Modeling Volatility Regimes of the BRVM10 Index: Hidden Markov and Markov-Switching GARCH Approaches

This paper investigates the nonlinear dynamics of the BRVM10 index, the leading stock market benchmark in West Africa, by combining Hidden Markov Models (HMM) and Markov Switching GARCH (MS-GARCH) frameworks. The analysis exploits daily data from March 2014 to May 2024 and is motivated by stylized facts typical of emerging markets: volatility clustering, structural breaks, heavy tails, and regime shifts. Preliminary diagnostics confirm non normal returns, strong ARCH effects, and multiple structural breaks, justifying regime switching volatility models. Three state HMMs under Gaussian, Student-t and Skew-t distributions reveal a consistent volatility hierarchy. The Gaussian HMM provides a transparent baseline with persistent calm and moderate regimes. The Student-t HMM improves tail risk modeling by capturing extreme shocks that quickly revert to stability. The Skew-t HMM achieves the best fit, reallocating probability mass toward the mid volatility regime while preserving a persistent crisis regime. The MS-GARCH(1,1) models, estimated with Student-t and Skew- t innovations, jointly capture regime dependence and volatility persistence. Both specifications show that calm and mid volatility regimes dominate (80 􀀀 90% of stationary mass), while high volatility states are rare but highly persistent once entered. Bootstrap confidence intervals confirm robustness yet reveal wide uncertainty around crisis transitions, reflecting their rarity and unpredictability. Economically, the results imply that investors benefit from predominantly stable conditions but must hedge against sudden turbulence. For regulators, persistent calm regimes highlight resilience, while the recurrence of fat tailed shocks calls for early warning systems and stronger macroprudential buffers. Overall, the Student-t HMM is best for crisis detection, while MS-GARCH excels in volatility persistence and regime durations. Their combined use provides a coherent toolkit for risk management and policy design in West African markets. AMS classifications : 62M05, 62F10, 62H20, 62M10, 62P20, 91B84, 62G07