|This document is available in three formats: this web page (for browsing content), PDF (comparable to original document formatting), and Word. To view the PDF you will need Acrobat Reader, which may be downloaded from the Adobe site. For an official signed copy, please contact the Antitrust Documents Group.|
DOJ/FTC Hearings on Single Firm Conduct and Section 2 of the Sherman Act
Alan H. Silberman
Assessing Monopoly Power in the Distribution System and Aftermarkets
The economists participating in today's panel emphasize the need for caution. They do so from the perspective of scholarly analysis and econometric modeling. As an attorney faced with advising clients and the prospect of trying cases before judges and juries, I reach the same conclusion, although I do so from a different perspective. I see complaints making Sherman Act §2 allegations and lower court decisions addressing §2 issues -- especially in private antitrust litigation -- in which the labels and keywords of the statute and precedent are being used (or misused in ways that I find problematic. In particular, the discussions concerning market power, exclusion and competitive effect are not presented in an integrated, coherent structure that can be understood with confidence.
Here, by way of example, are some principal concerns:
1. Exclusionary unfairness as seen in Conwood Companies, LP v. United States Tobacco Company, 290 F.3d 768 (6th Cir. 2002), is equated with monopoly power, based on the trier of fact's perception that the evidence indicates that it has been undertaken repetitively and with intent, without regard to whether the activity actually impairs competitive forces that should regulate behavior.
2. Market shares can be performance-driven and may not indicate any future ability to prevent the operation of competitive forces that would otherwise exist. Absence of new entry in these situations may not be proof of barriers to entry (other than the barrier that arises when would-be competitors do not want to challenge good performers.) This can be seen in examining wholesale distribution functions, in which the distributor acts for multiple sellers, but in which ultimate market control -- the ability of the upstream manufacturers or suppliers to circumvent the distributor if the distributor pursues strategies that reduce output and raise price -- rests with the upstream party. Exclusive dealing in such situations should not be a matter of concern, nor should these relationships give rise to "conspiracy to monopolize" A similar issue is presented when a manufacturer is able to maintain what appears to be a controlling market share (e.g., 65-80% of a product market) over a long period of time, notwithstanding the fact that at prospective competitior could develop a product ready-for-sale within a year at modest investment, but there is no entry because the seller meets price, quality and service demands of its customers. That seller should not have to worry that some misstep will be labeled an "attempt to monopolize".
3. Relational power is often created by contract or quasi-contract years before the conduct that is alleged to reflect an exercise of market power and potential monopolization. At the point where relational power is exercised, and alleged to be exercised in a manner that disadvantages a current competitor, it appears that there is no competition governing the activity, because, for example, the buyer is "locked-in". But that is probably not the correct time to assess market power: the more significant question would seem to be "Was the transaction that gave rise to relational power subject to competitive forces at the time the decision to enter into the transaction was made?"
4. A related concern involves demanding that a seller disaggregate a product offering into elements that can be sold separately so as to preserve the appearance of current competition among others offering those pieces, notwithstanding the fact that -- at the time the product or service offering was designed and first offered -- those other firms did not compete successfully by seeking to offer competitive packages (or establishing joint ventures that provided an equivalent). Where there is a reasonable basis for finding that design of a product or service was an effort to improve a product offering, that conduct should usually be seen as competitive behavior that should be rewarded. By making those decisions subject to later challenge (generally, in the name of a "civil rights" claim -- i.e., that other sellers have a right to succeed in the sale of one or more disaggregated elements) there is a substantial risk that we are chilling innovation and telling firms that they should not compete vigorously. If such conduct is later going to be labeled "monopoliy", or an "attempt to monopolize", it should require evidence of extraordinary clarity and weight, of a type that is best suited to expert administrative adjudication and prospective-only relief.
