The theory of intertemporal choice predicts that the cross-sectional varian
ce of the marginal utility of consumption is equal to its own lag plus a co
nstant and a random component. Using general preference specifications and
some assumptions about the nature of the random component, we provide an ex
plicit test of this hypothesis. Our approach circumvents the necessity to i
dentify a pure age profile of the cross-sectional variance of consumption a
nd yields a well-specified statistical test. This test is applied to data f
rom the United States, the United Kingdom, and Italy. The results are remar
kably consistent with the restrictions implied by the theory of intertempor
al consumption choices.