SME-SFI 学术讲座预告 20260817 - 20260824

同款资金，不同效应：投资者流向对股价的影响并非等同

Same Dollar, Different Impact:  Investor Flows Are Not Equally Price-Moving

讲座信息

Seminar Information

主讲人

Speaker

宋阳教授

华盛顿大学

Professor Yang Song

University of Washington

日期和时间

Date and Time

2026年8月21日（周五）

11:30 - 13:00

August 21, 2026 (Friday)

11:30 am - 1:00 pm

地点

Venue

综合教学楼A503会议室

Room 503, Teaching Complex A Building

讲座概述

Abstract

We demonstrate that how investor demand moves prices depends on the institution through which it is executed. Using novel holdings data on separate accounts, the dominant institutional investment vehicle, we show that separate account flows create demand comparable to mutual fund flows but generate essentially no price impact, fragility, or fire-sale dynamics. The contrast persists within mutual fund–separate account twins run by the same managers under identical strategies. Trade-level data from a major transition manager show that specialized institutional execution can substantially lower trading costs and price impact. The evidence reveals large heterogeneity in demand transmission: not all flows are equally price-moving.

主讲人简介

About the Speaker

宋阳教授

华盛顿大学

宋阳是华盛顿大学福斯特商学院的诺曼·J·梅特卡夫金融学教授。他于斯坦福大学商学院获得博士学位，其研究方向为资产定价、投资者行为以及金融市场动态演变。他的研究成果发表于众多顶尖学术期刊，包括《金融杂志》（Journal of Finance）、《金融研究评论》（Review of Financial Studies）、《金融经济学杂志》（Journal of Financial Economics）、《会计与经济学杂志》（Journal of Accounting and Economics）以及《管理科学》（Management Science）。除学术界外，他对市场机制的研究成果已被美国证券交易委员会（SEC）和美联储用于政策制定，并受到《华尔街日报》、《彭博社》和《金融时报》等全球金融媒体的广泛报道。

*Please scroll down for the English version

Yang Song is the Norman J. Metcalfe Professor of Finance at the University of Washington Foster School of Business. He earned his Ph.D. from the Stanford Graduate School of Business, where his research focused on asset pricing, investor behavior, and the evolving dynamics of financial markets. His research has been published in premier academic journals, including the Journal of Finance, Review of Financial Studies, Journal of Financial Economics, Journal of Accounting and Economics, and Management Science. Beyond academia, his findings on market mechanisms have been utilized by the U.S. Securities and Exchange Commission (SEC) and the Federal Reserve for policymaking and have received extensive coverage from global financial media such as The Wall Street Journal, Bloomberg, and the Financial Times.

当生成式AI成为销售的幕后教练

When GenAI Stays Backstage in Sales

讲座信息

Seminar Information

主讲人

Speaker

裴思琦教授

密歇根州立大学

Professor Siqi Pei

Michigan State University

日期和时间

Date and Time

2026年8月21日（周五）

10:30 - 12:00

August 21, 2026 (Friday)

10:30 am - 12:00 pm

地点

Venue

综合教学楼D504会议室

Room 504, Teaching Complex D Building

讲座概述

Abstract

In many customer-facing settings, consumers remain reluctant to interact directly with AI and continue to prefer human salespeople. Can GenAI still create value when it stays behind the scenes? We examine this question through a randomized field experiment in a large automotive retail network, where GenAI operates as a hidden coach, providing guidance to frontline sales consultants without ever interacting with customers. We find that access to GenAI improves sales performance and sales efficiency while shortening customer decision time. We further examine what types of AI guidance are most effective and how such guidance shapes the way salespeople interpret and respond to customer needs. The evidence is consistent with GenAI facilitating learning and more adaptive customer communication. Our findings identify backstage coaching as a distinct form of human–AI complementarity, showing how firms can capture the value of GenAI by augmenting frontline employees’ judgment and learning while keeping the customer relationship fundamentally human.

