International Debt Financing and Performance of Microfinance Institutions

This article analyzes the relationship between international debt financing and the performance of microfinance institutions (MFIs). The study uses data from 319 MFIs in 68 developing countries to examine whether there is a relationship between an MFI’s access to international debt and its financial and social performance. The study finds that access to commercial debt is related to strong financial performance, a high level of professionalization, and a low average loan size indicating outreach to poor customers. The targeting of women is not a priority for MFIs accessing international commercial debt. As for MFIs accessing subsidized international debt, they target female customers to a greater extent than other MFIs.

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/264343098
International Debt Financing and Performance of Microfinance Institutions
Article in Strategic Change · February 2013
DOI: 10.1002/jsc.1919
CITATIONS
21
READS
357
2 authors:
Some of the authors of this publication are also working on these related projects:
Livelihoods, Microfinance & Disability View project
Dataset combining financial and social ratings of Microfinance InstitutionsView project
Roy Mersland
Universitetet i Agder
87 PUBLICATIONS 1,677 CITATIONS
SEE PROFILE
Ludovic Urgeghe
Université de Mons
10 PUBLICATIONS 34 CITATIONS
SEE PROFILE

Strat. Change 22: 17–29 (2013)
Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/jsc.1919 RESEARCH ARTICLE
Copyright © 2013 John Wiley & Sons, Ltd.
Strategic Change: Briefi ngs in Entrepreneurial Finance
Strategic Change
DOI: 10.1002/jsc.1919
International Debt Financing and Performanc
Microfi nance Institutions1
Roy Mersland
University of Agder, Norway
Ludovic Urgeghe
Center for European Research in Microfi nance, University of Mons, Belgium
Introduction
During recent decades, the provision of microfi nance services to poo
micro-entrepreneurs has evolved to become a global industry. Until r
tions and subsidies have been the main source of funding for microfi
tions (MFIs). Lately, however, the growth of the industry and the pres
toward fi nancial sustainability has pushed MFIs to turn to intern
markets. Moreover, international funding is regarded by many as ess
the growth of the sector, arguing that only international capital mark
the estimated US$200 billion needed to reach the potential demand
nance services worldwide (Swanson, 2008). Recent academic researc
al., 2011) has also shown that internationalization, notably through in
can have an overall positive infl uence on the social performance of M
Th e development of specialized investment funds, called microfi
ment vehicles (MIVs), illustrates the emergence of this new spe
market. MFIs typically have both fi nancial and social objectives (Arm
Morduch, 2010) and attract funding from actors with varying degrees
motivation,from purelydevelopment-orientedto maximumprofi t-oriented
(Goodman, 2004). In 2010, the 95 MIVs in operation managed US$8 b
coming from public and private institutional investors (42%), individu
development institutions (21%), and others (3%), mostly invested in
loans to MFIs2 (MicroRate, 2011; Reille et al., 2011).
The international fi nancing of
microfi nance has become a new
specialized market which attracts
investors with varying degrees of
profi t motivation.
Investors lending at commercial
rates target MFIs with relatively
better fi nancial performance,
while those lending at subsidized
rates target fi nancially weaker
MFIs that focus on female
customers.
Commercial funding to microfi nance institutions (MFIs) seems to follow th
negative screening approach, being driven mainly by fi nancial performa
professionalization of the MFIs while subsidized funding seems to follow a po
approach, being driven mainly by targeting poverty alleviation and social inc
1 JEL classifi cation codes: G11, G23, L2, O16, O17.
2 Th e repartition of microfi nance assets invested by MIVs in 2010 was 82%
18% equity (MicroRate, 2011).

18 Roy Mersland and Ludovic Urgeghe
Copyright © 2013 John Wiley & Sons, Ltd. Strategic Change
DOI: 10.1002/jsc
Th is article examines the profi les of the MFIs receiv-
ing loans from MIVs. More specifi cally, using data from
319 MFIs in 68 developing countries, we study whether
there is a relationship between an MFI’s access to inter-
national debt and its fi nancial and social performance. Wefi nd that access to commercial debt is related to strongfi nancial performance, a high level of professionalization,
and a low average loan size indicating outreach to poor
customers. Th e targeting of women is not a priority for
MFIs accessing international commercial debt. As for
MFIs accessing subsidized international debt, they target
female customers to a greater extent than other MFIs.
Th e rest of this article is organized as follows. Th e next
section discusses how the fi nancial and social perfor-
mances of MFIs infl uence the type of funding received,
and outlines the hypotheses to be tested. Th e third section
explains the model, the methodology, and the dataset used
for estimations, while the fourth section presents and
discusses the fi ndings. Th e fi fth section concludes.
International funding and the performance
of MFIs
In this section we develop hypotheses on how interna-
tional funding is associated with the social and fi nancial
performances of MFIs.
The relationship between international funding and
MFI social performance
First, we investigate the link between the MFI’s social
performance and its access to international funding. As all
MIVs claim to off er social returns to investors, they belong
to the fi eld of socially responsible investments (SRIs).
Indeed, an SRI is “an investment process that integrates
social, environmental and ethical considerations into invest-
ment decision making” (Renneboog et al., 2008, p. 1). In
other words, we label “socially responsible” any invest-
ment that is linked to the corporate social responsibility
(CSR) of the target fi rm. In its modern understanding,
CSR not only involves the ethical obligations of fi rms
toward their stakeholders, but also requires inves
projects that yield social and economic benefi ts (Carr
1979; Porter and Kramer, 2002). In the microfi na
world, CSR would then mean that MFIs fulfi ll their soc
mission in an economically sustainable way.
Th ere are two approaches for responsible investm
selection: negative screening and positive screening
2007; Juravle and Lewis, 2008). Negative screening (a
called avoidance, or exclusion) involves a two-step pr
First, the investment manager excludes specifi c fi eld
activities that investors consider undesirable (for instafi rms involved in weapons, alcohol, or tobacco). T
investments are selected by a classical risk/return an
In contrast, with positive screening, nothing is exclude
beforehand but investments are selected primarily
non-fi nancial criteria (e.g., high environmental or soc
performance).
We will test two hypotheses. In the fi rst one, MIVs
use a positive screening approach and we expect to fi
positive relationship between the social performance
MFI and its access to international funding. In the seco
one, they use a negative screening approach and we
to fi nd a positive relationship with fi nancial performa
and none with social performance. Th e hypothes
negative screening in microfi nance is based on t
that MIVs consider microfi nance a social investment
se, as if they avoid or exclude any other activity which
not microfi nance, and then apply a typical fi nancial a
sis to the remaining potential investment projects.
Basedon the above,we proposethe following
hypotheses.
In the case of positive screening
H1a: Th e presence of international funding in an
is positively related to its social performance
In the case of negative screening
H1b: Th ere is no relationship between the presen
international funding in an MFI and its social
performance, but a positive relationship with fi nan
performance

