Hello All,

I am attempting to normalize my graphs by the unconditional average.

I run the following regressions

reg varlogwageinc i.agegrp9 i.year [aweight=wgt], robust noconst
estimates store time_effect

reg varlogwageinc i.agegrp9 i.cohort [aweight=wgt], robust noconst
estimates store cohort_effect

where agegrp9 is age groups by 5 year bands and cohorts are identified by year of birth.

I plot the two effects wanting the age profiles, I only keep the coeff on the agegrp9

coefplot (cohort_effect, recast(connected) ciopts(recast(rline) lpattern(dash)) ) (time_effect, recast(connected) ciopts(recast(rline) lpattern(dash))) ///
, vertical yline(0) keep(*.agegrp9 _cons) nooffsets ///
xtitle("Age") title("Variance of Log Wages")

I want to normalize the graphs by the unconditional average. So I need to multiple the coeff of the cohort_effect by [the unconditional mean of varlogwageinc which is (.72) divided by the mean of cohort_effect) So basically multiple all the coeff on the agegrp9 by 0.72/.37=1.9. Is it possible to multiple the coeff on agegrp9 by 1.9 and graph it?

Also I will do the same thing for the time_effect.

Thank you for your help