1
The commercialization of microfinance:
An empirical analysis determining the distinctive traits related to profit
status of microfinance institutions
Author: Renata Haseth, 10153489
Supervised by Prof. dr. Sweder Van Wijnbergen University of Amsterdam
Abstract. - This paper attempts to shed light on the trade-offs associated with the
commercialization of microfinance. Previous literature has not led to a clear consensus on the topic. In this paper fifteen regressions were run using panel data from 117 different countries to determine the crucial differences between for- and non-profit microfinance institutions. The results show signs of mission drift with non-profit institutions serving more female borrowers and having a lower average loan size per borrower and a lower yield. However potential reasons for these findings are discussed and the importance of not discarding the
commercialization and thus self-sustainability of microfinance is pointed out. My results have shown that commercial institutions pay higher salaries to their staff members, make higher costs per borrower and have lower profit margins. This is evidence against the return of loan sharks. Finally, both the for- and non-profit institutions reflect important qualities in the fight against poverty and as the industry matures and operating costs are lowered a healthy medium can be found.
2 Table of Contents Abstract……….. 1 Table of Contents……… 2 1. Introduction……….. 3 2. Literature Review………. 2.1 The effect of microfinance on poverty……….. 3
2.2 Trade-off……… 5
2.3 The commercialization of microfinance……… 2.3.1 Implications of commercialization………. 7
2.3.2 The challenges of commercialization………. 9
3. Hypotheses……….. 10
4.1 Data……….… 13
4.2 Variable Description……….. 13
5. Methodology………. 17
6. Regressions and discussion……….. 18
7. Conclusion……… 38
Bibliography………. 41
3 1. Introduction
Microfinance is the provision of credit, among other financial services to the poor. The poor are usually unsuitable candidates for traditional bank loans, due to their lack of collateral as well as the fact that they need very small loans. Microfinance banks have come in to fulfil this need. Programs around the world, using a variety of models, have shown that poor people achieve strong repayment records. Commercial players have been developing an interest in microfinance. This brings up an important question, are these high repayment rates result of a good business model or regular loan sharks?
Microfinance institutions have expanded rapidly over the last 10 to 15 years and according to recent estimates from the World Bank the industry has reached about 130 million clients. However this only constitutes to less than 20 percent of its potential market. Around three billion people still have little or no access to formal financial services, according to the International Finance Corporation. In recent years there have also been many changes in the orientation of microfinance institutions. One of the biggest changes comes in the form of a change in profit orientation, more and more microfinance institutions have evolved into for-profit institutions in order to attract commercial funding*. Although the industry was initially comprised of mostly non-profit institutions, an increase in commercialization has been clearly visible in recent years. Moreover, commercial banks have started to see potential in the industry and have entered the market either directly or through intermediaries (Bruton et al., 2011). This potential of making profits while helping the poor pull themselves out of poverty caught the attention of policy makers, investors and academics alike. However critics say, if it sounds too good to be true, it probably is. What is striking about this debate is the relative lack of evidence to inform it. There are several stories about both micro borrowers swimming in debt as well as successful micro-entrepreneurs, but these don't paint a clear picture about the impact microfinance has on the average borrower. In this paper I attempt to shed some light on the effect of the commercialization of microfinance.
The thesis is structured as follows; first off a review of the literature on microfinance is given. Both the effect of microfinance on poverty as well as the tradeoff between profit and poverty reduction is discussed. Section three focuses on the commercialization of microfinance. Section four gives the hypotheses and section five describes the collected data and variables. Section six contains the methodology, regression and gives the results. Finally in the last section the conclusion is given.
2. Literature Review
2.1. The effect of microfinance on poverty
Most of the literature on micro lending programs is oriented towards the effect microfinance has on poverty in developing countries. Even so, no clear consensus is reached on the effect microfinance has on poverty reduction. In 1998 Morduch (1998) examined the effect of microfinance on poverty in Bangladesh. A sample of 1800 households was used, some of which had access to microfinance and some of which did not. The results showed that the households with access to microfinance did
* Wagenaar (2012) documents that at least 59 not-for-profit microfinance institutions have transformed into for-profit organizations between 1996 and 2010. A famous example is the Mexican MFI, Banco Compartamos, which raised a hefty debate in the microfinance community by issuing initial public offerings in April 2007 (Cull et al., 2009).
4 not have particularly higher consumption levels than control households, and there was no notable difference in the children’s education. However, households with access to microfinance programs did have a much lower variation in consumption and labor supply across seasons. Thus although Murdoch (1998) did not find a notable effect of microfinance on poverty reduction, the
vulnerability of households was reduced by access to the programs.
A similar experiment was done in Hyderabad about five years later, in this case out of 105 slums, half was selected at random and in these slums a branch of the microfinance institution Spandana was opened, while in the other half of the slums Spandana did not open a branch. However, it was possible for other microfinance institutions to open a branch in these areas. After three to four years after the start of the experiment and after several of the control slums started receiving credit from Spandana and other microfinance institutions, the probability of borrowing from an microfinance institution in treatment and comparison slums was the same. Yet, on average households in the treatment slums have been borrowing for a longer period of time and are
borrowing larger amounts of money. There was no apparent change in consumption in the treatment slums, and the businesses were not more profitable, although there is an increase in profits at the top end. There were no changes in health, education, female empowerment or other development outcomes often believed to be affected by microfinance (Banerjee et al.,2003).
On the contrary, Khandker (2005) found that microfinance does contribute to poverty reduction. Using panel data from Bangladesh he examined the effect of microfinance on a participant level as well as on a aggregate level. The results indicate that microfinance does help reduce poverty, especially for female participants. Poverty is also reduced at a village level and he finds that microfinance thus leads to an improved local economy. This is corroborated by a study done in Bolivia. Using a small-sample survey on four microfinance institutions, evidence was found that when basing the reduction in poverty on measures such as income and asset holdings, microfinance appears to be successful and relatively cheap at reducing poverty of those close to the poverty line (Mosley, 2001). When looking at the impact of participation, by gender, in micro credit programs Pitt & Khandker (1998) showed that microfinance has a stronger effect on poverty reduction when the participant was female, although household consumption expenditure increased for both males and females. To correct for the bias from unobserved individual and village level heterogeneity, they used a quasi-experimental survey design. Differences in initial income may also play a role, while examining microfinance in Thailand and controlling for endogenous
self-selection and program placement, Coleman (2006) * found that participants of the microfinance program tend to be wealthier than non-participants, even before they started the program. In addition the wealthiest villagers are almost twice as likely to participate in the program as the poorer villagers. The wealthier villages are also more likely to become a program committee member and borrow more money than their poorer counterparts. However, local information on the creditworthiness of the villagers is used when selecting members. The results of this paper show that household welfare of the wealthier members is improved through microfinance, while the impact is largely insignificant for poorer members.
Even though a relationship between microfinance and clear poverty reduction cannot always be found, Microfinance critic Dichter (2007) admits that microfinance does lead to
consumption smoothing over periods of economic busts or unexpected crises. This positive aspect
5 should not be overlooked. If the ability to smooth consumption means children can stay in school during economic downturns, nutrition is more stable and medication can be purchased whenever needed, then microfinance is likely to have positive long- term effects on productivity.
