DKT/CASE NO.:P951201
TITLE: HEARINGS ON GLOBAL AND INNOVATION-BASED COMPETITION
PLACE: Washington, D.C.
DATE: October 25, 1995
PAGES: 902 through 1078
NOTE: This document has not been edited for transcription errors.
Donald S. Clark
Secretary of the Commission
Meeting Before the Commission
Date: October 25, 1995
Docket No.:P951201
FEDERAL TRADE COMMISSION
I N D E X
WITNESS: EXAMINATION
Presentation by: Page:
Sumanth Addanki 938
National Economic Research Associates, Inc.
Professor Dennis Carlton 925
University of Chicago
Professor Richard J. Gilbert 906, 982
University of California, Berkeley
Richard T. Rapp 916
National Economic Research Association, Inc.
Professor Dennis Yao 950, 1013
University of Pennsylvania
Michael Sohn 990
Arnold & Porter
Judy Whalley 1002
Howrey & Simon
FEDERAL TRADE COMMISSION
In the Matter of: )
)
HEARINGS ON GLOBAL AND ) Docket No.: P951201
INNOVATION-BASED COMPETITION )
Wednesday,
October 25, 1995
Federal Trade Commission
Sixth and Pennsylvania
Room 432
Washington, D.C.
The above-entitled matter came on for hearing,
pursuant to notice, at 9:30 a.m.
PARTICIPANTS:
FEDERAL TRADE COMMISSION:
ROBERT PITOFSKY
Chairman
ROSCOE B. STAREK, III
Commissioner
JANET D. STEIGER
Commissioner
CHRISTINE A. VARNEY
Commissioner
SUSAN S. DE SANTI
Director, Policy Planning
JOHN B. BAKER
Director, Bureau of Economics
WILLIAM E. COHEN
Project Director, Innovation Policy Planning
DEBRA A. VALENTINE
Deputy Director, Policy Planning
MELISSA HEYDENREICH
Staff Policy Planning
SPEAKERS:
SUMANTH ADDANKI
National Economic Research Associates, Inc.
PROFESSOR DENNIS CARLTON
University of Chicago
PROFESSOR RICHARD J. GILBERT
University of California, Berkeley
RICHARD T. RAPP
National Economic Research Associates, Inc.
PROFESSOR DENNIS YAO
University of Pennsylvania
MICHAEL SOHN
Arnold & Porter
JUDY WHALLEY
Howrey & Simon
P R O C E E D I N G S
COMMISSIONER VARNEY: Good morning, welcome back to day seven of our hearings. Today we are going to continue our discussion of antitrust enforcement and the role of potential competitiveness analysis or the concept of innovation markets in enforcement actions. It is a particular pleasure for me this morning to welcome back Dr. Richard Gilbert. He is a professor of economics and business administration at the University of California at Berkeley, again well represented at these hearings. We have to have a Berkeley school of thought.
PROFESSOR GILBERT: Absolutely. If we could agree, yes.
COMMISSIONER VARNEY: He is also a principal in the Law and Economics Consulting Group. From 1993 until May of '95 he was the Deputy Assistant Attorney General for Economics in the Antitrust Division, where I first met him and where he played a major role in making the Antitrust Guidelines for Licensing of Intellectual Property, and that was indeed a great experience. Before serving at the Department of Justice, Professor Gilbert was the director of the University of California Energy Institute and the associate editor of the Journal of Economics, the Journal of Economic Theory and the Review of Industrial Organization. He has written extensively on topics including antitrust economics, intellectual property and research and development, and your article with your colleague Steve Sunshine was the topic of great debate here yesterday, so we will start with you.
PROFESSOR GILBERT: Sure. Good morning. Good morning, Mr. Chairman, good morning commissioners, it's really a pleasure to be here to participate in this very important session. The question posed to our panel this morning is should antitrust enforcers rely on competition analysis for the concept of innovation markets and my answer is yes. It's not an either/or choice. Each approach has its place. It depends on the particular factual institutional circumstances of the case.
Innovation markets can be useful because innovation markets can identify competitive effects in situations where competitive effects cannot be adequately addressed using conventional product market analysis. I propose to describe today two situations that illustrate how innovation markets can be useful. I will not limit it to two situations, but those are the ones that I can discuss today, there may be others.
One situation is where a merger acquisition or joint venture has substantial spillover effects in markets where the parties to the transaction are not actual or potential competitors. A second situation is where the transaction creates a new product making conventional product market analysis or potential competition analysis either difficult or in some cases impossible.
As an example of the first situation, I will use the often-cited General Motors/ZF case, and since we have participants in the panel who know about that case, I will, if necessary, talk about a hypothetical General Motors/ZF case since I think what is important here is to show that there are situations where innovation market analysis is appropriate, whether or not they were appropriate to a particular case.
In the General Motors/ZF case, the situation that I will describe is one where General Motors and ZF were competing in Europe in the production of heavy duty automatic transmissions for trucks and buses. They were not competing in the United States except with the exception of isolated product markets involving only a small amount of customers. The merger presented typical product market difficulties in Europe; however, the antitrust enforcers in the U.S. could not be certain that the European antitrust agencies would, in fact, act the way the U.S. agencies would act and prevent an anti-competitive merger based on a product market analysis in Europe.
The main effect of this transaction on U.S. customers were the likelihood that the merger would slow the development of new products which would be available to U.S. consumers. In markets where General Motors sold transmissions, also in markets where General Motors competed with ZF, but mostly in markets where General Motors was a monopolist and ZF was not either an actual competitor or even a likely potential competitor.
Nonetheless, if a new product were developed, a new generation of transmissions were developed arising from innovation competition which was driven by the party's product market competition in Europe, that new product would be available to U.S. consumers. And if for some reason that new product was not made available, then U.S. customers would suffer as well.
As I said, the General Motors investigation presented the dilemma. The dilemma was that we did not have jurisdiction to attack the merger based on product market analysis in Europe. There was also an issue as to whether or not if we attacked the merger based on the isolated product markets in the United States, that could be fixed by a small divestiture in those product markets or it's possible the courts would have viewed that the U.S. product market effects were de minimis and thereby not be sufficient to sustain the challenge.
The bottom line, really, was that the innovation market analysis allowed the Department of Justice to focus on what were the true competitive impacts of the case. It seemed to us inappropriate to focus on minor product market effects when, in fact, the real concern about the case for U.S. customers was the likelihood that the transaction would have slowed the development of new products available to U.S. customers.
Now, of course, if we had, in fact, challenged this merger, the agencies would have to establish that there was an effect on innovation, on research and development expenditures and that that effect would translate into a reduction of innovation and that would hurt U.S. customers. I certainly believe that the agency would have to bear that burden and that was something we were prepared -- we meaning the Department of Justice at the time -- were prepared to bear that in court. Again, I don't know if the facts of the situation would have actually supported it or not. I think that's something for the courts to decide, but principally the concerns were valid concerns.
Now, another example that I will mention of the second situation where a transaction may effect the development of new products, I can use as an example of this example four in the joint Department of Justice/Federal Trade Commission intellectual property guidelines. That is an example of a research joint venture where the joint venture is organized to introduce a new product, in this case it was a new biodegradable plastic for disposable containers.
Now, in the transaction, of course, many -- most, if not almost all joint ventures in research and development are formed with procompetitive objectives and have procompetitive effects, but one can imagine a situation in which a research and development joint venture has an intent which is anti-competitive. It is possible to imagine a situation, for example, where the parties of the transaction are concerned that they may if they -- if the technology is proven, be required to introduce the new technology which would cause their existing assets to become obsolete and be written off, and this would be disruptive and possibly very expensive and reduce industry profits. Perhaps because they have to comply with environmental regulations. An example of this was the automobile manufacturers case in which this type of intent was alleged.
The difficulty of challenging a transaction like this using the potential competitive framework is clear because this is a situation in which the product itself doesn't exist yet. So, if you're talking about potential competition, you have to say it's potential competition to a hypothetical or inchoate product market. More over, the effect of the combination of the joint venture may be to delay the development of this new product so that the issue isn't whether or not the product gets developed or the state of competition in the new product once its developed, but rather will it get developed in 1997 or 1998 or 1999 so that one was now looking at competitive effects over some time period.
Again, I don't in any way view a research and development joint venture that is intended to slow innovation as at all typical. I think this is a very unusual situation, but I also feel that the tools of the antitrust have to be ready to be applied to unusual situations. I think we are fortunate in our economy that antitrust abuses are not common. And however when we do find an antitrust abuse, we have to have the tools available to analyze it properly.
As I mentioned at the beginning, of course not all types of competitive circumstances warrant an analysis based on an innovation market, many times potential competition theory would be perfectly adequate. For example, I'll use I think it's example 11 in the antitrust guidelines, there the example deals with a firm that has a dominant market position in a particular product. There is one likely competitor of that product. The dominant firm that has that product offers the potential competitor a license to use its product, the license may include a condition that says if you take this license, you are required not to compete in the development of any similar product.
I think that's an example of a competitive circumstance that can be analyzed quite adequately using the potential competition framework. And I don't think an innovation market is particularly necessary although it could be used. I don't think it's necessary, because the product exists and this other firm is a likely potential competitor of it.
Now, there's been a lot of discussion about whether economic theory and whether economic evidence can support an innovation market claim. And I would like to speak to that a little bit.
In our paper, Steve Sunshine and I briefly summarized the literature on the theory and the empirical evidence related to research and development and market structure in competition. And we acknowledge that the link between market structure and investment and research and development and the link between research and development and the pace of innovation is very complex. We also acknowledge that the theoretical evidence supporting the link is quite weak.
However, I do not believe it is fair to say that the evidence supporting a link between market structure, competition and innovation is nonexistent. I think that is an exaggeration, I don't think it's a fair statement.
In theory, there are a number of theoretical models. If we had a blackboard, I would be happy to put one up on the board that demonstrates under certain assumptions that the combination of firms is likely to reduce the incentives to invest in research and development. And the argument is very simple, the argument is that a monopolist, a firm with a lot of very concentrated market has an existing product or process and has to worry that the development of a new product or process will make its existing product obsolete. And that's often not a happy occurrence. Sometimes the monopolist benefits from it, but benefits only by the difference, the extent to which the new product or process adds value to the existing product or process.
Now, contrast that with a new competitor into the marketplace, the new competitor gets the whole benefit of the new product because it didn't have an old product. So, it has a greater incentive in that sense.
Now, I understand the argument that there are economies of scale often in the generation of research and development. I also understand the argument that has been made that often the incentive to do research and development is limited by the ability to approximate the benefits of that research and development. John Baker has written on that issue. A monopolist because a monopolist has the whole market approximate all of the benefits or almost all the benefits, at least the benefits that are specific to that industry.
There are those arguments, but there are also other arguments showing that there is a disincentive for a monopolist to invest in research and development. And as in any theoretical exercise, what it comes down to is do the facts of a particular situation fit one theoretical description or does it fit another theoretical description, which one is the right one. And I just think it's really not correct to say that there is no theoretical evidence.
On the empirical side, I acknowledge that the empirical side is quite weak. Industry-specific effects tend to dominate in any statistical analysis, but there are many anecdotes. We have Michael Porter's study of what contributes to innovation across different nations and the observation that innovation tends to thrive in competitive market structures. There's some interesting work by Nelson and Burgess on what happens when there's innovations that are freely licensed and there are multiple factors in the industry.
I think we all have our own anecdotal situations to look at and review and see that there is a sense that innovation thrives in competitive market circumstances. And on a more basic level, I'll also quote I believe it's Sir John Hicks, many people say it's Adam Smith, but quote that "a monopolist's greatest reward is a quiet life." Innovation is hard work, and you really have to work in an innovative -- in an environment where there's competition, you have to work to stay ahead, to continue to innovate and beat your customers. A monopolist doesn't have to do that.
I acknowledge it is difficult, I acknowledge it is complex, I acknowledge the empirical support is very, very weak. But I also stress that the agencies are extremely good at identifying particular factual circumstances related to a particular case. And I think the case specific investigation can deal with innovation market effects.
I caution that innovation market analysis should not be used in an indiscriminate fashion. I do think that the facts that we have weak empirical support and mixed theoretical models means that you have to be very, very careful and very deliberate, but I think as in all tools, if it's a tool that's appropriate for the job, even if that job only comes along on a very infrequent basis and very infrequently, then you use the right tool for the right job. And innovation is so important that we need to use the right tools.
COMMISSIONER VARNEY: Thank you, Richard. I think we'll do what we did yesterday, if that's acceptable to everybody, we'll go around and hear all the views before we start questions and dialogue. I would like to introduce Richard Rapp, who is the president of the National Economic Research Associates. He is a highly regarded writer and lecturer on a variety of antitrust issues, particularly relating to research and development, innovation, technology industries and health care. Mr. Rapp is also an extremely experienced expert witness. Mr. Rapp is a member of the American Economic Association, the Licensing Executive Society and on the Board of Editors of Antitrust Report. He is a member of the ABA Task Force on Market Power and Intellectual Property and a member of the New York Bar Association Committee on the U.S. in the Global Economy. It is a pleasure to have you here this morning.
MR. RAPP: Thank you, very much. I am delighted to appear before the commission this morning and present my views about the innovation market approach, particularly as it applies to mergers.
I want to begin my testimony by declaring my respect for the fair and even- handed presentation that Professor Gilbert and his co-author, Stephen Sunshine, have afforded us in their paper on incorporating dynamic efficiency concerns in merger analysis, which is the focus of my critique of the innovation market approach. And the presentation that Professor Gilbert just gave is I think a perfect example of the scholarly and constructive spirit in which their contribution has been made. Nobody could doubt that.
In a paper that I wrote that will be published in the Antitrust Law Journal later this year, I have raised a number of objections, nevertheless, to the innovation market approach. And my attempt to boil these objections down to their essence for the purpose of these hearings leaves me resorting to the language of scientific testing. And adapting that jargon to this particular circumstance, we might say that's a false positive, which is my main concern, that a false positive is a finding by the Federal Trade Commission or the Department of Justice in a merger inquiry that a merger will substantially lessen competition in a relevant innovation market when, in fact, the merger does not do that.
My main objection to the innovation market approach is that both the probability and the social cost of false positives are so high as to outweigh the real benefits that Professor Gilbert has pointed out in his remarks. The risk of false positives is high because we can't measure innovative output or innovation output the way that we measure output in product markets. If we could, an innovation market test based upon the restriction of innovation output might work, presumably would work as well as the merger guidelines approach works on output restricting implications of mergers in product markets. But because measuring innovation output is not possible, the innovation market approach takes recourse in examining market power issues by reference to research and development activities and restrictions in research and development activity. And in my view, this is the source of the problem.
A decrease in the number of firms engaged in related or overlapping R&D projects does not reliably signal whether total R&D activity or innovative output in the market will either increase or decrease as a result. Moreover, when R&D output is cut back, there is still no principled way of telling whether either the reasons for cutting back or the effects of cutting back are pro-competitive or anti-competitive. If you add those infirmities together, my point of view is that it simply makes the innovation market approach too dangerous to use.
The design of the innovation market approach to resemble the market power analysis in product markets where horizontal merger guidelines are used is based on a false analogy in my view. In product markets price increases and output restrictions are unequivocally bad for consumers. There's no debate about that.
By contrast, in innovation markets, the optimal amount of research and development is simply not known, and therefore when we observe a cutback in R&D, or the potential for a cutback arising in a potential merger or acquisition, we can't say whether competition or economic welfare is aided by discouraging that cutback, by predicting it in some sense.
We can investigate a little more deeply these assertions of mine by looking first at the need to police R&D activity in merger -- in the merger review process and then to look at the risk of doing so. In conventional merger analysis, the guidelines address the -- a merger's potential for increasing first the likelihood of anti-competitive coordination and second single firm market power. The first of these I think we can, if not dispense with, just mention briefly, and perhaps if it -- if the issue arises again, look at it in closer detail.
I think that there is a general agreement that the potential for collusive restriction in R&D is much less than the potential for output-restricting agreements in product markets. Briefly, this is because of the difficulty of policing R&D restricting agreements or coordination because R&D takes place in secret and also because the gains in cheating on some kind of R&D agreement are enduring in the way that the gains from cheating on a price fixing may not be.
I must say that I have the sense in reading the Gilbert and Sunshine paper and also the outline of testimony that Professor Gilbert supplied for a hearing this afternoon that -- as well as your remarks today, Professor Gilbert, that I think that it's fair to say that the main focus of the innovation market approach as defined by the Gilbert and Sunshine paper focuses on single firm market power and indeed merger to monopoly. I'm not certain that that reading is correct, but to the extent that it is, we can pass over consideration of what I'll call the collusion side of the story.
Where single firm market power is the issue, merger to monopoly or something close to it, then in my point of view, the doctrine of potential competition would seem largely adequate to the attack of dealing with it. And since my colleague, Dr. Addanki, is going to focus on that matter, that's another subject that I will skip over.
The reservation, however, is this: One of the cases that Professor Gilbert's testimony just described is the situation where potential competition analysis may falter. And that is the circumstance where there are no actual product markets for the goods or for the research that is the subject of inquiry in an innovation market analysis. Here in those settings, the factual analysis is all important and what the enforcement agency practicing the innovation market approach must do is it must predict whether or not the future product market will be less competitive as a result of the merger when future goods emerge.
In my view, this is more than just a tall order. Judgments -- about the only generalization I feel that I can make about it runs something like this: Judgments about the future course of innovation or the competitiveness of future goods markets are more likely to be speculative the further away from actual goods markets you are. So that we are left with a situation where if the development process of goods is far enough along, then it should be possible to anticipate their outcomes in conventional product market analysis by use of means like the potential competition analysis or supply side substitution or what have you. The further back you go from that condition to the point earlier in the research process where there are no goods to talk about, the more speculative and therefore dangerous the enterprise.
The more innovation intensive a market is, the less there is to worry about in terms of the monopolization or increasing concentration or restrictions in competition along the lines that the innovation market approach attempts to interdict. In markets where innovation is frequent, we can assume that the life span of these interferences is likely to be shortened.
In the 1950s when the cellophane case was being litigated, polyethylene and other plastics were already making inroads into the market for food wrapping, and a paper by Ray Hartman and others on the subject of leap frogging innovation in medical scanning devices is another good example of the effect that I had in mind.
All of those lead me to conclude that the need for the innovation market approach is not great. I don't want to be overly categorical about that. I wish to acknowledge that there are circumstances that are not completely covered by the innovation market -- by the -- sorry, by the potential competition approach, but again, my point of view is that the dangers of applying the approach and adopting it as a matter of policy outweigh the gains in analysis that arise in those few circumstances.
Following up on that point, the main risk, the risk of false positives, arises from the fact that R&D cutbacks are not a sign of reduced competition. In my paper, I use the pharmaceutical industry as an example of that and we know I think this story well. The advent of managed care in institutions like pharmacy benefit managers and mail order pharmacies has put downward pressure on the prices of pharmaceutical products. And if we assume for present purposes that that's so and that that will continue, then it follows as the day does the night that there will be research projects that under the old pricing regime would have been profitable that now will not be.
Pharmaceutical cutbacks in pharmaceutical research in that setting are inevitable, although the strongest firms are doing their best, as I understand it, to maintain and even increase their pharmaceutical budgets. The question is when that arises and merger or acquisition is part of the story, should those cutbacks in research and development arising from competitive clauses be interdicted by enforcement policy. My answer is no.
The implications of these observations is that R&D cutbacks can be pro-competitive in both their causes and their -- and their effects. In fact, economic theory teaches, I think, that we should expect periodic overinvestment or overcommitment to R&D by firms and the inevitable corollary that R&D cutbacks must occasionally happen. When you have goods -- factories for goods, those factories remain useful as long as the demand for the goods persists. When you have goods factories devoted to individual ideas, once those ideas are discovered and made public, the idea factories or at least most of them become redundant in some sense.
A last point, turning to practical policy issues, is the fact that merger enforcement in recent years seems to reside mainly with the Federal Trade Commission and the Department of Justice and less with the courts. It has become, and I think this has been well observed, a regulatory process. One of the chief dangers, therefore, in the innovation market approach is the ease with which agencies can impose divestiture and compulsory licensing and the willingness of merging parties to -- to divest or submit to licensing in response.
The reason for that assumed willingness is that in a big merger, any R&D project that happens to overlap can be presumed to -- to loom small in the sense that the effects or the benefits of R&D are out in the future and they are uncertain. And the implication of that perhaps overstatement or overgeneralization of the situation is that when these conditions arise, the merging parties are likely to be willing to sacrifice their interests in a research project for the sake of getting the merger through. To a degree that is not paralleled in the analysis and enforcement of mergers in product markets.
I conclude that since it is extremely difficult to distinguish good cutbacks from bad ones, and perhaps I should state that more strongly and say impossible in most circumstances, I conclude that the innovation approach to market power analysis of R&D competition even if it is performed studiously on a case-by-case basis and limited to the extreme cases is unpredictable and prone to error.
Thank you.
