Printer Friendly

Economics, the guidelines and the evolution of merger policy.

I. Introduction

The last 25 years has seen a dramatic evolution of IL& merger law moving from a situation in the mid-1960's where almost any significant horizontal merger could be illegal to a situation where only a few horizontal mergers are illegal. Most of the change occurred in the Reagan deregulation in the 1980's. However, unlike other forms of deregulation over the last 25 years, merger law evolved without any congressional action and with little Supreme Court guidance. Instead, the adjustment in merger regulation can be described as the endogenous response of the legal system to changes in both the announced public policy on mergers and the intellectual understanding of the competitive market system. By focusing on the language of the Clayton Act to proscribe only mergers that may substantially lessen competition or tend to create a monopoly, the courts have changed the case law to go beyond a narrow reading of the Department of Justice (DOJ) Merger Guidelines and approximate the current economic thinking. This evolution of merger policy through the courts has saved society the resources necessary to explicitly change the law and may have contributed to the growth in the economy during the 1980's.

After a brief presentation of the economic issues concerning mergers, this article evaluates the impact of the DOJ Merger Guidelines in the evolution of merger policy. The discussion starts with the historical approaches to market definition culminating in the hypothetical price analysis presented in the Guidelines. Then, the article reviews the attention given to this approach in recent merger cases and evaluates the Guidelines' effect on the reformation of the merger law. Next, competitive analysis within the relevant market is discussed. The article reviews the evolution of merger analysis and highlights the importance of the Guidelines in identifying when a merger is likely to substantially lessen competition. The bulk of the analysis concerns how these standards have been applied to recent merger decisions. A brief conclusion summarizes the results.

II. Economics in merger analysis

The economic theory used to predict the likely consequences of a horizontal merger has undergone a dramatic change over the last 25 years. In the 1960's, antitrust thinking was dominated by the "Harvard," or "structuralist" school, which considered concentration synonymous with market power. The perfectly competitive economy of atomistic competition was the ideal and concentration was only thought to be justified if clear economies of scale existed. Regulators and economists even went as far as to advocate a deconcentration policy, under which oligopolistic firms would be forced to divest large amounts of capacity.(1) This push for deconcentration invigorated a debate in which partisans of the "Chicago school" such as Robert Bork highlighted a number of underlying assumptions, which once exposed, were eventually rejected by the economics profession.(2)

Bork lists three crucial assumptions of the Harvard school and explains why none of them is likely to be true.(3) First, proponents of industry restructuring assumed that a significant reduction in output was the probable result of concentrated markets. Bork summarizes the statistical and case history evidence and observes that "the available evidence strongly suggests that oligopolies do not generally result in substantial or significant reductions in output."(4) Second, it was assumed that the dissolution of concentrated firms would be unlikely to destroy significant efficiencies. While a number of studies do show limits on plant-level economies of scale, this only represents one type of efficiency. Bork observes that firms obtained their large size in competition with their rivals so any size obtained by a firm (without predation) is efficient. Third, policy activists assumed that artificial barriers to entry served to prevent the growth of small firms. Bork states, "[I]t is clear that no such class of artificial barriers exist."(5) If firms are free to compete, price cannot be maintained above long-run cost. Of course, it is generally recognized that barriers can stem from other government action. This fact, however, argues for changing government policies instead of a general policy of deconcentrating industries. All three of Bork's points have been tested and generally accepted by economists, although the importance assigned to each observation may differ.(6)

While structuralist authors generally admit the truth of the Chicago points, they still appear to believe that there is usually a relationship between very high concentration and monopoly profits, while the Chicago approach suggests that the factors that allow cartelization in an oligopoly are much more complex. At the risk of oversimplification, the Harvard school appears to believe that high concentration can almost be considered a sufficient condition and the Chicago position would be characterized as high concentration is only a necessary condition for merger to generate an anticompetitive effect in a market exhibiting entry barriers.(7)

At the beginning of the 1980's, it was clear that official merger policy was mired in the 1960's.(8) As part of the general policy of regulatory reform, the Reagan administration's DOJ issued new merger guidelines that brought public policy more into line with economic thinking. As one shall see, even these reforms fell short of implementing all the Chicago policies.(9) Thus, the courts were faced with two decisions: first whether to validate the DOJ's reform of the merger laws, and second, whether to enforce a policy based entirely on the new learning in industrial organization. In the next two sections, we evaluate the court decisions to determine how these questions were, resolved.(10)

III. Market definition in merger analysis

It is generally recognized that market definition is crucial for defining an optimal concentration-based merger policy. The courts have struggled with the issue for years, while economists, with rare uniformity, have suggested that the analysis be based either directly or indirectly on the demand and supply elasticities. Supreme Court decisions in Dupont and Brown Shoe advanced legal concepts, but these decisions have served to further confuse the analysis.(11) The courts have been left with little guidance as to how to define markets and, as Posner describes, they have made numerous mistakes.(12)

In an effort to structure market definition analysis, the 1968 version of the Merger Guidelines introduced the idea that the sale of any product or service that is distinguishable as a matter of commercial practice from other products or services will ordinarily constitute a relevant product market. A similar idea was advanced for geographic market, although the market would be enlarged if no clear barrier hindered sales into the region. Posner notes this first DOJ attempt to standardize market definition was not particularly successful.(13) The suggested product market definition was overly vague as to when demand-side substitutes should be in the market. Likewise, the geographic market standard failed to actually define when a substitute product should be in the market. As Werden observes, the relevant market in any particular case could be "essentially as small or as large as one wanted."(14) Moreover, the analysis hardly considered supply-side substitution. Although the geographic market approach gave some consideration to supply-side shifts, it rejected firms that could compete in the market if the price increased by a small amount. Posner suggests that a firm's sales be excluded from the market only if the price charged by the firm plus the transportation cost into the area in question appreciably exceed (for example by more than five percent) the current market price in the local area.(15)

