Short Run Restrictions in SVAR Framework Question

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

drizzy71
Posts: 4
Joined: Mon Feb 06, 2017 4:42 pm

Short Run Restrictions in SVAR Framework Question

Postby drizzy71 » Tue Mar 28, 2017 2:22 am

Hey everybody,

I had a quick question about the structural factorization of a VAR object in Eviews. I have recently implemented the SVAR analyzed in Blanchard and Perotti (2002) by using the regular A and B matrix short run restrictions in eviews. I believe I understand this process sufficiently well. In their paper, their A matrix contains 3 previously estimated parameter set and the rest as zero restrictions. Then in their B matrix they estimate one free parameter and the rest are either zero restrictions or the diagonal '1' values. These two matrices are inputted into the Structural factorization option of the Var object and voila you have your SVAR.

However, I have been delving into the literature a bit more and have seen a different way to present short run restrictions in one matrix where the authors present just one n by n matrix of contemporaneous relationships of the variables with the restrictions shown in the matrix. Some examples are: Kim and Roubini (2000) [http://www.sciencedirect.com/science/article/pii/S0304393200000106] and this other paper: [https://www.nbp.pl/publikacje/materialy_i_studia/211_en.pdf].

My question is thus pretty simple. How do you input the short run restrictions matrix in these last two papers into eviews through the Structural factorization function of the VAR object? Is the matrix in Kim and Roubini simply the A matrix and the implied B matrix is a n x n matrix with zeros everwhere except the diagonal? I know this question is probably pretty simple to answer but I'm not too clear on this so any input would be supremely helpful !

Thanks and cheers.

EViews Matt
EViews Developer
Posts: 560
Joined: Thu Apr 25, 2013 7:48 pm

Re: Short Run Restrictions in SVAR Framework Question

Postby EViews Matt » Tue Mar 28, 2017 1:19 pm

Hello,

Yes, you have the right idea. Theoretically, you can use just one of either A or B to entirely capture the structural relationships, with the other matrix being an identity matrix. Both the papers you reference appear to use A for this purpose, effectively restricting B to be an identity matrix.

drizzy71
Posts: 4
Joined: Mon Feb 06, 2017 4:42 pm

Re: Short Run Restrictions in SVAR Framework Question

Postby drizzy71 » Tue Mar 28, 2017 1:23 pm

Yes, awesome, I just wanted to double check since I'm just now starting to understand the whole linear algebra required in identifying the restriction matrix. Thank you very much!


Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 33 guests