Microfinance can also help reach the World Bank’s sustainable development goals. These goals are set to improve the welfare of the world's neediest people. The main goal was to end extreme
poverty by 2015. These goals have now been replaced by the Sustainable Development Goals, which also state ending poverty as a priority. Microfinance can contribute to this goal through assisting the poor in their need for immediate cash to survive. Access to financial services forms a fundamental basis on which many of the other essential interventions depend. Moreover,
improvements in health care, nutritional advice and education can be sustained only when
households have increased earnings and greater control over financial resources. Financial services thus reduce poverty and its effects in multiple, concrete ways. Paired with the possibility of
microfinance reaching financial sustainability, these programs have the ability to reach far beyond the limits of scarce donor resources (Littlefield et al., 2003). While the quality of many studies could be improved, there is an overwhelming amount of evidence substantiating a beneficial effect of microfinance on poverty reduction.
2.2 Trade-off: Being profitable and eliminating poverty
During the first years of microfinance, charitable organizations with the goal to eliminate poverty gave out very small business loans. They relied on donations to keep up their work and help the poor. Big, commercial banks did not get involved, not seeing profitable opportunities considering the minuscule loans. Today, that has changed. After three decades of growth the microfinance industry has gained new players and become both more crowded as well as more complex. The Microfinance Information exchange reports microfinance institutions are now active in more than 100 countries and serve more than 92 million clients, of which about two thirds are borrowers and the rest are savers. Big banks have started to see the potential the industry has and the money that can be made at the bottom of the pyramid. Commercial investors are seeing profits too. And while once only philanthropists were interested in the industry, now microfinance institutions are
scrambling not to lose ground.
Due to this rise in the number of for-profit microfinance institutions, commentators are asking whether the sector benefits from the stronger profit orientations. Although there is a lot of discussion around the tradeoff between profitability and poverty outreach in the literature, there are few studies done on the subject. Hovi (2012) studied the trade-off between profitability and poverty outreach in the microfinance industry depending on the institution type. The main purpose of his study is to find out whether institutions can achieve highly profitable operations at the same time with large poverty outreach. He concluded that this was indeed possible. Furthermore he concludes that interest rates don't differ substantially between for and non-profit institutions and for profit institutions operate more cost efficiently while serving the poor. Christen & Drake (2002) look into the issues that come with an increase in commercialization of microfinance. They argue that
commercialization and an increase in competitiveness leads to diversification in lending products, lowered interest rates, as well as an increase in the quality of client services and a larger target
6 group. Although they also acknowledge disadvantages to commercialization, such as systematic risk for microfinance institutions and possible over-indebtedness on the part of clients, they conclude that increased commercialization would be good for the microfinance industry and necessary if the industry wants to gain higher growth rates.
In a study done by Rhyne (1998), eleven microfinance companies were examined to
determine whether there is a direct trade-off between outreach and financial sustainability. First of, no evidence of a clear trade-off between reaching the very poor and reaching large numbers of people has been obtained. Additionally it became clear that financial services for the poor can be provided on a financially viable basis. When trying to determine why some companies were more financially viable than others, there was no evidence found that loan size had anything to do with financial viability. However, the interest rate charged was found to be of significance. Profitable companies charged an effective real interest rate high enough to cover all costs, including potential devaluation due to inflation. The companies which held interest rates down chose to remain subsidy dependent. Although they may not have admitted it, these programs were subsidizing interest rates to clients. In her paper Rhyne (1998) argues that because pricing was such a direct determinant of viability, she realized the debate on the tradeoff between sustainability and outreach is ultimately about whether to subsidize interest rates. Those who choose to subsidize interest rates instead of reaching financial sustainability are thus claiming the poor would not be able to fully afford borrowing costs. This statement makes the debate a lot simpler since this can be and has been empirically tested. There is little or no evidence of the notion that increasing interest rates has significantly affected client demand for their loan products.
The switch to a more commercial approach has not had a significant effect on repayment rates, research has shown these have remained high. However, due to the high costs of
microfinance, profit-maximizing investors have little interest in institutions focusing on the poorest in the community, which are generally women. Although commercial institutions charge high interest rates, they also incur high transaction costs. In order to attract as many commercial investors as possible and build a sustainable business, a way must be found to keep these
transaction costs as low as possible. although the industry has overcome a lot of these problems, further innovation is necessary to overcome high costs (Cull et al, 2009).
The commercialization of microfinance might also be interesting for investors as a tool to diversify the portfolio and decrease portfolio volatility. Measuring the correlation between microfinance and international and domestic markets shows that portfolio volatility can be decreased for international investors, since the results show no correlation between microfinance companies and global capital markets, However there is evidence found of a correlation between domestic GDP and microfinance companies, which means that the reduction in volatility does not hold for domestic investors (Krauss & Walter, 2009).
The transition of microfinance to a more profitable industry has many consequences, such as an increase in competition which usually leads to benefits for consumers. On the other hand mission drift is one of the setbacks feared to go hand in hand with commercialization. In the following section I review consequences of commercialization.
7 2.3 The commercialization of microfinance
The commercial approach to microfinance is generally defined as “the application of market-based principles to microfinance” or “the expansion of profit-driven microfinance operations” (Poyo & Young, 1999). The commercialization of microfinance is distinguished by competition, profit and regulation. This has the positive effects of new product development, improvement of exiting products and strategy development. To attract commercial funds, microfinance companies may become more regulated, which is important for many operators to enter the formal financial system (Christen, 2001). There are two types of institutions paving the way in the commercialization of microfinance; the NGO’s and the large retail banks (include state-owned banks). The NGOs transform themselves into licensed non-bank financial intermediaries or into banks (Christen & Drake, 2002).
2.3.1 Implications of commercialization
As microfinance starts to commercialize and earn profits, an increasing amount of new players are likely to enter the market. This has led to heightened competition in the microfinance industry, with both new microfinance institutions as well as poor clients borrowing at different institutions. In general economic theory increased competition should lead to benefits for consumers. Although the microfinance industry is a bit trickier than most, several positive implications from
commercialization are apparent. The benefits that are associated with a commercial microfinance are varied. The Institutions schools of thought have argued that commercial microfinance has enabled institutions to be better regulated, provide cheaper and quality services to clients, access to domestic finance and increased outreach. (Woller et al. 1999).
2.3.1.1. Increased Outreach and Competition between microfinance institutions
Increased competition is expected to lead to lower consumer prices, more diversity and more supply of the product. Although in microfinance some seem to be less optimistic and caution for the negative effects competitive behavior may have on poor clients. Khandker et al. (2013) found evidence that contradicted these worries, the research shows the newer microfinance companies are not more risk-averse than their more mature counterparts. They also find evidence that the recovery rate does not decrease with increased borrowing from households. Although the Grameen Bank model, which focuses mainly on reducing poverty has been successful at reaching the poorest people using subsidies from the government and other donors, it is important to realize that “as a global solution to meeting microfinance demand, governments and donors cannot finance the hundreds of millions of people who constitute present unmet demand for microcredit services”. Through this realization, it becomes clear that the commercialization of microfinance should be seen as way to increase the supply of financial services to tend to the needs of people at all levels for sustainability (Robinson, 2001).
