Microfinance commercialization
and the need for
financial and social returns
03/07/2008
R.N. van den Brink
Faculty of Economics
Rijksuniversiteit Groningen
Supervisor: prof. dr. B.W. (Robert) Lensink
Groningen, July 2008
Abstract
1.1 Introduction
The United Nation (UN) reports that in 2005 still over 1 billion people, about one sixth of the world population, were living in extreme poverty.1 These people earn less than one dollar a day. To fight this poverty, and related catastrophes, the UN has set the United Nations Millennium Development Goals in September 2000, which aim at halving extreme poverty by 2015.2 According to the UN, microfinance can be a significant contribution to the achievement of the Millennium development Goals (MDG’s).3 Therefore, the UN declared 2005 to be the International year of Microfinance, to bring further attention to global poverty and the impact of microfinance.
In a very broad sense, Micro Finance is the provision of financial services to the poor. At first, the aim of micro finance was to provide very small loans (Microcredit) to the poorest on earth, to help them engage in productive activities or grow their tiny businesses (microenterprises), which could not have been financed otherwise. Over time though, Microfinance moved toward a broader range of services including loans, savings, insurance, transfer services and other financial products. Microfinance organizations and academics have come to realize that the poor require a variety of financial products, enabling “a world in which as many poor and near-poor households as possible have permanent access to an appropriate range of high quality financial services, including not just credit but also savings, insurance, and fund transfers” (Christen, Rosenberg and Jayadeva, 2004).
The practice of modern Microfinance, as it is known today, became active during the 1980s. Other attempts of lending to the poor, mostly done by development organizations, date back around ten years before in Bangladesh, Brazil and few other countries. However, these projects have failed to be sustainable projects to reach the poor. In 1983, a research project to examine the possibility of designing a credit delivery system to provide banking services targeted to the rural poor was transformed into an independent bank by government legislation. This bank, the Grameen Bank, was founded by Muhammad Yunus, and is remembered as a groundbreaking step in MicroFinance. Ever since new microfinance programs and institutions have been founded rapidly worldwide. The Grameen bank and its founder were jointly awarded the Nobel Peace Prize in 2006.
1
The UN Development Goals Report 2005; webite www.un.org/millenniumgoals/documents.html 2
The UN Millennium Development Goals (MDG’s) are: (1) eradicate extreme poverty and hunger; (2) achieve universal primary educations; (3) promote gender equality and empower women; (4) reduce child mortality; (5) improve maternal health; (6) combat HIV/AIDS, malaria, and other disease; (7) ensure environmental sustainability; and (8) develop a global partnership for development. 3
In the past decade, more and more microfinance organizations have claimed significant positive effects on society. In addition, several theoretical and empirical studies have pointed out the same perspective. It is indicated that microfinance reduces poverty and influences society positively further than only financially. As indicated by Murdoch et al. (2003), poverty is multi-dimensional and by providing access to financial services, many aspects of poverty may be reduced. Though, it should be noted that it is very important to address the drawbacks of microfinance impact assessments. Murdoch (1999) points out that the exhilarated win-win situation of promising poverty alleviation with suitable profits moved far ahead of its evidence in an early stage. Better evidence is therefore needed. To understand how microfinance became such an extended global movement, a general impression and the impact of microfinance are described in paragraph 2.
The main part of this paper contains a discussion on microfinance commercialization and its concerns. The key research question to be answered here is whether there may be potential investment opportunities with both financial and social returns. By showing the need for additional funding due to commercialization and the subsequent conflicts with the traditional aim of microfinance, the need for a different investment model is presented. This model should contribute microfinance by providing investors a financial perspective and in addition social responsible proceeds.
The social responsible side of the potential investments is indicated as outreach to the poor. An active debate is going on about the trade-off between outreach to the poor and efficiency. Where, due to commercialization, MFI’s tend to move to efficiency driven models, the traditional aim of microfinance was to reach the poor. Efficiency driven MFI’s have induced a shift from grouplending to individual lending practices, and towards more regulated MFI’s. Critics of microfinance commercialization argue that this will lead to a mission drift. Exorbitant payments to microfinance executives and extremely large capital gains for micro-investors have colored the microfinance sector the last couple of years. Therefore, it is argued that new models should be implemented to ensure sustainable access to financial services for the poor majority.
microfinance investments with a financial beneficial profile and the social responsible effects are those this paper seeks for.
2.1 A general impression of Microfinance
According to the MIX (MicroFinance Information Exchange), a not for profit private global microfinance information platform, the typical microfinance clients are low-income persons that do not have access to formal financial institutions. Microfinance clients are typically self-employed, often household-based entrepreneurs. In rural areas, they are usually small farmers and others who are engaged in small income-generating activities such as food processing and petty trade. In urban areas, microfinance activities are more diverse and include shopkeepers, service providers, artisans, street vendors, etc. Microfinance clients are poor and vulnerable non-poor who have a relatively stable source of income.
In addition, MIX points out that as Micro finance services are broadened, the potential market of microfinance clients also expands. As an example, poor farmers may not really wish to borrow but would rather like a safer place to save the proceeds from their harvest as these are consumed over several months by the requirements of daily living.
A distinctive feature of many microfinance providers is that a significant majority of clients are women, which is also strongly supported by the Grameen Bank. This originates from the traditional aim of microfinance, to serve the poor, because a high percentage of women tend to be among the poorest of the poor. Aguilar (undated) points out how, compared to men, women have greater barriers to overcome. Microfinance services could help them to extend their horizon and offer them social recognition and empowerment. Microfinance should enable women to produce a great impact, because they not only will increase the quality of their own lives but also at the same time the quality of their entire family.
The organizations that provide financial services to the poor are called Microfinance Institutions (MFI’s). The term MFI has come to refer to a wide range of organizations dedicated to providing these services: NGO’s, Credit Unions, Cooperatives, Private Commercial Banks and Non-Bank Financial Institutions, and parts of State-owned Banks.
A majority of MFI’s that offer microcredit also provide many non-financial development services and many commercial banks only have a small portion of its assets tied up in financial services to the poor. Though, these organizations are engaged in supplying financial services to the poor and so they do exist as MFI’s.
Traditionally, commercial banks have mostly avoided the provision of financial services to the poorest. Baydas, Graham and Valenzuela (1997) name three basic concerns. First, for commercial banks, microenterprises are recognized as bad credit risks. The perception is that small clients do not have stable, viable businesses for which to borrow and from which to generate repayment. Also, the poor have few or no assets that can be secured by a bank as collateral. In addition, the appropriate lending technologies to serve these clients, such as correct credit screening mechanisms, are often not available. Second, the operations to serve micro loans will be inefficient and costly because they are small and short-term. Managing a client’s account will incur costs for a bank regardless of the size of a deposit or loan, while the return on a larger loan or deposit compared to small ones will be much greater. Consequently, there is no incentive to make small loans or except small deposits. Third, there seems to be far-reaching social, cultural and language barriers for the poor that do not allow for an easy relationship with modern banking organizations.
