Pooled Mean Group - PMG (PSS, 1999) vs. CS-ARDL/CS-DL (CMPR, 2015)

[maragloria]

Hi there,

I would like to clarify a doubt I have concerning the PMG estimator in a panel setting. I'm using Eviews 9.5.

In the EViews 9 User’s Guide II (p. 838), it is stated that PMG (43.1 & 43.2) is appropriate in cases where there's correlation between the mean-differenced regressors and the error term.

However, in a recent paper by Alexander Chudik, Kamiar Mohaddes, M. Hashem Pesaran, and Mehdi Raissi (attached), the authors claim that PMG is not appropriate in such cases (neither is DOLS or FM-OLS). The authors suggests augmenting the regression with cross-sectional averages of the dependent and explanatory variables.

According to my understanding, the PMG as implemented by EViews (in a panel setting) do not augment the regression with cross-section averages. My question is: do you see a problem with adding these cross-section averages to a regression and continue to use PMG as the estimation method?

Many thanks!