Debt Financing and Performance 19
Copyright © 2013 John Wiley & Sons, Ltd. Strategic Change
DOI: 10.1002/jsc
The relationship between international funding and
MFI fi nancial performance
To propose hypotheses on the infl uence of an MFI’s fi nan-
cial performance on its access to international funding,
we make the assumption that the microfi nance invest-
ment landscape is as described by Goodman (2004): on
the one hand, development-oriented investors fi nance notfi nancially sustainable MFIs with grants, subsidized loans,
or donated equity while on the other hand, commercial
investorsfund fi nanciallywell-performingMFIs with
loans and equity at market prices. Th erefore, and as we
focus on debt investments, the distinction should be made
betweencommercialand subsidizedloans.Loansare
labeled “commercial” when the MFI has to pay interest at
the market rate, and “subsidized” if the interest rate is
below the market conditions.
Commercial funding and MFI performance
At its best, microfi nance has proven that it can generate
profi t and growth while being low risk (Swanson, 2008).
According to a study of MIV portfolios by Oehri and
Fausch (2008), microfi nance investments show low vola-
tility and low correlation to other asset classes, which
potentially makes microfi nance an interesting asset to
include in a portfolio for commercial investors.
Building on business lifecycle theory, which states that
the development of organizations depends on their capac-
ity to accessadaptedfundingsources(Little, 1974;
Channon, 2006), several authors (Kooi, 2001; de Sousa-
Shieldsand Frankiewicz,2004;Van Maanen,2005;
Bogan, 2008) argue that MFIs should be funded as
follows. In the youth phase, MFIs need highly risk-tolerant
subsidized capital in the form of grants and donated
equity to support the early years of operation as MFIs are
not sustainable enough to attract commercial funding. In
the growth phase, MFIs must increase their scale and gain
market shares with retained earnings and subsidized loans
as the main sources of funding. Th is stage is also when,
by complying with stricter banking regulations and trans-
parency standards, MFIs can make the transition from
non-profi t organizations to regulated institutions s
they can mobilize deposits and have easier access
mercial funding. Regarding this specifi c issue,
(2008) notes that this transition to a regulated ent
an expensive and diffi cult process that also requi
sidized funding. Consequently, many large and
lished MFIs continue to receive support to fi na
transition in the form of grants and subsidized loan
with risk capital provided primarily by socially orie
investors. Th e last stage of the lifecycle is maturi
when the MFIs are formal regulated banks with
structures similar to those of commercial banks (B
2008). Th us, mature MFIs should be funded most
deposits,local capitalmarkets,and commercialdebt
coming from international funds.
Takentogether,commercialinternationalfunding
should be positively related to the fi nancial perfor
of the MFI, as outlined in this second hypothesis
H2: Th e presence of international commercial
in an MFI is positively related to its fi nancial
performance
Subsidized funding and MFI performance
As for subsidized funding, the lifecycle theory
that MFIs in their early stages need subsidized fun
compensate for their lack of profi tability. We could
fore, expect that international subsidized funding i
tively related to the MFI’s fi nancial performance. H
the relationship might not be that clear cut. Th e S
erature provides insight into what type of MFIs the
oriented investors would typically target. As pr
outlined, social investors put their money into
that yield social benefi ts. However, socially orient
tors also intend to ensure good economic perf
from their investments (Porter and Kramer, 2002).
fore, MIVs claim to have “double bottom line” obje
and thus they invest in socially and fi nancially
MFIs. Moreover, De Schrevel et al. (2009) indicate
the rapid growth of MIVs between 2004 and 2

