Knowledge And Decisions

Antitrust laws, like all forms of third party monitoring, depend for their social effectiveness on the articulation of characteristics objectively observable in retrospect, which may or may not capture the decision-making process as it appeared prospectively to the agents involved. There is usually nothing in antitrust cases comparable to finding someone standing over the corpse with a smoking pistol in his hand. Objective statistical data abound, but its interpretation depends crucially on the definitions and theories used to infer the nature of the prospective process which left behind that particular residue of retrospective numbers. For example, merely defining the product often opens a bottomless pit of complexities. Cellophane is either a monopoly — if the product is defined to include the trademarked name, which only Dupont has a legal right to use — or has varying numbers of competing substitutes, depending on how transparent and how flexible some other brand of wrapping material must be in order to be considered the same or comparable. Under some definitions or demarcations of transparency and flexibility, cellophane is monopolistic for lack of sufficient substitutes. But by other definitions it is in a highly competitive market with innumerable substitutes. The controversies following the Supreme Court’s decision as to whether cellophane was a “monopoly” (no) suggests that there were other definitions which some (but not all) legal and economic experts found preferable. The point here is that there is no objective and compelling reason to take one definition rather than another, though the whole issue often turns on which definition is chosen. In more complicated products, there are often numerous variations on the same goods, and which of these are lumped together as “the same” product determines what the market is and how much the producer’s share of the market is, as variously defined. For example, Smith-Corona has a smaller share of total American typewriter sales than of electric typewriters sold in the United States, or of all portable electric typewriters made by American manufacturers. For many products, so much is imported that a firm’s share of American production is economically meaningless: any American producer of single-lens reflex cameras would have a monopoly by definition; all such cameras are currently imported. But the purely definitional monopoly has no effect on economic behavior, in the face of dozens of foreign competitors.

What is involved here is not a technicality of antitrust law but a far broader question about the use of knowledge, and the role of articulation. The basic problem in these definition-of-product issues is that substitutability is ultimately subjective and prospective, while attempts to define it must be objective and retrospective.

Even where a product seems unambiguously definable in some plain sense — a tangerine, for example — a question may still arise as to the economic significance of such a definition. If a worldwide cartel were to gain control of every tangerine on the planet, they could still not double the price of the monopolized product without ending up with millions of unsaleable and spoiling tangerines in their warehouses, while consumers switched to oranges, tangelos, and the like. In short, even where the physical demarcation of a product seems obvious and unambiguous, its economic demarcation may be difficult or impossible. The extent to which the price of one product affects the sales of another product is what is economically important. As a practical matter, sellers can acquire an unarticulated “feel” for this in an ongoing trial-and-error process, but that is very different from third-party observers of retrospective statistics being able to objectively document irrefutable results to courts. For one thing, the discrete time units in which data are collected by observers may be far longer than the almost continuous time dimensions of the actual transactors’ ongoing experience, so that the observers’ data are more likely to represent an amalgamation of highly disparate price and sales fluctuations during the time interval studied.