Yican Liu / 刘怡灿

Technological change can transform economies, generating both opportunities and disruptions that carry important political and policy consequences. Yet little is known about how people respond politically to new technologies at an early stage, when their eventual impacts remain uncertain. We study this question in the context of generative artificial intelligence (GenAI) through randomized online experiments in five Latin American countries with a total of 12,000 participants. Participants are asked to perform a series of office-related tasks with or without access to a GenAI-powered assistant and then complete a survey on their beliefs about GenAI's potential influence and their preferences across a range of policy areas. To identify how these policy preferences are shaped by the beliefs individuals form, we implement an additional information treatment that experimentally manipulates perceptions of GenAI's potential impacts. Overall, the project seeks to shed light on how people respond to novel technologies whose effects are uncertain and for which they have no prior experience—particularly in developing countries that may be more vulnerable to such disruptions.

We study how digital currency adoption through the rollout of crypto ATMs affects household finance and the real economy. Using a unique hand-collected dataset, we document a plausibly causal link between staggered ATM installation and higher local crypto market participation, broader equity market participation, and greater business creation, but also higher cybercrime and financial fraud. The effects are strongest in areas with limited banking access, suggesting that crypto ATMs act as shadow intermediaries for unbanked and low-income households. Overall, cryptocurrency adoption beyond FinTech-savvy users has significant real effects, involving tradeoffs among financial inclusion, privacy, and criminal activities.

As individual investors increasingly rely on smartphones for their daily lives, we construct a measure of real-time investor sentiment based on global amusement app downloads. We find that a one-standard-deviation increase in amusement app downloads over the past four weeks predicts a 49 basis point increase in stock market returns over the following four weeks. We also observe a 33 basis point reversal in stock market returns in another four weeks. Amusement app downloads are positively related to subsequent stock trading app downloads, suggesting that investor sentiment drives new stock market participation. Overall, we show that the digital footprint in app downloads predicts global market returns via investor sentiment.

We quantify the value of consumer data and privacy preferences using stock and option market reactions to major data protection regulations. Standard event-study methods substantially understate negative valuation effects on app-intensive firms. Option-based approaches estimate average abnormal returns of about 7 percent, corresponding to market capitalization losses of $0.20 trillion for GDPR, $0.32 trillion for CCPA, and $0.40 trillion for DSLC. We further exploit changes in mobile-app privacy permissions to value specific digital footprints, finding that fintech payment permissions account for 1.3 percent of firm value. Together, this evidence underscores the importance of consumer data for technology firm valuation.