Microfinance Institutions: Exploring the possible diversification benefits from an investor s perspective

Transcription

1 TILBURG UNIVERSITY Microfinance Institutions: Exploring the possible diversification benefits from an investor s perspective Master Thesis Tilburg School of Economics and Management Finance Department Natasha Faustin ANR: Supervisor: Dr. J.C. Rodriguez December 2014

2 Abstract This thesis aims to take a comprehensive look at microfinance institutions and explore the possibility of considering these institutions as an asset class for investors by evaluating their performance relative to global benchmarks. Their isolation from the formal economy along with their unique lending practices in comparison to traditional banks and credit facilities may suggest that these institutions are capable of being resistant to macroeconomic developments and may therefore be able to provide investors with some diversification benefits. The first part of the analysis considers the performance of individual microfinance institutions while the second part of the analysis takes a look at commercialized microfinance institutions. First, a fixed-effects regression analysis is conducted to explore the relationship between the performance measures of individual microfinance institutions in comparison to those of global benchmarks. In the second part of the analysis, the relationship between commercialized microfinance institutions and different global benchmarks during the 2008 financial crisis is explored. In addition, the Markowitz portfolio model is used to observe whether investments into commercialized microfinance institutions can increase the efficiency of an already diversified portfolio. The results from the first part of the analysis suggest that microfinance institutions exhibit, for the most part, a low and insignificant relationship with macroeconomic movements. When it comes to observing commercialized microfinance institutions during the 2008 financial crisis, an insignificant relationship was observed when regressed against the global benchmarks, specifically when considering fixed-income investments into microfinance. For the portfolio efficiency analysis, the results suggest that efficiency can be added to a global portfolio by investing into commercialized microfinance institutions. 2

3 Table of Contents Abstract Introduction Characteristics of MFIs History and Development Lending practices Additional characteristics MFIs and their impact on poverty alleviation Commercialization of MFIs Socially Responsible Investments Microfinance Investment Vehicles Investor profile Commercialization of individual MFIs Funding structure Examples of commercialization The risk profile of microfinance Potential benefits of commercialization Current state of literature Impact of macroeconomic environment on MFIs Analysis of publicly traded MFIs Methodology Analyzing individual microfinance institutions Analyzing current commercialized MFIs Data Analyzing individual microfinance institutions Analyzing current commercialized MFIs Empirical Results Analyzing individual microfinance institutions Analyzing current commercialized MFIs Conclusion Future Implications References

4 1. Introduction Microfinance is a powerful development tool one that can reach the poor, raise their living standards, create jobs and contribute to economic growth. The term microfinance describes the provision of banking services by poverty-focused financial institutions to the working poor in developing countries which are not served by mainstream financial institutions (Dieckmann 2008). Since the 1970 s, microfinance institutions (MFIs) have made efforts to provide microloans to the poor and are considered a possible tool for alleviating poverty in developing countries. Several characteristics and lending practices have been implemented by these institutions in order to adapt to the needs of their clients, who often live in rural areas and have no collateral. These characteristics and lending practices have allowed MFIs to differentiate themselves from most commercial banks. Microfinance institutions recognize that the working poor can act in an entrepreneurial matter and are, in principle, creditworthy. For these microborrowers, MFIs are often times the only alternative to paying excessive interest rates charged by unofficial moneylenders in developing countries. With the increase of international financial integration across advanced economies, emerging markets and developing countries, it is becoming more important to be able to identify investment opportunities that show low sensitivity to macroeconomic developments. The unique characteristics of MFIs along with their isolation from the formal economy thus merits further research into the potential that lies ahead for the continuing commercialization of these institutions. Due to their low integration with the formal economy, it can be argued that microfinance institutions should in principle exhibit low correlation to macroeconomic movements. Therefore, if microfinance institutions seem to display low volatility to global movements over an extended period of time, there may be some diversification benefits to adding these investments to an already diversified portfolio. Up to now, the empirical research conducted on the commercialization of MFIs and their relationship with the global economy has been very limited. While a few have conducted research on the relationship between microfinance performance measures and the domestic economic movements, only two research papers were found on the relationship between MFIs and global economic movements. In their paper, Kraus and Walter (2009) used fixed-effects regressions to evaluate microfinance performance relative to global benchmarks during the period of and found no statistically significant relationship with global market movements. A few years later, Galema, Lensink and Spierdijk (2011) use mean-variance spanning tests with short sale constraints to investigate whether adding microfinance funds to a portfolio of risky international assets yield diversification gains. Their study covered the period of and found that investments into MFIs can be a valuable addition to a broad portfolio of stocks and bonds. 4

5 The research conducted in this thesis will be performed on two different levels. First, the thesis will expand upon the previous work by combining aspects from both Kraus and Walter s (2009) and Galema, Lensink and Spiedijk s (2011) analyses and applying them towards an extended period of time. This analysis will help provide a better understanding of the impact that macroeconomic developments may have on the performance of these institutions. By looking at individual institutions, we are able to gather a greater amount of data with the goal of analyzing whether future commercialization of MFIs will prove to be beneficial for investors who are looking for an asset class that will add some diversification benefits due to a high resiliency with the market. In addition to evaluating the performance of the individual institutions, the second level of analysis will be conducted on the performance of MFIs that are already integrated into the bond and equity markets. The aim of the second part of this analysis is to observe any benefits that may exist from adding these commercialized institutions into a global portfolio. First, commercialized MFIs will be observed during the 2008 financial crisis to assess their relationship with global benchmarks during times of financial distress. Furthermore, an additional analysis will take place by applying Markowitz mean-standard deviation model and observing the efficiency of the portfolio both before and after the implementation of investments that are focused on MFIs. This type of analysis has not yet been conducted by academic researchers and can aid in providing clarity on the benefits of adding microfinance investments into a global portfolio. Overall, this thesis aims to provide a comprehensive analysis of both the future potential for further commercialization of microfinance as well as the present performance of commercialized microfinance investments by examining the individual institutions as well as the funds that currently participate in microfinance investments. This research is of importance as it allows for the possibility of bridging the gap between microfinance institutions and socially responsible investors who want to make a social impact through their investments. The findings in this analysis can aid in the improvement of the commercialization of MFIs which allows for greater access for interested investors. Furthermore, the increase in capital funding through commercialization can allow MFIs to increase their outreach to the poor, seeing how many people still remain unbanked to this day. This thesis is organized as follows. The characteristics of MFIs along with their history and development over time will be discussed in section 2. In the third section, a closer look will be taken at the true impact that MFIs have on poverty alleviation. Section 4 will discuss the commercialization of MFIs, while sections 5 through 7 will discuss the current state of literature, methodology and data, respectively. The empirical results will be shown in section 8. Lastly, the conclusion and future implication will be presented in section 9 and 10, respectively. 5

