Analysing Russian food expenditure using micro-data

Type Working Paper
Title Analysing Russian food expenditure using micro-data
Author(s)
Publication (Day/Month/Year) 1999
URL https://www.econstor.eu/bitstream/10419/28552/1/311675956.pdf
Abstract
Since the beginning of transition, the level and structure of average food consumption and expenditure of Russian households has changed substantially. This development has gone together with a steep increase in the share of food in total expenditure. Notable differences with respect to food expenditure are observed between distinct household strata. In this paper, food demand of Russian households is investigated. For this purpose, households are classified by sociodemographic characteristics, and differences between food demand patterns of various household types are described using data of a Russian household survey of 1996. Russian food demand is econometrically estimated for seventeen food commodities belonging to five groups using a two-stage linear approximation of the Almost Ideal Demand System (LA/AIDS). Total expenditure allocation on food and non-food is analysed using Working's Engel model. The basic models are extended by sociodemographic factors. In a first step, unit values of food commodities are adjusted for quality differences and Probit analyses are carried out to analyse the decision to purchase food commodities. In a second step, the Engel model and the LA/AIDS are estimated applying the Generalised Heckman procedure in order to account for estimation bias introduced from zero expenditures. The estimates are used to calculate total expenditure and own price elasticities for different household groups. The results indicate that sociodemographic characteristics exert an important influence on the level and composition of food expenditure and on food demand elasticities. Therefore, if demand analysis shall contribute to the design of comprehensive food and social policies, not only average estimates for the population as a whole, but estimates for specific population groups should be considered.

Related studies

»