Illustrations by John Hersey
Published October 24, 2008
Nobel Economics at its Best
In 2008, the Milken Institute Review excerpted a chapter from Cass Sunstein and Richard Thaler’s marvelous then-new book, Nudge: Improving Decisions About Health, Wealth and Happiness (Yale University Press, all rights reserved). Now that Thaler has (deservedly) won a Nobel Prize for his pioneering insights in the field of behavioral economics, we’re proud to make the excerpt available again on MilkenReview.org.
Carolyn is the director of food services for a large city school system. She is in charge of hundreds of schools, and has the authority to determine what hundreds of thousands of kids are offered in her cafeterias each day. Carolyn has formal training in nutrition, and she is a creative type who likes to think about things in non-traditional ways.
One evening she and her friend Adam, a management consultant who has worked with supermarket chains, hatched an idea. They would run some experiments in her schools to determine whether the way the food is displayed influences the kids’ choices. In some cafeterias, the desserts were placed first in line; in others, last. And in still others, they were placed in a separate line. The location of other foods was varied from one school to another. In some, the French fries (but in others, the carrot sticks) were placed at eye level.
From his experience in designing supermarket floor plans, Adam suspected that the impact would be dramatic. He was right. Simply by rearranging the cafeteria display (not changing the mix of foods offered), Carolyn was able to increase or decrease the consumption of many foods by as much as 25 percent. It seems that school children, like adults, can be greatly influenced by small changes in context.
The knowledge can be applied for better or for worse. For example, Carolyn knows that she can increase consumption of healthy foods and decrease consumption of unhealthy ones. With hundreds of schools to work with, Carolyn believes that she now has considerable power to affect what kids eat.
Here are some suggestions she has received from her usually sincere, but occasionally mischievous, friends and co-workers on how she should use her power:
1. Arrange the food to make the students best off, all things (meaning both taste and nutrition) considered.
2. Choose the food order at random.
3. Try to arrange the food to get the kids to pick the same foods they would pick without outside prompting.
4. Maximize the sales of items from the suppliers offering the largest bribes.
5. Maximize profits, period.
Option 1 has obvious appeal, yet it does seem a bit intrusive — even paternalistic. But the alternatives seem worse. Arranging the food at random could be considered fair-minded and principled, and it is in a sense neutral. But if the ordering were randomized across schools, the children at some would have less healthy diets than those at others. Should Carolyn choose that kind of neutrality, if she could just as easily make most students better off by improving their health?
Option 3, trying to mimic what the children would choose for themselves, might seem an honorable way to avoid intrusion. But a little thought suggests that it would be difficult to implement. Adam’s experiment proves that what kids choose depends on the order in which the items are displayed. What does it mean, then, to say that Carolyn should figure out what the students would choose on their own? In a cafeteria, it is impossible to avoid some way of organizing food.
Carolyn is honorable and honest, so she does not give Option 4 any thought. Option 5 would have some appeal if Carolyn thought that the best cafeteria was the one that made the most money. But should she really try to maximize profits, if the result would be less healthy children?
Carolyn has assumed the role of what we call a “choice architect,” someone with the responsibility for organizing the context in which people make decisions. Although she is a figment of our imagination (that’s right, we made her up), many real people turn out to be choice architects — most without realizing it.
If you design the ballot voters use to choose among candidates, you are a choice architect. If you are a doctor describing alternative treatments to a patient, you are a choice architect. If you design the form that new employees must complete to enroll in the company health care plan, you are a choice architect. If you are a parent describing possible educational options to your son or daughter, you are a choice architect. If you sell cars, you are a choice architect (but you surely knew that already).
There are many parallels between choice architecture and more traditional forms of architecture. One is that there is no such thing as a neutral form. Consider the job of designing a new college building. The architect is given some mandates — say, there must be space for 120 offices, 8 classrooms and 12 student-meeting rooms. Hundreds of other constraints, legal, aesthetic and practical, are also imposed. Yet, in the end, the architect must come up with an actual structure with doors, stairs, windows and hallways.
As good architects know, small and apparently insignificant details can have major impact on behavior. In many cases, the power of details comes from pointing users in a particular direction. A wonderful example comes from, of all places, the men’s rooms at Schiphol Airport in Amsterdam. There, the authorities etched the image of a black housefly into each urinal. “It improves the aim,” explains Aad Kieboom. “If a man sees a fly, he aims at it.” Indeed, Kieboom, an economist who directs Schiphol’s building expansion, found that the etchings reduced spillage by 80 percent!
Just as a building architect must eventually settle upon a specific design, a choice architect, like Carolyn, must choose a particular arrangement of the food at lunch — and by so doing, she influences what people eat. In a word, she can nudge.