In the 1982 revision of the Merger Guidelines, the DOJ generalized Posner's approach to apply to both product and geographic market definition. Simply put, firms should be considered in the relevant market if consumers can turn to their products in response to what the 1984 revision labels a "small but significant and nontransitory increase in price." This is generally taken to be a five percent price increase lasting 1 year, although others claim a 10% price increase has been used in practice.(16) To define a market, the government searches for evidence that firms outside the provisional market would or would not expand output sufficiently to render a small price increase unprofitable. If the evidence suggests that the price increase would be defeated, the analysis is repeated with a broader provisional market, while if the price increase is sustained, the provisional market is accepted. The Guidelines note that indirect evidence, such as buyer's perceptions, differences or similarities in price movements, differences or similarities in the products, and seller's perceptions should be considered in the analysis. Also, supply-side substitution should be evaluated in determining which firms are in the market. As a final consideration, narrow markets will be considered if a hypothetical monopolist could price discriminate against a group of users with inelastic demands for the product.

Overall, the revised Guidelines advance a relatively objective standard on which to evaluate the market evidence. Moreover, the Guidelines highlight some characteristics as useful in market definition and by omission implicitly reject others mentioned in the Brown Shoe decision.(17) Finally, the Guidelines recognize the importance of supply-side flexibility in market definition. Thus, the Guidelines generally call for the use of economically meaningful markets, that is markets that can be subject to the exercise of market power.

The specific Merger Guideline five-percent technique has not met with broad acceptance in the federal courts. Only Federal Trade Commission (FTC) decisions usually mention the Guidelines in their market definition analysis, while the district courts rely on market behavior in drawing product and geographic market boundaries. For example, in Calmar, the court ruled that the DOJ markets of specific types of spray dispensers were too narrow.(18) The DOJ's anecdotal testimony (of responses to price increases) was considered to reflect the inertia of any manufacturer when presented with a change in circumstances. The court found the market evidence that different manufacturers of similar products had chosen different dispensing technologies more persuasive. Likewise, in Owens Illinois the district court ruled that a few inelastic glass bottle users were not enough to support a narrow market.(19) Even in the narrow segments, the court noted the opinions of industry participants should be viewed in light of their market behavior. Donnelley represents a third example, with the court ruling "isolated segments with isolated customers do not make for a separate product market."(20) Instead, the court found that the business realities suggested that both offset and gravure printing technologies compete with each other. So what we have is judicial resistance to applying the five-percent test to define narrow markets based on the hypothetical testimony of buyers, sellers and economists.(21) of Course, this is only a problem when the historical behavior of an industry conflicts with the testimonial evidence presented by the government.

In many cases, the economic evidence is generally consistent, and the Guidelines appear to have had a positive impact on the market definition process. The government's quasi-empirical technique limits its ability to gerrymander the market definition to maximize concentration. By setting an example, the government's actions serve to shift the debate from legal to economic issues.(22) This should reduce the market definition controversy and lead to stipulations on market definition, as both the government and the respondents focus on the same issues. One can look at the data for confirmation of the idea that the parties do not contest every market definition. Moreover, it is possible to examine the government's success rate on the merits to determine if the courts ire receptive to the government's, economic-based analyses.

In 13 of the 33 product markets and 19 of the 33 geographic markets, the respondents either stipulated to the government's market or failed to provide any evidence that the judge felt was persuasive enough to even mention in the formal opinion. The likelihood of explicit or implicit stipulation does not vary significantly over the time period.(23) Thus, the government seems to have had reasonable success throughout the 1980's in defining consensus markets.

One can also examine the government's record on contested market definitions. We reviewed all the cases and identified when the government's market prevailed or was rejected by the court. In a few cases, the judge chose a compromise market, but these markets were always in line with the principles advanced by one of the parties. The government appeared to prevail on 12 of the 20 contested product markets and 7 of the 14 contested geographic markets. Again, the evidence shows no real difference between the government's overall success rate over the time period.(24) This brings the government's overall record to 25 successes for product market and 26 successes for geographic market. In conclusion, the data suggest that even though the courts have declined to explicitly use the Guidelines' methodology, the government still prevails on approximately 75% of the market definitions.(25)

In addition to advancing the price test, the Guidelines suggested that supply-side economic evidence should be introduced into the analysis of product market. This goal appears to have been accomplished. Supply-side substitution has played an important role in a few product market decisions. For example, in Waste Management the court considered the supply-side substitution between firms in the roll-off and front-end commercial trash collection business in defining a broad commercial trash market.(26) However, the market fell short of the respondent's preferred market of all trash, because there were sufficient distinctions between commercial and residential trash collection. Likewise, in Occidental Petroleum, the court declined to consider PVC homopolymer and PVC copolymer to be in separate markets, because firms could switch from one to another with at most a small investment.(27) Supply-side considerations have also obtained attention as a form of entry.(28) For example, in Baker Hughes, the court excluded many foreign firms from the U.S. market, but considered them to be potential entrants.(29) Likewise, in Syufy, the market was limited to first run movie exhibition, but the second run competitor was incorporated in the analysis as an actual entrant.(30)

In conclusion, merger decisions on government cases filed after 1982, while declining to explicitly implement the Guidelines' five-percent procedure, appear to have increased the importance of economic analysis. Overall, the government's arguments are successful roughly 75% of the time for both product and geographic market. Moreover, supply-side issues regularly play a role in the analysis.

