impulse question

[luxu1983]

(1)impulse(model=eq,step=20,result=impulses,decomp=%decomp(cc))

(2)impulse(model=eq,step=20,result=impulses,factor=cc)

what is the difference between the (1) and (2)

[moderator]

You have two issues here.

1) The option name. DECOMP and FACTOR do the exact same thing. DECOMP is a name used in older versions, which we replaced with the newer name FACTOR. However, both options still work, and still do the same thing.

2) FACTOR (or DECOMP) is used to supply your own factorization of the covariance matrix. So, which of these:

impulse(model=eq,step=20,result=impulses,decomp=%decomp(cc))
impulse(model=eq,step=20,result=impulses,factor=cc)

makes sense depends on whether CC is a covariance matrix, or a factor of the covariance matrix that has been computed previously. If CC is a covariance matrix, this:

decomp=%decomp(cc)

computes a Cholesky factorization of that covariance matrix and uses it as the factor matrix, which is certainly sensible (although it's the same as the default behavior if you omit the option).

factor=cc

probably does not make sense in that context (i.e. if CC is a covariance matrix).

Of course, if CC is already a factor matrix, rather than a covariance matrix, the opposite is true.