Research
My research studies the creator economy and digital platforms, with a particular focus on how content creators monetize their work and how platform design and AI shape creator behavior.
Research Interests
Substantive: Creator economy, Platform economics, Generative AI
Methodological: Causal Inference, Machine Learning, Multimodal Data
Working Papers
Democratizing the Creator Economy: Monetizing Through Subscriptions
with Kai Zhu
Platforms increasingly provide direct monetization opportunities through creator–viewer subscriptions, contrasting with traditional reach-driven advertising models. We study the impact of the Twitch Affiliate Program (TAP), which substantially lowered entry barriers to paid channel subscriptions and micro-donations. Using a time-shifted difference-in-differences design, we estimate causal effects on creator production, strategy, and revenue. TAP significantly increases content supply, especially among smaller creators, whose streaming hours more than double. Effort dynamics exhibit goal-gradient patterns, peaking prior to qualification and stabilizing at a higher level afterward. Monetization access also professionalizes creator behavior, leading to more deliberate content strategies and improved game performance. Revenue estimates show that TAP provides meaningful supplementary income even for median creators, demonstrating that low-barrier subscription models can broaden participation and complement advertising-based monetization.
Work in Progress
Network Effects and Platform Competition: Insights from Online Dating Markets
with Kai Zhu, Qiaoni Shi and Sara Valentini
Network effects play a fundamental role in shaping platform markets, yet empirical evidence on their magnitude and heterogeneity remains limited. Using a comprehensive dataset on dating apps across 100 Chinese cities and leveraging the sudden suspension of a dominant app as a natural experiment, we address four key questions: (1) How does a regulatory removal affect user participation on the focal platform and its rivals? (2) What is the magnitude of local network effects revealed by this shock? (3) How do local market characteristics shape the strength of these effects? (4) How do heterogeneous network effects translate into market concentration? To answer these, we first employ a Difference-in-Differences design to quantify the shock's propagation through new user inflows and existing user engagement. Second, we estimate a structural demand model to recover network effect parameters and simulate counterfactual policy scenarios. We find that network effects are strongly positive but highly localized, varying significantly with city-level demographics. Counterfactual analyses further demonstrate that markets with stronger local network effects exhibit higher concentration and faster recovery from supply-side shocks. These findings underscore the importance of accounting for localized heterogeneity in platform strategy and antitrust analysis.