8 2.3.1.2 Improved Efficiency of microfinance institutions:
The commercialization of microfinance drives the industry to lower costs, reduce interest rates and increase productivity. Both NGO’s as well as banks have improved operating efficiency which indicates a positive trend in the commercialization of microfinance (Charitonenko & Rahman, 2002). Efficiency is usually measured using the ratio of administrative costs to average gross loans. In recent years the administrative costs are decreasing, while average loan sizes increase, so the amount of money required to service a loan is less, which means operations have become more efficient. This increased efficiency has also benefited clients, with lower interest rates over the years. This decrease has been quite consistent which offers low-income clients a more effective and reliable environment. All in all, competition has led to a more efficient microfinance industry and a reduction in expenses for both the institutions as well as the clients (Bodnar, 2010).
2.3.1.3 Expanded Set of Microfinance Products
New entrants into the microfinance industry bring innovations and fresh ideas. Despite the
historical emphasis on the Grameen model of microfinance, other models are being used more and more. Until recently most micro-lending has been trough group loans, whereas now an increasing amount of microfinance companies are introducing individual lending, amongst other types of new loan products. Additionally, various types of flexible saving services have been introduced to fit the needs of the poor. Micro-deposits have been created in a cost effective way. Micro-insurance has also been gaining interest recently, as well as remittance services and leasing services. New technologies are also adding to the possibilities for new microfinance products. All these trends should improve the quality of and increase the access to microfinance dramatically (Charitonenko & Rahman, 2002).
2.3.1.4 Financial markets improvement
Financial markets are also effect by the commercialization of microfinance. Due to this
commercialization an increase in advanced regulatory framework takes place which builds up and develops the institutions and authority of the financial sector. On top of that, the commercialization of the microfinance has also transformed the structure of the country’s financial sector, which is described by González and Villafani (2006) for the case of Bolivia. They state that this
transformation has substitution as well as aggregate effects. Considering the former, the
microfinance sector has expanded the potential market for financial services in Bolivia. Clients, who are traditionally left out of the banking system, are now eligible for credit. Due to the fact that this untapped market can now be effectively served, the microfinance sector in Bolivia has been growing more rapidly than the traditional banking industry. As a result, the microfinance sector has contributed to the growth in the Bolivian financial sector as well as a more diversified banking system. Moreover, the substitution effect regards the replacement of informal funding with formal financial services for the poor. This gives the low-income clients a chance to use their money in a more productive way, by now not only avoiding the extremely high interest rates charged by informal money lenders but also gaining access to other financial services. This is a great achievement since the poor have typically been seen as too risky. As the success and potential
9 profitability of the microfinance sector has become more apparent, formal banks have begun to enter this market. Some have started to take advantage of the big pool of low income clients and are expanding their services to fit the needs of the poor as well (González and Villafani, 2006). Even though formal banks are relatively new players in the microfinance industry, their entry could lead to quite some benefits. The banking experience they bring to the table, the established
infrastructure, internal controls and widespread branch networks all improve the quality of the microfinance sector (Navajas and Schreiner, 1998; Baydas et al., 1997). Making use of these already existing systems and networks would provide a cheap way to expand and increase outreach in the microfinance industry. Furthermore, from these formal banks, low-income clients have the opportunity to obtain larger loans and for longer maturities than from microfinance institutions (Berger et al., 2006). These are beneficial options for clients who reach the need for larger, longer term loans after a start in microcredit (Bodnar, 2010).
2.3.2 The challenges of commercialization
With commercialization comes the pressure of higher financial returns. Critics of
commercialization of microfinance believe that institutions transforming into regulated entities will offer larger loans and provide loans to higher income clients and thereby exclude the poorest of the population. This phenomenon is referred to as mission drift, which is properly defined as “the de-emphasis of social mission in pursuit of higher financial return” (Woller 2002). Grameen bank founder Muhammad Yunus fears that with too much focus on financial sustainability, microfinance institutions become dreaded loan sharks, which they initially intended to replace. Despite the benefits commercialization of microfinance brings, supporters of the social mission of
microfinance view the ongoing process as a threat to the initial purpose of helping the poor. This outreach of microfinance can be measured in several ways, including average loan size and interest rates charged. Research is needed to clearly evaluate the effect of commercialization, to find the right balance between financial sustainability and sufficient outreach.
10 3. Hypotheses
Being financially sustainable and profitable will help mature the microfinance industry and as the industry grows, more people will be able to benefit from these financial services. Microfinance was developed to fight loan sharks, but with the commercialization of microfinance came the concern that these microfinance institutions turned into loan sharks themselves. In order to get a clearer vision on this problem, the following research question is formulated:
“What are the significant differences between for profit and nonprofit microfinance institutions?” Because this question is too general, the following sub questions have arisen:
H1: “The yield on gross portfolio is higher for for-profit institutions compared to their non-profit counterparts”
The yield on gross loan portfolio is seen as a proxy for the interest rate charged by the institutions. For profit institutions are associated with higher interest rates for clients (Roberts, 2013).
H2: “The percentage of female borrowers is lower for for-profit institutions compared to their non-profit counterparts”
Given that nonprofits more focused on social impact, rather than profitability per se, we would expect non-profits to have a higher proportion of women borrowers. Previous research has also indicated that this is indeed the case (Cull et al., 2011). Results from previous investigations (see Armendariz and Morduch, 2005) seem to suggest that women borrowers provide greater social impact than their male counterparts and also tend to have higher repayment rates.
H3: “The operating expenses to total assets are lower for for-profit institutions compared to their non-profit counterparts”
For profit microfinance institutions are expected to be more efficient than non-profit microfinance institutions and thus have lower operating costs.
H4: “The cost per borrower is lower for for-profit institutions compared to their non-profit counterparts”
With their social mission, nonprofits would be expected to provide more services per borrower. In addition, trying to reach as many poor borrowers as possible would likely increase costs for nonprofits. Thus, we would expect nonprofit MFIs to have higher attendant expenses than for-profits.
H5: “The amount of borrowers per staff member is lower for for-profit institutions compared to their non-profit counterparts”
Because the incentive for the profit orientated microfinance institutions to work productively is higher, we expect the borrowers per staff member to be lower for for-profit institutions.
11 H6: “The portfolio at risk is lower for for-profit institutions compared to their non-profit
counterparts”
Due to the higher incentive of for profit microfinance institutions to avoid making loans to very risky borrowers, we would expect nonprofits to have a larger proportion of loans with high risks. H7: “The financial expenses to total assets are lower for for-profit institutions compared to their non-profit counterparts”
With their social mission, we expect non-profit institutions to focus on serving as many services to as many people as well as to the poorest, who are generally the most costly to tend to. However for-profit institutions focus on both serving their clients as well as making a for-profit. Thus, we would expect nonprofit MFIs to have higher attendant expenses than for-profits.
H8: “The average salary is higher for for-profit institutions compared to their non-profit counterparts”
As the amount of profit orientated microfinance institutions increases and the industry continues to grow, a fight for the best banking talent has emerged. Non-profit institutions have an increased need for well-educated employees at a managerial level to handle the increasingly sophisticated financial services non-profit institutions are obligated to offer in order to survive. It is expected that commercial for-profit institutions are capable of offering higher salaries compared to their non-profit counterparts.
