Department of Justice/Federal Trade Commission Hearings
Single Firm Conduct: Predatory Buying Panel
FTC Headquarters Building
June 22, 2006
I am grateful to the Department of Justice and Federal Trade Commission for the invitation to participate on this predatory buying panel. It is a great honor to be here. I am especially grateful because I have been thinking for longer than I care to remember about how to support Section 2 of the Sherman Act, and yet reconcile it with less controversial, more accepted frameworks for prosecuting cartels and evaluating horizontal mergers.
I will offer a suggestion along those lines today. Although I believe that my suggestion will make deserving exclusion cases easier to bring, some aspects may be significantly different from established jurisprudence. For that reason, I particularly recognize the privilege of having a place at this distinguished table.
Before proceeding, I need to say that my statement today reflects solely my own opinions and does not represent those of the Competition Bureau or any of its staff.
For this complex topic, I offer a series of recommendations.
When first asked to participate in a panel on “predatory buying,” my response was to object to the title. We should recognize that “monopolization” entails two essentially different types of practices, one that for shorthand could be called “predation,” and the other “exclusion.” The most succinct distinction is that “predation” cases involve doing too much of a good thing to bring about a bad result later. There, the understandable concern is with deterring energetic competition—not discouraging firms from charging low prices, adding product features, and the like.
Exclusion cases, on the other hand, involve doing a bad thing now. One way or another, they come down to acquiring control and effective market power over supply or access to an input or service needed to compete, what economists call complements. The most explicit way to acquire such control would be through a series of exclusive contracts with the complement’s suppliers. It may involve overbuying inputs through explicit purchase or, as I’ll suggest below, bundling, rebates, or other forms of “leaving money on the table.” I call this practice “complement market monopolization, or CMM.
The major problem with single-firm conduct law is the failure to recognize the essential differences between these two types of conduct, leading to the counterproductive imposition of predation standards on exclusion cases. Perhaps the failure arises from a presumption that one statuteSection 2must imply one principle. Perhaps it follows from the persistent belief that Section 2 must be premised on harm to rivals. Since competition also harms rivals, Section 2 law is thus driven by fear of over-deterrence. Instead, exclusion cases should be recognized as different, where we can apply horizontal tools and not predation screens to the delineation and protection of complement markets.
I would have changed the title of this panel to “Exclusionary Buying,” because the leading cases involve creating of market power over complements. The recent DOJ/FTC certiorari petition in Weyerhauser v. Ross-Simmons illustrates an exception that proves the rule. The setting is unusual, in that the concern is not that a timber processor would acquire so much control over a relevant market in uncut trees to be able to raise their effective price. Rather, according to the petition, the allegation is that a mill would pay too much for trees to drive out other buyers, with subsequent recoupment by cutting prices paid for trees in the future.
I have little to say about which market power, price-to-cost, and recoupment tests are appropriate for preventing over-deterrence in these rare predatory buying cases. I do suggest that courts demand not only evidence appropriate for such tests. They should also demand evidence that specific assumptions behind strategic models are satisfied, i.e., that the alleged predator either has a reputation for non-profit maximizing behavior to protect, or benefits from identified asymmetric failures in capital markets. Theoretical possibility alone does not make a practice harmful.
Market power is often characterized not just as the ability to raise price but also as “the ability to exclude.” This is a mistake of imprecision. Ability to raise the price of X depends on entry barriers or other impediments to competition, but those do not depend upon the price of X. Higher X prices would, if anything, encourage entry. Rather, the ability to exclude depends upon control over the prices of Y, Z, W, or something else needed to enter and produce X.
Delineation of the relevant complement market should therefore be the first step in all exclusion cases. Taking Dentsply as an example, the case rested on the premise that the national distributors constitute what in merger contexts we would regard as a relevant market, in this case for the distribution of teeth to dental labs. The HMGs provide the useful framework for testing this premise. They ask whether teeth manufacturers would turn to other distributors, or whether there would be entry into that distribution market in response to a “small but significant non-transitory increase in price” (SSNIP) of using such dealers.
I do not know the facts of that case and thus the answers, but the HMGs ask exactly the right questions. Cases eventually turn to evidence of entry or substitution into the complement market, but they do not make such concerns centralthe best indicator being the continued identification of the relevant market as that in which the alleged monopolizer is already dominant, not that over inputs or services competitors need to compete. Control over such a complement market is not only sufficient to raise competitive concerns; it is necessary for anticompetitive exclusion.
Hence, plaintiffs should focus on identifying that complement market and showing that the practices at hand cover enough of it to raise the complement’s price. In effect, one should ask if one would be troubled if the complement providers covered by the alleged exclusionary practice merged. Unlike usual characterizations of monopolization cases, this is one we know how to answeruse the HMGs. If the answer is no, stop; if the answer is yes, go to the next step.
Barriers to entry cannot be raised, and competition impeded, by any more than the extent to which the price of the complement can be raised. Sometimes this higher price will be explicit, sometimes it will be only an inferred higher priceProf. Carlton has usefully called it a “shadow price”if the exclusionary practice so ties up the complement market that only higher priced substitutes, including self-provision, are available.
Explicit exclusive dealing contracts offer one such standard: Firms wanting to use those dealers would have to cover the cost of breaching the contract. Other alleged exclusionary practices, such as bundle discounts or loyalty rebates, may create a significant price increase—once one has established the first step.
