopolies, and Theft” explains that monopoly’s social costs include the resources spent securing and defending special privileges, in addition to the lost transactions caused by monopoly pricing.

None of these arguments assumes bad motives. Unions, platforms, professional associations, incumbent contractors, and regulators can all respond rationally to the incentives that legal rules create.

A new antitrust immunity would confer a valuable right on some competitors: permission to coordinate. Political combat would predictably follow over who qualifies. Incumbent contractor groups could seek definitions that exclude new entrants. Unions could pursue exclusive-representation rules. Platforms could bargain for provisions favoring their business models over those of rivals. Established firms might welcome compliance costs that smaller entrants cannot afford.

Consumers and prospective contractors would prove harder to organize because each would bear only a small share of the total cost. Our warning about special-interest capture—organized groups shaping rules for their own benefit—is therefore central to the analysis. It is a cost created by the exemption itself. A legal regime intended to offset private market power can end up manufacturing politically protected market power instead.

Benefits Without the Cartel

A better approach would address the concrete problems of independent work directly. Portable benefits offer the most promising example. Traditional benefits come bundled with employee status largely for historical and tax reasons. Nothing inherent in health coverage, retirement savings, disability insurance, or paid leave requires tying those benefits to a single employer.

Portable-benefit accounts would instead belong to workers. Multiple hiring entities could make voluntary or required contributions, and workers could keep the benefits as they move among clients. Palagashvili and Jonathan Wolfson describe this model in “Why Independent Workers—and the Companies That Hire Them—Need Portable Benefits.” By June 2026, they reported that eight states had enacted some form of portable-benefit reform and that others were considering legislation.

The central legal reform is a classification safe harbor—a rule ensuring that specified benefit contributions do not count as evidence of employment. Without one, firms may hesitate to contribute to contractors’ benefits for fear that a court or agency will later use those contributions to reclassify the relationship. States can remove that disincentive by declaring that qualifying portable-benefit contributions do not determine worker status. Congress can adopt the same rule for federal employment laws.

The proposed 2025 Unlocking Benefits for Independent Workers Act illustrates this safe-harbor approach. If conflicting state rules eventually create a serious obstacle to interstate contracting, Congress could narrowly preempt them, meaning federal law would displace only the state rules that penalize qualifying portable-benefit arrangements. That preemption should address the identified conflict without imposing a national employment model.

Classification policy calls for the same precision. Actual misclassification—labeling someone an independent contractor while exercising the control and creating the dependence that governing law treats as employment—should remain subject to liability. Yet “misclassification enforcement” should not become shorthand for forcing bona fide entrepreneurs into employee status.

Classification rules should account for control over schedules and work methods, opportunities for profit or loss, personal investment, service to multiple clients, and other signs that a person operates an independent business. The Department of Labor’s 2026 proposed rulemaking, which would replace Biden administration guidance on employee and contractor status, gives greater weight to those entrepreneurial factors. Whatever test the department ultimately adopts, it should be stable, clear, and easy to administer. Uncertainty alone can deter firms from offering contract work.

Policymakers should also remove regulations that restrict entry and worker mobility. These include unnecessary occupational-licensing requirements, which require government permission to work in certain fields, as well as artificial limits on new business models, rules that impede benefit innovation, and restrictions that shield incumbent firms from competitors.

A large body of research links excessive economic regulation to weaker productivity and growth. James Broughel and Robert Hahn’s survey and synthesis of research across countries, for example, finds substantial evidence that economic regulation tends to reduce welfare in otherwise competitive markets. The labor-market consequence follows readily. Businesses hire employees and contractors when doing so creates value. Rules that discourage entry, investment, experimentation, and firm growth eventually reduce demand for their work.

This framework applies the same principles to both sides of the labor market. Enforce antitrust law against employer wage-fixing and no-poach cartels, in which employers agree not to recruit one another’s workers. Pursue exclusionary conduct when it satisfies the elements of monopolization. Apply worker-protection laws to sham classifications.

At the same time, preserve competition among workers, distinguish bargaining-power complaints from antitrust injuries, and protect workers who prefer independent contracting over standardized employment. The goal is larger than expanding the reach of labor law, antitrust law, or any representative institution. Policy should maximize opportunities for mutually beneficial exchange while preserving competition and worker choice.

Competition Works for Workers

The labor-antitrust exemption is a political settlement that the United States is unlikely to abandon soon. Policymakers can accept that legal fact without elevating it into a general economic principle. Unionization replaces some labor-market competition with coordinated bargaining. That coordination may raise wages for insiders while reducing employment, investment, mobility, and opportunities for outsiders. Monopoly representation can even harm the workers it purports to protect. Palagashvili’s research is especially important because it makes a crucial distinction between the case for worker voice and the case for monopoly unionism.

Extending this exceptional immunity to independent contractors would compound the underlying problem. Liu and I explain why broad legislative and judicial routes would create unstable legal boundaries. Law & economics supplies another objection: A new exemption would authorize additional restraints in markets where flexibility, diverse arrangements, and easy entry create much of the value. It could raise contracting costs, reduce output and job opportunities, limit worker choice through standardized terms, and create new privileges worth lobbying to control.

Targeted reform offers a better pro-worker agenda. Portable benefits can follow people among jobs. Safe harbors can let firms support those benefits without risking reclassification. Stable classification rules can preserve bona fide independent work while holding firms liable for genuine fraud. Antitrust enforcement can target employer collusion and exclusionary conduct, while broader reforms remove regulatory barriers to entry and entrepreneurship.

Employees and independent contractors share an interest in a dynamic economy where firms compete to hire, workers can move among opportunities, and new business models can challenge incumbents. Durable gains in worker welfare are more likely to come from stronger competition and fewer harmful regulations than from one more legally protected restraint.

There is nothing pro-worker about making opportunity scarcer.

The post Cartels With Benefits: The Trouble With Extending Labor’s Antitrust Exemption appeared first on Truth on the Market.

That question is now before California’s Sixth Appellate District. The dispute arises from a trial court ruling in Haynie v. Google that could turn routine age-based ad targeting into a violation of California’s Unruh Civil Rights Act. The International Center for Law & Economics (ICLE) filed an Aug. 17 amicus letter supporting Google’s petition for a writ of mandate in Google LLC v. Superior Court.

If the ruling stands, it could cause widespread, unintended harm to consumers, small businesses, and the digital economy. Multisided platforms—online services that connect businesses with users—would likely restrict advertisers’ use of age data. The predictable result would be less relevant and less age-appropriate advertising for everyone.

When Knowing Your Audience Is the Problem 

The lawsuit asks whether using age criteria to target online ads violates California’s Unruh Civil Rights Act. The trial court drew a sharp distinction. California law may permit differential pricing based on age, the court concluded, while differential advertising based on age may violate the Unruh Act.

The court also rejected Google’s claim of immunity under Section 230 of the Communications Decency Act, the federal law that generally protects online platforms from liability for third-party content. Because Google allowed advertisers to use age criteria, the court reasoned, it helped develop the allegedly unlawful content.

The Unruh Act aims to eliminate “arbitrary, invidious or unreasonable” discrimination. As one court explained, the “fundamental purpose of the Unruh Civil Rights Act is the elimination of antisocial discriminatory practices—not the elimination of socially beneficial ones.” Consistent with that purpose, courts have upheld age-differentiated products and prices when those practices provide a social benefit. 

In this case, the trial court found it sufficient for the plaintiffs to allege that Google let advertisers use age criteria, causing them to miss certain ads for insurance and financial products. Yet it is hard to see anything invidious about showing student car-insurance discounts to college-age consumers or Medicare Part D prescription-drug plans to seniors. Making ads less age-appropriate serves no one. 

There’s a Price for Irrelevant Ads

ICLE’s amicus letter urges the court to consider how multisided platforms work before deciding whether age-based ad targeting provides a social benefit.

Services such as Google Search, YouTube, and Gmail connect users with advertisers. Advertisers pay Google to reach those users, and that revenue subsidizes free access to Google’s services. The arrangement generates tens of billions of dollars in annual consumer surplus—the value consumers receive beyond what they pay.

That value depends partly on advertising that people find useful. If ads become less relevant or helpful, users will engage less, advertisers will earn lower returns, and Google will have less revenue to fund free content and services.

The U.S. Supreme Court recognized in Ohio v. American Express that multisided platforms must balance the interests of each group they serve to maximize the platform’s value. Google can offer its services for free because targeted advertising works. Users of all ages benefit because they might otherwise have to pay for access. Restrictions on ad targeting would reduce the services’ value to consumers and advertisers alike. 

Consumers would likely see less relevant, less age-appropriate, and potentially more obtrusive ads. Advertisers with limited marketing budgets would waste more money reaching people unlikely to want their products. Online platforms would face greater pressure to impose paywalls, raise ad prices, or restrict features. 

These economic effects help explain why age-targeted advertising can provide a social benefit under the Unruh Act. Google should not face liability merely for giving advertisers the tools to reach age groups likely to value their offers.

A Chilling Forecast for Online Ads

ICLE also argues that courts should interpret the Unruh Act consistently with Section 230 and the First Amendment. 

When an advertiser uses age in an invidious way, liability should rest with the advertiser rather than with a multisided platform that provides neutral targeting tools. Advertisers decide how to structure their campaigns. Google simply lets them select age as one possible targeting criterion. 

Treating Google as a co-developer of every ad that uses age would give platforms a strong reason to eliminate age-based targeting altogether. That response would suppress a substantial amount of lawful advertising, which qualifies as protected commercial speech. 

Section 230 and the First Amendment guard against this kind of “chilling effect,” in which the threat of liability discourages lawful speech along with potentially unlawful conduct. A rule meant to prevent discriminatory advertising should not make useful, lawful advertising collateral damage. 

You Can’t Target Everyone

Product-liability lawsuits against major social media companies have drawn plenty of headlines. This case presents a quieter threat to the ad-supported internet. If the trial court’s ruling stands, platforms may respond by eliminating useful targeting tools and restricting lawful advertising.

The Sixth Appellate District should grant Google’s petition and make clear that a platform does not incur liability merely by allowing advertisers to target ads by age.

Attention is scarce. Making advertising less relevant will not help consumers. You cannot interest everyone in everything—and the law should not force advertisers to try.

The post Google, the Unruh Act, and the Legal Risk of Knowing Your Audience appeared first on Truth on the Market.

Adam Smith knew the type. In “The Theory of Moral Sentiments,” he warned against the “man of system,” who imagines society as a chessboard and believes he can move human beings as easily as pieces.

Today’s man of system has a data-center plan. Governors, legislators, regulators, and activists increasingly speak as though they can determine where enormous new electricity loads will locate, which power sources will serve them, how their owners will bargain with utilities, what labor terms they will accept, and how much support they will provide their communities—all while preserving low rates, grid reliability, environmental goals, and America’s lead in AI.

The chess pieces, it turns out, have a motion of their own.

In Pennsylvania, Gov. Josh Shapiro unveiled the Governor’s Responsible Infrastructure Development Standards, or GRID Standards, as the terms developers must meet to receive faster permitting, tax incentives, and coordinated state support. The Pennsylvania House voted 134-68 in June to codify them. When the Senate did not act, Shapiro told the Pittsburgh Business Times that he would consider executive action.

In Texas, Gov. Greg Abbott ordered regulators earlier this month to freeze approvals for data centers seeking grid connections until the state completes an audit. New York has imposed a one-year statewide moratorium on permits for the largest facilities.

The political pressure is clear. A July Quinnipiac poll found that 74% of Pennsylvania voters opposed an AI data center in their community. Republicans, Democrats, and independents agreed. The New York Times called the opposition perhaps “the most bipartisan issue since beer.”

Some of that resistance is exaggerated, emotional, and plainly hostile to growth. Other objections deserve a serious answer. Data centers are large industrial facilities, and residents are entitled to ask about noise, land use, water, air emissions, electricity bills, tax abatements, and the integrity of local decision-making. The secrecy surrounding some projects has deepened public distrust. A credible free-market case for data centers must acknowledge legitimate costs and concede that hyperscalers—companies that operate enormous networks of data centers—do not always strike defensible bargains.

Yet policymakers commit a grave error when they treat a project’s costs as grounds to stop an industry. That response treats scarcity as evidence of market failure, assumes public officials can identify the correct technical response in advance, and interrupts the decentralized adjustments already underway.

It also mistakes the visible building for the demand it serves. Demand for cloud computing, cybersecurity, medical research, financial services, logistics, streaming, and AI persists after a government prohibits a data center. The facility simply goes elsewhere and takes its investment, infrastructure, tax base, and accumulated knowledge with it.

Austrian economics focuses on how dispersed knowledge, prices, and entrepreneurial experimentation help people adapt to scarcity. Through that lens, the relevant question is whether moratoria and prescriptive mandates improve the process by which firms, utilities, communities, and consumers reconcile rapidly growing demand with limited supplies of electricity, water, land, labor, and capital. In reality, such policies obstruct that process.

The Grid’s Preexisting Condition

The data-center debate emerged amid longstanding grid problems. U.S. electricity demand remained comparatively flat for roughly two decades, and utility forecasts, investment patterns, and regulatory institutions adapted accordingly. Meanwhile, policy-driven plant retirements subordinated reliability to an environmental ideology. Slow permitting, transmission constraints, equipment shortages, inflation, flawed market rules, and interconnection backlogs further limited how quickly suppliers could add capacity. AI and cloud computing then produced a large, geographically concentrated surge in demand.

The popular account starts with the latest events. Data centers arrive, electricity prices rise, and officials decide to restrain the new demand. A Misesian account, named for Austrian economist Ludwig von Mises, starts earlier. One government intervention disrupts the market process and produces consequences that prompt another intervention. Government action reduces the electricity supply or raises production costs. Increased demand exposes those constraints. Officials blame data centers, the most visible new users, for all resulting price increases and then propose freezes or quotas that make new supply even harder to finance.

Regulation did not cause every grid problem. Still, the newest electricity user cannot explain the current scarcity by itself. PJM Interconnection (PJM), the country’s largest regional grid operator, recently held a capacity auction, in which electricity suppliers receive commitments to ensure that enough power will be available during periods of peak demand. The auction produced billions of dollars in additional charges. The New York Times reported a $6.3 billion increase affecting 13 states and the District of Columbia, with much of its coverage focused on data-center demand.

The evidence is serious, though incomplete. A detailed 2026 analysis by Energy and Environmental Economics (E3) estimated that peak-load growth, driven primarily by data centers, accounted for roughly half of the increase in PJM’s capacity price between 2024-25 and 2025-26. Retirements, program changes, supply constraints, and market-design problems accounted for the other half. PJM lost more than 9 gigawatts of generating capacity between 2023 and 2025, and utilities have announced further retirements. Its interconnection process, which connects new generators to the grid, has struggled to bring replacement capacity online.

Blaming data centers for the entire increase is like blaming a heavy truck for exposing the weakness of a neglected bridge. Years of neglect weakened the bridge. The truck revealed the cost.

A moratorium on data centers therefore amounts to political rationing masquerading as prudence. Moratoria leave the grid’s problems untouched. They build no generators, shorten no transmission-permitting timelines, manufacture no transformers, and make no improvements to PJM’s interconnection process. At most, they suppress one category of demand while leaving the supply shortage intact. If that demand relocates, the jurisdiction loses the associated investment while the national infrastructure problem remains.

Hayek Walks Into a Data Center

Friedrich A. Hayek’s “The Use of Knowledge in Society” begins with a simple claim. No single mind possesses the knowledge needed to coordinate an economy. That knowledge is dispersed among individuals and embedded in the “particular circumstances of time and place.” Prices communicate it without requiring anyone to understand every distant cause.

Electricity markets illustrate Hayek’s point unusually well. A decision to locate a data center in Luzerne County in northeastern Pennsylvania, rural Texas, northern Virginia, or a site no one has considered depends on thousands of facts. They include available generation, substation capacity, transmission congestion, fiber routes, cooling methods, water conditions, land prices, construction labor, regulatory timelines, requirements for uninterrupted service, future computing power per rack, and customers’ tolerance for delay. Officials cannot know many of these facts. Others have yet to emerge because investment and experimentation will create them.

A higher electricity price, longer interconnection wait, or more expensive water right conveys useful information. It may prompt a developer to reduce consumption, finance new supply, shift computing to off-peak hours, change its technology or location, or abandon a project that cannot justify its resource use. Developers choose among these responses, and their choices generate new information. Scarcity calls for economizing and building. Bans silence the signal.

A moratorium suppresses that discovery. “Wait” reveals nothing about which projects are genuine, which can cover their costs, which can shift their electricity use, or which sites have adequate capacity. It replaces countless project-by-project decisions with a single political decree.

Texas demonstrates the problem. Regulators reported a queue of roughly 1,800 proposed projects representing 474 gigawatts—more than five times the state’s record peak demand. Data centers accounted for about 90%. Many applications are plainly speculative. Only 28 of 377 companies responded to a state survey requesting details about their plans. Regulators cannot protect reliability using fictional projects.

The queue presents a screening and information problem. Treating every application alike fails to solve it. Regulators can require binding demand forecasts and deposits that developers forfeit when they reserve scarce queue capacity without making progress. They can demand collateral, minimum payments, construction milestones, and evidence that developers control their proposed sites. They can assign the costs of dedicated substations and transmission upgrades to the customers that cause them. They can also remove applications that miss their deadlines. These mechanisms separate serious projects from placeholders. A categorical freeze obscures that distinction.

Israel Kirzner extends Hayek’s argument by describing competition as a process of entrepreneurial discovery. Markets allocate known resources among known uses. They also allow entrepreneurs to notice overlooked opportunities, test judgments under uncertainty, and reveal information through profit and loss. That process becomes especially valuable when inputs, production methods, and products are changing simultaneously, as they are in AI infrastructure.

No governor knows the optimal power mix for data centers in 2035. Neither does any hyperscaler. Institutions determine who can test forecasts and who bears the cost of error. Firms can try different combinations, copy successful approaches, and pay for their mistakes. A mandate imposes one political forecast and distributes the cost of its errors among everyone subject to it.

Hayek’s fatal conceit appears here in industrial form. Intelligence and good intentions cannot replace a process that generates knowledge no one possesses at the outset.

Make the Hyperscalers Pay the Electric Bill

The strongest objection to data centers concerns electricity. A large, steady load may require new generation, substations, and transmission. If a project shrinks, stalls, or disappears, other customers could be stuck paying for facilities built to serve it. Poor cost allocation could leave residential ratepayers subsidizing some of the world’s largest companies.

Yet the evidence does not show that data centers inevitably raise residential rates. E3 found no quantitative evidence of systematic historical subsidies. The Data Center Coalition funded the report, and E3 cautioned that the limited research cannot predict future outcomes. Its analysis of four Amazon facilities found that they produced an average annual utility surplus of about $3.4 million. An Electric Power Research Institute analysis also found that greater electricity sales correlated with modestly lower retail rates between 2015 and 2024 because large customers helped spread fixed costs across more kilowatt-hours.

Recent examples support that finding. Georgia regulators approved a plan expected to reduce a typical residential bill by about $50 a year, with much of the reduction attributed to revenue from new large customers under a tariff designed to protect other ratepayers. Indiana Michigan Power has cited data-center growth in support of a base-rate reduction. In Louisiana, Entergy projects that its agreements to serve Meta’s Hyperion campus will provide customers with approximately $2.65 billion in benefits over 20 years.

These figures remain projections from utilities and regulators. They establish no universal rule. PJM shows that rapid demand growth can raise capacity prices when suppliers cannot add power quickly enough. Data centers can lower average costs when their payments cover the added expenses and their revenue grows faster than total system costs. They can raise costs when supply remains constrained or utilities shift expenses to other customers. Policy should create terms that favor the former result.

Utilities are already developing those terms. According to E3, utilities established at least 38 specialized large-load tariffs between 2018 and early 2026, including 30 in 2025 and 2026. These rate schedules use longer contracts, minimum bills, take-or-pay provisions that require payment for reserved power even if the customer does not use it, upfront construction contributions, credit support, exit fees, phased increases in electricity use, and cost-allocation reviews. American Electric Power Ohio, for example, developed a tariff intended to protect other customers from speculative projects and abandoned assets.

These mechanisms apply a straightforward principle known as cost causation. A data center should pay the additional costs it creates. If it receives continuous reliability, grid balancing, transmission, or backup service, it should pay for those services even when it generates some electricity on-site. If the project finances assets that benefit the broader grid, its contract should also assign those benefits accurately.

Cost causation cannot settle every dispute. Regulators must still determine how to measure shared expenses, and utilities retain incentives to overbuild or shift risk to captive customers. Transparent tariffs, auditable assumptions, and enforceable contracts allow regulators and utilities to revise their approach as evidence accumulates. A moratorium offers no similar means of correction. It turns a difficult pricing problem into a political veto.

Every Gallon Has an Address

Water gives the data-center backlash its most vivid imagery. Computers generate enormous amounts of heat, and many facilities use water to keep their servers cool. A server farm drawing millions of gallons near irrigated fields or residential wells naturally alarms neighbors.

Assessing the strain requires more than a single annual figure. A facility may withdraw water from a river or aquifer and return much of it, while consumption measures the portion lost to evaporation or otherwise removed from the local supply. Data centers also have an indirect water footprint because many power plants use water to generate the electricity they consume.

Timing and location matter just as much. Heavy use on the hottest day of the year can strain a local water system even when annual use appears modest, and ample water nationwide cannot replenish an overdrawn aquifer in a particular community.

U.S. data centers directly used roughly 17 billion gallons of water in 2023. That figure sounds enormous in isolation, though it represents a small share of national use and reveals little about any particular project. Evaporative cooling in an arid watershed may create a serious local problem. Air cooling, reclaimed wastewater, and recirculating systems in a water-rich region create different costs and risks. Efforts to reduce direct water use may also increase electricity demand or construction costs.

Officials should assess water use by watershed and during periods of peak demand. Comparable industrial users should report how much water they withdraw and consume. Prices should reflect local scarcity, authorities should enforce water rights, and developers should pay for dedicated infrastructure. Technology-neutral standards can identify a measurable local harm while allowing developers to choose among reclaimed water, dry cooling, recirculation, and other methods.

The American Enterprise Institute’s (AEI) review of the backlash makes a concession that strengthens the free-market case. National totals can appear reassuring even when local peak demand strains water systems. Research cited by AEI projects substantial additional water-capacity needs through 2030 if cooling efficiency remains unchanged. The same research finds that efficiency gains could sharply reduce those needs. Governments may reasonably require reporting and coordination, as well as “pipe-neutral” development that adds enough capacity or conservation to offset a project’s demands. Blanket bans disregard these differences.

Critics often overstate the water objection, but water remains scarce and costly in many communities. Authorities should measure, price, and address that scarcity where it occurs. A statewide moratorium based on a national statistic treats unlike communities identically and blocks technologies that could render yesterday’s estimates obsolete.

Spontaneous Order, Now With Turbines

Markets are already answering the claim that they cannot supply enough power for AI.

Two years ago, economist Lynne Kiesling framed the choice as “make or buy.” Would data-center operators continue purchasing electricity from utilities, or would high prices, delays, and reliability concerns lead them to generate their own? Her recent follow-up describes a third option. Firms can make, buy, or ally with utilities, power producers, equipment manufacturers, and investors. These arrangements are changing the commercial relationship between computing and electricity.

Some operators continue to buy power under ordinary or specialized utility tariffs. Others sign long-term power-purchase agreements, which commit them to buy electricity from a generator for a set period. Microsoft entered a 20-year agreement supporting the restart of Three Mile Island Unit 1 in Pennsylvania, while Meta signed a long-term agreement associated with the Clinton nuclear plant in Illinois. Google and Amazon are backing advanced nuclear-reactor projects. Other firms are developing power projects with gas-turbine manufacturers, utilities, and infrastructure investors. Some place generation “behind the meter,” meaning on the customer’s side of the utility connection, or next to an existing power plant.

Each choice assigns risk, control, financing, permitting duties, and technical responsibility differently. Delays also impose a price. A firm racing to deploy scarce computing capacity may lose more during a multiyear wait for a grid connection than it would spend on costlier on-site generation. One company may accept the complexity of a nuclear agreement to secure long-term reliability. Another may choose natural gas because turbines can begin operating sooner. A third may stay with a utility because producing power would distract from its main business.

Pennsylvania offers several examples. Prime Data Centers has proposed a 450-megawatt facility beside a gas plant in Hanover Township and says it can operate behind the meter. Amazon has discussed a possible data center at the former Homer City coal site in Indiana County, where developers envision a combined natural gas-generation and computing complex. GE Vernova is investing nearly $166 million in western Pennsylvania facilities that manufacture high-voltage and gas-turbine equipment. Data-center demand is prompting investment throughout the electricity supply chain.

These projects still warrant scrutiny. A behind-the-meter facility may remain dependent on the grid for synchronization, balancing, backup power, or other reliability services. Only a facility capable of operating independently during a grid outage avoids that dependence. The Hanover Township project has yet to resolve every water and permitting issue. Homer City still lacks a final customer. GE Vernova’s investment also receives state support. Each project carries costs, uncertainties, and unresolved questions.

They also provide evidence of discovery. Firms are testing different combinations because no one knows which model will prove most efficient, and the answer will likely vary by project. Experience will show which factors favor utility service, long-term contracts, partnerships, colocation, or on-site generation.

The Amazon-Talen colocation agreement at Pennsylvania’s Susquehanna nuclear plant shows how complicated that experimentation can become. Colocation places a large electricity user beside a generator, potentially allowing the facility to obtain power without moving all of it across the wider grid. Federal regulators rejected an amended interconnection agreement in 2024. They later directed PJM to develop clearer rules for arrangements that pair large data-center loads with power plants. Physical proximity does not erase the costs of shared grid services. The rules must assign those costs to the customers that incur them.

Law and contracts establish the terms for this experimentation, though no regulator, utility, or firm planned the full pattern. Many parties responding to prices and local constraints have developed specialized tariffs, nuclear restarts, joint ventures, flexible electricity use, colocation agreements, and on-site generation. Austrian economists call this spontaneous order, a system of coordination that emerges through many decentralized decisions rather than a central plan. A moratorium halts that experimentation when firms and regulators have the most to learn.

GRID Comes With Strings Attached

In a recent piece for National Law Review, I argued that Pennsylvania correctly recognized what AI data centers could bring. The projects could attract private investment, create demand for new power plants and electrical equipment, put former industrial sites back to productive use, and generate jobs and tax revenue. Pennsylvania’s energy resources, manufacturing base, and available industrial sites make it a plausible beneficiary. GRID’s error lies in assuming the state can prescribe the energy sources, labor arrangements, and community commitments that developers should use to produce those gains. As the backlash grows, the distinction between enforcing neutral rules and dictating production choices becomes more urgent.

GRID combines sound principles with central planning. Data centers should pay the costs attributable to their electricity use. Disclosure, early community engagement, water planning, and enforcement of generally applicable environmental laws are also defensible. GRID adds a prescribed “build, bring, or buy” framework, an increasing share of “clean firm” power that can operate regardless of weather, solar-ready construction, specified labor and compensation rules, community-benefit commitments, and continuing administrative oversight.

For now, GRID remains voluntary. A developer may reject its standards and proceed under ordinary permitting requirements and generally applicable laws, though it will lose access to tax incentives, expedited permitting, and coordinated state support. Pennsylvania can use those government-controlled benefits to steer investment toward its preferred production model. The House vote to codify the standards and Shapiro’s contemplated executive action show how readily an optional program can supply the blueprint for regulation.

Economist Murray Rothbard called this arrangement “triangular intervention.” The term describes government requiring private parties to conduct an exchange on terms they would not otherwise choose. Under GRID, a developer gains access to state support only after accepting politically selected obligations governing energy, labor, reporting, and community benefits.

Compliance costs are only part of the problem. The deeper difficulty concerns knowledge. Officials cannot know in advance which technologies, contracts, or operating practices will best address a project’s demands. Every prescribed input or quota prevents firms from testing at least one possible approach.

A data center might discover that shifting computing tasks that can tolerate delay to off-peak hours reduces grid costs more than dedicated generation would. A technology-neutral tariff can reward that choice by charging the customer for the costs it causes. A mandate tied to a specified technology or quota cannot accommodate the innovation until officials amend or waive it. Incumbent firms will usually handle that process more easily than new entrants. A policy advertised as discipline for Big Tech may end up protecting it from competition.

Economist James Buchanan’s principle of generality offers a better standard. Rules should group regulated parties according to relevant characteristics. A 450-megawatt chemical plant and a 450-megawatt data center may create similar risks for the grid. A large beverage plant may pose a greater threat to local water supplies than a server facility that uses dry cooling. Regulation should reflect electricity use, emissions, water withdrawals, credit risk, noise, and land-use effects. Officials’ views of the customer’s business model have no bearing on those harms.

Several reforms meet that standard. Texas can require every large electricity user to substantiate its place in the interconnection queue and demonstrate its financial capacity. Pennsylvania can require any customer responsible for a network upgrade to pay the added cost. Local governments can apply neutral rules governing noise, setbacks, traffic, light, wastewater, and emergency services. A special permission system aimed at one politically unpopular industry, by contrast, would replace general rules with administrative favoritism.

The Sweetheart Deal Comes Due

Louisiana’s agreement with Meta is often presented as proof that data-center growth can benefit ratepayers. Entergy says the project will support seven new combined-cycle gas plants, more than 5,200 megawatts of generation, major transmission upgrades, batteries, nuclear uprates, and solar development. The utility projects $2.65 billion in customer savings over 20 years, including $120 million for vulnerable customers and $140 million for energy efficiency. The project has already produced construction activity, local sales-tax revenue, training commitments, and infrastructure spending.

Those claims deserve attention, along with attribution. Parties to the deal made the projections, and most of the promised benefits have yet to materialize. A New York Times investigation also documented secret negotiations, nondisclosure agreements, potentially enormous tax breaks, the use of public land, political conflicts, and contractual risks that could fall on other customers if the assumptions prove wrong. The investigation found genuine local gains as well as rising rents and community disruption.

Austrian economics and public choice theory can account for both. A project this large can create substantial value, finance new electricity supply, and spread the grid’s fixed costs across more customers. Yet government control over tax privileges, public land, and regulatory approvals creates opportunities for rent seeking, in which firms pursue political favors instead of competing on equal legal terms. It can also foster favoritism, secrecy, and the transfer of private risks to the public. Corporate welfare deserves no free-market defense, and opposition to moratoria provides none.

Louisiana also shows how one intervention can prompt another. Officials grant a favored project customized tax and regulatory treatment, while some risks may fall on ratepayers or taxpayers. Secrecy and favoritism then provoke public anger. Rather than withdrawing the privileges and applying general rules, officials impose new restrictions on later projects. Privilege breeds backlash, and backlash breeds control. Alternating between political favors for selected firms and restrictions on an entire industry produces an interventionist spiral in miniature.

What the Moratorium Doesn’t Show

Frédéric Bastiat’s distinction between the seen and the unseen helps explain why data-center politics favors prohibition. Residents can see a proposed building, transmission line, cooling plume, or higher utility charge. The costs of rejection are harder to observe. They include investments canceled, suppliers that never expand, workers who never receive training, power plants that remain uneconomic, and innovations that scarce computing capacity makes more expensive or delays.

Critics often point out that a completed data center employs relatively few permanent workers compared with a manufacturing plant occupying similar acreage. That comparison is fair, though server-room employment captures only part of the economic activity. Data centers also require construction trades, electrical equipment, fiber connections, security, maintenance, engineering, and new power generation. They add tax revenue and demand for suppliers. PricewaterhouseCoopers’ (PwC) 2026 report estimates that the broader U.S. data-center industry supported about 5.5 million direct, indirect, and induced jobs and contributed roughly $927 billion to gross domestic product (GDP) in 2024.

Those figures require context. They do not represent 5.5 million new jobs or $927 billion in new output caused solely by data centers. PwC used a broad definition that included workers who support data-center functions and counted indirect jobs at supplier firms and induced jobs supported by workers’ spending. The report estimated roughly 1 million direct jobs and $302 billion in direct GDP. The Data Center Coalition commissioned the study, which measures activity associated with the industry rather than the benefits guaranteed by any proposed facility.

Even with those qualifications, the political imbalance remains. The benefits are dispersed among future workers, suppliers, taxpayers, and consumers, while nearby residents bear concentrated and immediate costs. Public choice theory predicts that organized residents who attend a hearing will command more attention than people who might benefit years later and may not know that the project would affect them. The residents’ concerns are real. So are the forgone benefits.

Moratoria can also reduce competition. They protect existing data centers against new entrants, increase the value of scarce computing capacity, and favor companies large enough to wait, lobby, litigate, or move elsewhere. Smaller developers and innovative energy suppliers have fewer resources to endure indefinite pauses and negotiate customized political deals. Policymakers worried that AI will concentrate power among a few technology companies should consider who benefits when the government restricts new computing and electricity supply. A policy advertised as a restraint on Big Tech may become Big Tech’s best moat.

A moratorium also imposes costs during the pause. A data-center project requires developers to coordinate land, fiber connections, interconnection rights, equipment orders, construction schedules, and customer commitments. Delay one part, and the entire arrangement may collapse. Investors compare states and communities, then move when another location offers greater certainty. A community may reject the facility while continuing to consume services produced elsewhere and paying a share of regional grid costs. It will forgo the investment and tax revenue, along with its ability to influence the project’s design.

Meter Readers, Not Master Planners

Government has a legitimate role in data-center development, one compatible with dispersed knowledge, property rights, and the rule of law.