COMMISSIONER VARNEY: Thank you very much. We are going to turn to Dennis Carlton now. Dr. Carlton is a professor of business economics at the Graduate School of Business at the University of Chicago. Prior to that he was a professor of economics at the University of Chicago Law School and before that he taught Chicago's economics department. Dr. Carlton is also executive vice president of Lexecon. He is the author of numerous articles and wrote along with Jeffrey Perloff the well-known textbook Modern Industrial Organization. Dr. Carlton is the co-editor of the Journal of Law & Economics and associate editor of the International Journal of Industrial Organization. Pleasure to have you here, Dr. Carlton.
PROFESSOR CARLTON: Thank you. Thank you, it's a pleasure to be here.
I will -- since I have prepared written testimony, I won't go through it all, just highlight what I think are the most important points.
I would say by and large I'm in close agreement with the comments of Dr. Rapp, and not in agreement with the analysis of Rich Gilbert, though I think his analysis is insightful and by being so clear in his presentation, I can be clear in my criticisms of what I disagree with him about.
It's obvious that technology changes have resulted in dramatic improvements in our standard of living and therefore it does seem appropriate that we should investigate whether antitrust policies should pay more attention than it has in merger analysis to the effect of a merger on R&D. And the recent suggestion is that the antitrust policy should use the concept of an innovation or R&D market to examine the effect of a merger.
I'm skeptical of the benefits of following such a suggestion. As a matter of logic, antitrust policy could be used to prevent mergers that would harm consumers by concentrating in innovation market. However, in practice, the ability of the antitrust authorities to reliably identify such instances is likely to be very low.
This low reliability is in stark contrast to the greater reliability of saying using the merger guidelines to identify and prevent anti-competitive mergers that lead to higher prices for existing products. A movement towards relying on innovation markets to prevent mergers could easily lead to a vast decline in the reliability of enforcement in improving welfare. And I think the reason is simple.
Current policy has focused mostly on the competitive harms that a merger would cause in the near future. A policy relying on potential competition in the far future in yet unknown products requires the analyst to predict the far future. And the far future is much harder to predict than the near future. Any active antitrust policy that foregoes certain efficiency gains in the near future to achieve speculative competitive gains in the far future is likely to harm and not help consumers.
What I would like to do today is just very briefly outline the chain of logic you need in order to use innovation markets to identify anti-competitive mergers. I would like to show that each link in the chain of logic is weak both theoretically and empirically.
Now, once one understands the weak links, it is, of course, possible to devise a narrow policy aimed only at those special cases where the logic applies. But I think that makes sense only if those special cases cannot be addressed already with existing doctrines and if the new policy were to be narrowly applied only to those special cases. And I'm not convinced of either possibility. And I hasten to add it's not that I don't think that people at the enforcement agencies don't work hard, don't do a good job and don't look at very fact-intensive investigations carefully. It's just that I think this problem is too hard and that the use of innovation markets will lead us down the wrong path.
There's virtually no theoretical dispute that a reduction in competition, all else equal, lead to higher prices and decreased output, and that's bad. There's several empirical studies of individual industries that show that the number of competitors matter, although there's disagreement at what level of concentration prices might start to rise. But it's fair to say there's a general theoretical and empirical support for an antitrust policy aims at mergers that concentrate an existing market. Current antitrust policy focuses on whether an anti-competitive harm will occur in the near future by and large.
If price, for example, is going to rise, and for the first two years after a merger, you're likely to stop the merger. Arguments that significant deficiencies are going to occur in year three and beyond, my hunch would be -- would fall on deaf ears. And I think there's good reason for that. It's pretty hard to predict the future. Future benefits from halting a merger would have to be discounted not only for the future, but for the likelihood that they would occur.
The anti-competitive harming contrast is immediate and highly predictable and I think the current policy has made antitrust enforcement much more reliable as a mechanism to increase social welfare. Now, it's a small step in logic to extend this antitrust policy to deal with a merger of two firms that do not currently compete but would compete in the absence of the merger. That's the potential competition doctrine. As a theoretical matter, the issues are identical to those I just stated when competing firms are merging. The only practical difference, of course, is that now you have to be predicting the future and since all mergers, even of noncompeting firms, often have efficiencies. If you stop a merger with a potential competition doctrine, you are trading off benefits today for speculative benefits in the future. And that can be hard to predict those benefits.
Now, you might think that the innovation market doctrine is just a small step removed from the potential competition doctrine. Yet it is no small step in logic removed, it's actually a giant step it seems to me. There are at least three claims that are required for the innovation market doctrine to make sense. First, reducing R&D expenditures is bad; second, if there are fewer firms performing the R&D, there will be less R&D and fewer new products; third, and this is what I find the most troublesome. There are not enough other firms to perform R&D to develop future products to compete with the future products of the merged firm. There's neither theoretical nor empirical support for the general validity of any of these three claims. Let me go through each of them briefly.
First, reducing R&D expenditures is bad. Well, that statement just is generally not true. We all know that R&D is an input, it's not an output. R&D can be made more efficient, you can -- by a merger, you could get rid of duplicative R&D, you could make R&D more productive by allowing a large group of scientists to communicate with each other. So, the simple point is that you have to allow for the fact that less R&D could actually be desirable and actually have no effect at all on the output of new products or, in fact, would actually increase the output of new products.
Second, fewer competing firms will lead to less R&D and fewer new products. There is absolutely no theoretical or empirical consensus that reduced competition leads to less R&D and fewer new products. When imitation is possible by existing firms, a more concentrated market allows the innovator to capture more of the value of his innovation. So, in this way, market concentration solves the appropriability problem. Indeed, patents are specifically designed to create market power in order to provide the incentive to innovate. It strikes me there's a tension, that a tension exists between an antitrust policy that's premised on the notion that market power is bad for R&D and an intellectual property policy or patent policy that's premised on the notion that market power is good for R&D.
Now, the various economic theories predict that competition can have an nor mouse effect on R&D activity, and I agree with what -- with what Rich Gilbert said, we do have economic theories that show that it can have enormous effect on our -- that competition can have an enormous effect on R&D. The trouble is depending on your assumptions, you can get any results you like.
For Schumpeter and the line of research in his following, big market concentration can aid innovative activity. For Arrow and for that line of research, markets concentration is bad. However, it's easy to change the assumptions in Arrow's model and reverse his results.
As Professor Gilbert was explaining, if the monopolist isn't worried about someone coming along, the monopolist may have a slower incentive to innovate. You start making that monopolist a little nervous about whose breathing down his neck and he now has a greater incentive to innovate with the competitive firm. It's not that you can model where competition affects R&D, it's just that verifying the assumptions of these various theoretical models strikes me as very hard.
In the more developed literature on patent races or type to innovate, you get the exact same theoretical ambiguity. It seems like a wonderful opportunity for empiric, empirical analysis to resolve the theoretical ambiguity, but the empirical analysis does no such thing. In fact, I'll just read you one quote, this is from Baldwin and Scott, "There is no unambiguous evidence of an important, generally valid, relationship between concentration and innovative activity."
In summary, neither theory nor empirical work provides any general justification for an antitrust policy aimed at preserving competition in innovation markets. Moreover, I want to add that even if you were concerned about controlling competition in R&D markets, you should, I think, recognize that collusion in R&D markets is not likely to be as much of a concern as it is in traditional product markets. Because of the nature of R&D, it's often secret, there are often large payoffs to the R&D. These are all factors that are likely to make collusion quite difficult. That suggests that if you do pursue such a policy, the levels of concentration you use should be different and you should tolerate higher levels of concentration probably than you traditionally do.
Let me turn to the third point, because I think this is the most troubling for me. The third requirement for a large underpinning for the innovation market doctrine. There are not enough other firms to produce R&D in the future. It strikes me that identifying future competitors for unknown and unknowable products is extremely difficult. The longer the time period between the R&D and when new products are coming out, the less reliable is the prediction. Moreover, in industries that are dynamically changing rapidly, your ability to predict who are the firms is going to be very low.
I don't have time now to go through all the examples, I go through several in my paper, where I show how difficult it is to predict not just the firms, but even the industries from which R&D will come that will effect various product markets.
If you do define an innovation market, you have to include logically the innovation activity of all those firms with R&D efforts that might be producing products that are going to be competitive in the future. Now, it's not obvious to me how you construct such a market measure. How do you weight the R&D of different firms, do you weight it equally, do you handicap some of the R&D depending upon who's ahead in the race, who's closer to fruition. Do you weigh it differently depending upon whether the likely output is going to be produced with a supply curve that's elastic or inelastic and in low supply.
The fact of the matter is that we don't know how to weight to get a good estimate of shares, and I suspect market shares in any innovation market that you define are going to be extremely crude, even cruder than they usually are.
Now, I think it's rare, though not inconceivable, that an analyst will be able to identify with a fair degree of confidence the firms who are likely to be pursuing R&D that will lead to competing projects several years in the future. Perhaps in industries where government approval is necessary, certain types of drug testing, or certain types of government funding are required, like defense contracting, it may be possible to identify who the firms are who are going to be competing in the future in new products.
In those rare cases -- well, even in those cases, I want to add, the longer the time period between the R&D and the new product coming out and the more dynamically changing the technology, there still are prediction problems. But even if I can put those aside and focus on those few rare cases where you may be able to predict which firms will be completing in the future, it would seem to me that the potential competition doctrine could be used to prevent an anti-competitive merger.
I understand that this might mean that you will have to apply the potential competition doctrine to new products that are not yet in existence but whose existence can be reliably predicted. The focus of the analysis in that case, though, would be the restriction of output of those new products, not a decline in R&D activity. It's really still focusing on an output market, and if that requires an extension of the potential competition doctrine, that should certainly be one that I would be very comfortable with since it seems to follow logically. I would much prefer that to using an innovation market doctrine because of the difficulty of that doctrine in its reliability.
Let me just turn briefly to application of the doctrine. Professor Gilbert made mention of the GM/ZF case, and I do want to comment on that, and I should also comment that I served as a consultant for GM and ZF on that case. The Justice Department prevented the merger between ZF -- the acquisition by ZF of the Allison Transmission Division of General Motors.
First, let me notice -- note that this product was truck transmissions, it is -- it was not I think an industry that anyone would claim is a new R&D type industry. So that if it's a notion that new industries are developing that's requiring use of innovation markets, that's certainly not exemplified by this particular case, I don't think.
But second, this was a case -- the more important point is this, this was a case where the Justice Department alleged that there were three markets that would be adversely affected, one a product market in the United States of truck -- of transmissions for transit buses; two, a product market in the United States of automatic transmissions for refuse trucks; and three, a worldwide market for innovations in automatic transmissions. Specifically, the Justice Department was concerned that ZF would not continue to engage in R&D in as vigorous a fashion after the merger as before.
Now, assume with me for a moment, because I don't want to get into the specifics of the case, that it would have been possible to structure a settlement, perhaps through an independent licensee, that would have completely allayed all fears of anti-competitive harm in the product market. And further suppose that consumers would have benefitted as a result of this settlement. Well, the transaction was stopped two years ago. In the intervening time, I understand that no new products have emerged from ZF in automatic transmissions, and moreover ZF has withdrawn from the market in refuse trucks.
I think it's important to follow this case and other such cases to see when exactly these benefits from R&D that the Justice Department thought likely will emerge. And when they emerge, they should be discounted back and we should see whether those benefits exceed the benefits that consumers have been deprived of and that were achievable by a well-structured settlement.
Well, let me just briefly conclude, antitrust policy to prevent mergers that reduce competition in existing product markets is based upon a well-accepted theoretical and empirical research. There's no such widespread theoretical or empirical support for an antitrust policy aimed at preventing innovation markets from becoming concentrated.
Although there's a clear chain of logic by which a reduction in R&D competition in innovation markets could harm consumers, it's a chain of logic that is not one for which there is any general theoretical or empirical support. I see the following practical problems with applying an antitrust policy towards mergers involving innovation markets:
One, the inability to determine whether a decline in R&D expenditures is undesirable; two, the inability to predict the total R&D and the resulting number of new products would decline as a result of the merger; three, the inability to identify the other firms engaged now or likely to be engaged in the future in R&D that will lead to products that would compete with the products of the merging firms.
Even if you can get over these three hurdles, I caution you on the following: A benefit today is more valuable than one tomorrow. Benefits in the future are more likely to be uncertain compared to immediate efficiency savings.
Two, in dynamically changing industries, the products from R&D are going to be hard to predict, so it's going to be especially hard to figure out who are the firms that are in the market.
Three, collusion in R&D is not likely to be as much of a problem as it is in traditional product markets.
Four, R&D competition or the R&D competition doctrine as based on the use of innovation markets is a more speculative doctrine than the potential competition doctrine because it requires more difficult and less reliable predictions.
So, I'm skeptical that a general antitrust policy aimed at preserving R&D competition in innovation markets will improve society's welfare. Application of existing doctrine, especially that of potential competition, can likely deal with mergers that harm society by reducing competition in R&D. If antitrust agencies do use the policy of preserving competition in innovation market to prevent mergers in certain industries, I urge that they follow those effected industries to see whether the predicted gains from increased R&D competition ever materialize, and if they do, whether it was worth the wait. Thank you.
COMMISSIONER VARNEY: Thank you very much. I think what we'll do is maybe go one more and then we'll take a break and come back for our last speaker before we start our round table.
I would like to introduce Dr. Sumanth Addanki, he is a vice president of National Economic Research Associates. During the past nine years he has specialized in the application of microeconomics and econometrics to litigation. Before joining NERA he worked at the National Bureau of Economic Research on the measurement of industrial productivity and on the role of research and development in mergers. He was an instructor in economics and a teaching fellow at Harvard where he received his Ph.D. in economics in 1986. Good morning.
MR. ADDANKI: Thank you. It's a pleasure to be here.
Do we need new antitrust tools to analyze high tech deals? Do we need an antitrust paradigm? Probably the most extreme form of a high tech merger is what you might call a merger of ideas. A merger between two firms with no products, no sales, busy research and development facilities, but not even any technology ready to be bottled and sold.
Can such a merger even raise antitrust problems? The answer is in principal yes. It could in principal slow the pace of innovation, but to show that a merger like this is going to slow the pace of innovation is a very different category from ordinary merger analysis. And in practice as it turns out, a merger like this is unlikely to result in slower innovation unless the merging parties are potential competitors in some product market.
In other words, if you are going to want to stop a merger because it's likely to result in slower innovation, the chances are you are going to want to challenge that merger anyway on more traditional antitrust grounds such as that it's going to interfere with actual or potential competition.
I think an example is going to help. Imagine with me that there are two firms that want to merge that are working to commercialize so-called micro motors which are very small electric motors, so small that a half a dozen would fit on your thumbnail. They are initially suspected to be used in medical implants, but their future uses are wide open. Both of the firms are well funded startups and they're exploring essentially the same technologies based on ceramic materials and microscopic manufacturing techniques. Neither firm has a commercially viable product so far.
A merger between these two firms is largely immune to the usual market definition-market shares-market power kind of formula. That's because there are no products, no sales, and no market price that may or may not be effected by the merger. In fact, the only principal basis on which an antitrust enforcement agency might challenge this merger is that it's going to result in reduced innovation or slower innovation. And as I said, to do that, they're going to have to follow a somewhat different course from a routine merger analysis because showing that something is going to reduce innovation is different from showing that it's going to reduce -- it's going to reduce output in a product market, which is what conventional merger analysis is largely about.
To begin with, as we've heard from almost everyone today, collusion is unlikely to be a problem in the pure R&D situation. I won't repeat what you've already heard, incentives, monitoring and the means to enforce a collusive agreement on R&D are likely to be simply lacking in the pure merger of ideas case. So, logically we should only be concerned about this merger if it's going to unilaterally reduce the pace of innovation. The merged firm is going to unilaterally reduce the pace of innovation.
Let's simply go to the extreme case of that. Let's assume this is a merger taking us from two to one. In other words, the only two firms in the U.S. known to be doing research on micro motors want to merge. If the agencies want to block this merger on the ground that it's going to reduce innovation, slow innovation, what do they have to show? They have to show two things, first that the merged firm will have the ability and second that the merged firmly have the incentives to reduce the pace of innovation.
Let's start with ability first. The agencies will have to show that the merged firm holds unique and specialized assets which are needed for the development of micro motors and that no one else could readily either have those assets or could readily acquire and copy those assets. And that's because no one else should be able to pick up the slack if the merged firm does, in fact, try to reduce the pace of innovation.
The second thing they have to show is that the merged firm has the incentive to reduce the pace of innovation. And what that means essentially is that you have to show that for each firm the activity of the other firm was the primary spur on innovative activity for this firm, and hence, the removal of one as an independent participant in this business and in the business of innovation is going to take away the primary incentive for the merged firm to be sustaining its innovative program.
So, you have to show that there are going to be unique assets which can't be copied, and second that the products of -- that the merged firms -- that the merged firm will be trying to develop, which is micro motors, are not going to compete with other products that might be in the innovation pipeline at other firms, simply because the potential of that kind of innovation is a pretty potent spur to your own innovative efforts. Showing these two things is hard.
Let's consider the assets first. Let's note first that when innovation is what's at issue, we have to look worldwide for people who could supply it. The supply of innovation is borderless in that sense. The question then boils down to this, is there no firm in the world that has the assets that are required to develop micro motors. For instance, Mabuchi is the world's leading producer of small although not micro motors. It makes four million motors a day. Would Mabuchi have the interest and expertise to develop micro motors?
Buehler and Faulhaber are the leading European manufacturers of high precision small motors, their products are very highly regarded, very well engineered. Might they have the interest and expertise to develop micro motors. The problem is that the very nature of R&D is such that firms are going to be secretive about their R&D efforts.
You are never going to be able to know for a fact who is or isn't engaged in a particular line of research and development or who might or might not be able to acquire the assets needed to do it. And finally, who might or might not be able to copy or duplicate or in some way acquire these assets.
Let me put aside the question of unique assets for a moment and go on to the question of incentives. The -- as I said earlier, what has to be established is that no products in the innovation pipeline at other firms are going to compete with micro motors. Well, the micro motors are expected to be used, as I said, initially in medical implants and that's because their ability to provide controlled mechanical motion in a very small package makes them ideally suited for this purpose. But is there no other nascent technology that could provide the same advantage and which could supplant or replace micro motors if they are too late in getting to market?
For instance, so-called Shaped Memory Alloys or nickel-titanium alloys are specially treated to have a very unique property: They expand and contract quite perceptibly when an electric current is passed through them. They have yet to be refined and have not been used yet in medical implants, but certainly it's not a huge leap to imagine that they would be, particularly if micro motors are delayed in getting to market.
This really points up a more fundamental problem. It's notoriously difficult to try to second guess the process of innovation. Funny things happen on the way to the patent office. Promising lands of research wind up leading nowhere and path-breaking innovations can come from the unlikeliest sources. You can't tell before the fact where the next major innovation in a field is going to come from and we've heard some mention of that here today.
The micro motor firms want micro motors to be the next path breaking innovation for intermechanical implants but the shaped alloy implants want to be that, too, and they very well may get there first. In circumstances like this, I find it hard to see why a firm's R&D department is going to be allowed to sit on its hands simply because it's acquired its nearest competitor in R&D. As Andy Grove said, the CEO of Intel, one of the most successful semiconductor companies in the world, "Only the paranoid survive."
Now, I acknowledge that the situation could be different if one of the parties had substantial sales, particularly if one of the parties had substantial sales of a product that was going to be supplanted or replaced by the micro motors. Well, everything I've said continues to apply. At least now that that party has the opportunity to trade off the possibility of being late to market with the micro motors with the prospect of increased profits on the existing product line, and Professor Gilbert referred to that issue. But when you have no sales as is postulated with the firms we're dealing with here, there seems no reason why the firms would delay, would risking being late to market and potentially losing the value opportunity.
Now, as we've heard, the enforcement agencies have been willing to challenge mergers on the grounds that they might result in reduced rates of innovation, and the vehicle primarily has been lessening of competition in the innovation market. I'm not going to debate the merits of the innovation market concept, but suffice to say that Professor Gilbert and Sunshine's article on how to apply the innovation market approach in practice really largely mirrors all of the points that I have made.
In other words, if you want to evaluate whether a merger is going to reduce the rate of innovation, whether you call it an innovation market approach or not, you're going to have to address two questions, will the merged firms have the ability to retard innovation, will the merged firm have the incentive to retard innovation. The ability depends on being able to control unique assets that are specialized assets that are needed to do the innovation, and which no one else can replicate or copy or acquire in some way. I think it's very unlikely in practice, but it's not inconceivable that that could happen.
The question of incentives I think is the more interesting one. I think ordinarily you wouldn't expect firms, at least firms in the pure R&D race, to have any incentive to reduce the pace of innovation. Uncertainty about the very process of R&D intrinsic to the process itself as well as very imperfect information about what other people are doing, your other potential competitors are doing, is going to provide a valuable adequate spur in most instances.
I think the interesting exception arises when you do have a situation where one of the firms has substantial amounts of sales, existing sales, and there you have to entertain the possibility that the merged firm might be willing to scale back the pace of innovation in order to more fully exploit an existing product.
