This document is available in three formats: this web page (for browsing content), PDF (comparable to original document formatting), and Word. To view the PDF you will need Acrobat Reader, which may be downloaded from the Adobe site. For an official signed copy, please contact the Antitrust Documents Group.

Why Bother?:
On Market Definition under the Merger Guidelines

William Blumenthal *

Statement before the
FTC/DOJ Merger Enforcement Workshop
Washington, DC
February 17, 2004

Let me begin with a confession: When conducting merger analyses and when handling the premerger process in the United States on behalf of notifying parties, I have not defined a market in more than fifteen years. From my private sector, defense-oriented perspective, the need to conduct a comprehensive, formal, Guidelines-style market definition exercise has not presented itself even once — not in defending several transactions in the courts, not in handling dozens of staff investigations, not in handling hundreds of premerger notifications, not in providing counseling to party and third-party clients in connection with literally thousands of transactions. Worse than unnecessary, any effort formally to define markets would have been unduly costly, time-consuming, and invasive, and it probably would have yielded less reliable outcomes than more streamlined techniques. The only instances in which I have expressed a contention as to "relevant market" have been (a) in private litigation on behalf of plaintiffs and (b) in premerger notifications submitted in foreign jurisdictions that require a statement of market definition in order to perfect the submission. In those instances the contentions have been less analytically satisfying than no definition at all.

None of this is intended to deny the tremendous contributions of both the hypothetical monopolist test and its adoption in the 1982 Merger Guidelines. For two decades I have noted their special significance.(1) It probably is not overstatement to say that William Baxter, Tyler Baker, and the other framers of the Guidelines saved the intellectual integrity of Section 7.

Time marches on, though, as do techniques of analysis and practices of enforcement agencies. From the vantage point of 1982, the hypothetical monopolist test and its use in market definition were essential — in a stepwise approach that still relied heavily on concentration statistics, Step II ("Market Definition and Measurement") was the primary locus of effects analysis. With the substantial revision and expansion of the succeeding competitive effects step in 1992, however, the significance and utility of a discrete, prior market definition step became muddier.(2) By 2002 the analysis of market definition and the analysis of competitive effects had become so commingled that it was fair to ask: Is market definition an input to competitive effects or an output? The answer is probably "Both" or "Neither" — it's a case of simultaneous determination.

This evolution — and the meaning of "relevant market" today for purposes of merger intervention or non-intervention decisions — probably is not understood by more than 500 people on the planet. For most private practitioners, for most business personnel practicing self-help, for most new agency staffers, for most foreign enforcement agencies implementing new merger control regimes, "Market Definition and Measurement" remains a discrete step that begins the analysis. If they err at that step, all ensuing steps are apt to be wrong.

And they err a lot. For the private sector, the error is generally in the direction of overbreadth. The Guidelines say that "the Agency will begin with each product (narrowly defined) produced or sold by each merging firm,"(3) but the meaning of "product (narrowly defined)" is not widely understood beyond Hart-Scott-Rodino aficionados. I have received several Second Requests in which the terms "relevant product" and "relevant geographic area" are defined in a manner that yields more than a million combinations. One of the Second Requests was so granular in its focus that a literal response to a particular interrogatory, printed out as demanded on paper in an Excel spreadsheet format in normal font filling a page, would have filled ten thousand cartons. (The staff granted limitations.) By contrast, most private lawyers and business personnel approach market definition with a more strategic perspective that looks to business units or lines of business or industry classifications — as they historically have done with senior management and Wall Street analysts. To do more would be contrary to training and intuition, as well as expensive and disruptive.

For new agency staffers and especially for foreign agencies, the error is generally different — they may faithfully apply the market definition provisions (that is, sections 1.1 and 1.2) of the Guidelines, but they then fail fully to apply the provisions on identification of participants, calculation of shares, and significance of shares (sections 1.3, 1.4, and 1.52) and assessment of competitive effect (section 2). Once a market has been defined, it is simply too easy to lapse into calculating concentration based on historical market shares and drawing an inference, à la the 1960s. At a minimum, the inference often results in informational filtering that affects all that follows.

Basically, to draw from the parlance of the FTC's Bureau of Consumer Protection, the 1992 Guidelines are deceptive. They may be literally accurate, and their meaning may have been properly understood at the time they were issued, but their meaning is misinterpreted today by a material percentage of readers. I ignore them as an operational tool, and I urge associates and clients to do likewise, except perhaps as background reading. If the uninitiated try to apply the Guidelines without detailed annotations explaining terms of art, they are likely to reach an erroneous conclusion.

