Continuous-time smooth ambiguity preferences

Abstract

This study extends the smooth ambiguity preferences model proposed by Klibanoff et al. (2005) to a continuous-time dynamic setting. It is known that the original smooth ambiguity preferences converge to the subjective expected utility as the time interval shortens so that decision makers do not exhibit any ambiguity-sensitive behavior in the continuous-time limit. Accordingly, this study proposes an alternative model of these preferences that interchanges the role of the second-order utility function with that of the second-order probability to prevent the smooth ambiguity attitude of decision maker from evaporating in the continuous-time limit. By utilizing the utility convergence results established by Kraft and Seifried (2014), our model is eventually represented by the stochastic differential utility with distorted beliefs so that most existing techniques in economics and financial studies can be made applicable together with these distorted beliefs. We give an asset-pricing example to demonstrate the applicability of our model.