To avoid this lack of practical access to the formal financial sector by the poor, new and innovative banking systems were developed during the 1980s. These innovations enable commercial financial organizations to make small loans to the poor in a sustainable way, even with acceptable interest rates, without requiring collateral. Baydas, Graham and Valenzuela (1997) name several principal characteristics of microlending technologies, including:
- lending based on character, rather than collateral - group loan mechanism as a collateral substitute
- interest rates considerably higher than those for larger bank customers to cover all costs of the microfinance program
- frequent repayment schedules to facilitate monitoring of borrowers
- staff drawn from local communities with access to information about potential clients
An essential distinction of microfinance, compared to individual-based lending in traditional banking, is the idea of group lending. This is also firmly supported by the Grameen Bank. Individual-based lending draws on traditional banking practices and involves a standard bilateral relationship between the bank and customer. The liability for repaying the loan rests with the individual borrower only. This means individual lending is most vulnerable to problems imposed by information asymmetries and weak enforcement capacities.
When grouplending by MFI’s in microfinance is mentioned, this usually refers to solidarity group lenders. The idea of solidarity group lending is based on a credit approach that uses the peer-pressure from joint liability within self-formed groups of clients (between 3 and 10 members), to ensure the borrowers follow credit discipline. This means when a group member declines to pay interest, the rest of the group will have to contribute or else the credit arrangements for all the group members will stop, and in the future new contracts will be refused for all members.
In addition, village banking may also be considered as group lending, since village banks rely on a variation of the solidarity lending methodology.4 A village bank is an informal self-help support group of 20-30 members, in which members are supported by small loans, trained, mentored and motivated. The village banking system relies on cross-guarantees, where each member is responsible for the loan of another, so that social pressure ensures total repayment. In this paper, when referred to grouplending MFI’s it means both solidarity as village banking MFI’s.
Research in the area of the effectiveness of grouplending has been lively, as indicated by Hermes and Lensink (2007). They point out that that in theory, group lending should reduce agency costs for a lender because it stimulates screening, monitoring and enforcement of contracts among borrowers. The joint liability creates an incentive for individual group members to enforce repayment in order to reduce the risk of having to contribute to the repayment of loans of others and to ensure access to future loans. The group lending structure is also expected to be more effective in providing such activities as compared to the lender, because group members usually live close to each other and/ or have social ties. They are therefore better informed about each other’s activities. Hence, high repayment performance is expected because of effective screening, monitoring and enforcement among group members. For a comprehensive overview of theoretical and empirical work on grouplending, the article of Hermes and Lensink (2007) should be referred to.
4
2.2 The impact of Micro Finance
The practice of Microfinance has become a tremendously popular anti-poverty strategy in the past decade. Many Microfinance information platforms report promising figures in the race for global poverty reduction. Not only in developing countries Microfinance practices are now being implemented, also developed countries such as the US tend to benefit from Microfinance programs. It could be stated with absolute certainty that the number of MFI’s and the number of clients who are being served has increased enormously worldwide.
Though, the true question remains whether the impact of microfinance on society does reach as far as some may claim. Research on the impact of microfinance activities on society is quite extensive, but is seriously biased by methodological assessment implications. In the section below, the possible impact of microfinance on five key problems of the poor will be discussed briefly and will be followed by some general implications of its assessment.
2.2.1 Possible impact on society
First, there are indications that Microfinance has a positive impact on the reduction of poverty. Hossain (1988) found that the members from a microfinance program (Grameen members) had a 43 percent higher average household income than target non-participants in comparison villages. Khandker and Pitt (1998) attempted to statistically prove how microloans to woman from a program in Bangladesh, give an 18 percent return to income from borrowing. In an additional study by Khandker in 2005 it is found that overall poverty in the villages declined significantly more in program areas than in non-program areas, among program participants the poverty rates declined by even more, and more than half of this reduction is directly attributable to microfinance.
Second, studies indicate that poor people invest their income from microenterprises in their children’s education. Todd (1996) shows much higher levels of schooling among Grameen member’s children compared to non-members (81% compared to 54% for boys, and almost 100% compared to 60% for girls).
Fourth, higher and more stable incomes may generally lead to better nutrition, living conditions and preventive healthcare. MkNelly and Dunford (1998) find that program clients had better breast-feeding practices and their one-year-old children were healthier than non-clients children in terms of height-for-age and weight-for-age. Pitt, Khandker, Chowdhury and Millimet (2003) found substantial impact on children’s health, as measured by height and arm circumference, from women’s borrowing, but not from male borrowing.
Fifth, Hashemi, Schuler and Riley (1996) point out that each year of membership increased the likelihood of a female client being empowered by 16 percent. Also, they suggest positive spillover from microfinance is affecting the norms in communities.
2.2.3 Implications of impact assessment
While much research, as pointed out above, indicates that microfinance has a positive influence on poverty reduction, education, health etc., it remains important to question whether these conclusions are truly valid. It is claimed that the impact is both economical and social and there is a wide range of evidence that microfinance programs can increase incomes and lift families out of poverty. However, the impact of Microfinance on society can be very broad and complicated, and therefore also difficult to assess.
Hossain (1988) points out the limitations of impact assessments by pointing out the differences between participants in lending programs and comparison households regarding age, education, gender etc., which is prevalent among many microfinance impact evaluations and can influence outcomes noticeably. In addition, Karlan (2001) points out that there is a lack of validity of many microfinance impact studies because some designs may yield biased estimates of impact since MFI’s may have originally started to work with different types of clients than they currently serve, and because clients who chose to enroll earlier may differ from those who chose to wait and see before joining. Also, Armendariz and Murdoch (2005) address the importance of suitable impact evaluations and claim that no study yet has determined if the social and economical missions of microfinance are being achieved.
3.1 Commercialization of Microfinance
This section will provide evidence that the market for microfinance is commercializing and that additional capital is needed for further growth. Especially the international capital market is expected to increasingly enter the microfinance sector, while new and innovating products are being developed to reduce risks and boost profits.
