Journal of Business Research 94 (2019) 1–17

Contents lists available at ScienceDirect

Journal of Business Research

journal homepage: www.elsevier.com/locate/jbusres

The effect of loan approval decentralization on microfinance institutions' T

outreach and loan portfolio quality☆
⁎
Hubert Tchakoute-Tchuigouaa, , Issouf Soumaréb
a
Department of Finance and Accounting, KEDGE Business School, 680 Cours de la Libération, 33 405 Talence cedex, France
b
Department of Finance, Insurance & Real Estate, & Laboratory for Financial Engineering (LABIFUL), Faculty of Business Administration, Université Laval, Building
Palasis-Prince, Office 1204, 2325, rue de la Terrasse, Quebec, Quebec G1V 0A6, Canada

A R T I C LE I N FO A B S T R A C T

JEL classifications: We study the impact of loan approval decentralization on MFI portfolio quality and outreach, and the effects of
G21 alignment mechanisms when loan officers combine information production and decision functions. Using an
G32 independently pooled cross-section of 374 MFI-year observations for 280 MFIs in 70 countries, we find that
G39 effective incentive schemes and internal control systems help mitigate agency problems within MFIs, and thus
Keywords: increase the outreach of MFIs without altering the quality of their loan portfolio. Our results are robust after
Loan approval controlling for alternative portfolio risk and outreach measures, outreach threshold effect, crisis period, selection
Decentralization bias and endogeneity.
Loan portfolio quality
Loan portfolio risk
Microfinance
Outreach

1. Introduction
Microfinance institutions (MFIs) are double-bottom-line organiza-
tions and hybrid organizations (Battilana & Dorado, 2010; D'Espallier,
Hudon, & Szafarz, 2013). They are now part of the financial landscape
of most developing and emerging countries, and their primary goal is to
provide financial services to low income people and to small and in-
formal businesses. As reported by some recent studies (Armendáriz de
Aghion & Morduch, 2010; Dixon, Ritchie, & Siwale, 2007; Tchakoute
Tchuigoua, 2012), decision-making authority is allocated to the loan
officer in many MFIs. Merging resource allocation and information
production functions provides the loan officer with incentives to pro-
duce and use soft information when approving loans (Stein, 2002). The
loan officer, whose role in the loan approval process is now evidenced
in microfinance literature (Agier, 2012; Agier & Szafarz, 2013), be-
comes more powerful when information production and loan approval
decisions are concentrated in his or her hands. In this study, we first
investigate whether allocating decisional authority to the loan officer
improves outreach without deteriorating loan portfolio quality.
However, a decentralized credit decision or a powerful loan officer
can expose the MFI to a principal-agent problem and thus induce
agency conflicts between the loan officer and the lending organization
(Berger & Udell, 2002; Stein, 2002). A loan officer may make credit
decisions contrary to the interests of the MFI or without complying with
current loan approval procedures. Implementing appropriate incentive
schemes (better human resource management), strong internal controls,
and audit procedures are thus crucial for microcredit portfolio perfor-
mance and the overall financial health of MFIs (Basel Committee on
Banking Supervision (BCBS), 2010). Appropriate incentive mechanisms
and internal control procedures may ensure that loan officers who ap-
prove loans undertake allocation decisions that improve MFI efficiency.
The effect of loan approval decentralization on risk and outreach will be
especially important if the agent who received authority is subject to
incentive systems and controls that would tend to limit the risk asso-
ciated with decentralization. The second objective of this study is to
investigate whether incentive schemes and internal control systems are
effective in reducing agency conflicts within an MFI.
To date, the microfinance literature has not sufficiently addressed
the issue of agency costs induced by delegation. More specifically, the
question of whether well-designed incentives and an effective internal
control system – whether perceived or assessed as such by a third party
– could limit the risk associated with loan approval decentralization is

☆
We are grateful to participants at the seminar organized by the Department of Finance, Insurance and Real Estate and the Research Chair in Governance of Laval
University in April 2016 for their helpful comments and suggestions, especially Jean Bédard. All errors and omissions are the authors' sole responsibilities.
⁎
Corresponding author.
E-mail addresses: hubert.tchakoute@kedgebs.com (H. Tchakoute-Tchuigoua), issouf.soumare@fsa.ulaval.ca (I. Soumaré).

