Donald S. Clark, SecretaryOffice of the SecretaryFederal Trade CommissionRoom 159-H600 Pennsylvania Avenue, N.W.Washington, D.C. 20580
Re: Submission of Views Relating to the Draft Antitrust Guidelines for Collaborations Among Competitors
Dear Mr. Clark:
The American Antitrust Institute applauds the joint action of the Federal Trade Commission and the Antitrust Division of the U.S. Department of Justice in preparing the proposed "Antitrust Guidelines for Collaborations Among Competitors." We believe that this document will prove to be extremely useful for businesses and for the antitrust community.
The proposed Guidelines have the virtue of collecting in one place, in a relatively easy to understand form, a tremendous amount of sophisticated, state-of-the-art analysis. This document contains practical information about the decisionmaking process of the enforcers and the courts that will be especially useful to lawyers and businesspeople who are not full time antitrust specialists. Its place in antitrust history will be similar to that of the 1968 Merger Guidelines - perhaps not a perfect document, but an excellent first step and a significant improvement over not having any Guidelines whatsoever.
We do, however, have some suggestions as to how this excellent document could be improved.
First, we believe that the document should more clearly spell out the types of anticompetitive harms that could result from collaborations among competitors. Section 2.2, "Potential Anticompetitive Harms" describes these potential harms in only two paragraphs. While these paragraphs do a fine job for certain purposes, they might inadvertently have been drafted too narrowly.
These paragraphs focus upon classically anticompetitive joint activity. In other words, the Guidelines are concerned with collaborative or collusive activity that could lead to a monopoly-like outcome, in a manner that directly raises price, restricts output, or divides markets. In addition, Footnote 5 states: "These Guidelines do not address the possible exclusionary effects of agreements among competitors that may foreclose or limit competition by rivals."
There is, however, a third category of potential harm from joint corporate activity. Collaborations can also fix the rules of competition. This can be anticompetitive even if the participants never agree to raise prices, to restrict output, to divide markets, or to exclude or otherwise disadvantage rivals. Even if the final decisionmaking is independent, the result can be anticompetitive. The reason is that the change in the rules alters the noncooperative outcome in the marketplace in a manner that is analogous to the unilateral effects doctrine of merger policy.
This is because firms can agree upon rules of competition that insulate the participants to some degree from hard competition and limit the scope of competition. This insulation gives each participant increased pricing freedom, similar to the benefits from traditionally attained market power. By increasing the space between collaborators and their nearest competitors, the collaborators achieve the power to raise prices even though they never get together to set prices. Rather, they compete less vigorously or in a restricted manner in the collusively altered environment. The colluders still compete along some dimensions, but the fight has been fixed. They have rigged the rules of competition so they do not have to compete as fiercely.
The attached draft article, "The Three Types of Collusion: Fixing Prices, Rivals, and Rules," by Robert H. Lande and Howard P. Marvel, describes "rule fixing" in more detail. It also shows how the collaborations involved in such important antitrust cases as California Dental Association v. F.T.C., National Society of Professional Engineers v. United States, Detroit Auto Dealers' Association, Bates v. State Bar Association of Arizona, and F.T.C. v. Indiana Federation of Dentists, can best be analyzed in terms of "rule fixing", as opposed to some form of price fixing or exclusion. Each involved collaborations between competitors that were anticompetitive. Yet none would be specifically and adequately described by the material in Section 2.2 of the Draft Guidelines.
If Section 2.2 is interpreted broadly, "rule fixing" collaborations would be covered by the Guidelines, and certainly nothing in the Guidelines excludes consideration of this type of anticompetitive effect. Nevertheless, we urge the enforcers to remove all doubt that the kinds of harms that occurred in the above cases are indeed among those of concern to the Guidelines. We recommend that you revise the proposed Guidelines by including another paragraph under Section 2.2 which spells out how collaborations among competitors can result in "rule fixing" that is anticompetitive.
Second, the tone of the Guidelines is too accepting of efficiency claims. The AAI of course applauds the efficiencies that often flow from collaborations among competitors. Indeed, many collaborations result in pro-consumer efficiencies that could not reasonably have been predicted at the time of the venture's formation.
While many collaborations among competitors do result in significant efficiencies, however, it is usually difficult at the time of their formation to predict which ones will succeed in achieving these ends. The enforcers therefore should approach all such arguments with an air of skepticism, and should be reluctant to permit potentially anticompetitive practices in the hope that a particular collaboration will succeed in achieving lower costs or producing new products.
