This article was published in the Harvard Law Review, and can be found at https://harvardlawreview.org/2017/03/on-the-relevance-of-market-power/.
This is a book-length all-out attack on the conception of “market power” as it is currently applied. Like most law and economics literature, it is concerned with how best to design laws in order to maximise the social benefit of law enforcement. While the argument is not easy to encapsulate, I think the main point is that greater market power – understood as “the degree to which price can profitably be elevated above a competitive level, often taken as marginal cost” – does not necessarily mean that greater anticompetitive effects must follow. Instead, market power can have a variety of different effects, depending on the specific impact of the relevant practices on welfare. As such – and unlike what we do today – the concept of market power should be derived backwards from the level at which an antitrust infringement should be found, based on the impact that market power may have on the balance of anti- and pro-competitive effects of any practice, as assessed on a case-by-case basis.
What follows is a rather long description of the article, but I’m afraid it’s the only way to do it justice. Also, I’m trying to make sense of the article, given my limited knowledge of economics:
- Section II analyses the role of market power in establishing an antitrust infringement. Official rhetoric is that infringements (by monopolists, one would assume) require a two-step analysis: (i) is there enough market power; (ii) if so, is there an infringement. As Kaplow points out, market power must exceed some threshold in order for an infringement to be found. And yet, as a practical matter, a decision-maker faces significant uncertainty about the location of that threshold and about how much market power is actually present.
The author reviews the sources of this uncertainty. it concludes that because anticompetitive effects arise from a mix of market power and corporate practices which effects varies from case to case, this uncertainty is unsurprising: even assuming that market power and the effects of a certain practice can be separated, the requisite level of market power for a finding of infringement of competition law will depend on market conditions and, more specifically, on the effects of the relevant practice.
- Section III then moves on to review the various ways in which market power can feed into assessments of antitrust infringements.
Market power is relevant because: (i) it can provide an indication of the likelihood of certain practices being anti- or pro-competitive; (ii) it can influence the magnitude of anti- and pro-competitive effects of an anticompetitive practice. Given both these variables, it is clear that the relevant market power threshold must be variable: “in basic settings, the force of any evidence is given by its associated likelihood ratio, and it is impossible to assess the value of a ratio without regard to its denominator. More concretely, as we will see, there are settings in which market power raises the anticompetitive effects side but, in a closely related (and sometimes even identical) fashion, raises the procompetitive effects side as well.” In other words, it makes no sense to look at market power and the effects of a practice autonomously – they are closely intertwined. This section also provides an overview of some situations when market power makes it more or less likely that a practice will be anticompetitive – and concludes that “the relationship between market power (…) and the magnitude of [anticompetitive harm] is substantially more heterogeneous than is generally appreciated.”
A connected argument is that separating the assessment of anti- and pro-competitive effects might make heuristic sense, but ultimately leads to a focus on whether the practice may be anticompetitive which disregards the magnitude of potential pro-competitive effects: “One often hears — in panels of economists, academic commentary, and decisions by agencies and courts — that the outcome of a case should depend on which explanation [i.e. theory of harm or theory of lack of harm] is “right,” which in context often refers to which type of explanation is more likely. Supporting this interpretation is the frequent lack of any explicit attention to either [expected harm] or [expected benefit]. (…)Yet it is obvious that these magnitudes can readily be as important as the relative likelihoods in deciding whether to assign liability”.
- Section IV expands the analysis of Section III and assesses the use of “market power” as a screening tool and in determining optimal sanctions. The use of market power for screening is criticised because, since the various effects of market power are not well understood, screening based on such a defective tool is bound to be inadequate. If questions regarding market power and liability are controversial and subject to intense disputes at the final decision-making stage, surely such an instrument is unfit for preliminary assessments as well.
Regarding optimal sanctions, sanctions deter potentially pro-competitive activities while mere prohibitions do not. Which one is preferable should depend on an empirically grounded assessment, for which a proper understanding of market power is relevant .
- Section V reflects on the implications of the analysis developed above for contemporary law doctrine and commentary.
In particular, it is argued that the independent treatment of: (i) market power and anticompetitive acts; (ii) anti- and pro-competitive effects does not reflect what happens in practice, is analytically inappropriate and is practically counter-productive. In other words, final decisions cannot really be made based on a process where each step of the analysis is taken in turn, independently from one another. Instead: “One suspects that a more holistic analysis might be undertaken; then, given the liability decision that seems optimal, the decision maker may back out conclusions on the constituent parts and draft a formal decision accordingly. Moreover, it is possible that this process is to a significant degree informal or subconscious.”
This conclusion may come as an obvious one to anyone who has ever dealt with a court or agency (it is, after all, almost trite for anyone familiar with the literature on decision-making in this area). Kaplow also links this to the “special duty of dominant companies” in EU law, and actually kind of endorses it as being more honest than a pure step-by-step analysis (with the caveat that “such conclusion should be derived analytically and empirically”).
I’m not sure how useful this is for practitioners (or practice), but I think it is an undoubtedly important piece. I am not qualified to assess his theory on economic terms, but the thrust of the argument seems persuasive to me. On the other hand, Kaplow does not really provide any analytical alternative to market power. Nor does he acknowledge that his approach requires a pure effects’ based approach be adopted across the board – which will be difficult to administer, or even delineate in terms of partially applicable rules. In other words, further work will have to be pursued before the intuitions outlined here are susceptible of being adopted.