What is nudge theory?
Nudge theory is four claims under one name. Only one of them can fail like an experiment, and it has shrunk under correction.
Answer to the question
What is nudge theory?
Nudge theory is the idea, set out by Richard Thaler and Cass Sunstein, that changing how options are presented can steer behaviour without banning choices or changing incentives much. It bundles claims about people, design, method and effect. The effect claim is the contested one: published averages shrink sharply after correction for publication bias.
Key takeaways
- Thaler and Sunstein define a nudge as a change in how options are presented that alters behaviour predictably, forbids nothing, leaves incentives largely intact and is easy and cheap to avoid; nudge theory bundles four claims, and only the effect claim can fail like an experiment.
- Mertens and colleagues (2022) pooled more than 200 studies and over 450 effect sizes and found a small to medium average effect, alongside moderate publication bias.
- After correction, Maier and colleagues (2022) reported a pooled effect of 0.04 and Szaszi and colleagues (2022) reported averages from minus 0.01 to 0.08.
- Stefano DellaVigna's 2022 analysis of 126 trials found 1.4 percentage points of extra take-up for two US nudge units against 8.7 in journals, with about 70 per cent of the gap attributed to selective publication.
In its final edition, Nudge carries a preface in which Richard Thaler and Cass Sunstein tell readers that the authors oppose presumed consent for organ donation, the policy many had assumed the book backed. That is a fair picture of nudge theory today: its authors have trimmed it, its critics have audited it, and trials far larger than the first studies have put it to the test. What is left is a narrower idea than the slogan that spread. The sections below take it from the definition, through what the authors took back, to the field trials and the corrections, and last to the four claims the phrase bundles together.
What Counts, and What Was Taken Back?
The authors define a nudge as any aspect of the choice architecture that alters behaviour in a predictable way without forbidding options or significantly changing economic incentives. To count, the intervention has to be easy and cheap to avoid. Fruit placed at eye level passes. A ban on junk food fails, and so do taxes, fines, subsidies and mandates, which is what separates nudge theory from ordinary regulation.
The default is the standard case. An employer enrols staff in a pension plan unless they decline, and anyone can still leave. Brigitte Madrian and Dennis Shea studied a large company that moved from opt-in to automatic enrolment and found that participation rose sharply. Nothing about the plan itself had changed, which is the point nudge theory makes about presentation. Defaults are also the member of the family that later reanalyses were most willing to except.
The cheap-exit condition carries the ethical argument, and the final edition says so. If leaving is hard, the authors write, the arrangement is not the liberty-preserving thing they set out to defend. They call it sludge and say its cost belongs in any assessment of the policy, a theme taken up in The Sludge That Keeps You Subscribed. On the charge of manipulation, they borrow a test from philosophy: an act is manipulative if it fails to respect a person's capacity for deliberation. Most nudges pass, they say, and they report studies finding that telling people about a nudge does not weaken it.
The authors have revised their own account in public. Their best-known intervention, the savings programme Thaler designed with Shlomo Benartzi, once bundled many favourable features. In the final edition they report that two of them, tying savings increases to pay rises and letting people decide now to join later, are attractive but not essential. They also concede that light-touch methods cannot solve climate change, and say that prices matter more.
The organ-donation chapter went further. Many readers had taken the two authors to support presumed consent, where everyone is a donor unless they opt out. The Final Edition rewrites the chapter. Their reasoning applies the method's own test, since a default is only liberty-preserving when the opt-out is real, and an opt-out that many people never learn about is not. That is why they now read the familiar comparison between countries as a measure of how defaults shape stated preferences and not as proof that presumed consent saves lives.
What Do the Trials and Corrections Show?
The largest test came from nudge units. Stefano DellaVigna and a co-author assembled 126 randomised trials covering 23 million people, every trial run by two US nudge units, and set them beside nudge trials published in academic journals. The journal trials averaged 8.7 percentage points of extra take-up. The nudge-unit trials averaged 1.4, still statistically significant.
The authors' model attributes about 70 per cent of the gap to selective publication in the academic sample, and most of the remainder to differences in the nudges themselves.
The journal literature was pooled the same year. Stephanie Mertens and colleagues report in their abstract a small to medium average effect across more than 200 studies and over 450 effect sizes, with changes to decision structure, such as defaults, outperforming changes to information or reminders. The same abstract notes a moderate publication bias toward positive results.
That last point became the opening for two reanalyses in the same journal. Maximilian Maier and colleagues corrected for publication bias and reported a pooled effect of 0.04, with no evidence for an average effect remaining. They added that most categories were heterogeneous, so some nudges may work even where the mean does not. Barnabas Szaszi and colleagues applied three other correction methods and reported averages between minus 0.01 and 0.08. They argued that the field should ask when and why particular nudges produce large effects, and treated defaults as one of the few likely exceptions. The longer account of that exchange is in Why You Never Change the Setting and in The Psychological Solution First.
