Characteristics

Separability may be justified between goods that satisfy widely different needs. However, for other goods, it is their close similarities that attract attention. It is not on the basis of their essence that a consumer will choose between goods, but on the basis of the satisfaction they will produce.

Switching from tomatoes to eggs in deference to his original audience for the topic—an agricultural economics seminar at the Iowa State College of Agriculture and the Mechanic Arts in 1956—Gorman entitled his first paper on the topic of characteristics “A Possible Procedure for Analysing Quality Differentials in the Egg Market” (this paper finally appeared in 1980). Ever concerned with the interests of the applied economist, he saw the paper as a response to the need of Iowan farmers to understand what drove price differ­entials for eggs of different qualities.

The^, basic idea is simple: consumers buy different varieties of eggs solely for certain measurable characteristics (for example, he suggests, their vitamin content). If only two characteristics are relevant, then, given arbitrary prices, we may expect that at most two varieties of eggs will be bought by any given consumer; if three characteristics are relevant, at most three varieties.^[167] Only in the “degenerate case”, where the relative prices happened to be just right, would the consumer be indifferent between three or more varieties. But—and here is where things get interesting—as soon as we consider market equilib­rium, the prices will not be arbitrary: the degenerate case will prove to be the normal one, as it is ‘the only case in which every type of egg could find a sale^, (ibid.: 844; italics in original). This degenerate case can be characterised by a shadow price q of characteristic j such that, if purchased variety i delivered quantity a_ij of characteristic j, the price p_it of each variety i at each time t should always equal the value of the sum of its characteristics measured at the shadow prices:

Building on this insight, essentially an argument from the assumption that market prices should not embody arbitrage opportunities, Gorman proposed an empirical research agenda. The specific quantity of each characteristic delivered by each variety, though measurable for the consumer, is unknown to the researcher, as are the shadow prices.

For all his warmth towards the challenges faced by applied econometri­cians, Gorman had little real interest in pursuing applied empirical work. His attempts to operationalise the characteristics model on an ambitiously large scale using quarterly regional data on the consumption of over a hundred categories of food for 1956-1971 proved somewhat inconclusive (see Boyle et al. 1977).

Yet the characteristics model has assumed an empirical life of its own: far from egg or tomato markets, this insight now underpins the most widely used asset-pricing models in modern finance theory.^[169] After all, most financial assets are closely substitutable, and investors’ choices between them are largely driven by their potential to deliver a relatively small number of yield charac­teristics. Whereas Markowitz (1959) asserted that investors were seeking to balance portfolio risk and return, measured by mean and variance, modern theories allow the goals of investors to be unmeasured characteristics of the stream of future returns. Market-clearing prices of the various assets must, in these theories, be adapted to the shadow prices of these characteristics in the market, just as Gorman saw. Thus, such price processes are estimated by factor analysis-type methods (Campbell et al. 1997). Even the famous option pric­ing model of Black and Scholes (1972) and Merton (1973) appeals to pre­cisely the same arbitrage logic so lucidly presented by Gorman more than fifteen years earlier.

3.2