Of course, losing on a market definition issue does not always prove fatal to the case, because even the market alleged by the respondents may be concentrated. In fact, only seven of the government's 17 losses can be attributed to the failure to prove a market and even in these cases the courts tended to find against the government even in their narrow market.(31) Thus, it seems that the choice of the market is not always crucial to the evaluation of challenged mergers. (Of course, under the Guidelines, if the government does not define a concentrated market, the merger will not be challenged)

IV. Competitive effects in merger analysis

In the 1960% be general hostility to horizontal mergers was illustrated in a number of Supreme Court decisions such as Brown Shoe (although vertical considerations played a role in the decision), and Von's Grocery.(32) These decisions condemned mergers between horizontal competitors with only minimal market shares and advanced the idea that incipient market power should be blocked even if the transactions are efficient. In the related and more reasonable Philadelphia National Bank decision, the Court found that high concentration established a rebuttable presumption of illegality.(33)

The 1968 DOJ Merger Guidelines formalized the strict antitrust policy of challenging acquisitions that involve even small competitive overlaps. For highly concentrated markets (four-firm concentration ratio over 75%) a firm with a share of 15% could not buy a rival with a share over one percent. Moreover, no two firms with shares over four percent could merge. The rules were only slightly less draconian for relatively unconcentrated markets and in light of the concern with incipiency, no antitrust safe harbor existed. For example, the acquisition of a four-percent competitor could be challenged if the acquiring firm had more than 10% of the market. Likewise, share combinations of 5 and 5, 15 and 3, 20 and 2, or 25 and I were all subject to challenge. Efficiencies, while not explicitly condemned, were only to be considered in extreme cases.

In the General Dynamics decision, the Supreme Court retreated from its extreme structuralist position by requiring an examination of nonmarket share issues.(34) The Court held that full consideration of a markets "structure, history and probable future" was necessary to evaluate the probable competitive effect of the merger. This could be considered a reaffirmation of the Philadelphia National Bank decision, however the identity and weight to be given to the other factors remained unclear.(35)

A review of the Philadelphia National Bank and General Dynamics decisions would suggest that the presumption of an anticompetitive effect begins when the acquiring firm's postmerger share exceeds 20%-30% and the two largest firms exceed 40%-60%. These market shares are dramatically higher than the figures from Brown Shoe and Von's Grocery, but compatible with some forms of oligopoly theory. One can translate the share figures into a Herfindahl range of 1,400-2,200.(36)

In the 1982 revision of the Merger Guidelines, the DOJ updated its antitrust policy and tried to quantify some of the factors that should be considered in interpreting concentration. The revision established a safe harbor for mergers in markets with postmerger Herfindahls of under 1,000 and a doubtful zone for mergers in markets with postmerger Herfindahls over 1,800 (these figures roughly correspond to four-firm concentration ratios of 50 and 70, respectively). Then the Guidelines followed the 1968 version by suggesting that the DOJ is likely to challenge mergers that increase the Herfindahl by 100 points in moderately concentrated markets (Herfindahl between 1,000-1,800) and 50 points in highly concentrated markets (Herfindahl over 1,800).(37)

The decision to challenge a merger would be reversed if other factors implied that the merger is not likely to substantially lessen competition. However, the Guidelines suggest that the other factors will only offset a decision to challenge a merger that increases the Herfindahl by more than 100 points to a level substantially above 1,800 in extraordinary circumstances. Thus, one could interpret the Guidelines as creating three regions, the first where mergers are allowed (Herfindahls under 1,000), the second where mergers are likely to be challenged unless the evidence suggests that the merger is not likely to substantially lessen competition (Herfindahls roughly 1,000-2,000 and change over 50-100 points), and the third where mergers are expected to be anti-competitive (Herfindahl over 2,000, change over 100).(38)

The revised Guidelines advance a number of factors that may affect the probability that a merger is likely to substantially lessen competition. Entry conditions are of particular interest. The Guidelines note the DOJ is unlikely (but not unable) to challenge a merger if entry is so easy that firms could not succeed at raising price for a significant period of time (i.e., 2 years).(39) A number of other factors are also given as affecting the likelihood of an anticompetitive effect, although it is clear that they do not comprise a complete list. They include the homogeneity of the product, the importance of substitutes excluded from the market, the spatial similarities of the merging firms, the market characteristics, the fringe supply elasticity (added in 1984), the conduct of the firms in the market and the historical performance of the market.(40) By evaluating these factors, along with conditions of entry, the Guidelines are designed to determine if the merger in question will create or enhance market power in a marginally concentrated industry. Moreover, in special cases, these factors are able to rebut the strong presumption of an anticompetitive effect in a highly concentrated industry.

The 1982 version of the Guidelines implicitly considered efficiencies by creating a large safe harbor for mergers. However, merger-specific efficiencies were only considered a mitigating factor in extraordinary cases. The 1984 revision of the Guidelines liberalized the standard by considering evidence of "clear and convincing" efficiencies in deciding whether to challenge a merger. The weight given to this case-by-case efficiency defense remained unclear, although considering the Guidelines' focus on price, it is possible to conclude the efficiencies must be sufficient to preclude a price increase.

