1. The first of the alleged impediments has been well documented Spectacular evidence for the truth of this proposition was discovered in the famous 'Peckham experiment'of 1935-9, where 64 per cent of the ersons examined had identifiable disorders but were unaware of them In 1964, for example, it has been estimated that there were 150, 000 unknow diabetics in Britain. It also appears that the problem has similar dimen- sions in other countries. Such evidence appears to violate a fundamental and necessary condition for the attainment of an optimum through open First, the problem of ignorance is not a problem characterizing markets only. Knowledge must be economized in all social systems, and by using patient ignorance as a stick with which to belabour private medicine attention is diverted from the more important problem of assessing the optimal amount of ignorance. There is little evidence that post-war British patients are any less ignorant of their state of health than their pre-war parents were or that a nationalized health care system devotes more resources to preventing sick as than other systems (Office of Health Economics [37]). This, of course, is not evidence for or against the efficiency of any particular system of provision, but it is evidence for the view that the description of a theoretical optimum does not tell one how it may be achieved. Secondly, the inference ignores the possibility that the degree of ignorance measured in experiments such as that at Peckham may, in fact, be optimal. If information about one' s health is costly to collect, it may be irrational to dispel all ignorance; i.e. it is perfect information ather than ignorance, that is a priori more likely to be inconsistent with the postulates of welfare economics. The fact that one set of individuals es a social benefit in reducing the ignorance of others is a problem of xternalities, to which we shall return later, but it does no damage to the conclusions here that the observation of ignorance is not sufficient to infer inefficiency in resource allocation and that the specification of an optimal distribution does not indicate the most appropriate form of social organiza tion for attaining or approaching that optimum, What is required if a of how individuals operating within the framework of constraints implied by that form of organization can be expected to act compared with their behaviour under an alternative form. This, however, is 1 See Israel and Teeling-Smith [19], and the references cited therein. seems clear that for any hich it may indeed be(Culyer [12]), though there are some possible in practic Reproduced with permission of the copyright owmer. Further reproduction prohibited without permission.Reproduced with permission of the copyright owner. Further reproduction prohibited without permission