6 2. Characteristics of MFIs 2.1 History and Development Microfinance institutions started developing around the 1970 s by means of experimental programs in Bangladesh, Brazil and other countries. These programs were implemented with the purpose of providing loans for income generating activities by targeting very poor (often female) borrowers. Some of the early pioneers of microfinance institutions include SEWA Bank, ACCION International and Grameen Bank. SEWA Bank started off as an organization formed by a group of self employed women in The organization, known as Self Employed Women s Association was registered as a trade union in Gujarat, with the main objective of "strengthening its members' bargaining power to improve income, employment and access to social security"( Members of this organization later founded an urban cooperative bank known as SEWA Bank, which served the purpose of providing banking services to the poor, illiterate and to self-employed women. In 1973, ACCION International began issuing small loans in Recife, Brazil and eventually managed to create a network of financial institutions with branches in 22 countries. The Grameen Bank was founded by professor Muhammad Yunus in 1976 and has managed to become one of the most well-known and successful microfinance institutions in the world. The bank gained worldwide recognition after winning the Nobel Peace Prize in 2006 for their efforts to create and improve economic and social development. Over some three decades, MFIs have experienced significant growth with around 10,000 MFIs evolving worldwide. From 2004 to 2008, microfinance enjoyed unprecedented growth in emerging markets, with annual average asset growth of 39%, accumulating total assets of over US $60 billion by December 2008 (Chen, Rasmussen and Reille 2010). In addition, the total number of borrowers served by MFIs grew by 21 percent per year on average during the period of 2003 to The popularity of microfinance has steadily been increasing over time and has reached a global audience. The year 2005 was proclaimed as the year of microcredit by the United Nations as an attempt to globally promote the benefits and potential of the microfinance industry. Unlike top-down development initiatives such as debt forgiveness or international aid, microfinance has managed to differentiate itself through its bottom-up approach. These institutions emerge locally and allow borrowers to improve their situation by means of their own initiative instead of relying on external development strategies. 2.2 Lending practices MFIs have distinguished themselves from most commercial banks by adapting different lending practices in order to efficiently and effectively provide financial assistance to the poor. These practices include the implementation of a group lending system where lenders are required to guarantee each other s loan payments, which allows poor borrowers to finance small business 6

7 ventures without the use of collateral. In his paper, Gonzales (2007) notes how some theoretical models have been able to support the position of solidarity groups being less risky than individual loans. The argument behind this position lies in the group s ability to provide incentives for peers to screen, monitor and enforce each other s loans, especially when one member of the group has suffered a shock. Thus, the effect of screening can allow for the reduction of adverse selection while the effect of monitoring reduces moral hazard, both of which reduce administrative costs. While Banarjee and Duflo (2010) recognize the possibility of group liability being effective in contexts where large temptation to default is present, they also point to the possible cost that might be imposed by implementing a group liability structure. Since group participants do not benefit from the upside of any risk investment but are on the hook for the downside, excessive risk aversion is imposed by the group on the members. This may hinder the efficient allocation of financial resources to the appropriate entrepreneurial projects. Others have also argued that the group liability model is only effective in certain situations such as in rural settings where social control is higher. Although the majority of MFIs implement the group lending system, there are still a few MFIs that prefer to lend to individuals without the shared liability aspect. Another unique lending practice includes dynamic incentives, where lenders are rewarded with a bigger loan upon each successful repayment, which encourages repayment and helps rectify the issue of moral hazard. The repayment frequency of MFIs is also another lending practice worth noting. Whereas banks often require a monthly repayment frequency, most MFIs implement shorter frequencies of weekly or biweekly repayments. This allows MFIs to carefully monitor their portfolio quality and to adjust their lending practices and liquidity if necessary. There s also the belief that frequent meetings can harness and build social capital by encouraging regular social contact (Feigenberg, Field and Pande 2010). In a field experiment conducted by Feigenberg, Field and Pande (2010), direct evidence was found that more frequent interaction increases trust and economic cooperation among clients. The results of their experiment showed that members who met every week were more likely to give each other lottery tickets when given the option to do so. This act implies that group members trust that the proceeds of the lottery winnings will be shared if they give others some of their tickets. Overall, the different lending practices that have been implemented by MFIs have allowed these institutions to effectively adapt to the informalities involved with providing credit to the relatively poor. The efficacy of these methods can be witnessed by the relatively high repayment rates that MFIs have been able to achieve despite the lack of collateral from the borrowers. On a global level, microfinance clients have been able to produce an excellent track record for repayment, with repayment rates averaging at about 97%. 7

8 2.3 Additional characteristics Aside from their lending practices, MFIs possess additional characteristics that differentiate them from typical commercial banks. As an example, MFIs tend to have higher administrative costs in comparison to commercial banks. However, their high annual rates of interest have been able to compensate for these high administrative costs. In general, MFIs interest rates can range from 18 to 60 percent, depending on the conditions in each microfinance institution s service area. Despite these high interest rates, the high marginal productivity of capital expenditures undertaken by microfinance borrowers appears to justify the materially higher interest rates (Krauss and Walter 2009). MFIs are also able to benefit from cost advantage by allowing the group leaders to be in charge of collecting the individual amounts for the loan agent. Since all loans from a particular group get repaid on the same schedule, loan officers are able to collect payments from a large number of borrowers every day. This technology of loan collection plays a big role in reducing the administrative costs per loan. (Banarjee and Duflo 2010). The customer base of microfinance institutions is also a bit different than those of commercial banks, with most MFIs focusing more on a specific type of customer. Microfinance customers generally consist of small entrepreneurs who provide essential services and products to their community and who tend to operate in the informal economy. These customers seek credit in order to set up or manage their own business, which ranges from making small handcrafts to running neighborhood shops. Examples of micro-borrowers include street vendors, trades people, small farmers, and service providers. Another characteristic of MFIs is that women make up the vast majority of borrowers, especially in Asia, where shares of female debtors are as high as 99%. The reason for targeting female borrowers stems from the belief that women are more reliable as debtors due to their stronger social and family ties, which in turn causes them to invest in safer projects and lower their chances of default. (Dieckmann, Speyer, Ebling and Walter 2007) 3. MFIs and their impact on poverty alleviation According to the 2012 World Development Indicators report, it is estimated that around 1.28 billion people are living below the extreme poverty line of less than $1.25 a day, with an additional 1.18 billion living on $1.25 to $2 a day (World Bank 2012). Clearly, poverty remains a serious social and economic issue for many developing countries around the world. For socially responsible investors, it is important to understand the possible role that MFIs can play in alleviating poverty in developing countries, since these investors are not only looking for a financial return, but are also interested in generating a social return as well. By providing credit to the poor, MFIs have been considered as a possible means to help resolve the issue of poverty. 8

