A test of the life-cycle-permanent-income hypothesis: Cohort analysis of household wealth and portfolio holdings from 1977 to 1989
Abstract
The major objective of this study is to reassess the life-cycle-permanent-income hypothesis as an explanation of wealth accumulation using microeconomic empirical evidence. The technique of cohort data and cohort analysis is suggested by Deaton (1992) who referred it as a promising tool for studying household consumption and saving behavior. This study analyzes household wealth accumulation and portfolio holdings by constructing cohort data from three independently repeated cross-sectional surveys, the 1977, 1983, and 1989 Survey of Consumer Finances. The main results of this research can be summarized as follows: (1) The plots of mean and median total net wealth with cohort data indicate that wealth is a monotonic function of age, and thus, reject the assertion of a hump-shaped saving profile by traditional life-cycle or permanent-income hypothesis. (2) The results of cohort analysis using ANOVA models suggest that age, period, and cohort all significantly contribute to the variations of household wealth and portfolio holdings. The hypotheses of the main effects are completely supported. (3) The results of intra-cohort heterogeneity models for each cohort suggest that the major force of variation in total net wealth holdings is household income. That is, wealth accumulation traces income closely, and this finding is contrary to the assertions of conventional life-cycle or permanent-income hypothesis that wealth traces consumption but is independent of current income. The overall results of this study support the prediction of the life-cycle-permanent income hypothesis, and suggest that cohort analysis will be more appropriate and informative than cross-sectional studies. Implications are drawn for public policy and financial educators to assure the financial well-being of baby boomers during retirement years when there exists uncertainty in public and private pensions. Future research using cohort analysis needs to use as comparable data as possible to overcome the possible biases in analyses due to the pooling of cohort data.
Degree
Ph.D.
Advisors
Hong, Purdue University.
Subject Area
Finance|Economic theory
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.