Hsiao's version of granger causality test

[Liv Pan]

Dear all,

I'm a new user of Statalist so probably I am creating this post in the wrong way/section. I am sorry for this in advance.
I'm struggling on the application of the Hsiao's methodology to test the causality between two variables. In particular, I don't find the way of obtaining the FPE from a VEC model. While it is displayed when using a VAR model (for not cointegrated variables), I cannot get it from the VEC model I am instead using when the variables are cointegrated. I know that proably the solution is in the use of the command VARSOC, but still I don't know how properly use it as poestimation command after the VEC model.
I am sorry if the question may appear confused, but any comment would be helpful.
Many thanks in advance,

Liv

[Carlo Lazzaro]

Liv.
welcome to the list.
Even though I cannot aswer your question, I would recommend to re-post in on the General forum, as this one is more practicing purposes only.
I would also take a look at the FAQ on how to post (more) effectively and give full reference of articles and textbooks you quote, being this one a multidisciplinary forum (by the way, I notice that you are referring to a famous textbook on panel data econometrics, but probably many on this list legally ignore that source altogether ).

[Liv Pan]

Thanks Carlo.
I re-posted it, hopefully, in the right section.
The Hsiao's (1981) procedure aims to determine the optimal lag length to then apply in the causality test. I honestly don't think that it is difficult to implement in Stata, but I'm defenitely missing something fundamental!
Many thanks for your kind advice,

Best regards,

Liv