You are on page 1of 20

International Journal of Managerial Finance

What drives credit risk of microfinance institutions? International evidence


Naima Lassoued,
Article information:
To cite this document:
Naima Lassoued, (2017) "What drives credit risk of microfinance institutions? International evidence",
International Journal of Managerial Finance, https://doi.org/10.1108/IJMF-03-2017-0042
Permanent link to this document:
https://doi.org/10.1108/IJMF-03-2017-0042
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

Downloaded on: 05 September 2017, At: 01:44 (PT)


References: this document contains references to 53 other documents.
To copy this document: permissions@emeraldinsight.com
The fulltext of this document has been downloaded 1 times since 2017*

Access to this document was granted through an Emerald subscription provided by emerald-
srm:425905 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald
for Authors service information about how to choose which publication to write for and submission
guidelines are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as
well as providing an extensive range of online products and additional customer resources and
services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the
Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for
digital archive preservation.

*Related content and download information correct at time of download.


The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/1743-9132.htm

Credit risk of
What drives credit risk of microfinance
microfinance institutions? institutions

International evidence
Naima Lassoued
Ecole Superieure de Commerce de Tunis,
Received 8 March 2017
Université de la Manouba, Tunisia Revised 22 July 2017
Accepted 24 July 2017

Abstract
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

Purpose – The purpose of this paper is to shed light on the factors that affect microfinance institutions’
(MFI) credit risk. These factors include MFIs’ characteristics and country-level indicators.
Design/methodology/approach – This empirical study uses an unbalanced panel data of 638 MFIs from
87 countries observed over a period ranging from 2005 to 2015. Random-effects models are used to estimate
the models.
Findings – The results reveal that group-lending methodology, percent of loan granted to women and
diversification activities reduce credit risk; credit quality is enhanced by the relevance of the information
published by public or private bureaus and law enforcement cost increases credit risk. Finally, credit risk
tends to be limited in a good institutional environment.
Practical implications – Several implications can be drawn in light of these findings. For MFIs’ managers,
using group lending or granting more credit to women and diversifying their activities enhance their credit
quality. Furthermore, authorities need to strength debt repayment institutions and reinforce institutional
environment to help MFIs to limit their credit risk.
Originality/value – Previous studies focus on specific MFIs’ practices that enhance repayment rate or on
country-level indicators. One of the contributions of this paper is the use of both types of indicators.
Keywords Risk management, Credit risk, Microfinance institutions, Debt enforcement
Paper type Research paper

1. Introduction
In order to boost for self-employment and income-generating projects, microfinance
provides financial services to low-income clients that are often excluded from traditional
banking services (Ledgerwood, 1999). It has grown in popularity since the success of the
Grameen Bank in Bangladesh in 1983. Indeed, the number of microfinance institutions
(hence denoted MFIs) was 618 in 1997 serving 13,478,797 borrowers throughout the
world[1]. By the end of 2012, the number of MFIs reached 3,713 institutions granting credits
to 203,509,307 borrowers. MFIs were promoted as an essential tool used by major
international organizations and major donors to fight poverty. It allows for the creation of
employment and income opportunities to increase productivity and income of vulnerable
groups. Therefore, it seems to be a tool that supplements the financial sector.
Despite the growth, success and popularity of microfinance, many MFIs are struggling to
achieve financial self-sufficiency. They mainly survive on grants given by various international
agencies (including the World Bank) because of their social role. In fact, the main problem for
these institutions is information asymmetry between lender and borrowers, especially in
underdeveloped market (Stiglitz, 1990). Indeed, unreliability of financial information and
absence of conventional collateral complicate the screening and the monitoring process of
borrowers. Moreover, because poor clients are too risky, MFIs tend to focus on this category
of customers as some commercial lenders are reluctant to lend for such highly risky groups
(Hulme and Mosley, 1996). To overcome this type of risk, many lending innovations have been
proposed. Among these, the group-lending practice is the most important instrument to International Journal of Managerial
Finance
address the problem of lack of collateral that group lenders are required to guarantee. © Emerald Publishing Limited
1743-9132
Furthermore, to maximize repayment rates, some MFIs award incrementally larger loans upon DOI 10.1108/IJMF-03-2017-0042
IJMF each successful repayment and deny defaulters access to future loans (Galema, 2011).
In addition to these practices, MFIs diversify their activities and do not focus uniquely on
granting loans to the poor. They shifted to savings and insurance services.
Like other financial institutions, MFIs are not safe from financial crisis as the subprime
crisis revealed the dangers of providing an increasing array of higher risk loans to higher
risk borrowers. Indeed, financial crises have exhausted capital suppliers and increased
over-indebtedness of micro-entrepreneurs. As a result, many of them failed or swarmed into
financial stress (Servin et al., 2012). Therefore, managing risks by identifying the factors and
practices that affect the payment default rate for MFIs seems to be a crucial step for them to
ensure financial sustainability while reaching their social goals. Accordingly, in this study,
we try to identify the determinants of credit risk for MFIs. Specifically, we consider two
different categories of determinants, MFIs’ characteristics (non-systematic) and economic
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

and institutional disparities between countries.


We study a sample of 638 MFIs from 87 countries over the period 2005-2015. Our results
indicate that the group-lending methodology, percent of loan granted to women and
diversification activities reduce credit risk; credit quality is enhanced by the relevance of
information published by public or private bureaus and law enforcement cost increases
credit risk. Finally, credit risk tends to be limited in a good institutional environment.
This paper contributes in a number of ways to the literature on credit risk management
for MFIs. First, previous studies focused on two aspects. First, a particular focus was
given to specific MFIs’ practices, especially group-lending methodology, which is likely to
increase repayment rate or performance (e.g. Morduch, 1999; Crabb and Keller, 2006;
D’Espallier et al., 2011). Second is country-level indicators (e.g. Gonzalez, 2007; Krauss and
Walter, 2009; Galema, 2011). To our knowledge, our study is the first empirical attempt to
examine the effect of both types of indicators. Second, our study examines a
comprehensive sample of MFIs observed during the crisis period. Finally, in light of our
findings, we should provide an insight into the characteristics and practices useful for
MFIs to manage risk and we should draw some policy implications that may be valuable
for policymakers and shareholders.
The remainder of this paper is structured as follows. In Section 2, we review the relevant
literature. Our hypothesis is developed in Section 3. In Section 4, we describe our data and
present our econometric approach. In Section 5, we present and discuss our results. The last
section proposes some policy implications and concludes the paper.

2. The literature review


Researchers initially focused on examining efficiency of MFIs in terms of the practices are
likely to mitigate risk in view of improving repayment rates. Morduch (1999) suggests that
group lending mitigates risk by reducing adverse selection and moral hazard problems.
Indeed, group-lending contracts effectively bind a borrower’s co-signers to loans, mitigating
problems created by information asymmetry between lender and borrower. Co-signers
become encouraged to monitor reciprocally and to exclude risky borrowers from
participation, promoting repayment rates even in the absence of collateral.
Similarly, Islam (1995) found that lenders using peer-monitoring systems enjoy lower
rates than conventional lenders and that with the same interest rate. The expected
repayment rate is higher. Moreover, Tchakoute-Tchuigoua and Nekhili (2012), examining
a sample of 148 MFIs over the 2001-2006 period, found evidence indicating that
group-lending contracts significantly improve portfolio quality. However, risk mitigation
is achieved at the expense of performance, suggesting that group lending leads MFIs to
arbitrate between risk and performance. However, group lending in practice has some
disadvantages such as lowering contagion risk if one of the contractors is unable to clear
repayments (Armendariz de Aghion and Morduch, 2000).
Armendariz de Aghion and Morduch (2000) highlighted several important practices Credit risk of
that generate high repayment rates from low-income borrowers without requiring collateral microfinance
and without using group-lending contracts. These mechanisms include the use of institutions
non-refinancing threats, regular repayment schedules, collateral substitutes and the
provision of nonfinancial services.
Crabb and Keller (2006) examined the key risk factors of loan portfolios, including
institutional size and macroeconomic factors, in a sample of 37 MFIs during 2001-2003.
They found that the group-lending methodology used by most MFIs reduces loan
portfolio risk. However, as greater lending to women is assumed to consistently raise
portfolio risk, female group lending is found to mitigate this risk. Indeed, Saravia-Matus
and Saravia-Matus (2015) gave additional evidence indicating that female repayment
performance is in fact better than that of males in a Nicaraguan sample of MFIs.
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

