Remarks Of Philip B. Nelson [1]

FTC Hearings on the Changing Nature of Competition in a
Global and Innovation-Driven Age

Introduction

This panel was asked to discuss how imports should be considered in an antitrust investigation and whether the Federal Trade Commission (FTC) should modify its current analysis of imports.[2] I will briefly address two issues: (1) Does the antitrust bar that regularly interacts with the FTC and Department of Justice (DOJ) on mergers and other antitrust matters perceive that there is a problem with the Antitrust Agencies' treatment of imports? (2) What inferences can be made from the presence or absence of imports?

Antitrust Bar's Perceptions Of The FTC's Treatment of Imports

I thought the FTC might be interested in knowing what the antitrust bar thought of the Antitrust Agencies' treatment of imports. As a result, I conducted a limited, nonrandom survey of active members of the antitrust bar who handled mergers that involved international issues. A questionnaire was sent to 176 lawyers who are either active members of the American Bar Association's antitrust section or who are known to practice before the Antitrust Agencies because they are clients of my consulting firm. Respondents were promised that I would not pass on their names or law firms to the FTC. The questions posed in the survey and the results are provided in the Appendix.

While only 14 lawyers (8% of the sample) responded, these lawyers together have made more than 370 merger filings since the new Merger Guidelines were adopted, over 220 of their filings led to an investigation by one of the Antitrust Agencies, and 17 of these matters involved disputes with the agency staff over geographic market definition issues that involved foreign-based manufacturers. Moreover, there is sufficient variation in their answers that the results provided some interesting insights into the interaction between the Antitrust Agencies and the private bar in cases that involved imports.

Findings that deserve particular emphasis include:

• When there was a disagreement between an attorney representing a merging firm and the investigating attorney about whether the geographic market was international or not, it was usually the case that the investigating attorney was advocating a narrower geographic market.

  - In written comments, one respondents elaborated on their view that the agency staff often used too narrow a market, commenting: "Ease of entry and expansion are virtually ignored by the agencies in foreign output analysis, but are often highly visible factors in a merger." Others suggested that the burden of proof was high if the merging parties wanted to show that imports were a realistic competitive threat.

  - However, in some cases, such as Nippon Sheet Glass's acquisition of an interest in Libbey-Owens-Ford (involving "flat glass"), the reverse has been the case. Here, the FTC was concerned about contractual restraints on Asian imports even though current imports represented a small percentage of U.S. purchases and the parties wanted to narrow the geographic market.

• The disagreements over the geographic scope of the market result from differences in opinion about the facts, the appropriate analytical approach, or both.
• Some attorneys believe that the presence of imports necessarily means that additional foreign capacity should be included in the market. Most believe that this is only "sometimes" the case.
• Some attorneys believe that the presence of imports indicates that all of the capacity of foreign manufacturers should be included in the relevant geographic market.
• Only one attorney believed that the absence of imports clearly signaled that foreign capacity should be excluded from the relevant geographic market.
• Three attorneys believed that the Antitrust Agencies' geographic market policies were poorly defined. Most thought that they were as well defined as is possible, while four thought they were well defined.
• The written comments by the respondents appeared to recognize that detailed, market-specific factual information was required to assess the competitive significance of foreign producers.
• Some written comments suggested that the Antitrust Agencies' staffs did not reveal their belief about the geographic extent of the market and the reasons for these beliefs (e.g., "Only rarely do I emerge from a transaction believing I understood what had driven the staff. Both sides tend to be pretty 'close mouth' -- the standard model is that of litigation, i.e., there is little candid exchange of views because there may be litigation. Most of the time I only guess as to what they are thinking. I think the process would be enormously advanced if both sides were expected to reveal what their concerns were--what their own analysis was. I just don't see how you reconcile that with a litigation model.").

Conclusions That Can Be Drawn From The Absence or Presence Of Imports

As I have indicated in an article that I wrote with George Hay and John Hilke,[3] the absence of imports does not necessarily imply that foreign producers can not limit the exercise of market power in the United States. It is also true that the presence of imports does not mean that foreign producers can effectively discipline a U.S. monopolist or cartel. Only careful analysis of the structural conditions of particular markets and the behavior of those markets over time will allow one to interpret import patterns correctly.

