Eric Posner provides his round-two comments on the draft Merger Guidelines.
Suppose that some number (n) of firms with similar market shares compete in a well-defined market and two of them plan to merge. Suppose further that the two firms charge price equal to 10 and incur marginal cost equal to 5, and it is predicted that the post-merger price will equal 8 and marginal cost will equal 2 at the combined firm. Market power increases but price declines. Should the Federal Trade Commission or Department of Justice Antitrust Division (the Agencies) block or permit the merger?
Plainly, competition in the ordinary sense declines. This is obvious if n=2 (merger to monopoly). It is obvious if n>2 as well. For some large n (n>6?, >10?), we might say the merger doesn’t “substantially” lessen competition. The Merger Guidelines until 2010 endorsed this view by saying that market power was the unifying theme.
In the usual models, efficiency gains could result in a price decline despite the reduction in competition, not because of it—just as long as the cost savings accrue to the consumers rather than the producers. Efficiency gains could “rebut” the presumption of lessening competition only if it disrupted collusion or in other ways reduced industrywide market power. Previous Guidelines never made this distinction clear—probably because economists and lawyers could not agree on what competition meant, I suspect—and the ambiguity lurks in the cases.
The 2010 Guidelines purported to resolve this ambiguity through a sleight of hand. They added a new sentence shortly after the “unifying theme” sentence: “A merger enhances market power if it is likely to encourage one or more firms to raise price … or otherwise harm customers as a result of diminished competitive constraints” (emphasis mine). The example with which I started—a merger that increases market power and reduces price—is ruled out by fiat (those words appear in no earlier version of the Guidelines). Such a merger would be permitted because of the price reduction. The distinction between a merger that reduces prices because of an increase in competition and the merger that reduces prices despite a reduction in competition is collapsed, the language of “rebuttal” (rather than defense) being the only clue that it once existed.
A merger between large competitors almost always increases market power, whether or not it reduces prices. The plain meaning of the statute would bar those mergers. A typical way to try to avoid the plain meaning of a statute is to subordinate the text to some theory of the statute’s purpose. For example, one could say that the point of Section 7 of the Clayton Act is to secure the “benefits of competition” for the public. Since the benefits of competition are said to be price reductions, the statute is converted into a command to block mergers that increase prices rather than mergers that lessen competition. But legal conventions forbid courts to evade a statute by replacing a legislative command with a command to achieve the assumed legislative purpose. Otherwise, every statute becomes a guessing game as well as an invitation to disregard the goals of the legislators who brokered deals. Verbal tics and weird circumlocutions replace plain language. Because “competition” now means “price reductions,” you have to find a new word for competition, for example, “rivalry per se.” If you read around in the economics literature, you’ll see that other economists will sometimes use the word “anticompetitive” to mean inefficient and thus be compelled to use terms like “direct competition” to mean “competition.” All of this jumble began with Robert Bork’s long discredited claim that the antitrust statutes meant efficiency when they said competition: a zombie that will never die.
But let’s say that the Agencies should care about consumer welfare, either as a matter of law or (more plausibly) enforcement policy. That’s what lowering the HHI thresholds does. It’s a different question whether the Agencies should also be required to calculate price effects of mergers they review on a case-by-case basis and prove price effects in court if the merging parties refuse to fold.
This cumbersome style of enforcement, favored by many symposium commentators, requires a more robust defense than anyone has given it. No other enforcement agency I am aware of considers itself bound to do consumer welfare-style or other cost-benefit analysis on a case-by-case basis when it enforces the law. Not the Securities and Exchange Commission, the Commodity Futures Trading Commission, nor the Federal Deposit Insurance Corporation when it shuts down a bank. Not the U.S. Attorney’s Office in New York when it investigates an insider-trading allegation. Many federal agencies, like the Environmental Protection Agency, for example, conduct cost-benefit analyses when they issue regulations; but not when they enforce them. The tradition in the U.S., and no doubt every other well-functioning country, is to give enforcers discretion so that they can enforce the law where enforcement is most needed. This will generally be based on judgment and enforcement experience, informed but not controlled by empirical research. If the 2023 draft looks like a loosening of a straitjacket, that’s what it is.
