Pre-Analysis Plan

Evaluation of Microfinance Activities of BRAC in Sierra Leone

C.E.M. van Zaal Wageningen University

Supervisors: M.J. Voors & B.W. Lensink,Wageningen University M. Nasir, BRAC Sierra Leone, B. Conteh, Njala University

Date: 26-02-2014

Contents List of Tables ...... 5 List of Figures ...... 5 Abbreviations ...... 5 1. Introduction ...... 6 1.1. Problem statement and Objectives ...... 6 1.2. Outline ...... 7 2. BRAC MF program in Sierra Leone ...... 8 2.1 Group lending scheme ...... 8 3. Research questions...... 11 4. Literature and hypothesis ...... 12 4.1 Impact assessment ...... 12 4.2 Introduction to microfinance ...... 12 4.3. Microfinance in West-Africa ...... 13 4.4. Research question 1: What are the determinants of participation in microfinance group? ..... 14 4.5. Research Question 2: What are the impacts of microfinance services on poverty and vulnerability? ...... 14 4.5.1 Do BRAC’s microfinance services contribute to the income of their clients? 14 4.5.2 What is the impact of microfinance services on vulnerability of their clients? ...... 15 4.6. Research question 3: What is the level of client satisfaction with the microfinance services of BRAC? ...... 16 4.7. Research question 4: What is the impact of microfinance loan group composition on financial behavior? ...... 16 4.7.1 What is the composition of groups? 16 4.7.2 How does group composition affect financial behavior? 17 5. Hypotheses ...... 19 6. Methodology ...... 20 6.1. Sampling Evaluation ...... 20 6.1.1 Sampling method 20 6.1.2 Block randomization of branches and members 21 6.2 Use of control groups ...... 22 6.3 Sampling size Phase I ...... 23 6.4 Sampling size Phase II ...... 23 6.5 Methods ...... 24 6.5.1 Individual survey – Group members & Group leader 24

Page 2 of 120

6.5.2 Individual survey – Loan officer 25 6.5.3 Individual survey – Non member 25 6.5.4 Network analysis 25 6.5.5 Coordination game 26 6.5.6 Public goods game 26 6.5.7 Dictator game 26 6.5.8 Methods per research question 27 6.6 Limitations ...... 28 6.6.1 Identification of causal effects 28 6.6.2 Time 28 6.7 Planning ...... 29 7. Data ...... 30 7.1 Outcome indicators ...... 30 7.2 Control variables ...... 31 7.3 Covariates ...... 31 References ...... 32 Appendix A: BRAC ...... 36 Appendix B: Linking research questions, outcomes and measurements...... 39 Appendix C: Sample selection ...... 49 Appendix D: Instruments ...... 50 Group Member Survey ...... 51 Group Leader Survey ...... 68 Loan Officer Survey ...... 85 Non Member Survey ...... 93 Control Group Census Form: Inside boundary ...... 103 Control Group Census Form: Outside boundary ...... 103 Group Payments Sheet ...... 104 Coordination Game Record Sheet ...... 106 Appendix E: Protocols ...... 108 Evaluation MF SL BRAC – Protocols ...... 109 Phase 1: Group composition study ...... 110 Phase 1: Group composition study ...... 112 Pre-visit activities ...... 112 Schedule of activities during visit ...... 112

Page 3 of 120

Group Meeting ...... 112 General comments ...... 113 Phase 2: Evaluation study ...... 114 Pre-visit activities ...... 114 Schedule of activities during visit ...... 114 Meeting with participants ...... 114 General comments ...... 115 Respondent Survey PART 1 ...... 116 Respondent Survey PART 2 ...... 116 Network Analysis ...... 116 Activity #1: Coordination Game ...... 117 Activity #2: Public Good Game ...... 117 Activity #3: Dictator Game ...... 118 Pay-slip payment ...... 118 Protocol - Control Group Census ...... 120

Page 4 of 120

List of Tables Table 1 Set-up date branches ...... 21 Table 2 Block randomization ...... 22 Table 3 Treatment and control group per research question ...... 23 Table 6 Sampling Phase I ...... 23 Table 4 Sampling branches Pahse II ...... 24 Table 5 Sampling Phase II ...... 24 Table 7 Methods per research question ...... 27 Table 8 Planning ...... 29 Table 9 Information branches ...... 37