The Guidelines represent a real reform in antitrust policy, expanding on the principles of General Dynamics, but fall short of a policy advocated by the Chicago school.(41) Herfindahl levels necessary to infer a competitive problem were raised dramatically in comparison to the 1968 Guidelines and barriers to entry were assigned an important position in all marginal cases. However, the Guidelines retain a strong presumption against mergers in highly concentrated industries, suggesting the parties to a transaction must present substantial evidence to prove that the merger will not injure competition. Likewise, the Guidelines remain skeptical of efficiency defenses. Given the Guidelines are only a policy statement by the DOJ, it would not be surprising if the courts either failed to follow the suggestions of specific nonconcentration factors that merit attention or extended the Chicago principles to their natural conclusion by requiring the government to bear the burden of proof with a credible anticompetitive theory.(42)

The table presents the basic data gleaned from a survey of the government merger challenges litigated from 1982 to 1991. The first column notes whether the highest court found for the government (1) or the respondents (0). The next two columns list the Herfindahl and change in Herfindahl.(43) Data are marked with a star if they are estimated from information in the decision and two stars if they involved an alternative finding by the court. Columns 4-7 contain binary variables for barriers, ease of collusion, cost savings and buyer power. The variable takes on the value of I if the factor was found to contribute to the analysis, 0 if not, and is missing if no finding was explicitly or implicitly made. For example, in a number of cases, courts found some type of entry barrier (1) and in other cases, the lack of barriers was observed (0). The collusion variable was set to I when the court suggested the market was conducive to noncompetitive behavior and 0 when the overall evidence suggested collusion was relatively unlikely. The efficiency and buyer power variables were defined in a similar manner.(44)

Our initial evaluation of merger policy is based on concentration. Only 8 of the 33 decisions involved estimated Herfindahl levels of under 2,000 and the courts blocked only two of these mergers.(45) In Warner, the preliminary injunction was issued, because concentration and barriers were thought to establish a likelihood of success on the merits, while in the VCM part of B. F. Goodrich, the FTC ruled that sufficient other factors existed to support the conclusion that the merger was likely to lessen competition.(46) In the remaining six cases, the courts ruled the mergers were unlikely to affect competition due to low concentration, easy entry, other market conditions that enhanced competition, buyer power and efficiencies. These results are generally compatible with the wording of the Guidelines, although the prominence of the buyer power and efficiency explanations is a little surprising.

In the remaining 25 cases, concentration levels were above 2,000 and in all but two cases the increases were significant (i.e., greater than 100 points). However, the government won only 14 of the 23 problematic cases. This hardly suggests the courts have validated the Guidelines' wording of "challenge in all but extraordinary circumstances" that implies most mergers in highly concentrated markets are anticompetitive. In the government's 14 victories, the courts found barriers to entry in all the cases and other factors to support collusion in all but one case. Efficiencies were noted in only three cases and buyer power in just one. Thus, in these cases, the courts found strong support for the conclusion that the merger is likely to substantially lessen competition.

On the other hand, in the government's nine defeats, lack of entry barriers were noted in seven cases and other factors to negate the inference of collusion were observed in seven cases, while both efficiencies and buyer power existed in five (somewhat different) cases. Overall, the evidence suggests that lack of barriers, standing alone, may be fatal to a case (Waste Management and Promodes), evidence sufficient to suggest that collusion is difficult in combination with efficiencies may be fatal to a case (Owens Illinois (alternative finding) and Carilion Health) and even merger to near monopoly may be allowed if sufficient evidence exists to support the inference of no anticompetitive effect (Syufy Enterprises).(47) Moreover, buyer power played a major role in concluding the market would remain competitive (Country Lake Foods, Baker Hughes and Donnelley).

Thus, the courts have applied a detailed analysis that goes beyond Herfindahl levels in even the most concentrated markets. While high concentration may establish a presumption of an anti-competitive effect, it appears clear that the courts are very receptive to evidence that rebuts the presumption even if it is not included in the Guidelines (i.e., buyer power). This result is broadly compatible with the Chicago school in that an explanation of the competitive performance of a market is able to defeat a presumption of anti-competitive behavior based on concentration. Overall the Guidelines do not appear to be enforced as written.(48)

An alternative approach to the evaluation of merger policy would start with entry conditions. In all the cases where no barriers to entry were found and in the two cases where the courts made no finding on entry, the government failed to. prevail on the merits. Thus, the courts appear to consider entry barriers to be a necessary condition for a merger to be an antitrust violation. This result is compatible with generally accepted wisdom of economists: entry barriers are required for a merger to have an anticompetitive effect. Overall, the observation suggests the courts have not accepted the cryptic wording of the Guidelines on entry as just another factor to consider in merger analysis.

If the courts found some form of entry barrier exists to limit competition, then further analysis of the merger is required. In particular, the government won 16 of the 20 high-barrier cases, suggesting that entry barriers are not a sufficient condition for a merger to substantially lessen competition in a relatively concentrated market. Concentration may play some role in determining if the merger challenge will be successful. We note the average Herfindahl associated with the government's defeats in 1,754 (change in Herfindahl 430), while the average Herfindahl associated with the government's successful cases is 4,055 (change in Herfindahl 1,402). This suggests that the courts are more likely to find for the government in high Herfindahl cases. In addition to the higher Herfindahls, the court decisions tend to find structural conditions compatible with noncompetitive behavior in cases where the merger was blocked and factors suggesting competitive performance in the cases where the merger challenge was rejected.(49) This suggests that the courts looked beyond the simple market share statistics. Finally, efficiency defenses were accepted as a relevant factor in two of the four unsuccessful cases, but only three of the 16 successful cases. This suggests that although an efficiency defense is not sufficient to prevent a merger from being enjoined, it appears to have an impact on the decision.

A probit model can help in highlighting the relative importance of the various factors, although our flexibility is restricted by the limited amount of available data. One simple model attempts to explain the decision to block the merger on the basis of the Herfindahl (HERF) and efficiencies (EFF). Basically, as the Herfindahl increased, a merger was assumed to be more likely to be anticompetitive. Alternatively, if efficiencies were present the merger was less likely to be found to be illegal. The Herfindahl was taken directly from the table, although similar results were derived if the change in the Herfindahl was used instead. The efficiency variable was assigned the value 1 if the court decision noted efficiencies from the transaction and 0 otherwise.

The estimated parameters are given below (with the t-statistics in parentheses).

