Stata：absdid半参数再加权双重差分估计
1、简介
absdid命令是Abadie(2005)半参数差分(SDID)估计的Stata程序。
absdid通过比较治疗组间结果变量随时间的变化来估计ATT，同时调整治疗组间在基线时与倾向评分相关的可观察特征上的差异。
absdid实现了Abadie(2005)的半参数双重差分估计。
估计比较治疗组(tvar(varname) == 1)和未治疗组(tvar(varname) == 0)的depvar(基线和随访结果之间的结果变化)。
为了解决对治疗组的非随机选择，估计根据控制变量的xvar列表(varlist)在基线上调整治疗组之间的可观察差异。
2、语法格式
语法格式：
absdid depvar [if] [in] [, tvar(varname) xvar(varlist) options ]
选项含义为：
tvar(varname)是表示处理变量
xvar (varlist)指定控制变量
3、案例应用
为了说明absdid是如何工作的，让我们根据该命令提供的官方数据进行操作演示：
我们估计了参加工会对加入工会的女工工资的影响。
数据来源：所使用的数据节选自当前的人口调查——美国政府对失业率和劳动力参与率的月度调查。
变量介绍：数据包括1996年观察到的女性工人和1997年重新审查的女性工人
被解释变量：dlwage
解释变量：age black hispanic married grade
处理变量：union97
3.1 案例1 ：Estimate the union-wage premium
absdid dlwage, tvar(union97) xvar(age black hispanic married i.grade)
Abadie's semi-parametric diff-in-diff Number of obs = 18466
------------------------------------------------------------------------------
dlwage | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
ATT |
_cons | .0350802 .0163554 2.14 0.032 .0030243 .0671361
------------------------------------------------------------------------------
3.2 案例2 ：加入sle选项Union-wage premium with the sle option
. absdid dlwage, tvar(union97) xvar(age black hispanic married i.grade) sle
Abadie's semi-parametric diff-in-diff Number of obs = 18470
------------------------------------------------------------------------------
dlwage | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
ATT |
_cons | .0356633 .0163639 2.18 0.029 .0035906 .0677359
------------------------------------------------------------------------------
3.3 案例3 ：Union-wage premium when the probability to be treated varies between 0.01 and 0.99
absdid dlwage, tvar(union97) xvar(age black hispanic married i.grade) csinf(0.01) csup(0.99)
Abadie's semi-parametric diff-in-diff Number of obs = 18419
------------------------------------------------------------------------------
dlwage | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
ATT |
_cons | .0350528 .0163598 2.14 0.032 .0029882 .0671174
------------------------------------------------------------------------------
3.4、 其他选项
Union-wage premium using a polynomial function of order 4 to estimate the propensity score
. absdid dlwage, tvar(union97) xvar(age black hispanic married i.grade) order(4)
Variation of union-wage premium across age and education
. absdid dlwage, tvar(union97) xvar(age black hispanic married i.grade) yxvar(age hschool college)
Interaction terms
. absdid dlwage, tvar(union97) xvar(c.age##c.age black hispanic married i.grade) yxvar(c.age##c.age hschool college)
4、参考资料
Abadie, A. 2005. Semiparametric Difference-in-Differences Estimators. Review of Economic Studies 72: 1--19
Hirano, K., Imbens, G.W. and Ridder, G. 2003. Efficient Estimation of Average Treatment Effects Using the Estimated Propensity Score. Econometrica 71(4):1161--1189
来源：The Stata Journal (2016) 16, Number 2, pp. 482–490
Abadie’s semiparametric difference-in-differences estimator
Kenneth Houngbedji
Paris School of Economics
Paris, France
kenneth.houngbedji@psemail.eu
转自：“PPdata Academy”微信公众号
如有侵权，请联系本站删除！
