A Tree-Based Algorithm for Risk Management of Interest Rate Derivatives Portfolios
Posted: 22 Aug 1998
This paper presents a method for maximising the expected utility of terminal wealth of a portfolio of interest rate derivative securities with constraints primarily on the portfolio sensitivities. The constraints can be time and state dependent and can be enforced over the whole portfolio horizon. In this way an exact risk-return specification is allowed which is particularly useful to the risk management departments because various trading limits can be imposed effectively. To the proprietary desks of the bank, the method provides a clear-cut way of quantifying the various risks they are willing to take in order to achieve the maximum level of return. Despite the fact that the problem is formulated in the usual stochastic control framework, fortunately the proposed algorithm requires no enlargement of the lattice used to evaluate the various derivative securities. In this way a Non Linear Programming problem is formulated, and solved, with two types of algebraic constraints: those that reflect the local consistency conditions of the portfolio S.D.E. and those that reflect bounds on the portfolio sensitivities, minimum portfolio appreciations rates and various frictions in the market"
JEL Classification: G13, E4
Suggested Citation: Suggested Citation