One-day-ahead value-at-risk estimations with dual long-memory models: evidence from the Tunisian stock market

Abstract

In this paper, we assess the one-day-ahead Value-at-Risk (VaR) performance for the Tunisian Stock Market (TSE). Using the ARFIMA-FIGARCH and ARFIMA-FIAPARCH models under three alternative innovation distributions: normal, Student and skewed Student, we show that the ARFIMA-FIAPARCH with skewed Student innovations outperforms the other models since it jointly considers the asymmetry, long-range memory and fat-tails in the TSE return behaviour. This model provides the better results for in and out-of-sample VaR estimations for both short and long trading positions.