As reproduced here the condition
if `id' -- `i'
looks unlikely to be what you want.
Your code can be trimmed down a bit.
use "C:\Program Files\Stata9\data\GEINDEX\EQ4_1985_2004.dta", clear
tempvar id
bysort SIC = gen `id' = sum(_n == 1)
gen pred = .
qui forval i = 1/`=`id'[_N]' {
tempvar pred
regress TA REV PPE if `id' == `i'
predict `pred' if `id' == `i', xb
replace pred = max(pred, `pred')
drop `pred'
}
save "C:\Program Files\Stata9\EQ4\EQ4_1985_2004.dta", replace
Also, do check out -statsby-.
Nick
[email protected]
[email protected]
> I am using the program below to complete industry level
> regressions and then
> to apply the coefficiencts from the industry regressions to
> firms within each
> industry in order to predict the level of TA. It works very well.
>
> I am wondering if there is a way to imput the regression
> coefficents into my
> data set as variables for each firm's observation in addition
> to or instead of the predictions.
>
> Alternately, can Stata use the coefficients from the first regression:
> TA on REV PPE in a second regressioin: TA on (REV-AR) PPE.
> The model I am trying
> to replicate changes between in first stage and second stage
> regressions.
> >clear
> >use "C:\Program Files\Stata9\data\GEINDEX\EQ4_1985_2004.dta"
> > tempvar id
> > sort SIC
> > by SIC: gen `id'=1 if _n==1
> > qui replace `id'=sum(`id')
> > local i=0
> > tempvar pred0
> > qui gen `pred0'=.
> > local n=`id'[_N]
> > while `i'<`n' {
> > local i=`i'+1
> > local j=`i'-1
> > tempvar pred pred`i'
> > qui regress TA REV PPE if `id'--`i'
> > qui predict `pred' if `id'==`i', xb
> > qui egen `pred`i''=rmax(`pred' `pred`j'')
> > drop `pred' `pred`j''
> > }
> > rename `pred`i'' pred
> > save "C:\Program Files\Stata9\EQ4\EQ4_1985_2004.dta", replace
*
* 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/
