FTC v. DOJ on Section 2: Just Different Priors?

Cite this Article
Joshua D. Wright, FTC v. DOJ on Section 2: Just Different Priors?, Truth on the Market (October 15, 2008), https://truthonthemarket.com/2008/10/15/ftc-v-doj-on-section-2-just-different-priors/

Turns out the Global Competition Policy issue on Reviewing the DOJ Report on Competition and Monopoly, in addition to the articles I pointed to in this post, has added a few more responses to the Report, the FTC Response, and what the schism might mean for antitrust enforcement over the next several years. So far I’ve read and very much enjoyed the articles from Tom Barnett (DOJ), Luke Froeb and Pingping Shan, and Sean Gates (the others look good too).

One of the emerging points from these top notch antitrust commentators is the hypothesis that, building on the error-cost framework which suggests that optimal antitrust liability rules should be designed to minimize the sum of Type I and Type II errors as well as administrative costs, perhaps the explanation of the schism between the FTC and DOJ is just about the different priors that these agencies had about the relative frequencies and magnitudes of such harms. Its an interesting point. And one that while not mutually exclusive with my opinion that the agencies have different views about the role of economics in antitrust enforcement, suggests that it might not be the whole story.

This “different priors” story is also something that he been used to explain divergence between the US and EU with some success. Its a story that is tempting — in part because it allows one to evaluate the debate in terms that do not assign fault to either side. In addition to being tempting, I’m sure there is some truth to the story that the agencies have very different priors about the competitive effects of certain types of conduct. But I’m not convinced that different priors is really the story here. It strikes me as one of those explanations that explains nothing by predicting everything. Here are a few questions about this explanation that comes to mind.

First,  what evidence is there in support of the different priors explanation?  If the rift is really about empirical evidence concerning the likely impact of competitive conduct, and therefore the relative frequencies and magnitudes of Type I and Type II errors, that discussion is taking place outside of the public documents.  Different priors about competitive conduct presumably respond to the available evidence.  But as I’ve pointed out that the FTC Statement is incredibly light on reliance on economic theory or evidence and heavy on rhetoric, e.g.  “the Department’s Report is chiefly concerned with firms that enjoy monopoly or near monopoly power, and prescribes a legal regime that places these firms’ interests ahead of the interests of consumers. At almost every turn, the Department would place a thumb on the scales in favor of firms with monopoly or near-monopoly power and against other equally significant stakeholders.”

Second, where did the different priors come from?  The rift is on Section 2 is somewhat new.  Did the FTC and DOJ just recently diverge on views on Section 2 conduct?  Or has this rift been a long time coming?  Its one thing to say that two different enforcers have different priors as an explanation for the rift.  But I think that some deeper digging into the formation of those priors is worth doing.Third, I think there is some evidence that the FTC and DOJ do not agree that the error-cost framework is the right approach to thinking about the design of antitrust rules.  If true, its a major deviation from a consensus view that has developed in the cases and scholarly commentary in favor of the error-cost approach.  The basic idea, derived from Judge Easterbrook’s classic:  The Limits of Antitrust, 65 TEX. L. REV. 1 (1984), is that liability rules should minimize the sum or error and administrative costs.   In the antitrust context, Easterbrook argued that the optimal rules should be biased towards false negatives because false acquittals are self-correcting and the identification of anticompetitive conduct is such a difficult task.  There is some evidence that the FTC Majority Commissioners do not think that the error-cost approach, contrary to the consensus view, is appropriate (with my comments in the paragraphs below each bullet):

  • In response to the DOJ Report’s adoption of the error-cost framework and explanation of its concerns over over-enforcement, the FTC states that the challenge of identifying anticompetitive conduct “is not unique to Section 2” and asserts its confidence that “the federal antitrust enforcement agencies and the private antitrust bar are up to that task.”

The confidence in the ability to distinguish pro-competitive from anti-competitive single firm conduct in a setting that redounds with complex welfare tradeoffs, static v. dynamic efficiencies, and innovation, its remarkably different from the “first, do no harm” principles espoused in the DOJ Report.  Certainly, this passage supports the “different prior” story as well.  But again, the defense of the more aggressive FTC position does not seem to be about evidence or our knowledge about competitive effects.  Rather, this is about confidence in lawyers and regulators to get the right answer.

  • In response to the Easterbrookian point that false positives are likely to have greater social costs than false negatives because the latter are self-correcting, the FTC says “Even if correct, however, this hypothesis does not adequately consider the harm consumers will suffer while waiting for the correction to occur.  Markets can and do take years, even decades, to correct themselves. For one reason or another, it may take a long time for rivals to surmount entry barriers or other impediments to effective competition. Indeed, the monopolist’s own deliberate conduct may further delay a market correction and prolong the duration of consumer harm.”

Of course, it isn’t correct that because markets take some time to correct themselves that Type I and Type II error costs are equal.  The claim is not that false negatives are costless, it is that they are less costly than false positives.  False negatives, everyone agrees, have self-correcting tendencies that operate with varying degrees of speed.  False negatives have no such self-correcting mechanism.  The point that false negatives might self-correct slowly is not sufficient to address the theoretical point.

Its certainly true that one can design rational responses to different priors that result in different monopolization policies without assigning error to any of the designers.  I’m willing to concede that different perceptions of the likelihood and magnitude of errors are part of the story here.  Unfortunately, I think its a small part.  Were it a larger part of the discourse, I think a very rational discussion could be had about what sorts of rules and enforcement activities are warranted by the available evidence and which are not.  Sadly, I don’t see too much of that type of discussion.  For the record, my opinions here track closely to my view of the convergence issue between the United States and Europe: too much deference to different theoretical priors without respectfully hashing out how those priors hold up to available evidence.