Estimating extreme value at risk of the stock exchange Thailand index with Markov regime switching

This study proposes a novel framework for estimating Value at Risk (VaR) by integrating regime-switching volatility modeling with Extreme value theory (EVT) to enhance the accuracy of tail risk assessment.Using daily returns of the Stock Exchange of Thailand (SET) Index from 2011 to 2024, volatility is modeled through the Markov regime-switching GARCH (MRS-GARCH), GARCH-type, and HAR-RV models, while tail behavior is captured via the Peaks over threshold (POT) method with the Generalized Pareto distribution (GPD).The results show that the MRS-GARCH-EVT model consistently outperforms both GARCH-type-EVT and HAR-RV-EVT models.Its regime-switching structure effectively captures transitions between low-and high-volatility regimes, reflects volatility clustering, and enhances sensitivity to extreme tail events.Backtesting confirms its superior stability and reliability, while most GARCH-type models fail to achieve comparable performance.Overall, the MRS-GARCH-EVT framework demonstrates strong potential as a robust and practical approach for tail risk management in emerging equity markets.