Calibrating Risk Aversion in Additive Multivariate Utility Functions

Argyris, Nikolaos; Jaspersen, Johannes G. und Richter, Andreas (2020): Calibrating Risk Aversion in Additive Multivariate Utility Functions. Munich Risk and Insurance Center Working Paper, Nr. 35.

Abstract

Additive multivariate utility functions are common in applications of economic decision-making. They exist in many areas of multi-attribute decisions and feature prominently in several behavioral economic decision models. For predictions or welfare analyses using such models, it is often necessary to calibrate both ordinal and cardinal preferences in them. One aspect of cardinal preferences is risk aversion. However, the concept of risk aversion in additive multivariate utility functions is poorly understood. In fact, it is impossible to compare two additive multivariate utility functions solely with respect to their risk aversion regarding one or more attributes - changing their risk aversion changes ordinal preferences. We introduce the class of contextual additive multivariate utility functions and consider a subclass of increases in Arrow-Pratt risk aversion, namely those that increase risk aversion in the sense of Ross. In this setting we show that risk premiums change monotonically in risk aversion regarding a single attribute which eases the process of calibrating preference functionals. Additionally, ordinal preferences change in a sensible manner when Ross risk aversion is increased. We apply our procedure to calibration and risk premiums in the Kőszegi-Rabin decision model.

• Dokumententyp: Paper
• Keywords: Multivariate Utility Functions, Risk Aversion, Preference Calibration, Behavioral Economics
• Fakultät: Betriebswirtschaft > Institut für Risikomanagement und Versicherung
• Themengebiete: 300 Sozialwissenschaften > 330 Wirtschaft
• Sprache: Englisch