Dynamic Estimation of the Consumer Demand System in Postwar Japan

Sasaki, K. & Fukagawa, Y. (1984). Dynamic Estimation of the Consumer Demand System in Postwar Japan. IIASA Research Report. IIASA, Laxenburg, Austria: RR-84-018

Preview

Text RR-84-018.pdf Download (941kB) | Preview

Abstract

This report explores the dynamic demand relations operative in Japan in the period 1951-80 in order to elucidate the dynamic nature and characteristics of the varied structures of consumer demand. The analysis was conducted at the subgroup level on the basis of time series of family budget data, using Powell's version of the linear expenditure system. A taste variable was incorporated into the expenditure functions and five alternative specifications of the taste variable were utilized to take account of recent structural changes in demand. The first two are based on current annual increase in income and on current annual rate of increase in income. The next two incorporate lagged annual increase in income and lagged annual rate of increase in income, and the last specification is based on the time trend.

The analysis was based on a 21-commodity breakdown and numerous individual segments of the total observation period were chosen for estimating the dynamic model. The taste variables had the effect of stabilizing the demand system as a whole and they considerably reduced the instability of important estimates, such as those for money flexibility, subsistence consumption levels, etc. Consumption patterns in Japan are considered to have changed substantially toward more "Westernized" living and eating habits since the beginning of the 1960s. Per capita consumption of rice and fish went down with the increase in deflated income, whereas the consumption of animal protein food, fruit, beverages, and food away from the home all increased rapidly. Owing to income and taste effects, transportation, recreation, and rent showed a notable upward shift in average shares, while rice consumption declined remarkably in terms of its reduced marginal share.

Broadly speaking, estimated average substitution elasticity in Leser's model is inversely proportional to estimated money flexibility, which itself has a close relation to price elasticities. High values of money flexibility were obtained for the lower levels of per capita income in the early years of the period studied. For periods of more rapid economic growth, money flexibility estimates dropped to some extent, while for recent years they rose appreciably, reflecting the smaller response of consumer demand to price changes.

Actions (login required)

View Item