Small Loans, Big Promises, Unknown Impact: An Examination of Microfinance

Finding the best, most effective way to enrich people and polities has always been the ultimate goal of economics. This is evident from Adam Smith’s An Inquiry into the Nature and Causes of the Wealth of Nations all the way up to modern policy debates over, for example, the best ways to reduce unemployment or respond to a financial crisis. While this quest for wealth through economic growth is important for any country, it is arguably most vital for countries in the developing world. It is also there that achieving growth has been most difficult. History is full of strategies for turning poor countries into rich ones: ‘Washington Consensus’-style free markets and privatization; Soviet-style central planning and redistribution; infant industry protectionism; large loans to finance infrastructure investment; heavy education spending; debt forgiveness – etc. In this article I will focus on just one of these development tools, one that is, depending on whom you ask, and how the data is interpreted, either an affordable, feasible, and promising weapon in the fight against poverty; or a failed, ineffective waste of resources better spent elsewhere. This tool is microfinance.

Muhammad Yunus is traditionally credited with inventing modern microfinance while working as a professor of economics during Bangladesh’s 1974 famine. As he tells it, it became too hard “to teach the elegant theories of economics when people [were] dying of hunger all around,” (qtd. In Mourdoch, 1999,1575) so he decided to find out how it was that his elegant theories had so failed these people. He did so by talking to the poor people outside of the university gates. What he found from this informal information gathering was that the greatest economic obstacle to these people was, according to them, the lack of available credit other than that which was offered by extremely usurious local moneylenders. With that in mind, he began to lend his own money to these people (his first loan was one of about 27 US dollars to a group of 42 female furniture makers) and remarkably, most of them paid the loans back. It seemed to him that he had an effective tool with which to fight poverty, and began expanding his lending operations as a formal research project. When the project proved successful, he expanded it into an official, government-sanctioned bank in 1983 – the Grameen Bank.

Since then, Grameen has become one of the major players in microfinance, and has extended its operations to many countries. A large part of Grameen’s success is due to its unique way of making loans, which stresses entrepreneurship, diversified revenue streams, and group lending. To get a loan from Grameen, one must first form a borrowing group with five borrowers who all cosign each other’s loans. The loans are then distributed in waves (two people will get loans and repay them, then the next two, then the final person). The loans are typically used for some sort of investment that will increase the borrower’s future income, e.g. starting a business or getting an education. Unlike other formal loans though, Grameen loans are repaid in installments which begin very soon after the loan is disbursed – usually far before the investment can begin paying off. This means that borrowers must either have other income besides what they expect from investing the loan, or they must have savings that they can dip into in order to repay the loan in the short term. A default on the part of any one member of the group serves to forever cut the entire group off from new Grameen loans, while repayment holds the promise of better terms on later loans (Mourdoch, 1999, 1582-83).

Each of these conditions serves a purpose. The focus on lending to rural borrowers and women ensures that loans go where they can do the most good. Individuals living in rural areas are more likely to be poor and less likely to have access to formal finance (“Rural Poverty in the Developing World,” United Nations) than their urban counterparts, while women tend to be marginalized, with less access to economic opportunity than men (“FAQ,” Grameen Bank). One major assumption goes into saying that loans to poorer people and to women do more good than those to richer borrowers and to men: the idea of diminishing returns; giving a dollar to someone with only a dollar to their name doubles their wealth, giving that person another dollar will only increase their wealth by one half. It is true that they are better off with three dollars than with two dollars, but the marginal benefit they get from that third dollar is less than that from their second dollar. The same principle is at work with economic opportunity. Giving more economic opportunity in the form of a microloan to an impoverished rural woman is better than giving it to a wealthier urban man because the former has less opportunity to start with, so the same resources will do her more good. This assumption appears to be backed up by the data. Khandkar and Samad report participants in microfinance having statistically significantly lower rates of extreme poverty and significantly higher rates of labor force participation and school attendance, especially for girls and women (35-36, 2014).

