Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Panel GMM in TSP
4 views
Skip to first unread message
nick
Oct 1, 2007, 11:47:35 AM10/1/07
to
Hi,
I am applying GMM (Arellano-Bond estimator) to estimate a dynamic
panel in TSP. However, according to TSP, many of my instruments are
linearly dependent and my data are singular.
Any suggestions to overcome the problem?
I am applying GMM (Arellano-Bond estimator) to estimate a dynamic
panel in TSP. However, according to TSP, many of my instruments are
linearly dependent and my data are singular.
Any suggestions to overcome the problem?
Clint Cummins
Oct 1, 2007, 7:52:57 PM10/1/07
to
How many instruments? Number of estimated parameters?
Are you using a MASK?
These models are complicated - it would probably be simplest to just
contact me directly.
Clint Cummins
TSP International
lysa...@comcast.net
Oct 10, 2007, 2:21:57 PM10/10/07
to
I do not used TSP but I use the DPD package for OX that was created by
Arellano. I think these packages have some similarities.
Try reducing your instruments. Sometimes using too many lags will cause
the matrix not to crunch. If you shorten the maximum lags used as
instruments for the lagged independent variable this usually fixes the
problem. I often got this message in OX as well when there were too many
lags and too few individuals or too short of a time series for some of
the individuals.
0 new messages