8

IMPROVING THE LIVES OF THE POOR

During World War II, the United States military was troubled by the recurrence of “wheels-up” crashes: after landing, pilots would retract the wheels instead of the flaps. And, as you can imagine, retracting a plane’s wheels while on the ground is not a good idea. To solve the problem, they brought in an expert. Lieutenant Alphonse Chapanis was a psychologist by training, ideally suited to get inside these pilots’ heads. Why were they so careless? Were they fatigued? Were they relaxing too soon, thinking they could “let go” after a stressful mission? Was it a problem of training?

One clue quickly surfaced: the problem was limited to bomber pilots, those flying B-17s and B-25s. Transport pilots did not make this mistake. This clue helped Chapanis break free of his own biases. He decided not to look inside the pilots’ heads but instead inside their cockpits. In these bombers, the wheel controls and the flap controls were side by side and looked nearly identical. Transport planes, by comparison, had very different controls. What separated the bomber pilots from the transport pilots were the cockpits. One type of cockpit made it too easy to make a mistake.

This experience transformed how cockpits are designed. Chapanis and others came to realize that many pilot errors were really cockpit errors. Until then, the focus had been on training pilots and ensuring alertness, on producing “excellent pilots” who make few mistakes. But Chapanis’s conclusions changed this. Of course pilots must be trained; of course you must select for the best. But no matter how well you train them or pick them, they will make mistakes, especially if put in confounding contexts.

Error is inevitable, but accidents are not. A good cockpit design should not facilitate mistakes and, more important, should prevent errors from becoming tragedies. Chapanis solved the bombers’ problem by placing a small rubber wheel on the end of the landing gear lever so the pilots could tell which lever they were touching. A good cockpit provides feedback in case one might make a mistake. A low-altitude alarm next to the altimeter helps to ensure that a low-flying pilot actually intends to fly low. Planes are much safer today not just because we have built better wings or engines but also because we have gotten better at handling human error.

POOR BEHAVIOR

Chapanis started off stymied by the pilots’ behavior. Many analysts are similarly stymied by the behavior of the poor. Low-income training programs in the United States, for example, suffer from absenteeism, dropouts, and a failure by the intended recipients to sign up. Microfinance programs in the developing world bemoan the fact that their clients do not invest enough in high-return activities: instead, loans are used to pay off other debts, to fight “fires” (like school fees that have come due), or simply to buy consumer durables. And vaccination programs suffer when people fail to show up to get vaccinated, with the result that debilitating but preventable illnesses still rage through much of the developing world.

We have seen this in our own work. We once served as advisers to a welfare-to-work program in the United States that sought to help men and women on public assistance find jobs. One of the biggest challenges were the clients themselves. Despite repeatedly being advised to report to the worksite in professional clothes, they would often show up not wearing the right clothing. Many had substandard résumés, badly formatted and with typos. While sometimes this was due to lack of knowledge or skill, much of it was a failure to follow through, to execute as planned. Even after receiving instructions, few would avail themselves of the computers on site to format their résumés or of the offers to procure more appropriate clothing. When interviews were finally scheduled, clients would arrive without résumés and would not bring their “A” game. In many cases they simply failed to show up.

But the designers of these social programs rarely take the perspective that Chapanis took. Rather than look inside the cockpit, they have assumed that the problem lies with the person. They assume the problem is a lack of understanding or of motivation. So they follow up with attempts to educate or to sharpen incentives. In developed countries, this leads to a discussion of a “culture of welfare.” One solution has been to place a lifetime limit on the number of years that a person can receive welfare. This is driven by a simple impulse: to motivate the unemployed to look for work. It has also led to the chastising of aid programs, and it has occasionally motivated public officials to move away from simple transfers—for example, by charging people for clean water rather than giving it to them for free. It has also occasionally led to programs with strong incentives, such as conditional cash transfer programs, where the amount of aid one receives depends on performing assorted “good” behaviors.

But why not look at the design of the cockpit rather than the workings of the pilot? Why not look at the structure of the programs rather than the failings of the clients? If we accept that pilots can fail and that cockpits need to be wisely structured so as to inhibit those failures, why can we not do the same with the poor? Why not design programs structured to be more fault tolerant?