But I think that case really does fit quite squarely into or ought to fit quite squarely into potential competition analysis because the firm without the existing product should be viewed there as the potential entrant into the market that is currently plated with the products that the firm has the products are selling, and this firm could be a potential entrant either in its own right or by licensing some third party. And I think that's the nexus between reduced innovation and potential competition. If the products being developed have the potential to compete with cannibalized sales from products that either of the merging parties is currently profitably selling, there I think you can see some incentives to reduce the pace of innovation, but otherwise I don't see it.
There is one other form of potential competition which we've heard of which I think we should touch on briefly, I think professor Carlton touched on it, but which doesn't fit as neatly into the current doctrine of potential competition, at least legal doctrine, and that is the situation that comes up when you have two firms selling -- that are about to begin selling a product for which there are no current, no good current or imminent substitutes, and I think the most realistic example probably is in the drug or pharmaceutical situation where you have very tight institutional constraints on entry.
So, if you have two firms that are at the final testing stage just before FDA approval for a powerful new treatments in a brand new therapeutic class, and moreover, if you don't have any other candidates close to final testing, then a merger between these two firms is going to threaten potential competition, but not in the sense of the legal doctrine.
The competition is potential here in that it's future competition in a potential or incipient market. But I should add, again, that this is a bit of a red herring as far as innovation is concerned, it's really much more of a conventional antitrust problem of higher prices and reduced output. It's just that the market hasn't happened yet in which those effects are going to be felt. I think for those specific instances where you can say with some certainty what the incipient market is going to be and what its participants are going to be, it might be worthwhile expanding the existing legal doctrine to encompass those situations.
But for all other mergers where the parties don't currently compete in any product market, I think the primary concern is going to be one of innovation. And there I think the inquiry boils down to two questions, the first being a threshold one, do the firms have unique assets specialized assets that are needed for the innovation in question and are those assets such that cannot be -- that they cannot be duplicated or applied by anyone else, and if the answer is no, the inquiry ends. If the answer is yes, I think we move on to the more interesting question, does the merged firm have any incentive to slow the pace of innovation.
And I think that question can be an answered by asking does the merger threaten potential competition, particularly are the products being developed products that are going to compete with any products that either of the firms is selling in a substantial way right now. So, whether you call it an innovation market analysis or not, I think the questions of potential competition are going to have to be pivotal to an analysis of a high tech deal.
Thank you.
COMMISSIONER VARNEY: Thank you very much. That was a very clear explanation. Why don't we take about an eight-minute break so that we reconvene at 11:00 on that clock and we will finish up with Professor Yao. Thank you.
(A brief recess was taken.)
COMMISSIONER VARNEY: If we could reconvene and welcome former commissioner Dennis Yao. Fortunately for me he decided he wanted to go back to teaching, so I had to be a commissioner for a while. It's a real pleasure to see Dennis back here, he was a commissioner from 1991 to 1994. Dr. Yao is an associate professor of Public Policy and Management in the Wharton School, University of Pennsylvania. He also has an appointment in the strategy group of Wharton's Department of Management and developed Wharton's MBA course in competitive strategies and industry structure. In addition, Dr. Yao is a principal of the Law & Economics Consulting Group. He has published numerous papers concerning economics and policy in the areas of antitrust, defense contracting, innovation and intellectual property. He is also on the board of the Strategic Management Journal and Antitrust Counsel and he is chair of the Legal Association of the ABA's -- Legal Education Committee of the ABA Antitrust Section. Thank you and welcome back.
PROFESSOR YAO: Thank you, Commissioner Varney, Chairman Pitofsky and the rest. It's a pleasure to be here. I just shared with you some of my views about the use of innovation markets. Since I'm the last speaker, a lot of the things that I had intended to say have been said and so I will gloss over some of these points.
One of the things that I would like to do, however, is to offer a few potentially useful tools to deal with innovation markets, if, in fact, it's the case that the commission and the Justice Department continue to pursue this analysis.
Okay, I would first like to start out with some general observations concerning innovation markets future and current product markets and I agree with the previous speakers about the problems that are associated as you get further and further from the product market. As the distance between R&D and the marketable product increased, obviously the uncertainty and the speculativeness associated with making assessments of the facts also increased. And it would be really -- it would be great if one could link these -- the R&D to the product market in all cases.
I don't think one can and I don't think one can in some cases that are important. And so I think that pursuing innovation markets makes sense, though it has to be done extremely carefully, and certainly one has to be very aware of the learning that goes on as one goes from case to case because of the lack of consensus in the economic literature about how to link innovation and I guess welfare effects.
It had been mentioned by some of the others that the pharmaceutical industry was a very good example of a case in which future -- I guess you could call future product market might be a good way of thinking about innovation problems, and I think that this is probably the best case of the types of mergers that should be looked at when the product is still far from the market and largely because of the regulatory approval process and testing requirements that are associated with it.
One of the questions in thinking about whether or not it makes sense to pursue innovation markets is really whether there are any other candidates of approaches to deal with the kinds of problems that could be dealt with at the early stage. It would be nice, of course, if you could fix problems downstream. Let's say that there may be an innovation market or an R&D competition issue and that would show up later in a product market. If one could then take care of the problem, that would be fine, but, of course, that won't work for two reasons.
The first reason is that dealing with this problem in -- at a later time in the product market would not allow you to deal with the question of whether innovation or the amount of innovation had been changed. And the second is it's very difficult once the merger has been consummated to go back and do anything about it.
And so, it may be the case that these -- looking at the R&D market will be the only chance that the agencies have to deal with this problem. I think it's important to consider. I will mention later that I think it may be worthwhile for the agency to consider whether or not there is some stepped process by which if they decide they wanted to take an action against a merging R&D market but are somewhat unsure as to the effects in the product market, maybe there's a way to have a sort of delayed action for a couple of years pending what might occur in terms of the R&D.
So, I don't think that one has to just limit one's self to you have to do it now and all the remedies have to be taken now. It may be possible to consider setting up a situation in which you allow something to happen, perhaps, and then on the condition that maybe in a couple of years you review it and decide whether or not at that point a license might need to be enforced on the merging parties. This is an idea that has come up before in efficiencies with a number of people, many around this table.
Let me go to a discussion, a short discussion of usefulness of potential competition theories for antitrust merger analysis involving innovation. I agree that potential competition theory can be used to address many of the R&D issues that will come up in the merger. And that's fine, I think the problem, as Rich Gilbert had mentioned, is that there will sometimes be cases in which there will be no identifiable current or future product market and therefore potential competition is just not going to work.
Now, in thinking about potential competition theory and using it as an enforcement tool, one thing that we haven't done very much with is to consider whether the current state of potential competition theory lends itself to the kind of enforcement actions you want to take with respect to R&D. It's easy enough to say that potential competition in some sort of vacuum can work, and the general way in thinking about whether one of the parties to the merger might be a potential competitor and therefore blocking or causing the merger to be changed in some way on that basis makes some sense, but there is an existing way in which one deals with potential competition in the law.
It is not necessarily a way that was developed, in fact I'm -- I don't think it was developed with R&D markets in mind. As I understand it, many of these -- many of the requirements, the elements for proving liability under these theories are fairly required for a fair amount of evidence. That may be difficult to get.
Now, if that's true, and then we add to it that in the R&D markets it's pretty hard to get firm evidence all the time. One may find that the use of potential competition theory will be very difficult and it may be that it won't be useful not necessarily because the idea is a bad idea, or a way to approach it, but maybe because the way the law is structured will cause you to jump through so many hoops and collusions that it just won't work.
Now, having said that, maybe the law is exactly right with respect to where potential competition should be, okay, but I only bring that up because sometimes you develop a way of attacking problems based on a set of cases and if these cases don't have much to do with R&D, then they in some sense haven't been molded to include that class. And therefore they could potentially be defective procedures for that class. And that's just something to consider. It's something that hadn't come up, so I thought I would mention it.
Okay, I believe that innovation market analysis is a useful supplement to the analysis based upon current and future markets, but I agree with the statements I guess by everyone thus far this morning that the existing theoretical and empirical literature in economics is largely inconclusive about the relationship between concentration and R&D intensity, and whether reducing the amount of R&D is welfare reducing.
Now, most of these studies that people have been talking about have tended to be let's say the empirical studies have been cross-industry studies. So, for the most part, we're talking about we can't find a general relationship. Professor Gilbert mentioned that there are specific theories, theoretical models that would apply in particular circumstances or that could apply to particular industries in particular circumstances. And I think that the lack of a general finding while it should -- what it means is that you should be very, very careful about trying to apply any general rule. I think it does not mean that one cannot find in any particular circumstance with particular facts problems that one can feel pretty sure exist and that the remedies that are available to one will work.
And so I think I share Rich's view that if you look at the facts, you may learn something that the general economists can't learn or can't find at this point.
Now, there have been a number of industry-specific case studies which suggest a number of relationships between market structure in that particular case and what occurred in innovation, or the -- or how the specific assets that are contained by particular firms affected their choice of innovation. And the fact that these studies exist suggests to me that when you look at a particular case, you're thinking the same thing and you may find the same sorts of relationships and may feel comfortable about them.
Now, I want to turn particularly to one particular tool that may be useful in thinking about innovation markets. Mostly people have thought about innovation markets and said well, okay, we should identify perhaps a pile of assets, I think that's Baxter's term, and I think that's a useful way of you identify a pile of assets for each firm, you see whether they are special in some way and then from that you determine whether or not there are limitations in who can effectively compete in some -- in R&D in ultimately taking some to market. And that's very useful in the near term.
If one wants to think a little bit more about the intermediate term, I think Dennis Carlton does not want to say too much about the intermediate term because of discounting, but I think it's worth thinking about. I would suggest considering looking at a firm's -- what we call in the strategy area core competencies. These core competencies -- well, core competence is a business strategy concept that is intended to enforce managers to understand what unique set of skills and technologies their company or organization possesses that will allow them to compete successfully in current and more importantly in future markets.
So, it's stepping back a little bit from the pile of assets to what generates that pile of assets. Examples of core competency would be I say Motorola's competence in wireless communication, Sony's competence in miniatures, Honda's competence in power train. Now, these are admittedly very vague categories, but the object here is to focus on what a company can take into its future competition, not what it's already accomplished. So, an existing patent is not a core competency, the ability to get patents or to defend them could be a core competency because it says something about the future.
So, I think that this approach could be useful for helping one understand sort of the big picture of competition in the industry. Now, one thing because core competencies are not linked specifically to product markets, they could also suggest the kinds of potential -- the other competitors that might be there. Okay. It's one step back, it sort of says which firms will be able to do something successfully, and I think because of that it will help identify who might be in a market, but it may also give you a sense of who might be successful in this market.
Okay, so what's one going to do with this? Fortunately, business goes up in training and there are managers to think about companies in terms of core competency. What that suggests is you will be able to go to these managers and ask them about their core competency. In some cases, you may find that some of these firms have already done an inventory of some of their core competencies. It's something to start with. I don't know that it will provide enough to make one field comfortable with a particular case, an innovation market problem in that particular case, but I think it's a useful complement to looking at specific assets and I recommend for further consideration on the part of the commission and the staff.
Okay, I had already mentioned as well that I think it's useful to consider the development of policies that might permit later review when one is concerned about how the evolution of R&D to a product market might be. The natural response, later review sounds very regulatory and that is a problem and I think that one would have to be very careful about trying to use that approach. Nonetheless, if uncertainty is something that is bothersome, there are ways to put off making a decision or at least making an irrevocable decision at an early point of time and maybe making that decision at a later point of time.
Okay, in summary, I wanted to mention again and underline my view that innovation market theory is a useful supplement to the antitrust analysis. I think dynamic competition are just too dynamic to ignore. Even for the sake of economic knowledge dynamic competition doesn't provide sort of an overall cross industry guidance that one would like. That doesn't mean that one can't looking at the facts and examining the various theories that are available feel comfortable, I think, with bringing an action based on an innovation market.
Thank you.
COMMISSIONER VARNEY: Thank you very much, Dr. Yao. Chairman Pitofsky, would you like to start us off?
COMMISSIONER PITOFSKY: First of all, I want to thank you all for what is one of the clearest and most interesting sessions that we have had in this set of hearings. First a comment and then I would like to leave you with a question. The comment is I'm not sure there's all that much distance between the people who are disagreeing. Professor Gilbert says there ought to be innovation markets, but recognizing all of the concerns and qualifications that others have identified, he says let's do it very cautiously. Others say there could be anti-competitive effects in an R&D -- in an innovation market, but they're so hard to identify and so speculative and all that we're better off leaving that alone and concentrating on anti-competitive effects in the product market. That's not a vast difference, but there is a difference.
Let me offer this hypothetical and see how you respond to it. As you well know, there are really only three companies in the world who make jet engines for wide-bodied aircraft, Rolls Royce, Pratt & Whitney, GE. Suppose they came, I don't think they would come here, they would go to the Department of Justice and say we would like to have a joint development venture because there's redundant innovation here, it's very expensive, it's vastly expensive.
It's also true that most of the -- much of the competition takes place at the innovation stage, the design stage. Let us put together a joint venture. It will do the research, it will develop a prototype, but after that, each of the three companies will produce the engine on their own and will market it on their own so that the likelihood that there are any output effects is reduced. Also, this is not a field where you could say well, but what about the fifth, sixth and seventh companies. If you haven't made an engine for a 747, you are not going to make an engine for a 777, it's just not plausible.
What would have -- what would we lose there? We could look at the market, but there are no market effects, or if there are, they are speculative spillover effects that are unreliable, but you would have lost the rivalry that leads to the possibility that there would have been one first class engine, one second class engine, one third class engine. And maybe if our theory of competition is right, there would have been a better engine as a result of rivalry than through the joint venture. Is the recommendation that the government turn its head away and not examine the anti-competitive effects of the joint venture to build the prototype, or is there some other way of getting at that? Would we permit that joint venture or would we qualify it, would we examine it under the antitrust laws at all. I leave that for any of you who want to respond to it.
COMMISSIONER VARNEY: Let's start with Professor Gilbert.
PROFESSOR GILBERT: I am delighted to. It's a great hypothetical, Chairman Pitofsky. First of all, I think it's my understanding that by statute, that the agency would have to identify the effects and specifically would have to identify the effects in relevant research and development markets that was the wording of the Research and Development Production Joint Venture Act, I have forgotten exactly what the terminology is, but in its original conception it had that wording. So, there is a statutory obligation. I certainly think that even without that statutory obligation, it is something that deserves analysis.
Now, of course, there may be very good reasons for a research joint venture, there might be very costly redundancies, there might be complementary capabilities among the jet engineer manufacturers, there might be all kinds of efficiencies. Then again, there might be anti-competitive effects and this discussion has emphasized and focused on the difficulty of sorting out those efficiencies and the anti-competitive effects.
And I certainly agree, I also agree with your summary, at least from my point of view, as to the comments we have heard, that it is difficult to sort them out. But I would also merely volunteer that if you had information, for example suppose you had information saying that as part of the joint venture we will require that all of the technology we develop be freely licensed to the whole industry, U.S. and foreign competitors. I don't know how we would work the foreign competitors into this, by the way, but suppose that there were that condition and suppose we also found in all of the participants' documents a discussion that said this is a very good provision because you know we kill each other in R&D in this business, that's what loses -- that's where we waste all our profits. And if we agree on a mandatory licensing provision, no one's going to have much incentive to develop these innovations because we all have to license it to everybody else.
Now, again, you would still have to work this through, you would have to look at it, it would be very difficult to sort out. But I for one would have a difficult time advising the agency as to merely walk away and ignore the competitive issues.
COMMISSIONER VARNEY: Do you have a different view, Mr. Rapp?
MR. RAPP: I'm not sure. As has been the case for a while, I find myself not disagreeing with much that Professor Gilbert says except its implications. The -- I guess I have -- I have to take the cowardly way out and ask a question in return. Is it not the case -- what does the innovation market approach as it's outlined full blown bring to this analysis? I certainly agree that this is an issue for the government to look at, but the way I believe you have set up the hypothetical, Chairman Pitofsky, we've a goods market to look at and to observe the impacts of the joint venture aspects of competition in that.
To the extent that the -- it seems to me, then, that the agency's task is to see how closely the anticipated activity of the joint venture affect prices and output in that goods market, and I sort of restate my nervousness and my apprehensions about inquiries that go further back to future goods.
COMMISSIONER PITOFSKY: Well, let me press a little bit on that. Without the joint venture, there would have been three firms bidding to United Airlines -- bidding to Boeing or whoever it is to sell them an engine. With the joint venture, there's still three firms bidding to sell an engine. The only difference is as a result of the R&D joint venture, it's one engine. The only thing that's lost is the competition to produce a better engine. I would even -- suppose we postulate, no effect on price or output. Would we still say that antitrust has no business looking at the joint venture?
MR. RAPP: No, I concede that it does and I concede the loss, and I -- and the question that I leave is whether or not the innovation market is a -- approach is required to analyze that.
COMMISSIONER PITOFSKY: Um-hum.
COMMISSIONER VARNEY: How about this side of the table?
MR. ADDANKI: I think I would just -- I would just -- I'm trying to place the quote in my mind, I can't unfortunately, it doesn't come to me, but I think the notion that when you have horizontal competition, that an important aspect of that might be product development, and research by a non -- excuse me, competition by a non-price means is a pretty established one.
I think I'm having more trouble with the notion, and I think if I could try to speak for more than just myself, I think people who have trouble with the innovation market concept perhaps are articulating really a problem with trying to interdict a situation where there is no horizontal competition at all among the parties.
You see, the incentives are a lot easier to understand when you have horizontal competition, and if you say I am going to remove this aspect, I am going to excise this aspect of competition, that's a loss. That's a loss of competition. It's a loss of competition in the product market. So, I mean I guess perhaps it just elaborates a little on what Dick Rapp said, but I think that's where I would come out on that.
COMMISSIONER VARNEY: Professor Gilbert, does innovation market analysis bring anything to the hypothetical additional?
I'll come back.
PROFESSOR CARLTON: Okay.
PROFESSOR GILBERT: I think it does, I mean this is a very important question. And I suspect it really may come down to how the courts defer to analyze the effects. As Chairman Pitofsky said, in the final analysis, you have three companies competing in price, perhaps with the same engine but with a different name attached to it, but still competing in price and output. The courts could take the position that there's been no effect whatsoever on a product market. There's actual competition without -- with and without the joint venture, the joint venture does nothing.
But we know that the joint venture could do something. You know, again, we don't know exactly what it is without a thorough investigation and even then there's likely to be some uncertainty in our conclusion. But what it would do would be to change the nature of the product that is available in the product market, perhaps change the timing at which the product was available in the product market. Those are effects that may be very difficult to analyze in a potential competition framework.
COMMISSIONER VARNEY: Okay, Professor Carlton, and then Commissioner Steiger has been waiting patiently.
PROFESSOR CARLTON: You have posed a very hard question, which doesn't surprise me. I think it's important, though, to emphasize that the ability to propose hypotheticals that -- in which you can show that the innovation market doctrine works shouldn't alone be the basis for using the doctrine unless you can say that I can reliably identify such hypotheticals and that there are going to be some cases that will slip through the cracks, and those are important cases, especially if the doctrine could be misapplied.
Having said that, I think what your hypothetical hones in on is that there are no other efficiencies from the transaction, it's joint venture just in R&D, so we don't have to worry about short run efficiency gains that would be foregoing if we don't allow the practice, which I think is a very important consideration.
But second, I don't think the way you posed the hypothetical that you really have abstracted from the output market. You really have focused on the quality of the good as well as the timing of the good. And those are -- I recognize difficult questions, I would not take the position that you can't think of hard hypotheticals in which these concepts and a reduction in innovation can harm consumers, it's just practically how important are they and are they important enough to have a general policy that we could apply.
I think it would be very hard to figure out is R&D going to be done more efficiently, less efficiently, are they going to have a greater incentive or less incentive, I might ask the customers what they think about it, that might be a helpful place to start and I think I would be a lot less concerned if it was two instead of all three of them.
COMMISSIONER VARNEY: Commissioner?
MS. STEIGER: I would like to turn to the utility of the potential competition theory, and I must preface that by saying that during my term here we have -- and Dennis', to use the potential competition theory, albeit with one exception that I can think of in the very industries that Dr. Carlton identified as most useful to analyze under potential competition defense and industries regulated, for example, by the FDA.
Having said that, I am somewhat amused. This potential competition theory has been widely criticized, warned against, cautioned about over recent years. I hear it somewhat restored to great respectability at this point, and having put up with the slings and arrows of various law reviews and other learned economic journals on how dangerous it was, I'm finding this very interesting. But it does lead me to two questions, basically for Rich Gilbert and Dr. Rapp. How do you respond to what appears to be a disagreement here?
Rich, I understood you to say that potential competition analysis doesn't work where a product doesn't exist. Why is that so? R&D is frequently by definition the development of an as-yet nonexistent end product. Dr. Rapp seems to indicate that the analysis falters where there is no product market for the future product. Why is that so?
