Hello,
Is there anyone who has implemented the generalised Hausman test (robust to
heteroskedasticity and autocorrelation) in Arellano's 1993 article "On the
testing of correlated effects with panel data" with Ox-DPD?
It is also described on bottom of page 9 in the DPD manual.
I know all the components are there, but I can't figure out how to get them
together properly. Shall I use Gmm and GmmLevel functions?
I have created the time averages of my RHS variables, but how to separate
the "differences" equations from the "levels" equations, without using
GMM-SYS estimation? And as far as I understand, it shouldn't be necessary
to use instrumental variables, but only the transformed y and x data.
So I wonder how to "insert" the levels observations (the time averages)
into each cross-section Wi.
==================================
Tom Petersen
Transport and Location Analysis (TLA)
Div. of Systems analysis and Economics (SE)
Dept. of Infrastructure
Teknikringen 78 B, 1st floor
KTH - Royal Institute of Technology
SE-100 44 STOCKHOLM
Sweden
Phone +46 (0)8-790 96 32
Mobile +46 (0)70-424 00 75
Fax +46 (0)8-790 70 02
http://www.infra.kth.se/tla/
|