To Profit or not to Profit? Assessing financial sustainability outcomes of Microfinance Institutions

Abstract

We used a multilevel approach on comparing for-profit and not-for-profit microfinance institutions (MFIs). The database was composed of 202 MFIs (in 52 countries) from the most widely used microfinance dataset (the Microfinance Information eXchange, Inc. Market), with 669 observations from 2010 to 2014. Four financial sustainability outcomes were considered: yield on gross portfolio; return on assets; portfolio at risk, 30 days; and operational self-sufficiency (OSS). Although for-profit MFIs had a higher Yield, there was no significant effect of profit orientation on ROA, PAR30, and OSS. Further analysis shows that although profit orientation has a significant effect on the yield of small MFIs, it does not have any effect on larger MFIs, which is consistent with the theory that larger MFIs can distribute its fixed costs better, requiring lower interest rates and allowing smaller yield on the gross portfolio. In addition, we show that the intrinsic characteristics of the MFIs account for the majority of the variance from the four outcomes.

Full text

Received: 18 August 2017 Revised: 22 December 2018 Accepted: 30 December 2018
DOI: 10.1002/ijfe.1718
RESEARCH ARTICLE
To profit or not to profit? Assessing financial sustainability
outcomes of microfinance institutions
Rodrigo de Oliveira Leite1Layla dos Santos Mendes2Luiz Claudio Sacramento3
1School of Administration and Finance,
Universidade do Estado do Rio de Janeiro,
Rio de Janeiro, Brazil
2Escola Brasileira de Administracao
Publica e de Empresas, Fundacao Getulio
Vargas, Rio de Janeiro, Brazil
3Economics Department, Pontifical
Catholic University of Rio de Janeiro, Rio
de Janeiro, Brazil
Correspondence
Rodrigo de Oliveira Leite, School of
Administration and Finance,
Universidade do Estado do Rio de Janeiro,
Rua Sao Francisco Xavier 524 Sala 8024-B
20550-013, Rio de Janeiro, Brazil.
Email:
rodrigo.de.oliveira.leite@gmail.com
JEL Classification: G21; D21; C33
Abstract
We used a multilevel approach on comparing for-profit and not-for-profit micro-
finance institutions (MFIs). The database was composed of 202 MFIs (in 52
countries) from the most widely used microfinance dataset (the Microfnance
Information eXchange, Inc. Market), with 669 observations from 2010 to 2014.
Four financial sustainability outcomes were considered: yield on gross portfo-
lio; return on assets; portfolio at risk, 30 days; and operational self-sufficiency
(OSS). Although for-profit MFIs had a higher Yield, there was no significant
effect of profit orientation on ROA, PAR30, and OSS. Further analysis shows that
although profit orientation has a significant effect on the yield of small MFIs,
it does not have any effect on larger MFIs, which is consistent with the theory
that larger MFIs can distribute its fixed costs better, requiring lower interest rates
and allowing smaller yield on the gross portfolio. In addition, we show that the
intrinsic characteristics of the MFIs account for the majority of the variance from
the four outcomes.
KEYWORDS
financial efficiency, microfinance, mission drift, multilevel model, profitability
1INTRODUCTION
In the last decades, microfinance has been achieving
increasing prominence in policymaking for the poor
entrepreneur (Morduch, 1999). Muhammad Yunus, the
microfinance pioneer, founded the Grameen Bank in 1976
and, for that, received the Nobel Peace Prize thirty years
later. However, there is not much agreement on how
microfinance institutions (henceforth MFIs) should be
structured in order to achieve the best results in alleviating
poverty. Nevertheless, little attention is given to sustain-
ability and mission drift of MFIs nor to which approach is
the best to achieve sustainability without drifting from the
main mission: help the poor.
This study investigates the following question: Are
not-for-profit MFIs as sustainable as for-profit ones? If
not, what causes this difference? Literature advocates
both approaches. The former is considered a traditional
approach while the latter is deemed to be modern (Hudon
& Traca, 2011). Morduch (2000) expressed those differ-
ent points of view. Those that advocate the profit-oriented
approach of microfinance claim that with profits an MFI
can achieve self-sustainability and expand its loan portfo-
lio to help more people. They also claim that not-for-profit
MFIs are less efficient and cannot survive without subsi-
dies. Nevertheless, Hudon and Traca (2011) provide empir-
ical evidence that up to a certain threshold, subsidized
MFIs reach higher levels of efficiency.
This paper investigates whether MFI profit orientation
determines some difference regarding sustainability and
revenues. The authors aforementioned focus on the effi-
ciency of these institutions, whereas we focus on sustain-
ability, which still is not fully explored in the literature.
This study has several political implications. Although
Int J Fin Econ. 2019;1–13. wileyonlinelibrary.com/journal/ijfe © 2019 John Wiley & Sons, Ltd. 1

