1 Abstract The impact of financial regulation on the efficiency of microfinance institutions:

(1)

The impact of financial regulation on

the efficiency of microfinance institutions Nguyen Ho Anh Khoa

The impact of financial regulation on the efficiency of microfinance institutions:

Evidence from Asia, Africa, and Latin America

Abstract

The number of microfinance institutions has increased significantly since the last decade. This rapid growth calls for financial regulation. The question is how to regulate microfinance institutions or whether the current financial regulation is appropriate. This paper examines the impact of financial regulation on the outreach, sustainability and technical efficiency of microfinance institutions. Based on the three-year-database of microfinance institutions in 26 countries across Asia, Africa, and Latin America, the empirical evidence indicate that financial regulation lowers the outreach of bank-type microfinance institutions but improves their sustainability. In contrast, the more restrictive financial regulation is associated with the higher technical efficiency of nonbank financial institutions and cooperatives/credit unions. The higher efficiency of these institutions, however, cannot offer guarantee for their sustainability since it declines under the pressure of restrictive financial regulation. Finally, the evidence suggest that the performance of NGO-type microfinance institutions is immune to the influence of financial regulation.

Key words: microfinance institutions, DEA, technical efficiency, financial regulation index, Tobit Model, Fixed Effects Model

Supervisor: Dr. Tra Pham Dr. Aljar Meesters

Nguyen Ho Anh Khoa Student number: 1791885 MSc.BA. Specialization Finance

Faculty of Management and Organization, University of Groningen

(2)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Contents

Abstract page 1

I. Introduction 5

II. Literature review 8

1. Theoretical framework 8

1.1. The theory of production for financial intermediations 8

1.2. Corporate governance and financial institutions 8

2. Empirical literature 12

2.1. Internal corporate mechanisms 13

2.2. External corporate mechanisms 13

2.3. Hypotheses 16

III. Methodology and data collection 19

1. Methodology 19

1.1. The measurement of technical efficiency by DEA model 19

1.1.1. The selection of inputs and outputs 19

1.1.2. DEA model 20

1.2. The composition of financial regulation index 22

1.3. Econometrics models 23

1.3.1. The Panel Tobit Model 23

1.3.2. The Fixed Effects Model 25

2. Data collection and variables 26

2.1. Outreach, sustainability, and technical efficiency 27

2.2. Financial regulation index 30

(3)

2. Banks 37

3. Nonbank financial institutions 39

4. Cooperative/Credit Unions 41

5. Nongovernmental organizations (NGOs) 42

V. Robustness check 47

VI. Conclusion and discussion 50

Appendix 1. The theory of production 53

Appendix 2. Descriptive statistics 57

Appendix 3. Robustness check with Random Effects Model 64

(4)

List of Tables, Graphs, and Figures

Table 1. The average technical efficiency by regions page 28

Table 2. The Average Technical Efficiency by top ten countries 29

Table 3. The average technical efficiency by types of microfinance institutions 30

Graph 1. The average financial regulation index by regions 31

Graph 2. The average financial regulation index by types 32

Table 4. The average restrictiveness of financial regulation index 33

Table 5. Tobit estimates to the entire sample 35

Table 6. Tobit estimates to Banks 38

Table 7. Tobit estimates to Nonbank financial institutions 40

Table 8. Tobit estimates to Cooperatives/Credit Unions 41

Table 9. Tobit estimates to NGOs 43

Table 10. The summary of the Tobit estimation 44

Table 11. Fixed Effects Model to Technical Efficiency 48

Figure 1. Technical Efficiency and Return to Scale 54

Figure 2. Input- Orientated and Output-Orientated Technical Efficiency 55 Table 12. The description of variables to Tobit model and Fixed Effects Model 57

Table 13. Descriptive statistics on variables of the entire sample 59

Table 14. Descriptive statistics on variables of the subsample of Banks 60

Table 15. Descriptive statistics on Nonbank Financial Institutions 61

Table 16. Descriptive statistics on Cooperative/Credit Unions 62

Table 17. Descriptive statistics on variables of the subsample of NGOs 63

Table 18. Correlation matrix of explanatory variables 64

(5)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa I. INTRODUCTION

Microfinance increasingly attracts public attention since its main objective is to provide credit to the poor. In 2007, microfinance institutions record important achievements of which over 80 million borrowers are served worldwide. The number of microfinance institutions reaches 3,529 at the same time1. Since microfinance refers to the provision of financial services to low-income borrowers, the operations of microfinance institutions are somehow similar to those of traditional financial intermediations that relates to the soundness of financial system as a whole. Therefore, financial authorities have a reason to put the activities of microfinance institutions under control. Relatedly, the recent developments of microfinance industry are called for regulation.

The debate on how to regulate microfinance institutions concerning its special characteristics is still on the table. Since most of the microfinance institutions work as credit providers, prudential regulation seems to be appropriate. The same regulations of risk management on banks, for instance, are applied on microfinance institutions (Jansson et al, 1998). From the nature of microfinance institutions perspective, these prudential regulations are necessary to insure the sustainability of microfinance institutions (Van Greuning et al., 1999). However, the question is whether the prudential regulation is still relevant when these microfinance institutions lend to the poor (Hardy et al., 2003). For instance, if joint liability is not accepted as collateral, more poor borrowers are obviously under credit rationing. Hence, microfinance institutions might not widen and deepen their outreach as they could (Cull et al, 2009). From the types of microfinance institutions perspective, these prudential regulations might be ineffective to all kinds of microfinance institutions. For example, NGOs and cooperative/credit unions are different from nonbank financial institutions or banks in terms of their operations. That might leads to the different exposures of these microfinance institutions to financial regulation (Cull et al., 2009).

(6)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa the sustainability of microfinance institutions. In the same vein, Mersland and Strom (2007) find the similar results that bank regulation has no significant effects on the sustainability and even the outreach of microfinance institutions. In contrast, Cull et al (2009) find the trade-off between outreach and sustainability of supervised microfinance institutions. In fact, profit-orientated microfinance institutions tend to lower the number and value of loans to women borrowers to maintain financial positions while nonprofit-orientated counterparties whose outreach remains unchanged find their sustainability deteriorating. Obviously, the empirical evidence partially shed a light on the possible impacts of financial regulation on the outreach and sustainability of microfinance institutions separately. Nevertheless, they still have not answered the question: “How those regulations affect the overall performance of microfinance institutions?”