Instead, for purposes of analyzing actual transactions, I teach inexperienced lawyers, clients, and others to use a reduced-form approach that focuses on three questions:(4)

  1. Where is the value in the transaction?
    • Margin improvement through cost reduction is fine.
    • Margin improvement through price enhancement is bad.

  2. What will customers say?
    • Customer support for the transaction is good.
    • Customer opposition is generally a source of significant antitrust concern.

  3. What will happen to prices?
    • Reduction in prices as a result of the transaction is good.
    • Artificial elevation of prices as a result of the transaction is bad.
    • The baseline price against which the post-transaction price should be compared is what the price would have been in the absence of the transaction. If prices are falling and the transaction will slow, but not stop the price reduction, that is treated as a price elevation.
    • If some prices will fall and other prices will rise, antitrust analysis will focus on the products and services for which price will rise.

The first two questions are intended primarily as background to aid understanding. The third question goes directly to ultimate effect. The focus on price disregards other types of potential adverse effects, but one can delve into output levels, innovation, quality, and the like with follow-up probes. The probes come at the same question several different ways. ("What's going to happen to prices? On which products do you think you'll have the most pricing flexibility? Are there any products where you might try to raise price? What's going to stop you from raising price?")

Operationally, this mode of analysis borrows from the approach commonly used in the medical profession to diagnose ailments — based on initial interviews, identify candidates for adverse effects, and rule them out sequentially with follow-up tests. If no conjectured adverse effect survives scrutiny (and assuming that our conjectures are sufficiently broad and creative), the transaction will be cleared. If a conjectured adverse effect cannot be ruled out after considering all possible marketplace responses, that becomes the bull's-eye. In essence, the "relevant market" of concern is that grouping of products and geographic areas in which an adverse effect is likely to occur. One then moves on to ask whether evidence can be adduced to support that market definition consistent with the standards applied by the courts.

This approach should be recognizable to the agencies. Several years ago, while serving as Director of the FTC's Bureau of Economics, Jonathan Baker was pilloried for his lawless suggestion of res ipsa loquitur merger analysis.(5) As an operational tool, the res ipsa approach has great value. The principal issue, as Baker recognized, was whether the approach can be reconciled with the case law that governs agency interventions in the courts.

Reconciliation with the case law should not be undersold. In a previous article examining the unique influence of the 1982 Merger Guidelines,(6) I identified six reasons for their success:

  1. The 1982 Guidelines had a credibility derived from substantial adherence to pre-existing law.
  2. While not rejecting prior law, the 1982 Guidelines filled in its interstices in ways that enhanced our understanding and appreciation.
  3. The 1982 Guidelines were fully specified.
  4. The 1982 Guidelines, aided by the legal standards of the time, struck a unique balance between simplicity and flexibility.
  5. The 1982 Guidelines were sufficiently operational to be of practical use.
  6. The 1982 Guidelines fairly portrayed contemporary government enforcement policy.

As we examine the current applicability of the 1992 Guidelines and their market definition provisions, the Guidelines probably fail on elements 3-6. The challenge facing revisers today is to correct those deficiencies without running afoul of elements 1 and 2.

The case law seems to require market definition, and it seems to require a stepwise approach in which market definition precedes analysis of competitive effect. One can only imagine the blizzard of quotations from Supreme Court cases that would confront any agency effort to do otherwise. In du Pont the Court wrote that "[d]etermination of a relevant market is a necessary predicate" to a Section 7 claim.(7) It also wrote that market definition is needed because the substantiality element of Section 7 "can be determined only in terms of the market affected."(8) In Brown Shoe the Court quoted du Pont and further explained that market definition is necessary in light of Section 7's directive that a substantial lessening of competition be shown "in any line of commerce in any section of the country."(9) In other Supreme Court cases(10) and in hundreds of lower court cases, the phrases have been repeated as a mantra.

How can revised Guidelines reconcile the case law with an operational framework that accurately reflects agency practice? Several possible approaches warrant consideration.

The first approach would be to adopt Guidelines that claim to be an analytical case selection tool for use in the agencies' administrative discretion — an exercise in applied economics that largely eschews the case law. Legal considerations would not be expressly reflected until a final step in the analysis — determination of whether an economically-derived conclusion can be translated into legal elements sufficient to support the drafting of a complaint.

