Mirko, Since you're interested in the effects of the structural shocks on your variables of interest, you should examine the structural impulse responses (ie, from SVAR). The IRFs from the 'standard VAR' are the effects of shocks that in general will be linear combinations of the structural shocks you're interested in, so they won't tell you what you want to know. Does that help? PS -----Original Message----- From: gretl-users-bounces(a)lists.wfu.edu [mailto:gretl-users-bounces@lists.wfu.edu] On Behalf Of mastro2387(a)tiscali.it Sent: Friday, June 08, 2012 9:50 AM To: gretl-users(a)lists.wfu.edu Subject: [Gretl-users] difference between SVAR and VAR IRFs Dear All, I have a question, which probably you will find stupid.. I have to study the structural shocks to output for changes in government spending and taxes in the Normandy Invasion of WWII (period identified by dummy variable) and computing the consequent multipliers. For multipliers it is not a problem since they are linked with the VAR coefficients but for IRF...which ones do I have to check? Better saying,..the option for IRFs in the standard Vector Autoregression is anyway based on the Cholesky decomposition but IRF obtained in the SVAR are different: so which one do I have to choose and why? Thanks for attention and help!! Mirko Invita i tuoi amici e Tiscali ti premia! Il consiglio di un amico vale più di uno spot in TV. Per ogni nuovo abbonato 30 € di premio per te e per lui! Un amico al mese e parli e navighi sempre gratis: http://freelosophy.tiscali.it/ _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users

On Sat, 9 Jun 2012, Muheed Jamaldeen wrote: You may also want to take a look at the pdf file documenting gretl's implementation of SVAR: go to the "Model" menu, choose "Time Series" and then "Structural VAR". Click on the "Help" button and the pdf file should open. It's mainly oriented to help the user set up his model in gretl, but it also contains a few theory bits and some literature pointers you may find useful. -------------------------------------------------- Riccardo (Jack) Lucchetti Dipartimento di Economia Università Politecnica delle Marche (formerly known as Università di Ancona) r.lucchetti(a)univpm.it http://www2.econ.univpm.it/servizi/hpp/lucchetti --------------------------------------------------

RJF, The Cholesky decomposition approach to identifying structural shocks is usually justified on the basis of information lags or some other timing issue. For example, in a system with inflation and the federal funds rate, inflation could be ordered first if one thinks (as most do) that changes in the funds rate can't affect inflation in the same period. Your point is exactly right though; such a timing (recursive) arrangement may be hard to justify, especially if there are more than a few variables. That's why people look for alternatives such as those outlined in Jack's SVAR package. PS -----Original Message----- From: gretl-users-bounces(a)lists.wfu.edu [mailto:gretl-users-bounces@lists.wfu.edu] On Behalf Of Dr RJF Hudson Sent: Saturday, June 09, 2012 5:50 AM To: r.lucchetti(a)univpm.it; Gretl list Subject: Re: [Gretl-users] difference between SVAR and VAR IRFs Hello, Dr Lucchetti I'm hoping for a leg up on your Cholesky comment please, because I'm a bit confused..... Let's think of the expression X^T * A* X > 0 where A= LL^T ( A is symetric and positive definite) and X is real and non-zero ie all is as commonly known (eg, ' T ' = transpose, etc) The question is how do we find L One way is to use Cholesky's procedure. Now this is merely a mathematical procedure. Obviously if a different matrix A is formed by re-arranging the variables then a different solution lower triangular matrix L will be obtained. My query is simple...what does a mathematical procedure ( Cholesky) know about cultural shocks ? Am I missing something ? RJF Hudson ----- Original Message ----- From: "Riccardo (Jack) Lucchetti" <r.lucchetti(a)univpm.it&gt; To: "Gretl list" <gretl-users(a)lists.wfu.edu&gt; Sent: Saturday, June 09, 2012 6:28 PM Subject: Re: [Gretl-users] difference between SVAR and VAR IRFs On Sat, 9 Jun 2012, Muheed Jamaldeen wrote: You may also want to take a look at the pdf file documenting gretl's implementation of SVAR: go to the "Model" menu, choose "Time Series" and then "Structural VAR". Click on the "Help" button and the pdf file should open. It's mainly oriented to help the user set up his model in gretl, but it also contains a few theory bits and some literature pointers you may find useful. -------------------------------------------------- Riccardo (Jack) Lucchetti Dipartimento di Economia Università Politecnica delle Marche (formerly known as Università di Ancona) r.lucchetti(a)univpm.it http://www2.econ.univpm.it/servizi/hpp/lucchetti -------------------------------------------------- -------------------------------------------------------------------------------- _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users