Closing the Door After the Storm?
Ask any parent, teacher or family doctor and you will hear the same thing: something has gone wrong with how our children feel. More of them are anxious, more of them are low, and more of them struggle simply to get through the school gates in the morning. The worry is not new, but the pandemic turned a slow-burning concern into something that now feels like an emergency. And over the past year, governments around the world have reached for the same answer: keep children off social media.
As a health economist, I find this both encouraging and a little nerve-racking. Encouraging, because children's mental health has finally reached the top of the political agenda. Nerve-racking, because we are placing a very large policy bet on evidence that is still thin, and we are not yet setting ourselves up to learn whether that bet pays off.
What happened to children's mental health
Start with the numbers, because they are striking. In England, the NHS has followed the same group of children since 2017. That year, about one in nine children aged 7 to 16 had a probable mental disorder. By 2020 it was one in six. By 2023 it had reached one in five children aged 8 to 16. For older teenagers the climb was steeper still: among 17 to 19 year olds, the rate went from around 10 per cent in 2017 to more than 23 per cent in 2023. Eating disorders in that age group jumped from less than 1 per cent to roughly one in eight.
The United States tells a similar story, with an important twist, as the first chart shows.

Look closely and three things stand out. First, the rise began well before COVID. Sadness among American teenagers was already climbing in the early 2010s, roughly when smartphones and image-based apps became part of everyday teenage life. Second, the pandemic acted as an accelerator. Schools closed, friendships moved online, sport and clubs stopped, and many families faced job losses, illness and grief. For a while the phone was not just one window onto the world; for many children it was the only one. The peak came in 2021, when 42 per cent of high school students reported persistent sadness, rising to 57 per cent among girls.
Third, and this is the part that rarely makes the headlines, things have started to improve. By 2025 the American figure had fallen to 33 per cent, close to where it stood a decade ago, although girls are still far worse off than boys (43 per cent against 24 per cent). In England the rates stopped climbing after 2022 but remain stubbornly high. Notice the timing: this partial recovery began years before any country banned anything. That matters a great deal when we come to judge whether bans work.
A wave of bans
Australia went first. On 10 December 2025 it became the first country in the world to stop children under 16 from holding accounts on the big platforms, including Facebook, Instagram, TikTok, Snapchat, YouTube and Reddit. The legal duty falls on the companies, not on children or parents, with fines of up to A$49.5 million for those that fail to take reasonable steps, and the government now wants to double that. Within about five weeks, platforms had shut down around 4.7 million accounts.
Others have followed quickly. France's parliament approved a ban for under-15s in July 2026, blocking new accounts from September and closing existing ones by January 2027. In June, the UK government announced its own under-16 ban, expected by spring 2027, after a consultation in which most responding parents backed a legal minimum age. Denmark, Greece, Malaysia and Turkey have moved in the same direction, Norway is preparing legislation, and in September the European Commission put forward proposals for an EU-wide minimum age of 13. By one news agency's count, more than 20 countries have now proposed or introduced some form of restriction.
Is it working?
In July 2026, Australia's online safety regulator, eSafety, published the first results from its long-term study of the ban. The second chart summarises what it found three months in.

The share of under-16s with their own social media account fell by ten percentage points, from 52 to 42 per cent. That is not nothing. But the share using an age-restricted platform in any way barely moved, from 86 to 81 per cent, and daily use hardly changed. About half of the children who kept their accounts said nobody had ever checked their age. Others had profiles that already said they were 16, or face-scanning tools that guessed wrong. The study also found little change in time spent on sport, music, friends and family, which were the very things the ban was supposed to give back.
Ministers in Canberra have defended the law by pointing out that drinking age laws are not perfectly obeyed either, yet nobody suggests scrapping them. That is a fair point: laws can shift what feels normal long before they change behaviour. Still, three months of usage data tells us nothing yet about the outcome that really matters, which is whether Australian children are less anxious, less depressed or less likely to harm themselves.
Here is the uncomfortable truth. After more than a decade of research, scientists still disagree about how much social media causes poor mental health, rather than simply travelling alongside it. Large reviews keep finding that the link between total time online and children's wellbeing is real but small and inconsistent, and that it is hard to rule out the reverse story, in which children who are already struggling retreat into their phones. The clearer signal comes from how children use social media rather than how long: compulsive use that eats into sleep and schoolwork, harsh comparisons about appearance, bullying, and feeds that push harmful content towards children who are already vulnerable. Averages also hide a lot. For some groups, such as teenage girls, the risks seem larger. For others, including isolated or LGBTQ+ young people, online communities can be a lifeline.
How an economist sees it
There is one argument for bans that I find genuinely persuasive, and it is an economic one. Parents face a classic coordination problem. Most would happily delay their child's first Instagram or TikTok account, but no parent wants their child to be the only one left out of the group chat. When every family decides alone, almost every family gives in. A national rule changes the default for everyone at once and gives parents a simple line to use at the kitchen table: it's the law. That benefit is real even when enforcement is patchy.
But good policy also asks about costs and alternatives. Age checks raise privacy questions for adults as well as children. Teenagers pushed off the big, regulated platforms may drift towards smaller apps where harmful content is even harder to police. And a ban is cheap for governments, because the bill lands on tech companies and families. The danger is that it becomes a substitute for the harder and more expensive work of helping children who are already unwell. In England, NHS services are supporting more young people than ever, yet many still wait far too long for specialist care. Taking a phone away does not shorten a waiting list.
Finally, we have to measure honestly. Because countries are introducing bans at different times and at different ages, we have a rare natural experiment. Researchers can compare children just above and just below the age limit, or countries that act early with those that act later, and see what really changes. That only works if governments collect good data before and after, and make it public. Otherwise we risk claiming credit for an improvement that, as the American figures show, was already under way.
So, are social media bans the answer? They are probably part of one. They can reset social norms, take some pressure off parents and send a clear signal to the tech industry. On their own, though, they are a blunt tool aimed at a complicated problem. The pandemic showed us how quickly children's wellbeing can unravel when their world shrinks. Rebuilding it will take safer platform design, more support in schools, faster access to treatment and, above all, careful evaluation of what works. Our children deserve policies judged by their results, not just by their good intentions.
Sources
NHS England (2023). Mental Health of Children and Young People in England, 2023 – wave 4 follow-up to the 2017 survey.
Newlove-Delgado T. (2023). Using the MHCYP survey series to understand trends in child mental health. UK Data Service.
CDC. Youth Risk Behavior Survey: Data Summary & Trends Report 2013–2023; 2025 YRBS national report (September 2026).
eSafety Commissioner (July 2026). First results of the long-term evaluation of the social media minimum age, reported by Xinhua/Bernama and Al Jazeera.
AFP, AP and Reuters reporting on social media age restrictions in France, the UK, the EU and other countries (June–September 2026).
Punukollu M. et al. (2026). Social media use and mental health outcomes in secondary school students: a systematic review. BJPsych Open.