(7)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa financial regulation. Not only can the aggregate financial regulation index determine whether microfinance institutions are regulated or not, it also quantifies to which degree they are regulated. Thereby, the sketches of financial regulation on microfinance institutions across countries can fit reality well. Consequently, the evidence of the impact of financial regulation on microfinance institutions’ performance is expected to be more reliable than that of previous studies.

Third, by using various estimation techniques, I examine the impact of financial regulation on the performance of each type of microfinance institutions including bank, nonbank financial institution, cooperative/credit union, and non-governmental organization (NGO). Thereby, my study can provide rich informative empirical evidence on the field relative to those of Hartarska (2005) or Mersland and Strom (2008) which do not examine each exposure of different kinds of microfinance institutions to corporate governance. The evidence are expected to answer the question of whether banking regulation is relevant to various kinds of microfinance institutions or not. Since one cannot cover the sky with a hand, the paper attempts to figure out the possible influential biases of banking regulation on the performance of each type of microfinance institutions.

(8)

II. LITERATURE REVIEW

1. Theoretical framework

There are two main theories underpinning this paper. First, the theory of production relates to the measurements of financial institution’s performance. One of them exploited in this paper is the concept of efficiency. Second, the theory of corporate governance is used to explain the relationship between financial regulation and the efficiency of microfinance institutions.

1.1. The theory of production for financial intermediations

In this section, I present the theory of production in the way that the concept of efficiency is applicable to financial intermediation. Since the difference in measurement of efficiency between financial institutions and firms comes from the selection of inputs and outputs, it is necessary to present the approaches to the specification of such the selection: production, intermediation, and user-cost approach.

The theory of production involves measuring the performance of firms, which use inputs (capital, labor, etc.) to produce outputs (goods, services, etc.). There are two traditional measurement of the firms’ performance: efficiency and productivity. Since efficiency accounts for the economics of scale, it outperforms productivity to assess the firms’ performance in comparative studies (Coelli et al, 2005)2. Given its advantage, academic researchers apply the concept of efficiency for financial institutions. However, this application triggers a vast critical literature on the field since financial institutions cannot be considered as traditional firm. Actually, the inputs (deposits, labor, plant, etc.) and outputs (loans, financial services, etc.) of financial institutions are different from those of production firms (Sealey and Lindley, 1977).

(9)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa efficiency score, literature on financial institutions’ efficiency debate on how to specify these selections. Ferrier et al. (1990) use the production approach to measure the efficiency of financial institutions. They consider that financial institutions use physical resources such as employees and plants to conduct transactions, take deposits, make loans, and so on. Under this approach, labor and assets are considered inputs while deposits, loans, and services are considered outputs. Since the production approach treats financial institutions like production firms, the unique function of financial institutions as intermediaries is ignored (Sealey and Lindley, 1977).

From the point of view of financial intermediaries, Sealey and Lindley (1997) propose the intermediation approach that embeds the intermediary function of financial institutions to measure their efficiency. Thereby, deposits and funds, which financial institutions obtain from the markets, are treated as inputs while placing loans they make are treated as outputs. Nevertheless, the intermediation approach has its own drawback since it does not capture the other performances of financial institutions such as payment services, financial consultancy, etc. (Favero et al, 1995). The debate on whether production or intermediation approach should be used is still ongoing. Specifically, academic papers still question whether deposits should be considered as input or output (Berger and Humphrey, 1997).

(10)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa In general, the treatment of deposits to measure the efficiency of financial institutions is on debate. It is a sensitive matter since different selections of inputs and outputs lead different outcomes of efficiency. As long as the specification of inputs and outputs are relevant to the particular objectives of each study, the theory of production can be applied to financial institutions (Nieto et al, 2007).

1.2. Corporate Governance and Financial Institutions

The concept of information asymmetry plays a central role in the theory of corporate governance. It refers to the situation in which stakeholders such as shareholders, government, etc. do not have full information about hired managers’ performance because of the separation of ownership and control in modern corporations (Jensen and Meckling, 1976). This leads to the so-called adverse selection and moral hazard problems. Adverse selection refers to the situation in which bad or risky projects are likely to be funded since stakeholders do not have full information about the projects. It plays a role ex ante since it happens before the investment is funded. Moral hazard involves the misbehavior of hired managers, especially when they act for their own benefits instead of the benefits of stakeholders. It plays a role ex post since it occurs after the investment decisions are made (Tirole, 2006). Therefore, when moral hazard and adverse selection occur, stakeholders bear the costs of running bad projects and being expropriated by managers. In corporate governance literature, these costs are called agency costs. Based on the concept of agency costs, Jensen and Meckling (1976) define corporate governance as external and internal mechanisms that are used to reduce those costs, thereby, enhance the wealthy of corporations.

(11)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa depositors and creditors bear the higher cost of bankruptcy if they do not have information about the financial institutions’ performance. In other cases, shareholders might find their returns from the investment to financial institutions lower than the opportunity cost of funds due to the lack of managers’ experiences. Likewise, government runs the systemic risks since the collapse of one financial institution might lead to the failure of other institutions and of the financial system (De Bandt and Hartmann, 2000). Second, moral hazard in financial institutions happens when managers misbehave, for example, in making loans. They place the loans to borrowers based on their private scheme but not on the rational approach since they want to obtain the social relationships or higher positions, etc. (Macey and O’Hara, 2003). The problems of moral hazard in modern financial institutions might be exacerbated because of financial innovation. For instance, under the pressure of shareholders, managers invest in excessively risky projects or financial derivatives to earn higher profit, thereby, their positions are guaranteed. From the point of view of creditors, these decisions damage the safety of their claims since creditors are known as fixed claimants (Tirole, 2006). Generally, these arguments based on the problems of asymmetric information indicate the need for corporate governance in financial institutions. The remaining issue is that whether existing corporate governance mechanisms fit the content of financial intermediation or not. If these mechanisms can be applied on financial institutions, how they work and what expected effects they might have on the performance of these institutions.