3.1.1 The need for additional capital
The number of MFI’s and the number of clients served worldwide is increasing rapidly. Now, more than 10.000 MFI’s in more than 85 countries, serve over 100 million micro entrepreneurs. This global boost of MFI’s and clients calls for additional funding. Driven by increasing access to commercial funding sources, the volume of microfinance loans has risen sharply in recent years, from an estimated USD 4 bn in 2001 to approximately USD 25 bn in 2006 (Deutsche Bank, 2007). Still, only a fraction of today’s potential borrowers’ demand is met, while the microfinance sector still faces a $250 billion funding gap. A 2004 survey of over 144 MFI’s indicated that scarce donor funding has been the principal factor in limiting growth (Christen, Rosenberg and Jayadeva, 2004). This means a significant investment opportunity for capital markets.The current trends in microfinance will lead to a more financially and efficiency driven Microfinance environment, where many MFI’s tend to transform themselves from mission-driven, often inefficient NGO’s, into regulated financial institutions funded by private capital.
3.1.2 Classes of commercialization
experience a growth rate of their client base of 26 percent per year (Reddy, 2007), and in 2006 there were already about 30 MFI’s with a loan portfolio in excess of USD 100 million.5
3.1.3 Financial performance
The commercialization of microfinance is reflected in strong financial performance. Christen (2000) shows that in a competitive environment in Latin America, MFI’s are more profitable than its peers from other non-commercial regions, and in addition, are even more profitable than commercial banks in its own region. Littlefield and Holtman (2005) find that worldwide, the top MFI’s are nearly twice as profitable as the leading commercial banks in their local environment.
In addition, studies have indicated that MFI’s show low default rates, which tend to fall between 1% and 3% (Easton, 2005; Kraus and Walter 2008). In combination with impressive growth rates and strong returns, MFI’s are potentially compelling for foreign investors.
3.1.4 International capital markets
While domestic savings are still the main funding source for MFI’s, representing 41 percent of all assets in 2005 (Galema, Lensink and Spierdijk, 2008), many MFI’s turn to international capital markets as financing alternative.
As Swanson (2007) points out, most of the MFI’s are not deposit taking institutions, and are unlikely to become so, given the cost and complexity with regulations typically applied to institutions taking deposit from the public. Consequently, future funding for MFI’s is unlikely to be sourced mainly from deposits. It is also not assumed that other domestic sources in emerging countries will generate more than a fraction of the enormous potential capital demand. This is because capital markets in developing countries are thin and the key institutional investors are averse to or legally constrained from significant investment in microfinance.
The longer maturity of international capital financing will strengthen the financial structure of MFI’s, and will make them less exposed to external factors such as bank runs, currency risks, and macroeconomic crises. Especially the top tier is increasingly attracting the interest of foreign investors, as these MFI’s are usually profitable, have a more experienced management, and are considered to most effectively absorb the commercial funding.
5
3.1.5 Types of foreign investors
The landscape of foreign capital in Microfinance consists of two main types of investors, namely the public investors known as International, or Development, Financial Institutions (IFI’s or DFI’s), and the private investors, which includes individual and institutional investors. Foreign investment in Microfinance more than doubled from USD 1.7 bn in 2004 to around USD 4.4 bn in 2006. In this period, public investments from IFI’s increased from USD 1.1 bn to USD 2.4 bn and private investments increased from USD 0.6 bn to USD 2.0 bn (Deutsche Bank, 2007). In 2005, foreign capital provided 22 percent of funding for the top 100 MFI’s (Swanson, 2007).
The funding of foreign capital to MFI’s is, for approximately 50 percent of all investments, channelled through specialised Microfinance Investment Vehicles (MIV’s). The main investors in MVI’s are individual investors. IFI’s share in MVI funding has declined from 36 percent in 2005 to 30 percent in 2006, while institutional investors are catching up.
The number of MIV’s have increased rapidly and there are now over 80 MIV’s (CGAP, 2008). Total investment of these MVI’s have doubled between 2005 and 2006, reaching USD 2 bn in 2006.
MIV’s have different funding sources, due to a diverse investment approach regarding social returns and financial returns. While the more socially focused are particularly funded by development agencies and private donors, the more financially focused are mainly funded by commercial investors and social responsible investors.
One of the first MIV start-ups mainly backed by capital from development agencies and private donors was the equity fund ProFund. In 1995, it raised $23 million to finance Latin American MFI’s (Deutsche Bank, 2007).
3.1.6 CDO’s and Securitisations
As indicated by Galema, Lensink and Spierdijk (2008), foreign funding primarily took place via debt-structured finance, of which collateralized debt obligations (CDO’s) and securitizations were most important. While CDO’s and securitizations have a lot in common, the impact on MFI’s financial structure is different. Whereas the CDO’s relies on MFI’s ability to repay the loan, securitizations relies on the abilities of underlying borrowers to repay their loans.
In July 2004, the first CDO based on microfinance funds, BlueOrchard Microfinance Securities I (BOMSI), was created by BlueOrchard and partnered with Developing World Markets (DWM). The first closing raised USD 40 million via international capital markets and a subsequent closing held an additional USD 47 million. In both offerings, investors bought seven year notes, with a single repayment of principal at maturity. The proceeds of the investment of 90 investors were used to fund loans to 14 microfinance MFI’s in nine countries. Thus, BOMSI implied the securitization of loans to 14 microfinance institutions in 9 countries, which means much less diversification than typical CDO’s in developed markets where the asset pool may comprise many hundreds or thousands of loans (Swanson, 2007). The returns of the investors are the repayments of these 14 loans; there was no asset substitution or active management.
An innovating aspect of BOMSI’s funding was that it enclosed five levels of risk: senior debt, three subordinated debts, and equity. This made it possible to reach investors with different risk strategies. Investors are paid according to the ‘cash waterfall’, meaning that senior debt is paid first, then the other debt levels and equity investors receive the residual cash left. The US government development agency OPIC purchased the most senior tranche of securities, which encouraged investors who otherwise might have been unwilling to consider the transaction.
Also, BOMSI’s investors have not made loans to the fund, they have purchased securities in the form of bonds and equity interests.
On the first closing date of BOMSI, 4% of the invested capital was financed by private investors, while on the second closing date this already amounted to around 41% (Swanson, 2007). This is a clear indication of growing participation of commercial investors.
In 2006, BlueOrchard also developed a CDO, BleuOrchard Loans for Development (BOLD) and raised the amount of USD 99,1 million. The CDO securitised loans to 21 MFI’s in 13 countries, and 5 different currencies (Euro, US dollar, Columbian peso, Mexican peso, and Russian Rubble). The Dutch development bank FMO did the underwriting of the entire subordinated note class (Galema, Lensink and Spierdijk, 2008). In addition, BOLD 2 was launched in 2007 by BlueOrchard partnered with Morgan Stanley, and was the first microfinance capital transaction to be rated by S&P.6 It financed 20 MFI’s in 12 countries, and 6 different currencies. The entrance of Morgan Stanley, a typical traditional investment bank, is another sign that microfinance funding is gaining credibility as a capital markets practice. Very recently in January 2008, Citygroup privately placed a USD 165 million microfinance CDO. The portfolio will be actively managed with a reinvestment period lasting until April 2014. The first CDO to finance MFI’s in local currency was the Global Commercial Microfinance Consortium, and consists of more than 40 deals in 21 countries with a value of USD 80.6 (Reddy, 2007). The funding consisted of a group of institutional investors and development agencies and was managed by Deutsche Bank. A part of the capital structure contains mezzanine finance (between senior notes and equity), which provides unsecured debt that can be converted into equity in case of default.