Block = -5.17 +.00372 HERF -420 EFF (-L66) (1.71) (-I.62)

Likelihood Ratio Test-13.3 Puesdo R-square-.667

The coefficients are marginally significant, but the overall model easily passes the likelihood ratio test. Assuming the Herfindahl takes on the sample mean of 3,617 and no efficiencies exist, the model generates almost 100% probability of a complaint. In fact, as long as the Herfindahl remains over 1,830, a complaint will have a 95% chance of success, when no efficiencies are present. If efficiencies are introduced, the Herfindahl must rise to 2,958 to maintain a 95% probability of success. If the Herfindahl was only 1,800 in a case with efficiencies, the merger would have less than a one percent chance of being blocked. Thus, the model suggests efficiencies play a significant role in merger analysis, but is also compatible with the idea that mergers that increase the Herfindahl over 1,800 are illegal (given entry barriers) unless other factors are present.

In theory, it would be possible to add another variable to the model to account for differences in structural conditions beyond concentration (and entry) that affect the likelihood that the merger would substantially lessen competition. However, it is impossible to estimate such a model with the small sample, because of collinearity between the variables. Instead, it is necessary to combine the variables that affect the likelihood that the merger will be anticompetitive into one variable NETCON. One starts with the number of structural conditions noted in the court decision as conducive to collusion and then subtracts the number of conditions compatible with competitive market behavior. A final correction is made by subtracting one for each case where efficiencies were found to exist. Although the aggregation assumes the impact of each variable is identical, it avoids the estimation problems with a simple generalization of the model. It is also necessary to create a new Herfindahl variable (DHERF) which takes on the value 0 in all low Herfindahl cases (Herfindahl under 2,000, but note no cases between 1,801-1,999 so the exact cutoff is arbitrary), and the actual numerical value of the Herfindahl in all other cases.

The equation is estimated below (with t-statistics in parentheses).(50) Block = -.811 +.466 NETCON + .000820 DHERF (-1.07) (1.99) (1.78)

Likelihood Ratio Test-12.0 Puesdo R-square-.600

The two independent variables are both significant at the 10% level, as is the likelihood ratio test for the entire model. The equation suggests that the government has a 21% chance of prevailing on the merits if the NETCON variable is a wash at 0 and the Herfindahl is under 2,000. If the variable takes on the value 2, the probability of success increases to 55% while if NETCON is -1, the probability of success falls to 10%. Thus, mergers that do not increase the Herfindahl above 2,000, but offer efficiencies would be unlikely to be blocked, unless strong evidence of an anti-competitive effect were present. On the other hand, if the postmerger Herfindahl were at the mean of 3,617, the merger would have a 98% chance of being blocked by the court if no other factors were present. An efficiency defense that generates a NETCON value of -1, lowers the probability of an order by 3 points from 98% to 95% relative to the situation in which no efficiency defense exists. Likewise, an efficiency defense would lower the probability of an order against the merger from 88% to 75% if the Herfindahl was 2,400. Of course, other factors such as buyer power could also reduce the probability of a merger injunction. For example, if the Herfindahl was 2,400, efficiencies and buyer power would reduce the government's likelihood of success to 59%. Thus, the generalization of the model to consider other structural conditions shows efficiencies, standing alone, are unlikely to defeat an inference of an anticompetitive effect in a highly concentrated market.

The generalized model suggests that the courts are sympathetic to the Chicago approach that requires analysis of the structural conditions in a market to determine if the merger is likely to lessen competition. Herfindahls are but one factor that goes into the analysis and do not appear to be dispositive, unless they are well above 1,800. Efficiencies and other factors that suggest competitive performance remain important considerations in merger analysis.

Both concentration and change in concentration are relatively high in almost all the cases won by the government. This suggests that some concentration effect is a necessary condition, but it makes it impossible to distinguish between the level of the Herfindahl and the change in the Herfindahl. What we can conclude is that mergers that do not increase the Herfindahl over 1,800 do not usually raise antitrust concerns. Likewise, small increases in the Herfindahl will probably not induce a court to block the transaction.

The court decisions clearly show that barriers to entry are a necessary condition for a merger to be considered anticompetitive. The government never wins unless barriers exist. Efficiencies appear to lower the probability that the government will prevail, with a bigger effect in relatively marginal cases. It is possible that the courts realize that the anticompetitive effects remain speculative without a collusion story to explain how prices will rise. If the efficiency explanation proves persuasive, the courts may accept a small risk of an anticompetitive effect for a likely procompetitive effect from the efficiencies. Thus, it is possible that efficiencies and anticompetitive effects are not being explicitly balanced in the decisions, but the courts are choosing a likely efficiency benefit over a speculative loss in competition.

The court decisions suggest evidence beyond concentration is useful in many merger decisions. One statistical model suggests if the concentration ratio does not exceed 1,800 (with barriers and no efficiencies) the merger is unlikely to be condemned, unless a strong structural case can be made. On the other hand, the model predicts that factors suggesting the market is likely to remain competitive, such as buyer power, may offset the anticompetitive presumption of concentration in the area of 1,800-2,400. This represents a significant generalization of the traditional idea accepted by the Guidelines that concentration in the range of 1,800 to 2,400, by itself, can support an inference of anticompetitive pricing, at least in the presence of barriers to entry. It is clear that the Chicago learning has prevailed in some courts, but unlike the analysis of barriers, other structural conditions compatible with collusion are not considered necessary conditions for any merger challenge to be upheld. If the courts completely reject DOJ Guidelines' reliance on concentration in the future, the Chicago theory will have a final triumph in the reform of the antitrust law.