The point is not that consumer welfare is irrelevant; it is that too strong of a focus on it can interfere with enforcement even when enforcement is designed to advance it. Bork understood this; in The Antitrust Paradox, he said he opposed an efficiency defense because of the complexity of measuring efficiencies. His view was uncharacteristically prophetic. Today, every merger challenge bogs down into trench warfare between experts where victory hangs on the randomly assigned judge’s empirical sophistication. For risk-averse agencies who bear the burden of proof, that can only ratchet up the standard for litigating cases. It is familiar from many other areas of the law, that even if all you care about is consumer welfare, you may not want to make consumer welfare a central part of enforcement, if poor data, complex facts, contestable methods, and easily confusable judges stand in the way. That was the real but forgotten lesson of Philadelphia National Bank, which was more notable for its formalism than for its economics.
The antitrust standing cases do not provide much support for Herb Hovenkamp’s argument that the Supreme Court has revised its Section 7 jurisprudence. Standing cases are oriented to the practical needs of private litigation and nearly always assume that the plaintiff’s substantive theory is correct. In Cargill, for example, the Court avoided expressing an opinion as to whether Section 7 is violated when a firm is driven out of business (“Even if respondent is correct that Congress intended the courts to apply § 7 so as to keep small competitors in business at the expense of efficiency…”). The Court reasoned that because price-cutting (as long as it is not predatory pricing) is not a violation of the antitrust laws, there is no antitrust injury to a firm that loses profits because competitors merge and charge lower prices. The Court thus left open the argument that a merger that eliminated a competitor would violate section 7.
The 1986 opinion seems to reflect a contradiction between Section 7 and the then-emerging doctrine of antitrust standing. The plaintiff complained that it would lose profits; normally we would think that means the merger would increase competition, a lawful result. But if the merger caused the plaintiff’s profits to increase, then it would not sue. Combine antitrust injury doctrine and Section 7, and it would appear that a competitor can never sue to block a merger even though the Clayton Act creates a private right of action. The Court escapes this catch-22 by ruling that a competitor has standing only if it is destroyed entirely—then it is clearly harmed and the merger has reduced competition, with n declining by all of 2. Profits should be as low as possible but no lower than that. Competitor standing lives on in the predatory pricing exception and in challenges by targets of hostile takeovers who have sometimes argued successfully in lower courts that they have standing to challenge the acquisition because n declines by 1 and they lose their freedom of action. No need to show consumer harm.
The confusion in Cargill probably reflects the Court’s growing ambivalence, soon-to-be hostility, about private enforcement, rather than a rethinking of its merger jurisprudence. The Court may have thought that the government, not a competing firm, is the better plaintiff to challenge mergers like the one in Cargill as the real worry is the political and social impact of monopoly, not the profitability of particular small firms.
All government enforcers take into account the impact of their enforcement decisions on people, and will typically give priority to violations that cause the most concrete and easily identifiable harm to the most people while never limiting themselves to such cases. The HHI thresholds implicitly count harm to consumers and other counterparties (like workers) as the major harms that justify enforcement priority for large-firm mergers rather than small-firm mergers. In that sense, consumer or counterparty harm remains central to the Guidelines, contrary to the suggestion of some commentators. The three mistakes that permeate discussions of the draft Merger Guidelines are that only harms to counterparties should matter, that those harms are required to be calculated on a case-by-case basis, and that those harms are a requirement of Section 7 rather than a factor to be used by the Agencies at their discretion as they allocate resources to the enforcement of the law.
Disclosure: Eric Posner worked on the draft Merger Guidelines while serving as counsel to the assistant attorney general of the Antitrust Division, Department of Justice, in 2022.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.