[Gretl-users] VECM and LM Test

[Allin Cottrell]

On Wed, 21 May 2008, Yuniarto Hadiwibowo wrote:

> I am starting to use Gretl for analyzing VECM.

Welcome to gretl!

> I use testuhat to check the normality of the residual and it 
> works. I tried lmtest, but it didn't work for VECM. How to check 
> the possibility of autocorrelation?

Not all of the lmtest options work for VECM, but you should be 
able to get equation-by-equation autocorrelation tests using

  lmtest --autocorr

However, having just tested this I see something is wrong with the 
output -- we'll fix that shortly.  In the meantime, you can get 
correct results using the graphical interface: estimate the VECM 
then select Autocorrelation under the Tests menu in the output 
window.

Allin Cottrell