(12)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa financial institutions entering the financial systems. Since these small institutions are vulnerable to the economics fluctuations, the barrier imposed by minimum capital policy can reduce the likelihood of default of these institutions. From the stakeholders’ perspective, it obviously alleviates the agency cost of adverse selection problem. From the point of view of financial institutions, the regulation protects them from the spillover effects; hence maintain sustainability, when low capacity financial institutions collapse (Jansson et al, 1998). Similarly, the agency cost created by moral hazard problems can be alleviated since some financial regulations are designed to align the incentives of managers. For example, the policy of risk management proposed by Basel II committee prevents the managers of financial institutions taking excessive risks. Thereby, the safety of depositors’ money is insured. Since risky investments are controlled, the losses of defaults can be reduced so that the sustainability of financial institutions is enhanced (De Bandt and Hartmann, 2000). However, financial institutions that bear higher cost to comply with financial regulation attempt to compensate those cost through higher interest rates or fees. It accidently creates credit rationing to borrowers, hence, reduce the outreach of those institutions (Jansson et al, 1998).

In conclusion, the theory of corporate governance is necessary in the context of financial institutions especially when the problems of asymmetric information are exacerbated in these institutions. As a result, it makes the external corporate governance mechanisms to be far more important to financial institutions than to corporation. It is demonstrated by the presence of financial regulations, which are considered as instruments to control for the problems of adverse selection and moral hazard in financial institutions. Specifically, restrictive financial regulation enhances the sustainability but accidently reduce the outreach of financial institutions.

2. Empirical Literature on the Impact of Corporate Governance on Microfinance Institutions

(13)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Internal corporate governance mechanisms

Hartarska (2005) presents the first evidence on the impact of corporate governance on outreach and sustainability of microfinance institutions in Central and Eastern Europe and the Newly Independent States. The paper finds that performance-based compensation of managers does not improve the performance of microfinance institutions. The lower the salary of managers in non-governmental organization-type microfinance institutions, the lower is the outreach of these institutions. In contrast, the evidence suggests that the managers’ experience enhance the performance. The paper also indicates that the tradeoff between outreach and sustainability appears when there is a limited employee participation in the boards of directors.

In the same vein, Mersland and Strom (2008) examine the effects of internal corporate governance practices such as the boards of directors, managers’ characteristics, the ownership structure, customer-microfinance institution relationship, competition and regulation on outreach and sustainability of microfinance institutions. The evidence suggest that sustainability of microfinance institutions is improved with local directors, an internal board auditor, and with female managers. Likewise, the number of borrowers increases when managers are also chairman in the institutions. Outreach is lower when microfinance institutions lend to individuals than is it in the case of group lending. There is no difference in outreach and sustainability between non-profit organizations and profit-orientated microfinance institutions.

External corporate governance mechanisms

(14)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Jansson et al (1997) suggests the potential impact of various financial regulations on microfinance in Latin America and the Caribbean. The author provides the two categories of financial regulations imposed on microfinance institutions. The first category involves with the prudential regulation that pursues the objectives of maintaining financial stability of the economy and of protecting the safety of depositor’s money. It concludes of the requirements of capital adequacy, standard loan documentation and provisioning. According to the Basel Committee in 1988 (Basel II), financial institutions have to maintain a minimum capital adequacy ratio of 8%. To calculate this ratio, assets are weighted with their associated risks. Since microloans are considered as the most risky assets, authority tends to impose the higher capital adequacy to microfinance institutions. Similarly, authorities might also require microfinance to create provision for the expected losses. This regulation is underpinned by the potential existence of credit risk, which is particularly high for microloans. However, it is irrelevant if the provision is set too high, which may burden the financial positions of the institutions. To collect data of the soundness of financial systems, authorities often requires that microfinance institutions to report their performance in every certain period. It might be difficult for the staff because microfinance involves with many small loans.

(15)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Obviously, on the negative side, interest rate ceiling policy not only makes microfinance institution difficult to cover their high cost, but also induce microfinance institutions to pay high cost for screening borrowers (Cull et al, 2009).

The Grameen Bank in Bangladesh promotes the method of using joint liability as loan collateral, which is especially meaningful in microfinance. Joint liability groups replace physical collateral successfully since each one of the participants loses both money and future access to credit if the group as a whole cannot meet its obligations. In some countries, authorities have not accepted joint liability as loan collateral in the microfinance industry since it is a new concept. MIC Survey proposes that financial authorities underestimate this form of collateral because there are no real assets of sufficient value behind the participants (Christen et al, 2003).

Minimum capital requirements determine the structure of financial system. Low capital requirement might allow a large number of microfinance institutions to join to the financial system. However, bank supervisors might be less enthusiastic about the prospect of having a large number of new microfinance institutions since their activities requires labor intensive and sophisticated supervision. In addition, authorities argue that microfinance institutions do not have enough assets diversification, therefore are more vulnerable to economic fluctuations (Jansson et al, 1997).

(16)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa women as well. The paper suggests that these microfinance institutions tend to curtail their outreach to maintain the financial positions. In contrast, non-profit orientated microfinance institutions do not reduce their loan sizes as well as less make loans to women borrower when they comply with prudential supervision. However, they have to pay the cost of which reduces their profitability.

These papers examine the impact of corporate governance on profitability and outreach separately. Profitability refers to return on assets while the number of borrowers or the gross loans portfolio often represents outreach. They find the evidence of the tradeoff between two dimensions. However, I suggest that sustainability and outreach affect each other simultaneously. It is necessary to have an aggregate measure of performance that combine the sustainability and outreach of microfinance institutions. In addition, the measurement of financial regulations should be scrutiny since various restrictions will have different impact on the performance of microfinance institutions. Therefore, the dummy variables that refer to whether microfinance institutions are supervised prudentially or not might be not sufficient. To filling those gaps on the field, I examine the impact of financial regulation on outreach, sustainability, and technical efficiency of microfinance institutions across regions and types. Thereby, I test the following hypotheses:

Hypothesis 1: With regardless to the type of microfinance institutions, I propose that restrictive financial regulations improve sustainability but curtail outreach of microfinance institutions. As a result, the technical efficiency as an overall performance is worsened under the pressure of financial regulation.

(17)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Hypothesis 2: Under the pressure of financial regulation, the sustainability of banks is improved but the outreach is reduced, hence, their technical efficiency is damaged consequently.

Although nonbank financial institutions have the same functions with commercial banks such as providing credit and leases, these institutions mainly, focus on long-term loans. Therefore, they have the tendency to increase the interest rate to compensate for the higher risk of long-term financing. It is irrelevant if financial authorities impose the policy of interest rate ceiling on these institutions (Hossain et al, 2005). It is equivalent with more restrictive financial regulation might reduce the sustainability of nonbank financial institutions. Consequently, it is likely that these institutions might also curtail their outreach, hence the technical efficiency, under the pressure of restrictions.

