To give one concrete example: if the goal is decreasing teenage pregnancy, the most effective strategy is the use of long-term reversible contraceptives such as an intrauterine device (IUD). Trials with a sample of sexually active young women have found a failure rate of less than 1%, much lower than with other forms of contraception. Once the device is implanted, no further action is needed. Those looking for behavioral interventions that have a high probability of working should seek out other environments in which a one-time action can accomplish the job. If no one-time solution yet exists, invent one!
In some cases, successful interventions are simply reminders to people who might otherwise forget to do something. Many examples of this type have been made possible by the technology of mobile texting, which shows that nudges need not be creative, elaborate, or hidden in any way; simple, straightforward reminders in the form of a text can be extremely effective. One example comes from the domain of health. In a study in Ghana, the nonprofit Innovations for Poverty Action ran a randomized control trial testing whether text message reminders to take malaria medication helped people follow through with the medical regimen. Not only did they find these texts to be effective, but they also found that the most effective messages were brief; it was the reminder, not any additional information, which mattered.
Similarly, a study in the realm of education highlights the efficacy and scalability of simple text reminders. The study measured the effectiveness of READY4K!, a program that sends parents of preschoolers regular texts containing tips for good parenting, including ways to help children learn reading and writing skills. The study showed significant increases in parental involvement in literacy activities both at home and at school, in turn increasing learning gains for their children.
Such simple reminders are a good example that nudges can truly be gentle and transparent, and still work.**
The BIT passed its built-in two-year review and was renewed by the Cabinet Office in 2012. Because the team had continued to grow rapidly, it was necessary to find it a new home. The stay in the drafty original quarters was mercifully brief, but the next home, in borrowed space within the Treasury Department, was too small for the growing team’s needs. So in 2014, a decision was made to partially privatize the BIT. It is now owned in equal parts by the Cabinet Office, its employees, and its nonprofit partner NESTA, which is providing the team with its current workspace. BIT has a five-year contract with the Cabinet Office, so it can make plans that are independent of the outcome of the general election in May 2015. The team has grown to nearly fifty and now supports a range of public bodies across the U.K., and increasingly helps other national governments too, including an exciting new tax compliance study in Guatemala.
While I was kibitzing the efforts of the U.K. Behavioural Insights Team, Cass was busy in Washington serving as the administrator of the Office of Information and Regulatory Affairs, known as OIRA (pronounced “oh-Ira”). Formally a part of the Office of Management and Budget in the White House, OIRA was formed in 1980 with the mission to evaluate the economic impact of new governmental regulations to assure they do more good than harm. Although he did not have a mandate or budget to run randomized control trials, to some extent Cass was able to serve as a one-man Behavioural Insights Team during President Obama’s first term.
After four years working for the government, Cass went back to teaching at Harvard Law School, where he had moved just before President Obama was elected. But the U.S. nudging agenda did not end with Cass’s departure. In early 2014, Dr. Maya Shankar, a former violin prodigy turned cognitive neuroscientist turned nudger, created a small unit in the White House. Maya, who makes the Energizer bunny look lethargic, has a knack for making things happen. On an American Association for the Advancement of Science fellowship, she served as an advisor in the White House Office of Science and Technology Policy. In this role, Maya made it her mission to create an American version of BIT. Miraculously, she accomplished this in less than a year and without a mandate or any funding from the government.
The team, officially called the White House Social and Behavioral Sciences Team (SBST), began as a small unit of just six behavioral scientists: Maya, two fellows on loan from universities, and three more on leave from not-for-profit think tanks, the North American branch of the Jameel Poverty Action Lab (J-PAL), which specializes in running RCTs, and ideas42, which has behavioral economics as its core strength.
In just the first year, the SBST embedded a dozen behaviorally-informed randomized control trials into federal programs, with policy objectives ranging from increasing uptake of veterans’ benefits to helping people pay off their student loans. And the team is growing too. The federal government recently responded to the team’s early successes by committing part of its budget to fund additional team members. Thanks to federal support and the continued support of outside partners, the team should have doubled in size by the time this book is published.
Other countries are also joining the movement. A study conducted by the Economic and Social Research Council published in 2014 reports that 136 countries around the world have incorporated behavioral sciences in some aspects of public policy, and 51 “have developed centrally directed policy initiatives that have been influenced by the new behavioural sciences.” Clearly word is spreading.
It is worth highlighting that the authors of the report chose the term “behavioral sciences” to describe the techniques being used. The work of the BIT has often been mischaracterized as being based on behavioral economics whereas, in fact, there has been, at least up to now, very little actual economics involved. The tools and insights come primarily from psychology and the other social sciences. The whole point of forming a Behavioural Insights Team is to utilize the findings of other social sciences to augment the usual advice being provided by economists. It is a slur to those other social sciences if people insist on calling any policy-related research some kind of economics.
