Nudge

In brief

A nudge is any feature of the way choices are presented that steers people toward a particular option without forbidding any of the others or changing the incentives much. Putting fruit at eye level in a cafeteria is a nudge; banning the candy is not. The idea was named and defended by the economist Richard Thaler (born 1945) and the legal scholar Cass Sunstein (born 1954) in their 2008 book Nudge. Their central claim is that someone always has to decide how options are arranged, and no arrangement is neutral, so the arranger, the choice architect, might as well arrange them to help people rather than harm them. They call this libertarian paternalism: paternalist because it tries to make people better off by their own standards, libertarian because it leaves them free to go their own way.

The full treatment

The problem it answers

Classical economics assumed people choose well when left alone, so the only question was whether to interfere. Behavioral research broke that assumption. People are influenced by defaults, by the order options appear in, by how a choice is framed, and by inertia. Given this, Thaler and Sunstein argue, the old debate is miscast. There is no option to simply leave people alone, because the way a choice is set up always tilts it one way or another. A form has to have some default. A cafeteria has to lay food out in some order. If arrangement is unavoidable and never neutral, then refusing to think about it does not preserve freedom; it just hands the outcome to accident or to whoever designed the form for other reasons.

How it works

The tools of the nudge are the same forces that behavioral research uncovered, turned to a purpose. The most powerful is the default: the option that takes effect if a person does nothing. Because people are prone to inertia and to loss aversion (they feel the cost of switching more than the gain), defaults are extraordinarily sticky. Set the default to the choice most people would want on reflection, and most people end up there, while anyone who wants otherwise can still opt out with a signature. Framing is a second tool: the same fact stated as a loss or a gain, or a rate of savings shown as a percentage or a figure, changes what people pick. Simplification, timely reminders, and making the better option easier all count. The test Thaler and Sunstein set is strict: to be a nudge and not a shove, an intervention must be easy and cheap to avoid. A tax on soda is not a nudge; rearranging the shelf is.

The key examples

Two cases carry most of the argument. The first is organ donation. In countries where donation is opt-in, a person must actively register, and consent rates are low. In countries where it is opt-out, everyone is a donor unless they register otherwise. Eric Johnson and Daniel Goldstein, in a 2003 paper in Science, showed that this single design choice tracks enormous gaps in effective consent, with opt-out countries reaching rates near ninety percent. Nobody is forced; the default does the work.

The second is retirement saving. Many people want to save but never get around to enrolling. Automatic enrollment flips the default: employees are signed up unless they opt out, and participation jumps. Thaler and Shlomo Benartzi went further with Save More Tomorrow, described in a 2004 paper, which invites workers to commit in advance to raising their contribution rate when their next pay rise arrives, so the increase never feels like a loss from current income. Enrollment and saving both climbed sharply.

Distinctions that matter

Libertarian paternalism sits between two familiar positions and borrows from each. It is not hard paternalism, which restricts or bans, because the escape hatch is always open. It is not pure laissez-faire either, because it accepts that the state and firms shape choices whether they mean to or not, and asks them to shape them well. Thaler and Sunstein also insist a nudge should serve the chooser's own welfare, judged as far as possible by what that person would want with full attention, not the architect's interests. That is the line, often blurred in practice, between a nudge and a trick.

Lineage

The nudge descends from the behavioral economics that Daniel Kahneman (1934 to 2024) and Amos Tversky (1937 to 1996) built from the 1970s, and in particular from loss aversion and the stickiness of the status quo, which explain why defaults bite. Thaler, who won the Nobel Memorial Prize in Economics in 2017, was the economist who carried those psychological findings into policy. The term libertarian paternalism predates the book: Thaler and Sunstein introduced it in a 2003 article. The deeper backdrop is the liberal tradition's long worry about when interference with a free adult is justified, framed most sharply by John Stuart Mill in On Liberty (1859), whose harm principle held that a person's own good is not sufficient warrant for coercion. The nudge is in part an answer to Mill: it claims to improve outcomes without the coercion Mill forbade.

The strongest case for it

The core argument is that the alternative to a good nudge is not no nudge but a careless one. Since choices must be arranged, arranging them to help is close to a free lunch: it costs little, forces no one, and leaves every option on the table. Defaults in particular achieve what exhortation and education rarely do, moving behavior at scale without spending political capital on bans or subsidies. And because the opt-out is preserved, the intervention respects autonomy in the way that matters: anyone who has a considered preference can act on it at almost no cost. On this view libertarian paternalism is the humane middle path between a state that does nothing while people predictably harm themselves and a state that orders them about.

The strongest case against it

The objections are serious and come from several directions. The psychologist Gerd Gigerenzer (born 1947) argues that the nudge program treats human irrationality as fixed and works around it, when the better response is to teach people to reason well, through statistical literacy and better-presented risk, so they need no minder. Nudging, he warns, keeps people dependent.

A liberty-based objection, pressed by the legal scholar Jeremy Waldron (born 1953) among others, targets the manipulation. Many nudges work precisely because they bypass conscious deliberation: they exploit inertia and framing rather than persuade. Even when the outcome is good, treating a citizen as an object to be steered, rather than a mind to be reasoned with, fails to respect them as an agent. The philosophers Daniel Hausman and Brynn Welch made this case in a 2010 paper, distinguishing nudges that inform from nudges that push people around beneath their awareness. The economist Riccardo Rebonato, in Taking Liberties (2012), argued that libertarian paternalism smuggles the architect's values in under cover of the chooser's own welfare, since deciding what people would want on reflection is itself a contestable judgment.

There is also the who-guards-the-guardians worry: the same defaults that raise savings can be turned to extract signups, harvest data, or serve the seller, and firms nudge far more aggressively than governments. The line between a welfare-improving nudge and a self-serving dark pattern is thin, and nothing in the theory guarantees which side an architect lands on.

Where it stands now

The nudge moved from a book to an instrument of government within a few years. The United Kingdom set up the Behavioural Insights Team, widely called the Nudge Unit, in 2010, and the United States created a comparable team in 2014; dozens of countries and agencies have followed. A revised final edition of Nudge appeared in 2021. At the same time the evidence has been chastened: a wave of replication work through the late 2010s found that some nudges have small or fragile effects, and that defaults remain the reliable heavy lifter while lighter framing nudges are hit or miss. The result is a field that is more sober than it was, still widely used, and still arguing about the ethics. Defaults work; whether steering people below the level of deliberation is compatible with treating them as free is a question the debate has not closed.

Test yourself

Think of a default you never changed: a pension contribution, a privacy setting, an organ-donor box you left as it was. Ask whether that outcome reflects what you actually want, or only what someone else set for you. Then ask the harder question the whole debate turns on: if the default happened to match your interests, does it matter that you never really chose it?