Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Sharooon <treeshar@yahoo.com> |
To | "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu> |
Subject | st: consistency of fixed effects and within estimator |
Date | Fri, 21 Sep 2012 09:35:10 -0700 (PDT) |
I want to run a regression using panel data where xit is correlated with epsilonit, however, xi(t-1) is not correlated with epsilonit. I was using the within estimator to do this, and I would have liked to include lagged yit, however I read this is inconsistent with the within estimator. My question, is whether or not using xi(t-1) is also inconsistent in this setting? Furthermore, if I were to use the dummy variable estimator, I get the same point estimates for my beta's but bigger standard errors. Do I also have consistency problems here? Can I used yit-1 in the dummy variable context? What about xit-1? * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/