Dear all,
I have an (unbalanced) panel and would like to estimate a system of 2 equations with cross-equation restrictions. I have two problems given this framework: 1) being able to use alternative variance estimators; 2) being able to use other estimators than the dummies-based fixed effects within-groups estimator.
*So far I have used -reg3- with the –ols- option, using hand-made dummies and –constraints-. However, this does not allow for alternative variance estimators like robust, nor can I use a between-groups estimator.
*Alternatively I could use –cnsreg- and stack the data for the two equations, using hand-made dummies and –constraints-, but this gives me the same problems (or is –vce(bootstrap)- accounting for possible heteroskedasticity and/or autocorrelation? I am not familiar with bootstrapping.).
*The panel data commands of Stata do not allow for restrictions.
Does anyone have a suggestion for how to proceed? How can I use both restrictions and some robust estimate of standard errors? And how can I move away from the within-group estimator using hand-made dummies when using restrictions?
Thank you very much in advance for your help.
Best regards,
Edwin van der Werf
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/