9 Thanks to these institutions, many who have no collateral are still able to obtain loans, allowing them to use the funds provided to start-up and run a successful microenterprise. Poverty is a complex issue that requires different tools and approaches in order to be addressed properly. Although MFIs have been able to help many around the world, there is no true consensus on the degree to which these institutions can help eradicate the issue of poverty in developing countries. In their paper, Morduch and Haley (2002) provide a comprehensive review of the empirical research available that focuses on the impact that microfinance institutions have on the poor. Several outcomes of the research conducted suggests that the poor can benefit from microfinance from both an economic and social well-being point-of-view, and that this can be done without jeopardizing the financial sustainability of the microfinance institution. The review of the literature provided by Murdoch and Haley (2002) contains evidence substantiating a beneficial effect on increases in income along with reductions in vulnerability for those who participate in microcredit lending. As an example, a significant increase in household income for families with access to credit was observed during case studies conducted in Asia and the Pacific. These case studies, conducted by Remenyi and Quinones (2000), showed a 12.9% rise in annual average income from borrowers while only 3% rise was reported from non-borrowers in India. In Sri-Lanka, a 15.6% increase in income was reported from borrowers while a 9% rise in income was observed among non-borrowers. Despite its potential benefits for the poor, it is recognized that microfinance may not be for everyone given the fact that entrepreneurial skills and ability are necessary to run a successful microenterprise and not all potential customers are equally able to take on debt. While these points will be true across all levels of poverty, it is assumed that they will have a greater effect on the very poorest. (Murdoch and Haley 2002) In another paper, Graham Wright (1999) provides a review of the research conducted to date on the effects of MFIs on the income of the borrowers. Wright provides information on the first detailed research conducted on the economic impact of a microcredit programme, which was conducted by Mahabub Hossain on Grameen bank in The most direct outcome that Hossain witnessed in his research was the effect that the Grameen Bank had on the accumulation of capital by the poor. He noted a threefold increase in the amount of working capital employed by members enterprises within a period of 27 months. Although MFIs seem to improve the income of those with access to credit, Wright notes the importance of recognizing the difference between increasing income and reducing poverty. Poverty in itself is not exclusively determined by an inadequate income level, but also entails the inability to sustain a specified level of well-being. The context to which poverty is defined can also determine whether a certain strategy is deemed efficient. When considering poverty- 9

10 reduction as a process of moving households from below poverty line to above poverty line, then strategies involving credit provision through self-employment may be considered as useful. When using a broader, less linear view of poverty which sees income levels fluctuating below and above the poverty line, then strategies that focus on providing savings options and emergency consumption loans may prove to be effective. Despite the conflicting views on the degree of impact that MFIs have on the poor, it seems that these institutions allow some households to sacrifice instantaneous utility to finance lumpy purchases, either for their home or in order to establish or expand a business. (Banarjee et al 2013) 4. Commercialization of MFIs 4.1 Socially Responsible Investments Socially responsible investments (SRIs) have been increasingly gaining the attention of investors during the years. According to the European investment forum of 2007, U.S. investments in SRIs more than quadrupled from 639 billion dollars to 2.7 trillion dollars between 1995 and As for European investors, this number doubled from 501 billion Euros to 1 trillion Euros during the period of Socially responsible investors are looking for investment opportunities that extend beyond the typical goal of maximizing shareholder value. For these investors, stakeholder value becomes more of a priority and attention is paid to how their investments affect the social and/or environmental well-being of all stakeholders involved. With this in mind, MFIs may provide the perfect investment opportunity for those looking for a more social return. In essence, microfinance investments have the opportunity to provide investors with potential benefits of dual nature: First, they allow investors to adopt a social investment strategy aimed towards poverty alleviation and social development in developing countries. Second, they have the potential to simultaneously offer an attractive risk-return profile that is characterized by largely stable financial returns, low credit default rates and low correlation to the mainstream financial assets. (Dieckmann et al 2007) 4.2 Microfinance Investment Vehicles Public investments made into microfinance occur mainly through microfinance investment vehicles (MIVs) and through listed equity of MFIs. MIVs are independent investment entities with more than 50% of their non-cash assets invested in microfinance. MIVs play a key role in microfinance capital funding and are considered a convenient tool to invest collectively in a wide and diversified range of MFIs. MIVs invest primarily in debt instruments, equity investments or a mixture of both. For fixed income funds, more than 85% of their total non-cash assets consists of investments made into debt instruments, while equity funds have more than 65% invested into 10

11 equity instruments. In a survey conducted by Symbiotics in 2013, 60% of the participating MIVs consisted of fixed income funds while 21% consisted of mixed funds and 19% were made up of equity funds. In contrast to debt and sub-debt instruments, equity investments are still less likely to be included in commercially-oriented MIVs due to the difficulty involved in placing equity stakes in MFIs at a reasonable price, as they are perceived to be more risky by investors. In general, MIVs can be categorized into three different investment approaches according to their degree of commercialization, which is depicted below in Figure 1. Figure 1: Types of Microfinance Investment Vehicles according to investment approach Types of Microfinance Investment Vehicles Commercial investors Socially responsible investors Development Agencies Private Donors MF Development Funds Quasi Commercial Funds Commercial Microfinance Funds Focus on social return Focus on financial return Source: Dieckmann et al 2007 First, microfinance development funds (MF Development Funds) are more concerned with the development of MFIs and are less concerned with seeking a financial return. These funds act as non-profit entities or cooperatives who grant capital at favorable financial conditions. Usually, funding is provided below market rates and often complemented by financial assistance (Dieckmann et al 2007). The investors in this category consist of development agencies, corporations and private donors. Microfinance development funds primarily target MFIs that are approaching financial sustainability. The second type of MIVs are quasi-commercial funds, which aim to strike a careful balance between social return and financial return. While these funds would like to realize a financial return eventually, they are generally satisfied with the return being below that of market-based returns. The investors that participate in these kinds of funds are made up of private donors, development agencies and socially responsible investors. Lastly, the third and most commercialized type of MIVs are commercial microfinance 11