D’Espallier et al. (2011) investigated a sample of 350 MFIs in 70 countries. The results
pointed out that more women borrowers are associated ceteris paribus with lower portfolio
risk, lower write-offs and lower credit-loss provisions.
Ayayi (2012) studied credit risk determinants in a selected group of Vietnamese MFIs
and East Asian and Pacific MFIs. The author found that while liquidity has a positive
impact, size of gross loan portfolio and operating inefficiency have a negative impact on
MFIs’ credit risk. Furthermore, the author showed that raising operational financial
sustainability is associated with an improved portfolio quality.
Another research line focused on the effect of country-level institutional and
macroeconomic data on MFIs’ risk. The idea is that the microfinance industry is subject to
strong cross-border influence from international capital providers and international
knowledge transfer (Mersland et al., 2011). This line of research bears on the work of
La Porta et al. (1998) who found that legal systems and creditor rights promote credit
markets. For instance, Djankov et al. (2007) found evidence indicating that law
enforcement quality, creditor rights and information sharing positively correlate with the
private credit to GDP ratio. Brown et al. (2009) found that information sharing correlates
with lower credit costs in transition countries and that this correlation is stronger in
countries with poor legal systems.
Among the few authors who focused on MFIs, we mention Gonzalez (2007)
who examined whether changes in GDP per capita significantly affect MFIs’ portfolio
risk (30-day and 90-day portfolio risk, loan loss rate and write-off ratio). The author
examined a sample of 639 MFIs in 88 countries observed during 1999-2006. He found no
evidence for a relationship between MFIs’ asset quality and growth. An exception was for a
30-day portfolio risk, where a statistically significant relationship was found. These tests
indicate that microfinance portfolios are highly resilient to economic shocks.
Krauss and Walter (2009) examined microfinance systemic risk using annual data of
325 MFIs operating in 66 emerging markets during 1998-2006. The studied MFIs did not
show any risk exposure to international capital markets. When exposed to GDP, these
MFIs display highly significant correlation indicating that MFIs are not detached from
their respective domestic economies.
Finally, Galema (2011) explored the effect of creditor rights, information sharing and
law enforcement quality on MFIs’ risk taking. The author found that contract enforcement
days negatively correlate with MFIs’ risk taking while in countries with strong creditor
rights contract enforcement days reduce MFIs’ risk taking. In addition, the author found
that only unregulated and nonprofit MFIs take less risk as a result of longer contract
enforcement procedures.
Overall, previous research focused on the effect of specific MFIs’ practices, especially
group lending and gender diversity, on repayment rates, performance (e.g. Morduch, 1999;
D’Espallier et al., 2011) and credit risk (PAR_30) (Crabb and Keller, 2006). Specifically, the
IJMF latter study investigated only the effect of gender diversity, loan size and group lending on
credit risk in a small sample of 37 MFIs. We extend this study in three ways: first, we
add other MFIs’ practices likely to affect credit risks like savings, insurance services and
non-interest income activities. Second, we introduce a set of country-specific variables
representing debt enforcement that could determine credit risk. Third, we use a more
representative sample in view of generating representative findings. Moreover, the second
strand of literature which examined the effect of country-specific indicators on MFIs’ risk
focused on a single pre-subprime crisis (e.g. Gonzalez, 2007; Galema, 2011). Unlike these
studies, we used a sample that covers a more recent period spanning from 2005 to 2015.

3. Research hypotheses
3.1 MFIs’ characteristics
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

3.1.1 Group lending. The group-lending methodology seems to be a key innovation that
expanded the poor’s access to credit in developing countries (Morduch, 1999). This
contracting instrument is likely to mitigate the risks associated with information
asymmetry. Indeed, group borrowers are related by a joint liability, when one of them
switches to risky projects (moral hazard), the probability that their associates will have to
pay the liability increases. Thus, group members are encouraged to screen other clients
(Madajewicz, 2011). However, other line of research argues that group lending may generate
additional costs like group contracting costs, training borrowers on group procedures,
higher degree of supervision and a higher frequency of installments. These costs increase
interest rates of such microcredit loans leading to higher repayment risks (Savita, 2007).
Despite the aforementioned disadvantages of the group-lending methodology, we expect
that it will negatively affect credit risk:
H1a. Group lending negatively affects credit risk.
3.1.2 Severe repayment monitoring. Microfinance practitioners impose regular repayment
deadlines in order to maintain high repayment rates in the absence of collateral.
This monitoring takes place on a weekly or a monthly basis (Armendariz de Aghion
and Morduch, 2000). According to Morduch (1999), this mechanism has many advantages.
First, it excludes undisciplined borrowers at an early stage before accumulation
of their unpaid debt. Second, severe repayment monitoring may guarantee the bank a
minimum level of liquidity. Third, MFIs that use this practice target a specific customer
category with an additional good-standing income since the repayment process begins
before their investments generate returns. We expect that this tool negatively affects
credit risk. Then, in light of this discussion, we expect that severe repayment monitoring
inhibits credit risk:
H1b. Severe repayment monitoring negatively affects credit risk.
3.1.3 Savings requirement. Savings is perceived as financial collateral provided by
borrowers to secure a loan. Indeed, many MFIs ask borrowers for savings to be qualified for
a loan (Armendáriz de Aghion and Morduch, 2010). Morduch (1999) argues that some MFIs
require deposits to reinforce contracts and these funds serve as insurance against loan
default, death or disability. Similarly, Tchakoute-Tchuigoua (2014) point out that many
MFIs ask from borrowers to pay an additional percent of the contracted loan. These
contributions are usually deduced from the contractors’ loans in addition to loan
amortization and thus are forced to deposit savings. These imposed savings can be
withdrawn upon loan maturity, but only after MFIs recover their debts. We expect savings
requirement to decrease credit risk:
H1c. Savings requirement negatively affects credit risk.
3.1.4 Insurance. Insurance consists in collecting relatively small premiums from a customer Credit risk of
and funding relatively large payouts to the small portion of that population that suffers microfinance
losses from a specific risky event. It can provide low-income households a greater degree of
protection against property, death, health and disability risks. In fact, this population
institutions
category is highly vulnerable to these risks. In case of an event, loss will be covered by a
large number of people, at a much lower cost or premium per person (Kumar and Qazi,
2016). Therefore, insurance could be an effective tool allowing for alleviating credit risk of
vulnerable borrowers. Therefore, a negative relationship between credit risk and insurance
is expected:
H1d. Insurance negatively affects credit risk.
3.1.5 Gender diversity. Empowering women is an important aspect of microfinance since
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