Foreign producers can affect the market power of domestic firms even where foreign producers make no sales in the U.S. The most obvious reason for this is that foreign producers, who do not find it profitable to sell in the United States at current competitive prices, may find it profitable to sell at prices that are only slightly about current levels. However, it is also true when: (1) U.S. firms find it profitable to serve foreign markets; (2) they cannot price discriminate between U.S. consumers and foreign consumers; and (3) they must charge competitive prices to compete with foreign suppliers in foreign countries.

The presence of imports does not necessarily imply that the threat of additional imports will be sufficient to prevent a domestic monopolist or cartel from raising prices above competitive levels. The simplest case where this is true is when foreign suppliers are subject to a binding import quota that prevents them from expanding sales in response to a monopolistic price increase. However, it is also true when other structural characteristics of the U.S. or foreign markets limit the foreign producers ability to expand sales in the United States.

Structural characteristics that will limit the ability of increased imports to discipline U.S. suppliers despite the presence of ongoing imports include:[4]

• The presence of quotas. As was suggested above, quotas that set a ceiling that are at or near the current import level restrict the ability of foreign firms to respond to increased U.S. prices. Quotas that require imports to be less than a fixed percentage of U.S. sales are particularly restrictive because, if the quota is binding, imports will have to be cut in response to a monopolistic price increase that results from a reduction in U.S.-based firms' sales.
• The presence of foreign governmental subsidies that support current import levels, but which will not be expanded to support incremental imports.
• Product differentiation may mean that imports can not discipline the market. In a differentiated product market, the presence of imports does not imply that foreign producers are well-positioned to compete at competitive prices if current prices have been raised by an incumbent monopolist to prices above the competitive level (the "cellophane fallacy"). Even in a competitive market, it may be that imports are not a serious threat to prevent a post-merger price if few U.S. consumers view imports as an adequate substitute (since imports may address unique tastes and not be popular with the core market).
• The presence of imports, particularly transitory imports, does not necessarily indicate that foreign manufacturers have the distribution network and other infrastructure that are needed to support increased sales. Even if an importer has a distribution network, this network may not be as cost-effective as U.S. manufacturers' networks, which will limit the competitive influence of the importer. In particular, importers may not be able to provide cost-effectively the delivery services that key U.S. retailers demand.
• The presence of imports need not signal a significant competitive effect of the imports on all U.S. suppliers if the imports are concentrated along the U.S. coast, since it may be that inland portions of the U.S. are insulated from imports by high overland transportation costs.
• If importers have increasing production costs that are steeper than those faced by U.S. manufacturers (or have capacity constraints) and there are significant transition costs associated with redirecting output from one country to another (including contractual restraints that commit foreign production to foreign customers), the import elasticity may not be large enough to prevent a monopolistic price increase.
• If importers are required to make significant "lumpy" incremental sunk investments to expand U.S. sales and U.S. firms can quickly lower their prices to respond to competitive incursions, importers may not find it profitable to expand their U.S. sales significantly if their is a monopolistic price increase in the United States.
• When foreign markets are highly concentrated, strategic considerations, such as the fear that U.S. firms will retaliate by entering foreign firms' home markets, may discourage foreign firms from aggressively expanding sales in the United States.

The observation that the presence of imports does not necessarily signal that significant additional imports will appear if there is a monopolistic price increase is important because it implies that the proposal made by William Landes and Richard Posner,[5] 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,"[6] is flawed. This observation (along with the observation that the absence of imports does not imply that imports are not a serious competitive threat) is also important because it implies that geographic market definition tests that are based on shipment patterns, such as the Elzinga-Hogarty Test,[7] can be misleading unless they are used with care.

While the survey I conducted for the FTC's Hearings (see above) suggests that much of the private antitrust bar recognizes that there are problems with focusing on shipment patterns to the exclusion of other structural characteristics of markets, there is not universal agreement on this. Indeed, some courts have not been clear on this point. For example, the District Court for the Southern District of New York recently uncritically cited both the Landes-Posner and the Elzinga-Hogarty articles in their Kodak decision. However, even in Kodak, the Court did more than rely on shipments pattern data. It dismissed the Department of Justice's allegation of longoing international price discrimination not only because of ongoing shipment patterns, but because it believed there was evidence of homogeneity of film products, evidence of similar retail prices for imports and U.S. products, and evidence of fairly price elastic demand. The Kodak Court dismissed the DOJ's evidence of international variations in prices largely because the court felt that the price data covered too limited a time period.