2LEITE ET AL.
microfinance is usually considered a self-sustained devel-
opment policy, very few MFIs are not dependent on
donations and subsidies (Hudon & Traca, 2011). When
policymakers have to decide where to allocate their
resources, they must consider profit orientation, among
other MFIs characteristics.
This study has found that small MFIs do differ concern-
ing how much interest clients are charged. Interestingly,
larger MFIs charge similar interest rates, regardless of
profit status, which is consistent with the idea of cost effi-
ciency. We thoroughly investigated the cost structure of
those institutions and found that smaller for-profit MFIs
present higher costs compared with large for-profit MFIs.
According to Caudill, Gropper, and Hartarska (2009), big-
ger MFIs do not need to charge high interest rates to be
sustainable as smaller MFIs do. In fact, we show that larger
for-profit MFIs can charge lower interest rates. Results also
show that a large part of variance comes from the MFI
level.
Our study contributes mainly to two fields of research:
(a) profit versus not-profit orientation of MFIs, which
includes the discussion of sustainability and mission drift
of these institutions; and (b) efficiency of MFI institutions
in terms of cost, outreach, impact, and capital structure,
which we can compare globally. In the next pages, we
present several robustness tests about our statement that
there is no best orientation about MFIs in terms of finan-
cial outcome.
2LITERATURE REVIEW
Two main views on both welfare and institutionalist
approaches are brought foward by Olivares-Polanco (2005)
and Cull, Demirguç-Kunt, and Morduch (2007). In sim-
ple terms, the former is more focused on achieving
their mission, giving more weight to the outreach, with
non-governmental organizations (NGOs) serving as the
main channel. The latter has a preference for the MFI's sus-
tainability, with commercial banks as the prime source of
credit and with an integrated range of products available
to customers.
Some behavioural aspects are noted by Norell (2001),
who mentions that a not-for-profit MFI can signal to their
clients that the key concern is their well-being and not
the financial return, which can lead to the opportunistic
behaviour of strategic default. Additionally, MFI managers
can behave opportunistically, feeding a long-term depen-
dence of donations (Hudon & Traca, 2011). However, those
who advocate for the not-for-profit approach express their
concerns that for-profit MFIs may exploit the poor with
higher interest rates and become the “new moneylenders”.
They also claim that as long as poverty exists, there will
be subsidies for MFIs, because reducing poverty is a key
element in the policymaking of all governments.
There are studies focusing on a term called “mission
drift” (Armendáriz & Szafarz, 2011; Cull et al., 2007;
Olivares-Polanco, 2005). This happens when an MFI starts
to move away from its mission, that is, to reduce poverty,
towards a more commercial perspective. Examples of MFIs
that were founded as not-for-profit and then became
for-profit enterprises are exposed by Schmidt (2013) who
cites two very well-known cases: Compartamos in Mexico
and SKS in India.
Compartamos was founded in 1990 as an NGO, but
in 2007, it released its Initial Public Offer (IPO). How-
ever, they did not create new shares in the IPO process;
therefore, they did not receive new funds. Several authors
(Ashta & Hudon, 2012; Rosenberg, Gonzalez, & Narain,
2009; Schmidt, 2013) have criticized the IPO and the inter-
est rates charged by Compartamos. The incentives align-
ment is controversial because the goal of publicly listed
companies is to maximize the shareholders' return, but
this can cause an ethical conflict because their clients are
mostly poor people from developing countries.
There is evidence for credit elasticity for poor clients
(Karlan & Zinman, 2008), and this may be the reason
why loans made by Compartamos, even with high inter-
est rates, do achieve some positive outcomes (Angelucci,
Karlan, & Zinman, 2015). It was found that those
loans had a modest positive impact on business size,
trust, and female decision-making and a modest neg-
ative impact on depression and reliance on or need
for aid.
An interesting case is SKS MFI, which was founded in
India in 1998 and released its IPO in 2010. SKS adopted,
as Compartamos did, the policy of high interest rates. This
policy is credited for the crash of microfinance in India
in 2010 (Wichterich, 2012). During the crisis, prices of
SKS shares dropped 77%. The crisis originated in Andhra
Pradesh and is even suspected of having caused a num-
ber of suicides (CGAP, 2010), which also resulted in the
“indebtedness of 82 percent of rural household” (Wich-
terich, 2012), and over 35,000 people losing their jobs in
Andhra Pradesh.
Therefore, there is some evidence of the advantages and
disadvantages of for-profit high-interest MFIs. Neverthe-
less, in their study, Mersland and Strøm (2012) show that
the hypothesis of mission drift claim cannot be accepted,
because their results indicate that the microfinance indus-
try, criticized by high profits, is actually an industry strug-
gling with high costs and low earnings.
Zeller and Meyer (2002) and Olivares-Polanco (2005)
argue for the “triangle of microfinance”, that is, financial
sustainability, outreach, and impact. Those three objec-
tives could only be achieved if the MFI had profits in order

LEITE ET AL. 3
to be self-sustainable and grow (reaching more people and
giving larger loans).
However, profit-oriented MFIs tend to have different
goals than not-for-profit ones. Cull et al. (2007) show
that profit-oriented MFIs tend to lend to the “richest of
the poor” in order to achieve a higher level of profitabil-
ity. Also, Baquero, Hamadi, and Heinen (2018) explain
that competition has different effects on for-profit and
not-for-profit microbanks in outcomes similar to this
study.
Regarding the efficiency of MFIs, Paxton (2007) com-
pared semiformal institutions and found that young
institutions can benefit more from technological invest-
ments and by diversifying between rural and urban loans.
Another example is Caudill et al. (2009) who argue that
MFIs become more efficient over time and less dependent
on subsidies. However, Cull, Demirgüç-Kunt, and Mor-
duch (2018) found a puzzling result: The more subsidized
MFIs were not the NGOs but the more commercialized
institutions.
Similarly, Caudill et al. (2009) study presents evidence
in which MFIs that receive fewer subsidies become more
effective over time in terms of total cost. They used a mix-
ture model to identify subpopulations and classify them as
“more cost-effective” and “not more cost effective”. Thus,
maybe, profit-oriented MFIs and, at the same time, those
that receive less or do not receive donations and subsidies
can be more cost-effective. Therefore, this could be a mech-
anism through which MFIs can reach a higher number of
clients. Hence, MFI efficiency is fundamental to reduce
those costs.
In summary, we have two objectives: first, to use a robust
model that allows us to better deal with the heterogeneity
of data in order to have a cleaner effect of profit orientation
on four well-known financial metrics for MFIs and, sec-
ond, to verify the intrinsic characteristics over Time, MFI,
and Region. Additionally, we also explore the cost effi-
ciency, outreach, and capital structure of for-profit versus
not-for-profit MFIs.
3METHODOLOGY
3.1 Sample
Using a database of 202 audited MFIs (52 countries), with
669 observations, from 2010 to 2014 in a multilevel model,
we expect to contribute to the discussion of those two
different approaches (not for profit vs. for profit), and
their impact on several performance outcomes. By using
this time frame, we use data from audited MFIs only,
increasing our confidence in the reported values after the
financial crises (and its externalities), which affected sev-
eral banks around the globe. For instance, during the
crisis, the volume of donations could have been reduced
for the not-for-profit MFIs, which could lead to biased
results.
We used a database provided by the Microfinance Infor-
mation eXchange, Inc. (MIX Market). Data is self-reported,
which can create a problem if an MFI reports untruthful
figures. Notwithstanding, all the dataset was audited in a
number of MFIs during 2013 and 2014. This assures that
the data used is truthful and representative of the MFI's
financial position at that time. This data is widely used
in other studies, for instance, in 1997 to 2007 in Hermes,
Lensink, and Meesters (2011), but given that this audit is
recent, we can avoid the “unreliability issue”.
In the original sample, there are 891 MFIs from 97 coun-
tries with nominal values adjusted by dollar parity for all
years. We excluded observations that were not audited
and observations that contained missing data according
to key variables (size, age, percentage of female borrow-
ers, borrower retention rate, and staff turnover rate). In
addition, we winsorize all variables in the model at the
first and 99th percentiles. That is, we replace any observa-
tion below the first percentile with the first percentile and
any observation above the 99th percentile with the 99th
percentile. Table 1 shows the sample distribution by each
region/country. In addition Table 2 shows summary statis-
tics for the main variables, with the correlations shown in
Table 3.
3.2 Measures
Armendáriz and Morduch (2010) list five frequently used
financial ratios that are important for an MFI to be
sustainable: operational self-sufficiency (OSS), financial
self-sufficiency (FSS), return on assets (ROA), portfolio
at risk after 30 days (PAR30), and yield on gross loan
portfolio (Yield). Nevertheless, Cull, Demirgüç-Kunt, and
Morduch (2011) show that the FSS ratio may be biased,
because grants, donations, and other alternative ways of
funding are not included in the FSS calculation, making
not-for-profit institutions score lower in this financial ratio
even though funding for those institutions has been sta-
ble for many years (Morduch, 2000). For this reason, we
decided not to use this variable. Notwithstanding, we used
an alternative measure for the FSS, which follows the work
of Cull, Demirgüç-Kunt, and Morduch (2009), which is
presented in Section 5.6.
The first measure is yield on gross loan portfolio, which
is fundamental for this study. This variable is measured
using the interest and fees on loan portfolio divided by the
average gross loan portfolio for that period. In addition,
the variable is in real economic terms, and it considers

