Unfriending Meta: How a Facebook-Instagram Breakup Could Affect Ads, Users, and Welfare

In their July 2025 working paper published by the National Bureau of Economic Research (NBER), economists Justin Katz of Harvard University and Hunt Allcott of Stanford University delve into the economic implications of digital media mergers, focusing specifically on the proposed separation of Facebook and Instagram. Drawing upon data from respected research institutions such as the Pew Research Center, the Stigler Center, and a series of randomized experiments, including the Facebook and Instagram Election Study (FIES), the authors evaluate how the merger or separation of digital platforms influences advertising loads, user behavior, and market efficiency. The study is particularly timely, as it engages directly with the U.S. Federal Trade Commission’s (FTC) antitrust case against Meta, which claims Meta's ownership of Instagram and WhatsApp constitutes monopolistic control over the “personal social networking” market. In response, Meta argues that it competes broadly with platforms like TikTok, YouTube, and LinkedIn, challenging the FTC’s narrow market framing. The paper’s blend of robust theory and rich empirical data makes it a landmark contribution to digital competition policy.

A New Model for Digital Ad Competition

At the heart of the paper is a novel economic model designed to capture the complexities of digital platform competition in a two-sided market. Users allocate time across platforms like Facebook and Instagram, while advertisers choose where to buy ad space based on expected click-through rates and prices. Importantly, platforms must decide the intensity of advertising, known as ad load, while balancing the risk that excessive ads may reduce user engagement. What sets this model apart is its incorporation of multi-homing behavior, in which users engage with multiple platforms simultaneously. This feature allows the authors to address an increasingly relevant inefficiency in the digital ad economy: ad duplication. When users see the same ad across different platforms, its effectiveness declines, reducing value for advertisers and creating a less efficient marketplace.

What the Data Say: Overlap, Elasticity, and Diversion

To populate and estimate their model, Katz and Allcott use five main empirical sources. First, data from Pew’s 2020 National Public Opinion Reference Survey shows that while 88% of Instagram users also use Facebook, only 56% of Facebook users use Instagram. This asymmetry means Instagram users are more likely to be exposed to ads on both platforms, amplifying duplication concerns if the platforms are separated. Second, findings from the 2020 Facebook and Instagram Election Study (FIES) reveal that users substitute very little between the two platforms when one is deactivated. Diversion ratios were only 5.4% from Facebook to Instagram and statistically insignificant in the other direction, contradicting the FTC's market definition and suggesting that the platforms do not function as close user substitutes.

Third, the Digital Addiction experiment reveals a high price elasticity of user time: when offered $2.50/hour to avoid social media, users reduced their time by over 40%, implying that a significant portion of social media engagement is of low personal value. Fourth, evidence from Brynjolfsson et al. (2024) shows that user time is largely unresponsive to ad load increases, suggesting that platforms cannot easily recover engagement by adjusting ad intensity. Finally, the authors conducted a new experiment to test diminishing returns in duplicated ad campaigns. Results showed that duplicated campaigns had 28% lower click-through rates than non-duplicated ones, highlighting the real costs of ad duplication when delivery coordination is lacking.

Simulating the Impact of a Breakup

Using these empirical foundations, the authors simulate what would happen if Facebook and Instagram were separated, both with and without coordinated ad delivery. In the merged scenario, Meta moderates ad load to maintain high ad prices and user satisfaction. In contrast, separation with coordinated ad delivery leads platforms to compete harder for advertisers, increasing ad loads, especially on Instagram, where more users are multi-homers. Ad prices fall and advertisers benefit, but users see more ads, leading to a slight drop in consumer surplus. The overall change in total welfare is marginal, just 0.1% lower than the merged equilibrium, indicating that such a breakup may be redistributive rather than efficiency-enhancing.

Duplication Risks in Uncoordinated Scenarios

The situation is more problematic under uncoordinated ad delivery, where each platform optimizes its ad targeting independently. In this scenario, the authors find a sharp rise in duplicated impressions, especially for Instagram users, leading to lower ad effectiveness and stronger incentives to increase ad loads to compensate. Instagram's ad load jumps by 46%, ad prices fall substantially, and advertisers gain more surplus. However, these benefits come at the cost of inefficiency: users see more redundant ads, and overall welfare drops by 1.2%. While advertisers enjoy short-term gains, the broader digital ecosystem becomes less efficient and less user-friendly.

A Measured Verdict on the FTC’s Case

Katz and Allcott conclude that the FTC’s proposed remedy of separating Facebook and Instagram is unlikely to improve market outcomes significantly. Rather than restoring competition in a meaningful way, such a move would mostly redistribute surplus from Meta and its users to advertisers, without yielding gains in total efficiency. The authors acknowledge that their model does not account for all effects, such as long-run innovation, privacy, or mental health costs of social media. Yet, even within its narrower scope, the analysis suggests that structural remedies like breakups may be less effective than commonly believed. This research ultimately offers regulators a nuanced framework grounded in economic theory and empirical data to guide decisions in the evolving digital marketplace.