The RATS Software Forum

Hi,

I would like to know what is the difference with the Fully Modified OLS procedures on the estima web site.
Actually, there is three based on Hansen (1992):

1 -The Carstenson(2006) fm_ols.src
2 - fm.src
3 - fmols.src=>Outdated

Basically, I run the first two programs with the same dataset and I get different estimators.

Is there some difference between the two programs that I missed?
Thank you for helping me.

Nadia Gagnon

Nadiarock1 wrote:Hi,

I would like to know what is the difference with the Fully Modified OLS procedures on the estima web site.
Actually, there is three based on Hansen (1992):

1 -The Carstenson(2006) fm_ols.src
2 - fm.src
3 - fmols.src=>Outdated

Basically, I run the first two programs with the same dataset and I get different estimators.

Is there some difference between the two programs that I missed?
Thank you for helping me.

Nadia Gagnon

Fully modified LS depends upon a calculation of a one-sided long-run variance matrix. Different methods of computing that will lead to different results (one hopes not that different). The fm.src procedure does a Bartlett (Newey-West) window with an input number of lags. Carstensen's fm_ols does a VAR prewhitening/recoloring, combined with a Bartlett window on the pre-whitened residuals.