Hypothesis 3: The restrictive financial regulations have negative impacts on the outreach, sustainability, and technical efficiency of nonbank financial institutions.

Cooperatives/ Credit unions differ from banks and other financial institutions in that the members who have accounts in the credit union are the owners of the credit union and they elect their board of directors in a democratic one-person-one-vote system regardless of the amount of money invested in the credit union3. This kind of microfinance institutions is considered to have a good internal corporate governance mechanisms since the agency problems are reduced significantly (CUNA Model Credit Union Act, 2007). Therefore, the more restrictive financial regulation might be negative influential on the outreach, sustainability, and technical efficiency of these institutions.

Hypothesis 4: The restrictive financial regulations have negative impacts on the outreach, sustainability, and technical efficiency of cooperatives/credit unions.

Nongovernmental organizations (NGOs) differ from the other financial institutions to the extent that they work on the not-for-profit basis. These organizations, which often

3

(18)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa receive funding from donations, pursue the objective of providing credit to the poor. The empirical evidence indicate that they are not affected by the government regulations in terms of outreach and sustainability (Cull et al, 2009).

(19)

III. METHODOLOGY AND DATA COLLECTION

1. Methodology

To examine the relationship between financial regulation and the outreach, sustainability, and technical efficiency of microfinance institutions, I use three steps of methodology. First, Data Envelopment Analysis is applied to measure the technical efficiency of microfinance institutions. Second, I compose the index of financial regulation on microfinance institutions for each country. Third, I use Panel Tobit regression model to explain the impact of financial regulation index on the technical efficiency from the Data Envelopment Analysis. In addition, I perform the robustness check to the results of Panel Tobit Model by using Fixed Effects Model.

1.1 The measurement of technical efficiency by DEA model

The selection of inputs and outputs

The selection of inputs and outputs plays a crucial role in the outcomes of efficiency scores. Based on the approach propose by Nieto et al (2007), I use of the input and output selection which is the combination of outreach and sustainability of microfinance institutions. First, the number of credit officers (outreach) and the ratio of expenses to total assets (sustainability) represent the inputs while the number of active borrowers (outreach), the gross loan portfolios (outreach), and the ratio of revenue to total assets (sustainability) delegate the outputs.

(20)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa microfinance institutions in their paper are non-governmental organizations, which are not allowed to take deposits from public. Despite the existence of banks in my paper, the approach from Nieto et al (2007) is still the best choice since the number of banks does not account for a high percentage in the data collection.

The DEA model

As presented in Section 2.1, the procedure to measure the efficiency using Data Envelopment Analysis begins with the estimation of frontier. There are many frontiers such as the production frontier, the cost frontier, the revenue frontier, etc. Since the paper uses technical efficiency as the performance of microfinance institutions, the production frontier is estimated. In addition, since the database used in this paper does not include the information on input prices, allocative efficiency cannot be measured. In this paper, production frontier is constructed under the assumption of constant return to scale. Under this constraint, all microfinance institutions are assumed to operate at the optimal scale. It implies that the size of microfinance institutions is not account for comparing the efficiency score. Small microfinance institutions can produce outputs with the same productivity ratios as those of large microfinance institutions. It is equivalent to the assumption of which economies of scale does not exist. Hence, the amount increases in inputs lead to the same amount increases in outputs.

(21)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Next, the technical efficiency is defined as the distance from the production point of each microfinance institution to the production frontier. However, as illustrated in Section 2.1, the input- and output-orientated technical efficiency are equal under the assumption of constant return to scale.

The Data Envelopment Analysis requires the information on input and output of microfinance institutions. Given its advantages, Data Envelopment Analysis has some drawbacks that might lead to the inappropriate outcomes. A major problem is of the selection of inputs and outputs in a DEA model since there is no statistical test for it. The correlation between inputs and outputs lead to the inconsistent efficiency (Parkin and Hollingsworth, 1997). In fact, Jenkins and Anderson (2003) find the evidence that the efficiency score increases when the more inputs and outputs are added in the model.

Based on the selection of input and output introducing by Gutierrez-Nieto, Serrano-Cinca and Molinero (2005), I calculate the technical efficiency by using the number of credit officers and operating expenses as inputs and income, gross loan portfolio and the number of loans outstanding as outputs. Given the unique characteristic of microfinance institutions, the selection of input and output should be the combination of the indicators of sustainability and outreach.

The following Data Envelopment Analysis model aims to maximize outputs from a fixed quantity of inputs (output orientation). It comprises a linear programming model:

Max θ,λ θn (1)

Subject to - θnyin + ∑λjyij ≥ 0 i = 1,…,I (2)

xkn - ∑λjxkj ≥ 0 k = 1,…,K (3)

λj ≥ 0 j = 1,…,N (4)

where 1≤ θn ≤ ∞, and θn – 1 is the proportional increase in outputs that could be achieve

(22)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa efficiency score is defined as 1/θ that varies between zero and one. N denotes the number of microfinance institutions that produce I different outputs (yin represents the

observed amount of output I for bank n) from using K different inputs (xkn represents

the observed amount of input k for bank n). The λj are weights applied across the N

banks.

The linear program is solved N times for every microfinance institution in the sample to obtain the technical efficiency score for the nth microfinance institution, θn*. It is the

largest number of θn that satisfy the three constraints listed above. These constraints

ensure that the technical efficiency cannot lie outside the feasible set that ranges between zero and one.

1.2 The composition of financial regulation index

In this part, I present the method to measure the financial regulation index, which is the proxy of external governance practice. The index is calculated for each country in the sample. It comprises six relevant financial regulations on microfinance institutions. The method is straightforward. I add one score to the index if the corresponding financial regulation is restrictive and zero otherwise. The definition of the restrictiveness of these financial regulations is presented later. Therefore, the country financial regulation index takes the value between zero (least restrictive) and six (most restrictive). Besides, the legal status of microfinance institutions determines the degree of restristiveness of financial regulation imposed on these institutions. Therefore, I measure the financial regulation index for each type of microfinance institutions in one country including bank, non-bank financial institution, cooperative or credit union, and non-governmental organization (NGO). The methodology is based on the studies of La Porta et al (1997, 1998, and 1999) and Gompers et al (2003) which provide the intuitive method to compose the corporate governance index. The components of financial regulation index are provided by of the study of Jansson et al (1997).

