In this study, it is investigated the impact of suddenly structural breaks on estimated GARCH-type models with normal and heavy-tailed distributions for daily oil futures market returns. More specifically, the multiple structural breaks in return variance over the whole sample period are detected by the Inclán-Tiao’s algorithm. The estimated results of the ICSS AR-GARCH models show that the volatility persistence decreases dramatically after controlling for such discrete breakpoints. The changing oil futures risk can be best captured by the ICSS AR-EGARCH-GED model. Specifically, the comparison of the in-sample model evaluation champions the AR-EGARCH-t model over competing models within each identified sub-period.
Relation:
Investment Management and Financial Innovations, Vol.12, No.2, 16-25