Access to Finance

FORUM

Reports by CGAP and Its Partners No. 8, February 2014

Greeneld MFIs in Sub-Saharan Africa

A Business Model for Advancing Access to Finance
Julie Earne, Tor Jansson, Antonique Koning, and Mark Flaming

This research was funded by the Partnership for Financial Inclusion, a joint initiative by The MasterCard Foundation and IFC that aims to scale up commercial micronance and advance mobile nancial services in Sub-Saharan Africa. An important objective of the Partnership is to contribute to the global community of practice on nancial inclusion, and to share research and lessons learned for the common good. The MasterCard Foundation advances micronance and youth learning in developing countries to promote nancial inclusion and prosperity

Authors
The authors are Julie Earne and Tor Jansson from IFC, Antonique Koning from CGAP, and Mark Flaming, independent consultant. Flaming served as IFC nominee on the board of MicroCred. After completing the research and writing for this publication he joined MicroCred as their COO.

Acknowledgments
This paper and the research behind it have been jointly produced by CGAP and IFC, with generous support from The MasterCard Foundation. IFC has direct investments in six of the holding companies and 17 greeneld MFIs that participated in the research. The authors are grateful to the many networks and MFIs who have contributed data and time in support of this project, including Access Holding, Accion, Advans Group, ASA, BRAC, EcoBank, FINCA, MicroCred Group, Opportunity International, ProCredit, and Swiss Micronance Holding. We would also like to thank members of our esteemed advisory committeeBlaine Stephens, Michael Mithika, Rene Azokli, and Alexia Latortuefor being so generous with their time and advice and Greg Chen and Kate McKee at CGAP and Momina Aijazuddin, Alexis Diamond, and Greta Bull at IFC for their reviews and insightful comments. We also thank Anna Nunan at CGAP for her professional editing. Special thanks goes to Adam Sorensen of IFC and Hannah Siedek for their dedication and signicant contributions in researching and organizing data central to this analysis and to Fahima Bille for supporting these efforts. Siedek is the former CEO of ProCredit DRC and assisted with interviews and market analysis for the DRC. The market analysis for the other two markets was done by Yaw Brantuo, an independent consultant, for Ghana, and a team of FTHM Conseils for Madagascar.

2014, Consultative Group to Assist the Poor (CGAP) and International Finance Corporation (IFC) 1818 H Street NW, MSN P3-300, Washington DC 20433 Internet: www.cgap.org Email: cgap@worldbank.org Telephone: +1 202 473 9594 Rights and Permissions This work is available under the Creative Commons Attribution 3.0 Unported license (CC BY 3.0) http:// creativecommons. org/licenses/by/3.0. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: AttributionCite the work as follows: Earne, Julie, Tor Jansson, Antonique Koning, and Mark Flaming. 2014. Greeneld MFIs in Sub-Saharan Africa: A Business Model for Advancing Access to Finance. Forum. Washington, D.C.: CGAP and IFC. License: Creative Commons Attribution CC BY 3.0 TranslationsIf you create a translation of this work, add the following disclaimer along with the attribution: This translation was not created by CGAP and IFC and should not be considered an ofcial translation. CGAP and IFC shall not be liable for any content or error in this translation. All queries on rights and licenses should be addressed to CGAP Publications, The World Bank Group, 1818 H Street NW, MSN P3-300, Washington, DC 20433, USA; e-mail: cgap@world bank.org.

Contents

Executive Summary1
SECTION 1. Introduction3

1.1 Landscape3 1.2 Sample Cohort5

SECTION 2.

Greeneld MFI Performance9

2.1 Balance Sheet Indicators10 2.2 Growth and Operational Performance12 2.3 Financial Performance14
SECTION 3.

The Holding Company17

3.1 Structure17 3.2 Investors18 3.3 Mission and Strategy21 3.4 Institutional Capacity and Knowledge Transfer22 3.5 Success Factors, Common Challenges23
SECTION 4.

The Role of Greeneld MFIs in Market Development25

4.1 Market Share25 4.2 Skills Building26 4.3 Product and Channel Diversication27 4.4 Standards and Good Practices28
SECTION 5. Conclusion31 ANNEX 1 Methodology33 ANNEX 2

List of People Interviewed35

ANNEX 3 Bibliography37

ii

Executive Summary

ub-Saharan Africa (SSA) has the lowest level of access to nance of any region in the world, with an average banked population of only 24 percent (Findex 2012). The regions banking systems are small in both absolute and relative size, and the micronance sector has been relatively slow to expand in SSA compared to other regions in the world. There is a range of strategies for extending the reach of micronance, including the transformation of existing institutions, the creation of stand-alone greeneld micronance institutions (MFIs) with and without a centralized management or holding structure, bank downscaling, and others. This Forum explores the contribution of greeneld MFIs to access to nance in the region. The greeneld business model is focused on expanding nancial services through two main elements: (i ) the creation of a group of greeneld MFIs dened as institutions that are newly created without pre-existing infrastructure, staff, clients, or portfolios, and (ii ) the central organizing bodiesoften holding companiesthat create these MFIs through common ownership and management. The holding company usually also plays a strong role in backstopping operations, providing standard policies and procedures, and co-branding subsidiaries in the network. The greeneld model has come a long way in a short time in SSA from seven greeneld MFIs in 2006 to 31 by 2012. These are spread over 12 SSA countries, including post-conict markets such as the Democratic Republic of Congo (DRC), Cote dIvoire, and Liberia. While there is a range of micronance providers in SSA, the proliferation of greeneld MFIs expands the commercial end of the spectrum with regulated, mostly deposit-taking institutions focused on microenterprises and small businesses. At the end of 2012, the 31 greeneld MFIs in SSA had more than 700,000 loan accounts, an aggregate loan portfolio of $527 million, and close to 2 million deposit accounts with an aggregate balance of $445 million. While many greeneld MFIs are still young, there are signs of solid institution building for the longer term. At the end of 2012, they already employed more than 11,000 staff and had 700 branches. The greeneld MFIs are becoming noteworthy collectively and, in some cases, individually in their markets. The time is right to look more closely at how these MFIs and their holding companies build retail capacity, promote market development, and ultimately advance access to -

nance in SSA. A good number of greeneld MFIs now have a sufficient track record to enable an analysis of their performance and role in the market. This publication includes some references to other types of nancial service providers in SSA in a limited way and only to provide an overview of the spectrum of providers, not to give a quantitative comparison between different models. This stocktaking of the greeneld experience should help inform decisions for various stakeholders. Specically the paper will inform the coming generation of investment in micronance, including how much and what kind of funding is necessary to support capacity building in new institutions as well as those ready to scale up. The paper may also help to promote regulatory consistency for institutions providing a full menu of micro, small, and medium enterprise services. Section 1 introduces the greeneld business model and the landscape in SSA. It also describes the sample of greeneld MFIs and holding companies that participated in the research for this publication. Detailed performance data were obtained from 10 holding companies on 30 greeneld MFIs in SSA. Section 2 provides an analysis of the operational and nancial performance of greeneld MFIs with a focus on their lifecycle, outreach, funding structure, and nancial performance. The life cycle of greeneld MFIs can be divided in three stages: foundation (preparation and rst year of operation), institutional development (year two through nancial breakeven, which typically occurs in year three, four, or ve), and scale-up (from nancial breakeven onward). The performance of greeneld MFIs largely reects these three stages, each of which is characterized by milestones related to management, product development, infrastructure build out, outreach, funding structure, and sustainability. The average initial funding package required for a greeneld MFI ranges from $6 million to $8 million over the rst 34 years of operations and comprises a combination of equity, technical assistance, and debt. Starting equity capital is about $3.5 million and is supplemented by technical assistance. For greeneld MFIs in the study, the technical assistance budgets ranged from $1.5 million to $9 million, but clustered around $4 million. Shareholders and donors typically raise $3 million in grant funding on average, while the MFI pays the rest.
1

A comparison to Micronance Information Exchange (MIX) data for young African MFIs (those with 47 years of operations)1 shows that greeneld MFIs have achieved, on average, very robust performance. By the time greeneld MFIs reach 60 months of operations, they have attained considerably larger size, greater reach, higher loan quality, and better protability than MFIs with no strong holding/network affiliation. The average greeneld MFI tends to be much better capitalized and to have more formal structures and deposit-taking infrastructure than the average young MFI reporting to MIX. Section 3 provides an overview of the structure and typology of holding companies that participated in the research, including how they are governed and funded and how they implement the creation of greeneld MFIs. The typology distinguishes between holding companies that are led by consulting rms (Access Micronance Holding, Advans, MicroCred, Procredit, Swiss Micronance Holding) and those that are helped by network support organizations (ASA, BRAC, FINCA, and Opportunity International) and a commercial bank (EcoBank). Holding companies and their shareholders apply the greeneld business model to address some of the primary challenges in advancing access to nance in SSA. The small size of nascent nancial markets, high costs of doing business, uneven regulatory frameworks, and inadequately skilled human resources in many SSA markets benet from an approach where practices can be standardized and costs can be shared. Through network structures and common practices, holding companies are able to transfer knowledge and learning from one greeneld MFI to another. Their structured approach and heavy focus on human resources development have been a key success factor in their ability to create sustainable institutions in some of the most frontier markets in SSA. Section 4 explores the contribution of the greeneld model in market development in three markets: the DRC, Ghana, and Madagascar. While it is difficult to attribute changes in a market or behavior of competing institutions to the intervention of one or more greeneld MFIs, the authors used quantitative and qualitative information (looking at market share and observed quality of services) to discern effects on the overall level of nancial inclusion and aspects of market building. Development of human resources in the nancial sector and innovations in new products and/or product fea-

tures together with market share were key factors considered. Particularly in the less developed nancial markets, it appears that greeneld MFIs play a pioneering role in expanding the nancial access of microenterprises, small businesses, and low-income households. Their sustainable performance illustrates to the traditional formal banking sector that underserved businesses and households are bankable and even protable market segments. Additionally, greeneld participation in credit bureaus, when available, helps to build the foundation for a strong credit culture and promotes responsible nance for the market as a whole. The most important effect on market building has come from greeneld MFI investment in staff training and development. In the Conclusion, the authors contemplate the future role of greeneld MFIs in SSA. It is likely that the rate of creation of greeneld entities, at least in SSA, will slow, as the most feasible markets have now largely been entered. Yet, there remain about 25 countries in SSA without any greeneld MFI presence, and typically without the presence of any sustainable MFIs at all. And in almost every country, peri-urban and rural populations still struggle to access nancial services. One challenge for greeneld MFIs and their holding companies is, therefore, to develop a delivery model that facilitates commercially viable and affordable access in smaller, more dispersed markets and rural areas. Alternative delivery channels (including agent networks, mobile nancial services, and related partnerships2) are an area of very large investment for greeneld MFIs as they enter the scaleup phase. The greeneld MFI model is a complement to other strategies for increasing access to nance in SSA, such as reform of existing institutions without a holding structure and bank downscaling. The next few years will be very telling about the ability of greeneld MFIs to leverage their foundation and achieve scale, and for holding companies to replicate and sustain the success of their model in other markets, particularly in a context of diminishing funding for technical assistance. Undoubtedly they will nd themselves compelled to develop new methods, capacities, and practices to stay relevant and competitive in the micronance space. At the same time, it is also likely that many of these greeneld MFIs will increasingly begin to compete with commercial banks for mass market customers and those in the small and medium enterprise space. The nancial landscape in Africa is poised to become much more interesting.

1. M  IX index for young MFIs in Africa comprises 58 institutions between four and seven years old in December 2011. All but ve of these institutions are deposit taking, with a deposit base ranging from $30,000 to $22.6 million ($2 million on average) and gross loan portfolio ranging from $2,000 to $24 million ($2.7 million on average). The greeneld MFI subjects of this paper were removed from the benchmark population. 2. F  or more information see Flaming et al. (2013). 2

SECTION

Introduction

ub-Saharan Africa (SSA) has the lowest level of access to nance of any region in the world, with an average banked population of only 24 percent (Findex 2012). The regions banking systems are small in both absolute and relative size (as measured by liquid liabilities and credit as percentage of gross domestic product) (Beck, Maimbo, Faye, and Triki 2011). The micronance sector has been relatively slow to expand in SSA compared to other regions in the world. According to the Micronance Information eXchange (MIX) landscape data, services are concentrated in larger urban centers, and service delivery in rural areas is meager (MIX 2011). Until a few years ago, the main providers of nancial services to base-of-the-pyramid customers were credit unions, savings and loans associations, and nonprot credit programs. Now, new players are entering the region, including specialized greeneld micronance institutions (MFIs), downscaling Pan African commercial banks, and mobile network operators. This paper explores the greeneld business model, which focuses on expanding nancial services through two main elements: (1) creation of a group of greeneld MFIs dened as institutions that are newly created without pre-existing infrastructure, staff, clients, or portfolios, and (2) central organizing bodiesoften holding companiesthat create these MFIs through common ownership and management. The holding company usually also plays a strong role in backstopping operations, providing standard policies and procedures, and co-branding the subsidiaries in the network. Given these commonalities, this model can also be considered a type of franchise where the sponsors inject a tested approach and sufficient patient capital to move new institutions past the difficult start-up phase and onto a growth trajectory in some of the most challenging markets. Three aspects of the greeneld business model are addressed: (i ) performance of subsidiary greeneld MFIs, (ii ) the holding company model, and (iii ) contribution of greeneld MFIs to market development. Performance of subsidiary greeneld MFIs. This section analyzes the operational and nancial performance of greeneld MFIs with a focus on the lifecycle, outreach, funding structure, and time required to break even. Detailed performance data were obtained from 11 holding companies on 30 greeneld MFIs in SSA, and interviews were conducted with more than a third of the chief executive officers of these

banks. This cohort represents 90 percent of all greeneld MFIs created in Africa between 2000 and 2012. The holding company model. This section provides an overview of the structure and typology of holding companies, including how they are governed and funded and how they implement the creation of greeneld MFIs. The sec tion is based on interviews with the management of the holding companies and review of annual reports and nancial statements. Contribution of greeneld MFIs to market development. This section explores the contribution of the greeneld model in market development in the Democratic Republic of Congo (DRC), Ghana, and Madagascar. These countries were selected because they each have a critical mass of greeneld MFIs, most with a reasonably long track record. While it is difficult to attribute changes in a market or behavior of competing institutions to the intervention of one or more greeneld MFIs, the authors used quantitative and qualitative information (looking at market shares and perception of quality of services) to discern effects on the overall level of nancial inclusion, the human resources skills base in the nancial sector, innovations in new products and/or product features, and the demonstration effect from the introduction of and adherence to internationally recognized good practices, e.g., those related to transparency and client protection.

