GEOGRAPHIC MARKET DEFINITION IN THE PRESENCE OF GLOBAL COMPETITION

Testimony of Eastman Kodak Company
Before the Federal Trade Commission
October 19, 1995

Presented by Robert B. Bell
Wiley, Rein & Fielding
Washington, D.C.

I. INTRODUCTION

The 1992 Merger Guidelines barely touch on how competition between U.S. firms and foreign firms should be accounted for in assessing the impact of a merger. Indeed, the entire subject is treated in two paragraphs in section 1.43 of the Guidelines. Similarly, we are aware of only a single appellate case -- United States v. Eastman Kodak Company, 63 F.3d 95 (2d Cir. 1995) -- that treats the issue of whether a geographic market broader than the United States is appropriate for antitrust analysis. This testimony examines the Merger Guidelines and some of the leading economic literature on geographic market definition and proposes how the Commission should treat claims that the appropriate geographic market is broader than the United States.

II. THE ECONOMIC LITERATURE

The leading journal articles that address geographic market definition include Kenneth G. Elzinga & Thomas F. Hogarty, The Problem of Geographic Market Definition in Antimerger Suits, 18 Antitrust Bulletin 45 (1973) and William M. Landes & Richard A. Posner, Market Power in Antitrust Cases, 94 Harv. L. Rev. 937 (1981). Elzinga & Hogarty contend that geographic markets should be defined according to the areas from which, and to which, a significant amount of the relevant product is shipped. Although their article does not address in detail market definitions broader than the United States, the authors would find a "market . . . (at least) national in scope" where plants within the United States do not account for 75 percent of sales within the United States. Elzinga & Hogarty, supra, at 74. Similarly where firms within a tentatively defined market ship less than 75 percent of their output to customers within the tentatively defined market, Elzinga & Hogarty would expand the market until at least 75 percent of the shipments by firms within the tentative market area are to customers within the area. Id.

Elzinga & Hogarty freely admit that "[t]he 75 percent figure is obviously somewhat arbitrary . . . . Those objecting to this benchmark can readily substitute a higher (or lower) one into the estimating procedure." Id. at 74-75. Indeed, the Elzinga-Hogarty test is too demanding in the sense that the relevant geographic market is often larger than the United States where imports account for less than 25 percent of sales, and in a follow-up article the authors advocated replacing the 75 percent test with a 90 percent test. Elzinga & Hogarty, The Problem of Geographic Market Delineation Revisited: The Case of Coal, 23 Antitrust Bulletin 1 (1978). Conversely, followed mechanically it could show a world market in circumstances where, for example, imports were fixed by a quota and therefore would have little ability to constrain price increases.

Nevertheless, the Elzinga-Hogarty test offers important insights. The purpose of market definition is to determine what is constraining prices. Obviously where there are no imports or exports, and no reasonable possibility of imports or exports in the event of a price increase, only competition from firms located in the United States constrains prices. Once there are some imports and exports, they begin to influence prices charged to U.S. consumers. The important contribution of the Elzinga-Hogarty test is to recognize that once imports or exports are significant, they influence domestic prices. While one can argue with the numbers they chose and while there are some circumstances in which their test may yield incorrect results, it represents a helpful and practical tool for looking at geographic market definition.

Landes & Posner argue for a general rule that, "if a distant seller has some sales in a local market, all its sales, wherever made, should be considered part of that local market for purposes of computing the market share of a local seller." Landes & Posner, supra, at 963. They contend that if the foreign firm can sell any "product in the domestic market, they ought to be able to sell many units at no appreciably higher cost, since they only have to divert output from other markets." Id. at 964. The Landes & Posner argument assumes that foreign producers can readily supply the U.S. market, id. at 963 and, "strictly speaking" that the goods sold by foreign producers are perfectly substitutable for domestically-produced goods. Id. at 965.[1] While objections have been raised to their model, it yields correct results under most circumstances. Applied correctly in the presence of imports, the Landes-Posner test yields market definitions broader than the United States where imports have been continuous and there are no significant trade barriers.