Other than CDO’s, microfinance is also commercially funded by securitization of microfinance loans, in which private investors had a substantial share. The first securitization transaction was conducted by ICICI in India.
In May 2006, ProCredit Bank Bulgaria sold USD 48 million of its loan portfolio to institutional investors. The transaction was arranged by the Deutsche Bank, partial guaranteed by the European investment Bank and KfW, and rated BBB by Fitch Ratings.
In addition, ProCredit Serbia was structured in 2007 and issued a senior loan participation note worth EUR 125 million.
Also in 2006, the Bangladesh Rural Advancement Committee (BRAC), closed a program backed by micro loans, and primarily lends to the extremely poor. The deal involved about 3.3 million short loans with an average principal of only USD 95 (Galema, Lensink and Spierdijk, 2008). In total, 12 tranches of USD 15 million 6-month maturity notes, will be issued every six months. The deal was arranged by Citibank, RSA Capital, FMO and Kfw, and was AAA rated by a local credit rating agency.
6
3.1.7 Equity investments
While funding microfinance is typically debt based, MFI’s also started to raise capital from the international equity markets. Though, the factor slowing the growing attraction of private equity, is the small number of exits to date. Private equity investors tend to be more focused on capital gains upon sale of their stake than dividends payments as the principal component of its return. Nevertheless, in 2006 the first IPO in Africa was executed when Equity Bank in Kenya listed its shares on the Nairobi Stock Exchange.
In addition, the Mexican Compartamos conducted a successful IPO in April 2007 and transferred into a commercial bank.
Also, Accion Investments has invested USD 12.4 million in equity in five MFI’s (Reddy and rhyme, 2006).
3.1.8 Avoiding currency risks
Christensen (2007) points out six categories of risk when investing in microfinance: 1) Financial Risk, 2) Operational Risk, 3) Market Risk, 4) Regulatory Risk, 5) Country/Political Risk and 6) Foreign Exchange Risk. Christen argues that he first three risk types are controllable while the last are for a large part uncontrollable.
Foreign exchange risk is associated with the transfer of international capital to microfinance countries. The commercialization of microfinance will increasingly attract foreign investors, so the need to handle currency risks grows in importance (Crabb, 2004). Microfinance clients in developing countries rely on loans in dollars or Euros, although their earnings are in local currency. Therefore, they are vulnerable to currency devaluations and crises, which is more likely to occur in most microfinance countries than in developed countries.
transaction capacity of around €1.2 billion. By contributing to the fund, international investors can offer long-term finance in local currency to its customers in developing countries. The Dutch Ministry of Foreign Affairs has made TCX possible by granting a subordinated loan, which reduces the exposure of other development banks and private investors.
3.2 Diversification effects
Since growth in microfinance markets is particularly improved due to the additional funding from capital markets, diversification effects for investors may be of significant interest.
Most studies analyze whether performance of MFI’s is correlated with macroeconomic indicators to prove whether microfinance investments might be attractive for portfolio diversification. When it could be shown that MFI’s show low sensitivity to macro economy, investors might benefit from adding microfinance investments to its portfolio.
Several case studies support the theory of an uncorrelated relationship between microfinance and macro economical factors. The differences in risk exposure of MFI’s compared to other asset classes with macro economy is attributable to variation regarding ownership and governance structure as well as international exposure of clients. Also operational and financial leverage as well as product and lending methodologies could have an impact (Krauss and Walter, 2008) Galema, Lensink and Spierdijk (2008) point out that there are only three papers that quantitatively analyze resilience to macroeconomic shocks for a panel of MFI’s throughout the world: the studies of Ahlin and Lin (2006), Gonzalez (2007), and Krauss and Walter (2008). In addition, Galema Lensink and Spierdijk also quantitatively contribute the diversification debate with their innovating study. These studies all use different versions of the same dataset, the MixMarket dataset. The studies of Ahlin and Lin (2006) make MFI comparisons with domestic markets, while Krauss and Walter (2008) and Galema, Lensink and Spierdijk also include international markets.
environment is a significant determinant of MFI performance, though also by MFI-specific factors.
The study of Gonzalez (2007) contains a sample of 639 MFI’s in 88 countries for the period 1999-2006, and uses fixed and random effects panel regression to indicate whether changes in domestic GNI per capita affect MFI portfolio risk. Four indicators of MFI portfolio are used: Portfolio at Risk over 30 Days (PAR-30), Portfolio at Risk over 90 Days (PAR-90), Loan loss Rate (LLR), and Write-off Ratio (WOR). The results show that only for Portfolio at Risk over 30 Days there exists a statistically significant relation between changes in GNI per capita and portfolio risk of MFI’s. For the other indicators of MFI portfolio, no significant evidence for a relationship between Mfi asset quality and changes in GNI per capita was found. This suggests that microfinance portfolios have high resilience to economic shocks and therefore provides evidence that investors might have diversification benefits from investing in MFI’s.
Krauss and Walter (2008) use a sample of 325 MFI’s in 66 countries for the period 1998-2006, and use fixed-effects panel regressions to test the relation of MFI’s returns with global market and domestic market risk. MfI returns are measured by five financial indicators: Return on equity, Profit margin, change in total assets, change in gross loan portfolio and loan portfolio at risk. The global market risk is measured by the S&P 500, MSCI world and MSCI emerging markets indices, and the domestic market risk is measured by domestic GDP. They also use emerging market institutions (EMI’s) and emerging market commercial banks (EMCB’s) as a benchmark in the regressions to determine the relative market risk comporad to MFI’s. The results show that MFI’s show no correlation with global markets, while MFI’s are significantly correlated with the domestic market. Regarding EMI’s and EMCB’s, MFI’s seem to be more independent from global markets. Concluding, Microfinance investments may have useful portfolio diversification benefits for international investors, however, for domestic investors it seems not to provide significant portfolio diversification advantages.
3.3 Concerns of microfinance commercialization
With respect to the impact of microfinance on society and the ongoing commercialization, it is often argued that microfinance is only helping the less poor and not the poorest. While this was traditionally the aim of microfinance, many believe that commercialization has enabled a mission drift among MFI’s. Commercialization will induce competition among MFI’s, which will increase efficiency, but this may also reduce the scope for lending to the poor. Though, many critics of commercialization admit that commercial capital is needed for growth.
