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/
