problem about svar_identification

[lyjessica]

Hi, guys, could you please guide me through my issue?
I have a problem with svar_identification.
In the example of DH10.mod, SVAR identification scheme is described :

svar_identification;
exclusion lag 0;
equation 1, KCFSI;
end;

but if I change the code into:

svar_identification;
exclusion lag 0;
equation 1, KCFSI;
equation 2, RGDP, KCFSI;
equation 3, RGDP;
end;

It also works. But I can't understand the expression of equation, for example, what's the difference among "RGDP", "KCFSI", and "RGDP, KCFSI"?

Thanks a lot in advance.

[lyjessica]

I seem to have the idea of the svar_identification.

svar_identification;
exclusion lag 0;
equation 1, KCFSI;
end;

The codes mean that there is no KCFCI(t) in equation 1 .

svar_identification;
exclusion lag 1;
equation 1, KCFSI;
equation 2, RGDP, KCFSI;
equation 3, RGDP;
end;

The codes mean that there are no KCFCI(t-1) in equation 1, no RGDP(t-1) and KCFSI(t-1) in equation 2, no RGDP(t-1) in equation 3.
Am I right?
But I have new question. Can our dynare deal with expection value in markov switching DSGE model, for example KXFCI(t+1) and RGDP(t+1) ?

[MichelJuillard]

Your interpretation of the exclusion instruction in SBVAR_IDENTIFICATION is correct.

It is not yet possible to solve Markov-Switching DSGE models in Dynare

Best

Michel