If, all things considered, you think that Carolyn should take the opportunity to nudge the kids toward food that is better for them (Option 1), we welcome you to our new movement: libertarian paternalism. We are keenly aware that readers will not immediately find the term endearing — both words, after all, are weighed down by stereotypes from popular culture and politics. Even worse, at first hearing, the concepts seem contradictory. But we believe that, properly understood, both are just common sense — and are far more attractive together than alone.
The libertarian aspect of our strategies lies in their insistence that people should be free to do what they like. To borrow a phrase from the late Milton Friedman, libertarian paternalists urge that people should be “free to choose.” We strive to design policies that maintain or increase freedom of choice.
Libertarian paternalists want to make it easy for people to go their own way. They do not want to burden those who want to exercise their freedom.
When we use the term libertarian to modify the word paternalism, we simply mean “liberty preserving.” And when we say liberty preserving, we really mean it: Libertarian paternalists want to make it easy for people to go their own way. They do not want to burden those who want to exercise their freedom.
The paternalistic aspect lies in the claim that it is legitimate for choice architects to try to influence people’s behavior in order to make their lives longer, healthier and happier. In other words, we argue for self-conscious efforts, in both the private and public sectors, to steer choices in directions that will improve peoples’ lives. A policy is paternalistic if it tries to influence choices in a way that will make choosers better off according to their own criteria. Drawing on well-established findings in social science, we show that, in many cases, individuals make pretty bad decisions — ones they would not have made if they had paid full attention and possessed complete information, unlimited cognitive abilities and complete self-control.
Libertarian paternalism is a relatively weak, nonintrusive type of paternalism because choices are not blocked or significantly burdened. If people want to smoke cigarettes, eat a lot of candy, choose an unsuitable health care plan or fail to save for retirement, libertarian paternalists will not force them to do otherwise — or even make them uncomfortable in the process. Still, the approach we recommend does count as paternalistic because choice architects are not merely trying to track or to implement people’s anticipated choices. Rather, they are attempting to move people in directions that will make their lives better. In a word, they nudge.
A nudge, as we use the term, is any aspect of choice architecture that alters behavior in a predictable way without forbidding alternatives or significantly changing economic incentives. To count as a mere nudge, the intervention must be easy and cheap to avoid. Nudges are not mandates: Putting apples at eye level counts as a nudge; banning Twinkies does not.
Many of the policies we recommend have been implemented by private business (with or without a nudge from the government). Employers, in fact, are important choice architects in many of the examples in this book. In areas involving health care and retirement plans, we think that employers can give employees some helpful nudges. Firms that want to do well by doing good can even benefit from environmental nudges – say, by helping to reduce the emission of greenhouse gases. But as we shall show, the same arguments that justify libertarian paternalism on the part of private institutions apply to government as well.
Why Nudges Can Help
Those who reject paternalism often claim that human beings do a terrific job of making choices for themselves. And if they are not terrific, they are certainly better than someone else would make on their behalf — especially if that someone else works for the government. Whether or not they have ever studied economics, many people seem at least implicitly committed to the idea of Homo economicus, or economic man, the notion that each of us thinks and chooses unfailingly well, and thus fits within the stylized picture of welfare-maximizing human beings offered by economists.
Indeed, if you look at economics textbooks, you will learn that Homo economicus can think like Albert Einstein, store as much memory as IBM’s Big Blue, and exercise the willpower of Mahatma Gandhi. Of course, real people have trouble with long division, sometimes forget their spouses’ birthdays and wake up with hangovers on New Year’s Day. They are not Homo economicus; they are Homo sapiens. (To keep our Latin usage to a minimum we will hereafter refer to these imaginary and real species as Econs and Humans.)
Consider obesity. Rates of obesity in the United States are now approaching 20 percent, while another 40 percent are considered overweight. There is overwhelming evidence that obesity increases risks of heart disease and diabetes, and frequently leads to premature death. It would be quite fantastic, then, to suggest that everyone is choosing a diet because they consciously choose to live shorter lives and experience more illness in order to eat the foods they like best.
Of course, sensible people do care about the taste of food, not simply about health. We do not claim that everyone who is overweight is necessarily failing to act rationally. But we do reject the claim that almost all Americans are choosing their diet optimally. And what is true for diet is true for other risk-related behavior, including smoking and drinking. And some of them obviously agree. Why else would so many smokers, drinkers and overeaters willingly pay third parties to help them make better decisions?
But our basic source of information here is the science of choice, which has emerged from research over the past four decades. That research has raised serious questions about the rationality of many decisions that people make. To qualify as Econs, people are not required to make perfect forecasts — that would necessitate omniscience. But they are required to make unbiased forecasts.
That is, their forecasts can be wrong, but they can’t be systematically wrong in a predictable direction.
Take, for example, the “planning fallacy” — the tendency toward unrealistic optimism about the time it takes to complete projects. Hundreds of studies confirm that forecasts are flawed and biased.