5. In distribution systems, measurement of market share maay nvolve another complication: if the products and sales of a single brand are aggregated, the result may not be an index of market power in any sense. For example, if price and output are determined by downstream parties in the distribution system, sales volume and market share may increase relative to other competitors, but there would be little reason to treat those statistics as an index of brand market power and, in assessing the "power" of the upstream party to impose terms or restrictive conditions, retail sales data may be irrelevant to what should be the relevant question, which focuses on what competitive forces govern the transaction. This type of confusion has not be seen signficantly in the United States; however, it is embedded in EU approaches to "dominance", and needs to be avoided here.
In the private practice of law, the starting point for analysis is not economics. It is to try to understand the real world facts and how they relate to what we understand to be the objectives of §2 of the Sherman Act. That requires clear identification and articulation of the factors that are expected to create limitations on particular competitive situationsk, and to require an assessment with regard to whether or not those factors have been impaired by the conduct being scrutinized. While §2 is an important tool in the arsenal of antitrust enforcement, and there are certainly situations in which conduct merits condemnation under §2, it is important to identify and clarify the limits of §2 in various situations where its application is less clear. These occur, particularly, in conduct that involves differing forms of distribution, changes in the distribution process and innovations in marketing tactics that disadvantage businesses who (contrary to basic competition principles) were relying on the continuation of "business as usual" or failing to recognize the need for innovation that matches current business and consumer needs. The risk in all of these situations is that -- if we do not frame the inquiry properly -- innovation and competition can be portrayed as opportunistic anticompetitive conduct.
From the perspective of a private practitioner, speaking about real world facts is always fraught with danger, and trying to work with those facts in private antitrust litigation before a jury involves confusion, at best, and a high risk of error. That is particularly true when the focus of the §2 claim is on marketing behavior or tactics within the distribution process. Because the allegations accompanying these claims often describe significant losses of business visited upon the individual plaintiff or changes in traditional distribution practices, as well as predictions of higher consumer prices and "misuse" of "market power", many trial court judges are wary of summary procedures or early dismissal. The semantic maneuverings and hyperbolic assertions of counsel tax the abilities of many trial court judges, not to mention juries. And that is precisely why "caution" is required, not just in terms of economic inferences, but also in terms of the nature and quantum of proof that ought to be required before private litigation resources (of courts, counsel and clients) are expended.
If we are not able to carve out a class of §2 private claims that do not belong in the courts, or do not belong in the courts without having the benefit of extensive expert guidance from the Department of Justice or the Federal Trade Commission so as to avoid excessive costs and Type 2 error, it would be particularly useful for antitrust regulators to articulate the essential steps applicable to most §2 claims in clear, lay-oriented terms, and to articulate principles that can be applied in a "decision tree" fashion to inform discovery and disposition in private §2 cases. Guidance from the Department of Justice and the Federal Trade Commission plays a significant role in informing judges about substantive law and, indirectly, how to structure discovery and pre-trial processes so that claims can be addressed efficiently. Identification of "safe harbors" is one technique to achieve that result. But there may also be "semi-safe harbors"--areas where we should require more substantial proof to support inferences of a §2 violation where those inferences are being made by a jury, or by a generalist judge. The harbor is only "semi-safe", in the sense that such categorization would never preclude Federal Trade Commission inquiry and even adjudication with prospective remedies.
The last suggestion is, of course, not one that can be achieved immediately. But enhanced guidance to courts and counsel can be achieved with less complication and delay. A principal step in that direction is recognition that words such as "market power" or "anticompetitive" or "exclusionary" and the like are not starting points for analysis, nor are they generalizations that are easily applied to all types of business behavior. They are conclusions that apply only under specific conditions, and must rest on reason, not rhetoric.
What does "framing the inquiry properly" entail? My simplistic, non-scholarly sense of "competition" is that we expect business conduct to be subject to constraints that will, over time, result in a proper allocation of resources, because those constraints will limit willful efforts to reduce output and raise price or to exclude actual or potential competitors if the presence of those other firms would provide the constraint we are seeking. By contrast, the "civil rights" theory of antitrust leaves off the last phrase ("if the presencewould provide the constraint") and argues that we ought to protect the "excluded" -- at least where there is no proof (by some standard) that consumer welfare is not well-served by the exclusion.
