significantly from other groups with respect to the mostly stock probability. Single
women are less likely than married women to choose mostly bonds. Using the same
data set, Jianakoplos et al. (2003) find that not marital status per se but differences in
financial endowments between married and nonmarried persons influence investment
behavior.
Using a large (22,000 respondents) German survey data set, supplemented with field
experiments, Dohmen et al. (2005) find that greater risk aversion is associated with
higher age and being female, and that these results are robust to income differences.
In an interesting addition, the German survey allows consideration of risk attitudes
across domains of career selection, sports and leisure, car driving, health, and
financial matters. Results of the study suggest that domain matters in that the level of
risk aversion or risk loving changes across situations, and financial lotteries do not
appear to be particularly good predictors of actual behavior. Women are more risk
averse than men in each domain.
Guiso, Haloassos, and Japelli (2002) contains studies on asset allocation choices in
several countries. The results presented reveal similarities but also differences across
national borders, giving weight to our consideration of both U.S. and German data.
Again, age, income, wealth, and gender turn out to be important factors.
According to the literature, therefore, gender, marital status, education, ethnicity,
wealth, income, and age all seem to affect risk-taking behavior.7
2.2 Optimal Asset Allocations and Individual Welfare Considerations - the
Normative Analysis
2.2.1 Optimal Asset Allocations
During the 1960s and 1970s, researchers gave significant attention to the problem of
optimal life-cycle allocation of resources between consumption and saving and
across types of assets (Phelps, 1962; Yaari, 1965; Mossin, 1968; Hakansson, 1969,
7 For further survey results see Barsky et al. (1997) and Eisenhauer and Ventura (2003).