List of Figures Figure 1 Radius of branch and group ...... 8 Figure 2 Diagram group BRAC ...... 9 Figure 4 Map branches BRAC ...... 36

Abbreviations BRAC Bangladesh Rural Advancement Committee C1 Control group 1 C2 Control group 2 CG Coordination Game DG Dictator Game MF Microfinance loan MFI Microfinance institutions PGG Public Goods Game RDD Regression discontinuity design SEP Small entrepreneur loan SL Sierra Leone WUR Wageningen University and Research Centre

Page 5 of 120

1. Introduction This document constitutes a pre-analysis plan for the evaluation of microfinance activities of BRAC, formerly known as Bangladesh Rural Advancement Committee (BRAC), in Sierra Leone and specifies research questions, methodologies and specifications to be employed in testing hypotheses concerning group lending to female borrowers in Sierra Leone. This plan is written prior to data collection and analysis.

Microfinance activities of BRAC in Sierra Leone are the MF (microfinance) program and the SEP (small entrepreneur) program. MF programs are group lending programs where female borrowers are jointly liable for the repayment of their loans. SEP programs are programs where female and male borrowers have individual loans. These individual loans in the SEP programs are bigger than the loans in the MF program. In this study we only focus on the MF program.

1.1. Problem statement and Objectives BRAC has been operating in Sierra Leone since 2009. They provide microcredit loans to female borrowers throughout Sierra Leone. Presently, BRAC is expanding their development work. Even though their business has been expanding in Sierra Leone, BRAC does not hold any records about the effectiveness of their microfinance services to the poor in Sierra Leone. This evaluation aims to address this knowledge gap.

The objective of the study is to assess (i) the determinants of clients, (ii) whether clients improve their wellbeing, and are less vulnerable and (iii) whether clients are satisfied with the services provided by BRAC and (iv) to understand the processes underlying the composition of the groups and the influence of group composition on financial behaviour. This study tries to provide an empirical investigation, consisting of surveys, a network analysis and behavioural games, of the group lending scheme of BRAC. Currently, as of January 2014, the MF program is operating in 30 branches. The research will take place in one third of the branches of BRAC, and is divided into two phases. Phase I will address research question 1, 2 3 and 4, and will be conducted in Mile 91. Phase II will be conducted in eight branches, of which four branches will be based in Freetown and four in the rural areas of Sierra Leone. We chose this distinction between rural and urban branches, due to the possibility that urban branches are more heterogeneous than rural branches. Phase II will be addressing research question 1, 2 and 3. We decided to only survey research question 4 in Phase I, because in order to survey research question 4, behavioural games are needed, which are time consuming and costly.

Page 6 of 120

1.2. Outline In this study we will evaluate the microfinance activities of BRAC in Sierra Leone, more precisely their MF program. The MF program consists of a group lending scheme where borrowers are jointly liable for the repayments of their loans. Every group of 25 people consists of smaller groups of five people and a president, treasurer and secretary. BRAC has been operating in Sierra Leone since 2009, however, they do not hold any records about the effectiveness of their MF program. With the help of this study we try to assess the impact of the MF program and to examine the influence of group composition on financial repayment.

The main research questions are:

1. What are the determinants of participation in microfinance groups? 2. What are the impacts of microfinance services on poverty and vulnerability? 3. What is the level of client satisfaction with the microfinance services of BRAC? 4. What is the impact of microfinance loan group composition on financial behavior?

The research team will consist of collaboration between Wageningen University and the research team of BRAC. We combine surveys, behavioral games and a network analysis in order to find an answer to the above mentioned research question. Phase I of the research will take place in Mile 91; Phase II will take place in eight other randomly chosen branches in Sierra Leone.

In terms of expected outputs, we expect not only to enhance our understanding of the impact of the MF program of BRAC in Sierra Leone, but also to learn how groups are formed, how networks within these groups work and how the group composition influences repayment rate and impact of the MF program. The knowledge obtained will be used by both BRAC and Wageningen University.