Thus, before we apply tests concerning "consumer welfare effects" or "profit sacrifice, no economic sense" we ought to recognize that there may be types of exclusionary behavior, engaged in by firms with substantial sales in relationship to others (which would typically be described as "market share" and "market power"), which do not reduce the constraints on which competition depends. In such instances, no matter how willful the conduct, or how great the loss of profit to another firm, there is no reason to proceed with a full-blown §2 inquiry.
To use one example, already noted above: a wholesaler distributes 95% of a product line in a significant geographic area, representing several competing sellers. Over the years it has purchased smaller firms that had provided the same function. It turns to the various suppliers whose product it distributes and seeks an exclusive with each, knowing that this arrangement will end the existence of the remaining "competitors" and intending that result. But the fact that small competitors exit the market, that they are excluded so long as exclusivity holds, and that the larger firm now has "100%" of all sales, does not establish any reduction in the constraints which prevented the firm from pursuing a "reduce output/raise price" strategy and will continue to prevent that type of conduct. In the case described, the wholesaler keeps sales only because it serves the objectives of its suppliers; if it tries to pursues a different strategy, its suppliers will (absent a conspiracy among several suppliers which would itself be actionable) be able to discipline it, obtain substitute distribution or within a short period establish direct distribution.
The example is, admittedly an uncommon case. But it highlights the fundamental question -- how does the conduct of the firm being questioned change, or threaten to change the constraints that existed but for the conduct?
By way of illustration, here are a few examples, parelelling the categories of concern identified earlier.:
"Vigorous" competition is supposed to be the hallmark of an unrestrained market. But antitrust principles do not define the sum total of all behavioral standards in business. Cheating and stealing are still cheating and stealing, and actionable as such, even if undertaken as a competitive strategy.
If cheating, stealing, lying to customers and the like results in the potential exclusion of a competitor, is it a §2 concern? If we focus on the exclusionary effect, and on the assumption that a seller with increased dominance (by sales volume and relative share) may be able to dictate price, we are induced to say "Yes". Indeed, that appears to be the underlying premise of Conwood Companies, LP v. United States Tobacco Company, 290 F.3d 768 (6th Cir. 2002) -- or at least it would have been the likely premise had the defendant not conceded "monopoly power" at the start of the case. Certainly, Conwood involves an effort to exclude rivals on a basis other than efficiency, and Conwood cannot point to any countervailing consumer benefit from its conduct. But asserting that exclusion without efficiency by a dominant firm is the equivalent of a §2 violation is dangerous, and unnecessary, ground, unless the record supports a finding that competitive forces have truly been impaired. (As I understand Conwood, that finding would have been highly questionable inasmuch as the number of competitors increased, the number of competing brands increased, the sales volumes of competitors increased, and no competitor was forced to exit.) There are, after all, non-antitrust doctrines that can be invoked to control excessive behavior, including tortious interference with business advantage, misrepresentation, unfair acts and practices, and §5 of the Federal Trade Commission Act. By using the language of monopoly and attempted monopolization, decisions such as Conwood give false signals to courts and counsel.
The competitive constraints that were supposed to be at work in Conwood were not impaired by the bad behavior at issue. Competitors and customers were both in a position to raise concerns. Customers were able to impose limits, if they wanted to devote their resources to this product line. Once the problem was brought to light by litigation (but not necessarily antitrust litigation), corrective action was likely. There was no need to employ the arsenal of §2 theories, nor was there any need to twist §2 jurisprudence to resolve the issues.
