The elasticity of intertemporal substitution: new evidence from 401(k) participation

A key parameter in economics is the elasticity of intertemporal substitution (EIS), which measures the extent to which consumers shift total expenditures across time in response to changes in the effective rate of return. In contrast to the previous literature, which primarily has relied on Euler equation methods and generated a wide range of estimates, we show how a life-cycle-consistent econometric specification of employee 401(k) participation along with plausibly exogenous variation in rates of return due to employer matching contributions can be used to generate new estimates of the EIS. Because firms often cap the generosity of the match, employer matching generates nonlinearities in household budget sets. We draw on non-linear budget-set estimation methods rooted in the public economics literature, and using detailed administrative contribution, earnings, and pension-plan data for a sample of 401(k)-eligible households from the Health and Retirement Study, we estimate the EIS to be 0.74 in our richest specification, with a 95% confidence interval that ranges from 0.37 to 1.21.