EViews.com

Hi Everyone,

I have a single Time series and want to test wehter the mean of this series is equal to zero. Since the observations of my series are overlapping until a certain lag (serially correlated) I want to use the Neway West corrected standard error for the test. How can I compute the corrected error for a sample mean?

ls y c
then choose Newey-West

Sounds easy. Thanks for the Reply.

One more question about Neway-West:

Is it possible to compute Neway_West correction for fixed lags (e.g. 5)? I think Eviews always chooses the lags automaticly.

Hit the HAC options button.

Well I only find the HAC-Options under GMM_Estimation where I have to put something in the instrument list. What do I have to put in there if I wanted to do the exact same thing like the Neway-West-correction in the LS scenario (I still want to estimate the mean of a single series). If I do Neway-West-Correction in the LS-Scenario it automatically sets the lag. In other words: I need simple LS on a constant with Neway-West but with fixed lag and I do not know how to operate with the GMM-Estimation especially with the instrument list?

What version of EViews are you using?

Version 5. You think that there are only hac options in ls estimation in version 6 or 7?

The option is there in version 7. I have no idea if it's there in 5.

It is not there in EV5. You could put an explicit equation in the system and estimate using the system estimator in EV5.