Credit from informal sources tends to be expensive. In the Udaipur survey, those living on less than 99 cents a day pay on average 3.84 percent per month (which is equivalent to an annual rate of 57 percent) for the credit they receive from informal sources. Even credit card debt in the United States, which is notoriously expensive, pales in comparison. Bank of America’s standard credit card has an interest rate of about 20 percent per year. Those who spend between 99 cents and $2 a day per capita pay a little less: 3.13 percent per month. There are two reasons for this difference in interest rates. First, the slightly less poor rely less on informal sources of credit and more on formal sources than the extremely poor, and the formal sources are cheaper. But second, the interest rates charged by informal sources tend to be higher for the poor than for the less poor. The average interest rate from an informal source drops by 0.4 percent per month for each additional hectare of land owned by the person taking out the loan.
Interest rates vary across sectors and countries, but the bottom line is always the same: Yearly interest rates in the 40 to 200 percent range (or even higher) are the norm, and the poor pay more than the rich. The implications of the fact that many people do borrow at these rates are quite staggering. There are millions of people willing to borrow at a rate that the average U.S. saver would dearly love to be paid. Why aren’t investors rushing to them with bags of money?
This is not for lack of trying. From the 1960s until the late 1980s, many developing countries had government-sponsored credit programs, usually with subsidized interest rates, targeted at the rural poor. For example, in India starting in 1977, for every branch that a bank opened in a city, the bank had to open four additional branches in rural locations that did not have a bank. Moreover, banks were directed to lend 40 percent of their portfolios to the “priority sector”: small firms, agriculture, cooperatives, and the like. Robin Burgess and Rohini Pande showed that where more bank branches were opened as a result of this policy, poverty decreased faster.2
The problem was that these forced lending programs didn’t work very well as lending programs. Default rates were staggeringly high (40 percent during the 1980s). Lending was often driven more by political priorities than by economic need (a lot of loans were made to farmers just before elections in districts where the contest was expected to be tight).3 And the money had a tendency to end up in the hands of the local elites. Even Burgess and Pande’s generally favorable study concluded that it costs much more than 1 rupee to increase the incomes of the poor by 1 rupee through opening bank branches. Moreover, further work suggested that in the longer term, regions that got more branches may in fact have become poorer.4 In 1992, in the wave of reforms that liberalized India, the requirement to start branches in rural areas was dropped and a similar trend of eroding government support for public lending programs can be seen in most other developing countries.
Perhaps the social banking experiment was a failure because the government should not be in the business of subsidized lending. Politicians find it too attractive to use the loans as giveaways, and there is no better giveaway than a loan one does not need to repay. But why don’t private bankers want to lend to small entrepreneurs? Given that they are willing to pay up to 4 percent a month, which is many times what a bank makes on its average loan, wouldn’t it make sense to try to lend to them? Some Web sites in the United States now enable potential lenders in rich countries to lend to entrepreneurs in poor countries. Could it be that they have finally understood something that everyone else missed?
Or alternatively, perhaps there is something that informal moneylenders can do that banks cannot. What could that be? And why is it cheaper to lend to richer people?
The (Not So) Simple Economics of Lending to the Poor
One standard explanation for why some people might have to pay high interest rates is that they are more likely to default. This is simple arithmetic: If a moneylender must get back on average 110 rupees for every 100 rupees he lends just to stay in business (for instance, because this is his cost of funds), without default, he could charge a 10 percent interest rate. But if half of his borrowers default, then he must get back at least 220 rupees from the half that actually do repay and hence charge a 120 percent interest rate overall. However, rates of default on informal loans, unlike those on government-sponsored bank lending, are not very high. Such loans are often repaid with some delay, but not repaying at all is actually rare. A study of rural moneylenders in Pakistan found that the median rate of default across moneylenders is just 2 percent, even though the average interest they charge is 78 percent.5
The problem is that these low default rates are anything but automatic; they require hard work on the lender’s part. Enforcing credit contracts is never easy. If the borrower is allowed to misspend the loan proceeds, or somehow has bad luck and has no available cash, there will be nothing to collect. At that point, there is precious little the lender can do to collect on the loan. Given this, it is tempting for the borrower to contrive to appear to have no money even when she does, which makes matters worse for the lender. If this is allowed to go unchecked, the lender will never be repaid, even if the borrower’s project actually succeeds.
The way lenders all over the world protect themselves against the different forms of willful default is by asking for a down payment, some collateral, or what is sometimes called the promoter’s contribution, which is the part of the firm’s capital that comes from the entrepreneur’s pocket. If the borrower defaults on the loan, the lender can punish her by seizing the collateral. The more the borrower has at stake, the less tempting it is to take off with the borrowed money. But this means that the more the borrower can pledge, the larger the lender can make the loan. And thus we have the familiar (at least before the go-go days of no-down-payment mortgages) rule that ties the size of the loan to the amount of money the borrower already has. As the French put it, “On ne prête qu’aux riches” (“One lends only to the rich”).
