In this book, The Measurement of Social Welfare, I have examined the possibilities for a useful, rather than merely a formally correct, welfare economics. I have focused on the question of what the fundamental criterion of social welfare must be like in order that a welfare analysis based on it be especially relevant to guide policy decisions. Three characteristics seem appropriate. First, the standard must be internally consistent. Second, it must enable us to compare a large proportion of the policy alternatives before use. Third, it must be compatible with relevant values prevailing in the community.
The point of departure of the study is the finding by Arrow that if the standard be fashioned to reflect a particular interpretation of "rational" evaluation and in addition bear the kind of correspondence to the values of the individuals in the community which we call consumer sovereignty, then the standard will necessarily be internally inconsistent. This finding has had a strong impact, largely in convincing economists that almost nothing in the way of practical help can be expected from welfare economics. Very little of the immersion of economists in matters of public policy in recent years has been significantly informed by the formal corpus of welfare theory. This book is partly an attempt to see exactly how restrictive Arrow's finding is, and how best to by-pass it so as to form a useful analytic apparatus.
Part I sets forth the basic problem and examines Arrow's analysis and implications in some detail. Parts II, III, and IV examine a wide variety of approaches in the welfare literature in order to discover how useful are the criteria of choice which they generate. The approaches falling under the rubric, "New Welfare Economics," are analyzed, as well as approaches which introduce varying degrees of cardinality into individual utility indicators, including two chapters on approaches based on the Von Neumann- Morgenstern expected utility hypothesis. In addition, approaches resting on specific group uniformities in individual preferences and on group value judgments concerning interpersonal comparability of social welfare impact, are analyzed. These sections of the book provide an extensive, reasonably complete survey of the welfare literature of the past twenty-five years. They incidentally provide as well an extensive survey of the non-normative field of utility and decision theory. The discussion is self-contained through-