20 Roy Mersland and Ludovic Urgeghe
Copyright © 2013 John Wiley & Sons, Ltd. Strategic Change
DOI: 10.1002/jsc
explained by a narrow targeting of the most profi table and
professional MFIs. Th is could indicate that there is a posi-
tive relationship between access to subsidized funding and
the fi nancial performance of the MFI.
To summarize, we propose the following two alterna-
tive hypotheses for the relationship between international
subsidized funding in an MFI and the MFI’s fi nancial
performance
H3a: Th e presence of international subsidized funding
in an MFI is negatively related to its fi nancial
performance
H3b: Th e presence of international subsidized funding
in an MFI is positively related to its fi nancial
performance
Data and methodology
Dataset and descriptive statistics
Th e dataset comprises up to fi ve years of data from 319
MFIs in 68 developing countries. Th e information has
been compiled from risk assessment reports prepared byfi ve rating agenciesspecializingin microfi nance:
MicroRate,Microfi nanza,PlanetRating,Crisil, and
M-Cril. Comparisons of the methodologies applied by the
rating agencies reveal no major diff erences in MFI assess-
ment relevant for variables included in this study. Th e
dataset has a certain sample selection bias as only rated
MFIs are included. Th ey represent internationally ori-
ented MFIs with the intention to practice microfi nance
in a business-oriented manner, and they have the greatest
likelihood of achieving the dual goal of social and fi nan-
cial performance.
Th e rating agencies diff er in their emphasis and in the
abundanceof availableinformation.Th us,diff erent
numbers of observations on diff erent variables in diff erent
years are reported. Th e rating reports comprising the data
used for this study are from 2001 to 2008, with the vast
majority from 2005 to 2008.
Variables
Dependent variables
We will test our hypotheses on three dependent varia
First, we use a dummy stating whether the MFI h
international debt at all (1 for yes and 0 for no) with
diff erence between commercial or subsidized debt. T
we split this variable in two: commercial debt
one side and subsidized debt only on the other side
on the interest rate reported in rating reports compar
to the market rate in the country.
Financial performance
To proxy the MFI’s fi nancial performance, we use
return on assets (ROA), the operating expense ratio, a
the 30-day portfolio-at-risk (SEEP Network, 2005).
Th e ROA indicates how well the MFI is able to gen
ate profi t from its assets and is calculated as (Net op
income — Taxes)/Average annual assets.
Th e operating expense ratio, calculated as Op
expenses/Average annual loan portfolio, assesses the
ciency of an MFI’s activities. A lower level of operating
expenses indicates that the MFI is more effi cient tha
with higher operating expenses.
Loan portfolio quality is crucial as it represents
quality of the MFI’s largest asset. Th e risk associated
poor management of the portfolio can be dramatic, es
since microloans are generally not backed with banka
lateral (Jansson, 2003). We use the 30-day portfolio
which measures the share of the MFI’s outstandin
portfolio with more than 30 days in arrears.
Social performance
Obtaining measurable and trustable MFI’s data on soc
performanceis diffi cult.Consequently,the following
measures have been used extensively in the microfi n
literature.
Th e average loan size (Cull et al., 2007; De Bru
2008; Mersland and Strøm, 2010; Lensink et al., 2011
According to Schreiner (2002), a lower loan size indic

Debt Financing and Performance 21
Copyright © 2013 John Wiley & Sons, Ltd. Strategic Change
DOI: 10.1002/jsc
that the MFI reaches out to poorer customers. To ensure
comparability between countries, we take the average loan
size as a percentage of per capita gross national income
(GNI).
Th e targeting of women (De Bruyne, 2008; Arm-
endariz and Morduch, 2010; Mersland and Strøm, 2010;
D’Espallier et al., 2011). We use a time-invariant dummy
that indicates whether the MFI has a conscious bias
toward lending to women as indicated in the rating reports
(D’Espallier et al., 2011).
Th e rural outreach (De Bruyne, 2008; Mersland and
Strøm, 2010). We use a dummy variable defi ning whether
the MFI serves rural markets. As rural areas are generally
in fi nancial need and more diffi cult for MFIs to penetrate,
better rural outreach can be considered an indicator of
higher social performance.
Controls
We also include a number of control variables that could
infl uence whether an international MIV would lend to an
MFI. First, we include institution-specifi c controls: size
(logarithm of MFI assets); age (number of years since
start-up of MFI); a dummy stating whether the MFI was
originated by an international initiator, as Mersland et al.
(2011) show international orientation can have an impact
on social performance of MFIs; a dummy indicating
whether the MFI mobilizes voluntary savings; and the
level of professionalization proxied by a dummy for the
presence of an internal auditor reporting to the board. We
also include contextual control variables. First, the human
development index (HDI) to control for development
diff erences across countries and second, regional dummies
to capture diff erences across geographical regions (Latin
America, MENA region, EECA region, Asia, and Africa).3
Summary statistics
A total of 65% of the MFIs in our sample have inter
tional debt. Of those having international debt 30%
only commercial debt, 42% have only subsidize
and 28% have both types of debt.
Table 1 provides descriptive statistics for all va
used in the study. Th e average ROA is 0.8%, whil
operating expense ratio is 35.7%, illustrating the h
of microlending. Indeed, the operating expense
calculated as (Personnel costs + Administrative
Average total loan portfolio, is always higher in mic
nance than in “classical” commercial banking, and
mainlydue to the decentralizedcreditmethodology
(microcredit offi cers go every day to clients’
for cash disbursements and collection of repaymen
the small size of the transactions involved, which m
scale economies diffi cult. Th e average PAR30
With respect to social performance, the average lo
represents, on average, 52.6% of the gross nationa
per capita in the country; 47% of MFIs have a
favor of targeting women and 18% operate only in
areas. Th e average MFI has been operating for ni
Only 19% of the MFIs collect voluntary savings, wh
suggests that sample MFIs are primarily non-re
institutions.As for geographicaldistribution,Latin
America represents 45% of the observations follow
Eastern Europe and Central Asia with 21%.
Table 2 shows the correlation matrix of the var
High correlations among explicative variables can
a multicollinearity problem which would bias the in
pretation of results. According to Kennedy (2008),
relationsmust be at least0.8 to detectpotential
multicollinearity problems between variables, and
tratedin Table2 we can rule out problemswith
multicollinearity.
Estimation method
To determine which type of performance is as
with MFIs receiving international investments, w
pooled probit regressions. In probit regressions, th
3 We are aware that regional dummies only to a limited degree
refl ect the political and economic risk of each specifi c country,
but controlling for each country would require a much larger
dataset. Moreover, with the inclusion of the HDI we do
control for individual country diff erences as the HDI captures
both the social and economic development of a country.