(23)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa investors rights. From the microfinance institutions’ perspective, the method is appropriate to my paper since financial regulations also determine the rights of microfinance institutions. From the authority perspective, my index reflects the restrictiveness of the financial regulation of one country compared to the others. The main advantage of this method is that it helps to examine better the relative impacts of different financial regulation than of each regulation. However, Larcker et al (2007) argue that the aggregate index might be difficult to interpret if the individual regulations do not measure the same underlying governance construct. On the other hand, if relevant financial regulation that has a significant impact is not included in the governance index, the aggregate index does not have sufficient explanatory power. The omitted variable problems seem to be relevant in this paper since there is only six financial regulations are used.

In sum, the aggregate financial regulation index of one country gets a score of one if its capital adequacy ratio is higher than 8% and zero otherwise. The same principle is applied on the case of minimum capital requirement and provisioning. If the capital requirements and provisioning exceed the average level of those in the sample, the index takes the value of one for each case and zero otherwise. Likewise, when a country applies interest rate ceiling policy on its microfinance institution, its financial regulation index will get a score of one and zero otherwise. The same calculation is to the case of loan standard documentation and the acceptance of joint liability as collateral. Therefore, the aggregate financial regulation index takes a value ranging from zero to six. The value of zero means that a country financial regulation is least restrictive compared to the other 26 countries. In contrast, the value of six implies the country has the most restrictive financial regulation.

1.3 Econometrics models to the relationship between financial regulation and the performance of microfinance institutions

The Panel Tobit Model

(24)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa performance of microfinance institutions using the data of 321 microfinance institutions from the period of 3 years. Tobin (1958) introduces this model to solve the problem of censored or truncated data. Since efficiency score lies between zero to one, Tobit model is appropriate to analyze the influence of potential factors to the efficiency score. Likewise, Tobit is also applied on the indicators of outreach and sustainability since they are not measure by random data4. It takes the form of the following latent variable model:

Yit* = αit + βFriit + ψXit + φCit + δEit + εit (1)

The dependent variable Yit represents the number of active borrowers (outreach), gross

loan portfolio (outreach), financial revenue ratio (sustainability), and technical efficiency of microfinance institution i at year t. If Yit* ≤ 0, the indicators of

performance of microfinance institution i at year t take the value of Yit = 0.

If Yit* ≥ a, Yit = a.

And if 0 < Yit* < a, Yit = Yit*

αit differs across cross-sectional units to capture specific microfinance institutions’

effects. Β, ψ, φ, and δ are the vectors of unknown coefficients, the ε/Xit are normally,

identically and independently distributed with mean, zero, and variance σ2.

(25)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa McDonald (2009) argues that efficiency score are not censored data, but fractional data. He suggests that using Tobit model in the case of efficiency score is not relevant. In response to the problem, McDonald (2009) and Hoff (2006) recommend using Ordinary Least Square (OLS) is sufficient to obtain the significant coefficients from the regression. Therefore, I will use OLS to check the robustness of the results provided by Tobit model.

The Fixed Effects Model

I use Fixed Effects Model to examine the impact of financial regulation on the outreach, sustainability, and technical efficiency of financial institutions separately. The outcomes are used to test whether outreach and sustainability are substitutes or complements as well as the representative role of technical efficiency as an overall performance of microfinance institutions. A fixed-effects analysis has the important advantage that it implicitly controls for unobserved microfinance institutions’ characteristics that affect the outreach and sustainability measure, but do not change over time. The model takes the form of:

Yit = αi + βFriit + ψXit + φCit + δEit + εit (1)

where the dependent variable Yit is the vectors of the technical efficiency, the number of

active borrowers and gross loan portfolio (outreach), and financial revenue (sustainability) of microfinance institutions. αi differs across cross-sectional units to

capture specific microfinance institutions’ effects.

Across the specifications, Friit is the vector of dummy variables, which represent the

(26)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa the country (Claessens et al, 2001). In contrast, the annual inflation rate captures the inefficiencies of financial institutions since it shifts up the costs of financial institutions such as the increase in interest margin (Grigorian et al, 2002). In addition, I also use the variable of total assets (Tait) to control for the effect of economics of scale on the technical efficiency. It is necessary since technical efficiency ignores the effect of firm’s size on its technical efficiency (Coelli et al, 2005).

A set of intercept dummy variables, Cit, is included to examine the differential effects of regions and specific year on the technical efficiency. Furthermore, I also use a set of interaction dummy variables, Eit, to account for the joint effects of financial regulation index and different types of intercept dummy variables on the technical efficiency. It is based on the argument that microfinance institutions expose differently to financial regulation due to their types, the regions, and the specific year they operate. Likewise, one might argue that microfinance institutions are not affected in the same way by financial regulation index with regard to their sizes. I use the interaction dummy variable of financial regulation index and total assets to account for that possibility5.

2. Data collection and Variables

The data collection consists of two main components. The first dataset is to calculate the performance of microfinance institutions while the second one is to obtain the aggregate financial regulation index associated with these institutions. There are 321 microfinance institutions of 26 countries across Asia, Africa and Latin America with four main types of microfinance institutions: bank, non-financial institutions, cooperative or credit union, and non-governmental organization (NGO). I collect the yearly data of the annual report of these institutions for three years from 2005 to 2007 from the website Mixmarket. This dataset is completely balance panel data.

(27)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa are often profit-orientated that want to attract more funds from the public. These problems might lead to the biased technical efficiency of microfinance institutions in the sample in comparison with of reality.

2.1 Outreach, Sustainability, and Technical Efficiency

Under the assumption of constant return to scale, Data Envelopment Analysis provides the technical efficiency of each microfinance institution in 26 countries (see detail description in Appendix). To compare the technical efficiency of microfinance institutions across three regions, then it is aggregated to the country level.

