SEARCH

SEARCH BY CITATION

Keywords:

  • Microcredit;
  • Microinsurance;
  • Impact evaluation;
  • Information asymmetry
  • F35;
  • O19

Abstract

  1. Top of page
  2. Abstract
  3. I. INTRODUCTION
  4. II. WELFARE IMPACT OF MICROCREDIT
  5. III. MECHANISM UNDERLYING HIGH REPAYMENT RATES
  6. IV. CHALLENGES OF MICROFINANCE INSTITUTIONS
  7. V. SUMMARY AND CONCLUSIONS
  8. REFERENCES
  9. Appendix

“Microfinance revolution” is the term often applied to the successful expansion of small-scale financial services to the poor with high repayment records in developing countries. The present paper investigates the extent to which the microfinance revolution is truly revolutionary. More specifically, it explores the impact of microfinance institutions on the poor, the mechanisms underlying high repayment rates and their innovations, and the new challenges microfinance institutions are currently facing. Different from the existing published survey literature, we focus on current topics and attempt to show recent theoretical developments in a comprehensive manner using simplified models with very similar settings. We contend that microfinance is developing in a promising direction but has yet to reach its full potential.


I. INTRODUCTION

  1. Top of page
  2. Abstract
  3. I. INTRODUCTION
  4. II. WELFARE IMPACT OF MICROCREDIT
  5. III. MECHANISM UNDERLYING HIGH REPAYMENT RATES
  6. IV. CHALLENGES OF MICROFINANCE INSTITUTIONS
  7. V. SUMMARY AND CONCLUSIONS
  8. REFERENCES
  9. Appendix

The eradication of poverty continues to be a central political issue in most developing countries. Despite serious efforts by local governments as well as bilateral and multilateral donor communities over the past few decades, many people still suffer from poverty. According to estimates by the World Bank, more than 1.2 billion people were classified as extreme poor in 1998, living on less than US$1 a day (World Bank 2000). To combat pervasive poverty, the United Nations declared as the first of its Millennium Development Goals (MDGs) to halve the proportion of people who suffer from extreme poverty by 2015 compared with the 1990 level, and called for unprecedented commitment from nations worldwide.

It is widely recognized that limited access to financial services is one of the major bottlenecks preventing poor households from improving their livelihoods. The underdevelopment of the financial markets is felt especially keenly by the rural poor who live in low-density population areas and in riskier environments subject to greater seasonal income fluctuations, and who generally demand smaller-sized loans and have smaller savings accounts, all of which lead to high transaction costs shouldered by financial institutions. Moreover, imperfect information significantly increases default risks caused by adverse selection, moral hazard, and strategic default (e.g., Stiglitz 1990; Ghatak 1999), which make formal banks reluctant to offer services to the poor who cannot supply sufficient collateral to secure their loans. As a consequence, poor households tend to be excluded from formal financial services, not only borrowing but also savings and insurance, which in turn prevents the poor from investing in various profitable projects, smoothing consumption, and improving their ability to cope with various unexpected shocks (Okten and Osili 2004; Conning and Udry 2007).

In the absence of sufficient collateral to pledge, the poor generally have to rely on loans from informal moneylenders at high interest rates or from friends and family, whose supply of funds is generally limited (Weiss and Montgomery 2004; Conning and Udry 2007). In order to supply cheaper credit, many developing countries intensively implemented large-scale subsidized credit programs between the 1950s and 1970s. These programs were originally intended to target rural farmers with the hope that subsidized credit at below-market interest rates would not only increase their income but would also promote investment in agricultural modernization, such as in the diffusion of fertilizer and high-yielding crop varieties, thereby increasing food production and, subsequently, stimulating overall economic growth (Yaron and Benjamin 2002). Nevertheless, a growing body of evidence suggests that the subsidized credit programs were unsuccessful in that they induced local political elites to take advantage of below-market interest rates, leading to rent-seeking activities, insufficient outreach to the rural poor farmers, very low loan repayment rates, and inefficiency in financial markets with excess demand (Adams, Graham, and Von Pischke 1984; Robinson 2001; Zeller and Meyer 2002).

As an alternative, microfinance has recently attracted growing attention as a means of overcoming such a situation (Morduch 1999). Most microfinance institutions (MFIs) provide collateral-free small loans to low-income households who have long been deemed to be unbankable. These loans are generally expected to be used for self-employment and income-generating activities. Even though many MFIs rely on financial support from the state and donors to expand their outreach, in response to the widespread failure of government programs in the past, financial self-sustainability has also been of major concern, and appropriate pricing of loans is viewed as crucial for sustainability (Zeller and Meyer 2002). Therefore, interest rates are generally set at the market level to cover operational costs, and most microfinance programs across the world have extremely high repayment rates at around 90% to 98%.

Although the provision of credit is by far the most important product of financial services, much progress has also been made by many MFIs offering a range of savings and insurance products, which has great potential for alleviating poverty and reducing vulnerability (Nourse 2001; Robinson 2001; Churchill 2002; Zeller and Meyer 2002; Armendáriz de Aghion and Morduch 2005). With enthusiasm for their ability to improve the welfare of the poor, the number of MFIs has rapidly grown, from 618 to 3,133 between 1997 and 2005, and, correspondingly, the number of clients who benefitted from microfinance increased from 13.5 million to 113.3 million over the same period (Daley-Harris 2006). Given this growth in the microfinance industry, the United Nations declared 2005 to be the “International Year of Microcredit” and attempted to link the microfinance movement with the achievement of the MDGs. Furthermore, a leading MFI, the Grameen Bank, and its founder, Professor Yunus, received the Nobel Peace Prize in 2006 “for their efforts to create economic and social development from below.” In this way, microfinance is increasingly gaining popularity as an effective tool for poverty reduction, and its evolutionary process is often termed the “microfinance revolution.”

Yet, microfinance is not without controversy. Many studies have examined whether microfinance is as effective as originally expected. For example, although most people have agreed that microfinance is good for the poor, it should be proved in a rigorous way. The recent development of impact evaluation techniques allows us to go in the direction of asking whether microfinance is actually good for the poor. Another strong belief bolstered by the success of the Grameen Bank has been that its lending scheme, characterized by non-collateralized, jointly liable, group-based lending, is effective in maintaining the high repayment rate. However, recent theoretical as well as empirical studies have cast doubt on this and show there are other mechanisms underlying the high repayment rates. In addition, the belief that the increase in the number of MFIs relaxes credit constraints of the poor has come under challenge. There is also a growing concern that MFIs, often operated by nongovernmental organizations, should put more effort into becoming regulated for-profit financial institutions that work to enhance financial sustainability instead of primarily pursuing services to the poor. This kind of commercialization of microfinance has gained momentum since the 1990s (Drake and Rhyne 2002; Cull, Demirguc-Kunt and Morduch 2007). However, there is concern that MFIs might shift away from serving poorer clients in pursuit of commercial viability. This is referred to as “mission drift.”

In this study, we review the recent advancements in the microfinance literature to highlight improved understanding of microfinance issues and explore whether the microfinance revolution is truly revolutionary. More specifically, we will examine first whether microfinance has really improved the welfare of the poor, and then turn to investigating mechanisms explaining the high repayment rates. We will then briefly explain the current challenges MFIs are facing, including increasing competition among themselves. Finally, we turn our attention to microinsurance as an innovative trial that can potentially protect the poor from risks but is currently a challenging issue for MFIs because it can generate substantial losses for insurers. Note that we do not cover all recent important issues in this paper, such as commercialization of microfinance and mission drift, but instead focus on the recent issues of selected topics. One of the major contributions of this study is to show recent theoretical developments in a comprehensive manner using simplified models with very similar settings.

The remainder of this paper is structured as follows. In Section II, we examine the impact of microfinance. In Section III, we discuss how high repayment rates can be achieved under asymmetric information. We discuss the current challenges facing MFIs in Section IV, and present a summary and conclusions in Section V.

II. WELFARE IMPACT OF MICROCREDIT

  1. Top of page
  2. Abstract
  3. I. INTRODUCTION
  4. II. WELFARE IMPACT OF MICROCREDIT
  5. III. MECHANISM UNDERLYING HIGH REPAYMENT RATES
  6. IV. CHALLENGES OF MICROFINANCE INSTITUTIONS
  7. V. SUMMARY AND CONCLUSIONS
  8. REFERENCES
  9. Appendix

Whether microcredit truly helps to improve the welfare of the poor is a fundamental question asked by both practitioners and scholars. In this section, we will examine this question while reviewing the existing published literature. Before discussing the empirical evidence in detail, we present key problems in the impact evaluation with the aim of helping the reader to gain insight into why most impact studies should be interpreted with great care.

A. Evaluating the Impact

Although anecdotal evidence of the welfare impact of microcredit is mounting in the literature, empirical evidence based on rigorous evaluation is scarce. As Karlan and Goldberg (2007, p. 1) point, the rigorous impact evaluation has to answer “How are the lives of the participants different relative to how they would have been had the program, product, service, or policy not been implemented?” This requires the comparison of two potential outcomes, such as income, business profits, or physical and human capital investment, of the same individual, i.e., one with the treatment and the other without it. Because we can never observe both statuses for a particular individual simultaneously, the existing reports often try to compare the states before and after microcredit programs to estimate their impact on individuals (before–after comparison). However, in most cases, this does not provide reliable estimates because other factors like macroeconomic shocks also affect the post-treatment outcome. In other words, this approach fails to isolate the impact of microfinance from the time trend that affects the result. This implies that it is almost impossible to measure the impact of a program on a given individual (Duflo, Glennerster, and Kremer 2006). However, it is possible to obtain the average impact of microfinance if the counterfactual outcome of the treatment group can be constructed from the pool of the remaining population who have a similar outcome to the treatment group in the absence of the treatment. This can serve as a comparison group. Therefore, a major challenge for impact evaluation is to create a good counterfactual through the use of appropriate techniques under a set of acceptable assumptions.

A line of studies have used non-clients to approximate the counterfactual of clients without treatment (with–without comparison). In this case, the average difference in outcome of interest between clients and non-clients is regarded as the impact of microcredit. This is a good strategy as long as the outcome is independent of participation in microcredit. However, under a situation where participation is voluntary, it may well be that clients who decided to participate significantly differ from non-clients in terms of their expected gains and other relevant characteristics explaining participation and outcomes. If there are systematic differences between them, the outcome of non-clients cannot represent those of clients without treatment, and, as a result, the evaluated impact is highly likely to be biased. Even if such self-selection does not matter, nonrandom program placement by MFIs can potentially yield a similar bias. For example, MFIs might implement a program in villages where impacts are expected to be high or where poverty is more pervasive. In the former case, the outcome of the selected villages might be higher than for nonselected villages even without microcredit programs, so that the impact tends to be biased upward. The opposite will hold in the latter case (Morduch 1999).

To illustrate, suppose we are interested in the average treatment effect (ATE) and the average treatment effect on the treated (ATT), two of the most widely known impact measurements. ATE measures what impact program participation would have on individuals/households drawn randomly from the population, whereas ATT measures what impact program participation has on individuals/households who actually participated. Symbolically, letting y1 be the outcome with treatment and y0 be the outcome without treatment, ATE can be expressed as

  • image

where E(·) is an expectation operator. Similarly, letting d denote the treatment indicator which takes the value 1 if treated and 0 otherwise, ATT can be written as

  • image

Provided that we can observe both y1 and y0 for any given individual, the average differences of E(y1y0) as well as E(y1|d= 1) −E(y0|d= 1) should be attributable to the difference in the access to microcredit programs, because factors other than the treatment status are exactly the same. As such, if E(y1y0) and E(y1|d= 1) −E(y0|d= 1) are positive, it can be safely claimed that microcredit has a positive effect on the outcome of interest and vice versa.

The common approach used by various research reports simply compares program participants with nonparticipants and treats E(y1|d= 1) −E(y0|d= 0) as the average impact. Yet,

  • image

The last equality holds if the treatment status is independent of outcome without treatment; that is, d?y0 and E(y0|d= 1) =E(y0|d= 0). If program participants are more entrepreneurial, or if the program is placed in a more promising area, it is likely that the expected gain of clients will be higher than that of non-clients even without treatment, such that E(y0|d= 1) > E(y0|d= 0), which overestimates the true ATT. By the same token,

  • image

if d? (y0, y1). Therefore, unless the condition of d? (y0, y1) is satisfied, the estimated impact will be exaggerated.

The experimental approach, known as randomized control trials or randomized field experiments, where program placement or eligibility to participate in programs is assigned randomly to the population, is one of the most powerful tools for solving selection biases because as long as the sample size is large enough, the law of large numbers will ensure the mean independence, i.e., E(y0|d= 1) =E(y0|d= 0), and, therefore, E(y1|d= 1) −E(y0|d= 0) gives a consistent estimate of ATT under the following assumptions: (i) that there are no spillover effects, i.e., the treatment effect is not spilled over to the untreated group, which is known as stable unit treatment value assumption (Wooldridge 2002); and (ii) that there is no repercussion on others through market mechanisms, called the general equilibrium effects (Heckman, Lochner, and Taber 1998).1

Although the experimental approach is being adopted by more and more studies on impact evaluation, it is not always feasible to do. In particular, strictly randomizing participation in the program is quite difficult because it needs to force people who are unwilling to participate to become part of the treatment group. Therefore, it would only be possible for policymakers to randomize eligibility. A problem of this approach is that eligibility is not equal to treatment, so that we cannot obtain ATT or ATE as long as some individuals opt out. In this case, we can estimate intention-to-treatment (ITT), in which the average impact of the availability of the program, rather than participation in the program, is evaluated. To best of our knowledge, there are two papers that use ITT based on randomized controlled trials to evaluate the impacts of microcredit, i.e., Banerjee et al. (2009) and Karlan and Zinman (2009a).

Although an ITT estimator is useful, it critically depends on the share of people who actually participate in the program, which again depends on the outside option in that society, which varies across location and time. Thus, a critical shortcoming of an ITT estimator is the difficulty generalizing the estimation results or derived implications to other societies (external validity) (Ito 2007).

Partly because of such a problem, there are still many studies that rely on a nonexperimental or quasi-experimental approach, which constructs the counterfactual from observational data. Several statistical methods based on these approaches have been developed. These include: (i) a matching method, (ii) an instrumental variables method, (iii) a regression discontinuity design, and (iv) a difference-in-difference approach. In the following subsection, we will briefly explain the basic concepts and assumptions underlying these methodologies.

