[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: RE: xtivreg2 and Wooldridge's correction for attrition bias in a panel
> -----Original Message-----
> From: firstname.lastname@example.org
> [mailto:email@example.com] On Behalf Of
> Patrick Van Horn
> Sent: Friday, September 29, 2006 3:56 PM
> To: firstname.lastname@example.org
> Subject: st: xtivreg2 and Wooldridge's correction for
> attrition bias in a panel
> I have a panel of banks during the Great Depression with
> quarterly observations. I am interested in seeing how they
> were managing their cash reserves during this time. Therefore
> I the have reserves to assets ratio as my dependent variable
> and my independent variables are bank characteristics, macro
> variables, and monetary variables.
> I was running the xtreg, fe command for a fixed effects
> model. However, my data set suffers from attrition as some
> banks go out of business and leave the data set for good.
> Since the attrition will produce biased estimates in the
> fixed effects approach, I was referred to Wooldridge's
> discussion of attrition in his panel data book. He recommends
> an approach that uses first differences and then an IV OLS
> regression. In his book, its on pages 585-590 if this is not clear.
> It seems that the xtivreg2 command does exactly what he
> recommends for correcting for attrition. Am I correct in this
> assumption? Or do I manually need to first difference and
> then run an IV estimation?
The -fd- option of xtivreg2 will run an IV panel regression with all
variables in first-differences - dependent variable, regressors,
instruments, the lot. Whether that is actually what Wooldridge
recommends is another story, since your description is very brief.
Sounds like it might be, though.
Prof. Mark E. Schaffer
Centre for Economic Reform and Transformation
Department of Economics
School of Management & Languages
Edinburgh EH14 4AS UK
> Forgive me if I am asking
> something that is extremely remedial or nonsensical. I am an
> economic historian by training and not very well versed in
> econometrics of this type.
> Any input would be greatly appreciated.
> Thank You
> * 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/
* For searches and help try: