st: re: unit root test for panels

[Kit Baum <baum@bc.edu>]

Claude said


I am trying to do an Augmented Dickey-Fuller test for covariance
stationarity on a panel. I understand that panelunit can do this. Could
anybody tell me what the syntax should be? As you can see below,  
panelunit
should be installed on my computer but STATA does not seem to  
recognize the
command. Also, I am not able to find panelunit.ado on my computer. I  
used
"ssc install panelunit" and the installation seemed to have gone through
correctly.


. sort id year

. tsset id year
       panel variable:  id (strongly balanced)
        time variable:  year, 2004 to 2005

. dfuller csp
sample may not include multiple panels
r(459);

. panelunit csp
unrecognized command:  panelunit
r(199);




First of all -panelunit- is no longer necessary unless you are  
running an older version of Stata. It was a hack to permit -dfuller-,  
-pperron- and -dfgls- to work on a single time series within a panel,  
which was always technically feasible, but not implemented in  
official Stata's commands. StataCorp has in the meanwhile implemented  
those fixes. There is no -help panelunit- because the commands  
dfuller2, pperron2 and dfgls2 worked exactly as their official  
versions did, with this one added capability (which was literally one  
line of code).

Second, being able to access a single time series within a panel for  
unit root testing does not a "panel unit root test" make. For that,  
as other respondents have indicated, you might use -madfuller-. Or  
you might -findit panel unit root test- to explore the menu of  
alternatives.

Kit

Kit Baum, Boston College Economics and DIW Berlin
http://ideas.repec.org/e/pba1.html
An Introduction to Modern Econometrics Using Stata:
http://www.stata-press.com/books/imeus.html


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/