Expected subjective value theory (ESVT): A representation of decision under risk and certainty

We present a descriptive model of choice derived from neuroscientific models of efficient value representation in the brain. Our basic model, a special case of Expected Utility The-ory, can capture a number of behaviors predicted by Prospect Theory. It achieves this with only two parameters: a time-indexed payoff expectation (reference point) and a free pa-rameter we call predisposition. A simple extension of the model outside the domain of Expected Utility also captures the Allais Paradox. Our models shed new light on the com-putational origins and evolution of risk attitudes and aversion to outcomes below reward expectation (reference point). It delivers novel explanations of the endowment effect, the observed heterogeneity in probability weighting functions, and the Allais Paradox, all with fewer parameters and higher descriptive accuracy than Prospect Theory.(c) 2023 Elsevier B.V. All rights reserved.

Automated summary

We propose a descriptive model of choice based on efficient value representation in the brain. Our model is a special case of Expected Utility Theory and can explain behaviors predicted by Prospect Theory. With only two parameters, it captures the Allais Paradox and sheds light on the computational origins and evolution of risk attitudes and aversion to outcomes below the reference point. It offers novel explanations of the endowment effect, heterogeneity in probability weighting functions, and the Allais Paradox with fewer parameters and higher descriptive accuracy than Prospect Theory.