B. Nonexperimental Evaluation Design

1. Matching method

The matching method attempts to find a comparison group that is identical or very similar, ideally in all aspects, to the treatment group except for the treatment status. Then, the average outcome of the selected comparison group is compared with that of the treatment group. Similarity between the two groups is generally evaluated by observable characteristics. In other words, this method relies on the assumption that conditioning on observable characteristics, participation in microcredit is independent from the outcome of interest, which can be expressed by (y0,y1) ?d | x for ATE and y0?d | x for ATT, where x is a set of observable characteristics. These assumptions imply conditional mean independence, i.e., E(yd|d= 1, x) =E(yd|d= 0, x) =E(yd|x). Therefore, for ATT,

  • image

Because the last equality holds by the assumption of E(y0|d= 1, x) =E(y0|d= 0, x), ATT can be estimated consistently, unless unobservable characteristics are important determinants of selection to participate. The same applies for ATE.

One of the shortcomings of this method is that if x is high-dimensional, it is quite difficult to find a good comparison group similar to the treatment group in alldimensions of x. Rosenbaum and Rubin (1983) show, however, that matching on a single index that captures the propensity to participate conditional on x gives consistent estimates of the treatment effect in the same way as matching on all the elements of x. Let (d =1 |x) =Pr(x) denote the probability of participation given observable covariates x. To obtain unbiased ATT, an important assumption is that conditional on the probability of participation, treatment is independent of outcome in the absence of treatment, y0?d |Pr(x). In addition, there should be a substantial overlap in covariates x between comparison and treatment groups, so that individuals with the same x have a positive probability of being both participants and nonparticipants, i.e., 0 < Pr(x) < 1. Then, if we can find matched pairs of the comparison and treatment groups over the region of common support, who have exactly the same propensity scores, the simple average difference between the matched comparison and treatment groups can be treated as the impact of microcredit. If propensity scores between these two groups are close, but not identical, we need some adjustment for weighting determined by the distance between propensity scores of the treatment and comparison groups.2

A modified matching method is the so-called “pipeline comparison,” where tenured participants are matched with incoming participants who have applied for but not yet received loans from MFIs. The incoming clients are seen as a good comparison group because they are also self-selected into the program presumably in the same way as the tenured clients. Thus, biases caused by self-selection are potentially controlled for and the resultant differences in the outcome of interest are attributable to microcredit. Although many studies have adopted this pipeline comparison, there are potential problems associated with this methodology. We will explain those problems and discuss several studies that overcome such problems in the next subsection C.

2. Instrumental variables method

The instrumental variables method is a statistical technique that controls for selection biases. It requires variables affecting participation to program but not outcome as identification. In a regression framework,

  • image

where yi is an outcome of interest of household i, e is an error term, α is a constant term equal to E(y0| d= 0), and β is a parameter equal to E(y1| d= 1) −E(y0| d= 0). As is well known, if d is correlated with e, the OLS estimator for β will be biased.

Now, suppose that there is an instrumental variable, z, that satisfies corr(zi, di) ≠ 0 and corr(zi, ei) = 0, and takes the form of

  • image

where corr(zi, ui) = 0. Combining the two equations gives

  • image

Because corr(zi, ei) = 0 and corr(zi, ui) = 0 by assumption, the above reduced-form estimation is unbiased. Then, dividing ββ2 by β2, we can estimate the main parameter β consistently. In a special case, where z is binary with a single index d and further d1id0i (monotonicity) holds, β becomes

  • image

which is known as the Wald estimator. Because this estimates the change in y relative to d when such a change is induced only by the change in z from 0 to 1, the estimated impact does not reflect the average impact on the whole population. Rather, it identifies the impact on the subpopulation who takes the treatment only when offered. In this sense, the instrumental variables estimator thus calculated is often termed the local average treatment effect (LATE).

3. Regression discontinuity design

The regression discontinuity design is an idea that individuals around some critical cut-off point for program eligibility are similar. For example, suppose that MFIs target individuals with less than 1 ha of land and that those with just above 1 ha of land are ineligible to participate. Because this eligibility criterion is exogenously determined by MFIs, it would be reasonable to assume that both observable and unobservable characteristics of households are uncorrelated with eligibility. In contrast, the probability of participation as well as outcomes of just below and above the cut-off point would be quite different due to eligibility. Based on these assumptions, the regression discontinuity design compares outcomes of individuals just below the cut-off point for eligibility with those just above the cut-off point for eligibility.

Formally, let c denote some cut-off point on a certain variable C, which governs the program eligibility and d= 1 if C > c and 0 otherwise. If this rule is deterministic (sharp regression discontinuity), the impact estimator can be written by E(y1| cC < c+e) −E(y0| ceC < c) for small e.

In contrast, if the eligibility condition C > c is enforced with error (fuzzy regression discontinuity), we must scale up the differences, by dividing the difference in the probability of treatment:

  • image

This is equivalent to the Wald estimate using a dummy for C > c as an instrument for treatment status. Therefore, this also assesses the mean impact on the selected subpopulation around the cut-off point rather than the mean impact on the population as a whole (Ravallion 2008).

4. Difference-in-difference method

The difference-in-difference (DID) method is basically a combination of before–after comparison and with–without comparison.3 It compares observed changes in outcome of participants with those of nonparticipants over time. As argued previously, although before–after comparison alone or with–without comparison alone generally fails to account for selection problems, the DID method can eliminate biases due to fixed unobservable characteristics. To express it more formally, DID in the regression form can be written as

  • image

where yit is an outcome of interest of household i in year t, β's are parameters to be estimated, T is the time period, which takes the value of 1 for the post-treatment and 0 for the pre-treatment period, and e is an unobserved error term. In this specification, the parameter β2 is the DID estimator. Estimation biases emerge if the error term is correlated with the treatment status, such that corr(eit, Di) ≠ 0. Yet, suppose the error term e comprises the time-invariant component ν and a mean zero time-varying component ε, i.e., eit=vi+εit, and suppose further that the time-varying component is independent of participation, i.e., εit?Di, for all households i and that εit is not serially correlated, i.e., corr(εit, εit+1) = 0. Taking the first difference, the above equation can be rewritten as

  • image

where Δ represents changes in corresponding variables over time. This equation clearly shows that the impact of time-invariant characteristics, vi, is effectively differenced out. Because Δεi is uncorrelated with Di by assumption, the DID estimator β2 is unbiased. Then, analogous to the above ATTs, it takes the form of

  • image

The last term of the right-hand side of this equation will cause potential biases if Ey0|D= 1) ≠Ey0|D= 0). Therefore, the validity of DID rests on “parallel time drift,” where the change in outcome in the comparison group between pre-treatment and post-treatment periods is identical with that in the treatment group in the absence of treatment.

The DID method so far assumes that the data on both pre-treatment and post-treatment statuses are available. A modification of this approach is that, instead of using differences in participants and nonparticipants between pre-treatment and post-treatment periods, a difference in eligible and non-eligible households in program villages is compared with the same difference in nonprogram villages. A modified assumption in this case is that a difference between eligible and non-eligible households in nonprogram villages is a good counterfactual of the same difference in program villages in the absence of treatment. If this assumption is satisfied, DID would give a consistent estimator.

As is shown, these methodologies estimate different impacts of microcredit, depending critically on specific assumptions. The validity of such different assumptions and resultant estimated impacts are not without controversy. In the next subsection C, we examine what the recent published empirical literature tells us about the impact of microcredit.

C. Empirical Evidence of Impact Evaluation

Of the estimation strategies explained earlier, the pipeline comparison has gained momentum among microfinance practitioners because of its simplicity and low financial burden. Indeed, this methodology requires neither panel data nor interviews with non-clients. Applications of this method include a series of Assessing the Impact of Microenterprise Services (AIMS) program publications, such as Barnes, Gaile, and Kibombo (2001) in Uganda and Dunn and Arbuckle (2001) in Peru, and others like Mosley (2001) in Bolivia, and UNCDF (2004) in Nigeria, Malawi, Haiti, and Kenya. These studies generally find positive and significant impacts of microcredit on enterprise profits and the welfare of tenured clients.

However, Karlan (2001) and Alexander-Tedeschi and Karlan (2010) argue that this approach is flawed. They identify possible biases brought about by dropout, timing of decision, and institutional dynamics. First, dropout biases emerge if the remaining tenured clients systematically differ from ex-clients who no longer receive microcredit at the time of survey. For example, successful clients, who can improve business profits sufficiently, will accumulate their own savings, no longer need microcredit, and eventually graduate from the program. In contrast, unlucky clients might fail to invest the money well, feel ineffective, and leave the program. In any case, the remaining clients may be different from ex-clients in their essential experiences, leading to biases in the estimation of the impact. Second, timing of decision problems occur if the incoming clients have some good reasons for not participating in the first place. For example, they may be less entrepreneurial or more risk averse than tenured clients. It is also possible that the incoming clients take advantage of being latecomers by observing the past experience of tenured clients in business management with credit. In both cases, the key assumption of pipeline comparisons becomes invalid because the incoming clients are self-selected into the program in a different manner from the tenured clients, and, therefore, the resultant difference of outcomes cannot be attributed solely to microcredit. Third, institutional dynamics biases emerge if MFIs expand their outreach strategically. For example, MFIs might first operate in promising areas that have good clients, then after achieving comfort with the local culture, economy and business practices move to poorer areas to serve the poor. If this is the case, the characteristics of tenured clients and incoming clients are highly likely to differ systematically.

Using data from Peru, Alexander-Tedeschi and Karlan (2010) show that the failure to take dropouts into account leads to a significant upward bias: Annual enterprise profit is positive and higher for tenured clients by 4,083 nuevos soles with the pipeline comparison, while it is surprisingly negative and lower by 588 nuevos soles with the inclusion of the dropouts in the sample. As for household income from all sources, the impact of microfinance remains positive, but drops from 6,569 nuevos soles in the pipeline comparison to 2,062 nuevos soles in their preferred approach. Similarly, Alexander-Tedeschi (2008) uses the two rounds of panel data in Peru to examine the potential biases caused by dropouts, timing of selection, and program placement, and find that the pipeline comparison significantly overestimates the benefit of microfinance due mainly to dropouts and timing of selection. Although the impact of microfinance on enterprise profit is still positive after controlling for such biases, these findings suggest that the results of pipeline comparison should be interpreted with caution.4

Coleman (1999) studies the impact of microfinance undertaken by village banks in Thailand using a similar approach to that of Alexander-Tedeschi (2008). He surveys not only tenured clients and non-clients in treated communities where banks were already in operation, but also incoming clients and non-clients in control communities where banks were not yet in operation, which allowed the author to apply the DID method (i.e., comparison of the difference between tenured clients and non-clients in treated communities against the same difference in control communities). One major advantage of Coleman's study over similar pipeline comparisons is that the order of program placement is random. Even though this method cannot perfectly control for biases caused by dropouts as pointed out by Montgomery (2005), it is considered to be more credible than purely comparing clients with non-clients or tenured clients with incoming clients (Karlan 2001; Karlan and Goldberg 2007).5 Like Alexander-Tedeschi (2008), the estimation result shows that the failure to account for the selection process significantly overestimates the impact of microfinance. Indeed, Coleman (1999) finds that the correct specification shows insignificant impacts on physical assets, production, sales, expenses, labor time, and expenditures on health care and education. Moreover, he finds that clients of microfinance are more likely to borrow from informal money lenders. He explains that this is because many clients joined the program due to social reasons, such as “being a part of group,” without having identified projects to invest in. Therefore, they tend to use loans for consumption purposes, and when they have to repay microcredit, they borrow money from informal moneylenders due to the lack of money at hand.

Coleman (2006) extends his previous analysis to include the dropouts in order to control for potential dropout biases suggested by Karlan (2001) and also to examine why microcredit has little impact. To explore the latter issue, he differentiates committee members from ordinary members of the village banks as the committee members constitute the relatively better-off segment in the communities. Similar to his previous study, Coleman (2006) finds that the impacts of microcredit on the ordinary members are largely statistically insignificant or sometimes even negative and significant, whereas those on committee members are mostly significantly positive in various outcomes, including income, savings, productive expenses, and labor time. This implies that the negligible average impacts of microcredit as found in the previous study are largely because the microcredit under study does not bring about benefits for the poor. Applying the method related to Coleman (2006), Kondo et al. (2008) also show that the benefits of microfinance are disproportionally captured by the wealthier households in the rural Philippines.

Using panel data collected in Indonesia in 2007 and 2008, Takahashi, Higashikata, and Tsukada (2010) also address a similar research question. Based on propensity score matching combined with the DID method, they ask the extent to which microcredit improves the welfare of clients and whether it helps the poor. The advantage of their method is that estimation biases arising from differences in observable characteristics can be controlled for by propensity score matching, while those arising from time-invariant unobservable characteristics can be controlled for using the DID method. Their results show that the impacts of microcredit on household income and profits of self-employed businesses are largely insignificant, whereas those on sales (revenues) of self-employed businesses are positive and significant, implying that microcredit contributes to enlarging business size, but not profits. However, once the sample is divided into poor and nonpoor households, the effect of the increased business sales is positive and significant only for the nonpoor. Besides, the poor can benefit from microcredit to increase their schooling investment in their children. Based on these findings, they conclude that, although microcredit can potentially contribute to the reduction of intergenerational poverty through schooling investment, it might not have immediate impacts on poverty alleviation.

A study by Pitt and Khandker (1998) in rural Bangladesh is among the most influential and frequently cited papers in the impact evaluation literature. They collected their sample from both program and nonprogram villages and include village fixed effects in estimating the impacts in order to mitigate program placement biases. Furthermore, they use exogenous eligibility criteria to identify the impacts. The NGO microfinance programs they study target households that own less than half an acre of land. This rule is exogenously determined by the NGOs, and, therefore, participation in microcredit can be treated as exogenous to some extent provided that households cannot frequently engage in land transactions, and this rule is strictly applied by the NGOs. Using this eligibility rule as an identification strategy, Pitt and Khandker present an application of regression discontinuity design, with an intricate econometric technique called weighted exogenous sampling maximum likelihood–limited information maximum likelihood–fixed effects (WESML–LIML–FE). Pitt and Khandker (1998) show, among other things, that every additional 100 taka to a woman increases household consumption by 18 taka, that the increased consumption is more apparent in food-shortage seasons, which is an indication that microcredit helps consumption smoothing, and that a 1% increase in credit from Grameen Bank to a woman increases girls' school enrollment by 1.86%. Furthermore, it is found that, when controlling for village fixed effects and other observable characteristics, microcredit in Bangladesh successfully reaches the poor and significantly contributes to poverty reduction.

Morduch (1998), however, argues that the assumptions underlying Pitt and Khandker's estimation are erroneous and that a revised estimation remarkably changes the results. According to Morduch, land markets in Bangladesh are rather active and many sample clients purchase/sell their land. In addition, the eligibility criterion of less than 0.5 acres of landholding is often violated. Then, applying the simple DID method, he finds little evidence of positive impacts on clients except for consumption smoothing.

