A new preference model that allows for narrow framing (Q2050985)

In this article is proposed a new model for narrow framing that overcomes the limitations of the prominent mathematical model of narrow framing presented in [\textit{N. Barberis} and \textit{M. Huang}, J. Econ. Dyn. Control 33, No. 8, 1555--1576 (2009; Zbl 1170.91318)] (referred in this paper as BH). For making multiple decision or for evaluating multiple risks, they tend to consider one of them at a time, isolating it from other choice or risks, and, at the end of each period are assessed according to prospect theory so that the utility of gains and losses experienced by the individuals in these risks is calculated by a linear addition eventually using some for the agentconstant weighs. In the second part of the paper is showed that the BH model is not robust in that the total utility process can be nonexistent or nonunique. To resolve the issue of nonexistence or nonunique of the total utility process in the BH model, in the 3rd Section of the paper is propose a new model of narrow framing, referred to as the GH model. This reformulation shows that the GH model experiencing a loss/gain in risks that are evaluated separately reduces (increases) the certainty equivalent of the total utility proportionally. In the framework laid out here for the GH model, the authors do not assume any form for the utility of gains under the risk. The only requirement is the Markovian structure for the agent's consumption and investment. In this Markovian setting, it is reasonable that the agent only considers Markovian strategies for dynamic programming, as is presented into the article. In the section named Discussions of alternative modeling of preferences with narrow framing, there are presented two alternatives to weight the utility of gains and losses, leading to two variants of the GH model.