she operation of a zero price for use of the service(as under the NHS or even a full insurance system with risk discrimination) will ensure that individuals will, under either type of institution, try to adjust so that their marginal valuations become zero. Any differences between alternative institutions must arise because of differences in behaviour that are implied about how individuals can adjust by, for example, coinsurance in the market, or by political voting processes in the NHS case, which reduce total supply below the rate of consumption implied by zero marginal valuations. These are, however, more or less empirical questions about the processes by which preferred positions are attained and have very little on of an optimal solution. Before a judgement about the relative merits of the market or the nhs can be reached one therefore requires far more information about these processes and about how the excess demands under either institutional framework are actually removed. Currently, we lack even a satisfactory positive theory of anagerial behaviour as a framework in which such an empirical investi- gation could be conducted. 1 Thus, at the microeconomic level, one's preference for one system or another will depend upon how resources in excess demand are actually rationed out. Since a prime health service pplying agency, the hospital, is charac ofit in eithe system, there must be substantial initial uncertainty about any behavioura differences between them, which further whittles away any a priori case favouring either one over the other. This problem is additional to another concerning the determination of the size of the excess demand, which in the case of the NHS requires some theory of public expenditure(since marginal valuations are not, in practice, equated with zero)and in the market case requires analysis of the means used to reduce the effects of moral hazard 5. Uncertainty about the quality of care received by consumers is more important in the health services than in many other areas of economic life, since the patient is frequently prevented, in the nature of his case, from shop ping around and learning about the quality of the service of rive suppliers by trial and error. Even if second opinions are feasibly obtainable as they frequently are, the patient may not be able to weigh one against the other. The typical case, however, is probably that second opinions are obtainable but that they are not sought because of the mystique associated with the medical profession and the assumption that'the doctor knows est. The difference in the amount of information available to doctors on the one hand and to patients on the other is not of the same type as occurs with most other goods and services. Typically the producer knows more about the technical methods by which a product (in this case, say, a course of But see the references cited in p. 193, n. 1. Reference 13 has some further references of a more general kind. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.Reproduced with permission of the copyright owner. Further reproduction prohibited without permission