Human decision-making is not so great either. Consider what is called the “status quo bias” in the research literature, a fancy name for inertia. For a host of reasons, people have a strong tendency to go along with the default option. When you get a new cell phone, for example, you have a series of choices to make, from the screen background, to the ring tone, to the number of times the phone rings before the caller is sent to voice mail. The manufacturer picks one option as the default for each of these choices. And research shows that people stick with defaults, even when the stakes are much higher than choosing the noise your phone makes when it rings.
Two important lessons can be drawn from this research. First, never underestimate the power of inertia. Second, that power can be harnessed in the effort to improve life. Setting default options and other similar seemingly trivial menu-changing strategies can have huge effects on outcomes in everything from increasing retirement savings, to improving health care, to providing organs for transplant surgery.
A nudge, you’ll remember, is any factor that significantly alters the behavior of Humans, even though Econs would ignore it. Econs respond primarily to incentives. If the government taxes candy, Econs will buy less. But they are not influenced by such “irrelevant” factors as the order in which options are displayed. Humans respond to incentives, too, but they are also influenced by nudges.
Alert readers will notice that incentives can come in different forms. Steps taken to increase people’s cognitive effort – as in relegating the Snickers bars to a lower shelf in the store — might be said to increase the “cost” of choosing the candy. Some of our nudges do impose cognitive (rather than material) costs, and, in that sense, alter incentives. But nudges qualify as libertarian paternalism only if such costs are low.
A False Assumption And Two Misconceptions
Many people who favor freedom of choice reject any kind of paternalism. The standard policy advice is to give people as many choices as possible, and then let them choose the one they prefer. The beauty of this way of thinking is that it offers a simple solution to many complex problems. Just Maximize Choices — full stop!
In some circles, Just Maximize Choices has become a policy mantra. Sometimes the only alternative is thought to be a government mandate, derided as One Size Fits All. The false assumption here is that almost all people almost all of the time make choices that are in their best interest — or, at the very least, make choices that are better than those that would be made for them. We do not think that, on reflection, anyone really believes this.
Consider a chess novice playing against an experienced player. The novice loses precisely because he makes inferior choices — choices that could easily be improved by helpful hints. In many areas, ordinary consumers are novices, interacting in a world inhabited by experienced professionals trying to sell them things. It seems reasonable that people make good choices in contexts in which they have experience, good information and prompt feedback — say, in choosing ice cream flavors. People know whether they like chocolate, vanilla, coffee or something else. They do less well in contexts in which they are inexperienced and poorly informed, and in which feedback is slow or infrequent — say, in choosing between fruit and ice cream (where the effects are only noticeable years later and feedback is poor).
The first misconception alluded to above is that it is possible to avoid influencing choices. In many situations, some agent must make a choice that will inevitably affect behavior. It is true, of course, that some nudges are unintentional. Employers, for example, may decide whether to pay employees monthly or biweekly without intending to create any kind of nudge. Indeed, we suspect that most would be surprised that workers save more if they get paid biweekly because twice a year they get three paychecks in one month.
It is also true that private and public institutions can strive for one or another kind of neutrality — as, for example, by choosing randomly or by trying to figure out what most people want and increasing the prospects they will get it. But unintentional nudges can have major effects, and in many contexts, these forms of neutrality are unattractive.
Some people will happily accept this argument in the case of private institutions, but strenuously object to government efforts to influence choice with the goal of improving lives. They worry that governments cannot be trusted to be competent or benign. They fear that elected officials and bureaucrats will place their own interests first, or accept the goals of self-interested private groups. We share these concerns. In particular, we emphatically agree that, for government, the risks of mistake, bias and overreaching are sometimes serious.
We favor nudges over commands, requirements and prohibitions in part for that reason. But governments, no less than cafeterias (which governments frequently run), have to provide starting points of one or another kind. In this respect, the anti-nudge position is unhelpful — a literal nonstarter.
The second misconception is that paternalism always involves coercion. In the cafeteria example, the choice of the order in which to present food does not force a particular diet on anyone. Yet Carolyn might select some arrangement on grounds that are paternalistic in the sense that we use the term. Would anyone object to putting the fruit before the baked goods at an elementary school cafeteria if it induced kids to eat more apples and fewer Oreos?
Is the answer fundamentally different if the customers are teenagers, or even adults? Since no coercion is involved, we think that some types of paternalism should be acceptable even to those who most embrace freedom of choice. Freedom to choose is the best defense against bad choice architecture.
Choice Architecture In Action
Choice architects can improve lives by designing user-friendly environments. Many of the most successful companies have helped people, or succeeded in the marketplace, for exactly that reason.
Sometimes the choice architecture is highly visible, and consumers are much pleased by it. The iPod and the iPhone are good examples because they are not only elegant, but make it easy for the user to get the devices to do what they want. Sometimes, though, the architecture is taken for granted and could benefit from careful attention.
