Robust Asset-Liability Management

TL;DR

This paper introduces a robust bond portfolio approach for asset-liability management that accounts for ambiguity aversion, constraints, and interest rate uncertainties, offering a computationally efficient and leverage-reducing solution.

Contribution

It develops a generalized immunization method based on ambiguity-averse preferences, with a simple least squares computation and improved out-of-sample performance.

Findings

The method effectively manages interest rate risk under various scenarios.

It reduces leverage through implicit regularization.

Numerical tests show superior performance over existing approaches.

Abstract

How should financial institutions hedge their balance sheets against interest rate risk when managing long-term assets and liabilities? We address this question by proposing a bond portfolio solution based on ambiguity-averse preferences, which generalizes classical immunization and accommodates arbitrary liability structures, portfolio constraints, and interest rate perturbations. In a further extension, we show that the optimal portfolio can be computed as a simple generalized least squares problem, making the solution both transparent and computationally efficient. The resulting portfolio also reduces leverage by implicitly regularizing the portfolio weights, which enhances out-of-sample performance. Numerical evaluations using both empirical and simulated yield curves support the feasibility and accuracy of our approach relative to existing methods.