Michael Schürle, University of St. Gallen
Abstract:
Increasing pressure from the market and regulatory requirements have forced banks to take the leap from a static view of the balance sheet towards a dynamic analysis of their positions. Multistage stochastic programming models are well suited in this context as decision support systems for the implementation of dynamic investment and hedging strategies. We discuss two applications to asset liability management problems of a typical bank: The first deals with the determination of replicating portfolios for positions without contractual maturity such as savings or sight deposits. The second application is the optimization of the hedging strategy for a bank's entire balance. Practical experience shows that the dynamic investment or hedging strategies, respectively, that are obtained from the stochastic programming models exhibit more favorable risk/return characteristics compared to traditional approaches based on simple decision rules.