4LEITE ET AL.
TABLE 1 Sample distribution by each region/country
Region Country Freq. MFI Freq. obs. Percent Cum.
Africa Burkina Faso 1 2 0.5 0.5
Cameroon 1 1 0.5 1.0
Congo 1 1 0.5 1.5
Ghana 1 1 0.5 2.0
Kenya 1 5 0.5 2.5
Malawi 2 5 0.99 3.5
Nigeria 2 6 0.99 4.5
Tog o 1 2 0.5 5. 0
Uganda 2 3 0.99 6.0
East Asia and the Pacific Cambodia 7 21 3.47 9.4
China 1 3 0.5 9.9
East Timor 2 7 0.99 10.9
Indonesia 1 4 0.5 11.4
Philippines 6 18 2.97 14.4
Vietnam 3 7 1.49 15.9
Eastern Europe and Central Asia Albania 1 3 0.5 16.4
Armenia 4 17 1.98 18.4
Azerbaijan 9 26 4.46 22.8
Bosnia and Herzegovina 7 27 3.47 26.3
Georgia 2 10 0.99 27.3
Kazakhstan 4 18 1.98 29.3
Kosovo 2 9 0.99 30.3
Kyrgyzstan 2 9 0.99 31.3
Macedonia 1 4 0.5 31.8
Mongolia 2 9 0.99 32.7
Serbia 1 2 0.5 33.2
Tajikistan 2 10 0.99 34.2
Turkey 1 2 0.5 34.7
Latin America and The Caribbean Argentina 1 4 0.5 35.2
Bolivia 10 38 4.95 40.2
Brazil 4 8 1.98 42.2
Chile 2 7 0.99 43.2
Colombia 7 24 3.47 46.6
Costa Rica 2 7 0.99 47.6
Dominican Republic 2 9 0.99 48.6
Ecuador 21 93 10.4 59.0
El Salvador 2 6 0.99 60.0
Guatemala 3 10 1.49 61.5
Mexico 13 36 6.44 67.9
Nicaragua 3 12 1.49 69.4
Paraguay 2 6 0.99 70.4
Peru 12 28 5.94 76.3
Middle East and North Africa Egypt 1 3 0.5 76.8
Jordan 1 3 0.5 77.3
Lebanon 1 4 0.5 77.8
Palestine 2 5 0.99 78.8
Tunisia 1 5 0.5 79.3
South Asia Bangladesh 2 8 0.99 80.3
India 19 69 9.41 89.7
Nepal 1 2 0.5 90.2
Pakistan 19 48 9.41 99.6
Sri Lanka 1 2 0.5 100.0
Total 202 669 100

LEITE ET AL. 5
TABLE 2 Summary statistics
Variables Mean Std. dev. Min. Max. N
Profit =1
Yield 25.96 19.02 −8.66 95.40 271
ROA 2.87 5.37 −34.17 28.52 272
PAR30 3.44 4.44 0 33.66 264
OSS 119.54 25.20 48.99 246.79 272
Small 0.17 0.38 0 1 272
Medium 0.25 0.43 0 1 272
Large 0.58 0.49 0 1 272
New 0.07 0.25 0 1 272
Young 0.19 0.39 0 1 272
Mature 0.75 0.44 0 1 272
Percent of female borrowers 0.66 0.27 0.03 1 272
Borrower retention rate 0.79 0.17 0.28 2.16 189
Staff turn over rate 0.24 0.20 0 1.07 272
Profit =0
Yield 22.07 12.68 −11.59 71.60 389
ROA 2.39 6.15 −60.33 17.06 389
PAR30 3.46 3.80 0 33.31 379
OSS 115.62 23.57 16.79 244.00 389
Small 0.38 0.49 0 1 389
Medium 0.27 0.44 0 1 389
Large 0.36 0.48 0 1 389
New 0.02 0.14 0 1 389
Young 0.07 0.25 0 1 389
Mature 0.91 0.29 0 1 389
Percent of female borrowers 0.68 0.24 0.21 1 389
Borrower retention rate 0.78 0.13 0.26 1.11 255
Staff turn over rate 0.19 0.24 0 2.97 389
TABLE 3 Correlation Table
Variables 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13.
1. Yield
2. Return on assets 0.143
3. Portfolio at risk (30 days) 0.059 −0.144
4. OSS 0.046 0.808 −0.215
5. Profit MFI (yes =1) 0.122 0.041 −0.001 0.079
6. Small −0.054 −0.108 0.145 −0.134 −0.224
7. Medium size 0.084 −0.039 0.078 −0.010 −0.017 −0.378
8. Large size −0.024 0.132 −0.201 0.131 0.219 −0.580 -0.535
9. New −0.051 −0.123 −0.077 −0.068 0.115 0.008 0.005 −0.011
10. Young MFI 0.076 −0.028 −0.103 0.003 0.180 −0.058 0.009 0.046 −0.074
11. Mature MFI −0.040 0.090 0.132 0.033 −0.221 0.048 −0.010 −0.034 −0.468 −0.846
12. % of female borrowers 0.248 0.030 −0.224 −0.004 −0.046 −0.195 0.007 0.171 0.031 0.203 −0.196
13. Borrower retention rate −0.107 0.003 −0.082 0.050 0.040 0.017 −0.077 0.052 0.153 0.051 −0.128 0.008
14. Staff turnover rate 0.160 0.040 0.033 −0.019 0.107 −0.042 0.022 0.018 −0.032 0.148 −0.114 0.142 −0.063
inflation. It measures how much the MFI is actually receiv-
ing in interest and other payments from its clients. This is
another way to measure how much the MFI is charging
their clients.
The second measure, return on assets, is represented by
the net income (after taxes, grants, and donations) divided
by the average assets . It measures how well the institution
uses its assets to generate net income.