A second approach would retain market definition as an initial step in the analysis, but would import much of the competitive effects analysis into that step. Market definition and competitive effects would be simultaneously determined.

A third approach would be iterative — beginning with a quick "virtual market definition" that identifies candidate problem markets, proceeds to competitive effects analysis, and returns to "confirmation of market definition" as a late step.

Under any of the three approaches, I would highlight the hypothetical monopolist concept — perhaps giving it even greater emphasis than in the current Guidelines — and would use it as an organizing principle. In the two decades since the 1982 Guidelines, courts in the United States (and elsewhere) have widely endorsed the concept. Repeatedly tying steps in revised Guidelines back to their role in the search for the hypothetical monopolist would provide a needed bridge to the case law.

Undoubtedly other approaches will be identified. Creativity will be needed, because the challenge is great, and good guidelines are tough to draft. But take heart by considering the challenge that faced the drafters of the 1982 Guidelines and by looking at what they accomplished.


FOOTNOTES

* Partner, King & Spalding LLP, Washington, DC. The author gratefully acknowledges helpful comments on an earlier draft from his colleagues Brian R. Meiners, Amy B. Metzel, Jeffrey S. Spigel, and Tamara J.Yakimetz.

1. See, e.g., William Blumenthal, Clear Agency Guidelines: Lessons from 1982, 68 Antitrust L.J. 5 (2000); William Blumenthal, Ambiguity and Discretion in the New Guidelines: Some Implications for Practitioners, 61 Antitrust L.J. 469 (1993); Donald I. Baker & William Blumenthal, Symposium: The 1982 Merger Guidelines — The 1982 Guidelines and Pre-Existing Law, 71 Calif. L. Rev. 311 (1983). Greg Werden has periodically accused me (with some basis) as having had inconsistent early views on the hypothetical monopolist test. See, e.g., Gregory J. Werden, The 1982 Merger Guidelines and the Ascent of the Hypothetical Monopolist Paradigm, 71 Antitrust L.J. 253, 253 nn. 2-3 (2003) (noting praise in Baker & Blumenthal, supra, and criticism in Joe Sims & William Blumenthal, New Merger Guidelines Provide No Real Surprises, Legal Times, June 21, 1982, at 17.) The shift was not from the difference in co-authors, but rather from the passage of time. Sims and I were offering an immediate reaction several days after release of the Guidelines, and we (like others) doubted that the test would be workable. By the time Baker and I were writing a half-year later, we and others had found techniques to make it work.

2. See, e.g., Ambiguity and Discretion, supra note 1; William Blumenthal, Thirty-One Merger Policy Questions Still Lingering after the 1992 Guidelines, 38 Antitrust Bull. 593 (1993).

3. 1992 Guidelines § 1.11.

4. I have attached two presentations designed for this purpose. The first, from which the treatment in the text is drawn, consists of slides used in various training sessions. See Antitrust Basics of M&A, at slide 4 (undated). The second consists of the slides used in the merger segment of the ABA's Antitrust Fundamentals course at the Spring Meeting. William Blumenthal, Antitrust Fundamentals: Mergers, before the ABA Antitrust Section 50th Annual Spring Meeting, at slide 8 (Apr. 24, 2002). This reduced-form approach is borrowed from Wayne D. ("Dale") Collins, whom I first saw use it a decade ago.

5. See Jonathan B. Baker, Antitrust in the Information Revolution: New Economic Approaches for Analyzing Antitrust Issues, Prepared Remarks before The George Mason University Law Review Symposium on Contemporary Empirical Merger Analysis, Rosslyn, Virginia (Oct. 11, 1996), published as revised in 5 Geo. Mason U. L. Rev. 347 (1997).

6. See Clear Agency Guidelines, supra note 1, at 15-20.

7. United States v. E.I. du Pont de Nemours & Co., 353 U.S. 586, 593 (1957), quoted in United States v. Marine Bancorp., 418 U.S. 602, 618 (1974).

8. du Pont, 353 U.S. at 593.

9. Brown Shoe Co. v. United States, 370 U.S. 294, 324 (1962).

10. E.g., United States v. Marine Bancorp., 418 U.S. 602, 618 (1974); United States v. General Dynamics Corp., 415 U.S. 486, 510 (1974).