Policy Implications

Because foreign producers may have the ability to increase U.S. sales either by increasing their output or by diverting sales from other parts of the world, a given level of imports may represent a more serious competitive threat to a dominant firm or collusive group than a similar level of "domestic fringe" sales. Moreover, foreign producers can discipline U.S. producers even without importing into the United States. For example, they can provide U.S. fringe producers with know how that will allow them to compete effectively. In addition, as was pointed out above, in markets where geographic price discrimination is not possible, nonimporting foreign manufacturers can limit U.S. prices by setting competitive prices in foreign countries where U.S. firms find it profitable to compete.

While imports often signal that foreign producers can discipline U.S. firms, this need not be the case. As was detailed above, numerous factors can limit the foreign elasticity of supply and reduce the competitive effect of imports on U.S. prices.

Because of the underlying economics, the Antitrust Authorities must look carefully at the institutional characteristics of the industries that they are investigating. Only by posing the right questions to customers and other industry participants, will the Antitrust Authorities be able to sort out the effect that imports have on competition. Otherwise, there will be more United States v. Baker Hughes cases (where courts find, contrary to the Antitrust Authorities view, that customers will turn to imports) or unwarranted consent agreements. Moreover, in compiling the results of their investigations, the Antitrust Authorities must be careful that key relationships are not lost in the distillation of the information. In particular, while it is helpful to define markets so that concentration levels can be calculated and compared to concentration benchmarks that define zones where mergers will not be challenged, it is crucial that the underlying elasticity relationships be preserved for additional scrutiny if a merger proves to be in what appears, at first blush, to be a concentrated market.

The survey reveals a limited number of complaints about the Antitrust Agencies' handling of mergers that occur in markets with imports. Indeed, one respondent even commented "if its not broke, don't fix it." However, there are some complaints. In particular, there is some sentiment that at least some staff investigators require too high a standard of proof when testing the hypothesis that foreign firms can discipline U.S.-based competitors. On the other hand, courts are still relying on articles that oversimplify the analysis of geographic markets. As a result, it appears that the FTC could perform a valuable services by researching and documenting more fully the economic circumstances under which shipments pattern data can be misleading. In particular, I think it would be very helpful for the FTC to consider how product differentiation might allow a dominant firm to price discriminate across different geographic areas even when it serves these areas from a single plant. Particularly if this research is undertaken outside of the litigation context, it may do much to clarify the weight that should be assigned to shipments data in antitrust investigations and cases.

Endnotes:

[1] Philip Nelson is a Principal at Economists Incorporated. The opinions expressed here are his own and do not necessarily reflect those of other economists at Economists Incorporated.

[2] Specific questions that were suggested include: H ow should antitrust assess the role of imports in markets? Does current FTC practice capture the dynamics of global markets or should modifications to current practice be considered? Which modifications and why?

[3] George Hay, John Hilke, and Philip Nelson, Geographic Market Definition In An International Context 64 Chicago-Kent L. Rev. 711 (1989).

[4] For additional explanation of many of these structural characteristics, see Id.

[5] William Landes and Richard Posner, Market Power in Antitrust Cases, 94 Harv. L. Rev. 937 (1981).

[6] Id. at 963.

[7] Kenneth Elzinga and Thomas Hogarty, The Problem of Geographic Market Definition in Antimerger Suits, 18 Antitrust Bulletin 45 (1973). These authors propose a test for geographic market definition that relies on import and export data.

APPENDIX

INSTRUCTIONS

This confidential questionnaire has been prepared to obtain information about the private bar's perspectives on the FTC's and DOJ's analysis of mergers in which the relevant geographic market may include foreign-based suppliers. It should take about half an hour to complete. Answers will not be attributed to particular lawyers or law firms. Philip Nelson will tabulate the results for presentation at upcoming FTC hearings. If you attach your business card to the response, a copy of the findings will be sent to you in return for your participation in the project.