III. THE MERGER GUIDELINES

The Merger Guidelines state that a geographic market will be defined as the region in which a hypothetical monopolist could impose a "`small but significant and nontransitory' increase in price, holding constant the terms of sale for all products produced elsewhere." Section 1.12. The Guidelines can therefore yield geographic market definitions broader than the United States. They contain several caveats with respect to the definition of geographic markets and the assignment of market shares to foreign competitors: narrower geographic markets may be defined in the presence of price discrimination, Section 1.22; longer time periods than one year may be used for calculating market shares where exchange rates fluctuate significantly, Section 1.43; foreign market shares will be "capped" if an import quota is in effect, id.; and a single market share may be assigned to a country or group of countries if they "act in coordination." Id.

While the Merger Guidelines are analytically sound in their quest to define geographic markets by including in the definition of the relevant market those areas that are exerting a constraint on prices, the emphasis they place on market shares is analytically unsound. Market shares are at best only a proxy for market power, and where market power can be analyzed directly market share analysis is unnecessary and can often be misleading. See, e.g., Ball Memorial Hospital, Inc. v. Mutual Hospital Ins., 784 F.2d 1325, 1336 (7th Cir. 1986) ("Market share is just a way of estimating market power, which is the ultimate consideration. Where there are better ways to estimate market power, the court should use them."); Robert Pitofsky, New Definitions of Relevant Market and the Assault on Antitrust, 90 Columbia L. Rev. 1805, 1810-11 (1990) ("[I]t does not necessarily follow that a firm accounting for 90% of sales in a properly defined market has substantial market power. . . ."); Roscoe B. Starek III & Stephen Stockum, What Makes Mergers Anticompetitive?: "Unilateral Effects" Analysis Under the 1992 Merger Guidelines, 63 Antitrust L.J. 801, 804 (1995) ("In cases before the FTC it is not unusual to observe firms with high market shares that do not appear to be able to exercise market power.").

IV. A PROPOSED METHODOLOGY FOR DEFINING MARKETS IN THE PRESENCE OF IMPORTS

The paradigm case the Commission deals with is a proposed merger between two firms, both of which manufacture a good in the United States. The good is also manufactured abroad and imported into the U.S. by one or more firms. Whether the relevant geographic market should be expanded beyond the United States is an issue only if the merger would be challenged by the Commission if it concluded that the relevant geographic market were the United States. Under this scenario the Commission should ask three questions. First, how much would imports need to expand to defeat a price increase? Under the Merger Guidelines, "the Agency, in most contexts, will use a price increase of five percent" to determine the effect of an anticompetitive price increase. Section 1.11. As the Commission has explained, assuming an elasticity of demand of one, a five percent increase would be defeated by a five percent increase in output. See Owens-Illinois, Inc., 115 F.T.C. ___ (1992) (FTC Docket No. 9212, Feb. 26, 1992), slip op. at 32 n.36; R.R. Donnelly & Sons Co., ___ F.T.C. ___ (1995) (FTC Docket No. 9243, July 21, 1995), slip op. at 63 n.154. In the absence of good data showing demand elasticity, the Commission can assume for most purposes that the elasticity of demand is one. The key point is that "the focus of the analysis must be on the margin, i.e., on the extent to which a margin of customers will [switch to imported goods and] make the hypothesized price increase unprofitable." Starek & Stockum, supra, at 805.

Second, can foreign producers expand their shipments to the United States by an amount equal to five percent of current sales within the United States? In other words, assume that the worst happens and all U.S. producers collude. Would they be able successfully to exercise market power, or would foreign producers constrain the exercise of market power? Shipments into the U.S. could be readily expanded if foreign producers have excess capacity. They could also be expanded if sales can be diverted from a country where they are less profitable to the United States. While it can be difficult to determine precisely the amount of excess capacity and divertable capacity, in most cases it is possible to determine whether foreign producers could expand their U.S. shipments by five percent or more of U.S. sales. Frequently this will amount to only a small percentage of foreign capacity.

