Interpreting Pooled Mean Group Diagnostics - Eviews10+

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

STERLING
Posts: 1
Joined: Thu Feb 17, 2022 5:26 am

Interpreting Pooled Mean Group Diagnostics - Eviews10+

Postby STERLING » Mon Feb 21, 2022 8:09 am

Investigating a panel/PMG estimation of Environmental degredation as a function of two globalisation regressors + GDP across panel of 45 countries 1975-2019:
PMG/ARDL: log(CO2pc) log(GLOB1) log(GLOB2) log(GDP$ppp)
Had a couple of procedural queries on PMG testing/results:
a. Pre-Testing - I presume that, as with single eq. estimation, all pre-testing - Panel unit roots, Cointegration etc - should be done in log() format ie same as the transformation used in estimation. ?
b. Panel Cointegration testing (Quick-Group Stats-JCT)
I presume I can specify the log operator in "List of series" ie: log(CO2pc) log(GLOB1) log(GLOB2) log(GDP$ppp)
rather than have to create/test explicit variables in log format first ?:"logCO2pc" "logGLOB1" "logGLOB2" "logGDP$ppp"
(so we can always import untransformed "raw" Panel data as a working rule) Any handy guidance available on when to include trend + intercept in panel/PMG ?
c. Hausman Test - How to confirm PMG is preferred to MG estimation ? - as no [seems available (yet ?) Any recommendation as to how to do HT "manually" ? WIll this omission change in near future ?
d. Non-negative error-correction (ECM) coeffs - how to interpret (a significant no. of) these in Cross Section results even though overall reported Average ECM is negative (and below -1.0 ) ? Is this "+ve ECM issue" typical in large N panels/PMG estimation - maybe suggests cross-sections are TOO diverse ?. And in utilising these "non-negative ECM" models (eg in running a model simulation outside of Eviews) could their +ve ECMs ever be simply overwritten by the (negative) Average ECM as a valid "correction" ?
Grateful for any pointers on the above

Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 10 guests