A Bayesian decision model based on expected utility and uncertainty risk

Abstract

Risk is caused by the uncertainty of state of nature and a decision maker's selection, and the result may appear to be an unfavorable outcome. Therefore, a decision maker wants to maximize an expected return with minimal risk exposures. In this paper, we propose an expected utility and uncertainty risk (EU-UR) model based on the reference prior, which extends the classical decision model under uncertainty. The EU-UR model is made by making a compromise between measures of expected utility and uncertainty. The model is empirically validated by applying to the Levy's case and the Allais paradox. (C) 2014 Elsevier Inc. All rights reserved.