22 Roy Mersland and Ludovic Urgeghe
Copyright © 2013 John Wiley & Sons, Ltd. Strategic Change
DOI: 10.1002/jsc
Table 1. Summary statistic
Obs. Mean Std dev. Min Max
Financial performance
ROA 785 0.008 0.13 −0.99 0.34
Operating expenses ratio 773 0.357 0.51 0.02 11.32
Portfolio-at-risk 763 0.065 0.11 0.00 0.97
Social performance
Average loan/GNI per capita 810 0.526 0.69 0.03 5.16
Women targeting 801 0.473 0.50 0 1
Dummy rural market 792 0.176 0.38 0 1
Control variables
Logarithm of assets 800 14.716 1.32 10.60 18.26
MFI age 810 9.142 7.00 0.00 42.00
Dummy international initiator 808 0.402 0.49 0 1
Dummy voluntary savings 810 0.194 0.40 0 1
Dummy internal auditor 715 0.456 0.50 0 1
HDI 810 0.710 0.12 0.37 0.87
Cross-table — Number of MFI fi rm years per type of debt and region
Region Latin AmericaAfrica Asia EECA MENA Total % of total
Commercial debt 97 38 13 40 4 192 20%
Subsidized debt 98 45 60 49 15 267 27%
Both types of debt 87 15 13 57 11 183 19%
No international debt 161 69 34 55 19 338 34%
Total 443 167 120 201 49 980
% of total 45% 17% 12% 21% 5%fi cients of the explicative variables cannot be interpreted
as marginal eff ects on the dependent variable, and their
signs show whether the corresponding variable infl uences
positively or negatively the likelihood for the dependent
variable to equal 1. Coeffi cients are estimated using the
maximum likelihood method (Stock and Watson, 2006).
As the data have a panel structure but the two dependent
variables (commercial debt and subsidized debt) were
reported only for the last year in the rating reports, we
assume them to be constant over time. Th is assumption
is natural as MFIs tend to keep international debt once
received. In addition, the assumption corresponds to the
reality behind investments as investors include historical
performance when making their funding decisions. Th ere-
fore, we run cross-section pooled regressions. Moreov
as robustness checks (unreported) we have run single
(ratingyear)and double-year(ratingyear + previous
year) regressions, and the fi ndings generally confi rm
results reported below. In all regressions, we use
standard errors to correct for heteroskedasticity. Data
also been tested and treated for outliers using Grubbs
(Iglewicz and Hoaglin, 1993).4 Finally, we run regressions
with and without the MFI and country control variable
All three regressions are detailed in the Appendix.
4 Th ough robustness checks show that outliers don’t infl u
the results much, we have trimmed the dataset and left ou
from the analyses MFIs with average loans/GNI below 0.1
and above 5.5 as these represent extreme cases.

