Kahneman is commendably candid on the limits of behavioral psychology. While his insights have revolutionized thinking in social psychology and economics, those same insights instill modesty about what can realistically be known of outcomes in complex conditions. The blending of biases in billions of individuals where the biases are both contradictory in real time and adaptable over time makes the art of composite predictive analytics harrowing. It’s one thing to know that a herd can turn and run; it’s another to know where or when the herd will do so.
These limitations have not stopped Thaler and Sunstein from molding social behavior to their liking. They have enlisted behavioral psychology in pursuit of what they call libertarian paternalism. There’s nothing libertarian about it. This is revealed in their book when they write: “Many of the most important applications of libertarian paternalism are for government.”6 How the augmentation of state power is libertarian remains a riddle. The word “paternalism” sufficiently describes their aim. The channel of choice for their paternal agenda is choice architecture.
Choice Architecture
The new social and economic science of choice architecture proceeds from the necessity of design. Thaler and Sunstein point out that human-made systems inevitably have a design. Autos can have three-point seat belts, lap-style seat belts, or no seat belts. Cafeteria lines can start with candy or carrots. Ballots can list Democratic candidates first or Republicans. Every system includes design elements that involve choices made by the designer about placement, ease of use, color, vocabulary, and other factors.
At the same time, the human user facing a designed system with embedded choices has choices of her own. As a consumer moves through a cafeteria line, she asks herself, “Do I want the cheesecake today? Should I have more carrots to help cure a vitamin deficiency?” Consumer choice meets design choice in everyday settings.
By presenting design choices in a certain way, you can predictably affect consumer choices by preying on biases unveiled by behavioral psychology. This interface between design choice and consumer choice is the domain of choice architecture. In Thaler and Sunstein’s words:
Decision makers do not make choices in a vacuum.7 They make them in an environment where many features, noticed and unnoticed, can influence their decisions. The person who creates that environment is, in our terminology, a choice architect …. Our goal is to show how choice architecture can be used to help nudge people to make better choices (as judged by themselves) without forcing certain outcomes upon anyone, a philosophy we call libertarian paternalism. The tools we highlight are: defaults, … feedback, structuring complex choices, and creating incentives.
This manifesto makes clear that Thaler and Sunstein’s agenda goes beyond science. Behavioral psychology is supported by an enormous body of empirical data derived from well-designed experiments on real human subjects. It is sound science, not mere conjecture or ideology masquerading as science. Yet the application of science is another matter. Nuclear physics is sound science; nuclear weapons are a matter for debate. Thaler and Sunstein have weaponized behavioral psychology.
Some examples illuminate how Thaler and Sunstein proceed in practice. The authors consider charitable giving, in which typical solicitations might include a range of suggested contributions in the form of $50, $150, $250, $500, and so on. Research shows that the larger the suggested amount, the greater the contribution given. Thaler and Sunstein would recommend the charity increase its suggested gift levels to $100, $300, $500, and $1,000. That simple change in choice architecture results in higher contributions. This suggestion makes use of the anchoring bias. The target of the solicitation anchors her behavior in the higher amounts used in the second example, and makes a larger gift as a result. This interaction of design choice and consumer choice is predictable and can be used to increase charitable giving.
Another example involves the overconfidence bias. Data shows that the five-year failure rate for new businesses is over 50 percent. Yet when entrepreneurs are asked in a survey about the chance of success of their business, the typical answer is 90 percent; some say 100 percent. Obviously entrepreneurs on the whole are overconfident with respect to the potential for the success of their businesses. Thaler and Sunstein have a remedy for overconfidence. They write, “If people are running risks because of unrealistic optimism, they might be able to benefit from a nudge.8 In fact, we have already mentioned one possibility: if people are reminded of a bad event, they may not continue to be so optimistic.”
