I added a gretl file (of simulated data),
When i try to perform a Arellano-Bond estimation of Y1~c+x1 and one AR term, GRETL crashes
(using linux ubuntu and windows xp pro and gretl 1.7.9)
Thanks for the report.
The problem is that with a time-series length of 100 you have a
monstrous number of instruments. I expect gretl is exhausting
system memory and so being killed by the OS. In the script
larbond command you have the option of setting a limit to the
lags used, as in
arbond 1 10 ; y1 0 x1
(though even 10 is probably a bit high).
The bug here is that there's currently no way to limit the
instruments used in the GUI dialog for arbond; we'll have
to add that (and probably set the default to a sane maximum).
Fixed in CVS and Windows snapshot.
Log in to post a comment.
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.