Third, are the foreign goods an acceptable substitute for products manufactured in the United States? This question will almost always be answered affirmatively where the foreign producers are currently selling in the United States, since by definition they have achieved market acceptance. See, e.g., Landes & Posner, supra, at 965 ("[E]ven in the case of a differentiated product, foreign output would be presumptively includable in the domestic market upon a showing that foreign imports were occurring."). The better the substitute that foreign goods are for domestically-produced goods, the more likely that foreign goods constrain prices in the U.S.

If foreign producers can ship goods that are acceptable to U.S. buyers in quantities sufficient to defeat any price increase by the merging firms, there is no need to compute market shares. By definition, the merging firms are unable to exercise market power because the non-merging firms are able to respond to the price increase "with increases in their own outputs sufficient in the aggregate to make the unilateral action of the merged firm unprofitable." Merger Guidelines Section 2.22.

Of course, the Merger Guidelines consider the possibility of coordinated interaction among firms as well as unilateral effects. In both Owens-Illinois and Donnelly the Commission explained that as long as firms accounting for only a very small part of industry output refuse to engage in coordinated interaction, any attempt to exercise market power will fail. Owens-Illinois, supra, at 32 n.36; Donnelly, supra, at 63 n.154. In the typical situation, coordinated interaction will be unlikely where there are imports from more than one country. Among other reasons, coordinated interaction among foreign firms is especially difficult because exchange rate fluctuations continually change the profitability of the terms of coordination that have been agreed to either implicitly or explicitly.

V. IMPLICATIONS OF ANALYZING THE ROLE OF FOREIGN SUPPLIERS DIRECTLY RATHER THAN THROUGH MARKET SHARES

It is important to note that this approach -- which is simply one way of determining whether supply is sufficiently elastic to defeat an exercise of market power -- counsels strongly against antitrust enforcement where products are characterized by large fixed costs. Products that require large investments in research and development, plant, and other sunk costs exhibit high price-to-marginal cost ratios. Supply of such products tends to be highly elastic because additional sales are highly profitable. Accordingly, foreign producers of such products have a strong incentive to respond to a U.S. price increase by increasing their U.S. sales. Indeed, they have a strong incentive -- even without a price increase -- to increase their output whenever possible. Looking directly at whether foreign firms constrain the exercise of marker power may, therefore, yield the conclusion that foreign firms exert a greater influence where products are differentiated rather than homogeneous.[2]

Another corollary of this method of analysis is that geographic price discrimination[3] outside the United States is largely irrelevant. Whether price discrimination is possible in some part of the world has little or nothing to do with whether foreign producers can constrain the exercise of market power within the United States.

Geographic market definition need not be symmetric. Indeed, there may be many instances where foreign producers constrain U.S. prices, but because of quotas, tariffs, or structural barriers U.S. firms do not constrain foreign prices. An example of this might be the automobile industry. Japanese automobile manufacturers were constraining U.S. prices even before many of them opened plants in the United States, but U.S. producers were not constraining prices in Japan. Only if margins were significantly higher in the United States, indicating that foreign producers were not constraining U.S. prices despite a profit motive to increase their imports, would price discrimination be significant. Otherwise the focus of the analysis should be on the sufficiency of excess capacity and divertable capacity to constrain U.S. prices, and not on the presence or absence of geographic price discrimination.

Endnotes:

[1] However, it is important to note that under their analysis, "even in the case of a differentiated product, foreign output would be presumptively includable in the domestic market upon a showing that foreign imports were occurring." Landes & Posner, supra, at 965.

[2] However, although producers of homogeneous products typically have much lower price-to-marginal cost ratios, i.e., lower profit margins, a five percent price increase may represent a very large increase in producers' profit margins, creating a strong incentive to expand or divert output.

[3] The 1992 Merger Guidelines incorrectly imply that price discrimination is "charging different buyers different prices for the same product." Section 1.12. Economists almost uniformly find no price discrimination where "differences in prices between consumers exactly reflect differences in the costs of serving these consumers. . . ." Jean Tirole, The Theory of Industrial Organization (1988) at 133-34. In other words, geographic price discrimination is charging prices that lead to different margins in different geographic areas.