McKernan (2002), based on the estimation method similar to that of Pitt and Khandker (1998), examines the sensitivity of estimates to eligibility criteria and finds that the results regarding the effects of participation in BRAC and BRDB are not sensitive, but those for Grameen Bank are somewhat sensitive. In addition, McKernan finds that because high-profit households are more likely to participate with Grameen Bank, the failure to control for the selection mechanism overestimates the impact on business profits by more than 200 percentage points.

Against such counterevidence, Khandker (2005) uses panel data and conducts a robustness check. The results show that positive impacts on consumption remain in the revised specification. Moreover, the results even show an increase in impact: every additional 100 taka leads to increased consumption by 20.5 taka.

The positive impact of microcredit on consumption smoothing, which is agreed upon by both Pitt and Khandker (1998) and Morduch (1998), is challenged by Menon (2006b). She draws her sample from eight Grameen thanas (Grameen is the only program that operates in these thanas) and estimates the impacts of consumption smoothing by nonlinear least squares. The results show that, although microcredit indeed helps to improve the recipients' ability to smooth seasonal shocks, its effect declines over time and it has virtually no impact after four years of participation. Extending the data to other MFIs, Menon (2006a) also reaches the same conclusion. In contrast, Chemin (2008), using the same data set as Pitt and Khandker, demonstrates that even the average impact on consumption smoothing, measured by variation of log expenditure, disappears if selection is controlled for by propensity score matching.

Roodman and Morduch (2009) revisit the evidence presented by Pitt and Khandker (1998), Morduch (1998), and Khandker (2005). As a first step, they return to the original survey to reconstruct the data for their study. Summary statistics on the mean and standard deviation show that their reconstructed data match well with Pitt and Khandker (1998) and Morduch (1998), although not with Khandker (2005). As a next step, they adopt the same estimation methodologies to replicate the results and also conduct some specification tests to explore the validity of assumptions underlying these papers. Surprisingly, their replication exercises and specification tests show that the impact of all three papers is weak with their reconstructed data and the results are not particularly reliable: the impact on consumption is even negative and significant in comparison with Pitt and Khandker (1998); the impact on consumption volatility is weaker than that which Morduch finds; the use of panel data, as used by Khandker, does not necessarily yield a more credible estimate because it cannot compensate for the lack of clearly exogenous variation in the treatment variable. Having obtained such results, they do not claim that microcredit has little impact; rather, they insist the essential difficulty inherent in exploring the causal relationship between the provision of microcredit and the improvement of household welfare with nonexperimental approaches.

To date, to the best of our knowledge, only two experimental studies exist in the field of microfinance evaluation. Banerjee et al. (2009) conduct a randomized controlled trial in Indian slums, while Karlan and Zinman (2009a) conduct a similar exercise in the Philippines. Broadly, neither study reveals a significant impact of microcredit on poverty reduction. Banerjee et al. (2009) show, among other things, that households with an existing business increase durable expenditures and their profits, but, importantly, the average impacts on expenditure per capita as well as expenditures on education and health are not statistically significant. Meanwhile, Karlan and Zinman (2009a) show that there is no statistical difference in household income, the probability of being under the poverty line, and the quality of food between treatment and control groups. Moreover, subjective well-being slightly decreases for the treatment households.

In sum, although microcredit has increasingly gained popularity as an effective tool for poverty reduction, there is no solid evidence that supports its positive impact. Moreover, the above-mentioned studies largely find that a naïve estimate, which fails to eliminate the selection biases, tends to overestimate the impact of microcredit. This, in turn, implies that omitted variables, which affect outcomes and participation in the microcredit, are positively correlated and that it is the better-off who can benefit more from the existence of microcredit.

III. MECHANISM UNDERLYING HIGH REPAYMENT RATES

  1. Top of page
  2. Abstract
  3. I. INTRODUCTION
  4. II. WELFARE IMPACT OF MICROCREDIT
  5. III. MECHANISM UNDERLYING HIGH REPAYMENT RATES
  6. IV. CHALLENGES OF MICROFINANCE INSTITUTIONS
  7. V. SUMMARY AND CONCLUSIONS
  8. REFERENCES
  9. Appendix

What attracted attention to microcredit were its remarkably high repayment rates. Grameen Bank, a leading institution in microcredit, continuously recorded repayment rates of more than 95%. This was a big surprise to many, especially considering the bad repayment records of agricultural state banks that had provided subsidized loans to rural farmers. As reviewed in the previous section, recent empirical evidence shows that microcredit does not necessarily contribute to the significant increase in clients' income and business profits, so this means that the high repayment rates are not caused by the improved financiability of clients. Then, what enables microcredit to achieve such good repayment records? In this section, we explain why lending to the poor is difficult and how microcredit has tackled problems by using simple models. In particular, we analyze whether and how the elements of microcredit, such as group lending and dynamic incentives, solve the problems of asymmetric information in the credit market. Although a number of theoretical works focus on the role of group lending and dynamic incentives in order to alleviate the problems of asymmetric information, there are no published papers or books providing mathematical models unifying these theoretical works. The main contribution of this section is to present simple models sharing similar structure across various kinds of problems of asymmetric information and to argue the effect of group lending and dynamic incentives under the same theoretical models. This will help to provide the reader with a unified picture of the mechanisms of group lending and dynamic incentives.

Before describing the model, we will preview the main results:

  • • 
    Group lending alleviates the problem of adverse selection. It also alleviates moral hazard if the group can coordinate its members' decisions, but cannot otherwise. It achieves higher repayment rates if the returns are sufficiently high, but might reduce repayment rates if not.
  • • 
    Dynamic incentives are good for preventing moral hazard and strategic default. However, they do not affect the problem of adverse selection.
  • • 
    Introducing flexibility into dynamic incentives and group lending might contribute to the achievement of higher expected borrower welfare.

A. Difficulty in Lending to the Poor: No Collateral and Asymmetric Information

Usually poor people have: (i) low incomes, (ii) no stable income sources, and (iii) no collateral. (i) and (ii) are by themselves prudent reasons for banks not to lend to the poor. Here, however, we focus on the issue of (iii), no collateral. Without collateral, banks cannot collect the debt in case of default. However, lack of collateral signifies more than this. Combined with informational asymmetry, it causes adverse selection, moral hazard, and strategic default.

With sufficient value of collateral, the borrower's cost of default is the loss of the collateral, which is comparable to the loan amount, but without collateral, the borrower's monetary cost of default is zero, although there may be some non-monetary costs such as the bank's perpetual pressure and loss of reputation.

1. Common structure of the models

Borrowers, who are risk neutral, have investment projects that require one unit of capital, but they cannot finance the capital by themselves and have no collateral. The bank, which faces capital cost c for one unit of capital, offers them a loan with interest rate r. The borrower's investment project might fail and yield nothing. The bank cannot collect anything if the project fails because there is no collateral. We assume that the credit market is competitive, and in equilibrium, the bank's profit is zero.

2. Adverse selection

Suppose there are two types of borrowers, Type A (safe) and Type B (risky). A Type A (B) individual has a project that yields YA (YB) with a probability of pA (pB) but nothing otherwise, where pA > pB, YA < YB, and Type A's project yields a higher expected return, pAYA > pBYB. The ratio of Type A individuals is α and that of Type B is 1 −α. Lending to Type A is socially optimal but lending to Type B is not: pAYA > 1 +c > pBYB.

When the project fails, the borrower loses nothing because there is no collateral and obtains zero payoff. Therefore, the expected payoff for Type A from taking the loan is pA[YA− (1 +r)], whereas that for Type B is pB[YB− (1 +r)]. If the expected payoff is nonnegative, they will apply for the loans. The bank will earn pA(1 +r) − (1 +c) when it lends to Type A and pB(1 +r) − (1 +c) when it lends to Type B.

If the bank knows the types of loan applicants, it can offer different interest rates for Type A and Type B. Let rA and rB denote the interest rate for Type A and Type B, respectively. The bank's zero-profit conditions imply that inline image and inline image. rA < rB reflects the lower riskiness of Type A. The expected payoffs for types A and B are pAYA− (1 +c) and pBYB− (1 +c), respectively. By the assumption pAYA > 1 +c > pBYB, Type A will always take the loan and Type B will not apply for the loan.

However, if the bank does not know who is Type A and who is Type B, it has to set a uniform interest rate for every applicant. Suppose both Type A and Type B apply. Then, the expected profit for the bank is (αpA+ (1 −α)pB)(1 +r) − (1 +c). The zero-profit condition implies that

  • image(1)

Because YA < YB, when Type A applies, Type B also applies. Note that when Type A does not apply, Type B will not apply because if there is only Type B in the market, the interest rate will be inline image. Therefore, only when Type A has an incentive to apply, that is, YA≥ 1 +r*, will the lending occur. This condition reduces to

  • image(2)

If α, YA, pA, and/or pB is not sufficiently high, the lending will not be feasible and a credit market does not exist.6

3. Moral hazard

We now consider the borrower's decision on investment effort. If the borrower exerts effort, then the project return will be Y with a probability of pe and 0 otherwise. If the borrower shirks, then the return will be Y with a probability of ps(<pe) and 0 otherwise. Exerting effort incurs a cost of d. We assume that peYd > 1 +c > psY, implying that exerting effort is socially optimal. We denote the difference in the expected net return between exerting effort and shirking by η= (peps)Yd. If the bank can observe and verify whether the borrowers exert effort, the bank can punish those borrowers who shirks maintaining the repayment rate of pe. However, if this is not the case and punishment is not feasible, then the lending should be incentive compatible for the borrowers to exert effort. Because there is no collateral and the bank cannot observe the effort, the borrower's expected payoff from exerting effort is pe[Y− (1 +r)]−d, whereas that from shirking is ps[Y− (1 +r)]. The borrower has an incentive to exert effort only when pe[Y− (1 +r)]−dps[Y− (1 +r)]. By using η= (peps)Yd, this condition can be rewritten as

  • image(3)

The interest rate is determined by the bank's break-even condition. Note that we assume each borrower is “small” and that his or her choice does not affect the equilibrium interest rate determined by the break-even condition. As long as the incentive constraint (3) is satisfied, the break-even condition implies

  • image(4)

Substituting this into (3), we can derive the condition that borrowers exert effort as

  • image(5)

Unless the expected return from exerting effort is sufficiently higher than that from shirking, borrowers will not choose to exert effort.7

4. Strategic default

Third, consider the borrower's decision regarding whether to repay or not. Suppose the investment return is Y with probability of p and 0 otherwise. If the return is 0, there is no choice but not to repay. When the return is Y, the borrower can choose to repay or not to repay.

If the bank can always observe and verify the borrower's investment return, then the bank can enforce repayment. However, if the bank can neither observe nor verify the return, the bank cannot prevent borrowers from choosing not to repay. Here, we consider the intermediate case: the bank undertakes auditing with cost z > 0 and succeeds in observing and verifying the return with a probability of q. In this case, the bank can force a borrower to repay to a certain extent by using a contract specifying that the bank confiscates all the returns when the bank finds that he or she has sufficient income but chooses not to repay. The borrower will choose to repay if and only if Y− (1 +r) ≥ (1 −q)Y, where the left-hand side is the payoff from choosing to repay and the right-hand side is the expected payoff from choosing not to repay (payoff of 0 if the bank detects the borrower's strategic default and payoff of Y if not). Therefore, the borrower will choose not to repay unless the return is sufficiently high such that

  • image(6)

The interest rate is determined by the bank's break-even condition. We assume that the bank prefers that borrowers do not commit strategic default over letting borrowers default and confiscate Y with probability pq (the probability that investment is successful and monitoring is successful).8 Given the satisfaction of the incentive constraint (6), the equilibrium interest rate is

  • image(7)

Substituting this into (6), we can derive the condition that the borrowers choose to repay as

  • image(8)

Given the capital cost c and auditing cost z, borrowers are more likely to choose to repay with higher Y, p, and q.

In sum, imperfect information combined with no collateral: (i) makes credit cheaper for riskier individuals, which raises the credit cost for safer individuals, and if the interest rate is too high for safer individuals, the credit market ceases to exist (adverse selection); (ii) makes exerting effort less attractive (moral hazard); and (iii) gives the borrowers incentives not to repay when the return (Y) or monitoring technology (q) is not sufficiently high.

B. Group Lending with Joint Liability

One innovative mechanism that most microcredit institutions employ is group lending with joint liability. In the typical group lending scheme: (a) each member is jointly liable for each other's loan, (b) if any members do not repay, all the members are punished (often in the form of denial of future credit access), and (c) prospective borrowers are required to form groups by themselves. In the following subsection, we briefly examine how group lending with joint liability affects the problems of adverse selection, moral hazard, and strategic default. The original theoretical models can be found in Ghatak (1999) and Van Tassel (1999) for adverse selection, Stiglitz (1990) and Che (2002) for moral hazard, and Besley and Coate (1995) for strategic default. How the denial of future credit access combined with group lending works will be discussed in subsection C.

Here, we suppose for simplicity that a group consists of two individuals and there are no correlations between their investment returns. They have full information on their partners' investment returns.

1. Group lending and adverse selection

First, note that because a member has to repay for his or her partner when he or she fails in the project, every individual wants to form a group with safe borrowers; namely, Type A individuals. Type B individuals, however, cannot form a group with Type A individuals because the latter will not assent to it. Therefore, in the equilibrium, Type A is tied with Type A, and Type B matches with Type B (assortative matching).9

The payoff for Type i (i=A, B) to take the loan is thus expressed as

  • image(9)

because their partners will fail with a probability of 1 −pi and they have to repay for their partners as much as they can.10 Type A will apply for the loan when the value of (9) is nonnegative. This condition reduces to

  • image(10)
  • image(11)

The analogous condition for Type B is

  • image(12)
  • image(13)

Because YA < YB, there are three possible cases: (i) YB > YA≥ 2(1 +r) (ii) YB≥ 2(1 +r) > YA, and (iii) 2(1 +r) > YB > YA.

First, let us consider case (i). Because YB > YA≥ 2(1 +r), conditions (10) and (12) are automatically satisfied. Because the expected profit for the bank from lending two units of capital to a group of two member is inline image, the break-even condition implies the equilibrium interest rate of

  • image(14)

Note that this is lower than the equilibrium interest rate under individual lending (1). This is because the bank can collect money from a successful borrower when his or her partner fails in investment. Furthermore, because a successful borrower has to repay his or her partner's share with probability 1 −pi, the borrower's expected repayment conditional on his or her investment success is Ri≡ 1 +r*G+ (1 −pi)(1 +r*G) = (2 −pi)(1 +r*G), i=A, B. Without group lending, the expected repayment conditional on the borrower's investment success is inline image. Therefore, we can consider RA and RB as effective interest rates for Type A and Type B, respectively. By using (14), we can obtain the equilibrium effective interest rates as

  • image
  • image

Group lending succeeds in differentiating the effective interest rates between safer borrowers and riskier borrowers, making it smaller for safer borrowers and larger for riskier borrowers compared to the case without group lending. Accordingly, A's expected payoff under group lending, inline image, is larger than that under no group lending, inline image. Because we are analyzing the case YB > YA≥ 2(1 +r), this effective interest rate differentiation does not affect borrower A's participation decision because with (14), the right-hand side of the condition YA≥ 2(1 +r) becomes inline image, being larger than that of (2), which represents borrower A's participation constraint without group lending. However, it can affect borrower A's participation decision if YA is not so high, YA < 2(1 +r).

