RE: st: RE: RE: feeding optimal lag lengths in levinlin andipshinwithin a loop

["Nick Cox" <n.j.cox@durham.ac.uk>]

Thanks. In fact, a closer scrutiny of your 
original reveals what may be the main problem, 
so my request for extra information should 
have been superfluous in this case. 

I link you need to call 

lags(A1 A2 <etc.> 

_not_ 

lags(`A1' `A2' <etc.>)  

I wouldn't do it that way, however: 

foreach k of varlist netprem disclm elr solv {
	preserve
	keep if variable == "`k'"
	forvalues x = 1/19 {
		sum lagrawnotrend if comcode == `x', mean
		local lags "`lags' `r(mean)'" 
	}
	levinlin `k', lags(`lags')
	restore 
}

I still don't buy the explanation for the -preserve-
and -restore-. Sounds as if you just need to modify 
the -if- condition. 

Nick 
n.j.cox@durham.ac.uk 

> -----Original Message-----
> From: owner-statalist@hsphsun2.harvard.edu
> [mailto:owner-statalist@hsphsun2.harvard.edu]On Behalf Of 
> Dev Vencappa
> Sent: 22 October 2003 14:45
> To: statalist@hsphsun2.harvard.edu
> Subject: Re: st: RE: RE: feeding optimal lag lengths in levinlin
> andipshinwithin a loop
> 
> 
> 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
> >>> n.j.cox@durham.ac.uk 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 
> n.j.cox@durham.ac.uk 
> 
> > -----Original Message-----
> > From: owner-statalist@hsphsun2.harvard.edu
> > [mailto:owner-statalist@hsphsun2.harvard.edu]On Behalf Of 
> > Dev Vencappa
> > Sent: 22 October 2003 14:15
> > To: statalist@hsphsun2.harvard.edu
> > 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/
*
*   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/