12 investment funds. The target investors for these types of funds are private and institutional investors, which suggests that commercial microfinance investment funds have a clearer objective than the other types of MIVs previously mentioned. These groups of targeted investors require a high degree of transparency and quality in the official documentation provided by commercial microfinance investment funds. Examples of commercial microfinance investment funds include BlueOrchard Microfinance Securities, Dexia Micro-Credit Fund and responsability Global Microfinance Fund. 4.3 Investor profile The foreign investors that provide the majority of the funding to MFIs consists mostly of international financial institutions as well as private sector investors made up of nongovernmental organizations (NGOs), individual donors, foundations, high net worth individuals, pension funds and other institutional investors. The two main types of foreign investors consist of international financial institutions such as the World Bank and the European Bank for Reconstruction and Development and private investors such as NGOs, foundations, individual and institutional investors. Retail investors currently represent a small group of the players involved in microcredit investments, with most retail investors based in Switzerland, Germany, Belgium and the Netherlands. Unfortunately, the stringent legal requirements currently in place have proven to be a hindrance to this group of investors. In Germany, for example, most registered mutual funds can only be offered on a private placement basis, which limits the access for retail clients (Dieckmann 2008). The Belgian cooperative structure also presents limitations to investors, where dividends are capped. However, despite the small amount of retail investors, this group has the potential to increase in volume over time. The introduction of private sector institutional investors seeking full market return, along with the success of various social investment managers like ResponsAbility, Dexia and Oikocredit shows the potential that retail investors have. By making the move towards greater access to fully regulated financial products in countries like Germany, retail investors will be able to benefit more fully from the integration into the capital markets for microfinance institutions. The development of transparent investment fund structures with clear development and financial objectives should continue to be launched in order to attract potential investors that are willing to invest in microfinance, but may not be comfortable with the existing structures or alternatives being offered. 12

13 4.4 Commercialization of individual MFIs Only a very small portion of MFIs around the world are oriented towards commercial sustainability. However, the microfinance sector is slowly making the transition from a donordriven NGO-dominated framework towards an increasing involvement in capital markets. MFIs can be classified into four categories based on their degree of commercialization. The classification of MFIs is depicted below in Figure 2. Figure 2: MFIs according to commercialization Types of MFIs according to commercialization Tier 1 Tier 2 Tier 3 Tier 4 Top 150 MFIs (1 to 2% of MFIs): mature and financially sustainable Smaller MFIs (8% of MFIs): less known, near profitability Majority NGOs (20% of MFIs): approaching profitability Start-up MFIs (70% of MFIs): mostly unprofitable Source: Dieckmann et al 2007 The first category with the highest degree of commercialization is made up of mature and well known MFIs, which consists of 1 to 2 per cent of all MFIs. Institutions in this tier have already transformed themselves into more formal structures and are increasingly attracting the interest of commercial banks as well as both institutional and individual investors. It is estimated that around 150 MFIs belong to this top tier. The second highest tier is made up of smaller, younger and less known MFIs, which are made up of around 8% of all MFIs. Tier 2 institutions are predominantly NGOs that are in the process of converting themselves into regulated MFIs. Around 300 to 400 MFIs can be found in this tier. The third tier consists mostly of NGOs approaching profitability and often struggle with lack of funding. This tier accounts for approximately 20% of all MFIs. Lastly, the least commercialized MFIs are made up of start-up MFIs or institutions where microfinance is not the primary focus, which make up around 70% of the overall population of MFIs. (Dieckmann et al 2007). Seeing how only around 10% of microfinance institutions have made the move towards commercialization, it is clear that there is room for expansion into the capital market. 13

14 4.5 Funding structure Although around ¾ of the estimated market value of $17 billion dollars is funded by domestic markets, this number is highly skewed by the impact of the few countries where MFIs are actually allowed to take deposits. In reality, most of the estimated 10,000 existing MFIs are not deposit-taking institutions, given the cost and complexity of complying with regulations. Furthermore, domestic commercial banks typically tend to be averse to lending to these institutions. This leaves room for funding from international sources. Non-commercial investors currently account for around 80% of the $4 billion dollars available from international sources. However, with the movement towards commercialization and its appetite for growth in capital funding, it is unlikely that government agencies and non-profit organizations will increase their flow of funding proportionately to that of commercial investors (Swanson 2007). Therefore, an opportunity exists for additional funding through commercialization. The funding structure of an MFI usually follows a certain pattern depending on the life cycle of the institution. While start-up MFIs have a larger dependency on donations, the more advanced MFIs are characterized by higher debt leverage through domestic or foreign borrowing. Besides the life cycle of the institution, other factors play a role in determining the optimal capital structure of a microfinance institution. These factors include the growth of loan portfolio, the regulatory framework, the availability of donors and commercial lenders and the openness of the domestic financial system. The cost of certain funding sources also plays a role in deciding upon the optimal funding structure for the institutions. Issuing equity for example, is the most costly source of finance for MFIs, followed by unsecured and subordinated debt. On the other hand, retail deposits are considered to be the cheapest funding source. 4.6 Examples of commercialization Although it is a relatively new development, the commercialization of microfinance has been increasing strongly over the years. Examples of attempts made to tap into the capital market for microfinance funding include the creation of investment funds such as Profund, the ACCION Gateway fund, as well as the AfriCap Microfinance fund. In April 2007, Banco Compartamos of Mexico was one of the first to issue a highly successful initial public offering (IPO). The IPO was well received by capital markets with the issue being over-subscribed 13 times on an initial price to book ratio of 12.8 and an initial price-earnings ratio of (Dieckmann et al 2007) The Equity Bank in Kenya is another example of a microfinance institution that was able to issue an IPO on the Nairobi Stock Exchange in In 2004, Developing World Markets structured the first international microfinance-backed bond issue, which consisted of a $40 million securitization of cross-border loans to nine MFIs in Latin America, Eastern Europe, and Southeast Asia. The same firm later issued another $60 million collateralized debt obligation in 2006 which involved 26 widely dispersed MFIs that had an average annual loan portfolio growth rate of 30%, a return on equity in excess of 20%, and a portfolio-at-risk level of 2.5% (Swanson 14