Utilities should adopt transparent large-load tariffs based on cost causation. Developers should fund dedicated upgrades, post collateral for speculative capacity reservations, make minimum payments for infrastructure built to serve them, and pay exit charges if they abandon a project. Behind-the-meter facilities should also pay for any backup power, real-time balancing, and other grid-stabilization services they use.

State and local governments should enforce neutral, measurable rules governing noise, setbacks, traffic, emissions, wastewater, lighting, and emergency services. Water standards should reflect conditions within the affected watershed and apply equally to comparable industrial users. Community review should begin early, disclose material commitments, and produce decisions within predictable timelines. A hearing should establish site-specific facts rather than become a referendum on AI itself.

Government has a role to play in policing cost shifting. Developers should decide which power sources best meet their needs. Policymakers can help by removing barriers to new generation and transmission, improving interconnection queues, and permitting on-site and colocated generation subject to safety, environmental, and cost-allocation rules.

Performance standards can set measurable requirements for reliability, emissions, water use, and incremental costs while allowing firms to test different ways of meeting them. Technology mandates and politically chosen quotas foreclose those choices. Pennsylvania’s clean-firm percentages illustrate the problem. They prescribe a particular input instead of charging for a measurable effect. A better rule would require each project to cover its added grid costs and meet defined reliability and environmental standards while allowing the developer to choose its technology.

Governments should also abandon selective tax favors as a tool of social planning. Hayek’s distinction between general rules and discretionary commands in “The Road to Serfdom” offers a better framework. Government should announce general, prospective rules that allow firms to make their own plans. Bargaining over individualized privileges and obligations after a favored applicant arrives produces uncertainty and favoritism.

If a data-center tax exemption cannot be justified under a neutral tax system, lawmakers should repeal it. Retaining the preference and conditioning it on an expanding list of energy, labor, and community-benefit mandates compounds the original distortion. Tax neutrality provides a cleaner remedy for corporate welfare than industrial planning does.

Officials also need honest baselines. Electricity rates reflect fuel costs, inflation, plant retirements, reliability spending, transmission constraints, grid upgrades, market rules, and demand from many sources. Data centers may raise costs in one location and spread fixed costs among more customers in another. Water may be abundant over an entire year yet scarce during the hottest week.

Regulators who attribute every increase to the most visible new customer conceal other causes. Industry advocates who deny genuine local burdens sacrifice their credibility. Sound rules begin with candor from both.

The Chess Pieces Keep Moving

The data-center backlash is understandable. AI is advancing quickly, its infrastructure is physically immense, and secret negotiations and rosy projections have given communities ample reason for skepticism. Political unease, though, supplies none of the technical knowledge needed to design the next generation of computing and electricity.

Firms are already financing new generation, negotiating long-term power agreements, developing specialized tariffs, restarting nuclear plants, testing advanced reactors, building gas-fired generation, shifting computing workloads, improving cooling systems, and inventing new contracts. No state plan anticipated the full range of responses. Some projects will fail, and some deals will deserve rejection. Profit, loss, and experience reveal which arrangements work and who should bear their costs.

Government should protect property rights, enforce contracts, require truthful disclosure, charge accurately for public services, and make developers pay the costs they create. Policy should reflect the limits of official knowledge through general rules, technological neutrality, open entry, transparent agreements, and predictable decisions.

Infrastructure anxiety calls for more investment guided by prices, property rights, contracts, and competition. Those institutions allow firms, utilities, and communities to discover solutions that none can fully anticipate alone.

A community may reject a data center, but demand for computing will persist and attract investment elsewhere. The chess pieces will keep moving.

The post The Data Center Chessboard Has No Pause Button appeared first on Truth on the Market.

On July 6, Nigeria’s Federal Competition and Consumer Protection Commission (FCCPC) announced investigations into Meta, Alphabet, X, and unnamed generative artificial intelligence (AI) companies. The announcement followed a petition to the Nigerian presidency from the Nigerian Press Organization (NPO). 

The FCCPC identified three concerns: market dominance, the use of copyrighted news content to train AI models, and the absence of “equitable commercial engagement.” 

The press release’s final paragraph makes the FCCPC’s model explicit. It claims that a similar inquiry in South Africa ended with Google agreeing to pay South African news organizations 688 million rand ($40 million) annually for three to five years. As I explained in a previous piece, regulatory ambition has its own politics: An agency’s next move often follows the path laid by its counterparts abroad. 

But the FCCPC has misread the South African precedent in two important respects. First, 688 million rand is the total Google committed over five years, not an annual payment. Second, Google negotiated that payment under a statutory market-inquiry regime, a formal process that may give the South African Competition Commission (SACC) real power to impose remedies. Nigeria has no comparable regime. 

Nigeria is importing South Africa’s answer without South Africa’s legal machinery. 

The Fine Print Behind Google’s Check

Google’s media package emerged from the SACC’s Media and Digital Platforms Market Inquiry (MDPMI), which ran from October 2023 through November 2025. A market inquiry begins as a diagnostic exercise. The SACC asks whether a market feature has an adverse effect on competition, a much lower threshold than proving a legal violation. Its findings do not establish that any firm broke the law. 

Yet the inquiry comes with considerable remedial power. Under the Competition Amendment Act 18 of 2018, the SACC may take any remedial action it considers reasonable and practicable to address an adverse effect on competition. The exception is divestiture, which requires a firm to sell part of its business and which only the Competition Tribunal may order. The MDPMI’s final report bases its remedies squarely on sections 43C through 43E of the act. 

That discretion is extraordinary, though not unlimited. The remedial power lacks an enforcement mechanism, and South African courts have already held that the SACC cannot disguise a coercive divestiture order as voluntary compliance. 

Google negotiated with those powers in view. In its February 2025 provisional report, the SACC recommended annual payments of 300 million to 500 million rand. It also signaled that platforms that voluntarily adopted the proposed remedies could avoid final penalties or a digital levy. 

Google then proposed a package comprising 71 million rand annually for five years for national media through Google News Showcase, 45 million rand annually for three years for an AI innovation fund, and 38 million rand annually for three years—plus matching funds—for a Digital News Transformation Fund serving small and community outlets. The full package totaled 688 million rand over five years, or about $40 million. 

Meta, TikTok, Microsoft, and OpenAI accepted commitments covering revenue tools, access to their platforms, and publishers’ ability to withhold content. None involved payments. The SACC reportedly could not reach an agreement with X. 

A Remedy in Search of Jurisdiction

The FCCPC lacks the powers of its South African counterpart. Sections 17(b) and 17(d) of the Federal Competition and Consumer Protection Act (FCCPA) allow it to review economic activity and report on market practices, while section 18(1)(g) allows it to publish studies.

The FCCPA provides nothing resembling South Africa’s “quasi-regulatory” regime. It establishes no procedure for initiating a market inquiry, no adverse-effects standard, no binding force for the resulting report, and no remedial powers tied to the process. To obtain a remedy, the FCCPC must leave the study format behind and prove that a specific business abused a dominant position. A targeted firm is unlikely to accept such a finding without protracted litigation. 

Of the FCCPC’s three stated concerns, only the first—market dominance and anticompetitive conduct—falls within its competition mandate. Even that concern must eventually become an abuse-of-dominance case, with all the evidentiary burdens that entails. The other two suggest an agency either untroubled by its jurisdictional limits or unaware of them. 

Consider the FCCPC’s investigation into the unauthorized use of copyrighted news content to train generative AI models. Nothing in the FCCPA turns possible copyright infringement into a competition or consumer-protection matter. Copyright claims fall under the Copyright Act 2022, and copyright holders must enforce their own rights. 

Even under copyright law, the claim remains unsettled. Section 20(1) of the Copyright Act 2022 replaced the repealed act’s closed list of permissible fair-dealing purposes with an open, U.S.-style provision. It imported the four-factor test that American courts are now applying to AI training cases, with mixed results. Whether using Nigerian journalism to train a commercial AI model constitutes infringement or fair dealing remains an open question. 

If the NPO and its members object to the use of their content for AI training, they should sue in court. Publishers elsewhere have followed precisely that course, including in the United States and India. 

The third concern fares no better. The FCCPC alleges that “affected media organisations have been denied meaningful opportunities to negotiate fair compensation or appropriate commercial arrangements for the use of their journalistic content.” Although sections 124 and 127 of the FCCPA allow the FCCPC to police unfair contract terms in some circumstances, those powers belong to its consumer-protection mandate. 

Section 167(1) defines a consumer to include anyone “to whom a service is rendered.” Unlike the parallel definition for goods, the provision covering services contains no exclusion for business use. On a literal reading, a Nigerian publisher using Google’s advertising tools or Meta’s monetization programs could itself qualify as a “consumer,” with the platform serving as the “undertaking,” or business, that supplies the service.

A careful reading of the FCCPA, though, shows that its consumer-protection provisions govern transactions. They police the terms and conduct of an existing supply relationship. The central grievance here—that Google crawls, indexes, and summarizes news while the publisher receives no payment and, increasingly, no click—involves no service supplied to the publisher and no price.

The press release inadvertently casts news organizations as suppliers of an input. Because the FCCPA regulates how businesses supply services, the FCCPC cannot use it to create a duty to transact or to require payment for an input that a platform uses outside a transaction. 

The underlying complaint concerns bargaining power. Australia and Canada have pursued legislative solutions to this same controversy, with consequences discussed below that should give Nigeria’s National Assembly pause. The NPO should not have routed its grievance through the presidency. That political involvement will cast a shadow over whatever the investigation concludes. 

Pay Up—or Link Out

The FCCPC’s compensation theory rests on a one-way account of value: Publishers supply content, and platforms take it. Yet search engines and social platforms are two-sided intermediaries that connect publishers with readers. They send publishers referral traffic—visitors who follow links from a platform to a news site—that publishers actively pursue through search optimization and distribution teams.

Economic research confirms that platforms also create value through a “market expansion effect”: By helping readers discover news, they increase its total audience. 

When Spain’s 2014 ancillary-copyright law prompted Google to close Google News in the country, economists Joan Calzada and Ricard Gil found that visits to Spanish news sites fell by 8% to 14%. Using individual browsing data, Susan Athey, Markus Mobius, and Jeno Pál found that Google News users’ overall news consumption fell by roughly 20%. Small publishers suffered most, while large outlets barely noticed the shutdown. 

Spain was no outlier. When a contract dispute removed Associated Press content from Google News for seven weeks, Lesley Chiou and Catherine Tucker found that news aggregation increased visits to publishers’ sites. They found no evidence that readers simply scanned headlines or summaries instead of clicking through, as publishers often claim. 

German publishers even lobbied for an opt-in law that allowed them to withhold content from Google News. After those that opted out experienced massive traffic losses, they returned. None ultimately chose to remain outside Google News. 

The same pattern appears in countries that have required platforms to negotiate with publishers. Australia’s bargaining code initially produced deals, but Meta declined to renew agreements worth an estimated 70 million Australian dollars annually in 2024. Meta said news carried little commercial value for its platform. 

Canada’s Online News Act sought to ensure “that dominant platforms compensate news businesses when their content is made available on their services.” Meta responded by removing news from its platform. 

Australia is now proposing a 2.25% charge on large platforms’ Australian revenue to force them back to the bargaining table. The rate would fall to 1.5% for platforms that make deals with publishers. 

As Dirk Auer and Ben Sperry have explained, these media bargaining codes could create “legally sanctioned cartels” that facilitate price fixing and group boycotts. Consumers would likely bear the cost through “paid tiers or increased ad loads.” 

Compelling platforms to pay for links gives them a reason to stop carrying links. The publishers most dependent on referral traffic then suffer the greatest losses. Enforcement becomes extraction, and extraction operates like a tax. If that tax applies only to U.S. companies, it may also violate international trade rules against discriminatory treatment.

The compensation theory also misdiagnoses journalism’s economic troubles. Newspaper decline largely reflects the unbundling of services that newspapers once sold together. Robert Seamans and Feng Zhu found that Craigslist’s entry reduced classified-ad rates at affected U.S. newspapers by more than 20%. Michael Gentzkow showed that online attention can command higher advertising prices than newspaper attention. Advertisers followed audiences to platforms. The display of news had little to do with it. 

Nigeria has little leverage and a history of overestimating it. A platform’s willingness to pay is limited by the profit it earns locally, and Nigeria’s digital-advertising market remains small in dollar terms relative to its population. 

Nigerian regulators have already brought Meta close to the exit. Penalties imposed during 2024 and 2025—including a $220 million FCCPC fine that Meta continues to contest—led the company to warn that it might have to leave the Nigerian market. 

Meta may stay regardless, but the possibility of exit is real. News is among the least costly content for a social platform to drop. The NPO and its members would bear the consequences of a Nigerian news blackout.

The FCCPC has based its expectations on a South African settlement figure inflated by as much as fivefold. It is pursuing firms with smaller local stakes and a demonstrated willingness to withdraw services. Nigeria may think it is following South Africa. It is instead courting the Canadian outcome. 

The High Price of Imagined Payments

Google may negotiate anyway. It has done so when the legal compulsion was genuine, even if contested. In Canada, Google agreed to pay 100 million Canadian dollars annually, indexed to inflation, under the Online News Act. In South Africa, statutory authority gave the SACC leverage, while the parties negotiated the final sum. 

But a payment extracted through an investigation with no legal foundation is not legitimate competition law enforcement. It is coercion that works only while a firm considers the local market worth the toll. It also teaches global companies that their regulatory exposure in Nigeria may depend on political pressure rather than law. 

Google may pay because its search business depends on offering comprehensive results. Meta has already shown that it will remove news instead, and X may make the same choice. The petitioners would then have exchanged real referral traffic for imagined compensation. 

Publishers already have lawful options. If Nigerian news organizations believe AI developers infringed their copyrights, the Copyright Act 2022 allows them to sue, and courts can decide the claims. If unequal bargaining power is the problem, the National Assembly may consider legislation—but it should first study the consequences in Australia and Canada. 

Nigerian news organizations also face the harder task of updating their business models, as some foreign publishers have done. The FCCPC cannot investigate a viable business model into existence.

The post Copy, Paste, Compensate: Nigeria’s Misguided Bid to Make Big Tech Pay for News appeared first on Truth on the Market.

That competition question starts with open-source AI models. These models make their source code, training methods, and trained parameters publicly available under a permissive license, allowing others to use and modify them. As Dirk Auer and I previously wrote, citing Susan Athey, then-chief antitrust economist at the U.S. Department of Justice (DOJ), a few strong open models may be enough to constrain proprietary large language models (LLMs): 

… it is important not to neglect the role that open-source models currently play in fostering innovation and competition. As former DOJ Chief Antitrust Economist Susan Athey pointed out in a recent interview, the AI industry “may be very concentrated, but if you have two or three high quality — and we have to find out what that means, but high enough quality — open models, then that could be enough to constrain the for-profit LLMs.” Open-source models are important because they allow innovative startups to build upon models already trained on large datasets—therefore entering the market without incurring that initial cost. Apparently, there is no lack of open-source models, since companies like xAI, Meta, and Google offer their AI models for free… 

The same reasoning applies to open-weight models. These models make their trained numerical parameters, or “weights,” freely available for download, even if their training data, full code, and methods remain private. The weights encode what a model has learned during training. 

Open-weight models allow startups to build AI products and services without bearing the substantial upfront training costs that well-resourced incumbents can more readily absorb. Open-source and open-weight models therefore offer one reason, among others, for optimism about competition in AI markets. 

That competitive role has made open-weight models a central target in the industry’s new manifesto war. Legitimate concerns about safety and potentially illegal conduct have prompted several private-sector regulatory proposals. The most concrete concern involves Chinese developers’ alleged large-scale distillation of proprietary models, a process in which one model learns from another model’s outputs. The proposals range from mandatory approval before release to targeted restrictions on open-weight and open-source development. 

As Kristian Stout has argued, such restrictions are more likely to create problems than solve them. This post examines the competitive consequences of the most prominent proposals. Several could distort an AI market that has proved more open and competitive than the prevailing regulatory narrative suggests. The manifestos promise safer AI. Their fine print may promise today’s leaders a safer market. 

Let a Thousand Manifestos Bloom

Google DeepMind CEO Demis Hassabis made the first move. He published a widely circulated Substack essay on July 14 calling for “urgent action” as developers approach artificial general intelligence (AGI), a still-hypothetical system capable of performing a broad range of intellectual tasks at or above human levels. Hassabis warned that increasingly autonomous AI systems could pose cybersecurity, nuclear, biological, and other risks. He proposed: 

…a new approach to testing frontier AI model capabilities that is dynamic, adaptable, and rigorous. The US is well positioned, given its economic and technical standing, to take the first step in developing such a framework. It could establish a new Standards Body modelled on a federally overseen public-private partnership or self-regulatory organisation, much like the Financial Industry Regulatory Authority (FINRA), with a board that includes independent leading technical experts and open-source representatives. Funding would need to be substantial and likely mostly come from industry, in order to attract world-class technical talent and provide the necessary compute resources for large-scale testing.

The Standards Body would be responsible for developing assessment protocols and working with appropriate federal agencies and the US National Labs to conduct testing in areas relevant to national security. A model would qualify as ‘Frontier-class’ if it meets certain thresholds on a set of benchmarks determined by the Standards Body and regularly updated to keep pace with evolving AI capabilities. Organisations with ‘Frontier Models’ as defined by those benchmarks would be deemed ‘Frontier Labs’, and be encouraged to adopt best practices, such as publishing model cards with technical details, maintaining strong internal cybersecurity, vetting key personnel, and providing sufficient resourcing for safety and security research, and more.

Hassabis’ proposal does not specifically target open-weight or open-source models. The Standards Body would evaluate frontier-class models “whether they are open or closed,” while exempting models from startups and academic researchers that fall below the frontier threshold.

Formal neutrality does not guarantee equal competitive effects. Even if the body applied its criteria evenhandedly—a considerable assumption when frontier labs would fund and staff it and exercise substantial influence over its benchmarks—the burdens would fall differently across the market. 

Frontier labs are, by definition, the current leaders. Chinese AI developers, U.S. firms that build on open weights and open-weight ecosystems, and other entrants are racing to catch them. Any mechanism that slows progress at the frontier gives an advantage to firms already ahead. It could also burden companies that build on open-source models and serve customers satisfied with models that fall just short of the latest capabilities. Many businesses and consumers need capable, affordable tools rather than the newest model money can buy. 

A model six months behind the frontier can still perform highly sophisticated tasks. Once regulators establish assessments for frontier models, “highly capable” models would present an inviting next target. Government efforts to control the pace of frontier development could thus impair the competition they purport to protect. 

Nvidia CEO Jensen Huang entered the debate on July 24 with an industry letter signed by 25 companies, including Meta, Microsoft, Google, and OpenAI. The letter defended open-weight models as essential to U.S. leadership in AI: 

Open weights expand access to the AI economy. Startups, established businesses, universities, and public institutions can build on advanced models without training one from scratch or paying frontier model prices for every task. Open weights let every organization match the right model to the right job at the right cost, reserving frontier-scale capability for genuine frontier problems and running efficient, specialized models everywhere else. That discipline is what will make AI economically sustainable as its use scales into the billions of everyday tasks. America wins the AI era by diffusing it into the workflows of factories, hospitals, farms, classrooms, and main street businesses.

Open weights also strengthen competition and competition is what keeps the gains of AI broadly shared rather than concentrated in a few hands. By allowing many organizations to build, adapt, and deploy advanced models, open weights create rivalry not only among model developers but across cloud chips, applications, and services. That competition spurs innovation, drives down costs, and distributes the benefits of AI broadly across our economy.

Anthropic CEO Dario Amodei outlined his position in a July 27 blog post. He stressed that “Anthropic has never advocated for a ban on open-weight models,” then called for mandatory safety testing of every sufficiently capable model and tighter controls on the computer chips and model-distillation techniques available to Chinese competitors. The proposal stopped short of a blanket ban on open weights, though only just short. 

Meta CEO Mark Zuckerberg answered on July 28 with a Wall Street Journal op-ed arguing that broadly distributed AI—or “personal superintelligence,” in his formulation—could prevent a few gatekeepers from concentrating power: 

It is surprising that the discourse from many of those who are developing artificial intelligence is so filled with doom. I don’t understand why anyone who believes that AI will eliminate most jobs and much of humanity’s relevance would rush to build that future. The notion that AI is so dangerous that the only safe path is an extreme concentration of power seems dangerous. Historically, hoping that an absolute power will benevolently provide for humanity if sufficiently enlightened hasn’t led to safe or positive outcomes. …

Rather than centralizing this power, we believe that delivering personal superintelligence to everyone is the way to answer this question [how to direct AI]. This has the potential to begin a new era of personal empowerment, in which individuals have greater freedom to pursue their interests and reach their full potential.

Zuckerberg later published an expanded version of the essay. 

More than 1,300 employees of OpenAI, Anthropic, Google DeepMind, Meta, and other frontier AI companies joined the fray on July 28. Their statement, “Pacing the Frontier,” asked the U.S. government to “support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development.” 

The signatories invoked a version of the prisoner’s dilemma, in which each participant’s individually rational choice produces a worse result for the group. They argued that “each company—and country—is under intense competitive pressure not to unilaterally slow that acceleration. And today, the world lacks the technical and governance tools to deliberately pace frontier-wide progress.” 

Axios called the exchange a “manifesto war.” The label fits. Most coverage has treated the dispute as a fight over AI safety, while paying far less attention to what these manifestos reveal about competition—or what their regulatory proposals would do to it. 

Baptists, Bootleggers, and Benchmarks  

The regulatory-economics literature offers a useful way to read the manifesto war. Bruce Yandle’s “Bootleggers and Baptists” theory holds that durable regulation often attracts two coalitions: “Baptists,” who support it for moral or public-interest reasons, and “bootleggers,” who stand to benefit financially. 

The theory does not assume bad faith. The Baptists may be entirely sincere, and those with a commercial stake in regulation may genuinely share their concerns. AI labs and their employees may honestly fear the risks they describe while supporting rules that also protect the labs’ commercial interests. Advocates of open weights have financial interests of their own. Yandle’s point is that regulation often protects the bootleggers more than the Baptists’ stated goals require. 

The safety concerns raised by the “Pacing the Frontier” signatories deserve serious consideration. These include recursively self-improving systems, which can enhance their own capabilities, and rogue systems that escape evaluation sandboxes, the isolated environments researchers use for testing. Governing systems that can accelerate their own development presents a real problem. The researchers raising these concerns include some of the field’s most technically capable people. 

This post does not question their sincerity. It asks whether the proposed regulatory frameworks would affect competition and investment beyond what their safety rationale requires—and whether those effects would predictably favor firms already at the frontier. 

The proposals’ timing also deserves attention. The “Pacing the Frontier” letter appeared during the same week that Moonshot AI—a Chinese startup that develops the Kimi family of AI models—released Kimi K3, an open-weight model that observers rated alongside the best publicly available frontier models of early 2026, at a fraction of their cost. The White House had disclosed days earlier that Moonshot allegedly built Kimi K3 by distilling Anthropic’s Fable model on an industrial scale, using export-controlled Nvidia hardware obtained through Thailand. 

Chinese open-weight models such as DeepSeek, Qwen, ERNIE, and Kimi have gone from marginal players to serious global competitors in less than two years. As I noted in a previous post, the performance gap between the best American and Chinese AI models has fallen to 2.7%, compared with 17.5 to 31.6 percentage points in May 2023, according to the Stanford AI Index 2026. 

Alibaba’s Qwen family has surpassed 700 million cumulative downloads on Hugging Face and produced more than 180,000 derivative models, making it the world’s most widely distributed open-source AI system. DeepSeek V4 Pro matches leading American models on most benchmarks for autonomous task performance while charging 97% less for its outputs than GPT-5.5. 

Open-weight models now pose a genuine threat to the proprietary frontier-lab business model, even though some frontier labs also release open-weight models. Proposals to slow development, restrict access, or impose compliance costs on all sufficiently capable models could weaken the competitive pressure facing the companies that support those proposals, regardless of their safety rationale. 

Self-regulation poses a particular risk. As the Organisation for Economic Co-operation and Development (OECD) has warned, it “may also lead to firms co-ordinating their activities and engaging in cartel-like behavior (e.g., price-fixing) and creating barriers to entry for new firms.” 

This analysis does not resolve the policy question. It does, however, suggest that policymakers should account for competitive effects before turning these manifestos into law. 

Pacing the Frontier, Protecting the Front-Runners

The “Pacing the Frontier” letter describes a prisoner’s dilemma: No company can safely slow down while its rivals race ahead, so restraint requires a government-backed international mechanism that makes compliance mutual and verifiable. While the argument has intuitive appeal, it simultaneously understates the private incentives that already push companies toward safety and overstates the need for regulation before a model’s release. 

Frontier AI companies already face substantial reputational, commercial, and legal consequences when their models cause harm. A company whose model escapes an evaluation environment, facilitates large-scale fraud, or causes a documented security breach risks losing customers and employees while inviting regulatory scrutiny and civil lawsuits. No new regulatory framework is needed to create those consequences. 

These constraints are real. OpenAI and Anthropic both recently disclosed incidents in which advanced models behaved unexpectedly during cybersecurity evaluations. Both companies investigated, disclosed, and addressed the incidents without a government mandate. As International Center for Law & Economics (ICLE) scholars have argued, strict-liability or product-liability rules that allow victims to recover for injuries caused by AI systems could strengthen these incentives without imposing the competitive costs of pre-release approval. 

The deeper problem with the manifesto war’s proposals is their poor fit with the harms they cite. As Kristian Stout has pointed out, the conduct that supposedly triggered the crisis—unauthorized distillation of proprietary models, smuggling export-controlled hardware, and fraudulently accessing application programming interfaces (APIs)—already violates existing law. 

Authorities can prosecute fraud and export-control violations. Intellectual-property and unfair-competition laws can address unauthorized distillation. If existing law already covers the alleged conduct, enforcement should target the harm and the wrongdoers directly rather than burdening an entire class of technology. Bad actors will not become law-abiding because regulators impose new compliance duties on legitimate developers. They will find other methods, while law-abiding firms bear the costs. 

Policymakers considering genuinely new regulatory tools should apply the OECD competition-assessment standard. They should ask whether less restrictive means could achieve the same safety objective and whether the proposed rules would create entry barriers beyond what that objective requires. The World Bank has identified regulatory barriers to entry as one of the principal ways governments inadvertently restrict competition. 

Sound AI governance can protect safety and competition at the same time. Each proposal should face three questions: Who bears the costs? Which harms does it address? And could a less restrictive alternative achieve the same result? The evidence supporting a rule should match the burden that rule would impose. 

Manifestos are cheap. Entry barriers are not.

The post Open Weights, Closed Ranks: The AI Manifesto War appeared first on Truth on the Market.

The theory fits within the law & economics tradition because it offers an economic account of political behavior. It focuses on the incentives facing every leader and treats public law as endogenous to those incentives—that is, as the product of survival strategies operating within a given institutional framework. 

Bueno de Mesquita et al.’s formal model has broad implications for public finance, yet they do not develop those implications into a positive fiscal theory, meaning a theory that explains how fiscal policy works rather than how it should work. This piece takes up that task, analyzing tax and spending laws as survival mechanisms that determine who is taxed and who receives benefits to keep the leader in power.

Political Survival Is a Numbers Game

In every political system, two groups within the total population (N) constrain the leader’s political survival. The selectorate (S) consists of everyone with a legal say in choosing the leader. In the United States, it includes all registered voters. The winning coalition (W) is the subset of S whose support the leader actually needs to retain office—the “essentials,” as Bruce Bueno de Mesquita and Alastair Smith later called them in “The Dictator’s Handbook” (2011). 

W might consist of a majority coalition of voters. In a U.S. presidential election, for example, W includes the minimum geographic distribution of voters needed to win the Electoral College, supplemented in practice by party elites and major donors. Yet W need not exercise democratic control. In plutocratic or autocratic states, the support of a small coalition of business or military leaders may be enough to retain power. 

Bueno de Mesquita et al. argue that leaders everywhere face a budget constraint. They must use a finite pool of revenue to provide members of W with economic benefits or risk replacement. The difference between democratic and authoritarian systems lies in the ratio of W to S, rather than in their leaders’ motives. Leaders in both systems are self-interested survivors. 

In democracies, W typically constitutes a large fraction of S—often a majority of voters in parliamentary systems, though considerably less in plurality and Electoral College systems. In authoritarian regimes, W is only a small share of S. The leader needs to keep only a small inner circle loyal to remain in power. As the authors put it: 

Polities can be characterized by more-or-less formalized mechanisms that allow some people to have the prospect of gaining access to valued goods dispensed by government while excluding others from even the chance of such access. (Bueno de Mesquita et al. 2003, 42)

Small-W regimes themselves exist along a spectrum. At one extreme, pure autocracies, such as monarchies and military juntas, combine a small selectorate with a minuscule winning coalition. At the other, institutionalized authoritarian regimes combine a large nominal selectorate with a restricted winning coalition. 

In the Soviet Union (USSR), most adults had nominal voting rights, so S represented a large share of N. Yet the essentials—those who wielded real power through the Communist Party hierarchy—comprised only about 3% to 5% of the population, making the operative W/S ratio tiny. As the model predicts, Soviet leaders retained power by granting privileges and private goods to these essentials, who functioned as a distinct caste within communist society.

In liberal democracies, S likewise represents a large share of N, but W constitutes a much larger share of S. Under majority rule, W can approach half of S. Even in plurality and Electoral College systems, the ratio remains roughly an order of magnitude greater than the autocratic baseline. A U.S. president who provides benefits to only 5% of the population and stops producing public goods would not win reelection. 

The large W/S ratio explains why democratic leaders provide benefits to a broad range of citizens and receive relatively modest compensation compared with their authoritarian counterparts. Dictators, by contrast, can direct most of the public budget to roughly 5% of their cronies through private payments, privileges, and immunities while providing little for the general population.

The W/S ratio thus shapes political behavior through what the authors call the loyalty norm. When that ratio is small—e.g., 3%—any given member of S has little chance of entering the winning coalition. Current members therefore have strong incentives to remain loyal to the incumbent. Defection will likely mean permanent exclusion from the flow of private benefits. 

Spend Broadly or Bribe Wisely

Putting the “law” in “law & economics,” the model’s central application concerns tax-and-spending laws. Bueno de Mesquita et al. explain and predict how governments extract resources from the productive population and distribute them as public goods, private goods for the essentials, or discretionary funds for the leader.

In systems with a large winning coalition (W), as in mature democracies, leaders must satisfy a broad share of the population to retain power. This alignment between leaders’ incentives and the interests of much of the population distinguishes democracies. The cost of providing private goods rises with the size of W, so supplying private benefits to millions of people becomes prohibitively expensive under any plausible budget constraint. Large-coalition leaders therefore turn to public goods. 

Public goods are nonrivalrous and nonexcludable. One person’s use does not reduce their availability to others, and the government cannot easily limit their benefits to particular people. They can therefore benefit the entire population without costs rising in proportion to the coalition’s size. 

Switzerland offers a striking example. Its government provides high-quality public goods and infrastructure while maintaining a comparatively low tax-to-gross domestic product (GDP) ratio of roughly 27%, among the lowest in the Organisation for Economic Co-operation and Development (OECD). Many other large-coalition democracies, including Sweden, France, and Germany, have considerably higher ratios. The basic model does little to explain this variation among democracies. 

Small-W regimes create different incentives. When a leader depends on only a few essentials—as in a military junta, a one-party state, or an autocracy operating behind rigged elections—investment in public goods is politically inefficient because the benefits extend far beyond the coalition that keeps the leader in power. The leader instead directs the budget toward private goods that can be limited to W, including bribes, direct transfers, and legal immunities. 

The loyalty norm makes this strategy cheap. Coalition members know that if they defect, they have little chance of entering a challenger’s W because the challenger can choose replacements from the much larger selectorate (S). The essentials therefore accept smaller private payouts than they would if defection offered a realistic path into a new coalition, leaving the leader with more discretionary funds. 

The state thus favors private corruption over public investment because political survival rewards that choice.

Tax Policy as a Survival Strategy

The logic of political survival applies to taxation as much as to spending. For tax policy, the model predicts that large-W systems will tend toward moderate rates applied to a broad base, typically income or consumption. In other words, the government taxes many people or transactions at relatively modest rates. This prediction tracks a general observation in comparative tax design (Baskaran 2014; Genschel & Seelkopf 2016). 

The model treats this design as a strategic necessity. Excessively high tax rates, or punitive taxes targeting people outside the coalition, would create an “expropriatory” environment that encourages people to favor leisure over productive labor (Bueno de Mesquita & Smith 2011). The resulting decline in economic activity would reduce the revenue available to fund the public goods that keep a large share of the population satisfied. Large-W systems therefore tend toward a high-growth, low-tax equilibrium because the leader’s tenure depends on maximizing total economic output rather than extracting more revenue from a shrinking economy.

Small-W systems create different incentives. Democratic regimes must generate a surplus. Autocratic regimes extract that surplus and distribute it to the winning coalition (Albertus 2015). Because the productive population overlaps little with the coalition, the autocrat and the essentials bear little of the deadweight loss caused by high taxation—that is, the economic harm that taxes cause by discouraging work, investment, or production. Selectorate Theory therefore predicts high tax rates in small-W regimes and helps explain why average rates in dictatorships are often much higher than in democracies. 

In small-W systems, taxation transfers resources from N to W. It enriches the essentials and, just as important, impoverishes those outside the coalition. Both effects help the leader survive. Transfers buy the essentials’ loyalty, while reducing outsiders’ wealth weakens potential challengers. Wealthy outsiders might otherwise have the resources to finance a rival coalition.