1.1Landscape In SSA, the greeneld model made its debut in 2000 when ProCredit Holding opened a bank in Mozambique (see Box 1). For a few years, ProCredit was essentially alone in pursuing this strategy; it opened up in Ghana in 2002, in Angola in 2004, and in the DRC in 2005. While other network operators and local institutions started nongovernmental organizations (NGOs) and cooperative micronance entities much earlier, the greeneld model of a centralized holding company providing investment and expertise for the development of commercial micronance entities began in earnest at the turn of the millennium. Between 2005 and 2006, Advans, Access, and MicroCred holding companies were formed with a structure similar to that of ProCredit and by the end of 2007 had collectively launched ve greeneld MFIs in SSA. Accion started its rst
3

BOX 1

ProCredit Holding
The rst greeneld MFI was established by Internationale Projekt Consult (IPC) in Bosnia in 1996, and was licensed as a deposit-taking micronance bank. IPC had gained a lot of experience in institutional strengthening of micronance providers by upscaling cooperatives and MFIs in Latin America and downscaling banks in Uganda and Russia. Not satised with being just the consulting company on these projects, IPC decided to invest in the institutions it was strengthening or creating. The main reasons for investing in greeneld MFIs at the time were (i ) to have more control over capacity building and the growth trajectory of institutions, (ii ) encouragement from development nance institutions (DFIs) for IPC to put some skin in the game, and (iii ) a way for IPC to invest in the long-term value of the company.
Source: Interviews and B. Fritz and K. Hujo (2005)

IPC created the investment company Internationale Micro Investitionen AG (IMI) with shareholders, including IPC, IPC staff, and development investors. The company became the main vehicle for expanding the successful Bosnian experiment, rst to other Eastern European countries and then to Latin America. In 2000 ProCredit expanded to SSA, with the encouragement of its DFI shareholders. ProCredit set up ve banks in SSA, in Angola, DRC, Ghana, Mozambique, and Sierra Leone. Since then it sold its banks in Angola and Sierra Leone and continues to have a network of 21 banks, of which three are in SSA. Over the past few years ProCredit has taken a global decision to change its target clientele moving from microenterprises toward very small and small businesses.

greeneld MFI in the same period in partnership with three commercial banks in Nigeria. As a result of this initial experience, Ecobank and Accion entered into a partnership and opened two greeneld MFIs in Ghana and Cameroon. From that point on, the Access, Advans, and MicroCred networks each created more or less one new MFI per year. Toward the end of the decade, ASA and BRAC from Bangladesh created new organizational structures that also allowed them to begin establishing greeneld MFIs in Africa. Over the six years from late 2006 to end of 2012, a total of 27 additional greeneld MFIs were launched. Meanwhile, FINCA and Opportunity International (OI) have used the holding company structure to upgrade their existing (largely NGO) affiliates to regulated deposit-taking institutions, and then integrate them into a common investment company. Like the other networks, the holding company has been a vehicle for mobilizing investment capital, expanding and backstopping operations, and establishing an ownership model for an international network of nancial institutions. There is a range of strategies for extending the reach of micronance, including the transformation of existing institutions, the creation of stand-alone greeneld MFIs without a centralized management or holding structure, bank downscaling, and others. When greeneld MFIs expanded in earnest at the end of 2006, the rst seven (Procredit Angola, ProCredit DRC, Procredit Mozambique, Procredit Ghana, FINCA DRC, Opportunity Ghana, and MicroCred Madagascar) had 107,887 loan

accounts, with an aggregate loan portfolio of $57.4 million, and held 220,377 deposit accounts with an aggregate balance of $50.7 million. Six years later, at the end of 2012, there were 31 greeneld MFIs3 in 12 SSA countries, with 769,199 loan accounts and an aggregate loan portfolio of $527 million, and with 1,934,855 deposit accounts and an aggregate balance of

FIGURE 1

SSA Countries in Which Greeneld MFIs were Created Between January 2000 and June 2012

Senegal Nigeria Cameroon Democratic Republic of Congo Uganda

Sierra Leone Liberia

Cte Ghana dIvorie

Tanzania

Angola Zambia Madagascar Mozambique

3. E  xcluding the former Procredit companies in Sierra Leone and Angola, which were sold in 2007 and 2010, respectively. Technically, there have been several additional greeneld MFIs launched between June 2012 and the publication of this paper (ASA Kenya, Tanzania and Uganda, Oxus DRC, Advans Nigeria), but they were too recent to include. 4

TABLE 1 G  rowth Greeneld MFIs No. of Staff No. of Branches No. of Loans Outstanding Gross Loan Portfolio ($ million) No. of Deposit Accounts Total Deposit Balance ($ million)

of Greeneld MFIs in SSA, 20062012

2006 7 1,564 37 107,887 57.4 220,377 50.7 2007 12 2,512 56 141,231 94.7 317,943 106.7 2008 18 4,856 261 332,349 144.5 595,008 177.9 2009 22 6,685 392 449,973 203.6 780,497 211.6 2010 27 8,009 514 570,017 285.8 1,050,087 291.3 2011 30 10,137 625 743,640 409.5 1,574,750 371.8 2012 31 11,578 701 769,199 527.0 1,934,855 445.5

$445 million. These 31 greeneld MFIs had 11,578 staff and 701 branches. These greeneld MFIs are becoming noteworthy collectively (See Table 1), and in some cases individually, in their markets which is further discussed in Section 4 on the role of greeneld MFIs in market development.

1.2Sample Cohort According to the denition used in this paper, there were 33 greeneld MFIs created before June 2012 (see Table 2). Institutions created through takeovers and mergers were not included, nor were many of the subsidiaries of OI and FINCA, which started as NGOs without strong central network bodies or holding company structures. The detailed performance analysis of the greeneld MFIs (Section 2) is based on information from 30 MFIs out of 33 (see Table 2). These MFIs belong to the 10 holding companies listed in Table 3. ProCredit Angola, Ghana, and Mozambique are not included (shaded in Table 2). ProCredit HoldTABLE 2 G  reeneld

ing did not provide performance data for these banks by institutional age, but allowed the authors to use publically available data for the aggregate gures as well as the data the authors had already collected for ProCredit DRC and ProCredit Sierra Leone.4 IFC has direct investments in six of the holding companies and 17 greeneld MFIs in this cohort. Greeneld MFIs include a variety of legal forms. A majority are licensed and regulated deposit-taking institutions, ranging in legal structure from commercial banks to savings and loan companies to specialized deposit-taking MFIs. Some started as credit-only companies and relicensed as deposit-taking institutions a couple of years after they were created (FINCA DRC, MicroCred Madagascar); a few remain credit-only companies (the four BRAC entities). The holding companies in this study represent a range of organizational structures that are explained in more detail in Section 3. Regardless of their structures, the holding companies play an important role in identifying new markets and

MFIs Created in SSA Between January 2000 and June 2012

Country Mozambique Ghana DRC Angola Ghana DRC Madagascar Madagascar Cameroun Nigeria Senegal Start of Operations 2000 2002 2003 2004 2004 2005 2006 2007 2007 2007 2007 Current License (June 2012) Commercial Bank Savings and Loan Company Deposit-taking MFI Commercial Bank Savings and Loan Company Commercial Bank Commercial Bank Commercial Bank Deposit-taking MFI Micronance Bank Deposit-taking MFI Continued

Greeneld MFI 1 2 3 4 5 6 7 8 9 10 11 ProCredit Mozambique ProCredit Ghana FINCA DRC ProCredit Angola Opportunity Ghana ProCredit DRC MicroCred Madagascar Access Madagascar Advans Cameroon Accion Nigeria MicroCred Senegal

4. D  ata for ProCredit Sierra Leone are included only for the period that it was owned by ProCredit from 2007 until 2010, when it was sold to Ecobank. Procredit Angola was sold in 2007. 5

 reeneld TABLE 2 G

MFIs Created in SSA Between January 2000 and June 2012 contd
Country Sierra Leone Tanzania Ghana Ghana Ghana Nigeria Tanzania Uganda DRC Liberia Liberia Nigeria Sierra Leone DRC Ivory Coast Cameroon Nigeria Nigeria Senegal Tanzania Zambia Ivory Coast Start of Operations 2007 2007 2008 2008 2008 2008 2008 2008 2009 2009 2009 2009 2009 2010 2010 2010 2010 2010 2011 2011 2011 2012 Current License (June 2012) Commercial Bank Commercial Bank Savings and Loan Company Savings and Loan Company NGO (transformed into a Savings and Loan Company in April 2013) Micronance Bank Credit-only Company Credit-only Company Commercial Bank Commercial Bank Credit-only Company NGOMFI Credit-only Company Deposit-taking MFI Deposit-taking MFI Deposit-taking MFI Micronance Bank Micronance Bank Deposit-taking MFI Commercial Bank Commercial Bank Deposit-taking MFI

Greeneld MFI 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 ProCredit Sierra Leone Access Tanzania EB-Accion Ghana Advans Ghana ASA Ghana Access Nigeria BRAC Tanzania BRAC Uganda Advans DRC Access Liberia BRAC Liberia ASA Nigeria (ASIEA) BRAC Sierra Leone Opportunity DRC MicroCred Ivory Coast EB-Accion Cameroon ASA Lagos (ASHA MFB) MicroCred Nigeria Fides Senegal Advans Tanzania Access Zambia Advans Ivory Coast

countries for expansion, funding and capital structure, and adhering to the vision and mission of the network. Each of the holding companies has a cadre of specialists who provide technical assistance (TA) to the MFIs, help manage the holding company operations, and in many cases are founding investors in the holding company. DFIs, socially responsible investors, and specialized micronance investment vehicles (MIVs) are the primary investors in the holding companies and greeneld MFIs.

Now the time is right to look more closely at how these MFIs and their holding companies build retail capacity, promote market development, and ultimately advance access to nance in SSA. A good number of greeneld MFIs now have a sufficient track record to enable an analysis of their performance and role in the market. A stocktaking of the experience should help inform decisions that will shape the coming generation of investment in micronance.

TABLE 3 H  olding

Companies
Sponsor Year Created Investors # MFIs Globally (June 2012) 7 # GF MFIs in Africa (June 2012) 5 Total Assets of Greeneld MFIs in SSA (US$, Dec 2012) 157,966,796

Holding Company

Access Micronance Holding AG

LFS

2006

CDC Group plc, European Investment Bank (EIB), International Finance Corporation (IFC), KfW Entwicklungsbank (KfW), the Netherlands Development Bank (FMO), LFS Financial Systems GmbH, MicroAssets GbR (MA), Omidyar-Tufts Micronance Fund (OTMF) EIB, CDC, FMO, IFC, KfW, Horus Development Finance, FISEA (Proparco) Catalyst MF Investors (Owned by Gray Ghost MF Fund, Sequoia, ABP, TIAA-CREF, CDC Group, responsAbility, private investors) BRAC International Government Employees Pension Fund, Asset Mgt Corp of Nigeria, Social Securuty and National Insurance Trust, Stanbic Nominees Nigeria, IFC and IFC capitalization fund, Interlink Securities ltd. FINCA International, IFC, KfW, FMO, responsAbility, Triple Jump PlaNet Finance, IFC, Socit Gnrale, AXA Belgium, French Development Agency (AFD), EIB, Developing World Markets (DWM) OI

Advans SA SICAR

Horus

2005

74,072,517

ASA International Holding

ASA International

2006

37,060,395

BRAC International Holdings, BV EcoBank International

BRAC NGO EcoBank International and Accion

1972 1985

6 3

4 3

60,006,805 37,633,098

FINCA Micronance Holding Company LLC MicroCred SA

FINCA International PlaNet Finance

2011 2005

21 6

1 4

33,854,511 24,893,865

Opportunity Transformation Investment ProCredit Holding AG & Co. KGaA

OI

2000

15

52,027,591

IPC

1998

IPC GmbH, IPC Invest GmbH & Co., KFW, DOEN, IFC, BIO, FMO, TIAA-CREF, responsAbility, PROPARCO, Fundasal, OmidyarTufts Financial Systems Development Services AG (Fides), Sobelnat, P.G.C. Suisse, Bank im Bistum Essen

21

168,872,564

Swiss Micronance Holding, SA

Fides

2007

4,482,212

SECTION

Greeneld MFI Performance

he life cycle of greeneld MFIs, as it is observed in SSA, can be divided in three stages (see Box 2): (i) foundation (preparation and rst year of operation), (ii) institutional development (year two through nancial breakeven, which typically occurs in year three, four, or ve), and (iii) scale-up (from nancial breakeven onward). The performance of greeneld MFIs largely reects these three stages, each of which is characterized by milestones related to management, product development, infrastructure build out, outreach, funding structure, and sustainability.
BOX 2

In the following sections, the performance of the cohort of greeneld MFIs is evaluated based on institutional age to achieve a coherent comparison and aggregation of data, regardless of the calendar year in which they launched operations. Data are presented as simple averages, unless otherwise indicated, to display the performance of a typical greeneld MFI. In a few cases, when outliers distort the simple average, a weighted average is used. A comparison to MIX data for young African MFIs (those with 47 years of operations)5 shows that greeneld

Three Stages of Greeneld MFIs Life Cycle

Foundation. The foundation stage includes the legal creation of the new entity, shareholder negotiations, the licensing process, and onsite operations preparation. Five of the 30 institutions included in this analysis are currently at this stage. In most jurisdictions a company needs to be created and partly capitalized before a preliminary approval can be sought from regulatory authorities. At that time initial management of the greeneld MFI is also selected, usually consisting of staff seconded from the holding company. The initial staff is responsible for tailoring policies and procedures to the local market, designing and adapting products, installing an information technology (IT) system, identifying and building or refurbishing physical space for branches, and managing the relationship with regulatory authorities. They also recruit one or two cohorts of loan ofcers (usually 2030 in total) and train them for several months, often with network MFIs in other countries. It usually takes 46 months to prepare for operations from the point of receiving the preliminary approval from the regulator. It can then take 24 more months until the central bank inspects the greeneld MFI and grants the nal operating license. Delays in the foundation stage can easily lead to substantial cost over runs, particularly if the holding company has placed staff on the ground too quickly. Institutional development. During the institutional development stage, greeneld MFIs (in partnership with holding companies) focus on building staff capacity and installing risk management systems that will create the core foundation for future growth. Eleven of the 30 institutions were in the institutional development stage at the time of this analysis. Typically the MFIs have only one product, group, or individual microenterprise loans, but in some cases small and medium enterprises (SME) lending is piloted. As operations grow, risk management systems are increasingly institutionalized, including policies and procedures for decentralized management, internal audit, cash and liquidity management, and regulatory compliance, including antimoney laundering measures. The assetliability committee at the board becomes more active as deposits increase and begin to account for a greater portion of funds for intermediation. Scale-up stage. As greeneld MFIs pass breakeven and become protable, they move into a scale-up phase. Fourteen of the 30 institutions in this study were in the scale-up phase at the time of the analysis. At this point, the focus tends to shift toward product diversication and delivery channel development to attract new clients as well as deepen existing client relationships and gain market share. New products are developed to target secondary market segments, for example, agricultural lending for rural clients. The expansion of small and medium enterprise (SME) lending, in particular, can be a critical driver of protability by off-setting the high cost of smaller microloans as institutions expand their footprint into more rural areas. Greeneld MFIs have recently introduced automated teller machine (ATM) channels, and some institutions are now rolling out the rst phase of various forms of agent networks enabled through point-of-service, cards, and mobile devices.

