st: Re: levinlin

[Kit Baum <baum@bc.edu>]

Naceur

The "t" variable you mention is not able to be used as a time 
indicator: describe shows

t               str5   %9s

String variables cannot be used in tsset. If you use (from SSC)

tsmktim my, start(1988m1) i(country)

Stata will create the proper time variable and panel tsset:

. tsmktim my, start(1988m1) i(country)
       panel variable:  country, 1 to 6
        time variable:  my, 1988m1 to 2003m11

The real problem with your lnQ variable is that it is a constant over 
countries:

. tabstat lnQ, by(country) stat(mean sd min max)

Summary for variables: lnQ
     by categories of: country

 country |      mean        sd       min       max
---------+----------------------------------------
       1 |  13.09589  .2453448  12.64071  13.60677
       2 |  13.09589  .2453448  12.64071  13.60677
       3 |  13.09589  .2453448  12.64071  13.60677
       4 |  13.09589  .2453448  12.64071  13.60677
       5 |  13.09589  .2453448  12.64071  13.60677
       6 |  13.09589  .2453448  12.64071  13.60677
---------+----------------------------------------
   Total |  13.09589  .2448085  12.64071  13.60677
--------------------------------------------------

You cannot do a panel unit root test on one series repeated six times 
-- the whole notion of the test is that it should take account of the 
similar variation across units of the panel, but if that variation is 
identical, the test breaks down. I will see about trapping that 
condition with a more informative error message, but to date no one 
else has tried to do this.

Thanks
Kit

Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.html

On Nov 29, 2004, at 2:19 PM, Naceur Khraief wrote:


> Dear professor
>
> The time variable is "t" : 1988m1 to 2003m11.
>
> Best,
*
*   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/