6LEITE ET AL.
TABLE 4 Definition of variables
Main variables
Operational self-sufficiency Financial revenue / (Financial expense on funding liabilities
+ Net Impairment Loss on gross loan portfolio + Operating expense)
Return on asset (Net operating income, less taxes) / average assets
Portfolio at risk after 30 days (Outstanding balance, portfolio overdue >30 days +
Renegotiated loans) / Gross loan portfolio
Yield on gross loan portfolio Financial revenue from loans / Average gross loan portfolio
Profit 1 if for-profit, 0 if not-for-profit
Size Large if number of borrowers >30,000; medium if
number of borrowers is between 10,000 and 30,000; and small if
number of borrowers <10,000
Age New (1–4 years); young (4 to 8 years); and mature (8+ years)
Percentage of female borrowers Number of active female borrowers/ Number of active borrowers
Borrower retention rate Active borrowers at the end of the period / (Active borrowers
at the beginning of the period + New borrowers during
the period)
Other variables (used at robustness section)
Cost per borrower Operating expense / Average number of active borrowers
Average loan size / GNI Average loan balance per borrower / GNI per capita
Return on equity (Net operating income, less taxes) / Average equity
Prime interest rate Lending interest rate of the country + 2 percentage points
The third common measure applied here is the portfolio
at risk for 30 days, when the client is considered entering
arrears. This measure captures the part of the loan port-
folio affected by delinquency as a percentage of the total
portfolio. Although in traditional lending the delinquency
is usually measured after a client is 6 months into arrears,
in microfinance, due to the fact that there is no collateral
in most loans, a more conservative estimation is used. In
addition, Norell (2001) cites that the industry's standard is
5% of PAR30.
Finally, OSS presents the received revenue over the cost
of raising capital. This measure uses sources of costs as the
financial expense, provision for loan-loss and operational
expense in the denominator. Hence, it shows whether
the MFI can afford its costs and perpetuate its opera-
tions. A value of 100% (considering that we multiplied
our variables by 100) means that the MFI can afford their
costs. Values higher than 100% are desirable because they
can offer a financial slack for the MFI to undertake new
projects. Nevertheless, a very high value may be an indica-
tor that the MFI is exploiting their clients.
Our interest independent variable is profit, which
received the value of 1 if the MFI was registered as profit
oriented, and 0 otherwise (not for profit). This variable will
be interpreted as the difference in the dependent variable
for the MFIs that are seeking to profit and those that do not.
We used a number of control variables: size (small with less
than 10,000 borrowers, medium with 10,000–30,000 bor-
rowers, and large with more than 30,000 borrowers), age
(new with 1–4 years, young with 4–8 years, and mature for
more than 8 years), clients (percentage of female borrow-
ers and borrower retention rate) and staff turnover rate. We
used the categories small and new as a baseline. Table 4
presents all the variables used in this study, and how each
one is calculated.
In order to represent the results in a more meaning-
ful way, we chose to multiply our continuous variables
by 100 to make our interpretation more straightforward.
With these four financial ratios, we were able to compare
the performance of profit and non-profit MFIs control-
ling for the variance between the levels: time, MFI, and
region. These levels are particularly important due to the
panel structure (variation over time) and MFI idiosyn-
crasies, as well as region political and macroeconomic
characteristics.
3.3 Method
We present three reasons for using a hierarchical equation
model (multilevel analysis) in our study. First, in order
to observe a net effect of having a profit orientation on
financial outcomes, we diminish endogeneity by separat-
ing the error term in parts. Second, the dataset comprises
several MFIs in many regions with their own idiosyn-
crasies, and by allowing multiple levels of analysis, we
can observe which one is more important to explain the
variance of the dependent variable. Third, Breusch–Pagan
tests indicated that a panel structure would be preferable
to pooled ordinary least square regressions by rejecting the

LEITE ET AL. 7
null hypothesis that the time-invariant component of the
error has zero variance.
We used a multilevel approach with a three-level hier-
archical model, Equation 1 shows the specification of the
first level. The dependent variable (Yijk)of the year i,MFI
j,andRegionkis a function of the mean of the depen-
dent variable of MFI jand Region K(𝛽0jk)plus a random
error (𝜖ijk)representing the variance across time, normally
distributed with mean zero and variance of 𝜎2
e(𝜖i𝑗k∼
N(0,𝜎
2
e)).
Level 1 (Time) ∶Yi𝑗k=𝛽0𝑗k+𝜖i𝑗k.(1)
Equation 2 shows the second level of analysis, where the
mean of dependent variables across time of MFI jof the
Region k(𝛽0jk )is a function of a mean dependent variables
of Region k(𝛾00k)plus a random error (r0𝑗k∼N(0,𝜎
2
r))
representing the variance between MFIs.
Level 2 (MFI) ∶𝛽0𝑗k=𝛾00k+r0𝑗k.(2)
Finally, Equation 3 formalizes the analysis of the third
level, where the mean dependent variables of the MFI jin
Region k(𝛾00k)is then a random variable that is a function
of the grand mean of the sample (𝛿000)plus the random
error of the third level (u00k∼N(0,𝜎
2
u)).
Level 3 (Region) ∶𝛾00k=𝛿000 +u00k.(3)
Equations 4 depict the multilevel model with our IV (Pij)
and the random intercepts at the MFI and at the region
levels.
Level 1 (Time) ∶Yi𝑗k=𝛽0𝑗k+𝛽1i𝑗Pi𝑗+𝜖i𝑗k
Level 2 (MFI) ∶𝛽0𝑗k=𝛾00k+r0𝑗k
Level 3 (Region) ∶𝛾00k=𝛿000 +u00k.
(4)
Equation 5 depicts the multilevel model with our inde-
pendent variable “Profit” (P0ij), the controls (Cnij),and the
random intercepts at the MFI and the region levels.
Full model: Yi𝑗k=𝛾000 +
7
∑
n=1
𝛽ni𝑗kCni𝑗k+𝛽80𝑗kP0𝑗k
+𝜖i𝑗k+r0𝑗k+u00k.
(5)
Because profit is time invariant, a “normal approach”
with a panel with time fixed effects is not possible because
it would cause perfect collinearity. Thus, using a hierar-
chical linear model, we can control the time level and also
estimate the effect of the profit dummy.
Another reason we chose to use the hierarchical model
is that it allows us to verify the heterogeneity of the firms
across several levels. The research question requires us to
use controls in the estimation because we know that the
hierarchical structure cannot sustain the assumption of
independence (Raudenbush & Bryk, 2002). For instance,
because MFIs are in the same region, they are suscepti-
ble to similar elements, as we will see in the size of ICCs
(intraclass correlations) for the region.
Moreover, this technique allows us to identify how much
variance can be explained by each one of the three hierar-
chical levels, as we will see in the next section. Also, since
all the variables are endogenous, the multilevel model with
random intercepts allows each MFI and each region to
have its own intercept that controls endogeneity at the MFI
and region levels, as Hanchane and Mostafa (2012) noted.
4RESULTS
The results can be observed in Table 5. The null model
describes the variance of the dependent variables discrim-
inated by three levels (region, MFI, and time). Those mod-
els are presented in columns 1, 4, 7, and 10. In Model 1, the
time level explains 7.39% of the variance of Yield, 78.87%
is explained by MFI level, and 13.74% by region level.
In Model 4, the result with ROA as dependent variable,
time explains 20.19% of the variance, 79.81% is explained
by MFI level, and the region level does not explain any
meaningful amount of the overall variance. The values for
the MFI level are quite similar to Model 1, which indi-
cates that ROA and Yield variables variances are driven
mainly from factors at the MFI level. For the PAR30 vari-
able, time becomes more relevant because it accounts for
32.20% of the variance. The MFI level loses weight going
to 63.53%, and region explains 4.27%. Finally, the OSS vari-
ance is 74.09% in the MFI level. Time explains 25.39% of
the variance, and region explains only 0.52%.
The more distinctive result is the higher variance expla-
nation for Yield (13.74%) at region level. Moreover, the
MFI level at PAR30 has the lowest variance when com-
pared with the other three indexes. Time has the highest
variance explanation for PAR30 and lowest for Yield. The
MFI level explains the most of variance for all the depen-
dent variables, which implies that a hierarchical model
can be useful to see the variability of each MFI. This is
consistent with our goal in order to observe the difference
between profit and non-profit driven MFIs. The region
level together with the lower variance allows us to con-
trol the heterogeneity in the data that can lead to biased
estimation results.
The results indicate that in all models the profit dummy
variable was not significant, except in Models 2 and 3.
Interesting to note is that in Model 2 the coefficient is
5.794 and significant at 1% level, though after including
the controls, it remained similar (3.791) and significant at
10% (Model 3). The economic effect is noteworthy, in our
sample, for-profit MFIs charge their clients around 3.8%

