The Real Third Way

About 30 years ago, my family spent a month in London--our first visit. Soon after we arrived, my wife took our 7-year-old daughter to the British Museum. When they left, talking excitedly about what they had seen, they stepped off the curb onto a busy street--and got hit by a car. The driver had managed to slow the car down enough so that there were no serious injuries. He was profusely apologetic, but he shouldn't have been. The fault was my wife's. She had looked the wrong way as she stepped off the curb, hand-in-hand with our daughter.

Unwanted collisions like this between drivers and pedestrians were all too common in England because visitors from abroad are not used to people driving on the left side of the road. Faced with this persistent problem, what should the state do? It could do nothing. Respect people's freedom and autonomy. Assume that people will pursue their rational self-interest. Not patronize. Or it could do a little something. It could post signs at busy intersections saying, "Look Before You Cross," in the hope that people would then pay attention before stepping off the curb. Or it could do a little something more. It could acknowledge the specific problem and post signs that say, "Look Right Before You Cross."

Nudge, by economist Richard Thaler and legal scholar Cass Sunstein, is about situations that are just like this one, scaled up to encompass health insurance, retirement planning, investment, credit, education, medical decisions, organ donations, marriage, and more. The book is an expanded and updated version of an influential law review article they published in 2003. Though I have a few bones to pick with it, Nudge is as important a book as any I've read in perhaps 20 years. It is a book that people interested in any aspect of public policy should read. It is a book that people interested in politics should read. It is a book that people interested in ideas about human freedom should read. It is a book that people interested in promoting human welfare should read. If you're not interested in any of these topics, you can read something else.

Nudge is the fleshing out of a few simple but very important facts. First, every choice situation has a structure--or as Thaler and Sunstein call it, a "choice architecture." Second, that structure matters; it influences the choices people make. Third, there is no neutral way to structure choices; any architecture will nudge choosers in one direction or another. Finally, people are usually unaware that the structure of the choices they face influences what they choose.

These findings are derived from almost 40 years of research in a field that economists call "behavioral economics" and psychologists call "behavioral decision making." The pioneers in this field were psychologists Amos Tversky and Daniel Kahneman, and Kahneman was awarded the Nobel Prize in economics for this work a few years ago (Tversky would surely have joined him had he not died prematurely). Both Thaler and Sunstein have collaborated frequently with Kahneman, and Thaler himself is one of the major figures in the field.

Nudge does a lovely job of providing a whirlwind tour of behavioral decision making. Here is some of what we know:

Choices are affected by how options are described or framed. Many more people elect surgery over radiation when they are presented with statistics expressed as survival rates than when they are presented with (the same) statistics expressed as mortality rates.

Choices are affected by "anchors." An $800 suit seems much more expensive on a rack with $400 suits than on a rack with $1,500 suits. Experienced real estate agents asked to price homes in neighborhoods they know well are influenced by the listed asking prices, though they swear they are not.

People hate losses. A coin flip in which heads wins you $10 and tails loses you $10 is a bad psychological bet. This "loss aversion" helps explain why people are reluctant to change retirement plans, insurance plans, jobs, or pretty much anything else. Inertia is partly laziness, to be sure, but it's partly that the prospect of giving up something good looms larger than the prospect of gaining something better. Finally, people organize their mental balance sheets into a variety of different "mental accounts." Even money is not completely fungible. This helps explain why people faithfully put money into savings accounts that pay 4 percent while slowly paying off credit-card debt at 18 percent.

The best current thinking about psychological processes like these is that they are automatic, inevitable, and nonconscious. Our cognitive machinery spits out answers to questions like "is this a good deal?," and though we have access to the answers, we don't have access to the processes that produced them. We can then turn our tools of rational deliberation loose on the answers to decide whether or not to act on them. But the tools of rationality do not start at a neutral place; they have been "nudged" or anchored by these automatic psychological processes. And the tools of rationality are frequently unable to overcome the starting nudge. People don't have enough time. They aren't paying enough attention. The choices they face are too complex. They lack many of the needed analytic tools. There are lots of ways to go wrong, and there is plenty of evidence, reviewed by Thaler and Sunstein, that people do go wrong, often very wrong. People make bad decisions about credit, investment, retirement pensions, health care, food, and the like.

Not everyone is equally enslaved by these automatic processes. People whom Thaler and Sunstein call "Econs" have the training and temperament to cast the automatic answers aside and use rational tools to make decisions. Such people tend to reside in economics departments. The rest of us, a group Thaler and Sunstein call "Humans," are much more vulnerable. But policy-makers, many of them Econs themselves, have acted as if we're all Econs. They have usually ignored choice architecture and just worried about choice--the more the better. Just as software engineers design applications as if every user will be a software engineer (maximize flexibility and power; don't worry about usability), so Econs have designed social and economic programs as if every user will be an Econ.

