Capturing fat-tail risk in exchange rate returns using SU curves: a comparison with the normal mixture and skewed Student distributions

ABSTRACT Evidence that the distribution of daily changes of exchange rates has fatter tails than the normal has led researchers to consider different alternative distributions to estimate quantile-based risk measures. In this paper we investigate the ability of SU curves to capture the fat-tail risk of exchange rate returns, comparing them against the mixture of normals and skewed Student models. The results indicate that although the skewed Student provides the best overall fit, in terms of VaR performance the SU curves can yield better or at least similar results to any of the models considered. This, together with their flexibility in modeling different degrees of nonnormality, suggests that the SU system can be a useful alternative for riskmanagement.