Tax law also becomes a discretionary tool for rewarding loyalists and disabling challengers. Coalition members receive tax exemptions and favorable audit treatment (Wang 2015; Blaydes 2011), while opponents and dissidents face higher rates and selective audits (Leon 2014). Small-W regimes thus replace the large-W structure of broad tax bases and uniform rates with targeted taxes and selective enforcement (Dodlova and Lucas 2021). 

Tax Law’s Democratic Guardrails

The book explains how political leaders survive. Institutional rules determine the size of W relative to S—the share of the selectorate whose support the leader must secure with economic benefits. Whether W is large or small shapes the resulting tax-and-spending system. 

In democratic regimes, W constitutes such a large share of S that public goods offer the most efficient means of retaining power. Governments fund those goods through moderate tax rates applied to a broad base, which limits deadweight loss and sustains the economic growth on which the leader’s tenure depends. 

When W is small, public goods do little to secure the leader’s survival. Leaders instead direct bribes and legal immunities to the narrow political caste that rules society. Fiscal policy becomes another tool for staying in power. It redistributes wealth from the public to the political class, rewards cronies with favorable exemptions and friendly audits, and punishes dissidents through higher rates and selective enforcement.

The model leaves legal scholars with a normative question. If tax law is part of the authoritarian playbook, how should the law constrain the power to tax to prevent further democratic backsliding? 

Further Reading

• Bruce Bueno de Mesquita and Alastair Smith, “The Dictator’s Handbook: Why Bad Behavior Is Almost Always Good Politics,” PublicAffairs (2011).
• Mary E. Gallagher and Jonathan K. Hanson. “Power Tool or Dull Blade? Selectorate Theory for Autocracies,” Annual Review of Political Science, Vol. 18 (2015).
• Yuhua Wang, “Tying the Autocrat’s Hands: The Rise of the Rule of Law in China, Cambridge University Press (2015).
• Steven Levitsky and Daniel Ziblatt, “How Democracies Die,” Crown (2018).
• Gerald M. Easter, “Capital, Coercion, and Postcommunist States,” Cornell University Press (2012).
• Mancur Olson, “Dictatorship, Democracy, and Development,” American Political Science Review Vol. 87, No. 3 (1993).
• Charles Delmotte, “Equality Before Tax Law,” Washington & Lee Law Review Vol. 83 (2026).

The post ‘The Logic of Political Survival,’ by Bruce Bueno de Mesquita, Alastair Smith, Randolph M. Siverson & James D. Morrow appeared first on Truth on the Market.

On Aug. 3, the Australian government finalized legislation establishing the News Bargaining Incentive (NBI). The government first proposed the NBI in December 2024, then put it on hold after calling a general election a few weeks later.

The NBI seeks to encourage digital platforms to enter into commercial agreements supporting Australian journalism. Large platforms that decline to enter into or renew qualifying agreements must pay a charge based on advertising revenue attributable to the Australian market. Platforms with qualifying agreements may reduce their liability through a nonrefundable offset.

The NBI’s unusual reach is deliberate. It applies to large platforms that provide significant social-media or search services, regardless of whether they carry news content. According to the government, this approach closes a major gap in Australia’s existing bargaining code, which allows platforms to avoid payment obligations by removing news.

The government has openly expressed frustration that the code has failed to secure the desired transfer of revenue from large platforms to news publishers, despite its mandatory bargaining process and binding final-offer arbitration. The NBI responds by compelling payment even when a platform makes no use of news content.

This latest change to Australia’s legal framework warrants attention for two main reasons.

The ‘Success’ That Needed a Sequel

First, Australia’s model has become influential in the long-running dispute between digital platforms and publishers over the online use of news content. That dispute has driven policy interventions in several jurisdictions in recent years, and other governments have looked to Australia for a possible solution.

Australia’s bargaining code grew from an acknowledgment of the limits of Europe’s copyright-based approach. Digital platforms’ use of article headlines and short excerpts, known as snippets, was unlikely to infringe copyright. Even if copyright protected that material, fair-dealing exceptions might permit its use. The Australian Competition and Consumer Commission (ACCC) therefore recommended a code of conduct that would encourage negotiations and allow the ACCC to address potential imbalances in bargaining power.

The ACCC initially designed the code as a form of industry self-regulation. Digital platforms would have nine months to develop it. If they failed to submit an acceptable code to the Australian Communications and Media Authority (ACMA), ACMA could impose a mandatory standard.

The ACCC then began working with Facebook, Google, and news-media businesses on voluntary codes. It soon advised the government that this process was highly unlikely to resolve the central dispute over payment for content. The government responded by directing the ACCC to develop a mandatory code, prompting intense debate.

The code’s central feature is binding final-offer arbitration. It requires the parties to negotiate in good faith. If they cannot agree, an arbitration panel resolves the dispute. Each party submits its final proposal for the amount the platform should pay, and the panel must choose one. If both proposals conflict with the public interest, the panel may amend the more reasonable offer. This process serves as a backstop to voluntary negotiations. It seeks to produce faster decisions, deter inflated demands, and encourage both parties to make reasonable offers.

After the code’s first year, the Australian government’s 2022 review declared it a success. The government reported that digital platforms had reached more than 30 commercial agreements with a broad cross section of Australian news businesses. In its view, the parties would have been highly unlikely to reach those agreements without the code.

The review also identified serious concerns. Smaller news businesses struggled to secure agreements, while the secrecy surrounding completed deals made it difficult to assess whether the code had achieved its objectives.

Platforms also pursued sharply different strategies. Meta showed little interest in negotiating, consistent with its public claim that news plays a minor role in its business model. Google engaged more readily, though it directed most agreements toward its own strategic content initiatives.

The government’s decision to revisit even the mandatory code suggests that its 2022 declaration of success came too soon.

Free Riding Without the Ride

Second, the Australian NBI exposes a deeper problem with the free-riding narrative that has shaped the debate over publishers and digital platforms. As I argued in a recent paper, the rationale for the NBI conflicts with that narrative.

Despite differences in legal design and institutional context, laws governing the online use of news tend to follow a common script. Policymakers first acknowledge that publishers and online intermediaries, such as search engines and social-media platforms, benefit from one another. They also recognize that these platforms have become major gateways to news and that publishers rely heavily on the traffic they provide.

But these laws generally omit a crucial comparison. They do not determine whether the benefits of referral traffic offset the losses from fewer direct visits to publishers’ websites. Policymakers proceed from the premise that platforms free ride on publishers’ investments in producing news and have helped cause the publishing industry’s structural decline. That premise persists despite empirical evidence casting doubt on it. Policymakers treat free riding as an accepted fact without first demonstrating it.

The debate then turns to bargaining power. Publishers depend on certain platforms, which policymakers portray as unavoidable trading partners. Yet the relevant economic effects remain uncertain. Referral traffic may expand publishers’ audiences, while headlines and snippets may reduce direct visits by giving users enough information without requiring a click. Policymakers generally presume that the second effect outweighs the first.

Publishers may rely heavily on platforms to reach new readers. Platforms may have little reciprocal dependence because news content plays only a minor role in their core businesses. That reasoning weakens the free-riding claim by acknowledging that platforms do not need news content. Publishers could also gain more from referral traffic than they lose through competition with platforms, regardless of any difference in bargaining power. If so, their claim to compensation becomes difficult to sustain.

Policymakers nevertheless infer harm from a commercial relationship they regard as unfair. They then require large platforms to negotiate over the use of press publications and pay publishers “fair remuneration,” departing from their initial acknowledgment that both sides benefit.

The free-riding narrative has therefore served as a rhetorical justification for a mandated transfer of revenue. Policymakers continue to pursue that transfer without clear evidence that platform business models have caused publishers economic harm. Michael Geist calls Australian-style mandatory negotiation systems a “shakedown subsidy model” because they require platforms to compensate publishers without a clearly established basis for payment.

The NBI weakens the free-riding claim further by requiring platforms to pay even when they use no news content—when, in other words, there is nothing to free ride on.

Another Chance to Learn the Same Lesson

For at least two decades, policymakers around the world have called for journalism to reinvent itself. The industry has struggled to do so. Revenue-transfer schemes have failed to help publishers develop sustainable commercial models, while protection from competition and continued subsidies leave the structural causes of the industry’s decline intact. Without substantial changes to publishers’ businesses, the next decade will bring more of the same unresolved debates over the fate of traditional media.

Technological change is already overtaking those debates. While policymakers continue searching for ways to make digital platforms pay for snippets, artificial intelligence (AI) has changed how users find information online. Features such as AI Overviews provide synthesized answers within search results, accompanied by links to the underlying sources.

Publishers fear that these summaries will reduce website visits and the advertising or subscription revenue those visits generate. AI Overviews have therefore become the latest perceived threat to publishers’ business models and the industry’s financial sustainability.

Policymakers have had two decades to learn the lesson. Their next response will show whether they finally have.

The post Much Ado About No News: Australia’s Latest Plan to Make Platforms Pay appeared first on Truth on the Market.

At its August 6 open meeting, the Federal Communications Commission (FCC) is scheduled to vote on a notice of proposed rulemaking (NPRM) titled “Unleashing Unlicensed Spectrum for Direct-to-Device.” This proceeding will examine whether devices operating across more than 225 megahertz of unlicensed spectrum in the 902-928 MHz, 2400-2483.5 MHz, and 5725-5850 MHz bands should be authorized for satellite communications. The item represents a commendable step in the Commission’s already-productive spectrum work. 

Firms have entered the direct-to-device (D2D) market with great speed since the Commission’s 2024 “supplemental coverage from space” framework first authorized satellites to serve ordinary smartphones over spectrum licensed to terrestrial carriers. Satellite operators have also sought to acquire their own standalone spectrum for mobile service. SpaceX acquired 65 megahertz of spectrum from EchoStar for $17 billion, while Amazon and Rocket Lab struck deals collectively approaching $20 billion to acquire Globalstar and Iridium, respectively, and SpaceMobile leased spectrum from Ligado in a deal values at about $120 million. FCC Chairman Brendan Carr counts nearly $50 billion in D2D spectrum transactions over the past year alone. SpaceX has launched approximately 650 satellites for such use, with authorization to deploy as many as 7,500 satellites. And just last week, Amazon’s Leo unit asked the Commission for authority to deploy a constellation of more than 5,100 satellites dedicated to direct-to-phone service. The Commission’s willingness to modernize legacy spectrum rules has been an important enabler of this investment.

The draft NPRM extends that approach to unlicensed spectrum. The item proposes to add Earth-to-space allocations in the three bands where Part 15 rules already permit higher-power unlicensed operation, and therefore require no change to existing power limits, out-of-band emission limits, or other technical rules. All satellite operations would proceed on a strictly non-interference basis: they could not cause harmful interference to authorized services and would have to accept any interference received.

And because devices would transmit at the same power levels they use today, the Commission tentatively concludes that the interference environment would not change.

The decision to examine low-, mid-, and high-band unlicensed spectrum together is particularly notable. Each band presents a distinct engineering tradeoff. The 902-928 MHz band offers superior propagation and building penetration, making it well-suited to low-power, wide-area uplinks from Internet-of-Things sensors and trackers. The 2.4 GHz band is globally harmonized and home to billions of Wi-Fi and Bluetooth radios, which is the installed base that Hubble’s operations already leverage. 

And while the item’s affirmative proposal is limited to the Earth-to-space direction, the NPRM seeks comment on space-to-Earth operations as well. That particular item is focused principally in the 5725-5850 MHz band, whose shorter wavelengths permit smaller, more directional satellite antennas that could confine a downlink beam’s footprint on the ground. 

Which combination of these bands ultimately proves suitable for satellite links is precisely the kind of question a rulemaking record exists to answer. Spectrum suitable for D2D service remains scarce relative to projected demand. The surest way to relieve the spectrum constraint is to keep multiple options open.

Indeed, the one respect in which the item could aim higher is precisely here. As drafted, the NPRM proposes allocations only for uplinks, while downlinks are considered only as questions for further consideration. An uplink-only framework would support telemetry-style applications such as sensors, trackers, and asset monitoring, whereas two-way service requires downlink capacity. Without it, D2D operators would continue to face the very bottlenecks this proceeding could help relieve: acquiring scarce mobile-satellite spectrum or leasing rights from terrestrial carriers. 

Ideally, the rules that ultimately emerge from this proceeding would therefore include space-to-Earth allocations as well, subject to the power-flux-density and beam-footprint limits on which the Commission has already sought comment. 

The record does contain some dissent. For instance, NextNav, a licensee in portions of the lower 900 MHz band, has pressed the Commission to drop the 902-928 MHz band from the item entirely. The context for that request is the company’s own pending 2024 petition for rulemaking, which asks the FCC to reconfigure the band around a terrestrial positioning, navigation, and timing (PNT) network that NextNav would operate alongside its own 5G service. That petition has drawn opposition from a broad coalition of incumbent users that includes amateur radio operators, utilities, and the security industry, as well as from analysts skeptical of its technical claims. 

Whatever the ultimate merits of the NextNav petition, it supplies no reason to narrow this proceeding. An NPRM decides nothing but only asks questions and gathers evidence. Removing the 900 MHz band before a single comment has been filed would effectively prejudge the record in deference to one licensee’s preferred business model. 

This is a familiar dynamic in telecom policy that Thomas Hazlett has documented in earlier spectrum fights, where incumbents use procedural leverage to forestall competing uses. Within the satellite arena, a similar dynamic unfolded in the 2 GHz band, with DISH opposing SpaceX’s application to use the band for D2D service before it was even accepted for filing. This obstructionism defeats the purpose of the regulatory process. If concerns about coexistence are well-founded, the rulemaking record is exactly the place to substantiate them.

This is not a dry technical matter, ultimately, as it has major implications for closing the digital divide. For consumers in rural and remote areas beyond the reach of both cellular networks and terrestrial Wi-Fi, satellite connectivity in unlicensed bands would offer an additional path to coverage using the billions of inexpensive, unlicensed devices already deployed. And in emergencies that redundancy can be lifesaving. 

The FCC should seek comment on all three bands and decide the 900 MHz question on the resulting technical record. Excluding it at the NPRM stage would deny commenters the chance to test coexistence claims before the Commission has developed the evidence needed to resolve them.

The post The FCC Is Right To Keep Every Band on the Table for Direct-to-Device appeared first on Truth on the Market.

The immediate dispute revives a familiar argument. Brussels says it is policing anticompetitive conduct under the Digital Markets Act (DMA). Washington says Europe is taxing American companies. After two years of enforcement, the second account has become much harder to dismiss—which makes the Trump administration’s escalation worth taking seriously.  

Last week, the European Commission fined Google €890 million under the Digital Markets Act. It imposed €460 million for favoring Google’s own shopping, hotel, transportation, and sports results in Google Search, and €430 million for restricting developers from directing Google Play users to cheaper purchasing options. The penalty was Google’s first under the law and the largest imposed under it to date. 

The next day, Trump responded on Truth Social. He announced that the United States would “immediately” open a Section 301 investigation into what he called Europe’s “robbing” of American companies. The United States, he wrote, is not Europe’s “PIGGYBANK.” He predicted that the penalties would be “entirely reversed” and followed by a substantial tariff. 

Some of Trump’s particulars do not survive contact with the record. The $15 billion attributed to Apple appears to include the Commission’s €13 billion Irish state-aid recovery order in a tally of “fines,” though it was not a fine. Nor can a Section 301 investigation “reverse” a legally binding Commission decision. Only the European Union courts can do that. 

But the faulty arithmetic distracts from the two questions that matter: Is the underlying grievance well-founded? And, if so, is trade retaliation a defensible response? 

On the first question, the evidence has become considerably harder to dismiss. On the second, the answer is probably yes—but for reasons rooted less in trade policy than in the political economy of European regulation. 

Europe’s Industrial Policy in Antitrust Clothing

When Geoffrey Manne and I examined European Union competition enforcement in 2019, we found that U.S. firms paid vastly larger fines than European ones. Still, we concluded that the disparity largely reflected the sectors the European Commission targeted and its sales-based method for calculating penalties—not deliberate discrimination. After the DMA’s enactment and enforcement, that conclusion is harder to sustain.

The design problem is familiar. The DMA’s thresholds capture firms based on aggregate size, not demonstrated market power. Applying those arbitrary metrics, the Commission has designated seven gatekeepers. Five are American, while only one (Booking.com) is European. That imbalance is no accident. When lawmakers adopted the DMA, Andreas Schwad—member of the European Parliament for Germany—said it should focus on “the top five” companies while avoiding European rivals.  

Enforcement has made the pattern starker. Every DMA noncompliance penalty so far—€500 million against Apple, €200 million against Meta, and now €890 million against Google—has fallen on an American company. 

The Commission’s specification decisions are even more revealing than the fines. On July 16, it told Google precisely how to satisfy two DMA obligations. Google must open 11 Android system-level features to rival artificial intelligence (AI) assistants. Beginning in January 2027, it must also share anonymized Search-ranking, query, click, and view data with competing search engines and AI chatbots on fair, reasonable, and nondiscriminatory (FRAND) terms. 

These requirements go far beyond ordinary prohibitions on anticompetitive conduct. They are product-design mandates that de facto transfer parts of Google’s property to its competitors. 

Nor is the Commission coy about the destination. In its first DMA review, it candidly conceded: 

The DMA contributes to the EU’s competitiveness, innovation and technological sovereignty objectives by addressing structural bottlenecks in CPSs that may hinder market entry and scaling. By promoting contestable and fair digital markets, it complements efforts to close Europe’s scale-up gap in line with the EU Startup and Scale-up Strategy. It facilitates competition through obligations on interoperability, data portability, access to business user data, etc. and through restrictions on self-preferencing… [I]t forms part of a coherent policy framework aimed at strengthening Europe’s innovation capacity and reducing strategic dependencies.

Put plainly, the objective is digital sovereignty. The Commission is trying to bolster European rivals by redesigning—and partly expropriating—American technology companies. It is therefore no surprise that the same review identified cloud computing and AI as the next frontiers of DMA enforcement. The European Union’s June Technological Sovereignty Package, with its cloud-eligibility tiers and procurement preferences, pursues the same goal. 

The Commission’s own documents thus frame the DMA as an industrial-policy instrument. It aims to reshape the internet according to European policymakers’ preferences and for the benefit of the firms they want to succeed. 

To be fair, the DMA also binds Booking, and several of its loudest beneficiaries—including Epic and, increasingly, OpenAI—are American. The point is not that the law amounts to naked protectionism. It is that a facially neutral regime, administered by an institution facing the incentives described below, has produced a systematically one-sided result. 

The costs also extend beyond U.S. shareholders and European consumers. Compliance consumes one of the scarcest resources at frontier technology companies: senior engineers’ time. Meta reports devoting roughly 600,000 engineering hours to DMA compliance. Google says it assigned about 3,000 engineers full time for two years to comply with a single article. The Commission’s impact assessment, by contrast, projected annual compliance costs of about €1.41 million per platform—a figure that now looks almost quaint. 

Every hour spent rebuilding a codebase for Brussels is an hour not spent building something else. Google’s response to last week’s decision made the engineering and consumer costs explicit. Compliance, the company says, will require removing live hotel, flight, and restaurant prices and availability from European search results, while rolling back Play Store safeguards. American users may not pay the fines directly, but they still pay for the engineering talent diverted from improving the products they use. 

Making Brussels Feel the Cost

The deeper problem is that European policymakers pay almost no political price for enforcing the DMA more aggressively. The benefits of DMA enforcement flow to organized, well-represented constituencies: firms seeking access to rivals’ platforms, publishers, telecom incumbents, and would-be national champions. The costs are diffuse, delayed, and largely shifted onto foreign shareholders and 450 million consumers across 27 member states, who experience product degradation simply as products getting “worse.” 

The fiscal incentives point in the same direction. Fines flow into the general EU budget and reduce member states’ gross-national-income contributions. That gives governments revenue without the political unpleasantness of raising taxes. 

The contrast is telling. When the Commission blocked the Siemens-Alstom merger, Paris and Berlin issued a joint manifesto—and successfully demanded changes to the merger rules. When the Commission fines Apple, nobody blocks a motorway. 

The result is a policy that is nearly free at the point of production. European officials intend to keep it that way. They describe the DMA as settled law and refuse to treat it as a bargaining chip in trade talks, despite the August 2025 joint framework in which the EU and United States agreed to address unjustified digital-trade barriers. 

If the force driving enforcement is a public-choice problem—policymakers can impose the costs because almost nobody at home feels them directly—then the only response likely to work is one that creates a European constituency with something to lose. 

That is what targeted tariffs can do, which is why the choice of target matters far more than the headline rate. French and Italian wine, agricultural products, German automobiles, and pharmaceuticals are obvious candidates. These sectors have the organizational muscle to pressure national capitals, which can then pressure Brussels. The EU is already lobbying Washington for tariff relief on wine, spirits, olive oil, and cheese. That tells us where the pressure points are. 

Tariffs are not costless, of course, and escalation could trigger the EU’s Anti-Coercion Instrument, which allows Brussels to retaliate against foreign economic pressure with tariffs, procurement restrictions, and other limits on market access. But the relevant comparison is not tariffs against some frictionless alternative. It is tariffs against the status quo, in which the political cost to European policymakers of discriminatory regulation is roughly zero.

Now imagine the shoe on the other foot. Would the European Union sit quietly while its most successful companies were fined billions abroad under rules calibrated to capture almost exclusively them? It created the Anti-Coercion Instrument in response to considerably less. 

Does the DMA Pass the Section 301 Test?

Section 301(b) of the Trade Act of 1974 permits action against a foreign act, policy, or practice that is “unreasonable or discriminatory” and “burdens or restricts” U.S. commerce. The 1984 amendments define an unreasonable practice as one that is unfair and inequitable, even if it violates no international legal obligation, and a discriminatory practice as one that denies national or most-favored-nation treatment. The statute also expressly contemplates foreign industrial targeting. 

Because the DMA is facially neutral, de jure discrimination will be difficult to prove. The de facto case is much stronger. Designations, fines, and specification decisions have fallen disproportionately—and, in enforcement cases, exclusively—on American firms. Compliance costs exceed the Commission’s estimates by orders of magnitude, while total losses for U.S. companies may reach €114 billion. 

The harder question is credibility. As of this writing, the Office of the U.S. Trade Representative’s (USTR) public docket lists no DMA-related investigation. It is already handling 18 investigations, from forced labor in China and elsewhere to German pharmaceutical pricing. Tariffs arising from the forced-labor investigation also face a challenge before the U.S. Court of International Trade. Whether USTR has the capacity to complete another major Section 301 investigation is therefore far from clear. 

Europe, meanwhile, treats Section 301 as unilateral coercion, which raises the domestic political cost of yielding to it. Turning that confrontation into a tractable dispute will require a narrow, carefully documented case—not another press release with capital letters. 

The DMA has become an instrument of industrial policy whose costs fall overwhelmingly on a small group of foreign firms and consumers around the world, many of whom have no political voice in Europe. That is precisely the kind of problem trade law’s nondiscrimination principle exists to address. 

The goal should remain a negotiated settlement, and success should be measured by whether Europe changes how it applies the DMA—not by how much tariff revenue Washington collects. Trump’s rhetoric may be blunt, but Brussels may not reconsider the cost of DMA enforcement until French wine, German cars, or Italian cheese starts picking up the tab.

The post From Google Fines to French Wine: The Trade Case Against the DMA appeared first on Truth on the Market.

Ireland assumed the rotating presidency of the Council of the European Union from Cyprus for the second half of the year and declined to recycle Cyprus’ flawed draft. Instead, it reopened negotiations over the European Commission’s proposal on pseudonymization, cookie consent, and the use of data for artificial intelligence (AI). 

Serious reform will still face resistance in the European Parliament. A joint draft report from its industry and civil-liberties committees leaves the Commission’s most contested GDPR proposals untouched for now. But the co-rapporteurs’ public statements—and the flood of amendments filed since—show that the legislation remains very much in play. 

The European Data Protection Board (EDPB) has also weighed in. Although it does not make EU law, it exerts considerable influence over national governments. Its new guidelines distinguishing personal from anonymous data accept, for the first time, the Commission’s central point: whether data is personal should depend on the entity holding or using it, not on whether anyone, anywhere, could identify the person concerned. 

The guidelines also show why procedural reform may matter even more. As usual, the EDPB could not quite bring itself to offer guidance useful for much beyond increasing lawyers’ billable hours. 

The Reform Fight Takes Shape

The reform effort is now moving on four fronts, with the Council showing unexpected signs of ambition, Parliament preparing for a fight, the EDPB conceding a key point while muddying the details, and a separate AI measure quietly becoming law. 

The Council Hits Reset

The Cypriot presidency of the Council of the European Union tried to secure agreement among national governments on a common position that could have left the GDPR worse off than no reform at all. One week before its term ended, Cyprus circulated another compromise draft, following the version I discussed previously. 

It failed. Several governments reportedly objected, including Denmark, Germany, Italy, Poland, and Sweden. The precise reasons remain unclear, but reports suggest that these countries thought the draft did too little to simplify the GDPR. If so, there may still be hope for the reform process. 

Ireland, which now holds the rotating Council presidency, chose not to carry over the Cypriot text. Instead, it sent national governments a questionnaire reopening debate over cookie exemptions, pseudonymization, AI data processing, and reductions in compliance burdens. 

Parliament Prepares for Battle

The European Parliament’s lead negotiators published their draft report in late June. Two committees share responsibility for the legislation: the Committee on Industry, Research and Energy, led by Aura Salla of the European People’s Party, and the Committee on Civil Liberties, Justice and Home Affairs, led by Marina Kaljurand of the Socialists and Democrats. The Legal Affairs Committee and the Internal Market and Consumer Protection Committee also have advisory roles and have produced draft opinions. 

The joint report reflects what Salla and Kaljurand could agree on quickly. It therefore leaves the most contentious questions untouched, including the European Commission’s proposed clarification of “personal data.” 

That restraint will not last. Parliament members have already filed a large number of amendments, and both Salla and Kaljurand have indicated that they will seek changes on issues omitted from the draft, including the definition of personal data. 

The EDPB Concedes—Then Complicates

The EDPB has published new guidelines on anonymization that bear directly on the Commission’s proposal. Most notably, it finally accepted that whether information counts as anonymous—and therefore falls outside the GDPR—depends on the position of the entity processing it. This is the entity-relative view.

That is broadly the clarification the Commission proposes for Article 4, which defines personal data. But the EDPB wrapped that central point in broader, vaguer qualifications than the Commission likely would have adopted. I will return to those complications shortly.  

The ‘AI Omnibus’ Takes Effect

Meanwhile, a less controversial companion measure concerning the EU AI Act entered into force this week. This “AI Omnibus” postponed several compliance deadlines, prohibited AI-generated child sexual-abuse material and nonconsensual intimate imagery, and expanded the legal basis for using sensitive personal data to detect bias in AI systems. 

Where the GDPR Reform Fight Gets Real

The next stage of the fight will turn less on abstract definitions than on whether the final law gives businesses rules they can actually use. The debate now centers on three familiar trouble spots: what counts as personal data, how cookie consent should work, and whether the GDPR will leave room for AI development. 

The Definition Fight Is Mostly Symbolic 

Reports from the final weeks of negotiations under the Cypriot presidency suggest that several national governments oppose the European Commission’s proposed clarification of “personal data” in Article 4 of the GDPR. The proposal also faces resistance in the European Parliament. 

I support the clarification, but its practical importance is easy to overstate. The EDPB has now formally accepted the entity-relative view: whether information counts as personal data depends on the position of the entity processing it. 

Some will argue that any amendment to Article 4 should include qualifications like those in the EDPB’s anonymization guidelines. Others will say that, because the EDPB has already accepted the entity-relative view, no legislative change is needed. 

That argument gives the EDPB too much credit. Its guidelines offer little operational clarity and often seem designed to make implementation difficult unless one simply assumes that all data is always personal. The more consequential reforms are therefore procedural: the Commission’s proposal would let the Commission, rather than the EDPB, adopt legally binding implementing acts that define concepts such as pseudonymization and personal data. 

Those powers should be stronger. As drafted, compliance with a Commission implementing act would count only as “an element” in the legal analysis. It should instead create a robust presumption of compliance. 

Guidance That Actually Guides

What matters most is whether organizations handling EU data can determine which concrete safeguards place information outside the GDPR because it no longer qualifies as personal data. The EDPB’s anonymization guidelines once again show that the board is institutionally ill-suited to provide that kind of practical direction. 

Substance alone is not enough. The guidance must also carry enough legal force to create a strong presumption that those who follow it are acting lawfully. Polish member of the European Parliament Piotr Müller has proposed an amendment moving in that direction—Amendment 398. 

The joint industry and civil-liberties committee draft does not yet revise the relevant provision, Article 41a. Its broader thrust, along with many of the amendments filed in Parliament, instead points toward preserving or even expanding the EDPB’s role. Brussels may yet respond to unworkable guidance by giving its authors more authority. That would be very on-brand. 

Cookies: Pick Your Poison

The weakest part of the Commission’s proposal concerns cookie consent under Article 5 of the ePrivacy Directive. National governments were understandably unimpressed, and the Cypriot presidency reportedly removed the provision from its final draft. 

Ireland has reopened the question, at least in part. It reportedly asked governments whether the law should expand the list of activities that do not require consent, though it is unclear whether the questionnaire also addressed consent managed through browser settings. 

The joint committee draft leaves the Commission’s proposal intact. Other parliamentary amendments cover nearly every imaginable option: replacing the ePrivacy rule with ordinary GDPR standards, adding more exemptions, narrowing the proposed exemptions, deleting browser-level consent, or expanding it. The one thing Parliament appears to agree on is that cookie banners have not yet consumed enough legislative attention. 

AI Rules Back in Play

The Commission’s proposal contains two provisions of particular importance to AI. Article 88c would clarify that legitimate interests may provide a lawful basis for developing and operating AI systems. Article 9(2)(k) would create a limited exception for sensitive personal data that appears incidentally in AI training datasets. 

In my early June comments, I noted that, in the Council drafts, “Article 88c has disappeared from the operative text, but much of its substance survives in Recital 33a,” while “Article 9(2)(k) survived, albeit in narrowed form.” Under the Irish presidency, both issues appear to be back on the table. 

As with cookie consent, the joint committee draft proposes no changes to these provisions. Other amendments run in opposite directions, with some seeking to delete the AI provisions and others seeking to broaden them. 

I am watching Article 9(2)(k) particularly closely because it may be the Commission’s most immediately consequential reform. As I wrote in the International Center for Law & Economics’ (ICLE) March comments: 

Article 9(2)(k) addresses a practical constraint of large-scale AI training: special categories of personal data will inevitably appear in training datasets despite efforts to exclude them. … Perfect ex ante filtering is technically impossible. … Without Article 9(2)(k), controllers face a binary choice: guarantee perfect exclusion of special-category data or abandon AI training in the EU.

Now Comes the Brussels Waiting Game

The European Parliament’s July 15 amendment deadline produced more than 1,000 proposed changes, so the initial joint committee report settles very little. The lead negotiators will spend the autumn bargaining over compromise amendments while awaiting a targeted impact assessment. At this pace, Parliament is unlikely to adopt its negotiating position before February 2027.

The EDPB’s consultation on its anonymization guidelines closes Oct. 30. That process may clarify the board’s position, though recent history counsels against expecting clarity to arrive unaccompanied by qualifications.

The Council may move faster. Ireland now has six months to do what Cyprus could not: secure an agreed position among national governments. Much will depend on how it uses the responses to its questionnaire. The next compromise text should show whether Ireland is serious about simplifying the GDPR—or merely preparing a more elegant route back to the same thicket.

The post Brussels Tries to Fix the GDPR Without Making It Worse appeared first on Truth on the Market.

That two-way relationship will take center stage tomorrow, when the U.S. Senate Commerce Subcommittee on Telecommunications and Media holds a hearing titled “Intelligent Networks: Powering Artificial Intelligence and Transforming Communications.” Networks carry the traffic AI generates, while AI is beginning to reshape how those networks are built, managed, and shared. 

The hearing will likely cover the familiar ground of investment, deployment, and network architecture. Those questions matter. But each rests on a more basic input that rarely gets equal billing: radio spectrum. 

A new International Center for Law & Economics (ICLE) issue brief examines that overlooked foundation in detail. It is the second installment in ICLE’s series on the infrastructure underpinning the next race for AI leadership. 

The Airwaves Won’t Manage Themselves 

Every wireless service depends on reliable access to radio frequencies where competing signals do not cause harmful interference. Spectrum is not scarce in the ordinary sense; nature does not issue a fixed number of frequencies and then close the shop. But physics limits how much data can travel over a given band, and the law limits who may use particular frequencies and under what conditions. 

Those constraints shape the performance of every wireless service. That makes spectrum policy a critical part of infrastructure policy for AI and other next-generation applications. 

Meeting growing data demand does not require tearing down the spectrum-management system and starting over. It does require targeted reforms. Policymakers should preserve a balanced mix of licensed, unlicensed, and shared access to meet the differing needs of AI, augmented reality (AR), and virtual reality (VR). 

They should also move underused federal spectrum into commercial use more quickly by streamlining the cumbersome Spectrum Relocation Fund, improve coordination between the Federal Communications Commission (FCC) and National Telecommunications and Information Administration (NTIA) around a common technical record, and engage more coherently in the international bodies that shape global equipment markets. 

Most promising for a hearing on intelligent networks, policymakers should replace reflexive worst-case interference analysis with risk-informed methods. AI can help here, too. Tools such as spectrum digital twins—virtual models that simulate how networks and signals interact—can estimate interference more realistically and allow more users to share the same frequencies safely. 

