The Impact of Social Media on Music Demand: Evidence from a Quasi-Natural Experiment
(A2026-134016)

AUTHORS

Nils Wl�mert, Vienna University of Economics and Business; Dominik Papies, University of T�bingen; Daniel Winkler, Vienna University of Economics and Business; Christian Hotz-Behofsits, WU Vienna; Jūra Liaukonytė, Cornell University

ABSTRACT

Social media platforms such as TikTok have become central to how music is discovered and consumed. We study the February 2024 removal of Universal Music Group’s (UMG) entire catalog from TikTok as a quasi-natural experiment to identify how disruptions to short-form video exposure affect downstream demand on monetized streaming platforms. Using global data covering over 200,000 songs and more than one trillion streams, we find that treatment effects vary sharply across the virality distribution. In log (percentage) terms, the average effect of TikTok removal on Spotify streams is slightly positive. However, level estimates reveal substantial absolute declines, driven entirely by the small set of highly viral songs that account for most streaming volume. This heterogeneity reconciles the apparent discrepancy between percentage and absolute effects: lower-virality songs exhibit modest relative gains, but losses for top-decile viral tracks dominate in the aggregate. Our findings highlight TikTok’s role in sustaining demand for high-exposure songs and underscore the economic trade-offs underlying licensing disputes between music labels and digital platforms.