(28)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Table 1. The average technical efficiency under the assumption of constant return to scale by regions

Outreach Sustainability Overall performance

Number of active borrowers Gross loan portfolio Financial Revenue Ratio Technical Efficiency Asia 2005 147,537 17,260,224 0.170 0.310 2006 188,374 21,131,822 1.180 0.330 2007 232,768 33,511,393 0.220 0.315 Africa 2005 39,190 18,571,349 0.240 0.260 2006 48,396 13,073,828 0.270 0.300 2007 62,025 24,221,370 0.330 0.280 Latin America 2005 40,649 47,337,914 0.270 0.380 2006 49,606 62,810,748 0.300 0.377 2007 65,145 84,019,623 0.340 0.390

(29)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Table 2. The Average Technical Efficiency under the assumption of Constant Return to Scale by top ten countries

Country 2005 2006 2007

High technical efficiency India 0.497 [94,091]* 0.548 [106,399] 0.547 [124,065] Ecuador 0.547 [51,117] 0.552 [57,354] 0.540 [65,192] Brazil 0.409 [68,255] 0.465 [68,498] 0.485 [74,982] El Salvador 0.302 [41,343] 0.317 [43,339] 0.443 [45,100] Mexico 0.436 [45,282] 0.453 [51,193] 0.437 [55,132] Low technical efficiency

Malawi 0.195 [19,602] 0.219 [19,954] 0.241 [21,133] Nigeria 0.230 [12,451] 0.259 [14,887] 0.234 [15,132] Sri Lanka 0.250 [17,292] 0.285 [20,094] 0.213 [23,662] The Philippines 0.190 [47,129] 0.208 [53,625] 0.209 [54,132] Zambia 0.214 [11,291] 0.232 [17,321] 0.188 [17,699] *The number of active borrowers is presented in bracket.

(30)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Table 3. The average technical efficiency under the assumption of constant return to scale by types of microfinance institutions

Number of active borrowers Gross loan portfolio Financial Revenue Ratio Technical Efficiency Bank 2005 294,337 75,003,179 0.212 0.210 2006 300,234 81,003,887 0.233 0.234 2007 331,009 80,165,006 0.236 0.242 Nonbank F. Ins 2005 47,063 22,198,367 0.209 0.351 2006 51,234 31,973,221 0.211 0.313 2007 47,112 22,093,184 0.256 0.345 Cooperatives 2005 41,112 61,229,000 0.118 0.431 2006 41,331 63,002,876 0.129 0.472 2007 42,001 64,118,981 0.230 0.500 NGOs 2005 23,543 6,119,088 0.201 0.276 2006 26,122 7,188,282 0.239 0.288 2007 29,879 7,677,091 0.292 0.300

In sum, the descriptive results draw a picture of the technical efficiency of microfinance institutions across Asia, Africa, and Latin America. We can see that microfinance institutions across countries perform differently with regarding to the outreach and sustainability they can create. However, it should be noted that technical efficiency in this study is only valid to the extent of the sample I collect. These efficiencies cannot be used to compare with others technical efficiencies, for instance, of other papers or of American microfinance institutions.

(31)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa In this section, I present the average financial regulation index across Asia, Africa, and Latin America. Then I examine the variation of the average index with regarding to the legal status of microfinance institutions. Finally, the descriptive statistics of financial regulation index are presented at both the country level and individual level. Since financial regulation index is the main explanatory in this study, I also perform the correlation matrix of the financial regulation index variable and the other control variables. This technique is necessary to detect the potential multicolinearity problems.

Graph 1. The average financial regulation index by regions

0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50

Asia Africa Latin America

2005 2006 2007

(32)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa and Latin America is of three and four, respectively.

Graph 2 represents the level of financial restrictions imposing on the four types of microfinance institutions. Banks suffer the most restrictive financial regulations since they are regulated by more than five restrictions. Next, nonbank financial institutions stand at the second position while cooperatives are at the third one with regard to the pressure of financial regulation. Finally, NGOs are the least regulated institutions since there are only one to two restrictions for their activities.

Graph 2. The average financial regulation index by types of microfinance institutions across three regions

0 1 2 3 4 5 6

Bank Nonbank Financial Institutions Cooperatives/Credit Unions NGOs 2005 2006 2007

(33)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Table 4. The average financial regulation index classified by the degree of restrictiveness

Country 2005 2006 2007

The most restrictive

Bangladesh 2.75 4.50 5.00

China 4.75 4.75 5.00

Colombia 4.50 4.75 5.00

Ecuador 4.50 5.00 5.00

Vietnam 3.25 3.75 4.75

The least restrictive

Mexico 3.00 3.00 3.00

El Salvador 2.75 2.75 2.75

Uganda 2.00 2.25 2.50

Peru 2.50 2.50 2.50

Sri Lanka 2.00 2.00 2.25

(34)

IV. EMPIRICAL RESULTS

Panel Tobit Models provide the estimation of the impact of financial regulation on microfinance institutions’ performance across Asia, Africa, and Latin America within the period of 2005 and 2007. The performance of microfinance institutions is measured by the indicators of outreach and sustainability as well as the by overall performance. Specifically, the number of borrowers and gross loan portfolio represent the width and depth of outreach, respectively, while financial revenue ratio is the proxy of sustainability. Being aggregated by the indicators of outreach and sustainability, technical efficiency is the representative of overall performance. Before analyzing the empirical results, it is necessary to explain the procedure of including explanatory variables to the models since it might affect the significance of outcomes. First, I check the pair wise correlation of explanatory variables by using correlation matrix. The technique aims to detect and prevent the potential multicollinearity problems. Second, the variables are inserted to the models one-by-one to check the variations of coefficients’ value as well as the significance level of these values. Finally, I test whether the data violate the assumption of Tobit model about normality and heteroskedasticity by performing the descriptive statistics of the error terms and making the white robust standard errors with statistical technique provide by Eview program. I addition, I check the relevance of the inclusion of variables by performing the likelihood ratio test6.

The entire sample

(35)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa though the sustainability is damaged under the pressure of financial regulation (Cull et al, 2009). It is likely the case since the number of NGOs contributes a high proportion in the data collection. As one can see in Table 5, the financial regulation has the negative impact on the technical efficiency of microfinance institutions at a 10 percent of significance level. It can be explained by the fact even though the outreach is improved, the gains of the outreach are not sufficient to compensate to the loss of the sustainability. Therefore, the overall performance, technical efficiency, is worsened.

Next, the evidence indicate that total assets positively affect the outreach, sustainability, and technical efficiency of microfinance institutions at a 1 percent of significance level. It implies that the bigger the microfinance institutions, the better are their performance (Mersland and Strom, 2007). Besides, the results of the interaction variable of financial regulation and total assets show that the outreach and sustainability of bigger microfinance institutions are more enhanced by financial regulation than those of smaller institutions are.

