More Than Good Intentions

by Dean Karlan

  These were huge differences we had uncovered. They suggested a radically different story about the way Arariwa’s loans actually helped poor Peruvians improve their lives. Not everyone, it turned out, was building a microenterprise from the ground up. A big portion of clients simply paid lip service to Arariwa’s vision of dignified success through entrepreneurship, then turned around and did what they pleased with the money. This is not uncommon. In Indonesia, Don Johnston and Jonathan Morduch of New York University found that over 50 percent of clients reported using loan proceeds for consumption.

  Programs that distribute money or other valuable resources should question the wisdom of imposing rules that are unlikely to be followed. Making policies you can’t enforce shows impotence. Besides, all those scofflaws—think of Philip enjoying his rice—might actually be onto something.

  But there’s another problem too.

  The Bigger Problem with Slippery Money

  I said there were two larger points here. Here is the second, and more compelling, reason why we need to see beyond restrictions that aim to limit the ways people can use programs and program resources to improve their lives: Even when they follow all the rules to the letter, the true impact of participation often pops up in an unexpected place. (This is why, even if it were feasible for lenders to do so, it wouldn’t suffice to track the borrowed funds themselves.) Like water, money finds its own level. When it falls on uneven terrain it tends to fill the deepest holes first, no matter what you say to it—or to the person holding it. That’s its nature.

  Sometimes people help the leveling process along deliberately, like when Philip bought his rice cooker. Once he’d paid off enough of the guesthouse bill to get the proprietor off his back, his debt had ceased to be the deepest hole; now the money could go elsewhere.

  But the same process often takes place without any conscious effort. Imagine water cascading down the tiers of an Italianate fountain. Gurgling out of the spout at the top, it enters the top bowl, which overflows and spills into the second bowl, which overflows and spills into the third, and so on. The water was added to the first bowl, but ultimately it’s filling the last. So it is with the swag from development programs: Something of value—maybe a milk-producing goat or a school uniform for a child or money to start a microenterprise—is given to meet a specific need, but its effects might ultimately be felt in some distant quarter of the recipient’s life.

  Here is an example. Imagine a woman who sells tomatoes in a busy street market. Each morning she buys $50 worth of tomatoes from a wholesaler and, over the course of the day, sells them for $55. At the end of the day she takes $5 for herself and puts the remaining $50 in her back pocket to buy tomatoes the next morning. One day she gets approved to take out a $100 microloan to grow her tomato business. She goes to the bank first thing in the morning, picks up an envelope containing $100 in cash, and heads straight for the wholesaler’s shop where, using the $100 from the envelope, she buys twice as many tomatoes as usual. Over the course of the day she sells them for $110. Voilà—a growing business!

  As she closes up her stall in the evening, she remembers the $50 that she had set aside the previous afternoon. It’s right where she left it, in her back pocket. Riding high on the excitement of the day, she decides to celebrate. So she stops on the way home and buys a DVD player for her family.

  Did you see the money spill from one tier to the next? When the bank asks her how she used her loan, she answers (honestly!) that she spent it all on inventory for her tomato business. But from our bird’s-eye view we can see that the true impact of the microloan was to allow her to buy $50 more of tomatoes and a $50 DVD player. Even when the slipperiness of the people is not a problem, the slipperiness of the money itself remains.

  We can prescribe specific solutions to the problems we see, but our prescriptions are usually nonbinding. Sometimes people (like Philip) intentionally flout the rules; other times people (like the tomato seller) genuinely try to keep their promises, but wind up doing the same thing.

  In the last chapter we saw evidence from the field suggesting that strictly entrepreneurial loans probably are not the answer for everybody. And now we see that, even if they were, the differences in people’s priorities and the slipperiness of money itself mean even the most concerted efforts to restrict borrowers’ choices are likely to be in vain.