In conclusion, the courts have dramatically reformed merger law, moving beyond a simple interpretation of the DOJ Guidelines. High concentration and entry barriers are now necessary conditions for a merger challenge to succeed. Moreover, efficiencies and other structural factors can defeat a presumption to challenge a merger when Herfindahls are well over 1,800. This reliance on entry conditions and the creation of efficiency and structural defenses to at least some mergers in concentrated industries goes beyond the explicit wording of the Guidelines and appears to be an independent assimilation by the courts of Chicago-school economic analysis. Of course, the Guidelines could be easily changed to match the courts by editing the section on entry barriers to state the government will not bring a case unless barriers are present and raising the Herfindahl cutoffs from 1,000-1,800 to 1,500-2,400. If the evolution of merger law continues, further changes to the Guidelines may be necessary for the Guidelines to conform to the law.

V. Conclusion

The Merger Guidelines have been used by the courts in the reformation of merger policy in the 1980's. However, it would be incorrect to assert that the courts have accepted the Guidelines' structure of the law. Instead, the courts appear to have rejected the Guidelines' approach to market definition and have gone beyond the Guidelines' ideas in evaluating the competitive effects of a merger.

Little evidence exists to suggest that the Guidelines' approach to market definition is controlling in the federal courts. The courts appear to use economic analysis and judicial precedent to define an alternative approach that looks at the behavioral evidence in a market. Under the judicial rules, the government has won its share of contested market definitions during the 1980's, but failed on a number of occasions to substantiate narrow markets generated by the Guidelines' analysis. Thus, while the specific approach to market definition defined by the Guidelines has not won the day, the telling criticism of certain market definitions from the 1960's appears to have generated more economically reasonable decisions in the 1980's.

The DOJ Merger Guidelines attempted to implement a number of small changes from the historical case law. Instead of simply validating these changes, the courts have gone beyond the basics of the Guidelines to define a merger policy more in line with economic realities. Barriers to entry are now a necessary condition for a merger challenge to succeed. Likewise, efficiencies are an important factor that appear to have an effect in merger litigation. Finally, other structural conditions, such a buyer power, play a role in the merger decision as long as concentration is not extremely high. The requirement that evidence be substituted for the presumption of an anticompetitive effect from concentration appears to represent the final step in the evolution of merger policy. (1) For example, deconcentration proposals were recommended in the White House report on Antitrust Policy. White House Task Force on Antitrust Policy, Report 1. Antitrust Trade and Regulation Report, supp. no. 415 (May 26, 1969). (2) For a comparison of the Chicago and Harvard schools of thought see Posner, The Chicago School of Antitrust Analysis, 127 U. Pa. L. Rev. 939 (1979) and Audretsch, Divergent Views in Antitrust Economics, 33 Antitrust Bull. 135 (1988). (3) R. Bork, The Antitrust Paradox: A Policy at War With Itself 180-95 (1978). Numerous other authors advanced similar views. For an overview, see Pautler, A Review of the Economic Basis for Broad-based Horizontal Merger Policy, 28 Antitrust Bull. 571 (1983). (4) R. Bork, supra note 3, at 180. (5) Id. at 195. (6) For example, compare the recent writings of White, Antitrust and Merger Policy: A Review and Critique, 1 J. Econ. Pers. 17-20 (1987) to Fisher, Horizontal Mergers: Triage and Treatment, 1 J. Econ. Pers. 30-33 (1987) and Schmalensee, Horizontal Merger Policy: Problems and Changes, 1 J. Econ. Pers. 43-51 (1987). They all agree that no general relationship exists between concentration and profitability (a point that will argue for permitting all mergers in unconcentrated industries), entry barriers must be present for a merger to be anticompetitive (suggesting analysis of entry conditions be required in every case) and efficiency considerations are a valid concern (mandating some consideration of efficiency). (7) The DOJ appeared to believe that analysis of entry conditions is all that is necessary to condemn a merger in a concentrated industry. See, Kleit & Coate, Are Judges Smarter Than Economists? Sunk Costs, the Threat of Entry and the Competitive Process (Federal Trade Commission Working Paper No. 190, 199 1). (8) Schwartz dates the decline of antitrust from 1977, although none of the cases he discusses are horizontal mergers. Thus, one could argue that merger law was caught up in the antitrust revolution in the 1980's. See Schwartz, Cycles of Antitrust Zeal: Predictability? 35 Antitrust Bull. 771 (1990).

For an interesting discussion of background of the 1982 revision of the Guidelines, see Rhoades & Burke, Economic and Political Foundations of Section 7 Enforcement in the 1980's, 35 Antitrust Bull. 373 (1990). (9) It is interesting to note that ex-FTC Chairman James Miller observed the ascendancy of the Chicago school now seems all but inevitable. This evolution in antitrust doctrine constituted not only the winning of the intellectual war but a response to a public sensitized to institutional arrangements that arguably restrained their standard of living. J. Miller, The Economist as a Reformer 47 (1990). (10) The data sample included the 32 government merger cases litigated between June 14, 1982 and May 31, 1991. The B. F. Goodrich matter is counted as two cases, because it involves different overlaps on which different decisions were made. The analysis focuses on the district court or Federal Trade Commission decision, but replaces any conclusions that were reversed on appeal. Three cases are excluded from the sample. Federal Trade Commission v. Midcon was basically moot (Midcon had sold the offending assets) when decided by the Commission, Federal Trade Commission v. Olin was on appeal in late 1991 and United States v. Archer-Daniels-Midland Company was under consideration for appeal at the end of 1991.