Nevertheless, commercial MFI’s are claimed to be focused too much on managing its business to benefit investors and not to benefit its borrowers. In the recent IPO of Compartamos, private Mexican investors, including the bank’s top executives, pocketed USD 150 million from the sale. Critics of Compartamos argue that they are making obscene profits off poor people and that they are in danger of ruining the rest of the industry by giving this example. Very recently, the CEO of the Grameen Foundation, addressed the need for a limitation on private benefit from microfinance (Counts, 2008). He agrees that microfinance executives who have received large capital gains from their work are controversial, and potentially undermine the public’s positive perspective of microfinance. The Grameen Bank gave the example by adopting the Bangladeshi government’s pay scale as a salary indication, which should keep executive salaries modest.
3.3.1 Different visions on microfinance
3.3.2 Efficiency versus outreach
Opponents of the financial system approach argue that the financially driven method leads to mission drift among MFI’s. They believe that commercial microfinance is in conflict with the traditional aim, because MFI’s tend to focus on financial efficiency rather than on outreach to the poor. Globally, this could induce a shift in microfinance activities to regions more profitable than others. A skewed distribution of microfinance funding will not be beneficial for many poor people.
Several studies have indicated there may be a trade-off between depth of outreach and achieving financial sustainability since unit transaction costs for smaller loans are high as compared to unit costs of larger loans (Hulme and Mosley,1996; Conning, 1999; Paxton and Cuevas, 2002; Lapenu and Zeller, 2002). The smaller loans are those loans to the poorest whereas larger loans are loans to the less poor. In a study by McIntosh, De Janvry and Sadoulet (2006) the negative side effect of efficiency driven MFI’s is also pointed out. It is shown that wealthier borrowers are likely to benefit from increasing competition among MFI’s, but that it leads to lower levels of welfare for the poorer borrowers. Traditional microfinance supporters claim that efficiency driven MFI’s are only focused on those individuals who have strong financial potential and those who have not will be left out. This will induce a shift from grouplending to individual-based lending, and means that MFI’s will be lending a small amount of larger loans instead of a large amount of smaller loans. This shift of lending method is outlined in a study by Christen (2000), which shows that in three of the most competitive microfinance markets in Latin America, a shift from group to individual lending has occurred.
strategy, period of entry into market, or natural evolution of the target group. Thus, regulated MFI’s with larger loan balances does not necessarily indicate microfinance mission drift.
In an extensive study by Cull et all. (2007), they contribute to the discussion on mission drift by providing systematical and empirical tests. It is tested whether more profitability is associated with lower depth of outreach to the poor, and whether there is an intentional move away from serving poor clients to wealthier clients to achieve higher financial sustainability. A sample of 124 MFI’s in 49 countries is used. The study explicitly makes a distinction between group, village, and individual-based lending. Average loan size and the share of loans extended to women are used as proxies for outreach to the poor. Smaller loan size is taken as an indication of better outreach to the poor, while a larger amount of woman borrowers should also indicate better outreach to the poor. The results suggest that individual-based MFI’s seem to perform better in terms of profitability, but the share of poor borrowers and female borrowers in the loan portfolio is lower than for group-based MFI’s. Also, it is suggested that individual-group-based MFI’s, especially if they grow larger, focus increasingly on wealthier clients, whereas this is less so for group-based MFI’s.
In a comparable study by Lensink, Meesters and Hermes (2008), in which the trade-off between efficiency and outreach is also being tested, they use a stochastic frontier analysis. The study tests for a trade-off between outreach to the poor and efficiency of MFI’s using a sample of more than 1300 observations. As proxies for outreach the average loan balance (ALB) and amount of woman borrowers are used. The study shows that outreach and efficiency of MFI’s are negatively correlated. More specifically, the results suggest that efficiency can only be reached if MFI’s focus less on the poor, and reduce the percentage of female borrowers.
3.3.3 High interest rates
It is commonly known in microfinance, that borrowers are exposed to high interest rates. For example, Fernando (2006) reports that MFIs in the Asia-Pacific region charge rates ranging from 30 to 70% a year. The average global rate is about 35 percent annually.7 Critics argue that these rates are inappropriate high and will lead to even more poverty since borrowers who cannot come up with the money will end up taking on more loans to repay the others, and so on. It is argued that high interest rates have the sole purpose of providing commercial investors with exorbitant returns, instead of having a goal of reaching as much as poor people as possible in a sustainable way.
Though, Fernando (2006) argues that microfinance interest rates are too often compared inappropriately. The rates are compared to those charged by commercial banks and excessively subsidized lending organizations. Commercial banks deal with large loans and thus have lower transaction costs per unit basis than MfI’s. Subsidized lending organizations charge lower rates than cost recovery levels, because losses are underwritten by large amounts of subsidies. Therefore these rates may not be compared. Microfinance interest rates are high since lending to the poor remains a high-cost operation. Furthermore, studies have shown evidence that the poor consider access to credit more important than the actual interest cost (GCAP, 2002). For example, a study in Chile, Colombia, and the Dominican Republic found that a 6 percent monthly interest rate represented less than 3.4 percent of a typical micro entrepreneur’s total costs. Also, studies covering India, Kenya, and the Philippines found that the average annual return on investments by micro businesses ranged from 117 to 847 percent. In addition, it may be noted that the MFI rates are still significantly lower than the 300 percent to 3,000 percent annual rates that many borrowers were previously paying to moneylenders and local credit card organizations in many developing economies.8 Fernando believes that charging prices high enough to cover costs is an essential practice for any business that intends to continue its operations beyond the short-term, thus also for MFI’s. He suggests that improved market competition and efficiency will reduce costs for MFI’s and will eventually enable them to charge lower interest rates.
7
Website Unitus; www.unitus.com
8
3.4 Social and Financial returns
The increasing commercialization has led supporters of the traditional aim of microfinance to conclude that there is additional need for investors with both financial as social preference. The market for socially responsible investments (SRI’s) has shown tremendous growth and microfinance should be an attractive sector within this market. The SRI market is significant, with over USD 4 trillion in assets, while emerging markets still only accounts for a scarce USD 5 billion (CGAP, 2008). It is expected that SRI’s in microfinance should increase in share, since it is shown that microfinance provides reasonable financial returns and is social responsible.
In Galema, Lensink and Spierdijk (2008), they point out a fund with both financial and social preference. The SNS developed the Institutional Investment Fund and is exclusively for investments by institutional investors in microfinance. The fund is mainly focused on investments in Tier 2 MFI’s.
4.1 Methodology
This paragraph aims to provide an overview of the methods used to determine any possible benefits from investing in microfinance regarding risk-return strategy.