New Technologies Need More Than One Lane

AI, AR, and VR will all increase the load on wireless networks, but they will not place the same demands on them. Immersive AR and VR require sustained capacity and extremely low latency—the delay between a user moving her head and seeing the image respond. AI traffic tends to arrive in bursts and, unlike most consumer applications today, can strain upload capacity as well as download capacity. No single frequency band or access model can meet all those needs. 

That makes a layered approach essential. Low-band spectrum travels long distances and penetrates buildings well, but it carries relatively little data. Mid-band spectrum offers a useful balance of coverage and capacity, making it the workhorse for wide-area AR and cloud-rendered VR. The unlicensed 6 GHz band is a natural fit for immersive traffic that remains indoors and travels only short distances. Millimeter-wave spectrum can deliver enormous capacity in dense venues and campuses. Low-Earth-orbit (LEO) satellites add another layer by reaching places that terrestrial networks cannot. 

The rules governing access matter as much as the frequencies themselves. Exclusive licenses give carriers the interference protection and certainty needed to justify billions of dollars in network investment. Unlicensed spectrum provides the permissionless, low-cost capacity that supports Wi-Fi and a connected-device market that generates trillions of dollars in annual value. Dynamic sharing, which coordinates users in real time to prevent harmful interference, opens frequencies that cannot be fully cleared of existing users. 

Each model serves a distinct purpose. Tilt too far toward any one of them, and policymakers will either leave valuable capacity idle or deprive some applications of the access they need. 

The larger problem is that the U.S. system for making spectrum available moves at bureaucratic speed. The FCC governs commercial use, while NTIA manages federal use, and no clear final arbiter exists when they disagree. Much of the prime mid-band spectrum needed for AI and immersive applications remains assigned to federal incumbents, especially the Department of Defense (DOD). 

The One Big Beautiful Bill Act restored the FCC’s auction authority and created an 800 MHz spectrum pipeline. But while Congress can set a target, it cannot force an agency to relinquish frequencies. If the federal portion of that pipeline stalls, pressure will shift to productive commercial, shared, and unlicensed bands—including the 6 GHz band and the Citizens Broadband Radio Service (CBRS), which Congress declined to protect. 

A hearing about powering AI should therefore ask a basic question: Can the reallocation process actually deliver the spectrum these applications will need? 

AI Can Help Spectrum Think Smarter

AI is not just another source of network demand. It can also help networks manage competing radio operations more efficiently—and more intelligently. 

Machine-learning systems can identify unused capacity, predict interference, and adjust channel access in real time. That allows secondary users to operate without disrupting incumbents, meaning users already authorized to occupy the band. Research on cognitive-radio techniques for 6G has found meaningful gains in spectral efficiency—the amount of data transmitted over a given slice of spectrum—and reductions in interference. 

As these methods mature and regulators gain confidence in them, spectrum policy can move beyond the blunt question of whether one user should receive exclusive access to a band. The better question is how efficiently several users can share it. 

The management gains may matter as much as the added capacity. For decades, agencies have relied on deterministic, worst-case interference analysis. These models often assume that every transmitter operates at maximum power, in the least favorable location, under the most adverse conditions, all at once. That caution imposed few costs when spectrum was lightly used. In today’s crowded bands, it can block productive entry based on scenarios that are technically possible but vanishingly unlikely—and give incumbents a potent weapon against competition. 

A risk-informed approach asks better questions: What interference could occur? How likely is it? How serious would the consequences be? 

AI makes that approach more practical. A spectrum digital twin—a continuously updated virtual model of a real radio environment—can run thousands of simulations using realistic combinations of device locations, power levels, and signal conditions. The result is a probabilistic assessment of interference risk, rather than a single alarming hypothetical. 

The FCC already used probabilistic analysis when it opened the 6 GHz band, and the D.C. Circuit upheld that approach as a reasonable and legally defensible basis for spectrum policy. AI-driven modeling could make such analysis the rule rather than the exception. It would preserve robust protections for safety-of-life and national-security systems, where the stakes are genuinely high, while making room for new services elsewhere. 

That is what an intelligent network should look like: not merely a pipe carrying AI traffic, but a system in which automated coordination, real-time sensing, and data-driven interference analysis allow more users to operate safely in the same frequencies. 

Smart Networks Need Smarter Spectrum Rules

If the subcommittee wants networks capable of powering AI, it should focus on five reforms. 

First, regulators should preserve a balanced mix of licensed, unlicensed, and dynamically shared spectrum, and judge allocations by their total economic value—not auction revenue alone. Second, Congress should streamline the Spectrum Relocation Fund so NTIA can study candidate bands earlier and move underused federal spectrum into commercial use faster. 

Third, the FCC and NTIA should continue building a common technical record so disagreements emerge early, rather than erupting at the end of a proceeding. Fourth, regulators should make risk-informed interference analysis the default for spectrum reallocation and sharing. 

Finally, the United States should present coherent positions at the International Telecommunication Union’s (ITU) World Radiocommunication Conference (WRC). That forum shapes the technical standards and global equipment markets on which future networks depend, and China increasingly treats it as an arena for strategic competition. 

Clear direction from the White House will be necessary to align agencies whose missions and incentives often pull in different directions. But the central point is simple, and the July 30 hearing is well positioned to make it: Intelligent networks require intelligent spectrum policy. 

Spectrum is the invisible infrastructure of the AI era. The United States cannot lead at digital speed while governing the airwaves at bureaucratic speed.

The post The AI Boom Still Runs on Old-Fashioned Airwaves appeared first on Truth on the Market.

State attorneys general play a valuable role when harms are local or federal investigators miss key facts. But serial challenges to nationally integrated conduct turn that safeguard into a standing invitation to relitigate.

The result is a system in which no settlement fully settles and no clearance fully clears. That weakens federal enforcement, rewards the most aggressive jurisdiction, and lets local politics reshape national markets.

State antitrust should supply evidence, police local misconduct, and check genuine federal failure. It should not operate as an endless appeals process with 51 possible vetoes.

When Federalism Becomes a Veto Stack

American antitrust federalism has real virtues. State attorneys general can uncover local cartels, represent residents whose individual claims may be too small to pursue, and alert federal officials to facts a Washington-centered investigation might miss. States can also provide a useful check when federal enforcers overlook a genuine competitive harm.

Those benefits do not require subjecting every national transaction or business practice to 51 independent and potentially conflicting antitrust policies.

The growing problem is sequential contradiction. The U.S. Department of Justice (DOJ) or Federal Trade Commission (FTC) investigates a national matter, settles it or closes the investigation, and a coalition of states then tries to keep the same transaction or conduct in court. The firm has not received a federal answer. It has merely cleared the first gate.

This “veto stack” turns federalism from a source of evidence into a mechanism for repeated litigation.

In an earlier Truth on the Market commentary, I argued for a sounder division of labor (see also related work by Babette Boliek). State enforcement should complement well-grounded federal initiatives and focus scarce resources on competitive harms particular to a state.

That account also recognizes that public-choice pressures and rent seeking can distort decisions at both levels of government. The point is institutional, not partisan. The proper allocation of authority is the one that best reduces the combined costs of anticompetitive conduct and mistaken intervention.

Federal Clearance, State-Sized Asterisk

Recent litigation reveals several forms of state-federal divergence. They should not be forced into a single narrative because each raises a different institutional concern. Together, though, they suggest that federal clearance increasingly comes with an asterisk.

Live Nation offers the strongest warning against claiming that continued state litigation is always wasteful. The DOJ announced a proposed settlement shortly after trial began, but most state plaintiffs stayed in the case and won a jury verdict.

One could still view that outcome as a policy error. Live Nation argues that combining promotion, venue operations, and Ticketmaster’s ticketing services creates an efficient vertical integration that lowers costs. But even assuming the jury reached the right result, one verdict cannot resolve the broader institutional question.

Firms bear the ex ante costs of overlapping authority in every case, including those in which a second prosecution produces no offsetting consumer benefit. Conflicting remedies can also persist long after a court determines liability.

HPE/Juniper raises a different concern. The DOJ first challenged the $14 billion acquisition, then settled for the divestiture of HPE’s Instant On business and licensing commitments covering Juniper’s artificial-intelligence (AI) operations source code.

The states did not file a separate Clayton Act suit. Instead, 12 states and the District of Columbia intervened in the Tunney Act proceeding, which allows courts to review federal antitrust settlements, and urged the court to reject or closely examine the decree.

John Yun’s analysis focused on a key aspect of competition to which the original complaint gave too little weight. Combining Juniper’s AI-powered networking capabilities with HPE could create a stronger rival to Cisco and increase competition. Greater international competitiveness could also advance national security interests.

The state intervention extended uncertainty over a completed integration whose competitive rationale depended on AI networking, scale, and the prospect of a more credible third platform.

Nexstar/Tegna and Paramount Skydance/WBD go further. In Nexstar, state plaintiffs sued after federal regulators approved the deal and the transaction closed.

In Paramount, the DOJ publicly ended an eight-month investigation after concluding that the asserted theories appeared to confuse harm to competitors with harm to consumers. A 12-state coalition filed a federal antitrust suit the next month.

The merits of those challenges remain distinct. The institutional message is harder to miss. A federal resolution may merely open the next round.

The High Cost of One More Veto

Antitrust enforcement is an exercise in managing error costs. False negatives allow market power to persist. False positives deter conduct that lowers costs, improves products, or creates new competitive constraints. The goal should be to minimize the expected social costs of both errors, along with the administrative and compliance costs of enforcement—not to maximize cases, remedies, or the number of officials with a vote.

Sequential state prosecution increases those costs in at least four ways.

First, it erodes the settlement value of federal enforcement. A firm cannot reliably exchange concessions for closure when non-settling states remain free to pursue conflicting relief. That uncertainty raises the price of settlement, weakens the government’s ability to secure efficient remedies, and encourages firms to litigate defensively.

Second, it delays integration. Merger efficiencies often depend on timing. Engineers leave, product road maps diverge, capital budgets freeze, and customers postpone adoption while a transaction’s legal status remains uncertain.

Third, a veto stack magnifies uncertainty. The relevant business question becomes whether any politically prominent jurisdiction might attack the conduct under a more elastic standard, rather than whether the conduct satisfies a coherent consumer-welfare rule. The most aggressive enforcer then determines expected liability, regardless of which agency has the best information.

National companies respond predictably. They design their conduct around the strictest state rule, abandon marginal investments, or avoid business models whose legality requires a fact-intensive assessment of efficiencies.

Fourth, fragmented enforcement creates remedy externalities. A state may seek divestiture, access mandates, data sharing, compulsory licensing, or conduct restrictions whose costs fall largely beyond its borders. State officials receive political credit for a visible intervention, while consumers, workers, and shareholders nationwide absorb the lost scale or innovation. The jurisdiction making the decision does not bear the full national cost.

These risks grow sharper in high technology. Dynamic competition depends on uncertain investments, complementary assets, learning effects, and rapid quality improvements that static market shares often measure poorly. My dynamic-consumer-welfare approach therefore focuses on innovation and long-run consumer gains.

The International Center for Law & Economics’ (ICLE) work on state laws with national spillovers makes a related point. State experimentation becomes costly when one jurisdiction exports the consequences of its policies to the interstate economy. Requiring every state to grow comfortable with an unfamiliar technology before firms may proceed will predictably protect incumbents and punish experimentation.

When ‘AG’ Means ‘Aspiring Governor’

Why do national technology firms make such tempting targets? Public-choice analysis offers a simple answer without accusing anyone of corruption or bad faith.

State attorneys general are elected officials or politically prominent generalists. Large technology cases generate headlines, coalition leadership, fundraising material, and a platform for higher office. Competitors, organized interest groups, and plaintiffs’ lawyers also have strong incentives to provide complaints, studies, and political support.

The beneficiaries of innovation are harder to organize. Future customers, small businesses, and users who value free or improved services rarely form a ready-made constituency.

Empirical research cannot prove that any particular case rests on political motives. It can, however, identify the incentives built into the system. Colin Provost examined when “AG” becomes shorthand for “aspiring governor” and found a relationship between multistate litigation and later campaigns for higher office. John Dove’s work on state antitrust enforcement likewise considers electoral cycles and institutional political pressures.

Those findings counsel restraint. Broader discretionary standards do not produce neutral public-interest enforcement by magic. They increase the rewards for political entrepreneurship.

Technology markets sharpen the problem. Their products are complicated, their alleged harms translate easily into populist rhetoric, and their benefits often appear as better quality, lower latency, stronger security, faster model development, or future competition rather than lower prices today.

A prosecutor can point to size, data, vertical integration, or an unhappy rival. Showing that integration improved performance or created a stronger competitor takes more work. The political reward arrives when the complaint is filed. The innovation lost to delay rarely makes the evening news.

Fifty-One Rulebooks for One National Market

Conflicting enforcement becomes more dangerous when states also apply different substantive rules.

For roughly four decades, the consumer-welfare framework has disciplined antitrust by asking whether challenged conduct is likely to reduce output, raise quality-adjusted prices, suppress innovation, or otherwise harm competition. The framework does not grant firms laissez-faire immunity. It separates competition law from a broader power to redistribute economic advantage among firms and political constituencies.

Several state initiatives point toward wider and less predictable mandates. New York’s Twenty-First Century Antitrust Act, which passed the state Senate in May and remains pending in the Assembly, would create an abuse-of-dominance regime, restrict certain vertical conduct, expand premerger requirements, and authorize private lawsuits. My critique of an earlier version warned that such a standard could protect competitors at consumers’ expense and interfere with national policy.

California’s pending AB 1776 would expand single-firm liability under the Cartwright Act. Babette Boliek’s recent ICLE critique argues that the bill would replace objective consumer-welfare analysis with protection for favored market participants. New Jersey’s S451 would make certain uses of algorithmic rent-coordination systems violations of the New Jersey Antitrust Act.

These proposals differ in scope and may change or fail. Together, they show growing support for state-specific theories that could govern conduct far beyond state borders.

The expertise concern is comparative, not absolute. Many state antitrust lawyers are excellent, and federal agencies make mistakes. But the DOJ and FTC employ larger teams of economists, technologists, merger specialists, and lawyers focused on national markets. They also retain more institutional knowledge from repeated investigations.

A state coalition formed for a prominent case may rely on federal evidence and outside counsel while pursuing a different political objective. When a dispute concerns AI infrastructure, cloud architecture, enterprise networking, algorithmic pricing, or multisided platforms, analytical mistakes become especially costly.

Broad fairness or dominance standards also invite rent seeking. Rivals can repackage requests for protection as antitrust complaints. Regulated intermediaries can seek rules that preserve their position. Local interests can impose costs on consumers nationwide.

Recent Truth on the Market commentary has described this tendency as the “nanny state” substituting regulators’ preferences for demonstrated consumer harm. The law & economics objection does not dismiss every nonprice concern. It simply observes that a standard detached from measurable competitive effects gives too much discretion to officials who face uneven political incentives.

Congress Holds the Preemption Pen

Reform cannot assume that Washington merely lends states their antitrust authority. States may sue under their own statutes, and Congress has expressly authorized state attorneys general to bring parens patriae actions for damages under federal antitrust law. These actions allow a state to seek recovery on behalf of its residents.

In California v. ARC America, the Supreme Court held in 1989 that federal limits on indirect-purchaser recovery did not preempt state laws permitting such claims. The Court reasoned that federal antitrust law did not occupy the entire field and that federal policy alone did not dictate what remedies states could authorize.

ARC America therefore poses a serious obstacle to casual claims of implied preemption. But the decision does not require Congress to preserve overlapping state remedies forever. It interpreted the statutes Congress had enacted.

Congress may regulate interstate commerce and expressly preempt conflicting state rules, so long as it speaks clearly and respects constitutional limits against commandeering state governments. The cleaner approach would establish a federal rule for private conduct and transactions in national markets, backed by an express preemption clause and a carefully drawn exception for genuinely local enforcement. Congress would govern private conduct rather than tell states how to govern.

Courts have less room to act without Congress. Modern dormant Commerce Clause doctrine generally resists invalidating nondiscriminatory state laws merely because they affect conduct beyond the state. The Supreme Court reinforced that caution in National Pork Producers Council v. Ross.

Conflict preemption remains possible when compliance with both state and federal law is impossible or when state law obstructs a clearly stated federal objective. Yet ARC America makes that objective difficult to infer from antitrust policy alone. Judicial reform would therefore require a clearer signal from Congress or a modest doctrinal shift toward giving nationally uniform federal resolutions greater finality.

How to Defuse the Veto Stack

The answer is not to push states out of antitrust enforcement. States can uncover local harms, contribute evidence, and challenge a plainly deficient federal process. But nationally integrated conduct needs a federal resolution that carries more weight than an opening bid.

A workable reform agenda should preserve state participation while limiting contradictory remedies and serial litigation. Congress can clarify when federal decisions preempt state action, courts can give greater weight to reasoned federal assessments, and federal agencies can improve coordination and transparency. AI markets make those reforms especially urgent because even a nominally local injunction can alter a national product.

Make Federal Resolution Mean Actual Resolution

Congress should make federal antitrust resolutions final for conduct or transactions with substantial, nonincidental effects across state lines. Once the DOJ or FTC completes a defined review and issues a reasoned disposition, federal substantive standards should govern exclusively.

A qualifying consent decree, litigated judgment, or published closing statement could preempt later state actions seeking inconsistent prospective relief. States should remain free to participate in the federal investigation, submit evidence, seek compensation for residents, and challenge conduct whose competitive effects are predominantly intrastate.

The dividing line should turn on function, not size alone. Relevant factors could include nationwide pricing or product design, a multistate network, integrated data or infrastructure, or a remedy that would necessarily alter operations nationwide.

A savings clause should preserve state cases involving local bid rigging, regional cartels, occupational restraints, hospital or funeral-home conspiracies confined to one state, and small mergers whose assets and customers are overwhelmingly local. The governing principle should be subsidiarity, which assigns authority to the lowest level of government capable of accounting for the relevant costs and applying adequate expertise.

Preclusion should apply only after a transparent federal process. The DOJ or FTC should explain the theories investigated, the competitive evidence considered, and why the disposition protects consumers.

States could receive a brief consultation period and limited judicial review of whether the statutory conditions were satisfied. They should not receive a second trial on the merits. Congress could also require periodic reports on matters in which a federal resolution displaced a proposed state action.

Police Conflicting Remedies

Even without full field preemption, Congress could direct courts to displace state remedies that materially frustrate a federal resolution of national conduct.

A state damages claim based on a distinct local injury may coexist with a federal decree. A state injunction requiring divestiture after the DOJ accepted a licensing remedy presents a different problem. The question should be whether the state remedy conflicts with the federal resolution, not whether state officials merely disagree with it.

Courts should ask whether the requested relief would undo the settlement bargain, destroy efficiencies the federal disposition preserved, or force changes to a nationwide operating plan that federal enforcers approved.

Courts should remain cautious without becoming passive when Congress has not spoken. If a state brings a federal antitrust claim after the DOJ or FTC issues a detailed competitive assessment, courts should give that assessment substantial weight on market definition, efficiencies, and remedy design.

That approach would preserve state standing while reducing the anomaly of one sovereign obtaining federal relief that defeats an expert federal agency’s stated national policy.

Put the Federal Case on the Record

The most immediate reform requires no new statute. The DOJ already files statements of interest in private and state litigation. Both the DOJ and FTC should adopt a published protocol for intervening when a state case threatens a federal settlement, relies on a theory the federal agency rejected, or seeks a remedy with substantial interstate spillovers.

The filing should set out the federal consumer-welfare analysis, explain the national costs of conflicting relief, and address technical questions that a generalist court might otherwise hear only through dueling experts.

Such filings would be especially useful in Tunney Act proceedings. States should have an opportunity to be heard, but the public-interest review should not become a chance to retry the government’s entire case or replace a federal settlement with a coalition’s preferred remedy.

A clear federal account of uncertainty, litigation risk, innovation benefits, and remedial tradeoffs would help courts distinguish genuine defects from political disagreement.

Make States Explain the Encore

Federal officials should say plainly that cooperation is welcome, but serial prosecution of nationally integrated conduct carries real economic costs.

Congressional hearings could require federal and state enforcers to explain their competing theories, staffing, outside-counsel arrangements, and estimates of interstate spillovers. Greater transparency would force state officials to defend both the alleged harm and the added value of a separate action after federal review.

The DOJ and FTC could also condition joint investigations on ex ante coordination agreements governing information sharing, settlement consultation, and remedy design. A state would remain sovereign and free to withdraw. It should not be able to use the federal investigation, then surprise the parties with a conflicting national remedy.

Published best practices could encourage state offices to conduct cost-benefit analysis and reserve independent action for demonstrable local harm or a clearly deficient federal process.

Give AI One National Rulebook

AI-related conduct will test this system first. Training, cloud computing, chips, networking, data centers, model distribution, and enterprise deployment cross state lines by design. A state injunction governing model access, data use, interoperability, or vertical integration could reshape a national product even when issued under a nominally local statute.

Congress could establish expedited federal review and presumptive federal primacy for AI cases with substantial interstate effects. States would retain their traditional authority over fraud, privacy, contracts, and local cartels.

Federal primacy should preserve vigorous enforcement against exclusion, collusion, and acquisitions likely to harm consumers. The error-cost analysis must also account for dynamic global competition.

As Mario Zúñiga recently argued, antitrust accounts that omit China ignore a major rival in AI development. HPE/Juniper likewise shows how integrating AI-enabled networking can create a stronger challenger. The United States will struggle to out-innovate strategic competitors if every national technology investment depends on satisfying the most economically untethered state theory.

Federal Primacy, With Guardrails

The strongest objection is that federal enforcers can be captured, politicized, or simply wrong. Live Nation’s jury verdict gives that concern real force. A system that turned any cursory federal closing letter into blanket immunity would merely trade one institutional failure for another.

Federal primacy should therefore depend on a credible process. Federal enforcers should conduct an adequate investigation, provide a reasoned public explanation, disclose the logic behind any remedy, and submit consent decrees to judicial review where required.

Preemption also should not erase state damages claims on behalf of residents where federal law expressly preserves them, unless Congress clearly chooses otherwise. Prospective injunctions and efforts to block transactions create the largest national spillovers and present the strongest case for uniform federal treatment.

States should retain broad authority to compensate localized injuries, prosecute local conspiracies, and participate in federal cases. The governing principle is matching jurisdiction to the geographic scope of the market and the remedy, rather than centralizing authority for its own sake.

A more disciplined federalism could also improve state enforcement. State offices would have more resources for procurement cartels, local licensing restraints, unlawful noncompete agreements, health-care consolidation with genuinely regional effects, and anticompetitive state regulation. In those matters, local knowledge offers a genuine advantage rather than a vehicle for imposing one state’s policy nationwide.

Federalism Without the Endless Encore

Antitrust federalism should improve information and accountability. It should not ensure that a national firm can never obtain repose.

When states continue litigating after a reasoned federal settlement or clearance, the costs extend far beyond legal fees. Integration stalls, investment falls, remedies conflict, settlements lose value, and unfamiliar innovations face a built-in handicap. Those costs hit hardest in AI and other technologies where speed, scale, and experimentation shape long-run consumer welfare and national competitiveness.

The answer is neither abolishing state antitrust nor treating federal expertise as infallible. Congress should preserve state authority over genuinely local restraints while restoring federal primacy over conduct with substantial interstate effects. Courts should police conflicting remedies when federal policy is clear. The DOJ and FTC should coordinate earlier, file statements of interest, and explain when state actions threaten consumer welfare. Congressional oversight should expose the incentives behind serial prosecution.

A national market cannot function efficiently with 51 final decision makers. States should lead where the harm and expertise are local. Federal enforcers should answer for markets and remedies that are national.

Federalism should divide the work—not multiply the vetoes.

The post Too Many Cooks Spoil the Settlement appeared first on Truth on the Market.

For four decades, the FCC has barred any company from owning television stations that collectively reach more than 39% of U.S. television households. The rule reflects an older theory of broadcast regulation—one that treats the airwaves as a scarce public resource and promotes competition, localism, and viewpoint diversity through bright-line ownership limits rather than case-by-case review.

That theory once had an intuitive logic. When the modern cap emerged, most Americans got their news and entertainment from stations affiliated with the “Big Three” networks. Limiting any owner’s national footprint could plausibly prevent too much editorial influence from accumulating in too few hands.

That media world has vanished. Broadcasters now compete not only with one another, but also with streaming services, virtual multichannel video programming distributors (vMVPDs), podcasts, and social platforms whose national and global reach dwarfs anything a station group could assemble under the 39% cap. The rule now binds the competitors least able to bear it while leaving their largest rivals untouched.

The FCC has signaled that it intends to repeal the cap and review broadcast consolidation case by case. As a matter of competition policy, that move is overdue. As a matter of law, it is messier. Congress wrote the 39% figure into an appropriations statute, raising the question of whether the FCC may erase it on its own.

The policy debate also cannot stop at ownership. Broadcast regulation operates as an interconnected system. Its other parts include retransmission consent, which governs the terms and fees under which distributors carry broadcast signals; must-carry rules, which can require carriage of qualifying local stations; and FCC standards requiring the parties to negotiate in good faith. Together, these rules divide bargaining power among many of the same companies.

Removing the cap would give larger station groups more leverage over the cable, satellite, and streaming distributors that carry their signals. The retransmission regime was not designed to offset that added power. Repealing the cap while leaving the carriage rules untouched would therefore do less to eliminate a distortion than to move it elsewhere.

The better course is comprehensive reform. Because ownership limits, retransmission consent, and bargaining standards all fall within the FCC’s jurisdiction, the agency should consider them together. Otherwise, repeal may simply reshuffle bargaining power among industry players while consumers keep paying the bill. 

How 39% Became Broadcast’s Magic Number

The national television ownership rule measures concentration by audience reach, not station count. A broadcaster may not hold an attributable interest—a stake large enough to count as ownership under FCC rules—in stations that collectively reach more than 39% of U.S. television households.

The rule grew out of broadcasting’s traditional “scarcity rationale.” Because the electromagnetic spectrum could accommodate only a limited number of signals, regulators treated licensees as stewards of a public resource with a duty to serve the public interest. The FCC therefore pursued competition, localism, and viewpoint diversity through fixed ownership limits rather than reviewing each transaction on its particular facts.

The theory was straightforward. Preventing any company from building a nationwide footprint would preserve a decentralized system of locally accountable stations and keep editorial influence from accumulating in too few hands. That concern carried more weight in the 1980s, when most Americans relied on stations affiliated with the “Big Three” networks for news and entertainment.

The 39% cap is only the latest entry in a much longer regulatory ledger. The FCC first imposed national ownership limits in the 1940s, initially by restricting how many stations one company could own. It later shifted to the audience-reach measure used today.

With the Telecommunications Act of 1996, Congress directed the FCC to raise the cap to 35%. When the agency tried to increase it in 2003 to 45%, Congress again stepped in. The Consolidated Appropriations Act of 2004 changed the figure to 39%, where it has remained.

That history now sits at the center of the legal dispute over whether the FCC may repeal the cap without further action from Congress. Whatever the answer, the number rests on the same premise that has long animated the rule: Limiting a broadcaster’s national reach will protect a diverse, locally oriented media system.

A Broadcast Cap in a Streaming World

In today’s media economy, that premise points in the wrong direction. Local broadcasters compete for audiences and advertising dollars with streaming services, vMVPDs, podcasts, and social platforms whose national and global scale dwarfs anything a station group could assemble under a 39% ceiling. A rule intended to prevent broadcasters from becoming too powerful now risks keeping them too small to compete effectively.

The FCC has often resisted this conclusion by treating another service as a competitor only if it substitutes for broadcasting in every part of the business. Cable and streaming services, for example, may be treated as complements rather than competitors because they do not bid for network affiliations or retransmission-consent fees.

That reasoning is circular. Cable operators do not bid for retransmission-consent fees because federal law puts them on the paying side of that transaction. The regulatory system creates the difference, then cites the same difference as proof that broadcasters and distributors occupy separate markets and deserve different treatment.

Consumer behavior tells a different story. By Nielsen’s measure, streaming now accounts for roughly 48.6% of television viewing time, compared with 21.5% for broadcast. Over-the-air radio’s share of audio listening likewise fell from 46% to 34% between 2018 and 2025 as streaming and podcasts gained ground. A market defined narrowly as “local broadcast television” no longer reflects how audiences choose content or advertisers spend money.

Against that backdrop, lifting the ownership cap would be pro-competitive. Broadcasters can promote localism and viewpoint diversity only if they remain strong enough to offer a meaningful alternative to digital platforms. YouTube and Netflix may reach the entire country. Broadcasters remain frozen at 39% of television households.

Removing that asymmetry would let station groups build some of the scale their digital rivals already enjoy. Scale can make broadcasters more effective competitors, not less. Larger groups can spread the fixed costs of investigative teams, weather forecasting, and production infrastructure across more stations, reducing the average cost of producing local journalism that national platforms cannot easily replicate.

Greater scale could also strengthen broadcasters when they acquire programming and negotiate carriage on the streaming services and vMVPDs that increasingly connect them with viewers. It could help finance the “must-have” local content that distinguishes stations in advertising markets dominated by technology companies.

On those terms, consolidation may preserve localism and viewpoint diversity rather than threaten them. The greatest danger to local broadcasting may not be excessive concentration in midsize markets. It may be the financial collapse of stations that outdated rules prevent from adapting. A cap that keeps broadcasters too small to compete efficiently starves them of investment and cedes more ground to the digital giants the rules leave untouched.

The cap also exposes a deeper problem: the public-interest standard through which the FCC regulates broadcasters. The Communications Act directs the agency to act in the “public interest, convenience, and necessity,” but gives those words no concrete definition.

As Eric Fruits has observed in revisiting Stephen Breyer’s classic critique, such vague standards are a familiar regulatory pathology. They give agencies sweeping discretion while preserving the appearance of rule-bound decision-making. Because no objective test can measure the “public interest,” different commissions can use the same words to justify sharply different results. 

The 8th U.S. Circuit Court of Appeals’ decision in Zimmer Radio shows how that vagueness feeds back into the ownership rules. The court interpreted the statutory term “competition” through the public-interest standard, treating it as a broad grant of authority with few meaningful limits. It then deferred to an FCC market definition that excluded broadcasters’ digital rivals. 

The FCC’s use of the public-interest standard to promote localism and viewpoint diversity may also threaten broadcasters’ First Amendment rights. A standard with few limiting principles allows the agency to regulate content, condition license grants, and demand concessions from merging parties without tying those demands to any harm caused by the transaction. 

Broadcasting’s reduced First Amendment protection rests largely on the spectrum-scarcity rationale of Red Lion, which upheld content obligations that would almost certainly fail if imposed on a newspaper. That rationale has crumbled in an era of digital abundance. The result is an increasingly hard-to-defend system in which the same speech receives full constitutional protection in print, on cable, and online, yet remains subject to special FCC oversight when transmitted by a broadcast station. 

Repealing the cap would not cure every defect in the public-interest standard. But it would, at least, allow the FCC to assess consolidation across the broader media market rather than inside an increasingly artificial broadcast silo. 

Can the FCC Kill Congress’ Cap?

The policy case for repeal, however strong, faces a threshold objection: Congress mandated the 39% figure. Congress placed no limit on the raw number of television stations a company may own nationwide, but it imposed a firm ceiling on their combined audience reach. Whatever the cap’s merits, an agency ordinarily must apply the law Congress enacted until Congress changes it.

Two features of the Consolidated Appropriations Act of 2004 bolster that reading. First, Congress excluded the national cap from the quadrennial-review process, under which the FCC reexamines its other ownership rules every four years. That omission suggests Congress reserved this particular number for itself rather than leaving it to periodic agency reconsideration.

Second, Congress expressly barred the FCC from using its Section 10 forbearance authority to avoid applying the cap. Forbearance allows the agency, in certain circumstances, to stop enforcing regulatory requirements it finds unnecessary. Section 10 applies by its terms to telecommunications carriers and services under Title II of the Communications Act, while broadcast licensing falls under Title III. The cross-reference may therefore reflect imprecise drafting or an abundance of legislative caution. Either way, opponents of repeal argue that Congress’ message was plain: The FCC could not simply set the cap aside.

Recent Supreme Court doctrine makes the agency’s position more precarious. In Loper Bright Enterprises v. Raimondo, the Court retired Chevron deference, under which judges often accepted a reasonable agency interpretation of an ambiguous statute. Courts must now decide for themselves what the statute means, and a reviewing court may reject the FCC’s interpretation outright. 

The major-questions doctrine may present another obstacle. Under that doctrine, courts expect clear authorization before an agency makes a decision of vast economic or political significance. Opponents of repeal could argue that eliminating a nationwide ownership limit qualifies and that Congress never clearly gave the FCC that power. On this view, the cure for an obsolete cap is legislation, not an agency decision to stop enforcing it. 

The FCC nevertheless has a serious argument that Congress never stripped it of authority to revisit the cap. The agency created the national ownership limit under the Communications Act’s broad grant of regulatory authority and revised it repeatedly before Congress intervened. 

When Congress set a 35% cap in the Telecommunications Act of 1996, it directed the FCC to “modify its rules” accordingly. In 2002, the D.C. Circuit held that this language preserved the agency’s discretion to alter or eliminate the cap. The court reasoned that Congress could have written the number directly into the statute had it wished to freeze the limit in place. 

Two years later, presumably aware of that ruling, Congress raised the figure from 35% to 39% while retaining the same “modify its rules” language. It also declined to enact competing House and Senate bills that would have expressly codified the cap. Preserving language that a federal court had already interpreted as leaving the FCC room to act looks less like a withdrawal of authority than a ratification of it. 

