时间:2019年06月27日(星期四)14:00-15:00
地点:学院南路,学术会堂 606
报告题目:Estimating EndogenousTreatment Effect Using High-Dimensional
Instrumentswith an Application to the Olympic Effect
报告人:钟威,厦门大学王亚南经济研究院和经济学院统计系教授、博士生导师,经济学院院长助理。
报告摘要:Endogenous treatments arecommonly encountered in program evaluations using observational data where theselection-on-observables assumption does not hold. In this paper, we develop atwo-stage approach to estimate endogenous treatment effects usinghigh-dimensional instrumental variables. In the first stage, instead of using alinear reduced form regression in the conventional two-stage least squaresapproach, we propose a new high-dimensional logistic reduced form model withthe SCAD penalty to approximate the optimal instrument. In the second stage, wereplace the original treatment variable by its estimated propensity score andrun a least squares regression to obtain the penalized Logistic-regressionInstrumental Variables Estimator (LIVE). We show that the proposed LIVE isroot-n consistent to the true average treatment effect, asymptotically normaland achieves the semiparametric efficiency bound. Monte Carlo simulationsdemonstrate that the LIVE outperforms the traditional TSLS estimator and thepost-Lasso estimator for the endogenous treatment effects. Moreover, in theempirical study, we investigate whether the Olympic Games could facilitate thehost nation's economic growth using data from 163 countries. The proposed LIVEestimator shows a strong Olympic effect on the host nation's economic growth.
报告人简介:钟威,现任厦门大学王亚南经济研究院和经济学院统计系教授、博士生导师,经济学院院长助理。2012年获得美国宾夕法尼亚州立大学统计学博士学位,2014年和2017年分别破格晋升副教授和教授,2018年入选厦门大学“南强青年拔尖人才”(A类)。2016年入选福建省高校杰出青年科研人才计划,主持国家自然科学基金面上、青年项目各1项,2019年获得福建省科技厅自然科学杰出青年基金。担任美国统计协会(ASA)期刊《Statistical Analysis and Data Mining》的副主编。主要从事高维数据统计分析和理论、统计学习和数据挖掘算法、计量经济学内生问题、统计学和数据科学的应用等领域的研究。在The Annals of Statistics, Journal of the American StatisticalAssociation, Biometrika, Journal of Business & Economic Statistics, Annalsof Applied Statistics, Statistica Sinica,中国科学数学等国内外统计学权威期刊发表(含接收)20篇论文,其中入选ESI前1%高被引论文2篇。主要讲授《数理统计》、《广义线性模型》、《计量经济学》、《统计数据分析》等本科和研究生课程,多次获得学院教学优秀奖(前10%),2016年获得厦门大学第五届英语教学比赛一等奖。
本次活动受威尼斯432888cam2019专题学术讲座项目资助。