This paper considers the macroeconomic implications of a set of empirical studies finding a high degree of dispersion in preference heterogeneity. It develops a model with both uninsurable idiosyncratic income risk and risk aversion heterogeneity to quantify their effects on wealth inequality. The results show that with the available estimates of the risk aversion distribution from PSID data the model can match the observed degree of wealth inequality in the U.S., accounting for the wealth Gini index in several cases. The model replicates well several features of the wealth distribution. However, the share of wealth held by the top 1% is still substantially underestimated. It is also shown that models without risk aversion heterogeneity underestimate the size of precautionary savings, and that the results are robust to both different income process specifications and to self-selection into risky jobs.
Wealth Inequality Heterogeneous Agents Incomplete Markets Computable General Equilibrium