Debt Financing and Performance 23
Copyright © 2013 John Wiley & Sons, Ltd. Strategic Change
DOI: 10.1002/jsc
Table 2. Correlations
1 2 3 4 5 6 7 8 9 10 11
1 ROA 1
2 Operating expense ratio −0.3592 1
3 Portfolio-at-risk (30 days) −0.1899 −0.0826 1
4 Average loan/GNI per capita0.0457 −0.142 0.0528 1
5 Women targeting 0.0415 −0.0176 −0.071 −0.1289 1
6 Dummy rural market −0.1972 0.151 −0.0026 0.1353 −0.0473 1
7 Logarithm of assets 0.2187 −0.1801 −0.0584 0.1535 −0.0704 −0.1068 1
8 MFI age 0.0453 −0.1474 0.2522 0.0335 −0.1223 −0.1168 0.2354 1
9 Dummy international initiator−0.1122 0.1181 −0.1766 −0.0309 0.2152 0.0139 0.0119 −0.2158 1
10 Savings −0.0334 −0.124 0.1656 0.0462 −0.0619 0.1002 0.2171 0.2726 −0.2033 1
11 Dummy internal auditor 0.0605 0.0332 −0.0412 0.1247 −0.1898 0.0381 0.2663 0.1723 −0.0481 0.008 1
Empirical results
Table 3 shows the general model for international
regardless of the type of debt. (Tables 3–5 are com
of the threementionedregressionsexplainedin the
Appendix.) Column 1 tests the fi nancial and social
formance variables only, column 2 includes MFI co
variables, while column 3 adds the country HDI an
regional dummies.
Table 3. Pooled probit regressions for international de
[1] [2] [3]
ROA 0.211 0.550 0.790
Operating
expense ratio
0.0583 −0.0500 −0.0758
PAR30 −1.192**−0.854 −0.724
Average loan/
GNI per
capita
0.173* 0.167 0.0778
Women
targeting
0.230** 0.127 0.107
Dummy rural
market
0.347** 0.496*** 0.442**
Logarithm of
assets −0.0374 −0.00259
MFI age 0.0110 0.0103
Dummy
international
initiator
0.406*** 0.374***
Voluntary
savings
−0.453***−0.618***
Dummy
internal
auditor
0.260** 0.246**
HDI −0.0367
Region
dummies
No No Yes
Constant 0.274** 0.544 0.512
Pseudo-R2 0.0222 0.0643 0.0757
Observations 667 597 597
*** p < 0.01, ** p < 0.05, * p < 0.1.
Notes:
Region dummies are included for Latin America, Africa, East
Europe and Central Asia,
Middle East and North Africa, and Asia.
A robustness check (unreported) has been conducted by run
the same regressions using a logit model, yielding almost ex
the same results with similar pseudo-R2.

24 Roy Mersland and Ludovic Urgeghe
Copyright © 2013 John Wiley & Sons, Ltd. Strategic Change
DOI: 10.1002/jsc
Table 3 shows that four variables signifi cantly explain
an MFI’s access to international debt: the orientation of
MFIs toward rural areas, the presence of an international
initiator, the presence of an internal auditor reporting to
the Board and when the MFI doesn’t mobilize voluntary
savings. In addition, the coeffi cients of several perfor-
mance variables have signs as expected: MFIs accessing
international debt are those with higher return on assets,
lower portfolios-at-risk, and those that focus on targeting
women. Th e signifi cant fi ndings are interesting and of
policy interest. Rural markets are interesting for interna-
tional investors, but at the same time such investors prefer
MFIs that professionalize and follow “best practices” (in
this case by having an internal auditor reporting to the
Board). Th e fi ndings also show that MFIs with interna-
tional initiators have easier access to international funds.
Finally, MFIs that mobilize savings don’t fund themselves
internationally, probably because local deposits can be a
cheap source of funds without exposing the MFI to
foreign exchange risks. However, these general results do
not tell us much about the relationship between the type
of funding received and the performance of the MFI (H1a
and H1b) as the eff ects could be very diff erent from one
type of funding to another. We therefore disentangle the
international debt variable into two distinct variables:
international commercial debt only and international sub-
sidized debt only.5
Table 4 shows the regressions for international com-
mercial debt.
Beginning with the relationship between access to
commercial debt and fi nancial performance (H2), our
expectations are supported. Indeed, higher ROA, lower
operating expense ratio, and lower PAR30 signifi cantly
increase the likelihood for an MFI to have international
commercial debt. Th is fi nding is consistent with the
notion that commercial investors target more robust and
profi table MFIs (Goodman, 2004; Bogan, 2008). T
also confi rms the observation made by many that MI
target the “niche” of fi nancially profi table MFIs (
Schrevel et al., 2009; Wiesner and Quien, 2010). Rega
ing social performance, we fi nd a signifi cant negativ
tionship between the presence of commercial funding
the targeting of women by the MFI. Th us, comm
MIVs do not consider reaching women a priority.
positive coeffi cient reported in Table 3 is thus d
totally by subsidized international debt (see Table 5).
for rural outreach, the coeffi cient remains positiv
is only signifi cant in one of the regressions. Res
5 MFIs with both types of debt have been left out of the
sample for regressions in Tables 4 and 5, which explains the
diff erent N between Table 3 and Tables 4 and 5.
Table 4. Pooled probit regressions for international c
mercial debt
[1] [2] [3]
ROA 1.588** 1.212 1.690**
Operating
expense
ratio
−0.0151 −0.443* −0.976***
PAR30 −2.487***−2.119***−2.072**
Average loan/
GNI per
capita
0.0568 −0.0444 −0.141
Women
targeting −0.387***−0.308** −0.276*
Dummy rural
market
0.212 0.239 0.492**
Logarithm of
assets −0.114* −0.110*
MFI age −0.00303 −0.00863
Dummy
international
initiator
0.162 0.164
Voluntary
savings −0.665***−0.921***
Dummy
internal
auditor
0.614*** 0.621***
HDI 0.0192
Region
dummies
Yes
Constant −0.396*** 1.239 2.159**
Pseudo-R2 0.0444 0.111 0.174
Observations 528 475 475
*** p < 0.01, ** p < 0.05, * p < 0.1.

