Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Michele Mancini <mikmancinistata@gmail.com> |
To | statalist <statalist@hsphsun2.harvard.edu> |
Subject | Re: st: ivreg2 and endogeneity |
Date | Thu, 23 Jun 2011 23:08:34 +0200 |
Thank you for your help. X1, X2, X3 and X4 should be very good instruments. The t-test for these in the first step are above 40 in every case. So they're not weakly identified and I look at the overid test of all instruments, saying that something's wrong because the null is rejected. I think that the exclusion restrictions are fine because they're also used in the literature. If I treat X1 as an included instruments I get: X1 is not significant, Hansen reject the null and the endogeneity is not true. If I drop X1 I get: the others instruments are not weak, Hansen accept the null and the endogeneity is true. Should I drop X1 and conclude that fd is endogenous? I can't manage to solve this question. Thanks again Michele Mancini 2011/6/23 Austin Nichols <austinnichols@gmail.com> > > Michele Mancini <mikmancinistata@gmail.com>: > First, think through theory--is X1 really a good instrument? Are X2, X3, X4? > > But also, as a rough guideline, look at stats in this order: > First, look at id stats. If not weakly identified, then overid test > may have good power. > Second, look at overid test. If can't reject null, maybe endogeneity > test has good power. > Third, look at endogeneity test. If reject exogeneity, we prefer IV. > > Note that an endogeneity test is a measure of whether IV differs from > OLS, so if you > are using bad exclusion restrictions, your IV may not differ from OLS > not because > there is no endogeneity problem but because your IV assumptions are wrong. > > It looks to me like X1 is simply not a valid excluded instrument and > should not be used as such. > Perhaps it should be an included instrument, i.e. a control variable? > > On Thu, Jun 23, 2011 at 4:28 PM, Michele Mancini > <mikmancinistata@gmail.com> wrote: > > Dear Statalisters, > > I have some questions regarding my ivreg2 estimation: > > > My question is the following: as you can see the endogeneity test says that > > fd in actually exogenous and the Hansen J statistic strongly reject the > > null; should I treat fd as exogenous and use OLS instead of IV? The problem > > is that if I remove X1 from the instruments > > > How can i interpret these results of the C statistics? Should I drop X1 and > > treat fd as endogenous? Or using OLS instead of IV? > > Thank you so much for your kind help. > > Regards, > > > > Michele Mancini > * > * 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/ * * 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/