We could ask the same question of anti-poverty programs. Consider the training programs, where absenteeism is common and dropout rates are high. What happens when, loaded and depleted, a client misses a class? What happens when her mind wanders in class? The next class becomes a lot harder. Miss one or two more classes and dropping out becomes the natural outcome, perhaps even the best option, as she really no longer understands much of what is being discussed in the class. A rigid curriculum—each class building on the previous—is not a forgiving setting for students whose bandwidth is overloaded. Miss a class here and there and our student has started a slide from which she is unlikely to recover. The programs’ design presumes that if people are motivated enough, they will make no mistakes. Those who cannot be bothered to get to class on time, goes the implicit argument, must not care: they do not “deserve” the training.

But the psychology of scarcity predicts that errors like this will be all too common, perhaps even unavoidable, no matter how motivated the person. Imagine you come home from a day at work, worried about where you will find the money to make this month’s rent, cover all the bills, and pay for your daughter’s birthday party. You have not been sleeping well. A few weeks ago, you signed up for a training program in computer skills that one day could help you move up to a better job. But this evening the benefits of such training are abstract and distant. You’re exhausted and weighed down by things more proximal, and you know that even if you go you won’t absorb a thing. Now roll forward a few more weeks. By now you’ve missed another class. And when you go, you understand less than before. Eventually you decide it’s just too much right now; you’ll drop out and sign up another time, when your financial life is more together. The program you tried was not designed to be fault tolerant. It magnified your mistakes, which were predictable, and essentially pushed you out the door.

But it need not be that way. Instead of insisting on no mistakes or for behavior to change, we can redesign the cockpit. Curricula can be altered, for example, so that there are modules, staggered to start at different times and to proceed in parallel. You missed a class and fell behind? Move to a parallel session running a week or two “behind” this one. Miss a module and you can get back on track on the next round. Sure, it will take you a bit longer to finish, but at least you will get there. As it is, training programs are built with no mistakes in mind, as if the participants are not expected or allowed to stumble. But the poor—even, or perhaps especially, when they are unemployed—have a lot going on. And much of it does not sit so well with being a student. Skipping class in a training program while you’re dealing with scarcity is not the same as playing hooky in middle school. Linear classes that must not be missed can work well for the full-time student; they do not make sense for the juggling poor.

It is important to emphasize that fault tolerance is not a substitute for personal responsibility. On the contrary: fault tolerance is a way to ensure that when the poor do take it on themselves, they can improve—as so many do. Fault tolerance allows the opportunities people receive to match the effort they put in and the circumstances they face. It does not take away the need for hard work; rather, it allows hard work to yield better returns for those who are up for the challenge, just as improved levers in the cockpit allow the dedicated pilot to excel. It is a way to ensure that small slipups—an inevitable consequence of the bandwidth tax—do not undo hard work.

INEFFECTIVE INCENTIVES

Remember the lifetime limits on welfare payments discussed earlier? They were based on a belief that cycling in and out of welfare was due to a lack of motivation on the part of the poor. People went on and off of welfare, it was said, because the system made it too easy not to work. To fix this, in the United States a lifetime cap was imposed for the primary welfare program (now renamed Temporary Assistance for Needy Families). A person could now only be in the program for a total of five years over her lifetime.

A lifetime limit may not be foolish. Limits create scarcity, the logic goes, which might lead to better management of how the resource is “used.” This almost relies on the psychology of scarcity. But it is flawed. We have seen that deadlines work when they are pressing, when they are top of mind. A long-term limit, like a distant deadline, becomes pressing only as it approaches, toward the end. To those who are currently juggling and tunneling, the limit, years away, will reside outside the tunnel, until it is very near. Until the limit becomes a pressing threat, it will be neglected and will rarely cross the person’s mind. And by then it will be too late. This is almost certainly not what was intended by those who devised the plan—years of neglecting the deadline, followed by last-minute panic and eventual failure to receive further aid. In a way, it is the worst of all possible arrangements: it penalizes but fails to motivate.

Limits can be made more effective once we understand tunneling. For a limit to affect behavior it must enter the tunnel. One way would be to send a salient reminder of the months that are remaining. By calling attention to it we can try to force this distant problem into the tunnel. Another way is to change the structure of the limit. We have seen that frequent interim deadlines have a greater impact than a single distant deadline. So a better solution would be to create smaller but more frequent limits. (Perhaps, instead of so many years in a lifetime, only so many months in a given few-year period.) And to make the consequences of going over the limit smaller but immediate, easy to detect and to survive—perhaps a drop in payments rather than cutting off welfare altogether.

