WSJ: The Anti-Poverty Experiment

[summer]

The Wall Street Journal

The U.S. and other wealthy nations have spent trillions of dollars over the past half-century trying to lift the world’s poorest people out of penury, with largely disappointing results. In 1966, shortly after President Lyndon B. Johnson declared war on poverty, 14.7% of Americans were poor, under the official definition of the U.S. Census Bureau. In 2013, 14.5% of Americans were poor.



World-wide, in 1981, 2.6 billion people subsisted on less than $2 a day; in 2011, 2.2 billion did. Most of that progress came in China, while poverty has barely budged in large swaths of sub-Saharan Africa, South Asia and Latin America.



Is it time for a new approach? Many experts who study poverty think so. They see great promise in a new generation of experimental programs focusing not on large-scale social support and development but on helping the poor and indebted to save more, live better and scramble up in their own way.



Linda Hanson of Duluth, Minn., 64 years old, works as an administrative assistant at a local organization for the disabled; her husband, Glenn, 65, is a retired city bus driver. Today, the Hansons have achieved some financial stability, but by early 2014, they were in trouble: Linda had lost her previous job, their catering business had failed and they had racked up about $28,000 on their credit cards.


Overwhelmed by the debt, they struggled even to make the minimum monthly payments, said Mrs. Hanson—until they heard about Pay and Win, an experimental program offered by Lutheran Social Services in Duluth to encourage struggling borrowers to manage their debts. Those who steadily pay down their loans each month are eligible for raffle drawings.



“When you know you have a hope of winning,” says Mrs. Hanson, “what a motivation!” The Hansons soon got their finances in order and felt less overwhelmed. In January, the Hansons got a surprise windfall: They won the program’s grand prize of $5,000, which they committed to use to pay down the principal on their debt. “We’re going to be OK now,” says

In the Butuan City area of the Philippines, a program with a similar objective was set up in 2002 by a team of economists including Dean Karlan of Yale University. It provided more than 700 people with a bank account that allowed them to lock their savings up for months at a time and to fill out a certificate labeling their accounts with specific goals like education or buying a house. Follow-up research found that participants saved an average of 82% more than those from similar households who saved in conventional bank accounts.




Just as health care has been modernized by “evidence-based medicine” and professional sports by the rigorous statistical analysis that inspired Michael Lewis’s book “Moneyball,” bands of upstarts in the world of anti-poverty policy are applying the scientific method to their own work.



Many of these poverty fighters call themselves “randomistas,” after the randomized controlled trials that are at the heart of their methods. In such field experiments, people are randomly assigned either to a treatment group that receives an “intervention” or to a control group that does not. The experimenters meticulously collect and analyze data, then try to replicate the results elsewhere to see if they hold up.



A report published in the journal Science in May found that the randomistas’ methods not only work but stick. In Ethiopia, Ghana, Honduras, India, Pakistan and Peru, 10,495 households—about half of them earning less than the equivalent of $1.25 a day—took part in two-year experiments intended to help them become more self-sufficient.



Launched at different times in the various countries, the studies ran from 2007 through 2014 and were designed by economists from Yale, the Massachusetts Institute of Technology and several other research centers in the U.S. and Europe. The programs included a grant of livestock or business assets, a short-term cash and food stipend, health services, vocational training and savings accounts that, in some cases, locked up their money temporarily. One year after the programs ended, participants were consuming 7.5% more food, had a 14% increase in assets and had saved 96% more, on average.



Other such experiments have sextupled the use of clean water in areas of Africa rife with waterborne diseases by positioning chlorine dispensers next to wells and pumps, and have roughly tripled the rate of childhood vaccination among villagers in rural Rajasthan, India by offering bags of lentils to parents. Some of these interventions cost only a few dollars per person.



Much of this research is being conducted by economists and psychologists working through Innovations for Poverty Action, or IPA, based upstairs from an audio-and-video-equipment store just off the campus of Yale University, and Abdul Latif Jameel Poverty Action Lab, known as J-PAL, based at the Massachusetts Institute of Technology.



The randomista movement resonates with a new generation of donors who believe in the power of data. “A lot of what we do comes from a scientific approach,” says Trey Beck, head of investor relations at the D.E. Shaw group in New York, which manages $39 billion, partly in mathematically driven hedge funds. “What IPA does in forming and testing hypotheses has strong affinities with what we do in our corner of finance,” adds Mr. Beck, who has donated to IPA personally.



