6. Horizontal Agreements: Mergers and Price Fixing

The last remaining intellectual stronghold of strict antitrust enforcement is the continuing regulation of horizontal agreements such as joint ventures, price agreements, and horizontal mergers, that have the probability of reducing market output and raising market prices. The general antitrust thinking on horizontal agreements is that most mergers and joint agreements should be judged by an economic rule of reason, while price collusion and division-of-market agreements should remain illegal per se.

The rule-of-reason approach implies that the antitrust authorities evaluate and act upon the probable social costs and benefits of a proposed merger or joint venture. Joint business agreements can promise substantial cost savings in production and distribution, as well as in financing, industrial research, and product development. In addition, they may allow the innovation of entirely new products and services not feasible without interfirm cooperation. On the other hand, there is always the possibility that mergers and joint ventures may lead to output restriction and the suppression of price rivalry. The antitrust authorities and the courts, therefore, must weigh the probabilities of increased social benefits against the risks and costs of potential output restriction. Under a rule of reason, mergers whose probable benefits exceed the probable costs would be allowed, and those whose probable costs exceed the probable benefits would be prohibited.

The rule-of-reason approach is not yet generally accepted for dealing with price-fixing or division-of-market agreements; the consensus is that they should still remain illegal per se.¹ Since it is widely assumed that the social benefits associated with price fixing are minuscule or altogether nonexistent, and since such agreements intend to restrict market output, they can be safely excluded from any rule-of-reason analysis and prohibited entirely. Finally, it is hoped that the legal certainty of the per se approach to price fixing will have a chilling effect on this activity in the future.

Although a rule-of-reason approach to mergers and joint ventures sounds appropriate, and although a flat prohibition on price and output restrictions—so-called naked agreements—also sounds appropriate, these positions are fraught with many significant difficulties. The essential problem is that both approaches assume that the antitrust authorities or the courts can have access to information concerning the future course of the market process that is simply unavailable to any regulatory authority or court. In addition, both approaches assume an ability to measure economic phenomena that, in principle, cannot be measured by any outside observer. Thus, while both approaches may give the appearance of science and objectivity, both are, in fact, pseudoscientific and cannot legitimize government antitrust intervention in this area.

The Rule of Reason: Social Costs

The rule-of-reason approach implies that the antitrust authorities ought to permit horizontal agreements when the social gains are expected to exceed the social losses. Social losses relate to the probability that the agreement could reduce market output, and that probability depends, in turn, upon whether the agreement creates any so-called market power or not. According to conventional theory, market power depends directly on the market share of the firms involved and whether the increase in market concentration makes effective output restriction more probable.

The Antitrust Division of the Justice Department and the Federal Trade Commission publish horizontal merger guidelines that are based on the Herfindahl Index of market concentration.² A business merger that raises the Herfindahl Index by more than a stated number of points or pushes the industry index above a stated level will likely trigger legal action by the government in opposition to the merger.

Merger guidelines rest on two crucial assumptions. The first is the notion that there is some scientific way to define the so-called relevant market under discussion in any merger. The second is the belief that there is some scientific way to determine precisely which levels of market concentration generate so-called market power and which do not. In fact, the relevant market can never be known with scientific accuracy. Further, it is impossible—theoretically and empirically—to know which levels of market concentration generate market power.

Relevant Market

A rational discussion of market concentration is premised on some acceptable definition of “relevant market.” Firms are said to compete in some relevant market, and presumably a merger or joint venture may threaten to create, or to increase, market power. If relevant markets are defined narrowly, almost any horizontal agreement will increase market concentration and threaten to create market power. If, on the other hand, relevant markets are defined very broadly (and the merging firms usually hope they are), there is almost no horizontal agreement that could threaten to create market power. The question, then, is how to define relevant markets in order to calculate the appropriate levels of market concentration.

In general, a relevant market includes all suppliers whose products are “reasonable substitutes” for each other and excludes all others. If one were evaluating a possible merger between two soft drink companies, for example, or between a soft drink company and a beer company, one would have to determine whether the products produced and sold by the different companies were reasonable substitutes for each other. Narrowly, one could consider products to be substitutes (and their suppliers to be competitors) when a price adjustment by one supplier directly affects the output and sales of another. For instance, if soft drink price changes directly affect beer sales, then, presumably, soft drinks and beer are reasonable substitutes and the relevant market in any possible merger would have to include at least both these products and suppliers.

