Skip to main content
The Continuing Legal Education Seminar, The Florida Bar
Orlando, Florida
Roscoe B. Starek, III, Former Commissioner

Good afternoon. It is a pleasure to participate in this program sponsored by the Florida Bar. I should say that the following remarks are my own and do not necessarily represent the views of the Federal Trade Commission or of any other Commissioner.

I. There Is Something Left To Say About Innovation Markets

My topic for today -- innovation markets in merger analysis -- has been a subject of considerable interest and discussion in antitrust circles for some time.(1) But the discussion seems to have reached a new level of intensity over the past year or two. During the FTC's recently concluded public hearings on Global and Innovation-Based Competition, for example, the merits of innovation market analysis were vigorously debated by distinguished lawyers, economists, and business leaders approaching the subject from a range of contrasting perspectives.(2) Last year, to cite another example, the Antitrust

Law Journal published a thought-provoking symposium on innovation market analysis.(3) Having followed with considerable interest the outpouring of articles and speeches about innovation markets in merger analysis, I will admit to having been somewhat surprised by the vitality of the debate on what I might have otherwise thought a rather arcane, if nevertheless important, subject. Rather than taking sides in this debate or duplicating past discussions of pertinent cases and consent orders, I offer today a perspective on some of the significant antitrust enforcement policy issues presented in the ongoing debate over innovation market analysis.

A. Open-minded Skepticism

If pressed to come up with an alternative title for this discussion, I would be tempted to call it "Innovation Markets and Merger Analysis -- A Guide for the 'Open-minded Skeptic.'" The open-minded skeptic's approach to innovation markets combines an appreciation of the critical importance of innovation with an awareness of the lack of consensus on what it is, how to measure it, and how to encourage more of it. I generally share the view, widely held among observers of our economy, that technological change -- "technological progress," to use a slightly battered term -- is central to the economy's current and future success in raising living standards.

In the analysis of a proposed merger or acquisition, I am quite prepared to give appropriate weight to innovation market analysis to the extent that it helps the Commission to identify aspects of a transaction that may reduce market participants' incentives and ability to innovate. But before embracing it, I want to be confident that innovation market analysis will help us (at a reasonable cost) to recognize previously unidentified anticompetitive conduct, without also leading to the condemnation of legitimate competitive activities. The ultimate test of any new form of antitrust analysis is whether the new approach serves the overarching goal of antitrust enforcement: the protection of competition in order to maximize consumer welfare. Because our experience with "innovation market" analysis is still very limited, it is too soon to tell whether such analysis will, on balance, enhance consumer welfare.

B. The Nature of the Debate

Whatever becomes of innovation market analysis in the next five or ten years, the debate has been for the most part constructive. In my view, there is no inconsistency between open-minded skepticism and an appreciation of the considerable merit in the positions on both sides of the innovation market controversy. There is substantial common ground regarding the ends of innovation market analysis. We all favor innovation; no one, to my knowledge, has declared himself or herself "anti-innovation" on either side of this debate.(4) It follows logically from this that most proponents and opponents of innovation market analysis would probably agree as a theoretical matter that antitrust policy should foster market conditions that enable firms to achieve whatever level of innovation consumers prefer.

With regard to the means of assessing the innovation effects of mergers -- by which I mean the structure, the procedure, and, in some cases, even the legitimacy of the innovation market inquiry -- there is a high level of disagreement and uncertainty. Both the pro and con sides of the innovation market debate make respectable arguments.

Innovation market proponents justifiably emphasize that, given the economic significance of innovation, the innovation effects of mergers are too important to ignore. In a sense, if innovation markets did not exist, someone would have to invent them. The innovation market concept commands attention not because of any demonstrated transcendent wisdom, but because it promises to address, at least in part, the critically important question of how to assess the effect of an acquisition, agreement, or course of conduct on future innovation -- something we all would like to be able to do accurately.(5) Proponents characterize innovation market analysis as a technical concept, stressing its value as a practical tool for assessing actual transactions.(6) Finally, and perhaps most important, proponents of innovation markets believe that the information relevant to assessing innovation effects is sufficiently accessible and measurable to allow them to make informed, reasonably predictive judgments about whether a proposed merger will foster or hinder innovation.