Consider Thaler’s employer, the University of Chicago. The university has an “open enrollment” period every November, when employees are allowed to revise their selections of benefit plans ranging from health insurance to retirement savings. Employees must make their choices online after receiving a package of materials explaining the alternatives along with instructions on the mechanics of making their picks.
But some neglect to make a choice. So default options matter — a lot.
To simplify, assume two alternatives: those who make no active choice can be given the same choice they made the previous year, or their choices can be set back to “zero.” Suppose that last year Janet contributed $1,000 to her retirement plan. If Janet makes no active choice for the new period, one alternative would be to default her to a $1,000 contribution; another would be to default her to a zero contribution. Call these the “status quo” and “back-to-zero” options.
How should the choice architect decide between these defaults? Libertarian paternalists would like to set the default by asking what reflective employees in Janet’s position would actually want. Although this principle may not always lead to a clear choice, it is certainly better than setting the default randomly, or by making either status quo or back to zero the default for everything.
For example, it is a good guess that most employees would not want to cancel their heavily subsidized health insurance. So, for health insurance, the status quo default seems strongly preferred to the back-to-zero default.
Compare this to a tax-exempt “flexible spending” account, in which an employee sets aside money each month that can be used to pay for specific services such as uninsured medical procedures or child care.
Money put into this account has to be spent each year, or it is lost. Note, too, that an employee’s qualifying expenditures might vary greatly from one year to the next. So here, the zero default probably makes more sense than the status quo.
This problem is not merely hypothetical. We once met with three top administrators of the University of Chicago to discuss similar issues, and the meeting happened to take place on the final day of the employees’ open enrollment period. We asked whether the administrators themselves had met the deadline. One said that he was planning on doing it later that day, and was glad for the reminder. Another admitted to having forgotten; the third said that he was hoping that his wife had remembered to do it!
The group then turned to the question of what the default should be for a supplementary “salary reduction” program (jargon for a tax-deferred savings program). To that point, the default had been the back-to-zero option. But since contributions to this program could be stopped at any time, the group unanimously agreed that it would be better to switch to the status quo default. We are confident that many absent-minded professors will have more comfortable retirements as a result.
A New Path
Some of the most important applications of libertarian paternalism are to government. Our hope is that our policy recommendations will appeal to both sides of the political divide. Indeed, we believe that libertarian paternalism can be embraced by Republicans and Democrats alike.
For one thing, many of those policies cost little or nothing, so they impose no burden on taxpayers. For another, many Republicans are now seeking ways to go beyond simple opposition to government action.
As Hurricane Katrina showed, government is sometimes required to act because it is the only practical means for mustering and deploying resources. Republicans want to make people’s lives better; they are simply skeptical — and legitimately so — about eliminating people’s options.
For their part, many Democrats are willing to abandon their traditional enthusiasm for aggressive government intervention. They have come to agree that freedom of choice is a good, even indispensable, foundation for public policy. There is thus real hope here for crossing partisan divides.
Actually, we have evidence that our optimism is more than just rosy thinking. Libertarian paternalism with respect to tax-deferred private savings plans has received enthusiastic bipartisan support in Congress. And in 2006, some of the key ideas for nudging people to save more were quietly enacted into law.
Libertarian Paternalism: An Application
Prescription drug coverage for seniors was a hot topic during the 2000 presidential campaign. Al Gore proposed a classic government mandate: add drug coverage to Medicare, assemble experts to work out the specifics, and then offer the single plan to all seniors. George W. Bush countered with an expensive new entitlement to drugs at subsidized prices — but one that featured a wide variety of alternative plans devised by private health insurers.
Three years later, Congress narrowly adopted Medicare Part D, a version of President Bush’s plan. “The reason why we felt it was necessary to provide choices is because we want the system to meet the needs of the consumer,” Bush told a group of Florida seniors in 2006 as the plan was about to be rolled out. “The more choices you have, the more likely it is you’ll be able to find a program that suits your specific needs … I believe in consumers, I believe in trusting people.”
President Bush’s confidence in the wisdom of seniors left them with a great deal of decision-making responsibility. But this was no laissez-faire system: the government did set minimum coverage requirements for plans. In fact, this system of constrained free choice might seem like a nice example of libertarian paternalism in action. As choice architecture, however, Part D suffered from a cumbersome design that impeded good decision making. It offered a menu with lots of choices — which is fine. However, the plan suffered from four major defects:
1. It gave participants little guidance to help them make selections from the menu.
2. Its default option for the majority of seniors was non-enrollment.
3. It mandated a randomly chosen default for the 6 million people who were automatically enrolled in Part D, and created barriers to third-party efforts to match people and plans based on their past consumption of prescription drugs.
4. It failed to enlist many among the most vulnerable population – the poor and the poorly educated.
Do not misunderstand: Part D has done a lot of good. But there is plenty of room here for better choice architecture.
