The Nudge is Not Enough! The Love Story Between Behavioral Science and Practical Applications
A couple of weeks ago I published this post on BehavioralEconomics.com Thank you Alain Samson for the invitation.
A romantic relationship goes through various stages from early dating to marriage and, in about half of all cases, divorce. It begins with flirting and continues with that essential first date. If that goes well, it is followed by more dates. If things go OK and the chemistry is good, the relationship will go to the next level: one partner offering the other a shelf in their closet. Sooner than many realize, this leads to the natural question of Why pay two rents? followed by a de-facto living together. After a while, one of the partners pops the BIG question: Will you marry me?
The relationship between academic or theoretical behavioral science (let’s call him THEORY) and applied behavioral science (let’s call her PRACTICE) is not much different from a romantic relationship.
It was quite hard for THEORY to get that first date with PRACTICE, but luckily it happened.
In hindsight, the seminal papers of Kahneman and Tversky on heuristics and biases and on prospect theory published in mid and late 1970s were not enough, at the time, to get PRACTICE to accept the first date.
Fortunately, after about 20 years of flirtation, that first date happened. It was in mid and late 1990s, when Thaler and Benartzi developed and analyzed early implementations of the Save More Tomorrow program which helped (American) employees to save more for retirement by bridging the intention-action gap. In very brief, at every pay raise a person’s savings rate automatically increased (e.g. from 3% to 4%). The automated escalation of savings rates helped most people keep their commitment to save more, while the coupling with pay raises eluded the miserable feeling of losing money out of one’s current paycheck (i.e. loss aversion identified by Tversky and Kahneman).
Occasional dates happened between THEORY and PRACTICE after that, but neither side was taking the relationship too seriously.
The book Nudge (2008) by Cass Sunstein and Richard Thaler showed that THEORY and PRACTICE have a shot at a serious relationship. The establishment of the Behavioral Insights Team (UK Nudge Unit) in 2010 was equivalent to PRACTICE offering a shelf in its closet to THEORY. As in any romantic relationship, THEORY brought in more and more of its things into PRACTICE’s apartment. Now in 2015, they have (almost) de-facto moved in together.
Throughout their relationship, THEORY and PRACTICE have enjoyed making nudges… those small, relatively inexpensive, supposedly irrelevant changes in choice architecture that lead to potentially large changes in behavior – tax collection, college enrollment rate, savings rate, sales etc. Simply put, nudges are small changes that have a large impact on behavior. The result of THEORY and PRACTICE’s union.
However, the Nudge is Not Enough!
Indeed nudges or behaviorally informed interventions have (considerably) improved several areas of public and private services. Most of the time, these small interventions are more than welcomed. Simplifying and structuring choice related information is great simply because everyone hates filling in endless forms and making complicated choices between things they are clueless about (such as Ethiopian food).
Nudges are, most often, great! Nonetheless they are not enough.
The shortcoming of nudges is that most often they are simply tweaks augmenting a pre-existing service or policy.
While they can be beautiful, intriguing and occasionally elegant, nudges are just augmenting (improving) an existing service / policy regardless of that service’s (policy’s) quality, appropriateness or fitness.
For example, an education institution optimizes the choice architecture of its forms in order to smooth the actual application and enrollment processes, resulting in more students joining the institution’s programs. This nudge does not change the service provided. The additional students will attend the exact same courses, go through the exact same stages (from enrollment to graduation) as before the nudge was applied. While for the additional students who joined because of the improved choice architecture attending more education might be beneficial, it is possible for them to be rather unhappy since the courses might be boring and irrelevant.Getting more people into schools or other forms of (adult) education is generally beneficial for everyone involved. We can use behavioral science insights to increase enrollment and decrease drop-out rates. But what if we could use the same knowledge to design better education services?
For example, night-school or other forms of evening-learning are rather popular among adults. However, after a full day at work, System 2 is fatigued and self-control resources are almost depleted. Therefore, it might be a good idea to adapt both the content and teaching methodology to this cognitive reality.
Applying nudges to traffic tickets in order to increase payment compliance (i.e. voluntarily paying the fine) will not solve the issue of traffic safety. If anything, it will continue to feed a carrots-and-sticks approach to influencing human behavior. What if we could use existing knowledge in behavioral science to design safer roads? OK. That would cost a lot of money and will take a lot of time. But what if we could (re-)design insurance services that encourage safer driving behavior?
It is time for THEORY and PRACTICE to take their relationship to the next level: from Nudging to Behavioral Design.
