st: bootstrap to correct standard errors

["Anh Nguyen" <anh@vt.edu>]

Dear Stata_listers,
 I am trying to use bootstrap to correct standard errors in two steps
regressions models as follows :
 PROGRAM :


 probit attempt  goldenpara eindexlessgp  l.size l.leverage1 l.mtb l.mabret
l.de_inc ceo65  yd* indu*, cluster(gvkey)
predict ptake, pr
regress lspread goldenpara ptake eindexlessgp  tacons size leverage1 roa
reswoth mtb sgrowth1yr volatility duration convex lage wio yd* indu*
,cluster(gvkey)
predict err1, resid
gen wcor = (err1 != .)
summarize wcor
drop if ptake == .
 
program twostep1a, eclass
quietly {
drop ptake 
probit attempt  goldenpara eindexlessgp  l.size l.leverage1 l.mtb l.mabret
l.de_inc ceo65  yd* indu*, cluster(gvkey)
predict ptake, pr
regress lspread goldenpara ptake eindexlessgp  tacons size leverage1 roa
reswoth mtb sgrowth1yr volatility duration convex lage wio yd*
indu* ,cluster(gvkey)
}
end
set seed 12345
bootstrap "twostep1a " _b, reps(1000) strata(wcor) cluster(gvkey) 



However, I got this message at the end :  "insufficient observations to
compute bootstrap standard errors. no results will be saved". Using the
noisily command, I got this message : 
" twostep1a not sorted 
captured error running (twostep1a), posting missing values" 


I tried to search for this problem but still can not get the answer so far.
I will greatly appreciate if you could give me any help for this.
 Best Regards,


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/