(36)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Table 5. Tobit estimates of the impact of financial regulation on outreach, sustainability, and overall performance of microfinance institutions

Outreach Sustainability Overall

performance Variables _{Number of} Active Borrowers Gross Loan Portfolio Financial Revenue Ratio Technical Efficiency Constant 190.0274 [0.0026]*** -6.1597 [0.0016]*** 0.3796 [0.0000]*** 0.2126 [0.0000]*** Total Assets 0.0000 [0.0000]*** 0.0000 [0.0000]*** 0.0000 [0.0008]*** 0.0000 [0.0000]*** Fin. regulation 68.9924 [0.0016]*** 2.4193 [0.0677]* -0.0485 [0.0000]*** -0.0237 [0.0585]* Frixta 0.0000 [0.0015]*** 0.0000 [0.0000]*** 0.0000 [0.0233]** GDP growth -12.0411 [0.0240]** 0.4185 [0.0552]* -0.0190 [0.0000]*** 0.0036 [0.2234] Inflation -2.3988 [0.1894] 0.0789 [0.5919] 0.0010 [0.4277] 0.0023 [0.3215] Africa -104.2603 [0.0013]*** -4.6839 [0.0443]** 0.0884 [0.0000]*** 0.0307 [0.1305] Latin America -215.6878 [0.0001]*** 4.6855 [0.0004]*** 0.0421 [0.0000]*** 0.1565 [0.0000]*** Year 2006 16.5722 [0.6344] 0.7247 [0.5834] 0.0141 [0.1665] 0.0135 [0.3702] Year 2007 12.6969 [0.7304] 2.1040 [0.1835] 0.0123 [0.2141] 0.0035 [0.8221] R2 0.2510 0.9860 0.2013 0.1973 Adjusted R2 0.2431 0.9859 0.1929 0.1897 Normality test 0.0000 0.0000 0.0000 0.0000 Observation 963 963 963 963

***, **, and * denote the significance of 1, 5 and 10 percent, respectively. P-value are presented in bracket.

(37)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa number of borrowers served but the higher gross loan portfolios, financial revenue ratios, and technical efficiency relative to those of Asian institutions. The econometric results confirm the validity of the descriptive findings about the pattern of technical efficiency of microfinance institutions across regions (Section 3.1). Finally, it should be noted that neither are the impact of inflation rate on the performance of microfinance institutions nor the difference in their performance through time founded. Besides, I do not include the interaction variable in the regression of technical efficiency since its inclusion neutralized the impact of other variables significantly.

In general, the findings suggest that restrictive financial regulation enhances the outreach but curtails the sustainability and technical efficiency of microfinance institutions across regions. Since one might argue that each type of microfinance institutions expose differently to financial regulation, I examine the impact of financial regulation on each type of microfinance institutions. Thereby, the heterogeneity of microfinance institutions is taken into account.

Banks

(38)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa banks curtail the outreach to maintain sustainability so that overall performance is also damaged. Put differently, by reducing lending to risky borrowers, banks avoid the loss of defaults so enhance the soundness of financial health (Jansson et al, 1998).

Table 6. Tobit estimates of the impact of financial regulation on outreach, sustainability, and overall performance of microfinance institutions as Banks

(39)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa In Table 6, inflation rate negatively affects financial revenue and technical efficiency of banks at 10 and 1 percent of significance level, respectively. It is similar to the finding of Grigorian et al, 2002 that shows that financial institutions bear the cost of higher margin in the inflationary environments. As one can see, the impact of GDP growth rate, total asset on the performance of banks as well as the regional differences in performance of these institutions have the same pattern with that of in Table 5 mentioned above. There is again no evidence on the difference in performance of microfinance institutions through time.

Nonbank Financial Institutions

Table 7 indicates that financial regulation has positive effect on the number of active borrowers but negative on the financial revenue ratio of nonbank financial institutions at 1 and 10 percent of significance level. The findings confirm the irrelevant application of banking financial regulation on the sustainability of nonbank financial institutions since they provide long-term financing. Therefore, the interest rate ceiling policy, for instance, makes these institutions difficult to compensate the higher risk of long-term loans. However, with respect to the increase in outreach, the possible explanation is that nonbank financial institutions deal with the restrictions by switching to retail banking and no focus on leases or long-term financing anymore. Hence, their outreach is shifted up as a result.

(40)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa almost outperformed by Asian counterparties except for financial revenue ratio. Similarly, nonbank financial institutions in Latin America have lower number of active borrowers than those in Asia. In addition, nonbank financial institutions have higher financial revenue ratio in 2006 and 2007 than that of 2005.

Table 7. Tobit estimates of the impact of financial regulation on outreach, sustainability, and overall performance of microfinance institutions as Nonbank financial institutions

performance Variables _{Number of} Active Borrowers Gross Loan Portfolio Financial Revenue Ratio Technical Efficiency Constant 99.9408 [0.0121]** -661.5323 [0.5603] 0.5240 [0.0000]*** 0.3785 [0.0000]*** Total Assets 0.0000 [0.0000]*** 0.0008 [0.0000] 0.0000 [0.0004]*** 0.0000 [0.0068]*** Fin. regulation 47.5743 [0.0024]*** 462.7147 [0.5264] -0.0568 [0.0673]* 0.0425 [0.3568] Frixta 0.0000 [0.0073]*** 0.0000 [0.8078] 0.0000 [0.0022]*** 0.0000 [0.1117] GDP growth -0.6540 [0.8233] 195.2458 [0.0874]* -0.0296 [0.0000]*** -0.0047 [0.4795] Inflation -4.2014 [0.0035]*** -40.6162 [0.3273] 0.0004 [0.8371] 0.0014 [0.7310] Africa -76.6390 [0.0002]*** -2,545.3770 [0.0000]*** 0.0425 [0.0302]** -0.1367 [0.0016]*** Latin America -152.3251 [0.0000]*** 246.6488 [0.7396] 0.0261 [0.2618] -0.0694 [0.1377] Year 2006 15.2366 [0.2469] 59.8689 [0.9254] 0.0344 [0.0791]* -0.0367 [0.2176] Year 2007 15.3027 [0.2729] 1,286.4270 [0.1030] 0.0360 [0.0467]** 0.0046 [0.8746] R2 0.3211 0.9855 0.3188 0.1701 Adjusted R2 0.2954 0.9850 0.2922 1.38E-01 Normality test 0.0000 0.0000 0.0000 0.0000 Observation 267 267 267 267

(41)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Cooperative/Credit Union

Table 8 shows that financial regulation has negative influences on the number of active borrowers and the financial revenue ratio of cooperatives at a 1 percent of significance level. Since cooperatives/credit unions are often small, the compliance to regulation might too costly for them, hence, leads to the decrease in their sustainability7. Therefore, the decrease in sustainability might negatively affect the ability to provide to more borrowers (outreach). However, the regulation positively affects the gross loan portfolio and technical efficiency of cooperatives at 5 and 1 percent of significance level. The increase in gross loan portfolio can be explained by the likelihood that these cooperatives make larger loans to low-risk borrowers under the pressure of restrictions. Consequently, these effects of financial regulation on the outreach and sustainability are translated into the increase in technical efficiency.

