STRUCTURAL REMEDIES IN SECTION 2 CASES
RICHARD A. EPSTEIN *
A CAUTIONARY NOTE ON ANTITRUST I very much appreciate the opportunity to present my views to this Joint Committee of the Department of Justice and the Federal Trade Commission on the use of structural remedies in Section 2 cases. To set the stage for the analysis, the Section 2 cases that I shall examine are those which are concerned with unilateral practices that are intended to, or have the effect of, creating monopolies within given industries. The common practices that are usually addressed under this heading include exclusive dealing and tie-in arrangements, and predation claims. The various types of remedies that may be imposed in these cases include damages, including treble damages, the invalidation of particular contractual provisions, and structural \changes in the dominant corporation, including its break up into smaller units that may compete with each other, at least in certain markets.
In dealing with this issue, I put aside the question of whether there should be any remedies in these monopolization cases at all. In general, I am very skeptical about the success of these cases, because they raise issues of efficiency that are usually far more difficult than those associated with section 1 cases that deal with cartels and the division of markets. In those cases the restraint in output and the increase in price is usually associated with a loss of overall social (consumer + producer) welfare which makes some form of relief appropriate. In addition, most Section 1 cases involve secret conduct, which makes it appropriate to think of treble damages as an offset for the difficulty in detecting the violation. The situation is in fact more difficult than this account might suggest, because there are often powerful efficiency justifications for certain kinds of horizontal arrangements, which means that a per se rule of illegality must in practice be tempered by a series of exceptions for those practices with demonstrated efficiency properties. Bank clearing services among competitors are one example. In addition, Section 1 cases often give rise to serious difficulties at the level of proof. There are in my view cases where gossamer evidence has been allowed to take matters to the jury, and there are other cases where powerful proofs of collusive behavior have been overlooked by judges. But for both this issues, on balance, it seems as though modest correctives should be sufficient to put the entire field on a firmer footing.
Section 2 practices are harder on the issue of liability because there is no clear theory as to why or how unilateral practices allow a single firm to increase the profits that it could obtain from its, often supposed, dominant position simply by raising the prices over its key goods and services over which it enjoys a monopoly position. In addition, many Section 2 cases create an odd form of discrimination such that certain practices are allowed to smaller players in the marketplace but denied to the dominant firm. Hence, even if the practice in question does have some potential to capture monopoly power, the effort to quash that practice exacts a higher toll in efficiency than is normally found in Section 1 cases. It is I think wise, therefore, to tread softly on Section 2 matters when it comes to the design of remedies.
FIVE CONSENT DECREES General Considerations This basic presumption is I think borne out by the particular government actions in the various consent decrees that I discuss in my recent book, Antitrust Consent Decrees in Theory and Practice. The general theme of that book is that the remedy in question should be narrowly tailored to the violation. It should not be use to enjoined practices that are not in themselves illegal. In addition, the use of various injunctive remedies should be limited to relatively short time periods--usually five years or under--lest they impede the flexibility of the regulated firm which has to labor under a set of restrictions that are not imposed on any of its competitors. And third, the break up remedy (which may make some sense in some Section 1 cases) should be used only in extraordinary circumstances, given the adverse consequences that can follow.
Meat Packers (22-29) The first case I review is the famous Meat Packers decree that dates from the early 1920s. That decree did not a Section 2 case as such, but it did involve the imposition of extensive line of business restrictions that did prevent the governed firms from entering into various wholesale and retail markets over which they enjoyed no monopoly power. The broad decree was supported by various chain stores that stood to benefit from suppressing new avenues of competition, and the consent decree itself continued to operate on acquisitions years later that bore no relationship to the underlying antitrust violations. The major legal proposition from this decree stands for the proposition that bound parties cannot easily plead changed circumstances to get out of restrictions to which it consented in the settlement process. But that does not answer the question of why it makes sense for antitrust officials to impose those stringent conditions in the first place, when their chief consequence is to hobble a downstream competitor long after market circumstances have changed.
ASCAP-BMI (30-39) The second study in the book examines the extensive litigation over the ASCAP and BMI decrees, which have been the source of constant litigation since 1941. The basic charge in these cases illustrates the efficiency/restriction trade-off noted above. The use of Performing Rights Organizations has matchless efficiency in allowing contracts to form between hundreds of composers and millions of customers. But the combination of parties on the producer side could allow for individual composers to gain economic rents by bundling their goods in a single package, which requires the same overall fee for any given time period, regardless of the amount of music played in that period. The effort to alter the compensation systems to avoid that foreclosure effect led to endless variations on a theme without coming up with an ideal solution. In addition, the separate settlements between ASCAP and BMI created some competitive imbalance between the two. It is very much an open question whether the cost of policing this monopoly was greater than the social gains it produced.