Page 7 of 120

2. BRAC MF program in Sierra Leone Founded in 1972 in Bangladesh, BRAC is now one of the largest non-governmental organisations in the global south. BRAC’s main focus is on the poor, enabling them to organise community human and material resources to create development opportunities. They do this with a holistic development approach with tools such as microfinance, education, healthcare, legal services and community empowerment. The principal activity of BRAC is the provision of microfinance services (BRAC, 2013).

In Sierra Leone, BRAC has been providing microfinance services since 2009. Currently, they have 30 branches throughout Sierra Leone and their customers’ base has been growing with 7 per cent between 2011 and 20121. The core activity of BRAC’s microfinance services is the provision of group loans (MF program) and small entrepreneurial loans (SEP program).

2.1 Group lending scheme BRAC works in Sierra Leone with a group lending scheme with joint liability. Under joint liability, individuals, in this case women, group together to apply for loans and are jointly accountable for the repayment of the loans (Westover, 2008).

BRAC has established 29 branches all over Sierra Leone, as depicted in Figure 3 in Appendix A. Each branch has three or four loan officers, each in charge of 250-300 borrowers. A branch has a minimum amount of 800 borrowers. These borrowers live all within a range of four kilometres of the branch (Figure 1).

-

Figure 1 Radius of branch and group

-

1 More information about BRAC can be found on www.brac.net and in their Annual Report of 2012. Page 8 of 120

When a branch is being established, clientele has to be found. Therefore, the loan officers first visit every potential female borrower in the area of four kilometres to inform them about the lending scheme of BRAC. The criteria of becoming a borrower are:

- The member must be a female - The member must be aged between 18-50 years - Only one member from each household is selected - The member must have lived in the area for three – five years - The member most not belong to other MFIs - Prospects should live within four km from the branch office location - The member should be living within the boundary of the village/community - Member should have a business

Then the loan officers drop by a second time to see who of the informed eligible females is interested. The potential female borrowers, who are interested and eligible, form groups of four to five females and apply for an individual loan, but as a group. Then the loan officer decides if they are eligible for the loan and if they can enter the program.

The small groups, of groups of four to five people, are organised in one big group of 25 to 30 people. This is depicted in Figure 2;

Secretary President Treasurer

Loan

Officer Big group (25-30p)

Small group (5p) Small group (5p) Small group (5p) Small group (5p) Small group (5p)

President President President President President

Figure 2 Diagram group BRAC

These small groups, which form the big group together, consists of females which already knew each other before entering the BRAC’s group lending program. They have jointly formed the group and therefore we can assume that certain levels of trust already exist within this small group. The concept of joint liability also accounts only for the small groups, therefore trust and networks are more important in small groups then in the bigger groups.

When everyone is accepted and the big group is constituted, a committee of five people is chosen by all members of the big group. This committee consists of a president, secretary, treasurer and two

Page 9 of 120 executive members. Each original small group also has a president. Every year, the committee is being re-chosen if necessary. But often the president of the big group and the small groups remain the same person throughout the years that the group exists. The president of the big group is often also the president of one of the small groups. And in most cases, the president is one of the oldest members of the big group. Everyone receives a passbook where the amount of the loan, the interest rate and the repayments are being noted. A pass photo of the owner of the passbook is added in the book. Whenever someone loses her passbook, a fine has to be paid.

Every two weeks two group members receive a loan. This is based on their behaviour, as in defaults of repayment and commitment to the meetings, and on their income performance. The size of the loan differs per member. When a member receives a loan of 20 weeks (five months), the interest rate is 13 per cent, when the loan which is received last for 40 weeks (ten months), the interest rate is 25 per cent.

When a group member is unable to pay their weekly repayment, consequences will occur. The first time this happens, the other members will pay for the defaulting member. But if it happens frequently, the member will be thrown out of the group and therefore out of the group lending scheme. The loan officer will then choose a new eligible female in the area and this new member will join the small group.

During the weekly meetings, which are held in a place maximum one kilometre from all group members, the repayment of the loan is done and questions can be asked. The president of each small group collects the repayments, and then hands them over to the treasurer and the loan officer of BRAC. When someone is unable to repay their loan, the group members of the small group pay the repayment of that person. Due to this arrangement, peer selection and monitoring over each other’s peers is an important aspect of the group.

In most urban group lending schemes, the female borrowers own a shop or are street vendors. They tend not to own their own dwelling, but rent them from landowners. In the rural group lending schemes, borrowers tend to engage more in agricultural occupation, but for the most part women own a shop or are street vendors too. They do own their own land and house, because it is cheaper in the rural areas than it is in the urban areas.

