Economic Big Data and Machine Learning: Applications and Implications
Paper Session
Sunday, Jan. 3, 2021 3:45 PM - 5:45 PM (EST)
- Chair: Will Cong, Cornell University
Data Privacy and Temptation
Abstract
This paper analyzes how different data-sharing schemes of a digital platform may affect consumer surplus and social surplus when a fraction of the consumers have weak self-control and suffers from targeted advertising of temptation goods, such as gambling and video games. While sharing consumer data with firms improves the efficiency of matching consumers with normal consumption goods, it also exposes weak-willed consumers to temptation goods. Despite the seeming appeal of the opt-in policy of allowing each consumer to opt in or out of data sharing, our analysis shows that this policy may not be effective in protecting severely tempted consumers. When other consumers, motivated by the improved access to normal goods, choose to share their data, their opt-in reduces the anonymity of the weak-willed consumers who choose to opt out. To alleviate this externality, privacy protection regulation needs to limit the bundling of the consumer authorization to share data with normal good and temptation good sellers.Persuading Investors: A Video-Based Study
Abstract
Persuasive communication is a function of not only content but also delivery, e.g., facial expressions, tone of voices and diction. This paper examines the persuasiveness of delivery in start-up pitches. Using machine learning (ML) algorithms to process full pitch videos, we quantify persuasion in visual, vocal, and verbal dimensions. Positive (i.e., passionate, warm) pitches increase funding probability. Yet conditional on funding, high-positivity startups underperform. Women are more heavily judged on pitch delivery when evaluating single-gender teams, but are neglected when co-pitching with men in mixed-gender teams. Using an experiment, we show that persuasion delivery works mainly through leading investors form inaccurate beliefs.Who Benefits from Robo-Advising? Evidence from Machine Learning
Abstract
We study the effects of a large U.S. hybrid robo-adviser on the portfolios of previously self- directed investors. Across all investors, robo-advising reduces investors’ holdings in money market mutual funds and increases bond holdings. It also reduces idiosyncratic risk by lowering the holdings of individual stocks and US and international active mutual funds and raising exposure to low-cost indexed mutual funds. It further eliminates home bias by significantly increasing international equity and fixed income diversification. Finally—over our sample period—it increases investors’ overall risk-adjusted performance, mainly by lowering investors’ portfolio risk. We use a machine learning algorithm, known as Boosted Regression Trees (BRT), to explain the cross-sectional variation in the effects of advice on portfolio allocations and performance. Investors who benefit from advice are those with little self-directed investment experience on the platform, those with prior high cash holdings, and those with high trading volume before adopting advice. Individuals invested in high-fee active mutual funds also display significant performance gains. Finally, we study the determinants of investors’ sign-up and attrition. Investors who benefit more from robo-advising are also more likely to sign-up and less likely to quit the service.Discussant(s)
Serhiy Kozak
,
University of Maryland
Dirk Bergemann
,
Yale University
Bradley Hendricks
,
University of North Carolina-Chapel Hill
Michael Reher
,
University of California-San Diego
JEL Classifications
- G0 - General