Mathematical Finance, Stochastic Analysis, and Machine Learning By Mihai Sirbu: On the Analyticity of the Value Function in Optimal Investment and Stochastically Dominant Markets

Speaker:

Mihai Sirbu, University of Texas at Austin

Title:

On the Analyticity of the Value Function in Optimal Investment and Stochastically Dominant Markets

Abstract:

We study the analyticity of the value function in optimal investment with expected utility from terminal wealth and the relation to stochastically dominant financial models. We identify both a class of utilities and a class of semi-martingale models for which we establish analyticity. Specifically, these utilities have completely monotonic inverse marginals, while the market models have a maximal element in the sense of infinite-order stochastic dominance. We construct two counterexamples, themselves of independent interest, which show that analyticity fails if either the utility or the market model does not belong to the respective special class. We also provide explicit formulas for the derivatives, of all orders, of the value functions as well as their optimizers. Finally, we show that for the set of supermartingale deflators, stochastic dominance of infinite order is equivalent to the apparently stronger dominance of second order. Based on joint work with Oleksii Mostovyi and Thaleia Zariphopoulou.

 

Mathematical Finance, Stochastic Analysis, and Machine Learning