Would Users Lose if Meta Sells Instagram?

In a new working paper released by the National Bureau of Economic Research, Justin Katz, a PhD student in business economics at Harvard’s Kenneth C. Griffin Graduate School of Arts and Sciences, and Hunt Allcott, professor in the Doerr School of Sustainability at Stanford University and a senior fellow at the Stanford Institute for Economic Policy Research, explore the possible consequences of the Federal Trade Commission’s (FTC) antitrust lawsuit against social media giant Meta—including the sale of its Instagram application. Katz and Allcott find that, while there may be good reasons to break up the company, the results would likely be good for advertisers and bad for users. 

Image

Justin Katz is a PhD student in business economics at the Harvard Kenneth C. Griffin Graduate School of Arts and Sciences.

/

Harvard Business School

Can you set the table for us a bit? What is the Federal Trade Commission’s issue with Meta regarding Instagram? 

The FTC sued Meta in 2020, alleging that it had illegally monopolized the market for personal social networking. The FTC defined that market as consisting of Facebook, Instagram, and Snapchat. One of the proposed remedies in the lawsuit was for Meta to unwind its 2012 acquisition of Instagram. 

There are a number of questions at play in the case, but the big, important ones include the claim that Meta’s market control increases the number of ads users see, leads to reduced investment in the user experience, and discourages potential new entrants from breaking in. There are also concerns about reduced incentives to invest in privacy protections and other aspects of the platform. 

What’s also interesting is that this case seems to be bipartisan. The investigation began under the first Trump administration, continued under Biden, and the trial proceeded despite Meta’s efforts during the more recent Trump administration. It appears to be a concern shared across party lines within the FTC. 

Which of the FTC’s concerns does your study specifically address? 

Many of the FTC’s concerns are hard to measure quantitatively. We focused on the most straightforward one: what would happen to ad load if Instagram were spun off. 

Just to clarify, ad load refers to the number of advertisements a user sees in a given amount of time—say, per hour. Meta makes money by showing targeted ads to users and selling those placements to advertisers, so ad load is a key outcome. Our study leaves aside more dynamic questions—like whether new competitors would enter the market—or qualitative issues like privacy. 

Your study projects what would happen if Meta spun off Instagram. How did you go about doing that? 

Ideally, you’d want to run the experiment multiple times and observe what happens on average—but obviously, that’s not feasible. 

Instead, we started by theoretically asking: What are the key incentives Meta would face if Facebook and Instagram were either jointly owned or forced to compete as separate companies? We then identified the key data points needed to assess the strength of each incentive and used those data to quantify how important these different forces are. That’s the high-level approach. 

In practice, this meant building a formal model of the industry to reveal what incentives each platform faces under joint or separate ownership. We focused on three main incentives: 

• User competition: If Facebook and Instagram are separate, they may need to compete for users who are active on both platforms.
• Advertiser competition: They may also compete for advertisers who want to reach the same users across platforms and will pay the lowest possible price to do that.
• Ad duplication: Advertisers generally don’t want to show the same ad to the same person too many times. One advantage of joint ownership is that Meta can coordinate across platforms to limit those duplicate impressions. 

So, if Facebook and Instagram were separated, there’s the potential for a less efficient ad delivery process, with more duplicated impressions to the same users. 

How did you measure user behavior and response? 

We used two existing experiments. The first, from Meta itself, asked how much users care about ads. About 0.5 percent of users have been randomly assigned to never see ads on the platform, and that’s been ongoing for around a decade. Last summer, Meta researchers released results showing that these users only spend about 9 percent more time on the platform than regular users. So, while people may not love ads, their presence doesn’t appear to significantly affect time spent on the platform. 

The second experiment was part of the 2020 Facebook and Instagram Election Study, which primarily looked at political outcomes. Participants were paid to stop using either Facebook or Instagram, and researchers then observed where their time went. Interestingly, the substitution was pretty diffuse. Facebook users who left didn’t shift heavily to Instagram, and vice versa. Instead, they went primarily to news apps and web browsers, although there's some evidence that deactivated Instagram users went to Youtube and TikTok. That suggests Facebook and Instagram aren’t close substitutes for most users. 

And what did you learn on the advertiser side? 

We used data on current ad prices and quantities to estimate how price-sensitive advertisers are. 

Then, to measure inefficiencies from ad duplication, we ran a new experiment on Meta. We randomized whether a user saw the same ad campaign once or multiple times and measured how that affected click-through rates. Duplicated campaigns had about a 28 percent lower click-through rate than non-duplicated ones. So, it’s a pretty significant source of inefficiency that advertisers and platforms care about. 

Image

Katz and Allcott's research showed that duplicated ad campaigns on Meta had about a 28 percent lower click-through rate than non-duplicated ones.

/

Katz and Allcott (2025)

So, paint a picture of how things would be different for an Instagram not owned by Meta. 

Sure. Just a caveat: when we talk about “user experience,” the only dimension we can really measure is ad load. There might be other aspects that users care deeply about, which could change, but we don’t study those. 

If you separated Facebook and Instagram, they wouldn’t really be competing for users. But they would compete aggressively for advertisers. And you’d lose efficiency due to duplicated ads being shown to the same people across platforms. 

Because of that advertiser competition, we estimate ad prices would fall. But for prices to fall, ad quantities would need to increase—hence, a higher ad load. Right now, the average Instagram user sees about 60 ads per hour. We estimate that would increase to around 85 ads per hour if the platform were independent. So that’s a shift from about one ad per minute to something notably higher. 

That’s a significant jump. But when you said users don’t care that much about ads, did you mean in terms of how they feel about them—or just that it doesn’t affect their platform use much? 

It’s more the latter. I’m referring to how ad exposure affects their behavior, specifically how much time they spend on the platform. 

So when I say users “don’t care,” I mean that their behavior is relatively inelastic. If seeing ads causes a 9 percent drop in time spent, that’s a measurable response—but it’s not dramatic, especially in response to a big change like going from 60 ads an hour to zero. In contrast, advertisers are highly sensitive to price. That’s why, from a revenue standpoint, the incentive for competition is stronger on the advertiser side than the user side. 

So, bottom line: Are you saying Meta is not an illegal monopoly and should be left alone? Or are there still valid reasons to break it up? 

I think the FTC’s case rests on three pillars. One is that Meta’s market power leads to increased ad load. Our findings don’t support that—if anything, joint ownership seems to result in fewer ads due to greater efficiency. But the other two arguments—reduced competition from new entrants and fewer incentives to invest in things like privacy protections—are still potentially valid. We don’t address those directly in our paper, so it’s up to the FTC and the courts to decide if those are strong enough reasons to justify a breakup. 

One last point: we often think of users as the only consumers in these markets, but advertisers are consumers too. Our estimates suggest the benefits to advertisers and the costs to users would roughly cancel out. So it’s not obvious that there’s a clear harm—or a clear benefit—to consumers overall from a breakup.