5. M  IX index for young MFIs in Africa, comprised 58 institutions 47 years old in December 2011. All but ve of these institutions are deposit taking, with a deposit base ranging from $30,000 to $22.6 million ($2 million on average) and gross loan portfolio ranging from $2,000 to $24 million ($2.7 million on average). The greeneld MFI subjects of this paper were removed from the benchmark population.

TABLE 4 P  erformance

60 Months

of Greeneld MFIs at 12, 36, and

Month 12 Month 36 318 22 7 47 25,009 9.2 37,460 8.7 4.0 53 4.3 -0.1 -0.3 Month 60 524 31 11 63 36,714 20.0 81,682 23.1 3.4 36 6.6 3.1 18.9 MIX Young Africa* 69 10 n/a n/a 11,255 2.7 18,127 2.0 9.5 113 1.2 -2.4 -3.4

No. Staff No. Branches No. Deposit-Taking Branches No. NondepositTaking Branches No. Loans Outstanding Gross Portfolio ($ million) No. Deposit Accounts Deposit Volume ($ million) PAR30 Op. Expenses / Avg Portf (%) Equity ($ million) Net Income / Avg Assets (%) Net Income / Avg Equity (%)

131 9 3 22 9,495 2.3 7,123 0.8 3.9 200 3.6 -12.4 -44.6

*n=58 African MFIs between 48 and 84 months a. The gures in this table represent simple averages, except for net income/assets and net income/equity; ow data in ratios (e.g., operating expenses/average portfolio) are based on annualized six-month data.

MFIs have achieved, on average, very robust performance. By the time greeneld MFIs reach 60 months of operations, they have attained considerably larger size, greater reach, higher loan quality, and better protability than MFIs with no strong holding/network affiliation. The average greeneld MFI tends to be much better capitalized and to have more formal structures and deposit-taking infrastructure than the average young MFI reporting to MIX. (See Table 4.)

2.1Balance Sheet Indicators An early key decision for investors in greeneld MFIs is how much to provide in start-up equity capital. The capital needs to comply with regulatory requirements, absorb earTABLE 5  Balance Assets ($ million) No. in sample Equity ($ million) No. in sample Equity / Assets (%) No. in sample Deposits / Loans (%) No. in sample 10

ly-stage losses, and yet be sufficient to attract lenders who can provide debt funding. On average, greeneld MFIs in SSA have started operations with equity capital of approximately $3.5 million. While the equity is typically partly eroded over the rst 2448 months as the MFIs make losses, Table 5 shows a moderately increasing equity level over time. This is explained by the fact that shareholders usually inject additional equity to support the solvency and growth of the MFIs, demonstrating their generally strong commitment to the endeavor. These capital injections have also been a response to increasing minimum capital requirements in many African countries during the past six years and the realization among investor/sponsors that $3.5 million has not generally been sufficient to absorb losses, comply with regulatory requirements (for banks), and raise enough debt funding to support loans to the point of breakeven. Initial capitalizations of around $4 million to $5 million are now more common. In comparison, the average equity of the MIX Young Africa MFIs is $1.2 million after 47 years. In addition to equity, most institutions start with external TA grants of $3 million on average. Without this funding, the institution would have to pay for TA out of its own funds (most greeneld MFIs in fact pay for some of the overall TA costs themselves, as will be explained below), incurring higher start-up losses that in turn would be absorbed by the equity through increased retained losses. Taking into account both the initial equity capitalization and the TA funding, the average initial funding package required for a greeneld MFI ranges from $6 million to $8 million over the rst 34 years of operations. But thats just to get started. Naturally the MFI will soon also need additional funding in the form of debt and/or deposits to support a growing asset base. The TA aspect plays a signicant enough role in the nancial and operational development of the greeneld MFIs to warrant a more detailed explanation (see, also, Box 3). The start-up TA program is designed to develop and build capacities of the institution. The average TA program includes a six month to one year preoperational phase and

Sheet Indicators of Greeneld MFIs

Month 12 6.0 29 3.6 27 59 29 25 24 Month 18 8.5 28 3.5 27 47 28 36 23 Month 24 11.4 27 3.6 26 39 27 40 22 Month 30 13.6 23 4.0 22 36 23 43 18 Month 36 16.5 22 4.3 21 33 22 42 17 Month 42 20.9 20 4.5 20 29 20 43 16 Month 48 25.8 17 5.3 17 24 17 43 15 Month 54 28.0 14 5.3 14 23 14 53 12 Month 60 34.9 13 6.6 13 24 13 60 11

a 3648 month operating phase intended to support the institution through the foundation and institutional development stages. The TA budget typically funds 34 long-term senior managers, installation and tailoring of IT application and management information systems, several short-term specialists in internal audit, risk management and product development, and technical backstopping from the holding company/technical service provider. Once the TA budget is established, shareholders and donors typically try to provide enough grant funding to defray a majority of the TA
BOX 3

costs. The institution contributes the balance, usually 20 40 percent of the total. For greeneld MFIs in this cohort, the TA budgets range from $1.5 million to $9 million, but cluster around $4 million. In those cases, shareholders and donors typically raise $3 million in grant funding on average, leaving the MFI to pay the rest. As expected, greeneld MFIs typically remain modestly leveraged over the rst 60 months of operations. They start out with an equity-to-asset ratio at 100 percent, though some of it is eroded by preparation activities and by initial losses

Impact of Technical Assistance on Financial Performance

Why do greeneld MFI projects generally include substantial amounts of TA grant funding? Fundamentally, it is related to internal constraints of investors, such as risk-return preferences and time horizons, and the belief among donors and investors (some investors provide both equity and TA funding) that there are potentially broad benets to well-run and (eventually) large MFIs that can offer a meaningful range of nancial services to microenterprises, small businesses, and low-income populations in SSA (elaborated in Section 4 on market development effects). On the rst point (internal investment constraints), it is of course possible that greeneld MFIs may turn out to be good investments for the initial investors. However, it is unlikely that this will happen over any reasonable time horizon, which most investors would consider to be 58 years. While development-oriented investors may accept lower expected returns for higher expected impact, they also face limits to how far this can be stretched. What does this burden-sharing look like, and how does it affect the nances of the MFIs? As noted earlier, greenTABLE B3-A Example Calculation
Month 12 Net Income for the period ($) Add'l Cost to MFI if no external TA funding ($) Net Income if no external TA funding ($) Equity ($) Equity if no external TA funding ($) Annualized ROAE (%) Annualized ROAE if no external TA funding (%) (410,387) (500,000) (910,387) 3,553,198 2,553,198 -24.4 -71.5 Month 18 (359,666) (500,000) (859,666) 3,510,502 2,010,502 -23.7 -78.9 Month 24 (178,023) (500,000) (678,023) 3,558,164 1,558,164 -15.1 -74.8 Month 30 (33,527) (500,000) (533,527) 3,839,706 1,339,706 -5.8 -72.3 Month 36 (26,280) (500,000) (526,280) 4,324,016 1,324,016 -1.5 -73.5 174,318 4,480,075 1,480,075 3.6 -25.0 419,463 5,267,880 2,267,880 12.4 33.1 553,862 5,284,887 2,284,887 19.9 51.7 395,553 6,558,059 3,558,059 16.1 32.6 Month 42 174,318 Month 48 419,463 Month 54 553,862 Month 60 395,553

eld MFIs receive on average $3 million in external TA grants for their start-up period. In addition, they typically pay about $1 million out of their own pocket, for a total TA budget of $4 million. Table B3-A attempts to illustrate what would happen if the full TA cost were borne by the MFI. This simulated example shows that typical greeneld MFIs would experience higher retained losses if paying fully for the TA. There is also a lot more volatility in the return on average equity (ROAE), fueled by higher initial losses and diminished equity. The time to reach the monthly breakeven point, however, remains the same at month 42. But since the retained losses are higher and will take longer to recover, the expected return to investors is lower. Without TA grants, the expected internal rate of return (IRR) at ve years is approximately 1 percent; with TA grants, it is approximately 14 percent. An IRR of 1 percent is too low for DFI investors to justify an investment, even if they consider the investments to have an important development effect on the local market. In fact, 14 percent is below what many DFIs and social investors would consider acceptable in a region like Africa.

The table builds on the following methods and assumptions: (i) the actual average net income and equity positions for the greeneld cohort were used as a starting point; (ii) the $3 million received in external TA grants is spread evenly across the rst 36 months of operations; (iii) the remaining $1 million funded by the MFI is already reected in the average net income and equity gures.

11

TABLE 6

Average Balances of Greeneld MFIs

Month 12 Month 18 784 28 214 22 Month 24 795 27 238 21 Month 30 743 23 202 17 Month 36 841 22 228 16 Month 42 896 20 251 15 Month 48 929 17 245 14 Month 54 1,003 14 251 11 Month 60 1,147 13 207 9 703 29 152 23

Avg. Loan Balance ($) No. in sample Avg. Deposit Balance ($) No. in sample

Note: The average loan balance ($1,147) and deposit balance ($207) at month 60 reect 101 percent and 22 percent, respectively, of gross domestic product per capita in 2011 of countries represented.

unless shareholders inject additional equity. Over time, as protability stabilizes, the equity-to-asset ratio is determined more by loan growth and deposit mobilization. In the cohort there is more variation among the greeneld MFIs created as nonprot organizations or credit-only companies (sometimes the ratio becomes very low or even negative), presumably because these entities are not as bound by formal regulations regarding minimum capital and possibly also because of capital constraints at the sponsor level. The deposit-to-loan ratio, which is an important indication of the ability of greeneld MFIs to gain the trust of local populations and raise stable resources for on-lending, shows a continuous improvement over time. Nevertheless, the data in this paper clearly show that it is easier for greeneld MFIs to sign up depositors than it is to raise substantial volumes of deposits. The low-income nature of the depositor base means that amounts are small and that greeneld MFIs typically have to rely on signicant amounts of borrowings during their rst 60 months of operations. See Table 6. The ability to mobilize deposits also depends on a couple of other factors, such as institutional license (it tends to be easier for banks than nonbanks to mobilize deposits), and local market conditions (e.g., the reputation of the micronance industry or even the banking sector). In some cases, particularly over time, greeneld MFIs can come to be perceived as safer than local banks because of high standards of service and/or the nature of
TABLE 7

their shareholders (mainly DFIs). Greeneld MFIs mobilize local resources in various forms but, given their lack of a local track record, they initially depend signicantly on debt funding from DFIs and specialized MIVs managed by entities such as responsAbility, Symbiotics, Blue Orchard, Triple Jump, and MicroVest. Local banks have so far been reticent to fund greeneld MFIs. A transition to local deposit and local debt funding is taking place with support from partial credit guarantees, etc., but this will require time: even after ve years of operations many greeneld MFIs battle to raise sufficient funding locally to support their (typically) rapid loan growth.

2.2Growth and Operational Performance Greeneld MFIs have generally achieved impressive growth, with the average MFI having 36,714 loans, $20 million loan portfolio, 81,682 deposit accounts, and $23.1 million in deposit volume on the books at 60 months (see Table 7). Nevertheless, the ranges shown in gures 2, 3, 4, and 5 hint at the diversity in lending and deposit mobilization strategies among greeneld MFIs. Some greeneld MFIs focus on institution building and prioritize the development of a full menu of credit, savings, and fee products early on, while others focus on expanding their footprint and prioritize the roll-out of a single loan product. For example, credit-only institutions focused primarily on group lending tend to reach a large number of

Growth Indicators of Greeneld MFIs

Month 12 Month 18 12,504 28 4.0 28 13,738 22 2.6 23 4.3% 28 Month 24 18,622 26 6.0 27 21,136 21 4.4 22 4.5% 27 Month 30 23,045 23 7.2 23 31,551 17 6.1 18 4.4% 23 Month 36 25,009 22 9.2 22 37,460 16 8.7 17 4.0% 22 Month 42 28,346 20 11.6 20 48,900 15 12.2 16 3.7% 20 Month 48 32,554 16 15.6 17 61,743 14 14.2 15 2.9% 17 Month 54 35,569 14 17.4 14 71,174 11 17.5 12 3.7% 14 Month 60 36,714 13 20.0 13 81,682 9 23.1 11 3.4% 13 9,495 29 2.3 29 7,123 23 0.8 24 3.9% 23

No. Loans Outstanding No. in sample Gross Portfolio ($ million) No. in sample No. Deposit Accounts No. in sample Deposit Volume ($ million) No. in sample PAR30 No. in sample 12

outstanding loans relatively quickly (though usually with a relatively modest overall loan portfolio volume). BRAC, in particular, has been very quick to scale up its lending programs and has managed to reach more than 100,000 outstanding loans in Tanzania and Uganda within 30 months of operation. The growth numbers in Table 8 are a reection of the ability of greeneld MFIs to build out distribution networks (mainly branches and outlets) and train staff while

maintaining robust operational control. At 12 months of operation, greeneld MFIs have on average 131 staff and nine branches; at 60 months of operation, they have on average 524 staff and 31 branches. It is important to note, however, that the rate of branch expansion varies greatly between credit-only institutions and regulated deposit-taking institutions. It requires much more planning and investment, and sometimes regulatory approval, to set up deposit-taking branches. Regulated deposit-taking institutions in the

FIGURE 2

FIGURE 4

Evolution of Number of Loan Accounts

120,000

Evolution of Gross Loan Portfolio ($US)

35,000,000 30,000,000 25,000,000 20,000,000 15,000,000 10,000,000 5,000,000 0 6 12 18 24 30 36 42 48 Time after launch (months) Smallest GLP 54 60

100,000 Number of loan accounts by MFI

80,000

60,000

40,000

20,000

12

18

24 30 36 42 48 Time after launch (months) Smallest number of loans

54

60

Gross Loan Portfolio by MFI ($US)

Largest number of loans

Median

Largest GLP

Median

FIGURE 3

FIGURE 5

Evolution of Deposit Accounts

Number of deposit accounts by MFI 140,000

Evolution of Deposit Volume ($US)

120,000,000

Deposit volume by MFI ($US)

120,000 100,000 80,000 60,000 40,000 20,000 0 6 12 18 24 30 36 42 48 Time after launch (months) Smallest number of accounts 54 60

100,000,000

80,000,000

60,000,000

40,000,000

20,000,000

12

18 24 30 36 42 48 Time after launch (months)

54

60

Largest number of accounts

Median

Largest deposit volume Smallest deposit volume

Median

13

TABLE 8

Stafng and Branches of Greeneld MFIs

Month 12 Month 18 196 26 14 26 4 17 33 9 59 26 Month 24 232 26 15 27 5 17 34 10 73 26 Month 30 303 22 21 22 6 13 43 9 73 22 Month 36 318 21 22 21 7 13 47 8 68 21 Month 42 351 19 24 19 7 11 48 8 72 19 Month 48 420 17 26 17 9 11 57 6 72 17 Month 54 445 14 29 14 10 9 63 5 71 14 Month 60 524 12 31 13 11 8 63 5 71 12 131 28 9 29 3 19 22 10 65 28

Number of Staff No. in sample No. Branches No. in sample No. Deposit-Taking Branches No. in sample No. Nondeposit-Taking Branches No. in sample Loans / Staff member No. in sample

cohort opened on average 11 branches in the rst ve years, whereas credit-led models opened 75. Staff development is critical for sustained growth. During the rst 34 years, successful loan officers are promoted to supervisors, branch managers, and regional managers, slowly replacing international staff (typically there will be only one, perhaps two, international staff left at 48 months). Several of the networks have created group-wide staff development programs that enable national staff to rotate to sister institutions to be trained or to train others. Staff exchanges are also used to support specic initiatives, such as SME lending or the launch of debit cards. Most greeneld MFIs recruit and train young adults who have little work experience and are new to the banking sector. Many new staff are recruited with basic high school math skills and are trained in cash-ow-based credit analysis and customer service, reecting the qualities that characterize the credit culture of most greeneld institutions. Marketing largely focuses on bringing banking services to clients, rather than having clients come to the bank. As such, employees are selected for their ability to relate to and communicate effectively with clients in markets and at the place of their businesses. Staff productivity levels have improved steadily among greeneld MFIs in SSA (as measured by the loans-to-staff ratio), but the cohort has nevertheless struggled to reach the same productivity numbers as in other parts of the world. This is probably due to many greeneld MFIs having a signicant number of noncredit personnel involved in banking operations, their relatively stronger focus on SME lending, and the continued rapid recruitment of new loan officers. However, it should again be noted that there is

quite a lot of variation in the greeneld MFI numbers, specically between entities based on group lending (which tend to be credit only) versus those based on individual lending (which tend to be deposit taking). Group lenders tend to have signicantly higher loans-to-staff ratios.

