Emmanuel K. Oseifuah, Carl H. Korkpoe
We studied regime-switching behaviour of the volatility of the returns from the ZAR/USD exchange rate for the period January 4, 2002 to December 31, 2017. The results showed that, contrary to mainstream approaches for estimating volatility using GARCH (1,1) there are clear regimes in the returns which necessitate regime switching models. The results further revealed that the Markov regime switching model GARCH (1,1) with skewed student-t innovations is superior in capturing the heteroscedasticity of the returns. The deviance information criteria were used as a selection metric from among six candidate models.