Cash-Invariant Hull Representation of Divergence Preferences

Battle Count: 76

The paper provides important theoretical foundations for risk-averse portfolio selection. The monotone mean-variance utility and its generalizations are directly relevant to portfolio optimization under model uncertainty. The cash-invariant hull representation offers a computationally friendlier formula for evaluating divergence preferences, which could simplify numerical implementations in portfolio allocation. The characterization of domains of monotonicity (Section 3.3) is practically useful for determining when classical mean-variance optimization coincides with its monotone modification. However, the paper is primarily theoretical and does not provide direct trading strategies or empirical results.

The theoretical framework requires deep knowledge of convex analysis, Fenchel duality, and functional analysis. Implementing the cash-invariant hull representation computationally would require solving optimization problems involving conjugate functions and subdifferentials. The case analysis in the proof (four subcases A-D) suggests that numerical implementations would need careful handling of boundary cases. The extension to L0 domains (beyond L-infinity) adds further complexity. However, the main formula (1.8) is conceptually simpler than the original divergence representation (1.1) for computation.

This is a purely theoretical mathematics paper with complete proofs. All results are self-contained with explicit references to standard convex analysis (Rockafellar). The main theorem (Theorem 1.1) and its proof are fully detailed, including all four subcases (A-D) for the optimizer's location. Corollaries, propositions, and examples are provided with complete derivations. No computational experiments are needed for verification.