Critics of innovation markets do not generally take issue with the importance of innovation. The critics diverge from the proponents primarily in the beliefs that (1) the state of our knowledge regarding innovation is woefully insufficient to permit the assessment of the innovation effects of mergers and (2) the risks of overdeterring efficient conduct significantly outweigh the likely benefits of innovation market analysis.

C. The Next Several Years

In a relatively narrow sense, the advocates of innovation markets have already prevailed. Innovation market analysis is now an element of the official enforcement policy of the Commission and the Department of Justice, as articulated in the Intellectual Property Guidelines(7) and in a growing number of consent orders settling merger investigations.(8) But it is perhaps more accurate to say that the antitrust enforcement agencies and the antitrust community are at the same time debating and experimenting with innovation market analysis, i.e., with the application of antitrust analysis and enforcement to a new dimension of competition. The innovation market concept, then, is an intriguing work in progress and not a fully-formed analytical system. While it is likely that innovation market analysis will be significantly improved and refined, I would not be completely shocked to see the approach abandoned within the next three to five years.

It is still unclear to me whether, and in what form, innovation market analysis will be effectively absorbed and integrated into the "everyday" merger analysis of the antitrust enforcement agencies. That outcome will depend on the experience of the agencies and on our ongoing dialogue with the bar, academics, consumers, and the business community.

II. Innovation Market Analysis Under The 1995 DOJ/FTC Intellectual Property Guidelines

The text of the I.P. Guidelines, drafted and issued jointly as the enforcement policy of both federal antitrust agencies after more than a year of preparation, theoretically presents an official explanation of current innovation market analysis as the agencies understand it.(9) The explanation of innovation market analysis in the I.P. Guidelines consists of three paragraphs of text in Section 3.2.3 and a hypothetical example in which innovation markets are considered in the context of a joint venture.(10) I think it worthwhile to quote theGuidelines' definition of "innovation market" at some length:

An innovation market consists of the research and development directed to particular new or improved goods or processes, and the close substitutes for that research and development. The close substitutes are research and development efforts, technologies, and goods that significantly constrain the exercise of market power with respect to the relevant research and development, for example, by limiting the ability and incentive of a hypothetical monopolist to retard the pace of research and development. The Agencies will delineate an innovation market only when the capabilities to engage in the relevant research and development can be associated with specialized assets or characteristics of specific firms.(11)

The third paragraph in this section of the Guidelines focuses specifically on evidence that the agencies may consider in connection with merger review of innovation markets.(12) This evidentiary discussion is substantially devoted to explaining how the agencies intend to weigh market share data -- or proxies for such data -- in assessing the "competitive significance of innovation market participants."(13) The agencies state that they will take all relevant evidence into account, using market share data (where available and reliable) as well as evidence of buyers' and sellers' "assessments of the competitive significance of innovation market participants."(14)

At least for the present, the legal scope and application of the enforcement agencies' innovation market analysis is unclear. To be sure, the agencies have expressly endorsed the use of innovation markets in the analysis of licensing arrangements that "may adversely affect competition to develop new or improved goods or processes . . . ."(15) When the agencies are presented with this type of licensing arrangement, the I.P. Guidelines provide that they "will analyze its impact either as a competitive effect in relevant goods or technology markets, or as a competitive effect in a separate innovation market."(16) However, the I.P. Guidelines shed no perceptible light on the extent to which the agencies have integrated or intend to integrate innovation markets into merger analysis.(17)

The I.P. Guidelines state that "[t]he Agencies will apply a merger analysis to an outright sale by an intellectual property owner of all rights to that intellectual property . . ." or to an exclusive licensing transaction that amounts to the effective sale of intellectual property.(18) The text of the I.P. Guidelines does not squarely address the applicability of innovation market analysis to acquisitions of assets other than intellectual property rights; nor, to my knowledge, is that question answered in any other official statement of antitrust enforcement policy.(19)

III. Assessing the "Innovation Effects" of Mergers

A. The Gilbert and Sunshine Approach

The I.P. Guidelines essentially present a conceptual outline applying innovation market analysis to a specific category of intellectual property licensing restraint.(20) By contrast, former Deputy Assistant Attorneys General Gilbert and Sunshine have formulated an innovation market analysis generally applicable to a broad range of transactions in which preliminary evaluation identifies "relevant R&D product and geographical markets where competition in research and development may be affected by a proposed merger or acquisition."(21) Although Gilbert and Sunshine have gone beyond the I.P. Guidelines' contribution to the innovation market debate, their approach is structurally similar to the innovation market analysis set out in the I.P. Guidelines and is therefore vulnerable to some of the same basic criticisms that have been raised in response to the I.P. Guidelines.

