Parameter restrictions in Switching Regressions

[Thomasms90]

Is there any way to restrict coefficients in different regimes? I've only been able to restrict transition probabilities as it is.

In my case I'm estimating a simple 3 regime markov process with only the dependant variable and a constant (switchreg y c - i.e. no exogenous regressors or AR terms). I would like to identify an extreme regime with positive returns and high volatility (for instance restrict regime 3 to be positive and have higher variance than regime 1 and 2), but since I'm running a program with many loops, the regime may switch places. And it's very cumbersome to go through them one by one.

[EViews Glenn]

Unfortunately not.

It's an interesting idea. I'd have to think about how one might implement the restriction, both in terms of interface, and internally.

[Thomasms90]

Alright. Was just in case there was a way to do it.

I'll manage without, no problem :-)

[FelipeP]

I'm trying to do the same.
How did you work around this issue, Thomasms90?

[Thomasms90]

I wrote a bunch of if's to compare the regimes and correctly classify them in a table, then used that table for lookups.

[FelipeP]

Thank you for such a prompt reply and your idea. I will do like you and use an IF.
I will use the IF to compare the coefficients, since I know a certain coefficient has to be higher in one regime than the other at all times.