Mastering Python for Finance - Second Edition

The Markov regime-switching model

To model nonlinear behavior in economic and financial time series, Markov switching models can be used to characterize time series in different states of the world or regimes. Examples of such states could be a volatile state, as seen in the 2008 global economic downturn, or the growth state of a steadily recovering economy. The ability to transition between these structures lets the model capture complex dynamic patterns.

The Markov property of stock prices implies that only the present values are relevant for predicting the future. Past stock-price movements are irrelevant to the way the present has emerged.

Let's take an example of a Markov regime-switching model with m=2 regimes:

ϵ_t is an independent ...