women face peculiar barriers that men do not face. Compared to men, women were found to
display better debt repayment performances (Yunus, 1999). Many arguments are given
to explain this better recovery rate for woman than for men. First, women are highly
aversive to risk than men (Croson and Gneezy, 2009). They are likely to choose relatively
less risky projects (Sharma and Zeller, 1997). Second, they generally enjoy hard work ethics
and financial discipline (Bhatt and Tang, 2002). Third, women depend more on MFIs since
they have fewer opportunities to escape poverty in developing countries as they have
limited access to education and are committed to their families and communities. This latter
limitation impedes their ability to take advantage of potential educational or professional
opportunities elsewhere (Boehe and Cruz, 2013). Despite the shared agreement on the high
repayment rate of women borrowers, some authors call into question this relationship.
For instance, Crabb and Keller (2006) argue that lack of access to capital for women induces
greater risk. For Phillips and Bhatia-Panthaki (2007), women have less access to funding
than men; therefore, they tend to invest in traditional projects that are not profitable or
competitive, enabling them to repay their loans at the due date.
We expect that female loan percentage has a negative effect on credit risk:
H1e. The percent of women borrowers negatively affects credit risk.
3.1.6 Income diversification. Theoretically, like other financial institutions, MFIs shift
toward non-interest income to compensate for their shortfall in interest margins on
lending activities. Then, they rely on alternative income types other than interest-based
income. We expect that for well-diversified financial institutions, where non-interest
income is important, credit risk should be lower than for less-diversified financial
institutions. Indeed, MFIs may charge lower rates on loans if they will gain additional fees
from borrowers. Moreover, MFIs have an information advantage about their borrowers
thanks to additional services.
Accordingly, we formulate the following hypothesis:
H1f. Income diversification negatively affects credit risk.

3.2 Country characteristics


Our focus among other things is macro-institutional and economic determinants. In fact,
judicial and institutional disparities seem to yield different bank-lending policies (Bae and
Goyal, 2009; Ge et al., 2012), since better institutional environments are able to overcome
information asymmetries in credit markets.
3.2.1 Law enforcement quality. Law enforcement quality refers to how law regulates court
proceedings. More explicitly, in reality litigations should be highly formalized, with formal
procedures that describe the steps of the insolvency process (Djankov et al. (2008). When
enforcing a debt contract becomes costlier, MFIs’ lending becomes riskier. MFI lenders are less
IJMF likely to be able to enforce repayment and in the event of default are less likely to recover the
full amount of the loan. Consequently, creditors are less willing to grant loans (Galema, 2011):
H2a. Contract enforcement costs positively affect credit risk.
3.2.2 Creditor rights. It is commonly accepted in the literature and empirical research that
strengthening creditor and investor rights matter in loan contracting and firm’s financing
decisions (Aghion and Bolton, 1992). Poor creditor rights lead to lower recovery rates in case
of default, thus increasing MFIs’ risk. Moreover, legal rights alleviate uncertainty and cost
while pursuing repayment. Therefore, MFIs operating in an institutional environment with
stronger legal creditor rights have more chances to be repaid. Arguably, creditor power
strengthens the ability of lenders to force borrower repayment by seizing collateral.
Therefore, we assume that creditor rights facilitate recovery of unpaid loans and reduce
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

credit risk:
H2b. Creditor rights negatively affect credit risk.
3.2.3 Information sharing. In economies where information asymmetry is high, lenders are
not able to evaluate borrowers’ true credit quality, thus adverse selection problems will
increase. Therefore, the existence of public or private credit agencies contributes to secure
defaulting borrowers, since adverse selection problems are mitigated. Jappelli and Pagano
(2002) and Padilla and Pagano (2000) argue that credit agencies may act as a borrowing
disciplinary device since borrowers will be more cautious about their repayment records if
lenders share their information on defaulting customers. In addition, information sharing
helps to reduce operating costs, including the cost of screening potential borrowers ex ante
and that of monitoring existing credit contracts. Hence, lending quality is improved by
ex ante mitigating adverse selection of loan contracts and mitigating moral hazard during
loan contracts (Zhao et al., 2011).This leads to the following hypothesis:
H2c. Information sharing negatively affects credit risk.
3.2.4 Governance quality. The governance index describes the legal and judicial framework
as compiled by Kaufmann et al., 2010). The index includes the following six sub-indices:
voice and accountability, political instability and violence, government efficiency,
regulatory quality, rule of law and corruption control. According to Godlewski (2004),
an inadequate regulatory and supervisory regime and a weak legal and institutional
framework characterized by bureaucracy and political instability negatively affect crediting
banks. Indeed, such a context reduces the efficiency of regulation. Moreover, many studies
explain that in emerging markets weak legal, institutional and supervisory devices hinder
either the process of granting credits or the process of loan control and recovery later.
Therefore, we expect that well-functioning institutions and good governance decrease MFIs’
credit risk. Hence, our last hypothesis is as follows:
H2d. Good governance quality negatively affects credit risk.

4. The research design


In this section, we present our sample and define the variables used in our study.

4.1 Data sources and sample selection


For the sake of our analysis, our data are selected from three different sources. First,
MFI-level data are drawn from the Microfinance Information eXchange “MIX data set”
covering more than 1,900 MFIs. The MIX database is a web-based microfinance platform
that provides data on individual MFIs.
Law enforcement quality, creditor rights index and information sharing are taken from Credit risk of
the Doing Business website. The governance index is taken from Kaufmann et al. (2010) by microfinance
adding different governance indicators available on the World Bank Word Governance institutions
Indicators website. Finally, we extracted macroeconomic data from the World Bank World
Development Indicators website.
We targeted an initial sample consisting of 1,900 MFIs observed over the 2005-2015
period. We excluded MFIs and countries with missing data. We ended up with
an unbalanced panel of 638 MFIs, totaling 5,379 MFI-year observations from 87 countries
examined over the 2005-2015 period. Sample distribution by regions is presented
in Table I.

4.2 Econometric model


Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

To examine determinants of MFI, credit risk, we estimate the following regression models
using the following general form:
PAR_30ijt ¼ b0 þ b1 GR_LENDijt þb2 REP_SHCijt þb3 SAV_DUMijt
þb4 INS_DUMijt þb5 WOM_BORijt þb6 DIVERijt
þb7 LAW_ENFijt þb8 CRED_RIGHijt þb9 INF_SHARijt
þb10 GOV_INDijt þControl variablesijt þeijt (1)
Credit risk is modeled as a function of various MFIs and country-specific factors and
i denotes MFI, t time period and j country.Where PAR_30 is the Portfolio at Risk30;
GR_LEND the group lending; REP_SHC the severe repayment evaluation; SAV_DUM the
savings; INS_DUM the insurance; WOM_BOR the women borrowers; DIVER the income
diversification; LAW_ENF the law enforcement cost; CRED_RIGH the creditor rights index;
INF_SHAR the information sharing; and GOV_IND the governance index.
Bearing on the literature on credit risk (e.g. Crabb and Keller, 2006; Gonzalez, 2007;
Krauss and Walter, 2009; D’Espallier et al., 2011; Galema, 2011), we include the following
control variables.
Capital adequacy ratio. MFIs often maintain a minimum level of capital required to meet
financial obligations and address unexpected losses. However, according to the moral
hazard hypothesis (Berger and DeYoung, 1997), highly capitalized MFIs generally grant
riskier loans. Therefore, the sign of the relationship between the capital adequacy ratio and
credit risk is expected to be positive or negative.
Size. We expect that size negatively correlates with credit risk because larger MFIs are
more diversified and likely to be more experienced to grant a wider variety of loans and deal
appropriately with defaulting borrowers.
Profitability. We suggest that it negatively affects credit risk. Indeed, profitable MFIs are
less pressured to generate income and thus less constrained to engage in risky credit
offerings. However, unprofitable MFIs are constrained to engage in more uncertain credits
to sustain their profitability.
Outstanding loans. A significant portion of outstanding loans in MFIs’ total assets
increases MFIs’ default risk. Since financial risk increases with leverage, a positive relationship
between a banking firm’s risk and leverage is expected (Tchakoute-Tchuigoua, 2015).
Outreach measured by average loan size. We expect a positive effect of outreach on credit
risk because for an individual MFI borrower, it is difficult to reimburse excessive high
amounts (Gool et al., 2011).
GDP growth. We include GDP growth to control for the macroeconomic cycle. During
periods of economic expansion, higher income is generated and borrowers’ repayment
capacity is improved, thus reducing credit risk. Conversely, during recession periods, level
IJMF Panel A: Africa
Angola 1 0.16%
Benin 9 1.41%
Burkina Faso 12 1.88%
Burundi 7 1.10%
Cameroon 12 1.88%
Chad 2 0.31%
Congo 6 0.94%
Cote d’Ivoire 5 0.78%
Ethiopia 3 0.47%
The Gambia 2 0.31%
Ghana 11 1.72%
Kenya 9 1.41%
Mozambique 6 0.94%
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