8LEITE ET AL.
TABLE 5 Multilevel estimation results
Yield on gross portfolio Return on assets Portfolio at risk—30 days Operational self-sufficiency
M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12
Profit (1 =yes) 5.79*** 3.79* 0.35 0.22 0.45 −0.07 3.65 3.41
(2.17) (2.13) (0.79) (0.91) (0.48) (0.52) (3.39) (3.77)
Constant 23.57*** 21.09*** 9.23 2.29*** 2.15*** 1.76*** 3.34*** 3.15*** 7.59*** 116.59*** 115.04*** 103.18***
(3.31) (3.65) (6.31) (0.39) (0.51) (2.17) (0.47) (0.52) (1.52) (1.93) (2.31) (10.37)
Variance components
Region level (%) 13.74 16.06 31.25 <0.01 <0.01 <0.01 4.27 4.49% <0.01 0.52 0.34 <0.01
MFI level (%) 78.87 76.48 60.85 79.81 79.79 81.22 63.53 63.30% 65.53 74.09 74.15 70.29
Time level (%) 7.39 7.46 7.90 20.19 20.21 18.78 32.20 32.21% 34.47 25.39 25.51 29.71
Total variance 282.06 279.54 253.11 35.11 35.07 34.08 14.61 14.60 11.75 668.76 665.42 612.45
Observations 668 668 449 669 669 450 651 651 439 669 669 450
Mean dep. var. 23.98 23.98 23.98 2.00 2.00 2.00 3.43 3.43 3.43 118.61 118.61 118.61
Control variables No No Ye s No No Yes No No Ye s No No Ye s
Note. Std. errors in parenthesis. ICCs are in bold. As control variables we used size (small, medium, and large), age (new, young, and mature), % of female
borrowers, borrower retention rate, and staff turnover rate. The categories small (size) and new (age) are baseline.
*p<0.10. **p<0.05. ***p<0.01.
to 5.8% more than not-for-profit MFIs. Compared with
the mean Yield of 23.5%, the dimension of the difference
becomes even more evident.
We can infer no difference between profit and non-profit
oriented MFIs for three out of four most common
financial measures used to compare the performance
of MFIs (Armendáriz & Morduch, 2010). Hence, the
profit-maximizing behaviour of firms is not contradictory
of the main goal of microfinance.
In order to further investigate the difference between
profit and non-profit MFIs in the Yield ratio, we subsam-
pled according to firm size. We were motivated to do this
analysis because this ratio was the only one that had a sig-
nificant difference, and according to Baquero et al. (2018),
the majority of the non-profit institutions faces interest
rate ceilings, which can hinder our analysis. However,
in their sample, the yield spread is higher for non-profit
banks, which means that they have more flexibility to
define higher interest rates. Thus, we wanted to explore
and find the mechanism that may explain this difference
between for-profit and not-for-profit MFIs.
It is natural to expect that small MFIs require more
profits (comparative to larger MFIs) to sustain their activ-
ity in order to be self-sustainable and grow its business,
given its revenues are smaller. Moreover, larger MFIs have
economies of scale due to fixed costs that can be dissipated
in a larger portfolio (Gonzalez, 2007), so they may charge
smaller interest rates (therefore reducing its yield on gross
portfolio). The results for these estimations are presented
in Table 6.
The results occurred as predicted. The difference in aver-
age yield ratio, which can be interpreted as interest rates
TABLE 6 Comparison of Yield across MFI sizes (small vs. large)
Yieldongrossportfolio
Small size Large size
Profit (1 =yes) 8.14** 9.90** 4.01 3.59
(3.91) (4.87) (3.01) (2.84)
Constant 20.95*** 5.28 21.71*** 21.23***
(2.15) (8.79) (3.93) (7.81)
Observations 193 124 299 210
Control variables No Yes No Ye s
Note. Std. errors in parenthesis. Variance levels omitted.
*p<0.10. **p<0.05. ***p<0.01.
charged, is significant only in smaller MFIs. This result
is theorized as follows: smaller profit-oriented MFIs need
higher yield on gross portfolio (therefore higher interest
rates) to maintain their growth and expand their portfolio
(i.e., making more loans). However, there is no differ-
ence between profit-oriented and not-for-profit MFIs in
the larger subgroup (right half of Table 6). The comparison
of small, medium, and large firms is presented in Figure 1.
What drives this effect? We theorize that smaller
not-for-profit MFIs can rely on alternative sources of
financing (such as grants and donations), whereas
for-profit MFIs need to “self-finance” themselves by
increasing their revenues because their fixed costs can
only be distributed to a small number of clients. The larger
profit-oriented MFIs can distribute their fixed costs in a
larger portfolio, which enables them to charge smaller
interest rates in comparison to the smaller ones (Gonzalez,
2007).

