|
Description
For a panel of OECD countries, we show that the magnitude of the estimate for the excess sensitivity of private consumption to current income cannot be explained by a model based on certainty equivalence.
JEL code: E21
Keywords: excess sensitivity, certainty equivalence, current income consumers, panel data, GMM.
WE INVESTIGATE THE IMPACT of the assumption of certainty equivalence on the estimation of the excess sensitivity of private consumption to current income for a panel of 17 OECD countries using annual data over the period 1981--2003. The contribution of the paper is theoretical and methodological. Theoretically, we derive a nonlinear consumption function from a model that allows for the presence of two consumer types: rule-of-thumb current income consumers and optimizing, forward looking but prudent consumers. While the presence of rule-of-thumb consumers has been extensively investigated in time-series studies (see Campbell and Mankiw 1990, 1991, and a large number of subsequent studies) and, on occasion, in panel data studies (see Evans and Karras 1998, Lopez, Schmidt-Hebbel, and Serv6n 2000), the novelty of this paper is that we explicitly consider forward-looking consumers with a precautionary savings motive which is captured by the possibility that their marginal propensity to consume out of wealth is lower than under certainty equivalence. In this paper, we argue that restricting the marginal propensity of forward-looking consumers to its value under certainty equivalence leads to a severe overestimation of the fraction of current income consumers. As noted by Zeldes (1989), under certainty equivalence and with time-separable utility, the growth rate of consumption depends only on the difference between the interest rate and the rate of time preference. If this difference is small or negative then the growth rate of consumption will be small or negative as well. In models with rule-of-thumb consumers the positive growth rate of consumption observed in aggregate data can then only be attributed to disposable income growth leading to an overestimation of the fraction of rule-of-thumb consumers and thus to an overestimation of the degree of excess sensitivity of private consumption to current income.
Methodologically, we note that Euler equation studies imply consumption equations that are either in first differences, in quasi-differences, or in growth rates. Panel data estimation in the presence of endogenous variables necessitates a first difference transformation to get rid of country-specific heterogeneity (i.e., the estimators of Anderson and Hsiao 1982 and Arellano and Bond 1991). (1) If the equation that is estimated is in first differences, in quasi-differences, or in growth rates to begin with, a first difference transform implies a large loss of information and potential finite sample biases if the instruments are weak. Therefore, unlike the previous papers on excess sensitivity that use a panel methodology, we use a generalized method of moments (GMM) estimator that employs moment conditions of the type suggested by Arellano and Bover (1995) and Blundell and Bond (1998). These moment conditions use information from the untransformed consumption function and not just information from the first difference transform of the consumption function. Further, the use of a macro panel implies a relatively large time dimension and therefore a large number of available moment conditions. Since GMM estimators can also be subject to finite sample biases when the number of moment conditions is too large (see Tauchen 1986), we employ linear combinations of the moment conditions instead of the actual moment conditions in the estimation. Our estimator is therefore a system GMM version of the "stacked" instrumental variables estimator as discussed for instance by Arellano (2003, p. 170).
Our results suggest, first, that about 37% of consumers in OECD countries are rule-of-thumb current income consumers. This fraction is in line with the existing literature. Second, our tests reject the certainty equivalence hypothesis for the remaining forward-looking optimizing consumers. Third, we argue that inappropriately imposing certainty equivalence in our model leads to a... |
More articles from Journal of Money, Credit & Banking
Erratum to "Why Has U.S. Inflation Become Harder to Forecast?".(Correc..., October 01, 2007
Looking for additional articles?
Click here
to search our database of over 3 million articles.
|