But there are important difficulties with this approach. In the first place, it may be impossible to determine, in practice, whether changes in the prices of soft drinks caused the change in beer sales. Second, changes in the prices of a soft drink—other prices remaining the same—may not be significant enough to affect beer sales appreciably; non-price elements of rivalry may be important, even more important than price. The fact that beer sales are not appreciably affected by some soft drink price change does not necessarily mean that soft drink and beer firms are not rivals. It may imply only that price changes are a poor measure of substitutability in markets where non-price competition is important.

What, precisely, is the relevant market for soft drinks? Is it only soft drinks, or does it also include fruit drinks, orange juice, milk, bottled water, wine, and beer? Does it include only U.S. manufacturers and sellers, or are producers in Canada and Holland to be included? Should the relevant market be defined nationally, or should it be divided into regional submarkets? How are these questions to be answered unambiguously?

These are not rhetorical questions. In traditional merger discussions, the definition of the precise relevant market can make all the difference. If the relevant market for soft drinks is restricted to soft drinks, then almost any merger between big soft drink companies can look potentially output threatening. If the market for soft drinks extends beyond soft drinks, however, many more horizontal agreements can be permitted. In the absence of an unambiguous definition of the relevant market, it would seem impossible to determine with scientific certainty whether changes in market share or in levels of concentration threaten competition and make effective output restriction more probable.

Market Share and Market Power

Even if relevant markets could be clearly defined, there is an additional threshold problem associated with any rule of reason. Government merger guidelines may be useful in indicating to business the likelihood of antitrust action, but they are of no scientific value in theoretical discussions of market power, and they cannot justify government intervention. Although the general public has the impression that there must be some good reason for the antitrust authorities’ choice of particular limits in the Herfindahl Index of market concentration, those limits are completely arbitrary. No one—and certainly not the antitrust authorities—can ever know whether a merger of firms that creates, say, a 36 percent market share, or one that raises the Herfindahl Index by 150 points, can create sufficient economic power to reduce market output and raise market price. No one knows, or can know, whether monopoly power begins at a 36 percent market share or a 36.74-percent market share. Neither economic theory nor empirical evidence can justify any merger guideline or prohibition.³

Even if relevant-market and market-concentration considerations were not ambiguous and arbitrary, there would still not be sufficient reason to legally restrict horizontal agreements. All the issues confronted in chapter 3 concerning the ability of firms in free markets to actually charge long-run monopoly prices and earn monopoly profits are relevant here and need not be repeated. It is enough to point out that antitrust theory cannot demonstrate that firms in free markets can earn long-run monopoly prices and profits, or that resources in free markets can be inefficiently misallocated. Free markets are always competitive and tend naturally to eliminate inefficiency.

Output Restriction

Yet, a further problem in applying a rule of reason to horizontal mergers and joint agreements is the difficulty of measuring any output restriction and determining that it is the result of monopoly power. Existing levels of production cannot be compared against the standard of pure competition (see chapter 3); nor can any premerger industry output level be considered the appropriate level of production. The premerger output level is a disequilibrium output level, established through interdependent pricing and output determinations rather than through independent pricing, as under atomistic competition. It cannot serve as a welfare benchmark because there is no difference, in principle, between market-price and output determination in the premerger and postmerger situations. Interdependent pricing and output determination exist under both circumstances, and both are part of an open-market discovery and adjustment process. It is true that legal monopoly could establish an unambiguous output restriction benchmark by prohibiting market entry. But, in the absence of some legal restriction on production and market entry, it is impossible to determine whether free-market outputs have been inefficiently restricted by any merger. Thus, even a merger resulting in “monopoly” could not be unambiguously condemned.

The Rule of Reason: Social Benefits

The social-benefit side of the rule-of-reason equation poses as many difficulties as the social-cost side. Generally, the social benefits associated with horizontal agreements include economies and efficiencies of interfirm production, financing, advertising, distribution, marketing, and research and development. Some of these benefits are measurable and objective; some are subtle and subjective. Some require financial and accounting skills to understand; others are anticipations and expectations based upon entrepreneurial experience in these matters. The question is whether the antitrust authorities or the courts can evaluate the probable benefits of horizontal agreement as accurately as the relevant entrepreneurs, or, even more fundamentally, whether parties outside an agreement can correctly evaluate future benefits at all.

