就是Stata did平行趋势检验，drop了政策前一期，但pre-_2期一直存在共线性,我换哪一年为基期，都是pre_2被omitted
destring id,replace<br>
gen policy=year-2018<br>
tab policy<br>
replace policy=-3 if policy<-3<br>
replace policy=3 if policy>3<br>
forvalues i=3(-1)1{<br>
gen pre_`i'=(policy==-`i'&treat==1)<br>
}<br>
gen current=(policy==0&treat==1)<br>
forvalues j=1(1)3{<br>
gen post_`j'=(policy==`j'&treat==1)<br>
}<br>
drop pre_1<br>
xtset id year<br>
reghdfe rd pre_* current post_* , a(id year) vce(cl id)<br>
outreg2 using 2.doc, tstat e(r2_a,F) bdec(3) tdec(2)<br>
ssc install coefplot<br>
coefplot, baselevels vertical keep(pre_* current post_*) omitted order(pre_4 pre_3 pre_2 current post_1 post_2 post_3 post_4) level(95) yline(0,lcolor(edkblue*0.8)) xline(3,lwidth(vthin) lpattern(dash) lcolor(teal)) ylabel(,labsize(*0.75)) xlabel(,labsize(*0.75)) ytitle("政策动态效应", size(small)) xtitle("政策时点", size(small)) addplot(line @b @at) ciopts(lpattern(dash) recast(rcap) msize(medium)) msymbol(circle_hollow)<br>