Now consider case (ii), YB≥ 2(1 +r) > YA. As in case (i), Type B's participation constraint (12) is automatically satisfied. Type A's participation constraint is given by (11). From the break-even condition inline image,11 the equilibrium interest rate is

  • image(15)

Therefore, A's participation constraint can be written as

  • image(16)

The right-hand side is smaller than that of (2), implying that group lending relaxes A's participation constraint and makes lending feasible with a wider range of parameter c.

Appendix A shows that in case (iii), 2(1 +r) > YB > YA, we can obtain a similar result where group lending relaxes A's participation constraint, making lending feasible with a wider range of parameter c. In all, group lending can alleviate the problem of adverse selection and make lending more likely to be feasible through differentiation of the effective interest rates.

2. Group lending and moral hazard

Assume that the members of the group act to maximize the total payoff of the two borrowers and that anyone who deviates will be punished with serious social sanctions.

Consider the condition that the group prefers to exert effort. The expected payoff from exerting effort is

  • image

whereas that from shirking is

  • image

We focus on the case Y≥ 2(1 +r). Appendix B deals with the case 2(1 +r) > Y and shows that the qualitative results are similar.

The group chooses to exert effort only when the payoff from exerting effort is higher than that from shirking. By using η= (peps)Yd, this condition can be written as

  • image(17)

The interest rate determined by the bank's break-even condition given satisfaction of the incentive constraint (17) is

  • image(18)

which is lower than the equilibrium interest rate under no group lending (4). Substituting this into (17), the condition that borrowers exert efforts becomes

  • image(19)

Because inline image, the right-hand side becomes smaller than (5), implying that group lending can induce borrowers to exert effort with smaller η.

The analysis above crucially depends on the assumption that the group can coordinate its members' effort choices. Appendix C shows that if the group cannot coordinate its members' effort choices at all and each member pursues his or her own interest, group lending will not induce efforts with smaller η and might induce more shirking relative to the case without joint liability. Whether group lending can improve repayment rates or not crucially depends on whether the group can coordinate its members' decision. If it is difficult for group members to monitor, group lending cannot alleviate the problem of moral hazard. Although Banerjee, Besley, and Guinnane (1994) point out that joint liability can increase peer monitoring and induce members to choose safer investment, the results crucially depend on the cost of monitoring effort and monitoring precision. Any empirical results that do not refer to these underlying factors might have less external validity.

3. Group lending and strategic default

Under group lending, if a borrower (whose investment return is Y) repays but another borrower does not, then the former needs to decide whether to shoulder his or her partner's repayment. If the borrower decides to shoulder, then he or she will obtain Y− 2(1 +r) and the game will end. However, if the borrower decides not to shoulder, then both of the two borrowers will be audited by the bank and the investment returns of each borrower (if any) will be confiscated with probability q. We assume that the probability of audit success is independent across the two borrowers. Note that even if the bank can find that a defaulting partner actually has no investment return (nonstrategic default), it requires his or her partner to shoulder for the defaulting partner, and if he or she does not shoulder, he or she will be audited and his or her investment return (if any) will be confiscated with probability q.

We assume that the members of the group act to maximize the total payoff of the two borrowers and anyone who deviates will be punished with serious social sanctions as in the previous sub-subsection. For the analysis, we should consider two cases: (i) both members have income Y, and (ii) only one of them has income Y.

In case (i), the group will choose to repay if and only if 2Y− 2(1 +r) ≥ (1 −q)22Y+ 2q(1 −q)Y, where the right-hand side is the expected payoff of choosing not to repay: the group can retain 2Y when the bank cannot observe or verify the returns of both members, which occurs with a probability of (1 −q)2, and can retain Y when the bank observes and verifies the return of only one of two members, which occurs with a probability of 2q(1 −q). This condition reduces to

  • image(20)

which is identical to (6), implying that, given r, group lending will have no impact on the borrowers' incentive to repay when both members succeed in their business.

In case (ii), the group will choose to repay if and only if Y− 2(1 +r) ≥ (1 −q)Y, or,

  • image(21)

The right-hand side is larger than that of (20). If Y is sufficiently high, group lending can induce a borrower to repay for his or her defaulting partner.

Now, we consider the interest rate determined by the bank's break-even condition. When (21) is satisfied, the bank's break-even condition is [p2+ 2p(1 −p)]2(1 +r) − (1 −p)22z= 2(1 +c).12 Therefore, the equilibrium interest rate is

  • image(22)

which is lower than the equilibrium interest rate under no group lending (7). Substituting this into (21), we derive that the bank can ensure the group repays both in case (i) and case (ii) if

  • image(23)

However, if (20) is satisfied but (21) is not satisfied, the group will repay only in case (i). In this situation, the bank's break-even condition (for two members) is 2p2(1 +r) + 2p(1 −p)qY− 2(1 −p2)z= 2(1 +c).13 Thus, the equilibrium interest rate is

  • image(24)

Therefore, the bank can ensure that the group repays in case (i) if

  • image(25)

whose right-hand side is larger than (8), implying that there exists a range of Y where individual lending can force borrowers to repay but group lending cannot.

In sum, group lending can induce repayment both in case (i) and case (ii) and record a repayment rate of p2+ 2p(1 −p) =p(2 −p) if inline image and can induce repayment only in case (i) and achieve a repayment rate of p2 if inline image. Under individual lending, a borrower will repay if inline image and achieve a repayment rate of p. Figure 1 depicts the relationship between Y and repayment rate under individual lending and group lending. Group lending can achieve a higher repayment rate than individual lending when the investment returns are sufficiently high, that is, inline image, but will suffer a lower repayment rateotherwise. This is because when the returns are not sufficiently high, the succeeding borrowers have no incentive to shoulder the loan repayment for their partners and they will choose not to repay their own loan as well.

image

Figure 1. Repayment Rate in the Model of Strategic Default

Download figure to PowerPoint

Appendix D shows that if the group cannot coordinate its members' repayment decisions, the repayment rate can be lower when inline image. The group's ability to coordinate its members' decisions is important when the investment returns are not sufficiently high. The existence of severe social sanctions and information sharing among the members is crucial for the group to have coordination ability. With perfect information sharing among the members, Ray and Sjostrom (2004) show that incorporating cross reporting into joint liability can achieve higher repayment rates and generates higher expected borrower welfare without social sanctions. However, such cross-reporting is not widely used in reality (Bhole and Ogden 2010).

C. Beyond the Myth of Joint Liability: Dynamic Incentives and Frequent Installments

In the previous subsection, we show that joint liability will alleviate the problem of adverse selection, but its effects on moral hazard are crucially dependent on the coordination ability of the group and it might worsen the problem of strategic default if borrowers' investment return Y is not particularly high.

Recently, other features of microcredit such as dynamic incentives and frequent installments have attracted the attention of researchers and practitioners. Dynamic incentives mean that if borrowers repay the loans, they can access future loans (often with larger amounts). Because the amount of microcredit loans are “micro,” they have substantial needs for further loans to develop their business. The considerable needs for future loans can be one of the factors that distinguishes microcredit from traditional bank loans. However, most MFIs require their borrowers to repay weekly. Some other MFIs mix weekly and monthly installments depending on the frequency of income flow of the borrowers. This also gives microcredit a distinction.

In the following, we first analyze the effect of dynamic incentives. Dynamic incentives consist of two factors: (i) contingent renewal (future loans are accessible upon repayment of the current loan) and (ii) larger loan amounts of future loans. For simplicity and to use the same framework as in the previous sections, we focus on the contingent renewal and assume that the loan amounts are invariant. After that, we discuss the flexible form of dynamic incentives and joint liability, and the other form of dynamic incentives, sequential lending. We then look at the function of frequent installment.

1. Dynamic incentives and adverse selection

First, we show that dynamic incentive has little impact on adverse selection. For simplicity, assume that investment projects financed by microcredit yield a return YA(YB) with a probability of pA(pB) only in one period. If a borrower or group of borrowers succeed in the investment and repay the loan, they can access the same amount of credit in the next period. We also assume that saving is impossible and borrowers cannot use the earnings in the previous periods to repay the current loans.

Under individual lending, the expected present value of accessing microcredit is inline image for Type A and inline image for Type B, where δ (< 1) denotes a common discount factor. These can be rewritten as

  • image
  • image

Borrowers will participate in the credit if inline image is nonnegative. Note that the condition that inline image is nonnegative is Yi≥ (1 +r), which is exactly the same as the case without dynamic incentive. Thus, introducing dynamic incentives into individual lending scheme cannot alleviate the problem of adverse selection.

Next, consider the case of group lending. There are three possible cases: (i) YB > YA≥ 2(1 +r), (ii) YB≥ 2(1 +r) > YA, and (iii) 2(1 +r) > YB > YA. The expected present value of accessing microcredit for type i=A, B in case (i) is inline image. Thus,

  • image(26)
  • image(27)

A group of Type i borrowers will participate in the credit if inline image is nonnegative. The condition for (26) and (27) being nonnegative is exactly the condition (10) and (12), respectively.

When Yi < 2(1 +r), the expected present value of accessing microcredit for Type i is inline image because if one of the group members fails, another member has to repay as much as they can on behalf of the defaulting member and the group cannot access further loans. inline image can be rewritten as inline image. The condition for inline image and inline image being nonnegative is exactly the condition (11) and (13), respectively.

Therefore, the introduction of dynamic incentives does not change the participation decision of either type in any of the cases of (i) (ii), and (iii). This result is robust to the introduction of a different discount rate between Type A and Type B, or the introduction of effort cost or running cost. The point is that participation decision depends only on whether the borrower's expected payoff is positive or not and she faces the same problem in each period. Thus, she has an incentive to participate with dynamic incentives if and only if she has an incentive to participate without dynamic incentives. One way that dynamic incentives can alleviate the problem of adverse selection is through asset accumulation. Here, we have assumed that saving is impossible. However, if saving is possible and MFIs can take the accumulating assets as collateral, it is possible that in some future periods MFIs can have a sufficient amount of collateral to make the expected payoff from the current loan negative for Type B individuals and positive for Type A individuals. But this requires the discount factor to be high and pA and pB not to be small so that borrowers place considerable weight on the value of future loans.

2. Dynamic incentives and moral hazard

With dynamic incentive, the expected present values of exerting effort and shirking under individual lending are inline image and inline image, respectively. These can be rewritten as

  • image(28)
  • image(29)

The borrower will exert effort when inline image. Note that the values in brackets in (28) and (29) are exactly the expected values from exerting effort and shirking under individual lending without dynamic incentives, respectively. Because inline image, dynamic incentives make the incentive constraint less strict.

For joint liability, we only consider the case where Y≥ 2(1 +r) and the group can force its members to act to maximize the total payoff of the two borrowers. The expected present values of exerting effort and shirking are inline image and inline image, respectively. These can be rewritten as

  • image(30)
  • image(31)

Again, the values in the brackets in (30) and (31) are exactly the expected values from exerting effort and shirking under joint liability without dynamic incentives, respectively. Because inline image,14 dynamic incentives make the incentive constraint less strict under group lending as well. In addition, Appendix E shows that when the group can coordinate its members' effort choices, joint liability with dynamic incentives performs better than individual lending with dynamic incentives in terms of preventing moral hazard.

If we consider the case where the group cannot coordinate its members' effort choice, then we need to allow for the repeated interaction between the borrowers: a borrower may able to punish his or her partner who chooses to shirk in the form of always choosing to shirk afterwards and to reward his or her partner who chooses to exert effort in the form of always choosing to exert effort afterwards. Itshould be explored whether and in what conditions joint liability with dynamic incentives without group coordination is desirable in reducing moral hazard relative to individual lending with dynamic incentives. To our knowledge, there is no theoretical study dealing with this issue.15 The experimental study by Gine et al. (2006) finds that joint liability with dynamic incentives increases moral hazard (in their case, moral hazard means investing in a risky project). However, they do not vary the value of η and, in fact, they set η to be negative (i.e., the expected return of risky investment is higher than that of safe investment), which is at odds with the framework of moral hazard because choosing risky investments is socially optimal. More theoretical and experimental research is needed to evaluate the performance of individual lending and group lending with dynamic incentives.

3. Dynamic incentives and strategic default

By introducing dynamic incentives, the problem of strategic default can be alleviated. First, we will consider individual lending. With dynamic incentives, the expected payoff from repaying the loan is Y− (1 +r) +δVIL, where VIL is the expected value of future credit access under individual lending. If the borrower chooses not to repay, his or her expected payoff is (1 −q)Y. Thus, the borrower will repay if and only if

  • image(32)

Because the probability of obtaining Y is p, VIL can be expressed as VIL= max{p[Y− (1 +r) +δVIL], p(1 −q)Y}. As long as (32) is satisfied, p[Y− (1 +r) +δVIL]≥p(1 −q)Y and, therefore, inline image. By substituting this and the equilibrium interest rate (7) into (32), we obtain

  • image(33)

Compared with (8), it is obvious that dynamic incentives reduce strategic default.

Next, we will consider group lending. As in the previous section on moral hazard, we only analyze the case where the group can force its members to act to maximize the total payoff of the two borrowers. Analysis of the case where the group cannot force its members to act in such a way can be found in Kono (2010).

We need to consider the following two cases: (a) the group will repay the loans when at least one borrower gains income Y, and (b) the group will repay the loansonly when both borrowers gain Y. Denote the expected value of future loans to the group in case (a) by inline image, and in case (b) by inline image, which are expressed as inline image and inline image, respectively. These reduce to

  • image(34)
  • image(35)

First consider case (a). The group will repay the loans when only one borrower has Y if inline image. It is easy to show that if this condition is met, it will repay the loan when both borrowers have Y. With (34) and the equilibrium interest rate (22), this condition reduces to

  • image(36)

whose right-hand side value is smaller than that of the corresponding condition without dynamic incentives (23).

In contrast, in case (b), the group will repay the loans when both borrowers have Y if inline image. With (35) and the equilibrium interest rate (24), this condition reduces to

  • image(37)

whose right-hand side value is again smaller than that of (25). Therefore, dynamic incentives alleviate the problem of strategic default under joint liability as well.

