FMOLS Wald Test

[paldejong]

Dear RATS Users,

First and foremost, Happy New Year to all of you!

For my latest revision of research, I am using the FMOLS procedure for cointegrated panels by Pedroni (Pedroni, Peter (2000) "Fully Modified OLS for
Heterogeneous Cointegrated Panels," Advances in Econometrics, Vol. 15, 93-130) to test for cointegration before and after a significant economic event. A reviewer suggested that I should not just rely on the point parameters (one before and after the event) but should report a statistic of potential difference between the parameters based on the Wald test. I am not quite sure on how to implement this. I could use your help.

Thank you,

Pieter de Jong

[TomDoan]

I assume from the description that you want to test for a difference in the cointegrating vector before and after. If you do a sample split, the estimators before and after should be uncorrelated, so the variance of the difference would be the sum of the variances. This would compute the test statistic which would be asymptotically chi-squared with degrees of freedom m-1, where m is the number of endogenous variables.

Code: Select all

@PANELFM(SMPL=%PERIOD(T)<=split,other options)
# list of endogenous variables
compute xx1=%xx,beta1=%beta
@PANELFM(SMPL=%PERIOD(T)>split,other options)
# list of endogenous variables
compute xx2=%xx,beta2=%beta
compute wald=%qforminv(xx1+xx2,beta1-beta2)

[atarca]

Hi Tom,
How can I compute this Wald Test statistic for each of the panels by using the %%ibetas and %%ixx? I also need to compare the individual betas of the two samples for each of the cross-section unit. I would be grateful if you could help.
Best Regards,
Atarca

[TomDoan]

Hi Tom,
How can I compute this Wald Test statistic for each of the panels by using the %%ibetas and %%ixx? I also need to compare the individual betas of the two samples for each of the cross-section unit. I would be grateful if you could help.
Best Regards,
Atarca

That's included in the Pedroni JAE 2007 replication