https://doi.org/10.1016/j.jbusres.2018.09.021
Received 9 May 2017; Received in revised form 19 September 2018; Accepted 20 September 2018
Available online 26 September 2018
0148-2963/ © 2018 Elsevier Inc. All rights reserved.
H. Tchakoute-Tchuigoua, I. Soumaré
not well documented.1 Some recent microfinance papers investigate the
risk associated with decentralization in microfinance lending in devel-
oping countries, including subjective preferences of the loan officer and
the impact of loan officer behaviors on risk and MFI outreach (Agier &
Szafarz, 2013; Labie, Méon, Mersland, & Szafarz, 2015; Sagamba,
Shchetinin, & Yusupov, 2013). Agier (2012) emphasizes the role of
credit officers on microloan performance and Tchakoute Tchuigoua
(2012) examines the effect of decentralization on loan contract terms.
This empirical literature on microfinance then suggests a link between
decentralization and adverse selection and provides evidence that
granting authority to loan officers may bias the selection process be-
cause of subjective loan officer preferences. The effectiveness of ex-
isting alignment mechanisms in MFIs can therefore enable MFIs to
overcome this principal-agent problem arising from decentralizing the
loan decision, and also align loan officer interests with those of the MFI.
To analyze the effectiveness of alignment mechanisms on the risk
and outreach of MFIs, we study an independently pooled cross-sectional
sample of 374 MFI-year observations for 280 MFIs from 2001 to 2012
across 70 countries and make at least two main contributions to the
existing microfinance literature.
First, we extend upon previous studies on loan officer role and be-
havior in MFIs. Contrary to Agier (2012) who limits the role of the loan
officer to information production, we assume loan officer duality and
the resulting principal-agent problem between the MFI and the loan
officer who both produces information and allocates loans. By assessing
the mitigating effect of incentive schemes and internal control me-
chanisms on loan officer behavior, our contribution also goes beyond
the previous literature, which focuses on loan officer subjectivity (Agier
& Szafarz, 2013; Labie et al., 2015; Sagamba et al., 2013). These studies
assume or demonstrate that discriminatory behavior of a loan officer is
partly because of inexistence of a control system or failures of existing
internal control systems. The quality of the internal control system and
the quality of incentive mechanisms are not taken into account from an
empirical point of view. Our study accounts for this missing piece by
considering the existence of effective internal control systems and in-
centive mechanisms. We measure the effectiveness of the internal
control system and the incentive system using the rating scores pro-
duced by a rating agency, Planet Rating in this case, while controlling
for the associated selection bias.
Second, by focusing on loan officer-MFI agency problems and on the
effectiveness of alignment mechanisms within MFIs, we extend upon
the previously mentioned microfinance studies and contribute to the
existing microfinance corporate governance literature. Indeed, some
studies have focused on the agency conflicts between MFIs executives
and owners and on the effectiveness of incentive schemes and control
mechanisms implemented by MFIs in order to align the interests of
these two groups (e.g., Hartarska, 2005; Hartarska & Mersland, 2012;
Mersland & Strøm, 2009; Tchakoute Tchuigoua, 2014). Based on pre-
vious studies in nonfinancial organizations (Adams, Almeida, &
Ferreira, 2005), MFIs (Galema, Lensink, & Mersland, 2012) and banks
(Pathan, 2009), some others have placed particular emphasis on CEO
power, that is, the merging of Chairman and CEO positions into a single
position, and examined its impact on performance and risk. Other
studies focus on agency problems between the interest of borrowers and
the interest of MFIs, and assess the effectiveness of incentive design for
aligning the borrower and MFI preferences (Armendáriz de Aghion &
1
The existing banking literature emphasizes lending practices of large and
small banks and their effects on credit availability, risk and profitability (Berger
& Black, 2011; Berger, Cowan, & Frame, 2011; Berger, Miller, Petersen, Rajan,
& Stein, 2005); loan officer's rotation policy in commercial banking and its
effects on moral hazard, that is, the officers' reporting behaviors (Hertzberg,
Liberti, & Paravisini, 2010); hierarchical distance of information use in large
multinational banks (Liberti & Mian, 2009); and gender bias in bank lending
markets (Beck, Behr, & Guettler, 2013; Bellucci, Borisov, & Zazzaro, 2010).
2
Journal of Business Research 94 (2019) 1–17
Morduch, 2010; Stiglitz, 1990). However, except for a few theoretical
studies that have examined the incentives designed for loan officers
(Aubert, de Janvry, & Sadoulet, 2009; Besley & Ghatak, 2005; Conning,
1999), little is known from an empirical standpoint about a principal-
agent problem involving MFIs and loan officers, and the effectiveness of
alignment mechanisms designed to avoid information problems within
MFIs.
Our empirical evidence supports the hypothesis that providing the
loan officer with incentives such as performance-based pay and putting
in place an effective internal control system may contribute to aligning
the interest of the loan officer with that of the institution. Hence, im-
plementation of human resource management practices and internal
control systems, which are perceived as effective, mitigates the possible
deterioration of loan portfolio quality following loan approval decen-
tralization, without altering MFI outreach. This result links our study to
the literature on organizational architecture (Berger & Udell, 2002;
Stein, 2002) and to microfinance-specific literature which focuses on
designing incentives for loan officers (Aubert et al., 2009; BCBS, 2010;
Labie et al., 2015). We account for the effect of MFI ownership type,
given that managerial discretion and profit distribution constraints are
likely to vary across different types of ownership among MFIs. We find
that incentive mechanisms and internal control system effectiveness
increase MFI outreach without altering loan portfolio quality when loan
approval is decentralized, in both not-for-profit and profit-oriented
MFIs; the effects are much stronger in not-for-profit organizations and
even tend to reduce risk.
The remainder of the article is organized as follows. Part two gives
an overview of prior literature and develops the research hypotheses.
Part three explains the research design. Part four presents the results
and robustness tests, and part five concludes with an acknowledgment
of research limitations and avenues for future research.
2. Background
2.1. Why is loan approval decentralization an important issue for
microlending?
The corporate governance literature provides an explanation of why
firms grant decision-making authority to agents (loan officers).
Incentive-based theories (Aghion & Tirole, 1997; Stein, 2002) suggest
that transferring the decision-making authority to agents who produce
soft information is a way to recognize and reward their expertise in this
field and enable the loan officers to make appropriate decisions. Sah
and Stiglitz (1986) explain the choice of a decentralized decision-
making structure in terms of the cost of acquiring and communicating
information. Given information asymmetries between the person who
gathers and processes information, and the one with the authority to
make decisions, the data communicated by the former to the latter can
be either partial or contaminated, thereby leading to flawed decision
making.
Investigating decentralization of the loan approval process is of
particular interest in the microfinance sector for at least two reasons.
First, close ties between MFIs and their clients are one of the main
features of microfinance (Stiglitz, 1990). Close proximity to the poor
not only makes it easier for MFIs to understand their clients' needs but
also enables them to develop trust with the communities in which they
operate and develop and offer products and services in line with the
financial needs of their intended target markets (see, for example,
Ledgerwood, Earne, & Nelson, 2013). Proximity also enables loan of-
ficers to produce soft information and to grant loans efficiently. Given
that loan officers usually live in the same local community as their
borrowers and maintain direct and personal contact with them, they
may build privileged ties with other small businesses and individuals
who hold relevant information about potential borrowers and their
businesses in the local community (Berger & Udell, 2002). In addition,
daily interactions and personal relations between loan officers and local
H. Tchakoute-Tchuigoua, I. Soumaré
borrowers provide easy access to soft information.
Second, according to the BCBS (2010), most MFI failures stem from
deterioration in the quality of loan portfolios. Christen, Lauer, Lyman,
and Rosenberg (2012) suggest that there may be agency problems in
loan allocation and that deterioration of an MFI's loan portfolio quality
may result from the ineffectiveness of the allocation process and the
internal governance system. The 2014 Centre for the Study of Financial
Innovation (CSFI) survey, Microfinance Banana Skins, ranks credit risk
(that is, risk that poor lending practices lead to loan losses) second
among risks facing MFIs and finds strong links with weaknesses in
management and internal control systems. In addition, the microcredit
market has become increasingly competitive, and conclusions of the
CSFI (2014) show that microfinance experts rank competition third
among the risks facing MFIs. The increased competition among MFIs in
the microcredit market pushes them to expand their credit portfolio and
to increase their market share. Competition enables MFIs to expand
their outreach but seems to exacerbate agency problems in loan allo-
cation. Some recent empirical studies evidence the negative effect of
competition on MFI loan portfolios (Assefa, Hermes, & Meesters, 2013;
Baquero, Hamadi, & Heinen, 2018; Guha & Chowdhury, 2013;
Tchakoute Tchuigoua, 2016). To address the loan portfolio effect of
competition, MFIs implement human resource policies among which
staff incentive systems to ensure that the lending activity will not jeo-
pardize the sustainability and survival of the MFI.
2.2. Hypotheses development
2.2.1. The effect of Loan officer approval on MFI outreach and loan
portfolio quality
Recent studies using household survey data in developing and
emerging countries document that proximity allows MFIs to produce
soft information and improves access to finance (Allen et al., 2014;
Allen, Demirgüc-Kunt, Klapper, & Martinez-Peria, 2016; Brown, Guin,
& Kirschenmann, 2016). Lending decisions based on the use of soft and
hard information enable MFIs to better screen borrowers. Simultaneous
use of soft and hard information may lead to accepting loan applica-
tions that could have been rejected solely on the basis of hard in-
formation, which is not necessarily reliable or relevant in the micro-
finance sector. Decentralization thus helps to increase access to credit
and improve loan conditions. We may thus expect that outreach in-
creases when the loan officer approves loans, which leads to the fol-
lowing Hypothesis 1.a.
H1a. Loan approval decentralization is positively associated with MFIs
outreach.
Using soft information as a complement to hard information thus
reinforces the ability of MFIs to select borrowers efficiently based on
their repayment capacity, which de facto limits the risk that the credit
portfolio will deteriorate. Decentralization contributes to limiting the
risk in the credit decision and may have a positive effect on the quality
of the MFI loan portfolio. We may thus expect the riskiness of the loan
portfolio to decrease when the loan officer approves loans (Hypothesis
1.b).
H1b. Loan approval decentralization is positively associated with MFIs
loan portfolio quality.
2.2.2. The mitigating effect of alignment mechanisms
Decentralization of the credit decision is not a sufficient condition to
overcome informational problems in MFIs. Indeed, a powerful loan
officer can create a principal-agent problem. Strong relationships be-
tween the loan officer and the MFI's clients (borrowers) and the re-
sulting lending decentralization gives considerable leeway to loan of-
ficers, especially when incentive schemes and internal control systems
are lacking, and the existing ones ineffective. For subjective reasons, the
loan officer may voluntarily grant non-performing loans by ignoring
3
Journal of Business Research 94 (2019) 1–17
qualitative information that could lead to the rejection of credit ap-
plications that he or she defends (Berger & Udell, 2002). The loan of-
ficer can thus voluntarily select borrowers with low ability to repay,
which may result in nonperforming loans and thus deteriorate the MFI's
loan portfolio quality. According to one stream of the governance lit-
erature (Berger & Udell, 2002; Brickley, Linck, & Smith, 2003; Stein,
2002) and the microfinance-specific literature (Aubert et al., 2009;
Jeon & Menicucci, 2011; Labie et al., 2015), providing the loan officer
with incentives such as performance-based pay and an effective internal
control system may thus contribute to aligning the interests of the in-
stitution with the loan officer's interest, and thus avoid information
asymmetry between the principal (the MFI's management or its credit
committee) and the agent (the loan officer).
Aubert et al. (2009) developed an ex ante information acquisition
model and show that setting up an incentive mechanism will tend to
align the interest of credit officer with that of the MFI. The credit officer
will thus select borrowers adequately based on their ability to repay
and their poverty levels. Assuming an agency conflict between the MFI
and the biased loan officer with loan decision-making authority, Labie
et al. (2015) show how designing wage incentives in a nonprofit MFI
may deter its loan officers from discriminating. They document that
offering incentives in socially oriented MFIs mitigates discrimination by
loan officers. Jeon and Menicucci (2011) demonstrate how incentive
schemes help avoid moral hazard and constrain loan officer behavior.
Incentive schemes limit the loan officer's ability to embezzle borrower
repayments. These theoretical studies suggest that well-designed in-
centive schemes allow an MFI to improve outreach, limit MFI dis-
crimination against disabled micro-entrepreneurs, and help improve
the MFI's loan portfolio quality.
From an empirical standpoint, except for exploratory studies,2
analysis of the effectiveness of incentive mechanisms for loan officers
and internal control systems in reducing agency conflicts within the
MFI has received very little attention in the microfinance literature.
Based on the previously described literature, we may expect that pro-
viding incentives to the loan officer and a better internal control system
are likely to mitigate the effects of powerful loan officers on MFI per-
formance. We thus expect that incentive schemes and internal control
systems may ensure that loan officers who approve loans make allo-
cation decisions that improve the MFI's outreach and the loan portfolio
quality, hence the following Hypothesis 2.
H2. The effects of loan approval decentralization on loan portfolio
quality and outreach are attenuated when alignment mechanisms are
effective.
2.2.3. The effect of MFI ownership types
MFIs offer financial services to the poor across several institutional
forms including cooperatives, credit unions and non-governmental or-
ganizations (NGOs), usually grouped in the category of nonprofit MFIs,
and banks and nonbank financial institutions (NBFIs), considered as
privately owned and profit-oriented MFIs. Since the transformation
movement of MFIs was initiated in the early 1990s, the debate over the
impact of choosing an institutional form and the impact of the evolution
of ownership type remains a subject of concern among practitioners and
scholars in the field of microfinance. Some studies have attempted to
answer the question of whether the financial and social performance of
2
An international survey on staff incentive scheme practices in the micro-
finance industry (McKim & Hughart, 2005) shows that the percentage of MFIs
implementing staff incentive schemes increased for > 12 years. They show that
the percentage of MFIs using staff incentive schemes grew from 6% to 63%
between 1990 and 2003. In 2005, they found 72% of MFIs had an incentive
scheme, with some differences across regions, by legal status and by MFI
characteristics. Incentive schemes for loan officers are individual monetary
schemes (bonuses) and are found in 83% of MFIs.
H. Tchakoute-Tchuigoua, I. Soumaré
MFIs varies across MFIs ownership type, and find consistent evidence.
Results do not evidence differences in accounting performance across
MFI ownership types and thus yield inconsistent evidence with the
transformation thesis. Mersland and Strøm (2008) and Tchakoute
Tchuigoua (2010), among others, show that NGOs are not more socially
oriented than privately owned MFIs, nor are privately owned MFIs
more profitable than NGOs. This is evidenced as well in a more recent
study by D'Espallier, Goedecke, Hudon, and Mersland (2017), which
applies an event study methodology to compare the performance of 66
transformed MFIs over the period 1993–2011, and investigates how
transformation influences the MFI's overall business model, including
the MFI's profit function, the mix of funding sources and the scope of
services provided.
MFIs clients' characteristics are one of the main distinctive features
of microfinance ownership type (Jansson, Rosales, & Westley, 2004;
Tchakoute Tchuigoua, 2010). As noted by Galema et al. (2012), mi-
crofinance NGOs borrowers tend to be low-income entrepreneurs, work
in informal sectors, with no physical capital and have limited formal
documentation. In these microfinance NGOs, loan approval is mostly
decentralized and based on the use of soft lending technology and on a
field evaluation of the client's ability to pay. Conversely, privately
owned MFIs target wealthier clients with greater ability to repay their
loans, and are more centered on the use of hard lending technology to
assess risk and the borrower's ability to repay. This would result in
better credit portfolio quality and larger average loan size in for-profit
MFIs than in other institutional forms. This seems to be confirmed by
some existing empirical studies. Indeed, D'Espallier et al. (2017) find a
significant increase in average loan size, both in absolute terms and
scaled by GNI per capita for transformed MFIs. Tchakoute Tchuigoua
(2010) finds that private microfinance companies have better loan
portfolio quality than nonprofit MFIs. He also finds no difference among
nonprofit MFIs, that is, between cooperatives and NGOs. The existing
differences between for-profit MFIs and not-for-profit ones may be ex-
plained by the fact that, privately owned MFIs adopt management
procedures and practices similar to those of conventional banks, and
they may be expected to implement similar sophisticated risk man-
agement practices, as opposed to those used by cooperatives and NGOs.
This finding may also be driven by differences in staff incentives and
internal control systems within MFIs.
In addition, as documented by Galema et al. (2012) and Servin,
Lensink, and Van den Berg (2012), each MFI ownership type is specific
in terms of managerial discretion (the level of ties between ownership
and control) and profit-distribution constraints (those who receive
profits earned by MFIs). NGOs are characterized by a non-distribution
constraint and cooperative/credit unions and privately owned MFIs
distribute profits, respectively, to members and owners. Privately
owned MFIs apply market-based principles that may enable them to
implement mechanisms, such as market or corporate control or per-
formance-based compensations, to effectively influence opportunistic
behaviors of loan officers. Perilleux, Hudon, and Bloy (2012) and
Hudon and Périlleux (2014) also provide supportive evidence for an
existing difference in profit distribution constraints across MFIs own-
ership type. They use a stakeholder-centered approach and answer the
question of how MFIs allocate their surplus to stakeholders and whether
the allocation process varies depending on the MFI ownership type.
They document that the surplus distribution is more in favor of provi-
ders and employees in cooperatives, whereas, NGOs and shareholder-
MFIs keep their surplus internally as a self-financing mechanism.
Moreover, according to Aubert et al. (2009), designing incentives to
align the interests of employees with the MFI's objectives is more costly
in socially oriented MFIs than in for-profit oriented MFIs. This may even
results in a trade-off between fighting discrimination against disabled
borrowers and extending outreach to the poor in nonprofit MFIs (Labie
et al., 2015). These latter two studies are consistent in the way that, in
the presence of agency costs, introducing incentives improves social
and financial performance in for-profit MFIs.
4
Journal of Business Research 94 (2019) 1–17
The above literature review allows us to conclude that for-profit
MFIs have better outreach and better loan portfolio quality than non-
profit MFIs. However, as argued previously, allocating the loan deci-
sion-making authority to the loan officer can lead the loan officer to
prioritize his subjective preferences and loan expansion can be seen as a
result of the loan agent's subjective preferences. In this case, it can be
expected that in for-profit MFIs, designing effective alignment me-
chanisms will help to mitigate the discretionary power of the loan of-
ficer and limits the effect of the loan officer managerial discretion on
the loan portfolio quality and the outreach of the MFI. We thus for-
mulate the following Hypothesis 3.
H3. The attenuating effect of effective alignment mechanisms will be
stronger for privately owned MFIs than for not-for-profit MFIs.
3. Research design
3.1. Model
The main question addressed in this study is whether incentive
schemes and internal control systems limit agency conflicts between the
MFI and the loan officer when the loan approval process is decen-
tralized. Incentive schemes and internal control systems will be effec-
tive if the riskiness of the loan portfolio decreases or remains the same
while outreach increases.
To address this research question, we first model MFI performance
(risk and outreach) as a cross-sectional function of the credit approval
decision, internal control effectiveness, and human resource manage-
ment effectiveness after controlling for some MFI-level variables,
country-level variables, year fixed effects and heteroscedasticity. The
estimated model is as follows:
MFI performance = α0 + αi Xi + βi Yi + γj ICVj + δt + ε, (1)
where i indexes MFIs, j indexes country and t indexes year. Xi is the
vector of our main MFI-specific variables: Loan officer approval,
Governance effectiveness, Internal control effectiveness, and the cross
products “Loan officer approval ∗ Governance effectiveness” and “Loan
officer approval ∗ Internal control effectiveness”; Yi is the vector of
other MFI-level variables: Profitability, Loan portfolio growth,
Individual lending, and the cross product “Loan officer
approval ∗ Individual lending,” and MFI ownership type. ICVj is the
vector of country-level variables: the Kaufmann, Kraay, and Mastruzzi
(2010) corruption index and GDP growth. δt is the year fixed effects
and ε is the error term.
To answer the question of whether the effects of governance effec-
tiveness and internal control effectiveness when MFIs decentralize the
loan approval process vary by MFI ownership type, we split the sample
into two subgroups (for-profit MFIs versus not-for-profit MFIs) and re-
estimated Eq. (1) for each subgroup.
3.2. Variables
3.2.1. Dependent variables
Our first dependent variable is MFI loan portfolio risk, which we
measure by three indicators. The main indicator is the Portfolio at risk
at 30 days, which measures the portion of the loan portfolio that is
30 days past due. For a robustness test, we use two alternative measures
to proxy loan portfolio risk, namely, Write-offs ratio and Loan loss
provision expenses.
We use three outreach indicators as dependent variables. The main
indicator is the Breadth of outreach, measured by the number of active
borrowers. The second outreach indicator is the Yield on loan portfolio
(real) which proxies the annualized interest rate charged on loans by
MFIs (Cozarenco, Hudon, & Szafarz, 2016; Cull, Demirgüç-Kunt, &
Morduch, 2007; D'Espallier et al., 2013). The third measure of outreach
is the Depth of outreach, which measures the outreach to the poor.
H. Tchakoute-Tchuigoua, I. Soumaré Journal of Business Research 94 (2019) 1–17

Table 1
Description of variables.
Identity of variable Definition and measurement

