Strategies with minimal norm are optimal for expected utility maximization under high model ambiguity

Laurence Carassus,Johannes Wiesel

2024-01-14

Abstract:We investigate an expected utility maximization problem under model uncertainty in a one-period financial market. We capture model uncertainty by replacing the baseline model $\mathbb{P}$ with an adverse choice from a Wasserstein ball of radius $k$ around $\mathbb{P}$ in the space of probability measures and consider the corresponding Wasserstein distributionally robust optimization problem. We show that optimal solutions converge to a strategy with minimal norm when uncertainty is increasingly large, i.e. when the radius $k$ tends to infinity.

Optimization and Control,Probability,Mathematical Finance

What problem does this paper attempt to address?

This paper attempts to solve the optimality problem of strategy selection in the expected utility maximization problem under high model uncertainty. Specifically, the author studies the expected utility maximization problem in a one - period financial market, introduces adverse selection within the Wasserstein sphere to capture model uncertainty in this market, and considers the corresponding Wasserstein distributionally robust optimization problem.

### Core Problems of the Paper

1. **Expected Utility Maximization Problem**: Given an initial capital \(x_0\) and a constraint set \(D\subseteq\mathbb{R}^d\), the goal is to find the optimal portfolio weights \(w\in D\) such that the investor's expected utility is maximized under the worst - case scenario of uncertainty. The mathematical expression is:
\[
u(x_0):=\sup_{w\in D}\inf_{Q\in B_k(P)}E_Q[U(x_0 + x_0\langle w,X\rangle)]
\]
where \(U:\mathbb{R}\to\mathbb{R}\) is a non - decreasing, concave utility function, and \(B_k(P)\) is the \(p\)-Wasserstein sphere centered at \(P\) with radius \(k\).

2. **Case of Increasing Uncertainty**: When the uncertainty becomes larger and larger (i.e., \(k\to\infty\)), the optimal solution will tend to the minimum - norm strategy. Specifically, the main result of the paper shows that:
\[
\lim_{k\to\infty}w_k\text{ exists and belongs to }D = \{w\in D:|w|=\inf_{w\in D}|w|\}
\]

### Main Contributions of the Paper

- **Existence and Convergence**: Under certain assumptions, it proves the existence of the optimal portfolio selection, and when the uncertainty increases infinitely, the optimal solution converges to the minimum - norm strategy.
- **Handling of Model Uncertainty**: By using the Wasserstein distance to model Knightian uncertainty, it provides a new method to handle the model uncertainty problem in the financial market.
- **Theoretical Analysis**: It provides detailed theoretical analysis and proof to ensure the validity of the results under different conditions.

### Relevant Background

- **Merton (1969)**: The earliest to study the optimal investment problem under a fixed probability model.
- **Gilboa and Schmeidler (1989)**: Proposed the max - min method to handle the optimal investment problem under model uncertainty.
- **Wasserstein Distance**: Has received extensive attention in the fields of mathematical finance and statistics in recent years and is widely used in robust optimization problems because it can represent non - parametric neighborhoods.

### Conclusion

Through strict mathematical derivations, this paper proves that under high model uncertainty, the optimal solution of the expected utility maximization problem will converge to the minimum - norm strategy. This conclusion provides an important theoretical basis for understanding and handling uncertainty in the financial market.