This means that poorer borrowers will be able to borrow less but it does not, by itself, explain why the poor should pay such high interest rates or why banks would refuse to lend to them. But there is something else that kicks in. In order to be able to collect on the loan, the lender needs to know many things about the borrower. Some are things the lender would like to find out before deciding to lend, such as whether the borrower is trustworthy. Others, such as the borrower’s whereabouts and nature of the borrower’s business, help in collecting on the loan should there be a problem. The lender may also want to keep an eye on the borrower, visiting her from time to time to make sure the money is being used as promised and nudging the business in the desired direction if need be. All of these efforts take time, and time is money. The interest rate has to go up to cover them.
Moreover, many of these expenses do not scale with loan size. There is no way to avoid collecting some basic information about the borrower, even if the loan is very small. As a result, the smaller the loan, the larger the monitoring and screening costs will be as a fraction of loan size, and because these costs have to be covered by the interest collected, the higher the interest rate will be.
To make matters worse, this creates what economists call a multiplier effect. When the interest rate goes up, the borrower has more reason to try to find a way not to repay the loan. That means the borrower needs to be monitored and screened more carefully, which adds to the cost of lending. This pushes the interest rate up even further, which necessitates more scrutiny, and so on. The upward pressure feeds on itself, and interest rates can skyrocket. Or, as often happens in practice, the lender may decide that it is not viable to lend to the poor: Their loans would be too small to make it worthwhile.
Once we understand this, many things fall into place. Because the main constraint on lending to the poor is the cost of gathering information about them, it makes sense that they would mostly borrow from people who already know them, such as their neighbors, their employers, the people they trade with, or one of
the local moneylenders, and that is exactly what happens. Strange as it might seem, this emphasis on contract enforcement could also drive the poor to borrow from those who have the power to really hurt them if they were to default, since such lenders would not need to spend as much time monitoring (their borrowers wouldn’t dare to stray) and the loans would be cheaper. In Calcutta in the 1960s and 1970s, a lot of the moneylenders were Kabuliwalas (men from Kabul)—tall men in Afghan clothes with a cloth bag slung across their shoulders who would go from door to door, ostensibly selling dried fruits and nuts, but actually mostly using that as a cover to carry out their lending operations. But why couldn’t someone more local do the lending? The most likely answer is that these men had a reputation for being fierce and implacable, a stereotype sealed by a story that all Bengali schoolchildren have in their textbooks, in which a good-hearted but violent Kabuliwala kills someone who was trying to cheat him. The same logic also explains why the Mob in the United States was for many people the “lender of last resort.”
A more baroque illustration of the power of threat can be seen in a story from London’s Sunday Telegraph, dated August 22, 1999, titled “Pay Up—or We Will Send the Eunuchs to See You.”6 The report describes debt collectors in India making use of the old social prejudice against eunuchs to collect from long-standing defaulters. Because people believe that seeing a eunuch’s genitals brings bad luck, the eunuchs were instructed to show up at the defaulter’s house and threaten a “showing” if they continued to be uncooperative.
The high costs of collecting information on a borrower also help explain why even when there are several moneylenders in each village, competition does not drive the price of credit down. Once a lender has paid the cost of vetting a borrower, and the borrower has established a good reputation with him, leaving is difficult. If the borrower did go elsewhere for credit, the new lender would have to do the due diligence all over again, which would be expensive and would drive interest rates even higher. Moreover, the lenders would be suspicious of such a new client: Why did someone feel the need to abandon an existing relationship, when it is obviously costly to do so? The moneylender would then be doubly careful, which could further raise interest rates. Thus, despite the apparent choice of lenders, borrowers are somewhat bound to the one they know already. And moneylenders can exploit this advantage to raise interest rates.
This also explains why banks do not lend to the poor. Bank officers are not very well placed to do all the necessary due diligence: They don’t stay in the village, they don’t know the people, and they rotate frequently. Respectable banks are not in a position to compete with Kabuliwalas. They cannot easily threaten to break people’s kneecaps or even send them eunuchs. The Indian branch of Citibank got into serious trouble when it was discovered that it was using “goondas” (local hooligans) to threaten borrowers who did not repay vehicle loans. And the courts are not really an option, either. In 1988, the Law Commission of India reported that 40 percent of the cases for asset liquidation (of bankrupt borrowers) were more than eight years pending.7 Think of what this means from the lenders’ point of view: They know that even if they are sure to win their case against a defaulting firm, they will only be able to claim the pledged assets in several years’ time (with plenty of opportunities for the borrower to divert the assets). Of course, this means that from the point of view of the lenders, the value of the borrower’s assets at the time the loan begins will be all that much lower. Nachiket Mor, who was then one of the vice presidents of ICICI Bank, once described to us what he had thought was an absolutely brilliant way to get farmers to repay their agricultural loan. Before disbursing each loan, he would ask them for a postdated check for the same amount. The great insight was that if the farmer did not repay, the bank would then be able to send the police to collect on the check, because not honoring a check is a criminal offense. This worked for a while, but then it began to unravel. When the police realized that they had hundreds of bounced checks to track, they politely told the bank that, really, this was not their job.
