by macro_man » Thu Aug 09, 2012 9:30 pm
After thinking this through, here is what I should have said in the previous post. I want to estimate a SVECM using BQ long-run restrictions using the short and long-run restrictions code you provide. It seems, then, I that I use CATS to get the cointegrated vector(s) and then bring them into RATS and apply the SR&LR restrictions. However, one can also apply LR identifying restrictions on the cointegrating vector(s) in CATS too.
Question:
(1) If all I am interested in doing is applying the BQ LR restrictions to the VECM, do I worry about applying LR indentifying restrictions to the cointegrating vectors in CATS before pulling the model into RATS?