A STUDY ON THE TRUE COST-OF-LIVING INDEX USING ALTERNATIVE UTILITY FUNCTIONS.

摘要

Study of the true cost-of-living index has interested economists for over a half century. The discussion has intensified during high inflationary periods but there have been comparatively few studies devoted to the topic. Of those undertaken, all have been done during periods of relatively stable prices. As a result, the biases found in the Consumer Price Index (CPI) have been consistently low.;In this study, the true cost-of-living index is derived from the traditional consumer utility maximization model and the substitution bias is estimated for the high inflationary period of 1972-I to 1982-I, by using alternative utility functions and disaggregated consumption data for the U.S. The weakly-separable utility function is initially assumed to be a three-level utility tree, but later modified as a two-level utility tree in order to allow more substitution among commodities. Two measures of the basis for making inferences about the relative contribution of the bias (namely, the measure of the change in relative prices and the degree of commodity substitution) are also explored.;This dissertation has obtained three main findings as follows: first, the substitution bias of the Laspeyres price index is large in the recent period of high inflation; second, the model used does make a significant difference in the size of substitution bias; and finally, the large bias is mainly caused by the large change in relative prices that are presumably induced by drastic increases in world-oil prices since the 1973 oil-embargo.;The bias of the Consumer Price Index occurs because the CPI--which is basically the Laspeyres Price Index--uses a fixed bundle of commodities in the base period as weights and, hence, does not allow the possibility of substitution away from those commodities that become relatively more expensive in a latter period. The CPI, therefore, contains an upward bias, called the substitution bias, in measuring the cost of living over time.