Following the last point, one could ask whether bundles, rebates, or other programs have to increase the effective price of the complement as much as would explicit contracts. I have no reason to believe it should. Were we to follow the HMGs, as we should for complement market delineation, we might only need ask if the practice leads to a SSNIP of the complement.
This tells us that whether a bundle is anticompetitive has nothing to do with a predation-like test. It does not depend on whether the incremental price of adding a good to a bundle, or of supplying more of a product given a discount, is less than some measure of marginal or average variable cost. Rather, it depends only on the extent to which such practices create market power in order to raise the price others must pay for the services provided by retailers, distributors, or other complement providers getting the discount.1
Even for predation, some commentators have noted that some or all of these screens need not increase competition and consumer benefit. Nevertheless, they may be appropriate to prevent over-deterrence of competition through low prices or added features. However, in exclusion cases, controlling a monopoly share of complement markets is not inherently pro-competitive, and thus need not have high bars for its protection.
The profit sacrifice or “no business sense” testthe two are equivalent if one assumes that “business sense” means “maximize profits”substitutes concern with intent and tactics for concern with effects, as if whether someone had been murdered depends on the price paid for the gun. Others have noted that it creates an absolute efficiencies defense, in that a penny of gain from a practice excuses untold anticompetitive harms. As Rick Warren-Boulton has said, the test is notably inappropriate when regulated monopolists do the excluding.
Although I have criticized “raising rivals’ costs,” mostly for its emphasis on “rivals,” Prof. Salop deserves enormous credit for pointing out long ago that predatory sacrifice and recoupment is unnecessary to carry out tactics that raise those costs. My difference is that I would focus directly on the complement market.
Ironically, the test also forgets that once upon a time, profit sacrifice implied previously unobserved efficiency, not anticompetitive harm. We learned that exclusive territories, exclusive dealing, tying, and even resale price maintenance must generate efficiencies because they reduce demand, making even monopolists worse off otherwise. That realization gradually reformed most vertical restraint law. Assuming now that a profit sacrifice must be anticompetitive forgets antitrust history and invites us to repeat mistakes that have not been fully undone after nearly a century.
On equally efficient competitors, I point out what should be obvious: Inefficient competitors hold down price. Complement market monopolization leading to their exclusion can raise price and harm consumers.
Having gone after two sacred cows, I may as well finish off the herd: The Grinnell prior possession of a monopoly test also can impede meritorious exclusion cases. It distracts attention away from the complement market, focusing instead on the characteristics of who monopolized it. Prior dominance could even be a defense, but once complement market monopolization is shown, it should be up to a defendant to claim it has no consequence because of monopoly elsewhere in the production chain.
Moreover, this test is counterproductive. Proving the cost, demand, and entry barriers necessary to establish prior dominance undercuts the argument that the alleged exclusionary practices make a difference. Using Richard Posner’s phrase, the monopoly should be “fragile” at worst. An exclusion case will be strongest if the sector would be competitive, but for the practice under scrutiny.
Ask whether we would apply these standards to mergers. Should all mergers be legal unless one could show they would be unprofitable but for anticompetitive harm? Should any merger, including to monopoly, be legal if a more efficient firm buys and eliminates a less efficient competitor? Of course not. Even prior dominance may make the incremental effect of a merger less troubling. If these tests would gut merger law, and if exclusion cases are akin to acquisitions in the complement market, they do not belong on this side of Section 2.
Analogy to mergers opens the door to more creative remedies. Generally, either a practice is OK, or it is not and should be stopped. We should instead take a share-based approach. Exclusive dealing contracts, bundles, or other alleged monopolizing practices might have efficiency benefits. The problem is not the practices per se, but their scalethat they pre-empt so much of the complement market to raise its price significantly. Rather, defendants should be allowed to retain the practice, but only over a non-dominant share of the complement market35%, 50%, or some appropriate number. If the practice is actually efficient, it will be kept. If it serves only to exclude, this remedy would lead to its discontinuance.
About two years ago, I gave a talk at the FTC on these ideas, entitled “Saving Section 2.” As I began, an economist there asked, “Why should anyone want to save Section 2?” My answer may not have satisfied him, but in short, it is that it can and should be saved. Were all Section 2, single-firm conduct cases about protecting a monopolist’s rivals by drawing vague or impossible lines between competing just enough and too much, I might have agreed with the questioner. However, exclusion cases are not about maintaining monopolies but creating new ones. In focusing on complement market monopolization, such cases can and should be no more controversial than merger and collusion cases are today.
Thank you again for the privilege of allowing me to share these observations. I hope I can clarify them through responses to any questions you have here, and as they arise in the future.
* Professor, Public Policy and Economics, UMBC; Senior Fellow, Resources for the Future. Email: firstname.lastname@example.org. 2006: T. D. MacDonald Chair in Industrial Economics, Competition Bureau, Gatineau, Canada; email email@example.com. These opinions expressed here are mine and do not reflect the views of the Competition Bureau or any of its staff.
Some, but not all, of the views expressed in this paper and citations to supporting authorities are in my "Saving Section 2: Reframing Monopolization Law," available from the AEI-Brookings Joint Center for Regulatory Studies at http://www.aei-brookings.org/publications/abstract.php?pid=1002. A version is also forthcoming as a chapter in Ghosal, Vivek and Johan Stennek (eds.), The Political Economy of Antitrust (Amsterdam: North-Holland, forthcoming 2006).
1 Time does not permit a full explanation, but it turns out that the discounter’s marginal cost may be relevant, but only to the degree that courts would use marginal cost in calculating damages for breach.