Does Nudging Work? A Critique of Sunstein and Thaler

Stories from New Scientist.
Nov. 12 2011 9:20 AM

Nudge No More

Benevolent meddling won’t help us make good decisions.

129265154
Would more people become organ donors if just given a nudge? Last month, Alexandra Burke and Jameela Jamil tried to sign up a record number of organ donors in one hour in London.

Photo by Ben Pruchnie/Getty Images

We have all cringed watching friends and family make terrible decisions, and been tempted by visions of the pain spared if we could only make them follow our advice. The same feeling motivates well-intentioned technocrats to take charge of the public: People are plainly making sad blunders they will regret.

Economist Richard Thaler and legal scholar Cass Sunstein (now a senior policymaker in the Obama administration) present the latest, and subtlest, version of this temptation in their influential work on "nudging" people into making wiser choices. They argue that wise decision-makers should tweak the options and information available so that the easiest choice is the right one. For example, this can guide people to donate their organs if they die unexpectedly by making organ donation an opt-out rather than an opt-in choice. And it can encourage people to plan for their pensions by making pension contributions automatic for everyone who does not explicitly opt out of the system.

"Nudging" is appealing because it provides many of the benefits of top-down regulation while avoiding many of the drawbacks. Bureaucrats and leaders of organizations can guide choices without dictating them. Thaler and Sunstein call the approach "libertarian paternalism": It lets people "decide" what they want to do, while guiding them in the "right" direction.

Much criticism of this approach comes, in fact, from libertarians, who see little difference between guiding a person's choices and eliminating them. A nudge is like a shove, they argue, only more disreputable because it pretends otherwise. The real problem, though, is that Thaler and Sunstein's ideas presume good technocrats can use statistical and experimental results to guide people to make choices that serve their real interests. This is a natural belief for scientists and intellectuals, especially those who see the awful ways scientific knowledge is abused politically, and think life would be better if scientists had more authority.

Advertisement

However natural, though, this won't work because libertarian paternalists are often wrong on the underlying social science. For example, Thaler and Sunstein's claims about the benefits of opt-out schemes are belied by little evidence it increases donations. According to Kieran Healy, a sociologist at Duke University differences in donation rates are better explained by differences in organizational effectiveness than differences in opt-in/opt-out. It is not clear that opt-out would increase donations; unsexy but crucial reforms to regional schemes would almost certainly work better.

This points to the key problem with "nudge"-style paternalism: presuming that technocrats understand what ordinary people want better than the people themselves. There is no reason to think technocrats know better, especially since Thaler and Sunstein offer no means for ordinary people to comment on, let alone correct, the technocrats' prescriptions. This leaves the technocrats with no systematic way of detecting their own errors, correcting them, or learning from them. And technocracy is bound to blunder, especially when it is not democratically accountable.

As political scientist Suzanne Mettler of Cornell University argues, libertarian paternalism treats people as consumers rather than citizens. It either fails to tell people why choices are set up in particular ways or actively seeks to conceal the rationale. When, for example, Obama's administration temporarily cut taxes to stimulate the economy, it did so semi-surreptitiously to encourage people to spend rather than save.

Mettler uses experiments to show how ordinary people can understand complicated policy questions and reach considered conclusions, as long as they get enough information. This suggests a far stronger role for democratic decision-making than libertarian paternalism allows. People should be given information, and allowed to reach conclusions about their own interests, and how to structure choices to protect those interests. By all means consult experts, but the dialogue should go both ways.

Results from agent-based modeling, evolutionary theory, network theory, and experiments in group decision-making also support Mettler. Take the "diversity trumps ability" theorem of Scott E. Page, from the University of Michigan: Groups of agents with diverse understandings of the world will solve difficult problems better than narrowly focused groups with higher expertise.

And models of evolutionary search, starting with the "genetic algorithms" of John Holland, also at Michigan, suggest higher diversity per se makes it easier to find paths to new fitness peaks. Research into the sociology of networks also finds innovation is most likely at points where different views intersect.

All this suggests democratic arrangements, which foster diversity, are better at solving problems than technocratic ones. Libertarian paternalism is seductive because democratic politics is a cumbersome and messy business. Even so, democracy is far better than even the best-intentioned technocracy at discovering people's real interests and how to advance them. It is also, obviously, better at defending those interests when bureaucrats do not mean well.

While democratic institutions need reform to build in dialogue between citizens and experts, they should not be bypassed. By cutting dialogue and diversity for concealed and unaccountable decision-making, "nudge" politics attacks democracy's core. We should not give in to temptation—and save our benevolent meddling for family reunions.

This article originally appeared in New Scientist.

Henry Farrell is a political scientist at the Elliott School of International Affairs in Washington, D.C.

TODAY IN SLATE

Medical Examiner

Here’s Where We Stand With Ebola

Even experienced international disaster responders are shocked at how bad it’s gotten.

Why Are Lighter-Skinned Latinos and Asians More Likely to Vote Republican?

A Woman Who Escaped the Extreme Babymaking Christian Fundamentalism of Quiverfull

The XX Factor
Sept. 22 2014 12:29 PM A Woman Who Escaped the Extreme Babymaking Christian Fundamentalism of Quiverfull

Subprime Loans Are Back

And believe it or not, that’s a good thing.

It Is Very Stupid to Compare Hope Solo to Ray Rice

Building a Better Workplace

In Defense of HR

Startups and small businesses shouldn’t skip over a human resources department.

How Ted Cruz and Scott Brown Misunderstand What It Means to Be an American Citizen

Divestment Is Fine but Mostly Symbolic. There’s a Better Way for Universities to Fight Climate Change.

  News & Politics
Politics
Sept. 22 2014 6:30 PM What Does It Mean to Be an American? Ted Cruz and Scott Brown think it’s about ideology. It’s really about culture.
  Business
Moneybox
Sept. 22 2014 5:38 PM Apple Won't Shut Down Beats Music After All (But Will Probably Rename It)
  Life
Outward
Sept. 22 2014 4:45 PM Why Can’t the Census Count Gay Couples Accurately?
  Double X
The XX Factor
Sept. 22 2014 7:43 PM Emma Watson Threatened With Nude Photo Leak for Speaking Out About Women's Equality
  Slate Plus
Slate Plus
Sept. 22 2014 1:52 PM Tell Us What You Think About Slate Plus Help us improve our new membership program.
  Arts
Brow Beat
Sept. 22 2014 9:17 PM Trent Reznor’s Gone Girl Soundtrack Sounds Like an Eerie, Innovative Success
  Technology
Future Tense
Sept. 22 2014 6:27 PM Should We All Be Learning How to Type in Virtual Reality?
  Health & Science
Medical Examiner
Sept. 22 2014 4:34 PM Here’s Where We Stand With Ebola Even experienced international disaster responders are shocked at how bad it’s gotten.
  Sports
Sports Nut
Sept. 18 2014 11:42 AM Grandmaster Clash One of the most amazing feats in chess history just happened, and no one noticed.