Steven Salop provides his round-two comments on the draft Merger Guidelines.
Merger enforcement has been underdeterrent in numerous markets (as noted here and here), and should be more interventionist. In the first round of the symposium, Professor Dennis Carlton and others raised questions regarding the justifications for the various presumptions in the Guidelines designed to encourage a more robust antitrust enforcement. These critiques also raise the related issues of evidentiary standards and rebuttals in the Guidelines. My Round II comments analyze and defend the economic justifications for more interventionist merger enforcement and legal standards. They also focus on the burdens of proof and presumptions used in various places in the draft Merger Guidelines and comment on several specific guidelines.
50% Foreclosure Share Anticompetitive Presumption
Professors Jonathan Baker, Nancy Rose, Fiona Scott Morton, and I have recommended that useful anticompetitive presumptions can be applied to foreclosure concerns, in which a dominant firm denies another firm access to an input good or service. In Round One of the symposium, Jennifer Sturiale and I proposed importing the anticompetitive presumption based on a 50% foreclosure share (i.e., the share of input purchases of downstream rivals that are accounted for by the upstream merging firm) into Guideline 5, while eliminating Guideline 6.
As a legal matter, Microsoft explained that a foreclosure share of 40-50% is commonly required to presume monopoly power in Section 1 exclusive dealing cases. Other courts (here and here) suggested 30-40%. The Department of Justice’s Section 2 report reported that 50% has been viewed as the minimum share for presuming monopoly power. A share of 60% was famously applied in Alcoa. These sources suggest that a 50% foreclosure share could be sufficient to shift the burden here to the defendants (the merging parties) since Section 7 is more interventionist than Sections 1 or 2. Or, if the term “presumption” makes readers (or courts) uncomfortable, one can equally well say that a 50% foreclosure share should be sufficient evidence of likely ability and incentive to foreclosure to shift the burden to the parties to rebut.
The economic analysis of input foreclosure illustrated in the appendix diagram in my paper with Michael Riordan—and vGUPPIu analysis—show that input price increases to rivals (i.e., partial foreclosure) typically are profitable for a merged firm with market power and lead to upward pricing pressure on consumers as a result. This is because it shifts sales to its downstream partner, exactly like the GUPPI in horizontal mergers. Elimination of double marginalization (EDM) efficiencies then are properly accounted for in the rebuttal stage, while taking account of merger-specificity, opportunity costs, coordination, and other impediments to passing-though EDM as lower downstream prices. (To be clear, substantial foreclosure incentives does not prove that EDM pass-through also will be high.)
Carlton expressed preference for “hard evidence” rather than concentration, market shares, and non-quantitative incentives analysis. But dismissing such evidence is a recipe for underenforcement. Economists may extol econometrics, but we also know that any empirical finding can be unreliable because of data limitations and imperfections, statistical significance tests focused only on false positives, and numerous econometric complications. These complexities suggest that most if not all econometric studies will be subject to criticisms that most courts cannot sensibly evaluate. Other quantitative methodologies also face limitations (e.g., vertical arithmetic applies only to total foreclosure not partial foreclosure, and only when the actual diversion ratio can be reliably compared to the calculated critical diversion ratio). Quantification with Nash Bargaining Equilibria, equilibrium simulation models, and vertical GUPPIs also will be challenging—if not baffling—for many courts. As a result, these quantitative methodologies would often be fought to a stalemate by economic experts or ignored by the court. In the absence of simple quantitative metrics and qualitative evidence from the government, courts might rely solely on the testimony of self-interested executives.
I am certainly not suggesting that these quantitative methodologies should be automatically excluded. And they surely should be analyzed rigorously by the Agency economists. So we may not really disagree that much. Carlton ultimately said that proxies and non-quantitative analysis will be useful to guide courts.
Carlton and others may want the burden placed on the Agencies with a high evidentiary standard of proof, perhaps because they are still following Judge Frank Easterbrook and placing a higher weight on false positives than false negatives. (Indeed, I once joked at an American Bar Association panel that some might want the evidentiary standard set at a level “just a little bit higher” than whatever evidence the plaintiff has presented.) But as Baker has explained, Easterbrook’s claims are outmoded if they were ever true. Andrew Gavil and I explained why allocating Easterbrook’s excessive burden on the Agencies is not going to prevent anticompetitive conduct and effects. That high burden also is inconsistent with the decision theory underpinnings of Section 7.
