IMPULSE() instruction code

Questions and discussions on Vector Autoregressions

IMPULSE() instruction code

Postby asmith05 » Sun Mar 27, 2011 11:14 am

Hello,

Does anyone have the underlying code for the IMPULSE() instruction? I am wanting to compute regime dependent impulse response functions from a MSVAR. Therefore, I can't use the IMPULSE() instruction since I must reference a MODEL=varmodel. I want to tell it which coefficient matrix and which covariance matrix to use when computing the impulse responses. Basically in the end I want 2 sets of impulse response functions, one set for each regime. I think I can accomplish this if I can tell RATS to compute 2 impblk matrices, one for each regime.

Thanks for any help!

Lee
asmith05
 
Posts: 5
Joined: Thu Mar 24, 2011 8:33 pm

Re: IMPULSE() instruction code

Postby TomDoan » Mon Mar 28, 2011 6:02 pm

The inputs to IMPULSE are the model, the coefficients and the covariance matrix or the factor matrix. The way that you would handle this is to use %MODELSETCOEFFS to reset the coefficients of the model. Note that they need to match up with the arrangement of the coefficients in the model, which isn't always the case for MSVAR's (where organizing regressors by lag tends to make the most sense, which the RATS VAR models are set up with the regressors organized by variable). The procedures VARCALC and VARLAGSELECT show how to rearrange the coefficients for use with %MODELSETCOEFFS.
TomDoan
 
Posts: 2720
Joined: Wed Nov 01, 2006 5:36 pm


Return to VARs (Vector Autoregression Models)

Who is online

Users browsing this forum: No registered users and 1 guest

cron