2.3Financial Performance
Financial performance is an integral aspect of the greeneld MFI model since many of its investors care almost as much about nancial returns as they do about development impact. For these investors, it is important to see a steady progression toward nancial sustainability through rising revenues, falling cost ratios, and improving margins and returns. Indeed, Table 9 shows that the greeneld MFIs in the cohort have been able to sustain fairly rapid revenue growth over their rst 60 months, increasing on average by $500,000 every six months and reaching $5 million by the ve-year anniversary. At the same time, they have managed to push operating expense ratios lower. However, this has not meant that the trajectory to nancial sustainability and protability has been entirely smooth. The averages in the table give the impression of a stable progression toward sustainability but, in fact, greeneld MFIs typically experience signicant swings from prots to losses and back to prots during this period.6 Many greeneld MFIs register substantial losses over the rst 24 months before achieving initial breakeven around 2436 months but then, as they begin to assume the full cost of any additional management service contracts, fall back into losses for the next 612 months. Only around months 4248 do they emerge fully self-sustainable.

6. T  he authors have not attempted to remove the TA support from the gures presented in this paper because the amount of the support is very difficult to precisely quantify and attribute among different accounting periods. 14

TABLE 9  Financial Total Revenue ($ million) No. in sample Portfolio Yield (%) No. in sample

Ratios of Greeneld MFIs

Month 12 0.62 28 59 28 200 28 (0.39) 28 -120 28 -12.4 28 -44.6 28 Month 18 0.98 26 55 25 108 26 (0.35) 27 -69 26 -8.8 27 -24.2 27 Month 24 1.59 25 56 23 82 24 (0.17) 26 -26 25 -4.1 26 -13.7 26 Month 30 1.98 23 56 20 57 21 0.01 23 -13 23 0.4 23 -0.3 23 Month 36 2.46 21 54 21 53 21 (0.03) 22 -11 21 -0.1 22 -0.4 22 Month 42 2.75 19 54 19 45 19 0.17 20 -5 19 1.8 20 -3.9 20 Month 48 4.03 17 55 16 38 16 0.42 17 10 17 3.3 17 20.0 17 Month 54 4.27 14 54 14 37 14 0.55 14 12 14 3.8 14 26.0 14 Month 60 5.02 13 52 13 36 13 0.40 13 8 13 3.1 13 18.9 13

Op. Expenses / Avg Portf (%) No. in sample Net Income ($ million) No. in sample Net Income / Revenue (%) No. in sample Net Income / Avg Assets (%) No. in sample Net Income / Avg Equity (%) No. in sample

Greeneld MFIs charge interest rates that typically generate portfolio yields around 55 percent for the rst ve years. The average operating expense ratio of greeneld MFIs shows steady improvement over time, falling to 36 percent at 60 months of operation. However, both portfolio yields and operating expense ratios are high compared to mature MFIs in other regions of the world, which ranged from 11 percent to 16 percent in 2011 (Rosenberg, Gaul, Ford, and Tomilova 2013). To align performance with MFIs in other regions, the greeneld MFIs in SSA will need to further reduce the operating expense ratio by 1020 percentage points, something that may not be easy given the high costs of doing business in the region. See Figure 6. Establishing new MFIs is a difficult endeavor, with many potential challenges and pitfalls. Some of the most common mistakes and problems that greeneld MFIs have experienced in SSA are highlighted in Box 4.

FIGURE 6

Evolution In Portfolio Yield of Greeneld MFIs

120 100 % Portfolio yield 80 60 40 20 0

12

18

24 30 36 42 48 Time after launch (months) Lowest portfolio yield

54

60

Highest portfolio yield

Average

15

BOX 4

Common Challenges and Pitfalls in Establishing Greeneld MFIs

The startup stage takes longer and costs more than expected. The preoperational stage is a complex undertaking that, among other things, includes interacting with regulatory authorities, identifying and renovating ofce space for branch operations, recruiting and training staff, tailoring policies and procedures, and conguring the IT platform. If cost overruns occur due to regulatory delays or poor planning or execution, less funding is available for the operational phase. Key management staff leaves prematurely. The CEO or other key staff sometime depart prematurely, in rare instances before the duration of their initial contract, but generally before the institution reaches the break-even point or scale up phase, either due to difcult living conditions, family reasons, or disagreements with the board or the technical service provider. Sometimes the problem is exacerbated by attrition of critical frontline staff, such as loan ofcers. Between 2006 and 2010 six CEOs of 12 greeneld MFIs in SSA departed prematurely. Given that management is one of the key success factors in greeneld MFIs, disruptions in the senior management team can create serious difculties and signicantly impact performance.
Source: CEO Interviews and Jansson (2010).

Overly aggressive expansion. Greeneld MFIs sometime undertake aggressive expansion, driven by the holding company, board, or management, in an effort to quickly reach a large loan portfolio and nancial breakeven. Such as the case for standalone MFIs, this generally leads to high levels of nonperforming loans. Sometimes this mistake simply involves making too many loans too quickly, but sometimes it also involves poorly managed or premature transition into SME lending before having built basic internal expertise. In some cases, it also involves establishing too many branches too quickly, which can also weigh on protability through higher asset depreciation. Growth exceeds funding. Signicant discrepancies are often discovered between projected and actual demand, particularly regarding deposit mobilization. Sometimes deposit growth takes off immediately, but more often it lingers at low levels for 1520 months before taking on a more expected trajectory. Rarely does it follow a smooth curve in line with loan portfolio growth. Clearly, a situation of slower than anticipated deposit mobilization can create a major bottleneck with regard to funding, so it is important that greeneld MFIs start out with one or two potential lenders closely associated with the venture (some DFI shareholders, such as IFC and FMO, can also ll the role of lender to greeneld MFIs).

16

SECTION

The Holding Company

ccording to the denition used in this paper, greeneld MFIs belong to a larger network or holding company, which through common ownership and management, plays a strong role in backstopping operations, providing standard policies and procedures, providing staff development and training, and co-branding the subsidiaries in the network. This section explains the holding company structure: who its investors are and how it is governed, its mission and key strategic choices in building out a network, and how management capacity is fostered and knowledge transferred.

3.1Structure Each of the holding companies is very much an extension of the historical operations and strategies of the group founders (the sponsors), reinforced by the investors who joined as initial shareholders. Three types of sponsor organizations are present in the holding companies covered in this study: specialized micronance consulting rms, micronance network support organizations (NSO), and a regional bank (see Box 5).7 Despite their distinct origins, the holding companies share several structural similarities, though there are some interesting variations. Sponsor investment in the holding company. The sponsor organizations are normally founding investors in the

holding companies. In the consulting-rm-led model, the sponsor (i.e., the consulting rm) typically holds a 320 percent minority ownership of the holding company. However, in the NSO-led model, the sponsors typically have a much larger stake in the holding company, often above 50 percent, as their shareholding in the holding companies came from contributing shares they held in a relatively large number of existing MFIs. BRAC and ASA are the only two sponsor organizations that have not invested in the holding companies, as Bangladeshi law prohibits nonprot organizations such as BRAC and ASA from owning shares in foreign companies, whether holding companies or operating companies. Sponsor management of the holding company. In several cases, the holding companies have been launched as investment companies with minimal if any administrative staff, managed by the sponsor organizations under a contract with the holding company and its shareholders. This is the case for networks led by consulting rms (Advans, Access, etc.), as well as some of the NSO-led ones (FINCA, Opportunity). The BRAC and ASAI holding companies have more inhouse administrative staff, presumably because they must function independently from the sponsor organizations. Management and technical support to greeneld MFIs. A primary role of the holding companies is to secure, directly or through an associated technical service provider, consis-

BOX 5

Typology of Holding Companies Creating Greeneld MFIs

Consulting rm led. Five of the holdings were founded by specialized consulting rms for the purpose of investing in and building a global network of subsidiaries. These initial sponsors, all based in Europe, wanted to be more than mere service providers; they wanted to be investors in branded micronance networks. The sponsor consulting rms and their related holding companies are LFS and AccessHolding, Horus and Advans, FIDES and Swiss Micronance Holdings (SMH), and IPC and ProCredit. MicroCred was started by PlaNet Finance, a nonprot organization with a consulting arm and relatively diverse micronance interests. NSO led. Four of the holding companies that create greeneld MFIs were established to consolidate the afliates of international micronance networks and expand with new greeneld MFIs. FINCA, BRAC, ASAI, and OI all belong to this category. Local bank led. Ecobank is a Togo-based bank (i.e., the holding company is based in Togo) with operations in 33 countries in SSA. It participated as a shareholder in Accion Micronance Bank in Nigeria and expanded its mass market operations by creating specialized MFIs in Ghana and Cameroun in collaboration with Accion International and Accion Investments in Micronance (AIM).

7. S  ome other large commercial banks that also provide micronance services, such as Equity bank, have expanded to new countries. However, these banks, or their subsidiaries, do not focus exclusively on the MSE market and therefore have not been included in the study. 17

tent and high-quality support to the greeneld MFIs during their different lifecycle stages. In the case of the consultingrm-led networks, this service is provided by the sponsor organization, i.e., the consulting rm behind the network. Several of the consulting-rm-led networks are considering (or may have already considered) migrating to an organizational arrangement where the technical capacity of the sponsor organization is merged with the holding company, to achieve even closer alignment of interests and a simpler organizational set-up that can more easily attract private investors in the future. Ownership of the subsidiaries. The holding companies in this study generally aim to hold at least 50 percent ownership in their greeneld MFIs. There are a few African greeneld MFIs where this is not the case (e.g., Accion Nigeria and Fides Senegal) but these cases are rare, and becoming rarer still. The key people behind the holding companies are increasingly focused on making sure that operational responsibilities and nancial incentives are appropriately aligned. Alignment would be weaker if the holding company, which is responsible for providing technical and management services to the greeneld MFIs (whether directly or through a linked service provider), does not have a shareholder stake in the greeneld MFI that is appreciably larger than any of the other investors, preferably a majority stake. However, if the holding company has limited nancial resources, it may face a trade-off between the degree of ownership control and the number of entities it can launch. It may therefore accept to hold less than 50 percent, at least for an initial period, if the minority shareholders are considered very like-minded. (See Figure 7.)

FIGURE 7

Holding Company Structure

Investors Sponsor/TA Provider

Shareholding

Investment Holding Company

Management Services MFI MFI

Lead Shareholder MFI MFI

Minority Investment & TA Funding

Note: This illustrates primarily the consulting-rm led model. In the case of others, the sponsor/TA provider would be together in the same box as the holding company.

3.2Investors DFIs have played a key role in creating and supporting most of the networks that launch greeneld MFIs today. For most of the consulting-rm-led networks, the holding companies began as a partnership between the sponsor and a core group of DFIs: EIB, IFC, KfW, FMO, and AFD. The group of investors has somewhat expanded over time, but it is still dominated by DFIs (see Table 3). DFIs also played a role in the transformation of the FINCA network into a holding company model, contributing signicant amounts of equity and debt funding to enable the transformation and continued expansion of their network MFIs. (See Box 6.) DFIs have actively supported the creation of the holding companies for several reasons. The holding company model has provided DFIs with a single vehicle for making larger investments in micronance and leveraging their participation with other investors. The holding companies are also seen as providing a relatively feasible exit route when DFIs believe their role has been completed, as shares in a geographically diversied holding company are thought to be easier to sell than multiple small investments in difficult countries. Equally important, the leading DFIs wanted to create commercial incentives to ensure the full engagement of the sponsors. The holding company arrangement has engaged the consulting rm and NSO sponsors as shareholders in the holding company, where they stand to gain or lose along with the other investors. The second largest group of funders in the holding companies and the greeneld MFIs is socially responsible MIVs. Organizations such as the OmidyarTufts Micronance Fund, Doen Foundation, Developing World Markets, responsAbility, Triple Jump, Incon, and Gray Ghost have all invested in the holding companies and/or in the greeneld MFIs. They tend to prefer the holding companies, for diversication and liquidity, and typically see their role more as providing expansion capital than venture capital. They therefore do not typically participate with equity in the founding stage of individual greeneld MFIs. However, they sometimes enter during the institutional development or the scale-up stage. Several MIVs also provide debt funding to greeneld MFIs. Private commercial investors have shown interest in the holding companies as well. Large companies such as Axa, TIAA-CREF, Sequoia, ABP, and Bank im Bustum Essen have made signicant investments. Small, individual investors are also present. It is possible to discern some different approaches to funding strategy and investor selection by the sponsors of the different networks. The consultant-led networks generally have DFI-dominated ownership, though Fides took a very different approach and sought out mainly individual investors with higher risk appetite and longer investment horizon than DFIs

18

BOX 6

DFI Funding for Holding Companies and Greeneld MFIs

Greeneld MFIs and their related holding companies have become an important vehicle for MFI investors. Many DFIs support the holding companies at both the holding and the greeneld MFI levels. As of the end of 2011, 7 percent of global DFI funding for micronance ($676 million) went to greeneld holdings, up from 4 percent in 2007. Most of the DFI investments at the holding level (87 percent) went to the consulting rm-led model, with the largest share going to ProCredit, while BRAC and FINCA essentially received the other 13 percent in 2011. KfW, OPIC, and IFC provided three-quarters of all the funding, with most of the remaining amounts coming from AFD, Proparco, FMO, EIB, BIO, and CDC.
FIGURE B6-A DFI Funding (US$ million, Dec. 2011)
11.3 11.8 17.9

Grant Debt

119.1 Greenfield MFIs 192.6 Holdings

Guarantee

21.5

Equity 0

43.6 100 200 USD million 300

353.3 400

Greeneld MFIs represented more than a quarter of the total number of all DFI direct investments in nancial institutions in SSA (25 of 91) at the end of 2011. Close to half of the funding directed at greeneld MFIs was invested as equity in 25 entities. In terms of debt, nine received a loan and eight a guarantee from a DFI. Eighty percent of the DFI funding went to MFIs of the consulting-rm-led holdings and the other 20 percent went to afliates of FINCA, OI, and Accion. At the end of 2011, IFC, AFD/Proparco, and KfW accounted for 64 percent of all direct DFI funding to the individual greeneld MFIs. Other public funders in addition to those invested in the holdings mentioned above include the African Development Bank and DCA USAID (for guarantees).
Source: 2012 CGAP Cross-Border Funder Survey

when it created Swiss Micronance Holding in 2007. DFIs may eventually be invited to support SMHs next (expansion) phase. It is also important to note that in most consultingrm-led networks there is substantial nancial participation by the individuals behind the sponsor companies.