LEITE ET AL. 9
FIGURE 1 Yield linear prediction for size (FP vs. NFP MFIs)
In addition, the ICC of firms at the MFI level is high
in all models presented at Table 6. It suggests that a large
percentage of variance of yield on gross portfolio is at the
MFI level. Further research must be done, but it is rele-
vant to hypothesize if the best approach for the MFIs is
to be not-for-profit when smaller and to change its nature
to profit-oriented after growth. This approach would allow
smaller yields in all sizes: large and small. Unfortunately,
few MFIs changed its profit status over time, which
makes it difficult to estimate its effects with a statistical
approach.
5ROBUSTNESS
5.1 Propensity score matching
The robustness of our findings was scrutinized with a
1-to-1 nearest neighbor propensity score matching with
replacement. Equation 6 shows the reduced form of
matching used in this study, where X is the vector of the
control variables in the previous models (we used non-
winsorized values to avoid bias). This robustness test was
performed in order to make small MFIs comparable with
large MFIs in all the observable variables except the profit
dummy variable.
ln(Y|Pro𝑓iti=1)=𝛽0+X𝛽+𝜖i.(6)
All the results from the robustness estimations are pre-
sented at Table 7. The interaction between the profit
dummy and a dummy for size (1 =large, 0 =small) was
significant at the 10% level and negative (𝛽=−7.72,p=
0.08)in model MA1 (Table 7), showing that gro wth in
profit-oriented MFIs have a larger effect in reducing yield
on gross portfolio, which means that those MFIs charge
lower interest rates.
5.2 Country level and random
coefficients
In order to test the robustness of the multilevel model, we
also added the country level in the multilevel model, with a
random intercept. One should note that we did not include
random coefficients at the country level because there are
some countries with a single MFI; thus, we cannot observe
the counterfactual, which makes the estimation of random
coefficients impossible. Therefore, the random coefficient
of the profit dummy was estimated at the region level, the
only level in which there was the variance of the profit
dummy. The equations are shown below.
Yi𝑗ck =𝛾0000 +
7
∑
n=1
𝛽ni𝑗ckCni𝑗ck +𝛽80𝑗c[k]P0𝑗c[k]
+𝜖i𝑗ck +r0𝑗ck +u000k,
(7)
where cdenotes country level and [k]denotes a random
coefficient at the region level.
Even in the model with a random profit coefficient at
the region level and a country level, the profit dummy was
still significant at 10% (𝛽=2.29,p=0.097), as shown
in model MA2 (Table 7). This is a rather robust test, with
a country level controlling different country intercepts for
Yield, and also the effect of regional heterogeneity in the
profit coefficient being controlled.
5.3 Cost variables
With this test, we look at a mechanism, which could
explain why growth impacts yield on for-profit MFIs
more than not-for-profit MFIs. We theorized that, because
for-profit MFIs cannot rely on alternative sources of
financing their activities, they have to charge higher inter-
est rates. However, as they expand, they might become
more cost-efficient (Caudill et al., 2009), which enables
them to charge interest rates that are similar to those
charged by not-for-profit MFIs.
To investigate if this theory has empirical support, we
tested whether for-profit MFIs become more cost-efficient
as they grow when compared with their not-for-profit
counterparts. If this is not the case, then the theory postu-
lated is not valid. To assess that, the main model depicted
in Equation 5 was adopted, so the dependent variable was
the Operating Expenses divided by Assets, and Operating
Expenses divided by Total Portfolio. A dummy for size (1 =
large, 0 =small) was also included, in addition to an inter-
action between size (small vs. large) and profit status (not
for profit vs. for profit).
In both models, the interaction was significant and neg-
ative (𝛽OE
assets
=−0.06,p=0.026;𝛽OE
port𝑓olio
=−0.10,p=
0.023), full results are presented in models MA3 and

10 LEITE ET AL.
TABLE 7 Results from robustness estimations
Yield OE/assets OE/portfolio Cost per borrower
MA1 MA2 MA3 MA4 MA5 MA6
Profit (1 =yes) 10.99 2.99* 0.05** 0.10*** 36.36 35.77
(7.34) (1.80) (0.02) (0.04) (31.87) (32.06)
Large MFI 1.82** 0.00 0.01 −48.22** −48.14
(0.91) (0.19) (0.03) (24.17) (24.38)
Profit x Large MFI −7.72* −0.06** −0.10** −5.08 110.86
(4.40) (0.03) (0.04) (38.07) (67.10)
Profit x Large MFI x BRR −145.77**
(69.10)
Constant 19.96*** 14.09** 0.12*** 0.16*** 305.90*** 281.77
(4.08) (5.79) (0.04) (0.06) (56.64) (57.63)
Observations 279 449 334 334 334 334
Matching Yes No No No No No
Control variables No Ye s Ye s Yes Yes Ye s
Country level No Yes No No No No
Random coefficient No Yes No No No No
Note. Std. errors in parenthesis. Variance levels omitted. BRR stands for borrower retention rate.
*p<0.10. **p<0.05. ***p<0.01.
MA4 (Table 7). Thus, when compared with not-for-profit
MFIs, growth improves the cost-efficiency of for-profit
MFIs more, what is in accordance with (Caudill et al.,
2009). Hence, we argue that this is the reason why
larger MFIs are able to charge as low interest rates than
for-profit MFIs.
5.4 Capital structure
We now proceed to investigate the capital structure of
for-profit versus not-for-profit MFIs. We do that by inves-
tigating the return on equity of the for-profit and the
not-for-profit MFIs. Moreover, we also address the effect
of the legal form of MFIs, which could affect their capi-
tal structures, on the four measures of financial outcomes
analysed in this paper (Yield, ROA, PAR30, and OSS).
Table 8 presents the results.
First, we show that the legal form, per se, does not seem
to impact the sustainability outcomes of MFIs (columns
L1–L4). Hence, we show that the differences in for-profit
and not-for-profit MFIs are mainly driven not by the legal
form but from the profitability status of those institutions.
Moreover, we show that for-profit institutions do not
have a higher return on equity when compared with their
not-for-profit counterparts (column L5). This corroborates
our result that it does not seem likely that there is a mission
drift of for-profit MFIs because they do not have higher
returns on their capital (hence return to the shareholders)
when compared with the not-for-profit ones.
5.5 Outreach
As previously explained, there is evidence that for-profit
MFIs tend to target the “richest of the poor” in order to
increase the profitability and its sustainability (Cull et al.,
2007; Cull et al., 2011). The results previously presented
show that we indeed found that for-profit MFIs are more
prone to charge higher interest rates because there is a
significant difference in the yield between for-profit and
not-for-profit MFIs.
Henceforth, although this seems to indicate that the
MFIs are targeting a different sector of the population, we
now test whether the outreach of the for-profit MFIs is
different from their for-profit counterparts.
We define the outreach of the MFIs as both the number
of outstanding loans and the average loan balance (nor-
malized by the GNI per capita). Table 9 shows the results.
Columns T1 and T2 show the coefficients are significant
and positive for both variables.
Hence, for-profit MFIs have a different outreach than
not-for-profit ones: They tend to target the “richest of
the poor” with higher loans, and also they grant more
loans as well. Notwithstanding, the not-for-profit ones
give fewer loans, which are also smaller, thus tending
to a different niche when compared with their for-profit
counterparts.
Thus, we show that there is a trade-off between financial
performance and outreach. Although serving the “richest
of the poor” enable MFIs to grant more loans, they can-
not impact the poorest. The not-for-profit ones can fill this