Presuming logically you can say there is a utility to whatever this thing is under development, why do you need to be able to identify a future market in particular, given Dennis' examples of the by-product markets that seem to spin out with great regularity, and I am referring to this Teflon example, for example, although he gives us others about products that started out as heart devices and wound up in textiles.
So, that's a long way around to trying to get an answer to two questions about the now hallowed potential competition theory.
COMMISSIONER VARNEY: Professor Gilbert and then Dr. Rapp?
MS. STEIGER: And is that just a case of the devil who knows better than the one you have phrased?
PROFESSOR GILBERT: Well, Commissioner Steiger, you correctly point out that much of this discussions from the basis of presumptions about what the potential competition theory is. In this case it's a devil we assume. And I think it's the case, obviously the other participants have to verify this, but I think it's the case that we are all saying that if there are adverse economic consequences, then the antitrust laws should attempt to deal with them whatever framework we have for that analysis. If the adverse consequences are clearly there, they should be dealt with.
Now, it is possible for at least in theory, I guess you don't know what the jurisprudence would allow, but in theory for a court to say that despite the fact that there is no existing product market here, we have two companies that are each the most likely potential entrant into this nonexistent product market.
The courts might also take the view, I just don't know, the courts might take the view that there is no product market, therefore there can't be a potential competition problem because the market doesn't exist. The outcome of that exercise clearly depends upon how the courts frame the issue.
There's also another point that I will provide a critique of my own comments here, just in case there hasn't been enough, which is that some of the effects that I am concerned about may not exactly pass a review of a competitive effect. If I take as an example a market in which let's go back to GM/ZF. ZF leaves the market, we are left with General Motors, they are the only company in the United States.
I would be concerned -- if it's the case, I would be concerned if innovations that are slowed down as a result of decreased competition in Europe, if those innovations don't make it to the United States, that to me is an economic effect that I would be concerned about. I could also see the courts taking the view that well, maybe there's an innovation market effect here, but there's no competition that is affected.
So, a lot of the outcomes are so heavily dependent on exactly how the courts view their enforcement -- their legal interpretation of what competition is and what effects are and what potential competition is. We're all making assumptions about that. I just don't know if the assumptions are valid ones. And so, my out is I'm not going to assume anything, I am going to do it on economic facts and what the courts follow.
MR. RAPP: I think my answer is going -- to your question, Commissioner Steiger, is going to be an exercise in psychoanalysis. I think that while the reference to the utility and adequacy of the potential competition approach that I made, I think that I had, and perhaps others, have used it because it brings us back, it is a way of bringing us back to real goods markets and to the integrity and power of the horizontal merger guidelines as a market means for market power analysis. And while that doesn't answer specific questions about how -- about what the failings of the potential competition doctrine are, it points up the fact that ambiguities that might -- that innovation market approach presents and that looser forms of analysis present are clarified once you bring things back to product markets.
For example, the difference between a product and a product market is important. And if I may say so, if we were able to discuss further the second of the examples that Professor Gilbert provided in his testimony about a new form of packaging material or bottle, something like that, I think one of the directions that that conversation might take is do you really mean that there is no -- that because there is no product that there is no product market that we could analyze, you see, and given the ability -- and if that new product as yet not on the market were likely to enter an existing product market consisting of glass bottles or cans or whatever, then I for one find myself comforted by being able to hang my hat on the horizontal merger guidelines. That's one person's opinion for why the potential competition approach is so comforting.
COMMISSIONER VARNEY: Any comments?
PROFESSOR CARLTON: If you want me to say something I will. I didn't mean to give --
MS. STEIGER: First of all, thank you very much, I just want to echo what our presiding officer has said, this is an extraordinary presentation, probably one of the finest we've ever seen, but I would tell you, Rich, it is not coincidental there is no blackboard. We put them away -- we put them away when the economists come in.
Dr. Carlton wanted to speak.
PROFESSOR CARLTON: I just wanted to say that the potential competition doctrine, although it's better than the innovation doctrine because it's more reliable, still has the problem that it's trading off certain future efficiencies in the near future that a merger usually can create against predictions of benefits in the future. And for that reason, I think it's correct that it's a doctrine that itself has the potential to do harm. I think the reason people prefer or some of the people here, or maybe I ought to just speak for myself, prefer the potential competition doctrine to the innovation market doctrine in that it is more reliable -- however unreliable we think potential competition doctrines are, it is more reliable than innovation markets.
COMMISSIONER STEIGER: That is the case of the devil, as you know. Thank you very much.
COMMISSIONER VARNEY: I think our director of the Bureau of Competition, Jonathan Baker, has a couple of questions. Bureau of Economics. Is competition a promotion from economics, Johnathan?
MR. BAKER: I thought I would develop a theory.
COMMISSIONER VARNEY: Excuse me, I didn't mean to insult your position.
MR. BAKER: When -- at a program recently Ann Malester, who is at the other end, and Mark Whitener were --
MS. MALESTER: Misidentified.
MR. BAKER: -- identified as deputy directors of the Bureau of Economics, I sent them a congratulatory note, too.
COMMISSIONER VARNEY: I think our resident economist who is also a lawyer has a couple of questions.
MR. BAKER: I have a question for Dennis, Dennis Carlton, we have multiple Dennises, and it's about the discussion in your talk about why you believe we should not pursue concentration increases in innovation markets because you said there was no theoretical justification for doing so and only weak empirical one. And you went through the logical links that would be required for a theoretical justification, and I had problems with that. And I -- when I was listening to you, I was worried that the argument that you were making proved too much because a similar logic might suggest that we should not -- we should equally not pursue concentration increases in goods markets, and you seem to be comfortable with pursuing those.
If I had a blackboard available, I could explain in a little more detail what I mean, but let me try a little bit verbally. You said how do we know that the reduction in R&D would be bad. Well, how do we know that a reduction in quantity is going to be bad. We could have -- it is possible as a matter of theory that we could have excessive entry in markets, business stealing effects could lead to that in some markets. We have some markets like health care where the agents are deciding both what services to provide for the patients and how -- to provide it themselves and there could be acquiring too many services for those patients. In other words, we could have excessive output in goods markets as well.
And you said the number -- how do we know that when the number of firms decline we will get less R&D and less output of R&D -- innovations as a result. Well, how do we know when we have less firms we will have less output of output in goods markets? After all, with goods firms we could have inefficiencies, less assets, lower costs as a result, and, in fact, just as you said with R&D markets, I would indicate the empirics of what happens between the number of firms and cross-sectional studies and what happens with price, the volume of output here as a weak one, and then you said not -- there aren't -- there might not be enough firms or how do we know there aren't enough other firms besides the ones we are looking at to step up for their lost R&D and produce R&D in the future.
After all it's hard to measure, et cetera, et cetera. Well, in the same way it's conceivable we could have a market where potential competition is very important, the entrants could be anyone in the economy, it might be difficult to -- and it might be difficult to measure a concentration as well in those settings or in a setting where market products are differentiated. What I was wondering is why do you think these two areas are so different and why do you think we should be looking at potential competitive problems and efficiencies and the trade off between them differently in the R&D area than in the goods market area?
PROFESSOR CARLTON: I think that's a very good question. I think the short answer is I believe we have a more reliable base of knowledge both theoretical and empirical when it comes to concentrating existing markets than R&D. In my paper, I actually deal exactly with that question. I think you are correct to say that cross sectional studies have been the basis for mostly "weak or lack of empirical findings in the R&D literature." And it's well known, as I talk about in my written testimony, that cross-sectional studies have lots of infirmities.
The same could be said, therefore, of cross sectional studies relating price to concentration and in fact has been said about that, and I actually have criticized such studies heavily in my textbook. However, there are also studies of individual industries which get around the infirmities of cross sectional studies that I think provide a more solid basis than we have in the R&D area for saying that increase in concentration can lead to higher prices.
There is a debate as to how important concentration is when it starts being a problem. And I think people can have legitimate debates about that. But I think there have been enough studies of individual industries where we have a sense, for example, that a merger to monopoly could lead to higher prices. In terms of the theory, I think the theory is much less ambiguous in the case of existing product markets than new product markets.
I for one would not rely heavily on potential entry stories to allow mergers that create high concentration in a current product market unless I could get pretty solid evidence that this potential entry story had some basis in fact. It's too easy to make those arguments, I think.
Now, so, I wouldn't say allow any merger because of potential entry, I would require hard evidence, and I think that's where current antitrust policy is now. In terms of predicting who might be in R&D, that's true, I agree, that could be heard, but all I'm pointing out is even if you know who is in R&D, predicting what products they are going to produce is a very difficult problem. And if you just do some back-of-the-envelope calculations about how much do I have to discount future benefits, if there's a probability each year that, you know, someone else is going to come in and we'll do it anyway, you have don't have to go out very many years before the discount rate is a very high, high rate.
Now, you raised one other question which had to do with whether more entry is good, whether competition is good, whether concentration wouldn't be all right in some markets, for example. That really raises a question about externalities. Let's suppose there's some product that we think is undesirable, but people are consuming it anyway. Should I allow a merger to -- should the FTC allow a merger to monopoly? The answer would be yes, price goes down, output goes up, you would have less consumption. Actually you could make such an argument, my hunch, though, my recommendation would be that policy elsewhere, but what's an externality or not should be handled somewhere else if the FTC or DOJ should take as their charge to say given the policies and how the externalities have been created, my goal is to preserve competition.
MR. ADDANKI: I was going to add one small remark to that. If there really is a negative externality from a good, it seems to me that the better way to deal with it from a policy matter is to have it be as competitive as possible and to attack it deeply so that at least you get to collect the rents on the product instead of rents going to monopolies. Pardon me?
COMMISSIONER VARNEY: Let me go to Dr. Rapp and then, Jonathan, unless you want to come back with anything, I would ask Dr. Rapp to wrap us up before lunch give us your thoughts.
MR. RAPP: Thank you. I have an overpowering urge to supplement Professor Carlton's answer having wrapped myself in the flag of the merger guidelines. It seems to me the answer to that question has to start with statute and include the distinction between the defense of competition and the defense of welfare maximization because a lot of effects and defects in the analysis of the relationship between structure and welfare gains that you mention in your question get you past the point of competition.
The question of whether competition is good or bad. If you take the analysis one step back and put it only in these terms, is there more of a basis for supposing that conventional horizontal merger analysis in product markets is more capable of discerning pro-competitive from anti-competitive outcomes rather than welfare enhancing versus welfare reducing outcomes. I think it's a clearer answer and I think it's clearer on theoretical, not empirical grounds. The literature as Professor Carlton said and as you well know is weak on the relationship between concentration, profitability and so forth and I don't think that it provides the basis for the enforcement activities that have been -- that merger guidelines have been the basis for. I think the -- it is the analytical underpinnings, the strength of the analytical underpinnings of the horizontal merger guidelines that gives them their power and I think that is what is missing from the innovation market approach more than the empirical side of the critique. That's my supplement.
COMMISSIONER VARNEY: Professor Yao, any comment on that?
PROFESSOR YAO: I just want to make sort of a general point, or leave you with a general question, which is it more useful for determining the projectory innovation in an industry. If it turns out that history is useful in some industries, then I think one will have a sense of where the innovations are going to come from, who is most likely to innovate and what is most likely to happen. Now, we can come up with many, many examples in which history doesn't help, and there we've got a problem, but there may be a number of industries in which history does help and I think that's just a return to focusing on the facts of the industry in trying to go forward with one's analysis of innovation rights.
COMMISSIONER VARNEY: Well, I want to thank our panelists for this morning and invite you to stay with us this afternoon when we reconvene at 1:30 to continue this discussion, albeit with some new people and slightly redefined questioning. Thank you very much for your input this morning, and we will continue at 1:30. Thank you.
(Whereupon, at 11:54 a.m., a lunch recess was taken.)
//
//
//
//
//
//
//
//
//
//
//
//
//
A F T E R N O O N S E S S I O N
(1:35)
COMMISSIONER VARNEY: We will go ahead and get started, get the record started. I am pleased this afternoon that our panelists have been able to stay with us and we have been joined by two new panelists, Mike Sohn and Judy Whalley from yesterday. I am going to go ahead and introduce Mike and Judy now and then go back to Rich who is going to make some remarks, then I think Mike and then Judy and then we will just go right into the discussion. And Michael, you weren't here yesterday, so I can tell you the score is about three-three on innovation markets, either for them or against them, so you get to weigh in this afternoon so we can sway the balance.
Mike Sohn is a partner in the law firm of Arnold & Porter where he heads their Antitrust Practice Group. Mr. Sohn's practice encompasses a broad range of antitrust and consumer protection matters with a particular focus on mergers and acquisitions. He has represented such clients as Allied Signal, American Home Products, Baker Hughes, General Electric and Occidental Petroleum. It doesn't say Boston Scientific here.
MR. SOHN: That's an older version.
COMMISSIONER VARNEY: From 1980 through 1987 he served as general counsel of the Federal Trade Commission. He was a member of the Administrative Conference of the United States from 1977 to 1981 and a member of the Executive Committee of the Regulatory Council of the United States from 1978 to 1980. Mr. Sohn has written various articles regarding antitrust and consumer protection issues. Welcome, Mr. Sohn.
Judy Whalley joins us this afternoon, she is a member of Howrey & Simon. Prior to joining that firm she spent fifteen years with the Antitrust Division serving as a trial attorney, Chief of the Chicago Field Office, Deputy Director of the Office of Operations and ultimately the Deputy Assistant Attorney General for Litigation. In 1988, President Bush named Ms. Whalley Distinguished Rank Executive, the highest award bestowed on senior government officials. She has written and lectured extensively on antitrust issues and teaches antitrust at Georgetown University. Welcome back.
Rich, would you like to start this afternoon?
MR. GILBERT: Yes, thank you, Commissioner Varney. This afternoon I understand we are going to address how to assess the likelihood of unilateral or coordinated conduct in R&D and also how to evaluate the likelihood of entry into R&D and the future generation of product markets.
As we discussed this morning, many of you may not have been around to hear this discussion, so at the risk of some redundancy, I will go through some of the issues. In order for a coordinated conduct, whether it's tacit coordination or actual explicit coordination, in order for it to be sustained, a number of conditions have to be satisfied and this is whether we're talking about research and development or whether we're talking about output markets.
First, there has to be some distribution of benefits such that each firm is better off when everybody, in fact, coordinates their behavior than when they act independently. So, in effect, there must be some gains from coordination. That's often the case in conventional product markets. It may not be the case in -- always in research and development because firms may have quite different preferred research and development paths, they might have different core competency, they might want to develop products that may take advantage of their complementary assets in other product markets.
So, they may not be able to rate what's really better for each and for all of them relative to their independent conduct outcome.
Even if there is something that's better, there might be several R&D programs that are better and they have to choose which of the best -- which of the better programs that they will actually coordinate on. There might be two entirely different research paths to develop say a new jet engine, but the -- in order for a coordinated outcome to succeed, it has to be the case that everybody agrees that it's either going to be path A or path B, you can't mix them up. If you mix them up, you lose the coordination. And that can be difficult in R&D, a little more difficult than in product markets.
You have to have a mechanism to monitor adherence to a collusive outcome. Again, whether it's a tacit collusive outcome or an explicit collusive outcome, you have to be able to check and see if people are behaving the way you want them to behave. If not, then it's in the interest of each member to act to satisfy their independent objectives because they know that nobody's watching, they can cheat.
Again, that can be difficult in R&D because much R&D is conducted under secrecy. You need a mechanism to punish anyone who, in fact, cheats on the collusive outcome. The problem in R&D is that the punishment is going to happen after a firm deviates, which -- and that may not be detected until after a firm is successful.
And so at that point, here you have a competitor that has achieved the benefits of independent -- of cheating on a cartel, watching everybody else slow down their research and development programs while this one cheater speeds up, develops a new product or a new process and then it can be quite hard for others to punish, unless they're in a very repeated relationship where things happen every two years or they can punish them at an output market. But you can see how in many circumstances once the cheating occurs, the damage is done and it's too hard to punish.
And finally, you have to prevent entry into this marketplace. R&D is a common entry path. I don't think this is as much of a problem as maybe some others on the panel think. There are circumstances where you know who the credible R&D competitors are and you just know who's in this game and who isn't in the game, and it's not easy to really gain the status of credible innovator. So, I think when you summarize all of this, it does lead me to conclude that concerns about research and development should focus on unilateral behavior rather than coordinated conduct.
Now, at the same time there are some possibilities, one thing I think that's worth being cautious about is arrangements in which participants either in tacit or explicitly might facilitate the exercise of unilateral behavior in ways that are anti-competitive. One of the -- I'll give an example which is not research and development, but the airline theory of publishing investigation where the concern was a particular reporting mechanism which once adopted led to incentives to engage in certain types of disciplining behavior and enforce -- possibly enforce -- having not been that close to it, I don't know all the facts -- but possibly enforce coordinated outcomes.
One can imagine something like that in an R&D context, I don't know exactly what it would be, but I mentioned earlier today the possibility of some elastic agreement about the licensing of technology from an R&D arrangement which might have the effect of causing each member of this group to have a unilateral incentive to reduce incentive -- to reduce investments in R&D. Coming not from the expectation that others are slowing down their own R&D program in response because of the mutual interdependence, but rather because a framework has been developed so that everybody has a less incentive to engage in R&D. That might be a concern. But tacit collusion, even explicit collusion I think would be considerably more difficult in most R&D environments than price coordination.
The second topic I would like to address here briefly is the likelihood of entry into R&D and how to evaluate that. Well, clearly innovation is an important route of entry for new competitors into an industry and all else equal, the more R&D that is going on, the more entry that you would expect, and as my colleague -- colleagues, Jorde and Teece have maintained that such entries probably not going to be very price elastic, meaning that if you apply the five or ten percent small but significant nontransitory increase in price, probably not much will happen, but that doesn't mean that the entry isn't out there. And I do think that that's a factor that the agency should consider the likelihood that there may well be important entry that is not particularly price sensitive.
Where I take issue, though, with this approach, is that I wouldn't -- I wouldn't put a lot of emphasis on this entry unless it is fairly close in time fairly likely, or unless it is likely to be -- it is more likely to occur if there's the -- an exercise of market power. Because in that circumstance, you might expect that the -- the entry would discipline the -- might discipline the exercise of market power and you might also have a situation which I think is really a -- should be a focus of this afternoon's discussion, and that is where the exercise of market power may make the entry more likely and be a good thing because it introduces a new product or a new process.
Another way -- so, what I'm focusing on now is how we might want to rethink certain aspects of the entry section of the merger guidelines to take into account certain R&D-related phenomenon. And I think there are two areas where some rethinking is advisable. One is that entry with R&D can be very drastic. The entry section of the merger guidelines is written largely I think with the expectation that entry is somewhat incremental, that is there might be a price increase and in response to that price increase there might be a flow of capital into the marketplace and that might neutralize that price increase. And if it happens more than two years out, there's a tendency in the guidelines to discount that.
Well, what if three years from now there is drastic entry, the likelihood that entry will occur that is so dramatic that it will just change the marketplace completely. We can ask whether that will be more important than incremental entry that occurs in two years and whether or not the guidelines should account for that to some extent. I will also say, though, I feel quite strongly about this, is that we should not allow mergers to create market power just because there is R&D entry happening at a future date. Why tolerate the exercise of market power just because entry is going to happen at some future time.
Now, where that evaluation is quite different, and I'll give you examples where I think the evaluation has been done, is where the exercise of market power may be a necessary consequence of creating a new product or a new process. An example I like to think of is the creation of -- is the accumulation of spectrum for cellular telephone operations where it is possible that the accumulation of spectrum may give rise to certain exercise of market power say in a particular part of the spectrum. So, maybe what's going to happen is you are going to take some spectrum and concentrate frequencies so that you can offer cellular services, efficient cellular services, but there might be a piece of spectrum where you have some users left over who are going to suffer from higher prices because they might actually get more concentrated services. The example is dispatch services for cellular dispatch.
Now, if it's the case that that concentration of spectrum helps bring along the entry of a new product and accelerates the entry of the new product, then it seems to me that that's a calculation that the agencies should address and should balance the anti-competitive effects in say a dispatch market against the likelihood of creating a more efficient cellular market. But the mere fact that entry is going to occur or that there's going to be a development, the very fact that the telecommunications industry is progressive and has drastic entry that occurs on a five-year basis, that alone I don't think is enough to justify transactions, acquisitions and mergers, joint ventures that create market power because that market power is a welfare loss for consumers and unless it promotes R&D in some sense, why should that -- why should consumers face that welfare cost without some benefit in technological progress.
So, while I think that there are possibilities for revising the merger guidelines and evaluating the entry section of the merger guidelines to accommodate certain drastic innovations and maybe rethink the entry aspect of the merger guidelines to some extent, I wouldn't go overboard with it. I don't see the point of accepting aggregations of market power unless you can -- unless you can establish the link between the market power and the pace of innovation.
Thank you.
COMMISSIONER VARNEY: Thank you, Professor Gilbert. You might think about when we come back to the discussion some of your colleagues yesterday suggested that we ought to think about entry not in what they refer to as arbitrary time increments like two years but more in terms of product life cycles, that that ought to be our entry. We'll come back to that.
