Note: Read the prior installments of this essay series here: Part I; Part II; Part III.
In Chapter 6 of Antitrust: The Case for Repeal, Dominick T. Armentano discusses antitrust policy’s approach toward horizontal agreements, such as mergers, price collusion, cartel attempts, and market division agreements – most of which are still prosecuted to the fullest extent. For some of these agreements, the rule of reason is intended to weight the costs against the benefits of the particular practice and thus to give discretion to the courts to decide whether the particular action is acceptable. But other practices, such as price collusion and division of market agreements, remain per se illegal. The advocates of this per se illegality defend it via the claim that the costs of such practices always outweigh their benefits.
Problems With the Rule of Reason
There exists a fundamental flaw with the rule of reason approach. Namely, evaluation of costs and benefits of a particular practice is subject to the Hayekian Knowledge Problem. How can a judge or any particular individual know fully what the costs and benefits of an action are? Economists generally recognize that the costs and benefits are subjective to the economic agent, and yet even many of them frequently ignore the implications of this idea.
Many judges, furthermore, know little about economics and are unaware of the difficulties involved in arriving at a genuine understanding of costs and benefits. The rule of reason presupposes an ability to measure benefits and costs, but in markets, benefits and costs are evaluated by the acting individuals. These actors are able to accurately judge benefits and costs, as their actions directly concern their own welfare. But the rule of reason uses the vague notion of “social costs,” usually measured on the basis of market output. Courts, under the rule of reason, tend to evaluate an activity based on how much output restriction it results in – which the courts extrapolate from market share data. The courts tend to reason as follows. As a result of increased market share of the firm, the firm has higher market power, which leads to a restriction of output. The Neoclassical monopoly model is used to make these causal leaps; using this model entails making massive assumption with regard to market factors such as elasticity of demand – but these assumptions are by no means necessarily realistic.
Furthermore, in order to know what is happening to market share, one must know what the relevant market is and how to define it. Yet the relevant market is different for each person and cannot be objectively defined. Courts, in providing such a definition, attempt to include all the substitutes for the good in question and nothing else, but substitutes can take virtually any form; they need not be direct. What is or is not a substitute depends on the definition of the relevant market – so it is folly for the judges to attempt to define the relevant market based on what is or is not a substitute.
Some economists have tried to offer the following criterion for determining what counts as a substitute of a good. This criterion uses the cross-price elasticity measurement. Assume that A and B are goods. Then cross-price elasticity = (%Change in quantity demanded of Good B)/(%Change in price of Good A). This measure can, according to some, determine whether A and B are substitutes or complements and thus help figure out what the relevant market is.
However, there also exist non-price aspects that determine competition among goods. Competition is not just a matter of price; non-price rivalry may bring about the existence of substitutes that the cross-price elasticity measurement cannot detect.
In the absence of an unambiguous definition of the relevant market, it is impossible to determine with scientific certainty what changes in concentration will do to consumers in a particular market.
Furthermore, there exists no strict relationship between market share and market power; the latter does not necessarily follow from the former. Moreover, how does one know whether a particular value of the Herfindahl-Hirschman Index (HHI) necessarily corresponds to a monopoly or a firm with “too much” market power? At what value of the HHI does a monopoly begin? Why is such a value (say, X), a superior choice to X+1 or X-1?
Social benefits of horizontal mergers can include economies of scale, economies of scope, and various efficiencies from financing, advertising, distribution, research, and development. Some of these are measurable; others cannot be; many are subjectively determined in the entrepreneurial judgment of firms’ managers. Entrepreneurs may have projections that may not be accessible to or justifiable by someone else. Only market competition can determine if these projections are ultimately valid.
The Staples Case (1997)
The Staples Case is, according to Armentano, one of the most disastrous recent antitrust cases. Staples and Office Depot sought a merger; there existed ample competition in the realm of office supplies, which were available in virtually every major store. But antitrust authorities insisted on a definition that confined the relevant market to office supply superstores – i.e., only Staples, Office Depot, and Office Max. Thus, antitrust authorities claimed that a merger between two of the superstores would greatly increase the market share of the merged firm. But if all the suppliers of office supplies were taken into consideration, Office Depot and Staples together would have comprised only about 5% of the relevant market – clearly insufficient for antitrust authorities to block a merger.
Problems With the Per Se Approach
While the rule of reason approach may have its flaws, per se illegality is worse; it presumes that the benefits of particular actions, such as price collusion and division of market agreements, can never outweigh the costs. But it is possible that the contrary is the case.
Price coordination among competing firms, especially in uncertain times, can often allow for higher output than would be produced otherwise – as historical examples demonstrate. Furthermore, there sometimes exist benefits from firms sharing information with one another without engaging in constant price adjustments, which are costly – sometimes even from the perspective of the buyers, who tend to dislike overly fluctuating and unpredictable prices. There is no justification for concluding that cartels and price collusion are never appropriate in any situation.
Often, horizontal cooperation can be a better form of organization than a horizontal merger, because the former is more flexible; temporary price agreements can change more readily with economic conditions. Furthermore, cartels are fundamentally unstable in a free market – so even if a given cartel is deleterious to consumers, it is unlikely to last for long. Why, argues Armentano, should antitrust authorities prevent horizontal cartels if some of them might be beneficial and most of them will not endure anyway?
Armentano concludes in Chapter 7 by summarizing his stance. Antitrust laws must be repealed because they are a violation of liberty: they infringe on individual private property rights and ability to act in non-coercive ways. Furthermore, antitrust laws are based on a nebulous notion of “social efficiency” that is difficult to measure, determine, or define; these laws often fail to achieve even their stated objectives. Finally, antitrust laws inhibit the voluntary, spontaneous individual plan coordination which often enhances people’s well-being in a free market.
Even Adam Smith, who lamented businessmen’s tendency to collude to the detriment of the consumer, recognized in the same passage from The Wealth of Nations that it is impossible for government to prevent such collusion in any manner consistent with liberty and justice. In short, antitrust laws are an utter failure theoretically and practically; they must be done away with before they inflict further damage.
Pongracic, Ivan. Second Lecture on Armentano’s Antitrust: The Case for Repeal. Hillsdale College. Hillsdale, MI. December 4, 2007.
All lecture material is used with explicit permission.