United Shoe Machinery (40-53) The third of these decrees, which deal with the United Shoe Machinery Company again was an epic struggle that started with the 1899 merger and ended only with the dissolution of the company after break-up was ordered in 1968. The merger itself had obvious efficiency advantages in bringing a huge number of upstream and downstream patents under common control so as to avoid the endless holdouts at every stage of the productive process. The original decrees all found specific exclusive clauses in these agreements to be against the antitrust laws, but once they were removed, the legal change had little impact on United Shoe's overall market share, for most customers were willing to pay a somewhat higher price in exchange for the convenience of dealing with a single vendor. There are several lessons to learn from this unfortunate application of the antitrust laws. First, the efficiencies in these vertical arrangements often matter a great deal. Second, the actual harm caused is often less than is supposed, so that the wise company declares unilateral surrender by removing many contractual restrictions at the first sign of trouble, lest courts and juries think they have greater significance than they do. And third, the break-up in this case doomed the firm, which could have come as no surprise to anyone who had followed the entire proceedings from the outset.
Bell System (54-73) The fourth of these decrees was in a sense surely the most important because it involved the break-up of the original Bell System in 1982 by the aggressive antitrust intervention from Judge Harold Greene. One illusion that dominated the entire bumpy course of litigation under the decree, as well as that under the 1996 Telecommunications Act that supplanted the decree, was that antitrust action could make this complex network industry behave as if it were perfectly competitive. The chief vice of the decree was that Judge Greene was confident that he knew the ideal structure of the industry--with the Regional Bell Operating Systems taking monopoly positions in the local exchange market, while competitive long line carriers facilitated calls between the various RBOCs. But the administrative costs of running this system proved astronomical, and did the distortions between carriers that did, and did not, fall under the consent decree. The more modest alternative of having the FCC order interconnections had been rejected in Judge Greene's initial decision, even though it would have spared many of these structural transformations. Yet 25 years later, the industry works extensively through vertically integrated firms that bear no relationship to the structure that Judge Greene envisioned. No one could have expected him to understand the changes in technology that rendered his decree obsolete. But, even if he could not predict the direction of these developments, he should have been aware of the pace at which these developments would take place.
Microsoft (74-111) The fifth of these decrees is that which involves Microsoft itself, which is the only decree of the set that avoided, if only barely, extensive structural breakdown in favor of a more modulated approach that sought to insure interconnections that created a measure of competitive balance between Microsoft and other competitors in the various application markets. One reason, I believe, that the Microsoft case avoided the calamitous results of the United Shoe and the Bell System break-up was that the various decrees were reviewed by the Circuit Court of Appeals for the District of Columbia and not the Supreme Court, which in both the aforementioned cases issued one-sentence affirmances of the initial major District Court opinions. The Microsoft decree has been attacked on the ground that it did not facilitate new entry into the operating system market, but that criticism is misplaced, given that the dominant operating system is likely, as under the essential facilities doctrine, to have strong efficiency justifications. The costs of duplication are likely to be very high, and there is no apparent advantage of having software writers have to make two different versions for each program they produce. So long as there is entry onto the basic operating system, the interconnection remedy should be preferred. Indeed, huge portions of the gratuitous dislocations in the Bell System case came from the subsequent statutory system to require the piecemeal sale of unbundled network elements, which raised pricing, valuation, and cross-subsidy problems that were never full resolved under the Act.
The overall moral is clear. In dealing with dominant practices under Section 2, caution in the fashioning of remedies should be the key watchword. There is no systematic reason to add aggressive remedies on top of what are often highly difficult antitrust cases even on the matter of liability.
* James Parker Hall Distinguished Service Professor of Law, The University of Chicago; the Peter and Kirsten Bedford Senior Fellow, The Hoover Institution. This remarks are prepared for the Joint Committee of the Department of Justice and the Federal Trade Commission, held on March 28, 2007 in Washington D.C. The paper is based on large measure on my recent book Antitrust Consent Decrees In Theory and Practice: Why Less is More (AEI 2007), which was written with the financial support of Microsoft Corporation. The above remarks represent only my opinion, not those of Microsoft. The page references in the text are to this book.