The Design of Part D
Before the enactment of the law, about half of all American seniors — some 21 million — had some form of prescription drug coverage through private plans or from a government agency such as the Department of Veterans Affairs. Government planners had high hopes of covering the rest through Part D. The working principle was to provide seniors with as many choices as possible. The result was an approach built around six key features.
1. For most people, Part D is voluntary; you receive benefits only if you enroll. The exception is some 6.2 million low-income seniors and disabled people who were covered by Medicaid, the medical insurance program for the poor. These two groups are supposed to choose from a subset of the approved private plans — namely, the cheapest, most basic plans meeting government mandates. In 2007, between five and 20 such plans were offered in each state. Any senior covered by Medicaid who does not make a choice is enrolled randomly in one of these approved basic plans.
2. Seniors who do not enroll when they become eligible, and who lack a comparable private plan, face a penalty on the premiums they pay when they do enroll.
3. Seniors can enroll in a stand-alone prescription drug plan or a joint Medicare-Prescription Drug plan. (Joint plans are for those enrolled in Medicare Advantage, special privately operated plans that generally provide more benefits than the traditional Medicare program, but limit doctor choice.) They are advised to “rely on advice from people you know or trust,” to “choose a plan you are already familiar with,” or to use the guide on the Medicare Web site.
4. Plans differ across states. Most offer between 50 and 60 stand-alone plans and between 15 and 142 joint plans. The total number of approved plans has increased since the law was enacted.
5. The government (with help from independent non-profit organizations) spent $400 million on a campaign to encourage those eligible to choose a plan during the initial enrollment period. Medicare officials traveled the country in a giant blue bus to promote the program.
6. Coverage starts with the first prescription a patient needs. But it then stops for a while after the patient has spent a threshold amount, only to start up again when another spending threshold is reached. This coverage gap is described as the “doughnut hole.”
“If consumers are up to this task, then their choices will ensure that the plans and insurers that succeed in the market are ones that meet their needs,” wrote Daniel McFadden, a Nobel Prize-winning University of California economist who has studied Part D extensively. “However, if many are confused or confounded, the market will not get the signals it needs to work satisfactorily.” And with so many complex plans to choose from, it should not be a huge surprise that seniors have had a difficult time sending those right signals.
Confusion Awaiting Clarity
As the initial six-month window for enrolling in Part D was closing, people were struggling to sign up. Consider the experience of seniors in McAllen, Texas. McAllen is a town of over 100,000 in the Rio Grande valley near the Mexican border — the kind of poor town (about one-fifth of the seniors live in poverty) intended to benefit hugely from Part D.
To obtain benefits, however, seniors in Texas needed to wade through 47 approved plans. “The new Medicare program is a full-time job,” explained Ramiro Barrera, a co-owner of Richard’s Pharmacy in nearby Mission, Texas. “We are swamped with requests for help from beneficiaries.”
The experience in McAllen was hardly unique. Saturday Night Live spoofed the maze of detail in a phony public service commercial, promising a plan for seniors that was easy to understand, provided they had already mastered their computers and satellite television controls. President Bush sympathized with the frustration, but said that the gain would ultimately be worth the pain. “I knew that when we … laid out the idea of giving seniors choices, it would create a little confusion for some,” he told the audience in Florida. “I mean, after all, up to now there hadn’t been … many choices in the system, and all of a sudden, [for] a senior who feels pretty good about things [here comes] old George W. — and all of a sudden 46 choices pop up.
“We encouraged all kinds of people to help. … AARP is helping; the NAACP is helping; sons and daughters are helping. …I readily concede some seniors have said, ‘there are so many choices, I don’t think I want to participate.’ My advice is, there is plenty of help for you.”
The impulse here was commendable, but you have now read enough to know that offering people 46 choices and telling them to ask for help is likely to be about as good as no help at all. And in Medicare Part D’s case, many of the groups meant to assist seniors were puzzled themselves.
The confusion spread to medical professionals, who agreed with their patients that the number of plans in the current program was bewildering. Others, including AARP, decided to offer their own plans as well as giving advice on selections – a pretty obvious conflict of interest.
In the end, getting seniors into a plan turned out not to be the biggest problem. As of January 2007, fewer than 10 percent of all Medicare beneficiaries – about 4 million — lacked drug coverage. Federal health officials touted the high enrollment as evidence of seniors’ capacity to navigate the complex Part D landscape. And there is at least some evidence they were right.
In November 2005, when seniors were getting a first taste of the system, half of 1,800 seniors surveyed had an unfavorable view of the program, compared with 28 percent who viewed it favorably. By November 2006 the unfavorable rating had fallen to 34 percent, while the favorable rating had risen to 42 percent. When asked about their own experiences, three out of four held a “very” or “somewhat” positive view of Part D.
Of course, it is true that education can make complicated choices easier. But we think that there has been a lot less learning about Part D than a casual look suggests.
Advertising campaigns and media coverage certainly boosted awareness. But no one should read the statistics and conclude that 38 million seniors filled out Part D applications because the government asked them nicely.
