People who labor in and around the law are naturally wary of commentary that begins with praise. They have learned, for example, that appellate court opinions that start with references to the "learned" or "distinguished" trial judge invariably are reversals. And a lifetime of exposure to parental guidance, lectures, sermons and other forms of address suggest that praise is often a prelude to complaint.
This is not the intention here. The "cheers" in the title are genuine and the "challenges" are just that, pleas for help rather than criticism. Not that the challenges are easy. What I am asking of antitrust economists is no less than better guidance on the limits of economic learning and suggestions on how to proceed when economic learning can no longer provide rigorous answers.
First, the cheers. People may not fully appreciate the extent to which economic analysis has become the dominant framework for resolving antitrust problems. Up to relatively recent times, there was a rigorous debate about possible alternative sources for antitrust decisions, like dispersion of political power, wealth transfer effects, and various social considerations,(2) but economic analysis of consumer welfare effects has swept the board. Many who are reluctant to accept welfare economics as the sole objective of public policy seem to recognize that it provides the most practical guidance and is least likely to lead to arbitrary results.(3)
This consensus has recently been brought home to me in a very personal way. Since late 1999, I have been privileged to work in an agency that is bipartisan by design and headed by people with varied views on general economic, social and political issues. We sometimes have differing views on the appropriate disposition of particular cases, but these differences are typically based on different views about the facts. It is noteworthy and gratifying to me that we almost invariably adopt the same framework for analysis - - and that framework is grounded in microeconomics or the economics of industrial organization, referred to here as "antitrust economics."
I will never forget the profound impact that the so-called "New Learning" had on the antitrust bar when it burst outside the academic walls in the early 1970s. Up to that time, antitrust lawyers and judges did not pay much attention to economics, and the economics they did apply tended to be wrong. Efficiencies were suspect, for example, and monopolists were not supposed to compete very hard.(4) In a remarkably short period of time,(5) the New Learning became the dominant paradigm. Antitrust practitioners learned that there was no simple correlation between the number of competitors and the intensity of competition; that efficiencies had a major role to play; that many "restraints," once deemed suspect, could be pro-competitive; that price predation is an extremely unlikely strategy; and that ultimately consumer welfare is advanced by policies that protect "competition" rather than the welfare of individual competitors.
There were two problems. The New Learning (later called "Chicago-School" economics) appeared to be so simple, powerful and elegant - - particularly, as outlined in Bob Bork's immensely popular and influential book(6) - - that some may have gotten the impression that antitrust cases were easy. They mistakenly thought that it was no longer necessary to get their hands dirty with the facts.(7)
Inevitably, there was a reaction. A lot of people were uncomfortable with the somewhat restricted role for antitrust that the Chicago-School of economics seemed to suggest. The bases for their discontent ranged from pure nostalgia for a by-gone era, through various levels of intellectual disagreement, to perhaps unacknowledged concerns that were more purely personal.(8) As mentioned, the objection based on social and political factors seems to have died out; this talk will focus on arguments more purely economic.
A number of economists have been closely identified with a critique of Chicago-School economics from within the economics profession itself. They have, for example, identified situations where vertical restraints may cause consumer harm or where predatory behavior might pay off. Terms like "raising rivals' costs" and "network effects" have gained common currency and engage the attention of the antitrust community. A new level of complexity has been introduced by so-called "post-Chicago" economic theories.
An appropriate analogy for me is the evolution of physics in the 20th century. Einstein did not demonstrate that Newton was wrong; Newton's equations will still yield the right answers in ordinary circumstances. But, there are certain extreme situations where Einstein's corrections have to be taken into account. Similarly, post-Chicago economists (or most of them) do not seem to claim that Chicago learning is wrong, but rather that it is incomplete.
Today, however, I do not want to address the more highly publicized post-Chicago arguments but would like to focus on perhaps even more profound issues involved when we attempt to apply economic learning. In order to continue the analogy with 20th century physics, I do not compare Einstein's relativity theories but rather theories of quantum uncertainty. Are there problems in antitrust economics where the solutions are not only unknown, but unknowable? And how should we address them?
Let me mention a few examples:
- The Sylvania court recognized that restraints on intrabrand competition can lead to more vigorous interbrand competition, and announced, as a general rule, that interbrand competition is the "primary concern of antitrust law."(9) But, do we really know - - are we even capable of knowing - - if and when it would be appropriate to overcome this ordinary presumption? Note also, that the balance between intrabrand and interbrand competition necessarily embodies assumptions about market definition; you cannot weigh things on a balance unless you can define what they are. But, market definition is an imprecise art. We will return to the market definition issue in a moment.