Group lending however, may be the most interesting, and perhaps the most integral component of Grameen’s success. Under a group lending system, borrowers must form groups to get loans, and while each member gets their own loan, repayment is done as a group. In effect, each member of the group cosigns all of the other group members’ loans. This is done as a remedy for the problems of moral hazard, adverse selection, and voluntary default that plagued pre-Grameen attempts to encourage economic development by providing subsidized credit. Moral hazard, in this context, is the worry that without collateral at risk (again, Grameen loans are uncollateralized); borrowers may not do all they can to ensure that their investment pays off. Adverse selection is the worry that only risky borrowers will take uncollateralized loans. Voluntary default is exactly what it sounds like: defaulting not because your investment fails, but because, by defaulting, you get to keep more of your revenue. All of these problems arise because the lender cannot reasonably hope to observe its borrowers closely enough to see how much work they are putting in, how risky they are, or whether or not they actually made any money on their venture. However, a local lending group can monitor borrowers that closely and can pressure them into working hard, exclude them if they are too risky, and collect payment if they default voluntarily. Adding on a dynamic payment incentive (cutting off the group for any default while promising better future terms for repayment) only serves to reinforce these incentives: if you have to pay for your neighbor’s mistakes, you will watch him; if his mistakes promise to sink your whole future, you will watch him especially closely.

While Grameen’s set of conditions seem to work, and have been widely adopted, there are other methods of microfinance. Some microfinance institutions (MFIs), such as Bank Rakyat Indonesia, make larger collateralized loans to slightly richer, more urban borrowers. It could be argued that by doing so, they are doing less good than other organizations that make loans to the very poor; however, they may also fill a vital niche. If microfinance is only available to the poor, and formal finance is only available to the rich, then there is a risk of creating an unbanked middle class, and with it, a new type of perverse incentive. Poor borrowers may only work to enrich themselves up to a certain point if they know that going beyond that point will make them lose their financial services. Other MFIs, like Bank Kredit Desa (or BKD, also in Indonesia) and various village bank systems do away with group lending, and combat moral hazard, adverse selection, and voluntary default by other means. BKD piggybacks on Indonesia’s federal chain of command by assigning loan officer duties to local village officials, who send information on borrowers up the chain to Jakarta, where it is given to BKD officials (Mourdoch, 1999,1576-1579). This system allows BKD to maintain both scale and individuality, but is not easily applicable to less-stable countries or to those with less clear chains of command. This is where village banks come in. Village banks are banks set up to serve a small geographical area. NGOs usually provide the banks with funding and support, and the villagers with financial training to run the bank (Mourdoch, 1999, 1579). The villagers are then given control over the bank, so they can decide to whom and how much to lend. Because the banks are close to the lenders, they can monitor them, allowing the banks to limit moral hazard, adverse selection, and voluntary default without utilizing group lending.

On paper, microfinance seems like an ideal tool for economic development. It can be targeted at the poorest, most disadvantaged segments of society, then distributed directly, thus reducing the risk of aid expropriation by middlemen. Microfinance is also not vulnerable to criticisms that it discourages work – a problem common to other forms of development aid such as monetary grants and in-kind assistance. Instead, it encourages work that may not have ever been done in the absence of a microloan. Additionally, anyone that simply consumes borrowed money instead of using it productively will most likely find themselves unable to repay. Going hand in hand with its encouragement of work, microfinance is not necessarily a losing proposition for the donors. The donors get their money back, and may even make a profit. All in all, it seems like a win-win proposition: the poor get help to pull themselves out of poverty with their own labor, and the donors end up at least as well off as they were. Why then, has microfinance not been widely adopted and poverty all but eradicated?

In short, it is because microfinance simply doesn’t seem to live up to its promise as a panacea for poverty. There are a variety of reasons why this is the case. First among them though, is the fact that the effects of microfinance on poverty seem to be modest at best. Much of the evidence of extreme, transformative effects of microfinance has been anecdotal and unrepresentative of microfinance’s overall effects. More rigorous testing of microfinance programs, such as those utilizing randomized controlled trials (RCTs), has shown far less dramatic effects. In fact, most microfinance programs, according to a paper by Khandker and Samad “find limited or no benefits of microfinance” (2014, 4). This is not to say that microfinance programs do not help anyone, but rather that they do not help the average person, or at least do not help them much. The same study reports no significant differences in per capita income or expenditure between borrowers and non-borrowers in any of the examined time periods. Other measures, such as significantly less “extreme poverty” (a measure not well explained in the paper) for borrowers in all periods, as well as higher net worth asset value in earlier periods might imply that microfinance only works for those that are extremely poor, i.e. those considered poor even by Bangladeshi standards. However, when one looks beyond monetary measures, the picture begins to look more promising. Borrowing households show male and female labor supplies that become significantly higher than those for non-borrowing households over time (Khandker and Samad, 2014, 35). Borrowing households also tend to send their daughters to school at a higher rate than non- borrowing households. Measures such as these lend some credence to the idea, often peddled by MFIs that, even if microfinance does not lift people out of poverty on a large scale, it can still improve their lives by giving them new labor and educational opportunities, as well as potentially valuable entrepreneurial experience.