In sum, under individual lending with dynamic incentives, the aggregate repayment rate is p if inline image and 0 otherwise. In contrast, under joint liability with dynamic incentives, the aggregate repayment rate is

  • image
  • image
  • image

Because inline image, the relationship between Y and the repayment rates under individual lending and joint liability is analogous to Figure 1. Unless Y is sufficiently high, individual lending works better than group lending even with dynamic incentives.

4. Flexible dynamic incentives, flexible group lending, and sequential lending

We have assumed above that dynamic incentives take the form of denial of any future loans for defaulting borrowers. However, given stochasticity of investment return, this scheme also punishes borrowers who are willing to repay but cannot afford to repay, so the complete denial of future loans might not be desirable. With this view, Alexander-Tedeschi (2006) shows that the optimal length of loan denial need not be a lifetime one to solve the problem of strategic default. For simplicity, here we focus on the problem of strategic default under individual lending with dynamic incentives and investigate the optimal probability of allowing a defaulting borrower to access future loans, ρ, instead of investigating the optimal length of loan denial. We also assume q= 0, implying that borrowers will not repay in the absence of dynamic incentives and that the bank will not do auditing.

Let inline image be the expected future payoff when the borrower repays in the current period and inline image be the expected future payoff when he or she does not repay in the current period. The expected payoff from repaying the loan can be written as inline image and that from not repaying as inline image. Because the borrower has no income and cannot afford to repay with probability 1 −p, inline image can be written as inline image. However, inline image can be written as inline image because he or she will have access to the loan with probability of ρ and he or she will choose the action which generates higher expected payoff, repaying (with payoff inline image) or not repaying (with payoff inline image). When inline image,16inline image and, therefore, inline image can be written as inline image, or

  • image

The condition that he or she will repay, inline image, can be rewritten as δ(1 −ρ)pY≥ (1 −δρ)(1 +r), or

  • image(38)

Given the satisfaction of the incentive constraint (38), the equilibrium interest rate is inline image.17 With this (38) implies

  • image(39)

Because inline image, as long as inline image, we can have ρ? (0, 1) and therefore, positive probability for defaulting borrowers to access the loans. Complete denial of future credit access is not necessary for inducing borrowers to repay. Because social welfare increases in ρ, the probability of allowing the defaulting borrower to access the loans, the optimal ρ, which depends on δ, is

  • image

As δ approaches 1, ρ* approaches 1, implying that when borrowers are quite patient, the bank can allow a defaulting borrower (either due to strategic default or unfortunate low return) to access future loans with high probability without causing strategic default. However, if borrowers are not sufficiently patient and inline image becomes negative, then no nonnegative ρ* exists, i.e., even complete denial of future loans is not enough to prevent strategic default.

The above analysis shows that MFIs can introduce some flexibility into loan denial without harming repayment rates dependent on the underlying economic conditions. Bhole and Ogden (2010) further incorporate flexibility into joint liability. They allow that total group repayment does not need to be 2(1 +r) if a borrower shoulders his or her defaulting partner. They also allow the probability of loan access denial to be different among: (i) borrowers who themselves repay and whose partners also repay; (ii) borrowers who shoulder a defaulting partner with total repayment less than 2(1 +r); (iii) borrowers shouldered by their partner, and (iv) borrowers who themselves do not repay and whose partners also do not repay. They show that this flexible group lending can prevent borrowers from committing strategic default under a wider range of parameters and also generates higher expected borrower welfare than individual lending. However, they restrict their analysis to history-independent strategies where borrowers do not punish their partner according to their past behavior. More theoretical analysis on flexible joint liability and individual lending with mutual insurance is warranted.

Another form of dynamic incentives is sequential lending, where only some of the group members receive the loans at first and the remaining members must wait until those members who received the loans first complete the repayment or keep a good record of repayment. Chowdhury (2007) shows that with joint liability and social sanctions, sequential lending is efficient in inducing borrowers to repay the loan. The traditional Grameen group lending takes the form of sequential lending: loans are initially provided to two members of the group and if they keep good repayment records, the remaining three members can obtain loans. Although sequential lending might be a good way to exploit social sanctions in order to induce borrowers to repay, in reality it is not adopted by many MFIs. One reason is that borrowers are not willing to wait for long time and prefer other MFIs' loans that do not require sequential lending. In particular, good borrowers who can easily obtain loans from MFIs and banks are less likely to participate in sequential lending. The types of borrowers that each lending strategy attracts should also be explored in future studies.

5. Frequent installments

Frequent installment is a very common feature of microcredit schemes. Most schemes utilize weekly installments, although some employ monthly installment depending on the borrower's frequency of income flow. Armendáriz de Aghion and Morduch (2005) argue that the merits of frequent installment are: (i) to allow credit officers to know their clients' problems at an early stage, which enables them to take immediate action; (ii) to select borrowers who have a steady income flow other than the investment project funded by the loans (and, therefore, are less likely to face repayment problems) by requiring weekly installments; and (iii) to provide borrowers a means to solve “savings constraints.”“Savings constraints” mean that people face difficulty in saving money due to self-control problems or to pressures from neighbors and relatives for handouts or from their spouse for spending on drinking and smoking.19 A repayment schedule with frequent installments takes the money out of the house soon after it is earned, making it less likely for the earned money to be used up by themselves, their spouse, relatives, or neighbors.

The problem of self-control is often incorporated into economic theory in the form of quasi-hyperbolic preference in which the present value of a flow of future utilities (us)s≥t as of time t is expressed as inline image, where β is the parameter of present-bias preference with 0 < β < 1. Basu (2009) shows that those who have quasi-hyperbolic preference have more demand for loans because they overvalue the increase in current consumption brought by the loans and they also value the regularly scheduled repayments enforced by the bank as a commitment device. This suggests that MFIs can make their credit more attractive by incorporating frequent installments.

That microcredit borrowers really do face savings constraints is suggested by a number of cases. One example is house repair. Many microcredit borrowers use the credit to repair their houses. Even though the credit does not help to raise income, usually such borrowers keep good repayment records. Given the fact that they repay the loans within a year, they could have saved the money within the year instead of borrowing it at interest. In a field survey in India, we interviewed a number of households who invested in house repair and most answered that being forced to repay is important. Because not repaying loans entails larger costs than not saving (such as pressure from field officers and group members, denial of future credit, and bad reputation), borrowing money from MFIs that implement frequent installment can work as a commitment device for saving more than would be saved by themselves. Thus, the interest paid can be regarded as the cost of being unable to save by themselves.

D. Empirical Evidence

We have provided theoretical explanations of the high repayment rates of microcredit focusing on joint liability, dynamic incentives, and frequent installments. In this section, we review the results of empirical studies on these issues.

Although joint liability has long attracted the attention of practitioners and researchers, there has been no reliable empirical investigation of the effectiveness of joint liability. As we argued in Section II, to generate a consistent estimator of the impact, we need a good counter-factual. Merely comparing the repayment rates of MFIs that employ joint liability and those that do not will not generate a consistent estimator because these MFIs and their clients will differ in many aspects, such as regional difference, differences in application criteria, differences in target clients, and differences in organizational ability.

Gine and Yang (2009) tackle this problem in a randomized experiment. They remove joint liability from existing group-screened lending groups selected in a random manner. Because the clients were in the same group lending scheme before the experiment, the detected difference in repayment rates between the group with group liability and the group without it can be solely attributed to the effect of joint liability. Note that although adverse selection occurs at the participation decision, moral hazard and strategic default occur after the participation. Hence, although the difference in repayment rates among the new clients (those who participated after the experiment) reflects the aggregate effect of adverse selection, moral hazard and strategic default, the difference in repayment rates among the old clients (those who had participated before the experiment) captures only the combined effect of moral hazard and strategic default. Surprisingly, their findings do not support the effectiveness of group lending in reducing adverse selection or a combined effect of moral hazard and strategic default: there are no significant differences in default rates among either the old or new clients one year and three years after this experiment. They also find a larger number of new participants under individual lending than under the group liability scheme. They argue that this is because group lending imposes punishment on the other members in the case of someone's default and it discourages existing clients from inviting their relatives and close friends to participate in fear that if they fail in their business, their relatives and close friends who they themselves have invited must shoulder the loan repayment for them.

Although the experiment of Gine and Yang (2009) succeeds in generating reliable estimates of the effectiveness of joint liability on adverse selection and the combined effect of moral hazard and strategic default, it is still not clear how joint liability actually affects moral hazard and strategic default separately. Gine et al. (2006) and Kono (2010) try to answer exactly this question. They use a lab-type field experiment to identify the effect of group lending on moral hazard (risky investment) and strategic default, respectively. Gine et al. (2006) find that group lending itself will induce borrowers to choose risky investment but improve repayment rates through risk sharing, and the introduction of self-group formation is important to make group lending function well. Kono (2010) focuses on the problem of strategic default and finds that group lending itself induces the free-rider problem (letting other members shoulder his or her own repayment). He also points out the possibility of risk sharing even under individual lending. In the real world, people often cope with income shock by sharing risk with their relatives and neighbors (Townsend 1994; Grimard 1997). Ghatak and Guinnane (1999) point out that if the group maximizes joint welfare, then members will always share net incomes and be voluntarily jointly liable for each other's loans regardless of whether the formal terms are those of joint or individual liability. However, in the theoretical prediction of Besley and Coate (1995) or experimental comparison of Gine et al. (2006) and Abbink, Irlenbusch, and Renner (2006), the possibility of risk sharing under individual lending is totally ignored. Kono (2010) allows voluntary transfers among group members under individual lending and finds that under individual lending borrowers actually share income risk (but less often than under joint liability) and individual lending achieves higher repayment rates than group lending. He also finds that introducing cross-reporting or social sanctions is important for keeping repayment rates high under group lending, but even with them group lending cannot significantly outperform individual lending.

Some other studies focus on the role of social capital or social connections to keep the repayment rates high under joint liability. Theoretically, as mentioned earlier, the effectiveness of group lending may depend on the social connections among group members and although Gine and Yang (2009) find that joint liability does not contribute to higher repayment rates on average, it is quite possible that group lending works well among borrowers with good social connections. The lab-type field experiments of strategic default games under joint liability in several developing countries reveal that social capital is an important determinant of repayment decision (Cassar, Crowley, and Wydick 2007; Cassar and Wydick 2008). Karlan (2007) shows that individuals with stronger social connections to their fellow group members (i.e., either living closer or being of a similar culture) have higher repayment rates in the microcredit program run by FINCA-Peru. However, we should note that when they have good social connections, it is also likely that they might voluntarily share income risks without joint liability. Whether incorporating joint liability into programs operated in socially well-connected areas is a better option or not should be clarified in future research.

Kurosaki and Khan (2009) use the installment level data and economic theory to show that the problem of strategic default did exist under group lending in Pakistan. They provide the theoretical prediction that under joint liability, the repayment decision of a individual borrower is correlated with that of other members and find that the repayment delay of each installment is correlated with other members' repayment delay even after controlling the individual fixed effects, beyond the level explained by possible correlation of project failures due to locally covariate shocks.20 Given the fact that after the implementation of a new system with strict enforcement of punishment against repayment problems, almost zero default rates were achieved, their study suggests the importance of program design for alleviating the problem of information asymmetry and achieving high repayment rates.

On dynamic incentives, the only available study based on randomized control trials to date is the one conducted by Karlan and Zinman (2009b) on consumer credit in South Africa. They randomly sent mail offers to the consumer credit company's previous clients with and without dynamic incentives (they can access loans with lower interest rates if they complete the repayment of the loan offered). They show that borrowers who are attached to dynamic incentives keep significantly better repayment records. A lab-type experiment carried out by Gine et al. (2006) also reveals that a dynamic incentive is very powerful in reducing moral hazard (risky investment).

As for the role of frequent installments, a study in rural India by Bauer, Chytilova, and Morduch (2008) shows that those who have hyperbolic discounting factors are actually more likely to participate in microcredit, which is consistent with the theoretical prediction in the previous subsection. Although this result suggests the importance of frequent installments for borrowers to solve savings constraints, it is still not clear how frequent the installments should be. Although high-frequency installment which closely follows the cash flowing into the households will provide a desirable commitment device for them, it will increase the transaction costs borne by borrowers and lenders. To investigate whether requiring less frequent installments makes it difficult for borrowers to cope with savings constraints and eventually leads to a decrease in the repayment rate, Field and Pande (2008) randomly assign monthly installments to a subset of the group who applied for the weekly installment scheme. They report no significant differences in repayment rates or repayment delay between the group subject to weekly installments and the group subject to monthly installments. This result suggests the possibility of substantially reducing transaction costs without affecting repayment rates through making the frequency of the installments monthly instead of weekly. However, as holds true of almost all other empirical works, it should be carefully considered before implementing this policy how well this result is applicable to other regions, which differ in many economic and social situations.

IV. CHALLENGES OF MICROFINANCE INSTITUTIONS

  1. Top of page
  2. Abstract
  3. I. INTRODUCTION
  4. II. WELFARE IMPACT OF MICROCREDIT
  5. III. MECHANISM UNDERLYING HIGH REPAYMENT RATES
  6. IV. CHALLENGES OF MICROFINANCE INSTITUTIONS
  7. V. SUMMARY AND CONCLUSIONS
  8. REFERENCES
  9. Appendix

We have discussed the mechanisms of microcredit that might contribute to high repayment rates. However, due to the expansion of MFIs and the expansion of microfinance services, these organizations now face new challenges. In this section, we focus on the rising competition among MFIs and the insurance services for the poor, microinsurance, because the theoretical framework provided in the previous section is useful for considering these two issues.

A. Competition among Microfinance Institutions

It is generally thought that an increase in MFIs will expand financial service opportunities for the poor, thereby contributing to the improvement of their welfare. Yet, if there are too many MFIs competing with each other in the same region, borrowers who did not complete repayments to a certain MFI might be able to obtain loans from other MFIs. In the framework of the previous section, this leads to an unintentional increase in ρ, the probability that defaulting borrowers can obtain loans in the next period. If ρ satisfies (39), then the lending is still feasible. Indeed, if MFIs stick to complete loan denial for defaulting borrowers based on some irrational reasoning, then the rising competition can make ρ closer to the optimal level. However, if ρ becomes too large to satisfy (39), the rising competition leads to lower repayment rates. To avoid such an undesirable consequence, it will be necessary to establish a unified credit information system where the credit history of each client is collected from all MFIs and shared among them in order to restrict lending to borrowers who have defaulted or who have taken multiple loans (McIntosh, de Janvry, and Sadoulet 2005).

It is also possible that borrowers of a certain MFI can borrow additional loans from other MFIs in order to repay the loans or to finance everyday needs. We have argued the role of frequent installment as a commitment device to solve saving constraints. If they can obtain loans whenever they want, those who face pressure from their spouse or relatives or face the problem of self-control might decide to repay their weekly installment by borrowing money. In this way, the existence of competing MFIs who generously offer loans will actually render the effectiveness of frequent installment as a commitment device ineffective. The final result will be that some borrow an unrepayable amount of money from multiple sources and enter the multiple debt trap.