  But that doesn’t keep lenders from trying. The most widely used tactic for keeping clients in line is creating incentives for borrowers to monitor one another. The reasoning goes: If they won’t adhere to rules imposed from outside, maybe an insider—or, better yet, a whole slew of insiders—could do the trick.

  In the next chapter we’ll look at the main vehicle lenders use to apply peer pressure—the group liability loan—and see how well it works after all.

  6

  TO COOPERATE IN GROUPS

  What About the Weakness of the Crowd?

  If you looked at just the Art Deco marquee of the Roxy Theater, jutting triangularly up against the sky, you might think of a tropical city in its colonial heyday. Linen suits; close, steamy night air stirred by a gentle breeze; lazy clatter of palm fronds; the sweet smell of plantains frying in roadside stalls; music issuing from the door of a nightclub called the Copacabana, a nightclub with little round tables and a bandstand and good imported gin.

  But you can’t look at just the marquee of the Roxy Theater. Your eye also takes in its crumbling facade, the elegant curved glass window of its ticket booth webbed with dust and traversed by a long diagonal crack, the dead bugs laid upside-down in its empty poster displays. You cannot ignore the fine chalky silt of the parking lot inhaled against the back of your throat like lime, caustic as cement dust. You cannot tune out the drone of idling semi trucks on the road behind you or the buzz-saw snorts of the motorbikes that tear past them up the narrow shoulder. And it isn’t a steamy night, but the blinding, white-hot morning of Tuesday, February 5, 2008.

  Accra, Ghana, is a city whose colonial heyday it is not.

  But it may be in some kind of a heyday yet; and the Roxy Theater is as good a symbol as any. Despite the absence of a projector, a screen, seats, or even a roof, the derelict place still fills up at least a few times a week. These days, though, its patrons are not moviegoers. They are microcredit clients.

  On that particular white-hot morning, Jake was at the Roxy to talk with “Community Leaders,” women who had distinguished themselves over years of successful borrowing from an Opportunity International, a major Ghanaian microlender. They were the pillars of the microcredit program. Many had taken a dozen loans or more, and most had served as executives of their borrowing groups. And there was one thing they all had in common: None had missed so much as a single payment. They were perfect clients.

  Jake had been briefed before he got to the theater, but still didn’t know exactly what to expect. He wondered as he climbed the uneven cement stairs to the balcony where everyone was seated: What will these women look like? He half imagined a corps of pants-suited business mavens. Pumps, pinstripes, and shoulder pads; seriousness and savvy. They were, after all, the cream of the crop.

  It was a short-lived fantasy. The stairs gave onto a flat cement balcony with metal folding chairs set in neat rows. You could tell how precisely they had been arranged because they had not been disturbed. Nobody was sitting in them. The women were up and moving and talking. They looked just like anybody else: long skirts of bright printed fabric, rubber flip-flops, secondhand silkscreened T-shirts, head wraps. Big smiles framing white, white teeth. And the sound of laughter.

  Levity, Jake learned that morning, was another characteristic shared by all these women. Which is not to say they weren’t serious—quite to the contrary. They took their debts very seriously. You don’t borrow and repay thousands of dollars over hundreds of months without ample reserves of self-discipline and drive. Their sense of humor came in where other borrowers were concerned; it’s what kept them from tearing their hair out.

&nb
sp; The loans these women had taken were group loans, so their obligations to the bank were bound up with those of their fellow group members. And while the women at the Roxy Theater had earned designation as Community Leaders by their exemplary track records, many others hadn’t acquitted themselves so well. In their vast experience, the Community Leaders had seen it all: people unable to pay on time, people unable to pay in full, recalcitrant borrowers determined to walk out on their debts, group members who simply up and disappeared. When these things happened, it fell to the rest of the group to pick up the slack. They reached into their pockets and made payments on the delinquents’ behalf. Over the years, those payments added up.

  It’s hard to put a number on these things, but Mercy’s sense of humor was worth better than a thousand dollars—or, to put it in context, about one and a half times Ghana’s per capita annual income. That, she told me, was how much she had paid during her eight years as a client to cover for other group members when they came up short.

