Estimating the Intertemporal Substitution Elasticity
37 Pages Posted: 16 Feb 2009
Date Written: February 16, 2009
We estimate the intertemporal substitution elasticity (EIS) under a simple proxy method and instrument variable (IV) methods, addressing the seasonality effect, wealth-related fluctuation in relative risk aversion (RRA), and the time aggregation problem. The empirical tests are based on the pooled data consisting of 24 countries and two sub-samples, namely, continental European and emerging countries. The proxy method generates theoretically unacceptable negative EIS everywhere and its underlying assumption of martingale inflation is rejected. TSLS does hardly better, but GMM produces significantly positive EIS estimates. Enriching the regression equation with consumption seasonals and effects from asset markets - modeled as both main effects and interactions with the real interest rate, as theory suggests - again produces significantly positive, and higher, EIS. The resulting RRA, 1 to 4, chimes well with the numbers found in unconditional market-portfolio studies within a mean-variance framework.
Keywords: seasonality, inflation rate, instrumental variable, time aggregation, varying relative risk aversion
JEL Classification: G12, F43
Suggested Citation: Suggested Citation