15 undated) In 2007, the credit rating agency Standard & Poor announced plans to begin rating MFIs, which would allow pension funds to invest in microfinance-backed securities. Even though it has occurred in the past, direct access to capital markets is still quite rare for MFIs. Most of the access to capital markets occurs indirectly through loans from structured investment vehicles. 4.7 The risk profile of microfinance Investors face two main types of risks when investing into microfinance. The first type of risk is associated with the nature of the microfinance sector while the second type of risk has to do with country specific risks that are typical of developing markets. When it comes to investing through MIVs, investors must keep in mind the lack of liquidity involved as well as the fact that MIVs do not often pay recurring dividends and yields. Also, holdings in MIVs usually have no market price as they are not listed on a regulated market. Therefore, net asset values are used for valuation purposes. Currency risk also poses a potential risk to investors as hedging instruments are often illiquid, too expensive or simply not available in developing countries. The repayment agreement between MFIs and MIVs determines who bears the currency risks involved. If the currency risk is borne by MFIs, then repayment is made in hard currency. In contrast, if the risk is to be borne by the MIVs, then the MFIs repay the loans to MIVs in a local currency (Dieckmann et al 2007). Moreover, European investors must keep in mind the additional currency risk involved since the majority of MIVs are denominated in U.S. dollars. The country specific risks that may arise include political instability, exchange rate controls, currency devaluations and restrictions on the transfer of private capital or on investments of foreigners. Lack of transparency also presents a risk for investors. Unfortunately, very few microfinance investment funds provide adequate and sufficient information on their financial return, their cost structure, their total expense ratio and their loan loss provision, which makes it difficult for traditional investors to compare such an investment with their mainstream investment (Goodman 2007). In addition, the lack of sufficient information also makes it difficult for investors to compare among different microfinance investment funds. However, as microfinance continues to make the move towards commercialization, commercial investors will be able to become more active in microfinance, which in turn will also improve the quality of information provided by microfinance investment funds. 4.8 Potential benefits of commercialization The combination of low default rates (between 1% and 3%), impressive growth rates and reasonable returns makes microfinance a potentially attractive investment opportunity. In addition, the potentially low systematic risk that MFIs may carry creates an advantage for these institutions when competing with other instruments as a potential asset class. Since clients of MFIs are less integrated into the formal economy, these micro-borrowers rely primarily on 15

16 domestic products and services, which causes them to be widely unaffected by currency fluctuations influencing the prices of imports. This low integration of borrowers into the formal economy seems to suggest that the performance of MFIs may exhibit low correlation to macroeconomic development, both in terms of the domestic macro economy as well as global macroeconomic events. The integration of MFIs into the formal economy can also provide benefits to the current and future clients of microfinance. By implementing a more formal capital structure, MFIs will be able to benefit from a positive impact on governance and management accountability. This, in turn, can improve the profitability of the institution, which opens the door to innovation, product diversification and more professional services for clients. Furthermore, the influx of capital funding that arises from commercialization can improve the outreach to the untapped demand that still exists among poor micro-borrowers. While the penetration rate in Bangladesh has already reached 35%, countries like India, Brazil and Nigeria have penetration rates as low as 2% to 3%. To satisfy this demand, it is estimated that a total funding mix of debt, subordinated debt, equity, deposits and guarantees for MFIs of approximately 250 to 300 billion dollars is theoretically needed. (Dieckmann et al 2007) Clearly, the current funding gap presents an opportunity for socially responsible investors to invest in microfinance, which in turn helps supply the capital needed to allow for the provision of loans to millions of people who still remain unbanked. Despite its possible benefits to investors, MFIs have yet to be fully integrated into the capital market. The potential of increasing the investments made into MFIs relies on the following assumptions. First, that microfinance will gradually evolve into a niche investment product that will increasingly attract retail investors and benefit from the popularity of socially responsible investments. The second assumption is that more MFIs will become capable of absorbing foreign funding in the near future. And third, we can assume that investing into microfinance will appeal to a wider range of commercial investors if it proves to be conducive to an investor s portfolio diversification. By further exploring the possibilities that these institutions may have, specifically when it comes to portfolio diversification, MFIs may begin to truly compete with other asset classes for the attention of institutional managers as well as retail investors who are looking to further diversify their portfolios. 16

17 5. Current state of literature 5.1 Impact of macroeconomic environment on MFIs There are a limited number of research papers that have provided some evidence on the possible attractiveness of microfinance for investors. Main examples include: Ahlin and Lin (2006), Galema, Lensink and Spierdijk (2011), Gonzales (2007) and Krauss and Walter (2009). Overall, their research has been aimed at empirically addressing the degree of sensitivity that MFIs show to the macroeconomy. The focus for Ahlin and Lin (2006) and Gonzales (2007) has been on microfinance resilience to domestic macroeconomic shocks, whereas Krauss and Walter (2009) have focused on both domestic and international shocks and Galema, Lensink and Spierdijk (2011) have focused exclusively on international shocks. Ahlin and Lin (2006) use within and between panel regressions to examine whether performances of MFIs are affected by country-level macroeconomic factors. Their sample includes 112 MFIs from 48 countries for the years As performance measures they use: self-sustainability, default rates, costs per borrower and growth in clientele. The result of their research indicates that microfinance success, at least in terms of financial sustainability, growth, and repayment rates, is significantly affected by the domestic macroeconomic environment in which they are situated. However, they also find that the majority of performance is left unexplained by the macroeconomy. Thus, the importance of institution specific age effects suggests that much of the success originates within the institution. Galema, Lensink and Spierdijk (2011) use mean-variance spanning tests with short sale constraints to investigate whether adding microfinance funds to a portfolio of risky international assets yield diversification gains. Annual data was collected on MFIs covering the period Since MFIs are not actively traded and hence have no market returns, the authors relied on accounting measures such as return on equity and return on assets to proxy for microfinance market returns. As benchmarks, global equity and bond indices were used to mimic an investor s broad investment portfolio. First, the spanning test was applied to all MFIs for which data was available, then the MFIs were split into different regions, and lastly, they were split into different types of MFIs. Their analysis suggests that, in general, MFIs can be a valuable addition to a broad portfolio of stocks and bonds. In particular, MFIs from Latin America improve the meanvariance frontier almost always, regardless of the benchmark, while MFIs from Africa never seem to improve the mean-variance frontier. When distinguishing between types of MFIs, they find that rural banks are the best performers, having a high mean return and only small variance. Gonzales (2007) uses four indicators to analyze whether changes in domestic GNI per capita significantly affect microfinance portfolio risk. The four indicators include Portfolio at Risk over 30 days, Portfolio at Risk over 90 days, Loan loss Rate, and Write-off Ratio. His sample contains 639 MFIs in 88 countries for the period Gonzales uses fixed and random-effects 17

