estimate markove regime switching

[za gh]

hi
I want to estimate a model with markov regime switching but I have a problem . However as soon as I move from two to three regimes my coefficients do not have any Standard Error, nor Z Stat and Prob (all indicate NA).
I also use tvtp ;I am facing the same problem. with repeated estimates of the number of times (10 times) come to fruition, but it takes a long time. What is the reason for this problem? Software problem?
can you help me ??????
Thanks

[EViews Glenn]

The following discussion is relevant.
http://forums.eviews.com/viewtopic.php?f=4&t=12310
Basically, Markov switching models by their nature are difficult to estimate and taking a long time is something that is the norm rather than the exception. Moreover, these problems are magnified as you increate the number of regimes.

[ecogulesce]

hello
ı want to estimation number of regime for markov switching model(for interest rate)
could you tell me step by step how can ı decide ıt ?
ı looking forward to hearing from you
thanks