irf for var with mgarch in mean

[luxu1983]

dear all
may you code the impulse-response function for a vector autoregression with multivariate GARCH-in-mean (John Elder 1995)?
Thank you

[TomDoan]

Again, that is not a well-defined "function". The effect of a time t shock depends upon the GARCH covariance matrix at t, producing a different "M" term for each subsequent entry.

[kayak]

Elder uses the unconditional variance for the size of the shock. There for the IRF as he defines should not be dependent on the point in time. The simulation then depends only the coefficient estimates.

[TomDoan]

The calculation of IRF's for a VAR-GARCH-M are covered in Elder-Serletis(2010).