题　目：Algorithmic Self-Preferencing on E-Commerce Platforms: Evidence from JD.COM

嘉　宾：张任宇，副教授，香港中文大学

主持人：江浦平 助理教授 上海交通大学安泰经济与管理学院

时　间：2024年5月16日（周四）10:00-11:30am

地　点：安泰楼A507室

内容简介：

E-commerce platforms that are both designers and participators of a marketplace, such as Amazon and JD.COM, might leverage recommendation algorithms to preferentially promote their own products, a phenomenon termed “algorithmic self-preferencing.” In response to the increasing scrutiny from both regulatory bodies and the academic community, our study introduces a comprehensive framework to define, predict, and detect instances of algorithmic self-preferencing within such e-commerce platforms. We define algorithmic self-preferencing as the platform's excessive promotion of its own products over equivalent ones sold by third-party sellers and establish the necessary and sufficient condition for such self-preferencing phenomenon. Our theoretical model predicts that a self-preferencing algorithm is characterized by higher consumer traffic (i.e., the number of clicks) but a lower conversion rate for the platform-owned product. We then leverage a large public dataset from JD.COM to empirically identify algorithmic self-preferencing. Utilizing coarsened exact matching for causal identification, our empirical findings reveal that platform-owned products receive 55.65% more clicks yet exhibit 23.77% lower conversion rates compared to their identical third-party counterparts. This disparity not only evidences algorithmic self-preferencing on JD.COM but also validates our theoretical predictions. Our extensive robustness checks confirm that the presence and magnitude of algorithmic self-preferencing are robust with respect to different identification strategies and model specifications. Finally, we further demonstrate the generalizability of our framework with the data from Amazon, revealing similar patterns of self-preferencing.

演讲人简介：

张任宇，香港中文大学商学院副教授（with tenure），快手经济学家&Tech Lead，主要研究数据科学方法论（包括数据驱动优化、机器学习和因果推断）及其在大规模在线平台运营策略的评估与优化中的应用。研究成果在Management Science, Operations Research, Manufacturing & Service Operations Management等顶级期刊发表并获得INFORMS, POM等多个学术共同体研究奖励。研究项目获得NSFC, SMEC, STCSM和HK RGC资助。担任学术期刊Production and Operations Management的Senior Editor和Naval Research Logistics的Associate Editor。在香港中文大学、纽约大学和快手内部讲授数据科学、运筹学和经济学课程。为快手平台开发经济学/数据科学方法论与框架，应用于评估并优化平台增长策略以及宏观流量与营收生态。个人网站：https://rphilipzhang.github.io/rphilipzhang/

欢迎广大师生参加！