So, with simple design choices as suggested by Thaler and Sunstein, consumers can be influenced to give more to charity and entrepreneurs can be cautioned to avoid excess risk. Sounds wonderful, but is it? If Thaler and Sunstein’s outcome is superior, why is the manipulation needed? Put differently, why do the academics maintain an a priori assumption that the uncovered bias is undesirable? Even if the list of suggested charitable contributions used the smaller dollar amounts, there’s nothing to stop a contributor from donating a larger amount. Are all charities automatically desirable? There’s nothing stopping an entrepreneur from taking a hard look at her business model and concluding that success is a longshot without being reminded that bad things happen to good businesses. Those possibilities aren’t good enough for Thaler and Sunstein because they don’t trust the judgment of typical donors or entrepreneurs. They feel compelled to give what they call a “nudge” to achieve a supposedly superior outcome.
With reference to their book title, Thaler and Sunstein state, “A nudge, as we will use the term, is any aspect of choice architecture that alters people’s behavior in a predictable way without forbidding any options or significantly changing their economic incentives” (emphasis added).9 Nudge is a euphemism; hard shove is a better descriptor given what we know about the power of behavioral science. Euphemism is the refuge of the authoritarian. The rationale for Thaler and Sunstein’s regimentation is that they’re smarter than you and me.
The twin social engineers are explicit about their superior smarts. They subscribe to a process described as “taking steps to help the least sophisticated people while imposing minimal harm on everyone else.”10 This is not that difficult to do because “humans are not exactly lemmings, but they are easily influenced.”11 With reference to framing, a kind of behavioral manipulation, Thaler and Sunstein write, “Framing works because people tend to be somewhat mindless ….”12 They acknowledge that deception is sometimes needed to put their smart ideas over on the less sophisticated. Even good is not good enough for Thaler and Sunstein. They advise, “If you want to nudge people into socially desirable behavior, do not, by any means, let them know that their current actions are better than the social norm.”13 The authors claim that “individuals make pretty bad decisions—decisions they would not have made if they had paid full attention and possessed complete information, unlimited cognitive abilities, and complete self-control.”14 Since Thaler and Sunstein are in the nudge business, presumably they are the ones with unlimited cognitive abilities, not you.
The field of choice architecture and nudges is shot through with arrogance, unexamined premises, and adverse consequences. Take the example of the charitable solicitation where more is more; higher suggested contributions result in higher actual contributions. What is the basis for the implicit assumption that institutionalized charity is good? Random acts of kindness are certainly good, and don’t need nudges. Consider the mass-marketing apparatus behind the kinds of solicitations that Thaler and Sunstein want to nudge. If you support gun control, do you favor Thaler and Sunstein’s technique to help the NRA Foundation increase tax-deductible donations used to promote assault rifles? If you are pro-life, do you favor the Thaler and Sunstein nudge to increase contributions to Planned Parenthood Federation, the nation’s largest abortionist? Thaler and Sunstein promote a pro-charity nudge as an uncontested good without examining what charities do with the money.
The pair’s suggestion that entrepreneurs should be reminded of failure to discourage
overconfidence in the face of a 50 percent flop rate is equally misguided. It is true that most new businesses fail within a few years. It is also true that entrepreneurs consistently overestimate their chance of success on average. So what? Thaler and Sunstein are compelled to apply their big brains to warn off entrepreneurs from their endeavors. Yet this ignores the asymmetry in the results of success and failure. An early entrepreneurial failure typically involves relatively few lost jobs and little lost capital because the business did not get that far out of the gate. Start-up business failures happen fast due to poor execution, poor quality, or simple lack of demand for, say, the original one-size-fits-all, waterproof, stain-resistant ice-cream glove. The opportunity costs of these efforts are low and the unsuccessful entrepreneur is on to her next enterprise in no time. In contrast, the few entrepreneurial successes can be huge. They can create thousands of jobs, generate large investments in plant and equipment, and produce prodigious profits for investors to support consumption or reinvest in new enterprises. Indeed, this pattern of many small failures offset by a few large successes is the bedrock assumption of the venture capital business. Who are Thaler and Sunstein to discourage entrepreneurs by reminding them of a “bad event”? Society should encourage entrepreneurship every step of the way. Thaler and Sunstein also ignore the fact that the entrepreneurial dream itself has value independent of ultimate financial success. We need dreamers, not academic Debbie Downers.