H9: “The profit margin is higher for for-profit institutions compared to their non-profit counterparts”
Since nonprofit organizations do not have the same profit-maximization motivation than their for-profit counterparts, nonfor-profits are expected to have lower for-profit margins than for-for-profit firms. In general one would expect nonprofits to operate very close to their costs, leaving little room for excess profits.
H10: “For-profit institutions are more likely to take deposits compared to their non-profit counterparts”
Because a regulated entity can collect deposits, we expect for-profit microfinance institutions to be more likely to take deposits. Microfinance institutions generally fall into four categories: Bank, Credit Union (CU), Nongovernmental Organization (NGO), or Non-Bank Financial Institution (NBFI). The first are usually for-profit institutions and allowed to take deposits while NGOs are traditionally non-profit institutions and are usually not permitted to take deposits. NBFIs are a bit more neutral, although they tend to mostly be for profit*.
H11: “The average loan size per borrower is higher for for- profit institutions compared to their non-profit counterparts.”
12 This hypotheses is seen as a way to measure outreach. It is expected that non-profit institutions serve larger numbers of the poor, which leads to them to giving out loans with smaller average loan size than their for-profit counterparts. Prior research (Cull et al., 2009) suggests that being a non-profit is associated with a greater outreach than a for-non-profit.
H12: “The debt to equity ratio is lower for for-profit institutions compared to their non-profit counterparts”
Non-profit institutions are generally dependent on donations and have a very difficult time raising debt financing. For this reason I expect the debt-to-equity ratio is higher for for-profit institutions. H13: “The risk coverage is lower for for-profit institutions compared to their non-profit
counterparts”
Because non-profit institutions generally take on riskier loans than for profit institutions, I expect the risk coverage to be higher for nonprofit institutions.
H14: “The return on assets is higher for for-profit institutions compared to their non-profit counterparts”
This hypothesis follows the same reasoning as hypothesis 9.
To add meaning to these research questions, the objectives of this research paper should be
formulated. These objectives draw a clearer image of the researches sense of purpose and direction. The objective of this research paper is to clarify the difference between the profit status of
microfinance institutions and to verify if there is any truth behind the stereotypes surrounding profit seeking microfinance institutions.
13 4.1 Data
In this paper, the data used is gathered from the Microfinance Information Exchange, known as MIX. This database was established in 2002, founded by Consultative Group to Assist the Poor (CGAP) and sponsored by several well-known organizations such as the Bill and Melinda Gates Foundation, Deutsche Bank Foundation and Citi Foundation. It is a non-profit global, web based microfinance information platform which promotes the transparency and the exchange of
microfinance information. MIX has financial and operational data, as well as data on social performance, products and funding structure data of over 2000 microfinance institutions. All data submitted to MIX is submitted on a voluntary basis. Microfinance institutions are chosen per region. Per region the institutions are chosen that best present the particular area. This means that although MIX does not include every microfinance institution in the world, MIX Market includes the institutions profiles from developing markets that are most representative of microfinance at each geographical level. Smaller institutions with either strong commitment to transparency or links to networks where MIX works also appear in the data set. To make sure the data is reliable and easy to use, MIX preforms an audit, cleans the data and applies more than 135 quality checks and audit rules. Audit rules make sure financial statements balance and check whether ratios levels are abnormally high or low.
To measure microfinance transparency and the quality of the data, MIX uses a ‘diamond’ ranking system. The higher the number of diamonds means more transparent and reliable data. Each profile on the site is ranked for its overall quality, as well as an annual ranking for each full year of data that is published on the website. The profile diamonds reflect the most recent
information available for this MFI *.The panel data set used is unbalanced.
I started with about 18000 observations before cleaning the data more thoroughly, first I dropped any duplicates. Thereafter, any incorrect observations and obvious outliers were also dropped**.The following dummy variables were created: AK1-3 (Age) , BV1-2 (Profit Status), BW1-6 (Region), DEP1-2 (Deposits) and BX1-2 (Regulated).
4.2 Variable description
Number of active borrowers (AG)
Outreach refers to MFI's ability to provide high quality financial services to a large number of clients. One way to measure outreach is by the number of active borrowers. This measure is defined as the number of people who currently have an outstanding loan balance with the microfinance institution or are primarily responsible for repaying any portion of the gross loan portfolio. In case of group lending, the amount of individuals in the group should be accounted for separately, rather than the group as a whole (Stauffenberg, 2003). Generally, due to economies of scale, costs are negatively related to size, while profitability has a positive correlation with regards to scale. The fixed costs of production are amortized across a larger number and value of outputs. Thus, it is expected that the number of borrowers is positively related to financial self-sufficiency (Nara, 2013).
14 Average loan balance per borrower / GNI per capita (P) The average outstanding loans contain only loan amounts that have not yet been repaid, and should not be confused with the total amount loaned. By dividing the average loans by GNI per capita the difference between countries can easily be compared. GNI per capita is not the ideal measure of income because it is skewed by high inequalities in income distribution, especially in Africa. Nevertheless, it is the best available option in this study. It is measured as followed: Average outstanding loans or savings balance per client / GNI per capita (Rosenberg, 2009).
Debt to equity ratio (N)
The debt to equity ratio indicates how much debt and how much equity the microfinance institution is using to finance its assets. The debt to equity ratio is measured by dividing its total liabilities by shareholders’ equity.
Capital to asset ratio (M)
The capital to asset ratio is measured as an approximation for capital adequacy. It is calculated by dividing total equity by total assets. It is the amount of money which a MFI has in the form of equity, shown as a percentage of its assets. This measure provides information about the ability of the company to meet long-term expenses and obligations as well as the amount of money that is available for any future foreseen costs. Internationally, this has been agreed to be minimum 8%. Gross loan portfolio to total assets (AZ)
Gross loan portfolio to total assets measures which part of a company’s assets is actually used to loan out. This measure is important in the microfinance industry due to the fact that these
institutions tend to have other activities involving the social side of the sector. To make sure the data on this variable is as reliable as possible; I have followed the tactics of Jørgensen (2012) by deleting all observations over 1. Although ‘gross loan portfolio to total assets’ is a relatively new term and therefore cannot easily be found in books of theory , research done by the World Bank has shown 0.689 to be the mean (Cull et al., 2005).
Operating expenses to gross loan portfolio (Z)
This ratio is similar to the operating expenses to total assets ratio mentioned above. The two differ in the denominator, the former divides operating expenses by gross loan portfolio, while the latter divides the operating expenses by total assets.
Portfolio at Risk / Arrears (AD)
Portfolio at risk is measured by the ratio of the microfinance institution’s total gross outstanding loan portfolio that is at default risk. It is an easily comparable measure, which does not
underestimate the risk. For the microfinance industry, a loan is considered at risk if a payment on the loan is more than 30 days late. In the commercial banking industry this rule is less strict, due to the fact that commercial banks generally obtain more collateral from borrowers, which reduces their risk. Portfolio at Risk is calculated by dividing the outstanding balance of all loans with arrears over 30 days, plus all refinanced (restructured) loans, by the outstanding gross portfolio as of a certain date (Jansson & de Desarrollo, 2003).
15 Write off ratio (AE)
Although portfolio at risk is a helpful measure, it does not tell the whole story. This measure can be manipulated, as can all performance measures. This is often done by writing off delinquent loans. For this reason portfolio at risk must always be analyzed in combination with the write off ratio. High write-offs point to of poor quality microfinance and usually clients have to pay the price for the company’s inefficiency (Jansson & de Desarrollo, 2003).
