Do Consumption and Income Have a Long Run Relationship?

Abstract

This paper provides some new empirical evidence on the consumption-income relation which is one of the most thoroughly studied subjects in economics. According to the recent literature in economics the two variables should be cointegrated for many theoretical results in economics, such as the permanent income hypothesis, to be meaningful. Our initial empirical results, however, show that cointegration between income and consumption is not well confirmed for U.S. quarterly data for extended postwar periods. This is an important problem that has to be addressed in the literature. In this paper we conjecture that failure of confirming cointegration for the consumption-income relation is due to nonstationary fluctuations in some relatively short period(s) although the relation prevails in the majority of data period. Our empirical result confirms our conjecture. Two periods of "short-run" nonstationarity are identified for an extended postwar era of the U.S. economy: One is the Volker era in the early 1980's and the other consists of the recent years of unusually low interest rate. Our result has important implication for empirical analysis in economics where consumption and income variables are involved.