For starters, the high enrollment rates were, in part, a consequence of the fact that approximately two-thirds of seniors were enrolled through routes in which they got a lot of guidance — employer or union plans, Medicaid, Veterans Affairs, federal employee coverage or Medicare Advantage. Advertising campaigns and media coverage certainly boosted awareness. But no one should read the statistics and conclude that 38 million seniors filled out Part D applications because the government asked them nicely.
Note, too, that many people are still not enrolled in the program, even though it is clear that they should be. Four million may be a relatively small percentage of seniors. But it is a large number in absolute terms, and studies suggest that this unenrolled group is probably dominated by poorly educated people living just above the poverty line. In addition, one-quarter of the 13.2 million seniors eligible for an extra subsidy because they have low incomes – again, most of them poorly educated and living alone — did not take advantage of the special offer. And since coverage for this group is practically free when the subsidy is added in, one can assume that most would have been better off in the program.
Even when people do elect to enroll, there is evidence than an abundance of choice can overwhelm them. After a year of experience in the drug program, only about one in 10 seniors said it needed “no real changes.” And in November 2006, 60 percent of seniors surveyed agreed with the statement that the government should “select a handful of plans … so seniors have an easier time choosing.”
This view is even more strongly held by the medical community. More than 90 percent of doctors and pharmacists, who had been bombarded with patient questions throughout the enrollment period, agreed that the program was too complicated.
Such responses suggest that consumer satisfaction could be a lot higher if the program were designed better. Complexity is the most glaring problem. But it is not the only one. In fact, two other pieces of Part D’s choice architecture are equally problematic.
Random Default Plans for the Most Vulnerable
In the introduction, we discussed the options faced by our fictitious cafeteria supervisor Carolyn. One was to display food items at random. We said that this option could be considered fair-minded. But it didn’t strike us as desirable because it would penalize some students by inducing them to consume a diet of pizza, egg rolls and ice cream.
Still, this is the option the government adopted for six million of its poorest and sickest citizens. It automatically assigned those in Medicaid who did not pick a plan to one randomly chosen from all plans with premiums below certain benchmarks for the region. As a result of plan restructuring, another 1.1 million people were eligible for random assignment in 2007. One state, Maine, shrewdly resisted this system in favor of an “intelligent assignment” process for 45,000 people. We will return to shrewd Maine shortly; for now, we focus on the other 49 states.
The poorest and sickest enrollees are those people eligible for both Medicare and Medicaid (“dual eligibles”). Dual eligibles are more likely to have diabetes and strokes than other Medicare beneficiaries, and they use, on average, 10 or more prescription drugs. They include the most severely disabled Americans, physically and cognitively handicapped men and women of all ages, and elderly patients suffering from dementia and requiring full-time care. The government has not said how many dual eligibles actively chose a plan, but the evidence suggests that very few did. Dual eligibles are able to switch plans at any time — but if few are actively choosing, we suspect few are taking advantage of the switching option.
Random assignment can cause random harm to people placed in plans that don’t fit their needs. For the drugs that dual eligibles take most often, and that are in categories covered by the law, plans varied considerably in their coverage, from as low as 76 percent to as high as 100 percent. This means that some dual eligibles were defaulted into a plan that did not cover the drugs they use most. They could switch, of course; but most stayed with the plan that had been lovingly picked at random for them. And given the patchy drug access, it is not surprising that random plan defaults impaired people’s health. In a recent survey of dual eligibles, 22 percent said they had stopped taking medications temporarily or permanently because of problems in managing the new plans.
The government’s official reason for rejecting intelligent assignment in favor of random assignment is that people’s prescription needs change: Someone’s past use is no guarantee of future needs. There has been a lot of head-scratching about this rationale in the health care community, since last year’s drug use is often an excellent predictor of next year’s, especially for the chronically ill. And certainly it is a better predictor than picking a plan out of a hat.
It seems somewhere between callous and irresponsible to assign plans without even looking at people’s specific needs. Random assignment is also inconsistent with the market-based philosophy of the plan. In markets, after all, better products get a higher share of sales. We do not think that every automobile manufacturer should get the same market share, any more than we think that families should pick their cars at random. Why should we want randomness for insurance plans?
How costly were the misallocations caused by this random assignment? One way to answer the question is to see how many people chose to switch plans after the first year. (Every November there is an open enrollment period when participants can switch.)
Unfortunately, we don’t know as much as we’d like to about plan-switching because the government has been unwilling to release the data in a timely manner. It did, however, announce that during the open enrollment period for 2007, about 2.4 million — 10 percent of Part D enrollees — changed plans. But of those who changed, 1.1 million were low-income beneficiaries, most of whom were moved by the government so that they would not have to pay increased premiums. This means that, excluding dual eligibles, only 6 percent chose to change plans.
