Full Insurance in the Presence of Aggregate Uncertainty

This paper tests implications of full consumption insurance. The object is to determine how much mileage can be obtained from a model with complete markets, with such features as private information or liquidity constraints omitted. The implication exploited is that individual consumption responds to aggregate risk but not to idiosyncratic risk. The test involves regressing the change in household consumption onto the change in aggregate consumption and other right-hand-side variables such as the change in household income and change in employment status. All variables other than the change in aggregate consumption are predicted to be insignificant in explaining the change in household consumption. With observations on consumption and income for 10,695 households from the Consumer Expenditure Survey, the results are mixed. The results for one specification (exponential utility) are mostly consistent with full consumption insurance; the results for the other specification (power utility) are not.