NSO-led companies have taken a mixed approach. When FINCA created FINCA Micronance Holding (FMH) in 2010 it focused heavily on identifying like-minded investors willing to weigh social and nancial performance equally. As such, FINCA International (NGO sponsor) maintained a majority stake in FMH, and a combination of DFIs and socially responsible investors (IFC, KfW, FMO, responsAbility, and Triple Jump) invested in the remaining shares. OI formalized its network in 1998 when all affiliated partner organizations signed a membership agreement with the OI network. OI programs are nanced through direct solicitation of funds from individuals, corporations, foundations, and religious organizations. OI created Opportunity Transformation Investments (OTI), a wholly owned subsidiary, to invest in and hold ownership positions in OI affiliates. ASAI and BRAC have chosen to partner mainly with MIVs and institutional investors. ASAI is 100 percent owned by Catalyst Micronance Investors (CMI), a Mauritius-based company whose shareholders include more socially responsible private investors, pension funds, and MIVs than DFI funding. Much of the funding for BRAC subsidiaries comes from the BRAC Africa Loan Fund, a Cayman Island vehicle established with funding from OPIC, MIVs, and socially responsible investors specically interested in supporting BRACs expansion. The case of Ecobank is naturally different, as the group was not created for the purpose of extending micronance services. Consequently, the investors of the holding company do not represent the same focused interest in micronance as can be found behind the other networks. Rather, they are fairly mainstream institutional investors with rather broad agendas, raising the question of how micronance will be accommodated and supported in the larger organization. All of the holding companies require at least a majority share in the network MFIs to exercise control, which they consider to be a critical success factor. The consulting-rmled holding companies own majority shares in the MFIs in partnership with minority investors, typically the same DFIs that have invested in the holding company. In the FINCA network, the holding company owns 100 percent of the MFIs. OTI is the majority owner of the OI subsidiaries, and the remaining shares are divided among the individual OI companies in Canada, Australia, and the United Kingdom. The BRAC and ASAI holding companies also strive to own 100 percent of subsidiaries but have made exceptions where legally required. Ecobank has opted to establish subsidiary nance companies to clearly separate micronance operations from Ecobanks core retail business. These subsidiaries are typically owned and controlled by the local Ecobank subsidiary together with the Ecobank International holding company, but Ecobank has also included a few strategic partners as minority investors, notably AIM and IFC. (See Box 7.)
19

BOX 7

Accion Investments in Micronance

Ecobanks micronance greenelding ventures were undertaken in collaboration with AIM and Accion International. The idea was to use Ecobanks footprint in the region andin the creation of greeneld MFIsand combine it with Accion Internationals technical expertise and AIMs investment capital. Under this model, Accion International would provide initial capacity-building services through three-year TA programs, while AIM supported the ventures nancially through signicant minority equity stakes.a This approach suited all parties: it gave Ecobank a chance to roll out micronance operations under experienced supervision; it allowed AIM to invest with a 58 time horizon (which coincided with its fund life), and it enabled Accion International to expand its activities in Africa through a high-visibility partnership with multiple replication opportunities. The collaboration has produced two greeneld MFIs so far: in Ghana and Cameroon.b Ecobank also participated as a minority shareholder in a greeneld MFI in Nigeria launched by Accion. The EcobankAccion model involves an important challenge in that the operational responsibility (which lies with Accion International) and the nancial responsibility (which lies primarily with Ecobank) are not completely aligned. It is to avoid this challenge that many holding companies are closely linked with the technical service provider and insist on taking majority stakes in the greeneld MFIs. Over time, Ecobank and Accion will need to address this challenge, particularly if they intend to create additional greeneld MFIs. Lately Ecobank and Accion have pursued new micronance projects in SSA independent of one another.

a. In 2012 the shareholders of AIM sold their shares to Bamboo Finance (a global private equity group specializing in micronance and social entrepreneurship) and, in the case of AIMs Africa portfolio companies, to Accions Gateway fund. b. While Ecobank has a majority stake in the projects in Ghana and Cameroon, the shareholding in Nigeria is more fragmented (it has a total of six shareholders). However, Ecobank is the largest minority shareholder with 24 percent, and Ecobank is an important investor with 18.5 percent.

Governance. Governance for greeneld MFIs is meant to support long-term growth and development. Viewed as one of the primary benets of this model, high-quality and consistent governance practices provide a strong foundation and enabling operating environment from inception and throughout the institutions lifecycle. Particularly critical to governance is the ownership structure and board composition, which inuence key strategic decisions related to the shareholders agreement, articles of association, and key operating principles. The consulting-rm-led model lends itself to a governance discussion as it accounts for just over half of the cohort analyzed in this paper, and it provides a homogeneous structure for analysis across a number of institutions. In consulting-rm-led greeneld MFIs, the ownership structure includes the holding company, likeminded investors comprised mainly of DFIs and, in some cases, local shareholders, who appoint directors based on their level of ownership. Generally the approach has been not to appoint executive directors; however, in some cases regulatory requirements for resident directors have necessitated the need, at least temporarily, for nonvoting executive directorships. Independent directors are identied as management, and shareholders become more familiar with expertise that exists in the market. Board committees are gradually expanded as the MFIs operations become more complex.
20

During the foundation stage, the shareholders agree on articles of incorporation and the shareholders agreement, which detail the mission and governing principles of the institution, board structure, and voting requirements. Majority and super majority votes are generally stipulated for strategic decisions, including the issuance of new shares, entry of new investors, changes to primary operating policy guidelines of the institution, and the procurement of technical and management services. The large number of decisions requiring a super majority vote by the board and/or the shareholders attests to the importance attached to shared objectives and like-mindedness in these projects. A management services contract (MSC) is often used to structure and budget TA services from the holding companies, and/or the associated consulting rm, to the greeneld MFIs. Backstopping from the holding company and the technical services provided by the related consulting rm are critical to the MFIs operating continuity, staff development, and adherence to core operating policies and procedures. The board and shareholders play an important role in governing the potential conict of interest that could otherwise arise from having staff from the consulting rm (acting on behalf of the holding company) involved in the decisionmaking process related to the procurement of technical services. Directors appointed by the holding company (which often is partly owned and managed by the consulting rm) typically recuse themselves when the board votes on the

MSC. In other cases, decisions regarding the MSC are referred to the shareholders of the greeneld MFI, which typically include several minority shareholders. Notwithstanding the potential conicts, the consulting rms ownership stake in the holding companies and, therefore, indirect ownership in the greeneld MFI, helps align nancial interests.

3.3Mission and Strategy Without exception, the networks in the study are driven by a mission to expand access to nancial services. They aspire to build retail mass market nancial institutions that serve populations neglected by mainstream banks, especially micro, small, and medium-sized enterprises (MSMEs). Not surprisingly, the extremely low-level of nancial inclusion in SSA has been a key consideration in the networks decision to enter and expand in the region. But they approach the market in different ways, in terms of country selection, preferred institutional license (bank vs. NGO vs. nonbank nancial institution), and commercial orientation. Country selection. The holding companies typically point to market characteristics, personal security of staff, political and economic stability, features of legal systems, and quality of nancial sector supervision as signicant factors in determining country selection decisions. At the same time, each of the holding companies has specic criteria for selecting a market conducive to its unique implementation model. Most of the holding companies seek markets capable of supporting an efficient scale of operations and some of these groups explicitly exclude smaller markets where they feel that the impact and potential scale do not justify the investment. On the other extreme, some holdings are taking a longterm approach to rural areas where they are less likely to face competition. Most of the networks want a market that supports diversication into the SME segment. They contend that the SME market is also underserved and that developing SME products enables MFIs to maintain a partnership with clients as they mature, allowing for continuity in the nancing relationship while protecting against attrition of the longest standing and most protable relationships. Finally, several groups deliberately seek a footprint that will allow for reasonably easy replication and sharing of expertise and resources, for example, by focusing on a particular subregion. ASAI and BRAC are focused specically on English-speaking, common-law jurisdictions because it facilitates transfer of the model they have developed in Bangladesh. MicroCred is increasingly focusing on French-speaking West Africa, where the regulatory environment is relatively uniform. While country selection is usually driven by the management and board of the holding companies, in some cases DFIs and MIVs have made efforts to steer networks to certain markets by promising signicant and consistent support

to the new greeneld MFIs. This was the case when Access Holding launched a greeneld MFI in Liberia, for example. Institutional license. With only one exception, all of the holding companies stress the importance of operating as an institution that is permitted to mobilize deposits. Some of them have been willing to start with credit-only institutions but, when doing so, with the clear expectation to acquire a deposit-taking license eventually. For these networks, deposits are both a necessary service to customers and an important funding source for growth. This approach does not apply to BRAC, since its model in Africa focuses on delivery of credit and social services. To secure funding for its greeneld MFIs, BRAC has created a debt-funding facility within its group structure. The networks typically have two options to obtain deposit-taking capabilities in a particular country: a deposit-taking micronance license or a commercial bank license. Some of the consulting-rm-led networks have opted for commercial bank licenses wherever possible, because they have the aspiration to create full-service banks capable of meeting most, if not all, nancial needs of MSMEs and low-income households. However, minimum capital requirements for commercial banks are increasing across Africa, leading some networks to opt for micronance deposit-taking licenses with the intent to transition to commercial banks over time. Commercial bank licenses are, of course, subject to signicant compliance cost and complexity, so not all networks are willing to go that route when there are deposit-taking micronance licenses available that can get the basic job done. Commercial orientation. The holding companies are double-bottom-line investors, but they articulate their investment return expectations with different levels of precision. Initially, the consulting-rm-led holding companies expected to generate rates of return centered around 1314 percent for their investors in a 68-year timeframe. The DFIs established this benchmark at the founding of the consulting-rm-led holding companies, and subsequent investors have entered with similar, if not higher, expectations. However, there have been few exits from the holding companies and it seems likely that the investment time horizon may be a few years longer than initially anticipated. It simply takes more effort and more time than anticipated to build successful MFIs of signicant size. As noted earlier, Fides has deliberately sought out investors with a longer return horizon (10 years) and a more modest return expectation (810 percent) for SMH, which it feels is necessary when operating in the rural market. The NSOled holding companies speak in more general terms about expecting reasonable returns, and stress their primary commitment to social objectives. A common thread across all networks, regardless of their nancial return expectations, is their aspiration to be so21

cially responsible. They all have a memorandum of association or articles of incorporation that dene access to nance for micro and small entrepreneurs as a primary mission. Some networks, in particular BRAC, take this further and attempt to provide or facilitate health and education services. Such services are generally linked to a creditonly approach to micronance.

3.4Institutional Capacity and Knowledge Transfer The holding companies and sponsors typically take a proactive role in building institutional capacity among their greeneld MFIs. They do so by providing technical and management services to the MFIs, which in turn are specied in a management services contract or a TA agreement. There are some differences in how the networks have arranged the delivery of these services. In the MicroCred network, for example, the MFIs procure management services and TA directly from the holding company. In the SMH, Advans, and Access Holding networks, MFIs procure much of this from the sponsors (Fides, Horus, and LFS, respectively). MFIs operating under BRACs Dutch holding company also procure services directly from the sponsor, BRAC Bangladesh. FINCA MFIs obtain TA, oversight, and services from FINCA International and FMH in exchange for dividend payments. The long-term vision for service delivery in the Ecobank model is less clear, as Ecobank has not yet developed comprehensive capacity to provide tailored expertise to its network MFIs. Regardless of these differences, the knowledge transfer from the holding company or the sponsor is usually channeled through three main mechanisms: (i) the international management team; (ii) the systems, policies, and procedures; and (iii) the training of national staff. Management team. Holding companies and/or sponsors usually provide anywhere from two to six managers for the rst three years of operations of a greeneld MFI. Typically this includes the CEO, the chief operating officer, the banking services manager (if the MFI takes deposits), and one or two temporary branch managers. These managers are meant to provide skill and experience during the initial stage of operations. But they are also expected to provide internal coherence and a strong culture that reect the values and mission of the network. They are as much role models and advocates of corporate culture as they are technical experts. This management team typically has a lot of authority in guiding the delivery of specic technical services to the MFI: what services are needed, when they are brought in, and how they are implemented. Over time, most of these managers are replaced by national staff who have risen through the ranks and proven their capacity to execute the required responsibilities. Systems, policies, and procedures. Another key advantage conferred by the holding company/sponsor is a stan22

dardized set of systems, policies, and procedures: the IT platform, internal control and risk management policies, and lending procedures, among others. Many networks have gone through an extensive process of identifying, vetting, and rening appropriate business practices for its MFIs. The success of greeneld MFIs is in no small measure connected to the effective implementation of such practices, and it is often cited as a critical success factor by the holding companies themselves. One holding company proudly calls itself the McDonalds of micronance, and contends that systematic implementation of standardized systems is the key to cost and quality control. In particular, the adaptation and implementation of a common IT platform across the network can yield signicant advantages compared to standalone MFIs, in the form of better support services (provided in part by the holding company or the sponsor), lower licensing costs, and greater ability to manage complex system requirements. This is particularly important for greeneld MFIs, which have plenty of other challenges to worry about. Almost all of the networks in this study have implemented or are in the process of implementing a single core banking system for their MFIs. Training national staff. Holding companies identify human resource development as the greatest challenge they face in SSA, more so than in other regions. The most common approach is, like in other regions, to hire young adults with little work experience and develop their skills and capacity over time. Given that the average greeneld MFI has about 300 employees after 30 months of operations, and more than 500 after 60 months, the training need is enormous. And many times the training must go back to the basics, particularly in post-conict countries. Often basic arithmetic skills need refreshing before staff can evaluate and present loan proposals. Most networks have relatively wellestablished training modules for every key aspect of the business, facilitated by standard systems, policies, and procedures. It is also common practice for networks to move employees among network MFIs as a way to build and transfer knowledge within the group. In addition to staff training that occurs at the MFIs and staff transfers between the MFIs, many networks are setting up regional hubs or regionally based staff that support MFIs through technical expertise, backstopping, and management oversight. FINCA was one of the rst to set up a regional hub, which in addition to providing technical backstopping, has direct management responsibility of MFI CEOs, effectively decentralizing management. Similarly, but with a less robust approach, OI and Accion have regional heads based in Africa. Other examples range from a training academy ProCredit set up in Ghana to the regional decentralization of IT systems housed in Ghana for Advans and in Senegal for MicroCred. (See Box 8.)

