Dennis Carlton provides his round-two comments on the draft Merger Guidelines.
To read more from the ProMarket Merger Guidelines Symposium, please see here.
The round-one comments from the other participants of the ProMarket symposium on the draft Merger Guidelines confirm one of the primary concerns of my prior comment that, unless significantly revised, the proposed Guidelines will fail to receive the broad-based support that recent prior Guidelines have achieved. The Guidelines will no longer be viewed as a statement of current economic thinking on how to identify and analyze relevant evidence regarding the competitive effects of mergers. Instead, the Guidelines will be viewed as a legal document full of highly selective case citations designed to justify to skeptical judges and a skeptical legal community that the current aggressive antitrust enforcement regime is based on the case law. But as my prior comment and those of several respected legal scholars in this symposium point out, many of those cases cited in the draft Guidelines are old, their principles have often been rejected in subsequent court decisions, and those cases are often based on economic reasoning that would be rejected today. A simple comparison of the comments of the antitrust scholars who question the draft Guidelines’ interpretations of the cited cases to the comments of the antitrust scholars who favor the draft Guidelines and claim that the draft Guidelines reflect what antitrust should really be about is sufficient to illustrate the existence of profound disagreement that strikes at the very heart of the draft Guidelines’ attempt to justify the aggressive antitrust agenda of the current enforcement agencies.
Even without taking a position as to which side is correct, there is no doubt that such disagreement—rather than consensus—is what one can expect going forward, unless the draft Guidelines are revised. If the enforcement agencies want to have a separate policy paper explaining to judges and the legal community why, despite recent rejections of their legal theories by several courts, their approach is the correct one according to the case law, then they should, by all means, go ahead and do that. But it is a mistake, in my view, to change the Merger Guidelines from a document that has achieved broad consensus and provided useful guidance to practitioners and judges as to what economic evidence is relevant and how to analyze it into a legal guide for aggressive antitrust enforcement. Please keep the Guidelines as they have been—a guide as to what is relevant economic evidence. The problem with the draft Guidelines won’t be fixed simply by moving the economic appendix into the main text, though that should be done. The problem will be fixed only if the Guidelines do not demote economic evidence in favor of strong presumptions often based on market shares and on old cases. The draft Guidelines read as a legal brief to judges and a demotion of economic analysis. The Agencies need to decide if that is what they want to do. I fear that it is.
The other major concerns I raised were the unclear focus in the draft Guidelines as to what exactly they think are the desirable economic effects of competition and whether the draft Guidelines have replaced economic effects evidence with evidence of market concentration based on a formulaic application of HHI thresholds. I still do not know the answers even after reading the comments of those symposium participants who especially praise the current draft Guidelines. Is having lower concentration and more competing firms desirable even if that results in higher prices? Are the draft Guidelines proposing the breakup of prior mergers into small firms regardless of the consequences simply because having less concentrated markets is the goal?
Since several of the other commentators have discussed the shortcomings in many of the 13 Guidelines that I did not discuss, I will not review them here except to say that I share many of their concerns. Let me focus instead on the comments related to the points I did discuss. I explained that it appeared that the draft Guidelines devoted insufficient attention to the possible efficiencies from mergers. I understand that one can disagree about the importance of those efficiencies for a particular merger. I also agree that one should carefully evaluate evidence of such claimed efficiencies before crediting them. But I don’t see how one can fail to pay close attention to the fact that aggressive merger enforcement risks depriving the economy of efficiency enhancing mergers that will benefit consumers. One cannot dismiss efficiencies from mergers, as one commentator appears to do, with the suggestion that there must not be substantial merger efficiencies since so many mergers fail. Many mergers involve no antitrust issues. Are those just a random pairing of firms unrelated to efficiency? Surely, there must be some recognition that efficiencies achieved through merger are a plausible source of benefits to consumers and that reducing that source could create harm.
I explained in my earlier comment the important omissions in the discussion of vertical mergers, primarily the lack of recognition of the clear and significant differences between a horizontal merger and vertical merger. One involves the combination of substitutes, the other of complements. That fact means the incentives for harms and benefits from a merger differ greatly between the two. For example, the elimination of double marginalization is a well-accepted possible source of benefit of vertical integration, among other benefits. Of course, not every vertical merger is procompetitive, and so one often needs a vertical model that takes account of all the possible harms, such as raising rivals’ costs, as well as the benefits in order to come to a conclusion about the net effect of the vertical integration. I endorse that idea. I have written about it and testified to it (e.g., in Illumina/Grail). But suppose there is a credible contract—perhaps one involving commitments to binding arbitration over readily observed variables—that promises that the price to a rival will not rise from what it would be otherwise (i.e., in the absence of the merger). In that case, all else equal, the elimination of double marginalization would be expected to lead to benefits to consumers since there would be no raising rivals’ costs. Do the commentators who favor the draft Guidelines accept that reasoning? I cannot tell. Do they want to prevent a competitive disadvantage to a rival even if that prevention leads to a harm to consumers?
