On the contribution of the Markowitz model of utility to explain risky choice in experimental research

It is becoming increasingly common to accept that heterogeneity of preferences is an appropriate approach to describe aggregate experimental data on risky choice. We propose a parametric form of utility consistent with Markowitz’s (1952) hypotheses as a useful model to consider. This value function exhibits the fourfold attitude to risk and can also capture different combinations of risk attitudes and higher-order preferences. Moreover, it can be combined with probability weighting functions as well as with other value functions as part of mixture models that capture heterogeneity of preferences. We employ data from three recent experimental studies and show that this model can contribute to the explanation of their findings.