Mr. Sohn?
MR. SOHN: Thank you, Commissioner.
As Rich has eloquently written, and I should say that I guess Rich has now achieved what all economists hope for, has written an article which is controversial, colleagues have replied to him and he has written some replies, and it shows possibility of going on for some time. I congratulate him.
On the other hand, being in private practice, I have considerably less opportunity and ability as well to think about these things abstractly. And so, when I read Richard's article and as I have tried to follow the literature since, I began to ask myself well, have I seen this in my practice, have I read about it in the case law. Is this happening in the real world. The it being coordinated behavior, unilateral effects to slow the pace of innovation or drastically change its path. And the answer is that I have not. And the literature that Rich and the others have contributed don't contain much in the cites to the case law as well. Everybody cites the same single consent order involving the Automobile Manufacturers Association in 1969 and an allegation that they acted collusively to restrain the development of pollution control equipment.
As I understand it, a grand jury was convened in that case, but only a civil consent came out of it, and even then the civil consent was modified several years later to take out specific constraints on certain collaborative behavior because the world had moved on and the conduct initially restrained was no longer viewed as unambiguously anti-competitive. So, that's not a rich history of experience under the Sherman Act.
I think it teaches what Rich and others have written, that one must approach this in a careful and focused way and not go overboard with it.
The markets that I would like to focus on and the terminology important are markets where I think a case can be made that the merger guidelines don't, at least without some modification or amplification, quite do the job. And these are I think future generation goods markets in which the next generation product is felt to completely displace in a relatively short period of time the existing goods market. And the second is the purer form of innovation market which comes in two varieties, I guess, you can have the kind of situation which you had in the consent order that the Commission entered in American Home Products and American Cyanide where the merging firms were two of what were allegedly very few competitors seeking to invent a new vaccine for rotovirus, there being no vaccine today.
You could go back even further I guess and say that the pooling of certain research skills which are scarce even where the product in mind is considerably less fuzzy than it was in the case of the rotovirus vaccine might be a problem as well, but I think as many have pointed out, the further back you go the more difficulty you have applying the theory in any meaningful way.
Let me turn to -- briefly to efforts to link concentration with anti-competitive effects in either future generation markets or innovation markets. Of course, as everyone knows, the horizontal merger guidelines does this in a quite pronounced way, and there is a body of empirical evidence which at least to many suggests that the link is real and beyond the body of empirical evidence, whatever you think of it, there's a robust oligopoly theory that suggests that collusion is possible, at least under certain circumstances where detection is possible and punishment is possible.
I think the same does not hold in innovation markets. I do not detect any empirical basis for the belief that there's a link between market concentration and the pace of innovation. And I think that as many have pointed out in the absence of a robust theory of oligopoly that would support such a link. I defer to others on the panel, but at least as I understand Arrow's work, which is often cited for their proposition that a monopolist may have less incentive to innovate and an assumption in his model is that the innovation occurs with respect to and in close proximity to a good that's being monopolized. And I think that it may well be unwise to generalize to innovation markets from that theoretical work.
The likely reasons for the absence of empirical and theoretical evidence linking concentration with competitive effects in these markets has been identified. I'll just tick them off briefly. Very hard to measure shares, how does this agency go about deciding the relative competency of firms doing R&D.
I can tell you again based on my own practice that when one looks at the documents of the merging firms, one is on unsure ground. There is what I have come to recognize as the rose projection phenomenon. Every firm these days is short on R&D funds. And that spawns boundless optimism in the mind of those seeking those R&D funds. When you compound that difficulty with a not unprecedented possibility that other firms in the market may seek to gain the commission process by downplaying their own place on the path to innovation, principally because they may be concerned about the efficiency gains of the merger, any kind of close effort to approximate shares with where one stands or what one has spent on innovation, I think is very suspect.
I think, and I think Commissioner Varney has suggested that one over end is probably the safest way to go here so long as you have one bear set of skills that can be brought to bear on the innovation market in question.
Let me turn to unilateral effects first. The way I think about this is to look at what the likely post merger incentives are going to be for the merged firm. Again, in a goods market, there's fairly common ground that a party with market power will follow the incentive to set a profit maximizing price which often is above the competitive price. But the mergers of a -- and the incentives of a monopolist in an innovation market is far less clear. Let me for the sake of time just give one example. Suppose you have three going to two in a vaccine market. Yeah, suppose that. What do you call it, a stylized version.
COMMISSIONER STEIGER: Let's call it a hypothetical.
MR. SOHN: Right. It's not clear to me that one ought to leap to the conclusion that the merged firm or the two firms will have differing incentives. If you are in a situation where the merged firm perceives an ability to appropriate the returns on its R&D investment either through patent or some other significant first move or advantage, it's not at all clear to me that the pace of innovation will be slowed even by three going to two. Indeed, you could argue that if you have -- particularly where you have three firms who are equally likely to reach the market prior to the innovation, prior to the merger, they will spend one-third of what they anticipate as the gains of getting there first and may spend one-half of what they anticipate to be the same gain after the merger. So, you may actually have increased R&D. And I think it's a very murky picture.
One word on the concern which has been expressed about the loss of a different research path. And here the concern is that the merged firm will choose which of two research paths it now owns is the more likely winner and put on the shelf the other one. Well, that may well be the case, but it's not clear to me that this is a bad thing. It may or it may not be. It is not clear to me that an enforcement agency is in a position to say that the merged firm that makes such a choice will likely choose wrong. Nor is it clear to me that it would be easier for an enforcement agency to judge whether in the case where the merging firm is going to achieve some efficiencies as a result of the acquisition and spend more on what it perceives to be the winning track, that spending more on the winning track than was spent premerger while shelfing what it perceives to be the losing track isn't the better thing for consumers.
Just a word about coordination because there seems to be a great deal of common ground here. Everyone seems to believe that it's very hard to do and I am not -- I'm not going to go through all the reasons. The research didn't take place in public, assuming you could agree on the terms of coordination, which is I think a considerably harder task for the would-be cartel since there's no higher price out there to agree on. You have to agree on such things as research paths or how fast or how slow you are going to go and your activity or your partners is not taking place in an open marketplace. So, that makes agreement on terms difficult, it makes detection difficult, and as Richard just pointed out, punishment is uniquely difficult because it may well not -- the deviation may not be discovered until far down the road.
Now, there may be exceptions. In the pharmaceutical area, most meaningful research has to go through clinical trial, these are supervised by FDA, they take place in the field. In the defense industry, as Jan is familiar, there's a lot of government funding and briefing back to various would be participants in the R&D pot as to why they did get it and didn't get it. And you can in some industries at least get a glimmer of whether someone is cheating on a hypothetical cartel. But I submit that risking, particularly in a context like defense where there are significant rewards for winning and there may never be another procurement, or at least not one for a long time, participate in a cartel even with that vague reflection of what's going on in the R&D market is a risky business.
A word on entry, and I hope I haven't taken too much time, I think one must define what constitutes entry in an innovation market. The guidelines note that to be effective in the context of a goods market at least, entry must deter or counteract the anti-competitive effective concern. So, in a goods market a perception that entry is going to be timely, likely and sufficient, would at least in theory deter an anti-competitive price increase post merger.
I think it's not a great leap to conclude that you can have a similar analysis in the R&D or innovation context. If the anti-competitive effect of concern is a slackening of the pace of innovation, a new R&D entrant can announce that he has made a considerable sunk investment and intends a major R&D effort and if it perceives slackening of the pace of innovation, and the mere possibility that there are folks out there who could do it would deter slackening of the pace of innovation.
In terms of timeliness, likelihood and sufficient -- sufficiency, I don't think there's a case to be made for less than two years in an innovation market. I frankly had not thought about Richard's point about innovation being drastic, and perhaps at least under certain conditions longer periods than two years may be useful because if the entry is perceived as coming in three years, but to be Earth shaking, then slackening of the pace of innovation may well be deterred.
In terms of likelihood, this is harder than a goods market, the 1992 guidelines talk about determining whether entry at minimum viable scale could be profitable at premerger prices without exceeding the premerger sales opportunity of the new entrant. Well, what everyone thinks of that task in a goods market, and I have always hired a good economist to help me think about it, there are at least some objective measures that you can point to in a goods market, which I think are largely lacking in an innovation market. And I think if we're going to have this theory, we need a lot of creative thought on how to think about innovation in this kind of market.
Some very preliminary thoughts, one possibility is to mimic the guidelines approach and think about the anticipated size of the new product market. And how many companies of minimum viable scale can it ascertain -- can it maintain once it arrives. This can be a particularly relevant factor it seems to me where you have an equal probability that each of the firms premerger can achieve success because the merger may create room for one more at minimum scale who is not presently doing the R&D.
I think there's kind of an uneasy balance between one's belief in the likelihood that collusion can take place because of the possibility of deviation being detected and punishment being meted out, and one's belief that the right signals about entry will be sent to deter anti-competitive behavior. If you think collusion is likely because these things are perceivable enough to permit enforcing of the cartel, then you should believe with equal fervor that that entry is likely to deter anti-competitive behavior.
I think with Rich that identifying firms who are likely potential entrants is not all that much of a mystery. Some likely sources would be companies whose existing products would be made obsolete by the R&D, companies with R&D projects currently aimed at different products but which involve similar skill sets to the innovation market at issue, and companies for whom the R&D effort would broaden an already existing product line and enable them to take advantage of economies of scope or scale.
Some conclusions, I think it's common ground and is likely to be after today that enforcement efforts should proceed cautiously here until we know more. I urge the commission in this kind of environment to publish clear enforcement guidelines respecting future generation in innovation markets. There is, I perceive, a critical need for guidance both in the business community where the entrails here are particularly hard to read, and deals that may be a waste of time as you see it are actively being pursued. And I think it would be useful to promote the colloquy between counsel and the staff to have guidelines out there that everyone could point to.
Related to this, much of the enforcement letter to date in this area has taken place in the context of consent orders which issue -- are much larger transactions, American Sound, $10 billion transaction, on overlap of R&D in the process of vaccine. I think it's inevitable, but I do think that the parties do have strong incentives to fix the problem and move on, and those incentives may be so long that the safeguard of the litigation alternative or even a very vigorous defense before the commission may not exist, so why rush to get it done. I think in that context it's very important that the commission clearly set forth and consistently apply its enforcement principals.
One final thought, I would seriously consider a safe harbor in that area. Mine's not going to be any better than anybody else's, but I do think it's fair to comment that if there's common ground or largely common ground, that anti-competitive effects are less frequently encountered in innovation markets and rather than to coordinate and maintain the coordination, then it strikes me that there should be a more hospitable safe harbor.
In a related context, one commentator suggested that if three firms remain or entry by three firms would be adequate, that should be enough to close the books without further analysis. I think that's a view that's worthy of serious consideration.
Thank you.
COMMISSIONER VARNEY: Thank you very much. I think that we might come back to that you might want to give a little thought to, yesterday and this morning we heard quite a bit, I will take liberty of paraphrasing it, quite a bit of encouragement, hey, look, you don't need to use innovation markets, if you have a real loss of competition, you can probably use a future competition market. And I would ask when we come back to the discussion, how does that fit with your unilateral anti-competitive effects if you've got two firms with the same research path merging while perhaps we don't -- we can use the horizontal merger guidelines and say you've got to put one out to license and you decide which one you are going to keep, but let the market decide whether or not the other one is worthwhile.
COMMISSIONER STEIGER: May I add to that list some amplification of the suggestion that a monopolist might have no interest in retarding R&D. What about the situation where, in fact, the monopolist has an obsolescent potentially good, why would he not, in fact, wish to retard R&D particularly if he could effect that retardation on an innovative market in order to maximize the profit from his potentially obsolescent product.
COMMISSIONER VARNEY: Ms. Whalley?
MS. WHALLEY: Thank you very much. I suspect that I am going to echo much of what has already been said, and I will try and keep those remarks brief, but I do think it's important to touch on the point that the potential for innovation being retarded by the exercise of market power has been identified for a long time in the case law, but not much has been done with it. And particularly at the agencies, there has not been much of a focus on the potential of retarding innovation until just recently. I would applaud the agencies for renewed focus on that issue because innovation is so critical to the success of the American economy and the ability of American companies to compete in global markets, that it is well worth the investment of thinking and investigating resources to try to ensure that the innovation is protected.
Having said that, though, I do also want to echo a concern that I have heard here today that the relationship between concentration and innovation is a very complex one. It is not clear that there is a direct and positive relationship between concentration and innovation, and that that uncertainty falls sort of in my mind into two categories.
One, the overall concern that there is not a direct relationship between the level of concentration and the case of innovation. And I would agree with Mike Sohn that there simply is not empirical evidence that's substantial enough to support that argument. There's certainly strong proponents on either side of the question. A second area weakness in our analytical understanding at this point is in individual markets themselves, what factors affect the pace of innovation. Unlike price competition where there has been a great deal of work done, both empirical and theoretical, to aid us to identify key factors that increase the likelihood of vigorous price competition increasing or being reduced by market power, that simply is not yet there in the analysis of the markets.
So, I think first we cannot transfer the presumption about concentration leading to a risk of diminished versions that we have on the price side, that discussion leads to a potential price competition. I also think that our ability to understand the factors in individual markets that would lead to reduced innovation is not a power as heavy on the price side.
That to me leads to echoing again what I've heard both Rich Gilbert and Mike Sohn and I suspect others this morning say that given that lack of understanding and ability to predict, it is particularly risky to try to do cases on coordinated effects basis. And I would recommend that such cases not be brought except in the most extraordinary of circumstances. Situations that I might conceive of where it would be appropriate to do that is where there's already substantial evidence indicating that coordination is going on between the merger and that the merger is likely to make it cleaner to reinforce the ability to coordinate, that's already been demonstrated.
A second situation that I might conceive of where coordinated effects could make sense would be where there is a maverick innovator, install the guidelines analysis of the price competitor. A price competitor who has been pursuant to the research path, who has been in other ways disruptive is now being acquired where there have been other evidentiary indications of stability in conducting research and development among the existing parties but more the maverick. However such circumstances might exist if they ever would, I would venture to predict, otherwise I would say they would be excessively less.
The factors that I have identified and won't reiterate make coordinated behavior with respect to innovation so difficult I think to outweigh the arguments that would support bringing coordinated effects cases. I think the risk that we would wind up deterring innovation and eliminating efficient mergers is simply too great given the uncertainty that or the unlikelihood that the result of the entrants would be -- result of the measurements would be coordinated at best.
One other point I would like to make here, you know, following up on this point about uncertainty and our lack of a full understanding of the role of market power in innovation, some have argued that that's a sufficient basis for walking away from concern about innovation entirely. And I do not think that was appropriate. As I said earlier, innovation is simply too critical to our economy to say we did not understand innovation well enough, therefore we shouldn't be concerned about innovation at all. I don't think that's an appropriate way to go. What I think is important is that the agencies adopt very clear and very rigorous standards for when innovation markets or potential competition markets, future generation product markets are used to ensure that the focus has been only on those cases that have the greatest likelihood of accurately predicting an adverse effect on competition.
So, I would argue not that we abandon the concern about innovation, but that great care be taken in setting out standards and rigorously applying those standards to identify the cases where the problem is very significant, where there's a higher degree of certainty than we would want to have in price-based cases.
I think with any new theory or approach in antitrust enforcement, there's also a risk of excessive adoption. I have seen that in my years of enforcement, a new theory comes to the front, it's an advance in learning and understanding, everyone becomes very excited, everyone says it's applicable to their case, and the result is that it winds up being used in cases far beyond its real ability to predict. I would point to what I feel was the excessive use of the analytical tool of Chicago School thinking in the early 80s as an advance of that. Clearly cases were being enforced that were important, but it became a marginal theory in its value. The point being there was a lack of actual usefulness in prediction, and it was not as rigorously applied as often as it had been in that fewer cases should have been brought, an opinion that the innovation theory, which is an exciting new advance, not fall into the same misuse here resulting in more cases being brought than should be bought in deterring efficient mergers that should be allowed to go forward.
Let me talk about some specific proposals for these more rigorous standards. One I have mentioned, I really don't think at this point it makes sense to proceed with coordinated effects cases except in extremely rare circumstances. Second, I think that such a presumption should be made in favor of worldwide geographic markets. The ability of information to flow worldwide is separate today and increasing every day as the computer linkages, the ability to communicate across borders increases. Ideas are generally not subject to the kinds of constraints which limit the flow of products and service. Tariffs, shipping, costs, availability of distribution or services, brand name recognition, they flow over borders. Ideas outside of the United States by firms not participating in the United States can be disseminated here a number of ways, by sale or transfer of the innovation to companies, fringe companies or new competitors, perhaps upstream or downstream participants in the U.S. market, or sponsorship of a new U.S. entrant. I think, again, only in extraordinary circumstances should the geographic markets or innovation markets be narrower than the world because it's simply unlikely that the flow of information is going to be prohibited.
One point I would like to add on, this coordinated effect point, which relates also to unilateral effects and entry, and that is that one of the difficulties that's been identified in pursuing with the coordinated effects theory is what kind of market share do you ascribe, how do you determine the concentration in the market, the relative positions of the companies in the market.
What I think is perhaps the greatest advance in thinking in the innovation market theory that's been propounded by Gilbert and Sunshine is this notion of looking to specialized assets to define the market. I think that notion should be used for more than defining the market. I think that the access to specialized assets is also an appropriate way to define participants in the market, and that the one over end approach based on having access, possession of those specialized assets is an appropriate way to define market share to be much more realistic than looking to past sales or try to speculate as to the likely success of future innovations, which I would say in most situations is going to be impossible.
I also think that that use of specialized assets is an appropriate way to evaluate entry and I would like to come back to that. And I think if one decides not to use an innovation market approach, but to use a potential potential competition or future generation markets or even in applying more traditional product marketing analysis in industries where there's rapid innovation that the use of the identification of specialized assets that are critical to innovation is an appropriate factor to use for identifying participants there and predicting their likely future role.
I would again say that when looking at unilateral effects, it is important to apply a higher standard of certainty of outcome about the unilateral effects than one might in a pure price product market analysis. Because of the risk of overdeterrence of innovation and the lack -- the lessened ability we have to predict outcomes. As an outsider observing the various cases that have been brought, my perception is that a fairly rigorous standard has been applied, perhaps more in some than in others, but I think that that is important in GM/ZF.
My understanding from the public record is that there was a clear evidence indicating the parties' consideration of reducing innovation as a factor involved in the transaction. That combined with the fact that as I understand it again from the public record, the agency was looking at essentially only two companies in the world that posed the specialized assets necessary to conduct innovation leads to a conclusion that the evidentiary basis for seeing that innovation might be impeded is very strong in that case. I think that is the kind of record that should be required before a decision is made to go forward and challenge an innovation market. There should be a very clear story that the government can tell about the likely anti-competitive effect.
One additional thought on evaluating unilateral likely -- unilateral anti-competitive effects, I think it is important that the commission continue to consider the question of the nature of incentives to due process innovation as opposed to product innovation. In my review of the literature, which certainly was not complete, but there is -- there is not clear work that has been done on the issue as applied to antitrust analysis of whether the incentives for process innovation may be substantially stronger and less subject to market power causing a reduction than product innovation.
Unless innovation -- process innovation is very costly or renders installed equipment obsolete, it seems to me unlikely to be -- likely to be to a company's advantage to improve its produces and reduce its unit cost. Even if it has more of a power in downstream use markets, product market innovation is more likely to be disruptive to the exercise of market power and the activities to impede it. It's not clear to me that the strength of the incentives is the same for process innovations and it may be appropriate to limit concern of innovation markets to product innovation.
Finally let me touch just for a moment on entry. I think entry is a much more difficult question here. I would suggest two things to consider in evaluating whether the existing job pool for entry is appropriate. The first is that the focus should be upon the ability of new companies, new participants to acquire the specialized assets necessary for entry. That gives you an easy rule of thumb guideline for focus point in evaluating the entry. And if -- if it has been determined that there is an innovation market, I would agree it ought to be because specialized assets are required in this industry and they are in the hands of a limited few companies. That gives you easier approach to evaluating entry.
The evaluation of entry should also be focused on the acquisition of those specialized assets. The second point on the timeliness of entry, it seems to me that an extended period for evaluating entry may be appropriate here and my thinking is the following: The anti-competitive effects the reduction of the R&D. And under the guidelines analysis, we measure the two years from the anti-competitive effect of concern, but the anti-competitive effect here does not play out in the marketplace in terms of actual sales itself for perhaps a year, two years, three years. If entry begins to respond to the anti-competitive effect of restricting innovation, it may not come to the marketplace for even longer, but I would argue that it would have the deterrent effect on withholding innovation even though the innovation may not itself come to the marketplace for five years.
For example, I complete my merger. As a result of the merger I begin to deter innovation, my innovative product is going to get to the marketplace in three years. Maybe it would have gotten there in one or two, but as a result of having withheld innovation, if that is detected, an important issue that Mike Sohn raises, then the other company is going to set to work and begin its innovation. It may take it three years to get there or four years to get there, longer than the current guidelines period, but having an impact in the market within the two-year time frame from the point that the merged parties products get to market. So, I think in that circumstance, it may be appropriate to extend the analytical time period. And I think at that point I will stop.
