Econometrica: Jul 2018, Volume 86, Issue 4
Dynamic Mixture-Averse Preferences
To study intertemporal decisions under risk, we develop a new recursive model of non‐expected‐utility preferences. The main axiom of our analysis is called mixture aversion, as it captures a dislike of probabilistic mixtures of lotteries. Our representation for mixture‐averse preferences can be interpreted as if an individual optimally selects her risk attitude from some feasible set. We describe some useful parametric examples of our representation and provide comparative statics that tightly link decreases in risk aversion to larger sets of feasible risk attitudes. We then present several applications of the model. In an insurance problem, mixture‐averse preferences can produce a marginal willingness to pay for insurance coverage that increases in the level of existing coverage. In investment decisions, our model can generate endogenous heterogeneity in equilibrium stock market participation, even when consumers have identical preferences. Finally, we demonstrate that our model can address the Rabin paradox even in the presence of reasonable levels of background risk.
Log In To View Full Content
Supplement to "Dynamic Mixture-Averse Preferences"
IN THIS SUPPLEMENT, we provide several supporting results that are used in the main paper. In Section S.1, we provide a general local expected-utility result for mixture-averse preferences that nests Proposition 1 in the main paper as a special case. In Section S.2, we establish the relationship between mixture-averse preferences and several prominent non-expected-utility theories, including rank-dependent utility, betweenness, disappointment aversion, and cautious expected utility. Section S.3 describes the implications of preference for diversification for insurance demand, and discusses how preference for diversification is equivalent to risk aversion for either rank-dependent utility or any preference that is quasiconcave in probabilities. Section S.4 establishes the existence of a value function for the optimal risk attitude representation. Proofs are contained in Section S.5.