LEITE ET AL. 11
TABLE 8 Results from robustness estimations (capital structure of MFIs)
Yield ROA PAR30 OSS ROE
L1 L2 L3 L4 L5
Bank 6.07 3.79 1.05 6.07
(18.44) (4.39) (2.55) (18.44)
Credit union/cooperative −2.57 2.03 1.83 −2.57
(17.85) (4.25) (2.47) (17.85)
NBFI 9.92 4.16 0.48 9.92
(17.50) (4.17) (2.42) (17.51)
NGO 4.87 3.65 0.53 4.86
(17.60) (4.18) (2.44) (17.60)
Profit 0.01
(0.04)
Constant 99.64*** −1.41 6.58** 99.64 −0.05
(19.24) (4.45) (2.70) (19.24) (0.13)
Observations 450 450 439 450 450
Control variables Yes Ye s Yes Yes Ye s
Note. Std. errors in parenthesis. Variance levels omitted. NBFI stands for nonbanking financial
institution. NGO stands for non-governmental organization.
*p<0.10. **p<0.05. ***p<0.01.
TABLE 9 Results from robustness estimations (Sections 5.5 and 5.6)
Avg. loan size/GNI ln(# of outstanding loans) Yield
T1 T2 T3
Profit 0.37** 0.74*** 3.73*
(0.15) (0.14) (2.06)
Prime interest rate 0.20**
(0.09)
Constant 1.12*** 7.72*** 5.26
(0.23) (0.35) (6.51)
Observations 450 450 447
Control variables Yes Yes Yes
Note. Std. errors in parenthesis. Variance levels omitted.
*p<0.10. **p<0.05. ***p<0.01.
gap by granting fewer loans targeting this parcel of the
population.
However, we have no evidence of a “mission drift” by
the for-profit MFIs, namely, they did not start to increase
the interest rate to become the new “moneylenders.” Evi-
dence of that is presented at Section 4, where we show that
there is no difference in the yield charged by for-profit and
not-for-profit MFIs (Table 6).
5.6 Exogenous control variable
We used the approach defined by Cull et al. (2009), in
which the prime interest rate plus two percentage points
is used as an exogenous control variable. We collected data
from the “lending interest rate” for each country (adding
two percentage points to account for the increase in the
perceived risk of microfinance loans) from the World Bank
website, and we rerun our estimate on the Yield depen-
dent variable, because it was the only variable in which
there was a significant difference between for-profit and
not-for-profits MFIs.
As column T3 of Table 9 shows, the use of this exoge-
nous control variable does not affect the profit coefficient,
which remains significant. Moreover, we also show that
this prime interest rate is positively correlated with the
yield, showing that in countries with greater costs of cap-
ital, which require higher interest rates, the MFIs also
increase their yield in order to remain sustainable.
Additionally, the use of the prime interest rate plus two
percentage points is regarded by Armendáriz and Mor-
duch (2010) as a “better measure” to account for the finan-
cial self-sufficiency of MFIs when compared with the FSS
ratio, which is regarded as being “imperfect as a guide
to sustainability”. Hence, we show that the MFIs ensure
their financial self-sufficiency by charging slightly higher
interest rates in countries with higher costs of capital.

12 LEITE ET AL.
6CONCLUSION,
CONTRIBUTIONS, AND POLICY
IMPLICATIONS
Our results show that not-for-profit MFIs can have as
good financial ratios (ROA, PAR30, and OSS) as for-profit
MFIs and also maintain lower interest rates for their
clients. However, as Cull et al. (2011) noted, this does
not mean that not-for-profit MFIs have as good an out-
reach as for-profit MFIs, because for-profit MFIs may have
better sources of financing itself in the market. In addi-
tion, contrary to what Baquero et al. (2018) noted, we find
no difference in portfolio at risk (PAR30) between profit
and not-for-profit MFI's, indicating that clients from both
institutions have similar levels of delinquency.
Our results have important implications. This study
shows that there is not an inherently better approach for
microfinance. Thus, policymakers can use the advantages
of both models of MFIs to achieve the goal of poverty
reduction. In addition, we show that growth influences
more the cost efficiency of for-profit MFIs than their
not-for-profit counterparts. These results corroborate with
other studies which show that there is no intrinsically
better microfinance approach between profit-oriented and
not-for-profit institutions (Morduch, 2000; Schmidt, 2013).
Each one serves a different purpose. Whereas for-profit
MFIs may aim at the “richest of the poor” (Cull et al., 2007;
Cull et al., 2009), thus being able to achieve higher out-
reach, and charge larger interest rates (Cull et al., 2011),
not-for-profit MFIs may aim to have a smaller outreach
and could tend to serve the poorer with lower interest rates.
For the majority of the outcomes, there is no differ-
ence in sustainability between for-profit and not-for-profit
MFIs. The duality between welfarists and institutionalists
is not opposite, at least in financial measures of perfor-
mance. Independent of poverty reduction, we found evi-
dence that both perspectives presented similar operational
sustainability (OSS).
Our results allowed us to reach the two objectives of this
study: to address the effect of being profit or not-for-profit
oriented and to observe the heterogeneity of the data over
different levels of analysis. Moreover, we identified where
this effect is evident.
We found that small profit-oriented MFIs have a larger
yield on the gross portfolio when compared with the
not-for-profit ones. We concluded this is true because
not-for-profit MFIs have alternative sources of financing
their activities, whereas for-profit MFIs need a higher
yield on the gross portfolio to maintain self-growth. Large
profit-orientedMFIs do nothaveasignificant different yield
when compared with not-for-profit ones, which shows that
larger MFIs are more efficient with their expenses and also
that they can distribute their fixed costs to a larger portfo-
lio, with a similar risk (by not having a different portfolio
at risk), allowing for smaller interest rates.
Additionally, this study was able to estimate the impor-
tance of time, institutional characteristics, and region on
the four sustainability indexes used in similar studies. On
average, the intrinsic characteristics of the MFI account for
70% of the variance, whereas time accounts for 25% of the
overall variance. The effects of different regions are shown
to be quite small, accounting for only 5% of the overall vari-
ance. This shows that the intrinsic characteristics of firms
have the biggest explanatory power, and region character-
istics explain a small portion of the financial indexes. This
study shows that MFIs in different regions, but with equal
intrinsic characteristics, are quite comparable. This result
gives robustness to experiments that are conducted across
different regions, such as Banerjee et al. (2015a), Banerjee,
Karlan, and Zinman (2015b), and Cull et al. (2009).
Regarding financial sustainability of MFIs, Cull,
Demirguç-Kunt, and Morduch (2007) focused their
research on analysing the sustainability of MFIs regard-
ing the lending's type (individual, solidarity, and village),
which is different than our main goal (compare for-profit
vs. not-for-profit MFIs). They found no significant differ-
ence in the profit status of MFIs and the yield; however,
we found that for-profit MFIs charge higher interest rates.
Our results are also robust to a propensity score match-
ing, which strengthens our causality claim. Moreover, by
using a hierarchical linear model, we can decompose the
variance in each level and find how much variance time,
MFI, and region levels. Thus, our results complement
those of Cull et al. (2007).
Despite the lack of exogenous variation, we diminished
the endogeneity effect by adopting three different method-
ologies. The first approach is the traditional control vari-
ables approach (age, size, percentage of female borrowers,
the retention rate of borrowers, and staff turnover rate).
The second approach used was a multilevel model allow-
ing for random intercepts, in which we break the variance
and the error term. This multilevel approach can lead
to more truthful and less biased conclusions (Hanchane
& Mostafa, 2012). The third approach uses a propensity
score matching in order to weigh the regression with the
probability of being a smaller or a larger MFI. A possible
limitation is that the data entries are self-reported, which
can result in misleading information given by the MFIs.
We address this issue by using only data from audited
MFIs. However, the only way to arrive at results that we
can claim strict causality is by using either an experiment,
a quasirandom setup (exogenous shock) or an instru-
mental variable approach. Nevertheless, none of these
approaches are feasible in this study, especially the experi-
mental approach, because we cannot randomly assign the
MFIs to be for profit or not for profit.
A final limitation is that many MFIs did not report
all the information that was used as controls variables,
which reduced our sample and consequently diminished
our statistical power. We adopted this perspective in order