Numerous other cases brought by the government were excluded from the sample, because they settled before trial. Although these cases were usually settled on the government's terms, it is inappropriate to give them any weight in the analysis, because they may only represent the enforcement policy of the government and not the underlying law. Respondents are likely to settle when the costs of further litigation (for example, direct expenses, probability of an adverse verdict, business costs of delays and management time) outweigh the expected benefits of further litigation. The merits of a case may play only a small role in the decision. Given the parties rarely litigate, the antitrust enforcement agencies play the role of regulators in addition to law enforcers. See Kauper, The Justice Department and the Antitrust Laws: Law Enforcer or Regulator, 35 Antitrust Bull. 83 (1990). (11) See, United States v. E.I. Dupont de Nemours and Co., 353 U.S. 586 (1957) and Brown Shoe v. United States, 370 U.S. 294 (1962). (12) R. Posner, Antitrust Law: An Economic Perspective 125-30 (1976). (13) Id. at 131-33. (14) Werden, Market Delineation and the Justice Department's Merger Guidelines, 1983 Duke L. J. 571 (1983). (15) R. Posner, supra note 12, at 133. (16) The DOJ is alleged to have used a 10% test. See Pitofsky, New Definitions of Relevant Market and the Assault on Antitrust, 90 Col. L. Rev. 1836 (1990). (17) For example, distinct customers, vendors and prices are all cited in Brown Shoe as useful in market definition, but are not mentioned in the 1982 Guidelines. See Brown Shoe, supra note 11. (18) United States v. Calmar Inc., 612 F. Supp. 1298 (D.C.N.J. 1985). (19) F.T.C. v. Owens Illinois, 681 F. Supp. 27 (D.D.C.), aff'd, 850 F.2d 694 (D.C.Cir. 1988). (20) F.T.C. v. R. R. Donnelley & Sons, Inc., 1990-1 Trade Cas. (CCH) [paragraph] 64,852 (D.D.C. 1990). (21) Part of the problem may be the myopic application of the five-percent test. If customers testify to a willingness to pay a five-percent price increase in the short run, the government may infer a narrow market even though long-run considerations such as an adverse effect on reputation, preclude a hypothetical monopolist from raising price.

It is interesting to note that the five-percent test has been recently used against the government. In the Country Lake Foods decision (United States v. Country Lake Foods, Inc., 1990-2 Trade Cas. (CCH) [paragraph] 69,113 (D.Minn. 1990)), the court found a broad geographic market was supported by testimony that a number of dairies would sell into the Minneapolis-St. Paul area in response to a 5%-10% price increase. This type of analysis will be most effective for homogenous products (milk) for which there is a well developed transportation system. However, the analysis might be applied more broadly in the future. (22) One can still claim the rules are picked to gerrymander the process. For example, the Guidelines' procedure appears to define narrow markets. Not surprisingly, the government alleges the narrow market in all but one of the product markets and all but four of the geographic markets. When the government alleges a broad market, it tends to increase concentration. Of course, one could always argue it is the respondents that gerrymander the market to minimize the change in concentration. Regardless of which side is at fault, the Guidelines' procedure tends to reduce the possibilities for strategically defining markets by committing the government to a procedure. It is interesting to note that none of the markets alleged by the government would clearly fail a "laugh test," which Franklin Fisher implicitly defines as a product market of "high-priced, frozen nonethnic entrees." See, Fisher, supra note 6, at 28. (23) Splitting the court decision sample almost in half at January 1989 reveals the Reagan authorities obtained stipulations in 39% of the product markets and 61% of the geographic markets, while the figures for the Bush administration are similar (40% and 53% respectively). (24) The government won 10 and lost 8 of the market decisions before 1989 (7 wins and 4 defeats for product market) and has won 9 and lost 7 during the 1989-1991 period (5 wins and 4 defeats for product market). (25) Legal commentators such as Prager, How Much Is Enough? Antitrust Developments Affecting Mergers and Acquisitions, 56 Antitrust L.J. 303 (1987) and Calkins, Developments in Merger Litigation: The Government Doesn't Always Win, 56 Antitrust L.J. 855 (1988) have noted the FTC is much more successful than the DOJ at winning cases. Calkins explained the difference based on litigation style, with the FTC litigating both the Guidelines and case law, while the DOJ focuses primarily on the Guidelines.