There are two possible interpretations of these low switching rates. One, favored by defenders of the plan, is that all is going well — that seniors have chosen the best plans for their needs. The second, more plausible interpretation is that inertia and the status quo bias are keeping people from switching.
How can we tell which interpretation is right? One way is by comparing the participants who actively chose their own plan with those who had a plan picked at random for them. For the latter group there can be no presumption that the plan they started with was the best one. And the fact that we find low switching rates for both groups suggests that most participants find the time and energy it takes to decide on the best plan is just not worth the effort.
Is it, in fact, worth that effort? The answer depends on how varied the plans are, and how costs differ depending on the set of drugs people use. Consider one study of the prices of drugs covered by basic plans (the kind that poor beneficiaries would be defaulted into) in three regions of the country. The researchers reported savings between $5 and $50 per drug per month when individuals are assigned to the lowest-cost, best-fitting basic drug plan.
More data comparing entire plans (as opposed to individual drugs) should be available soon, and we expect they will confirm results that other academic teams are beginning to find: Choosing the right plan, rather than a random plan, has the potential to save both seniors and the government a lot of money. If hundreds of dollars were at stake for every person, many seniors would find it worthwhile to spend at least an hour or two sorting out the best plan (much as they would in choosing a new washing machine or putter).
Unfortunately, spending an hour or two wouldn’t get the job done. The chief tool that people have to help choose a plan is the Medicare Web site. But there is an obvious problem with relying heavily on a Web site: Most seniors do not yet use the Internet, and those who do are rarely Web-savvy. Most, in fact, get their information about Part D passively from mailings from insurers, the government and non-profits like AARP. Those mailings are unlikely to contain personalized information. So the Web site, with its tools for personalization, is the best potential source for help. To whom does the job of navigating the site fall? To seniors’ adult children, of course.
An economist friend of ours, Katie Merrell, does research on health coverage and took it upon herself to choose plans for her elderly parents. She found that the task took hours, even for an expert. Katie allowed us to see how painful choosing a plan would be by providing a list of the drugs her mother takes. Thaler logged on to the Medicare Part D Web site and tried his luck.
What a nightmare! Just to give one example, the site does not have a spell-checker. If you type “Zanax” instead of “Xanax,” you don’t get any help. This is a problem because drug names resemble strings of random letters, so typing errors are to be expected.
Navigating dosages is also tricky. You need to know both the size of the dosage per pill and how frequently it is taken. By the same token, the Web site assumes you take a generic drug, if it is available, and gives you the option of keeping the premium brand drug. Many people, however, take generics while still calling them by the brand name, requiring close attention to every drug selection.
Once a user manages to get all the data entered, the Web site offers three plan suggestions, with annual cost estimates. Technophobic seniors can call 1-800-MEDICARE and have a customer service representative give them the three plan suggestions. But no explanation is offered for how these plans have been chosen.
Eventually (with help from Katie that bordered on psychotherapy), Thaler managed to get some answers — though not the same ones that Katie got. Still, because Thaler is nearing Medicare age himself, he thought perhaps someone younger would have an easier time of it. So we asked one of our graduate student research assistants to give it a try. Being younger and more patient helped, but he got yet another set of answers. We then pulled out all the stops and put the youngest, smartest member of our team on the job — our student intern (and Jeopardy! Teen quiz show whiz), who was headed for a top college that fall. Even she, who normally finds everything easy, was befuddled at times. And no two of us, though armed with the same data, ended up with the same cost estimates or plan choices.
At first, we were stumped. But later we found that we all got different estimates because prescription drug plans are constantly updating their drug prices. There is no guarantee that the cheapest plan for your mother today would be the cheapest plan for your mother tomorrow. In fact, Consumers Union has tracked price differences in five large states and found frequent changes. Sometimes these fluctuations are only a few dollars; sometimes much more. Nearly 40 percent of the 225 plans underwent changes of more than 5 percent, which can add up to several hundred dollars per year.
How hard is it to choose the right plan? The short answer: really hard. For the sake of argument, ignore decisions about whether to enroll in Medicare Part D, or whether to enroll in a stand-alone drug plan or a Medicare Advantage plan. Assume that you, like most enrollees, are picking a stand-alone plan.
You’ll need to compare plans along 15 dimensions. True, the Medicare Web site tries to help seniors sort plans across some of these dimensions. But even if you run the Web site gauntlet to find the three cheapest plans available, you shouldn’t breathe easy. You will not be able to tell whether prior authorization of prescriptions will be hard to obtain in your situation, or what the quantity limit on a particular drug will be. This information is probably available only after you sign up for a plan and attempt to fill prescriptions.
Figuring out whether seniors are making good choices would require information about their health and their plans. Given the obvious concerns about privacy, the government has not released these data. But it says, and apparently believes, that seniors are making good choices.
