We test the validity of three models of intertemporal consumption choice (the permanent income hypothesis, the myopic model, and the consumption insurance model) by means of mobility indexes for the underlying consumption distribution. Each of the three models delivers different transition laws for the consumption distribution. Estimating the transition laws requires knowledge of the evolution of the entire consumption distribution, not just of the conditional mean of consumption growth as in more standard procedures. We design an appropriate non-parametric test and confront these yet unexplored theoretical predictions with a panel drawn from the Bank of Italy Survey of Household Income and Wealth. We find substantial consumption mobility in the data. We then derive the predictions for consumption mobility of the theories of intertemporal choice and compare them with the actual mobility. The comparison rejects the theory of consumption insurance and the myopic model and is favorable to the permanent income hypothesis once we allow for measurement error in consumption.