Even when the bank manages to get its money back, things can backfire: Banks do not like headlines associating them with “farmer suicides.” And to cap it all off, when elections are around the corner, governments love to write off outstanding loans. Given all this, it is no surprise that banks find it easier to stay away from lending to the poor altogether, leaving the field to moneylenders. However, although the moneylenders have an advantage in getting their money back, they have to pay a lot more for the money they lend out than the banks. This is because we are happy to give our savings to the bank for safekeeping even if it pays us little or nothing in interest, but few people would think of depositing their savings with a moneylender. This, combined with the multiplier effect and the monopoly power that moneylenders often enjoy, explains why the poor face such high interest rates.
The innovation of people like Muhammad Yunus and Padmaja Reddy, then, was not just the idea of lending to the poor at more reasonable rates. It was figuring out how to do it.
MICRO INSIGHTS FOR A MACRO PROGRAM
From its modest beginnings with the Bangladesh Rehabilitation Assistance Committee (universally known as BRAC) and the Grameen Bank in the mid-1970s in Bangladesh, microcredit is now a global phenomenon. It has reached anywhere between 150 and 200 million borrowers, mainly women, and is available to many more. It is sometimes described, almost like a character in a Greek myth, as a beast with two teats—a profit mission and a social mission—and by all accounts it has known impressive successes on both fronts. On the one hand, the Nobel Prize for Peace, which was awarded to Muhammad Yunus and the Grameen Bank, crowned a series of public accolades; on the other hand, the Initial Public Offering (IPO) of Compartamos, a large Mexican MFI, in spring 2007 was a (controversial) triumph of the commercial side. The offering raised $467 million for Compartamos, although it also drew attention to the 100-percent-plus interest rates it charges. (Yunus publicly expressed his discontent, calling the CEOs of Compartamos the new usurers, but other MFIs are already following in their footsteps: In July 2010, the IPO of SKS Microfinance, India’s largest microfinance institution, raised $354 million.)
One can see why Yunus may not like the association with usury, but in a (good) sense microcredit is moneylending reinvented for a social purpose. Like traditional moneylenders, MFIs rely on their ability to keep a close check on the customer, but they do so in part by involving other borrowers who happen to know the customer. The typical MFI contract involves loans to a group of borrowers, who are liable for each other’s loans and hence have a reason to try to make sure that the others repay. Some organizations expect the borrowers to know each other when they come to borrow, whereas others bring them together by making them come to weekly meetings. The very act of meeting together every week helps clients know each other better and become more willing to help out a group member who faces a temporary difficulty.8
Like the moneylender, MFIs threaten to cut off all future lending to anyone who defaults outright and do not hesitate to use their connections within village social networks to put pressure on recalcitrant borrowers. Unlike moneylenders, their official policy is never to use actual physical threats.9 However, the power of shame seems to be sufficient. A borrower we met in Hyderabad was struggling to repay loans from several MFIs. But she said she never missed a payment, even if it meant borrowing money from her children or going without a meal for a day: She loathed the idea of having the credit officer come to her doorstep and “make a nuisance” in front of the whole neighborhood.
Where MFIs clearly diverge from traditional moneylending is in removing almost all flexibility. Moneylenders will allow their borrowers to choose how they borrow and the way they repay—some repay once a week, but others repay whenever they have money in hand. Some repay only the interest until they are ready to repay the entire principal. An MFI client, by contrast, typically has to repay a fixed amount every week, starting one week after the loan is given out, and,
at least for first loans, everyone usually receives the same amount. Moreover, the borrower has to make the payment at the weekly meeting, which is always at a fixed time for each group. The advantage of this is that keeping track of repayments is very easy: The loan officer just counts to see if he has the total amount he was supposed to get from that group and if he does, which is almost always the case, he is finished and can start on the next group. This allows a loan officer to collect repayment from 100 to 200 people every day, whereas a moneylender has to wait around not knowing when the money is coming in. Moreover, since the transaction is so simple, the loan officer does not need to be particularly well educated or trained, which also keeps costs down. In addition, loan officers are paid on steep incentive contracts, based on recruiting new clients and making sure that everyone repays.
All of these innovations contribute to reducing the administrative costs of lending, which, as we argued above, get blown up by the multiplier effect and make lending to the poor so very expensive. This is how most MFIs in South Asia manage to make money by lending to the poor at interest rates of around 25 percent per year, whereas the local moneylenders typically charge two to four times as much. Interest rates are higher in other parts of the world (one likely explanation is that loan officer salaries are higher), sometimes even higher than 100 percent per year, but remain much lower than the other alternatives for poor people. In urban Brazil, for example, MFIs offer microcredit at the rate of about 4 percent a month (60 percent a year), and the easiest alternative, which is credit-card-debt refinancing, costs between 12 percent and 20 percent per month (289 percent to almost 800 percent per year). Defaults, famously, are extremely rare, at least outside politically motivated crisis. The “portfolio at risk” (loans that may default, but not all will) was less than 4 percent in South Asia and no more than 7 percent in most Latin American and African countries in 2009.10 And so microfinance, with its 150 to 200 million clients, has earned its place as one of the most visible anti-poverty policies. But does it work?
Poor Economics Page 19