Faced with the facts of human fallibility, it is tempting to engineer choice out of the system and just offer what is in people's best interests: one pension plan, one health-insurance plan, one college curriculum, and so on. Thaler and Sunstein view this kind of paternalism as a nonstarter, both because there is usually no one thing that is best for everybody and because freedom of choice is itself a good thing. But we can do better, much better, than we currently do by acknowledging that choosers are Humans and by paying attention to choice architecture. Give people choice, but engineer the choice architecture so that most people, most of the time, will choose what is in their best interests. "Libertarian paternalism" is what they call this approach: choice with the right set of nudges. Libertarian paternalism, they argue, is the real "third way."

If England wants to stop people like my wife from getting hit by cars, the fully paternalistic option is out; you can't force people to look in the right direction before they cross the street, even if you want to. The nudge, "look before you cross" is a step in the right direction. But it's too small a step. "Look right before you cross" (which is what was actually done) is better: It's an explicit nudge to overcome the strong psychological habit of looking left. If you're going to create a nudge, why not create the most effective one you can?

Perhaps the most important nudge available to policy-makers is to pay attention to what happens when people do nothing, which in many situations is by far the most popular option. Our driver's license renewal forms invite us to do nothing when they ask (I am paraphrasing): "Do you want to be an organ donor if you die in a car accident? If so, check the box and sign below." The forms employees fill out when they start new jobs similarly invite them to do nothing by asking whether they want money taken out of their paychecks and put into 401(k) accounts. Is a failure to check off the box approving positive action in each case a reflection of people's true preferences? Suppose the form reversed the default condition and asked whether people object to being an organ donor or to having part of their wages put into a retirement account. This simple nudge triples the rate of organ donation and dramatically increases employee saving for retirement.

Thaler and Sunstein convincingly show that by switching defaults from "opt-in" (nothing happens unless you sign) to "opt-out" (something happens unless you sign), thousands of lives can be saved with harvested organs, and millions more people can have the financial resources to live decent lives in retirement. This simple and cost-free act of mindful choice architecture can improve our collective well-being--without compelling anyone to do anything. And in chapters on savings, credit, investing, organ donation, and energy conservation, Thaler and Sunstein show how other small, usually inexpensive changes in choice architecture can have very large effects.

Yet Thaler's and Sunstein's libertarian paternalism has plenty of critics. Libertarians criticize the paternalism. Create your choice architecture to maximize freedom, not welfare, some say. This may amount, in some circumstances, to eliminating defaults and making people choose--for example, check one box to be an organ donor and the other box not to be; other-wise, no driver's license for you. (Note the paradox of a libertarian position that forces people to choose.) Lefties criticize the libertarianism. Why always the need for choice, they wonder, especially in domains in which there are externalities--where innocent people end up paying for the bad choices of others?

Thaler and Sunstein respond to such criticisms, though not as carefully as I would have liked. They're enthusiasts of the magic of the market--as long as market choices are structured for use by Humans, not Econs. My own view is that this market enthusiasm makes sense in some domains but not in others. In health care, for example, the "market" is worse than imperfect. And as Robert Kuttner eloquently argues in his book, Everything for Sale, the assumption that getting as close as you can to a perfect market will always make things better (what Kuttner calls "the logic of the second best") is highly suspect. Similarly, I would sacrifice a fair bit of libertarianism for some well-designed, welfare-enhancing paternalism. But as Thaler and Sunstein point out, extreme paternalism would often be infeasible, ineffective, or repugnant. Their advice is to leave your ideology (left or right) at the door and figure out what works best in each situation. The standard for "working best" will sometimes be clear: Organ donation saves thousands of lives, and when asked, the vast majority of Americans approve of it. When the results are not clear, however, freedom should probably dominate welfare as a standard for "working best," unless there are major externalities.

Fair enough. But this brings me to my only real disappointment with this extraordinary book. I think the research in behavioral decision making on which Nudge is based has major implications for what we take "freedom" to be. In what sense are people "free" to choose when a host of psychological processes operate automatically and nonconsciously in shaping our choices? Even Econs are nudged, though they may be better than the rest of us at overcoming these nudges and more willing to put time and effort into doing so. If "nudge ability" is an ineliminable feature of human nature, then those with libertarian impulses need to tell us what "freedom" looks like for people who are in (cognitive) chains. We need a normative account of what freedom means for organisms that have the decision-making characteristics that humans seem to have.

I didn't expect Thaler and Sunstein to answer the question of what Nudge means for our understanding of freedom, but I wish they had raised and discussed it. And not just out of theoretical interest. We look to the market for solutions for two primary reasons--market efficiency and the value of freedom. Why and where markets fail to achieve efficiency is the subject of a great deal of sophisticated analysis, but the limits of the market in realizing freedom raise a less familiar and more difficult problem. If studies of decision-making change our conception of freedom, they may change our view of how well markets serve freedom. And that could have wide repercussions in American social and political institutions.