4.1.1 Mean-variance spanning
In its mission to reduce global poverty, microfinance could be considered as a socially responsible investment. Though, investments in microfinance could also be beneficial to investors who aim at a diversified portfolio strategy. Therefore, it should be tested whether investing in microfinance can help reduce the risk of an existing portfolio. In order to analyze whether or not adding microfinance funds to a portfolio of risky assets is beneficial for investors, a mean-variance test will be used in this paper.
Where previous studies, such as Krauss and Walter (2008), were primarily focused on whether or not financial indicators of MFI’s show low correlation with domestic or global benchmarks, this study aims to be much more in line with standard mean-variance investment theory. As indicated by Galema, Lensink and Spierdijk (2008), the disadvantage of previous research is that it does not simultaneously considers risk-return considerations. The mean-variance spanning technique will test, by assuming that investors choose portfolios based on the mean-variance characteristics, whether adding microfinance assets to an existing portfolio of risky assets improves this portfolio in mean-variance terms. So, will this microfinance investment allow investors to reach a mean-variance efficient frontier with a higher mean and a lower variance. By assumption, portfolios that are on the efficient frontier are mean-variance efficient. This means these portfolios have the highest return for a given mean-variance of all possible portfolios. Put differently, these portfolios have the lowest variance for a given level of return of all possible portfolios.
The first study, in which the mean variance technique is proposed to test diversification benefits, was by Huberman and Kandel (1987). Other studies in which this test is used are Harvey (1995), Bekeart and Urias (1996) and De Roon et al. (2001), who deal with the benefits of international diversification.
microfinance assets. Under the null hypothesis of spanning the portfolio of benchmark assets has the same mean but a lower variance than the portfolio of Microfinance test assets. So the benchmark assets dominate the test assets, and consequently investors are no worse of by only investing in the benchmark assets. In that case investing in Microfinance offers no diversification advantages. Conversely, if the null hypothesis of spanning could be rejected, investing in microfinance does offer diversification advantages. More extensively, Appendix A presents a more formal way of describing the mean-variance spanning method, as pointed out by Huberman and Kandell (1987), and also described in Kan and Zhou (2001).
4.1.2 Short sales restriction
Galema, Lensink and Spierdijk (2008) show that investing in microfinance could give investors a better risk-return profile. Though, in a number of cases the results imply that investors should go short in microfinance, which is not possible. They suggest that further research should add short sales restrictions in the spanning test to avoid this possibility. This paper will therefore include a short sales restriction in its mean-variance test. This is extensively discussed by De Roon et al. (2001) and for further details it is referred to their work. In this paper the following pooled regression will be conducted:
i t i t i t R e R2, =α+β 1, + , t = 1…..T i = 1….M (1.0)
which follows directly from the mean variance spanning test (See Appendix), where at time t the test asset returns and benchmark returns are given by the K- and N- dimensional vectors
i t
R2, and R1t,i respectively. In this case the test is restricted to N = 1 and implies testing:
(
−1)
min ≤0 + β η α tK (2.0)(
−1)
max ≤0 + β η α K ta rejection of spanning. Equation (1.0) can be tested by calculating the test statistic suggested by Kodde and Palm (1986):
( )
η
min(
α
ˆJα
J)
'Var[
α
ˆJ( )
η
]
1(
α
ˆJα
J)
,ξ
= − − − s.t.α
ˆ ≤J 0 (3.0)where
α
ˆ and Jα
J are the restricted and unrestricted estimates of Jensen’s alpha, respectively. The tests statistic is asymptotically chi-square distributed, and its distribution is given in Kodde and Palm (1986). The probabilities are determined by using numerical simulation, as proposed by Gourieroux, Holly, and Montfort (1982)4.1.3 Wald Test
The restrictions of the mean-variance spanning can be tested by using a Wald test (Johnston and Dinardo, 1997), which is limited to a normally distributed sample. However, this problem of not normally distributed sample returns in the mean-variance test is solved by the fact that it is asymptotically chi-square distributed with 2N degrees of freedom.
In addition, cross-sectional dimensions will have to be used for the spanning test to deal with the small amount of time series observations which the data sample provides (see part on data).This means that the mean-variance spanning test will be conducted in a panel setting as given in equation (1.).
5.1 Data
The dataset used in this paper contains annual data on MFI’s for the period 1997 to 2007, and is publicly available from MixMarket (www.mixmarket.org). The MixMarket is a global, web-based, microfinance information platform and seeks to develop a transparent information market to link MFI’s worldwide with investors and Donors and promote greater investment and information flows. All numbers are converted to US dollars at contemporaneous exchange rates. Participation of MFI’s in the MIX database is voluntarily, but data submission is closely monitored. MFI’s have to enclose substantiating documentation, such as audited financial statements and annual reports, that helps external analysts and researchers to understand the operations. As reported in Gonzalez (2007), MFI’s should have the availability of adequate information systems, which is driven by the potential exposure to investors and donors looking for investment opportunities among MFI’s. Therefore, the Mix database can be viewed as a random sample of the best MFI’s in the world, but definitely not a random sample of all MFI’s. Consequently, the data set should present the potential investment environment for microfinance investors looking for diversification benefits.
As pointed out in Walter and Krauss (2008), the standard approach to analyze the risk of an assets class is to use historical market returns. However, since MFI’s are not actively traded there is no market-to-market valuation, and the standard approach is not feasible. Also the fundamental approach of calculating alpha and beta is not possible since microfinance is an emerging asset class with no peer group of listed organizations. Walter and Krauss (2008) point out that using emerging market commercial banks as a listed peer group is not a valid approach, because microfinance behaves very differently in terms of risk and thus cannot be regarded as the same asset class.
The data, as reported by the MixMarket, is not adjusted for subsidies and this dilutes the real market risk of MFI’s. Walter and Kraus (2008) point out that, from an investor perspective MFI subsidies can be compared to a too-big-to fail (TBTF) support for commercial banks. This term is used to describe how large financial institutions are likely to be bailed out by national supervisory institutions in times of severe financial distress, given that their collapse could endanger the stability of the financial system. Though, subsidies constitutes an investment risk, since the frequency and size are neither predictable nor enforceable, which in this study could not be accounted for.
For the benchmark, the returns from the MSCI world (MSCI W) and MSCI emerging markets (MSCI E) are used, which have also been used by Walter and Krause (2008) and Galema, lensink and Spierdijk (2008).
In addition, two bonds indices are used as benchmarks; the JP Morgan Global Broad index (JPMG) and the JP Morgan EMBI Global Composite (JPME) Most investments in microfinance are typical debt-based and not equity-based. Therefore, a comparison based on the addition of microfinance assets to an existing portfolio of debt may be better than to a portfolio of equity.
