Scenario Simulation Method for Risk Management

Abstract The scenario simulation method described in this article is a computationally efficient alternative to the traditional Monte Carlo method. The method applies principal component analysis to select key factors, and approximates a multidimensional lognormal distribution of interest rates and exchange rates by a multinomial distribution of key factors. The computational efficiency is further enhanced by stratified sampling at multicurrency stage. Because of the efficiency, the model can afford a large sample size to obtain a satisfactory representation of return distribution for multicurrency fixed‐income and derivative portfolios. The method has wide applications in risk management.