Average salary / GNI per capita (AP)
This is defined as: Personnel Expense / Personnel, average / GNI per capita Profit status (BV)
A microfinance institution can either be non-profit (1) or for-profit (2)*. It is expected that
nonprofit institutions have a more social orientation and for profit firms are more driven by making profit. Although the status may seem to indicate that a non-profit firm does not generate profit, this is not necessarily the case. The difference between for- and non-profit institutions is the way the firm uses the profits. For-profit institutions may distribute the profits to shareholders, while non-profit firms are obligated to reinvest the non-profits in the company (Cull et al., 2009)
Age (AK)
This measure indicates the age of the microfinance institution. Every company falls into one of three categories, “New” which indicates the institution is 1 to 4 years old (1) , “Young” which indicates the company is 5 to 8 years old (2) or “Mature”, which indicates a company is more than 8 years old (3) . The count starts the year the microfinance started their operations and ends at the time the data is submitted by the company. A dummy variable has been created to indicate whether the entity is mature, new or young.
Profit Margin (BU)
The profit margin is measured as the company’s profits as a percentage of the total sales revenues generated. It is also called the return on sales ratio or gross profit ratio and it measures how much out of every dollar of sales a company actually keeps in earnings.
Yield on Gross Portfolio (CE)
The yield on gross portfolio includes all the incomes from loans, which aside from interest rate charged includes fees and other loan charges as a percentage of the gross loan portfolio
Percent of female borrowers (BL)
This is a percentage, calculated by dividing the number of active female borrowers by the total number of active borrowers.
Operating expenses to total assets (Y)
This is a percentage, calculated by dividing the operating expenses by the institutions total assets.
16 Region (BW)
Microfinance institutions are divided into five primary geographic regions: Africa (1), East Asia and the Pacific (2), Eastern Europe and Central Asia (3), Latin America and
the Caribbean (4), Middle East and North Africa (5), South Asia(6). Financial expense to total assets (W)
This is a percentage, calculated by dividing the financial expense by the total assets and taking the average
Percent of female loan officers (BM)
This is a percentage, calculated by dividing the number of female loan officers by the total number of loan officers.
Regulated (BX)
Indicator of whether a microfinance institution is regulated (2) or not (1).
Borrowers per staff member (AB) This is calculated by dividing the number of active borrowers by the total number of staff members. Cost per Borrower (AA)
This is calculated by dividing the operating expenses by the total number of active borrowers and taking the average.
Deposit to loan ratio (AW)
This variable contains all deposits, whether voluntary, compulsory, retail or institutional. The deposits to loan ratio is calculated by dividing deposits by the gross loan portfolio.
Assets (H)
This is the total of all net asset accounts. Financial revenue to total assets (W)
This is a percentage, calculated by dividing the financial revenue by the total assets and taking the average.
Risk coverage (BY)
This measure shows what percent of the portfolio at risk is covered by actual loan loss reserves. It shows how well prepared the institution is to handle the worst case scenario . It gives an indication of how prepared an institution is for a worst-case scenario. The risk coverage is about 80 to 120 percent of portfolio at risk for microfinance institutions, which is substantially higher than the coverage at commercial banks. This is mainly due to the immaturity of the industry and the lack of collateral.
17 Average deposit balance / GNI per capita (R)
This is calculated by dividing the Average Deposit Account Balance by gross national income per capita.
Return on Assets (S)
This is one of the most popular profitability ratios for the reason that it shows the rate of return for creditors as well as investors of the company. It indicates how well a company controls its costs and uses its resources.
Deposits (DEP)
This dummy variable shows whether the institution takes deposits (2) or does not (1). 5. Methodology
The method used in this research paper is panel data regressions with random effects. There are several reasons for choosing to use panel data. Most importantly, panel data controls for individual heterogeneity, while time series and cross section studies do not and thus run the risk of obtaining biased results. Additionally, I am aware of the fact that some variables needed for my research are either difficult to measure or hard to obtain so that not all the necessary variables may be included in the equation. Omission of these variables may lead to bias in the resulting estimates, however in contrast to time series study or a cross section research, panel data controls for state and time invariant variables. Instrumental variables may be a solution to the possible omitted variable problem, however the dataset does not allow this. Panel data give more informative data, more variability, less co-linearity among the variables, more degrees of freedom and more efficiency.
By blending the inter-individual differences and intra-individual dynamics panel data has several advantages over cross-sectional or time-series data, including more accurate inference of model parameters. Panel data usually contain more degrees of freedom and less multi-collinearity than cross-sectional data (Hsiao et al.,1995). In my research, I used the random effects model, which is a regression with a random constant term (Greene, 2003). The main reason for using random effects instead of fixed effects is the fact that the random effects model has the distinct advantage of allowing time-invariant variables to be included among the regressors, which enables me to include variables such as “gender”. Although I am aware of the fact that the coefficients are more likely to be biased when using random effects compared to a situation in which fixed effects are used, using random effects does often lead to smaller standard errors. However the estimates may be biased because we are not controlling for omitted variables. Thus, by using random effects we assume that the error terms are uncorrelated with the X variables but vary (randomly) across individuals and therefore can be considered part of the residual. Random effects assume a normal distribution and due to my large sample size, the assumption of normality is not critical (Yaffee, 2003).
Regressions with both robust and clustered standard errors were run. The reason for using cluster standard errors is that when standard errors are clustered OLS estimates are unbiased, but standard errors are not i.d.d. There are 117 different countries used in the dataset. Banks located in
18 the same country may not have independent standard errors, so clustered standard errors based on country are used. So the use of clustered standard errors assumes that observations within the same country are correlated in some unknown way, inducing correlation in the standard errors within each country, but that banks in different countries do not have correlated errors. When using clustered standard errors the significance level of the regressions decreases substantially. For this reason the regressions using clustered standard can be found in the appendix and the regressions with robust standard errors are ran in section 6.*
6.1 Regression
To answer my research question I have ran several regressions**. Regression I: Yield on Gross Portfolio
This measure is used as a proxy for the interest rate charged. Although it is inferior in the eyes of the borrower to the simply using the interest rate, it is highly likely that interest yield trends and interest rate trends would move roughly in parallel over a span of years. In determining the factors that drive the interest rate charged, we follow a report by CGAP (Rosenberg et al, 2013). In their paper the interest rate is determinant by the costs of funding (financial expenses), loan loss expense (risk coverage), operating costs and profits. Operating expenses make up the greater part of costs and this expense is the largest determinant of the interest rate the borrowers end up paying. The real yield on gross portfolio is chosen instead of the nominal yield in order to eliminate the influence of the difference in inflation rates between regions. The impact of a high profit margin on the yield is corroborated by Kneiding & Rosenberg (2008).
It is expected that for profit institutions charge higher interest rates than their nonprofit counterparts, due to the higher incentive for making profit.
