Previous Articles     Next Articles

Comparison of Covariate Balance Weighting Methods in Estimating Treatment Effects

ZHAN Mingfeng1, FANG Ying1,2, LIN Ming1,2   

  1. 1. Wang Yanan Institute for Studies in Economics and Fujian Key Laboratory of Statistical Sciences, Xiamen University, Xiamen 361005, China;
    2. Department of Statistics and Data Science, Xiamen University, Xiamen 361005, China
  • Received:2021-02-21 Revised:2021-08-05 Online:2022-11-25 Published:2022-12-23
  • Contact: LIN Ming,
  • Supported by:
    This research was supported by the National Natural Science Foundation of China under Grant Nos. 71631004 and 72033008, the National Science Foundation for Distinguished Young Scholars under Grant No. 71625001, and the Science Foundation of Ministry of Education of China under Grant No. 19YJA910003.

ZHAN Mingfeng, FANG Ying, LIN Ming. Comparison of Covariate Balance Weighting Methods in Estimating Treatment Effects[J]. Journal of Systems Science and Complexity, 2022, 35(6): 2263-2277.

Different covariate balance weighting methods have been proposed by researchers from different perspectives to estimate the treatment effects. This paper gives a brief review of the covariate balancing propensity score method by Imai and Ratkovic (2014), the stable balance weighting procedure by Zubizarreta (2015), the calibration balance weighting approach by Chan, et al. (2016), and the integrated propensity score technique by Sant'Anna, et al. (2020). Simulations are conducted to illustrate the finite sample performance of both the average treatment effect and quantile treatment effect estimators based on different weighting methods. Simulation results show that in general, the covariate balance weighting methods can outperform the conventional maximum likelihood estimation method while the performance of the four covariate balance weighting methods varies with the data generating processes. Finally, the four covariate balance weighting methods are applied to estimate the treatment effects of the college graduate on personal annual income.
[1] Rubin D B, Inference and missing data, Biometrika, 1976, 63(3): 581–592.
[2] Imbens G W and Wooldridge J M, Recent developments in the econometrics of program evaluation, Journal of Economic Literature, 2009, 47(1): 5–86.
[3] Liu Z, Cai Z, Fang Y, et al., Statistical analysis and evaluation of macroeconomic policies: A selective review, Applied Mathematics — A Journal of Chinese Universities, 2020, 35(1): 57–83.
[4] Tang S, Some recent developments in modeling quantile treatment effects, Applied Mathematics — A Journal of Chinese Universities, 2020, 35(2): 220–243.
[5] Rosenbaum P R and Rubin D B, The central role of the propensity score in observational studies for causal effects, Biometrika, 1983, 70(1): 41–55.
[6] Kang J D and Schafer J L, Demystifying double robustness: A comparison of alternative strategies for estimating a population mean from incomplete data, Statistical Science, 2007, 22(4): 523–539.
[7] Imai K and Ratkovic M, Covariate balancing propensity score, Journal of the Royal Statistical Society: Series B, 2014, 76(1): 243–263.
[8] Hansen L P, Large sample properties of generalized method of moments estimators, Econometrica, 1982, 50(4): 1029–1054.
[9] Zubizarreta J R, Stable weights that balance covariates for estimation with incomplete outcome data, Journal of the American Statistical Association, 2015, 110(511): 910–922.
[10] Chan K C G, Yam S C P, and Zhang Z, Globally efficient nonparametric inference of average treatment effects by empirical balancing calibration weighting, Journal of the Royal Statistical Society: Series B, 2016, 78(3): 673–700.
[11] Kitamura Y and Stutzer M, An information-theoretic alternative to generalized method of moments estimation, Econometrica, 1997, 65(4): 861–874.
[12] Owen A B, Empirical likelihood ratio confidence intervals for a single functional, Biometrika, 1988, 75(2): 237–249.
[13] Hansen L P, Heaton J, and Yaron A, Finite sample properties of some alternative GMM estimators, Journal of Business & Economic Statistics, 1996, 14(3): 262–280.
[14] Sant’Anna P H, Song X, and Xu Q, Covariate distribution balance via propensity scores, arXiv preprint arXiv: 1810.01370v4, 2020.
[15] Hainmueller J, Entropy balancing for causal effects: A multivariate reweighting method to produce balanced samples in observational studies, Political Analysis, 2012, 20(1): 25–46.
[16] Zhao Q, Covariate balancing propensity score by tailored loss functions, Annals of Statistics, 2019, 47(2): 965–993.
[17] Wang Y and Zubizarreta J R, Minimal dispersion approximately balancing weights: Asymptotic properties and practical considerations, Biometrika, 2020, 107(1): 93–105.
[18] Wan S K, Xie Y, and Hsiao C, Panel data approach vs synthetic control method, Economics Letters, 2018, 164(3): 121–123.
[19] González-Manteiga W and Crujeiras R M, An updated review of goodness-of-fit tests for regression models, Test, 2013, 22(3): 361–411.
[20] Donald S G and Hsu Y C, Estimation and inference for distribution functions and quantile functions in treatment effect models, Journal of Econometrics, 2014, 178(3): 383–397.
[21] Firpo S, Efficient semiparametric estimation of quantile treatment effects, Econometrica, 2007, 75(1): 259–276.
[22] Melly B, Estimation of counterfactual distributions using quantile regression, Discussion Paper, Universität St. Gallen, 2006,
[23] Hahn J, On the role of the propensity score in efficient semiparametric estimation of average treatment effects, Econometrica, 1998, 66(2): 315–331.
[24] Hirano K, Imbens G W, and Ridder G, Efficient estimation of average treatment effects using the estimated propensity score, Econometrica, 2003, 71(4): 1161–1189.
[25] Advani A, Kitagawa T, and Słoczyński T, Mostly harmless simulations? Using Monte Carlo studies for estimator selection, Journal of Applied Econometrics, 2019, 34(6): 893–910.
[26] Chen T, Ji Y, Zhou Y, et al., Testing conditional mean independence under symmetry, Journal of Business & Economic Statistics, 2018, 36(4): 615–627.
No related articles found!
Full text