Just like the innovation market analysis in the I.P. Guidelines, the Gilbert and Sunshine model draws substantially from the 1992Horizontal Merger Guidelines for both the structure and the substance of its innovation market inquiry.(22) For example, the authors define "innovation market" as "a set of activities and a geographical area in which a hypothetical monopolist would impose at least a small but significant and nontransitory reduction in R&D effort."(23) In their focus on market concentration, post-merger investment incentives, competitive effects, and efficiencies,(24) the questions posed by Gilbert and Sunshine to determine whether a proposed merger would likely diminish "R&D activity" in a "relevant innovation market" also track the factual inquiry of the Horizontal Merger Guidelines.(25)

The Gilbert and Sunshine innovation market inquiry can be reduced to five steps. The first step is to identify the overlapping R&D activities of the merging firms, which roughly corresponds to the relevant product market inquiry in the Merger Guidelines. The second step is to identify any alternative sources of R&D -- essentially the demand substitution inquiry. The third step in the Gilbert and Sunshine innovation market inquiry is the evaluation of actual and potential competition from downstream products that could make it unprofitable for a hypothetical R&D monopolist to raise price or reduce output. The fourth step is the assessment of potential competitive effects on investment and R&D that could result from the increased concentration brought about by the merger. The final step is to assess any efficiencies arising from the merger that would likely increase output and lower the postmerger price of R&D in the innovation market under review,(26) in order to determine whether such efficiencies would be sufficient to outweigh any likely anticompetitive effects of the proposed merger.

Consistent with the innovation market analysis set out in the I.P. Guidelines, the applicability of the Gilbert and Sunshine inquiry is theoretically limited to situations in which specialized "innovation assets" or capabilities can be identified in the process of reviewing a proposed merger. The Gilbert and Sunshine innovation market inquiry is not to be applied unless the enforcement agencies can identify "markets in which R&D [is] directed toward particular new products or processes [that] require specific assets that are possessed by identified firms."(27)

B. Problems with Innovation Market Analysis for Mergers

Innovation market analysis generally, and the Gilbert and Sunshine approach specifically, have received extensive criticism, much of which merits our consideration.(28) Rather than detail every objection to every controversial aspect of innovation market analysis for mergers, I have synthesized into seven issues the major concerns raised by some of the most cogent critics of innovation market analysis. I reiterate at this point that, as an open-minded skeptic, I am no more endorsing the critique of innovation market analysis than am I endorsing the use of innovation market analysis in merger review. I see promise in the concept of innovation markets, but it is still too soon to tell whether (and how) innovation markets might be usefully integrated into the agencies' standard merger review inquiry.

For the time being, I suggest careful consideration of the following issues by the agencies and by all participants in the innovation market debate as we assess the validity and utility of innovation markets in the merger context:

1. Both the I.P. Guidelines and the Gilbert and Sunshine model attempt to use "research and development" activity as a proxy for innovation. In practical application, then, the existing innovation market analyses do not actually identify or measure "innovation." Do the proxies come close enough?

a. What is "R&D activity?"

b. Why should we think that measurements of "R&D activity" will permit the reliable evaluation of "innovation" activity?

2. Can the level of "concentration" in research and development activity be reliably established in a practical and transparent way as a part of the enforcement agencies' merger review process?

3. Can we assume that a less concentrated "R&D market" will produce more innovation than a more concentrated "R&D market"?

a. There are many examples in recent business history of big firms with huge R&D "assets" losing the competitive race for innovation to much smaller firms with quite limited R&D resources.

b. Does innovation market analysis adequately address what we might call the "David and Goliath" problem?

4. Can we reliably associate innovation with certain specialized assets?

a. For example, would current innovation market analysis recognize and appropriately assess the specialized R&D assets that Steve Jobs and Steve Wozniak maintained in Mom and Dad's garage shortly before the appearance of the first Apple computer?

5. Current innovation market analyses are structurally similar to the merger review inquiry required by the 1992 Horizontal Merger Guidelines.

a. Do the 1992 Horizontal Merger Guidelines offer an appropriate model for the evaluation of how a proposed merger might affect the level and quality of post-merger innovation?

