Application of Markov switching models to modeling the nominal exchange rate of ruble to US dollar

This paper studies the dependence of the nominal exchange rate of ruble to dollar on the oil price using an error correction model with Markov switching and consumer price and producer price indices. It is shown that two different modes can be distinguished for the nominal exchange rate during 2003-2019, viz. with a high and a low rate of reversion to the long term equilibrium in response to oil shocks. The optimal Markov model is identified on the basis of information criteria. The hypothesis of an equal trend for the two modes is not rejected. At this moment, the first regime prevails, in which the dependence of the exchange rate on oil prices is lower, and the time of convergence to the long-term equilibrium is longer. The behavior of the model in 2020 is considered, and it is revealed that despite a significant drop in the oil price in 2020 (March-April), the exchange rate is likely to be in the first mode with a low rate of reversion, in contrast to the situation in 2014.