The other statutory provisions do not necessarily foreclose that reading. Excluding the cap from mandatory quadrennial review does not automatically eliminate the FCC’s independent authority to reconsider it. Nor does a prohibition on forbearance necessarily prevent the agency from repealing the rule through notice-and-comment rulemaking—the standard process in which an agency proposes a rule, receives public comments, and explains its final decision. Declining to enforce a rule and lawfully rescinding it are not the same thing. 

The demise of Chevron may even cut in the FCC’s favor. Loper Bright requires courts to determine the statute’s best reading, but it also recognizes that Congress may delegate policy choices to agencies. A reviewing court could conclude that Congress directed the FCC to modify its ownership rules while leaving the agency discretion over what those rules should become. 

The FCC would receive no automatic deference on the statute’s meaning. But a court could still find that Congress gave the agency authority to make the choice. The end of deference is not the end of delegation.

Pull One Lever, Move the Whole Machine

Lifting the cap cannot be treated as a stand-alone reform. The ownership rules, Title VI’s retransmission-consent and must-carry regime, and the FCC’s good-faith bargaining standards all govern the same firms and shape leverage across the same market. Changing one part while leaving the others untouched does not eliminate a distortion so much as relocate it.

The link between ownership and retransmission consent is bargaining power. Retransmission fees have grown to rival advertising as a source of broadcaster revenue, and a station group’s leverage rises with the size of its footprint. Removing the cap would therefore give consolidated broadcasters more leverage over the cable, satellite, and vMVPD distributors that carry their signals.

The retransmission-consent regime was not designed with that leverage in mind. When Congress created it, cable operators were widely viewed as local monopolists. Today, larger station groups can use their scale to demand higher carriage fees. Distributors may respond by passing those costs to subscribers or dropping smaller, independent networks to contain expenses.

The regime’s original rationale has also weakened on its own terms. Congress adopted mandatory carriage in the 1992 Cable Act to address the “cable bottleneck”—the fear that monopoly cable systems would refuse to carry local stations. But as the D.C. Circuit recognized in Comcast Corp. v. FCC in 2009, cable operators no longer possess the bottleneck power that animated Congress in 1992. Broadcasters can now reach viewers through their own websites, apps, and streaming services.

Once the bottleneck disappears, the case for compulsory carriage—and for the elaborate machinery of fee negotiations and blackouts layered on top of it—becomes much weaker. Repealing the ownership cap would not create that mismatch, but it would magnify it by removing one constraint on broadcasters while preserving a carriage regime built for another era. 

Revisiting retransmission consent could take several forms. The cleanest option would be to phase out both retransmission consent and must-carry, treat broadcasters like other content owners, and let copyright law and voluntary contracts govern distribution. As Geoffrey Manne argued in congressional testimony in 2013, that approach would also end the regulatory asymmetry between broadcasters and their streaming competitors. 

Short of repeal, the FCC could tighten its good-faith bargaining rules to prevent parties from using scale merely to delay negotiations, posture, or bundle unrelated demands. It could also limit automatic fee-escalation clauses that allow each acquisition to ratchet up rates across an entire portfolio. 

The agency might also consider final-offer, or “baseball-style,” arbitration for disputes during high-value programming windows. Each side would submit its best offer, and an arbitrator would select one rather than splitting the difference. Because an unreasonable proposal risks losing outright, the process discourages extreme demands and reduces the chance that viewers become collateral damage in blackout brinkmanship.

The broader point is simple: Broadcast regulation will shift bargaining power somewhere. The relevant question is not which industry gains leverage, but whether consumers receive lower prices, better programming, or more reliable service.

Repealing the cap by itself may simply transfer revenue from distributors to larger broadcasters without producing any consumer benefit. Because ownership rules, carriage rules, and bargaining standards all fall within the FCC’s jurisdiction and govern the same products and firms, this is one of the rare cases in which comprehensive reform is not just desirable, but achievable.

Don’t Stop at 39%

The national ownership cap was built for a media market that no longer exists, and the FCC is right to seek its retirement. But the case for repeal is also a warning against half-measures.

Whatever a court ultimately decides about the Commission’s authority, the broader lesson is clear: Broadcast regulation is a system, not a stack of unrelated rules. Ownership limits, retransmission consent, and bargaining standards all shape leverage among the same firms. Pull one lever, and the others move.

If the FCC repeals the cap but leaves the carriage rules untouched, it will not have deregulated so much as redistributed bargaining power. Larger broadcasters will gain leverage that the retransmission regime was never designed to check, with no guarantee of lower prices, better programming, or more reliable service for consumers.

The Commission should not merely lift the cap. It should finish the job.

The post Pulling the 39% Thread: Why the FCC Must Fix More Than the Broadcast Cap appeared first on Truth on the Market.

The International Center for Law & Economics (ICLE) submitted comments in response to the FTC’s request for input, and this post is, in part, a digest of those comments. The statement gets some important things right, including the risks of excessive regulation and a patchwork of state laws. But it offers little concrete guidance on deception, leans on dubious assumptions about consumer expectations, wanders into constitutionally protected editorial judgments, and treats federal preemption as more wish than doctrine. 

In other words, the AI Policy Statement needs considerable work. Then again, perhaps guiding enforcement was never quite the point. 

But first, some context. 

Just Asking Questions—and Testing the Limits

Way back in February 2025, the FTC issued a Request for Public Comment Regarding Technology Platform Censorship (RFI). There, as you may recall, the agency explained: 

FTC staff is interested in understanding how consumers have been harmed—including by potentially unfair or deceptive acts or practices, or potentially unfair methods of competition—by technology platforms that limit users’ ability to share their ideas or affiliations freely and openly. 

FTC staff was, as the kids say, “just asking.” Then again, perhaps it was not just asking. The RFI sought comment on whether technology platforms’ content-moderation practices might violate either the consumer-protection prong of Section 5 of the FTC Act—unfair or deceptive acts or practices (UDAP)—or its competition prong—unfair methods of competition (UMC). Some questions seemed loaded—and loaded for bear, at that. Comments from FTC leadership describing tech “censorship” as “un-American” did little to dispel that impression. 

ICLE’s comments on the RFI raised those concerns, as did submissions from the Program on Law, Economics, and Privacy at George Mason University’s Antonin Scalia Law School, the Foundation for Individual Rights and Expression (FIRE), the American Enterprise Institute (AEI), and the Cato Institute. ICLE offered “reasons why both the law and underlying economics support limiting how the FTC acts in response to the information gained in this RFI,” while acknowledging that “a call for diverse public comments is likely the best means at the agency’s disposal to gather preliminary information, and that no subsequent formal economic study or law-enforcement action is implied by such an inquiry.” 

Still, our comments were heavy on the limits and brief on the allowance. Some addressed the potential use of the FTC’s UDAP authority against technology platforms’ content-moderation policies and practices. The rest considered whether the FTC could use its UMC authority—or the federal antitrust laws more broadly—to challenge content moderation. 

That emphasis reflected a multipronged initiative by both the FTC and the U.S. Department of Justice’s (DOJ) Antitrust Division suggesting that content moderation might violate the antitrust laws. Beyond the RFI, FTC Chair Andrew Ferguson made informal comments about “tech platform censorship,” while Dina Kallay, deputy assistant attorney general for antitrust at DOJ, offered related remarks. 

DOJ also filed a statement of interest in a private antitrust suit brought against several major news publishers by Children’s Health Defense and other plaintiffs. U.S. Health and Human Services Secretary Robert F. Kennedy Jr. previously chaired Children’s Health Defense. The suit itself seemed dubious, and The Wall Street Journal was no more impressed by DOJ’s statement of interest, which it called “bizarre.” 

To be clear, our position was not that content-moderation policies could never support a viable antitrust or UDAP claim. It was that the space for such claims was narrow. Among the hurdles, platforms’ content-moderation decisions are themselves protected by the First Amendment to the U.S. Constitution. 

The Supreme Court’s decision in Moody v. NetChoice seemed especially relevant. The case concerned Florida and Texas laws that sought to regulate “large social media companies and other internet platforms.” As the Court explained, those laws would “curtail the platforms’ capacity to engage in content moderation—to filter, prioritize, and label the varied third-party messages, videos, and other content their users wish to post.” 

That holding did not foreclose every possible challenge to content-moderation policies or practices, but it narrowed the field considerably. Platforms, like newspapers, may still face liability for price fixing or unlawful monopolization in their business dealings, even when their content-moderation decisions qualify as protected speech. 

For more on this Scylla and Charybdis—and the dangerous route between them—we recommend Content Moderation Antitrust, a collection published by Concurrences and edited by Bilal Sayyed, former director of the FTC’s Office of Policy Planning. It includes our article, “Is There an Empty Set at the Intersection of Antitrust and Content Moderation?” (available free here). 

In the year that followed, the issue seemed to fade. Or, at least, neither the FTC nor DOJ took it to court. 

Credit Where It’s Due—Before the ‘But’

Concerns about artificial intelligence (AI)—and proposals to regulate it—are myriad and, in many cases, “out there.” That does not mean there are no legitimate concerns, or that policymakers could not design well-tailored rules to address durable market failures that harm consumers. We simply have not seen many such rules proposed or adopted.

As ICLE’s comments on the AI Policy Statement observe, “[t]he FTC gets much right in this proposed AI Policy Statement.” There is a “but” coming, but credit where it is due.

First, the statement understandably responds to White House executive orders and policy directives. As the Supreme Court recently reminded us in Trump v. Slaughter, the FTC is an executive agency, and the Constitution vests “the executive power” in the president of the United States.

Questions about the proper scope of that power aside, it is entirely appropriate for the FTC and its commissioners to respond to executive orders issued by the White House—and not merely because the Court has held that the president may remove a commissioner at will and without cause.

It is fitting, then, that the AI Policy Statement responds to several White House documents, including an AI Action Plan, a report on AI from the Council of Economic Advisers, and two executive orders (here and here). 

Four Things the FTC Gets Right

Several policy observations in the FTC’s statement seem salutary.

First, we agree that consumers and competition at every level of the AI stack may be better protected by uniform federal policy than by a motley collection of state regulations. That is not because federal law should generally displace state regulation—there is no good reason to think so—or because, to borrow Justice Louis Brandeis’ phrase, a state cannot serve “as a laboratory” for testing novel policies, assuming it collects the right data and conducts careful analysis.

Leave aside, for now, that some state proposals strike us as bonkers. The central problem is that several features of AI weigh against state-by-state experimentation: the rapid pace of AI development, the interstate reach of leading firms and their products, and the nationwide distribution of consumers. State AI policies are therefore likely to produce substantial spillover effects beyond their borders.

A thicket of potentially conflicting state regulations would also add considerable complexity—and other costs—to AI development. We are therefore inclined to agree that the country needs a “national AI framework” that will “protect innovation and competition by providing national regulatory clarity and certainty and avoiding a balkanized or patchwork regulatory approach driven by States—or, most dangerously, imposed by certain anti-innovation State governments on the rest of the country.”

That is not an endorsement of the effectiveness or efficiency of any particular federal policy. But it is not nothing.

Second, the AI Policy Statement declares that “[e]xcessive AI regulation would undermine American AI supremacy by deterring and suppressing the same ingenuity responsible for making American AI great.” That seems right, almost by definition. And while “supremacy” is not obviously a Section 5 concern, excessive AI regulation—whether federal or state—would plainly tend to impede innovation in ways that harm competition and consumers.

These observations may not provide much concrete guidance, but they could still support useful research and advocacy by FTC staff—work that former FTC Chair William Kovacic called the agency’s “policy R&D” mission.

Third, and perhaps more directly relevant to enforcement policy, the FTC is right that AI providers are not generally exempt from scrutiny under the FTC Act. Section 5 expressly excludes certain firms and types of commerce, including banks and common carriers, and Section 4 implies an exemption for nonprofits. But the statute contains no general exemption for AI providers at any level of the stack.

At a minimum, AI firms could violate Section 5’s deception prong by making false or misleading claims about their products or services when those claims are material, likely to mislead reasonable consumers, and likely to cause harm.

Whether any firms have done so is another question. Still, the statement’s citation to the FTC’s 1983 Policy Statement on Deception signals a grounded approach.

Fourth, while the AI Policy Statement tells us relatively little about the AI industry, it does recognize that AI is “an umbrella term covering a universe of different tools and systems.” That may be a feature rather than a bug, but it also points to the complications ahead.

Reasonable Expectations Require Actual Evidence

There remains the question of when AI firms might—and might not—face liability under the FTC Act. On that, the AI Policy Statement tells us far less than we might like.

The Commission is right that “AI” covers a wide and evolving range of products and services. It is also right to emphasize innovation in a field that changes at remarkable speed. But much of the FTC’s discussion of deception rests on quick, oversimplified assumptions about consumer expectations—“baseline consumer expectations,” “consumers’ reasonable expectations” about chatbots, and so on.

Those expectations are likely to vary across consumers, products, and services. They are also likely to shift as consumers gain experience with an ever-changing array of AI-based tools.

That is not to say that consumer expectations do not exist or should not inform Section 5 analysis. But beyond a few basic assumptions about how consumers interpret factual claims in advertising or other marketing assurances, the Commission should not presume much without careful, context-specific inquiry.

Even seemingly straightforward assumptions about advertising claims can prove controversial. For a sense of the complications, compare two statements by Maureen Ohlhausen in the FTC’s POM Wonderful case: the Commission opinion and her concurring statement. In the latter, she disputed several implied claims identified by the Commission majority, as well as its substantiation standards—concerns later shared by the D.C. Circuit, which rejected the majority’s substantiation standard. 

From Consumer Protection to the Ministry of Truth

Of course, AI firms—like other firms—may engage in commercial fraud. Nothing about AI makes its providers categorically incapable of violating Section 5 through false or misleading advertising claims, marketing materials, or other conduct. 

Newspapers may be the firms most strongly protected by the First Amendment’s Speech Clause, and the Supreme Court has specifically held that the Constitution protects their editorial discretion. But newspapers may still violate the antitrust laws—and, by extension, Section 5—when their business dealings, apart from speech or editorial discretion, are anticompetitive. We might similarly expect a UDAP claim unrelated to content to survive First Amendment scrutiny. Given the right facts, the FTC could establish a violation of the FTC Act. 

Apart from clear and relatively trivial cases, though, the AI Policy Statement does little of what agency guidance is supposed to do. How will—and should—the Commission evaluate AI-provider conduct under its deception authority? How will it approach cases in ways that reflect what the Commission and its staff learn about the industry? And how will it apply its established authority to that industry-specific knowledge?

An error-cost framework helps illustrate the shortcomings. Core UDAP cases involve clear consumer fraud. In those cases, overdeterrence poses little concern because the conduct has no procompetitive rationale, and there is no such thing as too little fraud.

Borderline cases are different. When enforcement targets conduct whose meaning, materiality, or likelihood of consumer harm remains unclear, two concerns arise. First, although such cases may help develop the law, they also create a risk of false positives. That raises questions about both the frequency and cost of false positives—not merely false negatives. Here, one might return to the early law & economics literature on efficient penalties or fines, comparing, e.g., Steven Shavell’s work with that of Louis Kaplow and Robert Cooter. 

Second, when liability turns on a firm’s decision to provide information in the first place, pursuing unclear cases makes disclosure itself risky. That can suppress useful information along with noise or harmful claims.

False promises about, e.g., an AI product’s performance might be actionable. But this is where the AI Policy Statement goes off track. As ICLE’s comments acknowledge:

Published AI policies and marketing materials could, in principle, mislead reasonable consumers. The representations cited in the Statement, though, appear limited. Most providers do not promise certainty or perfect accuracy. Instead, they expressly warn users that outputs may contain errors.

For example, the statement claims that “[i]n marketing their products as problem-solving tools, AI companies have represented explicitly and implicitly that their AI systems aim to produce the best output possible given technological and resource constraints.” Perhaps some have. But its first example is an OpenAI statement that reads more like a warning than an assurance: “At OpenAI, we’re working hard to make AI systems more useful and reliable. Even as language models become more capable, one challenge remains stubbornly hard to fully solve: hallucinations.”

There could, of course, be other statements, other materials, and demonstrable consumer harm. But if this is the statement’s illustration of the Section 5 problem, one wonders. Taken one by one, its citations to statements from OpenAI, Anthropic, and Grok look like strained examples of misleading speech, if not outright counterexamples to the Commission’s point. Statements from Google (about Gemini), DeepSeek, and GabAI seem similarly cautious.

Cases built on dubious readings of marketing statements give AI companies an incentive to tell us less, not more. Harmful omissions or “reasonable” consumer expectations? Maybe. Tell us more.

More troublingly, the statement focuses on “outputs that are distorted by undisclosed ideological objectives,” models with “ideological bias,” and a “hidden agenda.” Those phrases are exceedingly vague and seem like a frolic and detour from ordinary deception law. Worse, they point toward precisely the kinds of conduct entitled to the highest level of First Amendment protection, whether by newspapers, as the Court held in Miami Herald Publishing Co. v. Tornillo; technology platforms, as it held in Moody v. NetChoice; or, we expect, AI providers. 

There is much to say about how the First Amendment will likely protect AI chatbots. Chatbots answer questions—including, but hardly limited to, questions of objective fact—which is paradigmatically speech. AI companies also exercise considerable editorial judgment when selecting inputs, training models, and testing outputs for consistency with company policies and objectives. Users, in turn, have a First Amendment interest in “hearing” what chatbots have to say.

There is no reason to think AI chatbots lack a general right to participate in the “marketplace of ideas.” The FTC should be wary of becoming Oceania’s Ministry of Truth, charged with deciding whether particular AI outputs are “accurate” enough to survive government review. As with social media, government failure in regulating AI misinformation seems more likely than market failure.

Preemption by Wishful Thinking

The AI Policy Statement’s last major problem lies in its remarkably thin—and remarkably wrong—account of federal preemption. 

We will be brief, partly because the ICLE comments offer a fuller discussion and partly because we recognize that not everyone—even among our readers—wants to plunge into preemption doctrine, which stems chiefly from the Constitution’s Supremacy Clause, with an assist from the Necessary and Proper Clause.

The AI Policy Statement says: 

Although the FTC Act does not expressly preempt state law, state law is impliedly preempted to the extent it conflicts with a federal regulatory scheme. A state law that requires an AI firm to deceive its consumers obviously conflicts with Section 5’s express purpose of protecting consumers from such conduct.

A state law requiring firms to deceive consumers would have all sorts of problems, including constitutional ones. But implied preemption through “obstacle preemption”—on the theory that such a law conflicts with Section 5’s “express purpose”—is not among them. The statement’s preemption argument is not merely exceedingly brief. It is rubbish.

Many of the Commission’s UDAP enforcement actions under this administration have appeared well-grounded. That is a welcome development. The AI Policy Statement is something else: less a guide to enforcing Section 5 than an invitation to improvise beyond it.

The post The FTC’s AI Accuracy Statement Needs a Fact Check appeared first on Truth on the Market.

On July 16, Moonshot AI, a Chinese artificial-intelligence company, released Kimi K3. It is an open-weight model, meaning the numerical parameters that encode what the model learned are publicly available for others to download, modify, and run. Even skeptical observers rated K3 “pretty much on par” with the best publicly available models of early 2026. 

Five days later, OpenAI disclosed a very different milestone. During an internal cybersecurity evaluation, its models chained together several zero-day exploits, escaped their test environment, and achieved remote-code execution on Hugging Face’s production servers. A zero-day exploit targets a software flaw unknown to the developer or not yet patched. Remote-code execution allows an attacker to run commands on another computer. Hugging Face is a widely used platform for hosting and distributing AI models, datasets, and development tools. OpenAI called the result “unprecedented.” 

Then Washington entered the fray. On July 22, White House science adviser Michael Kratsios said the government had information that Moonshot built K3 by distilling Anthropic’s Fable model at industrial scale. Distillation is a technique for training a smaller or competing model on the outputs of another model. A distillation attack uses large volumes of unauthorized or deceptive queries to copy capabilities from a rival system. 

According to Kratsios, Moonshot used a platform designed to evade detection and relied on export-controlled Nvidia servers accessed through Thailand. Within hours, Treasury Secretary Scott Bessent warned that “open source is not open season on American IP” and said distillation attacks that “cross the line into IP theft” could put “sanctions and Entity List designations . . . on the table.” An Entity List designation subjects a person or company to U.S. export restrictions, often requiring licenses before American firms may supply specified goods, software, or technology. 

The three developments point in different directions. Kimi K3 suggests that the gap between leading proprietary and open-weight models may be measured in quarters rather than years. OpenAI’s disclosure shows that even one of the world’s best-funded laboratories struggled to keep its own model inside a sandbox, a controlled environment intended to prevent outside access or damage. 

The White House response points somewhere else again. Washington reached for its strongest trade and sanctions tools to address conduct that, by the government’s own account, involved fraudulent application programming interface (API) access and smuggled chips rather than open weights themselves. An API allows one piece of software to send requests to another and receive its outputs. 

These events identify the questions AI governance should confront. The policy debate nonetheless keeps looking elsewhere. 

That debate came into focus last week in a widely shared response to Kimi from Dean Ball, a former White House AI-policy adviser, current head of strategic futures at OpenAI, and author of the Hyperdimensional newsletter. Two of his claims deserve scrutiny. The first is that open-weight models are “inherently decelerationist” because they discourage investment in frontier-model development. The second is that a world dominated by open weights ends in “full AI communism,” with the state providing AI as “digital public infrastructure.” 

Both claims misunderstand how markets create value. They also direct policymakers toward the wrong tools for addressing AI misuse and security threats.

Policymakers should focus on three priorities. They should preserve a healthy mix of open and proprietary models so defensive AI tools can spread widely. They should abandon export controls that fail to achieve their aims while imposing serious unintended costs. And they should recognize that firms providing access to models are often best positioned to detect and stop abuse.

A Commodity Is Not Communism

The economics are straightforward. If an open model that trails the frontier can reach “pretty much on par” within a few quarters, parts of the model layer are becoming commodities. That is what happens when meaningful differences between competing products shrink. 

Technology analyst Ben Thompson, founder of the Stratechery newsletter, offers a useful refinement. A token from one model is not interchangeable with a token from another, so whether a model behaves like a commodity depends on the task. The relevant question is how much it costs to complete that task and how much demand the task creates. On this account, the commodity is less the model itself than the useful intelligence assembled from its tokens. 

That refinement cuts against Ball’s argument. U.S. firms can remain competitive even if rivals distill their models, provided they still offer better efficiency, lower costs, or a better fit for particular uses. 

A commodity is not communism. Nor does a public good in the economic sense—something nonrival and nonexcludable—require state provision. Open-source software is the standard example of a privately supplied public good. 

Commoditization also does not necessarily deter investment. Linux replicates many functions of Windows and macOS, yet it became the foundation for Android and much of the world’s cloud-computing business. Value moved to other parts of the technology stack, including data, distribution, applications, and integration. This is the familiar strategy of commoditizing a complement so that demand grows for the product you sell. 

If customers continue paying for frontier capabilities, free substitutes pose little threat. If customers switch, the market has decided that the frontier premium is not worth the price. That is competition, not communism. Ball cannot plausibly claim both that the frontier remains steep and valuable and that open weights will eliminate the investment needed to reach it. 

There is a more plausible version of the case against open weights, though it is a trade argument. Ball attributes China’s open-weight strategy partly to “the normal Chinese strategy of aggressive exports.” The White House now uses similar language. Secretary Bessent’s warning that “open source is not open season on American IP” frames Chinese models as a problem of dumping and theft, with trade sanctions as the answer. 

If a state-backed rival distributes a product below cost to damage a domestic industry, trade law calls that dumping. The claim is familiar, heavily contested, and subject to established remedies. It should be evaluated as an economic and national-security question, with evidence of underpricing, injury, and likely effects. There is no need to dress it up as a broader question about whether AI may be sold through ordinary markets. 

Even then, a dumping claim based on free models nearly defeats itself. Predatory pricing usually requires some plausible path to recouping the initial losses once competitors have been weakened. It is hard to see how a laboratory giving away open-weight models later raises prices enough to recover that investment. 

Ball predicts that the administration may instead manufacture diffuse “regulatory risk” around Chinese models. That would amount to protectionism without the usual burden of proving underpricing, injury, or any long-term theory of how Chinese laboratories expect to profit from open-weight releases. 

The economics matter most for what they imply about policy. If parts of the model layer are becoming commodities, banning open-weight models or stretching export controls to cover them will miss both the source of the alleged harm and the conduct causing it. 

The plausible complaint is about trade practices. The plausible control point is neither the model weights nor the hardware that runs them. 

Export Controls Meet the Copy Button

The “AI communism” argument is one front in the broader fight over export controls, which remain a form of economic policy. Consider an extreme example. In September 2024, the U.S. Department of Commerce extended export controls to quantum computing, a technology so immature that no one can yet identify its main commercial uses, forecast its cost curve, or say which hardware design will prevail. Scholars have aptly described the exercise as regulation under deep uncertainty. 

Export controls, assuming they work at all, have the best chance of success when they target physical technology that governments can identify and isolate. Quantum computing therefore presents a relatively favorable case. Quantum machines require dilution refrigerators, cryogenic systems, vibration isolation, and electromagnetic shielding. That bulky, traceable infrastructure creates natural enforcement chokepoints. 

Even there, the evidence calls for humility. Studying the 2007 U.S. “China Rule,” Ernest Liu, Yingyi Liu, Alexey Makarin, and Xuan Wen find that export controls reduced targeted imports in the short term. They also pushed affected Chinese firms and their suppliers to spend more on research and development and file more patents in the controlled technologies. China’s cryogenics industry may already be innovating around current restrictions. 

At best, export controls buy time. They do not guarantee permanent denial. Their value depends on how the United States uses that time, which is why the debate over chip controls ultimately turns on competing forecasts about AI development and China’s ability to produce substitutes. 

Model weights present a far harder target than quantum equipment. They are weightless, infinitely replicable at nearly zero marginal cost, and already distributed around the world. Open-weight models also generate no revenue stream to embargo. Trying to “starve” China of them would deny Chinese laboratories little while burdening global experimentation, safety research, and low-cost defensive uses. 

A ban would still create one clear beneficiary. It would give a small group of U.S. firms a legally protected moat, along with the rent seeking and regulatory capture that such protection invites. Restrictions on exporting or importing Chinese models would function chiefly as industrial protection. 

The irony is that Ball attributes China’s open-weight strategy partly to “an unintended byproduct of US export controls.” The controls helped produce the workaround innovation that policymakers now cite as a reason for more controls. 

Even the chip restrictions appear to function mainly as expensive speed bumps. U.S. policy has pushed Chinese firms to replace American suppliers. Chinese chip stocks have rallied on expectations for Huawei accelerators, while surveys show Chinese companies shifting away from Nvidia toward domestic alternatives. 

Chinese chips may remain a generation behind, but the gap is narrowing. China also has access to far more abundant energy. At sufficient scale, weaker chips paired with more electricity can still power highly capable systems. 

The greater long-term risk is that firms around the world adopt a Chinese-centered hardware and software stack. That would matter far more than modest pressure on the profit margins of U.S. chipmakers. 

Remote access makes the controls still less effective. Even if chip restrictions worked exactly as intended, illicit access to U.S. models is already widespread. Laboratories use fraudulent accounts both to distill models and for ordinary daily access. Anthropic disclosed that three Chinese laboratories, including Moonshot AI, had harvested more than 16 million Claude exchanges through roughly 24,000 fraudulent accounts. 

The United States has badly miscalculated its AI export controls. Policymakers should focus instead on AI governance, and governance must extend beyond model alignment. Firms that provide access to models may be the least-cost avoiders, meaning the parties able to prevent or limit misuse at the lowest cost. 

The legal tools available to carry out Secretary Bessent’s threat point in the same direction. Procurement bans, information and communications technology and services orders modeled on the restrictions against Kaspersky, and Entity List designations all target transactions, services, or distribution channels. 

An Entity List designation restricts exports to a listed party. It does not prevent a U.S. user from downloading a publicly available model. Freely published software also generally falls outside the Export Administration Regulations. 

An attempt to ban a free model file under the International Emergency Economic Powers Act (IEEPA) would face the statute’s exemption for informational materials. Courts relied on the same provision when they blocked parts of the 2020 TikTok restrictions. 

Whatever the merits of those measures, restrictions aimed at services and transactions have the strongest chance of surviving judicial review. Both the law and the economics point toward the capability layer, where firms provide model access and can observe how customers use it. 

Police the API, Not the Model

Export controls and bans on open-source models will create more problems than they solve. That does not mean policymakers should do nothing. It means policy should prepare for the likeliest future, one in which defensive AI must be widely available to firms and individuals. 

Governance should therefore focus on the capability layer, where model outputs become action. Consider distillation, in which an adversary harvests API outputs to train a substitute model. Security researchers have documented this attack method since at least 2016. 

The API owner is best positioned to detect industrial-scale extraction. It can see account creation, payment signals, proxy use, and query patterns long before any regulator can. The government’s own allegations against Moonshot describe precisely that route. Moonshot allegedly used fraudulent, detection-evading access to a U.S. laboratory’s API to conduct distillation at industrial scale. That would be a failure of API security, not a leak caused by open weights. 

The laboratory also has access, at least in theory, to the most advanced AI-security tools available. Restricting open models while tolerating weak API security would create moral hazard. It would shield incumbents from the competitive consequences of their own security failures, weaken their incentives to fix those failures, and leave the public bearing the remaining risk. 

The Hugging Face incident makes the problem hard to ignore. Before frontier laboratories ask policymakers to cordon off the world’s open models, they should show that they can keep their own agents inside their own evaluation sandboxes. Stronger laboratories, tighter APIs, and better security monitoring will move faster than legislation and target the actual vulnerability more precisely than a ban. They will also improve through repeated use. 

The framing of the news reveals another blind spot. OpenAI made headlines because its model escaped a sandbox and exploited outside systems. The public heard far less about whether the same model could build a system secure enough to resist that attack. The industry devotes enormous attention to demonstrating offensive capability and far too little to the defensive capabilities AI should provide. 

Open weights should therefore be treated as part of the defense, not merely as a source of risk. Distributed threats require distributed protection. A $200-a-month frontier subscription cannot serve as the security layer for billions of devices. 

Local, auditable, open models may offer the only economical way to provide phishing detection, log review, and endpoint triage on that scale. That is one reason the Defense Advanced Research Projects Agency (DARPA) released the cyber-reasoning systems developed through its AI Cyber Challenge as open source. Regulators should preserve a healthy mix of open and proprietary models so defensive tools can spread as widely as the threats they address. 

Open-weight competition constrains prices, shifts value across the technology stack, and broadens access to defensive tools. The governance agenda is less dramatic than a sweeping ban. Harden the laboratories. Police the APIs. Monitor the physical infrastructure. Litigate trade grievances as trade grievances. Then let competition do the rest.

The first rule of AI governance should be simple. Secure the door before banning the key. 

The post Open Models, Closed Minds: AI Policy Keeps Regulating the Wrong Thing appeared first on Truth on the Market.

The Commission imposed €460 million for “self-preferencing,” or giving Google’s services prominent placement in search results, and €430 million for limiting how app developers direct users to cheaper offers outside Google Play.

Those nine-figure penalties reflect a broader shift in European competition policy. The DMA treats Google’s size and commercial success as evidence of an unfair imbalance that regulation must correct. The rest of the story turns on what Brussels means by “fair.”

The Crime of Stocking Your Own Shelves

Self-preferencing is a routine form of competition. Supermarkets place their house brands at eye level beside rival products. Mechanics sell the parts they install. Restaurants pour their own house wine. Amazon lists AmazonBasics products alongside competing brands. Digital storefronts such as Steam, the PlayStation Store, the Nintendo eShop, and Apple’s App Store take a share of the transactions they facilitate. 

These practices can help consumers or harm them. The facts determine which. 

Traditional competition law required the European Commission to show that self-preferencing excluded rivals and harmed consumers. It spent much of a decade trying to make that case in Google Shopping. 

The DMA dispenses with that inquiry. Article 6(5) prohibits self-preferencing. Article 5(4) prohibits restrictions on “steering,” or directing users to offers outside a platform. The Commission need not show consumer harm, weigh efficiencies, or consider free-riding, which occurs when one business benefits from another’s investments without paying for them. Nor must it consider how the rule may weaken incentives to build new products or services.

The conduct is unlawful because a company the DMA designates as a gatekeeper engaged in it.

The Original Sin of Success

What changed was the yardstick. The DMA is less a conventional competition law than a redistribution regime built around its own definition of “fairness.” Traditional antitrust asks whether conduct harms competition and consumers. The DMA starts with the premise that large platforms possess too much bargaining power.

The European Commission’s own impact assessment makes that premise explicit. It describes gatekeepers’ “incomparable economic strength” as evidence that their relationships with business users are “imbalanced.” The document points to app stores that earned more than $83 billion in 2019, a social network that earned more than $70 billion, rising markups, and soaring post-pandemic stock valuations. Size and profitability appear less as evidence of commercial success than as proof of unfairness. 

Many of the supposedly dependent firms are hardly helpless. Amazon, Microsoft, Epic Games, Samsung, Sony, and Booking.com rank among the world’s largest companies. The DMA nonetheless casts them as the weaker party when they deal with a designated gatekeeper. 

With apologies to Hegel, one can call the mechanism “Dialectical Fairness.” The DMA seeks balance by weakening the stronger party and strengthening the weaker one. Because it treats the gatekeeper’s size, profitability, and bargaining power as the source of unfairness, reducing those advantages becomes part of the remedy. 

Seen through that lens, the Google decisions make more sense. Featuring Google’s hotel results or flight prices is unfair because it may strengthen a company the DMA already regards as too powerful. Consumer benefits do not settle the question. They may deepen the problem if they draw more users to Google and make it still more profitable. 