Namibia 1 0.16%
Niger 7 1.10%
Nigeria 11 1.72%
Rwanda 3 0.47%
Senegal 7 1.10%
Sierra Leone 3 0.47%
South Africa 5 0.78%
Sudan 1 0.16%
Tanzania 7 1.10%
Togo 4 0.63%
Uganda 10 1.57%
Zambia 4 0.63%
Zimbabwe 1 0.16%
Total 149 23.35%
Panel B: East Asia and the Pacific
Indonesia 5 0.78%
Vietnam 12 1.88%
Philippines 14 2.19%
Cambodia 17 2.66%
China 10 1.57%
Papua 2 0.31%
Samoa 1 0.16%
Tonga 1 0.16%
Total 62 9.72%
Panel C: Eastern Europe and Central Asia
Albania 5 0.78%
Armenia 6 0.94%
Azerbaijan 17 2.66%
Bosnia and Herzegovina 6 0.94%
Bulgaria 8 1.25%
Georgia 9 1.41%
Kazakhstan 8 1.25%
Kosovo 6 0.94%
Kyrgyzstan 3 0.47%
Moldova 3 0.47%
Mongolia 7 1.10%
Montenegro 3 0.47%
Romania 6 0.94%
Russia 10 1.57%
Tajikistan 11 1.72%
Table I.
Sample distribution (continued )
Turkey 3 0.47%
Credit risk of
Ukraine 2 0.31% microfinance
Total 113 17.71% institutions
Panel D: Latin America and The Caribbean
Argentina 10 1.57%
Belize 1 0.16%
Bolivia 8 1.25%
Brazil 12 1.88%
Chile 5 0.78%
Colombia 13 2.04%
Costa Rica 9 1.41%
Dominican Republic 7 1.10%
Ecuador 13 2.04%
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

El Salvador 14 2.19%
Guatemala 11 1.72%
Guyana 1 0.16%
Haiti 3 0.47%
Honduras 20 3.13%
Jamaica 5 0.78%
Mexico 21 3.29%
Nicaragua 17 2.66%
Panama 2 0.31%
Paraguay 6 0.94%
Peru 17 2.66%
Suriname 1 0.16%
Uruguay 1 0.16%
Venezuela 2 0.31%
Total 199 31.19%
Panel E: MENA region
Egypt 4 0.63%
Jordan 4 0.63%
Lebanon 3 0.47%
Morocco 6 0.94%
Tunisia 1 0.16%
Total 18 2.82%
Panel F: South Asia
Afghanistan 11 1.72%
Bangladesh 15 2.35%
Bhutan 1 0.16%
India 10 1.57%
Nepal 12 1.88%
Pakistan 30 4.70%
Sri Lanka 18 2.82%
Total 97 15.20% Table I.

of unpaid debts is expected to increase. We expect a negative relationship between GDP


growth and credit risk.
Interest rate. We expect that interest rate positively affects credit risk, since an increase
in interest rate leads to an increase in debt burden, leading to higher credit risk.
Inflation. An increase in inflation rate affects borrower solvency (in particular small
borrowers) because an increase in prices decreases households’ real income. The latter may
no longer be able to pay their commitments as a result of rising inflation. Consequently,
we expect that inflation has a positive effect on MFIs’ credit risk.
IJMF Table II describes in detail all the variables used in our study, and reports the expected
signs of the respective coefficients.
In this study, we chose the panel data methodology because it has the advantage of
increasing degrees of freedom and reducing collinearity between the independent variables
and consequently it leads to more efficient estimates. In addition, OLS estimation may not be
well appropriate to determine the relationships between the dependent and independent
variables. In fact, this estimation method assumes homogeneity of the endogenous variables
for all individuals in the sample. However, the credit risk variable seems to vary considerably
across countries and years. To estimate our model, we chose the random-effects model
because our model includes many country-level variables (like GDP growth, inflation, creditor
rights and so on) which are the same for all MFIs in a given country. Other variables vary in
time like group lending, insurance and savings. The standard errors of the estimated
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

coefficients are adjusted for cluster effects at country level as suggested by Peterson (2009).

Predicted
Variables Abbreviation Description sign

MFI-level data
Portfolio at Risk30 PAR_30 Outstanding balance. Loans overdue W 30 days/gross
loan portfolio
Portfolio at Risk90 PAR_90 Outstanding balance. Loans overdue W 90 days/gross
loan portfolio
Provision for Loan LLP Provision for loan impairment/assets
Impairment
WOR WOR Value of loans written off during period/average gross
loan portfolio
Group lending GR_LEND Dummy variable equal to 1 if the MFI uses group lending −
and 0 otherwise
Severe repayment REP_SHC Dummy variable equal to 1 if the MFI uses robust −
evaluation repayment evaluation and 0 otherwise
Saving SAV_DUM Dummy variable equal to 1 if the MFI accepts deposits −
and 0 otherwise
Insurance INS_DUM Dummy variable equal to 1 if the MFI accepts insurance −
and 0 otherwise
Women borrowers WOM_BOR Percentage of borrowers who are women −
Capital adequacy ratio CAP_ADQ Equity/total assets −/+
Size SIZE Natural logarithm of total assets −
Diversification DIVER Non-interest income to total income −
Profitability PROF Return on asset ratio −
Loan portfolio OUTS_LOAN Outstanding loan portfolio/total assets +
Outreach OUTREACH Natural logarithm of average loan balance per borrower +
Country-level data
Quality of law LAW_ENF Contract enforcement costs +
enforcement
Creditor rights index CRED_RIGH The strength of the legal rights index measures the −
degree to which collateral and bankruptcy laws protect
borrower and lender rights and thus facilitate lending
Information sharing INF_SHAR Dummy variable equals 1 if a public credit registry or a −
private credit bureau exists and zero otherwise
Governance index GOV_IND Governance index from Kaufmann et al. (2010) −
GDP growth GDP_GR Growth rate of gross domestic product on annual basis −
Inflation INFLATION Inflation rate (in percentage terms) +
Table II. Interest rate INT_RATE Real interest rate is the lending interest rate adjusted +
Variables description for inflation
5. Results Credit risk of
5.1 Descriptive statistics microfinance
Table III presents summary statistics for all variables. As we can see, the mean of PAR_30 institutions
is 0.058. It seems that loan portfolio in our sample is healthy and close to the average
reported by previous studies (Galema, 2011; Tchakoute-Tchuigoua, 2015).
For risk management indicators, we found that 17.4 percent of MFIs employ the
group-lending methodology. Most MFIs of our sample (86.1 percent) use repayment
schedules, 58.2 percent of MFIs offer their customers savings services and 44.6 percent
offer insurance services. It is worth to note is that 64.6 percent of borrowers are women.
This finding confirms that MFIs’ lending often targets women more than men (Crabb and
Keller, 2006). As for diversification, we found that non-interest income represents,
on average, 3 percent of income, suggesting that MFIs’ activities are highly focused
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

on lending.
The mean of the capital adequacy ratio is 0.321, indicating that MFIs maintain a level of
capital adequacy ratio above the minimum required in most countries. Moreover,
outstanding loans represent on average 75 percent of assets. Average profitability is 0.014
with a minimum of −0.161 percent and a maximum of 0.117.
Looking at the country-level variable, 88.9 percent of MFIs operate in countries
with an information-sharing agency. Average contract enforcement costs are
38.9 percent of claims. Galema (2011) reports a mean of 36.1 percent. This implies that
for a debt contract of $1,000, for instance, the creditor pays on average $389 in
enforcement costs.
The creditor rights index is 0.409 indicating that MFIs operate in countries with lower
creditor protection rights. In addition, the governance index indicates a weak legal and
institutional frameworks as the mean is 0.400, less than 0.5
For the economic context, we found that GDP growth in particular presents a
high disparity across countries with a minimum of 0.158 percent and a maximum of

