The Nudge, and What It Can Do

Take two European countries with similar cultures, similar religious profiles and near-identical survey responses about organ donation. In one, registered consent runs at around fifteen per cent. In the other, around ninety. The variable that tracks the difference is not attitude, campaigning or law in any strong sense. It is which box is ticked by default.

Nothing is forbidden in either system. Anybody can choose either way, at no cost, in a moment. And almost everybody goes with whatever was already set.

Defaults are powerful because three mechanisms act at once. Changing requires action, and action has a cost however trivial. The default reads as an implied recommendation; what a sensible person would choose. And once in place it becomes the reference point, so alternatives start to look like losses. Any one of these would produce an effect. Together they produce differences of fifty or sixty percentage points, which is larger than almost anything else in applied psychology.

Thaler and Sunstein built a policy programme on this, under the name libertarian paternalism. The argument behind the deliberately provocative phrase is that there is no such thing as a neutral design: somebody must decide what appears first, what is pre-selected, how options are worded, and that decision will affect behaviour whatever it is. Given that, they say, it may as well be made deliberately and in the chooser's interest. The standard reply is that choosing the arrangement is an exercise of power however many options remain formally available.

Two corrections have arrived since the book, and both matter. The first concerns size. Defaults are large and reliable. Most other nudges (reminders, reframing, social comparison messages) produce small effects, and the published average has fallen substantially as pre-registered trials and publication-bias corrections have accumulated. The literature followed the same arc as several others from the same decade: striking early results, enthusiastic policy adoption, downward revision once the file drawer was accounted for.

The second criticism is more interesting than the paternalism objection and gets far less attention. Nudges are cheap. Their effects are usually modest. That combination makes them extremely attractive to an institution facing a problem whose real solution is a tax, a regulation, or a large capital programme. Using a leaflet where the problem needs a tax is not a failure of behavioural science; it is a use of it, and an entirely predictable one.

Pension auto-enrolment is the counter-example and the strongest case for the whole approach: participation jumped by tens of points wherever it was introduced, dwarfing decades of information campaigns about saving. It is also, slightly awkwardly for the framing, a law. The intervention that worked was legislation requiring employers to set a particular default, which is not quite how nudging is usually sold.