Bootstrapped IRF confidence levels are used both because of problems in the use of analytic expressions when disturbances are normal and in the presence of non-normal distributions. To obtain a better understanding you should consult an appropriate text. There is material in Lutkepohl and Kratzig (2004), Applied Time Series Econometrics, Cambridge. or Neusser (2016), Time Series Econometrics, Springer..

John C Frain
3 Aranleigh Park

On 18 August 2017 at 18:41, lasses skola <lassesskola@gmail.com> wrote:
So what about my last question? Are IRF confidence intervals based on bootstrapping valid even if the VECM has non-normal residuals?


fredag 18 augusti 2017 skrev Allin Cottrell <cottrell@wfu.edu>:
On Fri, 18 Aug 2017, Riccardo (Jack) Lucchetti wrote:

On Fri, 18 Aug 2017, Sven Schreiber wrote:

- I just tried your (Allin's) example setup with UNEMP PRIME in the SVAR addon, and actually I get a very similar immediate effect of UNEMP->UNEMP. So I think I was wrong when I inferred from the SVAR doc that the shock has unit size. (There is a slight difference, but I think that's just the usual issue of d-o-f corrections or not in the sigma estimate.) Perhaps the docs for the SVAR add-on need to be clarified (note to self); or Jack could weigh in on this.

I'll try to contribute to the discussion once I'm back from holiday (end of the month). :-)

Enjoy your holiday! (Mine's about to expire.)

Allin
_______________________________________________
Gretl-users mailing list
Gretl-users@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users

_______________________________________________
Gretl-users mailing list
Gretl-users@lists.wfu.edu
http://lists.wfu.edu/mailman/listinfo/gretl-users