Under that theory, success converts ordinary business conduct into an offense. 

Fairness by the Click

Consider what Google is accused of doing in Search. It gives users direct answers such as live football scores, hotel listings, and flight prices instead of sending them through a page of blue links. Under ordinary competition principles, that looks like winning on the merits by building a better product. 

The Commission prefers a longer route. It wants Google to display more links to rival services, even when that makes Search slower or less useful, because those links redistribute traffic—and the advertising revenue that follows it. 

That is the DMA’s version of fairness. It judges competition by who receives the click rather than whether users receive a good answer. A seamless product becomes suspect when its convenience strengthens the gatekeeper instead of its rivals. 

I have written before about how this logic turns a law advertised as pro-competition into a mandate for fragmentation. Across the DMA’s enforcement record, users’ immediate interests—speed, convenience, and useful answers—receive little weight. The rules treat those benefits as a problem when they make Google’s product more attractive. 

Brussels Knows a Fair Price When It Sees One

The self-preferencing decision quietly recasts Google as a quasi-public utility. In related compliance proceedings, the European Commission has said Google must weigh “the interests of all participants in the market,” including merchants, hotels, airlines, train operators, and rival search services. Trade associations had complained that changes to Search reduced their visibility. 

Under that standard, Google becomes responsible whenever a product change costs another business traffic. A private company must protect its competitors’ commercial fortunes while improving its own service. Every change must leave merchants, intermediaries, and rivals at least as well off, even when users prefer the new product. 

The Google Play decision reveals the same impulse. The Commission accepts that Google may charge an “initial acquisition” fee for connecting a developer with a new customer. It then declares Google’s fee too high. 

That judgment turns the Commission into a price regulator. Markets usually determine whether a fee is sustainable. The DMA provides no benchmark for a lawful amount, no governing principle, and no answer to the obvious question: Too high compared with what? 

The Commission simply knows an unfair price when it sees one. 

When Winning Becomes the Violation

To be sure, digital platforms can harm consumers, and gatekeeper practices deserve scrutiny. The problem is that the DMA changes the question. Competition law asks whether conduct leaves consumers worse off. The DMA asks whether Google has gained too much and its rivals too little. 

Once that becomes the test, the usual signs of competition—a better product, a cleaner interface, a market-tested fee, and billions of returning users—become evidence against the winner. 

Competition policy once rewarded winning on the merits. The DMA now fines it.

The post The Crime of Winning: How Europe’s DMA Punishes Google for Competing appeared first on Truth on the Market.

Last month, Judge Brian Cogan of the U.S. District Court for the Eastern District of New York gave preliminary approval to the third attempted settlement in the two-decade antitrust fight between merchants and payment networks. In re Payment Card Interchange Fee and Merchant Discount Antitrust Litigation, the sprawling multidistrict case filed in 2005, has already produced two rejected settlements, one 2nd U.S. Circuit Court of Appeals vacatur, a $5.54 billion damages fund, and perhaps more economist hours than any private antitrust case in American history. 

The latest agreement, announced in November and valued by court-appointed experts at roughly $38 billion through 2031, would cut average credit-card interchange fees by 10 basis points for five years. A basis point is one-hundredth of a percentage point. The deal also would cap fees on standard consumer credit cards at 1.25% for eight years, a reduction of more than 25%, and freeze posted fees at their March 2025 levels. 

The most consequential change concerns the “honor-all-cards” rule, which generally requires merchants that accept a network’s cards to accept all cards in that category. The settlement would divide acceptance into three groups—commercial, premium consumer, and standard consumer—and allow merchants to accept or reject each group separately. It also would expand merchants’ ability to impose surcharges on credit-card transactions.

Big-box retailers remain unimpressed. The National Retail Federation and the Merchants Payments Coalition oppose the deal, arguing that merchants would still pay too much, especially on rewards cards. Some analysts expect appeals that could delay final resolution until 2029. 

That would be a pity. The settlement would distort the economics of card payments, and the litigation beneath it remains highly dubious. Still, it would end a case that has consumed more than two decades. It also compares favorably with the legislative “solutions” promoted by large retailers, including state efforts to exempt taxes and tips from interchange fees and the preposterously named federal  Credit Card Competition Act. 

Interchange Fees and the Art of Keeping Both Sides Happy

Start with the basic economics of payment networks and the role of interchange fees. In a four-party card system, the merchant’s bank pays a small share of each transaction—typically 1% to 2%—to the cardholder’s bank. That payment is the interchange fee. 

A common misconception treats interchange as little more than compensation for processing a transaction. Its role is much broader. Interchange arose as a practical solution to a coordination problem in the original BankAmericard system and later became a way to balance the interests of merchants and consumers. 

When BankAmericard, the forerunner of Visa, launched in 1958, it operated as a three-party network. Bank of America issued the cards, signed up merchants, and processed the transactions. 

That changed when Bank of America began licensing other banks to issue BankAmericard cards in 1966. At first, the amount an acquiring bank paid to the issuing bank depended on the merchant discount the acquirer charged, or claimed to charge, sometimes adjusted for processing costs. The arrangement proved difficult to administer and audit. 

Merchant discounts varied across banks and merchants, which gave acquirers an incentive to understate them. The formula also left acquirers with little or no margin on transactions involving cards issued by other banks. Issuers could receive too little, acquirers had weaker incentives to recruit and serve merchants, and the system encouraged strategic misreporting. 

In 1970, Bank of America and its licensees created National BankAmericard Inc., an independent company owned by participating banks. The new organization replaced the uncertain formula with a uniform interchange reimbursement fee. 

Set at 1.95% in 1971, the standardized fee separated the issuer’s payment from the merchant discount negotiated by the acquirer. The acquirer could retain the difference between the merchant discount and the interchange fee to cover its own costs. 

Interchange therefore began as a way to align the incentives of independent issuers and acquirers in an open payment network. The Interbank Card Association, which operated MasterCharge, the precursor to Mastercard, adopted a similar approach about a year later. Economists later developed a fuller account of why the mechanism worked. 

William Baxter’s seminal 1983 paper explained that interchange helps payment networks solve the central problem of a two-sided market. A card network must attract enough merchants and consumers at the same time. Each side becomes more valuable as participation on the other side grows. 

Consumers benefit from convenience, less need to carry cash, and the ability to spend beyond the money in their wallets. Credit cards can also provide short-term financing, rewards, and various forms of insurance. 

Merchants benefit from higher sales, lower cash-handling costs, and access to customers who value the convenience and liquidity that cards provide. Baxter argued that these gains form part of the system’s total value. Interchange helps distribute that value between merchants and consumers. 

His analysis also showed that the optimal fee depends on the value each side places on an additional transaction. The fee could, in principle, be positive, with merchants helping fund cardholder benefits, or negative, with consumers subsidizing merchants. Baxter found that a positive fee will often be efficient because it lowers consumers’ effective cost of using cards through rewards and other benefits, encouraging adoption and use. 

Later economists, including Jean-Charles Rochet and Nobel laureate Jean Tirole, extended this analysis. They showed that the optimal fee depends on how sensitive each side is to price, how participation on one side affects demand on the other, and the degree of competition among issuers, acquirers, and networks. 

When merchants are less likely than consumers to leave the network in response to a price increase, the usual result is that merchants bear more of the cost. Interchange then allows issuers to fund benefits that encourage consumers to carry and use cards. 

Interchange is therefore best understood as a tool for balancing participation and increasing the network’s value. Treating it solely as reimbursement for issuer costs misses most of its function and makes cost-of-service regulation a poor fit. 

Interchange helps fund rewards, up to 45 days of interest-free credit for cardholders who pay their balances in full, zero-liability fraud protection, chargeback rights, and continuing investment in authorization, tokenization, and fraud detection. Tokenization replaces sensitive card information with a temporary digital identifier, reducing the value of stolen data. 

Merchants benefit from each of these features. Studies consistently find that card acceptance increases average transaction size and total spending relative to cash. Consumers who lack enough cash at checkout simply buy less. 

Guaranteed payment also shifts credit and fraud risk away from merchants. When a cardholder defaults, the merchant has already been paid. Card acceptance reduces the costs of handling cash, including theft, armored transport, delayed access to funds, and register reconciliation. It also speeds checkout. 

For online commerce, cards have proved indispensable. E-commerce now accounts for roughly one-sixth of U.S. retail sales. Without reliable card payments, online retail would have developed more slowly and remained much smaller. 

Merchants Want the Cards, Just Not the Price

Large retailers have spent decades trying to shrink the cross-subsidy that helped make payment cards ubiquitous, even though they benefit heavily from the system. At first glance, they seem determined to have their card-generated cake and eat it too. 

Each successful fee cut makes cards less attractive to consumers and can reduce card use over time. That would eventually hurt merchants as well. Smaller merchants would bear more of the damage because the fixed costs of acquiring and serving them make up a larger share of the merchant discount. Large retailers understand that arithmetic. 

The first major legal challenge came from National Bancard Corp., or NaBanco, an acquiring bank whose interests closely tracked those of merchants. NaBanco sued Visa to eliminate the default interchange fee, arguing that issuing banks had engaged in per se price fixing. 

Drawing on William Baxter’s work, the 11th U.S. Circuit Court of Appeals rejected the claim. The court held that a collectively set default fee was necessary for a four-party card system to function and ruled for Visa. National Bancard Corp. v. Visa U.S.A. Inc. (11th Cir. 1986). 

Merchants scored a larger victory in Wal-Mart Stores Inc. v. Visa U.S.A. Inc., which settled in 2003 for $3 billion. The settlement allowed merchants to accept Visa and Mastercard credit cards without also accepting their signature-debit cards. The broader objective was already clear. Merchants wanted lower fees without giving up card acceptance. 

The current case, known as MDL 1720, began in 2005. It rests on several unproven and highly implausible claims. The plaintiffs contend that Visa and Mastercard operate as a cartel, that default interchange fees are cartel prices, and that rules governing card acceptance, surcharges, and customer steering help enforce the arrangement. 

When litigation moved too slowly, merchants turned to Congress. In 2010, Sen. Dick Durbin (D-Ill.) added a provision to the Dodd-Frank Act that imposed price controls on debit-card interchange fees. 

Retailers then sued the Federal Reserve for setting the cap too high in NACS v. Board of Governors of the Federal Reserve System. After losing, they revived the challenge a decade later in Corner Post Inc. v. Board of Governors of the Federal Reserve System. A district court ultimately vacated Regulation II, finding that the rule allowed issuers to recover too much. The case is now before the 8th U.S. Circuit Court of Appeals. 

Europe followed a similar path. A 1992 complaint by the British Retail Consortium helped launch the enforcement campaign that produced the European Commission’s 2007 Mastercard decision, the European Union’s 2015 Interchange Fee Regulation, and a series of English damages cases stretching from Sainsbury’s Supermarkets Ltd. v. Mastercard Inc. to Merricks v. Mastercard Inc. 

Large retailers have pursued the same objective through every available institution, including antitrust suits, legislation, and regulation. They want the government to force interchange fees lower. 

Their conduct tells a more complicated story. After two decades of insisting that interchange fees exceed the value of card acceptance, almost no major U.S. retailer has stopped accepting the cards. 

Kroger briefly dropped Visa credit cards at two regional chains in 2018 and 2019. The experiment lasted only months before the company reversed course. In opposing the current settlement, the National Retail Federation’s general counsel explained why retailers will not reject even the most expensive rewards cards. More than 80% of customers carry them, she said, and refusing them would cost merchants substantial business. 

Quite so. 

Merchants continue buying the service at the posted price, year after year, while insisting that it is worth far less. Their behavior suggests that card acceptance creates substantial value. Their campaign seeks to use state power to capture more of that value for themselves. 

That is rent-seeking with a consumer-protection soundtrack. 

The campaign has never aimed to restore a market price. Large retailers want legal rules that transfer a larger share of the card system’s gains to merchants, with little regard for the costs imposed on consumers, issuers, or smaller businesses. 

A Court-Ordered Discount on Other People’s Cards

Against that backdrop, consider the settlement’s design. 

The rate provisions amount to price controls. Average credit-card interchange fees would fall by 10 basis points for five years. Much of that reduction would likely come through the eight-year cap of 1.25% on standard consumer cards, roughly 25% below current average rates. All other posted rates would remain frozen at their March 2025 levels. 

The three-tier acceptance framework would largely dismantle the honor-all-cards rule. Merchants could accept standard consumer cards while rejecting premium rewards cards or commercial cards. 

Whether many merchants will do so remains doubtful. Rewards cards are common, and their benefits encourage cardholders to spend more. A merchant that rejects them risks losing those customers. Most consumers also carry only one credit card, so declining it may mean losing the sale altogether. 

The settlement would also give merchants much more freedom to impose surcharges, including different surcharges for different cards. Australia’s experience suggests that merchants with unusual offerings or captive customers—such as taxicabs, airlines, and event-booking companies—are the most likely to use that power to extract higher payments from consumers. 

Combined with selective card acceptance, differential surcharges could also steer customers toward cards with lower interchange fees. 

Large retailers would therefore keep much of the spending, convenience, and risk-shifting benefits of card acceptance while contributing less to the benefits that encourage consumers to carry and use cards. They would obtain that transfer through a court-supervised shakedown. 

When Swipe-Fee Savings Skip the Shopper

Contrary to large retailers’ claims, interchange fees do not create a pile of idle profits waiting to be redistributed. They allocate costs and benefits across the two sides of the card market. Merchants pay more than the narrow cost of processing a transaction, and issuers return part of that difference to consumers through rewards, fraud protection, interest-free credit, and other cardholder benefits.

When regulators force that difference lower, those benefits shrink.

The Durbin Amendment provides the best-studied example. After the Federal Reserve implemented it through Regulation II in 2011, debit-card interchange fees at covered banks—those with at least $10 billion in assets—fell by roughly half. Banks lost more than $6 billion a year in revenue.

They responded quickly. Most covered banks eliminated debit-card rewards programs, reduced the availability of free checking for customers with low balances, and raised monthly maintenance fees. Hundreds of thousands of lower-income customers then left the banking system.

Durbin also required all issuers, large and small, to place at least two unaffiliated payment networks on each debit card. That mandate sharply reduced interchange fees on personal identification number, or PIN, debit transactions because smaller networks were not bound by the default fees negotiated between issuing banks and the major national networks, Visa and Mastercard.

As I noted previously:

Following the introduction of mandatory “competitive routing” on debit cards, smaller PIN-debit networks saw a profit opportunity. But those networks were not focused on maximizing the value of the system, so they were willing to carry payment messages at a lower interchange rate than the major networks. Smaller issuing banks were forced to accept these lower PIN-debit interchange fees, and the major networks were forced to cut their PIN-debit fees to remain competitive. As a result, many smaller banks have experienced reductions in interchange revenue similar to their larger cousins and have responded similarly—by reducing the availability of free checking accounts.

The promised savings for consumers largely failed to appear. Surveys and empirical studies found that most merchants kept prices unchanged. Some raised them.

Consumers also changed how they paid. Many shifted from debit cards to credit cards used for routine purchases. That effect appeared mainly among consumers with strong credit scores who qualified for premium cards. 

A law promoted as consumer protection instead transferred billions of dollars each year from bank shareholders and customers, disproportionately lower-income customers, to large retailers and their shareholders.

Australia’s experience followed the same pattern. After the Reserve Bank of Australia imposed interchange-fee caps in 2003, rewards declined, annual card fees rose, and any reduction in retail prices proved too small or diffuse to detect.

Liberalized surcharging created another problem. Merchants imposed enough add-on fees that the Reserve Bank first capped “excessive” surcharges and, by 2025, proposed banning card surcharges altogether.

Australia also saw substitution. Banks began offering premium customers American Express-branded “companion cards” with generous rewards. American Express gained about 25% in market share before the Reserve Bank brought those cards under the same price controls.

The European Union’s experience after its Interchange Fee Regulation points in the same direction. Interchange caps reduced cardholder benefits and shifted costs elsewhere. The gap between credit-card annual percentage rates (APRs) and the European Central Bank’s base rate widened, suggesting that banks passed at least some of the burden to customers who carried balances.

Retail-price savings remained meager. As under Durbin and in Australia, merchants kept most of the gains.

The Savings Come With a Smaller Pie

There is little reason to expect this settlement to produce different effects, though their magnitude may be smaller. Issuers facing fee caps and possible rejection of premium cards will likely trim rewards, raise annual fees, tighten approval standards at the margin, and shift their portfolios away from newly unprofitable customers. 

Surcharging carries its own costs. It slows checkout, can prompt customers to abandon purchases, and annoys consumers by making the fee highly visible. Some customers will switch to payment methods they value less to avoid a surcharge that may exceed the merchant’s own card-acceptance cost.

Three-party networks such as American Express and Discover are outside the settlement. They will remain free to use merchant revenue to fund cardholder benefits and may respond by offering rewards that Visa- and Mastercard-branded cards can no longer match.

That competition could soften some of the harm to consumers, though at the price of added inconvenience and duplication. Consumers may need another card to retain the same benefits. Any shift toward American Express or Discover would also erase part of the savings merchants expect from the settlement.

Each response weakens cardholders’ incentives to carry and use Visa- and Mastercard-branded cards. Less generous rewards will reduce card use at the margin. Lower use will shrink the increase in transaction size associated with card payments and reduce merchants’ total sales.

Lower transaction volume would also spread the networks’ enormous fixed costs across fewer payments. That could reduce investment in payment technology. Recent advances such as tokenization, contactless payments, and real-time fraud scoring have made transactions faster and safer, especially online. Slower investment would make further improvements less likely.

Some spending will shift to three-party cards, but that substitution will provide only a partial offset. The likely result is less generous cardholder benefits, lower card use, less consumer spending, and slower payment innovation. Consumers and merchants would both lose some of the commerce that interchange-funded benefits now encourage.

The Least Bad Way to End a Bad Case

All that said, the proposed settlement would likely do far less harm than the legislative alternatives now on offer.

As we have documented elsewhere, the Illinois Interchange Fee Prohibition Act and similar state proposals would exempt taxes, tips, or other transaction components from interchange fees. Implementing those carveouts would require extensive changes to payment-network infrastructure and could fragment a national system into state-by-state payment rules.

The settlement may reduce issuer revenue more than those laws would, but it creates far fewer compliance costs. It also leaves the basic payment infrastructure intact. 

The proposed Credit Card Competition Act (CCCA) offers an even starker comparison. Like the settlement, the act assumes that payment networks lack sufficient competition and that interchange fees are therefore inflated. Starting from that mistaken premise, it would require issuers with more than $100 billion in assets to make every credit-card transaction routable over at least one unaffiliated network.

That would amount to the Durbin Amendment for credit cards, with added hazards. Credit transactions bundle underwriting, revolving credit, fraud guarantees, and dispute rights. The issuer and network allocate those risks through their contractual relationship.

A second network may not price the credit risk, bear the fraud liability, or support the same security systems. Requiring issuers to route transactions through such a network could create large and poorly assigned liabilities.

Durbin already showed what happens when routing mandates compress interchange fees. Rewards disappear, account fees rise, and access tightens. Applied to credit cards, the pressure would fall on rewards, credit availability, and fraud-prevention investment, where the potential losses exceed those associated with debit cards.

Merchants claim that payment networks suffer from too little competition. Both the settlement and the CCCA would weaken competition where consumers experience it most directly—among card issuers offering different rewards, rates, fees, and benefits.

Artificially lower interchange fees would make it harder for issuers to sustain that variety. The settlement would likely produce a modest contraction. The CCCA would impose a much sharper one by compressing interchange fees across the market.

By comparison, the proposed Settlement would likely have only a modest effect.

The settlement also has the virtue of an expiration date. Its rate provisions last five or eight years, after which networks regain pricing freedom. The CCCA would continue indefinitely unless Congress repealed it. The settlement bends the price mechanism. The act would come much closer to breaking it.

The settlement would still advance large retailers’ long-running campaign. It would thin the cross-subsidies that fund cardholder benefits, participation, and payment innovation. At the margin, it would reduce some of the commerce that payment cards generate for merchants and consumers alike.

Yet the settlement remains a defensible resolution. It is temporary, negotiated, and less destructive of the card system’s pricing structure than any legislative alternative under serious consideration. It may also give lawmakers a reason to leave the industry alone.

After 21 years of litigation, peace has value. The trick will be convincing Congress to let it stand.

The post The Price of Peace in the Swipe-Fee Wars appeared first on Truth on the Market.

On July 20, Gov. Mikie Sherrill signed the Forbidding the Algorithmic Inflation of Rent Act, or FAIR Act, declaring that landlords who use shared pricing tools are engaging in “collusion by algorithm.” The phrase is built for a press release. As an antitrust standard, it sweeps far too broadly.

Antitrust law already has a clear target. A software vendor can serve as the hub of a cartel by collecting competitively sensitive information, relaying rivals’ plans, pressuring users to accept common prices, restricting discounts, or helping participants detect and punish defections. If landlords use software to carry out an agreement that would be illegal around a conference table, the software offers no immunity. The U.S. Justice Department’s (DOJ) RealPage case and proposed settlements show the kind of conduct that warrants close scrutiny.

But shared pricing software, common data, and even some use of nonpublic information do not, by themselves, establish collusion. New Jersey has replaced a difficult, fact-intensive inquiry with a sweeping ban. The law may suppress tools that improve pricing accuracy, reduce costly errors, increase capacity use, and help firms respond to changing conditions.

The state has targeted a technology because it can facilitate unlawful coordination. Antitrust law should target the agreements and practices that suppress independent rivalry.

The rule is simple enough. Prosecute collusive agreements and the mechanisms that sustain them. Do not ban computation, common code, or nonpublic data merely because several firms use them.

When Similar Prices Become a Crime

The FAIR Act reaches far beyond a vendor telling two landlords to charge the same rent. It defines an “algorithmic device” broadly and bars rental owners from paying for or using the services of a “coordinator.”

A coordinating function includes collecting competitively sensitive information from multiple owners and using it to recommend rents, lease terms, or occupancy levels. It also covers setting terms based on another owner’s sensitive information. More vaguely, it reaches recommendations made to two or more owners through the same or a substantially similar algorithm when those recommendations facilitate “parallel pricing coordination.”

The law’s treatment of information is broader still. “Nonpublic” information generally means information unavailable to the public at no cost. A dataset that mixes public and nonpublic material counts as entirely nonpublic. The statute also reaches tacit coordination, which may be inferred from a pattern of parallel conduct.

Those provisions elevate two weak proxies into major grounds for liability. One is the use of data that costs money. The other is the appearance of similar conduct among competitors.

The FAIR Act does not ban every algorithm a landlord might use. A unilateral tool trained only on the owner’s own data, or a basic spreadsheet that requires human analysis, may fall outside its core. In practice, though, the law creates something close to a categorical ban on a large class of third-party revenue-management products.

A vendor that serves several landlords must now worry that individualized recommendations produced by similar code will be treated as parallel pricing coordination. That risk remains even when users can reject the recommendations and never see or learn their rivals’ data.

Other states have taken different approaches. New York’s 2025 law treats it as an unlawful agreement when a landlord knowingly or recklessly relies on a coordinating algorithm that collects data from multiple owners, processes that data, and recommends rental terms.

California chose a more disciplined rule in Business and Professions Code Section 16729. It makes use of a common pricing algorithm unlawful when the use forms part of a contract, combination, or conspiracy that restrains trade, or when one person coerces another to adopt the algorithm’s recommendation. California at least keeps agreement or coercion at the center of the offense.

That distinction is fundamental to antitrust law. Courts have long separated unlawful concerted action from lawful conscious parallelism. Competitors often respond in similar ways to common costs, demand shocks, regulation, interest rates, or publicly visible prices. Parallel conduct may support an inference of agreement, but it does not itself prove one.

New Jersey threatens to blur that line whenever software makes parallel behavior easier to detect.

The Code Is Not the Cartel

Former Federal Trade Commission (FTC) Acting Chair Maureen Ohlhausen made the central point in her 2017 speech, “Should We Fear the Things That Go Beep in the Night?” Algorithms can serve benign or malign ends, and familiar antitrust principles can usually tell the difference.

A cartel does not become lawful because its members communicate through code. Independent pricing does not become collusion merely because computers help firms process information faster.

Some skepticism remains warranted. Laboratory studies have found that reinforcement-learning agents can sometimes produce prices above competitive levels in repeated-game settings, even without explicit instructions to collude. Emilio Calvano and his coauthors, for example, reported in the American Economic Review that certain pricing algorithms learned strategies that sustained supracompetitive prices.

Other studies have reached different results depending on market structure, algorithm design, and the form of the recommendation. Algorithms are not a single species. Their competitive effects depend on how they work and how firms use them.

Antitrust analysis should focus on those details. Does the provider secure a common commitment from rivals? Does the system constrain or punish deviations? Can users monitor one another’s conduct? Does the vendor transmit firm-specific plans? Do firms communicate about pricing strategy? Is adoption so widespread in a concentrated market that the provider can discipline competition?

Those facts help distinguish software that facilitates an agreement from software that merely improves decision-making.

By contrast, little follows from the fact that several firms use the same vendor, receive recommendations from the same codebase, or respond similarly to market conditions. Businesses routinely rely on common accounting software, cloud infrastructure, payment processors, consultants, and data providers. Shared inputs do not ordinarily convert independent decisions into concerted action.

Pricing software should receive the same treatment unless its design or use supplies the missing agreement.

The Rent Is High for a Reason

The case for restraint goes beyond avoiding false positives. Algorithmic pricing can improve how markets work.

Friedrich Hayek’s classic essay, “The Use of Knowledge in Society,” explains that economic knowledge is scattered among millions of people and often concerns fleeting facts about time and place. Prices transmit that information without requiring any central planner to gather it all. A price change can reflect scarcity, abundance, shifting preferences, or higher costs, and it prompts people to adjust.

Rental housing depends on this kind of local knowledge. The economically relevant price of an apartment turns on the unit’s features, vacancy rates, lease length, expected turnover, seasonality, concessions, maintenance costs, nearby construction, neighborhood demand, and how much renters value moving now rather than later.

A human manager can process some of that information. An algorithm may process more of it, more consistently, and with less delay.

That does not turn the market into a centralized plan. A well-designed model can make prices more responsive. If demand weakens, it may recommend a lower effective rent or a larger concession before a unit remains vacant for months. If a local shortage emerges, it may reveal the value of adding units, renovating marginal properties, or directing investment toward the constrained area.

More accurate prices can reduce vacancies and shortages, improve matches between renters and units, and reveal opportunities for mutually beneficial exchange.

I recently made this point in Truth on the Market. Legal rules that make firms afraid to use pricing tools can reduce market efficiency at the expense of both producers and consumers. Economists generally define algorithmic pricing as the automated use of software to adjust prices in response to information. That automation can reduce decision costs, respond quickly to changes in inventory and demand, and correct the inertia and rough rules of thumb that often keep prices wrong for too long.

An efficient price is not always a low one. When housing is scarce, an accurate price may be high. Suppressing the price signal does not produce more apartments. It may instead lead to longer searches, arbitrary rationing, poorer maintenance, lower investment, or hidden nonprice terms.

New Jersey itself recognizes that its housing agenda must include more construction and land-use reform. The state’s builders are therefore right about the order of causes. Algorithms did not create zoning restrictions, construction costs, interest rates, or the shortage of developable land.

Blaming the messenger may create the appearance of action while leaving the shortage intact.

Private Data Is Not a Smoking Gun

The strongest economic critique of this emerging legal approach comes from economist Jay Ezrielev’s “Premature Antitrust Standards in Algorithmic Pricing.” Ezrielev focuses on common-data algorithms, which collect information from multiple firms and use a shared model to generate individualized recommendations.

Courts and lawmakers increasingly treat pooled nonpublic data as a near-conclusive warning sign. Ezrielev explains why that shortcut fails. The mere fact that information is nonpublic says little about how collusion would occur.

The traditional concern with sharing competitively sensitive information is straightforward. Rivals may use it to identify one another’s prices or output, monitor compliance with a cartel, and punish cheating. But a combined dataset visible only to an algorithm may give users no way to observe a rival’s conduct. The software may produce a recommendation without revealing any firm-specific data behind it.

Under the right conditions, aggregating nonpublic data can also improve competition. Sparse or noisy local markets are hard to forecast using public listings alone. Combined data on occupancy, renewals, cancellations, and transactions may help a model detect demand changes that no single owner could identify reliably.

Better forecasts can reduce vacancies, improve the timing of concessions, increase occupancy and output, and lower the cost of pricing errors. Those gains can benefit landlords and renters alike.

The same principle appears throughout the economy. Firms routinely entrust sensitive information to accountants, lawyers, investment bankers, insurers, cybersecurity vendors, and cloud providers. The quality of those services may improve as the intermediary gains broader experience. Antitrust law usually asks whether the arrangement weakens incentives to compete or enables coordination. It does not infer a cartel merely because several clients share private information with the same intermediary.

Ezrielev also identifies a paradox in the public-versus-nonpublic distinction. If legal risk depends on whether data are publicly available, firms may respond by publishing more of it. Yet public, firm-specific information can make tacit coordination easier by allowing rivals to detect price cuts and departures from a common pattern almost immediately.

A rule designed to prevent monitoring may therefore encourage the transparency that makes monitoring easier.

None of this makes nonpublic data irrelevant. Current, detailed, rival-specific information poses greater risks when a vendor reveals it to users, creates dashboards that allow reverse engineering, brings competitors together to discuss strategy, or uses the data to enforce a common objective.

Those are questions about design and conduct. They require examining access, aggregation, delay, anonymization, recommendation structure, and incentives. They do not justify New Jersey’s blunt rule that data unavailable to the public at no cost are presumptively suspect, much less its decision to treat any mixed dataset as wholly nonpublic.

The economically meaningful distinction is whether the system helps competitors coordinate and police a common plan or helps each user make a better independent decision.

A Ban Built for the Biggest Landlords

Broad bans also impose long-term costs that disappear in a debate focused on this month’s rent.

Pricing is a major commercial use of machine learning because it generates frequent feedback. Developers can test forecasts, learn from errors, improve data systems, and build better tools for inventory, logistics, capacity planning, and demand estimation. Rules that exclude useful data or expose a provider to liability because two customers receive recommendations from similar models reduce the payoff from that experimentation.

The burden will fall unevenly. Large landlords can build proprietary systems using their own portfolios, engineers, and legal departments. Smaller owners are more likely to rely on third-party tools that spread development costs across many customers.

A ban on shared platforms may therefore protect firms large enough to develop the technology in-house while denying smaller rivals comparable capabilities. A rule sold as a check on large corporate landlords could end up strengthening them.

A state-by-state patchwork makes matters worse. Software providers build products for national markets. Faced with conflicting definitions of “nonpublic data,” “coordination,” and “algorithmic device,” they may design every product to satisfy the strictest state or stop serving smaller markets altogether.

Startups will struggle more than incumbents with those fixed compliance costs. The likely consequences include less entry, slower model improvement, and fewer experiments with tools that might lower costs or expand output.

That result also conflicts with the federal push for leadership in artificial intelligence. The White House’s “America’s AI Action Plan” identifies faster innovation and adoption as central to economic competitiveness. One need not endorse every part of that plan to see the tension. The federal government wants firms to develop and deploy artificial intelligence, while states prohibit major classes of learning and decision tools without requiring proof of competitive harm.

Antitrust errors do more than raise compliance costs today. False positives can redirect research spending and determine which technologies reach the market tomorrow.

Police the Cartel, Not the Code

The answer is not to ignore algorithmic cartels. It is to adopt a federal framework that targets the mechanisms of coordination while leaving room for beneficial experimentation.

First, antitrust law should keep agreement and competitive effects at the center of the analysis. An explicit arrangement among competitors to accept a common price, restrict output, or use a vendor to enforce discipline should face per se condemnation. A novel software system that generates recommendations while preserving independent decision-making should ordinarily receive rule-of-reason analysis. Courts should examine market power, adoption, actual operation, and efficiencies.

Parallel outcomes and common software may support an inference of collusion when paired with plus factors, meaning evidence that makes independent conduct less plausible. They should not replace proof of concerted action.

Second, the DOJ and FTC should issue guidance on algorithmic pricing. The agencies have already opened a 2026 inquiry into updated competitor-collaboration guidance and identified algorithmic pricing and information sharing as subjects for review. Any resulting guidelines should distinguish dangerous features from meaningful safeguards.

High-risk features would include a common commitment to follow recommendations, coercion or penalties for deviation, access to current and detailed rival-specific information, tools that expose deviations, communications among users about pricing strategy, uniform limits on discounts or price cuts, and broad adoption in a concentrated market.

The analysis should also consider the provider’s incentives, how often users receive and accept recommendations, whether outputs are individualized, and whether the system influences output or occupancy as well as price.

The agencies should pair that list with a rebuttable safe zone for systems that preserve independent rivalry. Relevant safeguards would include no user access to rival-specific data, aggregation and anonymization, appropriate delays, individualized objectives and outputs, meaningful freedom to reject recommendations, no penalties for doing so, no vendor-facilitated meetings among competitors, strong firewalls and audit logs, and documented efficiency justifications. Nonpublic data should count as one factor, not a forbidden category.

Third, federal enforcers should use statements of interest to clarify doctrine, rather than only to expand liability. The DOJ and FTC filed a 2024 statement of interest in hotel-pricing litigation explaining that an algorithm cannot immunize conduct that would otherwise violate antitrust law. That principle is sound.

Future filings should add its necessary limit. A common algorithm does not erase Section 1’s agreement requirement, and pooled nonpublic data do not by themselves establish a naked restraint. Carefully chosen filings could steer courts away from per se treatment of unfamiliar arrangements before economic evidence and experience support it.