Businessmen and entrepreneurs, standing in for owners, have strong incentives to estimate interfirm benefits and costs correctly; their own success and even the very life of their corporations may well be at stake. One would assume, therefore, that businessmen considering horizontal agreements would be especially careful to obtain the most accurate information available concerning the possible financial effects of any merger or joint venture.

It is important to note, however, the inherently subjective element in all entrepreneurial calculation concerning future costs and benefits. Successful businessmen can discover and exploit profit opportunities that are not obvious and that others do not see, and this entrepreneurial process goes well beyond simple cost accounting and economic calculation.⁴ Successful business agreements appear to be like successful marriages; they work efficiently, but only the parties involved can understand the relative costs and benefits. Moreover, like marriages, their continued success depends more on tacit knowledge and understanding than on any objective cost-benefit calculations.

The only objective ex post test of the correctness of entrepreneurial decisions is the market process itself. If the businessmen who enter into a horizontal agreement are correct concerning probable costs and benefits, then the merged firm’s market performance will likely be enhanced and its market share and profits may well increase. On the other hand, if the entrepreneurs have miscalculated, then the organization will likely waste economic resources and lose market share to relatively more efficient rivals and competitors. In either case, there is no legitimate reason to believe that the antitrust authorities or courts can have direct knowledge of these costs and benefits or that their intervention can be a reasonable substitute for a working out of a market process.

The Staples Case

Many of the problems inherent in horizontal merger analysis just reviewed were evident in the Federal Trade Commission’s 1997 opposition to a proposed merger between Staples and Office Depot. The merger between the office supply superstores was eventually abandoned after the FTC convinced a district court to grant a preliminary injunction to halt the consolidation.⁵

The district court accepted the FTC’s argument that the merger created market power for the merged firms and allowed them to raise or maintain prices at “anticompetitive levels.” Indeed, the court uncritically accepted the FTC staff analysis that Staples and Office Depot had already raised prices 5 to 10 percent in cities where they faced no other superstore competition. Since the consolidation allegedly resulted in a 75-percent market share and left only one other independent superstore competitor (Office Max), and since the alleged cost savings associated with the merger were “unverified,” the court granted the FTC injunction.

The district court’s decision is a travesty of sound economic principles and reasoning. In arriving at its conclusions, the court accepted the outrageously narrow FTC-created definition of the relevant market (office supply superstores only), ignored the low barriers to entry into office supply sales, and completely distorted the true nature of rivalrous competition in the overall office supply market. The fact remained that besides Office Max, there were thousands of independently owned office supply stores in competition with Staples and Office Depot, including impressive national discounters such as Wal-Mart and Best Buy. Indeed, any reasonable definition of the relevant market would have concluded that Staples and Office Depot combined had only 5 percent of office supply sales in 1996.⁶

The FTC court sanctioned action against Staples-Office Depot was an unwarranted exercise in antitrust industrial planning. The power of government, not the voluntary choices of company shareholders or price conscious consumers, was employed primarily to decide the future course of industrial organization in the office supply industry. The fatal conceit associated with this exercise of governmental authority should be readily apparent.⁷

The Per Se Approach

The logic for an absolute prohibition of horizontal price agreements is that such “collusion” intends only output restriction—a social cost—but creates little, if any, opportunity for the generation of social benefits since it doesn’t involve an integration of facilities between firms. Economic analysis and so-called considerations of law enforcement efficiency dictate that they remain illegal per se.

But the difficulties of evaluating mergers with a rule of reason are also encountered when trying to bring a Per se perspective to bear on price-fixing agreements. One can argue, first, that efficiencies and cost savings to society may indeed be associated with such agreements. Second, although such agreements may intend to restrict production and increase group profits, they are generally not able to do so.