The rising competition among MFIs will also change their portfolios. Facing such competition (McIntosh and Wydick 2005), MFIs will find their profits reduced and will need to improve their portfolios. Note that clients who generate more profits are usually the nonpoor, because they are less likely to have repayment problems and they often apply for larger loans than the poor, which will decrease the transaction costs per dollar. Thus, the rising competition will shift the target of MFIs away from the poor and the poor's access to financial services will end up declining.

Bond and Rai (2009) argue the possibility of “borrower runs.” Because the effectiveness of dynamic incentives depends on the expectation of future loan availability, if the bank suffers severe losses due to the competition and borrowers expect that the bank may cease to lend, then they have less incentive to repay. If the repayment rate or the market condition for the bank worsens beyond a certain level, then borrower runs occur: borrowers choose not to repay because they expect other borrowers not to repay and the bank then withdraws. Therefore, any increase in default rates or worsening of the bank balance sheet caused by the competition might increase the likelihood of borrower runs.

B. Microinsurance

Microinsurance is insurance for the poor. It constrains its premium at low levels so that the poor can afford to pay it by limiting the coverage of the insurance. Along with savings and transfers, many large MFIs are considering offering microinsurance services to the poor.

Although poor people insure themselves from income risk to some extent through informal insurance among community members, their consumption is still significantly affected by fluctuations in income (Townsend 1994; Grimard 1997). In addition, this insurance strategy does not function against regional income shocks such as famine and flood, because all the community members suffer from the shocks. Credit can help people protect themselves from negative shocks, but some shocks such as serious disease or the death of cattle are persistent and can reduce income in subsequent years. Large medical expenses and the death of income earners followed by the selling off of productive assets are often the causes of falling into poverty. Given the importance of insuring the poor from negative shocks, some MFIs provide insurance for the poor at low costs.21

To date, however, microinsurance has suffered from low uptake rates and, in some cases, MFIs have ceased to provide microinsurance due to huge losses incurred by the schemes. The problem with microinsurance is the lack of well-designed contract schemes. Although microcredit successfully tackles the problem of asymmetric information through group lending and dynamic incentives, most microinsurance schemes have no effective measures to solve the asymmetric information problems inherent in the insurance market; namely, adverse selection and moral hazard. Below we discuss these problems in health insurance as an example; most of the argument is applicable to life insurance, cattle insurance, and crop insurance.

1. Adverse selection

Given the premium and insurance coverage offered, unhealthy people (who are more likely to use insurance) are more willing to purchase insurance. This leads to high claim rates, which then pushes the insurance premium so high that healthier people will not purchase the insurance. At equilibrium, only unhealthy people purchase insurance and healthy people are not covered.

Theoretically, offering a well-designed menu of insurance policies can alleviate this problem. For example, offerings include a low premium policy but with low coverage and a high premium policy with high coverage, it is possible that the healthy purchase the former and the unhealthy purchase the latter with the result that both the healthy and unhealthy are covered by the insurance. This is what private insurance companies do in developed countries. However, currently almost all of the microinsurance companies offer just one policy, and as such microinsurance fails to prevent the adverse selection problem.

What makes the situation worse is that almost all of the micro health insurance schemes do not exclude pre-existing illness. If pre-existing illness is covered and people can buy insurance whenever they want, the optimal strategy for them is not to purchase the insurance until they get illnesses that are covered by it. Furthermore, it is often the case that NGO staff rather than insurance company workers sell the insurance. When they find poor households whose members need surgery, they proudly explain that if they buy the insurance, they can have the surgery for free. Sometimes they try to find sick people in order to sell the insurance.

In a study of the participation decision in micro health insurance in Karnataka, India, Ito and Kono (2010) show evidence of adverse selection: households with a higher ratio of sick members were more likely to purchase the insurance.

2. Moral hazard

In the health insurance market, there are two kinds of moral hazard: ex-ante moral hazard and ex-post moral hazard. With cheaper access to health-care services thanks to health insurance, policyholders might be less worried about becoming sick and exert less health-care effort, resulting in a higher probability of getting sick. This is the ex-ante moral hazard, which means that having insurance increases the probability of getting sick. In contrast, the ex-post moral hazard is the increase in hospital visits given health shocks. With insurance, the costs of hospital visits are substantially reduced and people will visit hospitals with minor health problems with which they would not go to hospital if they did not have the insurance. These two kinds of moral hazard increase the claim rates, which in turn increases insurance costs and premiums.

3. An innovative scheme: Index insurance

There is one (and maybe only one) innovative scheme that microinsurance institutions have started to use to alleviate the problems of adverse selection and moral hazard, namely, index insurance. Index insurance provides a payout based on the realization of a publicly verifiable aggregate index that is correlated with household income, such as rainfall at a local rain gauge. Because payouts are only correlated with an aggregate index that is exogenous to individuals, having insurance will not markedly affect individuals' actions such as effort and, therefore, the problem of moral hazard can be alleviated. In addition, because the likelihood that a household receives the payout is the same in the same area, the problem of adverse selection can be minimized as long as the insurer can calculate the appropriate insurance premium for each area.

Gine, Townsend, and Vickery (2008) provide a good survey on index insurance and list five desirable features for the index as follows:

  • (i) 
    The index construction is transparent to policyholders, and the realization of the index is verifiable to them.
  • (ii) 
    The calculation of the index is free from tampering or manipulation.
  • (iii) 
    The distribution of the realization of the index can be accurately estimated, so that the product can be appropriately priced and the expected return can be estimated by potential policyholders.
  • (iv) 
    The index can be measured inexpensively and calculated in a timely manner.
  • (v) 
    The realization of the index, or a transformation of the index, is highly correlated with household income and consumption.

One of the desirable indices to use is rainfall, which is used in weather insurance. Because rainfall significantly affects agricultural productivity, criterion (v) above is mostly satisfied.22 Moreover, if there are reliable long-term rainfall data objectively recorded by agencies such as the Meteorological Department, the data are transparent, verifiable, and might be available at low cost, which satisfy criteria (i) to (iii) above. Furthermore, because rainfall data are available on an almost real-time basis, it is possible to calculate payouts and pay policyholders in a timely fashion, which satisfies criterion (iv).

Regardless of the desirability of weather insurance, its take-up rate is still low. Gine, Townsend, and Vickery (2008) analyze the participation decision in weather insurance in India but the take-up rate was just 4.6% at the time of survey. They found that wealthier households are more likely to purchase and households who face credit constraints are less likely to purchase. What is particularly interesting is that more risk-averse households are “less” likely to buy insurance, which contradicts conventional theory. However, this might be due to household uncertainty about the product itself. They find that measures of familiarity with the insurance vendor play a key role in insurance take-up decisions and that more risk-averse households who are familiar with the insurance vendor are indeed more likely to buy insurance. This implies that people who are not familiar with insurance consider buying insurance as a risky investment.

4. Bundling with microcredit

Facing low take-up rates, some microinsurance institutions sell the insurance bundled with microcredit. In the weekly or monthly installment, microcredit borrowers are forced to pay the premium for the insurance in addition to the regular repayment for the credit. The two main reasons cited by most of the microinsurance institutions are:

  • • 
    Bundling is the best way to increase the number of insured.
  • • 
    Bundling can alleviate the problem of adverse selection.

By bundling health insurance with microcredit, the insurer can attract healthy people who are able to run businesses funded by microcredit. However, bundling can reduce microcredit take-up itself. Because people do not buy insurance as they do not value it, bundling unnecessary goods with microcredit programs makes the microcredit programs less attractive. Gine and Yang (2009) conduct a randomized experiment in Malawi of bundling microcredit with weather insurance at nearly actuarially fair prices. They report that bundling weather insurance with microcredit reduces take-up by 30%. They argue that this is due to limited liability, by which the repayment amount in the case of failure is reduced and, therefore, the farmers were already insured to some extent. However, this result also suggests that bundling microcredit with microinsurance that is not popular by itself can substantially reduce the take-up of microcredit. In short, bundling might help to sell insurance, but an increase in the take-up of microcredit is sacrificed.

V. SUMMARY AND CONCLUSIONS

  1. Top of page
  2. Abstract
  3. I. INTRODUCTION
  4. II. WELFARE IMPACT OF MICROCREDIT
  5. III. MECHANISM UNDERLYING HIGH REPAYMENT RATES
  6. IV. CHALLENGES OF MICROFINANCE INSTITUTIONS
  7. V. SUMMARY AND CONCLUSIONS
  8. REFERENCES
  9. Appendix

The successful expansion of small-scale financial services to the poor with high repayment records by many microfinance institutions in developing countries is often referred to as the “microfinance revolution.” Enthusiasm for microfinance as an effective tool for poverty alleviation has accelerated since the Grameen Bank and its founder, Professor Yunus, received the Nobel Peace Prize in 2006. This paper investigates the extent to which the microfinance revolution is truly revolutionary.

A review of the literature demonstrated that, although there is much anecdotal evidence emphasizing the positive effect of microfinance, the empirical results from more rigorous impact evaluations are mixed. Some researchers argue that microfinance has a positive impact on clients (Pitt and Khandker 1998), some point out that when controlling for selection biases (which are caused by such factors as self-selection and program placement), the impacts of microcredit become weaker than naïve estimates, which do not take into account selection, although positive and significant impacts still remain (Alexander-Tedeschi 2008). Others find little impact of microcredit when controlling for selection biases (Morduch 1998; Coleman 1999). Moreover, a strand of the published literature shows that near poor or nonpoor households tend to become the beneficiaries of microcredit (e.g., Copestake, Bhalotra, and Johnson 2001; Navajas et al. 2000; Sharma and Zeller 1999) and that microcredit tends to bring benefits to wealthier households rather than to poorer ones (Coleman 2006; Kondo et al. 2008; Takahashi, Higashikata, and Tsukada 2010).

However, in contrast to the widespread failure of subsidized credit programs in the past, many MFIs achieve remarkably high repayment rates, which is a prerequisite for financial self-sustainability. We showed how group lending, which was adopted initially by the Grameen Bank and then by others, theoretically overcomes adverse selection, moral hazard, and strategic default caused by asymmetric information. We also discussed that group lending is not the panacea and showed how other innovative schemes, such as dynamic incentives and frequent installments, work well to increase repayment rates.

Although these new schemes have become emblematic of microcredit, most MFIs have not succeeded in creating any innovative and effective insurance schemes that solve the asymmetric information problems inherent in the insurance market. The emerging scheme of index insurance has the potential to mitigate these problems, but so far it has not significantly increased participation.

All in all, the microfinance revolution may be revolutionary with its relatively high repayment rates, which could not be achieved by past subsidized credit programs. In addition, the introduction of insurance products can be labeled as innovative, because virtually no institutions have ever tried this in developing countries. However, judging from the empirical evidence, there is still room to improve the performance of microfinance, especially in terms of impact on and outreach to the poor. With this empirical evidence in mind, we conclude that microfinance is developing in a promising direction but has yet to reach its full potential.