Concentration and Other Rebuttable Presumptions
Guideline 3 of the draft Merger Guidelines applies this 1,800 HHI anticompetitive presumption to coordinated effects. It also makes more explicit the implicit anticompetitive presumptions in the 2010 Guidelines for acquisition of a maverick or when there is a history of attempted collusion. Following Heinz, Guidelines Section 4 properly allocates to the defendants the rebuttal burden of showing that the market is not vulnerable to coordination. And, the draft Guidelines properly propose that executives’ well-rehearsed testimony about reputation, public image or corporate DNA as warding off post-merger anticompetitive behavior should be viewed very skeptically.
These anticompetitive presumptions should be rebuttable. Philadelphia National Bank set a high rebuttal standard of “clear showing.” This was relaxed in Baker Hughes to “showing.” But that relaxation should not be over-claimed. Both Baker Hughes and Heinz applied a “sliding scale” to the rebuttal burden, whereby a stronger prima facie case implies more and better rebuttal evidence. Although framed as a burden of production, this sliding scale means that “sufficient” rebuttal evidence is needed to satisfy the defendant’s burden. A “sufficiency” requirement amounts to a type of burden of proof (i.e., burden of persuasion), in fact, if not in terminology.
For example, efficiency rebuttal claims must be “verified.” To satisfy the ease of entry standard, potential entry must be shown to be timely, likely, and sufficient. Nor does the anticompetitive structural presumption explode like a “bursting bubble” if the defendant provides rebuttal evidence. Thus, the distinction between the burdens of production and persuasion is “elusive” in practice, a point that critics might overlook.
Suggestions for Several Specific Presumptions and Rebuttals
I read the draft Guidelines as embracing an economic approach to competitive effects analysis, though this could be made much clearer. While it often frames concerns in terms of higher concentration, it characterizes lessening competition in economic terms in various places, albeit not on page two when the conceptual frameworks are introduced. While the draft often does not use the term market power in describing lessening of competition, the concept of achieving, enhancing, or maintaining market (or monopoly) power is implicit. However, it would be better to make this explicit. Because some have raised questions about the role of economics, these economic concepts should be made more explicit in the Guidelines Overview section and elsewhere to avoid misunderstanding, confusion and criticism.
The Guidelines list numerous “binding legal precedents.” It would be useful for inexperienced judges to have a supplemental Commentary that reviews the economic literature supportive of more interventionist merger enforcement and provides pointed discussion of specific examples of cleared mergers (and insufficient remedies) that led to anticompetitive effects. This will make these judges more aware of enforcement limitations and open to a more skeptical view of mergers they are adjudicating and a lower burden of proof on the Agencies.
I also suggest including discussion of additional rebuttals in Section 4 of the Guidelines based on both (i) continued post-merger competition and (ii) weakness in pre-merger and likely post-merger competitive impact of a non-failing acquired firm. Both are common and potentially convincing rebuttal claims. Relevant evidence to evaluate these rebuttal claims and any skepticism over commonly proffered evidence also can be included.
Guideline 7 appropriately stresses potential harms from entrenchment. But referring to a 30% market share for a “dominant firm” presumption seems low. Thirty percent is commonly used as a trigger for market power, not dominance. The (circular sounding) definition (“the power to raise price [or] reduce quality, or obtain terms that they could not obtain but for that dominance”) sounds more like the definition of market power, not dominance. “Dominance” is commonly associated with monopoly power and typically proxied with a share of 50-60% or more. Replacing 30% with 50% also would be consistent with the current 50% foreclosure share presumption in Guideline 6.
“Frog in the pot” concerns justify the “retrospective” analysis in Guideline 9. This deals with the classic example of a firm that sequentially acquires numerous small (e.g., with 1% market shares) competitors until it reaches a total monopoly, all the while never increasing the HHI by 200 points.
But the “prospective” application of “frog in the pot” in Guideline 8, which involves predicting future high concentration and market power, is harder to justify. To illustrate, consider a market initially comprising 12 equal-sized firms. Suppose that over three years four of the firms each increase their market shares to 15% from cost-reducing internal growth, two others exit, and the remaining six firms achieve shares of 10%, thereby raising the HHI to 1,500. Suppose now that two of the 10% firms propose to merge, which would create the largest firm and increase the HHI by 200 points, thereby satisfying Guideline 8’s prima facie case. Is it reasonably likely that this merger necessarily raises problematic risks of higher prices or reduced innovation? Moreover, if the merging firms supply credible evidence that this 10-9 merger will allow them to reduce their costs and thereby expand output, this rebuttal evidence would be deemed non-cognizable (as noted above). Thus, this merger could be challenged, based solely on the fear of future mergers or unspecified adverse market developments.
