When Product Markets Become Collective Traps: The Case of Social Media
Individuals might experience negative utility from not consuming a popular product. For example, being inactive on social media can lead to social exclusion or not owning luxury brands can be associated with having a low social status. We show that, in the presence of such spillovers to non-users, standard measures that take aggregate consumption as given fail to appropriately capture welfare. We propose a new methodology to measure welfare that accounts for these consumption spillovers, which we apply to estimate the consumer surplus of two popular social media platforms, TikTok and Instagram. In large-scale, incentivized experiments with college students, we show that, while the standard welfare measure suggests a large and positive surplus, our measure accounting for consumption spillovers indicates a negative surplus, with a large share of active users deriving negative utility. We also shed light on the drivers of consumption spillovers to non-users in the case of social media and show that, in this setting, the “fear of missing out” plays an important role. Our framework and estimates highlight the possibility of product market traps, where large shares of consumers are trapped in an inefficient equilibrium and would prefer the product not to exist.
We thank Peter Andre, Luca Braghieri, Simon Cordes, Joshua Dean, Sarah Eichmeyer, Eliana La Ferrara, Matthew Gentzkow, Arkadev Ghosh, Alex Imas, Ro’ee Levy, Matthew Lowe, Shakked Noy, Ricardo Perez-Truglia, Aakaash Rao, and Gautam Rao, as well as many seminar participants for useful comments. Caroline Boone, Pietro Ducco, Paul Grass, Anastasiya Nebolsina, Ewan Rawcliffe, Christoph Ziegler, and especially Beyza Gulmezoglu provided excellent research assistance. Roth acknowledges funding from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy EXC 2126/1-390838866. The research described in this article was approved by the University of Chicago Social and Behavioral Sciences Institutional Review Board and pre-registered in AsPredicted (#137878, #142247, and #144630). The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.