- The same issue of the balance between interbrand and intrabrand competition is involved in the analysis of joint ventures. Some ventures involve large business organizations; others are a lot more homely. The people who read this paper probably practice in law firms and economic consulting groups that restrict competitive options in various ways. Doctors do the same thing. The evolution of the Health Care Guidelines(10) demonstrates that it is not at all easy to define what is and what is not a pro-competitive restraint, even for small-scale enterprises. We still have a lot more questions than answers.(11)
- The two examples mentioned above involve what we might call a balancing of effects over space, that is, effects inside and outside the walls. But, what of disparate impacts over time - - the situation where present effects may be pro-consumer or largely benign and the long term effects may be harmful? How do we balance these things? More precisely, what is the appropriate discount for the time value of money coupled with growing uncertainties the further you project into the future? The intractability of this problem is perhaps best illustrated by the presence of so many eminent personages, of common philosophical persuasion, on both sides of the Microsoft case.
- The problem of market definition has been mentioned above. I discuss some particular aspects of the issue in some depth in a recent paper called The Significance of Variety in Antitrust Analysis,(12) and I will not dwell on it in any length here. The argument, in summary, is that economic output is moving progressively further away from the assumptions of product homogeneity on which so much antitrust analysis is predicated. We are living in an economy characterized by the sale of differentiated products, services and experiences, in various combinations. Costs, prices and outputs may no longer be the only (or even the most) significant variables, and traditional market definitions based on price elasticities may lead to conclusions that are either intolerable or simply irrelevant. The paper really does nothing but ask questions; I certainly do not have any answers.
So, what are the challenges?
For me, the challenges were best expressed in a remarkable lecture that Professor Harold Demsetz gave at George Mason University Law School about ten years ago. The lecture titled "100 Years of Antitrust: Should We Celebrate?" with commentary by four distinguished scholars, was never published other than in a booklet circulated to the attendees,(13) and never was given the notice that it deserved. The likely reason was that the commentators and other observers focused unduly on Demsetz' profoundly skeptical views about the value of antitrust enforcement,(14) and did not pay enough attention to the arguments on which he based his ultimate conclusions.
Imagine, if you will, the stunned reaction when one of the vocal proponents of what came to be known as Chicago-School antitrust - - a school of thought that was popularly supposed to believe in the supremacy of simple, straightforward theories - - opened his lecture with the shocking statement that: "We do not yet possess an antitrust-relevant understanding of competition."(15)
When addressing conflicting effects in different areas (or competition in "space," as mentioned above), he said: "It is the mix of competitive forms that is subject to our influence, not the absolute level of competition. The practical policy goal is not to increase competition but to select the "best" mix of competitive forms."(16) He went on to say, however, that we do not have an objective way to make the selection: "[p]art of our disagreement about competition policy derives from differences in the subjective values we attach to various forms of competition."(17) The same issues arise when it is necessary to balance present and future competitive effects (the balance over "time" mentioned above): "Persons with opposing views on this issue implicitly believe in different rates of transformation between competition now and competition in the future."(18)
My first challenge to antitrust economists is to help lawyers and policymakers to navigate at the boundaries of your knowledge. The issues that Professor Demsetz raised ten years ago are still profoundly important, and we need to better understand whether they are being solved or whether they are even capable of being solved. This is in part an appeal for further research, if that would help, and for public discussion of the issues in ways that are more accessible to the non-economists among us. Dumbed-down economics is of no particular use, but translators and popularizers are not to be despised.
The Federal Trade Commission recently(19) held a conference in Washington for the benefit of its own staff, to discuss this general topic. The panelists included some of the leading industrial organization economists in the United States, who offered their own suggestions for further research. Two topics in a lengthy list were of particular interest to me, because they speak directly to the uncertainties identified above.
First, many "post-Chicago" theories of exclusionary contracts are based on reasonable models that show how exclusive contracts and foreclosure can produce anticompetitive outcomes.(20) The outcomes, however, obviously depend on a number of assumptions that are built into the models, such as the existence of economies of scale or whether one firm's decision to enter an exclusive contract imposes an externality on other competitors. Therefore, it is not clear from the current economics literature how robust the conclusions of anticompetitive exclusion happen to be. The models really only show that there are reasonable possibilities and plausible situations where anticompetitive effects may arise. Although it is clear that these anticompetitive effects can be present in many settings, it is less clear whether these effects can be overwhelmed by other offsetting factors. Empirical research is very limited in this area, and a lot more is needed before a policy-maker can apply these exclusionary theories with any confidence.
