Modeling and forecasting extreme commodity prices: A Markov-Switching based extreme value model

- A regimen switching model with a multifractal structure for extreme events is proposed.
- The approach captures the key stylized facts at extreme levels successfully.
- Capture the presence of short- and long-term memory patterns of extreme events
- Provide the most accurate VaR estimates for three main categories of commodities

We propose a Markov-Switching Multifractal Peaks-Over-Threshold (MSM-POT) model to capture the dynamic behavior of the random occurrences of extreme events exceeding a high threshold in time series of returns. This approach allows introducing changes of regimes in the conditional mean function of the inter-exceedance times (i.e., the time between two consecutive extreme events) in order to admit the presence of short- and long-term memory patterns. Further, through its multifractal structure, the MSM-POT approach is able to capture the typical stylized facts of extreme events observed in financial time series, such as temporal clustering of the size of exceedances and temporal behavior of tail thickness.We compare the performance of the MSM-POT model with competing self-exciting models and a GARCH-EVT approach in an in- and out-of-sample VaR forecasting exercise based on the extreme returns of six daily commodity futures prices (i.e. Brent and WTI crude oil, cocoa, cotton, copper, and gold). Empirical results suggest that the VaR estimates generated by the MSM-POT model for the returns analyzed produce the most accurate forecasts.