We therefore believe that the Guidelines should contain a strong general caveat that efficiency claims are easy to assert but difficult to demonstrate with any degree of confidence or to predict reliably. The Guidelines should state that the enforcers' skepticism about the net benefits of efficiency gains will increase whenever the potential anticompetitive effects could be especially significant, and whenever the anticompetitive effects are thought to be more likely (in other words, we recommend a sliding scale approach).
The Guidelines also should note that it is difficult to predict reliably whether similar efficiencies could have been achieved without the joint corporate activity, or through a less potentially anticompetitive form of the joint corporate activity. Section 3.36(a) does contain some skeptical language, but we believe that this language does not go far enough. Either the language in Section 3.36(a) could be strengthened, or an efficiency-skeptical caveat should be placed at the end of Section 2.1.
Third, efficiencies that improve competition and consumer welfare most are those that are in whole or in part passed to consumers in the form of lower price or greater choice. By contrast, efficiencies that entirely benefit the collaborators (because, for example, they only affect short-term, fixed costs) are less desirable from a consumer perspective. The Guidelines should give greater weight to efficiencies that are likely to result in lower prices or greater choice for consumers. Sections 2.1 and 3.36(a) imply this, but the Guidelines should make this point more explicitly and emphatically.
Fourth, the Guidelines should state that the primary purpose of the antitrust laws is to prevent certain arrangements that result in supracompetitive prices for consumers or otherwise deprive them of the choices that the free market would provide. This idea is present somewhat in Sections 2.1, 2.2, 3.37 and 4.3. The Guidelines should explicitly state, however, that although the enforcers welcome the efficiencies that collaborations sometimes bring, the enforcers are centrally concerned with preventing practices that could lead to supracompetitive prices or diminished consumer choice.
Fifth, we believe that the Antitrust Safety Zones are extraordinarily helpful. They bring as much clarity as is feasible to this field. They are especially useful for non-experts, and will help to give non-experts the same information on the potential legality or illegality of collaborations that is already possessed by full time antitrust specialists.
We originally had thought that the safety zones should be different for different types of collaborations (i.e., that the Guidelines should have a smaller safety zone for marketing joint ventures since marketing collaborations are the most suspect type). We ultimately concluded that the Guidelines' choice of a single safety zone is appropriate at this time. When these Guidelines are eventually revised, however, the next set of drafters can reassess this issue and evaluate how the safety zones have been working in practice. The next set of drafters can then decide whether certain types of collaborations deserve even larger safety zones. (The current Guidelines already take a step in this direction. They state that marketing collaborations "may be" procompetitive, that production collaborations "are often" procompetitive, while "most" R & D collaborations "are" procompetitive.)
Sixth, Section 3.2 of the Guidelines states that integration agreements that otherwise would be per se illegal will be analyzed under the rule of reason if they are "reasonably related to the integration and reasonably necessary to achieve its procompetitive benefits...." We believe that the "reasonably necessary" standard is the correct standard, and one that is consistent with the better case law on the subject. The "reasonably related to" language, however, is so weak, vague and elastic as to be virtually meaningless. It adds nothing and could even be misleading. It should be dropped.
Seventh, just because a collaboration was procompetitive or benign at the time it was formed does not mean that it will forever be procompetitive or benign. Section 2.4 says this, but this section should specifically state that high technology collaborations in particular might lend themselves to different assessments over time. The competitive status of a collaboration in a fast-changing high technology market is even more changeable and more difficult to predict. It is especially important that the legal status of high technology collaborations not be fixed forever at formation.
Eighth, collaborative activity designed to produce a new product or engage in innovation could be structured or directed in a manner that allows one or more of the collaborators to steer the venture towards their own parochial interests, rather than in the interests of the venture as a whole. In other words, the "independent" collaboration might not be truly independent. If this bias caused the collaboration to produce products that were complementary to products produced by one of the collaboration's members, for example, this could affect competition in some market other than those in which the collaboration directly operates. Parts of the Guidelines, including Section 3.34(d), already deal with this issue, but the point could be made more explicitly.
The first Merger Guidelines were revised once the enforcers gained experience using them in a variety of real-world situations, and we assume they will be further revised at some times in the future. We expect that, after the Guidelines for Collaborations Among Competitors have been in place for a few years, the enforcers will evaluate how they have been working, and then will revise them in light of this experience to make an even better document. We hope that the enforcers will at that time also expand the Guidelines to cover certain subjects that were deemed beyond the present document. We hope that the second version of the Guidelines will deal with such issues as exclusion, membership, access, and standard setting.
Despite our suggestions, we reiterate our belief that the Guidelines constitute a very high quality achievement. We are grateful to the Federal Trade Commission and to the Antitrust Division of the U.S. Department of Justice for giving us this opportunity to provide you with our comments.
Robert H. LandeDirector andSenior Research Scholar