Second, we are becoming increasingly aware of dynamic competition issues in merger investigations. Many parties will argue that any potential anticompetitive effects will be short-lived because the rate of innovation in the particular industry is very rapid, or alternatively they will argue that particular market definitions do not capture the competition or innovation that threatens to disrupt or has already altered the existing market structure. On the other side of the investigation, FTC staff often evaluate the effect of the transaction on R&D competition and may argue that network effects prevent the adoption of disruptive innovations. Unfortunately, the economic literature does not adequately suggest ways to assess the current state of dynamic competition, or the evidence to support or refute the arguments. At this time, economists have not yet provided us with the means to determine the likelihood of disruptive innovation or competition, although Schumpeter suggested the importance of the issue nearly sixty years ago.(21) Similarly, there is very limited evidence of the empirical importance of network effects,(22) despite the voluminous theoretical work on the issue.
A further, and purely personal, appeal is for the highest level of candor in economic presentations. I am not particularly impressed by an eminent economic advocate who draws unequivocal conclusions about matters that may inherently involve some value judgements - - particularly, if there is an equally eminent advocate who draws diametrically opposite conclusions. I am more likely to be impressed by advocates who identify the issues with clarity, state what their science allows them to state, and clearly identify the point where their expertise ends.(23)
And yet, it is possible that an economist's expertise does extend further, even in the realm of the unmeasurable and incomparable. This is the second challenge for you. Harold Demsetz, with academic detachment, could take an agnostic stance in the utility of antitrust; we do not have that luxury. Antitrust laws are not likely to be repealed anytime soon; some of us have taken an oath to enforce them, to the best of our ability; and all of us deal with them on a day-to-day basis. We have to muddle through somehow.
In my experience, even though economists have to admit that their science does not allow for a purely objective resolution of various issues, they are nonetheless able to shed some light on perhaps non-obvious economic consequences of decisions more subjectively arrived at. To illustrate with an example from an entirely different field of law: I can remember Demsetz himself telling a class of federal judges that economists are not the people who finally determine whether minimum wage legislation is desirable, but they can point out what is likely to happen if the legislation is passed. That way, at least, the decision makers have a better appreciation of the stakes.
Take a closer example. As decision makers, we worry a lot about the risks of over-enforcement and under-enforcement - - what you call Type I and Type II errors. The conventional wisdom is that under-enforcement is a less serious problem because the market place is likely to eventually discipline anticompetitive effects, whereas efficiencies lost through over-enforcement are lost forever. I am not sure that either statement is universally true. In close cases, it might be a lot more helpful to have uncertainties acknowledged and the decisional factors identified and addressed in these terms rather than to provide glib conclusions that a particular transaction is or is not anticompetitive. Some of you do it already; I would encourage more of it.
I personally find that a lot of our cases are hard - - quite frankly, much harder than I expected them to be on the basis of my experience in the private sector, when it was not my job to strike balances but rather to advocate a point of view. (The cases are hard sometimes, even when everyone agrees about the nature of the problems, because it is difficult to evaluate the adequacy of the proposed solutions.(24)) I am asking for your help in these uncharted areas because I am profoundly uncomfortable if my only fallback is the notion that "I know an anticompetitive effect when I see it."(25)
The problem will not go away because we cannot ignore potential antitrust issues just because we do not have, and may never have, the economic tools to resolve them with confidence. Antitrust economics has come a long way - - in a real sense, it has won the field. Three cheers! But, I feel compelled to say to you who practice it that the antitrust laws are primarily informed, but not bounded, by your present knowledge and capabilities.
1. Commissioner, Federal Trade Commission. This paper is based on talks given on November 15, 2000 at a conference sponsored by the economic consulting firm of Charles River Associates and July 7, 2001 at the meetings of the Western Economic Association, and more recently updated. I would like to acknowledge the assistance of attorney advisor, Thomas J. Klotz. The ideas expressed are my own and not necessarily shared by any other Commissioner.
2. The alternative arguments are summarized in Lawrence A. Sullivan and Warren S. Grimes, The Law of Antitrust: An Integrated Handbook 9-19 (2000).
3. I have personally argued that producer and consumer "freedom" is the real ultimate value, but that economics provides the tools for mediating conflicting claims of freedom. Thomas B. Leary, Freedom as the Core Value of Antitrust in the New Millennium, 68 Antitrust L.J. 545 (2000).
