subgroups with relatively large welfare losses through suboptimal asset allocation
decisions. Specifically, individuals in the United States with low financial wealth and
in Germany with low income or medium wealth would benefit most from better asset
allocations. Further analyses exhibit the impacts of gender, education, and age.
The paper is organized as follows: A review of the literature in section 2 is followed
in section 3 by the definition and calibration of the normative benchmark model.
Section 4 describes the data used for our empirical investigation, and results of our
econometric analyses are presented in section 5. Building on the results of sections 4
and 5, a welfare analysis—the utility-based comparison of theoretically optimal and
actual behavior—is done in section 6. We summarize our results and derive policy
implications in section 7.
2 Relationship to Existing Literature
As summarized by John Y. Campbell in his 2006 presidential address to the
American Finance Association (Campbell, 2006), there are two general approaches
to research on household finance: positive research, which investigates the actual
behavior of people, and normative research, which aims to derive how people should
behave according to a set of rational criteria. Our contribution confronts the former
with the latter by measuring differences in utility (measured using a normative
model) for individuals whose asset portfolios we observe deviate from the
normatively given optimal asset allocation.
2.1 Empirical Evidence - the Positive Research
Because of their fundamental role in decision making under uncertainty, risk
attitudes are of significant interest to policymakers and economists. Variations in
attitudes across demographic and socioeconomic characteristics are relevant because
of the ultimate implications in setting public policy, particularly with regard to
financial, occupational, and similar decisions. Empirical evidence suggests that
gender, marital status, education, ethnicity, wealth, income, and age are among the
relevant factors associated with risk-taking behavior.
Halek and Eisenhauer (2001) use the University of Michigan Health and Retirement
Study to test the relationship between risk attitudes and a variety of demographic