Vertical Integration and the Informational Content of Prices
| Peter Klein |
Many years ago, when I was taking Williamson’s Economics of Institutions class at Berkeley and fishing around for dissertation topics, I had the idea to do some empirical work on the relationship between inflation and vertical integration or conglomerate diversification. The basic idea is that monetary expansion not only raises price levels, but also increases the dispersion of relative prices — introducing “noise” into the price mechanism — giving entrepreneurs an incentive to internalize transactions, on the margin, they would have otherwise conducted in the market. My interest was partly piqued by an off-hand remark by Dick in a review of Chandler’s Scale and Scope:
Things began to go wrong in the 1960s with the wave of conglomerate diversification, that is, with diversification by companies into areas wholly unrelated to their “core competence.” ITT was the paradigm of this phenomenon. Originally an international maker of telephone switching equipment, it bought, among other things, an insurance company and the maker of Hostess Twinkies. Chandler sees this as an inefficient practice, with many of the disbenefits of overextended British personal capitalism. There is no historical precedent for such unrelated diversification, he notes, except for German Konzerne during the hyperinflation of the 1920s. What is interesting — and what Chandler doesn’t mention — is that it is precisely inflation, in this case the Lyndon Johnson inflation of the 1960s, to which many have pointed as the cause of the wave of conglomerate mergers. The conglomerate is in effect an “internal capital market” that invests in a diversified portfolio of unrelated interests. But why? The stock market is much better at diversifying away risk than is such an arrangement, and it has many other advantages as well. In a time of inflation, the argument goes, price signals become distorted as managers find it difficult to disentangle changes in relative prices (that is, real prices) from changes in the price level. In such a world, the internal information and control within a conglomerate may have advantages that outweigh the disadvantage.
But, in any case, the trend in the less-inflationary 80s was the opposite one, the breaking apart of corporate holdings. . . .
The idea that conglomerate diversification, and “hierarchies” more generally, are responses to conditions in external markets has proven very useful in my own work; it also appears in Amar Bhidé’s neglected 1990 paper on diversification. Dick’s review cites a 1989 paper by Don Boudreaux and Bill Shughart linking US inflation rates and a measure of vertical integration but I couldn’t find such a relationship for diversification, and ended up going in a different direction.
My interest has been rekindled, however, by a couple of papers from Bob Gibbons, Richard Holden, and Michael Powell, “Rational-Expectations Equilibrium in Intermediate Good Markets” and “Integration and Information: Markets and Hierarchies Revisited.” These papers combine an incomplete-contracting model à la Grossman and Hart (1986) with various price-information models in the spirit of Grossman and Stiglitz (1980). Let me quote from the introduction to the second paper:
[W]hile Coase was explicit that the “price mechanism” is the chief alternative to internal organization, and Williamson’s (1975) title famously emphasized “Markets”as the alternative to hierarchy, over the next 35 years, the market disappeared from the literature on firms’ boundaries. Instead, the literature focused on non-integration versus integration at the transaction level, rather than the functioning of the price mechanism at the market level.
Omitting the price mechanism from the theory of the firm could be problematic. In particular, suppose (as seems plausible in the world and is being explored in a growing theoretical literature) that agents choose fi rm boundaries and internal control structures in part to affect incentives to gather and communicate information. Agents’ interests in choosing governance structures to strengthen these incentives will depend in part on how well market prices already perform this function. For example, if market prices are very informative, then agents will choose governance structures to improve incentives for other activities (say, cost reduction), effectively free-riding on the informativeness of the price mechanism. On the other hand, as Grossman and Stiglitz (1980) noted long ago, if everyone free-rides, then there will not be any information in prices. Thus, analyzing the choice between integration and non-integration for one dyad in isolation (rather than in the context of a market of analogous dyads making analogous choices) potentially commits two errors. First, each dyad takes the informativeness of the price mechanism as an important parameter in its choice of governance structure, but this parameter is hard to discern in most models of the integration decision. And second, this important parameter is endogenous: agents’ governance-structure decisions affect the informativeness of the price mechanism.
We view fi rms and the market not only as alternative ways of organizing economic activity, but also as institutions that shape each other. In this paper we explore how the informativeness of the price mechanism and firms’ integration decisions interact. To do so we analyze an economic environment that includes uncertainty. Formally, the uncertainty concerns consumers’ valuation of final goods, but we discuss other interpretations below. Parties can resolve this uncertainty at a cost. As in other rational-expectations models, the price mechanism both clears the market and conveys information from informed to uninformed parties. The fact that the price is not perfectly informative provides the requisite incentive for some parties to pay the cost to resolve the uncertainty.
Of course, this way of modeling uncertainty and the information content of prices doesn’t capture the Hayekian concept of tacit knowledge (Thomnsen, 1992; Kirzner, 1997, pp. 60-85), but the exercise is informative and highly recommended for students of the theory of the firm.