6. Does innovation market analysis take adequate account of the potentially procompetitive role of some forms of competitor collaboration in the context of innovation?

a. If we assume for the purposes of argument that there may be procompetitive collaborations in research and development under some circumstances, does (or could) innovation market analysis allow us to distinguish anticompetitive collaboration from collaboration that fosters innovation?

7. What are the costs of the likely "false positive" determinations arising from the application of innovation market analysis in merger review?

a. How significant are the risks of overdeterring efficient and potentially innovation-friendly transactions in the everyday use of innovation market analysis in antitrust merger review?

* * * * * *

At this point I do not know the answers to all the questions I posed, but I think the questions deserve some thought before we all rush to embrace innovation market analysis. When new approaches to antitrust enforcement are advanced, we sometimes see an overenthusiastic response. A stampede to utilize innovation market analysis in far too many cases could limit procompetitive activity and lessen consumer welfare by deterring the very innovation we are trying to maintain.

I want to thank you all once again for the opportunity to share some of my thoughts with you this afternoon. I would be happy to try to answer your questions.

(1)Analysis of something akin to innovation markets predates the common usage of the term "innovation markets," seee.g.United States v. Automobile Manufacturers Ass'n, 307 F. Supp. 617, 618 (C.D. Cal. 1969) (approving consent decree settling charges of conspiracy "to eliminate competition in the research, development, manufacture and installation of motor vehicle air pollution control equipment . . . " in violation of Section 1 of the Sherman Act), aff'd in part and appeal dismissed in part, 397 U.S. 248 (1970), as does the theoretical application of the concept to mergers. See William F. Baxter, The Definition and Measurement of Market Power in Industries Characterized by Rapidly Developing and Changing Technologies, 53 Antitrust L.J. 717 (1984) (merger analysis of "R&D markets").

(2)Seee.g., Sumanth Addanki, Remarks Before FTC Hearings on Global and Innovation-Based Competition, Washington, D.C. (Oct. 25, 1995), at 938 (transcript); Dennis W. Carlton, Remarks Before FTC Hearings on Global and Innovation-Based Competition, Washington, D.C. (Oct. 25, 1995), at 925 (transcript); Richard J. Gilbert, Remarks Before FTC Hearings on Global and Innovation-Based Competition, Washington, D.C. (Oct. 25, 1995), at 906, 982 (transcript); Michael Sohn, Remarks Before FTC Hearings on Global and Innovation-Based Competition, Washington, D.C. (Oct. 25, 1995), at 990 (transcript); Judy Whalley, Remarks Before FTC Hearings on Global and Innovation-Based Competition, Washington, D.C. (Oct. 25, 1995), at 1002 (transcript); Dennis A. Yao, Remarks Before FTC Hearings on Global and Innovation-Based Competition, Washington, D.C. (Oct. 25, 1995), at 950 (transcript).

(3)Symposium: A Critical Appraisal of the "Innovation Market" Approach, 64 Antitrust L.J. 1 (1995).

(4)An example of this common ground appears in the innovation market symposium cited in n.3, supra. Professor Hay takes issue with innovation market analysis in general -- mostly questioning its originality -- and he specifically finds fault with the variation proposed by former Deputy Assistant Attorneys General Richard Gilbert and Steven Sunshine, two proponents of innovation market analysis for mergers. George A. Hay, Innovations in Antitrust Enforcement, 64 Antitrust L.J. 7 (1995). Professor Hay nevertheless acknowledges that "[i]nnovation is important. Innovation is good. The incentives to innovate should be protected and possibly strengthened." Id. at 17.

(5)This implicates a perennial "Schumpeterian" hypothesis. Schumpeter theorized that an industrial organization in which firms possess market power -- rather than a regime of perfect, atomistic competition -- would produce the highest levels of innovation. Joseph A. Schumpeter, Capitalism, Socialism and Democracy 106 (1942). After decades of empirical research into this question, the debate has not been resolved. Some empirical studies support Schumpeter's theory, but many others reach equivocal or opposite conclusions. F.M. Scherer, Schumpeter and Plausible Capitalism, 30 J. Econ. Lit. 1416 (1992); F.M. Scherer, Innovation and Growth: Schumpeterian Perspectives (1986).

