ARAKI Masuo Aoyama Gakuin University, Research Institute, Researcher, 綜合研究所, 特別研究員
GOTO Fumihiro Aoyama Gakuin University, Department of Economics, Professor, 経済学部, 教授
HONGO Shigeru Aoyama Gakuin University, Department of Economics, Professor, 経済学部, 教授 (90082867)
|Budget Amount *help
¥2,000,000 (Direct Cost : ¥2,000,000)
Fiscal Year 1997 : ¥800,000 (Direct Cost : ¥800,000)
Fiscal Year 1996 : ¥1,200,000 (Direct Cost : ¥1,200,000)
As for the pconsumption function, two characteristics have been pointed out. The fist is that the cross-section consumption function based on macro-economic data has been stable for a long period of time, approximately a line from the origin, that is, C=bY.The second is that the consumption function for micro-data, which is also approximated by a straight line with positive intercept, C=a+bY(a>0) and this intercept has been increasing as time gies on. Economic interpretation for these phenomena is so-called "relative income hypothesis."
We have found, contrary to the above statement, that cross-section consumption function has been quite stable in the post-war period in Japan. Naimely, the relationship between real income and real consumption is almost a straight line with exception for several years around 1973. This fact strongly conflicts with the traditional relative-income-hypothesis. Hence the following two problems arise :
(1) what mechanism kept cross-section consumption function stable, and (2) why consumer behaved quite differently around "oil-crisis" period.
As for the second point, Funaoka's paper, that discusses the effects of expected inflation on consumer's behavior, seems to give a theoretical background. Yet, if we apply Funaoka's estimation procedure now, we cannot obtain a similar result due to strong collinearity among economic variables.
In our research we tried to clarify the recent consumption structure using household survey data, in particular micro-data from Households Income and Expenditure Survey. We showed a possibility of resolving the paradox by applying a processing of the time series data, combined with a model that incorporates the effects of the assets.
Another contribution we made was a survey and theoretical examination of "bootstra method, " as one of the most promising computer-intensive methods that can be applied to econmic analyzes.