Outreach Breadth of outreach Natural logarithm of the number of active borrowers

Yield on loan portfolio
Depth of outreach
Loan portfolio quality indicators Portfolio at risk 30 days
Write-off ratio
Loan loss provision expense
ratio
Decentralization Loan officer approval
Credit committee approval
Governance ratings Governance effectiveness
Internal control ratings Internal control
effectiveness
Type of loan contract Individual lending
Profitability Return on assets
MFIs growth Loan portfolio growth
Ownership type For-profit MFIs
Country level variables Country corruption index
Country economic growth
Number of active borrowers (NAB)
If NAB > 30,000, high outreach;
If 10,000 ≤ NAB ≤ 30,000, medium outreach; if NAB < 10,000, low outreach
(Yield on Gross Portfolio (nominal) − Inflation Rate) / (1 + Inflation Rate)
Yield on Gross Portfolio = Interest and Fees on Loan Portfolio / Loan Portfolio
Average loan size per borrower scaled by the per capita gross national income (GNI).
(Outstanding Balance on Arrears over 30 days + Total Gross Outstanding Refinanced (restructured) Portfolio)/
Total Gross Portfolio
Measurement of portfolio quality. It shows the part of the portfolio affected by outstanding payments when
there is a risk that they might not be repaid. The threshold is < 10% given that financial guarantees in
microfinance are not always sufficient.
Write Offs/Loan Portfolio, gross, average
Total amount of loans written off during the period. A write-off is an accounting procedure that removes the
outstanding balance of the loan from the Loan Portfolio and from the Impairment Loss Allowance when these
loans are recognized as uncollectable.
Net loan loss provision expense/Average gross outstanding portfolio
Net loan loss provision expense = Loan loss provision expense and write-off minus Recovery from Loans written
off
Dummy: 1 if loans are approved by the loan officer, and zero otherwise (the credit decision is taken at the
branch committee level or at headquarter).
Dummy: 1 if loans are approved by the MFI credit committee, and zero otherwise.
Value comprised between 1 and 5; 5 being the highest (better governance).
Value comprised between 1 and 5; 5 being the highest (better internal control).
Individual loans as a percentage of the outstanding loan portfolio.
Loan is granted to a single borrower.
Net Operating Income/Assets, average
Relative change of the gross loan portfolio
Privately owned MFIs included microfinance banks and non-bank financial institutions (NBFI), Binary variable:
1 if the MFI is a privately owned MFI; 0 otherwise.
Control of corruption from the World Governance Indicators (WGI): The index reflects perceptions of the extent
to which public power is exercised for private gain, including both petty and grand forms of corruption, as well
as “capture” of the state by elites and private interests. Estimate of governance (ranges from approximately
−2.5 (weak) to 2.5 (strong) governance performance).
Source: Kaufmann et al. (2010); World Bank.
Annual growth rate of the real GDP per capita

Table 2
MFIs sample distribution by region and year.
This table provides the distribution of MFIs sample by region and by year.
Panel A: MFIs distribution by region

Region Number of MFIs % of the sample

Africa & Middle East 149 39.84

Europe & Central Asia 56 14.97
Latin America & Caribbean 141 37.70
South Asia, East Asia & the Pacific 28 7.49
Total 374 100

Panel B: MFIs distribution by year

Year Number of MFIs % of sample

2001 2 0.53
2002 11 2.94
2003 24 6.42
2004 28 7.49
2005 54 14.44
2006 60 16.04
2007 46 12.30
2008 36 9.63
2009 33 8.82
2010 37 9.89
2011 30 8.02
2012 13 3.48
Total 374 100.00

5
H. Tchakoute-Tchuigoua, I. Soumaré Journal of Business Research 94 (2019) 1–17

Table 3
Summary statistics.
This table displays summary statistics for the variables used in the regressions. We provide definitions of the variables in Table 1.
Number of observation Mean Standard deviation Minimum Maximum Median

Portfolio at risk 30 days 373 0.05 0.06 0.00 0.45 0.03

Loan Loss provision expense 367 0.02 0.04 −0.19 0.34 0.01
Write-off ratio 326 0.03 0.06 0.00 0.56 0.01
Breadth of outreach 374 21,983 44,882 23.00 477,267 10,284
Depth of outreach 322 0.75 1.28 0.00 9.52 0.32
Yield on loan portfolio 343 0.38 0.19 0.02 1.10 0.33
Loan officer approval 374 0.52 0.50 0.00 1.00 1.00
Credit committee approval 374 0.28 0.45 0.00 1.00 0.00
Governance effectiveness 374 3.43 0.82 1.00 5.00 3.00
Internal control effectiveness 374 3.37 1.00 1.00 5.00 3.00
Individual lending 371 0.61 0.41 0.00 1.00 0.84
Profitability: return on assets 374 0.03 0.09 −0.68 0.52 0.03
Loan portfolio growth 334 0.50 0.97 −0.74 15.94 0.35
For profit MFIs 374 0.38 0.49 0.00 1.00 0.00
Corruption index 373 −0.46 0.43 −1.51 1.47 −0.46
Annual GDP growth 373 0.06 0.03 −0.05 0.17 0.05

3.2.2. Explanatory variables
Loan officer approval: When the allocation decision is decentralized,
MFIs face at least two agency conflicts: (1) one between the MFI and the
loan officer when he or she approves loans and (2) another one between
headquarters and the branch when loans are approved at the branch
level.
We initially focused on the first agency conflict, the one between the
MFI and the loan officer. To build the variable Loan officer approval,
we conduct an exploratory study of rating reports to answer the ques-
tion of whether the loan officer has decision-making authority or not.
The exploratory analysis of rating reports enabled us to identify three
situations.
In the first situation, the loan officer has the authority to make the
final credit approval decision. Resource allocation and information
production are associated. Some cases, such as Benefit and Mikrofin in
Bosnia & Herzegovina, Miselini in Mali, JV MFO Microinvest in
Moldova and MFO Alliance in Georgia, illustrate this situation. For
example, in the case of JV MFO Microinvest in Moldova, the loan officer
approves loans up to US$7000. The loan approval process in this case is
fully decentralized and the loan officer is the decision maker.
In the second situation, the branch-level credit committee that
comprises the branch manager and loan officers approves loans. Loan
officers who act as an interface between the credit committee and
borrowers produce information, and the final authority to approve
loans rests with the credit committee. By participating in the credit
committee at the branch level, it is more likely that the loan officer
influences the credit decision. The loan approval in this case is semi-
decentralized, and the loan officer is associated with the loan approval
decision. The case of JV MFO Microinvest in Moldova illustrates this
situation. The branch credit committee approves loan of between US
$7000 and US$30,000.
In the third situation, the headquarters approves loans. The loan
approval decision is centralized, that is, all loan applications are first
completed by loan officers, reviewed by the branch manager, and then
sent to the relevant officer at headquarters toward credit committee
approval. Nor Horizon in Armenia, Nachala in Bulgaria, Narodnyi
Kredit in Russia and LEAD in Egypt are illustrative cases of this third
situation.
We code the variable Loan officer approval as 1 if the loan officer
both uses information production and also approves credit, and 0
otherwise. This corresponds to the first case described.
As a robustness test for the decentralization of the loan approval
decision, we use the alternative measure Branch-level credit committee
approval stemming from the second situation described above, where
the loan officer is associated with the loan approval decision taken at
the branch level. In this case, the variable Branch-credit committee
approval takes a value of 1 if the loan approval decision is made by the
branch credit committee and 0 otherwise.
3.2.2.1. Internal control effectiveness. We proxy the internal control
system effectiveness using the aggregate scores provided by Planet
Rating. Indeed, Planet Rating uses a five-point alphabetical rating scale
(from e to a) to assess each of the main areas of rating: governance,
information, risk management, activity, funding and liquidity, and
efficiency. We thus measure the effectiveness of MFI internal control
systems using the risk management ratings provided by Planet Rating.
The risk management ratings cover three main areas: risk management
systems, internal controls and procedures, and internal audits. We
match the ratings (e to a) with values between 1 and 5, with the value
of 1 corresponding to grade e, 2 for grade d, 3 for grade c, 4 for grade b,
and 5 for grade a. The higher the value the better the perceived
effectiveness of the MFI's risk management and internal control system.
3.2.2.2. Governance effectiveness. Here also, we use Planet Rating
governance effectiveness scores as a proxy for the effectiveness of
human resource management. The corporate governance score
provided by Planet Rating, in addition to the effectiveness of the
governance structure and decision-making processes, captures the
relevance of the strategy, the effectiveness of the management team
and the effectiveness of human resource management practices
(notably the recruitment process, training of loan officers and
existence of a well-designed and motivating remuneration scheme).
The governance effectiveness score is measured on the same scale as the
internal control system and takes values between 1 and 5, with 5 being
the highest value (better governance).
3.2.3. Other control variables
MFIs use either individual loan contracts or joint liability contracts,
which consist of group solidarity lending and village banking lending.
Following previous studies that link the choice of joint liability contract
to MFI social efficiency and loan portfolio quality (Cull et al., 2007;
Mersland & Strøm, 2009; Tchakoute Tchuigoua, 2012), we account for
the type of lending contract applied by the MFI. We include Individual
lending as an indicator of MFI lending methodology, and the cross
product “Individual lending ∗ Loan officer approval” to assess whether
loan decentralization influences the effect of the lending methodology
on MFIs' risk and outreach. We measure Individual lending as average
loan size approved under individual loan contracts as a percentage of
the overall gross loan portfolio. Because of the lack and insufficiency of
data, we were unable to account for the loans granted by product type
(consumption loans, micro-enterprise loans) or by industry (agri-
cultural loans, services loans, trade loans and manufacturing loans).

6
H. Tchakoute-Tchuigoua, I. Soumaré Journal of Business Research 94 (2019) 1–17

Ownership type: We measure Ownership type as a dummy that

takes a value of 1 if the MFI is profit-oriented (for-profit) and zero
Annual GDP
growth
otherwise. Profit-oriented MFIs include microfinance banks and non-
bank financial institutions (NBFI) (Tchakoute Tchuigoua, 2010).

1.00
We measure Profitability as the return on assets (ROA) and include
Loan portfolio growth to account for the growth of the MFI.
Corruption index

Country-specific variables: Based on previous country case studies

(Patten, Rosengard, & Johnston Jr., 2001) and existing empirical papers

−0.14⁎⁎
such as Ahlin, Lin, and Maio (2011), who associated country macro-

1.00
economic conditions with MFI social performance and efficiency, we
control for the country's economic growth or economic performance
(GDP growth measured as the growth rate of the real per capita GDP).
For profit

We also control for the institutional framework using the Country

MFIs

1.00
0.04
0.01
corruption index.
Table 1 summarizes and describes the variables used in the study.
Loan portfolio

3.3. Sample
growth

0.16⁎⁎⁎
−0.06
1.00
0.02

Data used in this study comes from two different sources. Our main
source of data is Planet Rating's website (www.planetrating.org).
Profitability

Organizational structure data, such as governance, internal control and

This table presents the Pearson matrix of the explanatory variables used in the regressions. We provide definitions of the variables in Table 1.

loan approval decisions, and also other MFIs level variables, come from
−0.04
0.11⁎⁎
1.00
0.02

0.03

assessment reports produced by Planet Rating, a rating agency specia-

lized in rating microfinance institutions. We chose Planet Rating data
for at least two reasons. First, alternative data sources such as the
Microfinance Information Exchange (MIX) database do not enable us to
Individual

gather core and useful data to conduct this kind of study on the choice
lending

0.15⁎⁎⁎

−0.03

of decentralized versus centralized loan approval decisions, the effec-

1.00
0.01
0.02

0.04

tiveness of internal MFI control systems, and the effectiveness of MFI

governance practices. Planet Rating reports provide detailed informa-
tion about the loan approval process and MFI organizational structure.
Internal control system

Second, among rating agencies specialized in MFI ratings, Planet Rating

seems to be the only rater that assigns scores to each of the areas as-
sessed. Indeed, Planet Rating, through its GIRAFE (governance, in-
effectiveness

formation, risk management, activity, funding and liquidity, efficiency)

0.28⁎⁎⁎
0.16⁎⁎⁎
−0.02

−0.08

methodology rates and assigns scores to internal MFI control systems

0.11⁎⁎
0.10⁎⁎
1.00

(risk management) and governance practice effectiveness (including

human resources management). Rating reports for rated MFIs between
2003 and 2015 are available on the Planet Rating website.3 During the
2003–2015 period, we count about 435 rating reports available on the
Planet Rating website. However, we were unable to extract some of the
effectiveness
Governance

reports because they remained confidential and were not freely avail-
0.69⁎⁎⁎

⁎⁎⁎

0.14⁎⁎⁎

0.25⁎⁎⁎
0.15⁎⁎⁎
−0.05

−0.07

able, specifically, recent rating reports published between 2013 and

1.00

0.14

2015. Additionally, some other rating reports by Planet Rating were not
available on the website of Planet Rating. The Rating Fund website
(www.ratingfund2.org) enables us to exploit assessment reports pro-
Credit committee

duced by Planet Rating but unavailable on the Planet Rating website,

such as rating reports during the period 2001–2003, and thus comple-
−0.15⁎⁎⁎
approval

ments data gathered from the Planet Rating website. We exclude rating
−0.04
0.09⁎

0.10⁎
1.00
0.04

0.00
0.05
0.09

reports in which precise information about the loan approval process

was missing. We end up with a final sample comprised of 374 assess-
ment reports for 280 MFIs in 70 countries from 2001 to 2012.
Country-level data come from the World Bank's World Development
Loan officer

Indicators (WDI)4 and Worldwide Governance Indicators (WGI)5 data-

−0.66⁎⁎⁎
approval

0.14⁎⁎⁎
−0.02

−0.08
−0.03
−0.04

−0.04

bases.
1.00

0.02

0.01

Table 2 presents the distributions of the sample by region (panel A)

and by year (panel B). As panel A of the table shows, the sample in-
cludes MFIs from the following four main regions: Africa and the
Credit committee approval
Governance effectiveness

Middle East (149 rating reports); Eastern Europe and Central Asia (56
Internal control system

Loan portfolio growth

Loan officer approval

Annual GDP growth

rating reports); Latin America and the Caribbean (141 rating reports);
Correlation matrix.