18 panel regressions and finds that there only exists a statistically significant relationship between changes in GNI per capita and portfolio risk of MFIs with respect to Portfolio at Risk over 30 days. Therefore, the study seems to provide some evidence that microfinance may provide attractive opportunities for portfolio diversification. Krauss and Walter (2009) use fixed-effects panel regressions to perform an empirical analysis of the systematic risk of MFIs by regressing key fundamental parameters and ratios of the leading MFIs against the S&P500, MSCI Global, and MSCI Emerging Market indexes (as proxies for global market risk) as well as against domestic GDP (as proxy for domestic market risk). Their data set comprises of annual data covering the period and includes 325 MFIs from 66 emerging market economies. The fundamental parameters used to evaluate microfinance performance include return on equity, profit margin, change in total assets, change in gross loan portfolio and loan portfolio at risk. As a result, Krauss and Walter find that, with the exception of profit margin against the MSCI World Index, MFIs in their study display no statistically significant relationship with global market movements. Regarding exposure to domestic GDP, MFIs display apart from net operating income highly significant correlation with all parameters analyzed, demonstrating that MFIs are not detached from their respective domestic economies. In his paper, Gonzalez (2012) explores the resiliency of MFIs during the recent 2008 financial crisis. The level of resiliency was tested by looking at the changes in portfolio quality relative to the changes in domestic GDP. During his analysis, Gonzalez provides evidence suggesting that the microcredit sector was less resilient during the 2008 economic recession in comparison to previous years, and that this lack of resilience was in part due to an increase in the share of nonmicroenterprise lending, combined with an expansion of microcredit in countries with more formal economies (as measured by the percentage of salaried workers). Gonzalez thus found that, the higher the level of formalization, or integration of microfinance loan portfolios with the domestic economy, the worse the decline in portfolio quality during economic recessions. 5.2 Analysis of publicly traded MFIs With the available research on MFIs being limited, it is no surprise that even less literature is available that focuses on the performance and possible resilience of microfinance investment funds. Janda and Svarovska (2010) are one of the few that have been able to perform such a research. Their paper presented an original case study of selected microfinance investment funds, where certain risk/return characteristics were compared against given world and emerging markets equity and fixed income indices. The advantage that their study had over the previously mentioned research has to do with their ability to use monthly returns as opposed to the yearly returns. By using the monthly net asset values of the microfinance investment funds, Janda and Svarovska were able to better observe the developments of these funds due to the increased frequency of the data utilized. Their study comprised of 11 microfinance investment funds, 18

19 which consisted of the most developed funds with transparent portfolio structure inherent to developed financial markets. The data was collected from January 2006 to March Overall their analysis revealed that investments in microfinance investment funds that focus especially on debt instruments showed no positive correlation with global or emerging capital markets. In a more recent study, Briere and Szafarz (2013) were able to conduct an analysis of the performance of publicly traded MFIs while examining the risk factors associated with microfinance and the possible diversification benefits that come from microfinance investments. By constructing microfinance country equity indices and an international Global Microfinance Index, they were able to assess the changes in these indices in comparison to national and financial sector indices. Their study has depicted results that are a bit different from the previous research conducted. According to their research, microfinance equity has become less risky, but is also much more closely correlated with the financial sector, and yields fewer diversification benefits. Unfortunately, empirical research has yet to be conducted on the efficiency of microfinance investments by means of a mean-standard deviation portfolio model. So far, the most comparable analysis involves research on the efficiency of commodities when added to an already diversified portfolio. In their paper, Satyanaran and Varghis (1996) examine the diversification benefits of adding commodities to a portfolio consisting of developed and emerging market assets. As a benchmark for commodity performance, the monthly return of the Goldman Sachs Commodity Index is used for the period of 1982 to In some aspects, commodities possess certain characteristics that are relatively comparable to those of microfinance investments. Both commodities and microfinance investments, for example, have been argued to carry a low or negative correlation with the market. Although commodities have been around longer and have already established a high degree of commercialization into capital markets, they can still be considered as relatively new investments and can be looked as an example of how new investments with a low relationship with the global economy can add efficiency to a portfolio. So far with the exception of Briere and Szafarz the research conducted seems to suggest that investing in MFIs can possibly provide diversification benefits to a portfolio of risky assets due to the relatively low volatility with macroeconomic shocks. However, the fact that the research conducted so far has not expanded beyond 10 years slightly reduces the reliability of the analysis provided. In order to allow investors to make an informed decision on the appropriate microfinance investment opportunity according to their own preferences, a deeper evaluation of these microfinance institutions is needed. Therefore, further analysis on the volatility of MFIs to macroeconomic developments may prove to be beneficial. In addition to conducting an analysis on the individual institutions, it is also helpful to analyze the performance of current investment funds that are aimed at microfinance. By doing this, we will be able to get a better understanding of how the microfinance investment funds are performing relative to the market. 19