5.2 Descriptive analysis of the data
The main goal of this paper is to find potential investment opportunities in microfinance with both a social return and a financial return. Social returns for investors are those investments that add some social responsible input in its environment. This could mean that it helps to reduce the poverty or increase the chance of education, better healthcare etc., for the people involved. In microfinance research, the social aspect is often specified by outreach to the poor, as also discussed in paragraph 3. To determine this outreach to the poor the existing literature typically uses two key indicators. These indicators, also used by Cull et al. (2007) as pointed out above, are Average Loan Balance (ALB) and percentage of female borrowers (Woman). ALB is considered as an indicator for lending to the poor. The higher ALB, the less focus on the poor by a MFI. For Woman, it is considered that a higher percentage means more focus on the poor since woman tend to be among the poorest. So evidence for better outreach to the poor is associated with lower ALB and higher Woman. The next section tries to point out MFI’s with good outreach indications regarding ALB and Woman. Several different approaches to distinguish between groups of MFI’s are being used to determine potential investment opportunities.
5.2.1 The complete dataset
In Table 1, a description of the distribution of MFI’s over the years, from the complete dataset, is provided. In 1997 the dataset contains only 43 observations while the amount of observations increase rapidly during the years. For 2007 the amount of observations is also small since most MFI have not yet completed the financial statements. The last two columns show that for most MFI’s only for 2, 3 or 4 years of observations is available. For six MFI’s there are observations over the entire period of 11 years. In total, the dataset contains 4355 observations from 997 different MFI’s.
Table 1. Description of the panel
Year Observations ALB Woman Nr of years available Nr of MFI's
mean mean 1997 43 952 0,7615 1 87 1998 81 615 0,6819 2 221 1999 117 499 0,7064 3 210 2000 172 622 0,6633 4 204 2001 268 574 0,6309 5 97 2002 417 584 0,6422 6 59 2003 630 704 0,6420 7 43 2004 748 824 0,6460 8 28 2005 877 821 0,6589 9 27 2006 838 956 0,6686 10 15 2007 146 478 0,8287 11 6 Total 4337 871 0,6600 Total 997
5.2.2 Criteria for group distinguishing
The dataset allows to distinguish between groups of MFI’s based on four different categorizations; 1)Loantype; 2)Legal status; 3)Regulation; 4)Regions. The information to distinguish between these categories is obtained from the MIX Microbanking Bulletin.
The following loantype classification is made: 1) Individual (Individual lending); 2) Mixed (the MFI uses individual lending and solidarity lending or individual lending and village bank lending); 3) Solidarity (Solidarity group lending); 4) Village (the MFI uses village group lending). As also discussed in paragraph 3, different lending techniques may result in different approaches towards the stragey of MFI’s in terms of commercialization. Cull et al. (2007) point out how MFI’s with different lending techniques show different patterns of its social mission, target clients and location. Their study shows that grouplending is less profitable than individual lending and serves the poorest clients. The average loan balance is significantly smaller for the grouplending MFI’s than for individual lending MFI’s. This means it has better outreach to the poor. Also the clients from grouplenders are more likely to be women than for individual lenders, which also indicates better outreach to the poor. It is therefore expected that in this paper, grouplending MFI’s will indicate better outreach than individual lending MFI’s. The group of MFI’s with mixed lending techniques will not be examined very closely in this paper, since it is not known in which proportion both lending techniques are practiced for each MFI.
different locations to obtain the desired returns. Some regions may be further in microfinance development than other regions, and may therefore be more commercially driven. Also, cultural differences may have influence on the success of microfinance programs.
The MFI’s from the dataset can also be divided in groups according to a different legal status. The following statuses are being distinguished; 1) Bank; 2) Cooperation; 3) Non-Bank financial; 4) Non-profit (NGO’s); 5) Rural Bank; 6) Other. A different legal status could mean a diverse approach for MFI’s regarding credit and savings policy, management operations, strategy implementation and so on. These approaches could all have significant impact on the ability to reach the poor or to strive for financial efficiency. It may be expected that, for instance, MFI’s with a Non-profit status have a dedicated policy to reach the poor. On the other hand, MFI’s with a Bank status are expected to be more commercial. Therefore, the MFI’s with a Non-profit status are expected to score better on outreach than Banks.
The MFI’s from the dataset are distinguished based on a different state of regulation; 1)Regulated; 2)Non-regulated. In paragraph 3, it is explained how regulated MFI’s have more chance to attract commercial capital and to collect savings from its clients, and therefore tend to operate on a more commercial basis than non-regulated MFI’s. For this reason, it may be expected that non-regulated MFI’s will have better outreach than regulated MFI’s.
An overview of the outreach indicators and the financial returns are presented in tables 2, 3, 4 and 5. The columns present the mean and variance for ALB and Woman, the ROE and ROA, and the amount of observations. The calculations for ALB and Woman is based on all the available observations per focus group, however, the amount of observations as indicated in the last column may differ slightly from this number due to missing data. The observations as indicated respond to the amount used for ROE and ROA calculations, and is generally a little bit less than for calculations of ALB and Woman.
Table 2. Outreach and financial indicators for different loantypes
Loantype ALB Woman ROE ROA Obs.
Mean Var. Mean Var. Mean Var. Mean Var.
Individual 1281 1681974 0,4697 0,0573 0,1884 0,2334 0,0431 0,0021 345 Mixed 635 782167 0,6615 0,0667 0,0990 0,2095 0,0279 0,0051 483 Solidarity 138 12078 0,7722 0,0573 0,0064 0,4294 0,0119 0,0173 151 Village 168 44940 0,9201 0,0188 0,0635 13,2417 0,0179 0,0169 147
Table 3. Outreach and financial indicators for different legal status
Legal status ALB Woman ROE ROA Obs.
Mean Var. Mean Var. Mean Var. Mean Var.
Bank 2180 7824857 0,5229 0,0715 0,1942 0,5161 0,0270 0,0048 288 Co-op. 1217 3813735 0,5158 0,0647 0,3114 36,7270 0,0171 0,0073 488 Non-Bank Financial 1018 5010513 0,5876 0,0688 0,0346 0,5831 0,0112 0,0132 933 Non-profit 423 1392591 0,7793 0,0586 0,0514 4,9589 -0,0101 0,0377 1428 Rural 707 486452 0,5306 0,1198 0,1943 0,0177 0,0330 0,0007 151 Other 738 815128 0,6602 0,0840 -0,1089 3,3595 -0,0739 0,1013 88
Table 4. Outreach and financial indicators for different state of regulation
Regulation ALB Woman ROE ROA Obs.
Mean Var. Mean Var. Mean Var. Mean Var.