During the fieldtrips, we asked the borrowers for the biggest challenges and benefits of receiving a loan. When asking for the biggest challenges, the borrowers answered; (i) transportation, (ii) increase of prices, (iii) rainy season, (iv) when business goes bad to meet the demand of the loan and (v) the loan ceiling. The biggest benefits they mentioned were; (i) education, (ii) assets, such as the purchase of an okada2 and (iii) increase in business.

2 An okada is the Sierra Leonean word for a motorbike. Okada’s are motorbike taxi’s. Page 10 of 120

3. Research questions This research encompasses several research questions;

1. What are the determinants of participation in microfinance groups? 2. What are the impacts of microfinance services on poverty and vulnerability? 3. What is the level of client satisfaction3 with the microfinance services of BRAC? 4. What is the impact of microfinance loan group composition on financial behavior?4 a. What is the composition of groups? b. How does group composition affect financial behavior? c. How does group composition affect the impact of microfinance services?

3 Client satisfaction in terms of size of loans, interest rate, amount of meetings, accessibility of BRAC when needed, loan officer 4 Research question 1-3 are formed by BRAC, research question 4 is formed by the research team Page 11 of 120

4. Literature and hypothesis In this section we discuss the literature about microfinance and impact assessments, concerning the two research questions. First, a short introduction about impact assessments is being given. Second, we introduce microfinance and the concept of joint liability. Third, we discuss the history of microfinance in West-Africa. And last, we elaborate on each of the research questions.

4.1 Impact assessment Development programs of microfinance institutions are designed to change a certain outcome in a certain context, for example increase income, increase business activities and decrease deaths. Whether these outcomes are changed or not is a question often raised by actors such as donors, governments or other critics. Since the 1990s impact assessment has become an increasingly important aspect of development activity, as being a part of evidence-based policy making (Gertler et al, 2011), and have become popular with donor agencies and therefore also have become a significant activity for recipients’ agencies (Hulme, 2000).

When donors and policy makers are demanding results and accountability from programs, impact evaluations can provide robust and credible evidence on performance and on whether the program reached its desired and predetermined outcomes. Another use of impact evaluations is that they can find out which parts of a program do and do not work (Gertler et al, 2011). Therefore, impact evaluations incorporate both proving impacts and improving interventions (Hulme, 2000).

Even though the question whether a program changes a certain outcome is being raised, not often the actual effectiveness of a program has been examined. Empirical evidence concerning the impact and effectiveness of impact evaluations is little. Westover (2008) reviewed over 100 articles, of which only six can be called quantitative and academically rigorous. The other remaining articles are all qualitative and specific case studies. There are various impact studies which have shown the financial viability of microfinance programs, but the question of the effectiveness and impact on the poor of such programs is still in question.

4.2 Introduction to microfinance Even though the concept of microfinance was not anything new (Srnec and Svobodová, 2009), the idea of microfinance of how we know it nowadays originally started in Bangladesh in the 1970s with the Grameen Bank and its creator Muhammad Yunus. The concept of microfinance refers to offering financial services to low social economic status individuals which are excluded from the traditional fincial system. They are considered “unbankable”, due to lack of collateral, steady employment and a verifiable credit history. Microfinance is designed to provide small amounts of start-up capital, e.g. loans, for entrepreneurial projects, which help individuals generate income, build wealth and most importantly exit poverty (Westover, 2007).

One of the most common methods of microfinance, to provide these loans and to be sure of repayment is the use of group loans with joint liability. Group lending refers to the practice of working with clients in small groups (typically of three to seven members). Loans are distributed to individuals, but the group as a whole is held jointly liable should repayment difficulties arise (Armendáriz de Aghion and Morduch, 2000). Joint liability is a concept where groups of individuals, together apply for loans and are jointly accountable for repayment (Westover, 2008). It is stated that joint liability can deal effectively with the four major problems that lenders face. These problems are:

Page 12 of 120 i) adverse selection – to explore what kind of a risk-taker a potential borrower is; b) moral hazard – to make sure a borrower utilizes the loan properly and therefore is able to repay the loan; c) auditing costs – to find out how a borrower’s business performs in case of inability to repay; and d) enforcement – to force the borrower to repay the loan.