Thank you.
COMMISSIONER VARNEY: Go ahead.
COMMISSIONER STEIGER: I hope that you and others perhaps would address a question since you have raised a world market for ideas, if you will, in innovation. For later discussion, would your belief that collective collusive behavior is unlikely in innovative markets be any less certain if you were analyzing joint venture or a merger involving a foreign partner given the alleged anti-stress, alleged history of cartel behavior within certain overseas industries?
COMMISSIONER VARNEY: And before we get to that answer, we would like to hear from Dr. Yao and then we will take a break and start our round table discussion.
DR. YAO: Thank you. Being last again, much of the good material is already taken.
COMMISSIONER VARNEY: You've got the home team advantage here.
DR. YAO: Well, that leaves me with all the controversial things because the only things I can talk about are things that I disagree with or things that somehow inadvertently slipped through, which I don't believe would have happened. So, let me spend a little bit of time going over some things that you have heard before, but hopefully with a little bit of a different spin.
Competitive effects from a merger obviously can include a reduction in the amount of R&D or reduction in the quality and diversity of R&D activity and then you would want to take that to a welfare effect. We are not done when you get to that reduction, but I will use that as a starting point for thinking about it, and these effects that we were talking about can occur through coordinated interaction or single firm behavior.
I agree with what has been said about the extreme difficulties in coordinating a reduction in R&D amongst firms in an industry, however I do have a caveat, actually a few caveats about this which might suggest some possibilities for coordination to actually occur. And you should take this as if I have just gone through a laundry list of why it won't, because that's really the starting point.
Okay, and I'm done with that laundry list and now I want to just poke at a few things which in particular cases might be worth at least checking out to see and if they lead you somewhere, well, okay, but I think they will have a hard time against some of the other points. Okay, the first is although R&D takes place in secret, it is also the case that the employees of these firms, the engineers and scientists and what not, particularly in I suppose places like Silicon Valley where they all go to Forty-niners games together or something like this, talk and they talk a lot about what they're working on. Maybe they shouldn't talk a lot about what they are working on, but I would be surprised if a fair amount of leakage did occur.
Now, of course that's an empirical question, that's a factual question. But if there's the case that there is a fair amount of leakage because of these professional networks, then it may be possible for firms to detect what other firms are doing and what is supposed to be pretty secret business. Now, it might be pretty hard to detect reductions in the amount or intensity, although that's possible, but it may be a little bit easier to detect the direction or research path or track that that company is pursuing.
So, I just sort of note that in thinking about whether to dismiss totally a possibility of coordinated interaction, one should at least think or pause and wonder whether there is some network that might exist for this particular industry, that in fact would make some of these problems less than you would initially think. Okay.
The other two -- well, actually the next one I think is also important and Judy Whalley had mentioned something along these lines. If there's a history of past coordination in the product market, however one could figure that out, let's see, there's a suspicion of this, or perhaps there was a case that had gone on before, then one might think that a vocabulary of coordination or perhaps some underlying understanding might make it easier to coordinate, and I think that's worth at least considering.
Finally, there may be in some perhaps fairly rare circumstances the adoption of some facilitating device that might promote coordination. I could imagine, for example, that some companies getting together to set a standard -- standards are very pro competitive in my view, okay, so again we go through the laundry list, A, B, C, D, E, F, G, but when you get to the bottom, you might ask yourself well, if you're making a standard, does that give you an opportunity to talk about some things which might help coordination.
Okay, again I should reinforce that I think most of the -- most, the vast majority of plausible competitive effects are likely to be out of the unilateral effects side, but because we haven't talked about these other things, I thought putting them out for discussion might be a good idea. And you know, these things ultimately become questions of fact and looking at the history and what's gone on. And it may be that when you look at this, it just isn't there. All right.
Now, looking at the unilateral effects side, in terms of a reduction in R&D activity, I think this morning and then again this afternoon we talk a lot about the weaknesses in -- or the inconclusiveness of the economic literature with respect to the relationship of concentration and the reductions of R&D, and then the next step whether these things are welfare producing. It is, of course, theoretically possible that some mergers will increase R&D and that could be good. By the same tone, it could be bad, according to what we have been talking about before. That might happen in particular if there are appropriability issues that are solved by the merger or perhaps some blocking patents or some other property rights issues are somehow removed as a result of the merger, and so those things need to be taken into account.
However, let's take it on an individual basis, it can also be, I think, the case that the overall amount of R&D might decrease and one might be able to figure that out from the facts. Having now participated enough of those factual inquiries, I have not much to add to that.
Now, I did want to talk a little bit about what I would call reduction in the quality or diversity of R&D activity. Mike Sohn mentioned something about the research tracks and I wanted to again bring I guess to the table a couple of points that I think are useful in thinking about what possible and competitive let's say welfare reducing effects might occur as a result of a merger in terms of R&D capabilities.
Now, obviously -- well, I don't know, maybe it's not obvious, to my way of thinking, if the reason that the merger is taking place is that they want to somehow match up some complementaries in R&D and other things and that looks pretty compelling, then I feel pretty comfortable.
However, if this is a merger that is not about that but is about some other -- for some other reason, then one wonders when you take the two R&D departments and do something with them, what will happen. They may not even be part of the overall strategic plan for this merger. And one of the things that could happen, of course, is that you take these R&D activities, maybe they were pursuing different tracks, they now might make a decision to eliminate one of the tracks. They might redirect them.
Now, that could be good, for all we know, that could be a greater focus, more intensity, more exchange of knowledge, better outcome. Better outcome for following that track, at least, but I want to suggest that the way companies may make these decisions, may actually reduce the diversity in a way that is not so good and it has to do with a lot of the internal incentives that are -- that go on within the firm. I would argue that there are direct and indirect pressures that might actually push different research tracks that might be contained within one company in the same direction.
Now, what might some of these things be? Well, one would be that the overall decision maker is making an overall decision over the entire set of R&D choices and there's something to be said for we want variance in R&D. Well, I guess we want a high need, but we want that upper tail, and one of the things in having a single decision maker might cause is some conformity or convergence in the kind of R&D that is pursued.
In addition to that, a firm usually alters its R&D in response to marketing, manufacturing, other issues that come in from other parts of the company. It's not just these guys in some room -- well, I hope it's not, thinking about trying to improve the product. And to the extent that these outside influences impact or should direct the direction of the tracks, these are going to be the same outside inputs for both of the now same tracks, let's say, if they're continuing two tracks in that firm. And that could lead to some convergence as well.
Another piece that goes along with that is to the extent that a company has a strategic mission or plan and sees a certain set of core competencies that it wants to take advantage of, the research tracks in their direction should respond to what these competencies are. Now that we have put two companies together, we now have a different set of competency, which means they may float towards the same direction, the research tracks might again float towards the same direction.
Okay, having said all of that, and I think there is a fairly strong literature and organizational theory and business strategy that will support a lot of this, for example there's a paper that I recently read by let's see, Rosenblum and Christensen that talked about how hard disk drive manufacturers pursued particular disk drives depending on who their customers happened to be. And the ones who had locked in the current generation customers tended to not try to make smaller disk drives that later became useful in like moving from minicomputers to PC's to laptops, and they were very much influenced by who they were serving at the time. So, that would be one of the kind of influences that I was talking about.
Having said all this, the natural question is okay, if managers -- now that I have said all this, managers probably know all this, so shouldn't they somehow design or organize their company in a way to avoid some of these problems and I think to some degree, companies have tried to do that. A lot of companies have tried to take units and basically isolate them from the influence of the rest of the company in order to avoid some of the things that I just said are negative. Of course a lot of companies haven't, a lot of companies are incapable of doing that.
So, I guess that comes down to again a question of how good are the managers and then there's that question of should the enforcers second guess that. But I think it's important to note that things are not that easy in the corporate world in terms of setting up incentives. And that when you are under one roof, you have a burden sometimes that makes it very hard to be equivalent in terms of diversity with two separate needs.
There are also some comments I wanted -- I guess I should just say that having said all of those sort of negative things, that obviously putting together R&D can lead to taking advantage of lots of complementaries that can well overwhelm the kind of things that I just talked about. Obviously reduction in duplication is very important and valuable.
Okay, the last thing, I just wanted to make a couple of comments about entry and then we can get to our discussion. I think with respect to entry, one of the problems with thinking about R&D and innovation markets with respect to entry again is the observeability question. It isn't that easy, perhaps, to observe what happens after a merger takes place. So, as a result, perhaps the firms that are considering entering, they have nothing more than this knowledge of these firms merge and maybe that's not enough to go ahead with. A second problem related to the first is let's assume a company did enter -- there's some questions sometimes as to whether the merging companies even knew that they decided to enter. A lot of this depends upon how much information is flowing back and forth, but I think it compounds that the issue of trying to -- to analyze entry before R&D markets.
I just stand by again reinforcing I think the basic point, which is unobserveability seems to be the key characteristics and this unobserveability means despite a lot of things that I said earlier, coordination among firms in an industry is very difficult, and so I would think that most of the vast majority of the problems would occur having to do with unilateral tests.
COMMISSIONER VARNEY: Okay, thank you, Dr. Yao. I think we will take a short ten-minute break, and when we come back, perhaps Professor Rapp and Carlton would like to make some comments on what they heard this morning or this afternoon and we will go from there.
(Pause in the proceedings.)
COMMISSIONER VARNEY: Dr. Rapp?
DR. RAPP: These effects are visible, the uncertainty associated with interferant with the research process when you are already down the road in stage three clinical trials or something like that, it seemed to me as though a much lower risk. My advice further is don't use R&D cutbacks as a synonym for anti-competitive effects. Don't count too much on specialized assets, unless you can make that term less elastic, not in an economist's sense of the word, but it strikes me that there is a -- it was a useful, an important element in the Gilbert and Sunshine formalization of the -- of the innovation market approach, but because it is not rigorously defined, there is a danger that a specialized asset can be found everywhere. Nobody knows at this moment whether in the next drug manufacturer down the line a two -- a lab with a two-year lead time or a six-month lead time could not be defined as a specialized asset which if the implication is that everybody else doesn't have one and that's an entry barrier, could lead in my view to just the wrong outcome.
So, I think we have to be careful in our reliance on that. If you add those considerations that were discussed in this afternoon's sessions, with my continuing fear of the -- of the danger of false positives that arise in the innovation market approach, then my advice remains that the innovation market approach as we now understand it, a parallel inquiry very similar to the inquiry that we observed in product markets where a small but significant reduction in R&D -- a small but significant increase in price to define product markets is substituted for a small but significant reduction in R&D to define innovation market.
I must persist in my point of view that that's an approach that should be abandoned. But I think the search for -- for competitive effects in correctly defined goods markets where innovation or R&D is involved should continue because fundamentalists I do agree with Professor Gilbert and others that dynamic efficiency is an important, possibly even more important than allocated efficiency and that it is -- it's appropriate that these should be watched in the context of competition policy.
Thank you.
COMMISSIONER VARNEY: Thank you, Dr. Rapp. Professor Carlton?
PROFESSOR CARLTON: I will try to keep my remarks brief because most of what I had to say I said this morning and the additional material I wanted to say Dr. Rapp eloquently stated. My basic point this morning was don't be too quick to trade off certain benefits from an efficiency enhancing merger in order to get speculative future gains based on R&D and innovation markets that hopefully will lead to new products. That doesn't mean I think that R&D is not important, or that if you can identify such cases, you shouldn't try to go after them, but that the potential competition doctrine struck me as a more reliable way to do it.
Hypotheticals are very easy to construct, in which noncompeting firms merge, yet consumers in the United States are harmed. The examples that are used like that hypothetical to justify innovation markets in which U.S. consumers get harmed are just that, a hypothetical. They have nothing whatsoever from a logical point of view to do with innovation markets. I could dream up 100 such hypotheticals having nothing to do with innovation and simply mergers are occurring between noncompeting firms and for a variety of reasons, perhaps the firms get re-organized internally or the like, consumers are harmed in a particular market, even though there's no competition between the merging firms. I don't think I would formulate antitrust policy on the basis of hypothetical examples.
What I keep stressing is you must show that these are realistic examples that arise regularly enough that they can be reliably predicted. Otherwise, we will be back to a situation in which everybody will have to examine every case "carefully," and if I find the exceptional case in which, even though it doesn't look like these firms are competing, I can think of some way because Harry is no longer in charge of the production line that costs are going up, I am going to enjoin this merger, that strikes fear in my heart because I don't know a principled way to do that examination reliably and that is a fear I have.
The notion that there are some cases as I mentioned this morning in which you can identify perhaps -- and I said rape cases -- who are going to be the firms who are going to be competing in the future, because, for example, the drug industry, I think I gave an example because there are special -- maybe you could call that specialized assets.
I want to underscore something that both Dr. Rapp said and Judy Whalley. It would be hard to define what you mean by a specialized asset, although Judy mentioned that in the ZF case, it looked to her by what was publicly available that those were specialized assets. I can assure you from the point of view of ZF and GM, they did not concur in that opinion, and as I said this morning, I think the best test of the value of the doctrine in that case is that right now, two years have passed, we have developed no new products, I hope we can continue watching this industry to see if new products come out. If they don't, we have given up a lot of years of benefit of not much pay-off.
That leads me to another point again, I think it was maybe Judy or maybe Mike mentioned is that one of the dangers of a new antitrust doctrine is its overuse, and that is a concern that I am worried about because I would think that businesses that are merging that have R&D are going to be worried what in the world is going to be thrown at them, how can they tell if they have a problem.
So, I thoroughly endorse some notion of safe harbors and again, because collusion is less of a problem than in an ordinary product market, you might want to make the safe harbors very safe. The point that this raises is exactly what evidence you look at when you look at these cases. If you cannot -- don't have a long enough history for an industry that you can really get a sense as to what's going on, what you often are left with is going through the documents.
Now, I don't want to demean going through the documents, but I am worried about an analysis that only relies on some memos that might have been written by what some person thinks they might do in the future. As I think Mike said, people always get overzealous, perhaps on both sides, and I would be worried if that was all the evidence we had to rely on. And I'm sure that if that became the evidence we relied on, pretty quickly an antitrust council for these firms would make sure that the right memos either weren't written or were written.
COMMISSIONER VARNEY: Would that they could.
PROFESSOR CARLTON: And that worries me an awful lot. So, I guess I would just summarize by saying I think we should be concerned about R&D, but we should be especially concerned about our inability to reliably predict when something will harm competition and let's not give up a bird in the hand for two in the bush.
COMMISSIONER VARNEY: Okay, I have just I think a couple of very brief questions, it might even be one-word answers. Mike, on the unilateral effects, could we do a potential competition analysis for lying on the horizontal merger guidelines when you've got two?
MR. SOHN: The answer has got to be semantics. I'm convinced that there probably is a set of cases where at least as they're written the guidelines don't fit comfortably. I can imagine a situation where several years from a product, but the goal line is clearly enough to be defined and I wonder whether -- I mean that doesn't mean that I would be concerned about it, actually special circumstances, but I think it's a sufficient position to take that the guidelines don't fit comfortably there as well.
COMMISSIONER VARNEY: Richard, you looked at this argument and you are familiar with the thought of the way we ought to be looking at barriers to entry is not an arbitrary two-year, whatever period, but more the product life and what are your thoughts on the product generation?
MR. GILBERT: I am vaguely familiar with it. I know that in the work that Teece and others, Hartman have done, that they like to think about various product life cycles, for example, medical diagnostics, medical equipment, and out comes something new and peaks, and then another generation passes by it. I don't see anything fundamentally inconsistent with viewing the world that way and then also viewing it, I don't see how that is inconsistent with the merger guidelines view of the world.
I will take the opportunity, though, just to make one main clarification. I don't -- I know a number of people react to innovation markets analysis as being a -- an extreme view of we have to be concerned about the hazards of concentration and the activities of research and development, and I don't view it this way. I view it as a framework for evaluating competitive effects in innovation, not a framework for reaching conclusions that certain concentrations are necessarily anti-competitive.
And moreover, I think when you think about research and development and you think about it in a dynamic competition in a more dynamic way, I think you must simultaneously look at the entry side of the picture and whether you think about that in terms of product life cycles or just the likelihood of spontaneous entry or drastic entry, the entry of a major competitor that really upsets the industry, it leads me to say that maybe our product market boundaries are fuzzier than we thought they were. Maybe the linkage between present pricing and non-price aspects of competition in the product market today is better governed by the likelihood of entry in two or three years. I think there are firms out there, I don't have much doubt that there are firms out there that despite being in a concentrated market are very concerned that if they don't maintain the pace of research and development, that a new discovery will come along, a break-through technology will come along and will eliminate them. Even if it goes beyond the two-year guidelines test, and that effects their present behavior, so I view this as both sides that you have -- if you can look at innovation and product concentration side, you also need to look at innovation as a deconcentrating effect that can neutralize market competition -- concentration as well.
COMMISSIONER VARNEY: Okay. Debra, you had a follow up?
MS. VALENTINE: I actually wanted to follow up on that with you, Rich and Judy, because obviously some of the reasons -- well, not obviously, I think each of you were telling us that there were reasons for thinking about extending the entry time frame with innovation markets. You because of this drastic change and Judy because of the theory of products not coming to market for a while and the timeliness and sufficiency wouldn't nonetheless take place somehow before we actually had some impact on competition in the market. What we were hearing yesterday was a proposal for a four-year time frame in current generation markets where there -- where the markets nonetheless are ones characterized by a fair degree of change, high technology, innovation and churning. And I guess the real question is would you make the same arguments for current generation markets like that or do your arguments that we've heard today for innovation markets apply to those current generation product markets?
MS. WHALLEY: I think it's important to go back to the underlying reason that we're concerned about entry, and that is because the likelihood of entry is going to deter the competitive effects occurring or erode any anti-competitive effects that do occur. And picking two years is really someone's idea of this is a number that purports to represent the point at which entry will deter the anti-competitive effect. I think even if the guidelines as written, they say that in specific markets conditions may cause you to vary that amount to more accurately represent at what point entry coming in would have the appropriate deterrent effect.
So, I think it's appropriate to look at this issue and say in a particular market, innovation markets I do think it may be appropriately longer to use a longer time period. What I heard being said yesterday I thought was something very different, which I don't agree with, which is that you ought to look to the -- maybe I'm not going to accurately paraphrase this, but what I was hearing was you ought to look to the life cycle time because it is the life cycle that indicates when competition will take place. And if in a particular industry it takes four years or eight years, but that's when competition takes place, that's what your period for entry ought to be because now you're accurately reflecting competition. I don't see a connection of that with our reason for being concerned about entry. And that's why I'm not comfortable with that approach at all. But I do think adjustments that reflect the ability of entry to deter anti-competitive effects is appropriate.
COMMISSIONER VALENTINE: I would very much agree that your theory would still perhaps deter. Okay.
MR. GILBERT: I have a real problem, but I'm going to agree with what Judy says here, but I have a real problem with the idea that we should not be concerned about anything that happens if there's going to be entry after four years. And at the same time, we shouldn't be concerned about any anti-competitive effects that might happen after five years because it's too speculative leaving us with a one-year window from four to five in which to evaluate all possible anti-competitive antitrust policies or all antitrust policies.
Now, I agree -- I think you stated very well, Judy, that the merger guidelines can be applied in a flexible manner here. There are some industries for which the possibility of drastic entry three or four years from now is a serious disciplining effect on their current behavior. There are other industries who could probably care less about the prospect of that entry. It's not going to change their behavior one whit.
And if there isn't any linkage between those two, then that disciplining entry, that entry isn't disciplining anything in the short run and there's just now a present value calculation of do you care what happens over the next four intervening years, and I think one should care what happens over the next four years. Even if that's not just because that's a convenient increment for political life cycles, or five years or eight years.
So, it really is a question of whether or not there's a linkage between the likelihood of entry, the magnitude of that entry and its effect on present pricing and no price innovation decisions.
MS. WHALLEY: Can I ask a follow-up question? Rich, what do you see as the connection between using specialized assets to identify participants in your market and the ability of firms to predict drastic entry such that it would have a chilling effect on their decisions to innovate or how much to innovate?
MR. GILBERT: They're clearly related, I mean this is -- there's no magic cures here, so it's not a magic formula that just because you can recite specialized instances that you now know exactly what the contours of entry of competition are going to be in the industry, and I think I agree with Dick's concerns about that.
To the extent that's really what specialized assets are, the bottom line of specialized assets just says that there are some firms that are competitively advantaged due to those assets, and they can't change very much that situation, they can't change very much -- that situation cannot change very much over a reasonable time frame, that is the acquisition of those assets is difficult over say a one or two or three-year time frame. There can always be surprises, particularly in more basic research.
One comment that we kept skirting around this morning, and I don't know if we focused on it or communicated, which is that much of innovation competitive analysis is likely to be relevant in industries where the innovation is incremental. The break-through technologies are probably very hard to predict. You don't know if the next semiconductor measurement technology will come from x-rays or from optic or from some sort of chemical process. You don't know which one is going to be the next break-through. But it might be easier to predict that someone who is now using optics technology might be able improve that optics technology for the next generation.