Thaler and Sunstein claim their work is designed solely to protect people from their own mistakes. Since the mistakes are predictable due to innate biases, why not structure choices in ways intended to lead the consumer-citizen to good choices instead of bad ones? Thaler and Sunstein place great weight on the absence of coercion. While they admit to using choice architecture to increase the likelihood of certain outcomes, they emphasize that the consumer still has a choice; nothing is forced. This is what they mean by their grandiose claim of libertarian paternalism. Libertarian in the sense that the consumer has a choice, yet paternal in the sense that economists are looking out for you and protecting you from your own childlike impulses.
Thaler and Sunstein claim that design choices are inevitable; no system works without some design, conscious or random. Not making a design choice is a choice. You are left with whatever design the originator chose, including the results of her biases. This produces a likely result based on that particular design. Why not be proactive? Why not identify desirable outcomes, design the form or interface to encourage that result, and reap the benefits of better choices at little or no cost? It seems the closest economics has to offer to the proverbial free lunch. Significant gains in welfare are achieved with no more effort than rearranging a few boxes on an application.
The best-known operation of the Thaler-Sunstein approach is to revise application forms from opt-in to opt-out elections. For example, a company’s new employee information questionnaire might offer the new employee a chance to join a company benefit program. The form has a box for that choice. If you check the box, you’re in the program. If you do not check the box, you’re not. Checking the box is called opting in to the program. Behavioral research shows that people don’t like filling out forms and are averse to checking boxes, independent of the substance or benefits of the choice. Many new employees leave the box blank and skip the program because of this bias. What if the benefit program is a 401(k) plan that offers tax deferral on current income, tax deferred compounding of plan earnings, and employer matching funds? The choice architects reason that those features are desirable, so they redesign the form. The new questionnaire informs you that you are automatically enrolled in the benefit program unless you check the box. If you do nothing, you’re in the program. If you check the box, you’re not. The result is that far more people enroll in the program because of the human bias against checking boxes. With this simple change in the form design, Thaler and Sunstein can greatly increase plan participation. That’s their nudge. Welfare is improved because more employees are making the “right” choice, as defined by Thaler and Sunstein. No one is forced to make a particular choice. Everyone has full information about the consequences of the choice. Coercion is nonexistent and the costs of changing the form are low. (The biggest cost is probably the consulting fees that choice architects earn to make these recommendations, yet when spread over thousands or millions of applications even that cost is trivial.)
What’s wrong with this enlightened, low-cost, noncoercive approach to human behavior? Almost everything.
The first problem is that the Thaler-Sunstein method seeks to displace an older solution to behavioral bias, called learning. Humans have enormous capacity to learn based on imitation, imagination, and the harder technique of learning from your mistakes. As a young driver in the late 1960s, I once hopped out of my car, shut the door behind me, and left my lights on. When I returned to my car the battery was dead because of the drain from the lights. I had no money and no membership in AAA. What followed was an ordeal. I walked to a pay phone to call my father—no mobile phones in those days. He arrived eventually with a set of jumper cables. We opened the hoods of the two cars, his and mine, connected the cables, and jump-started my battery. Then we were both on our way. The point is, I never accidentally left my lights on again. On separate occasions I have locked my keys in the car. That’s a more challenging situation than a dead battery, because you have to call the police or a locksmith to open the car door, using the same tools and techniques as a professional car thief. Again, I learned a lesson. Today, I don’t need a beeping sound to remind me of what to do when I get out of my car. Importantly, the dead-battery and locked-door lessons have carried over to other endeavors. In time, I learned to stop and think before making sudden moves.
