version 16 cap log close set more off clear* log using bootstrap_estimates.log, replace ****First update the standard errors in the profit regression (Table A7) use bootstrapestimatesprofit.dta, replace drop if run>109 for var beta*: egen est_X=median(X) for var beta*: egen se_X=sd(X) for var beta*: egen topci1_X=pctile(X), p(99.5) for var beta*: egen lowci1_X=pctile(X), p(0.5) for var beta*: egen topci5_X=pctile(X), p(97.5) for var beta*: egen lowci5_X=pctile(X), p(2.5) for var beta*: egen topci10_X=pctile(X), p(95) for var beta*: egen lowci10_X=pctile(X), p(5) ****Table Appendix A7 *lni *Industry FE forvalues k=5(5)10{ sum betafirmlni`k' *_betafirmlni`k' sum betainvlni`k' *_betainvlni`k' sum betalatlni`k' *_betalatlni`k' sum betaconslni`k' *_betaconslni`k' sum obslni`k' sum r2lni`k' } *Firm FE forvalues k=5(5)10{ sum betafirmlnife`k' *_betafirmlnife`k' sum betainvlnife`k' *_betainvlnife`k' sum betalatlnife`k' *_betalatlnife`k' sum betaconslnife`k' *_betaconslnife`k' sum obslnife`k' sum r2lnife`k' } *tobin q *Industry FE forvalues k=5(5)10{ sum betafirmtob`k' *_betafirmtob`k' sum betainvtob`k' *_betainvtob`k' sum betalattob`k' *_betalattob`k' sum betaconstob`k' *_betaconstob`k' sum obstob`k' sum r2tob`k' } *Firm FE forvalues k=5(5)10{ sum betafirmtobfe`k' *_betafirmtobfe`k' sum betainvtobfe`k' *_betainvtobfe`k' sum betalattobfe`k' *_betalattobfe`k' sum betaconstobfe`k' *_betaconstobfe`k' sum obstobfe`k' sum r2tobfe`k' } ****Now get the standard errors and the point estimates for the moving regression (Table A5, figure 3) use bootstrapestimatesmove.dta, replace drop if run>109 for var beta*: egen est_X=median(X) for var beta*: egen se_X=sd(X) for var beta*: egen topci1_X=pctile(X), p(99.5) for var beta*: egen lowci1_X=pctile(X), p(0.5) for var beta*: egen topci5_X=pctile(X), p(97.5) for var beta*: egen lowci5_X=pctile(X), p(2.5) for var beta*: egen topci10_X=pctile(X), p(95) for var beta*: egen lowci10_X=pctile(X), p(5) ***Table A5 forvalues k=5(5)10{ **linear sum betamoveinv`k' *_betamoveinv`k' sum betamovefirm`k' *_betamovefirm`k' sum betamovecoworker`k' *_betamovecoworker`k' sum betamovetenure`k' *_betamovetenure`k' sum betamovecons`k' *_betamovecons`k' sum obsmove`k' sum r2move`k' **by decile matrix betamoveinv`k' = J(9,3,.) forvalues d=2(1)10{ sum betamoveinv`k'dec`d' *_betamoveinv`k'dec`d' quietly { local v = `d'-1 sum betamoveinv`k'dec`d' matrix betamoveinv`k'[`v',1] = r(mean) sum lowci5_betamoveinv`k'dec`d' matrix betamoveinv`k'[`v',2] = r(mean) sum topci5_betamoveinv`k'dec`d' matrix betamoveinv`k'[`v',3] = r(mean) } } sum betamovedecfirm`k' *_betamovedecfirm`k' sum betamovedeccoworker`k' *_betamovedeccoworker`k' sum betamovedectenure`k' *_betamovedectenure`k' sum betamovedeccons`k' *_betamovedeccons`k' sum obsmovedec`k' sum r2movedec`k' *Figure 3: coefplot (matrix (betamoveinv`k'[,1]) , ci((betamoveinv`k'[,2] betamoveinv`k'[,3]))), vertical /// graphregion(color(white)) bgcolor(white) legend(off) ylabel(, nogrid) xtitle(Inventor Effect Decile) title(`k' year rolling window) ytitle(Beta) /// coeflabels(r1 = "2" r2 = "3" r3 = "4" r4 = "5" r5 = "6" r6 = "7" r7 = "8" r8 = "9" r9 = "10") graph save movefullper`k'decile.gph, replace } ****Now get the standard errors and the point estimates for the matching regression (Table 4) ***Table 4 use bootstrapestimatesmatch.dta, replace drop if run>109 for var beta*: egen est_X=median(X) for var beta*: egen se_X=sd(X) for var beta*: egen topci1_X=pctile(X), p(99.5) for var beta*: egen lowci1_X=pctile(X), p(0.5) for var beta*: egen topci5_X=pctile(X), p(97.5) for var beta*: egen lowci5_X=pctile(X), p(2.5) for var beta*: egen topci10_X=pctile(X), p(95) for var beta*: egen lowci10_X=pctile(X), p(5) forvalues k=5(5)10{ sum betamatch1*`k' *_betamatch1*`k' sum obsmatch1`k' sum r2match1`k' sum betamatch2*`k' *_betamatch2*`k' sum obsmatch2`k' sum r2match2`k' sum betamatch3*`k' *_betamatch3*`k' sum obsmatch3`k' sum r2match3`k' sum betamatch4*`k' *_betamatch4*`k' sum obsmatch4`k' sum r2match4`k' sum betamatch5*`k' *_betamatch5*`k' sum obsmatch5`k' sum r2match5`k' } cap log close clear