This paper examines whether there is a discrepancy in estimating the elasticity of intertemporal substitution using food consumption instead of nondurable consumption data. We show that a discrepancy exists and that the discrepancy increases as nondurable consumption increases.