A slight pause before an unthinking action is invaluable. While mountain climbing on fixed ropes there are situations where one rope ends and another begins, with both ropes anchored to the mountain using ice screws or other fixed points. The climber is attached to the first rope through a system of slings, carabiners, and a climbing harness. In passing from one rope to another, the technique is to clip on to the second rope using a second carabiner before unclipping from the first. It’s just as easy to reverse the technique by unclipping from the rope before clipping onto the next rope, with one important difference. When you unclip first you are momentarily unattached to either rope before you clip on to the second. If you slip or get hit with a wind gust in that moment and lose balance, the result can be fatal. Seasoned climbers learn to move slowly and think clearly in these situations. Fortunately, Thaler and Sunstein have not yet pursued choice architecture for mountaineers at twenty thousand feet. In life, a few mistakes go a long way. Once you’ve been impulsive at some cost, you learn to be less impulsive in future. That’s how hominids survived 3 million years of evolution to become university professors. Thaler and Sunstein’s efforts to replace hard learning with soft nudges moves us backward. They prefer automatons to autonomy. Experience is a better teacher.
The second deficiency in the Thaler-Sunstein approach is that effects of nudges are not lasting. The two professors ignore another aspect of human nature—adaptive behavior, and the ability of people to ignore their nudge after a few interactions.
This is the case with my Audi. There might have been a time shortly after I bought the car in 2008 when I took a minute to figure out what the beeps were telling me. If so, I pushed that out of my brain immediately. The beep begins. Seat belts? Lights? Engine running? Door open with the engine running? Keys left in the car? I have no idea. I am numb to nudges. I only know the beeps are annoying; I’m relieved when they finally go away.
Frequent flyers are familiar with this phenomenon. Air travelers are barraged with audio and video nudges from the time they’re dropped off at the terminal until the time they land at their destination.
“Take your laptop out of your bag.”
“Remove your belt and shoes.”
“If you see something, say so
mething.”
“Empty your pockets.”
Once they’re on the plane, the cannonade continues. Does anyone really watch those cutesy video safety presentations? (They’re cutesy because travelers long ago tired of the by-the-book versions.) A flyer’s reaction to the in-flight seat belt sign is optional, a decided “If I feel like it” response. The only nudge flyers pay attention to is the final “ding” at the gate that signals it’s okay to get out of your seat. That’s more of an anti-nudge, because passengers are already primed to pop out of their seats and would do so even sooner but for the ding.
The screener’s and airline’s shouts, orders, and prods might have served some purpose at some time, but are long lost on frequent flyers. It’s not clear why a silent video demonstration near security checkpoints would not serve just as well for first-time flyers. But flyers’ response to the cattle punching has quickly gone from grudging acceptance to cold compliance to complete insensitivity. Adaptive behavior trumps the nudge effect over time.
A third flaw in the Thaler-Sunstein conceit is that the costs of nudging are low. They’re not. The cost of changing a choice on a single application may be trivial. The marginal manufacturing cost of making my car beep at me is practically zero. Yet even trivial costs add up when replicated in thousands of applications imposed on busy humans. A practical accounting should consider the weight of imposing thousands of nudges per week on busy individuals just trying to get through their day.
An alarm clock is more than a nudge; it’s like a bucket of cold water in the face, but it’s how we start our days. You make coffee. The coffee machine beeps when the coffee’s done, as if you didn’t know from the aroma. You log on to your computer. An uninvited pop-up nudge asks if you want to update your apps. The nudge won’t take no for an answer; that’s not one of the choices. You can say yes, in which case you have to restart the computer you just started, or you can say “Remind me later today,” or “Remind me tomorrow.” You can check one or the other, but they’re both lies, because you really don’t want to be reminded at all; you just want to be left alone to scan the news or email.
Aftermath Page 11