When "market power" is assessed at the moment where an alleged injury is felt by the plaintiff, the focus of the inquiry is the ability, at that moment, to exclude actual or potential competitors. Kodak -- Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451 (1992) -- is an example. However, using the "on what constraint does competition depend" lens, we can understand Kodak and other aftermarket tying cases in a different way, and avoid the risk that theories of market power based on perceived "lock-in" factors become the basis for a range of §2 claims that are then applied in other cases.
Long-term contracts, including true franchise agreements or quasi-franchises in which the inferior party is "invested" (financially or even emotionally) in the attributes of a particular brand and therefore averse to change, create relational power. But relational power is not necessarily market power of the type that poses antitrust laws. The constraint we expect to be in effect in such matters is one that is present at the inception of the relationship. If, in fact, the contract creating relational power is the product of a process that was then subject to competition, the fact that such power is later exercised should not be the occasion for a wholly new analysis. The conduct was the subject of competition.
However, in order to assert that competition was a constraining factor at the inception of a transaction, it is appropriate to inquire as to the factual basis for the assertion. This leads us to look at the nature and amount of information disclosed. Where there has been disclosure sufficient to put the other party on notice that power may be exercised post-execution and may limit choices in what would otherwise be an "aftermarket", use of the "market power" concept to the subsequent conduct is inapt. See Queen City Pizza, Inc. v. Domino's Pizza, Inc., 922 F. Supp. 1055 (E. D. Pa. 1996, affirmed, 129 F. 3d 724 (3d Cir. 1997), cert. denied, 118 S. Ct. 1385 (1998); Wilson v. Mobil Oil Corp., 984 F. Supp. 450 (E. D. La 1997); PSI Repair Services, Inc. v. Honeywell Corp., 104 F. 3d. 811 (6th Cir.), cert. denied, 117 S. Ct. 2434 (1997). Indeed, in Queen City, the court relied on the generalized disclosure made at the time of contract and found that to be sufficient without becoming involved in an inquiry about lifecycle pricing, etc.
What is the market constraint applicable at the inception of these transactions? In the franchise situations, it is not simply the availability of other, similar, franchises (although that is often sufficient). It can be the broad range of alternatives for use of capital and entrepreneurial energy available to a proposed contracting party.
In PSI, and in many of the technology cases, the evidence reflects a decision, at the inception of a product offering, to configure a product (or product/service combination) so as to compete more effectively in a single business. The competitive warp and woof of that business provides a competitive constraint. The fact that, thereafter, firms may want to participate in a segment of the seller's product offering by demanding that an element be disaggregated ought not be seen as a "market" in which the "market share" is equated with monopoly power and the firm's original innovation undone. Should a seller that develops a "self-correcting" machine be sanctioned because it is using the "power" of its manufacturing position to destroy the business of repair services? Should General Motors (if it had the market share it once possessed) be subject to criticism because it incorporates "On-Star"® which has the ability to open my locked car when my keys are inside the vehicle, thus reducing demand for locksmiths? The answers are obvious. The answers should be no less obvious when we are dealing with complex products in the high-tech field or where intellectual property is involved. See, e.g., Digital Equipment Corp. v. Uniq Digital Technologies, Inc., 73 F.3d 756 (7th Cir. 1956): just because there is proof of demand for disaggregation, it does not mean that a seller is attempting to eercise market power in a separate market; it may be nothing more than trying to make a product more competitive.)
Of course, it may be that not all cases lend themselves to this analysis or to the conclusion noted above. Microsoft may be one that called for different treatment. However, as noted at the end of these observations, the existence of complex cases that may warrant further or different inquiry can be used to support expert administrative proceedings with prospective relief, thereby minimizing the dangers of overgeneralization that are inherent in litigation advocacy.