So, I think the specialized assets are more useful to identify those short-run incremental improvements than they would be for the real break-throughs, who knows where they are going to come, but break-throughs don't happen all that often either.
PROFESSOR CARLTON: I just want to add that the issue about the time period and the effect of entry depends on the termed linkages on the demand side is perhaps the supply side, but on the demand side anyway is interval linkages for durable goods versus nondurable goods. And one of the things you want to look at is the size of the market changes. If every four years there's a big competition and that's when it was and then in the intervening years nothing happens, then obviously it's every four years you want to look at.
And for durable goods, since you can delay your consumption of the item, that may allow a durable good market to -- in a durable good market entry to have an effect more immediately than in the nondurable good market, which was really Richard's point. If you have monopoly power for four years, then somebody comes in, who cares if you are not four years getting the monopoly power.
COMMISSIONER VARNEY: Okay, notwithstanding your emanations about hypotheticals, one of the things that we are going to do is our policy planning staff here is responsible at least in the first instance for pulling together these hearings in a way that makes sense for the commissioner to think about what we ought to be doing down the road, either in regards to policy or other recommendations. And it's very helpful for us for staff particularly to be able to pose series of hypotheticals and have specific questions. Not to, you know, come up with the ultimate hypothetical that disproves you, but to really get a sense of where everybody is on these issues and to try to figure out where there's consensus and where there's not.
So, what I would like to do now is turn to our director of policy and planning, Susan DeSanti, and ask her to run us through some basis on her staff on some of the issues that they've got to get some concrete recommendations for. For those of you who went to Catholic school, on Friday the nuns always brought the priest in to answer the really tough questions and you try to think up the question that the priest couldn't answer. My personal favorite being can God make a rock so big that he can't lift it. So, that was up there with the wonderful, my other all time favorite, the international date line question. You have to receive communion once a week. You haven't received communion, you're on a boat, it's one time, you cross the international date line, boom.
MS. DeSANTI: Commissioner Varney, that is a perfect introduction, because in the sense that it is exactly in those specific factual situations that the hard questions come to pass and the rubber hits the road.
COMMISSIONER VARNEY: There's no priest.
MS. DeSANTI: Yes, there is no priest.
PROFESSOR CARLTON: Certainly not here.
MR. BAKER: That's why God created antitrust commissions.
MS. DeSANTI: That's John's version at least.
MR. GILBERT: He gathers those rocks.
MS. DeSANTI: We can make a rock large enough.
MR. BAKER: We will have a minimum on that on Thursday.
MS. DeSANTI: And we will have ours on Tuesday before.
And what I wanted to do was ask some people on my staff and also Ann Malester, who has been involved in several of these types of cases, to put some factual hypotheticals out. I think it's -- one thing I would like to say, I think it can be very frustrating for people inside the agency and for people outside of the agency to have these discussions because there are other facts that make -- that staff and commission members may be privy to do that wouldn't make a difference to how you all think about these issues that we can't share with you. And similarly, you have experiences that you can't share with us.
And that's why I want to assure you, Dennis, that many of our hypotheticals are not hypothetical in some sense or other. They may be facts that we are transferring to different types of situations, but I was struck yesterday by the number of people who were saying that well, pharmaceuticals is really a unique, different, unusual, rare type of situation. That may be true when you are thinking about the entire world of possibilities that you may be presented with, but this commission has been presented with a number of cases that raised issues in those markets.
So, we really have had a lot of time to wrestle with some intricate and difficult facts. But with that --
MR. SOHN: Can I just comment on what Susan said because it certainly resonated to me from the standpoint of practicing before you, particularly in innovation markets, you know, the staff says well, we have called the others in the industry and your client and the merging firm are one and two in the race to innovate.
Now, how do I verify that? Should I go back to my client and say I don't think the firm I'm buying is fifth. And I don't think research is worth very much and that's an objective view. And they'll give you chapter and verse on why it is not and why from what they can tell when they go to see what you and other people are doing in the clinics and so forth, they think that it's five and not three. This is a problem that doesn't occur in a goods market, at least not in that form.
So, it does create, it does create, I think, grounds in which people get unduly cynical, which comes back to my thought that if you had some clear, as clear as can be, enforcement policy statement out there, people would have greater confidence in the process even though we still may not be able to talk freely about the facts.
MS. DeSANTI: I think that's a very good point, and I -- what I would like to do is run through a few hypotheticals with different people and then go through some issues specifically that -- and just determine among all of you where there is consensus and where there's disagreement, and if there's disagreement, why, in terms of possible consideration of what kinds of factors could be listed as relevant to these kinds of issues. And I also wanted to say that, you know, maybe we can get some customer point of views -- points of view in the discussion. And maybe we can start with Ann Malester since she's done some of these cases.
COMMISSIONER VARNEY: Can I ask Mike one question, before you jump in there. It's something you brought up that specific example because that's something that I have had a lot of trouble with, too, but I think you can go a couple of different ways here based on some of the volume of argument that we have heard in the last two days.
One is to say that you really can't measure this stuff, so, you know, be careful when you tread in there. The other is to say well, what's the best way to measure it. And it seems to me, and I may be completely wrong here, that there are a diversity of sources of information when you are looking at who's number one, who's number two, three, in the industry, in an industry in R&D. One is the companies themselves and their internal documents, and often times, you know, we will find a lot of documents that says this guy is number two, we need to go get them. Other times we will find other company's documents where they are pretransaction talking about what's going on in the industry in their estimation, just their opinion.
It also seems to me in the industry like biotech, we have financial analysts, I mean more than you can count, both on the west coast and the east coast, you have got people who do nothing but watch this industry for a living and have some pretty good -- or not good, but pretty solid views on who the industry leaders are.
So, it's not in R&D and in specific paths. You've also got clinical trials and we have access sometimes to FDA or other documentation on where they think, so I guess my question is yeah, it's really hard to get a handle on this and we shouldn't over put -- put over reliances, I think that we sometimes do, on competitors staying on any track. I think just that sometimes it's easier to rely on that than it is to keep going on out, but is the implication of what you're saying it really can't be done with any -- it usually can't be done with any degree of certainty or the way you are doing it -- the way it appears that we are doing it right now is just is too dicey?
MR. SOHN: That's an excellent question. No, I don't reach the conclusion that it can't be done. To me it teaches humility and a need to really go only for the very clearest cases, which is the point that Judy and others have made. In terms of reliance on the various sources that you can rely on, financial analysts, I personally have always been cynical about that because what they know they hear from the various firms. You have one eye on the stock market and the likelihood that the financial analysts will hear what the firm wants the market to hear is considerable. So, I would be careful about that source, although it might be slightly better than a competitor. I think in a defense industry context, particularly where DOD is running an R&D program and critiquing people, pretransaction, as to -- as to where they stand, if the Defense Department who after all is the customer takes you through why it ranked merging firms one and two or one and five, that strikes me as something I would put some agreements on.
COMMISSIONER VARNEY: That's a perfect subway to end.
MS. MALESTER: This morning and sometimes this afternoon, also, people were talking a lot about different potential anti-competitive effects and I wanted to try to talk about one and get some reaction about one that I really didn't hear any discussion of, which is the effect on prices when the goods reach the market. And just to take the hypothetical that Mike Sohn brought up, let's just assume we have three companies researching for a new vaccine, and two of them plan to merge and Dennis and Mike were talking about the pluses and minuses of taking one research track and eliminating it.
And clearly there is some debate about that might be good, that might be bad, but I'm wondering why there didn't seem to be a concern about the fact that when these vaccines reach the market, having three firms selling them rather than two is likely to make the prices of those vaccines lower. And if people on the panel feel that is a competitive effect we should be concerned about, is that an actual competition case, a potential competition case or an innovation market competition case?
PROFESSOR CARLTON: My view is first I think that's a very good question, in general I am not opposed to hypothetical questions, I think they are a good way of testing your theories, but what I think I was reacting to was maybe the use of the hypotheticals ZF/GM case where we assumed that ZF was not in the United States when it was, in fact, and I think those type of hypotheticals aren't -- they truly are hypotheticals and they give an ear that is an empirical reality to a hypothetical when there's not.
Maybe I was being overly defensive because I was involved with that case, but I think hypotheticals are a good way to proceed and I think the question you asked is the right one. I think that the effect on prices is precisely the competitive effect you are concerned about, holding constant the quality of the good and the time the good comes on the market, I am very concerned with the number of firms that are competing and whether that has an effect on the price.
And if I feel that the price would be lower, that would be something that I would take into account. I think the way I would handle that, as I described this morning, is rather than be focusing on the reduction in R&D, I would instead be focusing on the reduced output and higher price that occurs in their future market. That logically seems to me much closer to a potential competition doctrine in which I am worried about future firms competing in the future. The fact that it's in the future in a future product that doesn't exist today, I understand maybe a distinction with how we usually think about the potential competition doctrine, but to me as a logical matter it shouldn't be a principal to distinction, the logic is exactly the same. Competition in the future among a larger number of firms will result in lower prices. As long as I can reliably predict that this product will occur in the future, it seems to me the potential competition doctrine or the minor extension I have given to it would be the right approach.
MR. SOHN: I agree with all of that. I would just add that the less clarity there is as to whether there's ever going to be a goods market, the less we would worry about a price effect in that goods market. And in this hypothetical vaccine case, people have been looking for many vaccines for many, many years, and lots of promising R&D programs have not proved out in clinical. So, I think you need to factor in that difference.
COMMISSIONER VARNEY: Judy?
MS. WHALLEY: I guess I would be concerned about both aspects in that hypothetical. I would definitely be concerned about the price effects down the line of having three companies in the market rather than two. I think that there ought to be room in the law for what I sort of cavalierly call the potential potential competition case. I think it's very important in that type of case that some of the rigor of the potential competition doctrine not be lost in terms of identifying, you know, that these are the companies that are the most likely potential entrants and their reasons why others would not quickly and easily enter. And I find this specialized asset approach one way to do that. But I think I would also be concerned about the potential for loss of innovation -- innovation paths in there and that there could be an effect, not just a price effect, but also again depending on the relative position of the third company and its ability to sort of absorb and take the place of the company that's being lost, I could imagine innovation effects in that market, too.
PROFESSOR CARLTON: Excuse me, Judy, would that be different than a potential competition effect in a market? In other words, you are talking about the quality of the good and, you know, it's a nonpriced competition among the three.
MS. WHALLEY: I think that --
PROFESSOR CARLTON: I don't know if I am allowed to ask questions. I'm sorry.
COMMISSIONER VARNEY: Yes.
PROFESSOR CARLTON: In Chicago everybody asks questions all the time.
COMMISSIONER VARNEY: Only if they are not hypothetical questions.
MS. WHALLEY: I do think that part of this is a semantic issue. I mean, I said yesterday, and I feel this really strongly, the important thing is to devise the analytical tool for figuring out if there is a problem.
PROFESSOR CARLTON: Right.
MS. WHALLEY: And what the criteria for applying that are and to make sure that they are rigorous enough to exclude false positives. And what you call it, whether it's potential competition case or innovation market case, it seems to me it doesn't make a lot of difference except as to the theory of the courts which seems to want to move in increments and not drastic leaps of innovation.
COMMISSIONER VARNEY: Rich?
MR. GILBERT: One way to think about this hypothetical is that we have markets for existing goods, we have markets for future goods, and then we have the process of getting from here to there.
When Dennis responded, I think the panelists so far have made very good, very important points, but Dennis said I am concerned about competition in this future goods market holding the timing of that product market and the quality of those products constant. Well, what about the effects of competition in changing the time at which that product might be available or the quality of the products that might be available. It's difficult to predict, but I think Dennis would agree that it's a factor that goes into this analysis that's not just a competition of the future product markets. And suppose I change, I don't know if we are allowed to do this, but suppose I change the hypothetical a little bit.
COMMISSIONER VARNEY: You must be from Berkeley to do that.
MR. GILBERT: Let's suppose that there's going to be a patent for the vaccine and it's established that there's only going to be one winner, in which case we know what the future product market concentration is, it's a monopoly in this -- well, it's a single firm in this product class. That might not be a relevant product monopoly, of course, but a single firm.
Now, there might still be an issue, though, of is it going to be developed this year or next year or the year after or the year after that. So, it's certainly an important issue, and I would agree with the analysis that says you want to look and be concerned about product market competition in the future product market, but I would also be concerned about when you get the future product as well.
MS. DeSANTI: I would like to change the hypothetical so that one version is Richard's change, it's a winner-take- all based on you know there is going to be a monopolist at the end of it, and then a second possibility that you have a number of customers coming in to complain about the likely effects of the merger of the R&D efforts on product diversity. And there are customers coming in and saying well, you know, these are different R&D tracks and typically what comes out of this process are treatments that may have different degrees of efficaciousness for different patients. And we're concerned with product diversity. So, I would be interested in responses on both of those issues.
Dennis, you can start.
COMMISSIONER VARNEY: Let's go to Dr. Rapp next.
DR. RAPP: I think my instant reaction is that the assumption of customers who can foresee the outcome in a conventional therapeutic goods market means at least in my perhaps crabbed interpretation of things that this is a pure merger guidelines issue and that while it might require a little innovation and interpretation to -- in the writing of the complaint or something like that, but that there's no default to the overall mechanics of the innovation market approach. Unless I'm missing something, I mean, I just see that the insertion of the customers and the relatively easy prediction of the outcome once the approvals have been granted and so forth makes this -- puts this in the conventional analysis camp as far as I can tell.
COMMISSIONER VARNEY: How about for the original competition that Ann posted, no customers, no monopolist, are we in an innovation market or are we in a potential competition?
DR. RAPP: Well, if we are in an innovation market, I'm nervous about it in other words.
COMMISSIONER VARNEY: We knew that.
DR. RAPP: Nothing new there, I admit. What I'm saying is we come back to the issue of timing. The further we are to have the modified version, the further we are from an actual goods market in months ought to be a measure of the humility that is required to even look into the -- to make the inquiry because what it means is that the odds of doing a bad thing come closer to 50 percent, the further you are from that condition.
MS. DeSANTI: I would like to pursue that a little bit farther with you. Part of the reason I'm asking about the customers is we talked some this morning about the degree to which economic theory gives us a basis for being concerned about competitive effects. We talked some about the degree to which there's empirical evidence, I'm interested in whether as a conceptual matter you would have any -- you know, regardless of whether it's pursued under potential competition or innovation market theory, is there any reason that you would think that that's not something that the antitrust agencies should take a look at.
DR. RAPP: As a matter of theory, no. In other words, if I understand the question correctly, I don't think that there is any reason to absolutely rule that out.
MR. SOHN: I'm sorry, Dennis is going.
DR. YAO: Let me respond to the initial hypothetical, I guess I agree with Rich and Judy that one should look at both the future product market, the price effect in that market as well as the innovation market, but let me add a little twist here. Suppose you look at this and it looks like the price is going to go up in the future product market. I suppose the parties are going to come in and argue about efficiencies, they're going to say look, we're going to get together, there's going to be some complementary -- complementaries and maybe we're going to get this thing sooner, maybe it's going to be better.
Now, I ask you, how can you avoid dealing with the innovation market if this is the argument that they're using to sort of come back? And it may be the case that you can't completely dismiss the innovation market part from the product market part because the parties may, in fact, make the argument that pulls you into making an assessment about the effectiveness of R&D. Of course, it could be a little more clever and come in and talk about distribution and manufacturing differences in order to offset this price effect, but I wouldn't be surprised if you would hear a big story about complementaries. And so I think you might be pushed into this anyway.
PROFESSOR CARLTON: I think, though, that that depends on what you define to be the innovation effect. If you define that in innovation market, if you define that to be examining the efficiencies of R&D, again I don't want to get -- whether we call that, you know, innovation market or inefficiency in input I think is a detail. When you distinguish between the innovation market approach and the more traditional say potential competition approach, I distinguish between whether you know what product are likely to be coming out of the R&D process from cases in which you do not and you have to be making some sort of applications.
I think there's a theoretical matter to get to something you asked, Susan, all the economists and lawyers here would probably agree that hypothetically you can deal with slightly every case if you give me discount rates and probabilities of assessment. I think the hard question is when you don't have products that can be reliably predicted from the R&D, what should you do. That's a hard question. And I think some of us have come out on the case that it's too speculative to pursue while others have said well, maybe there are some exceptions that we can go forward at. But I don't think say the original hypothetical you posed, I don't think -- I think you made it simple enough so that you were keeping constant the time and quality and just were asking about the price effect.
MS. MALESTER: And in this case there was a specific product that we were looking at. But let's assume for a second to shift to make it a little more complicated, and I would ask you to look at that first. Let's take, for example, the defense industry that we have talked about a little bit and assume that hypothetically there are three companies that have put in a lot of their own funds and have received a lot of Defense Department funds to develop stealth technology, and at the moment there is nothing yet in the defense budget for purposes of building a stealth fighter aircraft or a stealth bomber, but we're assuming that all of this funding has been done not for the fun of it but because the Defense Department hears at some point it is going to need a fighter bomber that has stealth capabilities, and these three firms, two of them decide to merge, today.
How would you analyze that? Do you say because we don't have a product we can identify today, that that should be the end of the inquiry?
PROFESSOR CARLTON: No, I wouldn't say that. I think I might say that, you know, it's like a produce-to-order industry. Just because there's not a product that's specified because the government hasn't asked for it yet doesn't mean that once they ask for it that, you know, people wouldn't bid for it. In other words, I think there's a case for bidding where people have the capability, you can define a product.
It's kind of like let's suppose there are three builders of homes. Well, I haven't yet specified my home, but I still would say that they are in competition for the building of the home that I am going to specify next year. So, that doesn't -- that's not -- the case I'm worried about where I knew who has the capability to bid on the project and I am going to specify, what I am worried about is when people are doing R&D and it's very hard for me to predict what products are going to be coming out of that process, at that stage of uncertainty, I am worried about using innovation markets to stop a merger because I would be quite unreliable in my predictions. That's my concern.
MS. MALESTER: In the hypothetical I posed and the stealth technology one, are you still comfortable calling that a potential competition case or an --
PROFESSOR CARLTON: It strikes me as potential competition in the future for a product the government will specify, and that when the government specifies the product, these three firms would bid. And if there are only two of those firms who are bidding, then that would be a reduction in the number of people bidding for that product. Whether that's significant or not would be a separate question, but that's the question you face all the time where three going to two is a problem.
MS. MALESTER: And is there any problem with calling that a reduction in actual competition in the development of stealth technology?
PROFESSOR CARLTON: Well, as I understand the way you set it up, it would be a reduction in competition for the number of participants who would bid to develop, you know, the next product the government specifies. I don't know, you know, I don't know if it matters semantically what we call it. I think in my view the distinction between what I think is a minor extension in my mind of the potential competition doctrine to cover future products that did not exist today. That I understand. That seems much different than the innovation approach which is looking at R&D at this time when the products have -- are unspecifiable. That's the distinction I make between the two.
MR. SOHN: Excuse me --
MR. BAKER: Excuse me, may I just respond to that? Do you see a difference in the difficulty in defining the product market in your two ways of getting at this? That is what exactly is the product market bounds for five products that have not yet been specified by the Defense Department? Don't we have an easier time thinking about what the product would be if we formed it in the innovation market context?
PROFESSOR CARLTON: I thought the question was the government budget hasn't been specified. We know in one year they are going to need a new product, we just don't know how to describe what's in this new airplane and that whatever they put in it, three firms are the ones who are going to bid on it. That doesn't seem like a problem I would have difficulty dealing with under a potential competition doctrine or the notions that I would associate with that. I don't think I have to go to an innovation approach for that.
COMMISSIONER VALENTINE: Can I in that context ask the two litigators who we finally have now since all morning the economists -- well, several of the economists were taking the argument that you have just heard Dennis make, that when there is a nearly or likely predictable future product it may be best to resort to an extension of the potential competition? If you went into court on Ann's case, let's say you have to do Ann's side of the hypothetical now as opposed to your potential client's side, would you rather argue a potential competition theory or an innovation market theory in her stealth hypothetical?
MR. SOHN: Knowing what I know about potential competition precedents, I think I would try something new. I mean that quite seriously. I think you are going to set up a lot of hurdles that in the context of the last several shifts in Ann's hypothetical would be difficult to establish, although the case where the product is defined would be the easiest under potential competition doctrine. I think at the end of the day if the motivation for trying to turn us into potential competition is so you can come within the rubric of is that a line I would rather set out or is that a line of exercise, I would rather set out what I was concerned about, define it as an innovation market. I think -- I would rather think that you come out better on that rather than trying to cram this into the potential competition precedents. And I leave to you to decide whether the precedents are meritorious or not.
If I could make one comment on what I think is the last of Ann's hypotheticals, whereas I understood it the program hasn't been defined or announced but there's been some preliminary R&D funding and it appears to be three going to two, it strikes me in that hypothetical, I would be a little cherry about concluding that there aren't other defense firms with the requisite skill sets who will come out of the woodwork given the scarcity of defense projects generally once that program is announced. So, I will be on the competitive effects side and particularly on the entry side very alert to whether you really have a situation in which three are going to two.
