A Test of the Relative Income Hypothesis

Abstract

The original permanent income and life cycle models were constructed around the view that savings rates are independent of the level of long-run income [19]. Since then, it has been well documented using household data that savings rates rise with long-run income.' To be consistent with the empirical evidence, Modigliani suggests modifying the life cycle hypothesis to include a bequest motive. His hypothesis, however, is that the share of lifetime resources that a household plans to devote to bequests is an increasing function of its lifetime resources relative to others in the same age cohort [17]. The share of resources devoted to bequests is made a function of the relative, and not the absolute, level of resources so as to be consistent with Menchik and David's empirical study of the bequest function [16]. But if saving for bequests is influenced by relative income, why not other forms of saving? A simple model in which all types of saving are influenced by relative income was proposed by Duesenberry in 1949 [4]. His basic hypothesis was that households are concerned about community consumption standards and that this concern leads to savings rates being an increasing function of a household's position in the income distribution. Although a per se desire for status is seen by some economists as inconsistent with rational behavior, it is clear that a relative advantage is often instrumental in achieving absolute goals [6; 7; 10; 11; 22]. While several policy implications of the relative income hypothesis have come to light [2; 6], it has rarely been analyzed empirically. The purpose of this paper is to present a test of the relative income model. The hypothesis is that if income rank is really the variable driving savings rates, then it should be possible to hold the level of income constant and still see the average propensity to save (APS) moving with income rank. Section II presents a simple model of the savings decision in which concern for relative standing in the current consumption hierarchy is integrated into a permanent income framework. The model is then used to explore the issue of functional form. Section III discusses the data, the equation to be estimated, and the regression results. Section IV presents a summary and concluding remarks.