Nick
Sorry for missing some details in the explanation. This is how the data looks like. Let us say in one panel I have variables A B C D E F and suppose variable A is the optimal lag length for each company in that panel. This one panel however refers to an identifier variable called netprem. Now I append to the first panel a second panel which is for for identifier variable disclm, which takes the same values for variables B C D E F but has different values for variable A (as the optimal lag length will be different for a different identifier and each company for that diferent identifier). Same applies to the panel for identifiers elr and solv. This is why I have to use preserve to ask stata to run the levinlin test for one identifier, and then restore the data to run it for the second identifier and so on.
When using this loop, the error message I get is
option lags() required
which seems to indicate that stata cannot read these scalar values A1, A2, etc in the levinlin option.
Hope that helps in finding a solution to the problem.
Many thanks
Dev
Dev Vencappa
School of Economics
University of Nottingham
University Park
Nottingham
NG7 2RD
U.K.
Tel : +44 (0)115 951 5608
Fax: +44 (0) 115 951 4159
>>> [email protected] 10/22/03 2:26 PM >>>
You don't say how this fails to work, but a lot may
hinge on
keep if variable == "`k'"
which suggests either an unusual data structure
or a misunderstanding of -keep-.
However, none of the -preserve-, -restore- or
-keep- seems essential for the rest of the code.
Also, are the results of -summarize- all integers?
Nick
[email protected]
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]]On Behalf Of
> Dev Vencappa
> Sent: 22 October 2003 14:15
> To: [email protected]
> Subject: st: RE: feeding optimal lag lengths in levinlin and
> ipshinwithin a loop
>
>
> Stata users, I have a slight problem with panel unit root test.
> Thanks to an earlier suggestion by Nick, Ernest and others,
> I managed to replicate what I am doing below for simple
> time series.
> My current problem is that I have four variables recorded
> in a panel of 19 companies over 18 years. For each
> variable,I have worked out an optimal lag length for each
> company, and would now like levinlin to be able to read
> this optimal lag length for each company rather than me
> having to feed it in the lags option of the command every
> time. I have come up with the following loop, which does
> not seem to work unfortunately. Can anybody spot the
> problem please and suggest a shortcut to that if possible?
> The same thing will also be used for ipshin. Does anybody
> also know how to save the results of ipshin and call them
> into a variable?
>
> foreach k of varlist netprem disclm elr solv {
> preserve
> keep if variable=="`k'"
> forvalues x=1/19{
> sum lagrawnotrend if comcode==`x', mean
> scalar A`x'=r(mean)
> }
>
> levinlin `k', lags(`A1' `A2' `A3' `A4' `A5' `A6' `A7' `A8'
> `A9' `A10' `A11' `A12' `A13' `A14' `A15' `A16' `A17' `A18' `A19')
> restore
> }
>
*
* 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/
*
* 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/