Microfinance institutions which use joint liability can perform well, in some social contexts, for two reasons. First, members of a community know more about one another and what kind of ‘type’ another community members is than an outside institution. This sort of information can be helpful for solving the problems a, b and c. Second, in comparison to banks, which cannot apply (financial) sanctions against defaulting, poor borrowers because in definition they are poor, community members can. Community members may be able to impose powerful non-financial sanctions at low costs, which helps solve problem d. In short, an institution that gives poor people the proper incentives to use existing information on their neighbours and to apply non-financial sanctions can out-perform a conventional bank (Ghatak and Guinnane, 1999).

The use of group loans with joint liability has been broadly contested throughout the years. For example work of Wenner (1995) and Wydick (1999) provide evidence that joint liability works as supposed, namely in their analysis the variables social cohesion and better information flow among group members imply improved repayment rates (Ghatak, 1999). Stiglitz (1990) and Besley and Coate (1995) claim that the idea that joint liability contracts can mitigate moral hazard in project choice by creating social costs to default is central to the idea of group lending in microfinance. Another author is Fischer (2008) who found, with the help of a laboratory experiment, that joint liability facilitates informal insurance among borrowers, allowing them to make profitable but risky investments. Even though several well-known models of joint liability lending predict a reduction in the rate of risky project choice under group lending, Giné et al (2010) found that joint liability increases rates of risky investment choice when borrowers are able to communicate freely (what is the case in reality), and has little effect on behaviour in the absence of communication. And with the help of data from experimental repayment games Kono (2014) found that borrowers were motivated to free-ride under the joint liability scheme.

In this study we will not focus on the question whether joint liability is feasible or not. In the next section we will briefly highlight the history of microfinance in West-Africa, then we will discuss the literature about the topics we highlight and the research question we assess in this study.

4.3. Microfinance in West-Africa Microfinance already exists for centuries in Africa and in the rest of the world. There are many global examples of ‘historical’ usages of microfinance. These range from informal, small-scale, rotating savings-and-loans clubs in England, Ireland and Germany in the 18th century to savings and credit cooperatives in Indonesia in the 19th century. In West-Africa, and especially Nigeria, microfinance dates back to the 15th century (African Development Bank, 2006) in the form of rotating savings and credit associations. They are called susu, which is a Yoruba term in Nigeria, and were transported during the slave trade to the Caribbean islands, where both the institution and the term still exist (Seibel, 2005).

History of microfinance in West-Africa can be distinguished by two separate periods. The period before 1985-1990, and the period that followed. In the period before 1985-1990, savings and credit cooperatives were promoted. These cooperatives promoted the traditional objectives of mobilizing

Page 13 of 120 savings and educating people. The first cooperatives operated in prosperous areas and most often in cash crop zones where members generally earned enough to cover expenses and repay the loan. But these cooperatives were criticized for only working with the middle class, and results were not very good in terms of repayment rates and cost recovery.

A turning point was reached towards the end of 1980s, when new types of microfinance began to emerge. This new generation of microfinance institutions did not abandon the notion of savings mobilization, but also emphasized access to credit and was committed to poverty alleviation. They did this with the help of solidarity group programs, which were also focused mainly on women. Also microfinance institutions emerged which were more flexible and responsive to the demands of their members.

By the middle of 1990s the microfinance sector really began to change. The position of MFIs in West- Africa changed from being a monopoly to a competitive market. In order to keep running, they had to grow, extend their outreach and respond to the needs of poor populations. Today, microfinance institutions become more hybrid; they try to make profit while staying sustainable (Chao-Béroff et al, 2000).

4.4. Research question 1: What are the determinants of participation in microfinance group? There have been many studies examining the characteristics of clients who decide to join microfinance services. In this study two characteristics are already defined. Namely, all clients are female and they live within the four kilometre boundaries of the branch.

Still, there are other determinants to be examined. When addressing the literature, some authors have shown that women tend to self-select themselves into groups with members who have identical status and identical ethnic identities, mitigating possibilities for solidarity (Rankin, 2000). Other studies have states that women self-select into groups with members who have significant assets, such as (husbands with) high income, therefore concentrating among those with access to capital and excluding the poorest of the poor (Ackerly, 1997) (Fernando, 1997).