LEITE ET AL. 13
to have a global comparison of MFIs, privileging exter-
nal validity. Notably, even with smaller statistical power,
we were able to achieve significant statistical effects and
important theoretical contributions.
ACKNOWLEDGEMENT
We thank participants and audience at the 2016 Annual
Meeting of the Brazilian Academy of Management and the
2016 USP International Conference in Accounting. This
study was financed in part by the Coordenacao de Aperfe-
icoamento de Pessoal de Nivel Superior - Brazil (CAPES) -
Finance Code 001. We thank Prof. Rafael Goldszmidt and
Prof. Fabio Caldieraro for their suggestions regarding pre-
vious versions of this paper.
ORCID
Rodrigo de Oliveira Leite https://orcid.org/
0000-0003-3504-4639
REFERENCES
Angelucci, M., Karlan, D., & Zinman, J. (2015). Microcredit impacts:
Evidence from a randomized microcredit program placement
experiment by compartamos Banco. American Economic Journal:
Applied Economics,7(1), 151–182.
Armendáriz, B., & Morduch, J. (2010). The Economics of Microfi-
nance. Cambridge, Mass: MIT press.
Armendáriz, B., & Szafarz, A. (2011). On mission drift in microfi-
nance institutions, The Handbook of Microfinance (pp. 341–366).
London, England: World Scientific.
Ashta, A., & Hudon, M. (2012). The Compartamos microfinance IPO:
Mission conflicts in hybrid institutions with diverse sharehold-
ing. Strategic Change,21(7-8), 331–341.
Banerjee, A., Duflo, E., Goldberg, N., Karlan, D., Osei, R., Parienté,
W., ... Udry, C. (2015a). A multifaceted program causes lasting
progress for the very poor: Evidence from six countries. Science,
348(6236), 1260799.
Banerjee, A., Karlan, D., & Zinman, J. (2015b). Six randomized eval-
uations of microcredit: Introduction and further steps. American
Economic Journal: Applied Economics,7(1), 1–21.
Baquero, G., Hamadi, M., & Heinen, A. (2018). Competition, loan
rates, and information dispersion in nonprofit and for-profit
microcredit markets. Journal of Money, Credit and Banking,50(5),
893–937.
CGAP (2010). Global implications of the crisis in indian microfi-
nance. Consultative group to assist the poor (CGAP), Focus Note
67.
Caudill, S. B., Gropper, D. M., & Hartarska, V. (2009). Which micro-
finance institutions are becoming more cost effective with time?
Evidence from a mixture model. Journal of Money, Credit and
Banking,41(4), 651–672.
Cull, R., Demirguç-Kunt, A., & Morduch, J. (2007). Financial per-
formance and outreach: A global analysis of leading microbanks.
The Economic Journal,117(517), F107–F133.
Cull, R., Demirgüç-Kunt, A., & Morduch, J. (2009). Microfinance
meets the market. Journal of Economic Perspectives,23(1),
167–192.
Cull, R., Demirgüç-Kunt, A., & Morduch, J. (2011). Does regula-
tory supervision curtail microfinance profitability and outreach?
World Development,39(6), 949–965.
Cull, R., Demirgüç-Kunt, A., & Morduch, J. (2018). The microfinance
business model: Enduring subsidy and modest profit. The World
Bank Economic Review,32(2), 221–244.
Gonzalez, A. (2007). Efficiency drivers of microfinance institutions
(MFIs): The case of operating costs. Microbanking Bulletin,15,
37–42.
Hanchane, S., & Mostafa, T. (2012). Solving endogeneity problems in
multilevel estimation: An example using education production
functions. Journal of Applied Statistics,39(5), 1101–1114.
Hermes, N., Lensink, R., & Meesters, A. (2011). Outreach and effi-
ciency of microfinance institutions. World Development,39(6),
938–948.
Hudon, M., & Traca, D. (2011). On the efficiency effects of subsidies
in microfinance: An empirical inquiry. WorldDevelopment ,39(6),
966–973.
Karlan, D. S., & Zinman, J.(2008). Credit elasticities in less-developed
economies: Implications for microfinance. American Economic
Review,98(3), 1040–1068.
Mersland, R., & Strøm, RØ (2012). Microfinance: Costs, lending
rates, and profitability, Handbook of key global financial mar-
kets, institutions, and infrastructure (pp. 489–499). Cambridge,
Massachusetts, United States: Elsevier.
Morduch, J. (1999). The microfinance promise. Journal of Economic
Literature,37(4), 1569–1614.
Morduch, J. (2000). The microfinance schism. World Development,
28(4), 617–629.
Norell, D. (2001). How to reduce arrears in microfinance institutions.
Journal of Microfinance/ESR Review,3(1), 115–130.
Olivares-Polanco, F. (2005). Commercializing microfinance and
deepening outreach? Empirical evidence from Latin America.
Journal of Microfinance/ESR Review,7(2), 5.
Paxton, J. (2007). Technical efficiency in a semi-formal financial
sector: The case of Mexico. Oxford Bulletin of Economics and
Statistics,69(1), 57–74.
Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models:
Applications and data analysis methods, Vol. 1. Thousands Oaks,
CA: Sage.
Rosenberg, R., Gonzalez, A., & Narain, S. (2009). The new moneylen-
ders: Are the poor being exploited by high microcredit interest
rates?
Schmidt, R. H. (2013). Microfinance et éthique. Revue D'économie
Financière,112(4), 243–276.
Wichterich, C. (2012). The other financial crisis: Growth and crash
of the microfinance sector in India. Development,55(3), 406–412.
Zeller, M., & Meyer, R. L. (2002). The triangle of microfinance: Finan-
cial sustainability, outreach, and impact. Baltimore: Intl Food
Policy Res Inst.
How to cite this article: Leite RO, Mendes LS,
Sacramento LC. To profit or not to profit? Assess-
ing financial sustainability outcomes of Micro-
finance institutions. Int J Fin Econ. 2019;1–13.
https://doi.org/10.1002/ijfe.1718