Dependent Variable Yield on Gross Portfolio Independent Variables Profit Status (for-profit)
Financial expense to total assets Operating expenses to total assets Risk coverage
Portfolio at risk / Arrears
Average loan balance per borrower / GNI per capita
Profit margin
**For more information on the variables and abbreviations see section 4.2 and Appendix 6. *See Appendix 8
19
CE Coef. Robust Std. Err.
z P>I z I (95% Conf. Interval)
BV2 0.0442205 0.009498 4.66 0.000 0.0256047 0.0628363 W 0.1273987 0.2881057 0.44 0.658 -0.4372782 0.6920756 Y 0.4593937 0.2347778 1.96 0.050 -0.0007623 0.9195496 BY -0.0000257 0.0000198 -1.30 0.194 -0.0000645 0.0000131 AD 0.014637 0.007337 1.99 0.046 0.0002566 0.0290173 P -0.0014129 0.0014586 -0.97 0.333 -0.0042716 0.0014459 BU 0.0935767 0.0298507 3.13 0.002 0.0350703 0.1520831 _cons 0.1323227 0.0482085 2.74 0.006 0.0378359 0.2268095
The profit status of the microfinance institution has a significant effect on the yield on gross portfolio. A for-profit institution has a 4.42% higher yield, with a standard error of 0.95%. Higher operating expenses to total assets are significant at a 5% level and also lead to a higher yield, which makes sense since the operating expenses make up the largest part of the interest rate (Guntz, 2011). A higher profit margin is also reflected in a higher yield. Another interesting find is that a riskier portfolio has a higher yield on gross portfolio. The other variables are not significant at a 10% level.
Regression II: Percent of female borrowers
Microfinance has the clear mission of using credit, savings and other financial products to increase the quality of living for households and help them better cope with risk. Microfinance targets the poorest in the world, in particular women. Microfinance has the potential to make a significant contribution to gender equality and promote sustainable livelihoods and better working conditions for women. One way to measure whether a microfinance institutions suffers from mission drift, is to measure the number of female borrowers.*
Previous research has shown that microfinance companies with more female clients are associated with lower portfolio-at-risk, lower write-offs, and lower credit-loss provisions, ceteris paribus. This corroborates the common belief that women are less risky borrowers. There is a higher percentage of female borrowers among microfinance companies that target the poorest and are situated in low trust countries. If the microfinance company is located in Latin America, the percentage of female borrowers decreases (Aggarwal et al., 2012). When estimating the effect of profit status on percent of female borrowers “percent of female loan officers” to account for the effect of female loan officers choosing female borrowers. Finally, Aggarwal et al. (2012) found that nonprofit microfinance companies serve a higher percent of female borrowers. This leads to our assumption: “Non-profit microfinance institutions serve a higher percentage of female borrowers compared to microfinance institutions with a for profit status.”
*Mission drift is better measured by looking at the character of the villages, towns, and neighborhoods where the microfinance institution is opening its new branches. However this is a very expensive and time consuming procedure.
20 It is expected that for non-profit institutions serve a higher percentage of female borrowers than their for-profit counterparts, due to their more pronounced social mission.
Dependent Variable Percent of female borrowers
Independent Variables Profit Status (for profit)
Percent of female loan officers
Region (BW) Africa (1) East Asia and the Pacific (2) Eastern Europe and Central Asia (3) Latin America and the Caribbean (4) Middle East and North Africa (5) South Asia (6) BL Coef. Robust Std. Err. z P> I z I (95% Conf. Interval) BV2 -0.0439004 0.0169886 -2.58 0.010 -0.0771975 -0.0106033 BM 0.107201 0.0189032 5.67 0.000 0.0701513 0.1442507 BW1 -0.3356201 0.0338792 -9.91 0.000 -0.4020221 -0.269218 BW2 -0.1676502 0.0359012 -4.67 0.000 -0.2380152 -0.0972852 BW3 -0.4614445 0.0326298 -14.14 0.000 -0.5253977 -0.3974912 BW4 -0.3452572 0.0289631 -11.92 0.000 -0.4020237 -0.2884906 BW5 -0.4015907 0.0423278 -9.49 0.000 -0.4845517 -0.3186297 _cons 0.9147397 0.0285607 32.03 0.000 0.8587618 0.9707175
The profit status of a microfinance institution is significant at a 10% level. A for profit institution has 4.39% less female borrowers than a non-profit institution. This is in line with my hypotheses. A higher percentage of female loan officers has a highly significant positive effect on the percentage of female borrowers. As for the region the institution is located in, South Asia (BW6) has the highest percentage of female borrowers, followed by Asia and the Pacific, Africa, Latin America and the Caribbean the Middle East and North Africa and finally Eastern Europe and Central Asia.
Regression III: Operating expenses to total assets
This measure is used to capture good management. Due to the fact that microfinance is such a labor intensive market, operating costs are the main driver of microfinance institutions’ interest rates. Operating expenses cover mostly administrative and personnel costs and these costs are largely transferred to borrowers in the form of interest rates (Guntz, 2011).
21 For example, if a microfinance institution has a reduction of 5% in operating expenses, while losses from defaults and financial expenses don't change, the institution has the choice of either decreasing the interest rate with 5% or increase profits with 5%, or any combination of the two that adds to a total change of 5% (Gonzalez,2007). Salaries make up the largest part of operating expenses and fixed costs are relatively low compared with variable costs (Rosenberg, et al., 2009). Operating expenses make up close to 50 percent of nominal interest yields. Countries with lower operating expenses generally have lower interest rates as well. The age of a microfinance institution also has an effect on operating expenses. For microfinance institutions less than six years old, an additional year in the market expected to reduce the operating expense ratio between two and eight percentage points (Gonzalez, 2007). Additionally, loan sizes have an effect on operating costs. High operating costs are observed in Mexico, where the operating costs make up almost 45 percent of gross loan portfolio. The reason for these high operating causes is most likely the small average loan sizes (Navajas & Tejerina, 2006), relatively high costs of qualified labor and high transport expenses since the microfinance industry in Mexico serves mostly rural areas with low population density (Kneiding & Rosenberg, 2008).
It is expected that the operating expenses to total assets are lower for for-profit institutions compared to their non-profit counterparts due to their increased incentive to work efficiently.
Dependent Variable Operating expenses to total assets Independent Variables Profit Status (for-profit)
Average loan balance per borrower Borrowers per staff member Average salary / GNI per capita
Region (BW) Africa (1) East Asia and the Pacific (2) Eastern Europe and Central Asia (3) Latin America and the Caribbean (4) Middle East and North Africa (5) South Asia (6) Yield on gross portfolio
Age (AK) New (1) Young (2) Mature (3)
22
Y Coef. Robust Std. Err.
z P>I z I (95% Conf. Interval)
BV2 -0.0201087 0.0138702 -1.45 0.147 -0.0472939 0.0070764 P -0.0015531 0.0008449 -1.84 0.066 -0.0032091 0.0001029 AB -0.0000601 0.0000268 -2.24 0.025 -0.0001127 -7.50e-06 AP 0.0048081 0.0013819 3.48 0.001 0.0020997 0.0075162 BW1 0.0179876 0.0127998 1.41 0.160 -0.0070995 0.0430747 BW2 0.0237165 0.008767 2.71 0.007 0.0065335 0.0408996 BW3 0.0001508 0.0067056 0.02 0.982 -0.0129918 0.0132934 BW4 0.0852576 0.0263912 3.23 0.001 0.0335319 0.1369833 BW5 -0.004513 0.0096776 -0.47 0.641 -0.0234807 0.0144547 CE 0.2648698 0.0382619 6.92 0.000 0.1898778 0.3398617 AK2 0.0245606 0.0061659 3.98 0.000 0.0124757 0.0366454 AK3 0.0086883 0.0032278 2.69 0.007 0.0023619 0.0150146 _cons 0.080016 0.0086496 9.25 0.000 0.0630631 0.0969688
The effect of the profit status on the operating expenses is not significant. Variables which are interesting and significant include the negative effect of Average loan balance per borrower, which is expected. However surprisingly, the more borrowers per staff member, the higher the operating costs. This may be due to staff members taking on a lot of borrowers not because they are efficient and can handle manage the amount of borrowers but because there is not enough staff. The average salary has a positive effect on operating costs, which can be interpreted as the higher the salary offered the better the staff works. Higher salaries usually attract higher educated people and harder workers. A higher yield is also positively related to operating costs. This is expected because operating costs make up the largest part of the yield, so the higher the operating costs the higher the interest rate.
