Page 1 of 1

SVAR short and long-run restriction

Posted: Thu Jan 11, 2024 9:10 am
by ksa
Dear, All.

I am currently working on replicating Bjornland and Dag Henning Jacobsen 2013("House prices and stock prices: Different roles in the U.S. monetary transmission mechanism)(http://home.bi.no/a0310125/BH_SJE_2013.pdf) by running an SVAR model with both short-term and long-term constraints. Despite applying the constraints as depicted in the image, I am encountering some issues. If I don't check the 'normalize sign', I face a 'near singular matrix' error. Moreover, in its current state, the model is yielding under-identified results, and the impulse response is not observable. I would be immensely grateful for any guidance or suggestions on how to address these challenges for a more accurate estimation.

Re: SVAR short and long-run restriction

Posted: Thu Jan 11, 2024 11:17 am
by EViews Matt
Hello,

I believe the restrictions you've currently applied to the B matrix should instead be applied to the S matrix. The A and B matrices will then be unrestricted.