Fourth, Congress should consider targeted preemption. A federal law could displace state and local rules that impose liability solely because firms use the same pricing algorithm, receive similar recommendations, or contribute nonpublic data without proof of agreement, coercion, or a likely anticompetitive mechanism.

States would remain free to enforce antitrust and consumer-protection laws against actual collusion, deception, discrimination, and unfair practices. The goal would be a coherent national baseline for interstate software and data services, not immunity for algorithms.

Preemption makes sense because fragmented state bans can impose costs beyond state borders. A prohibition in one large jurisdiction may dictate product design nationwide, discourage entry, and deny consumers elsewhere access to useful tools. Congress often adopts national rules when interstate commerce and networked technologies make 50 conflicting regimes unusually costly. Algorithmic pricing warrants the same consideration.

Finally, enforcement should remain empirical and open to revision. Agencies should study actual effects on prices, output, vacancies, quality, and entry before imposing permanent design rules. Remedies should address demonstrated sources of harm, such as rival-specific disclosures, coercive acceptance requirements, or monitoring tools, rather than banning entire categories of data.

Sunset provisions, regulatory sandboxes, and retrospective reviews would allow regulators to revise their approach as the technology and evidence develop.

Collusion Still Requires Collusion

New Jersey’s FAIR Act begins with a legitimate concern and ends with the wrong legal design. Algorithms can facilitate collusion. So can trade associations, consultants, phone calls, and private meetings. Antitrust law should target the agreement and the mechanisms that sustain it, rather than the general-purpose tool that carries the information.

The costs will extend beyond landlords losing a convenient product. Broad bans can make prices less informative, shield large incumbents from smaller technological rivals, encourage inefficient public disclosure, slow AI research, and fracture a national software market into incompatible state regimes.

Renters may pay through fewer available units, poorer matches, higher operating costs, and reduced investment. Meanwhile, zoning restrictions, construction costs, interest rates, and limited developable land will keep doing what they were doing before the algorithm arrived.

A sound policy would prosecute genuine hub-and-spoke cartels, provide clear federal guidance, examine actual conduct, create safe harbors for independent decision tools, and preempt state laws that mistake data processing for agreement.

Preserve rivalry. Demand evidence. And do not confuse better arithmetic with a cartel.

The post New Jersey’s War on Pricing Software Won’t Build More Apartments appeared first on Truth on the Market.

That is the logic running through two decisions the European Commission adopted last week involving the same company, under the same law, on the same day. Yet read side by side, they seem to come from different legal universes. 

The first measure, issued under Article 6(11) of the Digital Markets Act (DMA), requires Google to share anonymized search data with rival search engines and AI chatbots. The Commission surrounded that dataset with an elaborate system of safeguards. Identifiers must be removed. Access is delayed by at least a week. Rare or revealing queries are excluded. Eligible firms must pass screening, undergo an independent audit before receiving any data, and submit to annual audits thereafter. They must also use ring-fenced processing environments, comply with purpose and retention limits, and clear checks for sanctions and control by high-risk third countries. 

The second decision, issued under Article 6(7), requires Google to give rival AI assistants the same deep access to Android that Gemini receives. That includes ambient sensors, on-device app data, screen contents, and the ability to control other applications. The data concern identified users, include content, and arrive in real time. The main safeguard is a consent prompt. 

One dataset gets an armed escort. The other gets a checkbox. The DMA’s internal logic can explain the difference. A consequentialist analysis has a harder time doing so. 

A Walled Garden for Data, an Open Door for Devices

Understanding the mismatch requires a closer look at what each decision does. 

The search-data decision implements Article 6(11), which requires Google to provide rivals with ranking, query, click, and view data on fair, reasonable, and nondiscriminatory (FRAND) terms. The Commission begins with technical anonymization. Google must remove direct identifiers and timestamps, suppress rare terms and unusually long queries, and place each user in a group of at least 1,000 people who share the same location, device type, and language. In the Commission’s telling, the result is a “haystack” of disconnected queries. 

The decision goes much further. Because technical safeguards can only reduce the risk of reidentification, the Commission also regulates who may receive the data. Only genuine search businesses qualify. 

A firm must have operated in the European Union for at least two years. A newer entrant may qualify if it has raised more than €50 million and serves at least 50,000 monthly European users. The decision excludes sanctioned entities and firms controlled by third countries that pose structural cybersecurity or data-protection risks. Before sharing any data, Google may also assess whether a particular recipient presents serious cyber or privacy risks. 

Recipients must pass an independent audit before gaining access and submit to annual audits afterward. They must process the data in ring-fenced environments, use it only to improve search, refrain from training general-purpose AI models with it, and delete it on schedule. 

These safeguards largely turn on the recipient’s identity and conduct. The result resembles a walled garden, the same basic architecture gatekeepers built for themselves and that the Commission is dismantling elsewhere. Yet Article 6(11) itself devotes only one word to protection: “anonymized.” 

Article 6(7) takes a different approach. Its text expressly allows gatekeepers to adopt “strictly necessary and proportionate” measures to protect system integrity and security, provided they justify those measures. Yet in its March 2025 decisions involving Apple, the Commission reduced that protection to a narrow allowance. Interoperability solutions must work as well as Apple’s own and, according to the decision’s summary, “must not require more cumbersome system settings or additional user friction.” 

Under Article 6(7), any developer may request access, and Apple must process those requests within fixed deadlines. The framework does not allow firms to screen applicants based on identity, business model, or data-handling history. Draft joint guidelines from the European Commission and European Data Protection Board (EDPB) would go further by barring gatekeepers from considering an applicant’s record of violations under the General Data Protection Regulation (GDPR). 

The permitted safeguards focus on the product rather than the recipient. Gatekeepers may use consent prompts, encryption, and measures that preserve end-to-end encryption. They may not decide who gets through the door. 

Against that background, last week’s Android AI decision comes as little surprise. Google must support any integrity measure with “objective and verifiable evidence” of risk and apply it equally to its own services. As I argued in May, that standard makes precaution against genuinely novel threats almost impossible. 

The discrepancy is hard to miss. Where the statute says little about screening recipients, the Commission created an extensive vetting regime. Where the statute expressly permits protective measures, the Commission reduced them largely to prompts. 

To be sure, the provisions are worded differently, and that difference carries legal weight. Article 6(11) imposes an affirmative duty to anonymize the data, which the Commission had to translate into operational rules. Query logs also contain personal data, so the GDPR applies directly. Article 8(1) of the DMA requires compliance measures to respect those obligations. 

Article 6(7), by contrast, frames security protections as a limited exception available to the gatekeeper. The Commission therefore reads them narrowly to prevent evasion, as regulators often do with legal exceptions. The method is coherent enough. Commands receive broad effect. Exceptions receive little room to breathe. 

The Best Case for the Double Standard

The Commission’s different readings of the two provisions are defensible, at least up to a point. The strongest case rests on user control and scale. 

Start with user control. Under Article 6(7), each data flow begins with an identifiable user action, such as pairing a watch or tapping “allow.” That choice may provide a legal basis for the transfer through consent or contractual necessity. It also creates a transaction-specific safeguard through prompts, device-level permissions, and the ability to disconnect the service. 

Article 6(11) works differently. No individual user chooses to share anything. Query data from millions of people moves in bulk, and no workable consent mechanism could cover it. The only path to lawful disclosure is to remove the data from the GDPR’s reach through anonymization. 

Because anonymization carries so much legal weight, any threat to it becomes central. The most obvious risk is that a recipient could try to identify users again. Vetting, ring-fencing, and audits follow from that concern. The safeguards therefore track the legal basis for each form of access. Article 6(7) relies on the user’s choice. Article 6(11) relies on the recipient’s trustworthiness. 

Scale offers a second defense. Article 6(11) separates the amount of data disclosed from the recipient’s own size. Even the smallest eligible firm receives data derived from nearly the entire population of Google Search users, then keeps a copy in its own environment for up to five years. 

If anonymization fails, the harm can reach millions of users and cannot be undone. A leaked dataset cannot be unpaired, as AOL learned in 2006. Under Article 6(7), a third party receives data only from users who choose its service. Access grows with adoption, and users can revoke it. 

These arguments make a stronger case for the Commission’s distinction than the decisions themselves do. They still do not carry the day. 

When the Checkbox Becomes the Security System

That defense fails twice on its own terms, then a third time on a question it never asks. 

Start with consent. Nothing moves until a user pairs a device or taps “allow.” But Article 6(7) requires Apple and Google to open their interfaces and process requests before any user makes that choice. By the time the prompt appears, the requester already has technical access and may look no different to the user than a legitimate accessory maker. 

Attackers also work upstream of genuine consent. They can phish the prompt, spoof the accessory, or exploit the pairing process. As Miko?aj Barczentewicz warned before the DMA took effect, one determined bad actor can cause immense harm by exploiting an interoperability mandate. That risk weakens the case for treating privacy less carefully under Article 6(7) than under Article 6(11). 

The contrast becomes sharper when the regimes are placed side by side. Under the Google Search decision, data recipients must earn trust. Under the Apple and Android AI decisions, the Commission presumes it. Anyone may apply, and firms may not reject applicants based on who they are. The remaining check is a prompt that users have little basis to assess. 

We already know what an unvetted queue can contain. When Amazon reviewed applicants for DMA data access, more than 75% were based outside the European Union. Many appeared to be data brokers with murky privacy practices. Apple has reported interoperability requests broad enough to read every message and email on a user’s device. 

The regimes also differ in what changes hands. A sensible privacy framework should match safeguards to the sensitivity of the data. More revealing data should receive stronger protection. 

These decisions reverse that logic. Article 6(7) interoperability can expose the most sensitive material on a phone. The Android AI decision goes further by covering microphones, cameras, screen contents, and the ability of AI agents to imitate taps and complete transactions. Article 6(11), by contrast, covers week-old queries in which each individual is hidden among at least 1,000 others. The Commission has placed its lightest safeguards around the most sensitive data. 

Nor is it clear that Article 6(11) always creates greater risks at a larger scale. Imagine a popular smartwatch app with five million users that gets hacked or turns out to have been malicious all along. Five million people could have their messages exposed at once, tied to their identities. Article 6(7) requires no audit, certification, or background check to prevent that outcome. 

Now multiply that risk across hundreds of companies entitled to connect, each creating another point of failure. The unvetted crowd may pose a greater danger than the small number of heavily audited firms receiving Google’s search data. A regime that tracked actual risk would impose some checks everywhere and stronger ones where more people could suffer harm. 

Recipient screening is hardly a Brussels invention. Platforms used gated access, screened counterparties, and revocable permissions long before the DMA. They built those protections, presumably, to meet consumer demand for privacy and security. The Commission now overrides those design choices without adequately accounting for the functions they served. Those functions must then be rebuilt through regulation. 

Under Article 6(11), the Commission preserved that inherited machinery. Under Article 6(7), it dismantled it and replaced it with the one tool the rest of the statute treats with suspicion—a consent prompt. Elsewhere, the DMA assumes that user choice is weak and defaults are sticky. That is why it requires choice screens and restricts dark patterns. The same concern helps explain why markets often provide privacy and security through platformwide rules rather than one prompt at a time. 

The consequences are concrete. Siri AI will not launch in the European Union, and Gemini on Android could meet the same fate. When openness mandates and privacy law press in opposite directions, rational firms reduce functionality to limit their exposure. 

Fit for Purpose, Blind to Consequences

Why does Europe’s flagship digital law guard anonymous search queries like state secrets while leaving people’s messages, microphones, and screens behind a single tap? 

The two-speed approach makes sense only if consequences do not count. There is a legal explanation for the gap. The provisions use different language, and each decision follows its own internal logic. 

That may be defensible as statutory interpretation. It is nonetheless quite obviously poor privacy policy. The most sensitive data on a phone may receive the weakest protection, while less sensitive data comes wrapped in audits, screening, and access controls. No one designing a privacy regime from scratch would choose that result. Platforms had already built many of the missing safeguards under Article 6(7), and the Commission dismantled them. 

The stakes are rising. AI assistants can read screens, press buttons, access microphones, and spend money. Article 6(7) now helps determine who may wield those powers. The costs are already visible. Rather than open their most sensitive features to any requester, firms are withholding products from European users. 

The Commission’s own review called the DMA “fit for purpose.” That judgment depends on what the law actually produces. The search-data decision shows that the Commission knows how to protect users when it sees the risk. Its harder task is explaining why platforms may not do the same.

The post Europe’s Privacy Paradox: Fort Knox for Search Data, a Checkbox for Your Phone appeared first on Truth on the Market.

Sens. Amy Klobuchar (D-Minn.) and Eric Schmitt (R-Mo.) introduced the Securing Enforcement of Americans’ Right to Competition at Home (SEARCH) Act earlier this month. Billed as a bipartisan effort to protect consumers and curb Big Tech’s power over online search, the proposal would write into federal law many of the sweeping remedies sought in recent monopolization cases against Google.

The SEARCH Act would bar Google and other search engines from paying browsers and operating-system providers for default placement. It would require browsers and devices to present users with a “choice screen,” force search engines to provide rivals with data and search results at marginal cost, and restrict technology acquisitions and joint ventures more aggressively than current antitrust law does.

Protecting competition is a worthy goal. The bill’s economics are much shakier. By replacing the United States’ traditional focus on consumer welfare with European-style ex ante market design, the SEARCH Act would likely raise costs, weaken incentives to innovate, and give federal enforcers broad power to favor some firms over others.

Europe Already Tried This Search

The SEARCH Act borrows heavily from the European Union’s Digital Markets Act (DMA), which imposes broad, upfront restrictions on large digital platforms. Under the DMA, regulators need not prove that a company has market power or that a particular practice harms consumers. The law bans certain conduct categorically.

The DMA has now been in force for two years, and its promised gains in innovation and consumer choice remain elusive. In several cases, Europe’s ex ante rules have made familiar services more cumbersome.

After the European Union barred Google from integrating hotel and flight search tools in ways regulators considered “self-preferencing,” users had to click through more pages to complete basic bookings. Studies estimate that those extra steps cost European users millions of euros in lost time, while direct hotel-ad bookings fell 36%.

European rules also require “choice screens” that ask users to select a search engine or browser. These prompts may offer more options on paper, but many users ignore them, dismiss the pop-ups, and return to familiar services. More prompts do not necessarily produce more meaningful choice.

Regulatory uncertainty has also delayed new technology. European consumers waited longer for AI-powered search, live translation, and smart-assistant tools because platforms feared large fines under unclear DMA requirements. Those delays also left European technology startups without tools that their American competitors could already use.

Importing the same regulatory philosophy into U.S. antitrust law would make the American technology sector less competitive in the name of promoting competition.

There’s No Such Thing as a Free Default

Section 3 of the SEARCH Act would bar search engines from paying browsers, device makers, and other distributors for default placement. Yet default arrangements long predate the internet and often reflect ordinary, pro-competitive business conduct.

Supermarkets, for example, routinely give their private-label products prime shelf space. That arrangement can reduce “double marginalization,” which occurs when firms at different stages of a supply chain each add their own markup. Fewer markups can mean lower prices for shoppers.

In digital markets, payments for default status can serve as a form of “competition for the market.” Search engines compete for distribution by offering browsers and device makers better terms. The resulting revenue helps fund free software, including open-source browsers such as Mozilla Firefox, as well as mobile operating systems and privacy-focused services.

A ban would strip independent platforms of a major source of revenue. They would then have to cut features, add more advertising, charge users, or some combination of the three. Consumers would pay the price through higher costs, weaker products, and a worse browsing experience.

A Search Engine on Someone Else’s Dime

The SEARCH Act’s data-sharing rules are just as troubling. Sections 5 and 6 would require covered search platforms to provide “qualified competitors” with access to search indexes, ranking signals, and user-query data at marginal cost.

Building a modern search engine takes billions of dollars and years of continuous work. Companies must crawl the web, organize vast amounts of information, and refine the algorithms that decide which results users see.

Requiring firms to share those assets at cost creates a free-rider problem. Rivals gain access to expensive infrastructure without making comparable investments, which gives them less reason to develop better crawling tools or new search designs. Incumbent search engines also have less reason to improve their products when regulators can require them to hand those improvements to competitors.

Mandatory data sharing also creates serious privacy risks. Even when firms remove obvious identifiers, distributing search-query data across more companies gives hackers more targets and creates more opportunities to identify users by combining datasets.

The SEARCH Act would trade user privacy and long-term investment for a government-designed shortcut to competition.

Regulating the Search War That AI Already Changed

The SEARCH Act arrives just as generative artificial intelligence is reshaping online search more quickly than any law or lawsuit could. The rapid growth of conversational tools such as ChatGPT, Claude, and Perplexity shows how fast technology markets can change when new products challenge established ones.

These services have already brought real, market-driven competition to traditional search engines without Congress designing the terms of entry.

The SEARCH Act would still pull “AI search” into its regulatory scheme. Section 3(e) would require covered platforms to obtain written approval from the Federal Trade Commission before making certain acquisitions, investments, or joint ventures involving search or AI competitors.

That preapproval requirement would add delay and uncertainty to venture funding, strategic partnerships, and product development. It could slow the deployment of American AI models and weaken U.S. technological leadership just as competition from China and other geopolitical rivals intensifies.

Why Courts Need Room to Judge

Flexible U.S. antitrust law has one major advantage over rigid ex ante mandates. Judges can weigh competitive harms against benefits, account for commercial and technological realities, and tailor remedies to the conduct proved at trial. They also know that an appeals court can reverse them if they misapply the law or give too much weight to one side of the evidence.

Consider the recent Google Search case. U.S. District Judge Amit Mehta found that Google’s default agreements with Firefox and Safari helped it unlawfully maintain its search monopoly. Those contracts limited rival search engines’ ability to reach “minimum efficient scale,” the level of activity needed to compete economically.

According to Mehta, Google’s default placement generated more searches, clicks, and user data, which helped the company refine its index and ranking algorithms. Rivals received less data and therefore struggled to produce comparably strong search results.

When fashioning a remedy, though, Mehta confronted tradeoffs that the SEARCH Act largely ignores. He recognized that simply banning Google’s distribution agreements could harm competition among both search engines and browsers. He also concluded that requiring Google to provide its data to rivals for free or at marginal cost could weaken incentives to create and improve those assets, a concern long reflected in U.S. antitrust law.

Mehta instead required Google to offer certain data to competitors on “commercial terms.” That approach restricts Google’s control over its proprietary resources while still requiring rivals to pay market-based compensation. It seeks to preserve incentives for Google and its competitors to collect data, improve search quality, and develop competing tools.

The EU’s DMA offers no comparable flexibility. Neither would the SEARCH Act’s categorical ban on default agreements.

Google has appealed the liability ruling, which legal scholars have also criticized. Appellate review gives courts another chance to test whether a practice’s anticompetitive harms truly outweigh its pro-competitive benefits. Rigid ex ante rules sharply limit, or eliminate, that safeguard.

Europe’s Mistakes Need No U.S. Sequel

America’s technology sector became a global engine of growth under antitrust laws that focus on consumer welfare, rather than firm size or the protection of competitors from vigorous competition.

Existing laws, including the Sherman Antitrust Act, already give federal courts power to police anticompetitive conduct case by case. Recent platform litigation shows that judges can examine evidence, weigh competitive harms against benefits, and craft targeted remedies without imposing broad rules that may leave consumers worse off.

Congress should learn from Europe’s experience before importing its mistakes. The United States does not need antitrust rules that assume the verdict before the evidence arrives. It needs enforcement that follows the facts wherever they lead.

The post The SEARCH Act Copies Europe’s Homework—and Its Mistakes appeared first on Truth on the Market.

Competition authorities often warn that procedure should not become an obstacle to enforcement. Fair enough. Antitrust investigations can be slow, document-heavy, and vulnerable to delay tactics. A firm with deep pockets may try to turn “process” into attrition.

But that is only half the institutional problem. An agency with broad discretion, weak disclosure duties, and limited independent review can turn enforcement into regulation by accusation.

That concern is economic as well as legal.

Due process helps competition authorities separate harmful conduct from vigorous competition. Notice, access to evidence, a meaningful chance to respond, protection for privileged and confidential information, reasoned decisions, and independent review all improve the quality of enforcement. When those safeguards are weak, agencies make more errors, firms face greater uncertainty, and interest groups gain more opportunities to seek favorable treatment.

The costs do not stay inside the hearing room. They appear in lower investment, less experimentation, and fewer innovations.

My recent Antitrust Chronicle article on the “next step forward” examined the substantial international framework already in place. It includes the Organisation for Economic Co-operation and Development’s (OECD) recommendation on procedural fairness, the International Competition Network’s (ICN) Framework for Competition Agency Procedures, and the growing use of due-process commitments in trade agreements.

The problem is no longer a shortage of principles. It is the gap between stated commitments and actual practice.

That gap reflects institutional incentives. Competition agencies do not operate as detached maximizers of social welfare. Like other institutions, they respond to political pressure, resource limits, internal priorities, and organizational self-interest.

A serious reform program must therefore do more than urge agencies to be fair. It must make sound procedure visible, reward compliance, and impose some cost when authorities ignore it.

Due Process Is an Error-Cost Problem

The law & economics case for due process begins with uncertainty. Outside the easy case of a naked cartel, the same conduct may help or harm competition depending on facts that are difficult to observe.

Exclusive contracts may shut out rivals, or they may encourage relationship-specific investment. Product integration may exclude complementary products, or it may lower transaction costs and improve quality. A merger may reduce rivalry, or it may combine the assets needed to bring a new product to market.

The error-cost framework emphasized by scholars at the International Center for Law & Economics (ICLE) asks decisionmakers to minimize the combined social costs of false condemnation, false acquittal, and administration. Scholars usually apply that framework to substantive rules, burdens of proof, and legal presumptions. It applies just as readily to procedure.

Procedural safeguards improve the information available to the decisionmaker. Early, specific notice tells the investigated party which theory it must answer. Access to inculpatory and exculpatory evidence reduces the risk that an agency will build its case on a partial record. The right to counsel and adequate response time allow technical and economic claims to be tested rather than merely asserted.

A reasoned decision requires the authority to connect the evidence to the governing legal standard. Independent review adds an external check on confirmation bias and the gradual expansion of an agency’s mission.

These protections cannot guarantee correct outcomes. They can, however, reduce both the likelihood and the cost of error. That is especially valuable in innovation-driven markets, where competitive effects are hard to predict and a mistaken remedy may shape business decisions for years.

In “Innovation and the Limits of Antitrust,” Geoffrey Manne and Joshua Wright explain why authorities may condemn novel business practices before their benefits become clear. Sound procedure can slow that rush to judgment while preserving enforcement against genuine exclusion.

The economic value of procedure also extends beyond individual cases. Predictable enforcement lowers the risk associated with investment and allows firms to plan transactions, contracts, and product launches with greater confidence. It also encourages voluntary compliance because firms can better understand what the law requires.

Fair procedures can also strengthen an agency’s legitimacy. Firms are more likely to accept adverse decisions when they can see how the agency reached them, rather than dismissing them as political acts.

Christopher Yoo, Thomas Fetzer, Shan Jiang, and Yong Huang make the broader case in their comparative study, “Due Process in Antitrust Enforcement.” They connect procedural protections to accuracy, government credibility, economic growth, limits on bureaucratic abuse, competitive reform, and lower corruption.

Those benefits are central to the institutions that support a market economy.

Due Process, Terms and Conditions Apply

Internationally, the normative work is largely complete. The OECD’s Recommendation on Transparency and Procedural Fairness in Competition Law Enforcement calls for predictable rules, impartial institutions, nondiscrimination, proportionality, timely resolution, meaningful engagement with parties, protection of privileged and confidential information, independent review, and periodic reassessment of enforcement practices.

The ICN’s Framework for Competition Agency Procedures turns similar commitments into an agency-led, voluntary framework. Participating authorities publish templates describing their procedures, agree to core principles, and may take part in interagency dialogues. The current CAP 2.0 work plan aims to reaffirm those commitments, draw attention to neglected areas, and recruit more signatories.

That is genuine progress. Soft law can shape expectations, help reform-minded officials press for change within their own governments, and create a shared vocabulary across legal systems. The CAP templates also make national procedures easier to compare.

Yet soft law has familiar limits. The framework creates few enforceable rights for the firms and individuals directly affected by agency action. Interagency dialogues are generally confidential and depend on officials’ willingness to participate. Authorities may invoke national law to justify departures from the framework. Periodic reviews tend to celebrate collective progress more readily than identify specific failures.

As James Rill and Jana Seidl argue in their analysis of the CAP’s “promise of accountability,” genuine convergence requires a mechanism that reveals whether jurisdictions have actually implemented the norms they accepted.

Other contributions to the CPI Antitrust Chronicle’s June 2026 issue on due process make the same point. J. Mark Gidley and Daniel Sokol argue that formal rights mean little if parties cannot invoke them when they matter most, including during searches, evidence collection, formulation of charges, interim relief, and review.

Ian Forrester and Pablo Trevisan compare administrative and adversarial systems while stressing the practical value of being heard by a neutral decisionmaker. John Taladay and Christine Ryu-Naya warn that market investigations can expand beyond fact-finding and become vehicles for industrywide remedies without procedural protections equal to that power. Abbott (Tad) Lipsky points to the painfully slow movement from decades of discussion to measurable improvement.

The recurring defects are well known. Agencies may rely on vague or shifting theories of harm, restrict access to the file, withhold exculpatory evidence, impose compressed response periods, or conduct intrusive searches with weak ex ante or ex post controls. They may use confidentiality rules to obscure their reasoning, allow investigations to drag on until the process itself becomes punishment, impose interim measures before the facts are settled, or face judicial review too deferential, slow, or narrow to offer a useful remedy.

A consensus on paper does not produce compliance on its own. Explaining the gap requires asking a less comfortable question. Who benefits from weak process, and who pays for it?

Why Agencies Prefer Procedural Wiggle Room

The first obstacle is a collective-action problem. Better procedure produces broad benefits, including greater legal certainty, fewer errors, more investment, and stronger public confidence. Consumers, businesses, and the wider economy share those gains. Enforcement agencies and political institutions bear the immediate costs.

Disclosure consumes staff time. Hearings slow decisions. Independent review can overturn agency victories. Clear standards constrain discretion. An agency may therefore see procedural reform as surrendering authority in exchange for benefits that accrue mostly to others. Six factors help explain the resistance.

First, public-choice analysis predicts this tension. My recent Truth on the Market essay on why administrable antitrust presumptions can become “too useful to kill” explains that agencies naturally prefer rules that reduce evidentiary burdens, conserve resources, and improve win rates. The same logic applies to procedure.

An authority gains an institutional advantage when it can formulate allegations late, restrict access to the file, negotiate remedies under severe time pressure, or combine investigation with the initial decision. It has little reason to surrender those advantages voluntarily.

Second, weak process creates opportunities for rent seeking. Less-efficient rivals, domestic incumbents, organized labor, sector regulators, and political officials may all benefit when an agency investigates a successful firm, even if the competition case is thin. Competition law can become a tool for redistributing rents, pursuing industrial policy, or burdening foreign businesses.

The targeted firm bears concentrated costs. Political beneficiaries can claim to defend “fairness,” small businesses, jobs, or national sovereignty. Broad substantive standards and opaque procedures make consumer protection difficult to distinguish from competitor protection.

Third, agency-led international institutions inherit their members’ incentives. The ICN’s strength comes from its practical expertise and freedom from the slow machinery of treaty negotiations. Its weakness follows from the same design. An organization composed of enforcement agencies is unlikely to build a demanding system for publicly censuring those agencies.

Confidential peer dialogue and consensus-based recommendations are the politically feasible price of participation. They also help explain why the network’s implementation mechanisms remain modest.

Fourth, institutional capacity varies widely. A mature agency may be able to employ hearing officers, separate investigative and decisional staff, issue detailed decisions, and support expedited judicial review. A newer authority with a small budget may struggle to translate documents, preserve digital evidence, and meet basic deadlines.

Requiring every authority to adopt an elaborate procedural model could divert scarce resources from cartel enforcement and competition advocacy. Yet limited resources can also become a permanent excuse for indefinite investigations and minimal disclosure.

Fifth, due process becomes most vulnerable when political pressure makes it most necessary. High-profile cases against large or unpopular companies create demands for speed and visible action. Officials who defend procedural protections may be accused of serving monopolists or foreign interests.

The costs of a mistaken intervention often emerge slowly and are difficult to trace back to the agency. The political rewards of an aggressive announcement arrive immediately. That mismatch encourages agencies to act first and learn later.

Sixth, legal tradition matters. Some systems place considerable trust in integrated administrative decisionmaking. Others rely more heavily on adversarial courts. Critics may therefore portray procedural convergence as an effort to export American litigation practices.

That objection has force when reformers demand institutional uniformity. It weakens when the goal is functional equivalence. Whatever the institutional design, parties should receive timely notice, meaningful access to the case against them, an opportunity to respond, protection against conflicts of interest, and effective independent review.

Putting Teeth in Procedural Fairness

A workable reform agenda should begin with modesty. Due process cannot become a license for endless delay, tactical discovery, or obstruction of legitimate investigations. National systems also need not copy U.S. federal litigation. The aim is narrower—to improve the accuracy and legitimacy of decisions at a reasonable administrative cost.

First, make performance visible. The Framework for Competition Agency Procedures templates should evolve beyond descriptive questionnaires and include a limited set of comparable indicators. Agencies could report median investigation length, when they disclose theories of harm, rules governing access to inculpatory and exculpatory evidence, the use and duration of interim measures, access to privilege review, publication of reasoned decisions, and the time required to obtain judicial review.

These indicators need not produce a crude league table. They should reveal where formal commitments diverge from actual practice.

The indicators need not produce a simplistic league table. Their purpose is to reveal where formal commitments diverge from operational reality.

CAP reviews could also publish anonymized case studies and aggregate statistics on interagency dialogues. A process that identifies persistent patterns without turning every dispute into diplomatic theater would raise the reputational cost of noncompliance. Reform-minded agency leaders would also gain an external benchmark when seeking legislative authority or additional resources.

Second, give affected parties a structured voice without handing them control of the framework. Firms should not direct peer review, but they should be able to submit documented procedural concerns to an independent CAP contact point or the OECD secretariat.

Bar associations, economists, consumer groups, and nongovernmental advisers could help screen complaints and identify recurring problems. Requiring specific allegations and exhaustion of domestic remedies would discourage tactical filings. The process could produce a confidential referral, an anonymized thematic review, or a public recommendation when the evidence shows a persistent pattern.

Third, strengthen internal checks that cost less than a full institutional redesign. Agencies can appoint hearing officers or procedural ombudsmen to resolve disputes over access, privilege, deadlines, and confidentiality. They can require written approval and prompt review for interim measures and dawn raids—that is, unannounced searches of business premises.

Agencies can also separate investigative staff from initial decision makers where practical, adopt presumptive deadlines with public explanations for extensions, and require decisions to address material defense evidence rather than merely repeat the agency’s theory.

These reforms can fit agencies of different sizes. A small authority need not create a separate court. It could rely on a rotating external panel, a ministry-wide administrative judge, or a regional cooperation arrangement. Technical assistance should support these institutional safeguards alongside training on how to bring cases.

Fourth, attach limited procedural conditions to international cooperation. Evidence sharing, confidentiality waivers, and investigative assistance expand enforcement power. Those tools should come with assurances that shared information will receive adequate protections for privilege, confidentiality, permitted use, and review.

Countries need not suspend cooperation whenever procedures differ. But particularly sensitive assistance should depend on verified safeguards and should be restricted when the receiving authority cannot provide them.

Fifth, use trade agreements with greater care and credibility. Modern agreements already offer models for a procedural baseline. Chapter 21 of the United States-Mexico-Canada Agreement includes commitments on notice, representation by counsel, the opportunity to present evidence, written decisions, and review. OECD research documents the growing use of competition provisions in trade agreements. The African Continental Free Trade Area Competition Protocol points toward a more integrated model that includes state-to-state dispute settlement.

Future agreements should make a narrow set of procedural commitments subject to consultation and, in cases of systematic violation, dispute settlement. Remedies should be targeted and graduated. Transparency requirements, corrective-action plans, suspension of discretionary cooperation, and prospective compliance commitments will usually serve the goal better than tariffs. Tariffs burden consumers and can turn a rule-of-law dispute into a protectionist contest.

Section 301 of the Trade Act of 1974 and similar tools may provide pressure in extreme cases. They should remain a last resort.

Sixth, improve judicial review when timing matters most. Review that arrives after a transaction has collapsed or a technology has been redesigned offers little practical protection. Courts should be able to examine procedural irregularities, test the evidentiary basis for interim measures, and grant expedited relief when irreversible harm is likely.

Deference to agency expertise should not extend to undisclosed evidence, shifting legal theories, or remedies disconnected from the agency’s findings.

Seventh, evaluate procedure as economic policy. Agencies routinely assess the effects of business conduct. They should also assess the effects of their own processes.

Ex post reviews could examine whether preliminary theories survived, how often remedies were modified or reversed, how long investigations lasted, and whether compliance costs matched the stakes. The aim is to encourage institutional learning and identify patterns that ordinary case-by-case review may miss, rather than punish agencies for reasonable losses.

When Bad Process Becomes Bad Economics

The international consensus on procedural fairness is broad. The incentives to implement it remain weak.

Arguments for better process often focus on dignity, legality, and fairness. Those values matter. So do the economic consequences of getting procedure wrong.

In dynamic markets, enforcement can redirect innovation before courts have a chance to review it. Interim restrictions, design mandates, forced access, limits on integration, and merger delays can alter investment decisions long before the evidence is complete.

