Not So Timely: Screen-On Push Notifications and User Engagement
We study how the timing of mobile push notifications affects user engagement. Industry practice assumes that delivering messages precisely when users are active—such as at screen-on moments—maximizes effectiveness. Using large-scale data from a leading push notification platform in China, we challenge this assumption through two complementary quasi-experiments. A regression discontinuity design around ...
Zachary Zhong, University of Toronto
Wednesday, Nov 26, 2025 | 10:30am-12:00pm | Room 335
More Details
Platform Leakage: Disintermediation and Decentralization in Two-Sided Markets
Leakage happens when buyers and sellers coordinate outside the platform, usually to avoid fees. Despite platforms’ concerns about losing revenue, leakage---by its very nature---is difficult to measure and mitigate. Using geolocation data from an on-demand delivery service, we identify offline transactions that are hard to track in online marketplaces. We exploited a quasi-experiment that gradually ...
Yingkang Xie, Washington University
Thursday, Nov 20, 2025 | 10:30am-12:00pm | Room 335
More Details
The Small Subset Effect in Preference Elicitation: Evidence from Check-All-That-Apply Tasks
Manufacturers, digital platforms, and marketing researchers frequently seek to understand consumer preferences to inform product design, customer profiling, and personalized recommendations. A common method for eliciting such preferences is to present consumers with a list of predefined options and ask them to “check all that apply.” Although this approach is widely used in surveys, online registrations,...
Leilei Gao, The Chinese University of Hong Kong
Wednesday, Nov 19, 2025 | 2:30pm-4:00pm | Room 335
More Details
The Growing Index Effect in the Corporate Bond Market
Leveraging transaction data and AI-powered price estimates, we show that index-driven trading—fueled by the rapid growth of investment funds in the corporate bond market—has reshaped market dynamics and liquidity. Whereas intraday trading was once evenly distributed, by the end of 2024, 10% of daily volume occurred within one minute of index closing. Exploiting Bloomberg Index’s closing-time change,...
Qifei Zhu, National University of Singapore
Wednesday, Nov 19, 2025 | 2:00pm-3:30pm | Room 333
More Details
Simulating Public Opinion with Large Language Models: An Evidence-based Exploration
As a new medium of communication, large language models (LLMs) learn values and attitudes from human-generated data, giving them the potential to capture and reflect public opinion. This talk will examine the capacity of LLMs to simulate public opinion from two perspectives. First, we will investigate the extent to which LLMs can represent public opinion across different nations and social groups, ...
Baohua Zhou, Fudan University
Friday, Nov 14, 2025 | 2:00pm-3:30pm | Room 333
More Details
Intergenerational Welfare Assessments
This paper studies welfare assessments in economies with rich demographics. We introduce the notion of demographically disconnected economies, those with no date at which all individuals are concurrently alive. We identify the unique class of units that enables meaningful welfare comparisons in such economies: those based on perpetual consumption, that is, consumption at all dates. Using this metric,...
Eduardo Davila, Yale University
Wednesday, Nov 5, 2025 | 9:30am-11:00am | Room 337
More Details
From Text to Image to Video: Exploring Generative AI for Public Interest Communication
Generative AI is often viewed as a threat to truth and trust, yet it can also be harnessed for the public good. This talk examines how text-, image-, and video-based AI tools shape public interest communication across three studies. The first study tests AI-generated vaccine correction messages tailored to personality and belief profiles, showing both the promise of scalable targeting and the risks ...
Hang Lu, University of Michigan
Wednesday, Oct 29, 2025 | 10:30am-12:00pm | Room 333
More Details
How Does Active Involvement Benefit Investors? Evidence from 85 Billion Cell Phone Signals
I investigate the reputation effects of active involvement, specifically examining how venture capitalists’ (VCs’) on-site meetings with portfolio companies affect VCs' reputations and future deal flow. By analyzing cell phone signals collected around VC and startup office buildings from 2018 to 2023, I measure VCs’ involvement intensity and deal flow quality. Using exogenous variation in travel ...
Xiaoyong Fu, The University of Hong Kong
Wednesday, Oct 29, 2025 | 2:00pm-3:30pm | Room 339
More Details