Price Coordination and Efficiency

Over the last few years several commentators have argued that the reduction in risk associated with horizontal price coordination in open and uncertain markets could increase market output and enhance consumer welfare⁸; it may also lead to information-cost and price adjustment-cost savings.⁹ Indeed, some analysts now consider a rule-of-reason approach toward all horizontal price agreements to be justified.¹⁰

A number of efficiencies are possible from price and output coordination among firms operating under conditions of market uncertainty and imperfect information—the normal conditions. Price adjustments can be costly to both buyers and sellers; price coordination could limit price changes and reduce price-adjustment costs. Price information can in some markets be costly to obtain; price coordination could lower information costs. Price uncertainty for risk-averse buyers could reduce their purchases, and price uncertainty for risk-averse sellers could reduce their market output; price coordination could reduce risk and increase sales and market output. Price coordination could stabilize output and inventory fluctuations in the short run and lead to greater market outputs and lower costs in the long run. The uncertainties of market entry could be reduced, and entry encouraged, if potential entrants could make price and output agreements. Price uncertainty could restrict non-price rivalry; price coordination could lead to additional research and innovation.

Support for a per se illegality approach to price-fixing schemes presumes that it is possible to know beforehand which business combinations will generate net social efficiency and which will not. This presumes that the very information provided by the working out of the market process can be known before that market process is allowed to operate. The point here is not that cartels and market division agreements are always appropriate. But, absent an open market process, there is no scientific way to conclude that such arrangements are always socially inappropriate or should be illegal Per se.

The knowledge problem and the discovery principle with respect to price coordination were exemplified in the needless controversy over the expiration of the Reed-Bulwinkle Act (1948), which had generally exempted the trucking industry from Sherman Antitrust Act jurisdiction. The industry maintained that it required a partial continuation of that exemption so that industry rate-bureau organizations could legally continue to coordinate routes and prices for member truckers.¹¹ Antitrust enthusiasts maintained, on the other hand, that collective ratemaking constituted horizontal price collusion, a judgment that would continue to frustrate the welfare advantages associated with deregulation and increasing competition in the trucking industry.¹²

Clearly, the only way to discover whether collective ratemaking, on balance, served shippers or not, was to permit the activity to continue in a free and unregulated transportation market. Industry rate bureaus performed a price- and route-coordination function in trucking for many decades, and it is unclear, under conditions of free entry, why that activity should be legally restricted or prohibited. If certain carriers employ rate-bureau services and achieve efficiencies, then those carriers may gain business and market-share relative to other carriers that price and route independently. If, on the other hand, the costs to shippers exceed the value provided by price coordination, the carriers may lose business and market-share to carriers that price independently. In either case, there is no ex ante logic for a Per se prohibition.

Furthermore, it made no sense to replace Interstate Commerce Commission (ICC) regulation in transportation with regulation by the Antitrust Division of the Justice Department. ICC rate regulation, route control, and entry restrictions stood in the way of a truly competitive open market process in trucking for fifty years, and the industry was deregulated in order to discover how it should be organized for efficient service to customers.

It should be apparent that neither regulators nor economists—nor attorneys, nor judges—can know beforehand which market institutions are socially efficient and which are not. That is precisely the purpose of open markets. If industry rate bureaus, on balance, are inefficient, then they will become ineffective and they will dissolve. If they are efficient, they will continue to function. But these are questions that can be answered only in a completely deregulated transportation market with total antitrust immunity for all carriers.

Those who support the Per se illegality of price agreements argue that whatever social advantages might result from those agreements can be achieved more readily and more acceptably through direct contract integration. If firms are really serious about achieving coordination efficiencies, it is argued, they can always merge or enter formal joint venture agreements. But why, one might ask in response, should all business coordination be forced to take the path of formal integration? Contract integration may well provide additional, and significant, economies, but surely the firms involved—not the Antitrust Division—can determine whether additional economies do exist, given the risks associated with formal integration. If contract integration can easily lead to additional net-efficiency gains, firms will be only too anxious to pursue it. On the other hand, the advantages of price coordination may well represent all the net gains expected from interfirm agreement. In that case, loose price coordination is socially efficient.

In an uncertain world, loose price coordination may be a far more flexible device for achieving possible economies than formal contract integration. The risks and costs of formal integration under conditions of uncertainty may be considerable, and loose associations that allow firms to be rivalrous at any moment may well represent a near-optimal social organization. To mandate that firms be completely rivalrous at every moment may cost the economy the efficiencies associated with interfirm coordination. Yet, to press firms anxious to coordinate some activities into formal integration agreements and their attendant risks may make just as little sense from an efficiency perspective. The appropriate solution is to permit the firms themselves to select the degree of coordination appropriate to the problem to be solved. Again, no antitrust regulation is warranted.

