Value-at-Risk and Expected Shortfall Estimation for the Moroccan Stock Market: A Comparative EVT Approach with GARCH Filtering

Abstract

Portfolio management and regulatory requirements rely on the ability to accurately measure extreme market risk, and such events are more evident in emerging markets. In this paper, the Value-at-Risk (VaR) and Expected Shortfall (ES) for the Moroccan All Shares Index (MASI) are estimated between 2006 and 2026 using four estimation methodologies: Historical simulation, Normal model, Block-Maxima (BM), and Peaks-over-Threshold (POT) with GARCH filtering. Backtesting shows that the Normal model underestimates tail risk systematically, and the BM method provides the most conservative tail risk estimates (VaR  = -6.07% at 99.9%, ES  = -11.45% at 99.9% ), making it appropriate for regulatory stress-testing. The only method not rejected by the Kupiec test at all confidence levels is POT-GARCH, which yields more moderate estimates (VaR  = -4.61% at 99.9%, ES  = -5.61% at 99.9%), making it the most appropriate for daily risk monitoring. These findings are useful for extreme risk management in emerging African markets and contribute to the body of literature on EVT applications in this context.