BOX 8

Cross-Fertilization in the Holding: Example of Advans

Holding companies increasingly rotate national staff among their network MFIs as a motivational measure, to build a consistent group culture, and to share knowledge on methods, standards, and procedures. Younger institutions benet from expertise and culture built up in more mature greeneld MFIs at low or no added cost to them. The rotating staff remain either fully paid by the more mature MFI or the host MFI takes on the direct costs of salary and travel. These rotations offer interesting growth opportunities for national staff, who start seeing the possibility of an international career. One example is the assistance of Advans Cameroun, which was launched in 2007, to Advans Cote dIvoire, which opened its rst branch in March 2012. Advans Cote dIvoire relied extensively on support from Advans Cameroun to prepare and launch operations. Two Cameroonian branch managers worked in Cote dIvoire, and a group of Ivorian loan ofcers got 4.5 months on-the-job training in Cameroun. One of the internal trainers from Advans Cameroun coached client ofcers in Cote dIvoire. The exchange proved to be a winwin solution. Clearly the arrangement helped Advans Cote dIvoire get off to a better start. But the staff exchanges also had a positive effect on staff loyalty in Advans Cameroon, according to its CEO. Attrition rates have declined in recent years as staff recognize international growth opportunities within the network, which are not offered by other MFIs or even local commercial banks in Cameroun. In addition to staff exchanges, procedures and training material are shared across the Advans network. The CEO of Advans Cote dIvoire estimates that it saved two to three months in international staff time compared to earlier greeneld MFIs, because of materials available from Advans Cameroun. Other benets include pilots with new products and delivery channels, which are tested in one MFI and then shared with others; this spreads the development costs over more entities.

Source: Interviews with CEOs of Advans Cameroun and Cote dIvoire (December 2012).

3.5Success Factors and Common Challenges Standardization of operational policies, procedures, and systems are the critical operational success factors most often cited by the holding companies themselves. They also cite close oversight and consistent reinforcement of internal controls as critical to achieving a sustainable institutional culture in greeneld MFIs. In the long term, the holding companies contend that human resource development is key to success. In addition to these factors, core strategic factors include alignment of shareholder interests, clarity of vision and mission, sponsor commitment, the regulatory and business environment, and sufficient resources. When transparent, these core strategic factors lead to success; how ever, when muddled, they become signicant challenges to MFI sustainability. Shareholder alignment. It is important that the shareholders involved, at the holding level and at the MFI level, have a similar long-term vision for the network and its MFIs. If the long-term vision is shared and agreed to, most disagreements will be about tactics and can be easily resolved. If the long-term vision is not shared, or if it is unclear, fundamental disagreements can arise and can require a lot of effort and energy to resolveeffort and energy that could be better used to improve the MFIs operations. Some networks have learned the hard way. Clarity of vision and mission. Clarity of vision and mission that is centered on creating and managing MFIs leads to

clarity in organizational priorities. Clear organizational priorities enable coherent long-term decision making for building appropriate expertise and capacity at the holding company, including human resources, IT system capabilities, tools and methods, and means of sharing knowledge. Capable networks tend to be guided and supported by a holding company that has a focused, long-term perspective, that enables its management and staff to invest in the right expertise, organize teams purposefully, and institutionalize effective knowledge transfer practices. Holding companies or sponsors that are not focused on creating and managing MFIs are not very likely to be effective in greeneld MFIs, given the difficulty and complexity of this business model. Sponsor commitment. As noted earlier, strong commitment can often come from a strong sense of mission. However, to reinforce this commitment, holding company structures typically aim to make sure that the sponsor and key decision makers and managers assume a nancial stake in the network. This becomes particularly important when things are not going as well as hoped. In those instances, the MFI may not be able to pay for all services required, underscoring the importance of having a sponsor and key staff that are willing to go the extra mile without necessarily being immediately compensated for it. For example, some sponsors will send additional junior consultants, or consultants with strong technical skills but limited eld experience, to support the early operational stage of greeneld MFIs without charging for it. In other cases, staff from mature network
23

MFIs are seconded on a subsidized or direct-cost basis to younger MFIs. Not only does this support the operations of the MFIs, but it also builds the experience and expertise of network and holding staff, which is likely to yield additional returns to the network. Enabling and predictable regulatory and business environment. Several networks identify the regulatory and supervisory regimes as a challenge to their activities. There is broad frustration with protracted licensing procedures, difficult communication with regulatory authorities, and a general lack of supervisory competence that, in some cases, has undermined the sectorss credibility and stability. The greeneld model discussed in this paper relies less on external supervision, since it is designed with very strong internal controls and governance structures. In addition, greeneld MFIs in Africa face signicant challenges related to the business environment. Political instability has been a constant threat and, in several cases, has imposed heavy costs on greeneld MFIs and holding companies. Personal security is also a concern, as it adds to cost of operations and makes it more difficult to attract international talent as necessary. Several networks also cite the challenges of their international staff in adapting to local

conditions. These problems are particularly acute in postconict countries. Sufcient resources. For many of the aforementioned reasons, the holding companies report that the cost of doing business in SSA is decidedly more expensive than the cost of their operations in other parts of the world. As a very simple example, the cost of preparing a deposit-taking branch in Eastern Europe is estimated at $50,000 whereas in SSA it ranges from $150,000 to $400,000. But other operating costs are also higher, including communications, transport, and security. The NSO-led holding companies, which have decades of experience in other parts of the world, also point to the challenge of creating MFIs in SSA under ever increasing expectations about nancial performance. Available grant funding is more modest compared to when many Latin American MFIs launched. At the same time, investors are more demanding about reaching breakeven in a three- to four-year timeframe. Several of the holding companies signaled some concern about overly ambitious expectations from DFIs about nancial performance in a short timeframe. They see MFIs concentrating in urban centers and a trend toward consumer credit products that is elevating the danger of over indebtedness in some markets.

24

SECTION

The Role of Greeneld MFIs in Market Development

he following section provides insights into market developments associated with the start-up of greeneld MFIs in the DRC, Ghana, and Madagascar.8 These three markets each have several greeneld MFIs and represent different country contexts in terms of nancial sector development. In each country at least two greeneld MFIs have been operational for more than ve years, increasing the likelihood that effects of their interaction with the market can be observed. This research shows that greeneld MFIs play various roles in the development of the market for nancial services for those at the bottom of the pyramid. In addition to improving access to nance they also increase the level of skills in the nancial sector, introduce new products and channels to the market, and expand the number of access points for clients. The analysis in these markets is largely based on interviews with stakeholders, direct data collection from the greeneld MFIs, and secondary data sources for the micronance and broader nancial sector in each country. While it is difficult to attribute changes in a market comprised of many institutions to the intervention of one or more greeneld MFIs, the authors used mostly qualitative and some quantitative information to examine the effect they have had. Throughout the text, it is indicated where the anecdotal discovery method supports assertions, but also where this may not be the case.

4.1Market Relevance The greeneld MFIs in the DRC, Ghana, and Madagascar represent only a small portion of total nancial sector assets, but they are signicant players in terms of numbers of households and enterprises served. In some cases, they also manage a signicant number of branches and employ a signicant number of employees relative to the nancial sector overall. See Table 9. These effects are observed most clearly in countries with a less developed nancial sector. In post-conict DRC, the

four greeneld MFIs served 89,942 microenterprise borrowers and 265,714 depositors at the end of 2011, representing approximately 50 percent of all borrowers in the micronance sector and 65 percent of the depositors served by MFIs.9 In that year, the micronance sector counted 146 cooperatives and 16 local MFIs in addition to the four greeneld MFIs. When looking at the mainstream nancial sector, the two greeneld MFIs with banking licenses represent a signicant market share of loans and deposits. In 2011, ProCredit DRC and Advans DRC accounted for more than 20 percent of all deposit accounts in the commercial banking sector and 13 percent of all commercial bank loans. When adding loan accounts from the other greeneld MFIs, the market share increases to 61 percent of all loans in the nancial sector.10 Their deposit volume and loan portfolio represented about 7 percent of the nancial sector. However, more remains to be achieved as the penetration rate remains low overall with only 5 percent of the Congolese adult population banked.11 In Madagascar AccsBanque and MicroCred together held US$33 million in deposits and US$35 million in loan portfolio at the end of 2011. This represented only 2 percent and 5 percent of the banking sector, respectively, but in number of deposit and loan accounts the two greeneld MFIs accounted for 25 percent and 17 percent, respectively, of the 11 banks in the system. When compared to the 36 other MFIs in the country12 the two greeneld MFIs held almost half of the entire microcredit portfolio and close to a quarter of MFI deposit balances. AccsBanque and MicroCred have clearly contributed to the rapid growth of micronance in Madagascar, which reached a penetration rate of 21 percent in 2012 compared to 14 percent in 2008. The banking sector (excluding the two greeneld MFIs) reached only 3 percent of the population in 2012 (up from 1.8 percent in 2008). In Ghana, a more advanced, competitive, and diverse nancial sector compared to DRC and Madagascar, greeneld MFIs represented ve of the 19 savings and loans com-

8. G  reenelds included in the analysis were in DRC: Advans (2009), FINCA (2003), ProCredit (2005), and OI (2010); in Ghana: EB Accion (2008), Advans Ghana (2008), ProCredit (2002), OI (2004), and ASA (2008); and in Madagascar: ABM (2007) and MicroCred (2006). Market analysis has been performed by Yaw Brantuo for Ghana, Hannah Siedek for DRC, and FTHM Conseils for Madagascar. 9.  Fund for the Financial Inclusion in DRC (FPM), Synthesis of the political, economic, and nancial evolution of DRC. Facility for nancial inclusion in DCR. Third trimester 2012. Kinshasa, DRC. 10. Data based on the International Monetary Fund (IMF) Financial Access Survey (FAS) database, which includes information for commercial banks only. 11. FINDEX, http://siteresources.worldbank.org/EXTGLOBALFIN/Resources/8519638-1332259343991/N4ssaEN_08202012.pdf, accessed January 2013. 12.  For this market analysis the two greeneld MFIs in Madagascar are compared to the 36 MFIs (ve etablissements nancier and 31 institutions de micronance agrees) as their clientele, and methodology is more aligned than with the other nine commercial banks in the country. 25

TABLE 9

Market Share of Greeneld MFIs in DRC, Ghana, and Madagascar, 2011

Deposits
Total deposit volume (USD) Number of deposit accounts 317,217 21.2 514,258 7.2 133,538 25.2 Gross loan portfolio (USD) 71,005,502 6.7 90,424,087 1.5 35,276,058 4.8

Loans
Number of loan accounts 93,640 60.9 149,669 21.7 37,072 17.3

DRC Greeneld MFIs % of commercial banks and Greeneld MFIs Ghana Greeneld MFIs % of commercial banks and Greeneld MFIs Madagascar Greeneld MFIs % of commercial banks and Greeneld MFIs

141,917,898 7.1 82,619,026 0.9 33,351,084 23.4

Notes: For the DRC and Ghana, the FAS database includes data for commercial banks only and not data for other nancial institutions such as those in Madagascar. To calculate market shares, the data of greeneld MFIs that are not included in FAS have been added to commercial bank data to establish the denominator (FINCA and OI in the DRC and all greeneld MFIs for Ghana). For Ghana the share of loan accounts for greeneld MFIs could be somewhat overestimated as FAS data included number of borrowers for commercial banks only, which tends to be lower than the number of loan accounts. Sources: Greeneld MFIs and IMF FAS database, 2011.

panies.13 Together the greeneld MFIs served more than 150,000 borrowers and over 540,000 depositors at the end of 2011, which represents around 5.1 percent of all adults or 2.8 percent of the entire Ghanaian population (see Figure 8).14 Their deposit volumes remain modest as a percentage of the nancial sector (in this case, dened as commercial banks plus savings and loan companies) at 0.9 percent but they hold 7 percent of deposit accounts. Similarly, their
FIGURE 8

outstanding loan portfolio is modest at 1.5 percent of the nancial sector while their loan accounts represent a substantial 22 percent (see Figure 9).

4.2Skills Building By their own account, the greeneld MFIs most signicant effect on market development is through their contribution to the professional development of staff in the banking
FIGURE 9

Number of Deposit Accounts, December 2011

8,000,000 7,000,000 6,000,000 5,000,000 4,000,000 3,000,000 2,000,000 1,000,000 0

Number of Total Loans Outstanding, December 2011

800,000 700,000 600,000 500,000 400,000 300,000 200,000 100,000

Madagascar

Ghana

DRC

Madagascar

Ghana

DRC

Greenfield MFIs

Commercial banks excl greenfield MFIs

Greenfield MFIs

Commercial banks excl greenfield MFIs

13.  At the time of this study, ASA was in the process of transforming from an NGO to savings and loan. For the purpose of the study it has been included among the greenelds, which together are compared with industry data for deposit-taking banks in Ghana (including commercial banks, rural banks, and other banking and quasi-banking institutions). 14.  To put this in perspective, in 2005, all mainstream commercial banks together had an estimated penetration of 5 percent of the overall population. No more current comparative information was available. 26

and micronance sectors. According to market participants, the greeneld MFIs have introduced superior human resource practices that positively impact the nancial sector. With the exception of a small number of international staff, all 11,600 employed in greeneld MFIs as of December 2012 are nationals. The number of people employed and trained by the greeneld MFIs is therefore becoming signicant in relation to the overall nancial sector in many countries. In Ghana, greeneld MFIs employed more than 2,000 staff in 2011 while the mainstream banking sector employed 16,000 staff. The two greeneld MFIs in Madagascar have more than 1,000 staff, which represents 23 percent of staff in the micronance sector and almost 19 percent of banking sector employees. The employees of greeneld MFIstypically young adults who have little or no previous work experiencereceive extensive training in several topics and skills related to credit and banking. Eventually they become attractive candidates for mainstream banks, and their skills gradually are incorporated into the larger market as their careers bring them to other institutions. The positive results to the nancial sector from the large investments in staff training and development by greeneld MFIs reduces the potential market distortion from providing TA grant funding to individual institutions. Greeneld MFIs typically have an intensive and systematic approach to staff selection, recruitment, and training. They spend 35 percent of their operating budget on staff development. This is where a signicant portion of the initial TA resources is invested. Most greeneld MFIs have company-specic training facilities that offer courses for induction and professional development. They also offer intensive on-the-job training. Many of the people interviewed in the DRC, Ghana, and Madagascar commented on the high quality of the training offered by greeneld MFIs. Mainstream banks and other nancial institutions appear to agree because they frequently try to poach staff from greeneld MFIs. Some holding companies calculate that they will train two to three times the number of required staff to address expected attrition to local nancial institutions. Staff turnover rates reported by the greeneld MFIs varied between 8 percent and 20 percent.15 For example, Fidelity Bank Ghana hired staff of ProCredit to support its branch roll out. Several of the banks that entered the SME space in the DRC are also run by former ProCredit staff. (See Box 9.) Staff compensation varies across markets. The greeneld MFIs in the DRC tend to offer less attractive packages than mainstream banks and some MFIs. In contrast, salaries of-

BOX 9

ProCredit Young Bankers Program

ProCredits Young Bankers Program is an intense, half-year training on banking and nance for university graduates with little or no practical work experience. ProCredit reported that its recruiters for the program look for individuals with good analytical, organizational, and communication skills who are capable of solid quantitative analysis, have demonstrated ability to think logically and critically, and can work effectively in teams. Such individuals are also required to show a clear desire and ambition to learn and develop in the profession. Those selected to participate in the program are taught basic mathematics, accounting, and other relevant banking subjects. But more importantly, the training emphasizes both theoretical and practical training with a view to instilling good banking practices and the ProCredit culture and methodologies for doing business. This sort of investment in staff training is complemented, for example, by advanced training in ProCredits regional training center for middle managers in Macedonia and its international training center in Germany, which provides a three-year part-time training course for senior managers.
Source: Yaw Brantuo from interviews and ProCredit website [www.procredit-holding.com]

fered by greeneld MFIs in Madagascar appear to be higher16 than the average for the micronance and banking sectors. According to those interviewed, this has had a notable effect on some of the MFIs in Madagascar, especially ACEP and BNI-CL, which have seen staff leave for better opportunities at the two greeneld MFIs. The greeneld MFIs appear to provide a career bridge between the less formal micronance sector and the more formal banking sector.

4.3Product and Channel Diversication Greeneld MFIs tend to be at the forefront (compared to other MFIs) of introducing innovation in low-income retail banking. Greeneld MFIs have introduced new products, credit policies, and service standards that have been replicated by other nancial institutions. For example, in the DRC, ProCredit introduced free savings accounts without a minimum deposit requirement at a time when most banks had minimum requirements of more than US$1,000. ProCredit attracted large numbers of savers and demonstrated that the Congolese population was able and willing to save. Following this exam-

15.  This compares with turnover rates of only 23 percent for other MFIs reporting to MIX in the DRC and Madagascar. In Ghana, turnover rates appear to be high (1020 percent) for most MFIs that reported these data to MIX. 16.  In Madagascar this is reported to be 20100 percent higher, depending on the function. For example, a loan officer with two years of experience and a baccalaureate received an average salary of US$120 in an MFI, US$150 in a mainstream bank, and US$210 in a greeneld MFI. 27

ple, other banks, such as Rawbank and BIAC, relaxed their account-opening requirements, and the number of deposit accounts in the DRC has grown from 30,000 in 2005 to 1 million in 2012. Similarly, Malagasy MFIs adapted their internal procedures, processes, and IT systems to keep up with the new greeneld competition, evidenced by the reduction in loan processing times from weeks to ve days. In Ghana and the DRC, greeneld MFIs were the rst to introduce new technologies in banking for low-income populations. EB-Accion, Opportunity, and ProCredit introduced ATMs in Ghana, which were previously available only at commercial banks. EB-Accion Ghana and Advans Ghana, for their part, have introduced mobile deposit collection (using cell phones for instant verication) to offer clients additional convenience and to compete with traditional susu collectors. In the DRC, ProCredit established the rst ATMs, and mainstream banks soon followed. Clients in the DRC now have access to point-of-sale (POS) devices at over 300 locations, facilitating the withdrawal of funds and cashless purchases. Some greeneld MFIs have pioneered the development of nancial services that are perceived as particularly risky and challenging in their markets, such as microinsurance and ag-

ricultural nance. OI started an agricultural nance program in Ghana in 2010 with a pilot credit scheme for cocoa farmers. It now serves 9,000 farmers and has introduced geographic information system (GIS) technology to more accurately map the smallholder farmers (see Box 10). In a few cases, the success in SME lending of greeneld MFIs has attracted other providers into this market segment. According to observers in the DRC, the performance of ProCredit, which reached nancial sustainability in three years, triggered banks such as BIC and TMB to downscale and serve the SME segment. To acquire the necessary expertise, they relied on former employees from ProCredit to roll out these services. In Madagascar, Bank of Africa (a shareholder of MicroCred) and BFV-Societe Generale (a shareholder of AccsBanque Madagascar) have started to modestly downscale in the past two years. Whereas ve years ago Malagasy SMEs were not served at all, there is now an increased offer of services from a range of nancial institutions.

4.4Standards and Good Practices Greeneld MFIs can play an important role in market development by demonstrating professionalism and good practices. Evidence of this is not easy to establish but some signs

BOX 10

Greeneld Banks Pioneering with Smallholder Finance

Smallholder nance is typically avoided by mainstream banks and considered too risky. Several of the more mature greeneld MFIs, such as ABM in Madagascar and OI in Ghana, have piloted agricultural nance products to address this gap. After four years in operation, AccessBanque Madagascar launched an agricultural loan product Agro loan for smallholder farmers who mix subsistence farming with some production of cash crops. ABM uses a cash-owbased methodology and adapts its loan terms to the farmers needs with variable monthly payments (according to clients cash ow) and a principal grace period, if needed. The pricing of the loan is similar to a microenterprise loan. By 2012, ABM had extended 1,700 agro loans, which represented around 7 percent of its clients and 2.4 percent of the loan portfolio. The average loan size was around US$419. ABM offers this product in a radius of 25 km around the branches because loan supervision requires close monitoring. ABM has a dedicated team of agricultural loan ofcers that receives in-house and external training on the topic and receives slightly higher compensation than the enterprise loan ofcers because of their specialization. This
Sources: Agrin (2012) and Yaw Brantuo.

methodology is now in the process of being transferred to AccessBank Tanzania. OI Ghana piloted an agricultural nance program in 2010 with an input credit scheme using 536 cocoa farmers in the Ashanti region. By 2012 OI Ghana was serving about 9,000 farmers across six administrative regions of Ghana. OI Ghana is introducing the use of GIS technology to more accurately map the locations of smallholder farmers to enable them to accurately determine the areas under cultivation and avoid over or under use of agro-inputs, which has negative impact on the yield. In Ghana, where agricultural nance is only 6 percent of the loan portfolio of the entire banking sector and the quality of these loans are below industry average, a breakthrough in this area could impact the sector. While AMB and OIs experiences are still at a small scale, their performance is watched carefully by others. Experiences such as these of ABM and OI could potentially provide results that would widen the overall contribution of the banking sector to the domestic economy and positively impact the lives of the rural population, which is still severely underserved.

28

are nevertheless visible. Greeneld MFIs generally apply high standards related to transparency with clients and are also often active contributors to national credit reference bureaus. Also, greeneld MFIs have on several occasions advocated changes on behalf of the micronance sector, to enhance transparency, raise standards, and improve the quality of regulations. See Box 11. Several holdings endorse the Client Protection Principles,17 and train their staff to operationalize these principles. In the DRC, Advans and ProCredit led the way in more transparency toward their clients, and now Rawbank and BIC, two traditional commercial banks, also publish their prices and terms on their websites. In Ghana, market actors interviewed found greeneld banks to be more open and transparent in their dealings with their clients. This view is supported by the availability of client-oriented material on the websites of greeneld MFIs and the clearly visible pricing information posted in the banks. In Madagascar, AccsBanque Madagascar is one of only two MFIs that publish effective interest rates for their clients. Greeneld banks comply with reporting to the credit bureau or are, depending on the country context, actively participating in exchange of credit references among institutions operating in the same market. The holding companies highlight the importance of this for sound credit risk management and responsible nance.18 In Cameroun, for example, where Advans built its rst greeneld MFI, the bank started a credit information exchange that now involves 12 MFIs. In Ghana greeneld MFIs are all active providers and users of credit reference information, as required by law. In Madagascar, greeneld MFIs (which are licensed as commercial banks) willingly report to two credit bureaus: one for banks and one for MFIs. This dual reporting creates a bridge between the larger micronance sector and the banking sector, enabling the clients of these two greeneld

MFIs19 to have potential access to nancial services from entities in both sectors. Finally, some greeneld MFIs contribute to the development of market infrastructure and a more favorable regulatory environment by participating in the banking or MFI associations in their country. ProCredit in the DRC was instrumental in negotiating a liquidity ratio in favor of the micronance sector with the Central Bank.20 Advans Cameroun contributed training material on know-your-customer requirements to the Central Bank for a workshop provided to all MFIs in the country. In Ghana, greeneld MFIs played a role in lobbying for measures to allow savings and loans companies to clear checks and engage in foreign currency denominated transactions.
BOX 11

ASA International, an Impact Investment Pioneer in Transparency

ASA International is among the rst funds and holding companies to be rated by the Global Impact Investment Rating Service (GIIRS). The rating provides an independent judgment of social and environmental impact, practices, policies, and achievement and covers the holding, investment fund (CMI), as well as the underlying investments, ASAIs greeneld MFIs in Asia and Africa. While the rating is preliminary and not public, it provides ASAI a benchmark relative to other impact investing companies. In 2012 ASAI received a rating of 161 points out of a total of 200, which ranks it among the top quintile of the early GIIRS-rated companies, the so-called Pioneers. ASAI envisages to regularly update its rating and is currently undergoing a follow-up rating.
Source: GIIRS (2012) and GIIRS website (www.giirs.org).

17. T  he Smart Campaign website (www.smartcampaign.org) lists Access, Accion, Advans, BRAC, FINCA, MicroCred, OI, and Swiss Micronance holding as endorsers as well as some of their individual affiliates in SSA. 18.  In Ghana XDS Data Ghana noted that the greeneld MFIs were among the few nancial institutions that undertook direct and extensive due diligence of the credit reference system before signing on. Executives of greeneld MFIs undertook onsite visits to observe reliability, safety, and adequacy of the equipment and related processes of their facilities. 19. Other MFIs do not have access to the bank credit bureau. 20. The liquidity regulation in the DRC was dened in such a way that current and savings accounts needed to be covered by the same liquidity even though savings accounts were four times less liquid. ProCredit DRC demonstrated this with actual data in discussions with the Central Bank following which the ratio was revised to 60 percent for current accounts and 40 percent for savings accounts. 29

30

SECTION

Conclusion

he greeneld model has come a long way in a short time in SSA. While there is a range of micronance providers in SSA, the proliferation of greeneld MFIs expands the commercial end of the spectrum with regulated, deposit-taking institutions focused on microenterprises and small businesses. In most countries there are gaps in every segment of the market. Many greeneld MFIs address the broader micro and small business segments, while many existing MFIs tend to cater to microenterprise alone. Holding companies have established promising institutions in very difficult markets, including post-conict markets such as the DRC, Cote dIvoire, and Liberia. While many greeneld MFIs are still young, there are signs of solid institution building for the longer term and positive effects on local markets. Holding companies and their shareholders have found ways to leverage the greeneld business model to address some of the primary challenges in advancing access to nance in SSA. The small size of nascent nancial markets, high costs of doing business, uneven regulatory frameworks, and inadequately skilled human resources in many SSA markets benet from an approach where practices can be standardized and costs can be shared. Through network structures and common practices, holding companies are able to transfer knowledge and learning from one greeneld MFI to another, leveraging the investment in human resources and skills across borders. Their structured approach and heavy focus on human resources development has been a key success factor in their ability to create sustainable institutions in some of the most frontier markets in SSA. Greeneld MFIs have kept risks and losses relatively low thanks to rigorous training of staff, consistent application of tested methodologies, and strong commitment to strict quality and service standards. The sustainable performance of greeneld MFIs illustrates to the traditional formal banking sector that underserved businesses and households are bankable, and even protable, market segments. Market analysis in the DRC, Ghana, and Madagascar has shown that greeneld MFI practices have often been transferred to other market participants by example and through staff movement. Even though attribution remains a difficult issue, it appears, especially in less-developed nancial markets, that greeneld MFIs play a pioneering role in expanding the nancial access of microenterprises, small businesses, and low-income households.

Where does the model go from here? The number of greeneld MFIs has grown rapidly in the past decade, and new entities are still being added to this segment of the micronance industry. It is likely, however, that the rate of creation of greeneld entities, at least in SSA, will slow, as the most feasible markets have now largely been entered. But this leaves about 25 countries in SSA without any greeneld MFI presence, and typically without the presence of any sustainable MFIs at all. Some have populations too small to sustain a commercial institution (many island nations), some are unstable and in conict (Somalia), some have a combination of these issues (the Central African Republic), and some have limitations on foreign ownership in the banking sector (Ethiopia). And in almost every country, peri-urban and rural populations still struggle to access nancial services. One challenge for greeneld MFIs and their holding companies is, therefore, to develop a delivery model that facilitates commercially viable and affordable access in smaller, more dispersed markets and rural areas. Indeed, some of the more mature greeneld MFIs that have achieved breakeven are now exploring alternative delivery channels, such as agent banking and mobile nancial services, to extend their reach in markets with low population densities that present challenges for traditional bricks-and-mortar expansion models. Likewise, their product development is pushing the boundaries at both ends of the spectrum with a focus on SME lending at one end and payment, credit, and deposit services for the mass market at the other. Another possible development is that some holding companies will develop greater appetite and capacity to acquire existing MFIs and small business banks. So far this has been a rare occurrence indeed, as the holding companies are usually apprehensive about the costs and risks involved in the acquisition of other entities. However, it is possible that a dearth of attractive greeneld opportunities will create incentives for some holding companies to consider such options, particularly if they feel they have a strong network of MFIs that could support the operational aspects of such an approach. At the same time as the holding companies and greeneld MFIs face signicant operational challenges (and opportunities), they will also have to manage their investors expectations, particularly those of the DFIs. Proof of concept now has to give way to mass market reach and shareholder returns. Apart from the pressure that this
31

creates, it will also generate questions about the continued role of DFIs in greeneld networks, with some arguing that there is a continued role (e.g., by helping achieve mass market reach) and others arguing that the DFI role has largely been fullled and that there is time to realize some nancial returns. The shareholding of greeneld MFIs has been very stable so far, but it is possible that the market will see more movement in the ownership of these entities going forward, particularly if local investors start taking a greater interest. It is also possible, though not very likely, that the market could see sales of entire greeneld entities if they are not deemed to t the future strategy of the network. In some cases, DFIs have pursued a dual-stage exit, swapping shares from the local subsidiaries into the holding. The diversication provided by a holding company portfolio presumably makes the investment more liquid and potentially easier to exit through a sale of shares to new holding investors or the potential exit through an initial public offering (IPO) in the capital markets. Finally, there is the (even more remote) possibility that a major reshuffle

could occur as a result of an IPO at the holding company level, at which point many DFIs may exit their investments in both the holding company and greeneld MFIs. The greeneld MFI model is a complement to other strategies for increasing access to nance, such as reform of existing institutions without a holding structure and bank downscaling. The next few years will be very telling about the ability of greeneld MFIs to leverage their foundation and achieve scale, and for holding companies to replicate and sustain the success of their model in other markets, particularly in a context of diminishing funding for TA. Undoubtedly they will nd themselves compelled to develop new methods, capacities, and practices to stay relevant and competitive in the micronance space. This is already evident in the increasing level of investment directed toward technology-based solutions and alternative delivery channels. At the same time, it is also likely that many of these greeneld MFIs will increasingly begin to compete with commercial banks for mass market customers and those in the SME space. The nancial landscape in Africa is poised to become much more interesting.