Only attorneys who represented private clients that filed HSR papers with the FTC or DOJ after April 2, 1992 (when the current Merger Guidelines were announced) are to respond to this questionnaire. All answers are to reflect your experiences with the HSR process during the 1992-1995 period.

Please return your questionnaire (using the enclosed envelope) to:

Philip Nelson
Economists Incorporated
Suite 600
1233 20th Street, N.W.
Washington, D.C. 20036

If you have any questions please contact Philip Nelson at (202) 223-4700.

QUESTIONS

[Note: Small letters at left of essay responses are used to distinguish answers from different attorneys.]

1. During the April, 1992-1995 period, have you worked as a private attorney on an acquisition or joint venture that involved a Hart-Scott-Rodino (HSR) filing? (Note: If no, please do not complete this form.)

Yes 14 [100%]

No 0

2. During the April, 1992-1995 period, how many HSR filings did you work on?

___374 [Avg. 26.71] ______

3. Of these filings, how many were cleared for investigation by the FTC and how many were cleared for investigation by the DOJ?

____146 [Avg. 10.44] _____ FTC Clearances

____177 [Avg. 5.46] ______ DOJ Clearances

4. Of the filings identified in Question #2, how many involved a manufactured product that you believed was sold in a relevant geographic market that included foreign-based manufacturers who should be included in any HHI calculations?

__112 [Avg. 8]_______

5. Of the filings identified in Question #2, how many involved a dispute in which you believed that the relevant geographic market and HHI calculations should include foreign-based manufacturers and the lead investigating attorney disagreed?

____17 [ Avg. 1.21]_____

6. Of the filings identified in Question #2, how many involved a dispute in which the lead investigating attorney believed that the relevant geographic market and HHI calculations should include foreign-based manufacturers and you disagreed?

___0_____

7. For those filings where there was a disagreement between you and the lead investigating attorney over the definition of the relevant geographic market and calculation of HHIs, what was the source of the disagreement?

Different understanding of facts 2 [22.22%]

Different standards for market definition or other analytical difference 2 [22.22%]

Both different factual understanding and different analytical approach 4 [44.44%]

Other If "other," explain in space below: 1 [ 11.11%]

b. How to weigh significance of imports/supply response

c. This kind of basic disagreement just didn't arise. Many of our deals involve non-U.S. companies. Perhaps the answer is just obvious.

8. If the relevant product is currently being imported into the U.S. from a foreign country, should additional output/capacity of foreign manufacturing firms that are currently exporting to the U.S. be included in the relevant geographic market and in any HHI calculations?

Yes 5 [35.71%]

No 0

Sometimes 9 [64.29%]

9. If you answered the preceding question with "sometimes," what factors do you believe should be considered to determine whether to include additional output/capacity of foreign manufacturing firms that are currently exporting to the U.S. in the relevant geographic market and in any HHI calculations?

b. Capacity constraints abroad/excess capacity; opportunistic behavior; U.S. based distribution network. -- existence or a lack thereof; exchange rates/export response to fluctuations

c. Dumping orders; sensitivity to price increases of 5%-10%; industry sector history; currency issues; # user sector restrictions; if any, or use of foreign products

d. Generally, conditions of entry and ability to expand output, pricing trends and regulatory issues, if any. Ease of entry and expansion are virtually ignored by the agencies in foreign output analysis, but are often highly visible factors in a merger analysis.

e. Ease of supply shift to U.S.; readily available unused capacity; produceable at relatively low marginal cost.

f. Extent to excess capacity that could readily be applied to U.S.; barriers; track record of foreign suppliers in U.S.; volume swings for imports.

h. Whenever the product is being sold in the U.S. to customers for which U.S. producers compete

k. Restrictions, both legal and otherwise, on additional imports into the United States

m. Likelihood that additional product actually would be exported to U.S. (depends on ease of additional entry, degree of excess capacity, etc.) in the event of some percentage price increase.

n. If it can switch capacity to supply U.S. within reasonable time frame

10. If the relevant product is not currently being imported into the U.S. from a foreign country, should any foreign-based manufacturers' output/capacity be included in the relevant geographic market and in the calculation of HHIs?

