Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Bee Wah TAN <bewahtan@gmail.com> |
To | statalist <statalist@hsphsun2.harvard.edu> |
Subject | st: 0 lag command for Pooled Mean Group (PMG) Estimator |
Date | Thu, 21 Nov 2013 12:07:50 +0800 |
Dear All Statalist User, I am running the Pooled Mean Group (PMG) estimator. The best lag length - AIC, BIC for my ARDL model is (1,2,1,1). Thus, the PMG command is written as this:- xtpmg d(1).lrgdpc d(1/2).lglob d(1).lpri d(1).lsmc, lr(l.lrgdpc lglob lpri lsmc) ec(ec) replace pmg After determine the ARDL best lag is (1,2,1,1) and run out the PMG results, I continue with diagnostic checking, and the command for serial correlation diagnostic check is as below:- First, I run the regression base on the best lag:- reg dlrgdpc lglob lpri lsmc dlrgdpc(-1) dlglob dlglob(-1) d.dlglob(-2) dlpri dlpri(-1) dlsmc dlsmc(-1) Then, I test the serial correlation test by putting the xtserial command:- xtserial dlrgdpc lglob lpri lsmc dlrgdpc(-1) dlglob dlglob(-1) d.dlglob(-2) dlpri dlpri(-1) dlsmc dlsmc(-1) My question is:- 1. What is the command for 0 lag? For lag 1, we put d(1).lglob for example. If lag 0, is that we need to write d(0).lglob for 0 lag?? 2. I am running the xtserial diagnostic checking but I am not sure whether it is the correct way/command? Can someone advice me further? Thanks -- *Best Regards, *Bee Wah TAN *[image: Description: http://ybpguide.com/2008/04/22/go-green-checklist/]SAVEPAPER - Please do not print this e-mail unless absolutely necessary* * * 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/