Younger institutions also have significantly higher operating costs than more mature institutions, which makes sense because older institutions have had more time to increase efficiency and lower costs.
Regression IV: Cost per borrower
This variable is another way to measure good management. There are different aspects which affect the cost per borrower. The first is a difference in salaries and costs of living across regions. For this reason the location of the microfinance institution should be taken into account when estimating the cost per borrower. Secondly the average loan size has an effect on the cost per borrower. The smaller the loan, the more expensive it is per dollar lent. To take this into account “Average Loan Balance per Borrower/GNI per capita” is used. Socially responsible microfinance
23 institutions who want to be sustainable businesses and deliver their services in the future must charge enough to cover costs.
Additionally the age and the scale of the institution cannot be ignored. If the loan sizes, age and scale of the microfinance institution are not taken into account, smaller, younger institutions or institutions providing smaller loans, usually to the poorest in the community show higher costs, this can put these institutions is a bad light and appear to be very inefficient institutions. Finally there are several other variables associated with operating costs which cannot be accounted for in this paper. A difference in lending methodologies can lead to differing operating costs, for example group lending is likely to be less costly than with individual lending methods. Unfortunately this data is not readily available. Lastly, the provision of other financial on non-financial services are also linked to efficiency and operating costs. Many case studies have analyzed these issues, including training for micro entrepreneurs, education and remittances, but there has been no consensus on the impact of these services on efficiency (González, 2011). Regulation also plays a role in the cost per borrower, Smith (2011) researched the effect of regulation on the microfinance industry and concluded that competition negatively affects the performance of microfinance institutions, in terms of both cost expended per borrower and profit margin. Regulation enhances these negative effects of competition.
It is expected that the costs per borrower are lower for for-profit institutions compared to their non-profit counterparts due to their increased incentive to work efficiently.
Dependent Variable Cost per borrower
Independent Variables Profit Status (for-profit)
Average salary / GNI per capita
Region (BW) Africa (1) East Asia and the Pacific (2) Eastern Europe and Central Asia (3) Latin America and the Caribbean (4) Middle East and North Africa (5) South Asia (6) Average loan balance per borrower / GNI per capita Age (AK) New (1) Young (2) Mature (3) Regulated (BX) No (1) Yes (2)
24 AA Coef. Robust Std. Err. z P > I z I ( 95 %Conf. Interval) BV2 93.5618 24.37661 3.84 0.000 45.78453 141.3391 AP 9.998001 7.683786 1.30 0.193 -5.061943 25.05795 BW1 92.05105 55.33051 1.66 0.096 -16.39475 200.4968 BW2 75.05906 14.696 5.11 0.000 46.25543 103.8627 BW3 416.6565 40.52889 10.28 0.000 337.2214 496.0917 BW4 266.5384 21.01895 12.68 0.000 225.342 307.7348 BW5 148.812 25.0711 5.94 0.000 99.67354 197.9504 P 20.02822 16.72236 1.20 0.231 -12.747 52.80344 AK2 -34.07305 15.53743 -2.19 0.028 -64.52585 -3.620258 AK3 -51.27662 7.524592 -6.81 0.000 -66.02455 -36.52869 BX2 1.255949 19.9361 0.06 0.950 -37.81809 40.32999 _cons -33.48987 14.99591 -2.23 0.026 -62.88131 -4.098427
Profit status has a significant positive effect on the cost per borrower. Newer institutions have lower costs per borrower.
Regression V: Borrowers per staff member
Productivity is seen as combination of both outreach and efficiency. It is usually measured in terms of borrowers per staff member. The goal of productive microfinance institutions is to maximize services with minimal resources, including staff and funds (Lafourcade ,et al. 2005). Productivity is important because of its effects on profitability. Woller and Schreiner (2002) found that productivity is a determinant of financial sustainability. Furthermore, through a study on rural microfinance in Tanzania, Ganka (2010) found a negative and strongly statistically significant relationship between number of borrowers per staff and financial sustainability. He stated that because of the inefficiency of rural microfinance institutions they risk becoming unsustainable when the business grows. He justified that microfinance staff for rural MFIs are not efficient as a result they fail to manage borrowers when their number grows causing microfinance institutions’ unsustainability. On the contrary, Christen et al. (1995) found no association between productivity and financial sustainability.
The region the institution is located is taken into account, because previous research has found firms operating in a more competitive environment exhibit higher measured productivity (6) (Schiffbauer & Ospina, 2010). Wages and firm size have also been found to effect productivity (Fallahi et al., 2010; Leung et al.,2008; Batt, 2002). In this paper we use “number of active borrowers” as a proxy for firm size. It should also be taken into account that the productivity of new firms that just entered a market tends to be lower than average young firms generally have lower than average productivity. After entering the market, these firms usually catchup with more mature firms within a few years, or they exit the market (Brouwer et al., 2005).
25 It is expected that the borrowers per staff member are lower for for-profit institutions compared to their non-profit counterparts due to their increased incentive to work productively.
Dependent Variable Borrowers per staff member
Independent Variables Profit Status (for-profit) Age (AK)
New (1) Young (2) Mature (3)
Number of active borrowers
Region (BW) Africa (1) East Asia and the Pacific (2) Eastern Europe and Central Asia (3) Latin America and the Caribbean (4) Middle East and North Africa (5) South Asia (6) Average salary / GNI per capita
AB Coef. Robust Std. Err.
z P>I z I (95% Conf. Interval)
BV2 -9.94805 8.865164 -1.12 0.262 -27.32345 7.427352 AK1 2.94423 5.715812 0.52 0.606 -8.258556 14.14702 AK2 -2.774201 2.640577 -1.05 0.293 -7.949637 2.401235 AG 0.0000407 0.0000126 3.22 0.001 0.0000159 0.0000654 BW1 -92.01517 15.19005 -6.06 0.000 -121.7871 -62.24321 BW2 -82.59334 11.742 -7.03 0.000 -105.6072 -59.57944 BW3 -136.1836 9.950061 -13.69 0.000 -155.6854 -116.6819 BW4 -76.98929 15.25811 -5.05 0.000 -106.8946 -47.08394 BW5 -88.31566 12.36559 -7.14 0.000 -112.5518 -64.07956 AP 3.396053 1.348906 2.52 0.012 0.752246 6.039859 _cons 196.6173 12.96434 15.17 0.000 171.2077 222.027
Profit status does not seem to have a significant effect on borrowers per staff member. The more borrowers an institution has, the higher the amount of borrowers per staff member, although at an increase of 0.004% this effect seems negligible. Additional the average salary is positively related
26
* Portfolio at risk / arrears 30 days is used. See Appendix 7 for the regression using Portfolio at risk / arrears 90 days.
to the amount of borrowers per staff member, which can be interpreted as the higher quality labor the institution attracts (due to higher salaries), the more efficiently they work and the more borrowers an employee manage.
