Singluar covariance Matric/How to work with sub sample

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Singluar covariance Matric/How to work with sub sample

Toti
This post has NOT been accepted by the mailing list yet.
Dear Statalist,
I have two questions regarding xtabond2 regression in Stata 13.
I have red both Roodman's papers on how to run the regression and the dangers of too many instruments.
1. I keep on getting the warning which says, "Warning: Two-step estimated covariance matrix of moments is singular. Using a generalized inverse to calculate optimal weighting matrix for two-step estimation.
Difference-in-Sargan/Hansen statistics may be negative." How can I overcome this problem? What problem does a singluar covariance matrix of moments cause?
2. I would like to run the regression on sub-samples and compare the results.For instance, on pooled cross-country, then for the High-income and Low-income countries separately. Although I attempted to use the [if..] option, it says it's a wrong command.
I truly appreciate any help you may provide me.
Toti.
Oklahoma State University