Whenever anyone asks me to sign a copy of Nudge, I always add the phrase “nudge for good.” Nudges are merely tools, and these tools existed long before Cass and I gave them a name. People can be nudged to save for retirement, to get more exercise, and to pay their taxes on time, but they can also be nudged to take out a second mortgage on their home and use the money on a spending binge. Businesses or governments with bad intentions can use the findings of the behavioral sciences for self-serving purposes, at the expense of the people who have been nudged. Swindlers did not need to read our book to know how to go about their business. Behavioral scientists have a lot of wisdom to offer to help make the world a better place. Let’s use their wisdom by carefully selecting nudges based on science, and then subjecting these interventions to rigorous tests.
I am proud to say that my hometown, Chicago, has just launched its own behavioral insights team with the help of ideas42. Encourage your own governments to do likewise. The failure to do so amounts to serious misbehaving.
________________
* Richard has now moved to the United States and has a post at the Brookings Institution in Washington, DC.
† One can think of this strategy as appealing to people who are “conditional cooperators” as discussed in chapter 15.
‡ Of course, there is some ambiguity about what is meant by the word “transparent.” If a salad bar is placed in a prominent location in the cafeteria (as it is at the Chicago Booth School of Business, I can proudly report), I do not think that it is necessary to post a sign saying that the goal of its prominent location is to nudge you to choose the salad over the burgers. The same goes for the language in the letter. It is not necessary to point out the key sentence and say that we have inserted it to increase the chance that you send us a check promptly. That is what the entire letter is trying to do, after all. So by my definition, transparency means that nothing is hidden, and that eventually the results of all studies will be released to the general public. (This topic is explored at length in a recent article by Cass Sunstein [2014], entitled
“The Ethics of Nudging.”)
§ You might ask what is magic about twenty-three days? It turns out that in the administrative system, if the bill has not been paid by then, another letter goes out, because the HMRC computers are set up to monitor payment on that date. Running experiments in government requires a good deal of accepting the limitations of what is already being measured.
¶ For example, to my knowledge there has never been a randomized control trial test of Save More Tomorrow. The reason is that we could never get a company to agree to pick some employees at random to be offered the plan and not offer it to others. The closest we came was when we were able to get one company to run different tests at two of its plants, with the other twenty-six plants serving as controls. These trials were not perfect, but we still learned things, for example about the value of educational sessions, but interpretations had to be cautious since the employees selected themselves into the educational sessions. When it comes to running experiments in both government and business, you cannot afford to be a purist.
# This design is not pure random assignment because with only three neighborhoods, it is reasonable to worry that there could be subtle differences in the neighborhoods that might confound the results.
** Reminders are another example of how in many cases nudges are inherently transparent. There is no reason to add, “By the way, the purpose of this text message was to remind you to take your medicine.” Duh!
Conclusion:
What Is Next?
It has now been more than forty years since I first began writing the beginnings of the List on my office blackboard. Much has changed. Behavioral economics is no longer a fringe operation, and writing an economics paper in which people behave like Humans is no longer considered misbehaving, at least by most economists under the age of fifty. After a life as a professional renegade, I am slowly adapting to the idea that behavioral economics is going mainstream. Sigh. This maturation of the field is so advanced that when this book is published in 2015, barring impeachment, I will be in the midst of a year serving as the president of the American Economic Association, and Robert Shiller will be my successor. The lunatics are running the asylum!
But the process of developing an enriched version of economics, with Humans front and center, is far from complete. Here I will say a bit about what I hope will come next, with an emphasis on “hope.” I know better than to forecast how a discipline will change over time. The only sensible prediction is to say that what happens will surely surprise us. So, rather than make predictions, I offer a short wish list for the field’s progress in years to come. Most of the wishes are aimed at the producers of economic research—my fellow economists—but some are aimed at the consumers of this research, be they managers, bureaucrats, football team owners, or homeowners.
Before looking forward to what economics might become, it seems sensible to look back and take stock. Much to everyone’s surprise, the behavioral approach to economics has had its greatest impact in finance. No one would have predicted that in 1980. In fact, it was unthinkable, because economists knew that financial markets were the most efficient of all markets, the places where arbitrage was easiest, and thus the domain in which misbehaving was least likely to appear. In hindsight, it is clear that behavioral finance has thrived for two reasons. First, there are tightly specified theories, such as the law of one price. Second, there is fantastic data that can be used to test those theories, including daily data on thousands of stocks going back to 1926. I don’t know of any other field of economics that would allow for as clear a refutation of economic theory as the story of Palm and 3Com.*
Of course, not all financial economists have renounced their allegiance to the efficient market hypothesis. But behavioral approaches are taken seriously, and on many issues the debate between the rational and behavioral camps has dominated the literature in financial economics for over two decades.
The linchpin for keeping this debate grounded and (mostly) productive is its focus on data. As Gene Fama often says when he is asked about our competing views: we agree about the facts, we just disagree about the interpretation. The facts are that the capital asset pricing model has clearly been rejected as an adequate description of the movements of stock prices. Beta, the only factor that was once thought to matter, does not appear to explain very much. And a pile of other factors that were once supposedly irrelevant are now thought to matter a great deal, although the question of why exactly they matter remains controversial. The field appears to be converging on what I would call “evidence-based economics.”
