2013/3/18 Riccardo (Jack) Lucchetti <r.lucchetti@univpm.it>
On Mon, 18 Mar 2013, Gabriela Nodari wrote:

Thank you for the answer. I gave  a look to the code but I fear I need some
time to understand it completely.

As for the Svar I know that the user can choose the number of replications.
Instead, I would like to know what does gretl do when I ask to calculate
and plot Irfs from the model window without using the Svar pack!

Thank you!

Both the SVAR module and the C version use the simple bootstrap for computing the confidence intervals for IRFs, as per Runkle (1987) "Vector Autoregression and Reality", JBES.

I'll probably make the "bootstrap-after-bootstrap" method (Kilian 1998)
avalaible in the SVAR module soon, as an extra option.
This would be a nice feature. So far I am only aware of JMulti (Hall and Efron percentiles ) and RATS which offer different procedures. I think Eview has implemented only a standard MC-procedure.

  Riccardo (Jack) Lucchetti
  Dipartimento di Scienze Economiche e Sociali (DiSES)

  Universitą Politecnica delle Marche
  (formerly known as Universitą di Ancona)

Gretl-users mailing list