Hypothesis 1: Clients self-select into groups with identical status and identical ethnic characteristics

Hypothesis 2: Clients self-select into groups with significant assets

4.5. Research Question 2: What are the impacts of microfinance services on poverty and vulnerability? This research question entails two sections, namely poverty and vulnerability. We will determine the impact of microfinance services on poverty by determining the poverty outreach of BRAC’s microfinance services. And we will determine the impact on vulnerability by determining the impact on education, health, housing, savings and food security.

4.5.1 Do BRAC’s microfinance services contribute to the income of their clients? One of the most important goals of microfinance institutions is to reach the poorest of the poor. Another important goal is to be financial sustainable. But due to the many challenges facing microfinance institutions, it is difficult to achieve both these goals at a high level. Challenges the microfinance institution face are; First, high competition among MFIs has led to lower interest rates

Page 14 of 120 and lower costs. Second, commercial banks have become interested in microfinance and started their own services, which forces MFIs to reduce their interest rates. Third, donors from developing countries have become interested in financing MFIs, which lowers the need of MFIs to become financial sustainable. Fourth, new technologies, such as cell phones, entered the business which lowers costs and fifth, some developing countries have liberalized their financial market, which improves the financial market (Hermes et al, 2008).

Due to these challenges and changes, it has become easier for MFIs to be financial sustainable and harder to reach the poorest of the poor (Gulli, 1998). Small loans, which are preferable for poor people, become costly due to e.g. monitoring costs, screening costs. Therefore it becomes more costly to serve the poorest of the poor (Hermes et al, 2008).

Other factors which determine the outreach of poverty are; i) the institutions’ policy and mission (Henry et al, 2003) (Montgomery, 2004), ii) contextual factors such as local infrastructure or presence of other MFIs and iii) client related factors, in terms of their needs and constraints of their enterprise or household (Gulli and Berger, 1999). Explaining the latter point, one could think of the presence of a selection-bias of the less poor clients. Poorer clients may not want or may not able to get involved due to fear of debt or cultural reasons (Rosenberg, 2006) (Banerjee and Duflo, 2003) (Hentschel and Waters, 2002).

Hypothesis 3: High financial sustainability of a MFI will lead to a low poverty outreach

Hypothesis 4: Bad infrastructure and presence of other MFIs will lead to a low poverty outreach

4.5.2 What is the impact of microfinance services on vulnerability of their clients? Whether a household is poor is an important determinant of a households’ wellbeing. Even though a household may not be poor, when the household is vulnerable it can turn into a poor household within a day or even hour, when occurring a large shock (Chaudhuri et al, 2002). This shock can be a natural disaster, death of a household member or a bad crop. Therefore we can see vulnerability as the risk a household has to become poor. The more chances a household have to ‘survive’, the less vulnerable a household is. These chances to survive can be increased by e.g. better education, dwellings which are better resistant to natural disasters, better health.

Chaudhure et al (2002) define vulnerability as “the ex-ante risk that a household will, if currently non- poor, fall below the poverty line, or if currently poor, will remain in poverty”. We will not use this definition, but will focus more on the idea of vulnerability as a multidimensional concept as Eakin and Bojórquez-Tapia (2008) do, and see vulnerability as the chance a household has to survive. We here divide the concept of vulnerability in level of education, health, housing, savings and food security.

Literature (Wright et al, 1999)(Garikipati, 2008)(Zaman, 2004) state that microfinance services reduce the vulnerability of poor households, and therefore reducing the risk that a household becomes poorer after an external shock.

Hypothesis 5: Microfinance services lead to less vulnerable households

Page 15 of 120

4.6. Research question 3: What is the level of client satisfaction with the microfinance services of BRAC? With the growing amount of microfinance institutions worldwide clients have more choice of choosing the services of a microfinance institution which suits them best. Thus, many MFIs are responding on this growth and to the changing landscape (Churchill, 2000) by seeking to better understand their clients’ demands, preferences and opinions (Anyango et al, 2002). The benefits of such a strategy are that it enhances customer loyalty and reduces drop-outs and therefore increases the profitability of an MFI (Churchill and Halpern, 2001).