Yes 0

No 1 [ 7.69%]

Sometimes 12 [92.31%]

11. If you answered the preceding question with "sometimes," what factors do you believe should be considered to determine whether to include foreign-based manufacturers' output/capacity in the relevant geographic market and in the calculation of HHIs?

a. Currency changes/ Import restraints

b. Prior history; exchange rates; distribution requirements (US.); product recognition, e.g. generic, commodities, specialized.

c. Rarely - but you would want to understand how and why they would enter. What about 5%-10% price increase etc. This could only be a "make-weight" argument - if this is all you have, you've got a problem.

d. See response to No. 9.

f. Past imports from that country; strengths of the supplies of that country; extent of excess capacity; profit opportunity for importers from that country.

h. Whenever the foreign producer is an uncommitted entrant and there is evidence that entry is likely in response to a domestic price increase

i. If the foreign manufacturers could profitably begin selling into the U.S. in response to a SSNIP.

j. See answer to 9 above

k. Reason why imports not now occurring; whether post-merger, such imports could occur

l. Announced intentions of foreign companies; expressions of concern by domestic companies over possible entry of foreign imports; lack of significant legal barriers; freight costs; degree to which industry is new or rapidly changing

m. Same as #9 (although the likelihood that the product would be exported to the U.S. would generally be greater under the facts presented in #8 rather than #10).

n. same answer as 9

12. If, in your view, a firm that manufactures the product in a foreign country is properly included in the relevant geographic market, should the output/capacity of all of the manufacturers of the relevant product in that country be included in the relevant geographic market and in the calculation of HHIs?

Yes 2 [ 15.38%]

No 1 [7.69%]

Sometimes 10 [76.92%]

13. If you answered the preceding question with "sometimes," what factors do you believe should be considered to determine whether to include all of the manufacturers' output/capacity in the relevant market and in any HHI calculations?

b. Different marketing incentives; distribution (U.S.); Characteristics of product: see 11, above; import history

c. You have to understand why some do and some don't. -- What determines what column they are in, [is:] would a 5%-10% increase bring them in? If there is no history - it's going to be a tough argument.

d. See response to No. 9

f. Capacity, efficiency and incentive of each manufacturer

h. See #11

i. If they could begin selling in U.S. in response to a SSNIP.

j. See answer to 9 above

k. Depends upon whether other offshore manufacturers could export to USA if price of product to rise, etc. post-merger.

m. See #9

n. Same as #9

14. Do you believe that the antitrust agencies' geographic market definition policies are poorly defined so that staff attorneys do not have clear guidance about how to define relevant geographic markets?

Yes, poorly defined 3 [21.43%]

As well defined as is possible 7 [ 50%]

No, not poorly defined 4 [28.57%]

15. Do you believe that the Federal antitrust agencies' analysis of the relevant geographic markets differs from yours:

(1) when the issue is the inclusion of additional output/capacity from foreign manufacturers that are currently importing into the U.S. (Qs 8 & 9 above);

(2) when the issue is the inclusion of foreign-based output/capacity in the relevant geographic market when there currently are no imports (Qs 10 & 11 above); and

(3) when the issue is the inclusion of all manufacturers located in a foreign country when at least one such manufacturer is properly included in the relevant geographic market (Qs 12 & 13 above)?

If so, how? If your experiences with respect to these three geographic market definition issues differs significantly between the two Federal antitrust agencies, please describe these differences.

