In 1976, when Pearte and I showed that a well-behaved consumer preference function could be expanded as a Taylor series to represent the money-metric solely in terms of the parameters of ordinary demand functions, we wondered why this possibility had not been perceived earlier. Indeed, several colleagues and friends suggested that the general idea was well known, but when suitably challenged they were unable to come up with convincing references. It may very well be the case that many diligent readers who have struggled with the arguments contained in the previous chapters will be of the same opinion. Certainly, most first-year graduate students will be aware of the tools of analysis that have been used in this volume. The marginal utility of money emerges directly from the assumption that consumers maximize their satisfaction subject to a budget constraint. The cost-of-utility function is now well ensconced in the current generation of textbooks dealing with microeconomic theory. The numerical methods used are of a very basic kind. The only difference between the approach taken here and the conventional literature of applied welfare economics is the manner in which the constituent parts are assembled and developed.

Yet the question remains: If this is so, why have consumer surplus and related index-number techniques continued to attract such widespread attention and use? In the Preface to this book, and subsequently in Chapter 4, we noted that almost a century ago Alfred Marshall had become increasingly skeptical of consumer surplus.