Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Fredrik Vest <vest.fredrik@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: Reproducing results from -xtpmg- |
Date | Fri, 7 Mar 2014 15:40:24 +0100 |
Hi, I am not able to reproduce the results from the -xtpmg- command. Take this example: Suppose I am interested in the following ECM: dX=b0 + k*(l.X - l.Y) + b1*dY + error, where d and l denote the difference and lag operator. Why don't a) anb b) below give the same results? a) -xtpmg d.x d.y, lr(l.x l.y) dfe cluster(region)- b) (the way I was taught ECM-models in the pure time-series case) -xtreg x y, fe cluster(region)- -predict xb- -gen resid=x-xb- -xtreg d.x d.y l.resid, fe cluster(region)- Is b) only valid in time-series without panels? Consulting the article in the Stata journal belonging to -xtpmg-, I note the following: "For the FE model, xtpmg is simply a wrapper for Stata's xtreg, fe command (see [XT] xtreg). The underlying parameters of (8) are calculated with nlcom and stored as EC. [...]" How would I estimate my equation first using -xtreg,fe- and then -nlcom-? -Fredrik- * * 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/