Debt Financing and Performance 25
Copyright © 2013 John Wiley & Sons, Ltd. Strategic Change
DOI: 10.1002/jsc
voluntary savings and internal auditor are upheld when
only commercial debt is considered (Tables 3 and 4 yield
similar signifi cant results). It should also be noted that
those MFIs accessing commercial debt are signifi cantly
smaller than other MFIs. Finally, we see that the dummy
for the international initiator is no longer signifi cant in
the subsample where only commercial international debt
is considered. Th us, the international initiator is fi rst
and foremost helping the MFI to access subsidized debt
(see Table 5) and not commercial debt.
Table5 showsthe regressionsfor international
subsidized debt.
Th e diff erences between Tables 4 and 5 are s
While commercial international debt goes to MFIs w
solid fi nancial performance (high ROA, low ope
expense ratio, and low portfolio-at-risk), subsidized
national debt goes to MFIs with weaker ROA,
costs, and higher portfolio-at-risk.6 Moreover, contrary to
commercial debt, subsidized debt is associated wit
targeting women. We also see that subsidized deb
older and internationally initiated MFIs that don’t h
internal auditors reporting to the Board. Not surpri
the fi nding that voluntary savings now has a positfi cient indicates that when inexpensive funding is
able also, MFIs that mobilize savings are intere
surprising result is the diff erence between Tables
when it comes to average loan. Subsidized debt is
cantly associated with higher average loan while thfi cient signs for commercial debt (Table 4) are
(in the models including controls). Th e most
reason for this is that lending to the poor can indee
good business for the MFI — low average loans
strong fi nancial performance can be combined (M
and Strøm, 2010) — and that MIVs providing subsi
debt are most concerned about supporting wea
especially when these reach out to women. Th
mean that the targeting of women, and not necess
targeting of the poor, is what attracts subsidies infi nance. Moreover, it could mean that the way sub
are distributed in the microfi nance industry sh
reconsidered.
At fi rst glimpse the results for the rural dummy
strange. While this variable shows strong signifi ca
in Table 3, only one of the regressions in Tables 4
gives a signifi cant association between access to
tional debt and outreach to rural markets. Howeve
tional analyses (unreported) show that the sign fi ndings reported in Table 3 to some extent are dr
Table 5. Pooled probit regressions for international subsi-
dized debt
[1] [2] [3]
ROA −0.908 −0.345 −0.493
Operating
expense ratio
0.111 0.243 0.440*
PAR30 0.495 0.234 0.202
Average loan/
GNI per
capita
0.0760 0.147* 0.201**
Women
targeting
0.544*** 0.336*** 0.310**
Dummy rural
market
0.231 0.257 0.0332
Logarithm of
assets −0.0287 −0.0193
MFI age 0.0184* 0.0198**
Dummy
international
initiator
0.285** 0.319**
Voluntary
savings
0.169 0.245
Dummy
internal
auditor
−0.399***−0.418***
HDI −0.656
Region
dummies
Yes
Constant −0.860*** −0.630 −0.807
Pseudo-R2 0.0467 0.0620 0.0920
Observations 528 475 475
*** p < 0.01, ** p < 0.05, * p < 0.1.
6 Th ough the coeffi cients for the fi nancial variables in
are not signifi cant, the diff erences between the results
Tables 4 and 5 allow our interpretation.

26 Roy Mersland and Ludovic Urgeghe
Copyright © 2013 John Wiley & Sons, Ltd. Strategic Change
DOI: 10.1002/jsc
those MFIs that have taken both commercial and subsi-
dized debt (these MFIs are, as mentioned, left out from
the analyses presented in Tables 4 and 5). Moreover, all
regressions in Tables 4 and 5 show positive coeffi cient
signs, indicating that international lenders do indeed care
for rural outreach, and probably the commercial lenders
prefer rural markets even more than subsidized lenders
(signifi cant result in the full model in Table 4).
In sum, this analysis suggests that even if the interna-
tional funding to MFIs comes from socially responsible
investors, we need to distinguish between commercial and
subsidized funding to understand MIV practices. Com-
mercial funding seems clearly to be driven by fi nancial
performance and the level of professionalization of MFIs,
while the special targeting of women is not a priority. Th is
seems to match the negative screening approach — micro-fi nance is considered a social investment per se so MIVs
off ering commercial debt can concentrate on analyzing the
level of professionalization and fi nancial performance of the
MFI. On the other hand, subsidized funding seems clearly
to target institutions focusing on women without prioritiz-
ing level of professionalization or fi nancial performance.
Th us, subsidized providers of debt seem to follow a positive
approach but mainly limited to the targeting of women.
Conclusion
Starting with the statement that international funders of
microfi nance claim to pursue both fi nancial and social
bottom lines through their investments, this article tests
what type of characteristics and performance in an MFI
actuallyattractsinternationalinvestments,segmented
into commercial and subsidized debt. Th e overall conclu-
sion is that commercial funding seems to match the nega-
tive screening approach as it is driven mainly by fi nancial
performance and the level of professionalization of the
MFIs, while subsidized funding is driven mainly by the
targeting of women and not by the level of professional-
ization or fi nancial performance of the MFI. Th us, subsi-
dized loan providers seem to follow a positive approach
in their investments.
By applying fi nancial criteria to select MFIs, co
mercial MIVs seem to consider those institutions per s
part of the social investment fi eld. From a pragmatic
of view this seems reasonable. After all, even if an MF
doesn’t specifi cally focus on women, normally half of
customers will in any case be women (D’Espallier et a
2011). As a result, the commercial MIVs can concentr
on identifying MFIs that can demonstrate a good leve
professionalization combined with sound fi nancial res
and effi cient operations.
Two important policy implications can be draw
from this article. First, MFIs should professionalize the
operations and assure good fi nancial performance in
to attractinternationalcommercialfunding.While
Mersland and Strøm (2009) indicate that having an in
nal auditor reporting to the Board is one of the few go
ernance mechanisms that can improve an MFI’s fi nan
performance, we now show that it is also associated w
better access to commercial funding. Moreover, w
Mersland and Strøm (2010) show that MFIs with
most effi cient operations are those with the best pot
to reach poor customers, we now fi nd that such MFIs
also those attracting commercial funding.
Second, MIVs providing subsidized funding need to
rethink their targeting strategy. Even though the
dized MFIs target women to a larger extent than
subsidized MFIs, it may easily lead to a dependency tr
clued by the fact that older MFIs still receive subsidies
found in the article. Moreover, it looks like the subsidi
funds go to MFIs with good international connecti
instead of MFIs with professional and effi cient operat
Our results should motivate researchers to study whe
MIVs providingsubsidizedfundingare hinderinga
needed professionalization of the industry, and wh
the targeting of women has become an excuse for ine
cient operations.
Th is article is only a fi rst step in understand
drivers of international microfi nance investments,
has some limitations which should motivate more res
First, rough dummies are used to distinguish between

