Coeffplots and labelling

[Yanis Bekhti]

Hi all,

I seek to produce a relatively descriptive analysis of the share of employment within different manufacturing sectors over a period of interest. In this case, the commodity boom period (2002-2014). To capture the heterogeneity that exists between my countries, I have regressed the share of employment within each sector against total employment over my period of interest (a dummy that takes the value of 1 for the years of interest) and included an interaction term between my period of interest and a dummy that takes the value of "1" for each country in my sample (9). My code is as follows and works.

levelsof Country_ABV, local(Country_local)
foreach x of local Country_local {
gen `x'= 0
    replace `x'=1 if Country_ABV == "`x'"
    eststo gie`x': reg Share_Employment_Sectors_WP $Controls i.GP##`x' i.Country i.Sectors, cluster(Country)
    margins, dydx(i.GP) over(`x')
    eststo mge`x': margins if `x'==1, dydx(i.GP) post
}

In addition to my results, I wanted to produce a graphical representation using Coeffplots and the margins of error calculated using the Taylor approximation method (i.e., code above). However, I have a problem because I cannot link each of my estimates with my labels which I would like to place, not in the graph region, but on the y-axis. Here is my code:

coefplot ///
(mgeBOL, ms(o) mcolor(black) ciopts(recast(rcap) color(black))) ///
(mgeBRA, ms(o) mcolor(black) ciopts(recast(rcap) color(black))) ///
(mgeCHL, ms(o) mcolor(black) ciopts(recast(rcap) color(black))) ///
(mgeCOL, ms(o) mcolor(black) ciopts(recast(rcap) color(black))) ///
(mgeCRI, ms(o) mcolor(black) ciopts(recast(rcap) color(black))) ///
(mgeECU, ms(o) mcolor(black) ciopts(recast(rcap) color(black))) /// 
(mgeMEX, ms(o) mcolor(black) ciopts(recast(rcap) color(black))) /// 
(mgePER, ms(o) mcolor(black) ciopts(recast(rcap) color(black))) /// 
(mgeURY, ms(o) mcolor(black) ciopts(recast(rcap) color(black))) ///
, ///
ysize(5) legend(off) ylabel(none) ///
xsize(5) xlabel() ///
xline(0, lcolor(black) lpattern(solid)) ///
ylabel(1 "BOL" 2 "BRA" 3 "CHL" 4 "COL" 5 "CRI" 6 "ECU" 7 "MEX" 8 "PER" 9 "URY", angle(0)) ///
title("{bf: - Heterogeneity in Formal Employment -}", color(black) margin(bottom) size(3.5)) ///
horizontal ///
graphregion(color(white))
graph save "$dirrfigs\Model 2 - Employment Heterogeneity.dta", replace 
graph export "$dirrfigs\Model 2 - Employment Heterogeneity.png", replace

I tried several methods, but no way to link my "mgeBOL", "mgeBRA" with my ylabel "BOL", "BRA", etc. So I get the following output (below). Could I have some help?

[Andrew Musau]

This thread should be helpful: https://www.statalist.org/forums/for...ot-with-string. coefplot is from SSC (FAQ Advice #12).