Regulated 1025 4106462 0,6078 0,0806 0,1321 10,2746 0,0094 0,0178 1990
Non-regulated 659 2526585 0,7256 0,0678 0,0514 4,05919 -0,0058 0,0329 1386
Table 5. Outreach and financial indicators for different regions
Region ALB Woman ROE ROA Obs.
Mean Var. Mean Var. Mean Var. Mean Var.
Table 3 indicates that only Non-profit MFI’s (NGOs), score well on outreach regarding both ALB and Woman. For ALB and Woman these MFI’s show significant better results than for the entire dataset. This is corresponding with the expectations that Non-profit organizations are less commercially driven and more outreach focused. The MFI’s from this group will be tested for spanning. Table 4 shows that Non-regulated MFI’s reach better to the poor and the female borrowers than Regulated MFI’s, which is also in line with the expectations. The results from Table 5 point out that particularly South Asia has good outreach in terms of both ALB and Woman. These MFI’s will be tested for spanning to determine if this is a potential investment group in terms of social and financial return. In table 5 it is also noticed that the MFI’s from Eastern Europe score significantly less good on both outreach indicators.
Overall, the results from these outreach indications are in line with the expectations, since grouplending scores better than individual, regulated better than regulated and non-profit better than the rest of the legal statuses. Also, it is remarkable that in general the mean and variance for ROE tend to give more extreme values than for ROA. One possible explanation could be that for those MFI’s with more divergence between ROE and ROA, this is influenced by the Debt-to-equity ratio. Where MFI’s with a higher ratio tend to inflate its ROE and deflate its ROA.
with a beneficial financial profile. The same distinguishing as for individual lenders will be made, based on different regions, legal status and state of regulation.
Table 6. Outreach and financial indicators for individual versus group lenders
Loantype ALB Woman ROE ROA Obs.
Mean Var. Mean Var. Mean Var. Mean Var.
Individual 1281 1681974 0,4697 0,0573 0,1884 0,2334 0,0431 0,0021 345
Group 152 19054 0,8448 0,0438 0,0346 6,7271 0,0148 0,0171 298
In table 7A and B, a description of outreach variables and financial return variables are given for different MFI groups within group lending and individual lending. In table 7A it is directly clear that the majority of groups score high on ALB, and thus low on outreach. Only Africa, the Middle East and South Asia tend to have a low ALB, but the results from the Middle East and South Asia are biased due to the small amount of observations. For the outreach indicator Woman, not any group has the same or higher average than the average of the entire dataset. Therefore, it may be concluded that within the individual lending MFI’s, it is only attractive to test spanning for the region Africa.
In table 7.B it is shown that for grouplenders every subgroup has a lower than total mean ALB. In addition, regarding female borrowers, every group has a higher than total mean percentage. Therefore, the spanning test should be conducted for every subgroup within the group lenders to seek for positive risk-return profiles.
Further remarks could be given by the results from table 7.A, where non-regulated MFI’s indeed score better on outreach, for both ALB and Woman, compared to regulated MFI’s. This is more or less the same in table 7.B, but less significant. Also Non-profit scores best on outreach in terms of ALB and Woman compared to the other legal status. However, this is not confirmed by the results for ALB of individual lenders, but is does for female borrowers. In table 7.B, for legal status the data only allows to distinguish between Non-bank financial and Non-profit organisations since the other subgroups lack observations.
Table 7.A Outreach and financial indicators for subdivided individual lenders
Subgroups ALB Woman ROE ROA Obs.
Mean Var. Mean Var. Mean Var. Mean Var.
Africa 785 628639 0,3388 0,0271 0,1710 0,0191 0,0243 0,0005 32 East Asia 1006 919860 0,3206 0,1264 0,3089 0,4636 0,0575 0,0009 49 East Europe 2223 2954686 0,4340 0,0434 0,1542 0,0128 0,0527 0,0018 52 Latin America 1228 1500892 0,5558 0,0314 0,1940 0,2352 0,0432 0,0023 204 Middle East 692 33508 0,2230 0,0004 -0,0477 0,0038 -0,0140 0,0003 6 South Asia 512 509 0,0890 0,0028 -0,9868 1,9081 -0,0721 0,0082 3 Regulated 1512 2202917 0,4479 0,0406 0,1920 0,3607 0,0343 0,0015 217 Non-regulated 940 685157 0,5067 0,0841 0,1806 0,0188 0,0575 0,0028 129 Bank 982 730602 0,3988 0,0380 0,3944 0,9997 0,0338 0,0016 56 Co-op. 1163 1075832 0,3660 0,0840 0,0690 0,0193 0,0147 0,0008 40 Non-Bank Financial 1446 1203637 0,4804 0,0139 0,1655 0,0295 0,0372 0,0016 101 Non-profit 1432 3297328 0,5926 0,0550 0,1115 0,1635 0,0615 0,0032 106 Rural 1204 1005112 0,2370 0,0907 0,2528 0,0076 0,0543 0,0009 37
Table 7.B Outreach and financial indicators for subdivided group lenders
Subgroups ALB Woman ROE ROA Obs.
Mean Var. Mean Var. Mean Var. Mean Var.
6. Results
In the previous paragraph the different groups of MFI’s with positive outreach indications have been pointed out. These are the potential investment opportunities from a social point of view. The goal is to find spanning test results among these groups of MFI’s for which the outcome rejects the hypothesis (the hypothesis is that adding an additional microfinance asset to a existing portfolio of risky assets will have no positive value in terms of risk-return profile). This indicates that investing in the associated MFI’s could be beneficial in terms of risk-return. These different groups are tested for spanning with and without short sales restrictions and the results are presented in this paragraph. All tables with the results are presented in Appendix B.
The mean variance spanning test with and without short sales restriction is conducted for groups of MFI’s with different loantypes, as presented in table 8 (Appendix B). The spanning test is based on the values of Alpha and Beta. The spanning test is rejected if the p-value is below 0.05. The results for the spanning test with short sales restriction directly indicate a difference between individual lenders and group lenders, with regard to risk-return. For the spanning test without short sales restriction, the results are not completely in line. Without short sales restriction, the test is more often rejected and thus indicate a positive risk-return profile more often for group lenders. With the short sales restriction, both solidarity and village lenders cannot be rejected for the spanning test in all four cases. For the ROE and the ROA the tests shows that for international as for domestic investors and for bond based investors, the group lending MFI’s provide no risk-return benefits. The individual lenders directly indicate financial benefits, which is all in line with the expectations. In Table 9 solidarity and village lenders are merged into one dataset, grouplending, and is additionally testes for mean variance. For tests with short sales restriction, the merged group confirms the results from table 8, in which both groups are also not rejected in all cases. Without the restriction, the test with the merged group is again not in line with tests with restriction.
