SVAR Long-run & Short run restrictions

[Vladimir Emelianov]

Hi,

So i'm currently doing my dissertation where i've decided to follow the methodology of this paper:https://www.snb.ch/n/mmr/reference/w..._2013_05.n.pdf. They essentially use US data to build a SVAR which as four shocks: two short and two long-run shocks. I'm currently trying to follow this methodology but apply the analysis to Norway. However, i've run into a few problems. I haven't been able to estimate both short-run and long-run shocks in either stata or Eviews. Is there some way of doing this in Stata? If not, what would my alternatives be? I'm not entirely sure I understand the paper's restrictions on both the short and long-run matrices of the SVAR. I've tried to estimate the short run shocks by themselves but my impulse response functions look really wrong. The paper estimates a few of the SVAR equations using 2SLS and I honestly have no idea how to implement this in a SVAR framework through software packages such as Stata and eviews. Any guidance with respect to this will be greatly appreciated. I've been stuck on this issue for a few weeks and my supervisor is unavailable. Please let me know if any more information is required.

[Eric de Souza]

Go to http://blog.stata.com/
and search for svar (use the search box in the top right hand corner). You should find two detailed explanations, one for svar and one for long run restrictions

[Vladimir Emelianov]

Neither one of the articles answers my questions, though I assume this means that both LR and SR shocks cannot be estimated simultaneously in stata?

[Eric de Souza]

No, they cannot. As page 760 of the manual states "These options specify the long-run constraints in an SVAR. To specify a long-run SVAR model, you must specify at least one of these options. The list of options specifies constraints on the parameters of the long-run C matrix (see Long-run SVAR models for the definition of C). None of these options may be specified with any of the options that define short-run constraints.

[Vladimir Emelianov]

Would you happen to know of any software that is able to do this?

[Eric de Souza]

An add-in has been written for EViews which is called svarpatterns. I have no idea what it's worth. You will find it by Googling.
Imposing both short- and long-run restrictions is a tricky business both statistically and numerically.
I haven't looked at RATS to see what it does.

[Vladimir Emelianov]

Great, i'll take a look at those. Thanks for your help!