THE IMPORTANCE OF FAT-TAILED AND SKEWED DISTRIBUTIONS IN MODELING VALUE-AT-RISK

Abstract

Most of the Value-at-Risk models assume that financial returns are normally distributed, despite the fact that they are commonly known to be left skewed, fat-tailed and excess kurtosis. Forecasting Value-at-Risk with misspecified model leads to the underestimation or overestimation of the true Value-at-Risk. This study proposes new conditional models to forecast the daily Value-at-Risk by employing the new fat-tailed and skewed distributions to GARCH models. Empirical results show that the fat-tailed and skewed distributions provide superior fit to the conditional distribution of the logreturns among others. Backtesting methodology and loss functions are used to compare the out-of-sample performance of Value-at-Risk models. We conclude that the effects of skewness and fat-tails are more important than only the effect of the fat-tails on accuracy of Value-at-Risk forecasts.