Nazarov L.V.

The paper describes a modified scenario method for risk estimation of financial instruments. It is based on the scenario method proposed by Jamshidian F. and Zhu Y. allowing to significantly reduce computation time of risk indicators on large portfolios compared to the Monte Carlo method. It is proposed to change the way of choosing scenarios and points of approximating distribution. It allows to improve the quality of approximation. More than that, the new method makes possible to remove restrictions on the type of distribution of portfolio price factors allowing to expand the scope of its application. The comparison of VaR (value at risk) estimation using original and modified method was performed on a financial portfolio consisting of an interest rate swap for cases when the values of the price factors have normal distribution, gamma distribution and Student’s t-distribution.