Norms Can Be More Powerful Than Law
New Data on Kids and Social Media
1
I am going to be presenting data from Australia, where a social media ban is not working. In particular, norms among teenagers in general, and FOMO in particular, are effectively counteracting the social media ban. What is required is a tipping point, and Australia isn’t close.
But I am not going to tell you about that data for a little bit. Might you stay with me?
2
When people do not steal, is it because of norms, or because of law? The answer is both, of course. But for many people, norms are a lot more powerful than law.
Cigarette use has declined dramatically in the United States. Changing laws have certainly played a big role. But so have changing norms. If you smoke in many places, you look like a big loser. Decades ago, smokers looked like big winners. Shifting law and shifting norms have marched hand-in-hand.
Law has an expressive function. It can affect norms. Sometimes law fuels changes in norms. But sometimes law runs up against norms. What happens then?
3
People buy some things that they wish did not exist. You might get a tattoo, contingent on the existence of tattoos, but you might not be glad about the existence of tattoos. You might buy a tie, and wear a tie, given the existence of ties, but you might deplore the existence of ties. You might wear high heels, given the existence of high heels, but you might wish that there were no high heels.
I call these goods “Barbies”: goods that people pay for (with money or time), but that they would wish out of existence, if only they could. See https://www.cambridge.org/core/journals/behavioural-public-policy/article/goods-that-people-buy-but-wish-did-not-exist/96796CDA397016B3FF839212BE75886D
(Apologies to Barbie enthusiasts.)
4
Along with an all-star team led by the great Leonardo Bursztyn, I have been lucky enough to be a coauthor of a paper on the effects of Australia’s effort to stop or reduce social media use by young people under the age of 16.
(You can find a draft of the paper here: https://www.nber.org/papers/w35162; you can find an oped here: https://www.nytimes.com/2026/05/08/opinion/social-media-teens-phones-australia-solutions.html?searchResultPosition=1)
Our simplest finding is that the effort isn’t really working. We surveyed 835 Australian teenagers four months after the ban took effect, and find that only about one in four 14–15-year-olds comply.
We also find:
Most banned teens believe their peers are still using banned platforms, and most non-compliers point to social forces—friends still on the platforms and the fear of missing out—as the reasons why they continue.
Most teenagers describe circumvention as easy: false birth dates, a parent’s account, a sibling’s login, a VPN. Any one of these works.
About 78 percent of the teenagers believe there are no personal consequences for violating the ban. (They are right.)
Compliers find it harder to keep up with friends and report feeling more bored.
When asked whether peers who comply with the ban are more popular or less popular than those who do not, 45 percent said less popular. Only 4 percent said more (!).
Asked about the share of peers required for them to comply themselves, respondents give a mean stated threshold that always lies around three-quarters, well above observed compliance (27%).
5
The simplest conclusion is that Australia is witnessing a contest between social norms and law, and social norms are winning.
A more precise statement would look like this. Australian’s ban imposes no punishment on teenagers, and to that extent, the cost of noncompliance is zero. At the same time, the cost of compliance is well above zero, because noncompliers are excluded from some kind of network, and face the equivalent of sanctions. They give a signal to their peer group, one that marks them as losers.
What is needed is a tipping point — if about 70 percent of 14 or 15 years old stay off, the tipping point will be reached. (Actually the percent is a bit higher than that.)
To get more teenagers to comply, the cost of noncompliance would have to increase, or the cost of compliance would have to decrease. Australia could tweak its policy to produce those effects. For example, it could work on social norms and incentives directly. It could also work with social media companies to do more to make its policy work. (A lot could be said on these topics, and we say some of it in the NBER paper.)
6
Sometimes a law is ineffective because it runs up against norms. More specifically, it might run up against FOMO. When people buy products or engage in activities that they wish did not exist, there is a real opportunity for desirable change - but the structure of the problem can make change pretty hard to engineer. The comparison between smoking and social media use offers a host of lessons.