Prediksi Volatilitas Saham KINO dan MRAT menggunakan Model BEKK-MGARCH

This study analyzes the volatility prediction of KINO and MRAT stocks using the BEKK-MGARCH model during January 2019 to December 2024. Both stocks exhibit high price fluctuations, with volatility significantly influenced. Persistence effects are more dominant, with asymmetric spillover where KINO's influence on MRAT is stronger. Conditional correlation shows a shift from positive to negative in the 30-day forecast. Model evaluation demonstrates low RMSE values of 2.05×10⁻⁵ (KINO), 3.38×10⁻⁵ (MRAT), and 4.42×10⁻⁵ (covariance), indicating excellent predictive performance and confirming the reliability of the BEKK-MGARCH model with exponential smoothing in forecasting the volatility dynamics and relationship between these two stocks with high precision. However, the Jarque-Bera test rejects residual normality (p < 2.2×10⁻¹⁶), and the Ljung-Box test detects autocorrelation, suggesting the need for more complex models such as Student-t distribution or asymmetric models. These findings provide important insights for investors in managing risk and portfolio diversification strategies in the cosmetics sector.