Chapter Summary

Potential parties to an economic exchange often share many common goals, but in an important respect they must be viewed as adversaries. In both product and labour markets, both buyers and sellers face powerful incentives to misrepresent their offerings.

For messages between potential adversaries to be credible, they must be costly to fake. A firm with extensive sunk costs, for example, can communicate credibly that it offers a reliable product because if it fails to satisfy its customers, the firm stands to lose a lot of money. In contrast, a street vendor, for whom the costs of going out of business are very low, has a much more difficult time persuading buyers he offers high quality.

Messages between potential adversaries must also satisfy the full-disclosure principle, which means that if one party is able to disclose favourable information about itself, others will feel pressure to disclose parallel information, even if considerably less favourable. The producer of a low-quality product does not want to signal his product's inferior status by offering only limited warranty coverage. But unless he does so, many buyers will make an even less favourable assessment.

When a trading opportunity confronts a mixed group of potential traders, the ones who accept it will be different-and in some way, worse-on the average, than those who reject it. Cars that are offered for sale in the secondhand market are of lower quality than those that are not for sale; participants in dating services are generally less worth meeting than others; and so on. These are illustrations of the lemons principle, and are also sometimes referred to as examples of adverse selection.

Because of the problem of adverse selection, firms are under heavy competitive pressure to find out all the information they possibly can about potential buyers and employees. These pressures often translate into the phenomenon of statistical discrimination. In insurance markets, people from groups with different accident rates will often pay different premiums, even though their individual driving records are identical. This pricing pattern creates an understandable sense of injustice on the part of individuals adversely affected by it. In competitive markets, however, any firm that abandoned this policy could not expect to survive for long.

An important analytical tool for dealing with choice under uncertainty is the von Neumann-Morgenstern expected utility model. This model begins with a utility function that assigns a numerical measure of satisfaction to each outcome, where outcomes are defined in terms of the final wealth to which they correspond. The model says that a rational consumer will choose between uncertain alternatives so as to maximize his expected utility, a weighted sum of the utilities of the outcomes, where the weights are their respective probabilities of occurrence.

The central insight of the expected utility model is that the ordering of the expected values of a collection of gambles will often be different from the ordering of the expected utilities of those gambles. The differences in these rankings arise because of nonlinearities in the utility function, which in turn summarize the consumer's attitude toward risk. The concave utility function, any arc of which always lies above the corresponding chord, leads to risk-averse behaviour. Someone with such a utility function will always refuse a fair gamble, which is defined as one with an expected value of zero. A person with a convex utility function, any arc of which lies below the corresponding chord, is said to be a risk seeker. Such a person will always accept a fair gamble. A person with a linear utility function is said to be risk neutral, and is always indifferent between accepting and refusing a fair gamble.

Insurance purchased in private markets is generally an unfair gamble, not only because of the administrative costs included in insurance premiums, but also because of adverse selection and moral hazard. The fact that most people nonetheless buy substantial amounts of insurance is taken as evidence that risk aversion is the most empirically relevant case. This observation is further supported by the pervasiveness of risk-sharing arrangements like joint stock ownership.

The appendix to this chapter discusses search theory and the winner's curse.