sm.tsa.MarkovAutoregression: Switching-Variance for Handling Heteroskedasticity

Yes, the switching-variance in sm.tsa.MarkovAutoregression accounts for heteroskedasticity. It's a model that can fit time series with different variances. The model can switch between different states, each with its own variance. This approach is effective for handling heteroskedasticity because it allows the model to fit different variances in different states.