In his book “Slim by Design” Brian Wansink says: it is better to work with human nature than against it. The main thesis of his book is that instead of emphasizing on counting calories and self-control reliant diets, it is much better to (re-)design eating spaces, homes and shops. This way, eating better (healthier) is the natural thing to do and not an eternal fight between temptation and self-control.
In the same line of thought, we can design public policies and (private) services that work with human nature and not against it. While nudges add a (thin) layer of human-friendliness, these behaviorally designed policies and services incorporate behavioral science knowledge in their very core.
Insurance companies truly and deeply hate when their clients have car accidents, leading to expensive repairs, because insurers have to pay the bills. Although this is the very nature of the insurance business, your insurer would love to take your risk of minor accident from 2% to 1.9% and/or have to cover the damage of a broken bumper than that of a full-frontal collision, while at the same time keep on charging you the same $400 / 6 months.
To some extent, behavioral design can create a car-insurance service that promotes safe(r) driving behavior. Part of the risk is purely random, while another part is (to some extent) related to behavior. Behavioral design can address the latter. Car manufacturers are already doing a lot to prevent drivers from not wearing a seat-belt or driving way above the speed limit. Insurance companies, too, can contribute to encouraging preventive behavior.
For example, in Europe cars need to go through regular maintenance and mandatory checks. An insurance company has the possibility of sending out customized reminders (nudges) when the check date is near. Such an approach will decrease the risks associated with unfit vehicles on the roads. Similarly, insurance companies can provide as a default option tracking devices that monitor driving behavior, provide real-time feedback, implement social-benchmarking on risky driving (e.g. 63% of drivers drive safer than you) and offer financial incentives (i.e. lower rates) for safe driving behavior. Moreover, the device can locate the car if it is stolen.
Health and Well-being
Health is a broad area in which nudges are popular — and for very good reasons. There are many examples of nudges for hand-washing, treatment adherence, in-store interventions for purchasing vegetables etc. The major challenge is to design health insurance and health-care services that incorporate behavioral science knowledge in a systemic manner.As in the car-insurance situation, health-insurance companies (public authorities) hate having to pay large bills on treatments for conditions that could have been prevented or are delivered in a sub-optimal manner (e.g. emergency rooms overcrowded by non-emergencies).
For example, the treatment for diabetes is quite expensive and has to occur for a lifetime. Promoting more appropriate eating behaviors in order to prevent the disease actually makes (economic) sense for health-insurance companies (authorities). Nudges can be useful and are welcomed. However, things are a bit more complicated; simple augmentations of pre-existing frameworks might not do. Rather complex preventive programs that have behavioral science at their core are needed. Texting individuals in high-risk (of diabetes) populations reminding them to eat more fruits might be useful. However, a more direct approach such as fruits for junk-food exchange program might do more. Services that provide regular home-delivery of easy to prepare (eat) healthier food already exist (e.g. Hello Fresh) and can be an inspiration for preventive health services provided by insurance companies.
Another behavioral design approach to decreasing health-related expenses and increasing health well-being is to improve the financial well-being of the most vulnerable population groups. It may seem a bit awkward for a health insurance company (authority) to care about the financial well-being of the poor. However, health and financial wellbeing are inter-related to some extent. Moreover, a critical situation in one will lead (sooner or later) to serious problems in the other.
Unlike middle class or more affluent people, the (very) poor cannot absorb financial shocks such as car-repairs, replacing a broken fridge etc. Short-term money lenders are eager to offer loans for such emergencies, but the interest rates are skyrocketing (e.g. 500% per year). Since most of these individuals live from one pay-check to another it is virtually impossible to repay the loan, leading to a vicious cycle of debt and misery. When caught in such a debt-trap, it is very likely that some will neglect their health, eat cheaper and less healthy food, work 16 hours a day etc. All of these behaviors will, ultimately, result in health problems and high healthcare bills.
Offering financial safety-nets for vulnerable categories might be a good idea for preventing serious health-problems and subsequent large medical bills. One solution would be to offer emergency small loans (e.g. up to $1000) with zero interest that would be repaid throughout one year in the health-insurance bill.
Applying nudges – augmenting existing service or policy frameworks – constitutes considerable progress similar to that of going from dating to de-facto living together in a romantic relationship.
Since the relationship looks and feels good, there is no reason for not taking it further. Behavioral science is so rich in potential applications that we should not restrict ourselves to highly effective, yet superficial, applications.Soon the time will come to ask the BIG question:
Will you do Behavioral Design with me?