Events enhance crude oil price risk measurement: embedding breakpoints in GARCH-VaR

The international oil market is complex, and the oil price fluctuation is always non-linear and chaotic. As the frequency of uncertain events rises and the interconnectedness among markets strengthens, structural breakpoints on behalf of uncertain events become increasingly prevalent, and their impact on structural changes in time series continues to grow. To accurately capture the market volatility, structural breakpoints are becoming important factors that need to be taken into consideration. This study developed a comprehensive framework for time series volatility and risk assessment incorporated with breakpoints. We presented an objective approach for identifying the number and location of structural breakpoints based on Wasserstein distance. Then extended the GARCH-VaR models by introducing the breakpoints as dummy variables to analyze the volatility of the conditional variance. Compared with the conventional models, the breakpoints-based GARCH-VaR model performs better, which can reduce the failure rate and improves the accuracy of the risk measure. The WTI and Brent oil price from January 2005 to January 2024 were used in the empirical study. During the period, major incidents such as economic crises, geopolitical conflicts, and large-scale public health events occurred. The results showed that there are 12 breakpoints in WTI oil price time series and 17 in Brent oil price time series which are related to the major historical events that have impacted the oil market. This demonstrates the necessity of considering breakpoints and the importance of detecting breakpoints accurately. We conducted a powerful test in the empirical research, through comparative analysis of the risk measurement failure rates in different GARCH-VaR models before and after incorporating breakpoints, the effectiveness and robustness of the extended model are validated.