Price Agreements and Output Restriction

The other major argument against the per se illegality of price fixing agreements, which is especially relevant if social efficiencies are indeed associated with them, is that there is little reason to expect such agreements to be harmful to society. Market-division agreements are, in the absence of direct government support, tenuous at best and tend to break apart in open markets when they are inappropriate. Genuine output-restricting agreements appear generally to be short-lived and unable to withstand changing market conditions. When markets are legally open to entry and rivalry, market conditions will normally neutralize attempts to simply reduce market output and raise market prices.

The public can easily be misled on the effectiveness of price conspiracy, often accepting conviction in price-fixing antitrust cases as evidence of effective output restriction and higher prices. This inference, is not, however, normally warranted. Under the law, price agreements themselves are illegal per se; merely to have made an agreement—whether it works or not—is sufficient to violate the antitrust statutes. Whether market outputs were actually restricted or prices were higher during the conspiracy is usually immaterial in government price-conspiracy cases. Indeed, firms indicted for price fixing under the Sherman Antitrust Act often enter nolo contendere pleas because they recognize that the existence of an agreement, notwithstanding its effectiveness, is sufficient for conviction.

The Addyston Pipe Case

The classic Addyston case illustrates some of the traditional difficulties associated with the Per se approach to price-fixing agreements.¹³Addyston concerned a conspiracy of six cast-iron-pipe companies that attempted to rig the bid prices for pipe sold to water departments in municipal governments in the 1890s. The precedent set in Addyston is one of the most important in all antitrust law. Circuit Judge Taft, in his review of the Addyston case on appeal from the district court, argued that the bid-price agreements were illegal in and of themselves; they appeared to intend only a suppression of competition among the firms, to their mutual advantage. Hence, no economic analysis—rule of reason—of the prices actually charged was necessary to condemn the arrangement.¹⁴ Justice Peckham, writing for a unanimous Supreme Court in 1899, agreed with Taft’s analysis and decision, as have many important antitrust scholars. Robert Bork, for instance, once called Taft’s decision “one of the greatest, if not the greatest, antitrust opinions in the history of the law.”¹⁵

An analysis of Addyston, however, raises serious questions about such evaluations. After all, the Per se approach purposely obscures the very economic issues that may be relevant in such cases. In Addyston, for example, Judge Taft, and later Bork, assumed that there were no important economic efficiencies associated with the price agreement and that the conspiracy restricted market production and raised the market price for cast-iron pipe—assumptions that may well have been false.

George Bittlingmayer’s detailed account of the Addyston case makes it clear that some “cooperative solution to market allocation” was necessary to achieve reasonably efficient cast-iron pipe production in this industry in the late 1890s. There was simply no “competitive” equilibrium in the pipe market, he argues, that was consistent with industry cost and demand conditions. With a very cyclical and unstable market demand for pipe, and with decreasing long-run average and marginal costs associated with producing pipe, “competitive” (marginal cost) pricing would not have allowed the various firms to fully recover their costs. Market prices would have been perpetually below average cost, and most of the firms would have eventually become insolvent. Thus, he reasons, some degree of interfirm cooperation—the bid-rigging scheme, for example—was required to keep plants operating efficiently and to maintain capacity during slack periods of demand.¹⁶

The available evidence also suggests that pipe prices, even with extensive collusion, were not generally sustained above marginal costs—let alone average costs. Inevitable rivalry within the conspiracy and from firms with significant pipe-production capacity not party to the conspiracy, as well as wide and unanticipated fluctuations in the demand for pipe, made effective long-run price conspiracy impossible.¹⁷ The presumption, therefore, that these agreements allowed the recovery of monopoly prices and monopoly profits with no offsetting social benefits appears totally unwarranted. Such presumptions are probably unwarranted in several other classic price-fixing conspiracies as well.¹⁸

Conclusions

Criticism of the Per se approach to large mergers and horizontal price agreements should not be misinterpreted as unqualified support for a rule of reason. Although any rule of reason would be an improvement over Per se illegality, the rule-of-reason approach itself is fatally flawed. No antitrust policy can be scientifically rationalized for mergers and other horizontal business agreements; the law should neither help nor hinder them.