There is a general lesson here for how (and how not) to structure incentives. Incentives that fall outside the tunnel are unlikely to work. Imagine you are trying to encourage the vaccination of children whose parents are struggling to make ends meet this month. Which is more attractive to them, a payout in a month or two or a payout now? In one study in rural Rajasthan, India, a mere kilogram of lentils proved particularly effective in getting people to come in and get vaccinated. Rewards and penalties in some distant future are less effective for those who tunnel. A hefty subsidy in a savings program that pays out years from now is nice, but it renders those savings an “important but not urgent” matter, one that falls outside the tunnel and can be neglected indefinitely. For an incentive to work, people must see it. And most incentives, unless designed well, risk falling outside the tunnel, rendering them invisible and ineffective.

BANDWIDTH COMES AT A PRICE

Conditional cash transfers are an increasingly popular way to transfer money to the poor: the amount of cash a person receives depends on the good behaviors she exhibits. Studies show that these programs work; clients respond to the cash incentives. But that’s only one side of the coin. The other side is that many potential clients fail to respond. Here again, the incentives often fall outside the tunnel; the payments come in the future and the desired behaviors are not what is tunneled on now. But this raises another question: Even if we could bring those incentives into the tunnel, should we? Each additional incentive taxes bandwidth. To capitalize on a bonus payment for a child’s medical checkup, a parent must set up the appointment, remember to keep it, find the time to get there and back, and coerce the child to go (no child likes the doctor!). Each of these steps requires some bandwidth. And this is just one behavior. Conditional cash transfer programs seek to encourage dozens, if not hundreds, of these good behaviors. Just understanding those incentives and making the necessary trade-offs—deciding which are worth it for you and which are not, and when—requires bandwidth.

We never ask, Is this how we want poor people to use their bandwidth? We never factor in this cost in deciding which behaviors are most worth promoting. When we design poverty programs, we recognize that the poor are short on cash, so we are careful to conserve on that. But we do not think of bandwidth as being scarce as well. Nowhere is this clearer than in our impulse to educate. Our first response to many problems is to teach people the skills they lack. Faced with parenting problems, we give parenting skills programs. Faced with financial mistakes—too much borrowing at too-high rates—we provide financial education classes. Faced with employees whose social skills are lacking, we offer “soft skills” classes. We treat education as if it were the least invasive solution, an unadulterated good. But with limited bandwidth, this is just not true. While education is undoubtedly a good thing, we treat it as if it comes with no price tag for the poor. But in fact, bandwidth comes at a high cost: either the person will not focus, and our effort will have been in vain, or he will focus, but then there is a bandwidth tax to pay. When the person actually focuses on the training or the incentives, what is he not focusing on? Is that added class really worth what little quality time he managed to spend reading or with his children? There are hidden costs to taxing bandwidth.

And even when we do decide that educating is the right thing, there can be ways to do so and still economize on bandwidth, as illustrated in a study by the economist Antoinette Schoar and her coauthors. They had been working with a microfinance institution in the Dominican Republic called ADOPEM, whose clients run small enterprises—general stores, beauty salons, food services—usually with no employees. ADOPEM felt that its clients were making mistakes in their accounting books and generally didn’t understand finance as well as they should. The solution seemed simple: financial literacy education. So Schoar procured a standard financial literacy training module, of the kind typically given to microentrepreneurs worldwide. Her reaction upon seeing the material: Wow, how tedious! (And she’s a finance professor at MIT.) The course was several weeks long and focused on traditional accounting techniques, teaching daily recordkeeping of cash and expenses, inventory management, accounts receivable and payable, and calculating profits and investment.

In a world of unlimited bandwidth, all this would be worth knowing. But in the real world, Schoar believed that she could do better for her clients. She gathered together a group of the best local entrepreneurs to look at how they managed their finances. They, too, were not engaged in complex accounting, but they did what the less successful entrepreneurs did not do: they followed good rules of thumb. For example, several would put the cash from their store in one register and pay themselves a fixed salary. This prevented the commingling of home money and business money that makes it difficult to determine how much they were spending at home versus how much the business was earning. (Some of the women kept one wad of cash in their bra’s left cup, and the other in the right cup.) This is not quite double-entry bookkeeping, but it was effective and simple. It economized on bandwidth and preserved most of the benefits.

Schoar collected the best rules of thumb and designed a different “financial education” class based on them. Her class was shorter and much easier to grasp. It used a lot less bandwidth, and this showed up in the data. Attendance was much higher, and at the end of the rules-of-thumb class, clients were ecstatic and asking for more; many even said they would pay for another class themselves. Normally, you have to cajole people to come back to a class on financial education.