The table indicates that there is no evidence of the impact of GDP growth rate on the performance of these institutions. The impacts of inflation rate and total assets on performance are the same with those of the case of banks. However, the interaction variable shows that the restrictive financial regulation negatively affects bigger cooperatives in term of their gross loans portfolio (outreach). Besides, cooperatives in Africa and Latin America obtain higher financial revenue ratio and technical efficiency but lower number of active borrowers than do Asian cooperatives. There is almost no difference in the performance of cooperative across years except for the higher financial revenue ratio in 2007 relative to that of 2005.

Table 8. Tobit estimates of the impact of financial regulation on outreach, sustainability, and overall performance of microfinance institutions as Cooperatives/Credit Unions

performance Variables _{Number of} Active Borrowers Gross Loan Portfolio Financial Revenue Ratio Technical Efficiency 7

(42)

Constant 51.1490 [0.0016]*** -836.3918 [0.9269] 0.2448 [0.0000]*** 0.0296 [0.8928] Total Assets 0.0000 [0.0000]*** 0.0008 [0.0000]*** 0.0000 [0.0000]*** 0.0000 [0.0000]*** Fin. regulation -8.5897 [0.0011]*** 3860.6490 [0.0113]** -0.0545 [0.0000]*** 0.2598 [0.0000]*** Frixta 0.0000 [0.0000]*** -0.0001 [0.0000]*** 0.0000 [0.0000]*** 0.0000 [0.0000]*** GDP growth -1.4327 [0.2158] -115.0479 [0.8540] 0.0001 [0.9652] 0.0058 [0.7259] Inflation -0.6315 [0.3597] -595.5380 [0.0642]* -0.0060 [0.0025]*** -0.0153 [0.1104] Africa -32.6984 [0.0001]*** 4438.1180 [0.0484]** 0.1826 [0.0000]*** 2.5621 [0.0000]*** Latin America -31.7351 [0.0008]*** 4687.1790 [0.1777] 0.0322 [0.0926]* 0.3331 [0.0006]*** Year 2006 -1.5970 [0.5716] 1455.8090 [0.6563] 0.0026 [0.7911] -0.0127 [0.8148] Year 2007 -6.5904 [0.2480] 774.2884 [0.8153] -0.0190 [0.0775]* -0.0517 [0.3523] R2 0.9890 0.9939 0.5794 0.5063 Adjusted R2 0.9876 0.9931 0.5240 4.41E-01 Normality test 0.0000 0.0008 0.0000 0.0000 Observation 87 87 87 87

***, **, and * denote the significance of 1, 5 and 10 percent, respectively. P-values are presented in bracket.

Nongovernmental Organizations (NGOs)

(43)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa than do Asian counterparties. However, the financial revenue ratio they obtain is higher than that of Asian NGOs. The regional dummy variable also indicates that Latin American NGOs can obtain higher technical efficiency relative to those in Asia. Finally, there is no evidence of the difference of the performance of NGOs across years.

Table 9. Tobit estimates of the impact of financial regulation on outreach, sustainability, and overall performance of microfinance institutions as NGOs

performance Variables _{Number of} Active Borrowers Gross Loan Portfolio Financial Revenue Ratio Technical Efficiency Constant -14.9442 [0.4530] 1.6268 [0.9986] 0.4817 [0.0000]*** 0.2155 [0.0000]*** Total Assets 0.0000 [0.0008]*** 0.0009 [0.0000]*** 0.0000 [0.0000]*** 0.0000 [0.0180]** Fin. regulation -12.3486 [0.1052] 594.9161 [0.1222] 0.0453 [0.1070] -0.0071 [0.8602] Frixta -0.0001 [0.0007]*** 0.0000 [0.0010]*** 0.0000 [0.0048]*** GDP growth 3.9025 [0.0781]* -36.7783 [0.7159] -0.0317 [0.0000]*** -0.0016 [0.7495] Inflation 0.5165 [0.4256] -15.1791 [0.6627] -0.0008 [0.7590] 0.0028 [0.3248] Africa 1.7925 [0.8382] -594.8556 [0.0926]* 0.1076 [0.0001]*** 0.0413 [0.1272] Latin America -24.0547 [0.0030]*** 200.3699 [0.5658] 0.0436 [0.0671]* 0.1215 [0.0000]*** Year 2006 1.2646 [0.8046] -56.9077 [0.7927] 0.0256 [0.1394] 0.0254 [0.2823] Year 2007 5.6141 [0.3663] -267.8468 [0.2729] 0.0318 [0.0737]* 0.0327 [0.1986] R2 0.4340 0.9864 0.3224 0.1542 Adjusted R2 0.4196 0.9860 0.3032 1.30E-01 Normality test 0.0000 0.0000 0.0000 0.0000 Observation 363 363 363 363

(44)

the efficiency of microfinance institutions Nguyen Ho Anh Khoa Table 10. The impact of financial regulation on the performance of microfinance institutions across Asia, Africa, and Latin America

performance Variables Number of active borrowers Gross loan portfolio Financial Revenue Ratio Technical Efficiency Entire sample + + - - Banks - - + -

Nonbank Fin. Inst. + 0 - 0

Cooperatives - + - +

NGOs 0 0 0 0

“0”, “+”, and “-” denote no evidence, positive, and negative impact

Generally, the empirical evidence suggest that financial regulation has different impact on the outreach, sustainability, and technical efficiency of microfinance institutions across regions and types. First, restrictive financial regulation improves the outreach but damages the sustainability and technical efficiency of microfinance institutions with regardless to their types. The evidence contradict with the implications of Janssen et al. (1998) who argue that financial regulation is likely to enhance sustainability but curtail outreach. The possible explanation is based on the conclusion of Cull et al (2009) which suggests that non-profit orientated microfinance institutions keep their outreach growing despite the decrease in their sustainability. Since the number of NGOs contributes a high proportion in the collection of data (46%), this explanation is likely the case.