Regression VI: Portfolio at risk / Arrears*
The most popular way to measure the quality of a portfolio in the microfinance sector is portfolio at risk. There are several variables that should be considered when estimating portfolio at risk. The first is gender of the loan officers. After doing research on the effect of differences in gender on loan officers performance, Beck et al. (2013) found that female loan officers incur less difficulty with handling loans compared to the male loan officers, because of the better capability of female loan officers to build trusting relationships with borrowers. Another important factor in the estimation and comparison of portfolio at risk is differences in lending methodology. Some institutions offer group lending, which usually works best in rural areas where social control is high, while other microfinance institutions choose to only lend to individuals (Dieckmann et al., 2007). Unfortunately data on lending methodology is scarce and therefore I was unable to take this into account, other relevant variables affecting portfolio risk include both microeconomic and macroeconomic factors. The microeconomic variables include the number of loans made and the size of these loans. The number of active borrowers in the institution may proxy for the ability to monitor borrowers, and institutions that lack sufficient scale will have higher risks. Another microeconomic factor affecting portfolio at risk is the interest rate, which effects the ability of the borrowers to repay the loan. It is expected that microfinance institutions that charge high interest rates face higher risks of repayment. At the macroeconomic level differences in countries or regions affect the portfolio at risk. An economic crisis in a particular country or region tends to increase the portfolio at risk for all institutions in the area (Crabb & Keller, 2006). The age of the institution is also taken into account due to the effect experience, existing history with clients and a change in their internal procedures for stimulating repayment as they grow older, can have on the portfolio at risk (D’espallier et al., 2011).
It is expected that the portfolio at risk is lower for for-profit institutions compared to their non-profit counterparts.
27
Dependent Variable Portfolio at risk
Independent Variables Profit Status (for-profit)
Percent of female loan officers Age (AK) New (1) Young (2) Mature (3) Region (BW) Africa (1) East Asia and the Pacific (2) Eastern Europe and Central Asia (3) Latin America and the Caribbean (4) Middle East and North Africa (5) South Asia (6) Yield on gross portfolio
Average loan size per borrower / GNI per capita Number of active borrowers
AD Coef. Robust Std. Err.
z P>I z I (95% Conf. Interval)
BV2 0.0121412 0.0069187 1.75 0.079 -0.0014193 0.0257017 BM -0.008189 0.0101188 -0.81 0.418 -0.0280215 0.0116435 AK1 0.0159753 0.0061386 2.60 0.009 0.0039439 0.0280068 AK2 0.0054288 0.0108327 0.50 0.616 -0.0158028 0.0266605 BW1 0.0403294 0.0139313 2.89 0.004 0.0130246 0.0676343 BW2 0.0072761 0.0147123 0.49 0.612 -0.0215595 0.0361116 BW3 -0.0042399 0.0112269 -0.38 0.706 -0.0262441 0.0177644 BW4 0.01149224 0.0114395 1.00 0.315 -0.0109287 0.0339135 BW5 -0.0098092 0.0127182 -0.77 0.441 -0.0347363 0.015118 CE -0.0332711 0.0132859 -2.50 0.012 -0.0593109 -0.0072313 P -0.0006851 0.0003186 -2.15 0.032 -0.0013096 -0.0000606 AG 5.90e-09 1.08e-08 0.54 0.587 -1.54e-08 2.72e-08 _cons 0.051173 0.0111524 4.59 0.000 0.0293147 0.0730313
For profit microfinance institutions have a higher portfolio at risk compared to their non-profit counterparts. The higher the average loan balance per borrower, the lower the portfolio at risk.
28 However, the results also show that the higher the average loan balance per borrower, the lower the portfolio at risk, which may seem puzzling since for-profit institutions are shown to have a higher loan size per borrower.
Regression VII: Financial expenses to total assets
Microfinance institutions generally have high costs, especially in comparison to other credit industries. The interest rate charged to customers cover these financial costs, the operating costs and the profit margin. Financial costs are more difficult to be controlled by the microfinance institutions than operating costs (González, 2011). The financial costs measure the total interest expense the microfinance institutions incurs to fund its lending portfolio. These costs determine the minimum interest rate a microfinance institution has to charger to fund its financial expenses. Deposits play a role in these costs, microfinance that collect deposits generally have a lower financial expense ratio compared to a microfinance institution that borrows through commercial funds. Subsides also lead to lower financial costs.
It is expected that the financial expenses to assets are lower for for-profit institutions compared to their non-profit counterparts due to their increased incentive to work efficiently.
Dependent Variable Financial expenses to total assets Independent Variables Profit Status (for-profit)
Deposits
Region (BW) Africa (1) East Asia and the Pacific (2) Eastern Europe and Central Asia (3) Latin America and the Caribbean (4) Middle East and North Africa (5) South Asia (6) Yield on gross portfolio
29
* See appendix 2
W Coef. Robust Std. Err.
z P>I z I (95% Conf. Interval)
BV2 0.0007311 0.0022774 0.32 0.748 -0.0037325 0.0051947 DEP1 -0.002735 0.0015768 -1.73 0.083 -0.0058255 0.0003555 BW1 -0.0402647 0.0041969 -9.59 0.000 -0.0484905 -0.0320389 BW2 -0.0214397 0.0038052 -5.63 0.000 -0.0288978 -0.0139816 BW3 0.0093289 0.0045372 2.06 0.040 0.0004361 0.0182217 BW4 -0.0170093 0.0046788 -3.64 0.000 -0.0261795 -0.0078391 BW5 -0.0467236 0.0038677 -12.08 0.000 -0.0543041 -0.0391431 CE 0.0265513 0.0238792 1.11 0.266 -0.0202511 0.0733537 _cons 0.0635452 0.003375 18.83 0.000 0.0569304 0.0701601
There is no significant effect of the profit status on the financial expenses. However, a firm not taking deposits has 2% lower financial expenses compared to a firm taking deposits. The region in which the institution is located also has a significant effect on financial expenses, institutions in Europe and Central Asia have the highest financial expenses, followed by South Asia, Latin America and the Caribbean, Asia and the Pacific, Africa and finally the Middle East and North Africa.
Regression VIII: Average salary/ GNI per capita
The average salary/ GNI per capita differs per region as seen in the table below and is the lowest in the Middle Eastern and North African countries*. The size of the institution is also said to affect the staff’s pay (Zhou, 2000). Total Assets are used as a proxy to measure the size of the institution. It is expected that the average salary is higher for for-profit institutions compared to their non-profit counterparts.
Dependent Variable Average salary / GNI per capita Independent Variables Profit Status (for-profit)
Region (BW) Africa (1) East Asia and the Pacific (2) Eastern Europe and Central Asia (3) Latin America and the Caribbean (4) Middle East and North Africa (5) South Asia (6) Total assets
