Graph plotting outcome variable before after treatement

[Marco Errico]

Dear all,

I am running a staggered difference in difference and I would like, after the regression estimation, to plot a graph as follow.
Y axis: Dependant variable
X axis: time period relative to treatment introduction.

The graph should be more or less as the example below. Given that I have 3 cohorts and each of them receive treatment in 2009,2010,2011 I cannot have x-axis with absolute year. But just t=0 t+1 etc.
Moreover: 1)as below, vertical bands should represent 95% confidence intervals for the point estimates in each period and 2)I need to omit the indicator for year t-1, which serves as the benchmark period with an OLScoefficient and standard error of zero.

Do you have any advice on how to address this issue?

Many thanks

ow:

[Øyvind Snilsberg]

use https://friosavila.github.io/playingwithstata/drdid/mpdta.dta, clear

gen rel_time = cond(first_treat, year - first_treat + 100, 0)

eststo event: xtreg lemp ib(99).rel_time i.year, fe i(countyreal) r

coefplot event, keep(*rel_time) base vert yline(0) xline(4.5)