Thank you for the reply. I think you understood me correctly, but just to elaborate, 8 regimes with each looking like (high constant, high AR(-1), high variance) and (H,H,L), (H,L,H) etc.
I would envision that the 8 regimes would be restricted to pdfs which are function of 6 coefficients and the likelihood is maximized w.r.t. the transition parameters and the 6 coefficients (high constant, low constant, high AR(-1) etc.)
I would envision that the 8 regimes would be restricted to pdfs which are function of 6 coefficients and the likelihood is maximized w.r.t. the transition parameters and the 6 coefficients (high constant, low constant, high AR(-1) etc.)