The FTC appears more successful at winning market definition decisions. In particular, the FTC obtained explicit or implicit stipulations in 9 of 19 product markets and 13 of 19 geographic markets. The comparable figures for the DOJ were 4 of 14 and 6 of 14, respectively. A similar pattern was found for disputed market definitions, with the FTC winning 7 product and 4 geographic markets, while the DOJ won only 5 product market and 3 geographic market decisions. Overall, the FTC prevailed on 84% of the market definitions and 89% of the geographic markets. The comparable figures for the DOJ are 64% for both product and geographic markets. (26) Waste; Management v. United States, 743 F.2d 976 (2d Cir. 1984h (27) F.T.C. v. Occidental Petroleum Corp., 1986-1 Trade Cas. (CCH) [paragraph] 67,071. (28) Supply-side substitution involves redeploying assets into the product market. If this can be done in 1 year, the Guidelines consider the new firms as market players and if the switch takes up to 2 years, the new firms are considered entrants. The courts do not appear to make this distinction. (29) United States v. Baker Hughes Inc., 731 F. Supp. 3 (D.D.C. 1990), aff'd, 908 F.2d 981 (D.C.Cir. 1990). (30) United States v. Syufy Enterprises, 712 F. Supp. 1386, aff'd, 903 F.2d 659 (9th Cir. 1990). (31) This result appears to rebut Pitofsky's (supra note 16, at 1809) concern that broad markets will mask market power. The decisions suggest no market power exists even in the narrow markets defined by the government. One could also note that the government appears to be able to use the Guidelines to define narrow markets, so Pitofsky's general concern that the Guidelines procedures generate broad markets is misplaced. (32) United States v. Von's Grocery, 384 U.S. 270 (1966). (33) United States v. Philadelphia National Bank, 374 U.S. 321 (1963). (34) United States v. General Dynamics Corp., 415 U.S. 486 (1974). (35) R. Posner, supra note 12, at 110-11. (36) For example, assume one 10% firm buys another 10% firm in an industry with one 20% firm and six other 10% firms. This would generate a postmerger Herfindahl of 1,400 with an increase of 200, while the postmerger two-firm concentration ratio is 40. Likewise, if two 15% firms merge in an industry with a 30% market leader and four 10% firms, the postmerger Herfindahl would be 2,200 while the two-firm concentration ratio is 60. Of course, these calculations serve only to define examples and other fact situations could give rise to either higher or lower Herfindahls. (37) The 100-point increase is slightly higher than the implicit 50-90 point increase defined by the 1968 Guidelines for most markets and the 50-point increase is slightly higher than the roughly 30-point increase defined for highly concentrated markets. (38) The figure of 2,000 is taken from Leddy, Recent Merger Cases Reflect Revolution in Antitrust Policy, Legal Times, November 3, 1986, at 2, but really stands for a number substantially over 1,800. Our discussion of the Guidelines tends to focus more on the wording of the 1984 revision of the Guidelines, which appear to be a clarification of the 1982 Guidelines. (39) The DOJ argued that the Guidelines require clear evidence that sufficient entry would occur to defeat a price increase before the merger is immunized from challenge. This interpretation was clearly rejected by the appeals court judge Clarence Thomas in Baker Hughes, supra note 29. Regardless of how entry is considered, the Guidelines (1984, paragraph 3.3) note sunk costs, stagnant markets and economies of scale as factors that make entry less likely to occur. (40) It is interesting to note that buyer power, a factor that plays a prominent role in some merger decisions, is not really mentioned in the Guidelines as a condition that makes collusion less likely. The Guidelines suggest that collusion will be more difficult in markets where large orders are the norm. Although large buyers can create large orders, they can attempt to defeat collusion in numerous other ways. (41) It is interesting to note that at least the FTC did not follow the strict wording of the Guidelines. See Coate & McChesney, Empirical Evidence on FTC Enforcement of the Merger Guidelines, 30 Econ. Inquiry 277 (1992). (42) The basic idea of the Guidelines that mergers should not be permitted to create or enhance market power should prove acceptable to a broad spectrum of Harvard and Chicago school economists. However, the details incorporated in the Guidelines on how to reach this conclusion are likely to be controversial. (43) Two cases required special treatment. Goodrich (F.T.C. v. B. F. Goodrich Co., 5 Trade Reg. Rep. (CCH) [paragraph] 22,519 (F.T.C. 1988)) involved two overlaps, with only one leading to a divestiture order. We handled this situation by separating the analysis into a PVC part where the merger was allowed and a VCM part where divestiture was ordered. IMO Industries (F.T.C. v. IMO Industries, Inc., No. 89-2955 (D.D.C. Nov. 11, 1989) involved confidential data. Thus, we took the information in the table from public sources. (44) It is interesting to note that the government prevails in 47% of the cases. This number is close to the predicted success rate of 50% for litigation in which the parties have symmetric incentives. (See Priest & Klein, The Selection of Disputes for Litigation, 13 J. Legal ST. 1 (1984).) (45) In four cases, Virginia National Bank, Occidental Petroleum, Owens Illinois and Country Lake, the court ruled against the government on market definition, but made alternative findings for the government's market which we use in the analysis. (46) F.T.C. v. Warner Communications, 742 F.2d 1156 (9th Cir. 1984). (47) See, F.T.C. v. Promodes, 1989-2 Trade Cas. (CCH) [paragraphs] 68,688 (N.D. Ga. 1989) and United States v. Carilion Health System, 707 F. Supp. 840 (W.D. Va. 1989). (48) The Guidelines tend to be cited more often when the court is going to find against the government. In 7 of the 11 cases where no barriers were found, the courts cited the Guidelines on entry in comparison to just 7 entry cites in the 20 high-barrier cases. Likewise, in 8 of 13 cases where collusion was thought to be difficult, the Guidelines' collusion analysis was cited. This compares to 6 of 14 cases where collusion was considered likely. (49) An average of 2.1 conditions compatible with collusion were found in the 16 blocked mergers, while only 1.5 in the 4 mergers that were allowed. The compatible figures for procompetitive conditions were .5 in the blocked mergers versus 2.5 in the approved mergers. (50) A summary variable for a DOJ case can be added to the model, but the coefficient, although negative, is not statistically different from zero. Thus, for those cases where barriers to entry exist, the DOJ does not appear to do worse than the FTC. In particular, the DOJ's record is four wins and one defeat in the high-barrier cases. This compares to the FTC record of 12 wins and 3 defeats (9 wins and I defeat in federal court). Of course, the DOJ lost 8 low-barrier cases and 1 case where no barrier finding was made to bring their overall record to 4 wins and 10 defeats. This does not compare favorably with the FTC record of 12 wins and 7 defeats (or excluding the administrative complaints where the FTC serves as the judge, 9 wins and 4 defeats).

MALCOLM B. COATE Bureau of Economics, Federal Trade Commission. AUTHOR'S NOTE: The analyses and conclusions set forth in this article are those of the author, and do not necessarily reflect the views of other members of the Bureau of Economics, other Commission staff or the Commission. I would like to thank Andrew Kleit, Pauline Ippolito and John Lopatka for helpful comments and Rene Bustamante for research assistance on this project.
COPYRIGHT 1992 Sage Publications, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Coate, Malcolm B.
Publication:Antitrust Bulletin
Date:Dec 22, 1992
Words:8956
Previous Article:An economic evaluation of federal antitrust activity in the manufacturing industries: 1980-1985.
Next Article:Mergers and acquisitions in Europe 1993: the new EC merger control regulation and its effects on national merger control in Germany.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |