Do Decision Variables Improve Microfinance Efficiency? A Stochastic Frontier Analysis for African Countries

• Date: 01 March 2017
• Published date: 01 March 2017
• Author: Sandra Kendo
• DOI: http://doi.org/10.1002/jsc.2118

RESEARCH ARTICLE

Strategic Change 26: 159–174 (2017)

Published online in Wiley Online Library

(wileyonlinelibrary.com) DOI: 10.1002/jsc.2118

Copyright © 2017 John Wiley & Sons, Ltd.

Strategic Change: Briengs in Entrepreneurial Finance

Strategic Change

DOI: 10.1002/jsc.2118

Do Decision Variables Improve Micronance

Efciency? A Stochastic Frontier Analysis

for African Countries1

Sandra Kendo

Department of Economics – CESEM, Neoma Business School, France

Micronance institutions that aim to increase their asset size and nancial

performance boost their efciency when they make nancial innovations, improve

client follow‐up, and monitor the projects they support to limit the risk of failure.

e micronance sector in Africa is characterized by a high degree of segmentation

among informal and formal lenders. is fact is illustrated by an asymmetric

market distribution, shared between an oligopoly of lenders grouping large assets

and a multiplicity of lenders with small assets. e high number of micronance

institutions (MFIs) have limited nancial and social performance. Only lenders

with signicant assets achieve better results in terms of product portfolio quality,

borrower diversication, risk minimization, and increased nancial protability.

Only MFIs with signicant assets can reach a high enough level of eciency to

enable them to achieve their twofold objective, namely nancial and social sustain-

ability. MFI eciency refers to their ability to produce a maximum number of

outputs from a set of available inputs (Farrell, 1957). In this case, MFIs must be

able to reduce their costs and achieve economies of scale.

Micronance eciency studies use diverse approaches to measure the eciency

level of individual MFIs. Firstly, some micronance eciency studies take a statisti-

cal approach, employing nancial and social ratios that appear limited for a robust-

ness check analysis (Yaron, 1992; Navajas et al., 2000). To improve this, authors

apply data envelopment analysis (DEA) or a stochastic frontier approach (SFA) to

assign eciency scores to MFIs. For example, Gutierrez‐Nieto et al. (2007) use a

DEA approach to measure MFIs’ eciency. eir approach clearly species the

inputs and outputs of each MFI and how some specic variables can impact the

eciency score. ey identify a country eect on eciency, mainly relying on the

status of MFIs [non‐governmental organizations (NGOs), micronance banks,

credit and saving cooperatives]. Hermes and Lensink (2011) use stochastic frontier

analysis to examine the link between micronance institutions’ outreach to the

1JEL classication codes: C24, C67, G21, O55.

Using a stochastic cost frontier,

this study evaluates the efciency

score of African micronance

institutions.

Moreover, we employ a Tobit

model to investigate how decision

variables impact the efciency of

micronance institutions in

African countries.

Results show that an increase in

asset size and nancial

performance, which positively

impacts technical efciency, is

more effective for nancial

cooperatives and non‐bank

nancial institutions with MFIs.

160 Sandra Kendo

Copyright © 2017 John Wiley & Sons, Ltd. Strategic Change

DOI: 10.1002/jsc

poor and their eciency. e outreach indicators they

consider are average loans per borrower and percentage of

female borrowers in the total loan portfolio of the MFI. A

high average loan per borrower indicates less deep out-

reach, since in this case the MFI is expected to provide

fewer loans to poor borrowers (Hermes and Lensink,

2011). A negative relationship has been established

between the average loan balance (outreach depth measure)

and eciency (Hermes and Lensink, 2011), which indi-

cates that MFIs are less ecient in this case.

Each of these methodological approaches has its

advantages and limits. In this study, we choose an SFA

developed by comparing the results of an intermediation

approach to micronance development activities with

those of a production approach. Considering both

approaches, we make an analysis related to the SFA by

investigating whether the initial results related to outreach

and nancial performance impacts are reliable, in terms

of the technical eciency of micronance. us, the

objective of this study is to evaluate the impacts of deci-

sion variables (size/nancial performance/outreach/risk

level) on micronance eciency. First, we assume positive

links between the size and eciency of micronance insti-

tutions. Second, we assume a negative relationship between

average loans and micronance eciency. ird, we

assume a positive relationship between nancial perfor-

mance and micronance eciency. Last, we assume a

negative eect between the risk level and eciency of MFIs.

We applied panel data on 172 MFIs over the period

2004–2011 and found that asset size and return on equity

(a measure of nancial performance) had a positive inu-

ence on technical eciency. Moreover, we observed that

average loans and risk premium (a measure of risk level)

had a negative impact on the technical eciency of micro-

nance. ese eects as results depend on nancial coop-

eratives and non‐bank nancial institutions. Our four

hypotheses are validated by some specic aspects of the

institutional organization of MFIs.

is article is organized as follows. First, we present a

literature review focusing on micronance eciency.

en, we explain the methodology applied to evaluate

technical score eciency and its determinants. Last, we

present and discuss the results and conclude with some

important points.

Literature review

e ability of micronance institutions to achieve eco-

nomic eciency through control and management of

transaction costs is related to their internal governance,

which is partially responsible for agency problems within

MFI organizational structures. Good governance requires

the existence of a clear denition of the institutional

framework and a transparent and ecient decision‐

making mechanism. e accumulation of functions by

some MFI sta members makes it dicult to accurately

assess the organization and functioning of MFIs. Agency

conicts may also occur between borrowers and lenders

in institutions such as credit and saving cooperatives and

mutual organizations, where clients’ dual status may

sometimes lead to decisions being taken that are less than

optimal for the institution. Micronance institutions that

focus on governance issues minimize risks, secure their

nancial investments, and reduce agency costs arising

from conicts between managers and shareholders

(Shleifer and Vishny, 1997).

Good governance can improve institutions’ eciency

by reducing transaction costs. e micronance sector

features ve interrelated governance components: the

quality and reliability of information technologies; the

clarity of organization principles; the denition of a clear

strategic vision shared by all members of the organization;

the implementation of legitimate forms of power that are

suited to the institution; and the organization’s rm estab-

lishment in society at large (Lapenu, 2002). Some particu-

larities of MFI governance are based mainly on the

stakeholders involved in wealth creation and the way

resources are used (Lapenu and Pierret, 2005). e parties

involved here are not only shareholders, ocers, and

owners; they are also borrowers, investors, and credit