Toshiaki Takigawa and Dirk Auer recently made this point in discussing artificial intelligence regulation. Preventive intervention imposes the costs of error early. The faster technology changes, the less useful a remedy may become by the time the record is complete, and the more damage a mistaken early order may cause.

Global enforcement compounds the risk. A multinational firm may adopt the rules of the most restrictive jurisdiction worldwide because maintaining separate products or business models is costly. A procedurally defective decision in one country can therefore reshape conduct far beyond its borders.

Smaller firms and startups face even greater pressure. They have fewer resources to endure lengthy investigations, contest confidentiality claims, or pursue appeals in several jurisdictions. Legal uncertainty can discourage the entry and experimentation that competition law is supposed to protect.

Due process should therefore be treated as neither a concession to defendants nor a box-checking exercise for international organizations. It improves the information available to decisionmakers, constrains discretion, and reduces the social cost of enforcement errors.

The OECD and the ICN have built a valuable consensus. The next step is to align incentives with that consensus through measurement, independent feedback, conditional cooperation, targeted trade commitments, and effective review.

No reform will eliminate political pressure, institutional self-interest, or differences among legal systems. A realistic agenda can still make procedural failure more visible, make sound practices easier to adopt, and attach credible consequences to persistent noncompliance.

Competition law protects rivalry and discovery. Its own procedures should show the same humility about what authorities know and the same caution about shutting down experimentation.

When process becomes an afterthought, competition can become collateral damage.

The post Competition Law’s Fairness Gap: Why Better Procedure Requires Better Incentives appeared first on Truth on the Market.

The European Commission’s latest Digital Markets Act (DMA) decisions capture that tension. Europe wants more innovation, investment, and globally competitive digital platforms. Yet when a company assembles the data, technology, distribution, and complementary services needed to compete worldwide, Brussels increasingly treats those advantages as inputs to be shared with rivals.

On July 16, the Commission adopted binding measures requiring Google to give eligible search engines, including AI chatbots with search functions, access to anonymized Google Search data. A parallel decision requires Google to open 11 categories of Android functionality to competing AI assistants. These include sensor inputs, app data, operating-system controls, screen automation, on-device computing resources, background execution, and hotword activation.

The Commission says these mandates will promote contestability, consumer choice, and innovation. The immediate gains are easy to see. Rivals receive valuable inputs, developers gain access to Android features, and consumers may encounter more services competing within Google’s platform.

The costs are harder to spot and easier to ignore. Forced access may weaken incentives to build proprietary datasets, reduce returns on risky platform investments, increase cybersecurity risks, and push competing systems toward the same design. It may also lead companies to delay, degrade, or withhold future products in Europe.

These decisions therefore deserve scrutiny as more than access mandates. They alter the rules governing who bears the cost of innovation and who receives the reward.

The Algorithm by Other Means

The Commission stresses that Article 6(11) does not require Google to disclose its source code, algorithms, or technology. Formally, that is true. The mandate instead covers anonymized ranking, query, click, and view data generated through paid and unpaid search.

Google must generally provide information it collects and uses to improve Search, including entered queries, language and device metadata, viewed URLs, user interactions, and result positions. The public measures contemplate a delay of at least seven days and allow each eligible recipient to receive the data for up to five years.

That source code remains private does not resolve the intellectual-property problem. In a data-driven business, valuable know-how appears in more than written algorithms. It also emerges from the feedback generated when those algorithms operate across billions of searches.

Query reformulations, click patterns, result positions, dwell times, and other behavioral signals can reveal how a system reacts to particular inputs. Combined with a recipient’s own models, experiments, and datasets, that information can reduce the cost of approximating aspects of the incumbent’s ranking behavior.

The meaningful distinction concerns direct access to technical instructions and access to the behavioral data from which competitors may infer parts of those instructions. The second form of access will not recreate Google Search. Search quality also depends on crawling, indexing, engineering talent, experimentation, brand, distribution, computing infrastructure, and other complementary assets.

The International Center for Law & Economics (ICLE) has therefore cautioned that shared data may not transfer much value and that the returns from additional search data may decline. Even so, mandated access can capture part of the return on Google’s past investment by lowering rivals’ costs of experimentation, error correction, and imitation.

Search data did not materialize on its own. Google generated it through decades of investment in crawling, indexing, infrastructure, cybersecurity, user interfaces, quality testing, and fraud prevention. Individual facts may not qualify as intellectual property, but the organized stream of information produced by the platform remains an economically valuable asset. Its value reflects both Google’s technological investment and the trust of users who chose to submit their queries.

The Commission’s pricing rule deepens the problem. Google generally may recover only the incremental costs of preparing, storing, and transmitting the dataset, plus a return on incremental capital capped at its weighted average cost of capital. An additional margin is available only in exceptional circumstances.

That formula largely excludes the sunk costs of creating the underlying information asset and the option value of keeping it exclusive. For information goods, those omissions are substantial. Innovation-intensive products often require high fixed and sunk costs, face uncertain commercial prospects, and cost little to reproduce once created.

A price based mainly on delivery expenses therefore ignores the investments that created the value being shared. As ICLE has warned, a fair, reasonable, and nondiscriminatory (FRAND) access rule can become a subsidy for competitors.

The decision does not amount to uncompensated expropriation. Access remains limited, Google must anonymize the data, recipients must meet eligibility requirements, and some compensation is available. Economically, though, the mandate resembles a compulsory license priced without full regard for the investment that produced the licensed advantage.

It transfers part of Google’s informational returns to rivals while leaving Google responsible for the costs, liabilities, and security risks of creating and maintaining the asset.

The Mandate That Keeps on Mandating

The Commission argues in its search-data Q&A that Google will retain ample incentives to innovate. Recipients will receive only a modified subset of its data, may not systematically reproduce Google’s search results, and may not use the dataset to train general-purpose AI models. Those safeguards limit the mandate’s reach, but they do not establish that it will leave investment unchanged.

Investment incentives operate at the margin. A rule need not capture every return to alter a company’s behavior. It need only reduce the expected payoff from the next dollar spent improving search quality, collecting data, developing behavioral signals, or launching a new feature.

A company that expects regulators to make valuable feedback available to competitors at a controlled price may invest less in producing that feedback. It may also reserve its most commercially sensitive innovations for products or jurisdictions where it can retain more of the resulting value.

The effects reach beyond Google. An aspiring platform must now account for the possibility that success in Europe will turn the scale economies it created into duties to assist rivals. Investors will discount expected returns to reflect that risk. Some firms will still invest, but they will demand higher projected returns before doing so. Marginal projects will lose funding, ambitious forms of integration will become less attractive, and business models that depend on long-term cross-subsidies will become harder to finance.

The mandate also creates a predictable regulatory ratchet. If recipients fail to gain market share, they can argue that the dataset is too narrow, delayed, anonymized, or expensive. If privacy protections make rare queries less useful, beneficiaries can press the Commission to relax them. If access does not allow rivals to match Google’s quality, they can demand more data fields, shorter delays, or technical assistance.

ICLE warned in its comments on Article 6(11) that judging “effective compliance” by competitors’ commercial results rather than the availability of workable access would encourage exactly this progression. Each disappointing competitive outcome could become evidence that the previous mandate did not go far enough.

The Commission’s own Q&A gives that concern some force. It rejected Google’s initial implementation in part because the company removed 90% to 100% of unique queries and attracted little meaningful uptake. Uptake may help show whether rivals find the data useful, but it does not measure consumer welfare. A dataset can satisfy an access requirement even when competitors decide it lacks commercial value. A highly valuable dataset can also impose greater costs on privacy, security, and future innovation.

Once competitor use and market share become measures of regulatory success, the Commission moves beyond opening access. It begins designing the competitive outcome.

Anonymized, With an Asterisk

The Commission has taken privacy seriously. Its system removes direct identifiers and some metadata, suppresses rare or unusually long queries, generalizes locations, aggregates interaction times, and strips out advertising URLs. Contracts, independent audits, limits on onward transfers, retention rules, and use restrictions add further protection.

Those safeguards matter. The decision does not authorize an unrestricted transfer of identifiable search histories. It also does not automatically open the data to every Chinese, Russian, or other foreign-controlled company. The Commission allows Google to exclude sanctioned entities and firms controlled by countries that pose serious and structural cybersecurity or data-protection risks. It also permits public-security exemptions and requires processing in the European Economic Area or under protections deemed essentially equivalent.

These measures reduce the danger. They do not make it disappear.

Search queries can reveal sensitive information about health, finances, travel, employment, political activity, intimate relationships, business strategy, and personal vulnerabilities. Removing direct identifiers may not prevent recipients from combining query data with their own logs, public records, advertising data, leaked datasets, or inference models.

Mikolaj Barczentewicz’s analysis of the preliminary proposal made the central point. Anonymity depends on more than what appears in the dataset itself. It also depends on the auxiliary information and technical capabilities available to each recipient.

That concern follows directly from Article 6(11)’s competitive premise. Recipients are expected to possess the complementary tools needed to extract commercial value from the data. The same expertise may also help them reconstruct sensitive information. Contracts and annual audits can deter abuse, but they may detect it only after data have been copied, inferred, compromised, or transferred.

National-security concerns also require precision. The public materials do not show that raw, identified searches by U.S. military officials or diplomats will reach hostile governments. But searches by American personnel in Europe may still form part of the larger pool from which the shared dataset is drawn. If the data remain sufficiently granular or are poorly anonymized, they could reveal patterns involving travel, facilities, vendors, technical problems, operational interests, or personnel concerns.

The Commission’s high-risk-country and public-security provisions address some recipient risks, but they do not necessarily match U.S. national-security judgments. They also cannot eliminate cyber intrusions, hidden beneficial ownership, acquisitions after certification, insider threats, or inferences drawn from multiple lawful datasets.

Google’s July 16 response argues that the measures expose private searches without adequate anonymization, knowledge, or consent and create risks to privacy, business secrets, and national security. That remains Google’s position rather than an adjudicated finding. Even so, the costs it identifies deserve analysis rather than dismissal as the grumbling of a regulated firm.

Privacy and security are part of service quality. A competition rule that expands rival access by weakening users’ expected confidentiality may produce more competitors and a worse product.

A Remedy With Guardrails—and One Without

The American search-remedies litigation offers a useful comparison. The U.S. district court imposed its remedy in United States v. Google only after finding specific exclusionary conduct. Its user-side data mandate covers the underlying information used the Generalized Learned User Embeddings (GLUE) and RankEmbed models and requires at least two disclosures. The court will determine the final number and timing after consulting a technical committee.

The judgment also draws firm boundaries around what Google must share. It excludes algorithms, ranking signals, scores, post-trained large language models, intellectual property, and trade secrets.

Recipients must satisfy court-approved security standards, pass regular audits, show a credible plan to invest and compete, undergo annual recertification, and pose no threat to U.S. national security. Google may object to a recipient’s initial or continued eligibility and receive a hearing. The court also recognized that anonymized user-side data could remain highly sensitive, so it restricted sales and onward sharing.

The American remedy still invites criticism. It prices access at marginal cost and may weaken incentives to invest. Yet it remains more closely tied to adjudicated conduct, defined datasets, judicial oversight, trade-secret protection, national-security screening, and recipient-by-recipient review.

The DMA decision reaches further in both duration and ambition. It does not respond to a finding that Google unlawfully acquired a particular dataset. It treats Google’s continuing accumulation of search data as a structural advantage that regulators may redistribute for as long as Search remains designated.

The decision also extends eligibility to AI chatbots with search functions and seeks, where technically feasible, to approximate the methods and speed with which Google uses the data itself. The U.S. judgment remains an antitrust remedy, even if an interventionist one. The European model looks increasingly like continuing public-utility regulation of information.

Leveling the Platform Down

The Android decision presents a related but distinct problem for innovation. The Commission requires Google to give third-party AI services free access, on equally effective terms, to Android functions involving invocation, context, actions, and computing resources.

That access is extensive. Third-party assistants may receive real-time input from microphones, cameras, screens, and speakers. They may interact with apps and Google services, automate tasks in virtual windows, change system settings, use on-device AI models, run in the background, and activate through always-on hotwords. Google must document the interfaces, provide testing and technical support, extend future functionality to third parties, and avoid unnecessary friction for users.

These rules could help independent AI developers. An assistant with deep device access can do far more than one confined to an ordinary app sandbox, the restricted environment that limits what an app may see and control. Consumers may gain more choice, and developers may create features Google would never have pursued.

The longer-term cost is weaker competition between integrated systems. Tight integration between Gemini and Android is itself a product feature. It gives Google a way to challenge rival AI providers and distinguish Android from Apple’s model. Requiring Google to reproduce those advantages for competitors may increase competition within Android while reducing competition between rival AI systems.

ICLE’s comments on Article 6(7) identify this neglected distinction. Giving equivalent Android access to leading AI providers could soften competition across the broader AI market by limiting Google’s ability to use operating-system integration as a competitive response. The main beneficiaries may be powerful incumbents rather than scrappy startups. They receive a regulated route into a platform they did not build.

Security risks make equal treatment harder still. Google can inspect its own code, impose internal development rules, revoke credentials immediately, test security throughout the development process, and coordinate responses across Android. It cannot exercise the same control over every outside provider. Equal technical access does not produce equal risk.

The Commission permits integrity measures that are strictly necessary and proportionate. For several sensitive functions, it also allows objective eligibility requirements and independent certification. Yet rules demanding transparent, objectively verifiable, and broadly symmetric restrictions may fit poorly with new threats.

Security teams often must act before they possess conclusive proof. Waiting for demonstrated exploitation can mean waiting until users have already suffered harm. ICLE therefore warns that Google may face a stark choice. It can expose sensitive functions broadly or remove them from its own services so that equal-treatment duties no longer apply.

That creates a predictable form of defensive leveling down. Google may narrow functionality, reduce the sensitivity of available application programming interfaces, delay integrated features, or withhold some services from Europe. Smaller platform developers watching the result may avoid close integration altogether.

When Brussels Becomes the Product Manager

My July 8 analysis of the Court of Justice of the European Union’s (CJEU) Google Android judgment offers a broader frame for evaluating these decisions. Intervention becomes especially risky when regulators treat successful platform design as presumptively suspect and fail to ask what would have existed without the challenged integration.

Android’s compatibility rules, monetization arrangements, defaults, and complementary services helped coordinate handset manufacturers, developers, advertisers, and users. Weakening one part can impose costs on the others.

The July 16 decisions extend that logic beyond retrospective antitrust liability. They prescribe data fields, anonymization methods, recipient eligibility, prices, interface design, access quality, documentation, technical support, timelines, and future functionality. This goes well beyond barring exclusionary conduct. It amounts to continuing administrative control over platform technology.

My related analysis of the DMA’s rule-of-law problems describes this shift from competition law toward product management and industrial administration. Because the DMA does not require the same case-specific proof of competitive harm as traditional antitrust law, procedural discipline and economic evidence become more important.

The formal specification proceedings ran from Jan. 27 to July 16, with preliminary search measures released April 16. The Commission says the process followed two years of discussions with Alphabet and included consultations, testing, and expert input. Google therefore received a process, but the deeper concern remains.

A compressed administrative proceeding under a statute designed to bypass much of traditional antitrust’s effects analysis now determines technical, privacy, security, and intellectual-property questions whose consequences may extend far beyond one product cycle.

The asymmetry in error costs should give regulators pause. The state can revise a mistaken rule. Google cannot fully reverse a data disclosure, undo a security breach, or recover innovation returns once regulators have transferred them to competitors. Under those conditions, “move fast” makes for a poor governing principle.

A Mandate With Brakes

A more economically defensible approach would treat dynamic competition, privacy, security, and firms’ ability to retain returns on investment as core design constraints.

Search-data access should begin narrowly. The Commission should tie it to evidence of specific competitive bottlenecks and initially limit disclosure to synthetic, sampled, or heavily filtered data. It should cap the frequency and duration of access, then expand the mandate only when evidence shows consumer benefits. Competitor uptake alone is a poor measure.

The ban on systematic replication also needs teeth. Recipients should face enforceable limits on reverse engineering, model extraction, combining the data with outside datasets, onward transfers, and changes in ownership or control.

Eligibility should include a recipient-specific national-security review conducted with relevant allied authorities. General country-risk categories cannot capture hidden ownership, personnel access, cybersecurity capabilities, prior incidents, acquisition risk, or the sensitivity of a proposed use. Google should also have a meaningful chance to challenge access before disclosure, when harm can still be prevented.

Pricing should reflect the fixed and sunk investments that created the dataset, the value of the access granted, and the effect on future investment. A rule confined to incremental delivery costs tells prospective innovators that regulators may convert successful investments into inputs for rivals without paying for their creation.

Android interoperability should rely on risk tiers. Ordinary invocation presents different concerns than persistent microphone access, screen automation, system controls, aggregation of app data, or background execution. Regulators should permit different treatment of first-party and third-party services when their governance, auditability, revocation procedures, and security responsibilities differ. New capabilities should undergo controlled testing before general release.

Both decisions also need meaningful sunset provisions and empirical review. The test should focus on consumer outcomes. Did users receive better products, lower quality-adjusted prices, stronger security, more innovation, and greater competition among business models? Giving rivals better inputs answers a different question.

When Success Stops Paying

The Commission sees Google’s search data and Android integration as barriers that keep rivals from competing. A dynamic law & economics analysis sees something else as well—the accumulated returns to investment, experimentation, coordination, and risk-taking.

Both views contain part of the story. Data and operating-system access can ease entry, and dominant platforms can use control over important assets to obstruct competition. Yet the mere existence of a valuable advantage does not show that sharing it will improve consumer welfare. It certainly does not show that rivals should receive it at a price largely detached from the cost and risk of creating it.

The July 16 decisions define competition too narrowly, as rival access to the fruits of an incumbent’s investment. They also define innovation too statically, as the number of firms able to use assets that already exist.

Sound innovation policy must ask who will build the next set of assets. A legal regime that promises successful platforms years of administrative redesign, compelled technical support, regulated access prices, and recurring disclosure of sensitive data will not end innovation. It will push firms to design around Europe, delay launches, collect less useful data, integrate less ambitiously, and invest where returns are more secure.

Europe does not lack rules requiring successful platforms to help their competitors. It lacks confidence that success will remain worth the trouble.

The post Europe Wants Tech Champions, Then Makes Them Share the Trophy appeared first on Truth on the Market.

That is a substantial change in how competition law works. The bill would allow CADE to regulate self-preferencing, default settings, interoperability, and data use in advance rather than through ordinary case-by-case enforcement. 

By the government’s estimate, five to 10 companies would be designated, most of them likely American. The stakes therefore extend beyond competition policy. Foreign regulation of U.S. technology firms has become a trade and foreign-policy flashpoint, and Brazil’s proposal could add another point of friction. 

The bill also tests a broader claim now shaping digital regulation around the world. Can traditional antitrust enforcement police dominant platforms, or should governments supplement it with pre-emptive rules modeled on the European Union’s Digital Markets Act (DMA)? 

Brazil offers a revealing test. Its competition authority has recently shown what existing law can accomplish. Settlements with Apple and Google produced DMA-style outcomes, including greater openness in Apple’s iOS operating system, through ordinary case-by-case enforcement. Early evidence from Europe, meanwhile, increasingly points to consumer harm and weaker incentives to innovate rather than the promised surge in competition. 

Brazil’s choice could influence other emerging economies weighing similar regimes. That gives the debate in Brasília significance well beyond Brazil. 

In “Digital Overreach: A Premature Turn to Ex Ante Regulation in Brazil,” Geoffrey Manne, Dirk Auer, and I argued that Brazil neither needs nor would benefit from a new ex ante regime for digital markets. The narrower question here is whether the substitute text now before the Chamber of Deputies improves on the government’s original proposal and how seriously it addresses the criticisms advanced by the International Center for Law & Economics (ICLE) and others. 

The substitute adopts many of the procedural and institutional safeguards we recommended. It leaves the proposal’s doctrinal core intact. That core remains the strongest argument against enacting the bill. 

Fast-Tracked, Then Stalled

The executive branch submitted Bill 4,675/2025 in September 2025. The measure would amend Brazil’s Competition Law, Law 12,529/2011, to let CADE designate large technology platforms as having “systemic relevance” in digital markets and impose “special obligations” on them. 

Two recent procedural developments pushed the proposal closer to a vote. 

The first came March 18, when the Chamber approved an urgency motion, REQ 4612/2025. Under the Chamber’s rules, urgency allows a bill to bypass the standing committees and proceed directly to the plenary floor. That compresses the process into negotiations among party leaders and the bill’s rapporteur, stripping away the procedural checkpoints where legislation is usually tested, amended, delayed, or quietly buried. 

The second came July 8, when the rapporteur, Congressman Aliel Machado of Paraná’s Green Party, filed his report and a substitute text after roughly three months of review. Party leaders had agreed that the report would clear the way for a floor vote, and the Chamber’s website now lists the bill as ready for the agenda. 

The substitute matters because it is, for practical purposes, the text the Chamber will consider. Members may still offer amendments on the floor, but under urgency, the rapporteur’s draft becomes the baseline for every negotiation that follows. 

The timing remains uncertain. An attempt to bring the bill to a vote during the week Machado filed his report failed to win consensus among party leaders. On July 15, the Chamber held its final deliberative session before the winter recess and again left Bill 4,675 off the agenda. 

That leaves a narrow window for action this year. Once lawmakers return to their states, the Chamber is expected to concentrate its legislative work before Brazil’s general elections in October and November 2026. Some expect a vote during the first week of August, after the recess. Brazilian politics advises against putting that date in ink. 

Even Chamber approval would settle only half the matter. Brazil has a bicameral Congress, so the bill would then move to the Senate, where the debate would begin again. 

New Guardrails, Same Bureaucracy

The substitute gives the bill’s new enforcement unit a longer name. It would now become the “Special Superintendence for Systemic Relevance, Free Competition and Consumer Protection in Digital Markets.” The change has no direct legal effect, but it fits Machado’s claim that the bill “regulates markets, not companies.” 

The more consequential revisions limit administrative rulemaking. Article 18-A bars the resolution defining the superintendent’s duties from creating new sanctioning powers, expanding the grounds for designation, or imposing obligations without a statutory basis. CADE’s complementary regulations would be limited to implementing the deadlines and procedures established by law. They could not create additional obligations, sanctions, or grounds for designation. 

That change responds directly to a concern running through ICLE’s analysis. The original bill paired open-ended designation criteria with broad administrative discretion. Brazil’s institutional constraints make that combination especially risky, even if similar powers appear in some of the foreign regimes the bill seeks to emulate. 

The substitute leaves the institutional structure intact. ICLE recommended placing the digital-markets function within CADE’s existing General Superintendence, following the approach used by the European Commission, the United Kingdom’s Competition and Markets Authority (CMA), and Germany’s Bundeskartellamt. Instead, the bill retains a parallel superintendence that duplicates CADE’s architecture and creates room for jurisdictional conflict. 

Worse, the substitute provides no new appropriations to fund it. The bill therefore preserves both the institutional fragmentation and the resource constraints that ICLE identified in the original proposal. 

More Choice, Less Discipline

Article 47-B sets out the new regime’s objectives. The original bill listed three goals: reducing barriers to entry, protecting the competitive process, and promoting freedom of choice. The substitute keeps all three and simply moves “freedom of choice” to the front of the line. 

That reshuffling does not cure the underlying problem. In the “Digital Overreach” white paper and an earlier Truth on the Market post, we argued that Article 47-B departs from the consumer-welfare framework embedded in Brazil’s Competition Law. None of its three objectives contains a limiting principle tied to measurable effects on consumers. 

Barriers to entry may reflect scale economies, quality investments, or network effects that make products better or cheaper. “Protecting the competitive process,” as Herbert Hovenkamp has put it, is little more than a slogan. It is circular, vague, and difficult to measure. 

“Freedom of choice” may be the weakest standard of all. Joshua Wright and Douglas Ginsburg have shown that a choice-based test detaches antitrust from consumers’ actual preferences and the tradeoffs among price, quality, innovation, and variety. It treats more options as inherently better, even when fewer options produce lower prices or better products. 

CADE’s own decisional record points in the same direction. “Freedom of choice” appears only once as a proxy for welfare. Price appears 94 times. 

We recommended making consumer welfare the regime’s explicit objective and limiting principle. The bill’s three stated goals could then serve, at most, as indicators of possible harm. The substitute declines that recommendation. 

That omission creates a deeper statutory tension. The bill would amend Brazil’s Competition Law rather than establish a separate regime. Article 36, Section 1, of that law recognizes that dominance achieved through efficiency and the “natural process” of competition is lawful. Article 88, Section 6, conditions merger approval on benefits reaching consumers. 

An ex ante regime centered on “freedom of choice” fits awkwardly within that framework. Sooner or later, CADE will have to decide which principle governs. 

A Better Gate, the Wrong Key

The designation provisions receive the substitute’s most consequential revisions. These rules determine which companies CADE may classify as having “systemic relevance in digital markets,” Brazil’s term for digital gatekeepers. In an interview, the rapporteur acknowledged that the original test was, to some degree, open-ended. 

The original bill allowed CADE to designate a company based on seven listed characteristics considered “non-cumulatively” and “among others.” The list was therefore nonexhaustive, and any one factor could, in theory, justify designation. The company also had to exceed either R$50 billion in annual global revenue, roughly US$10 billion, or R$5 billion in Brazilian revenue, roughly US$1 billion. 

The substitute replaces that framework with a two-step, closed test. The revenue thresholds become a threshold condition. CADE must then conduct a “joint and reasoned analysis” (análise conjunta e fundamentada) of six listed characteristics. The phrase “among others” is gone.

The substitute also changes how the thresholds may be adjusted. The original allowed ministers to revise them by joint act. The new text indexes them annually to Brazil’s IPCA inflation measure. That removes a political dial and prevents inflation from quietly pulling more firms into the regime, as has happened with Brazil’s merger-review thresholds. 

A new provision also makes explicit what the original bill left unclear. Designation alone would not impose any special obligation. Each obligation would require a separate process and justification. 

Most of these revisions move in the direction ICLE recommended. The “Digital Overreach” white paper criticized the original framework for combining open-ended criteria, politically adjustable thresholds, and a lengthy designation period. The substitute closes the list, fixes the threshold adjustment, requires a reasoned assessment of multiple factors, and, as discussed below, shortens the designation period. 

Two problems remain. First, the six characteristics are still linked by “or,” even though the statute requires a “joint and reasoned” analysis. The drafting points in two directions at once.

Second, some supporters argue that a higher bar would weaken the regime. That objection ignores the cost of designation. A designated firm, along with its entire corporate group, could face years of regulatory obligations, compliance reports, and heightened sanctions risk. 

An error-cost approach weighs the consequences of mistaken intervention against mistaken restraint. In fast-moving digital markets, false positives can be especially costly and hard to reverse. A designation carrying such consequences should therefore require substantial evidence.

The deeper problem remains unchanged. The test still does not require proof of market power. 

The rapporteur’s official release says the bill “does not start from the premise that large companies are a problem.” The statutory test suggests otherwise. ICLE recommended requiring proof of durable market power, consistent with the United Kingdom’s Digital Markets, Competition and Consumers Act (DMCC), which requires “substantial and entrenched market power,” and Germany’s Section 19a, which requires a finding of dominance. 

The substitute instead retains the “systemic relevance” framework. Its six factors remain structural proxies, including multisided business models, network effects, vertical integration, and access to data. Those features may reflect competitive success rather than durable dominance. 

The test also omits the questions a competition economist would ask first. Can new firms enter? Are new technologies disrupting the market, including artificial intelligence? Can users switch to substitutes? 

Designation still applies to the entire corporate group, although the substitute’s service-specific obligations reduce some of the practical breadth. The government also continues to estimate that five to 10 firms would qualify, most of them U.S.-based. That prospect carries obvious geopolitical costs and suggests that company size, rather than demonstrated consumer harm, remains the test’s center of gravity.

On its own terms, the substitute improves the designation test. It is narrower, more predictable, and more demanding. Its central conceptual flaw nonetheless survives.

Fewer Blunt Instruments, More Fine Print

Article 47-E lists the special obligations CADE may impose on designated firms. The substitute reorganizes them into three categories: transparency and reporting duties, positive obligations, and abstention obligations. The changes are substantive, and they cut both ways. 

The clearest improvement concerns merger review. The original bill required designated firms to submit every transaction to CADE, regardless of whether it met the turnover thresholds in Article 88 of Brazil’s Competition Law. That would have created a parallel merger-control regime. 

We argued that the rule was disproportionate. It would capture harmless deals, threaten the startup exit path on which venture investment often depends, and duplicate authority CADE already has under Article 88, Section 7, to call in below-threshold transactions. CADE has recently shown that it will use that power in AI acqui-hire investigations. 

The substitute scales the rule back. Designated firms must now notify the superintendence of below-threshold transactions, following the approach in Article 14 of the DMA. Any review would proceed through CADE’s existing call-in authority. The notice requirement remains blanket rather than risk-based, but the bill drops its most burdensome feature. 

The second improvement is mandatory tailoring. Each special obligation must apply only to specified services, products, or business practices identified in the proceeding. CADE must calibrate the obligation to “the competitive risks associated with each product or service,” and the Tribunal must identify the covered services and set an implementation timetable. 

That change partially adopts ICLE’s recommendation that remedies attach to particular services rather than entire corporate groups. It should prevent the kind of companywide obligations that have made the DMA so costly to administer. The corporate group would still receive the designation, but each obligation would target a particular digital product or service. 

The substitute also strengthens the balancing clause. The original text said CADE “may consider” information security, a firm’s other legal duties, and the functioning of the relevant product or service. The substitute says CADE “shall consider” those factors, as ICLE recommended. 

It also adds two new considerations: innovation, especially for small firms, and the “public interest in the competitive development of digital markets.” The latter phrase sheds little light. Still, mandatory consideration gives CADE a stronger duty to explain how it weighed competing concerns. 

The larger omission remains. The bill still provides no efficiencies defense, a problem discussed below. 

Against those gains, the substitute adds several new mandates. One provision would allow CADE to require “neutral and non-discriminatory” choice architecture, including choice screens, for selecting, installing, or setting third-party products as defaults. Dirk Auer and I have explained to the UK CMA why such remedies can raise consumers’ search costs without producing measurable gains in competition. 

A new abstention rule would also bar firms from worsening conditions for users who exercise rights created by the special obligations. That includes “subversion of user autonomy” through interface design, an anti-circumvention and anti-dark-patterns provision. 

The remaining changes are mixed. The vague ban on “predatory or abusive strategies” survives, though it now applies only where a firm “exploits the situation of dependence” of users. The anti-steering rule grows broader, extending to restrictions “of any other nature.” 

That expansion ignores the role some anti-steering rules play in protecting platform investments in product discovery, trust, and security. CADE’s settlement with Apple showed that those tradeoffs can be assessed through case-by-case enforcement. 

The self-preferencing ban also remains unchanged. Economic research does not support treating self-preferencing as presumptively harmful, and CADE has never based a digital-market conviction on it. 

More Process, Less Ambush

The procedural revisions work best as a package. Together, they make the regime slower, more participatory, and harder to rush. They also respond to ICLE’s concerns about legal certainty, institutional capacity, and rent-seeking. 

The designation period drops from as long as 10 years to six. A firm may seek review after two years if market conditions change significantly. Special obligations may also be revised after two years, and they would take effect within 90 days rather than 60. 

ICLE recommended designation terms of three to five years with mandatory periodic review. Six years, with review triggered only by changed conditions, still exceeds the five-year limits under the UK’s DMCC and Germany’s Section 19a. It is nonetheless a marked improvement over 10 years. 

Participation also expands. CADE must open a 30-day public-comment period when a designation proceeding begins. The consultation period doubles from 15 to 30 days, and the final opinion must expressly address the submissions received. 

A new voluntary-proposal process under Article 87-C allows a designated firm to submit implementation plans, technical parameters, monitoring mechanisms, and timetables. The proposals would not bind CADE, but the agency would have to consider them and explain its response. 

That is a modest step toward the negotiated, case-specific approach ICLE has argued better suits digital markets. CADE has limited experience prescribing product-design remedies. A rule developed with technical input from the firm will often work better than a rigid order that engineers must retrofit after the fact. 

The changes are not uniformly positive. “Any interested party” may still participate, and several public bodies can force CADE to open a proceeding. Referrals from CADE’s Tribunal, its General Superintendence, the Finance Ministry’s Secretariat for Economic Reforms, or any federal body responsible for digital markets or “diffuse and collective rights” trigger immediate proceedings and automatic intervenor status. 

The superintendence may now dismiss unsupported private complaints, which helps. But the bill still needs stronger safeguards against firms and advocacy groups seeking regulatory favors at a rival’s expense. 

The Alternative Becomes an Add-On

The substitute also creates two institutions absent from the original bill. Article 87-K establishes a nonpunitive market-study process. CADE could publish reports to inform designation decisions and evaluate, ex post, whether existing special obligations worked as intended. 

The bill also creates a nonbinding Advisory Council. At least half its seats would go to academics and nonprofit civil-society representatives, with appointment rules meant to reduce the risk of capture. 

ICLE recommended market studies as an alternative to an ex ante regime. The substitute instead adds them to one. Both institutions are sensible on their own. Neither answers the objections to the regime they would support. 

Better Guardrails, Same Destination

The substitute adopts many of the recommendations in the “Digital Overreach” white paper. It closes the designation list, requires reasoned analysis, shortens the designation period to six years, makes clear that designation does not automatically trigger obligations, and requires CADE to tailor each obligation to specific services. It also strengthens the balancing clause, reduces the proposed parallel merger regime to a notice requirement, limits rulemaking, and expands public participation.