Publication Date


Document Type


First Advisor

Campbell, Carl M., III

Degree Name

Ph.D. (Doctor of Philosophy)

Legacy Department

Department of Economics


Cost and standard of living--United States; Consumption (Economics)--United States


Benchmark models of optimization, in the spirit of the Permanent Income Hypothesis, present a strong theoretical case for a smooth consumption regime in which households do not let consumption fluctuate with anticipated variations in income. However, numerous studies in the past have empirically rejected the primary underpinning of the theory. This study attempts to provide substantial insight into the nature of household consumption dynamics over a life cycle and to examine the cause for the failure of the baseline theory. The present study is divided into three parts. First, it provides a summary of the extant literature on household consumption. The review is followed by an extensive empirical exercise based on the most exhaustive US household panel dataset, the Consumer Expenditure Survey (quarterly, 1991 to 2000). The findings suggest that somewhere between forty-seven percent and seventy-six percent of US households significantly deviate from consuming their permanent income. The results are robust with respect to multiple specifications and instrument sets. Furthermore, the results indicate that consumption changes are highly sensitive to income declines, but not to income increases. Under myopia one would expect a symmetric breakdown, and under liquidity constraints consumption should be more sensitive to income increases than to income decreases. From the current results one can neither confirm myopia nor argue in favor of liquidity constraints as valid causes of theory breakdown. In the third part of this study an empirical investigation is first carried out to understand diverging patterns of sensitivities of some important components of household consumption. Results indicate that food and utilities consumption (both necessities) show no statistically significant sensitivity to income fluctuations. On the other hand, both alcohol consumption and expenditure on entertainments (both of which can be provoked by binges) show high sensitivity. Secondly, a strategy has been developed to test the habit persistence hypothesis, and tests have been undertaken to evaluate its validity. Estimates indicate strong habit persistence in US households, with an average strength parameter of 0.65. The projection bias generated by such strong habit persistence is capable of explaining the observed hump shape of household consumption over a life cycle and the declining savings rate in the decade of nineties.


Includes bibliographical references (pages [101]-103).


vii, 113 pages




Northern Illinois University

Rights Statement

In Copyright

Rights Statement 2

NIU theses are protected by copyright. They may be viewed from Huskie Commons for any purpose, but reproduction or distribution in any format is prohibited without the written permission of the authors.

Media Type