Regime Switching - Sample

[papel]

Hi,
I estimate a regime switching model for the period from 1990M11 to 2012M11. Then re-estimate by using a different sample period (by changing the sample period from the equation object) but I get the results for the full period (1990 M11 2012M11).
I attach the file I work.

Regards,
Timotheos.

PS: For the beta version of Eviews 8 you have created a forum that is not available now. I had made a question how can we make constraint optimization (maximize the Sharpe ratio by setting the sum of the weights to be equal to 1) and you upload a program that can do it. Is it possible to send the program or post it on the forum?

[EViews Glenn]

You've found a bug. Our apologies.

I've checked the fix into the codebase and it will appear in the next update. For now, if you change the workfile sample, then enter the name of the equation following by the switchreg method *without* a spec in the command line, it will reestimate using the existing specification and the new sample.

Code: Select all

smpl 1960 1984 msar4.switchreg

reestimates the specification in MSAR4 using the sample "1960 1984".

[erikgill]

has a solution to this been addressed?

I have a workfile I am running multiple regressions on but need different sample sizes. Sometimes it works and other times it reverts to the whole sample of 1992-2012. I'd like to run all equations in one program but right now it is not consistently letting me set the sample within the program.

[EViews Gareth]

It should be in the latest patch

[EViews Glenn]

Sorry for not following up to the forum. The fix was posted in August.