What if, as in Eastman Kodak v. Image Technical Services, Inc., 504 U.S. 451 (1992) , the exercise of relational power is not disclosed at the time of contract, and there is less of a basis for asserting that the contract, at its inception, was subject to competitive forces? The analysis must then move to identify whether there is a further set of constraints on which we would otherwise rely to prevent the willful exercise of market power (here, transferring power from one "market" to what -- at least de facto -- was a second market.) We expect the overall market in goods and services to create a constraint that prevents opportunistic behavior post-contract signing. (In contract terms we call that "good faith contract performance"). Where the conduct at issue is designed to fully destroy that restraint by creating a barrier, we have reason for concern -- although that concern is moderated by the fact that the alleged exercise of monopoly power to exclude an aftermarket competitor could have been addressed at the inception, and thus the monopoly or market power concern must be carefully framed so as not to chill lawful competitive innovation from others. That principle should hold true whether the innovation as a matter of initial product design or a post-sale enhancement or improvement. There is also a risk of rewarding other firms for competitive indolence. Why, for example, was Le Page's unable to team with other manufacturers and create bundles that could compete effectively with 3M?
This is an area where economic expertise and real world facts can be brought together through empirical analysis. One is tempted to ask how many Kodak copiers are sold today, what their market shares are, and how long they remain in service. In fact, technology has reduced many products to "no service" items, not because power is being used to link product and service, but because reductions in replacement cost and in the number of skilled service people have combined to the point where repair service is going the way of the buggy whip.
To continue with overly-simplistic assertions, it ought to be self-evident that the "relevant market" is that market that yields calculations that shed light on the issue at hand. If the market description meets all sorts of tests of precedent, but yields a result that is not seen as efficacious, it ought to be ignored. One may call the number "market share", but it hardly connotes "market power" and is thus not useful.
In the Block Exemption Regulation regarding vertical agreements adopted by the EC, "market power" of a franchise or selective distribution system is measured by aggregating the total sales of the brand, at retail. The resulting calculation may, or may not, provide insight into whether the firm has the ability to interdict competitive constraints. For example, local sellers may be responsible for their own decisions as to price and output. In such instance, increased sales in certain localities reflects the competitive process. Aggregating such sales to support the conclusion that the upstream firm has market power is not only illogical; it creates a false positive -- mistaking competitive success for market power. We do not make this error in the United States. But it illustrates the kind of analysis that belongs in an assessment of market power and market share. The sales of a franchise or selective distribution system are relevant, in the aggregate, only if an increase in those sales vis-à-vis others suggests a possible increase in the ability to restrain competitive forces. If it does not tend in that direction, it should be ignored.
How would we apply the principle to franchises selective distribution in local markets? Assume a distribution point "X" that draws 90% of its customers from a two-mile radius. Assume that there are 50 such outlets in the applicable television market (ADI) and that newspapers have the same reach. But in area "A", there is only one competing location open in the same line of business and zoning laws create a barrier to new entry (allowing grandfathering). "X", the local firm, acquires the property and converts it into a site for a different line of business. In the two-mile radius, X now has 100% of sales, and its action was intended to exclude the competitor. If the constraints on price and output are local, X would (interstate commerce to the side) have a problem. But price and output in this line of business are driven by brand advertising and promotional activity, of X's brand and of competing brands. Thus the consumer obtains information as to competitive price and promotions from the broader geographic area, even if he or she trades only in a narrow locality. Market power must be assessed in relation to the constraining factors on which we are relying, and the assessment must rest on whether those factors are, or are not changed, not on the loss to a competitor (or the intent of the monopolist.) In this regard, looking at the profit sacrifice test would also give a questionable result. Clearly, X will increase sales and make more money, but the increase may not offset the cost of acquiring the property used by the competitor. Still X has not been able to raise price or impair consumer welfare.
And, in the end, the discussion supports the statement at the beginning: this is an area that requires caution. In the first instance, the Federal Trade Commission and the Department of Justice can provide broad guidance, suitable to generalist judges and counsel. But it can also go further, and engage in focused advocacy to suggest how guidelines should be applied. Indeed, responsible judges (and litigators) ought to invite such advocacy.