Variable Observations Mean SD Min Max

PAR_30 5,379 0.058 0.064 0 0.240


PAR_90 5,379 0.039 0.047 0 0.179
LLP 5,379 0.016 0.018 −0.003 0.066
WOR 5,379 0.016 0.024 0 0.088
GR_LEND 5,379 0.174 0.380 0 1
REP_SHC 5,379 0.861 0.346 0 1
SAV_DUM 5,379 0.582 0.493 0 1
INS_DUM 5,379 0.446 0.497 0 1
WOM_BOR 5,379 0.646 0.269 0 1
CAP_ADQ 5,379 0.321 0.264 −0.319 0.998
SIZE 5,379 15.712 2.132 3.899 24.468
DIVER 5,379 0.030 0.113 −0.124 0.300
PROF 5,379 0.014 0.063 −0.161 0.117
OUTS_LOAN 5,379 0.750 0.167 0.365 0.979
OUTREACH 5,379 5.400 1.359 2.819 7.880
INF_SHAR 5,379 0.889 0.314 0 1
LAW_ENF 5,379 0.389 0.260 0.001 1.518
CRED_RIGH 5,379 0.409 0.225 0 1
GOV_IND 5,379 0.400 0.158 0.033 1
GDP_GR% 5,379 5.257 2.648 0.158 10.260
INFLATION% 5,379 6.657 3.918 1.014 15.525 Table III.
INT_RATE% 5,379 7.090 3.941 0.150 19.585 Summary statistics
IJMF 10.260 percent. A similar pattern is observed for inflation, which ranges between 1.014 and
15.525 percent while interest rate ranges between 0.150 and 19.585 percent.
Correlations between the independent variables are not high, suggesting that
multicollinearity is not a serious concern in our estimation[2].

5.2 Empirical results


Table IV reports the results of the baseline model, which examines the determinants of
credit risk. In model (1), we estimate the effects of MFI-specific variables on PAR_30 and in
model (2) we add country-level indicators. The empirical results are reported with robust
standard errors clustered by country.
In model (1), the estimated coefficient of group lending (GR_LEND) is negative and
significant at the 5 percent level, indicating that group lending has lower credit risk
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

compared to individual lending. This finding confirms H1a, suggesting that the group-
lending practice alleviates information asymmetry (Madajewicz, 2011) and moral hazard
risks (Kono and Takahashi, 2010). More specifically, the probability that group members
will have to pay the liability increases when one of them switches from safe to risky projects.
Therefore, group members are incited to monitor each other. This finding confirms previous
results in different contexts. For instance, Crabb and Keller (2006) report a negative effect of
group lending on PAR_30 in a sample of 37 MFIs, using quarterly data. Others show that
group lending improves repayment rates (e.g. Zeller, 1998).
The coefficient of WOM_BOR is negative and statistically significant at the 1 percent
level. This result is in line with H1e, assuming that women have less credit risk compared to
men. This finding can be explained by the dependence of women to MFIs. In fact, women
have fewer opportunities to escape poverty in developing countries as they have limited
access to education and are committed to their families and communities, which impedes
their ability to take advantage of potential educational or professional opportunities
elsewhere (Boehe and Cruz, 2013). Our result supports the findings of many studies

Variables Model (1) Model (2)

GR_LEND −0.003 (−2.177)** −0.006 (−2.457)**


REP_SHC 0.006 (1.504) 0.004 (1.185)
SAV_DUM −0.001 (−0.040) −0.003 (−0.879)
INS_DUM 0.001 (0.391) −0.002 (−0.550)
WOM_BOR −0.042 (−3.561)*** −0.052 (−4.652)***
DIVER −0.037 (−2.034)** −0.046 (−1.739)*
LAW_ENF 0.062 (3.831)***
CRED_RIGH 0.079 (1.233)
INF_SHAR −0.002 (−2.255)**
Control variables
CAP_ADQ 0.013 (1.281) 0.007 (0.702)
SIZE −0.001 (−0.509) −0.001 (−0.265)
PROF −0.240 (−4.334)*** −0.218 (−3.150)***
OUTS_LOAN −0.020 (−1.095) −0.028 (−1.720)*
OUTREACH −0.003 (−2.503)** −0.003 (−2.190)**
GOV_IND −0.038 (−2.339)**
GDP_GR −0.254 (−4.547)***
INFLATION −0.025 (−0.471)
INT_RATE 0.025 (1.013)
Table IV. Observations 5,379 5,379
Determinants of R2 0.1093 0.1978
MFI’s credit risk Note: *,**,***Significance at 10, 5 and 1 percent levels, respectively
(Mersland and Strøm, 2010; Sharma and Zeller, 1997; D’Espallier et al., 2011) indicating that Credit risk of
female MFI borrowers positively affect debt repayment. However, Crabb and Keller (2006) microfinance
report the opposite result. institutions
The coefficient of diversification (DIVER) is negative and significant at the 5 percent
level. As H1f assumes, MFIs relying on alternative income types other than interest income
have lower credit risk. In fact, MFIs are exposed to income generation pressure and thus are
constrained to engage in risky credit offerings. By diversifying their activities, they
compensate for losses in some products by gains in others.
Model (2) includes country-level indicators. R2 improved from 0.1093 (model (1)) to 0.1978
(model (2)), implying that country-level variables improve the variance explained in the first
regression. When we add macroeconomic and institutional variables, the coefficients
estimated previously are still significant and maintain their signs.
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

The coefficient of LAW_ENF is positive and significant at the 1 percent level.