Individual lending

Corruption index
effectiveness

South Asia, East Asia, and the Pacific (28 rating reports).
p < 0.01.
p < 0.05.
For profit MFIs

p < 0.10.
profitability

3
Table 4

www.planetrating.com/FR/rating-girafe.html.
4
⁎⁎⁎

http://data.worldbank.org/data-catalog/world-development-indicators.
⁎⁎
⁎

5
http://info.worldbank.org/governance/wgi/index.asp.

7
H. Tchakoute-Tchuigoua, I. Soumaré Journal of Business Research 94 (2019) 1–17

Table 5a
Baseline results.
This table reports the results of the pooled OLS with risk measures as dependent variables. Standard errors are heteroskedastic robust. We control for year fixed
effects. All independent variables are defined in Table 1. We report estimated coefficients in the first row.
Governance effectiveness Internal control system effectiveness

Portfolio at risk at Write-offs ratio Loan loss provision Portfolio at risk at Write-offs ratio Loan loss provision
30 days 30 days

Loan officer approval −0.0762⁎⁎ −0.0717⁎⁎ −0.00345 −0.0455⁎ −0.0392 −0.00160

(−2.041) (−2.493) (−0.168) (−1.693) (−1.498) (−0.0756)
Governance effectiveness −0.0315⁎⁎⁎ −0.0188⁎⁎⁎ −0.00191
(−3.671) (−2.973) (−0.378)
Loan officer approval ∗ governance effectiveness 0.0201⁎⁎ 0.0200⁎⁎⁎ −0.000168
(2.062) (2.755) (−0.0308)
Internal control system effectiveness −0.0210⁎⁎⁎ −0.0123⁎⁎⁎ −0.00715
(−3.433) (−2.909) (−1.633)
Loan officer approval ∗ internal control system 0.0110 0.0107 −0.000808
effectiveness (1.599) (1.610) (−0.147)
Individual lending 0.0181⁎⁎⁎ −0.00842 −0.000416 0.0168⁎⁎ −0.00899 0.000516
(2.743) (−0.923) (−0.0861) (2.564) (−0.954) (0.112)
Loan officer approval ∗ individual lending 0.0000986 −0.000374 −0.0000231 0.000206 −0.000354 −0.0000129
(0.405) (−1.629) (−0.103) (0.925) (−1.509) (−0.0604)
Profitability −0.180⁎⁎⁎ −0.166⁎⁎⁎ −0.144⁎⁎⁎ −0.182⁎⁎⁎ −0.165⁎⁎⁎ −0.139⁎⁎⁎
(−5.604) (−3.504) (−3.102) (−5.325) (−3.466) (−3.086)
Loan portfolio growth 0.00513 −0.00155 0.00494⁎⁎⁎ 0.00585 −0.00110 0.00494⁎⁎⁎
(1.099) (−0.896) (3.356) (1.055) (−0.585) (3.891)
For profit MFIs −0.000993 0.00162 0.00848⁎ −0.000181 0.00210 0.0117⁎⁎
(−0.169) (0.260) (1.752) (−0.0305) (0.329) (2.399)
Corruption index 0.0118 0.00730 0.00748 0.0111 0.00789 0.00891⁎
(1.555) (1.057) (1.524) (1.534) (1.194) (1.801)
Annual GDP growth −0.115 −0.115 −0.110 −0.138 −0.134 −0.123
(−1.177) (−0.759) (−1.143) (−1.419) (−0.865) (−1.302)
_cons 0.142⁎⁎⁎ 0.0893⁎⁎⁎ 0.0245 0.105⁎⁎⁎ 0.0720⁎⁎⁎ 0.0504⁎⁎
(3.928) (3.168) (1.131) (3.691) (3.304) (2.495)
Year fixed effects Yes Yes Yes Yes Yes Yes
Number of observations 330 319 328 330 319 328
R-sq 0.241 0.199 0.173 0.224 0.193 0.202
Adj. R-sq 0.189 0.142 0.117 0.171 0.136 0.147
F 10.55 . 10.25 10.84 . 5.069

t-Statistics are in parentheses.

⁎
p < 0.10.
⁎⁎
p < 0.05.
⁎⁎⁎
p < 0.01.

4. Empirical results and discussion 4.2. Multivariate analysis

4.1. Summary statistics
Summary statistics presented in Table 3 show that the average value
of loans over 30 days past due is 0.05, below the 0.1 threshold (Bruett,
2005). We can conclude that our sample loan portfolio is healthy. In
52% of the cases, the loan officer approves loans. Our sample average
governance effectiveness exhibits a value of 3.43, indicating that the
MFIs in our sample obtain a governance rating grade between c and b.
The average value of internal control system ratings is 3.37, indicating
that the average sample grade is also between c and b. On average,
outstanding individual loans account for 61% of loans granted by MFIs.
Privately owned MFIs (NBFIs and microfinance banks) represent 38% of
the observations.
Prior to estimations, we assessed the presence of multicollinearity
among our explanatory variables by computing the correlation between
our variables (Table 4). Our diagnostic reveals that, except for the
correlation between ratings of governance effectiveness and the in-
ternal control effectiveness (r = 0.69; p < 0.01) and the correlation
between loan officer approval and branch credit committee approval
(r = −0.66; p < 0.01), the intensity of the relationship among the
other explanatory variables is relatively weak. To avoid multi-
collinearity, we include the indicators of governance effectiveness and
internal control effectiveness separately in the regressions.
The distribution of MFIs per year, given in panel B of Table 2, shows
that the data structure looks like a panel in that the data is observed
over 12 years (from 2001 to 2012). However, the sample size is not the
same from one year to another, and the data contains different statis-
tical units at different points in time. The data structure is thus an in-
dependently pooled cross-section (IPCS). No MFI renewed its rating 12
times during the sample period. Among the 280 MFIs in our sample,
only two MFIs renewed their ratings four times; five MFIs renewed
three times; nine MFIs renewed twice; and finally only 53 MFIs renewed
once. Therefore, we use the pooled sample robust OLS estimation
technique with controls for year effects as suggested by Wooldridge
(2010).
Tables 5a and 5b report the results of the pooled OLS with robust
standard errors and control for year fixed effects for the whole sample.
We assess whether governance effectiveness and a better internal con-
trol system mitigate agency problems within an MFI, especially when
the MFI decentralizes the loan approval process. We assess the effect of
these mitigating instruments on risk (portfolio at risk 30 days) and
outreach (breadth of outreach). The table also presents results for the
alternative risk measures (write-off ratio and loan loss provision) and
the alternative outreach measures (yield on loan portfolio and depth of
outreach).
The results show that allowing the loan officer to have decision-
making authority has a significant reducing effect on portfolio risk, and

8
H. Tchakoute-Tchuigoua, I. Soumaré Journal of Business Research 94 (2019) 1–17

Table 5b
Baseline results.
This table reports the results of the pooled OLS with outreach measures as dependent variables. Standard errors are heteroskedastic robust. We control for year
fixed effects. All independent variables are defined in Table 1. We report estimated coefficients in the first row.
Governance effectiveness Internal control system effectiveness

Breadth of Yield on loan Depth of Breadth of outreach Yield on loan Depth of outreach
outreach portfolio outreach portfolio

Loan officer approval 2.978⁎⁎⁎ −0.186⁎⁎ 0.508 1.985⁎⁎⁎ −0.112 0.176

(4.626) (−2.077) (0.655) (3.876) (−1.537) (0.320)
Governance effectiveness 1.024⁎⁎⁎ −0.0282 −0.306⁎⁎
(7.251) (−1.423) (−2.002)
Loan officer approval ∗ governance effectiveness −0.656⁎⁎⁎ 0.0506⁎⁎ −0.102
(−3.700) (2.109) (−0.515)
Internal control system effectiveness 0.736⁎⁎⁎ −0.0202 −0.219⁎⁎
(6.163) (−1.315) (−2.103)
Loan officer approval ∗ internal control system −0.362⁎⁎ 0.0300 −0.0153
effectiveness (−2.517) (1.550) (−0.109)
Individual lending −1.082⁎⁎⁎ −0.113⁎⁎⁎ 1.181⁎⁎⁎ −1.049⁎⁎⁎ −0.113⁎⁎⁎ 1.156⁎⁎⁎
(−6.134) (−4.178) (6.740) (−5.949) (−4.163) (6.587)
Loan officer approval ∗ individual lending 0.00776 0.00145⁎ −0.00519 0.00425 0.00138⁎ −0.00251
(1.206) (1.872) (−1.613) (0.729) (1.779) (−0.889)
Profitability 1.039 −0.0492 −0.355 1.047 −0.0438 −0.564
(1.489) (−0.311) (−0.410) (1.518) (−0.276) (−0.580)
Loan portfolio growth −0.133⁎⁎⁎ 0.0379⁎⁎⁎ −0.0373 −0.156⁎⁎⁎ 0.0387⁎⁎⁎ −0.0308
(−3.116) (3.214) (−1.127) (−2.839) (3.524) (−0.888)
For profit MFIs 0.0542 0.0282 0.178 −0.000421 0.0296 0.168
(0.356) (1.352) (1.318) (−0.00270) (1.391) (1.243)
Corruption index −0.275 0.0260 −0.218 −0.270 0.0299 −0.266
(−1.631) (1.383) (−1.045) (−1.490) (1.582) (−1.273)
Annual GDP growth −4.825⁎ −0.772⁎⁎ −1.637 −3.960 −0.806⁎⁎ −1.395
(−1.669) (−1.986) (−0.587) (−1.374) (−2.077) (−0.499)
_cons 6.172⁎⁎⁎ 0.598⁎⁎⁎ 2.389⁎⁎⁎ 7.116⁎⁎⁎ 0.584⁎⁎⁎ 1.828⁎⁎⁎
(9.593) (6.662) (3.945) (12.74) (7.888) (3.985)
Year fixed effects Yes Yes Yes Yes Yes Yes
N 331 329 309 331 329 309
R-sq 0.303 0.192 0.246 0.299 0.187 0.224
Adj. R-sq 0.256 0.137 0.191 0.251 0.132 0.167
F 12.09 7.961 . 9.765 8.254 .

t-Statistics are in parentheses.

⁎
p < 0.10.
⁎⁎
p < 0.05.
⁎⁎⁎
p < 0.01.

thus has a positive effect on the quality of the loan portfolio. However,
this effect is only observed for two of those measures (portfolio at risk
30 days and write-offs ratio). The effect on the MFIs' outreach is positive
and significant. Indeed, the relationship between decentralization of
loan approval and the breadth of outreach is positive and significant in
the governance and internal control models, suggesting that access to
credit is greater in MFIs when the loan officer approves loans. This is
confirmed when we use the alternative outreach measure, the yield on
loan portfolio. This finding is consistent with Hypothesis 1a and seems
to support the assumption that proximity improves financial inclusion
and access to loan (Allen et al., 2014; Allen et al., 2016; Brown et al.,
2016). In addition, the effect on the quality of the portfolio is positive
for the two risk measures portfolio at risk 30 days and write-offs, and
non-significant for the risk measure loan loss provision. Therefore, the
quality of the loan portfolio is either not affected or improves, thus
supporting Hypothesis 1.b. This result suggests that decentralizing loan
approval decisions or allocating decision-making authority to the loan
officer increases the outreach but does not necessarily alter the loan
portfolio quality, especially when there are better governance me-
chanisms and internal control systems in place.
The deterioration of the loan portfolio quality that may have re-
sulted from credit expansion to more borrowers, in our sample, is mi-
tigated by the effectiveness of the governance mechanism and the ef-
fectiveness of the internal control system. These findings support
Aubert et al. (2009) and Labie et al. (2015), who show that setting up
these incentives may enable loan officers to select borrowers
adequately.
The implementation of good governance (better human resource
management) practices and internal control systems, perceived as ef-
fective, improves the social performance of the MFIs (outreach) while
mitigating portfolio quality deterioration. Analyzing the cross product
of the loan approval variable and governance or internal control in-
dicators, our results show that better governance quality and internal
control systems improve MFI efficiency in selecting borrowers, and
hence improve outreach. Indeed, we observe that better governance
and internal control systems have an attenuation effect on credit ex-
pansion when the loan officer approves loans, which explains why the
effect on risk is non-significant or reducing. This result suggests that the
implementation of appropriate performance pay, audits, and incentive
schemes constrains the loan officer's behavior, limiting the risk of se-
lecting bad borrowers, thereby improving the quality of screening. This
evidence thus supports Hypothesis 2. Providing the loan officer with
incentives such as performance-based pay and putting in place an ef-
fective internal control system may contribute to aligning the interest of
the institution with the loan officer's interest when loan approval de-
cisions are decentralized. This result is in line with the governance
literature (Berger & Udell, 2002; Stein, 2002) and BCBS (2010) and
Jeon and Menicucci (2011).
For further analysis and to test Hypothesis 3, we divide our sample
by ownership type to explore whether the effects differ depending on
the commercial orientation of the MFIs. Tables 6a and 6b present the
regression results by ownership type (profit-oriented MFIs versus not-

9
H. Tchakoute-Tchuigoua, I. Soumaré Journal of Business Research 94 (2019) 1–17

Table 6a Table 6b
Ownership type results. Ownership type results.
This table reports the results of the pooled OLS with the loan portfolio at risk This table reports the results of the pooled OLS with Breadth of outreach as
at 30 days measures as dependent variable. Standard errors are heteroskedastic dependent variable. Standard errors are heteroskedastic robust. We control for
robust. We control for year fixed effects. All independent variables are defined year fixed effects. All independent variables are defined in Table 1. We report
in Table 1. We report estimated coefficients in the first row. estimated coefficients in the first row.
For profit MFIs Not for profit MFIs For profit MFIs Not for profit MFIs

