Same Rule, Different Result: How the Narrowing of Product Markets Has Altered Substantive Antitrust Rules

It has long been recognized in antitrust cases that market definition is typi cally malleable and frequently outcome determinative. In United States v. Grinnell, a Section 2 case, Justice Abe Fortas dissented from the definition of a market so narrow he called it a “strange red-haired, bearded, one-eyed manwith-a-limp classification.” In more recent years, commentators have argued both that the Court in Grinnell defined “excessively narrow submarkets” and that those submarkets were “consistent with the evidence as to demand substi tution.” In other words, the market could be both implausibly narrow and correct, particularly if judged by today’s standards, when product markets often require multiple adjectives.

The breadth of the relevant market mattered in Grinnell, as it does in merger challenges brought under Section 7 of the Clayton Act, because—as the Supreme Court recognized many years ago—“market definition generally determines the result of the case.” Former U.S. Federal Trade Commission Chairman Robert Pitofsky likewise called it “the most important single issue in most enforcement actions.” And former FTC Chief Economist Jonathan Baker has said the issue has determined “the outcome of more cases . . . than. . . any other substantive issue.” Market definition continues to play a determinative role in merger challenges today, especially as alleged markets have become narrower.

The primacy of market definition in antitrust analysis—at least in the courts—reflects the large number of substantive legal rules that rely, either explicitly or implicitly, upon it. For merger matters brought under Section 7 of the Clayton Act, how the decision-maker defines the ambit of the market determines both whether the merging firms are deemed competitors in the first place and whether their merger would substantially diminish competition. Since at least 1963, when the Supreme Court decided United States v. Philadelphia National Bank, courts reviewing Section 7 claims have assessed the competitive effects of a transaction within, rather than across, markets.

Despite the importance of market definition, the rules that govern it are flexible enough to support a range of permissible choices. As the U.S. Department of Justice Antitrust Division (together with the FTC, the Agencies) observed in a brief filed in 2015, “[f]requently, the government alleges narrow markets, the defendants describe broad markets, and the court must choose between the competing approaches.” The choice is often outcome-determi native, leading many to charge that market definition is “an essentially ex post choice” designed  “to  achieve  the  desired  results  in  calculating  market shares.” This gripe is long-standing; in the 1960s, commentators charged that “the Government has not been averse to shifting its market theories from case to case, seemingly with little justification other than making the relevant percentages more favorable to its cause.”

Given this broad discretion, market definition can vary not just from one case or judge to the next, but also over time, as new tools and theories gain purchase. These changes, in turn, may affect the way substantive antitrust rules are applied, even if those rules have not themselves changed. In the 1980s, for example, Robert Pitofsky objected to the Merger Guidelines of that era because they—at least in his estimation—“have tended to expand relevant markets and thus diminish apparent market power.” Other commentators took the opposite view of the same Guidelines, predicting narrower markets. More recently, Jan Rybnicek and Josh Wright argued the 2010 Horizontal Merger Guidelines would lead to narrower markets and fewer cognizable efficiencies than was the case under the preceding Guidelines.

Despite the importance of market definition and occasional predictions about how policy may affect it, whether markets have actually and systematically changed in scope over time is ultimately an empirical question, and one that existing surveys have not squarely addressed. This article tries to fill the gap.

Part I tests the hypothesis that product markets are frequently, and perhaps systematically, narrower today than those used by the Supreme Court to set the foundational legal rules. It finds that, of the 12 product markets defined in Section 7 cases decided by the Supreme Court, half have narrowed over time, with six product markets (used in 12 Supreme Court cases)—banking, beverage containers, energy, footwear, groceries, and spices—narrowing markedly. The remaining six product markets (used in seven Supreme Court cases)—automotive paint, beer, electrical conductor, natural gas, sodium chlorate, and spark plugs—have remained more or less the same. Remarkably, none of the 12 product markets has broadened since then. In other words, whereas previously the Supreme Court and lower courts defined a mix of broad and narrow markets, courts today usually define narrow ones, leaving litigants to debate whether the market is narrow or very narrow, with Ohio v. American Express Co. constituting the rare exception.

Part II describes four likely causes for the narrowing of product markets. First, the methodologies used to define markets have changed substantially over the years. As enforcers shifted their focus to differentiated products and unilateral effects, the models and tools used to define markets began to change, with the hypothetical monopolist test, diversion ratios, and upward pricing pressure becoming more common features of the market definition exercise. Second, as confidence in measures of demand-side substitution improved, courts and enforcers began to de-emphasize supply-side substitution, which has been excluded entirely from the market definition exercise since the 1992 Horizontal Merger Guidelines. Third, the Guidelines have emphasized some concepts, like price-discrimination markets, that can result in narrower markets. Fourth, the narrowing of product markets may reflect—at least in some cases—real changes in the underlying economy, like greater product differentiation. For example, the “premium natural and organic supermarkets” at issue in Whole Foods did not exist when the Court decided United States v. Von’s Grocery Co.; neither did “broadline foodservice distribution to national customers.” Yet it is unlikely that greater differentiation can explain all, or even most, of the narrowing. Some products were already differentiated in the 1960s; there were many kinds of children’s shoes, even though the Court consciously chose to lump them all together. Likewise, the courts have narrowed commodity product markets like coal and spices markedly, even though the products’ physical properties have not changed.

Whatever the cause, when a court defines product markets more narrowly today than in yesteryear, it necessarily applies substantive legal rules in a systematically different way than when those rules were first announced. Part III considers three legal rules: (1) the exclusion of out-of-market merger efficiencies; (2) the structural presumption, i.e., the market share at which a transaction becomes presumptively unlawful; and (3) the traditional emphasis upon mergers involving “overlapping” horizontal competitors. Narrowing product markets have altered rules (1) and (2) in ways that favor plaintiffs, while narrowing product markets have altered rule (3) in ways that can favor either plaintiffs or defendants.

Part IV uses two examples from the banking industry, whose dual-enforcement structure provides a useful comparison, to examine whether narrower product markets have in practice altered the way enforcers apply substantive rules. Whereas the DOJ has adopted ever narrower markets during its antitrust review, the Federal Reserve Bank (FRB) has retained the 1960s-era “broad market” approach, often by broadening the geographic markets further to account for suburban sprawl. In at least two bank mergers, CoreStates/ FirstUnion (1998) and BB&T/SunTrust (2020), the agencies explicitly acknowledged that this difference in methodology has led to different substantive results, with the DOJ estimating significantly greater harm and fewer cognizable efficiencies and therefore demanding larger divestitures than the FRB. These examples therefore illustrate both how antitrust product markets have narrowed and how that change affects the application of legal rules like market share thresholds.

The narrowing of markets also has important policy implications. Two merit brief mention. First, because the narrowing of markets has the effect of making antitrust enforcement more stringent, at least on average, it cuts against the narrative that antitrust rules have become “overly lenient” since the 1980s. Nor is this effect fully offset by higher market share thresholds as in recent years product markets have continued to narrow even as thresholds have remained unchanged. This conclusion is underscored by the case studies described in Part IV, which demonstrate that, ceteris paribus, bank merger enforcement is more stringent when markets are drawn narrowly. Second, and relatedly, proposals to return to 1960s-era antitrust rules—which, to be clear, we do not endorse—should do so wholesale, reverting to both lower thresholds and broader markets. Or, to borrow a phrase from Justice William Rehnquist, these proposals should avoid cherry-picking, instead taking “the bitter with the sweet.”

Continue reading the full text of this article and citations in PDF format.