Which Four Claims Are Bundled Together?
The phrase nudge theory carries four separate claims, and the audits touched only one of them. The first is about people: they depart from the rational-agent picture in economics in predictable ways. The second is about design: every presentation of options steers, so no neutral arrangement exists. The third is about method: steer, but without banning options or repricing them, and keep the exit cheap. Libertarian paternalism, the term the authors introduced in 2003, names that position, and in 2021 they tightened it.
| Claim | Kind of claim | Where it stands |
|---|---|---|
| About people | Empirical, from earlier behavioural economics | Broad body of findings, uneven in quality |
| About design | Analytic argument | Not tested as a result |
| About method | Position on ethics and policy | Argued, and tightened by the authors in 2021 |
| About effect | Empirical | Disputed, and smaller after correction |
The second and third claims are arguments, not results. Someone has to decide how options are laid out, so declining to design is itself a design, and no experiment can refute that. The fourth claim is about effect: small, inexpensive changes in how options appear shift behaviour at scale.
Only this one can fail the way an experiment fails, and on our reading it survives in a smaller form. The unit-level trials show an average above zero, and the corrected literature shows an average close to it, which means each nudge needs its own trial. A figure from a journal study is not a forecast for a rollout.
Common questions
What is nudge theory?
Nudge theory is the idea, set out by Richard Thaler and Cass Sunstein, that changing how options are presented can steer behaviour without banning choices or changing incentives much. It bundles claims about people, design, method and effect. The effect claim is the contested one: published averages shrink sharply after correction for publication bias.
Who created nudge theory?
Richard Thaler and Cass Sunstein developed it. They introduced the term libertarian paternalism in 2003 and published Nudge in 2008, with a revised Final Edition in 2021. Thaler's earlier work in behavioural economics supplied the research on how people depart from the rational-agent model.
Is nudge theory the same as choice architecture?
No. Choice architecture is the design of how options are presented, and it is unavoidable. Nudge theory adds a rule about method: steer without banning options or changing incentives significantly, and keep opting out cheap. A design that fails the cheap-exit condition is sludge in the authors' terms.
Does nudge theory work?
Results differ by setting. Published meta-analyses report a small to medium average effect before correction for publication bias and an average near zero after it. Trials run by two US nudge units averaged 1.4 percentage points of extra take-up. Defaults are the kind of nudge the reanalyses were most willing to except.
Sources
Thaler, R. H., & Sunstein, C. R. (2021). Nudge: The Final Edition. Seed book; vault hub Books/Nudge.
Thaler, R. H., & Sunstein, C. R. (2003). “Libertarian paternalism.” American Economic Review, 93(2), 175-179. doi:10.1257/000282803321947001 https://doi.org/10.1257/000282803321947001
Madrian, B. C., & Shea, D. F. (2001). “The power of suggestion: Inertia in 401(k) participation and savings behavior.” Quarterly Journal of Economics, 116(4), 1149-1187. doi:10.1162/003355301753265543 https://doi.org/10.1162/003355301753265543
Mertens, S., Herberz, M., Hahnel, U. J. J., & Brosch, T. (2022). “The effectiveness of nudging: A meta-analysis of choice architecture interventions across behavioral domains.” Proceedings of the National Academy of Sciences, 119(1), e2107346118. doi:10.1073/pnas.2107346118 https://doi.org/10.1073/pnas.2107346118
Maier, M., Bartos, F., Stanley, T. D., Shanks, D. R., Harris, A. J. L., & Wagenmakers, E.-J. (2022). “No evidence for nudging after adjusting for publication bias.” Proceedings of the National Academy of Sciences, 119(31), e2200300119. doi:10.1073/pnas.2200300119 https://doi.org/10.1073/pnas.2200300119
Szaszi, B., Higney, A., Charlton, A., Gelman, A., Ziano, I., Aczel, B., Goldstein, D. G., Yeager, D. S., & Tipton, E. (2022). “No reason to expect large and consistent effects of nudge interventions.” Proceedings of the National Academy of Sciences, 119(31), e2200732119. doi:10.1073/pnas.2200732119 https://doi.org/10.1073/pnas.2200732119
DellaVigna, S., & Linos, E. (2022). “RCTs to scale: Comprehensive evidence from two nudge units.” Econometrica, 90(1), 81-116. doi:10.3982/ECTA18709 https://doi.org/10.3982/ECTA18709