Sign up to our newsletter for the latest news,updates and offers


Lecture: Local Polynomial Order in Regression Discontinuity Designs

Lecturer:  Zhuan PEI, Associate Professor, Cornell University 

Moderator: Prof. Zhong LIU, School of Business Administration

Time: 09:30-11:00 Thursday, Jan. 7, 2021

Platform: Tencent Meeting 993 828 832 

Organizer: School of Business Administration 

Speaker Profile

Zhuan Pei joined the Department of Policy Analysis and Management at Cornell University in July 2015 as an assistant professor. His fields of interest including: Labor Economics, Applied Micro-econometrics and Public Policy. Zhuan Pei has already published several articles in the journal of AER、Econometrica、Advances in Econometrics and so on. In his research, he investigates the effect and design of social and employment programs and studies applied micro-econometric methods in causal inference. Prior to Cornell, he was a postdoctoral economist at the W. E. Upjohn Institute for Employment Research from 2012 to 2013 and an assistant professor of economics at Brandeis University between 2013 and 2015.

Lecture Preview

Treatment effect estimates in regression discontinuity (RD) designs are often sensitive to the choice of bandwidth and polynomial order, the two important ingredients of widely used local regression methods. While Imbens and Kalyanaraman (2012) and Calonico, Cattaneo and Titiunik (2014) provide guidance on bandwidth, the sensitivity to polynomial order still poses a conundrum to RD practitioners. It is understood in the econometric literature that applying the argument of bias reduction does not help resolve this conundrum, since it would always lead to preferring higher orders. We therefore extend the frameworks of Imbens and Kalyanaraman (2012) and Calonico, Cattaneo and Titiunik (2014) and use the asymptotic mean squared error of the local regression RD estimator as the criterion to guide polynomial order selection. We show in Monte Carlo simulations that the proposed order selection procedure performs well, particularly in large sample sizes typically found in empirical RD applications. This procedure extends easily to fuzzy regression discontinuity and regression kink designs.