20 6. Methodology 6.1 Analyzing individual microfinance institutions The first part of the analysis revolves around the belief that MFIs are capable of exhibiting a high resilience to macroeconomic movements due to their low integration with the formal economy. The first hypothesis is thus as follows: The performance measures of microfinance institutions will exhibit an insignificant relationship with returns of global benchmarks To test this hypothesis, an analysis will be conducted on the performance of MFIs in comparison to the performance of different global equity and bond indices. The typical approach to analyzing the risk of an asset class is to calculate the historical market beta by regressing the returns of an asset class over a certain period of time against the return of a benchmark beta. However, as Krauss and Walter (2009) point out, this approach is not feasible due to the fact that MFIs are virtually almost all non-listed companies with no mark-to-market valuation. Consequently, the only feasible approach in the case of microfinance institutions requires reliance on accounting earnings to arrive at an estimate of the accounting beta. Although this approach suffers from the fact that it compares backward-looking data with forward looking data, studies have found significant correlations between accounting betas and market betas. Examples include Ball and Brown (1968) and Beaver and Manegold (1975), who have empirically investigated the association of accounting betas and conventional market betas and found significant contemporaneous correlation between these two measures. Although Beaver and Manegold (1975) recognize that the accounting beta is not the most important predictor of market betas in their research, they attribute this issue to the multicolinearity with other accounting measures included in their regression analysis. Also, when using aggregation to reduce measurement error, Beaver and Manegold (1975) found that the aggregate accounting beta explained at least fifty percent of the contemporaneous variance in the aggregate beta in their research. While the findings in these studies provide an incentive for using accountingbased risk measures, it is still important to recognize that there are some limitations involved in comparing backward-looking data with forward-looking data. To measure the performance of the individual MFIs, the yearly return on assets (ROA) and return on equity (ROE) were used. Additionally, global equity and bond indices were used to mimic an investor s broad investment portfolio. As a benchmark for the equity indices, the returns for the Morgan Stanley Capital International (MSCI) World index and the MSCI Emerging Markets (MSCI EM) equity index were used. As for the bond indices, the JP Morgan Global Bond Index Broad (JPM GBI) and JP Morgan Emerging Markets Global Composite 20

21 (JPM EM) indices were used as a benchmark. First, a regression was performed on all MFIs over the specified period, and then additional regressions were performed to observe regional and institutional type differences between the MFIs. A fixed-effects regression model was used to run the ordinary least squares (OLS) regressions of microfinance performance on the different benchmarks. Fixed-effects regression models are considered a standard approach when dealing with panel data and controls for differences in the levels of variables associated with individual institutions. In very general terms, the linear model is specified as: = β +ε, where denotes the different benchmarks and measures the partial effects of in period t for institution i. However, since this model is too general, additional assumptions must be implemented. In this case, the added assumption is that is constant for all i and t, except for possibly the intercept term. The new model can thus be written as: = + β +ε Where is a K-dimensional vector of explanatory variables, not including a constant. The thus captures the effects of those variables that are peculiar to the i-th institution and that are constant over time. As an additional assumption, ε is assumed to be independent and identically distributed over institutions and time, with a mean of zero and variance. The model is thus considered to be a fixed-effects model by treating the as fixed unknown parameters. As stated by Krauss and Walter (2009), other models such as the random effects estimator are available and can yield more precise results but only at the expense of stronger assumptions which in this case are certain to be excessively restrictive and contain a high risk of misspecification bias in the results. For example, a random effects model relies on the assumption that either no omitted variables exist or that the omitted variables are uncorrelated with the explanatory variables in the model which is meant to produce unbiased estimates of the coefficients. In reality, however, it is quite likely that omitted variables will produce at least some bias in the estimates. Throughout the analysis, only partial information has been available for all institutions during the entire period. As a consequence, the data provided in the analysis consists of an unbalanced panel data set. When dealing with an unbalanced panel data set, selection bias can often become an issue. In this case, since participation in the MIX database is voluntary and requires substantiating documentation, the institutions in this database do not represent a random sample 21

22 of all MFIs in the world. However, as Krauss and Walter (2009) and Gonzales (2007) have pointed out in their analysis it can be argued that the institutions are a random representation of the world s top MFIs, which represents the potential investment universe for portfolio investors. 6.2 Analyzing current commercialized MFIs The financial developments that have occurred during the recent years provide an added opportunity to observe the performance of MFIs during periods of financial crises, and might help give some insight into the extent to which MFIs might be resilient to specific macroeconomic shocks. Currently, the research conducted on the resiliency of MFIs to financial crises is still in the preliminary stages. As stated by Gonzalez (2012): more analysis is necessary to understand the full trade-offs and synergies between the formalization of the microfinance industry, including the development of better indicators for the informality and flexibility of borrowers and their risk management strategies when facing recessions. To add some more perspective on the research related to the resiliency of MFIs to financial crises, analysis will be conducted on the performance of commercialized MFIs during the recent financial crisis of As a result, the second hypothesis states that: The performance of microfinance investment vehicles will exhibit an insignificant relationship with global benchmarks during the financial crisis of 2008 The reason why microfinance investment vehicles will be taken in to account instead of individual MFIs has to do with the frequency of the data available. For individual MFIs, only yearly data is available for the most part, resulting in a limited amount of observations. In contrast, microfinance investment vehicles produce monthly net asset values (NAVs) which allows for a more meaningful data output. By performing the analysis on microfinance investment vehicles during the crisis, we will be able to better analyze the data by taking into account the variations in performance that have taken place throughout the years. Whereas Gonzalez (2012) took the domestic GDP into account in his research, this analysis will focus more on comparing microfinance performance to that of the global economic development during the 2008 financial crisis. To assess the impact of the 2008 financial crisis on all commercialized MFIs, an OLS regression will be performed for the period of by using global bond and equity indices as benchmarks for macroeconomic development. The output of this analysis will aid in providing better insight into whether or not MFIs can truly be regarded as resilient to severe macroeconomic shocks. 22