Loan officer approval −0.0109 −0.0201 −0.116⁎⁎ −0.0566⁎ Loan officer approval 3.685⁎⁎⁎ 0.452 2.062⁎⁎ 2.266⁎⁎⁎
(−0.179) (−0.305) (−2.255) (−1.722) (3.855) (0.451) (2.430) (3.580)
Governance effectiveness −0.0148 −0.0395⁎⁎⁎ Governance effectiveness 1.066⁎⁎⁎ 0.988⁎⁎⁎
(−1.026) (−3.603) (4.556) (5.305)
Loan officer 0.00512 0.0345⁎⁎ Loan officer −1.027⁎⁎⁎ −0.392
approval ∗ governance (0.305) (2.461) approval ∗ governance (−3.738) (−1.600)
effectiveness effectiveness
Internal control system −0.0224 −0.0208⁎⁎⁎ Internal control system 0.340 0.850⁎⁎⁎
effectiveness (−1.393) (−3.018) effectiveness (1.596) (6.050)
Loan officer 0.00766 0.0170⁎ Loan officer −0.0788 −0.464⁎⁎
approval ∗ internal (0.430) (1.957) approval ∗ internal (−0.287) (−2.494)
control system control system
effectiveness effectiveness
Individual lending 0.0211 0.0245⁎ 0.0263⁎⁎⁎ 0.0243⁎⁎⁎ Individual lending −1.395⁎⁎⁎ −1.235⁎⁎⁎ −1.132⁎⁎⁎ −1.140⁎⁎⁎
(1.494) (1.683) (3.080) (2.867) (−3.182) (−2.722) (−5.055) (−5.096)
Loan officer −0.0301 −0.0338 0.000306⁎ 0.000450⁎⁎⁎ Loan officer 0.974⁎ 0.636 0.00513 −0.00120
approval ∗ individual (−1.168) (−1.362) (1.687) (2.627) approval ∗ individual (1.682) (1.102) (0.697) (−0.179)
lending lending
Profitability −0.150⁎⁎ −0.155⁎⁎ −0.209⁎⁎⁎ −0.216⁎⁎⁎ Profitability 0.823 0.128 1.234 1.238
(−2.617) (−2.345) (−4.706) (−4.657) (0.857) (0.117) (1.091) (1.177)
Loan portfolio growth −0.00920 −0.0105⁎ 0.00740⁎⁎ 0.00862⁎⁎ Loan portfolio growth −0.0531 −0.0422 −0.134⁎⁎⁎ −0.172⁎⁎⁎
(−1.549) (−1.795) (2.355) (2.176) (−0.235) (−0.175) (−2.997) (−4.415)
Corruption index 0.0114 0.0111 0.0150⁎ 0.0171⁎ Corruption index −0.288 −0.283 −0.224 −0.314
(0.961) (0.961) (1.688) (1.786) (−0.939) (−0.807) (−0.963) (−1.248)
Annual GDP growth −0.143 −0.161 −0.0487 −0.0707 Annual GDP growth −2.240 −0.730 −9.587⁎ −9.097⁎
(−1.224) (−1.430) (−0.308) (−0.443) (−0.556) (−0.175) (−1.953) (−1.949)
_cons 0.149⁎⁎ 0.196⁎⁎⁎ 0.151⁎⁎⁎ 0.0869⁎⁎⁎ _cons 5.707⁎⁎⁎ 7.894⁎⁎⁎ 6.307⁎⁎⁎ 6.997⁎⁎⁎
(2.272) (2.673) (3.319) (2.686) (4.952) (6.948) (8.061) (9.903)
Year fixed effects Yes Yes Yes Yes Year fixed effects Yes Yes Yes Yes
N 129 129 201 201 N 129 129 202 202
R-sq 0.252 0.295 0.311 0.270 R-sq 0.291 0.205 0.386 0.419
Adj. R-sq 0.121 0.172 0.235 0.189 Adj. R-sq 0.168 0.066 0.318 0.355
F . 17.74 15.68 F . . 14.04 14.40

t-Statistics are in parentheses. t-Statistics are in parentheses.

⁎ ⁎
p < 0.10. p < 0.10.
⁎⁎ ⁎⁎
p < 0.05. p < 0.05.
⁎⁎⁎ ⁎⁎⁎
p < 0.01. p < 0.01.

for-profit MFIs). This table reports the results of the pooled OLS with
robust standard errors and controls for year fixed effects. We assess
whether the effect of better internal control systems and better human
resource management on MFI performance – in terms of loan portfolio
quality (panel A) and outreach (panel B) – differ by ownership type
when the loan approval process is decentralized.
The results in this table show that loan approval decentralization
has a positive effect on the outreach of not-for-profit as well as profit-
oriented MFIs (panel B). Moreover, as in the full sample case, having
effective governance mechanisms and internal control systems in place
has an attenuating effect on a powerful loan officer's impact on MFI
outreach in both subgroups (profit-oriented MFIs and not-for-profit
MFIs). However, the interaction effect between the loan officer ap-
proval and the governance effectiveness is significant only among
profit-oriented MFIs, whereas the cross product loan officer
approval ∗ Internal control effectiveness is significant only among not-
for-profit oriented organizations.
As for the loan portfolio quality (panel A), we find significant dif-
ferences across MFI ownership types with respect to the direct impact of
loan approval decentralization. Indeed, loan approval decentralization
reduces risk only among the not-for-profit MFIs, whereas its coefficient
is non-significant in the profit-oriented subgroup. Effective governance
and internal control systems significantly improve not-for-profit MFIs'
loan portfolio quality; the coefficients are not significant in the for-
profit MFIs subgroup. Thus, in nonprofit-oriented MFIs and profit-
oriented MFIs, although loan approval decentralization increases MFI
outreach, it does not alter the loan portfolio quality. The existence of
effective governance mechanisms and internal control systems even
contributes to significantly reducing loan portfolio risk in not-for-profit
microfinance organizations. Hence, with respect to our third hypothesis
(H3), we find the effectiveness of incentive controls and governance
mechanisms to positively impact both nonprofit MFIs and for-profit
MFIs, with no specific attenuating intensity for profit-oriented MFIs as
posit by our hypothesis.
The other control variables have the expected signs. For instance,
individual lending, as opposed to joint liability contracts, reduces the
outreach and the portfolio quality, and more so among not-for-profit
MFIs. The presence of a loan officer with loan authorization power
tends to slightly reduce (10% significance level) the negative impact of
individual loan contracts on outreach in for-profit oriented MFIs.
Therefore, joint liability contracts can be effective as risk management
tools in MFIs. Profitability, measured by the return on assets (ROA),
significantly reduces the loan portfolio risk and has no significant im-
pact on outreach. High loan portfolio growth reduces portfolio quality
and outreach in not-for-profit MFIs. High levels of corruption dis-
courage the MFI from allocating credit to more borrowers and are
detrimental to portfolio quality, especially among not-for-profit or-
iented MFIs. Finally, GDP growth improves the loan portfolio quality,
and corruption worsens it.

10
H. Tchakoute-Tchuigoua, I. Soumaré Journal of Business Research 94 (2019) 1–17

Table 7a
Robustness checks.
Analyzing the crisis effect.
This table reports the results of the pooled OLS with portfolio at risk at 30 days as dependent variable. Standard errors are heteroskedastic robust. We control for
year fixed effects. All independent variables are defined in Table 1. We report estimated coefficients in the first row.
Pre-crisis period Crisis period Post-crisis period Pre-crisis period Crisis period Post-crisis period
Y01–Y06 Y07–Y09 Y10–Y12 Y01–Y06 Y07–Y09 Y10–Y12

Loan officer approval −0.0786 0.0590 −0.214⁎⁎ −0.0540 −0.0189 −0.106

(−1.443) (1.177) (−2.288) (−1.343) (−0.323) (−1.511)
Governance effectiveness −0.0356⁎⁎⁎ -0.00593 −0.0543⁎⁎
(−2.644) (−0.570) (−2.151)
Loan officer approval ∗ governance effectiveness 0.0209 −0.0172 0.0553⁎⁎
(1.491) (−1.359) (2.119)
Internal control system effectiveness −0.0189⁎⁎ −0.0238 −0.0391⁎
(−2.257) (−1.572) (−1.788)
Loan officer approval ∗ internal control system effectiveness 0.0126 0.00409 0.0267
(1.367) (0.264) (1.292)
Individual lending 0.0269⁎⁎ 0.00605 0.0360 0.0205 0.00812 0.0338
(2.083) (0.514) (1.227) (1.419) (0.683) (1.136)
Loan officer approval ∗ individual lending −0.00394 0.000529⁎⁎⁎ 0.00722 −0.00179 0.000642⁎⁎⁎ 0.00532
(−0.236) (2.853) (0.233) (−0.100) (4.127) (0.172)
Profitability −0.145⁎⁎⁎ −0.188⁎⁎⁎ −0.0872⁎⁎ −0.162⁎⁎⁎ −0.190⁎⁎ −0.0993⁎⁎
(−3.403) (−2.786) (−2.233) (−3.518) (−2.517) (−2.444)
Loan portfolio growth 0.00722⁎⁎⁎ −0.0177 −0.0167 0.00838⁎⁎⁎ −0.0158⁎ −0.0339
(3.064) (−1.614) (−1.241) (2.911) (−1.666) (−1.603)
For profit MFIs −0.0106⁎ 0.0170 −0.00293 −0.0136⁎⁎ 0.0180 −0.000223
(−1.660) (1.330) (−0.253) (−2.149) (1.477) (−0.0194)
Corruption index 0.0171 0.0201⁎ −0.00454 0.0129 0.0159 −0.00488
(1.256) (1.663) (−0.367) (1.018) (1.344) (−0.389)
Annual GDP growth 0.127 −0.181 −0.0347 0.0622 −0.194 −0.0779
(0.859) (−1.289) (−0.170) (0.395) (−1.372) (−0.346)
_cons 0.157⁎⁎⁎ 0.0840⁎ 0.236⁎⁎ 0.109⁎⁎ 0.140⁎⁎ 0.184⁎⁎⁎
(2.699) (1.838) (2.643) (2.403) (2.374) (2.671)
N 154 99 76 154 99 76
R-sq 0.288 0.253 0.294 0.238 0.286 0.224
Adj. R-sq 0.238 0.168 0.186 0.184 0.205 0.104
F 6.438 4.978 4.379 5.315 6.147 5.486

t-Statistics are in parentheses.

⁎
p < 0.10.
⁎⁎
p < 0.05.
⁎⁎⁎
p < 0.01.

4.3. Robustness checks
4.3.1. Use of alternative portfolio quality and outreach measures
To strengthen the validity and robustness of our results, we con-
ducted a number of robustness analyses. First, we use two alternative
loan portfolio quality or risk measures in addition to our main risk
measure, portfolio at risk 30 days, namely write-offs ratio and loan loss
provisions. Using these two additional risk measures to run our re-
gressions does not contradict our main findings mentioned above and
shown in Tables 5a and 5b. Indeed, we find similar signs for the coef-
ficients of the write-offs ratio measure as in the case of our main risk
measure, portfolio at risk 30 days. With the loan loss provisions vari-
able, although the coefficients of interest are not significant, the find-
ings do not contradict our results.
As alternative measures for outreach, we use the yield on loan
portfolio, measured as the total interest and fees on the gross loan
portfolio, and the depth of outreach, obtained as the average loan size
per borrower scaled by the per capita gross national income (GNI). The
results obtained and presented in Tables 5a and 5b are consistent with
previous findings. In sum, granting loan approval authority to the loan
officer combined with effective governance mechanisms and internal
control systems increases MFI outreach without jeopardizing loan
portfolio quality.
4.3.2. Controlling for the 2007–2009 crisis period effect
Second, to control for the potential effect of the 2007–2009 crisis
period, we split the sample into three subsamples: the pre-crisis period
(2001–2006), the crisis period 2007–2009 and the post-crisis period
(2010−2012). The results provided in Tables 7a and 7b show that there
are differing results depending on the sub-period considered. Indeed,
the results for the pre-crisis period and the post-crisis period are con-
sistent with our findings above for the whole sample period. However,
over the crisis period, effective governance mechanisms and internal
control systems had positive impacts on outreach, and no significant
effect on portfolio risk. Thus, overall, our results are robust even after
controlling for the crisis effects, since the results show that loan officer
approval and effective governance mechanisms and internal control
systems increase outreach without reducing portfolio quality. We were
expecting this, because no special behaviors were observed in terms of
the coefficients of the year dummies for the crisis years in the full
sample regressions presented in Tables 5a and 5b above.
4.3.3. Controlling for selection bias and endogeneity
Third, we control for the endogeneity of governance and internal
control ratings, and of the loan approval decentralization. Indeed, only
using data on rated MFIs in our study can create potential selection
bias, in that ratings are only assigned if the MFI decides to be rated. The
same holds for loan approval decentralization, as this can be a self-
selection phenomenon. Hence one may argue that only MFIs that are
confident about the indicators covered in the rating will go for a rating;
and, those only feeling strongly about governance mechanisms and
internal control will implement them effectively. Therefore, pooled OLS
estimations may yield biased and inconsistent results.
To address the selection bias and endogeneity issues, we implement
the two-step procedure developed by Heckman (1979). In the first step
of this procedure, we estimate two probit selection models (a rating

