Mixture Attitudes of Expectation-Based Loss Aversion

Koszegi and Rabin (2006, 2007) formulate loss-aversion models called Preferred Personal Equilibrium (PPE) and Choice-acclimating Personal Equilibrium (CPE) so successfully that many papers have been applied their models to a variety of economic fields in this decade. In this paper, without assuming any specific functional form, I show that these two loss-aversion models satisfy strong mixture aversion in the sense that a non-degenerate lottery is strictly preferred to any mixture between the lottery and its certainty equivalent. This property distinguishes these loss-aversion models from many other standard models such as the expected utility theory and disappointment aversion so that this result allows experimental economists to test a class of these models in a simple setting.