COMMISSIONER VARNEY: Judy?
MS. WHALLEY: On the point of would you rather go in on potential competition or this innovation market, I think I agree with Mike. I haven't thought a whole lot about it, but I think you are better off arguing the direct competition today to innovate and trying to address the line of customers question by talking about the ultimate effect in in-use product markets that should bring you within the rubric of section seven. I don't know whether you can succeed with that or not.
I mean, I think it would be something to see if you wind up litigating one of these cases what the court's acceptance of it is. But I would agree that you are better off trying it by just laying out the problem and arguing that that's a current competitive problem rather than this potential competition in the future market.
Can I make a point back on your point about customers and should the commission be concerned if they hear from customers who are concerned about the loss of research approaches. Clearly that's something that should be of concern, but I'm probably preaching to the choir here, but I think it's also important to be very careful about that because it's easier for customers to have, for lack of a better word, simplistic reaction to looking at the fact that company A was doing this approach, and company B was doing that approach, and gee, wouldn't it be great if we had both approaches and get to market, because the efficacy of the product might be slightly different for different customer masses.
To the extent that's true, it seems to me in most circumstances, not all, but most circumstances, the incentives of the companies are going to be to continue both approaches. And a concern about the loss of that on the part of the customers is just the indicator for further investigation as to what the incentives of the companies would be. I mean, it may be that the cost of conducting both outweigh the potential sales to different customer groups that have different benefits from the drugs, but that's certainly not a priority of the case. It may well be that it would be worth it for the companies that they would have strong incentives to continue both research approaches.
MR. SOHN: Could I add a point on that just briefly. I agree with what Judy said about being careful about which customers you listen to and some having more simplistic approach than others. I would give more credence to sophisticated customers judged by objective standards, to return to the defense industry for a moment, to me it's useful to know in a given case how the Defense Department has allocated its R&D firms -- funds rather. It's not a limited part for any particular R&D development program, do you observe that more often than not they would rather have the firm fund the two firms with what they judge from a distance to the product to be the most promising programs rather than split that money five or six ways in the interest of diversity. I think my sense of it is that it's more likely to be for the manufacture of the former than the latter, but you would have closer observation.
MS. DeSANTI: What I would like to do now is shift the discussion away some from potential competition versus innovation markets. I think we have heard a lot on that issue, but I would like to explore in the limited amount of time that we have left some of the key issues that I think have been raised. Mike, I think your safe harbor point is one of them. Several people have spoken to the issue of whether economics indicates there's any likely anti-competitive effect in particular situations. I would like to clarify some of those issues as well.
And finally, I would like to deal with the extent to which people are saying coordinated interaction is unlikely various saying coordinated interaction is impossible and I would like some clarification on those issues. Maybe we'll take them in reverse order. The coordinated interaction issue is very interesting, especially in light of for someone who has sat here for a couple of days and to my surprise had industry people come in and say well, some industries, you know, R&D is really observeable, everybody knows what's going on in the R&D.
Suppose, assume for a moment, we'll take the Dennis Carlton approach, assume for a moment that that is actually the case, that R&D is observeable. Then it seems to me there's less of an issue about whether cheating can be observed, and maybe this is a simplistic analysis, and you can jump in and correct me when it's your turn, but you might still have an issue of how you would go about punishing the cheating if you observed it. Have any of you considered the possibility that the coordinated interaction would involve not just the R&D market, but possibly some other current markets, current production markets in which these companies are also competing. So, in other words, if cheating were observed in an R&D market, could it then be punished in another current product market? Has anyone considered that possibility, thought about it at all?
MR. SOHN: Wouldn't it depend on the relative gains from cheating in the innovation market? I mean, if, for example, you're a relatively small factor in the goods market, you know, punish the hell out of me, I'm going to leap frog you to the next generation.
MS. DeSANTI: Yes, of course, there could be any kind of -- any number of factual variations to this, the question is if the R&D is observeable and much of the discussion about the little likelihood of coordinated interaction has focused on the fact that our R&D tends not to be observeable and is conducted in secret. If you change that one fact, to what extent does it change the analysis and move you closer to a paradigm that you might consider to be reasonable?
MR. GILBERT: Well, it's certainly a key fact and you would see in terms of these punishments assuming about in other markets. I mean you see that, for example, airlines where airline A enters airline B's market, B might respond by cutting its prices in another market that A -- airline A depends on and where B is not a big player. So, we get this mutual interdependence and these cross punishments going on. So, it's a very important factor.
What I find -- I have always found very frustrating about antitrust policy, particularly on the enforcement side, is that antitrust policy tends to deal with exceptional cases, the cases that you see often have some bizarre element to them. And then coming out as a result, coming out with guidelines that say we're going to give a safe harbor to three firms or more or we're never going to deal with a coordinated behavior issue in R&D. Since we're dealing with these exceptional cases and there probably aren't that many to begin with, the fraction of those exceptional cases that might have some further exceptional property you are now ruling out becomes trouble some.
DR. RAPP: I degree with that, the point in fact that when you get the cases really bizarre enough then they go to the Supreme Court. That's the way things seem to work in antitrust. But I have to say on the other side of the coin in making reference to a point that Michael zone made a while ago, that there is an element of at least potential fans fullness about this if we know looking back either in the case law or in economic history, in the history of technology of the 20th century, if there are no revealed examples of this, if we don't see coordinated interaction at the R&D level, then we have to -- I mean, that's not saying it could never happen. And the bizarre case might emerge, but it again, to set up procedures to interdict it becomes problematic. I can't claim to have made that exhaustive investigation, but I did take a pass through at least my old -- the bookshelf in my office that contains the history of technology and while there are good examples around of technology markets being tied up in one way or another either through coordination or, you know, patent pooling or whatever, it's hard to observe examples of the sort of thing that you are -- of other collusive behavior or highly coordinated behavior in R&D. So, I'm skeptical.
MS. DeSANTI: Let me move to the issue of when there might be an economic effect, a unilateral effect, leaving aside the issue of coordinated interaction for a moment and I'm trying to get at the extent to which people see evidence beyond the economic theory here as a basis for agency action. Assume you don't have any cannibalization issue in a particular market, in other words there are two firms who are competing for a totally new product, this totally new product will not cannibalize any of their existing sales. Is there anyone here who -- but it is a merger from two to one, okay, and we'll leave aside for a moment the question of whether entry is possible, I'm just doing the competitive effects. Is there a unilateral competitive effect here? Is there anyone here who maintains that there is no possibility of any anti-competitive effect from such a transaction?
MR. SOHN: You're presuming that no one in the whole wide world has the same skill sets?
MS. DeSANTI: I am leaving that aside.
PROFESSOR CARLTON: Could you repeat that, Susan.
MS. DeSANTI: I am really trying to focus in on the unilateral competitive effect. What extent is there support for a belief that there could be a potential unilateral effect, leaving aside the substantial likelihood that this is a rare case that I am talking about and in most cases you would find indeed that there was entry and loss of substitution.
PROFESSOR CARLTON: Let me transcend the hypothetical. Two firms merging, two R&D.
MS. DeSANTI: Two R&D.
PROFESSOR CARLTON: They are not currently competing, is there the possibility of being an anti-competitive effect?
MS. DeSANTI: Um-hum.
PROFESSOR CARLTON: The answer to that question as to your cartel question there is always the possibility that that could occur. Maybe I should stop there.
MR. GILBERT: I think to clarify what I think Susan said, I think she said there's no cannibalization so that there's no existing product to be cannibalized, but they are competing in the development of some new product, either one will get it or the other will get it. Is that right? So, that's R&D competition to bring this product to market.
PROFESSOR CARLTON: I think you could have a hypothetical in which either outcome occurs. There's either more R&D or less R&D. So, the answer to your question is I could cook up a theory that would support that possibility. I just think you -- you know, in answer to your previous question, too, it's theoretically possible. I think, though, that I would take some issue with what was said in answer to the previous line of questions that the exceptional cases are the ones you see and therefore you should be guarded in putting forth safe harbors.
Because I can think of theoretical examples, as I said, when noncompeting firms merge in which consumers are harmed, yet everybody believes, I think around this table anyway, I would hope, that there are current merger guidelines where they gave safe harbors. So, if two wheat farmers want to merge, there's not an exception to this instance of something is going to happen, it makes sense. So, that's all I will add.
MS. DeSANTI: We'll get to the safe harbor issue next, because I agree with you that that raises a lot of judgment call issues as to how would you weigh, you know, your false positives problems versus a lot of other, but let me go back to my question. Is there anyone here who in that circumstance would say that there's no basis for any theory that this would be a potential anti-competitive effect from any combination? I want to make sure we're clear here.
PROFESSOR CARLTON: That's different from saying there's no theory from an anti-competitive effect. I would say there are theories that would justify the opposite effect and then you ask me does that mean there's support or not, that's how I would characterize it.
COMMISSIONER VALENTINE: I have a question about the theory of the economists and I hope you haven't written about that, John, and I think you may have. You all talk about the great ambiguity theoretically and empirically in concentration and but if you let the concentrated end of the spectrum fall off and you focus on the very high ends of concentration, is there anything more consistent out there on what tends to happen in the empirical literature when you've got an increase in concentration and let's say you're going from two innovators to one, does there tend to be a reduction in innovation or a sort of slowing of the process of innovation?
PROFESSOR CARLTON: Well, you can go first, Richard, if you like.
MR. GILBERT: I can tell you what the empirical literature says that I have seen which is the early -- some of the early empirical literature suggested that while there was no obvious effect going from relatively unconcentrated markets to moderately concentrated markets that there was a reduction in R&D expenditures for a firm going from moderately concentrated to highly concentrated markets. However, when people go back to these data and adjust for things like industry-specific effects and then you look at this cross sectional econometric statistical data that you tend to lose those relationships and they tend to get swamped by the industry-specific effects.
PROFESSOR CARLTON: All I was going to say is the quote I made this morning was from a survey of going through all these studies and they specifically looked at these old effects that Rich mentioned stopped up and basically no one can find systemic effects anywhere.
MS. DeSANTI: And what about when you add into the analysis should we be looking at things like, say, Michael Porter's theory of the importance of competition in innovation or Porter's theory about the importance of localized competition? I, you know, maybe it's not fair to call it just theory, it's mixed theory and facts, I guess some judge would say, but should we also be taking that into account in making these judgment calls?
DR. RAPP: I think the distinction that you just threw at the end of that question is an important one. Porter's observations about the sources of competitive advantage of nations are instructive, but it is -- I think it's more than casual empiricism on this point, but that's a very large study where one of the points that he makes more or less by inference after a lot of fact gathers about, you know, why Italian firms manage to compete on international markets and other firms do not, is he observes that if you have aggressive control competitors and if you have competition, active competition in the home markets, that by a process that we can all understand intuitively it makes firms in your home country that must compete in foreign markets robust. And that contributes to their competitiveness in some meaningful way, so it's part of that larger story. That's the way I think we've been distinguishing theory from fact or from empirical inquiry here.
I would put that in the latter camp and I think to just try and complete the answer to your question, although I'll defer to either of the gentlemen on my right, I think so far as pure theory is concerned, I think you can even when you -- when the numbers are -- Debra's question, even when the numbers are small, you can make it happen either way. You can have in highly concentrated markets you can make the -- with the change in the assumptions you can change the outcomes.
DR. YAO: With respect to going from Porter to something more general, I think that the case studies are very important, they often times are the source of identifying interactions that later one can find to be more general when one takes them to other industries. The fact that you may at this point be left with a number of case studies that tell you something doesn't mean that you haven't learned something. It's just that you haven't gotten to the point where you can apply it more generally. You don't understand it fully, but it's often the case that this is the source for it was a later good idea that gets shown to be correct through a more extensive cross-industry analysis or these segmental industry analysis.
MR. BAKER: Since Michael Porter's analysis has come up, I would like to pursue it for a second even though it's a little bit of a tangent here. Dennis, you talked a while back about the way professionals are going to Forty-niner's games and talking to each other.
MR. SOHN: He said it's possible.
MR. BAKER: Those aren't your clients, Michael.
MR. SOHN: You never know, Johnathan.
MR. GILBERT: Those come apart now.
COMMISSIONER VARNEY: Johnathan, they may be potential clients.
MR. SOHN: Incremental gains.
MR. BAKER: It was heard a lot today -- I'm going immediately after this to close down the merger between the wheat farmers, Dennis, it's over. So, the professionals are going to bars in San Jose.
PROFESSOR CARLTON: To see if they've paid me yet. You've insulted them.
MR. BAKER: The terrific wonderful things that they are learning, et cetera, et cetera, and the point is, and Michael this is the lesson, Michael Porter -- one of the lessons Michael Porter seems to draw from some of his anecdotes about localized competition is that when efficiencies are near each other, inhibition spills over quickly and it spurs aggressive competition among neighboring firms. And the striking anecdote that is repeated again and again in different ways in parts of the book is that the world class firms of this industry that are all located within two blocks of each other in some small town in rural Italy, and it's a bit of a tangent, but it is and it isn't a tangent from what we are doing.
What do you all make of this? Is this something we should be paying attention to? Should we be protecting localized competition? Should we think it's efficiency when they acquire a localized competitor because they suddenly have access to these spillovers and should we be concerned when localized competitors merge? What do you all make of that?
DR. RAPP: Let me start the ball rolling with a quick comment that if I had to guess without really knowing, a lot of it has to do with hiring people without making them move long distances. So, you really don't have to talk in bars or in football stadiums, you just have to hire one another's scientists or professionals or something like that. And if that's what it's all about or part of it is what it's about, then there is some price at which they move further and further and further. So, I don't think it -- I mean, that's the way I -- I would analyze the problem that you are raising.
MR. GILBERT: It's also the case that only in Silicon Valley does the local heavy metal rock station run adds for gallium arsenide field effect engineers, so there are certain labor market firms.
PROFESSOR CARLTON: I think your question could be, I am asking a different question, and that is that when you have what are called -- these are called conglomeration economies that you economists used to study, maybe now it comes back in fashion, now the economists are on to agglomeration of economies, but I think your question is this:
If there is an agglomeration and there is going to be a merger involving a large firm that takes resources from the center and presumes moving them to Europe, should you be concerned because of the reduction of agglomeration economies in the United States? And that strikes me as a difficult antitrust -- a difficult policy question for antitrust authorities to get into, because it's presuming that there are externals in other markets and now you are looking at the related effects having to nothing to do with the restriction of output in the particular merging market, but you are looking at the ancillary effect. So, there's a theory of these are technical firms second best which basically means that if all of the markets aren't perfect and you do something about one market and it effects another market, welfare effects get hard to question, to figure out. So, let me give an example.
If there were a merger of tennis manufacturers and they jacked up the price of tennis rackets, that would effect the demand for tennis balls. We usually in merger analysis don't look at what's happening in the supply and demand of tennis balls and that I think is what your question goes to. I would be kind of nervous going down that route.
DR. YAO: I think one of the questions is whether or not we have evidence of case studies that are basic -- that would look like Porter but turns out with very different outcomes. And unfortunately I don't know the literature, so I can't tell you, but if one did I think one could really learn from that because one could try to figure out what's common and what's not. And a handful of examples is good, but --
MS. DeSANTI: I have two last questions, one more competitive effects question, which is suppose you have merger going from three to two or four to three, leaving aside the possibility -- and this is an R&D merger, okay, leaving aside the possibility of coordinated interaction, is there a theory of unilateral -- anti-competitive competitive effects there, and if so, what is it?
PROFESSOR CARLTON: I think this is the same question we asked earlier when you were going from two to one and that is can you think of a theory, is there a theory that would provide the result that when you go from three to two things get worse in the social welfare. And I think the answer to that is yes, there's a theory that conversely, if I tweak a few of the assumptions that are hard to verify, you can get the results, so I think it would be the same answer.
MS. DeSANTI: I take it the basic answer is economists are creative.
PROFESSOR CARLTON: Can be.
COMMISSIONER VARNEY: Good question.
PROFESSOR CARLTON: It's a broad innovation market.
COMMISSIONER VARNEY: I see Professor Gilbert is going to have to leave.
MR. GILBERT: The questions are getting difficult. I have a plane to catch.
COMMISSIONER VARNEY: We will take your answer as innovation market analysis would work in that situation, whatever the question, right?
MR. GILBERT: Absolutely.
COMMISSIONER VARNEY: Thank you so much, we enjoyed having you.
Susan, you have the last question?
MS. DeSANTI: The last question is simply the same safe harbor issue that Mike raised. I would like to hear from each of you as to how you would approve that idea. You have a number in mind, let us know what number you would think about, and if you don't have a particular number in mind but you want to say what you would rate most well, that would be useful.
MR. SOHN: Can we all write on a piece of paper?
MS. DeSANTI: Then I get to draw it out and we all guess which answer belongs to who?
COMMISSIONER VARNEY: Mike, do you want to elaborate since it was your proposal?
MR. SOHN: No, there's not much more there. The basic thought was that three was a number to start thinking about, not simply because Bill Baxter used it, although he's very thoughtful on these issues, but because of any noneconomic distinction that these effects of concern are considered to happen in innovation markets, and there ought to be therefore a more liberal safe harbor than we have in the current merger guidelines.
COMMISSIONER VARNEY: Judy, what do you think?
MS. WHALLEY: Well, I note in my article that the standard by Baxter had for three which I also found appealing. I think that that's appealing if you are to try to do some sort of coordinated effect analysis, but I also think it's appealing in the unilateral effects, but that there's obviously some further questions to be asked and the critical question to me is what are you proposing that you are losing by the merger? Are you losing alternative paths? Are you losing incentives to innovate quickly? That could make a difference, what your concern is in that particular market could make a difference in terms of how many other people you would want to have.
The second one sort of goes back to the differentiated product's unilateral effects analysis in the merger guidelines about repositioning. And it seems to me that an important question is of the remaining person or persons in the market how able are they to reposition. Is it necessary that they reposition in order to replace the competition that's lost from the merger, and that would affect whether there's enough. Or four, I mean, I think that's a very market-specific question.
COMMISSIONER VARNEY: Do you want to comment, Professor Carlton?
PROFESSOR CARLTON: I really don't have much to add to what's been said, other than presuming that you go forward with an innovation market concept, I think a safe harbor is very desirable.
COMMISSIONER VARNEY: Okay.
DR. YAO: I have sort of a nonanswer. Absent having a lot of facts of, you know, what you have seen, it's very hard to know, for me, personally, whether there's a good number or four is a good number or what.
The second thing is three or four in what market? We haven't actually solved what's the market, and that makes a huge difference as to whether three is sensible or four is sensible or what not. And finally, as a matter of policy, putting out something and then finding out that you are wrong, you're kind of in an awkward position. Starting with nothing right now, granted you've got a problem in terms of guidance, but you have to just consider you could find out that you're wrong. Of course you might not because you've decided it's a safe harbor, so you never investigate it.
MR. SOHN: But Dennis just challenged that last statement, I mean the fact is that the agencies are bringing these cases. And we may agree as a panel that there's an insufficient empirical basis for them, but if they are going to bring them, at a minimum it seems to me that they have some obligation to say what they are doing, and if they feel that even as an overtly tentative matter, until we know more, they are going to establish a safe harbor, and I think that would aid certainty.
DR. YAO: I think certainty would be a good thing, particularly with respect to the analysis. And secondly, if it's understood that it's somewhat conditional, I think that would be all right with me, too.
COMMISSIONER VARNEY: I think that's hardly ever understood. Dr. Rapp?
DR. RAPP: I wish to end up being a hard case, as I have been throughout. The safe harbor necessarily implies a well-defined market, and in this case in using the innovation market approach, it seems to me that that is sufficiently problematic, so that I can't see myself advocating any sort of a kind of comprised safe harbor for the sake of creating certainty when the market definition process is an uncertainty and imperfect world in my view.
COMMISSIONER VARNEY: Anything else? Well, I thank you all for these thoughts. If you have any further thoughts on the subject we have raised this afternoon, please feel free to send us a note and we will take it into account. I again thank you all for coming, I know some of you traveled great distances and we really appreciate it. It's been very enlightening, and tomorrow we will be talking about efficiencies in these markets. I think Commissioner Starek is going to lead us through those discussions, and we will pick up tomorrow morning. Thank you all very much.
(Whereupon, at 4:30 p.m., the hearing was adjourned.)
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
//
C E R T I F I C A T E
DOCKET/CASE NUMBER: P951201
CASE TITLE: HEARINGS ON GLOBAL AND INNOVATION-BASED COMPETITION
HEARING DATE: October 25, 1995
I HEREBY CERTIFY that the transcript contained herein is a full and accurate transcript of the notes taken by me at the hearing on the above cause before the FEDERAL TRADE COMMISSION to the best of my knowledge and belief.
DATED: October 25, 1995
SIGNATURE OF REPORTER
Sally Jo Bowling
(NAME OF REPORTER - TYPED)