# Australia’s Social Media Ban Did Not Change Norms Adherence to Australia’s social media ban was low and declined over time, suggesting that restrictions can be easily circumvented and may do little to reduce use when teenagers’ peers remain online. September 23, 2026 • Research Briefs in Economic Policy No. 502 By Leonardo Bursztyn, Aaron Leonard, Angela L. Duckworth, Rafael Jiménez-Durán, Filip Milojević, Christopher Roth, & Cass R. Sunstein In December 2025, Australia became the first country to ban under-16s from holding accounts on major social media platforms. The ban covers popular platforms, including Facebook, Instagram, Kick, Reddit, Snapchat, Threads, TikTok, Twitch, X, and YouTube. Platforms can face substantial civil penalties of up to 49.5 million Australian dollars if they fail to take reasonable steps to prevent underage users from creating accounts, but the ban imposes no penalties on individual users or their parents. Similar policies are now being considered in more than a dozen countries and numerous US states. These policies address widespread concern that heavy social media use may harm adolescent mental health. Yet unlike tobacco or alcohol, where enforcement operates at a physical point of sale, social media can be reached from any device at any time. Such bans are easy to circumvent, and being off a platform is individually costly when one’s peers remain on it. An easily circumvented ban, therefore, may succeed only if adherence crosses a tipping point beyond which quitting becomes self-reinforcing. Whether that point is reached depends on the share of peers each teenager requires to have quit social media before they themselves quit. When observed adherence falls short of most teenagers’ thresholds, adherence is predicted to decrease rather than increase. Four months after the law took effect, we surveyed 507 Australian teenagers aged 14–18 and 300 US teenagers facing no comparable policy. We also surveyed an additional 464 Australian teenagers about peer behavior and social norms. To understand whether Australia crossed the tipping point, we asked what share of their peers they believed were adhering and how many peers would need to adhere before they would adhere themselves. Two months later, we returned to the field and surveyed 792 teenagers, 423 of whom we surveyed a second time. Our findings reveal that adherence was low. In our initial survey, nearly 64 percent of Australian 14- and 15-year-olds reported using at least one banned platform during the previous week, compared with 87 percent of Australians aged 16–18. We used the older group to estimate how many younger teenagers would have used social media without the ban, and only about 27 percent of affected teenagers appeared to adhere to it. The ban imposed few practical costs on teenagers who wanted to continue using social media: 75 percent described circumventing the restrictions as easy or very easy, and 60 percent said their accounts had not been removed. Common workarounds included entering a false birth date, lying about age during verification, using a parent’s or older sibling’s account, and connecting through a virtual private network. Of nonadherents, 22 percent said an adult helped them create an account after their original account was deactivated. Social pressures also encouraged continued use. Among teenagers still using banned platforms, 52 percent gave a social reason, most often that their friends were still on the platforms or that they feared missing out. Use was substantially higher among teenagers who believed that more of their closest friends were still online. Adherents reported less pressure to check their phones, but they also found it harder to keep up with friends and felt more bored. These social pressures help explain why the existing adherence rate may not last. On average, affected teenagers said that roughly two-thirds of their peers would have to adhere before they would adhere themselves. Yet they believed that only about 30 percent of their peers were adhering, close to our estimated rate of 27 percent. Even current adherents looked fragile: They overestimated peer adherence at 48 percent and named a required threshold of 62 percent, and 44 percent said they would go back to using banned platforms if only a third of their peers had stopped. Based on the distribution of these stated thresholds, our theoretical model places the only stable level of adherence at approximately 17 percent. Unless enforcement or social norms changed, we therefore expected adherence to decrease rather than increase. We were able to test that prediction. When we surveyed Australian teenagers again in June 2026, six months after the ban took effect, 92 percent of 14- and 15-year-olds reported using a banned platform in the previous week, up from nearly 64 percent in April. Estimated adherence fell from about 27 percent to about 6 percent. Among the teenagers we interviewed in both surveys, adherence fell from 25 percent to 7 percent, and half said they had stopped using the platforms at some point and then resumed once they saw that their friends had not stopped. Enforcement tightened over these two months rather than loosened, so the collapse in adherence does not appear to reflect platforms giving up. Who adheres may matter as much as how many adhere. Of respondents, 45 percent viewed adherents as less popular than nonadherents, 51 percent considered the groups equally popular, and only 4 percent viewed adherents as more popular. If socially influential teenagers remain online, they may keep the platforms attractive and make nonuse socially costly. The law communicates that teenagers should stay off social media, but the behavior teenagers observe continues to communicate that most of their peers are online. Our evidence covers only the first six months of the ban, and teenagers who are too young to have ever held an account may behave differently as they age into the restriction. The evidence is also largely descriptive and based on self-reported behavior and hypothetical adherence thresholds. Although the trend could reverse, the decline does not appear to reflect teenagers misjudging their peers’ behavior. They accurately estimated how many of their peers adhered, yet adherence still fell. Ultimately, our findings suggest that other policies might be more effective at reducing social media usage among teens, either as an alternative or a complement to a ban like the Australian one. Possible approaches include coordinated time limits (which teenagers themselves prefer) instead of a complete ban, campaigns that make nonuse visible and socially rewarded, incentives for reduced social media use, and organized activities that replace time and interaction lost online. Policymakers should also consider alternative policy designs, such as age-tiering of platforms, rather than a ban. **Note** This research brief is based on Leonardo Bursztyn et al., “[Tipping Points and Australia’s Social Media Ban](https://www.nber.org/papers/w35162),” National Bureau of Economic Research Working Paper no. 35162, July 2026. ##### About the Authors ##### Leonardo Bursztyn University of Chicago ##### Aaron Leonard University of Chicago ##### Angela L. Duckworth University of Pennsylvania ##### Rafael Jiménez-Durán Bocconi University ##### Filip Milojević Massachusetts Institute of Technology ##### Christopher Roth University of Cologne ##### Cass R. Sunstein Harvard Law School [![Creative Commons License](/build/cato_2020/images/creative-commons.svg)](http://creativecommons.org/licenses/by-nc-sa/4.0/) This work is licensed under a [Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License](https://creativecommons.org/licenses/by-nc-sa/4.0/).