MODELING DYNAMIC BEHAVIOR OF CONDITIONAL VARIANCE IN CRUDE OIL PRICES: THE BEKK AND CCC-GARCH FRAMEWORK

Abstract

This study investigates the dynamic behavior of conditional variance and covariance among benchmark crude oil prices; Brent, Dubai  and West Texas Intermediate using multivariate GARCH frameworks of Diagonal BEKK-GARCH and Constant Conditional Correlation (CCC-GARCH) models. The data used in this study were obtained from US Energy Information Administration (EIA). The data set consists of weekly prices of three major oil benchmarks: Brent crude oil, Dubai crude oil, and West Texas Intermediate. The Diagonal BEKK-GARCH results reveal statistically significant short-term responses to shocks and strong persistence in volatility, highlighting the clustering effect of typical crude oil markets.  Positive definiteness in the BEKK model’s covariance matrix confirms model stability and appropriate specification. The CCC-GARCH model identifies sustained volatility with near-unit root behavior and confirms strong static correlation between Brent and WTI, while showing moderate interdependence with Dubai/Oman. However, the CCC-GARCH model offers deeper insights by capturing the time-varying nature of variances, if the correlation between series remains constant over time.  The findings reveal major implications for financial risk management, energy market forecasting, and macroeconomic policy. The study emphasizes the need for investors to adopt adaptive hedging techniques and recommends the use of dynamic correlation models for better prediction of volatility and asset co-movements. Policymakers in oil-exporting nations can leverage these insights to inform the design of stabilization mechanisms, sovereign wealth strategies, and counter-cyclical fiscal interventions. The study provides a robust empirical foundation for financial and policy decisions in volatile global energy markets.