32

ANNEX

Methodology
Denition of Greeneld MFI Greeneld MFIs are dened for the purpose of this publication as institutions that are newly created without pre-existing infrastructure, staff, clients, or portfolios and use standard operating procedures disseminated by a central group, often a holding company. The holding company usually also plays a strong role in backstopping operations, providing standard policies and procedures, and co-branding the subsidiaries in the network. Takeovers and mergers were not included, nor were some subsidiaries of OI and FINCA that started as NGOs many years ago and did not have data available from the start of their operations. Also, at that point in time, they werent part of a network governed by a holding company. Ecobank is the only commercial African bank included in the sample. Some other large commercial banks, such as Equity Bank and UBA, which also provide micronance services, have expanded to other countries. However, these institutions do not focus exclusively on the MSE market and have not created a separate structure to distinguish their micronance operations as Ecobank has made it more complicated to separate comparable date. Therefore, these banking groups have not been included in the study. Data from African Greeneld MFIs and Life Cycle Performance Analysis Eleven holding companies provided detailed performance data throughout the life cycle for 30 of the 33 African greeneld MFIs created in SSA since 2000. ProCredit Angola, Ghana, and Mozambique are not included in this detailed analysis as ProCredit Holding did not provide performance data for these banks by institutional age. The holding allowed the authors to use data available for two of their banks (DRC and Sierra Leone) that were readily available. ProCredit Ghana was included in the market development research (see below). Data for ProCredit Sierra Leone are included only for the period in which it was owned by ProCredit, from 2007 until 2010, when it was sold to Ecobank. The 30 greeneld MFIs included in the research represent various types of institutions. A majority are licensed and regulated deposit-taking institutions, ranging in legal structure from commercial banks to savings-and-loan companies to specialized deposit-taking MFIs. A few started taking deposits a couple of years after their creation, and a few re-

mained credit-only MFIs, mobilizing compulsory savings, but not taking voluntary deposits from clients other than their members. Performance of the greeneld MFIs is evaluated by institutional age (as opposed to calendar year) to evaluate progress and maturation from start-up, regardless of the year operations were initiated. This way, conclusions can be drawn at different stages of institutional development: foundation (preparation and rst year of operations), institutional development (generally year two through breakeven), and scale-up (from nancial breakeven onward). Among the 30 institutions in the sample, ve are in the foundational stage, 11 in institutional development stage, and 14 in scale-up stage. The number of greeneld MFIs in the sample for the different performance indicators and ratios therefore gets smaller closer to the month 60 timeline. Performance data are presented as simple averages unless otherwise indicated to display the performance of a typical greeneld MFI. When outliers appeared to distort the simple average, weighted averages have been used (e.g., in the case of return on assets and return on equity).

Performance Benchmark with MIX Young Africa In Section 1 the performance of greeneld MFIs has been benchmarked to the MIX index for young MFIs in Africa, which comprises 58 institutions, four to seven years old. All but ve of these institutions are deposit taking, with a deposit base ranging from $30,000 to $22.6 million ($2 million on average) and gross loan portfolio ranging from $2,000 to $24 million ($2.7 million on average). The greeneld MFI subjects of this paper were removed from the benchmark population. Market Share Analysis of the DRC, Ghana, and Madagascar Greeneld MFIs that were included in the market level analysis in the DRC, Ghana, and Madagascar presented in Section 4 are (with their year of creation): DRC: Advans (2009), FINCA (2003), ProCredit (2005), and OI (2010); in Ghana: EB Accion (2008), Advans Ghana (2008), ProCredit (2002), OI (2004), and ASA (2008); and in Madagascar: ABM (2007) and MicroCred (2006). To determine the market share of these greeneld MFIs in terms of deposits, gross loan portfolio, number of depositors, and number of loan clientsthe greeneld MFI data
33

collected for this study were compared to data available in the IMF FAS database 2011 data. In addition, some comparisons were made with the micronance sector or nancial sector as a whole, depending on data availability.

DRC For the DRC the FAS database includes data for commercial banks only and not for the rest of the nancial institutions. To calculate market shares, data of greeneld MFIs that are not included in FAS (i.e., FINCA and OI) have been added to commercial bank data to establish the denominator. For deposit accounts, FAS contained data for all deposit-taking institutions, so this market share reects the share of all greeneld MFIs of the entire nancial sector. For the comparison with other MFIs, the market share analysis relies on data collected by the Fonds pour linclusion nanciere en RD Congo (FPM). FPM reported 146 cooperatives, 16 local MFIs, and six international MFIs and micronance banks (including the greeneld MFIs in the sample) and three universal banks providing micronance. Ghana Also for Ghana the FAS database includes data only for commercial banks data so no data were available for

greeneld MFIs. Their data have been added to commercial banks data to establish the denominator for the calculation of the market shares. For Ghana the number of loan accounts is likely overestimated as FAS data included only number of borrowers, which tends to be lower than the number of loan accounts. In Ghana the greeneld MFIs are ve of the 19 savingsand-loans companies that represent only a very small part of the nancial sector. At the time of the study ASA in Ghana was in the process of transforming from an NGO to a savingsand-loan company. In Ghana, data for the greeneld banks are compared with industry data for so-called deposit money banks, including commercial banks, rural banks, and other banking and quasi-banking institutions.

Madagascar The FAS database has disaggregated gures for Madagascar facilitating the calculation of the market shares. Both greeneld MFIs are commercial banks. However, for the purposes of the market analysis, they are compared with the 36 Malagasy MFIs (ve etablissements nancier and 31 institutions de micronance agrees) as their clientele and methodology are more aligned than with the other nine commercial banks in the country.

34

ANNEX

List of people interviewed

Holdings Access Holding, Thomas Engelhardt, Management Board Member Access Holding, Christoph Diehl, Management Board Member Accion International, Brian Kuwik, Regional Head Africa Accion Investments, John Fisher, Vice President Advans, Claude Falgon, CEO AKAM, Mwaghazi Mwacho, CEO ASA International Holding, Martijn Bollen, General Counsel ASA International Holding, Mischa Assink, Senior Accountant BRAC, Ishtiaq Mohiuddin, Director Micronance BRAC, Tanwir Rahman, Director of Finance FINCA, Helen Lin, Africa Finance Manager FINCA, Mike Gama-Lobo, Vice President and Regional Director Africa MicroCred, Arnaud Ventura, CEO Opportunity International, Colin McCormack, Head of Africa Operations Opportunity International, Jean-Philippe Nefve, CFO of Global Micronance Operations EcoBank, Francis Adu-Mante, Managing Director EB-Accion Ghana ProCredit Holding, Helen Alexander, Management Team Member Swiss Micronance Holding, Thi Hanh, Operations Manager FIDES Swiss Micronance Holding, Christian Baron Swiss Micronance Holding, Konrad Ellsasser Affiliates AccesBanque Madagascar, Philip Acton, AccessBank Tanzania, Roland Coulon, CEO, Advans Cameroun, Frank Snieders, CEO Advans Cote dIvoire, Gregoire Danel-Fedou, CEO Advans Bank DRC, Francois Lecuyer, Directeur Generale Advans Ghana, Cedric Henot, CEO Advans Tanzania, Peter Moelders, CEO ASA Ghana, Mohammed Aourongjeb, Executive Director EB-Accion, Ghana, Frances Adu-Mante, Managing Director FINCA DRC, Ed Greenwood, CEO MicroCred Madagascar, Barnabe Francois, CEO

Opportunity International DRC, Gilbert Lagaillarde, Directeur Generale Opportunity International Ghana, Kwame Owusu Boateng, Deputy/Acting CEO

DRC Jean Claude Thetika, Directeur General, Fonds pour linclusion nanciere en RD Congo (FPM) Michel Losembe, Directeur General BAnque Internationale pour l`Afrique Au Congo, President of ACB (Association Congolaise des Banques) Ghana Philip Cobbinah, Deputy Head, Banking Supervision, Bank of Ghana William Asare, Bank Examiner, Banking Supervision Department, Bank of Ghana Yaw Gyima-Larbi, Head, Micronance Unit, Banking Supervision Depart, Bank of Ghana Gloria Quartey, Head, Center for Training and Professional Development, Bank of Ghana Yvonne Quansah, Director, Financial Institutions Sector, Ministry of Finance Bernard Joe Appeah, Principal Consultant, Pentax Management Consulting Emmanuel Owusu, Managing Director/President of Association, Global Access Savings and Loans/ Ghana Association of Savings and Loans Companies Yaw Gyam, Executive Secretary, Ghana Network of Micronance Companies Raymond Mensah, M&E Specialist, Rural and Agricultural Finance Programme (IFAD/Ministry of Finance) Richard Amaning, Executive Secretary, Ghana Association of Micronance Companies Vera Geraldo-Stephenson, Assistant Sales Manager, XDS Data Ghana Limited Madagascar Antoine Rakotondrasoalimangarivelo, Directeur des Portefeuilles, TITEM Bakoly T. Rafanoharana, Expert National, PAFIM Blaise Francis Rajoelina, Coordonnateur National de la Micronance, CNM Brillant Rakotoarison, Directeur Gnral, SIPEM Sa.
35

Charlot Razakaharivelo, Directeur Gnral, FIDEV Fanjaharivola Rakotomaharo, Secrtaire Gnral, APIMF Jean Herley Ambinitsoarivelo, Directeur, CEFOR Jos Serge Rajaonarison, Directeur Gnral, CECAM Jules Thodore Rakotondramanga, Secrtaire Gnral, CSBF Liva Claude Herimanana, Directeur Gnral Adjoint, ACEP Mahefa Edouard Randriamiarisoa, Directeur Gnral, ACEP Ndriana Ralaimanisa, Directeur Commercial et Marketing, BNI-CL

Randrianiaina Rakotoarivao, Directeur du Rseau, OTIV TANA Thomas Rasolonjatovo, Prsident Conseil dAdministration, MECI Youssouf Mahamoud, Directeur des Oprations, OTIV DIANA

Investment officers AFDB, Robert Zegers, Rafael Jabba, Barnett Douglas and Timo Teinila FMO, Andrew Shaw, and Maurice Scheepens IFC, Adam Sorensen KfW, Matthias Adler, Monika Beck, Simon Bleidiesel, and Karl-Heinz Fleischhacker

36

ANNEX

Bibliography
AccesBanque. 2012. Financial Statements 20082012. Madagascar: AccesBanque. Advans Ghana. 2012. Advans Ghana Competitor Analysis. Accra: Advans Ghana. Agrin. 2012. Banking Madagascars Small Farmers: ABMs Cash Flow-Based Agricultural Credit Analysis Methodology. Washington, D.C.: Agrin. Ahouissou, B., and M. Andriamahenina. 2012 Final Evaluation Mission Report of the National Micronance Strategy, SNMF 20082012. Antananarivo: UNDP and UNCDF. https://www.lamicronance.org/les/34227_ le_evaluation_SNMF_Madagascar.pdf Assouline, N., and C. Poursat. 2010. Malagasy MFIs Facing the Crises: What Lessons to Learn? AFD. Bank of Ghana. 2010. Bank of Ghana Annual Report 2010. Accra: Bank of Ghana. Bank of Ghana. 2011. Bank of Ghana Annual Report 2011. Accra: Bank of Ghana. Banque Centrale Du Congo. 2012. Aggregated Financial Sector Data. Kinshasa: Banque Centrale Du Congo, December. BCM (Banque Centrale de Madagascar). 2012a. Regular Note on the Evolution of the Activity of Micronance Institutions at the End of December. Antananarivo: BCM. . 2012b. Bulletin of the Central Bank of Madagascar. Number 21. Antananarivo: BCM, June. . 2012c. Regular Note on the Evolution of the Activity of Banks and Financial Establishment at the End of December, 2012. Antananarivo: BCM. Beck, T., S. M. Maimbo, I. Faye, and T. Triki. 2011. Financing Africa: Through the Crisis and beyond. MFW4A. Washington, D.C.: World Bank, BMZ, and AfDB. Buchs, T. and J. Mathisen. Competition and Efficiency in Banking: Behavioral Evidence from Ghana. IMF Working Paper 2005. Washington, D.C.: IMF. Cabinet ECR. 2011. Study and Analysis of the Legal, Reglementary and Fiscal Framework of the Micronance Sector in Madagascar. Antananarivo: Cabinet ECR. CNMF. 2007. National Micronance Strategy 20082012. Antananarivo: CNMF. . 2012. Consolidated Statistics of Micronance Activities, December 2012. Preliminary version. Antananarivo: CNMF.

EB-ACCION. 2013. EB-ACCION 2013 Status Report. Accra: EB-ACCION. FinMark Trust. 2010. Finscope Ghana Survey 2010 Report. Accra: Finscope. FPM. 2012. Synthesis of the Political, Economic and Financial Evolution of DRC. Facility for Financial Inclusion in DRC. Third trimester 2012. Kinshasa : DRC. Fritz, B., and K. Hujo. 2005. Economy under Latin American Conditions. Frankfurt am Main: Vervuert, pp. 95122. Gaul, Scott. 2011. Mapping Africa Financial Inclusion Results Review and Next Steps. Washington, D.C.: MIX. h t t p : // w w w. t h e m i x . o r g / p u b l i c a t i o n s / m i x microfinance -world/2011/09/africa-financialinclusion/results#ixzz2pNipbrQi GIIRS. 2012. Preliminary Rating ASA International Holding. Wayne, Penn.: GIIRS. www.giirs.org IMF (International Monetary Fund). 2011. Ghana: Financial System Stability Assessment Update 2011. Washington, D.C.: IMF. http://www.imf.org/external/pubs/ft/ scr/2011/cr11131.pdf INSTAT. 2011. The Labour Market in the Agglomoration of Antananarivo in 2010: A Decenium Perspective. Antananarivo: INSTAT. Jansson, T. 2010. Success Factors in Micronance Greenelding. IFC SmartLessons. Washington, D.C.: IFC, May. McKee, Katharine, Estelle Lahaye, and Antonique Koning. 2011. Responsible Finance: Putting Principles to Work. Focus Note 73. Washington, D.C.: CGAP. MicroCred Madagascar. 2012. Activity Report 2011. Antananarivo: MicroCred. Moyart, M., et al. 2005. Country-Level Effectiveness of Aid Report (CLEAR) for Madagascar. Washington, D.C.: CGAP. Nair, A., and A. Fissha. 2010. Rural BankingThe Case of Rural and Community Banks in Ghana. Agriculture and Rural Development Discussion Paper 48. Washington, D.C.: World Bank, May. Porteous, David. 2006. Competition and Microcredit Interest Rates. Focus Note 33. Washington, D.C.: CGAP, February.
37

ProCredit Holding. 2010. Annual Report. Frankfurt am Main: ProCredit Holding. . 2012. Annual Report ProCredit Bank Congo 2011. Frankfurt am Main: ProCredit Holding. Rosenberg, Richard, Scott Gaul, William Ford, and Olga Tomilova. 2013. Microcredit Interest Rates and Their Determinants, 20042011. Forum 7. Washington, D.C.: CGAP, June.

Schmidt, Reinhard H. The View of the Participating Observer: An Outline of the IPC/IMI Story from Neoinstitutional Perspective, in Economy under Latin American Conditions. Frankfurt am Main: Vervuert, pp. 95122. Transition Presidency. 2013. Law n2002-021 covering the laws for nance for 2013. Government of Madagascar.

38