One channel through which real interest rates affect aggregate demand is consumer expenditure on durable goods. This paper examines empirically the link between interest rates and consumer durables. Solving for the decision rule relating income and interest rates to consumer demand is in general an intractable task. This paper avoids this problem by examining the first-order conditions necessary for maximization by the representative consumer. The estimated model suggests that expenditure on consumer durables is far more sensitive to changes in the interest rate than is expenditure on nondurables and services.
I. Introduction
THE interaction between interest rates and consumption decisions plays an important role in a variety of macroeconomic models. In the textbook IS-LM model, the interest elasticity of consumer spending is one determinant of the monetary and fiscal policy multipliers. In macro models based on intertemporal substitution, the real interest rate is the relative price that adjusts to equate desired expenditure with the endowment or production level. One channel of these effects is consumer expenditure on durable goods. The purpose of this paper is to examine empirically the link between interest rates and consumer durables expenditure.'
Understanding fluctuations in consumer purchases of durables is vital for understanding economic fluctuations generally. Durables might at first seem easily ignored, since expenditure on non-durables and services is about seven times expenditure on durables. Durables expenditure, however, is the most volatile component of consumer expenditure. The standard deviation of the growth rate in non-durables and services, fourth quarter to fourth quarter, is only 1.4%, while the standard deviation for durables is 8.5%. Durable goods industries thus play a central role in the business cycle.
Despite their importance, there has been little recent work on consumer durables. Moreover, the studies that do exist largely ignore the effects of interest rates. In Mankiw (1982), I examine the implications of the permanent income hypothesis for durable goods. In a more general model, Bernanke (1985) emphasizes the role of adjustment costs. Both studies provide striking evidence contradicting the permanent income hypothesis, yet neither allows a variable and uncertain real rate of return. These rejections are possibly attributable to unwarranted assumptions regarding the real interest rate. Indeed, expenditure on durables is often thought to be very interest sensitive. One goal of this paper is to measure the interest sensitivity of consumer expenditure both on durables and on non-durables and services.
In his critique of econometric policy evaluation, Lucas (1976) criticizes the use of a standard reduced-form consumption function. Hall's (1978) seminal paper initiates one approach to estimating structural parameters while avoiding the problems Lucas points out. Instead of estimating the reduced-form decision rule relating income and interest rates to consumer demand, Hall examines only the first-order condition necessary for an optimum (the Euler equation). Much recent work uses this approach to study consumption and asset returns in the case of non-durable goods.2 In my paper with Rotemberg and Summers (1985), we examine intertemporal substitution in both consumption and labor supply. All these studies ignore