REFERENCES

  1. Top of page
  2. Abstract
  3. I. INTRODUCTION
  4. II. WELFARE IMPACT OF MICROCREDIT
  5. III. MECHANISM UNDERLYING HIGH REPAYMENT RATES
  6. IV. CHALLENGES OF MICROFINANCE INSTITUTIONS
  7. V. SUMMARY AND CONCLUSIONS
  8. REFERENCES
  9. Appendix
  • Abbink, Klaus; Bernd Irlenbusch; and Elke Renner. 2006. “Group Size and Social Ties in Micro-Finance Institutions. Economic Inquiry 44, no. 4: 61428.
  • Adams, Dale W.; Douglas H. Graham; and J. D. Von Pischke. 1984. Undermining Rural Development with Cheap Credit. London: Westview Press.
  • Alexander-Tedeschi, Gwendolyn. 2006. “Here Today, Gone Tomorrow: Can Dynamic Incentives Make Microfinance More Flexible? Journal of Development Economics 80, no. 1: 84105.
  • Alexander-Tedeschi, Gwendolyn. 2008 Overcoming Selection Bias in Microcredit Impact Assessments: A Case Study in Peru. Journal of Development Studies 44, no. 4: 50418.
  • Alexander-Tedeschi, Gwendolyn, and Dean Karlan. 2010. “Cross Sectional Impact Analysis: Bias from Dropouts. Perspectives on Global Development and Technology (forthcoming).
  • Armendáriz de Aghion, Beatriz, and Jonathan Morduch. 2005. The Economics of Microfinance. Cambridge, Mass.: MIT Press.
  • Banerjee, Abhijit V.; Timothy Besley; and Timothy W. Guinnane. 1994. “Thy Neighbor's Keeper: The Design of a Credit Cooperative with Theory and a Test. Quarterly Journal of Economics 109, no. 2: 491515.
  • Banerjee, Abhijit; Esther Duflo; Rachel Glennerster; and Cynthia Kinnan. 2009. “The Miracle of Microfinance? Evidence from a Randomised Evaluation.” Poverty Action Lab Working Paper no. 101. Cambridge, Mass.: Massachusetts Institute of Technology.
  • Barnes, Carolyn; Gary Gaile; and Richard Kibombo. 2001. The Impact of Three Microfinance Programs in Uganda. Washington, D.C.: USAID-AIMS.
  • Basu, Karna. 2009. “The Provision of Commitment in Informal Banking Markets: Implications for Takeup and Welfare.” University of Chicago: Mimeo.
  • Bauer, Michal; Julie Chytilova; and Jonathan Morduch. 2008. “Behavioral Foundations of Micro-credit: Experimental and Survey Evidence from Rural India.” IES Working Papers 28/2008. Prague: Institute of Economic Studies, Faculty of Social Sciences, Charles University in Prague.
  • Besley, Timothy, and Stephen Coate. 1995. “Group Lending, Repayment Incentives and Social Collateral. Journal of Development Economics 46, no. 1: 118.
  • Bhole, Bharat, and Sean Ogden. 2010. “Group Lending and Individual Lending with Strategic Default. Journal of Development Economics 91, no. 2: 34863.
  • Bond, Philip, and Ashok S. Rai. 2009. “Borrower Runs. Journal of Development Economics 88, no. 2: 18591.
  • Cassar, Alessandra; Luke Crowley; and Bruce Wydick. 2007. “The Effect of Social Capital on Group Loan Repayment: Evidence from Field Experiments. Economic Journal 117, no. 517: F85F106.
  • Cassar, Alessandra, and Bruce Wydick. 2008. “Does Social Capital Matter? Evidence from a Five-Country Group Lending Experiment.” University of San Francisco: Mimeo.
  • Che, Yeon-Koo. 2002. “Joint Liability and Peer Monitoring under Group Lending. Contributions to Theoretical Economics 2, no. 1: Article 3.
  • Chemin, Matthieu. 2008. “The Benefits and Costs of Microfinance: Evidence from Bangladesh. Journal of Development Studies 44, no. 4: 46384.
  • Chowdhury, Prabal Roy. 2007. “Group-Lending with Sequential Financing, Contingent Renewal and Social Capital. Journal of Development Economics 84, no. 1: 487506.
  • Churchill, Craig. 2002. “Trying to Understand the Demand for Microinsurance. Journal of International Development 14, no. 3: 38187.
  • Coleman, Brett E. 1999. “The Impact of Group Lending in Northeast Thailand. Journal of Development Economics 60, no. 1: 10541.
  • Coleman, Brett E. 2006 Microfiance in Northeast Thailand: Who Benefits and How Much. World Development 34, no. 9: 161238.
  • Conning, Jonathan, and Christopher Udry. 2007. “Rural Financial Markets in Developing Countries.” In Handbook of Agricultural Economics Vol. 3, ed. RobertEvenson and PrabhuPingali. Amsterdam: North-Holland.
  • Copestake, James; Sonia Bhalotra; and Susan Johnson. 2001. “Assessing the Impact of Microcredit: A Zambian Case Study. Journal of Development Studies 37, no. 4: 81100.
  • Cull, Robert; Asli Demirguc-Kunt; and Jonathan Morduch. 2007. “Financial Performance and Outreach: A Global Analysis of Leading Microbanks. Economic Journal 117, no. 1: F107F133.
  • Daley-Harris, Sam. 2006. “State of the Microcredit Summit Campaign Report, 2006.” Microcredit Summit, Washington D.C.: Mimeo.
  • Drake, Deborah, and Elisabeth Rhyne, eds. 2002. The Commercialization of Microfinance: Balancing Business and Development. Bloomfield, Conn.: Kumarian Press.
  • Duflo, Esther; Rachel Glennerster; and Michael Kremer. 2006. “Using Randomization in Development Economics Research: A Toolkit.” MIT Department of Economics Working Paper no. 06-36. Cambridge, Mass.: Massachusetts Institute of Technology.
  • Dunn, Elizabeth, and J. Gordon Arbuckle, Jr. 2001. “The Impacts of Microcredit: A Case Study from Peru.” Washington, D.C.: Assessing the Impact of Microenterprise Services (AIMS).
  • Dupas, Pascaline, and Jonathan Robinson. 2009. “Savings Constraints and Microenterprise Development: Evidence from a Field Experiment in Kenya.” NBER Working Paper no. 14693.
  • Field, Erica, and Rohini Pande. 2008. “Repayment Frequency and Default in Micro-Finanace: Evidence from India. Journal of the European Economic Association 6, no. 2–3: 5019.
  • Ghatak, Maitreesh. 1999. “Group Lending, Local Information, and Peer Selection. Journal of Development Economics 60, no. 1: 2750.
  • Ghatak, Maitreesh, and Timothy W. Guinnane. 1999. “The Economics of Lending with Joint Liability: Theory and Practice. Journal of Development Economics 60, no. 1: 195228.
  • Gine, Xavier; Pamela Jakiela; Dean Karlan; and Jonathan Morduch. 2006. “Microfinance Games.” Center Discussion Papers no. 936. New Haven, Conn.: Economic Growth Center, Yale University.
  • Gine, Xavier; Robert Townsend; and James Vickery. 2008. “Patterns of Rainfall Insurance Participation in Rural India. World Bank Economic Review 22, no. 3: 53966.
  • Gine, Xavier, and Dean Yang. 2009. “Insurance, Credit, and Technology Adoption: Field Experimental Evidence from Malawi. Journal of Development Economics 89, no. 1: 111.
  • Grimard, Franque. 1997. “Household Consumption Smoothing through Ethnic Ties: Evidence from Côte d'Ivoire. Journal of Development Economics 53, no. 2: 391422.
  • Guttman, Joel M. 2008. “Assortative Matching, Adverse Selection, and Group Lending. Journal of Development Economics 87, no. 1: 5156.
  • Heckman, James J; Lance Lochner; and Christopher Taber. 1998. “General-Equilibrium Treatment Effects: A Study of Tuition Policy. American Economic Review 88, no. 2: 38186.
  • Ito, Seiro. 2007. “A Practical Guide to the Program Evaluation Methods.” In Health Service and Poverty: Making Health Services More Accessible to the Poor, ed. HirokoUchimura. Chiba: Institute of Developing Economies, JETRO.
  • Ito, Seiro, and Hisaki Kono. 2010. “Why Is the Take-up of Microinsurance So Low? Evidence from a Health Insurance Scheme in India. Developing Economies 48, no. 1: 74101.
  • Karlan, Dean. 2001. “Microfinance Impact Assessments: The Perils of Using New Members as a Control Group. Journal of Microfinance 3, no. 2: 7585.
  • Karlan, Dean. 2007 Social Connections and Group Banking. Economic Journal 117, no. 517: F52F84.
  • Karlan, Dean, and Nathanael Goldberg. 2007. “Impact Evaluation for Microfinance: Review of Methodological Issues.” World Bank Doing Impact Evaluation Series no. 7.
  • Karlan, Dean, and Jonathan Zinman. 2009a. “Expanding Microenterprise Credit Access: Using Randomized Supply Decisions to Estimate the Impacts in Manila.” BREAD Working Paper no. 235.
  • Karlan, Dean, and Jonathan Zinman. 2009b Observing Unobservables: Identifying Information Asymmetries with A Consumer Credit Field Experiment. Econometrica 77, no. 6: 19932008.
  • Khandker, Shahid R. 2005. “Microfinance and Poverty: Evidence Using Panel Data from Bangladesh. World Bank Economic Review 19, no. 2: 26386.
  • Kondo, Toshio; Aniceto Orbeta; Clarence Dingcong; and Christine Infantado. 2008. “Impact of Microfinance on Rural Households in the Philippines. IDS Bulletin 39, no. 1: 5170.
  • Kono, Hisaki. 2010. “Is Joint Liability a Good Enforcement Scheme for Achieving High Repayment Rates?: Evidence from Framed Field Experiments in Vietnam.” Institute of Developing Economies: Mimeo.
  • Kurosaki, Takashi, and Hidayat Ullah Khan. 2009. “Vulnerability of Microfinance to Strategic Default and Covariate Shocks: Evidence from Pakistan.” The Institute of Economic Research, Hitotsubashi University, Tokyo: Mimeo.
  • McIntosh, Craig; Alan De Janvry; and Elisabeth Sadoulet. 2005. “How Rising Competition among Microfinance Institutions Affects Incumbent Lenders. Economic Journal 115, no. 506: 9871004.
  • McIntosh, Craig, and Bruce Wydick. 2005. “Competition and Microfinance. Journal of Development Economics 78, no. 2: 27198.
  • McKernan, Signe-Mary. 2002. “The Impact of Microcredit Programs on Self-Employment Profits: Do Noncredit Program Aspects Matter? Review of Economics and Statistics 84, no. 1: 93115.
  • Menon, Nidhiya. 2006a. “Long Term Benefits of Membership in Microfinance Programmes. Journal of International Development 18, no. 4: 57194.
  • Menon, Nidhiya. 2006b Non-Linearities in Returns to Participation in Grameen Bank Programs. Journal of Development Studies 42, no. 8: 1379400.
  • Montgomery, Heather. 2005. “Meeting the Double Bottom Line: The Impact of Khushhali Bank's Microfinance Program in Pakistan.” ADBI Policy Papers no. 8.
  • Morduch, Jonathan. 1998. “Does Microfinance Really Help the Poor? New Evidence from Flagship Programs in Bangladesh.” Working Papers no. 211. Princeton, N.J.: Research Program in Development Studies, Princeton University.
  • Morduch, Jonathan. 1999 The Microfinance Promise. Journal of Economic Literature 37, no. 4: 1569614.
  • Mosley, Paul. 2001. “Microfinance and Poverty in Bolivia. Journal of Development Studies 37, no. 4: 10132.
  • Navajas, Sergio; Mark Schreiner; Richard L. Meyer; Claudio Gonzalez-Vega; and Jorge Rodriguez-Meza. 2000. “Microcredit and the Poorest of the Poor: Theory and Evidence from Bolivia. World Development 28, no. 2: 33346.
  • Nourse, Timothy H. 2001. “The Missing Parts of Microfinance: Services for Consumption and Insurance. SAIS Review 21, no. 1: 6170.
  • Okten, Cagla, and Una Okonkwo Osili. 2004. “Social Networks and Credit Access in Indonesia. World Development 32, no. 7: 122546.
  • Pitt, Mark M., and Shahidur R. Khandker. 1998. “The Impact of Group-Based Credit Programs on Poor Households in Bangladesh: Does the Gender of Participants Matter? Journal of Political Economy 106, no. 5: 95896.
  • Ravallion, Martin. 2008. “Evaluating Anti-Poverty Program.” In Handbook of Development Economics 4, ed. PaulSchultz and JohnStrauss. Amsterdam: Elsevier.
  • Ray, Ashok S, and Tomas Sjostrom. 2004. “Is Grameen Lending Efficient? Repayment Incentives and Insurance in Village Economies. Review of Economics Studies 71, no. 1: 21734.
    Direct Link:
  • Robinson, Marguerite S. 2001. The Microfinance Revolution: Sustainable Finance for the Poor. Washington, D.C.: World Bank.
  • Roodman, David, and Jonathan Morduch. 2009. “The Impact of Microcredit on the Poor in Bangladesh: Revisiting the Evidence.” Working Paper no. 174. Washington, D.C.: Center for Global Development.
  • Rosenbaum, Paul R., and Donald B. Rubin. 1983. “The Central Role of the Propensity Score in Observational Studies for Causal Effects. Biometrika 70, no. 1: 4155.
  • Setboonsarng, Sununtar, and Ziyodullo Parpiev. 2008. “Microfinance and the Millennium Development Goals in Pakistan: Impact Assessment Using Propensity Score Matching.” ADB Institute Discussion Paper no. 104.
  • Sharma, Manohar, and Manfred Zeller. 1999. “Placement and Outreach of Group-Based Credit Organizations: The Cases of ASA, BRAC, and PROSHIKA in Bangladesh. World Development 27, no. 12: 212336.
  • Shoji, Masahiro. 2010. “Does Contingent Repayment in Microfinance Help the Poor during Natural Disasters. Journal of Development Studies 46, no. 2: 191210.
  • Stiglitz, Joseph E. 1990. “Peer Monitoring and Credit Market. World Bank Economic Review 4, no. 3: 35166.
  • Takahashi, Kazushi; Takayuki Higashikata; and Kazunari Tsukada. 2010. “The Short-Term Poverty Impact of Small-Scale, Collateral-Free Microcredit in Indonesia: A Matching Estimator Approach. Developing Economies 48, no. 1: 12855.
  • Townsend, Robert M. 1994. “Risk and Insurance in Village India. Econometrica 62, no. 3: 53991.
  • UNCDF. 2004. Microfinance Programme Impact Assessment, 2003. New York: UNCDF.
  • Van Tassel, Eric. 1999. “Group Lending under Asymmetric Information. Journal of Development Economics 60, no. 1: 325.
  • Weiss, John, and Heather Montgomery. 2004. “Great Expectations: Microfinance and Poverty Reduction in Asia and Latin America. Oxford Development Studies 33, no. 3–4: 391416.
  • Wooldridge, Jeffrey M. 2002. Econometric Analysis of Cross Section and Panel Data. Cambridge, Mass.: MIT Press.
  • World Bank. 2000. World Development Report, 2000/2001: Attacking Poverty. New York: Oxford University Press.
  • Yaron, Jacom, and McDonald Benjamin. 2002. “Recent Developments in Rural Finance Markets.” In Zeller and Meyer (2002).
  • Zeller, Manfred, and Richard L. Meyer, eds. 2002. The Triangle of Microfinance: Financial Sustainability, Outreach and Impact. Baltimore, Md.: Johns Hopkins University Press.
Footnotes
  • 1

    The same applies to ATE.

  • 2

    For more detail, see, e.g., Ravallion (2008).

  • 3

    This sub-subsection relies largely on Takahashi, Higashikata, and Tsukada (2010).

  • 4

    Another related study is Copestake, Bhalotra, and Johnson (2001) on Zambia, which shows that profit estimated by the DID method is positive but significantly lower than that by pipeline comparisons, suggesting that the pipeline comparison will be biased upward.

  • 5

    Although Montgomery (2005) recognizes potential bias due to dropout, he or she also applies a method similar to Coleman (1999) in her study of Khushhali Bank in Pakistan. This methodology is, however, challenged by Setboonsarng and Parpiev (2008). Using the propensity score matching method, the authors find that the microcredit program by Khushhali Bank in Pakistan has lower impacts than those studied by Montgomery (2005).

  • 6

    In this paper we focus on the condition under which the credit market exists. As shown below, given Type B's higher return and the uniform interest rate, we cannot have the situation where only Type A borrows and Type B does not borrow. Whenever the credit market exists, lending to both Type A and Type B generates higher aggregate payoffs than non-lending.

  • 7

    Another example of moral hazard is the borrower's decision on investment type, whether risky or safe. A model for risky investment is available upon request, but the results are qualitatively similar to the case of effort choice analyzed in the present paper.

  • 8

    Under the setting in this section, this assumption reduces to p(1 +r) − (1 −p)zpqYz, where the left-hand side is the expected profit when borrowers do not commit strategic default (get repaid 1 +r when their projects succeeds and not get repaid and are audited when their projects fail) and the right-hand side is the expected profit when they choose strategic default. In the following subsections, we maintain the assumption that the bank prefers that borrowers do not commit strategic default over letting borrowers default.

  • 9

    Chowdhury (2007) and Guttman (2008) point out that if repeated play is incorporated in the model, assortative matching is not necessarily achieved.

  • 10

    Because there is no problem of strategic default in the model of adverse selection, the bank can collect 2(1 +r) from successful borrowers if their partners have no income.

  • 11

    Because there is no problem of strategic default in the model of adverse selection, the bank can collect YA(< 2(1 +r)) from successful Type A borrowers if their partners have no income.

  • 12

    Note that the bank needs to monitor only when the group does not repay 2(1 +r), which occurs with probability (1 −p)2 (when both borrowers have no income) in this case.

  • 13

    Note that the bank needs to monitor except when both members have income Y, which occurs with probability (1 −p2), because in other cases the borrowers will not repay and the bank needs to monitor in these cases in order to give them incentives to repay when both have income Y.

  • 14

    This is because inline imageinline image.

  • 15

    Che (2002) analyzes the repeated games of moral hazard under individual lending and joint liability, but in his setting, no matter whether the borrowers repay the loans or not, lending is never terminated. In this sense, his model does not incorporate dynamic incentives.

