Estimation of Complete Demand System Under Data Limitation

Abstract

A common difficulty in estimating demand functions for developing countries is the lack of time series data. With the available cross-sectional data resulting from extensive surveys on households, most researchers have concentrated on the estimation of expenditure elasticities by ignoring the possible impact of prices. Obviously, the results of this partial kind of analysis may not be very reliable for policy design. Deaton (1988) introduced a methodology for using household survey data to estimate price elasticities by making use of spatial variation in prices, but its application requires certain data conditions that may not always be met, as, for example, in the case of Turkey.This article provides a practical solution for such cases by observing that regularly collected data on the cost of grouped commodity bundles across regions can reflect spatial variation in prices, and, thus, can be used as a proxy for the prices. As compared to previous studies estimating only the expenditure elasticities for Turkey, our results differ,in some cases by large margins. Thus, incorporation of prices into the demand analysis seems to be important not only in obtaining price elasticities but also in obtaining more reliable estimates of expenditure elasticities.