The following is an excerpt from Louis Kaplow’s new book,“Rethinking Merger Analysis,” now out at MIT Press.
Why rethink merger analysis? Merger policy is unquestionably one of the most consequential domains of competition regulation throughout the world, and it has advanced greatly over the past half-century due in large part to advances in industrial organization economics. Merger analysis is thus quite important, and its broad features seem to be well settled.
But are they? And should they be?
When I contemplated launching this project in 2017, I had already written about fundamental defects in the market definition paradigm that underlies the so- called structural presumption (with recent efforts to reinforce and extend it) and other key features of modern merger guidelines and court precedents. But many of these ideas have not fully penetrated economic research, policy advocacy, and everyday merger practice, in part because it was not appreciated just how the key criticisms fit in and what their implications were. Hence the need to rethink, or at least to think some more. It turns out that the problems are worse than most imagine. An illustrative ΔHHI challenge threshold in a prominent, recent article would change by more than two orders of magnitude simply by varying the posited elasticity of demand within the range consistent with the HMT ratifying a narrow, homogeneous goods market. The pertinent formula also contradicts the use of the ΔHHI in the “relevant” market whenever the homogeneous market is broadened. Moreover, existing protocols can misorder mergers— presumptively challenging one and giving a pass to another— even when the latter merger would raise prices dozens of times more than would the former. Because defects in the market definition process are logical and absolute— market definition analysis can only degrade inferences, regardless of the information set—continued reliance on it, much less doubling down, has to be a mistake, and the errors can be large.
Existing analysis of efficiencies needs rethinking because so much has never been analyzed in the first place. Merger efficiencies must be merger specific, which usually means that they cannot be achieved through contractual arrangements short of a merger. Yet relevant research in industrial organization economics, including that applied to merger analysis, is almost entirely divorced from literature on the theory of the firm, contract theory, and organizational economics that directly addresses such questions (and, incidentally, is associated with several Nobel Prizes). Nor is relevant business and industry expertise consulted in analyzing merger efficiencies. Lacking an analytical framework and hence not examining much of the pertinent evidence, it is no wonder that efficiencies seem inscrutable.
Efficiencies (and entry) are also sequentially siloed in official merger protocols, relegated to the end of the inquiry, if one gets that far. But, taking a decision analytic formulation and using for concreteness the odds ratio formulation of Bayes’ rule, how can one form the likelihood ratio for updating one’s priors when there is no denominator? (Here, the likelihood ratio is the probability that the merger is anticompetitive given all the evidence divided by the probability that the merger is efficient given all the evidence.) Moreover, inferences from the merging parties’ rationality constraint require an assessment of efficiencies even if one cares only about whether anticompetitive effects, in a vacuum, exceed some threshold. In this way and others, standard protocols are patently irrational.
Entry is also fundamentally misanalyzed. Inquiries into whether an otherwise anticompetitive merger will induce fully corrective entry are misguided except in extreme cases (some of which exist) because higher profitability, caused by higher equilibrium prices, is usually required post-merger to induce entry that was unprofitable beforehand. Instead, the prospect of (partially mitigating) entry may indicate that a merger motivated by anticompetitive considerations is unprofitable ex ante, shifting the appropriate inference to efficiencies—again reasoning from the merging parties’ rationality constraint. Ex post entry also has important welfare consequences of its own, as established by famous but now neglected literature (associated with additional Nobel Prizes) from half a century ago.
Perhaps more importantly, entry that may be induced by the prospect of a subsequent buyout has been neglected until recently. And much contemporary analysis that does address such acquisitions takes entrants’ existence and capabilities as given, whereas the most important impact of a merger regime in this domain involves effects on ex ante investment incentives. Reduced incentives from a more stringent regime can greatly reduce welfare—because many forms of innovation are undersupplied—but sometimes welfare rises because imitative investment and entry are often socially excessive. Hence, the direction of ex ante innovation is a first-order consideration in setting the relative stringency of merger regulation in important classes of cases.
In these ways and others, merger analysis needs substantial rethinking. This book seeks, often relentlessly, to ask all relevant questions without regard to where the answers may lead or whether they can directly be implemented in merger review. This course of investigation is the appropriate way to set research agendas, formulate policy, and determine how best to analyze proposed mergers. Proxies and shortcuts are necessary, but much contemporary policy debate puts the cart before the horse: How can one determine which protocols are better than others without first performing the underlying analysis correctly for broad classes of cases? That is, how can we know what counts as a good shortcut if we have not figured out where we should be going or checked whether the proposed shortcut takes us over a cliff?
This book also departs from much merger policy advocacy by analyzing how optimally to order mergers, from the most dangerous to the most beneficial, rather than advocating for more or less stringency. If we know, for example, that a certain population is underserved by the medical system, pulling out all the stops to give more drugs and perform more surgery on this population is not merely a blunt prescription but a truly dangerous one: additional, intrusive medical treatment on people who don’t need it does not help offset but instead seriously adds to the harm from the failure to treat those who would actually benefit. The first- order problem is to figure out which mergers are more harmful or more beneficial, and to what degree. To address that problem, we need to rethink merger analysis.
Excerpt from “Rethinking Merger Analysis” by Louis Kaplow. Reprinted with Permission from The MIT Press. Copyright 2024.
Author Disclosure: The author reports no conflicts of interest. You can read his conflicts of interest disclosure here. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