This finding confirms H2a and implies that higher contract enforcement costs are
associated with more credit risk. In line with our expectation, when enforcing a debt
contract becomes costlier MFI’s lending becomes riskier. This finding validates that of
Galema (2011).
The coefficient of (INF_SHAR) is negative and significant at the 5 percent level. This
result supports H2c, suggesting that information sharing reduces credit risk through
alleviating asymmetric information problems by disciplining borrowers seeking to maintain
a good reputation with most lenders. In addition, information sharing reduces the moral
hazard problem since it restrains excessive lending when each borrower may opt for several
financial institutions. Our result, however, contradicts that of Galema (2011) for a sample of
300 MFIs examined over the 2000-2008 period. Such a finding could be attributed to the use
of two different study periods. Other studies examining the effect of information sharing on
bank credit confirm our finding. For instance, Godlewski (2004) found that the existence of
private credit offices reduces credit risk for a sample of more than 2,000 banks from
30 emerging economies during 1996-2001. Similarly, Houston et al. (2010) found that less
information sharing increases bank’s risk taking for a sample of 2,386 banks across
69 countries examined during 2000-2007.
The coefficient of GOV_IND is negative and significant at the 5 percent level. We confirm
H2d, indicating that better institutional quality reduces credit risk by solving potential
conflicts of interest between MFIs and borrowers. Breuer (2006) found the same results in
the banking sector and argue that it is not banking regulation and supervision alone that
influence the extent of loan-granting problems but other institutional factors such as
corruption may matter. Godlewski (2004) found evidence indicating that a more efficient
judicial system where laws are well enforced and where corruption is controlled may
improve the credit-granting process.
For the control variables, the coefficient of profitability (PROF) is negative and statistically
significant at the 1 percent level. Consistent with the bad management hypothesis, greater
performance reduces credit risk because poor management quality affects the loan-granting
process when customers’ credits are not thoroughly evaluated.
The coefficient of OUTREACH indicates that loan amount negatively relates to credit
risk (significant at the 5 percent level). The result confirms our prediction and shows that
smaller loan amounts are associated with higher credit risk since small loan amounts are
mostly granted to the poorest people or to business beginners who lack experience.
The coefficient of GDP_GR is negative at the 1 percent level indicating that in period of
economic growth credit risk decreases. Indeed, an increase in GDP generally reflects higher
income flows for small borrowers and an increase in the profitability of firms and therefore
better solvency. Among the few studies examining the effect of economic growth on MFIs’
credit, we mention Ahlin et al. (2011) who show that growth boosts financial sustainability
IJMF by reducing default. However, Gonzalez (2007) found that portfolio-at-risk positively
correlates with growth.
The other coefficients of REP_SHC, SAV_DUM, INS_DUM, CAP_ADQ, SIZE and
OUTS_LOAN fail to gain significance.

5.3 Determinants of credit risk during the 2007-2009 financial crisis


An interesting question is whether factors affecting MFIs’ credit risk had the same influence
during the 2007-2009 financial crisis. This issue is important because financial crises
aggravated the poverty problem (Milana and Ashta, 2012). Moreover, Wagner and Winkler
(2013) argue that the crisis has deeply affected MFIs, especially those operating in countries
experiencing a severe post-crisis recession.
To investigate this issue, we separately rerun the model during the crisis period (2007-
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

2009) and out of the crisis period. The results, displayed in Table V, show that during the
crisis (model (1)) country-level indicators seem to matter more than MFIs’ characteristics.
For the formers, we found that only the coefficient of SAV_DUM is statistically significant
with a negative sign (at the 1 percent level) suggesting that savings can be an effective risk
management tool to face the crisis. More specifically, when borrowers are unable to pay
their debt, MFIs use savings to compensate the due debt amount.
Turning to country-level variables, we found that the coefficient of INF_SHAR is
negative and significant in accordance with our previous findings reported in Table IV.
Furthermore, the coefficient of CRED_RIGH is positive and statistically significant at the 5
percent level, implying that during the crisis period stronger creditor rights lead to higher
credit risk. In fact, creditor protection encourages MFIs to take more risk by lending
customers with financial difficulties, which are amplified by the crisis.
Model (2) displays the results estimated by excluding the crisis years. We confirm
our previous estimation for the baseline model (Table IV ). In fact, we found the same
significant coefficients.

Variables Model (1) Model (2)

GR_LEND −0.009 (−1.336) −0.007 (−2.311)**


REP_SHC 0.015 (1.513) 0.001 (0.0448)
SAV_DUM −0.032 (−2.887)*** 0.001 (0.077)
INS_DUM −0.008 (−1.027) −0.003 (−0.718)
WOM_BOR −0.011 (−0.377) −0.065 (−5.017)***
DIVER −0.002 (−0.991) −0.001 (−1.776)*
LAW_ENF 0.018 (0.547) 0.070 (6.737)***
CRED_RIGH 0.331 (2.223)** 0.086 (1.460)
INF_SHAR −0.005 (−2.550)** −0.002 (−1.874)*
Control variables
CAP_ADQ −0.036 (−1.458) 0.009 (0.818)
SIZE 0.002 (0.660) −0.001 (−0.418)
PROF −0.172 (−1.258) −0.185 (−3.357)***
OUTS_LOAN −0.037 (−0.804) −0.0214 (−1.228)
OUTREACH −0.006 (−1.515) −0.003 (−1.656)*
GOV_IND −0.017 (−0.652) −0.044 (−2.531)**
GDP_GR −0.278 (−1.990)** −0.220 (−3.528)***
INFLATION 0.005 (0.049) 0.026 (0.426)
Table V.
Determinants of MFI’s INT_RATE 0.088 (2.733)*** 0.025 (1.003)
credit risk during the Observations
2
1,531 3,848
crisis and out of the R 0.1727 0.2487
crisis period Note: *,**,***Significance at 10, 5 and 1 percent levels, respectively
5.4 Robustness check Credit risk of
To ensure the robustness of our results, we run several sensitivity tests. First, we assessed microfinance
the results’ sensitivity to a change in the measures of credit risk indicators: PAR_90, LLP institutions
and WOR.
The results, reported in Table VI, show that determinants of credit risk measured by
PAR_90 are the same like those of PAR_30. However, for the other risk measures, we found
the same disparities. In fact, WOM_BOR is still relevant for the other risk measures.
However, the coefficient of GR_LEND is negative and significant for only the loss loan
provision model. CRED_RIGH has positive and significant coefficients for the loss
loan provision and the write-off ratio specifications, implying that larger MFIs or those with
larger outstanding loans or having creditor protection rights have more incentive to lend
riskier borrowers.
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

Second, we test whether credit risk determinants differ by legal status. In fact,
shareholders or investors of profit MFIs want to maximize returns while shareholders or
donors in nonprofit MFIs want to further the institution’s social goals. In both cases, MFIs
should manage their risk in a sustainable way.
We also rerun our model for regulated and unregulated MFIs since regulated MFIs are
usually subject to capital regulations, reducing owners’ risk taking by forcing them to invest
more capital (Kim and Santomero, 1994).
Our results, displayed in Table VII, indicate that profit and nonprofit MFIs have
approximately the same specific determinants; GR_LEN and WOM_BOR matter for both
groups. However, the coefficient of DIVER is negative and significant for profit MFIs,
suggesting that the former reduce their risk by using non-interest income activities. As for
the nonprofit group, the coefficient of outreach is negative and significant indicating that
credit amounts decrease credit risk. This finding implies that nonprofit MFIs could reduce
credit risk by lending large amounts to safer borrowers.

Variables PAR_90 LLP WOR

GR_LEND −0.006 (−2.792)*** −0.003 (−2.221)** −0.001 (−0.049)


REP_SHC 0.005 (1.370) −0.001 (−0.338) 0.001 (0.045)
SAV_DUM −0.002 (−0.810) −0.003 (−1.583) −0.004 (−1.375)
INS_DUM −0.002 (−0.616) −0.001 (−0.380) 0.001 (0.322)
WOM_BOR −0.028 (−3.739)*** −0.007 (−1.774)* −0.005 (−2.197)**
DIVER −0.031 (−2.424)** 0.008 (0.902) −0.0147 (−1.011)
LAW_ENF 0.016 (1.749)* −0.001 (−0.338) 0.005 (0.696)
CRED_RIGH 0.075 (1.466) 0.076 (2.418)** 0.077 (1.734)*
INF_SHAR −0.002 (−2.000)** −0.001 (−0.261) −0.001 (−0.073)
Control variables
CAP_ADQ 0.005 (0.589) 0.003 (1.042) 0.009 (2.168)**
SIZE 0.001 (0.542) 0.002 (3.644)*** 0.002 (2.140)**
PROF −0.178 (−3.247)*** −0.086 (−5.905)*** −0.099 (−5.619)***
OUTS_LOAN −0.006 (−0.495) 0.018 (2.837)*** 0.019 (2.523)**
OUTREACH −0.002 (−1.611)* −0.001 (−1.266) −0.001 (−0.866)
GOV_IND −0.011 (−1.852)* 0.005 (0.706) 0.014 (1.356)
GDP_GR −0.164 (−4.161)*** −0.128 (−2.801)*** −0.094 (−2.508)**
INFLATION −0.039 (−0.838) −0.007 (−0.331) −0.036 (−1.782)*
INT_RATE 0.017 (0.895) 0.003 (0.373) 0.007 (0.609)
Observations 5,379 5,379 5,379
2
R 0.1223 0.1600 0.2147 Table VI.
Note: *,**,***Significance at 10, 5, 1 percent levels, respectively Robustness tests
IJMF Variables Profit Nonprofit Regulated Unregulated

