Markov switching autoregressive model for WTI crude oil price

In this study, we aimed to test the nonlinear structure of crude oil prices with Markov Regime Switching Autoregressive Models. In the study of weekly prices covering the period from May 06, 1990 to April 11, 2018, a two-regime Markov switching model was applied. In the case of two regimes, we proved the that the probability the process will be in regime 1 or 2 is given by steady-state probabilities. As a result, it can be seen that the predictions made by the Markov switching autoregressive model were succesful.