(6)Cf. Richard J. Gilbert and Steven C. Sunshine, The Use of Innovation Markets: A Reply To Hay, Rapp, And Hoerner, 64 Antitrust L.J. 75, 82 ("[I]nnovation market analysis is simply a tool to aid in the analysis of competitive effects.").

(7)U.S. Department of Justice and Federal Trade Commission, Antitrust Guidelines for the Licensing of Intellectual Property § 3.2.3 (Apr. 6, 1995) ("I.P. Guidelines").

(8)Seee.g.Silicon Graphics, Inc., FTC Docket No. C-3626 (Nov. 14, 1995) (consent order); Wright Medical Technology, Inc., FTC Docket No. C-3564 (Mar. 23, 1995) (consent order); United States v. Flow International Corp. & Ingersoll Rand Co., Civ. No. 94-71320 (S.D. Mich. filed Apr. 4, 1994), 6 Trade Reg. Rep. (CCH) ¶ 45,094.

(9)I.P. Guidelines § 3.2.3 ("If a licensing arrangement may adversely affect competition to develop new or improved goods or processes, the Agencies will analyze such an impact either as a separate competitive effect in relevant goods or technology markets, or as a competitive effect in a separate innovation market.").

(10)See I.P. Guidelines § 3.2.3 and Example 4.

(11)I.P. Guidelines § 3.2.3 (emphasis added).

(12)Id. (13)Id.

(14)Id. This portion of the I.P. Guidelines continues: "The Agencies may base the market shares of participants in an innovation market on their shares of identifiable assets or characteristics upon which innovation depends, on shares of research and development expenditures, or on shares of a related product. When entities have comparable capabilities and incentives to pursue research and development that is a close substitute for the research and development activities of the parties to the licensing arrangement, the Agencies may assign equal market shares to such entities."

(15)I.P. Guidelines § 3.2.3.

(16)Id. (emphasis added).

(17)Although the use of innovation markets is not inconsistent with the 1992 Horizontal Merger Guidelines, neither the term "innovation market" nor the substance of innovation market analysis appears in the Merger Guidelines text. Until recently, FTC merger analysis under Section 7 of the Clayton Act and the 1992 Horizontal Merger Guidelines did not typically attempt to assess the likely "innovation effects" of a proposed merger. But cf. 1992 Horizontal Merger Guidelines § 1.0 n.6.

(18)Id. § 5.7.

(19)But seee.g., Susan S. DeSanti, Remarks Before the Symposium on Innovation, American Bar Association Spring Meeting 16 (Apr. 6, 1995) ("[T]he Intellectual Property Guidelines do not apply to licensing transactions that are analyzed as mergers and acquisitions," but merger-related innovation market principles have influenced the development of the innovation market inquiry in the I.P. Guidelines); Richard T. Rapp, The Misapplication of the Innovation Market Approach to Merger Analysis, 64 Antitrust L.J. 19, 22 (1995) ("The 1995 Intellectual Property Guidelines govern agency review of licensing practices, not mergers."). If innovation market analysis is within the ambit of Section 7 of the Clayton Act -- as has been persuasively argued -- then this silence may be of no great consequence. See, e.g.,Richard J. Gilbert and Steven C. Sunshine, Incorporating Dynamic Efficiency Concerns in Merger Analysis: The Use of Innovation Markets, 63 Antitrust L.J. 569, 597-601 (1995) ("Gilbert & Sunshine"); but see also Robert J. Hoerner, Innovation Markets: New Wine In Old Bottles?, 64 Antitrust L.J. 49 (1995) (questioning Section 7 authority for innovation markets).

(20)See I.P. Guidelines § 3.2.3.

(21)Gilbert & Sunshine, 63 Antitrust L.J. at 594-97.

(22) at 594-95.

(23)Id. at 594.

(24)Id. at 595.


(26)Id. at 596-97.

(27)Id. at 596; see also id. at 588 ("The sources of innovation may not be reliably identified unless innovation would be unlikely in the absence of specialized assets.").

(28)Seee.g.Symposium: A Critical Appraisal of the "Innovation Market" Approachsupra n.3; see also Remarks of Dennis W. Carlton, Dennis A. Yao, Richard T. Rapp, and Lawrence J. White Before FTC Hearings On Global and Innovation-Based Competition (Oct. 25, 1995). Gilbert and Sunshine offer responses to elements of the critique of their innovation market analysis in The Use of Innovation Markets: A Reply to Hay, Rapp, and Hoernersupra n.6.