Risk Tolerance and Asset Allocation.

Abstract

Economic theory assigns a central role to risk preference in asset allocation. This dissertation includes three papers that investigate this relationship empirically. The first paper uses panel data on hypothetical gambles over lifetime income in the Health and Retirement Study to quantify changes in risk tolerance over time and differences across individuals. The maximum-likelihood estimation of a model with correlated random effects draws on detailed information from 12,000 respondents in the 1992-2002 HRS. The results support constant relative risk aversion and career selection based on preferences. While risk tolerance changes with age and macroeconomic conditions, persistent differences across individuals account for 73% of the systematic variation in preferences. The measure of risk tolerance also relates to actual stock ownership.

The second paper develops a measure of relative risk tolerance using responses to hypothetical income gambles in the HRS. In contrast to most survey measures that produce an ordinal metric, this paper shows how to construct a cardinal proxy for the risk tolerance of each survey respondent. The paper also shows how to account for measurement error in estimating this proxy and how to obtain consistent regression estimates despite the measurement error. The risk tolerance proxy is shown to explain differences in asset allocation across households.

The third paper investigates whether the characteristics of household labor income can account for the observed heterogeneity in financial portfolios. Households differ substantially in the riskiness of their labor income and in the magnitude of their labor income relative to their financial assets; however, the results of this paper suggest that households do not integrate their human capital in their financial asset allocation. This analysis uses direct, household-level comparisons between actual stock allocations and predicted allocations in three economic models with different assumptions about labor income. When labor income is excluded from the model, the correlation between actual and predicted stock allocations is 0.16. The inclusion of certain or risky labor income in the model leads to negative correlations of -0.12 and -0.06 respectively. There is no evidence that households view their wealth broadly and diversify risks across their financial assets and human capital.