  • 16

    The condition inline image is identical to the condition where the expected payoff from repaying the loan is not less than that from not repaying, inline image.

  • 17

    Note that the bank does not do auditing and, therefore, does not incur monitoring cost z because q= 0.

  • 18

    Note that the condition δ(1 −ρ)pY≥ (1 −δρ)(1 +r) implies inline image, ensuring inline image.

  • 19

    Dupas and Robinson (2009) find in a randomized control trial of opening interest-free savings accounts with substantial withdrawal fees (which works as a commitment device) that it had substantial positive impacts on productive investment levels and expenditure levels, which supports the importance of savings constraints.

  • 20

    Individual fixed effects capture the type of individual (adverse selection) and individual investment decision (moral hazard) and, therefore, help them purge any effects other than strategic default and covariate shocks.

  • 21

    MFIs can provide insurance against such regional income shocks in the form of rescheduling loan repayments. Shoji (2010) finds that in Bangladesh, MFIs' loan repayment rescheduling against the event of nationwide flooding helps keep borrowers away from borrowing high-interest money from moneylenders.

  • 22

    The primary disadvantage of index-based rainfall insurance is basis risk, that is, rainfall is imperfectly correlated with household income fluctuation. First, crop yields are not perfectly correlated with rainfall. Second, rainfall measured at the local weather station is not perfectly correlated with rainfall for an individual plot. Third, crop revenues are affected by non-weather factors such as pests, disease, and crop prices that are not closely correlated with measured local rainfall.

  • 23

    We use the condition peY > 1 +c to prove this.

  • 24

    If the bank expects that the borrowers will shirk, the interest rate determined by the break-even constraint is inline image. With this interest rate, the borrower will not take up the loan because the expected payoff will be psYps(2 −ps)(1 +r) =psY− (1 +c) < 0.

  • 25

    We use the assumption peY > 1 +c for proving this.

  • 26

    We make this assumption to make the argument consistent with Besley and Coate (1995).

  • 27

    Because a borrower will shoulder his or her partner's repayment only when inline image, whose right-hand side is larger than 2(1 +r), we do not have to consider separately the case Y≥ 2(1 +r) (where a borrower can shoulder the full amount of his or her partner's repayment) and the case Y < 2(1 +r) (where he or she cannot) as we did in previous sections.

  • 28

    Note that inline image.

Appendix

  1. Top of page
  2. Abstract
  3. I. INTRODUCTION
  4. II. WELFARE IMPACT OF MICROCREDIT
  5. III. MECHANISM UNDERLYING HIGH REPAYMENT RATES
  6. IV. CHALLENGES OF MICROFINANCE INSTITUTIONS
  7. V. SUMMARY AND CONCLUSIONS
  8. REFERENCES
  9. Appendix
A. Adverse Selection under Group Lending When 2(1 + r) > YB> YA

When 2(1 +r) > YB > YA, the borrowers' participation constraints are given by (11) for Type A and (13) for Type B. From the break-even condition inline image, the equilibrium interest rate is

  • image((A1))

Type A's participation constraint can be written as

  • image((A2))

Some algebra shows that the right-hand side of this expression is smaller than that of (2):

  • image

Therefore, group lending relaxes A's participation constraint and makes lending feasible with a wider range of parameter c.

B. Moral Hazard under Group Lending When Y < 2(1 + r)

Assume that the members of the group act to maximize the total payoff of the two borrowers and anyone who deviates will be punished with serious social sanctions. When Y < 2(1 +r), the group will choose to exert effort if inline image, or

  • image((A3))

Because the bank's profit from a group (two members) given satisfaction of the incentive constraint (A3) is inline image, the bank's break-even condition implies

  • image((A4))

which is again lower than the equilibrium interest rate under no group lending (4). Substituting this into (A3), we can derive the condition that the group of the borrowers chooses to exert effort A as

  • image((A5))

whose right-hand side is again smaller than that of (5).23 Thus, we have shown that group lending can induce borrowers to exert effort with smaller η if the group can coordinate its members' investment choice.

C. Moral Hazard under Group Lending When the Group Cannot Coordinate Its Members' Effort Choices

Suppose that the group cannot coordinate its members' effort choices at all and each member pursues his or her own interest. The payoff of a borrower choosing effort with his or her partner choosing j=e, s is expressed as

  • image

and that from choosing to shirk with his or her partner choosing j=e, s is

  • image

We need to consider two cases: (i) Y≥ 2(1 +r) and (ii) Y < 2(1 +r).

Case (i): Y≥ 2(1 +r)

The payoff matrix table can be written as in Appendix Table 1.

Table APPENDIX TABLE1.  Group Lending and Moral Hazard Without Coordination among Group Members
 Borrower 2's choice
EffortShirking
  1. Note: In each cell, the upper value expresses Borrower 1's payoff and the lower value expresses Borrower 2's payoff.

Borrower 1's choice  
 EffortpeYpe(2 −pe)(1 +r) −dpeYpe(2 −ps)(1 +r) −d
peYpe(2 −pe)(1 +r) −dpsYps(2 −pe)(1 +r)
 ShirkingpsYps(2 −pe)(1 +r)psYps(2 −ps)(1 +r)
pAYpe(2 −ps)(1 +r) −dpsYps(2 −ps)(1 +r)

The condition that the borrower exerts effort when the partner exerts effort is peYpe(2 −pe)(1 +r) − d ≥psYps(2 −pe)(1 +r), which reduces to

  • image((A6))

Analogously, the condition that the borrower exerts effort when his or her partner shirks can be written as

  • image((A7))

Because 2 −ps > 2 −pe≥ 1, we need larger η than that in (3) to satisfy these incentive constraints given r, implying that shirking is more likely to be chosen under group lending. Under group lending, the return from the effort will be used for shouldering his or her partner with positive probability (1 −pe or 1 −ps) while under individual lending, the return from the effort will accrue exclusively to the borrower. Thus, under group lending, there is an “externality” in exerting effort, which makes exerting effort less attractive for the individual borrower.

However, the equilibrium interest rate under group lending (18) is lower than under no group lending. By substituting (18) into (A6) and (A7), we obtain

  • image((A8))

in the case where his or her partner exerts effort and

  • image((A9))

in the case where his or her partner shirks. Condition (A8) is exactly the same as (5). Hence, as long as his or her partner chooses effort, group lending without coordination among the group members will generate the same result as individual lending in terms of moral hazard. However, because inline image, condition (A9) is stricter than (5), implying that group lending will worsen the problem of moral hazard when his or her partner chooses to shirk.

From these conditions, we can derive the following equilibrium outcomes. If inline image, the only equilibrium outcome is that the bank provides loans with interest rate (18) and the group members exert effort. If inline image, a borrower will exert effort when his or her partner exerts effort and will shirk when his or her partner shirks (multiple equilibria). Thus the equilibrium outcomes are either: (i) the bank provides loans with interest rate (18) and the group members exert effort, or (ii) the bank provides loans with interest rate inline image and the group members do not take up the loans.24 Finally, if inline image, then the only equilibrium outcome is that the bank provides loans with interest rate inline image and the group members do not take up the loans.

Case (ii): Y < 2(1 +r)

The condition that the borrower exerts effort when the partner exerts effort is inline image, which reduces to

  • image((A10))

Analogously, the condition that the borrower exerts effort when his or her partner shirks can be written as

  • image((A11))

In this case, the equilibrium interest rate is (A4). By substituting (A4) into (A10) and (A11) with rearrangement, we obtain

  • image((A12))

in the case where his or her partner exerts effort and

  • image((A13))

in the case where his or her partner shirks. Condition (A12) is exactly the same as (5), and condition (A13) is stricter than (5).25 Thus, as in case (i), as long as his or her partner chooses effort, group lending without coordination among the group members will not affect the problem of moral hazard but will worsen it when his or her partner chooses to shirk. The equilibrium outcomes are as follows. If inline image, the bank provides loans with interest rate (18) and the group members exert effort. If inline image, either (i) the bank provides loans with interest rate (18) and the group members exert effort, or (ii) the bank provides loans with interest rate inline image and the group members do not take up the loans. If inline image, the bank provides loans with interest rate inline image and the group members do not take up the loans.

D. Strategic Default under Group Lending When the Group Cannot Coordinate Its Members' Repayment Decisions

Consider the problem of strategic default under group lending when the group cannot coordinate or enforce its members' decisions and each member pursues his or her own interest. The borrowers simultaneously choose to repay or not to repay before actually repaying to the bank. If a borrower chooses to repay but another borrower does not, then the former needs to decide whether to shoulder his or her partner's repayment. If he or she does not want to shoulder, he or she can decide not to repay his or her own loan eventually.26 Because both borrowers will be audited if the group does not repay 2(1 +r), it is not optimal for a borrower to repay eventually if he or she chooses not to shoulder. Thus, a successful borrower who decides not to shoulder (and, therefore, not to repay eventually) will obtain Y if the bank audit is not successful (with probability 1 −q).

Appendix Figure 1 depicts the game tree of this game when both borrowers have investment returns Y. Note that the borrower will shoulder his or her partner's loan repayment if and only if Y− 2(1 +r) ≥ (1 −q)Y, or inline image.27Appendix Table 2 describes the payoff matrix at the timing of both borrowers choosing to repay or not to repay in this case.

  • image(Appendix Fig.1.)

[ Game Tree: Strategic Default with Group Lending When Both Borrowers Have Successful Investment Return ]

Table APPENDIX TABLE2.  Group Lending and Strategic Default without Any Enforcement Mechanism When inline image
 Borrower 2's Choice
Repay and ShoulderNot Repay
  1. Note: In each cell, the upper value expresses Borrower 1's payoff and the lower value expresses Borrower 2's payoff. Because the strategy of “Repay and non-shoulder” is dominated by the strategy of “Repay and shoulder” in the case of inline image, the former does not appear in the table.

Borrower 1's choice  
 Repay and shoulderY− (1 +r)Y− 2(1 +r)
Y− (1 + r)Y
 Not RepayY(1 −q)Y
Y− 2(1 +r)(1 −q)Y

If both borrowers choose to repay, then their payoffs will be Y− (1 +r). If one borrower chooses to repay but another borrower chooses not to repay, then the latter will obtain Y while the former's payoff will be Y− 2(1 +r) because he or she will choose to shoulder his or her partner's loan repayment. If both borrowers choose not to repay, their expected payoffs are (1 −q)Y. Note that when his or her partner chooses to repay, it is better for him or her to choose not to repay. This is because he or she expects that if he or she chooses not to repay, his or her partner will shoulder his or her loan repayment because inline image. Therefore, he or she has an incentive to free-ride when his or her partner chooses to repay. When his or her partner chooses not to repay, he or she will choose to repay if and only if inline image. However, this condition is necessarily met because we are considering the case inline image (where a borrower has an incentive to shoulder his or her partner's repayment). Thus, given the loan provision, the Nash equilibrium for the two group members will be (repay and shoulder, not repay) and (not repay, repay and shoulder).

If one borrower has investment returns Y but another does not, the latter has no option but default. Appendix Figure 2 depicts the game tree of the case where borrower 1 has Y but borrower 2 does not. Borrower 1 has an incentive to shoulder for borrower 2 when inline image, and only when he or she prefers to shoulder does he or she have an incentive to repay. Thus, if only one borrower succeeds in investment, the credit will be repaid only when inline image.

  • image(Appendix Fig.2.)

[ Game Tree: Strategic Default with Group Lending When Borrower 1 Has Successful Investment Return But Borrower 2 Does Not ]

Next, we turn to the case inline image, where a borrower has no incentive to shoulder. If one chooses not to repay, then it will lead to default because his or her partner will not shoulder. Thus, when one of the borrowers has no return, Y= 0, then his or her partner will choose not to repay.

Appendix Table 3 presents the payoff matrix of the case where both borrowers have returns Y. Given his or her partner's choice to repay, he or she has an incentive to repay if and only if inline image. When his or her partner chooses not to repay, he or she will never choose to repay. In this case, if inline image, we have multiple Nash equilibria of (repay, repay) and (not repay, not repay). Note that inline image ensures that the renegotiation-proof equilibrium of this game is (repay, repay) only. However, if inline image, we have only a Nash equilibrium of (not repay, not repay).

Table APPENDIX TABLE3.  Group Lending and Strategic Default without Any Enforcement Mechanism When inline image
 Borrower 2's Choice
Repay and ShoulderNot Repay
  1. Note: In each cell, the upper value expresses Borrower 1's payoff and the lower value expresses Borrower 2's payoff. Because the strategy of “Repay and shoulder” is dominated by the strategy of “Repay and non-shoulder” in the case of inline image, the former does not appear in the table.

Borrower 1's choice  
 Repay and non-shoulderY− (1 +r)(1 −q)[Y− (1 +r)]
Y− (1 +r)(1 −q)Y
 Not Repay(1 −q)Y(1 −q)Y
(1 −q)[Y− (1 +r)](1 −q)Y

The analysis above shows that as long as renegotiation is possible, the aggregate repayment rate will be p2+ 2p(1 −p) =p(2 −p) when inline image, p2 when inline image and zero when inline image. Note that this is the same as the case where the group can force its members to act to maximize the total payoff of the two borrowers (see conditions (21) and (20)). The equilibrium interest rate is also the same. So, the aggregate repayment rate is

  • image
  • image
  • image

If renegotiation is not possible, the aggregate repayment rate will be lower than p2 when inline image.

E. Comparison of Individual Lending with Dynamic Incentives and Joint Liability with Dynamic Incentives (Moral Hazard)

Compare the individual lending with dynamic incentives and joint liability with dynamic incentives. With the equilibrium interest rate (4), the condition inline image reduces to

  • image((A14))

However, with the equilibrium interest rate (18), the condition inline image reduces to

  • image((A15))

We denote the threshold value of η under individual lending with dynamic incentives and joint liability with dynamic incentives by

  • image
  • image

respectively. Because 1 +cδpepsY > 1 +cδpeps(2 −ps)Y, if inline image, then we should have ηIL > ηJL; that is, joint liability with dynamic incentives can induce effort with a wider range of parameters than individual lending with dynamic incentives. The condition inline image reduces to 1 −δ(2 −peps+peps) ≥ 0. Whenever 1 −δ(2 −peps+peps) ≥ 0, we should have ηIL > ηJL.

Now, consider the case 1 −δ(2 −peps+peps) < 0. Let inline image and B ≡ 1 −δ(2 −peps+peps) < 0. Then, by using peY > 1 +c, we can derive

  • image

Thus, we have proved that we always have ηIL > ηJL.28 Under the assumption that the group can force its members to act to maximize the total payoff of the two borrowers, joint liability with dynamic incentives performs better than individual lending with dynamic incentives in terms of preventing moral hazard.