Apparently ortho = TRUE/FALSE
serves to control a possible autocorrelation between variables (or their errors), making the marginal effect "cleaner". This is accomplished internally by adjusting impulses by a matrix that results from the cholesky decomposition of the covariance matrix of errors.
Already cumulative = TRUE/FALSE
allows the marginal effects to be accumulated or not (a classic cumsum()
).
As for the impulse, it is apparently predefined, and in that function it would not have to be changed. You can see, in the function Phi
, which is defined as diag(K)
, where K
is the number of variables.
Further information in (English): https://stats.stackexchange.com/a/143296/77852
You can view function codes using:
methods(irf)
getS3method("irf","varest")
vars:::.irf
vars:::Psi
methods(Psi)
getS3method("Psi","varest")
getS3method("Phi","varest")