  Mercy sold dry goods and canned goods in Makola, one of Accra’s largest outdoor markets. She stocked things like spaghetti, matchbooks, instant coffee, tomato paste, and canned herring. When she took her first loan she was a lowly “tabletop” vendor: one of the innumerable women who arrives early in the morning, walks through the market aisles with a collapsible card table and a cardboard box of goods balanced on her head, sets up shop on the sidewalk, and repeats the whole process in reverse at sundown. Now, in the midst of her twelfth loan, she had made great strides. She had upgraded to a cinder-block stall with a metal door and a heavy padlock, so she didn’t have to carry her wares to and from Makola every day. No longer limited by the amount she could carry on her head, she stocked more, and more varied, products than before. She bought her inventory in bigger bulk and at lower cost from a distributor.

  There was no doubt about it—Mercy had prospered over her career as a microcredit client. Of course, just because her business prospered at the same time that she borrowed does not mean the credit caused her to achieve this growth. But regardless of whether or how much the loan helped her, the question remains: Did she really have to pay such penalty (the thousand dollars spent covering for defaulters) for borrowing? Is there a better way?

  Up there in the still air of the Roxy Theater balcony, Mercy told Jake she thought this loan would be her last. “As for my own, I don’t mind at all. I can pay. My business is coming on. But I won’t cover for those ones”—and here she drew the corners of her mouth down into a grimace, shut her eyes, and shook her head, thinking maybe of her thousand dollars—“again.”

  The Vaunted Group Lending Model

  It’s no wonder Mercy was ready to quit. If anything, we might ask why she hung around as long as she did. But the big question here is not about her; it is about group lending in general. In the last chapter we saw that, for the poor as much as for anyone else, people’s unique needs, priorities, and ways of pursuing happiness are hard to suppress. That’s why Oti passed up business opportunities to watch movies with his girlfriend and why Philip spent his borrowed money on a rice cooker. Does it really make sense to bind up all these diverse strands, all pulling in separate directions? Aren’t some people—particularly the good eggs, like Mercy—bound to get burned and drop out? Where is the sense in a lending program that effectively penalizes the best clients by making them cover for the deadbeats?

  Strange as it seems, this has been a feature of modern microcredit since Muhammad Yunus, the godfather of the movement, made his first loan to a group of Bangladeshi bamboo craftswomen in the late 1970s. In the intervening three decades, most of the thousands of microcredit organizations that have sprung up around the globe were cast in the mold of Yunus’s Grameen Bank. Instead of lending to individuals, they grew by lending to groups.

  In the standard arrangement, the group as a whole is liable for each member’s loan. So, for instance, when ten clients borrow a hundred dollars each, the microlender sees it as a single group loan of a thousand dollars. The group swims (or sinks) together. If the group makes its payments on time and in full, then everyone is eligible for another loan once the current one has been repaid. If they fall short, then all the members—the good apples along with the bad—are barred from further borrowing. (Or at least it’s supposed to work that way. Many microlenders threaten to bar the entire group when only a couple members default, but relatively few follow through. It is a common practice to pull aside the well-behaved borrowers from a defaulting group and continue lending to them.)

  This setup might seem like a raw deal for good clients like Mercy, but in reality it’s a compromise. In theory, group liability is the reason why microlenders can operate at all. The group lending model solves three problems that historically kept banks from serving the poor. Think of these problems as questions that a bank has to answer for lending to be viable.

  First, who is this person who wants to borrow? Second, how can we be sure she will make her payments? And third, what can we do to get our money back if things go sour?

  In most developed countries, vast information networks and powerful legal mechanisms help address these questions. But potential lenders in the rest of the world have so few resources to draw on that they often simply refuse to operate. The great innovation of the group lending model was to bridge these gaps by harnessing the knowledge and power of the borrowers themselves.

