A Regime Switching for Dynamic Conditional Correlation and GARCH: Application to Agricultural Commodity Prices and Market Risks

Abstract

© 2018, Springer International Publishing AG, part of Springer Nature. Time varying correlations are often estimated with dynamic conditional correlation, generalized autoregressive conditional heteroskedasticity (DCC-GARCH) models which are based on a linear structure in both GARCH and DCC parts. In this paper, a Markov regime-switching DCC-GARCH (MS-DCC-GARCH) model is proposed in order to capture the time variations and structural breaks in both GARCH and DCC processes. The parameter estimates are driven by first order Markov chain. We provide simulation study to examine the accuracy of the model and apply it for empirical analysis of the dynamic volatility correlations between commodity prices and market risks. The proposed model is clearly preferred in terms of likelihood, Akaike information criterion (AIC), and likelihood ratio test.