GR_LEND −0.005 (−2.397)** −0.006 (−2.399)** −0.008 (−2.730)*** −0.007 (−1.744)*


REP_SHC −0.001 (−0.194) 0.007 (1.623) 0.003 (0.733) 0.007 (1.344)
SAV_DUM −0.004 (−1.010) −0.003 (−0.446) −0.008 (−1.422) 0.003 (0.491)
INS_DUM −0.007 (−0.763) 0.001 (0.0235) −0.003 (−0.527) 0.002 (0.303)
WOM_BOR −0.039 (−2.444)** −0.056 (−3.504)*** −0.055 (−3.790)*** −0.035 (−2.050)**
DIVER −0.223 (−3.009)*** −0.040 (−1.595) −0.009 (−1.808)* −0.001 (−2.000)**
LAW_ENF 0.094 (7.864)*** 0.012 (0.716) 0.092 (6.781)*** 0.001 (0.072)
CRED_RIGH −0.0742 (−0.694) 0.1900 (2.864)*** 0.018 (0.182) 0.171 (2.489)**
INF_SHAR −0.001 (−0.665) −0.001 (−2.859)*** −0.002 (−1.321) −0.004 (−2.955)***
Control variables
CAP_ADQ 0.020 (1.136) 0.002 (0.146) −0.002 (−0.172) 0.025 (1.394)
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

SIZE 0.001 (0.242) 0.001 (0.093) 0.001 (0.200) 0.004 (0.018)


PROF −0.234 (−2.550)** −0.233 (−3.378)*** −0.241 (−2.401)** −0.222 (−3.141)***
OUTS_LOAN −0.014 (−0.708) −0.036 (−1.615) −0.016 (−0.752) −0.032 (−1.333)
OUTREACH −0.002 (−0.888) −0.005 (−2.324)** −0.004 (−1.010) −0.003 (−2.098)**
GOV_IND −0.036 (−1.463) −0.037 (−1.872)* 0.036 (1.572) −0.042 (−2.033)**
GDP_GR −0.349 (−3.533)*** −0.173 (−3.485)*** −0.266 (−3.362)*** −0.211 (−2.922)***
INFLATION −0.112 (−1.308) 0.035 (0.583) −0.104 (−1.287) 0.160 (1.488)
Table VII. INT_RATE 0.020 (0.439) 0.052 (2.177)** 0.013 (0.229) 0.059 (3.6)***
Credit risk Observations 3,130 2,249 3,914 1,465
2
determinant: profit vs R 0.2368 0.2093 0.2490 0.1964
nonprofit MFIs Note: *,**,***Significance at 10, 5, 1 percent levels, respectively

As for the country-level indicator, we found that only the coefficient of LAW_ENF is
significant for profit MFIs. For nonprofit MFIs, the country-level indicator seems to be more
relevant. Indeed, information sharing (INF_SHAR) and institutional environment
(GOV_IND) decrease risk, though creditor rights (CRED_RIGH) increase credit risk.
In light of these findings, it seems that credit risk for nonprofit MFIs depends more on
economic and institutional factors since these institutions have mainly social goals.
Columns (3) and (4) display the results for regulated and unregulated MFIs. It seems
that regulated and profit MFIs (nonprofit and unregulated) have the same determinants.
This similarity seems logical since according to Cull et al. (2011) nonprofit MFIs are in
general unregulated.

6. Conclusion
A large literature looking at cross-country development highlights the importance of MFIs
in alleviating poverty. Because of their importance, there is widespread interest in
understanding the factors that affect MFIs’ risk management. At the same time, the recent
financial crisis has revived interest in how the institutional and regulatory environment
affect credit risk. In this regard, the purpose of this paper is to explore the main
determinants of credit risk in MFIs around the world. In addition to specific factors, we
considered institutional variables as determinants of credit risk allowing for a more
comprehensible view of how to manage credit risk.
Examining a sample of 638 MFIs from 87 countries over the 2005-2015 period, we found
that the group-lending methodology, percent of loan granted to women and diversification
activities are the important MFI features that reduce credit risk.
Furthermore, for the country-level indicators, we found that information published by
public or private agencies affects, while law enforcement cost increases it. Our results
indicate that credit risk also tends to be limited in a good institutional environment.
These findings are found to be robust during and out of the crisis period and using several Credit risk of
sensitivity tests. microfinance
Several implications can be drawn in light of these findings. MFIs’ managers should not institutions
fear to use the group-lending methodology or to grant more credits to women. Moreover,
they should diversify their activities by entering into new business lines to reduce risk.
The results send strong signals about the role of institutional quality in limiting credit
risk of MFIs. It is obvious that strengthening debt enforcement institutions is crucially
needed in order to enable MFIs to limit their credit risk. Therefore, authorities need to
encourage setting up credit agencies, if they do not exist and find adequate solutions
to reduce enforcement costs. Finally, a prerequisite seems to be a strong institutional
environment that helps MFIs to limit their credit risk.
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

Notes
1. www.microworld.org
2. Correlation matrix is not reported because of space constraints.

References
Aghion, P. and Bolton, P. (1992), “An incomplete contracts approach to financial contracting”, Review of
Economic Studies, Vol. 59 No. 3, pp. 473-494.
Ahlin, C., Lin, J. and Maio, M. (2011), “Where does microfinance flourish? Microfinance institutions
performance in macroeconomic context”, Journal of Development Economics, Vol. 95 No. 2,
pp. 105-120.
Armendariz de Aghion, B. and Morduch, J. (2000), “Microfinance beyond group lending”, Economics of
Transition, Vol. 8 No. 2, pp. 401-420.
Armendáriz de Aghion, B. and Morduch, J. (2010), The Economics of Microfinance, 2nd ed., MIT Press,
Cambridge, MA.
Ayayi, A.G. (2012), “Credit risk assessment in the microfinance industry”, Economics of Transition,
Vol. 20 No. 1, pp. 37-72.
Bae, K.H. and Goyal, V.K. (2009), “Creditor rights, enforcement, and bank loans”, Journal of Finance,
Vol. 64, No. 2, pp. 823-860.
Berger, A.N. and DeYoung, R. (1997), “Problem loans and cost efficiency in commercial banks”, Journal
of Banking & Finance, Vol. 21 No. 6, pp. 849-870.
Bhatt, N. and Tang, S.-Y. (2002), “Determinants of repayment in microcredit: evidence from programs
in the United States”, International Journal of Urban and Regional Research, Vol. 26 No. 2,
pp. 360-376.
Boehe, D.M. and Cruz, B.L. (2013), “Gender and microfinance performance: why does the institutional
context matter?”, World Development, Vol. 47, pp. 121-135.
Breuer, J.B. (2006), “Problem bank loans, conflicts of interest and institutions”, Journal of Financial
Stability, Vol. 2 No. 3, pp. 266-285.
Brown, M., Jappelli, T. and Pagano, M. (2009), “Information sharing and credit: firm-level evidence from
transition countries”, Journal of Financial Intermediation, Vol. 18 No. 2, pp. 151-172.
Crabb, P.R. and Keller, T. (2006), “A test of portfolio risk in microfinance institutions”, Faith and
Economics, Nos 47/48, pp. 25-39.
Croson, R. and Gneezy, U. (2009), “Gender differences in preferences”, Journal of Economic Literature,
Vol. 47 No. 2, pp. 448-474.
Cull, R., Demirguc-kunt, A. and Morduch, J. (2011), “Does regulatory supervision curtail microfinance
profitability and outreach?”, World Development, Vol. 39 No. 6, pp. 949-965.
IJMF D’Espallier, B., Guérin, I. and Mersland, R. (2011), “Women and repayment in microfinance: a global
analysis”, World Development, Vol. 39 No. 5, pp. 758-772.
Djankov, S., Hart, O., McLiesh, C. and Shleifer, A. (2007), “Private credit in 129 countries”, Journal of
Financial Economics, Vol. 84 No. 2, pp. 229-329.
Djankov, S., Hart, O., McLiesh, C. and Shleifer, A. (2008), “Debt enforcement around the world”, Journal
of Political Economy, Vol. 116 No. 6, pp. 1105-1149.
Galema, R. (2011), “Debt enforcement and microfinance risk”, working paper, University of Groningen.
Ge, X., Kim, J.B. and Song, B.Y. (2012), “Internal governance, legal institutions and bank loan
contracting”, Journal of Corporate Finance, Vol. 18 No. 3, pp. 413-432.
Godlewski, C.J. (2004), “Bank capital and credit risk taking in emerging market economies”, Journal of
Banking Regulation, Vol. 6 No. 2, pp. 128-145.
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