Until recently, most governments, development agencies and nongovernmental organizations aiding the poor didn’t rely on the scientific method to figure out what works, what doesn't and why.



So, says MIT economist Esther Duflo, “We are trying to promote a culture of learning that will permeate governments and NGOs and businesses to such an extent that it will become par for the course.”



Much of the randomistas’ fieldwork is built on the findings of modern psychology about the limitations of the human mind, gleaned largely from experiments in the world’s richest countries.



“The daunting realization is that we don’t know what the hell we’re doing in most fields of life, especially the ones that involve people,” says Richard Thaler, professor of behavioral science and economics at the University of Chicago Booth School of Business and author of the new book “Misbehaving: The Making of Behavioral Economics.” He adds, “The alternative to guessing is to run experiments.”



Figuring out what makes and keeps people poor isn’t a trivial challenge. Anyone who has ever traveled in the developing world knows that even the planet’s most desperately poor people aren’t lazy or stupid. And if poverty were caused by economic geography (hot climates hostile to agriculture and hospitable to disease) or a lack of modern infrastructure, then bombarding poor countries with money would long ago have ameliorated the problem.



Behavioral economics, which documents the many ways in which the human mind falls short of perfect rationality, offers some novel explanations. In theory, people consider all relevant information to make objective judgments about risk, reward and the best choices they can afford. In fact, however, people in all walks of life:



• love to gamble regardless of the odds of winning, leading them to bet on underdogs, play the lottery or hit the slot machines;



• value distant risks and rewards much less than current gains and losses;



• often lack the motivation to pursue their own best interest unless it can be done almost effortlessly—declining to get vaccinated, for instance, until they are given a map of where the clinic is located;



• procrastinate, putting off until tomorrow just about everything they can get away without doing today;



• put money into arbitrary buckets, creating different “mental accounts” for various purposes, even though all dollars should be equal;



• want the comfort of knowing that others agree with them;



• are impatient to their own detriment, preferring to take pension payments in one lump sum rather than as an annuity spread over time or to start receiving Social Security payments as soon as possible instead of deferring them.



If you are wealthy or middle class, these frailties of the human mind do not usually get you into severe trouble. If you are poor, however, even minor mental mistakes can have severe consequences.



Most Americans, for instance, might be annoyed—but are seldom disrupted—by the fact that banks may delay access to deposits for a couple of days or longer, whereas debits and withdrawals are usually deducted immediately.



“Now imagine that you have a checking-account balance of $6.73,” says Eldar Shafir, a psychologist at Princeton and co-author of “Scarcity: The New Science of Having Less and How It Defines Our Lives.” He adds, “Avoiding overdraft charges becomes a crazy challenge, a massive cognitive tax that makes thinking about trade-offs a constant job.”



Simplifying such problems is a priority of the randomista movement.



With mobile phones widespread among the poor, even in the developing world, researchers are testing the effectiveness of using text messages to remind people to save a portion of each paycheck, to check when their bills are due or to contribute to a retirement account. Regular text reminders among savers in Bolivia, Peru and the Philippines have increased their bank balances by an average of about 10%, at very low cost. Borrowers are much more likely to repay their debt when they receive a text message from the loan officer reminding them that it is due.



In some poor countries, governments spend 1% to 2% of gross domestic product annually on subsidizing fertilizer for farmers who couldn’t otherwise afford it. But the randomistas may have found a simpler, cheaper way to encourage farmers to fertilize their fields: by tapping into the human preference to defer spending.

Even middle-class savers in the U.S. are much more likely to set money aside for retirement if they don’t have to start saving immediately. Beginning to save one year from now sounds a lot more appealing than having to tighten your belt today. Among the working poor, such a “commitment savings” program has an even more powerful appeal.



Farmers in Kenya used 47% to 70% more fertilizer, increasing the yield of their crops, when they committed upfront, before the next growing season, to having it delivered. Those who waited until later in the year to order fertilizer, even when they got a 50% discount on the price, bought much less.



As anyone who has ever used a piggy bank knows, it also helps not to be able to get your hands on your money anytime you want it. Rural Kenyans who were offered a locked box in which to save, or permitted to earmark an account specifically for health expenses, were able to spend 66% to 138% more on preventive health care over the following year than those who set money aside informally.