Even though most microfinance institutions possess a product orientation approach, which states that success will increase by bringing goods and services to the market of which they are convinced are good for the public, recent studies (Woller (2002), Wilson (2001), Churchill (2000), Bashir et al, (2012)) show that market orientation (and therefore customer satisfaction) has a stronger positive relationship with institutional performance, such as profitability and market share. Market orientation entails that success will increase by determining the perception and needs of their clients and satisfy their clients through the design and delivery of their products (Wright et al, 2005), thus increasing customer satisfaction.

Customer satisfaction and customer loyalty are strongly interlinked. Bowen and Chen (2001) and Kandampully and Suhartanto (2000) both state that a small increase of customer satisfactions leads to a strong increase in customer loyalty. This strong correlation can lead to a recommendation of customers to other people, and therefore a rise in total amount of clients (Siddiqi, 2011). There are many opportunities for the potential growth of MFIs, but they must provide the products, service and convenience that their borrowers desire (Bashir et al, 2012).

Hypothesis 6: High customer satisfaction will lead to a growth in clients and profits of MFIs

4.7. Research question 4: What is the impact of microfinance loan group composition on financial behavior? The second research question is divided into two sub-questions. First, we elaborate on the composition of groups. And second, we will discuss with the help of the literature whether this group composition affects financial behavior.

4.7.1 What is the composition of groups? Many MFIs make use of the group lending mechanism. This mechanism begins with a lender, such as BRAC, deciding to enter a new location. This lender announces to the residents of this new location that in order to access credit, potential clients must form groups. In the case of BRAC clients need to form small groups, restricted to the size of five people, but this size varies across programs and regions. Clients obtain the loans and invest them independently. Access to the loans continues until the loan cycle is completed and as long as all loans are repaid. But if any member fails to repay, all group members are denied future credit. In this sense, group members are jointly liable for the loan repayments of their peers (Giné et al, 2006).

Discussion concerning this process of group composition has evolved; especially how groups form themselves and what types of people ‘choose’ each other. This is also called the process of assortative matching. Assortative matching means that if groups are allowed to form themselves, risky and safe borrowers will sort themselves into relatively homogenous groups. Thus safe

Page 16 of 120 borrowers will choose to form a group with other safe borrowers and therefore risky borrowers are forced to form a group with other risky borrowers (Guttman, 2008). One of the studies which stated that homogenous groups are better for repayments, are the models of Ghatak (1999, 2000) and van Tassel (1999.). Here the result found is that borrowers, if allowed to form their own groups, will sort themselves into relatively homogenous groups of “safe” and “risky” borrowers, which eventually lead to ‘positive’ assortative matching and higher benefits. Other authors state that heterogeneous groups lead to a higher profit, because risky borrowers help safe borrowers and otherwise. Contrary to most theoretical predictions, papers like Sadoulet and Carpenter (2001), and Lensink and Mehrteab (2001) find risk heterogeneity among borrowing groups in Guatemala and Eritrea respectively. Despite the risk heterogeneity, these programs have high repayment rates. The literature on group lending has suggested many theories whether homogeneity or heterogeneity may be better for repayment rates. However, empirical work testing the relatively importance of group composition for financial behaviour has lagged behind theoretical work on this topics (Ghatak, 1999).

The role that group members have in screening, monitoring and enforcement reduce problems of adverse selection and moral hazard, thereby improving repayment performance of groups. Next to the group members, there is also a group leader chosen. This group leader is chosen by the other members and has a leading role. Based on these we conclude that the group leader has a prominent role in the group (Hermes et al, 2006).

Another aspect of the group lending scheme are the loan officers. Loan officers attend every weekly meeting of the group. Loan officers are therefore the major link between MFIs and their clients and are central for service delivery. Because they mediate transactions between MFIs and clients (Dixon et al, 2007), loan officers are an important component of the group, even though they are not an actual member of this group.