Gonzalez, A. (2007), “Resilience of microfinance institutions to national macroeconomic events:


an econometric analysis of MFI asset quality”, MIX Discussion Paper No. 1, Washington, DC.
Gool, J.V.V., Sercu, W.P. and Baesens, B. (2011), “Credit scoring for microfinance: is it worth it?”,
International Journal of Finance and Economics, Vol. 17 No. 2, pp. 103-123.
Houston, J.F., Lin, C., Lin, P. and Ma, Y. (2010), “Creditor rights, information sharing and bank risk
taking”, Journal of Financial Economics, Vol. 96 No. 3, pp. 485-512.
Hulme, D. and Mosley, P. (1996), Finance Against Poverty, Routledge, London.
Islam, M. (1995), “Peer monitoring in the credit market”, Journal of Contemporary Asia, Vol. 26 No. 4,
pp. 452-465.
Jappelli, T. and Pagano, M. (2002), “Information sharing, lending and defaults: cross-country evidence”,
Journal of Banking and Finance, Vol. 26 No. 10, pp. 2017-2045.
Kaufmann, D., Kraay, A. and Mastruzzi, M. (2010), “The worldwide governance indicators: a summary of
methodology, data and analytical issues”, World Bank Policy Research Working Paper No. 5430.
Kim, D. and Santomero, A.M. (1994), “Risk in banking and capital regulation”, Journal of Finance,
Vol. 43 No. 5, pp. 1219-1233.
Kono, H. and Takahashi, K. (2010), “Microfinance revolution: its effects, innovations, and challenges”,
Developing Economies, Vol. 48 No. 1, pp. 15-73.
Krauss, N. and Walter, I. (2009), “Can microfinance reduce portfolio volatility?”, Economic Development
and Cultural Change, Vol. 58 No. 1, pp. 85-110.
Kumar, R. and Qazi, M. (2016), The Essential Microfinance, Notion Press, Chennai.
La Porta, R., López-de-Silanes, F., Shleiffer, A. and Vishny, R. (1998), “Law and finance”, Journal of
Political Economy, Vol. 106 No. 6, pp. 1113-1155.
Ledgerwood, J. (1999), Microfinance Handbook: An Institutional and Financial Perspective, World Bank
Publications, Washington, DC.
Madajewicz, M. (2011), “Joint liability versus individual liability in credit contracts”, Journal of
Economic Behavior and Organization, Vol. 77 No. 2, pp. 107-123.
Mersland, R. and Strøm, R.Ø. (2010), “Microfinance mission drift?”, World Development, Vol. 38 No. 1,
pp. 28-36.
Mersland, R., Randøy, T. and Strøm, R.Ø. (2011), “The impact of international influence on microbanks
performance: a global survey”, International Business Review, Vol. 20 No. 2, pp. 163-176.
Milana, C. and Ashta, A. (2012), “Developing microfinance: a survey of the literature”, Strategic Change:
Briefings in Entrepreneurial Finance, Vol. 21 Nos 7/8, pp. 299-330.
Morduch, J. (1999), “The promise of microfinance”, Journal of Economic Literature, Vol. 37 No. 4,
pp. 1569-1614.
Padilla, A.J. and Pagano, M. (2000), “Sharing default information as a borrower discipline device”,
European Economic Review, Vol. 44 No. 10, pp. 1951-1980.
Peterson, M. (2009), “Estimating standard errors in finance panel datasets: comparing approaches”, Credit risk of
Review of Financial Studies, Vol. 22 No. 1, pp. 435-480. microfinance
Phillips, C. and Bhatia-Panthaki, S. (2007), “Enterprise development in Zambia: reflections on the institutions
missing middle”, Journal of International Development, Vol. 19 No. 6, pp. 793-802.
Saravia-Matus, S.L. and Saravia-Matus, J.A. (2015), “Gender issues in microfinance and repayment
performance: the case of a Nicaraguan microfinance institution”, Encuentro: RevistaAcadémica
de la Universidad Centroamericana, No. 91, pp. 7-31.
Savita, S. (2007), “Transaction costs in group microcredit in India management decisions”, Vol. 45
No. 8, pp. 1331-1342.
Servin, R., Lensink, R. and Van den Berg, M. (2012), “Ownership and technical efficiency of
microfinance institutions: empirical evidence from Latin America”, Journal of Banking and
Finance, Vol. 36 No. 7, pp. 2136-2144.
Downloaded by UNIVERSITY OF ADELAIDE At 01:45 05 September 2017 (PT)

Sharma, M. and Zeller, M. (1997), “Repayment performance in group‐based credit programs in


Bangladesh: an empirical analysis”, World Development, Vol. 25 No. 10, pp. 1731-1742.
Stiglitz, J. (1990), “Peer monitoring and credit markets”, World Bank Economic Review, Vol. 4 No. 3,
pp. 351-366.
Tchakoute-Tchuigoua, H. (2014), “Institutional framework and capital structure of microfinance
institutions”, Journal of Business Research, Vol. 67 No. 10, pp. 2185-2197.
Tchakoute-Tchuigoua, H. (2015), “Capital structure of microfinance institutions”, Journal of Financial
Services Research, Vol. 47 No. 3, pp. 313-340.
Tchakoute-Tchuigoua, H. and Nekhili, M. (2012), “Gestion des risques et performance des institutions
de microfinance”, Revue d'économie industrielle, Vol. 138 No. 2, pp. 127-146.
Wagner, C. and Winkler, A. (2013), “The vulnerability of microfinance to financial turmoil – evidence
from the global financial crisis”, World Development, Vol. 51 No. 11, pp. 71-90.
Yunus, M. (1999), Banker to the Poor: Micro-Lending and the Battle Against World Poverty, Public
Affairs, New York, NY.
Zeller, M. (1998), “Determinant of repayment performance in credit groups: the role of program design,
intragroup risk pooling, and social cohesion”, Economic Development and Cultural Change,
Vol. 46 No. 3, pp. 599-621.
Zhao, T., Murinde, V. and Mlambo, K.? (2011), “How does the institutional setting for creditor rights
affect bank lending and risk taking?”, available at: www.management.stir.ac.uk/research/
economics/workingpapers (accessed November 20, 2016).

Corresponding author
Naima Lassoued can be contacted at: naima.lassoued@sesame.com.tn

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com

You might also like