11
H. Tchakoute-Tchuigoua, I. Soumaré Journal of Business Research 94 (2019) 1–17

Table 7b
Analyzing the crisis effect.
This table reports the results of the pooled OLS with Breadth of outreach as dependent variable. Standard errors are heteroskedastic robust. We control for year
fixed effects. All independent variables are defined in Table 1. We report estimated coefficients in the first row.
Pre-crisis period Crisis period Post-crisis period Pre-crisis period Crisis period Post-crisis period
Y01–Y06 Y07–Y09 Y10–Y12 Y01–Y06 Y07–Y09 Y10–Y12

Loan officer approval 2.886⁎⁎⁎ 1.746 3.449⁎ 1.857⁎⁎⁎ 1.264 2.606

(3.719) (1.440) (1.818) (2.682) (1.228) (1.616)
Governance effectiveness 1.102⁎⁎⁎ 0.816⁎⁎⁎ 1.198⁎⁎⁎
(5.944) (2.645) (3.934)
Loan officer approval ∗ governance effectiveness −0.671⁎⁎⁎ −0.342 −1.004⁎⁎
(−2.847) (−0.944) (−2.168)
Internal control system effectiveness 0.644⁎⁎⁎ 0.662⁎⁎ 1.128⁎⁎⁎
(4.313) (2.458) (3.976)
Loan officer approval ∗ internal control system effectiveness −0.338⁎ −0.168 −0.813⁎
(−1.790) (−0.501) (−1.954)
Individual lending −0.861⁎⁎ −1.306⁎⁎⁎ −2.191⁎⁎⁎ −0.663⁎ −1.340⁎⁎⁎ −2.297⁎⁎⁎
(−2.247) (−4.314) (−4.098) (−1.773) (−4.564) (−4.477)
Loan officer approval ∗ individual lending 0.199 0.00848 1.453⁎ 0.111 0.00558 1.582⁎⁎
(0.398) (1.576) (1.838) (0.226) (1.010) (2.079)
Profitability −0.265 1.635 0.0587 0.0573 1.439 0.313
(−0.234) (1.281) (0.0277) (0.0500) (1.070) (0.174)
Loan portfolio growth −0.127⁎⁎⁎ 0.111 −0.480 −0.165⁎⁎⁎ 0.0667 −0.0255
(−3.219) (0.540) (−1.298) (−3.741) (0.318) (−0.0593)
For profit MFIs 0.188 −0.225 −0.0989 0.220 −0.321 −0.143
(0.805) (−0.816) (−0.302) (0.922) (−1.083) (−0.421)
Corruption index −0.170 −0.180 −0.973⁎⁎ −0.0794 −0.144 −0.962⁎⁎
(−0.806) (−0.548) (−2.586) (−0.313) (−0.411) (−2.625)
Annual GDP growth −7.679⁎ 0.668 −4.310 −5.290 0.996 −3.963
(−1.778) (0.192) (−0.697) (−1.253) (0.270) (−0.621)
_cons 5.678⁎⁎⁎ 6.606⁎⁎⁎ 5.907⁎⁎⁎ 6.936⁎⁎⁎ 7.266⁎⁎⁎ 6.235⁎⁎⁎
(9.024) (5.544) (4.557) (11.31) (8.085) (4.956)
N 154 100 76 154 100 76
R-sq 0.358 0.297 0.364 0.314 0.308 0.356
Adj. R-sq 0.313 0.217 0.267 0.266 0.230 0.256
F 16.44 10.32 4.711 11.19 9.740 7.645

t-Statistics are in parentheses.

⁎
p < 0.10.
⁎⁎
p < 0.05.
⁎⁎⁎
p < 0.01.

choice decision model and a loan approval decentralization decision
model) with robust standard errors.
Ideally, controlling for the endogeneity of corporate governance
effectiveness and internal control systems requires building a control
sample. To do that, we use a control sample of 561 MFI-year observa-
tions from 59 unrated MFIs with unbalanced and reliable data, over
2001–2012. This sample was extracted from the MIX database. Only
MFIs rated four diamonds or higher by the MIX are considered.
Financial statements in this category are certified by auditors, and for
some of them, by the Big Four accounting firms (PwC, KPMG, Ernst &
Young, and Deloitte). Based on the existing literature on rating choice
(Adams, Burton, & Hardwick, 2003; Hartarska & Nadolnyak, 2008), we
model the rating as a function of some MFI-specific characteristics, such
as maturity (proxied by age), size of the loan portfolio, number of active
borrowers, asset size, performance, ownership type and portfolio risk.
The estimated probit rating decision model is as follows:
Rating decision = β0 + β1 Age + β2 Gross loan portfolio
+ β3 Portfolio at risk 30 days + β4 ROA
+ β5 Cooperatives + β6 Privately owned
+ β7 ln(Active borrowers) + ε.
(2)
To estimate the loan approval decentralization decision model, we
follow Berger and Udell (2002), Stein (2002), Beck, Demirgüç-Kunt,
and Martinez-Piera (2011) and Berger, Klapper, Martinez-Piera, and
Zaidi (2008) who consider variables such as size and ownership type,
and country macroeconomic variables, as determinants of the choice to
decentralize the loan approval process. The estimated probit decen-
tralization decision model is as follows:
Loan approval decentralization decision = β0 + β1 ln(Assets)
+ β2 Average loan size per borrower
+ β3 Depth of outreach + β4 Cooperatives
+ β5 Privately owned + β6 Group lending
+ β7 Village banking lending
+ β8 Annual GDP growth + ε. (3)
We use the results of Eqs. (2) and (3) to compute the inverse Mills
ratios (IMR), which we introduced as control variables in Eq. (4) below
to control for selection bias and endogeneity effects. We then obtain the
following model:
MFI performance = α0 + αi Xi + βi Yi + γj ICVj + IMR + δt + ε, (4)
where i indexes MFIs, j indexes country and t indexes year. Xi is the
vector of our main MFI-specific variables and Yi is the vector of other
MFI-level variables described previously in Section 3. ICVj is the vector
of country-level variables, δt is the year fixed effects and ε is the error
term.
The results for Eqs. (2) and (3) are reported in Table 8 (stage 1). We
observe that the chi-squared are statistically significant, which suggests
that the decisions to seek the rating or to decentralize the loan approval
decision are significantly associated with age of MFI, outstanding loans,
MFI ownership type, number of active borrowers, asset size, ROA,
portfolio risk, average loan size per borrower, depth of outreach and
economic performance, and therefore support the findings of previous
studies in microfinance (Hartarska & Nadolnyak, 2008; Tchakoute
Tchuigoua, 2012). We use the results of Eqs. (2) and (3) to compute the
inverse Mills ratios, which we introduced as control variables in Eq. (4)

12
H. Tchakoute-Tchuigoua, I. Soumaré Journal of Business Research 94 (2019) 1–17

Table 8
Results of the Heckman two steps procedure estimation. Controlling for the endogeneity of ratings and Loan decentralization to the loan officer.
Stage 1.A. Estimation of the probit rating decision model with robust standard errors.

Dependent variable: rating decision Coefficients t-Stat

⁎⁎⁎
Constant 3.47 9.50
Age (maturity) −0.49⁎⁎⁎ −6.55
Gross loan portfolio −1.51⁎⁎⁎ −4.53
Portfolio at risk at 30 days −0.97⁎ −1.68
Return on assets (profitability) 2.09⁎⁎ 3.30
Ln (active borrowers) −0.13⁎⁎⁎ −5.41
Cooperative 0.61⁎⁎⁎ 3.68
Privately owned −0.20⁎⁎ −2.09
Pseudo R2 14.15%
Chi2 154.25⁎⁎⁎
Log-likelihood value −515.0896
Number of observations 881

Stage 1 B. Estimation of the probit loan approval decision model with robust standard errors.

Dependent variable: the choice of decentralization Coefficients t-Stat

Constant 0.28 0.33

Size 0.31⁎⁎⁎ 4.36
Cooperative 0.76⁎⁎⁎ 3.13
Privately-owned 0.23 1.28
Average loan size per borrower −0.24⁎⁎ −2.07
Depth of outreach 0.02⁎⁎ 2.38
Group lending 0.20 0.65
Village banking lending 0.18 0.62
Annual GDP growth 5.07⁎ 1.79
Pseudo R2 12.31%
Chi2 44.79⁎⁎⁎
Log-likelihood value −191.51457
Number of observations 316

Stage 2.A

Governance effectiveness Internal control system effectiveness

Portfolio at risk at Write-offs ratio Loan loss provision Portfolio at risk at Write-offs ratio Loan loss
30 days 30 days provision

Loan officer approval −0.0220 −0.0495 −0.00667 −0.00978 −0.0208 −0.00359

(−0.690) (−1.456) (−0.311) (−0.383) (−0.726) (−0.146)
Governance effectiveness −0.0187⁎⁎⁎ −0.0149⁎ −0.00452
(−2.823) (−1.907) (−0.831)
Loan officer approval ∗ governance effectiveness 0.00571 0.0137⁎ 0.000838
(0.685) (1.658) (0.147)
Internal control system effectiveness −0.0126⁎⁎ −0.00955⁎ −0.0108⁎⁎
(−2.253) (−1.914) (−2.037)
Loan officer approval ∗ internal control system 0.00171 0.00543 −0.000459
effectiveness (0.250) (0.784) (−0.0713)
Individual lending 0.0180⁎⁎⁎ −0.00566 −0.00116 0.0169⁎⁎⁎ −0.00641 −0.000221
(2.720) (−0.621) (−0.221) (2.606) (−0.685) (−0.0441)
Loan officer approval ∗ individual lending 0.0000903 −0.000364 −0.0000368 0.000184 −0.000327 0.0000228
(0.408) (−1.322) (−0.192) (0.892) (−1.200) (0.122)
Profitability −0.190⁎⁎⁎ −0.140⁎⁎⁎ −0.155⁎⁎⁎ −0.196⁎⁎⁎ −0.141⁎⁎⁎ −0.150⁎⁎⁎
(−4.546) (−3.310) (−2.901) (−4.373) (−3.227) (−2.976)
Loan portfolio growth 0.00663 −0.00175 0.00588⁎⁎⁎ 0.00695 −0.00138 0.00634⁎⁎⁎
(1.472) (−0.897) (3.210) (1.387) (−0.611) (4.054)
For profit MFIs −0.000776 0.00228 0.00684 −0.0000913 0.00290 0.00973⁎
(−0.129) (0.334) (1.340) (−0.0149) (0.414) (1.957)
Corruption index 0.0108 0.00934 0.00857⁎ 0.00920 0.00926 0.00950⁎⁎
(1.497) (1.423) (1.737) (1.313) (1.458) (1.979)
Annual GDP growth −0.143 −0.0826 −0.117 −0.148 −0.0947 −0.126
(−1.323) (−0.453) (−1.068) (−1.370) (−0.516) (−1.209)
Inverse Mills ratio 1 (rating) 0.00625 0.0117 0.00959 0.00524 0.0121 0.0166⁎
(0.641) (0.791) (1.118) (0.552) (0.855) (1.955)
Inverse Mills ratio 2 (decentralization) 0.00134 −0.00478 0.00152 0.00100 −0.00375 0.00277
(0.105) (−0.565) (0.127) (0.0825) (−0.476) (0.245)
_cons 0.0898⁎⁎⁎ 0.0664⁎ 0.0282 0.0683⁎⁎⁎ 0.0527⁎⁎ 0.0543⁎⁎⁎
(3.483) (1.943) (1.274) (3.083) (2.158) (2.601)
Year fixed effects Yes Yes Yes Yes Yes Yes
N 304 292 302 304 292 302
(continued on next page)

13
H. Tchakoute-Tchuigoua, I. Soumaré Journal of Business Research 94 (2019) 1–17

Table 8 (continued)

Stage 2.A

Governance effectiveness Internal control system effectiveness

Portfolio at risk at Write-offs ratio Loan loss provision Portfolio at risk at Write-offs ratio Loan loss
30 days 30 days provision

R-sq 0.221 0.159 0.191 0.213 0.156 0.236

Adj. R-sq 0.157 0.087 0.124 0.148 0.084 0.172
F 8.541 . 5.528 9.238 . 10.18

Stage 2.B

Governance effectiveness Internal control system effectiveness

Breadth of Yield on loan Depth of outreach Breadth of Yield on loan Depth of outreach
outreach portfolio outreach portfolio

