Bootstrap on residuals using Control function

[Jae Li]

Hello everyone! I would like to bootstrap samples to estimate residuals based on two first-stage regressions by the control function, the following is my code but showed errors:

insufficient observations to compute bootstrap standard errors
no results will be saved
r(2000);

Do you possibly know where it got wrong? Any comments will be highly grateful! Many thanks!

xtset gvkey fyear
        program boot_cf, eclass
        capture drop resid
        gen newid4 =gvkey
        xtset newid4 fyear
        reg inve iv $control i.sic3 i.year, vce(robust)
        predict resid1, res
        reg inve_dc iv iv_dc $control i.sic3 i.year, vce(robust)
        predict resid2, res
        reg ret inve_dc inve $control  resid1 resid2  i.sic3 i.year, vce(robust)
        drop resid1 resid2
        end
        bootstrap, reps(50): boot_cf