Skip to content
This repository was archived by the owner on Mar 30, 2023. It is now read-only.

Latest commit

 

History

History
68 lines (54 loc) · 3.94 KB

drdid.md

File metadata and controls

68 lines (54 loc) · 3.94 KB

drdid Doubly Robust Difference-in-Differences Estimators

Syntax

  • drdid depvar [indepvars] [if] [in] , [ivar(varname)] time(varname) treatment(varname) [noisily method rc1]

drdid implements the locally efficient doubly robust difference-in-differences (DiD) estimators for the average treatment effect proposed by Sant'Anna and Zhao (2020). The estimator combines inverse probability weighting and outcome regression estimators (also implemented in the package) to form estimators with more attractive statistical properties.

Options

Parameters

Parameter Description
ivar Variable indexing groups, e.g., country. When ivar is ignored, repeated cross section data is assumed.
time Variable indexing time, e.g., year
treatment Dummy variable indicating treatment, e.g., reform

Methods

method is one of

Method Description
drimp (default) Sant’Anna and Zhao (2020a) Improved doubly robust DiD estimator based on inverse probability of tilting and weighted least squares
dripw Sant’Anna and Zhao (2020a) doubly robust DiD estimator based on stabilized inverse probability weighting and ordinary least squares
reg Outcome regression DiD estimator based on ordinary least squares
stdipw Abadie (2005) inverse probability weighting DiD estimator with stabilized weights
ipw Abadie (2005) inverse probability weighting DiD estimator
ipwra Inverse-probability-weighted regression adjustment (via teffects)
all Compute all of the above

Options

Option Description
rc1 When using repeated crossection data, the option rc1 requests the doubly robust, but not locally efficient, drimp and dripw estimators.
nosily Request showing the estimation of all intermediate steps.

Remarks

The command may create additional variables in the dataset. __att__ stores the recentered influence function for the estimated statistic and __dy__ stores the change in the outcome for an individual (when panel data is used). These variables are overwritten everytime the command is run.

The command also returns, as part of e(), the coefficients and variance covariance matrixes associated with all intermediate sets. See ereturn list after running the command.

Examples

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

Panel estimator with default drimp method

drdid re age educ black married nodegree hisp re74 if treated==0 | sample==2, ivar(id) time(year) tr(experimental)

Repeated cross section

drdid re age educ black married nodegree hisp re74 if treated==0 | sample==2, time(year) tr(experimental)

Authors

  • Fernando Rios-Avila (Levy Economics Institute of Bard College), maintainer
  • Asjad Naqvi (International Institute for Applied Systems Analysis)
  • Pedro H. C. Sant'Anna (Vanderbilt University)

License and Citation

You are free to use this package under the terms of its license. If you use it, please cite both the original article and the software package in your work:

  • Sant’Anna, Pedro H. C., and Jun Zhao. 2020a. “Doubly Robust Difference-in-Differences Estimators.” Journal of Econometrics 219 (1): 101–22.
  • Rios-Avila, Fernando, Asjad Naqvi and Pedro H. C. Sant'Anna. 2021. “DRDID: Doubly Robust Difference-in-Differences Estimators for Stata.” [Software] Available at https://github.com/friosavila/csdid_drdid/tree/v0.1

References

  • Abadie, Alberto. 2005. “Semiparametric Difference-in-Differences Estimators.” The Review of Economic Studies 72 (1): 1–19.
  • Sant’Anna, Pedro H. C., and Jun Zhao. 2020a. “Doubly Robust Difference-in-Differences Estimators.” Journal of Econometrics 219 (1): 101–22.
  • Sant’Anna, Pedro H. C., and Jun Zhao. 2020b. “DRDID: Doubly Robust Difference-in-Differences Estimators.” [Software] Available at https://cran.r-project.org/package=DRDID