Loan officer approval 1.878⁎⁎⁎ −0.208⁎ 0.0333 1.655⁎⁎⁎ −0.138 −0.452

(2.882) (−1.928) (0.0417) (3.034) (−1.446) (−0.757)
Governance effectiveness 0.680⁎⁎⁎ −0.0315 −0.268⁎
(4.623) (−1.274) (−1.746)
Loan officer approval ∗ governance effectiveness −0.481⁎⁎⁎ 0.0579⁎⁎ 0.00215
(−2.701) (2.055) (0.0107)
Internal control system effectiveness 0.540⁎⁎⁎ −0.0240 −0.192⁎
(4.449) (−1.182) (−1.655)
Loan officer approval ∗ internal control system −0.404⁎⁎⁎ 0.0382 0.131
effectiveness (−2.741) (1.564) (0.885)
Individual lending −0.539⁎⁎⁎ −0.124⁎⁎⁎ 1.305⁎⁎⁎ −0.512⁎⁎⁎ −0.124⁎⁎⁎ 1.286⁎⁎⁎
(−2.999) (−4.163) (6.755) (−2.910) (−4.158) (6.630)
Loan officer approval ∗ individual lending 0.00339 0.00123 −0.0113⁎⁎⁎ 0.00136 0.00119 −0.0102⁎⁎⁎
(0.688) (1.316) (−3.036) (0.281) (1.240) (−2.868)
Profitability 2.196⁎⁎⁎ −0.128 −0.547 2.456⁎⁎⁎ −0.128 −0.816
(3.370) (−0.679) (−0.615) (3.743) (−0.690) (−0.854)
Loan portfolio growth −0.0596 0.0357⁎⁎⁎ −0.0899⁎⁎ −0.0715 0.0368⁎⁎⁎ −0.0915⁎⁎
(−1.044) (3.204) (−2.231) (−0.987) (3.556) (−2.156)
For profit MFIs −0.0821 0.0437⁎⁎ 0.227 −0.0962 0.0443⁎⁎ 0.205
(−0.655) (2.081) (1.645) (−0.728) (2.085) (1.478)
Corruption index −0.364⁎⁎⁎ 0.0254 −0.239 −0.344⁎⁎ 0.0291 −0.279
(−2.619) (1.344) (−1.197) (−2.474) (1.540) (−1.406)
Annual GDP growth −8.017⁎⁎⁎ −0.846⁎ −2.345 −7.886⁎⁎⁎ −0.881⁎⁎ −2.125
(−3.414) (−1.920) (−0.774) (−3.328) (−2.004) (−0.707)
Inverse Mills ratio 1 (rating) 1.899⁎⁎⁎ −0.0651⁎ −0.651⁎⁎⁎ 1.971⁎⁎⁎ −0.0628⁎ −0.779⁎⁎⁎
(8.729) (−1.853) (−2.663) (9.369) (−1.750) (−2.894)
Inverse Mills ratio 1 (decentralization) −1.600⁎⁎⁎ 0.00671 −0.571⁎⁎ −1.552⁎⁎⁎ 0.00671 −0.645⁎⁎
(−6.471) (0.193) (−2.431) (−6.188) (0.187) (−2.592)
_cons 7.798⁎⁎⁎ 0.647⁎⁎⁎ 3.035⁎⁎⁎ 8.232⁎⁎⁎ 0.634⁎⁎⁎ 2.626⁎⁎⁎
(11.13) (6.210) (4.344) (12.70) (6.761) (4.534)
Year fixed effects Yes Yes Yes Yes Yes Yes
N 304 302 300 304 302 300
R-sq 0.504 0.218 0.287 0.504 0.214 0.274
Adj. R-sq 0.463 0.153 0.228 0.463 0.148 0.214
F 19.16 7.425 . 17.32 7.782 .

Stage 1.A table reports the results of the rating decision model (probit). From this model, we generate the Inverse Mills ratio 1 (rating). Stage 1.B table reports the
results of the decentralization decision model (probit). From this model, we generate the Inverse Mills ratio 2 (decentralization). t-Statistics are in parentheses.
Stage 2.A reports the results of the second stage with risk measures as dependent variables. All independent variables are defined in Table 1. We report estimated
coefficients in the first row. t-Statistics are in parentheses.
Stage 2.B reports the results of the second stage with outreach measures as dependent variables. All independent variables are defined in Table 1. We report estimated
coefficients in the first row. t-Statistics are in parentheses.
⁎
p < 0.10.
⁎⁎
p < 0.05.
⁎⁎⁎
p < 0.01.

to control for selection bias and endogeneity effects.
The results of the re-estimated models reported in Table 8 (stage 2)
yield consistent evidence and improve the reliability of our previous
results; that is, granting loan approval decision making to the loan of-
ficer increases the MFI's outreach without altering the loan portfolio
quality. With this selection bias correction model, we observe that loan
officer approval has a non-significant direct impact on portfolio risk,
but a positive significant impact on outreach, in all the regressions.
However, effective internal control systems and governance mechan-
isms have a risk-reducing effect. In addition, better governance and
internal control mechanisms have an attenuating effect on excessive
credit growth. Overall, the results show that granting loan decisions to
loan officers increases outreach without altering portfolio quality when
more effective governance mechanisms and internal control systems are
in place.
4.3.4. Alternative measure of loan approval decentralization
Fourth, as we have discussed above, it may be the case that the loan
approval decision is made at the branch credit committee level with the
participation of the loan officer in the decision process. Loan officers

14
H. Tchakoute-Tchuigoua, I. Soumaré Journal of Business Research 94 (2019) 1–17

Table 9
The credit committee approved loans.
This table reports the results of the pooled OLS with loan portfolio at risk and Breadth of outreach as dependent variables. Standard errors are heteroskedastic
robust. We control for year fixed effects. All independent variables are defined in Table 1. We report estimated coefficients in the first row.
Portfolio at risk at 30 days Breadth of outreach

Credit committee approval 0.00178 −0.00276 −0.849 −0.883

(0.0515) (−0.0811) (−1.041) (−1.336)
Governance effectiveness −0.0185⁎⁎⁎ 0.614⁎⁎⁎
(−3.441) (5.080)
Credit committee approval ∗ governance effectiveness −0.00543 0.188
(−0.592) (0.820)
Internal control system effectiveness −0.0144⁎⁎⁎ 0.530⁎⁎⁎
(−3.729) (6.172)
Internal control system effectiveness ∗ credit committee approval −0.00279 0.152
(−0.302) (0.831)
Individual lending 0.0129 0.0134 −1.105⁎⁎⁎ −1.127⁎⁎⁎
(1.463) (1.536) (−4.865) (−4.968)
Individual lending ∗ credit committee approval 0.0188 0.0144 −0.229 −0.100
(1.213) (0.957) (−0.615) (−0.277)
Profitability −0.176⁎⁎⁎ −0.179⁎⁎⁎ 0.887 0.924
(−5.635) (−5.183) (1.226) (1.284)
Loan portfolio growth 0.00560 0.00607 −0.131⁎⁎⁎ −0.146⁎⁎
(1.176) (1.108) (−2.964) (−2.467)
For profit MFIs −0.00197 −0.000715 0.0681 −0.00520
(−0.325) (−0.122) (0.429) (−0.0330)
Corruption index 0.0135⁎ 0.0118 −0.269 −0.233
(1.741) (1.583) (−1.505) (−1.250)
Annual GDP growth −0.164 −0.176⁎ −3.795 −3.302
(−1.588) (−1.716) (−1.276) (−1.107)
_cons 0.100⁎⁎⁎ 0.0809⁎⁎⁎ 8.000⁎⁎⁎ 8.381⁎⁎⁎
(3.453) (3.710) (12.42) (17.10)
Year fixed effects Yes Yes Yes Yes
N 330 330 331 331
R-sq 0.225 0.215 0.235 0.243
Adj. R-sq 0.172 0.161 0.183 0.192
F 4.735 4.999 9.912 8.639

t-Statistics are in parentheses.

⁎
p < 0.10.
⁎⁎
p < 0.05.
⁎⁎⁎
p < 0.01.

who act as an interface between the credit committee and borrowers
produce information, and the final authority to approve loans rests
within the credit committee. However, by participating in the credit
committee at the branch level, it is more likely that the loan officer
influences the credit decision. The loan approval in this case is semi-
decentralized, and the loan officer is associated with the loan approval
decision. We therefore use the branch-credit committee approval as an
alternative measure of loan approval decentralization. The regression
results given in Table 9 with this alternative loan approval decen-
tralization indicator support our findings.
4.3.5. Accounting for the non-linearity of outreach
Fifth, above we have assumed outreach to be continuous, but it is
possible that outreach is non-linear and there may be threshold effects.
We control for possible threshold effects by first running a multinomial
regression, where high outreach (NAB > 30,000) is coded 1, medium
outreach (10,000 ≤ NAB ≤ 30,000) is coded 2 and low outreach
(NAB < 10,000) is coded 3. The unreported results obtained do not
show any significant differences relative to our main results.6
Additionally, we split the sample into subsamples of high outreach
(NAB > 30,000), medium outreach (10,000 ≤ NAB ≤ 30,000) and
low outreach (NAB < 10,000) and rerun the regressions on each sub-
sample. The results provided in Table 10, although similar to the gen-
eral trend observed with the whole sample, seem to indicate a much
stronger effect in the low outreach subsample.
6
For conciseness, we do not report those results tables, but they are available
from the authors upon request.
5. Conclusion
In this article, we examine the effect of powerful loan officers on
MFI outreach and loan portfolio quality, and given that powerful loan
officers may exacerbate a principal-agent problem, we also investigate
whether alignment mechanisms—incentive schemes and internal con-
trol systems—in place are effective mitigating tools when loan officers
combine information production and decision functions.
We use an independently pooled cross-sectional sample of 374 MFI-
year observations from 2001 to 2012 for 280 MFIs active in 70 coun-
tries. Our results suggest that decentralizing loan approval decisions or
allocating decision-making authority to the loan officer increases the
outreach but does not necessarily alter the loan portfolio quality,
especially when there are better governance mechanisms and internal
control systems in place. We also find that incentive schemes and in-
ternal control systems help avoid agency problems within MFIs and
thus increase the outreach of MFIs without altering the quality of their
credit portfolio. These findings remain valid irrespective of the MFI
ownership type. These results are robust after controlling for alternative
portfolio risk and outreach measures, outreach threshold effect, crisis
period effect, selection bias and endogeneity. The results are also con-
firmed when we use branch-level credit committee approval as alter-
native measures of loan approval decentralization.
Overall, we have shown empirically that the alignment mechanisms,
namely, the effectiveness of the governance and internal control sys-
tems, are effective and tend to constrain the discretionary power of loan
officers in MFIs. However, the results obtained must be interpreted with
caution for at least two reasons. First, the analysis was based on MFI
data, which did not allow us to explore whether the loan officer could,

15
H. Tchakoute-Tchuigoua, I. Soumaré Journal of Business Research 94 (2019) 1–17

Table 10
Threshold level of outreach.
This table reports the results of the pooled OLS with Breadth of outreach as dependent variable. Standard errors are heteroskedastic robust. We control for year
fixed effects. All independent variables are defined in Table 1. We report estimated coefficients in the first row.
Governance effectiveness Internal control effectiveness

Low outreach Medium outreach High outreach Low outreach Medium outreach High outreach

⁎
Loan officer approval 1.401 0.286 −1.528 0.729 0.337 −0.488
(1.940) (0.864) (−0.826) (1.199) (1.412) (−0.619)
Governance effectiveness 0.553⁎⁎⁎ 0.156⁎⁎ −0.359
(3.582) (2.210) (−0.923)
Loan officer approval ∗ governance effectiveness −0.350 −0.0622 0.252
(−1.637) (−0.682) (0.623)
Internal control system effectiveness 0.252⁎ 0.118⁎ 0.134
(1.903) (1.943) (0.668)
Loan officer approval ∗ internal control system effectiveness −0.0913 −0.0745 −0.0321
(−0.532) (−1.121) (−0.159)
Individual lending −1.364⁎⁎⁎ 0.0000673 −0.245 −1.189⁎⁎⁎ 0.0133 −0.557
(−5.195) (0.000866) (−0.516) (−4.682) (0.160) (−0.827)
Loan officer approval ∗ individual lending 0.214 0.00407⁎⁎ 0.292 0.00379 0.00295 0.535
(0.609) (2.122) (0.546) (0.0111) (1.504) (0.751)
Profitability 0.675 0.270 1.457 0.689 0.430 0.783
(0.964) (0.972) (0.851) (0.907) (1.507) (0.445)
Loan portfolio growth −0.0621⁎⁎ −0.0217 −0.0331 −0.0735⁎ −0.0199 0.230
(−2.064) (−0.320) (−0.123) (−1.821) (−0.292) (0.790)
For profit MFIs 0.0682 0.00587 −0.597⁎⁎ 0.0704 0.0000780 −0.617⁎⁎
(0.366) (0.0933) (−2.399) (0.381) (0.00117) (−2.157)
Corruption index 0.180 −0.0440 0.0107 0.191 −0.0205 −0.113
(1.065) (−0.434) (0.0683) (1.017) (−0.207) (−0.777)
Annual GDP growth −7.334⁎ −1.885⁎⁎ 0.760 −7.970⁎⁎ −1.856⁎⁎ −0.762
(−1.793) (−2.097) (0.147) (−2.015) (−2.109) (−0.146)
_cons 7.839⁎⁎⁎ 9.105⁎⁎⁎ 12.65⁎⁎⁎ 8.940⁎⁎⁎ 9.207⁎⁎⁎ 10.57⁎⁎⁎
(9.293) (31.44) (6.583) (12.96) (26.41) (11.12)
Year fixed effects Yes Yes Yes Yes Yes Yes
N 159 115 57 159 115 57
R-sq 0.386 0.192 0.405 0.344 0.162 0.389
Adj. R-sq 0.297 0.020 0.074 0.249 −0.016 0.050
F 8.918 3.351 12.34 5.502 3.214 2.628

t-Statistics are in parentheses.

