Copula-Models in Foreign Exchange Risk-Management of a Bank

The paper deals with the optimization problem aiming to maximize the expected return given the amount of the bank’s open currency positions subject to the level of foreign exchange risk. The goal of the paper is to compare the efficiency of problem-solving assuming either multivariate normality, or using copulas to semi-parametrically simulate the empirical joint distribution. Back-testing is used both to support the choice of a proper copula‑model and to compare the different optimization approaches efficiency. Based on the research results it is shown that copula‑models should be preferred as enabling to receive higher yield given the same level of foreign exchange risk