Such ex ante enforcement based on these weakly predicted future harms can be imprudent. Scientific evidence suggests it is only brainless frogs that get boiled. Frogs with brains work to jump out of the pot. Agencies similarly can engage in “watchful waiting” and attack a future merger that more likely tips the market. Importantly, permitting one merger in a market does not mean that subsequent similar mergers must also be permitted. Guideline 8 thus might be deleted and evidence of trends can be folded into Guidelines 2, 3 and 7.
I also suggest deleting any non-cognizability of such variable cost-savings rebuttals. While the agencies and the law may not want to allow cost-savings to justify a merger-to-monopoly, the non-cognizability in Guidelines 6 and 8 is much broader. I recommend instead applying the Heinz “sliding scale” approach instead to create an extraordinarily high burden on the parties for mergers to monopoly.
Guideline 4: Potential Competition Mergers
As I wrote previously, the “noticeably greater” than “more-likely-than-not” evidentiary standard of proof in Meta/Within seems woefully excessive. To illustrate with an extreme example, suppose there are three (and only three) potential entrants into a monopoly market, each of which has a 20% independent probability of entering. A monopolist would be permitted to acquire all three entrants even under a “more-likely-than-not” (i.e., 50.1%) standard, even if there are no likely efficiency benefits. This is because the likelihood of entry by at least one of these three is only 49%. Yet, these three acquisitions obviously would raise unacceptable competitive risks. This example makes it clear that the legal standard instead should be “noticeably less” than “more-likely-than-not,” not “noticeably more.”
There is other good economic and legal support for the more interventionist standard and enforcement. Dominant firms have the economic incentive to preserve monopoly profits by neutralizing potential entrants by acquiring them—and by outbidding rivals if necessary. In decision theory terminology, the Section 7 “incipiency” standard places greater weight on false negatives than false positives, as reflected in Philadelphia National Bank. This analysis even suggests that a rebuttable anticompetitive presumption might be applied to such acquisitions by dominant firms. While the Guidelines cannot mandate a change in the legal standard, enforcement policy can encourage it.
Carlton urged readers to investigate whether Proctor and Gamble entered the bleach market after being prohibited from buying Clorox, suggesting that a failure to enter implied that the case was a false positive. I accepted his challenge. My investigation revealed that P&G subsequently did attempt to enter, but the entry failed because of anticompetitive “signal jamming” and implicit threats by Clorox in P&G’s test market.
P&G CEO A.G. Lafley explained Clorox’s strategy:
Do you know what Clorox did? They gave every household in Portland, Maine, a free gallon of Clorox bleach — delivered to the front door. Game, set, match to Clorox. We’d already bought all the advertising. We’d spent most of the launch money on sampling and couponing. And nobody in Portland, Maine, was going to need bleach for several months. I think they even gave consumers a $1 off coupon for the next gallon. They basically sent us a message that said, ‘Don’t ever think about entering the bleach category.’
Interestingly, P&G subsequently engaged in tit-for-tat retaliation against Clorox. As Lafley explained, “We certainly learned how to defend leading brand franchises. When Clorox tried to enter the laundry detergent business a few years later, we sent them a similarly clear and direct message — and they ultimately withdrew their entry.”
Thus, in answer to Carlton’s query, the enforcement error was the failure to apply Section 2 to anticompetitive exclusionary test market conduct, not the Section 7 case.
Guideline 13.C and the Ovation Theory
Guideline 13.C concerns “a merger that would dampen the acquired firm’s incentive or ability to compete due to the structure of the acquisition or the acquirer.” In my view, this is a useful generalization of the concerns in FTC Commissioner J. Thomas Rosch’s Ovation Concurrence. An acquisition of a product can lessen price competition if the acquiring firm has an incentive to raise prices because it lacks the reputational (or other) constraints that deterred higher prices by the seller firm. While this lessening of competition does not flow from elimination of direct competition between the two firms, it is merger-specific and covered by Section 7 because the reduction in competition does flow directly from the acquisition. This analysis might be applied to an acquisition by a private equity firm with a track record of a high price “milking” business strategy, in contrast to the low price, growth strategy of the seller. I also wonder if the concerns about a risky leveraged buyout in footnote 25 might flow from this Guideline.
Author Disclosure: Steven Salop is currently consulting with parties that have a financial stake in agency Merger Guidelines enforcement.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.