Even gambling can be repurposed to put money into the pockets of the poor instead of plucking it out. At the end of last year, President Obama signed into law the American Savings Promotion Act, which permits banks in the U.S. to offer savings accounts linked to a lottery-like prize. Savers become eligible for a raffle drawing in which they may win cash or other awards. Research in many countries has shown that the chance of hitting the jackpot can be a powerful incentive for people who would otherwise struggle to save.



Mary Bunch, 32, of Lakeland, Fla., and her husband, Willis, had to move into a motel for several months after he lost his job as a landscaper last year. Mrs. Bunch, who now works as a manager at a McDonald’s, signed up for SaveYourRefund, a nationwide program that direct-deposits federal tax refunds into a savings account before the recipient can spend the money.



Participants are eligible for a $25,000 grand prize. “We’re human,” she says. “I was hoping that we would win, but I didn’t really think it could happen to us.”



In May, the Bunches found out they won the $25,000. “Just knowing you have some money in the bank makes your heart beat normally,” says Mrs. Bunch. “It takes some of that pressure off.”



Like those who are wealthier, the poor also take comfort in being part of a crowd. In India, borrowers in “repayment groups” that convened weekly to pay off a portion of their debts were three times less likely to default on their next loan than those who met monthly, presumably because their more-frequent meetings deepened their social bonds, reinforcing peer pressure to pay down the loans.



How far can the randomista revolution go? The World Bank launched a small unit this year to study how behavioral insights can be integrated into policy. But “other methods should not be put to rest just because we are so elated with this method,” says Kaushik Basu, the bank’s chief economist. “If you put all your eggs into the basket of randomized controlled trials, you will neglect some important problems. It would be a very large mistake to treat this method as the Holy Grail.”



Development economist Jeffrey Sachs, director of the Earth Institute at Columbia University and a creator, with the United Nations Development Program, of the Millennium Villages Project, has long contended that impoverished nations are locked in a “poverty trap,” in which meager agricultural production and widespread disease make it all but impossible for people to amass the savings that are essential for economic growth. The solution, he argues, is a “big push” in foreign aid targeted at such specific goals as making farmers more productive, improving schools and providing better health care.



Prof. Sachs says that “many, almost surely most, of the cutting-edge breakthroughs in actual development in recent years did not result from [randomized controlled trials].” He believes that tackling problems at the level of communities or entire societies, rather than just households, is likely to be more effective—though, he adds, randomized controlled trials should be “a part of a diverse arsenal of analytical and policy tools.”



Retorts Prof. Duflo of MIT: “The big difference between Jeffrey Sachs and us is that he knows what needs to be done, and we don’t. We’re trying to learn it.”



The learning is accelerating. Innovations for Poverty Action, the group based in New Haven, now has 500 full-time employees and an annual budget around $40 million. Its motto is “More evidence, less poverty.” Says Prof. Karlan of Yale, the group’s founder: “That’s exactly what we’re trying to do, no more and no less. It’s important to be realistic about what any one approach can do.”

[bugandbibs]

This is fascinating. There has been good success with the housing first movement. What "we" are doing hasn't been working. It's nice to see people looking to research to see what does.

The saving tied to prizes makes sense. It's kind of like diet bet.

[5thofjuly]

This reminds me of a podcast (Freakonomics?) I heard that about a bank that was giving away what was essentially changes in a drawing for every X dollars customers put into a CD. There were like 20 winners/month, 19 small and one large, and all tied to the total amount of customers contributions. So, if they sold more CDs that month, the small prizes might be $100 each and the larger $2500. If the sold fewer CDs, the prizes could be half that. What they found was that savings shot up exponentially among low-income individuals. Everyone wanted the opportunity to get the prize, no matter how long the odds. Unfortunately, the state lottery commission was not on board with this approach and it got shut down, but it was interesting while it lasted.

[LadySusan]

I got lazy and only skimmed the article even though I typically enjoy reading about behavioral economics. I think the word "randomistas" was just too pretentious.

[Deleted]

I like the idea of more evidence-based methods of helping/empowering people achieve. What we have is a mess. Better to give money away as incentive than throw it down the drain in to ineffective, shitty programs. I do wish there could be a focus on looking at existing programs for measures of success and cutting the ones that show minimal gains. Employment assistance programs would be a good place to start/clean up.