Using coefplot to plot estimates from different methods (ie.TWFE and IW estimators) together on one graph

[Julie Guo]

I am using a panel data, try to estimate dynamic policy effects by two methods: one is the classic dynamic-TWFE specification and the second one is IW estimators provided by Sun & Abraham (2021).
The ultimate ideal graph that I would like to have are similar as follow: Basically, event-study plot by two methods on the same graph.

I made the graphs for these two methods separately, but cannot combine them together.

* dynamic TWFE regression
reghdfe y4 g_11-g_2 g0-g9 $controlsAll , absorb(Li year) cluster(Li)
est store TWFE

* make graph
coefplot TWFE, keep(g_* g*) vertical ///
yline(0,lcolor(edkblue*0.8)) ///
addplot(line @b @at, lcolor(edkblue*0.8)) ///
ciopts(lpattern(dash) recast(rcap) msize(medium)) ///
xlabel(,labsize(*0.75) angle(45)) 


** IW estimator by Sun & Abraham
eventstudyinteract y4 g_11-g_2 g0-g9, cohort(_nfd) control_cohort(never) covariates($controlsAll) absorb(i.Li i.year) vce(cluster Li)

* make graph
matrix C = e(b_iw)
mata st_matrix("A",sqrt(st_matrix("e(V_iw)")))
matrix C = C \ A
matrix list C

coefplot matrix(C[1]), se(C[2]) vertical ///
yline(0,lcolor(edkblue*0.8)) ///
addplot(line @b @at, lcolor(edkblue*0.8)) ///
ciopts(lpattern(dash) recast(rcap) msize(medium)) ///
xlabel(,labsize(*0.75) angle(45))

The two graphs looks like follows: Top one is from TWFE, bottom one is from IW.

I tried addplot. Failed. seems this code only apply for twoway style graphs.
I am not sure does it to do with the second graph is made through a matrix and the first one is from a regression ?

Btw, this is my first post, how do I adjust pictures size smaller (it looks so much bigger than the text now)?...😅

I would really appreciate any help from you!

Best,
Julie

[Andrew Musau]

There does not appear to be any conceptual difficulty in plotting stored estimates together with those from a matrix. So just check where in the process you are going wrong. Note the importance of a reproducible example in illustrating a point, as below. Review FAQ Advice #12 for suggestions on how to post more effectively.

sysuse auto, clear
regress mpg weight i.rep78
mat res= r(table)["b".."se", 1...]
est sto A
set scheme s1color
coefplot (matrix(res), drop(_cons) se(2)) (A, drop(_cons)), saving(g3, replace)