Debt Financing and Performance 27
Copyright © 2013 John Wiley & Sons, Ltd. Strategic Change
DOI: 10.1002/jsc
with or without subsidized or commercial international
debt. More information on the relative importance of each
debt type, as well as more information about the individual
MIVs, could potentially improve considerably the analyses.
Th us, researchers could build a dataset where they combine
variables from MIVs and MFIs. Second, we should be cau-
tious in the way we measure social performance. Even
though the three variables applied in this study (average
loan size, targeting women, and rural outreach) are widely
used in academic and practitioner studies, they are still only
rough proxies of social performance. Social performance
has a more qualitative nature and embraces many other
aspects of the MFI’s activity, such as social responsibility
and the interactions with various stakeholders of the MFI.
Th us, how investors actually assess social performance in
MFIs remains to a large extent a “black box” for future
research to open. In addition, researchers should assess to
what extent international investors consider operational
effi ciency to be a social variable as this can potentially drive
down interest rates. Finally, the causality direction could be
reversed for variables such as, for example, the internal
auditor where an MIV can demand that MFIs hire an
internal auditor as a condition of their funding. Event
studies where ex-ante and ex-post performance is compared
in relation to the installation of new governance mecha-
nisms, like an internal auditor, could bring interesting new
knowledge.
Acknowledgments
Th e authors are very grateful to professors Marek Hudon,
Marc Labie, Ariane Szafarz, and Hugues Pirotte for their
comments on earlier versions of this article. We also thank
Bert D’Espallier for his methodological support and many
other colleagues for their useful comments each time we
have presented diff erent stages of the article at conferences
and seminars.
Appendix
Here are the three regressions corresponding to Tables 3,
4, and 5, respectively:
(1) Pr(International debt = 1) = Φ (β0 + β1
ROA + β2 Opexp + β3 Par30 + β4 Avloan + β5
dmWomen + β6 dmrural + β7 Size + β8
Age + β9 dmIntInit + β10 dmSavings + β11
dmaudit + β12 HDI + β13 dmLatAm + β14
dmMena + β15 dmEECA + β16 dmASIA)
where Φ is the cumulative normal distribution.
(2)Pr(International commercial debt = 1) = Φ
(β0 + same variables)
(3)Pr(International subsidized debt = 1) = Φ
(β0 + same variables)
References
Armendariz de Aghion B, Morduch J. 2010. Th e Econom
Microfi nance, 2nd edn. MIT Press: Cambridge, MA.
Bogan V. 2008. Microfi nance institutions: Does capital s
ture matter? Working paper of the Department of
Economics and Management, Cornell University. Availa
http://ssrn.com/abstract = 1144762.
Bollen N. 2007. Mutual fund attributes and investor beh
Journal of Financial and Quantitative Analysis 42: 683–
Carroll AB. 1979. A three-dimensional conceptual m
corporate performance. Academy of Management Rev
497−505.
Channon D. 2006. Life-cycle strategy. In Th e Blackwell
clopedia of Management, Vol. 12, Strategic Manageme
edn, McGee J (ed.). Blackwell: Oxford; pp. 195–199.
Cull R, Demirguz-Kunt A, Morduch J. 2007. Financial per
mance and outreach: A global analysis of leading micro
Economic Journal, Royal Economic Society 117: 107–1
D’Espallier B, Guérin I, Mersland R. 2011. Women and re
ment in microfi nance: A global analysis. World Develo
39(5): 758–772.
De Bruyne B. 2008. Summary of Social Performance Ind
Survey. European Dialogue No. 1, June 2008, edite
European Microfi nance Platform.
De Schrevel JP, Labie M, Urgeghe L. 2009. Blue Orc
Connectingmicrofi nanceto capital markets — Sequel.

