Abstract

We propose a model of a loss averse investor who aims to maximize his expected wealth under certain constraints. The constraints are that he avoids, with high probability, incurring a (suitably dened) unacceptable loss. The methodology employed comes from the theory of large deviations. We explore a number of fundamental properties of the model and illustrate its desirable features. We demonstrate its utility by analyzing assets that follow some commonly used nancial return processes: Fractional Brownian Motion, Jump Diusion, Variance Gamma and Truncated Lévy.