It would be natural to wonder what other kind of economics there could be, but most of economic theory is not derived from empirical observation. Instead, it is deduced from axioms of rational choice, whether or not those axioms bear any relation to what we observe in our lives every day. A theory of the behavior of Econs cannot be empirically based, because Econs do not exist.
The combination of facts that are hard or impossible to square with the efficient market hypothesis, plus the strong voice of behavioral economists within the field, has made finance the field where claims about the invisible handwave have received the most constructive scrutiny. In a world where one part of a company can sell for more than the entire company, it is clear that no amount of handwaving will suffice. Financial economists have had to take seriously the “limits of arbitrage,” which could just as easily be called the limits of handwaving. We now know more about how and when prices can diverge from intrinsic value and what prevents the “smart money” from driving prices back into line. (In some cases, investors who are aspiring to be the “smart money” can make more money by betting on riding the bubble and hoping to get out faster than others, than by betting on a return to sanity.) Finance also illustrates how evidence-based economics can lead to theory development. As Thomas Kuhn said, discovery starts with anomalies. The job of fleshing out the evidence-based version of financial economics is hardly over, but it is very much under way. It is time for similar progress in other branches of economics.
If I were to pick the field of economics I am most anxious to see adopt behaviorally realistic approaches, it would, alas, be the field where behavioral approaches have had the least impact so far: macro-economics. The big-picture issues of monetary and fiscal policy are vitally important to any country’s welfare, and an understanding of Humans is essential to choosing those policies wisely. John Maynard Keynes practiced behavioral macro, but that tradition has long since withered. When George Akerlof and Robert Shiller, two distinguished scholars who are keeping the behavioral Keynesian tradition alive, tried for several years to organize an annual behavioral macroeconomics meeting at the National Bureau of Economic Research, it was hard to find enough good macroeconomics papers to complete a program. (In contrast, the behavioral finance meeting that Shiller and I coordinate, which is held twice a year, attracts dozens of solid submissions for each meeting, and the process of picking the six to include is difficult.) Akerlof and Shiller eventually abandoned the enterprise.
One reason we are not witness to a thriving group of behavioral economists doing work on macroeconomics may be that the field lacks the two key ingredients that contributed to the success of behavioral finance: the theories do not make easily falsifiable predictions, and the data are relatively scarce. Together, this means that “smoking gun” empirical evidence of the sort that exists in finance continues to elude us.
Perhaps more importantly, this also means that economists do not agree on even the most basic advice about what do to in a financial crisis like the one we experienced in 2007–08. Those on the left take the Keynesian view that governments should have taken advantage of the combination of high unemployment rates and low (or negative) interest rates to undertake infrastructure investments. Those on the right worry that such investments will not be well spent and fear that increasing the national debt will create budgetary crises or inflation down the road. These economists believe that tax cuts will stimulate growth, w
hile the Keynesians think that public spending will stimulate growth. Both sides blame the other for the slow recovery: it is due to either too much or too little austerity. Since we are unlikely to get governments to agree to let recession-fighting policies be picked at random, in order to run randomized control trials, we may never settle this debate.†
Yet the lack of consensus on what constitutes the core “rational” macroeconomic model does not imply that behavioral economics principles cannot be profitably applied to big-picture policy issues. Behavioral perspectives can add nuance to macroeconomic issues even in the absence of a clear null hypothesis to disprove or build on. We should not need smoking guns to get busy collecting evidence.
One important macroeconomic policy begging for a behavioral analysis is how to fashion a tax cut aimed at stimulating the economy. Behavioral analysis would help, regardless of whether the motive for the tax cut is Keynesian—to increase demand for goods—or supply side—aimed at getting “job creators” to create even more jobs. There are critical behavioral details in the way a tax cut is administered, details that would be considered SIFs in any rational framework. If Keynesian thinking motivates the tax cut, then policy-makers will want the tax cut to stimulate as much spending behavior as possible. And one supposedly irrelevant detail these policy-makers should consider is whether the cut should come in a lump sum or be spread out over the course of the year. Without evidence-based models of consumer behavior, it is impossible to answer that question. (When the goal is to stimulate spending, my advice would be to spread it out.‡ Lump sums are more likely to be saved or used to pay down debts.)
The same questions apply to a supply-side tax cut. Suppose we are contemplating offering a tax holiday to firms that bring money home to the U.S. instead of keeping it stashed in foreign subsidiaries to avoid taxation. To design and evaluate this policy we need an evidence-based model that will tell us what firms will do with the repatriated money. Will they invest it, return it to shareholders, or hoard it, as many U.S. firms have been doing since the financial crisis? This makes it hard to predict what firms would do if they found themselves with a greater share of that cash held domestically. More generally, until we better understand how real firms behave, meaning those run by Humans, we cannot do a good job of evaluating the impact of key public policy measures. I will have a bit more to say about that later.
Misbehaving: The Making of Behavioral Economics Page 37