5 respondents agreed with (1); 4 respondents agreed with (2); and 2 respondents agreed with (3). Some of respondents agreed with more than one proposition, at least sometimes.

b. [The third point] hasn't really come up. Usually talk about current & prospect of others; but not some, therefore all. DOJ is usually more open-minded; Pitofsky may help.

c. When you are trying to argue that the HHI numbers do not reflect actual market power, you tend to be more expansive or aggressive in your thinking than the staffs. My experience is that the staffs seldom make their thinking explicit. You either get a Second request or you don't -- an opposition or a pass. It is not easy to really understand what the analytic process needs

d. Yes as to each subpart of the question. Both agencies tend to be highly physical if not hostile to "speculative" foreign competitive impart. The staffs generally do not view foreign "entry" as impact unless it is demonstrated, actual U.S. sales activity.

e. Don't give sufficient weight to shiftable production imported from abroad; good example is steel merger initially opposed by [DOJ]

f. Generally, I have come to expect staff to count only actual imports in the calculation of market shares. If the agencies are open to a broader concept, it is not well reflected in complaints.

i. No

j. It does not differ, it is merely a sliding scale about which reasonable people can disagree

l. No opinion

n. No consistent policy on agencies part.

16. What suggestions, if any, do you have for improving the antitrust agencies' analysis of relevant geographic markets when foreign manufacturers might be able to supply U.S. consumers?

b. Markets work, sometimes not in the short term.

c. We have a high percentage of non-U.S. clients. I am not clear that the system is "broke" i.e., I'm not sure it needs "fixing".

d. Two thoughts (1) better discovery tools for foreign conduct. The staffs find it simpler to conduct hundreds of U.S. interviews than a handful of foreign ones. (2) more realistic assessment of likelihood of entry, expansion, import restrictions etc. The staffs seem to believe that foreign producers that are not selling in the U.S. are not doing so because of a cost/regulatory barrier of some type and therefore entry is dismissed out of hand

e. Get better data on supply substantiality in particular industries.

f. You are dealing with a subject that the agencies rarely talk about. Their treatment of markets is obscured by the sparse, conclusory language of their complaints and their failure to address the issue in competitive impact statements.

h. Not enough problems in any experience to evaluate

j. better info is always preferable

m. A comprehensive listing of factors that may be considered in determining the likelihood that product (or additional product) would be exported to the U.S., e.g.: degree of excess capacity; economies of scale; currency fluctuations; trade barriers; cultural barriers and mobility barriers; U.S. Government and state regs.; buyers convenience/ preference; ability to price discriminate; etc.

n. Many of your prior questions need to distinguish between a world market and a U.S. market in which imports are counted. If world-wide, current importing practices may not be relevant. If U.S., then importing practices and potential are crucial. Under Section 7 ("... in any section of the country"), is a "world" market statutarily recognizable?

17. Should the antitrust agencies consider different economic factors when defining the "relevant geographic market" than they consider when determining if "entry is easy"?

No 7 [58.33%] Yes If yes, explain how entry analysis should differ from geographic market definition analysis below. 5 [41.67%]

b. When I think of entry, I think of capital commitments to develop a product and expertise to do same. When I think of imports, I think of comparable quality, price at existing prices, distribution needs. The capital commitment, while it may be substantial is different.

c. The latter is the "who" of it. - The former is the "how" or the "when". The analysis does tend to converge as you seek to expand the geographic market beyond "current incumbents" - i.e., if, by definition, you are trying to "expand" the geographic market - the issue is largely "what would it take them to enter ." Only rarely do I emerge from a transaction believing that I understood what had driven the staff. Both sides tend to be pretty "closed mouth" - the dominant model is that of litigation i.e., there is little candid exchange of views because there may be litigation. Most of the time, I can only guess as to what "they" are thinking. I think the process would be enormously advanced if both sides were expected to reveal what their concerns were -- what their own analysis was. I just don't see how you reconcile that with a litigation model.

e. "Need also to look at potential for trade law protectionism"

h. Bottom line is probably the same if one ignores distinction between committed and uncommitted entrants.

i. The 1992 Guidelines have it right in distinguishing between "uncommitted entry" for purposes of market definition and "committed entry" for purposes of entry analysis, the distinction being the existence of sunk costs.

l. Arguably in import context. If imports have not yet come in and never have, and freight is significant, foreign firms may not be "in this market." But entry may be easy if they can easily set up shop in (or near) U.S. and eliminate freight issue.

m. No, except perhaps in the rare situation in which foreign production and/or capacity has a demonstrably price-limiting effect in the U.S., even without actual entry into the U.S.

n. Isn't the answer the difference between uncommitted supply response and entry de novo -- which primarily is one of time periods. (1 yr vs. 2 yrs)?