⁎
p < 0.10.
⁎⁎
p < 0.05.
⁎⁎⁎
p < 0.01.

for subjective reasons, ration credit. Second, some MFIs seal the
amounts when the credit decision rests with the loan officer. This limit
varies from one MFI to another in our sample, so there could be some
typologies/levels of credit allocation authority that we did not take into
account. These limitations are good avenues for future research.
Finally, another limitation is related to the measurement of the effec-
tiveness of governance mechanisms and the effectiveness of internal
control systems. We use ratings and assume them to be objective
measures of alignment mechanisms. However, we were unable to
control for raters' behavior (their independence) and thus assume zero
collusion between rating agencies and rated MFIs, which is not always
the case. The solution to that would have been to use ratings from other
rating agencies. To that end, we investigated whether alternative ag-
gregate measures of microfinance governance and internal control
mechanisms exist. Unfortunately, such data is not available in the mi-
crofinance sector for at least two reasons: (1) Although governance and
internal control are central in the rating process of most rating agencies
specializing in microfinance, Planet Rating is to date the only one that
assigns a score to each of the main rating areas. (2) Very few MFIs are
rated by more than one rating agency during a year. For those who seek
ratings from at least two different rating agencies, the governance and
internal control scores were not available when the MFI rating was
assigned by a rating agency other than Planet Rating. Unfortunately,
these data are not available in the microfinance sector. The lack of data
does not enable us to conduct additional sensitivity analyses.
References
Adams, M., Burton, B., & Hardwick, P. (2003). The determinants of credit ratings in the
United Kingdom insurance industry. Journal of Business Finance and Accounting,
30(3–4), 539–572.
Adams, R. B., Almeida, H., & Ferreira, D. (2005). Powerful CEOs and their impact on
corporate performance. Review of Financial Studies, 18(4), 1403–1432.
Aghion, P., & Tirole, J. (1997). Formal and real authority in organizations. Journal of
Political Economy, 105, 1–29.
Agier, I. (2012). The role of credit officers in the performance of microloan. Evidence
from Brazil. The Economics of Transition, 20(2), 271–297.
Agier, I., & Szafarz, A. (2013). Subjectivity in credit allocation to micro-entrepreneurs:
Evidence from Brazil. Small Business Economics, 41, 263–275.
Ahlin, C., Lin, J., & Maio, M. (2011). Where does microfinance flourish? Microfinance
institutions performance in macroeconomic context. Journal of Development
Economics, 95, 105–120.
Allen, F., Carletti, E., Cull, R., Qian, J., Senbet, L., & Valenzuela, P. (2014). Improving
access to banking – Evidence from Kenya, CEPR discussion paper no. DP9840.
Allen, F., Demirgüc-Kunt, A., Klapper, L., & Martinez-Peria, M. S. (2016). The foundations
of financial inclusion: Understanding ownership and use of formal accounts. Journal
of Financial Intermediation, 27, 1–30.
Armendáriz de Aghion, B., & Morduch, J. (2010). The economics of microfinance (2nd ed).
Cambridge, MA: MIT Press.
Assefa, E., Hermes, N., & Meesters, A. (2013). Competition and the performance of mi-
crofinance institutions. Applied Financial Economics, 23(9), 767–782.
Aubert, C., de Janvry, A., & Sadoulet, E. (2009). Designing credit agent incentives to
prevent mission drift in pro-poor microfinance institutions. Journal of Development
Economics, 90, 153–162.
Baquero, G., Hamadi, H., & 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.
Basel Committee on Banking Supervision (BCBS) (2010). Microfinance activities and the
core principles for effective banking supervision. Basel, Switzerland: Bank for
International Settlements.
Battilana, J., & Dorado, S. (2010). Building sustainable hybrid organizations: The case of

16
H. Tchakoute-Tchuigoua, I. Soumaré
commercial microfinance organizations. Academy of Management Journal, 53,
1419–1440.
Beck, T., Behr, P., & Guettler, A. (2013). Gender and banking: Are women better loan
officers? Review of Finance, 17, 1279–1321.
Beck, T., Demirgüç-Kunt, A., & Martinez-Piera, M. S. (2011). Bank financing for SMEs:
Evidence across countries and bank ownership types. Journal of Financial Services
Research, 39, 35–54.
Bellucci, A., Borisov, A., & Zazzaro, A. (2010). Does gender matter in bank-firm re-
lationships? Evidence from small business lending. Journal of Banking & Finance, 34,
2968–2984.
Berger, A. N., & Black, L. K. (2011). Bank size, lending technologies, and small business
finance. Journal of Banking & Finance, 35, 724–735.
Berger, A. N., Cowan, A. M., & Frame, W. S. (2011). The surprising use of credit scoring in
small business lending by community banks and the attendant effects on credit
availability, risk, and profitability. Journal of Financial Services Research, 39, 1–17.
Berger, A. N., Klapper, L. F., Martinez-Piera, M. S., & Zaidi, R. (2008). Bank ownership
type and banking relationships. Journal of Financial Intermediation, 17, 37–62.
Berger, A. N., Miller, N. H., Petersen, M. A., Rajan, R. G., & Stein, J. C. (2005). Does
function follow organizational form? Evidence from the lending practices of large and
small banks. Journal of Financial Economics, 76, 237–269.
Berger, A. N., & Udell, G. F. (2002). Small business credit availability and relationship
lending: The importance of bank organizational structure. Economic Journal, 112,
32–53.
Besley, T., & Ghatak, M. (2005). Competition and incentives with motivated agents.
American Economic Review, 95(3), 616–636.
Brickley, J. A., Linck, J. S., & Smith, C. W. (2003). Boundaries of the firm: Evidence from
the banking industry. Journal of Financial Economics, 70, 351–383.
Brown, M., Guin, B., & Kirschenmann, K. (2016). Microfinance banks and financial in-
clusion. Review of Finance, 20(3), 907–946.
Bruett, T. (2005). Measuring performance of microfinance institutions: A framework for re-
porting analysis, and monitoring. USAID: SEEP Network.
Christen, R. P., Lauer, K., Lyman, T., & Rosenberg, R. (2012). A guide to regulation and
supervision of microfinance: Consensus guidelines. CGAP.
Conning, J. (1999). Outreach, sustainability and leverage in monitored and peer mon-
itored lending. Journal of Development Economics, 60(1), 51–77.
Cozarenco, A., Hudon, M., & Szafarz, A. (2016). What type of microfinance institutions
supply savings products? Economics Letters, 140, 57–59.
CSFI (2014). Microfinance banana skins. 2014.
Cull, R., Demirgüç-Kunt, A., & Morduch, J. (2007). Financial performance and outreach:
A global performance of leading microbanks. Economic Journal, 117, F107–F133.
D'Espallier, B., Goedecke, J., Hudon, M., & Mersland, R. (2017). From NGOs to banks:
Does institutional transformation Alter the business model of microfinance institu-
tions? World Development, 89, 19–33.
D'Espallier, B., Hudon, M., & Szafarz, A. (2013). Unsubsidized microfinance institutions.
Economics Letters, 120, 174–176.
Dixon, R., Ritchie, J., & Siwale, J. (2007). Loan officers and loan delinquency in micro-
finance: A Zambian case. Accounting Forum, 31, 47–71.
Galema, R., Lensink, R., & Mersland, R. (2012). Do powerful CEOs determine micro-
finance performance? Journal of Management Studies, 49(4), 718–742.
Guha, B., & Chowdhury, P. R. (2013). Microfinance competition: Motivated micro-len-
ders, double-dipping and default. Journal of Development Economics, 105, 86–102.
Hartarska, V. (2005). Governance and performance of microfinance institutions in Central
Eastern Europe and the Newly Independent States. World Development, 33(10),
1627–1643.
Hartarska, V., & Mersland, R. (2012). Which governance mechanisms promote efficiency
in reaching poor clients? Evidence from rated microfinance institutions. European
Financial Management, 18(2), 218–239.
Hartarska, V., & Nadolnyak, D. (2008). Does rating help microfinance institutions raise
funds? Cross-country evidence. International Review of Economics and Finance, 17(4),
558–571.
Heckman, J. (1979). Sample selection bias as a specification error. Econometrica, 47,
153–161.
Hertzberg, A., Liberti, J. M., & Paravisini, D. (2010). Information and incentives inside the
firm: Evidence from loan officer rotation. Journal of Finance, 65(3), 795–828.
Hudon, M., & Périlleux, A. (2014). Surplus distribution and characteristics of social en-
terprises: Evidence from microfinance. The Quarterly Review of Economics and Finance,
54(2), 147–157.
Jansson, T., Rosales, R., & Westley, G. (2004). Principles and practices for regulating and
supervising microfinance. Washington, DC: Inter-American Development Bank.
Jeon, D. S., & Menicucci, D. (2011). When is the optimal lending contract in microfinance
state non-contingent? European Economic Review, 55(5), 720–731.
Kaufmann, D., Kraay, A., & Mastruzzi, M. (2010). The worldwide governance indicators: A
summary of methodology, data and analytical issues. World bank policy research working
17
Journal of Business Research 94 (2019) 1–17
paper no. 5430.
Labie, M., Méon, P.-G., Mersland, R., & Szafarz, A. (2015). Discrimination by microcredit
officers: Theory and evidence on disability in Uganda. The Quarterly Review of
Economics and Finance, 58, 44–55.
Ledgerwood, J., Earne, J., & Nelson, C. (2013). The new microfinance handbook a financial
market system perspective. Washington D.C: The World Bank.
Liberti, J. M., & Mian, A. R. (2009). Estimating the effect of hierarchies on information
use. Review of Financial Studies, 22(10), 4057–4090.
McKim, A., & Hughart, M. (2005). Staff incentive schemes in practice: Findings from a
global survey of microfinance institutions. MicroFinance network. Available at:
http://www.mfnetwork.org/working_groups/staff_incentives.html.
Mersland, R., & Strøm, R.Ø. (2008). Performance and trade-offs in microfinance
organisations—Does ownership matter? Journal of International Development: The
Journal of the Development Studies Association, 20(5), 598–612.
Mersland, R., & Strøm, R.Ø. (2009). Performance and governance in microfinance in-
stitutions. Journal of Banking & Finance, 33(4), 662–669.
Pathan, S. (2009). Strong boards, CEO power and bank risk-taking. Journal of Banking &
Finance, 33, 1340–1350.
Patten, R. H., Rosengard, J. K., & Johnston, D. E., Jr. (2001). Microfinance success amidst
macroeconomic failure: The experience of Bank Rakyat Indonesia during the East
Asian crisis. World Development, 29(6), 1057–1069.
Perilleux, A., Hudon, M., & Bloy, E. (2012). Surplus distribution in microfinance.
Differences among cooperative, nonprofit, and shareholder forms of ownership.
Nonprofit and Voluntary Sector Quarterly, 41(3), 386–404.
Sagamba, M., Shchetinin, O., & Yusupov, N. (2013). Do microloan officers want to lend to
the less advantaged? Evidence from a choice experiment. World Development, 42,
182–198.
Sah, R., & Stiglitz, J. (1986). The architecture of economic systems: Hierarchies and
polyarchies. American Economic Review, 76, 716–727.
Servin, R., Lensink, R., & Van den Berg, M. (2012). Ownership and technical efficiency of
microfinance institutions: Empirical evidence from Latin America. Journal of Banking
& Finance, 36, 2136–2144.
Stein, J. C. (2002). Information production and capital allocation: Decentralized versus
hierarchical firms. Journal of Finance, 57(5), 1891–1921.
Stiglitz, J. (1990). Peer monitoring and credit markets. World Bank Economic Review, 4,
351–366.
Tchakoute Tchuigoua, H. (2010). Is there a difference in performance by the legal status
of microfinance institutions? The Quarterly Review of Economics and Finance, 50(4),
436–442.
Tchakoute Tchuigoua, H. (2012). Active risk management and loan contract terms:
Evidence from rated microfinance institutions. The Quarterly Review of Economics and
Finance, 52, 427–437.
Tchakoute Tchuigoua, H. (2014). Performance of microfinance institutions: Do board
activity and governance ratings matter? Finance (Review of the French Finance
Association), 35(3), 7–52.
Tchakoute Tchuigoua, H. (2016). Buffer capital in microfinance institutions. Journal of
Business Research, 69(9), 3523–3537.
Wooldridge, J. M. (2010). Econometric analysis of cross section and panel data (2nd edi-
tion). Cambridge, London: MIT Press.
Hubert Tchakoute-Tchuigoua is Professor of corporate finance at KEDGE Business
School. He holds a degree in economics and finance (Catholic University of Central
Africa), a Ph.D. in management sciences (Bordeaux University), and an “Habilitation à
Diriger des Recherches” (qualification to supervise doctoral dissertation) from Sorbonne
Graduate Business School, University Paris 1-Panthéon-Sorbonne. His research activities
are in the field of development finance with a specific focus on the offer side, that is, on
very small banking organizations in developing and emerging countries also called mi-
crofinance institutions. He is the author or co-author of more than twenty articles pub-
lished in national and international high quality journals. His current publications cover
four main areas namely, financing policy, ratings, financial information quality, and loan
contract.
Issouf Soumaré is Professor of Finance and the Director of International Relations of the
Faculty of Business Administration at Université Laval in Canada. He is also the Director
of the Laboratory for Financial Engineering of Université Laval. His research and teaching
interests include international finance, risk management, financial engineering and nu-
merical methods in finance. His theoretical and applied financial economic works have
been published in leading international economics and finance journals. He worked for
the African Development Bank (ADB) at the former Risk Management Unit from 1996 to
1998. Prof. SOUMARÉ holds a PhD in Business Administration in Finance from the
University of British Columbia (UBC, Canada) and an MSc in Financial Engineering from
Université Laval (Canada).