HDFE Linear regression Number of obs = 600
Absorbing 2 HDFE groups F( 8, 29) = 6.41
Statistics robust to heteroskedasticity Prob > F = 0.0001
R-squared = 0.9408
Adj R-squared = 0.9347
Within R-sq. = 0.1408
Number of clusters (pro) = 30 Root MSE = 0.1765
(Std. Err. adjusted for 30 clusters in pro)
------------------------------------------------------------------------------
| Robust
lnci | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
pre6 | .2076109 .0939738 2.21 0.035 .0154129 .3998089
pre5 | .1051003 .057062 1.84 0.076 -.0116047 .2218053
pre4 | .0801337 .0550019 1.46 0.156 -.0323577 .1926252
pre3 | .1080589 .0453272 2.38 0.024 .0153543 .2007634
pre2 | .0842464 .0227338 3.71 0.001 .0377507 .1307422
current | .1776746 .0542038 3.28 0.003 .0668155 .2885337
post | -.2033604 .0704668 -2.89 0.007 -.3474811 -.0592397
post1 | .1328196 .0520221 2.55 0.016 .0264224 .2392168
post2 | 0 (omitted)
post3 | 0 (omitted)
post4 | 0 (omitted)
post5 | 0 (omitted)
post6 | 0 (omitted)
_cons | .989278 .0099459 99.47 0.000 .9689364 1.00962
------------------------------------------------------------------------------
Absorbed degrees of freedom:
-----------------------------------------------------+
Absorbed FE | Categories - Redundant = Num. Coefs |
-------------+---------------------------------------|
year | 20 1 19 |
pro | 30 30 0 *|
-----------------------------------------------------+
* = FE nested within cluster; treated as redundant for DoF computation
.
end of do-file
代码:*事件研究法///多期did要生成个体维度和政策维度的虚拟变量
gen tpilot=2013 if (pro_1==4|pro_1==7|pro_1==1|pro_1==12)
replace tpilot=2014 if pro==21
replace tpilot=2014 if pro==27
replace tpilot=2016 if pro==24
gen policy=year-tpilot //生产政策时点前后期数,tpilot是政策实施的时点
table policy //了解数据情况
replace policy = -6 if policy < -6
///根据政策时点前后期数event,生成表示政策时点前后各期的虚拟变量,*第一步,生成变量d_j、dj、current。
forvalues i=6(-1)1{
gen pre`i'=(policy==-`i'& treat==1)
}
gen current=(policy==0 & treat==1) /////第一步,生成变量prej、postj、current。
forvalues i=1(1)6{
gen post`i'=(policy==`i'& policy==1)
}
drop pre1
///丢掉一期作为基准组
reghdfe lnci pre* current post* lnpopi lnpgdp tegdp lntech lnfin lnte si, absorb(year pro) vce(cluster pro) //回归呈现政策前后各期的系数变化