The reduced bandwidth also made the class easier to absorb and more effective. In follow-up surveys, students were more likely to implement the rules of thumb than the complex rules of accounting. And this showed up in the bottom line. Revenues—actual business sales—went up for the rules-of-thumb graduates, especially in bad weeks when improved practices can matter most: they had 25 percent higher revenues in those bad weeks. Traditional financial literacy training, in contrast, had no impact. The lesson is clear: economizing on bandwidth can yield high returns.

Whether it is in the trade-offs that people are led to make, the way education is structured, the incentives that are created, or how we handle failure, understanding the psychology of scarcity can dramatically alter the way social programs are designed. Of course, none of this provides a magic bullet to end poverty. The problems are deep. But an awareness of the psychology of scarcity and the behavioral challenges it yields can go some way toward improving the modest returns of anti-poverty interventions.

BANDWIDTH CAN BE BUILT

You are a working single mother who holds down two jobs. You have a lot to juggle. Besides the financial juggling we talked about already, you must also juggle daycare for your kids, which is expensive. You know of one program that is highly subsidized, but it will accept only one of your kids, and it closes much too early to help with your second job. So you use a patchwork of solutions. You arrange for your younger child to stay with your grandmother. You must also arrange transportation from school to your grandmother’s for one child and from daycare for the other. And because you work in the service sector, your child care needs depend on the hours your staff supervisor gives you. She is nice and tries to help, but there is inevitable volatility.

Now imagine that we offer you a highly subsidized daycare program. What exactly are you getting for it? Surely we are saving you time shuttling your kids back and forth. We might be saving you money as well, either explicitly (this program is cheaper than your previous one) or implicitly (if we account for your grandmother’s time). But we would be giving you something else, even more precious. Something you could spend on many things. We would be giving you back all that mental bandwidth that you currently use to fret, worry, and juggle these arrangements. We’d be taking a cognitive load off. As we’ve seen, this would help your executive control, your self-control more broadly, even your parenting. It would increase your general cognitive capacity, your ability to focus, the quality of your work, or whatever else you chose to turn your mind to. From this perspective, help with child care is much more than that. It is a way to build human capital of the deepest kind: it creates bandwidth.

Typically, when experts evaluate this child-care program, they will look at narrow outcomes: Was the mother able to work more hours; was she less tardy? This, however, may be far too narrow a perspective. What the program produces is freedom of mind, greater bandwidth, not something that’s easy to measure. If the program is successful, its benefits should show up in many contexts. All else being equal, one ought to be able to look directly and see the mental impact of this program. Does working memory improve? Do impulse control and self-control improve? Some of our pessimism about existing programs might come from a failure to appreciate and therefore measure such impact. If we look too narrowly at this child-care program, we will miss many of its broader benefits. Taken together, a successful intervention may yield much more than a modest return. But if we fail to look where the deepest needs are and where the benefits accrue, we are bound to underestimate its impact.

There are, besides child care, many examples from around the world of how bandwidth might be built. The first comes from finance. Recall that a great deal of juggling among the poor comes from fighting everyday fires. If we can help people fight these fires, we will create new bandwidth. What is inherent to these fires is that they are acute—there is an immediate need for cash. The need is not for big investments; it is for small amounts—to buy a school uniform, for example. Put differently, the poor most want what the moneylender can easily offer: a small amount of money, provided quickly and repaid quickly to help out with an urgent need. Instead, the kind of finance that is offered to the poor is often built on the opposite principle: modest to large amounts of money provided judiciously and slowly. Such loans can be helpful for investing. But if people are busy fighting fires, they will not have the bandwidth for investments. Is it any surprise then that despite the presence of respectable microfinance institutions, people still prefer to go to moneylenders? In India, we tested one very short-term small loan product with KGFS, a full-service financial institution that serves the rural poor. And we were amazed by the high demand for loans that averaged less than $10. The product does not help build wealth; it does not turn people into entrepreneurs. On the surface, it does not look like the kind of sum that can transform a life. Yet it might do just that. The scarcity trap begins with firefighting and with tunneling, doing things that have tremendous costs lurking outside the tunnel. Change that and we can change the very logic of poverty.

We can also go back to the source. Income flows are often lumpy and volatile in the developing world, because workers lack formal, steady employment. Even in developed nations, many low-income individuals who are employed face a great deal of volatility in incomes and earnings. As we saw earlier, income volatility is a major source of the eventual need to juggle. Why not try to mitigate it? A greater focus on the creation of dependable jobs and stable incomes for the poor across the world could be psychologically transformative.