Multiple other papers point to similar outcomes from microcredit. One, by Banerjee, Karlan, and Zinman collates the results of six randomized control trials in six different countries (Bosnia, Ethiopia, India, Mexico, Morocco, and Mongolia) that sought to test the effects of microcredit on both individuals and on communities. The big takeaway from this paper is that they found little evidence of major, “transformative” effects on the average borrower, for better or for worse. It does appear that access to microcredit increases “micro-entrepreneurial activity” (Banerjee, et al. 2015, 11) however, an increase in activity is not the end goal of microfinance. Entrepreneurship is a means to an end – that end (or those ends) being economic development and poverty alleviation. It is less clear that increased entrepreneurial activity is accomplishing these goals. None of the studies found any statistically significant increases in either income or household consumption, despite the uptick in entrepreneurial activity (Banerjee, et al., 2015, 12).  The authors postulate that as workers take microloans and start businesses, they either quit old jobs or work less on them, thus decreasing their wage incomes, but supplanting it with business income. This also explains the lack of effect on household consumption – consumption is a function of income, so static income levels should translate into static consumption levels. However, Banerjee et. al also find that borrower households spend less on things like ‘temptation goods’ and recreation, implying that while their level of consumption may not change, its composition does, possibly hinting at more disciplined spending.

Just as in the Khandker and Samad paper, Banerjee et al. finds that microfinance’s lack of effect on monetary well-being hides other positive effects. They do not find significant effects on school enrollment, but do note increases – albeit small ones – in subjective measures such as “female empowerment…happiness, trust in others… [and] occupational choice” (Banerjee et al., 2015, 13-14). So it seems that microcredit may make people’s lives better, even if not in a tangible, material way.

So far, the results of this paper line up almost exactly with those of Khandker and Samad. What sets this paper apart is their analysis of the statistical problems in their results in particular and in evaluating microcredit in general. Banerjee et al. note that rates of take-up for microcredit are relatively low – in the neighborhood of 17-34% (2015, 2) – which presents a problem for econometric analysis. Such low take up not only implies that people are sorting themselves into treated or non-treated groups (the bane of randomized trials); it also implies that the experiments may not be looking at the right people. It is very possible that the people who stand to benefit from microcredit (for example, the most talented, entrepreneurial people) are already using it and benefiting from it. Thus, the marginal borrowers available for treatment in an RCT may be the ones less able to benefit from microfinance, for whom there is no significant, objective benefit for taking microloans. If the effects of microfinance on inframarginal borrowers (those already using microcredit before the RCT) are significant, it would vindicate microfinance, but without a way to control for other influences on poverty, the effects of microcredit on inframarginal borrowers are nigh-impossible to determine.

Evaluating microfinance also runs into time frame problems. Even the most immediate effects of receiving a microloan take time to appear, as microloans are designed to create an eventual increase in the flow of borrowers’ wealth ( i.e. their income), not in their stock of wealth – they are not cash transfers. This has been cited – as in Khandker and Samand – as a possible explanation for the disappointing results of most RCTs, which usually only look at the immediate effects of microcredit provision (2014, 2). While Khandker and Samand bring that issue up only to argue that their twenty year panel of data better captures long-run effects, they, too may be guilty of looking at too narrow a time frame. Twenty years may be enough to judge that microfinance does not have economic benefits for a generation of borrowers, but economic development is a long-run phenomenon, and one generation’s access to microfinance may have effects that percolate down through future generations. If the children and grandchildren of borrowers end up earning and consuming more than those of non-borrowers, then we may be able to conclude that microfinance is a useful development tool. Only time and further research will bring us better data on microfinance’s intergenerational effects.