28 Roy Mersland and Ludovic Urgeghe
Copyright © 2013 John Wiley & Sons, Ltd. Strategic Change
DOI: 10.1002/jsc
KennedySchoolof GovernmentCase,HarvardUniversity,
C 14–04–1762.1D.
De Sousa-Shields M, Frankiewicz C. 2004. Financing Microfi -
nance Institutions: Th e Context for Transitions to Private Capital.
Micro Report No. 32, Accelerated Microenterprise Advance-
ment Project, USAID.
Goodman P. 2004. Microfi nance Investment Funds: Objectives,
Players, Potential. 2004 KfW Financial Sector Development
Symposium, Berlin.
Iglewicz B, Hoaglin DC. 1993. How to detect and handle
outliers. ASQC Basic References in Quality Control, Vol. 16,
American Society for Quality Control, ASQ Quality Press
Milwaukee, WI.
JanssonT. 2003. FinancingMicrofi nance.Inter-American
Development Bank Sustainable Development Department
Technical Paper Series, MSM-118, Washington DC.
Juravle C, Lewis A. 2008. Identifying impediments to SRI in
Europe: A review of the practitioner and academic literature.
Business Ethics 17(3): 285–310.
Kennedy P. 2008. A Guide to Econometrics, 6th edn. Blackwell
Publishing: Malden, MA.
Kooi P. 2001.Raisingcapitalthroughequityinvestments
in MFIs: Lessons from ACLEDA, Cambodia. Input paper
for the UNCDF/SUM and UNDP Africa Global Meeting,
May 30–June 1.
Lensink R, Hermes N, Meesters A. 2011. Outreach and effi -
ciency of microfi nance institutions. World Development 39(6):
938–948.
Little AD. 1974. A System for Managing Diversity. Arthur D.
Little, Inc.: Cambridge, MA.
Mersland R, Strom RØ. 2009. Performance and governance in
microfi nance institutions. Journal of Banking & Finance 33(4):
662–669.
Mersland R, Strøm RØ. 2010. Microfi nance mission drift?
World Development 38(1): 28–36.
Mersland R, Randøy T, Strøm RØ. 2011. Th e impact of inte
national infl uence on micr obanks’ performance: A glo
survey. International Business Review 20(2): 163–176.
MicroRate. 2011. State of Microfi nance Investment 201
Microrate’s 6th Annual Survey and Analysis of MIVs. Micro
Luminis (www.MicroRate.com).
Oehri O, Fausch J. 2008. Microfi nance investment fund
Analysisof portfolioimpact.Universityof Liechtenstein,
Gevena papers on inclusiveness, World Microfi nance Foru
Geneva.
Porter ME, Kramer MR. 2002. Th e competitive advantage o
corporatephilanthropy.HarvardBusinessReview80(12):
56−68.
Reille X, Forster S, Rosas D. 2011. Foreign capital investme
microfi nance: Reassessing fi nancial and social returns
Focus Note No. 71, Washington DC.
Renneboog L, Ter Horst J, Zhang C. 2008. Socially responsib
investments: Institutional aspects, performance and inves
behaviour. Journal of Banking and Finance 32: 1723–1742
Schreiner M. 2002. Aspects of outreach: A framework
discussion of the social benefi ts of microfi nance. Jou
International Development 14: 591–603.
SEEP Network. 2005. Measuring Performance of Microfi
Institutions — A Framework for Reporting, Analysis and Mo
toring. SEEP Network: Washington DC.
Stock J, Watson MW. 2006. Introduction to Econometrics, 2n
edn. Addison-Wesley: New York.
Swanson B. 2008. Th e role of international capital market
microfi nance. In Microfi nance: Emerging Trends and
lenges, Sundaresan S (ed.). Edward Elgar: Cheltenham.
Van Maanen G. 2005. L’avenir du fi nancement du microcre
Techniques Financières et Développement, No. 78.
Wiesner S, Quien D. 2010. Can “bad” microfi nance practice
be the consequence of too much funding chasing too
microfi nance institutions? ADA discussion paper 2, Appui
Développement Autonome (www.lamicrofi nance.lu).

Debt Financing and Performance 29
Copyright © 2013 John Wiley & Sons, Ltd. Strategic Change
DOI: 10.1002/jsc
BIOGRAPHICAL NOTES
Roy Mersland is an associate professor at the
University of Agder in Norway. He has extensive
international management, consulting, and research
experience. He has published extensively in journals
such as World Development, Journal of Management
Studies, Journals of Development Studies, and Journal
of Banking and Finance. He is the director of the
Norwegian Centre for Microfi nance Research and is
head of the PhD program in International
Management at the University of Agder.
Ludovic Urgeghe is a permanent researcher at th
Center for European Research in Microfi nance and
currently a PhD candidate at the Warocqué School
of Business and Economics (University of Mons,
Belgium) where he is also a teaching assistant in
Management. His PhD research, situated in the
context of microfi nance commercialization, aims a
exploring the role of socially responsible investors
the microfi nance sector.
Corresponding author:
Ludovic Urgeghe
Center for European Research in Microfi nance
Warocqué School of Business and Economics
University of Mons
9 rue de Houdain, 7000 Mons
Hainaut, Belgium
e-mail: [e-mail]
View publication statsView publication stats