But we can go further. We tend to focus on big shocks, such as medical emergencies or rainfall insurance. Surely these are important. Yet when one is juggling, small shocks can have equally large effects. For a poor farmer, a sick cow can reduce daily income enough to cause a slide into a scarcity trap. We should therefore look to insure the poor against these apparently “small” shocks. In the United States, something as simple as inconsistent work hours (this week you work fifty hours, but next week you get only thirty) can cause juggling and perpetuate scarcity. A solution would be to create the equivalent of unemployment insurance against such fluctuations in work hours, which to the poor can be even more pernicious than job loss.

We have seen how most of the shocks that come from juggling and induce tunneling are generally quite predictable. On the one hand, suddenly needing money for fertilizer counts as a shock. On the other hand, it is entirely predictable. It happens every year, but when you are busy juggling, you do not see it coming. This points to the great potential value in finding ways to buffer against such shocks. One way is to create financial products that help the poor build savings slack. We could do that using some of the techniques for managing scarcity we discussed earlier. For example, we can use tunneling to our advantage. Offer high-fee loans to deal with current fires. These loans will be attractive in the tunnel, and we can use the high fees to build a savings account.

Better yet, create products that prevent the firefighting. We saw how scarcity traps and juggling often follow lax management during times of relative abundance. Why not help then? Build a financial product that takes a farmer’s harvest payment and smooths it out, effectively yielding a monthly income. This is but one example. More broadly, we spend enormous resources on financial planning for retirement. Helping the poor escape a continuous life of juggling and firefighting could be similarly transformative.

All this reflects a deeper, and somewhat different, perspective on poverty. It focuses not just on the poor’s obvious scarce resource, income, but on that other, less palpable but equally critical resource, bandwidth. Considerations of bandwidth suggest that something as simple as giving cash at the right time can have big benefits. If done correctly, giving someone $100 can serve to purchase peace of mind. And that peace of mind allows the person to do many more things well and to avoid costly mistakes. One cash transfer program in Malawi showed a 40 percent reduction in the psychological distress of low-income participants. Understanding how to provide transfers at the right time and measuring these broader impacts are more ways to move toward bandwidth-sensitive policies.

All this is a radical reconceptualization of poverty policy. It forces us to recognize the many ways in which different behaviors are linked. We understand that rent and food and school fees all form part of a household’s budget. Now, rather than looking at education, health, finance, and child care as separate problems, we must recognize that they all form part of a person’s bandwidth capacity. And just as a financial tax can wreak havoc in one’s budget, so can a bandwidth tax create failure in any of several domains to which a person must attend. Conversely, fixing some of those bottlenecks can have far-reaching consequences. Child care provides more than just child care, and the right financial product does much more than just create savings for a rainy day. Each of these can liberate bandwidth, boost IQ, firm up self-control, enhance clarity of thinking, and even improve sleep. Far-fetched? The data suggest not.

A PERSISTENT PROBLEM

The fight against poverty has been an uphill struggle. Program after program has proved either unsuccessful or at best modestly successful. Social safety nets tend to be sticky. In the United States, once a person has fallen into the social safety net, she is bound to return to it again and again. And training programs appear to be only moderately effective. Researchers who have sought to estimate their impact have found some benefits: they are worth the investment, but they are not able to alter the course of poverty. Changing neighborhoods also only helps a bit. One experiment in the United States moved thousands of families from low-income to higher-income neighborhoods, and found modest impacts, primarily on stress and quality of life, but the underlying patterns of poverty did not change.

Internationally, the results are similar. Microfinance—providing small loans to help start small businesses—has been touted as highly transformative. While the impact of microfinance is likely positive, several studies now suggest that it is unlikely to change the fundamental logic of poverty. Feeding programs show some impact on children’s learning. Education has a robust but quite limited return. For years, nonprofit organizations have tried to provide a variety of holistic packages to address the varied needs of the poor. Surely they are doing good work. But they, too, have observed only modest returns.

This is certainly not intended as a critique of current programs. Poverty is a difficult problem. Even modest returns can make for worthy social investments. This is, however, a suggestion for how we might do better. When we encounter programs that have had limited success, we may be tempted to infer that they deliver something people do not want or do not consider important. But perhaps the problem is not in what these programs are trying to deliver but with the actual delivery. Like the bomber cockpits of World War II, these programs might achieve greater success through better design. And a better design will have to incorporate fundamental insights about focusing and bandwidth that emerge from the psychology of scarcity.