It is worth explaining a simple concept that has sometimes led to confusion in vertical cases, resulting in a benefit being mistakenly labeled as an anticompetitive effect. Suppose initially Firm A supplies, say, steel to Firms B and C at $100 per ton and that the marginal cost of steel is $50. Now suppose that there is a merger between Firms A and B. After the merger, assume that the price that Firm A charges Firm C either remains at $100 or falls below $100 but remains above steel’s marginal cost of $50. The fact that the merged firm will ordinarily treat the shadow price of steel to itself at $50 is a benefit of vertical integration and not an example of raising rivals’ costs, despite the fact that the rival now faces a higher shadow price for the input than the merged firm. The cite in the draft Guidelines referring to the preservation of “fair” competition raises the specter of making this error.
The draft Guidelines’ and commentators’ discussion of the use of contracts in vertical mergers raises an important issue that applies to all mergers. When should the enforcement agencies be allowed to dismiss contracts as meaningless because, according to the Agencies, they are easy to breach? Would it matter if there were severe penalties for breaching the contract and failure to adhere to the commitments were easy to observe? Would it matter if these contracts were similar to other existing contracts in the industry? Would it matter if these types of contracts had been used to address antitrust concerns in prior cases and that even the enforcement agencies had said they worked successfully (as, e.g., in the AT&T/Time Warner case)? Is there an inconsistency in the enforcement agencies’ apparent positions that contracts to “fix” an alleged antitrust concern are not credible but that contracts to enable non-merging parties to achieve the same benefits as a proposed merger are credible? How can both positions always be true? Faced with this inconsistency, a reasonable analyst might conclude that the goal of the enforcement agencies is to ignore a contract when it interferes with the enforcement agencies’ ability to stop a merger (as could occur if the parties claim the contract fixes the antitrust concern) but applaud a contract when it enables the Agencies to stop a merger (as could occur if the enforcement agencies claim that the possibility of using contracts eliminates all merger-specific efficiencies). The draft Guidelines should be revised to include an expanded discussion of the enforcement agencies’ views of the role of contracts.
As I said in my prior comment, there are some additions to the draft Guidelines that are useful. Yet even here there is some confusion that the supporters of the draft Guidelines did not clearly address. In particular, the discussion of labor correctly points out that a merger can create market power and harm in the labor market. Yet the draft Guidelines go on to explain that these harms will not be balanced against other benefits elsewhere, such as in the product market. One gets the sense that the draft Guidelines are making the fundamental error that a merger that creates simple monopsony and drives down wages will lead to lower consumer prices. That is, of course, wrong as a matter of basic economic theory. The monopsony leads to a restriction of labor and lowers wages but that act does not lower the shadow price for labor to the firm when the firm is deciding how to price. (I doubt the drafters are making this error, but I fear that the draft Guidelines have the potential to mislead readers on this point.) But there is a more fundamental error. If there were some efficiency from the merger—say a technological one—would that not matter if it led to lower consumer prices despite the creation of monopsony? Surely some tradeoff must be relevant but apparently not according to the draft Guidelines. What if a merger harms some workers in one location but benefits more workers in another? Should the harm to the few be allowed to stop the merger? The Department of Justice should look at how it has examined past airline mergers. It used to take the position that a harm in one market can be balanced against a benefit in another market when the two markets are “inextricably linked.” Otherwise, one would never allow any airline merger in which route structures shift. Has it abandoned that position? See my article forthcoming in George Mason Law Review that explicitly deals with this issue for further details.
There is one point that, to my knowledge, no one has made but that would be a useful addition to the Guidelines. The enforcement agencies should promise to undertake studies of their past merger decisions. They should keep track of what their models, economists, and lawyers said and what the merging firms’ models, economists, and lawyers said and see who was right. In horizontal cases, they should ask what happened to prices, wages, output and innovation. (We do now have several retrospective studies but still not enough and not in the detail that I have just described.) In potential competition cases, they should ask whether their claims as to which firms were the potential competitors turned out to be true or false. In vertical cases, they should see whether their foreclosure concerns were mere speculation or turned out to be true. Otherwise, the ability of lawyers and economists to intelligently debate antitrust policy will not move forward much.
Finally, the goal for the Guidelines should be that they provide a coherent economic framework for analysis so that an analyst that applies the economic methodology in the Guidelines to the economic evidence, assuming that sufficient evidence exists, comes to the same conclusion regardless of the analyst’s prior beliefs about whether mergers in general are good or bad. Antitrust analysis should not be a religion but an exercise in economic analysis of the effects of a merger.
Author Disclosure: Dennis Carlton is a senior managing director for Compass Lexecon, an economic consultancy that specializes in competition practice and law. He consults for and against many major firms, some of which have a financial stake in the Merger Guidelines. Particularly relevant for this article, he has served as an expert for defendants in the AT&T/Time Warner, Illumina/Grail, Microsoft/Activision, and Meta/Within mergers. For more information on his consulting work, please see his Compass Lexecon profile.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
