Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Yogesh Uppal <stata.list.queries@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: Testing for serial correlation in panel data when using IV methods |
Date | Wed, 20 Mar 2013 22:41:52 -0400 |
Hi Statalisters, I am using xtivreg2 with multiple instruments for one endogenous variable in a data with 50 panels and 28 time periods. I need to test for serial correlation. The methods such as xtserial do not work in case of IV methods.The ivactest command gives me an error saying there are multiple panels. The abar command works when I use ivreg2 and include dummies for fixed effects, but does not work when I use xtivreg2. Does that mean abar is not suitable for panel data? Even the discussion in Wooldridge's book on testing for serial correlation for panel data (by including lagged errors in the regression and doing a t-test on it) has nothing to say when using an IV method. Is there a method to test for serial correlation with panel data when using IVs? Thanks for any suggestions. regards, Yogesh * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/faqs/resources/statalist-faq/ * http://www.ats.ucla.edu/stat/stata/