What would happen, I wonder, were we to rescind the smoking ban as Nigel Farage wants? My guess is not much. Most restaurants would keep the existing rules. Some pubs might set aside a room for smokers. Casinos, comedy clubs and jazz clubs might revert to the status quo ante. But would we return to a time where people routinely smoked everywhere? Unlikely. People have had the chance to experience a new version of normal, and in large part they prefer it.

You wouldn’t expect me to say this, but I think the legislation has to be considered good precisely because, even if it were abolished, much of the behaviour it created would stick. The same goes for seatbelt legislation. The law was normative — it served to create new habits and conventions which then became self-sustaining.

Good laws can make a habit easier to adopt by making it universal (the Greek word for ‘law’ — nomos — also means ‘custom’ or ‘social norm’. And even obviously sensible behaviours can be hard to adopt when they are abnormal. The only time I never wear a seatbelt is in the back of a London taxi: this is because buckling up in a cab is counter to the norm. Yet if there is one place you should wear seatbelts, it is in a black cab, where any sudden deceleration will hurl you five feet forwards into the glass partition.

In the same way, many norms work only when everyone conforms to them. Think of the white lines in a car park. Though purist Libertarians might claim these interfere with your right to park at the diagonal, everyone generally obeys them. This conformity matters, since if only one person parks across two bays, the entire system breaks down. Often such ‘rules’ are enforced more by social pressure than law.

An interesting instance where a useful norm has been established in one country but has failed to take hold elsewhere is the case of the German Reißverschlussverfahren, or ‘zipping directive’. This is a practice established in Germany where, at a contraflow or anywhere else where two traffic lanes converge into one, car drivers queue in parallel before taking strict turns to merge at a single defined point. This significantly reduces delays and accidents. Yet the Dutch have tried to introduce it and repeatedly failed. No one is quite sure why — especially since Dutch drivers have no problem adopting the habit in Germany. I have a hunch that inculcating this habit in Britain might be harder still: Brits have a strong cultural urge to merge lanes early to avoid the appearance of queue-jumping (which ranks only slightly below paedophilia in our pantheon of shame).

I learned about the Reißverschlussverfahren from reading Complexity and the Art of Public Policy by David Colander and Roland Kupers. They suggest that complexity theory and agent-based models — which take account of things like norms and culture in a way that standard economic models don’t — should be used to break the deadlock between the ‘market fundamentalist’ and ‘government control’ narratives in politics.

The interesting thing about a complexity-based approach is that, while it sounds geeky, it is actually creatively liberating. A complex model of the world is not only more accurate, it is also more open-minded about where it draws inspiration for its ideas — not only from standard economics, but from everywhere: the humanities, evolutionary biology, even traffic management.

The most commonly voiced complaint about politics is that ‘politicians sound like they don’t live in the real world’. Complexity theorists would say there is a good reason for this: they don’t.