Page 2 of 2
Re: three state variance regime switching model
Posted: Mon Nov 05, 2012 9:42 am
by TomDoan
I can't help you fix the problem if I don't understand what the problem is. You wanted to know when the process is in the high variance state. As I said, the Markov process only tells you the probabilities, and if you want a yes or no answer, the best guess is to select the data points where the probability of the high state is >.5. If that's not what you want, then you will have to figure out what it is that you do want.
Re: three state variance regime switching model
Posted: Mon Nov 05, 2012 10:37 am
by superper2008
Hi Tom,
Sorry not to make my question clear.. I want to select the data points where the probability of the high state is >.5...Could you help me fix the problem please?
Thanks
Regards
Re: three state variance regime switching model
Posted: Mon Nov 05, 2012 7:18 pm
by TomDoan
As written, the HIGHVAR series has 1's in the entries where the probability of the high-variance state is >.5. What else did you want?
Re: three state variance regime switching model
Posted: Fri Nov 30, 2012 7:55 pm
by superper2008
Hi Tom,
Could I know the probability of variance regimes in each date point? I read a working paper written by Dr. James Hamilton. He also posted his RATs procedures, one of which shows "full sample smoother probabilities". So I am wondering whether I could export the probabilities too. Thanks
Regards
Re: three state variance regime switching model
Posted: Sat Dec 01, 2012 8:22 am
by TomDoan
Your standardized returns aren't done right; you want:
set stdu = return/sqrt(variance)
graph(footer="Figure 4.11a Plot of standardized stock returns")
# stdu
HIGHVAR is a 1-0 dummy for the entries where the probability of the high variance state is > .5. If that's not what you want, what do you want?