  One way to see how this works is to put on a credit officer’s hat and compare the process of reviewing a loan applicant in the United States with the same process in a developing country—let’s say at Mercy’s bank in Ghana.

  Who Is the Borrower?

  For a bank in the United States, answering the first question is simple. A social security or tax identification number, for instance, illuminates a constellation of reliable information, from addresses to vehicle and voter registrations. Beyond that, credit reporting agencies like Experian, Equifax, and TransUnion compile detailed histories of our lives as consumers and distill them into a simple three-digit number—a credit score. A credit score lets a bank know how attentive a prospective borrower has been to her financial obligations in the past, and is a powerful indicator of how she’ll behave in the future. All of this, of course, appears instantaneously when the bank officer enters an applicant’s name into his computer.

  Now let’s see about Ghana. A bank starts by asking an applicant’s name, but what’s in a name? Not much, if that name is not unique, consistent, or verifiable. Most Ghanaians actually have four names: a family (last) name, a Christian given name, a local given name, and a nickname for the day of the week on which they were born. Spellings and order vary, even on official documents—if there are any official documents in the first place. Addresses are even harder. Instead of a house number, street name, and zip code, how about: “Go to Agona Junction, walk about four hundred meters toward Tema, turn right opposite Ebenezer Church. Go on the dirt path behind Quincy Chop Bar near the football field. Find the house with a white compound wall and a green gate.” Good luck.

  Knowing a name or address wouldn’t get you very far, anyway, since it doesn’t link you to a web of relevant information. The single biggest gap is the lack of credit reporting agencies. Ghana, like most developing countries, has none. It is virtually impossible to learn about a person’s economic history. And when a bank can’t learn about its potential clients, it can’t screen out the bad seeds. Then lending at all means running the risk of choosing untrustworthy borrowers.

  The group lending model addresses the problem by passing the burden of screening onto the borrowers. Since the group (and not the bank) pays the penalty when members act up, they are the ones who have an incentive to learn who is trustworthy. In some sense, they are better equipped to do the job anyway, given the dearth of information available to the bank. As neighbors, relatives, parishioners, and friends of fellow group members, borrowers know more about each other than the lender ever could.

  How Do We Know the Borrower Can Pay
?

  Suppose we manage to convince ourselves that a potential borrower is genuine. How can we then be sure she will make her payments once she gets the money?

  In the United States, loan applicants have to put their money where their mouth is. People secure loans with collateral like houses, cars, and jewelry. Or, without putting existing assets at risk, a pay stub or a tax return can prove that an applicant earns enough to cover her payments. At the very least, people seeking entrepreneurial loans are often required to submit a detailed business plan that shows how borrowed funds will be used and how the business will generate enough money to enable repayment.

  Meanwhile, collateral is unlikely to be an option for a Ghanaian credit applicant. Most people don’t have many sizable assets to begin with, and weak property laws hamstring those who do. In a common scenario, an inherited piece of land is offered up to secure a loan. When the bank investigates, it finds half a dozen claims on the little parcel. That’s half a dozen people who also believe they own the land—half a dozen people who’d probably stand in the way of a repossession, if it came to that. Understandably, banks usually avoid such entanglements.

  Verifying income is no easier than proving ownership. Since most employment in Ghana is informal, people usually get paid in cash. This is especially true for microcredit clients, who tend to be self-employed. So most applicants cannot present a pay stub to demonstrate earnings. Precious few people keep detailed records of their enterprises’ purchases and sales, so business plans and projections about future profitability are usually rudimentary if they exist at all.

  All this adds up to a lot of consternation among lenders. What they would really like to do is look over clients’ shoulders—that way, they could ensure that people are putting loans to good use and exerting sufficient effort to make their payments—but they can’t. They just don’t have the manpower to keep a close eye on everyone. And so some people wind up doing things their creditors wouldn’t approve of. Like Philip’s buying a rice cooker instead of paying his rent, for instance.