Questions on the variance decomposition

[arnedam]

Two quick questions on the variance decomposition module in Dynare:

1. Is it possible to get results over various horizons? (if not directly, is there a 'doable' solution we can code ourselves?)

2. How do we specify which endogenous variables we are interested in? For now we get the 'varobs' subset only.

Thanks,
Arne

[StephaneAdjemian]

Two quick answers :

1. It is possible to get conditional variance decomposition but you have to code it yourself with matlab from the state space representation of the dsge model given by dynare.

2. I am not sure but you should try to declare the endogenous variables after the estimation command, writting something like:

estimation(mode_compute=4, , , ) var1 var2 var3;


Best, St