Professor Chen Qiang’s Co-authored Paper Published Online in Journal of Econometrics

Recently, the co-authored paper Quantile Control via Random Forest by Professor Chen Qiang from the School of Economics, Shandong University was published online in Journal of Econometrics, a Class A international top journal. Professor Chen Qiang is the first author, and the co-authors are Professor Xiao Zhijie from Boston College and Assistant Professor Yao Qingsong from Louisiana State University.

The synthetic control method and regression control method are popular causal inference methods in panel data, especially suitable for situations where there is only one or a few individuals in the treatment group. At present, the statistical inference of the synthetic control method and regression control method mainly relies on placebo tests, but placebo tests require strong assumptions (such as random grouping) and are not easy to meet in practice. To this end, Chen, Xiao and Yao (2024) use quantile regression to estimate the 2.5% and 97.5% quantiles of the counterfactual outcomes of the treated individuals to construct 95% confidence intervals for the treatment effect in each period. However, the traditional linear quantile regression has a slow convergence speed and may have the bias of functional misspecification. Chen, Xiao and Yao (2024) propose to use the machine learning method of "quantile random forest" for quantile regression to construct confidence intervals for the treatment effect, and call this method the "quantile control method (QCM)". As a non-parametric ensemble learning method (for example, random forest is the average of 1000 decision trees), QCM is still robust under heteroscedasticity, autocorrelation or functional misspecification, performs well in small samples, and has great application potential. In the statistical software Stata, the quantile control method can be easily implemented through the command qcm, and the download and installation method is "ssc install qcm, all replace".

Chen Qiang is a Professor at the School of Economics, Shandong University, and a doctoral supervisor in quantitative economics. His research fields include econometrics, machine learning and economic history. He was selected into the Program for New Century Excellent Talents in University of the Ministry of Education in 2010. He has published papers in leading domestic and foreign journals such as Journal of Econometrics, Oxford Economic Papers (Lead Article), Journal of Comparative Economics, Economica, Stata Journal (Lead Article), Management World, Economic Research (Quarterly) and The Journal of World Economy, and authored best-selling textbooks Econometrics and Stata Applications, Advanced Econometrics and Stata Applications, Machine Learning and R Applications and Machine Learning and Python Applications.