Just as there are reasons to believe that microfinance is more effective than its detractors proclaim, there are also reasons to believe that it is less effective than its proponents would have one believe. The traditional argument against microfinance is that the repayment schedules trap farmers in debt such that taking out microloans impoverishes borrowers further, especially when they have to take out high- interest, informal loans to pay off microloans. While there may be cases of this happening, informal borrowing among microcredit borrowers falls, pointing away from debt traps as a common occurrence, however, average effects can hide heterogeneous outcomes, and it is possible that certain groups are more susceptible to debt traps than others (Banerjee, et al., 2015, 10-11).

Other objections to microcredit center on how the funds are used. Most loans go towards setting up a small business from which the borrower can make more money than they would otherwise. Intuitively, this seems like it should work. MFIs ease credit constraints, so small businesses pop up. And because the borrowers started off so poor, rates of return to capital should be very high. In fact, it has been confirmed that small businesses in developing countries have high rates of return – in the neighborhood of 60% per year (de Mel, McKenzie, and Woodruf, 2008, 1333). However, de Mel et al. found that they may also fall prey to the time frame problem. Pessimists would argue that the kinds of businesses that microfinance props up are ones that are likely to operate in a perfectly competitive environment, especially when MFIs tear down barriers to entry such as a lack of credit. If this is the case, then as time goes on, and more microfinance-backed enterprises come into the market, the high profits and returns to capital will be competed away, leaving late borrowers no better off for having started a business. As with so many aspects of microfinance, only time and further study of microenterprise profits will give us definitive answers to these questions.

Other detractors might ask why – if rates of return are as high as they are – do we need microfinance in the first place? Why aren’t investors lining up to throw money at Bangladeshi furniture makers?  The big answer to that lies with where microfinance borrowers are. Most live in rural areas in poor countries – which places them in the least financially developed regions of the least financially developed countries in the world. Without the correct infrastructure (both physical and financial) it is hard to assess creditworthiness, monitor money use, and enforce repayment. This leaves a gap in formal financing among the very poor, which is the very problem that Muhammad Yunus identified in 1974, and sought to solve with his Grameen Bank.

If the question is simply one of whether or not microfinance has been able to extend credit to the poor, then microfinance has been a success. But providing credit is not the end all of microfinance. Rather, it was a means to the end of poverty elimination, and while poverty has not been eliminated, microfinance may have made a dent in it. But it is hard to say for sure. There are reasons to doubt microfinance’s effectiveness, but the data seems to reveal these objections as far from damning. The data seems to point to microfinance doing at least some good, even if that good is hard to quantify – as is the case with claims that microfinance borrowers are happier or more economically independent. Problems with the data may suggest that microfinance is more effective than it appears, while a lack of data pointing at microfinance leading to debt traps and further impoverishment make it unlikely that microfinance is less effective than it appears. Thus, microfinance has been far from vindicated. But it has also not been totally disproven, and in the face of inconclusive data, evaluation problems, and a lack of other tools rivaling microcredit’s visibility and promise, it seems foolish to write it off as a failure just yet.


Works Cited:

Banerjee, Abhijit, Dean Karlan, and Jonathan Zinman. “Six Randomized Evaluations of Microcredit: Introduction and Further Steps.” American Economic Journal: Applied Economics 7.1 (2015): 1-21. Print.

De Mel, Suresh, David McKenzie, and Christopher Woodruff. “Returns to Capital in Microenterprises: Evidence from a Field Experiment.” The Quarterly Journal of Economics 123.4 (2008): 1329-372. Print.

“FAQ.” Grameen Bank Bank For Small Business RSS. Grameen Bank. Web. 5 July 2015.

“History of Grameen Bank.” Grameen Bank. Web. 28 June 2015.

Khandker, Shahidur R., and Hussain A. Samad. “Dynamic Effects of Microcredit in Bangladesh.” Policy Research Working Papers (2014). Print.

Morduch, Jonathan. “The Microfinance Promise.” Journal of Economic Literature (1999): 1569-614. Print.

“Rehabilitation and Attack.” The Economist. The Economist Newspaper, 19 Apr. 2014. Web. 25 June 2015.

“Rural Poverty in the Developing World.” UN News Center. United Nations. Web. 5 July 2015.