主讲人简介

About the Speaker

裴思琦教授

密歇根州立大学

裴思琦是美国密歇根州立大学布罗德商学院市场营销系助理教授。她的研究主要关注数字经济、网红营销、汽车营销以及人与人工智能的交互，并广泛关注新兴技术如何影响消费者行为、企业战略与社会福利。她的研究综合运用田野实验、因果推断和机器学习等方法。相关成果发表于《Information Systems Research》、《Management Science》等国际学术期刊，并获得包括WISE最佳论文奖在内的多项学术奖励。她还与日产汽车、寺库、小红书等企业开展合作研究，关注数字化转型、人工智能与消费者决策等问题。

*Please scroll down for the English version

Siqi Pei is an Assistant Professor of Marketing at the Eli Broad College of Business, Michigan State University. Her research focuses on the digital economy, influencer marketing, automotive marketing, and human–AI interaction, with a broader interest in how emerging technologies shape consumer behavior, firm strategy, and social welfare. Her research combines field experiments, causal inference, and machine learning. Her work has been published in Information Systems Research and Management Science. She has also collaborated with industry partners, including Xiaohongshu, Nissan, and Secoo, on projects related to digital transformation, AI, and consumer decision-making.

用于个体化处理效应估计的因果树精度极限

Accuracy Limits of Causal Trees for Individualized Treatment Effects

讲座信息

Seminar Information

主讲人

Speaker

Matias D. Cattaneo 教授

普林斯顿大学

Professor Matias D. Cattaneo

Princeton University

日期和时间

Date and Time

2026年8月24日（周一）

15:30 - 17:00

August 24, 2026 (Monday)

3:30 pm - 5:00 pm

地点

Venue

综合教学楼D904会议室

Room 904, Teaching Complex D Building

讲座概述

Abstract

Recursive decision trees are widely used to estimate heterogeneous causal treatment effects in experimental and observational studies. These methods are typically implemented using CART-type recursive partitioning, with splitting criteria designed to identify variation in treatment effects across covariate-defined subgroups. We study causal tree estimators based on adaptive recursive partitioning and establish lower bounds on their estimation accuracy. The class we analyze includes versions with and without sample splitting, based on common treatment effect and squared-error splitting criteria. Even in a constant-effect benchmark with randomized treatment assignment, causal trees constructed via standard CART-type splitting rules can have uniform-norm errors that decrease more slowly than any power of the sample size. The underlying mechanism is that greedy recursive partitioning selects highly imbalanced splits with nonvanishing probability, producing terminal nodes containing very few observations and leading to large estimation variance. We further show that sample splitting, often called “honesty,” does not remove this limitation. As a consequence, causal tree estimators may converge arbitrarily slowly uniformly over the covariate space. At the same time, these estimators can have small integrated mean squared error, showing that average accuracy can mask local inaccuracy. Our results also clarify the role of balanced partition assumptions in existing theoretical guarantees for causal forests and related ensemble methods.

主讲人简介

About the Speaker

Matias D. Cattaneo 教授

普林斯顿大学

Matias D. Cattaneo 教授现任普林斯顿大学经济学教授，并担任亚马逊学者（Amazon Scholar）。他的研究领域横跨计量经济学、统计学、应用数学、机器学习与人工智能，主要聚焦于数据科学的数学与统计基础。Cattaneo 教授致力于发展统计与计算方法，并将其应用于社会科学、行为科学和生物医学科学等领域，尤其关注项目评估与因果推断。他曾荣获2026年古根海姆数据科学奖学金，是国际统计学会当选会员，同时也是美国统计协会、数理统计学会和国际应用计量经济学会当选会士。其学术贡献获得广泛认可，曾多次获得论文奖项和期刊荣誉，并受邀发表重要学术演讲，拥有多篇高被引论文。Cattaneo 教授获得加州大学伯克利分校经济学博士学位和统计学硕士学位，托尔夸托·迪特利亚大学经济学硕士学位，以及布宜诺斯艾利斯大学经济学学位。

*Please scroll down for the English version

Matias D. Cattaneo is a Professor of Economics at Princeton University and an Amazon Scholar. His research lies at the intersection of econometrics, statistics, applied mathematics, machine learning, and artificial intelligence, with a focus on the mathematical and statistical foundations of data science. Professor Cattaneo develops statistical and computational methods for applications in the social, behavioral, and biomedical sciences, with particular emphasis on program evaluation and causal inference. He received a 2026 Guggenheim Fellowship in Data Science and is an elected Member of the International Statistical Institute, as well as an elected Fellow of the American Statistical Association, the Institute of Mathematical Statistics, and the International Association for Applied Econometrics. His scholarly contributions have been widely recognized through multiple paper awards, journal distinctions, invited lectures, and highly cited publications. Professor Cattaneo earned a Ph.D. in Economics and an M.A. in Statistics from the University of California, Berkeley, a Master’s degree in Economics from Universidad Torcuato Di Tella, and a Licentiate degree in Economics from Universidad de Buenos Aires.