---

¹Robert H. Bork, The Antitrust Paradox: A Policy at War with Itself (New York: Basic Books, 1978), chap. 13. See also the classic article by idem, “The Rule of Reason and the Per Se Concept: Price Fixing and Market Division,” Yale Law Journal 75 (January 1966): 375–475.

²For an explanation of the Herfindahl Index, see David S. Weinstock, “Using the Herfindahl Index to Measure Concentration,” Antitrust Bulletin 27 (Summer 1982): 285–97. For a critical analysis of how the guidelines are applied in practice, see Robert A. Rogowsky, “The Justice Department’s Merger Guidelines: A Study in the Application of the Rule,” in Richard O. Zerbe Jr., ed., Research in Law and Economics (Greenwich, Conn.: JAI Press, 1984), vol. 6, pp. 135–66.

³Paul A. Pautler, “A Review of the Economic Basis for Broad-Based Horizontal-Merger Policy,” Antitrust Bulletin 28 (Fall 1983): 571–651.

⁴See, for example, Israel M. Kirzner, Discovery and the Capitalist Process (Chicago: University of Chicago Press, 1985), pp. 15–39. Even Robert Bork has admitted that the efficiencies associated with agreement are impossible to measure quantitatively and that the most important ones may be “intangible and develop gradually over time.” See Bork, “Rule of Reason,” p. 386.

⁵FTC v. Staples, Inc., 970 F. Supp. 1066 (D.D.C. 1997).

⁶William A. Niskanen, “Welcome to the FTC Follies! Kicking Around the Staples-Office Depot Merger” Legal Times, June 16, 1997, p. 26.

⁷For an excellent criticism of the antitrust assault on mergers, see William F. Shughart II, “The Government’s War on Mergers: The Fatal Conceit of Antitrust Policy,” Cato Institute Policy Analysis No. 323, October 22, 1998.

⁸See, for example, S.Y. Wu, “An Essay on Monopoly Power and Stable Price Policy,” American Economic Review 69 (March 1979): 60–72.

⁹Donald Dewey, “Information, Entry, and Welfare: The Case for Collusion,” American Economic Review 69 (September 1979): 587–94. See also H.B. Malmgren, “Information, Expectations, and the Theory of the Firm,” Quarterly Journal of Economics 75 (August 1961): 399–421.

¹⁰”Fixing the Price Fixing Confusion: A Rule of Reason Approach,” Yale Law Journal 92 (March 1983): 706–30.

¹¹ See, for example, Jerry A. Hausman, “Information Costs, Competition, and Collective Rate Making in the Motor Carrier Industry,” paper prepared for the Motor Carrier Ratemaking Study Commission, August 16, 1982.

¹²For an interesting discussion of these issues, see Paul R. Duke, “The Impact of the Removal of Antitrust Immunity on Collective Ratemaking in the Motor Carrier Industry,” statement before the Motor Carrier Ratemaking Study Commission, Boston, March 19, 1982. See also Andrew Popper, “The Antitrust System:.An Impediment to the Development of Negotiation Models,” American University Law Review 32 (Winter 1983): 283–334.

¹³Addyston Pipe and Steel Company et al. v. United States, 175 U.S. 211 (1899).

¹⁴United States v. Addyston Pipe and Steel Company et al, 84 F. Supp. 293 (1897).

¹⁵Bork, The Antitrust Paradox, p. 26.

¹⁶George Bittlingmayer, “Decreasing Average Cost and Competition: A New Look at the Addyston Pipe Case” journal of Law and Economics 25 (October 1982): 201–29; idem, “Price-Fixing and the Addyston Pipe Case,” Research in Law and Economics 5 (1983): 57–130. When this simple price agreement was declared illegal per se by the Court in 1899, the conspirators in Addyston formally merged to form a corporation called U.S. Cast Iron Pipe and Foundry.

¹⁷This condition may have been initially suggested in Almarin Phillips, Market Structure, Organization, and Performance (Cambridge, Mass.: Harvard University Press, 1962).

¹⁸See, for example, the discussion of the electrical-equipment conspiracy and other noted price-fixing cases in Dominick T. Armentano, Antitrust and Monopoly: Anatomy of a Policy Failure 2nd ed. (Oakland, Calif.: Independent Institute, 1990), chap. 5.