We are not so sure. In an experiment, the economist Daniel McFadden and his team attempted to evaluate how good (or bad) seniors’ choices turn out to be. McFadden’s team radically simplified the simulated selection process. Seniors were offered only four options. To make the choice even easier, a person’s particular economic circumstances were also ignored. The four plans offered were worth the same amount of money. They differed only in the level of protection provided as drug bills rose.
Even in this simplified environment, a high percentage of seniors made poor choices among the four available plans because they failed to connect their choices to their actual health, prescription use and attitude toward risk. In all, nearly two-thirds of enrollees failed to choose the plan that minimized their out-of-pocket costs.
Possible Nudge: Intelligent Assignment
Random default assignment is a terrible idea. If a poor person is assigned to a bad plan and does not switch, she may decide to stop taking an expensive drug. This may save the government money in the short run. But it will be costly in the long run, especially for diseases such as diabetes, for which a failure to stay on the regime can lead to serious complications. The government also pays more if it assigns someone to one plan if another covers all that person’s drugs, yet costs less.
Yet, according to the Government Accountability Office, under random assignment only one-third of the beneficiaries were placed in plans that covered all of their recently used drugs, and one-quarter ended up in plans that covered fewer than 60 percent of those drugs.
Maine, as noted earlier, went its own way. To match each eligible participant with a plan, the 10 plans meeting state coverage benchmarks were evaluated according to three months of historical data on prescription use. Participants in plans covering fewer than 80 percent of their required drugs were switched automatically (with participants retaining the option to cancel the reassignment).
Another set of participants received letters informing them that better matches existed, and were advised to contact state officials for more information. Intelligent assignment switched more than 10,000 people — 22 percent of all the dual eligibles — and produced dramatic results. Maine officials now say that every dual eligible is in a plan that covers 90-100 percent of her required drugs.
Maine was not the only state interested in intelligent assignment. In 2005, two leading pharmaceutical groups, the National Association of Chain Drug Stores and the National Community Pharmacists Association, collaborated with a Tampa-based health care information technology company, Informed Decisions, to develop software that matched people with plans.
The consortium’s presentations to federal government officials were met politely but coolly. (Perhaps its advocates should have called it “intelligent design.”) As a result of skepticism from Washington and legal challenges from insurers, intelligent assignment is used to place dual eligibles only in Maine. Other states should be encouraged to experiment with similar methods. More important, the law mandating random assignment should be revised.
Possible Nudge: RECAP
Seniors could be helped a lot if our own RECAP (Record, Evaluate, Compare Alternative Prices) system were applied to Medicare. RECAP would also make using the Medicare Web site a snap (well, relatively speaking).
Here’s how it would work. Once a year, just before the Part D enrollment period opened, insurers would be required to send seniors a complete, itemized list of all the drugs they used over the previous year and all the fees they incurred. Insurers would also have to provide an electronic summary of their complete pricing schedule to anyone who wanted it. The information would be made available online, so it could be imported into both the Medicare Web site and comparison pricing programs that could now easily be offered by third parties. The purpose of the information would be to nudge seniors away from a status quo bias and encourage comparison shopping by making prescription drug costs as easily discovered as possible.
Because the costs of delay are high for large majorities of seniors, similar nudges could be used on eligible non-enrollees. Price disclosures could be sent to those seniors who delayed enrollment, highlighting the recent and current premiums for a sample of popular plans. One goal would be to show seniors how much money a delay costs them.
We believe requiring providers to offer RECAP reports would stimulate the development of an industry to facilitate plan evaluation. One Massachusetts company, Experion Systems, has already developed an online prescription drug plan assistant tool that is a more user-friendly version of the government Web site’s form. Experion has joined with the pharmacy chain CVS to make it possible to download the sort of information we would want in RECAP reports.
The RECAP information could also be used to improve intelligent assignment programs. One research team has produced some preliminary evidence that a RECAP-style nudge has promise.
In a study of Wisconsin Medicare beneficiaries, the team estimated that if people moved from their current plans to the lowest-cost plans that met their needs, they could save, on average, about $500 a year.
To see whether people would take advantage of these savings if they were nudged, the researchers mailed a personal letter to a random sample of study participants who had agreed to share their personal drug histories. The letter explained the costs in their current plan, the cheapest comparable plan, and the savings they could realize by switching plans. Another random sample of participants received generic Part D brochures instead. Both mailings contained the Internet address of the Medicare plan finder Web site and information about how to use it.
The personal letters appear to have nudged more people to pick lower-cost plans. The overall switch rate among seniors receiving personal letters was 27 percent, 10 percentage points higher than among those receiving brochures. More than three times as many letter receivers picked the cheapest plan — the one mentioned in the letter — although the overall percentages were still in single digits. These results are consistent with other studies showing that people are erring in their choices among plans, and that clear information can reduce those errors.
As a general rule, the more choices there are and the more complex the issue, the more important it is to have enlightened choice architecture. It’s hard to imagine a government program that fits the category better than the Medicare drug benefit.