23 Aside from observing the relationship that MFIs have with macroeconomic movements, it is imperative that we observe whether any efficiency can be added to a diversified portfolio after investing into microfinance. To conduct the analysis on the efficiency of microfinance investments, the third hypothesis is formulated as follows: MFIs that are currently commercialized will increase the efficiency of a portfolio when added to an already global and diversified portfolio If it is true that MFIs show a low sensitivity to the developments of the market, then it can be hypothesized that adding commercialized MFIs to a global portfolio may result in a more efficient portfolio with higher means and lower standard deviations. To test the third hypothesis, an efficient frontier will be derived by utilizing Markowitz s meanstandard deviation portfolio model. First, an efficient portfolio frontier will be constructed by using global equity and bond indices in order to represent a global diversified portfolio. Just as in the previous analysis, the MSCI World index and the MSCI Emerging Markets index will be used as benchmarks for equity indices, while the JP Morgan Global Bond Index Broad and JP Morgan Emerging Markets Global Composite will represent the global bond indices. In addition, the S&P 500 is also included as a benchmark in order to represent a diversified portfolio. First, an efficient portfolio frontier will be constructed using only the global benchmarks, and then an additional efficient portfolio frontier will be constructed after including commercialized microfinance institutions into the portfolio. To represent MFIs that are already included in capital markets, a microfinance equity index and a microfinance bond index will be utilized. As a benchmark for investments into microfinance equity, the SMX Emerging Sustainable Finance Index (SMX-ESFI) will be used and the SMX Symbiotics Microfinance Index (SMX-SMI) will be used to represent bond holdings in microfinance. The efficient frontier of a portfolio without investments in microfinance will be compared to the efficient frontier of a portfolio that does contain microfinance investments in order to observe the difference in efficiency between both portfolios. According to the Markowitz mean-standard deviation portfolio model, a portfolio is considered efficient when there is no other portfolio that can offer an equal or higher expected return with a lower standard deviation. The mean-standard deviation frontier is derived by minimizing the standard deviation of the portfolio at different given levels of return, subject to different constraints. 23

24 Following the portfolio model theory, the expected return, E(Rp), is calculated as the weighted average of the individual asset returns in the portfolio. Thus: E R = w E(R ) where w i = weight of asset i in the portfolio, and E(R i )= expected return of asset i For the standard deviation of the portfolio, a weighted average of the standard deviation of the individual asset returns is calculated by using the following equation: N σ p = w 2 i σ 2 i + w i w j σ ij i=1 N N i=1 = w i w j σ ij i=1 N j=1 N j=1 Certain assumptions are made under Markowitz s portfolio theory. For example, it is assumed that all investors are rational and risk averse, with the aim of maximizing economic utility. Furthermore, asset returns are assumed to be normally distributed. As a measure for expected returns, predictions are made based on historical measurements of asset return and volatility. Unfortunately, the Markowitz model does have some limitations. For example, because ex-ante returns are being estimated by using ex-post returns, the expected values very often fail to take into account new circumstances that did not exist when the historical data were generated. The optimal investment proportions are found using an optimizing software package in Microsoft Excel which was subject to the following constraints: minimizing the standard deviation while making sure all weights in the portfolio sum up to 1. For this specific analysis, no short sale restrictions were put into place. Allowing for short sales to take place results in the possibility of riskier investment strategies being undertaken. However, since the focus of this 24

25 analysis is on the possible added efficiency that microfinance investments can add to a portfolio, the output of the analysis is still capable of providing meaningful results. The aim of this research is thus threefold: first I would like to get an overall view of the sensitivity of individual MFIs to macroeconomic development over a longer period than previously researched, second, I would like to analyze the performance of commercialized MFIs during the 2008 financial crisis and lastly, I would like to analyze the performance of currently commercialized MFIs when added to an already diversified portfolio. 7. Data 7.1 Analyzing individual microfinance institutions The performance measures for all individual MFIs were retrieved from the MixMarket database ( of the Microfinance Information Exchange, Inc. (MIX). The MIX is a not-for-profit private organization supported by the Consultative Group to Assist the Poor (CGAP), the Citigroup Foundation, the Open Society Institute, the Rockdale Foundation, and other private foundations. MFIs that choose to participate in this database must submit substantiating documentation, such as audited financial statements, annual reports, ratings, institutional appraisals, and other materials that help external analysts understand their operations and increase data quality. In addition to monitoring data, the MIX also applies certain analytical adjustments, such as accounting for inflation, loan loss provision and subsidies in order to standardize the data provided. The return on assets was calculated by dividing the net operating income (after taxes) by the average assets while the return on equity was calculated by taking the net operating income (after taxes) and dividing it by the average equity. The calculations are performed by the MIX based on the raw data that is collected from each microfinance institution. Each key ratio is in turn calculated by the industry standards included in the SEEP Network s framework and financial industry reporting standard. The SEEP Network represents the largest and most diverse network of international development organizations and global, regional, and country-level practitioner networks that promote market development and financial inclusion. As for the performance measure for the global benchmarks, the information for the yearly return of the equity and bond indices was retrieved from the Thomson Datastream database. As mentioned earlier, the MSCI World Index as well as the MSCI Emerging Markets index were used as benchmarks for equity indices. The MSCI World Index captures large and mid cap performance across 23 developed countries while the MSCI Emerging Markets Index does the same for 23 emerging markets countries around the world. Both equity indices cover approximately 85% of the free float-adjusted market capitalization in each country. The JP 25

26 Morgan Global Bond Index Broad tracks fixed rate issuance from high-income countries spanning North-America, Europe and Asia while the JP Morgan Emerging Markets Global Composite index tracks the same information for emerging markets. The OLS regressions will be conducted using STATA. To increase the quality of the dataset, only MFIs with at least 3 consecutive years of financial data have been included. In total, 1377 MFIs have been included in the dataset. The period chosen spans from the years 1999 to 2013, and all numerical data was converted to US dollars at contemporaneous exchange rates. To increase the value of the research, MFIs were further split into the following regions: Africa, East Asia and the Pacific (EA & P), Eastern Europe and Central Asia (EE & CA), Latin America and the Caribbean (LA & CAR), Middle East and North Africa (ME & NA) and South Asia (SA). This regional categorization allows for a deeper understanding of the relationship that different regions have with global developments. In addition, in order to differentiate between the different types of MFIs, the institutions were split into: Banks, Cooperatives/Credit Unions (Co-op/CU), Non-bank Financial Institutions (NBFIs), Non-profit (NGOs), Rural Banks and Other. The categorization will thus provide potential investors with an idea of how MFIs from different regions and legal statuses have performed, which may be useful for future investment strategies. Figure 3 provides an overview of the MFIs included in the dataset, categorized into different regions and legal statuses. Figure 3: Classification of MFIs per region and legal status MFIs per region MFIs per legal status ME & NA 4% South Asia 19% LA & Car 28% Africa 18% EE & CA 19% EA & P 12% Bank 9% Other 1% Rural bank 6% Credit Union / Coop 15% NGO 37% NBFI 32% The region with the highest amount of MFIs is Latin America and the Caribbean, with a total of 388 institutions in the dataset, while the regions with the lowest amount of MFIs belong to East Asia and the Pacific with 12% and Middle Eastern and North Africa with 4%. In their paper, 26