4. E.g. Crown Zellerbach Corp. v. FTC, 296 F.2d 800, 825 (9th Cir. 1961), cert. denied, 370 U.S. 937 (1962); United States v. Aluminum Co. of America, 148 F.2d 416 (2d Cir. 1945).
5. The new learning first attracted significant attention at the famous Airlie House Conference in 1974. See Industrial Concentration: The New Learning (Harvey J. Goldschmid, et al. eds. 1974). Only three years later, it had a significant impact on the Sylvania opinion. Continental T.V., Inc. v. GTE Sylvania Inc., 433 U.S. 36 (1977).
6. Robert H. Bork, The Antitrust Paradox: A Policy at War with Itself (1978).
7. Matsushita Elec. Indus. Co. v. Zenith Radio Corp., 475 U.S. 574 (1986), was cited to support this view. This reading of Matsushita may not be correct but, for present purposes, the important point is that some endorsed it. The opinion in Eastman Kodak Co. v. Image Technical Servs., 504 U.S. 451 (1992), whatever its merits, can be viewed either as a retreat from this reading or as clarification of a misunderstanding.
8. It would not surprise an audience at an economics conference that Chicago theories were generally more appealing to corporate counsel who pay bills than they were to defense counsel who mail them out.
9. GTE Sylvania Inc., 433 U.S. at 52 n.19.
10. Department of Justice and Federal Trade Commission Statements of Antitrust Enforcement Policy in Health Care, Trade Reg. Rep. (CCH) ¶ 13,153 (1996).
11. See Antitrust Guidelines for Collaborations Among Competitors, Issued by Federal Trade Commission and the U.S. Department of Justice, Trade Reg. Rep. (CCH) ¶ 13,161 (2000).
12. Thomas B. Leary, The Significance of Variety in Antitrust Analysis, 68 Antitrust L.J. 1007 (2001).
13. Harold Demsetz, 100 Years of Antitrust: Should We Celebrate?, Brent T. Upson Memorial Lecture, George Mason University School of Law, Law and Economics Center (1991) (including comments by Robert Pitofsky, Richard Schmalensee, Robert H. Bork and Ernest S. Gellhorn).
14. He concluded: "Because of these opposing consequences of antitrust, I see little cause to rejoice greatly or to be remorseful over the 100-year history of the Sherman Act." Id. at 12.
15. Id. at 1 (emphasis in original).
16. Id. at 4 (emphasis in original).
17. Id. at 5.
18. Id. at 8.
19. The conference was held on September 11, 2001, and the fact that it continued throughout that terrible day is a tribute both to the panelists and to their audience.
20. See, e.g., Douglas B. Bernheim & Michael D. Whinston, Exclusive Dealing, 106 J. Pol. Econ. 64 (1998); Daniel P. O'Brien & Greg Shaffer, Vertical Control with Bilateral Contracts, 23 Rand J. Econ. 299 (1992); Oliver Hart & Jean Tirole, Vertical Integration & Market Foreclosure, Brookings Papers Econ. Activity 205 (1990); Phillipe Aghion & Patrick Bolton, Contracts as a Barrier to Entry, 77 Am. Econ. Rev. 388 (1987).
21. Joseph Alois Schumpeter, Capitalism, Socialism and Democracy (1942).
22. See, e.g., S.J. Liebowitz & Stephen E. Margolis, Path Dependence, Lock-In, and History, 11 J.L. Econ. & Organization 205 (1995); S.J. Liebowitz & Stephen E. Margolis, Network Externality: An Uncommon Tragedy, 8 J. Econ. Perspectives 133 (1994).
23. I had a comparable issue in the days when I was an employee of a major corporation. Business people, with whom I was in day-to-day contact, would sometimes ask for an opinion on purely business issues. Before answering, I believed it was very important to make clear that I was not drawing on any "expertise" as a lawyer.
24. See Statement of Commissioner Sheila F. Anthony, General Mills, Inc./Diageo plc/Pillsbury Co., FTC No. 001-0213 (Oct. 23, 2001), available at /os/2001/10/gmstmtant.htm Statement of Commissioner Mozelle W. Thompson,General Mills, Inc./Diageo plc/Pillsbury Co., FTC No. 001-0213 (Oct. 23, 2001), available at /os/2001/10/gmstmtthomp.htm Statement of Commissioners Orson Swindle and Thomas B. Leary, General Mills, Inc./Diageo plc/Pillsbury Co., FTC No. 001-0213 (Oct. 23, 2001), available at /os/2001/10/gmstmtswinleary.htm.
25. Cf. Jacobellis v. Ohio, 378 U.S. 184, 197 (1964)(Stewart, J., concurring) (describing obscenity).