Hypothesis 7: Homogeneous groups lead to a higher repayment

4.7.2 How does group composition affect financial behavior? As described above, self-selection of group can lead to homogeneous groups and heterogeneous grousp. Many authors have claimed that homogeneous groups will lead to a higher repayment rate (Huppi and Feder, 1990) (Ghatak, 1999, 2000) (van Tassel, 1999) (Gangopadhyay et al, 2005), due to matching of similar risk types and due to the facts that members screen and monitor the other members and enforce repayment, because individuals want to reduce the risk of having to contribute to the repayment of loans of others and because individuals wanto to ensure access to future loans (Hermes et al, 2006). Other authors state that heterogeneous groups lead to a higher profit (Sadoulet and Carpenter, 2001) (Lensink and Mehrteab, 2001), because risky borrowers help safe borrowers and otherwise, and therefore maximize the potential gain of mutual insurance.

The process and nature of group composition is crucial for improved performance of groups not only in terms of homogeneity, but also in terms of inherent social capital. Social capital can be of great benefit to both lenders and borrowers, due to social integration and linkages. These are important intra- and extra-group relationships which constitute an important source of social capital and this can have considerable influence on group performance (Olomola, 2002). Intra-group relationships, also known as relational social capital, in the form of personal trust between individuals and social

Page 17 of 120 homogeneity within groups have a positive effect on borrowing group performance (Cassar et al, 2007).

Another feature of group composition is the group leader. In the article of Hermes et al (2006), and this study, group leaders are elected after a group is formed. This group leader plays an important role as intermediary between the group and the program. The authors of this article have found, with empirical evidence, that social ties of group leaders have a positive effect on repayment performance of groups. They however did not find that monitoring activities of the group leader have a stronger positive impact on repayment performance. The study of Hermes et al (2006) is the first study to include the influence of the group leader. But they have only grasped one of the aspects of group leaders. Taking the leader-member exchange theory in consideration, which states that “the quality of leader-member relations determines the linkage between trust” (Gómez and Rosen, 2001), we expect that better quality of leader-members relations will lead to a higher trust in the group and therefore higher repayment rates.

Loan officers are a major link between MFIs and their clients; they therefore have some sort of an influence on the groups they supervise. Not much attention has been raised on these loan officers, but Dixon et al (2007) are one of the few who have. They have examined the role and challenges of loan officers in Zambia and have found that, coherent with the theory of Schreiner (1999) that loan officers play a key role in building and maintaining the social bonds between them and the group, the success of groups depends on loan officers. If loan officers fail, groups fail with them. This failure or success depends on how best the loan officers play and manages emotions and relationship with clients (Dixon et al, 2007). Here we will also make use of the leader-member exchange theory, and therefore expect that the better the relationship is between loan officer and group, the higher the trust between the two parties.

All three parties, e.g. the group, the group leader and the loan officer, are linked together through a network. In the study of Oke et al, (2007) they found that the more a clients is committed to group networking, for example by means of a social activity, the better the chance of repayment of microcredit. But as the group members are somehow ‘equal’ in the group, we expect the group leader and loan officer will not be. Mehra et al (2006) have found in their field-based study, that the embeddedness of leaders in the friendship networks of their subordinates, peers and supervisors has implications for objective group performance and leader reputation. Leaders are social architects who carefully manage, sustain and build their informal ties with others to enhance group performance and their own reputation.

Hypothesis 8: Higher trust relations between group members lead to a higher repayment

Hypothesis 9: Higher trust relations between group members and group leaders lead to a higher repayment

Hypothesis 10: Higher trust relations between group members and loan officers lead to a higher repayment

Hypothesis 11: Higher trust relations between group leaders and loan officers lead to a higher repayment

Page 18 of 120

5. Hypotheses All hypotheses are based on the literature discussed above.

1. H0/Ha: Clients self-select into groups with identical status and identical ethnic characteristics

2. H0/Ha: Clients self-select into groups with significant assets

3. H0/Ha: High financial sustainability of a MFI will lead to a low poverty outreach

4. H0/Ha: Bad infrastructure and presence of other MFIs will lead to a low poverty outreach

5. H0/Ha: Microfinance services lead to less vulnerable households

6. H0/Ha: High customer satisfaction will lead to a growth in clients and profits of MFIs

7. H0/Ha: Homogeneous groups lead to a higher repayment

8. H0/Ha: Higher trust relations between group members lead to a higher repayment

9. H0/Ha: Higher trust relations between group members and group leaders lead to a higher repayment

10. H0/Ha: Higher trust relations between group members and loan officers lead to a higher repayment

11. H0/Ha: Higher trust relations between group leaders and loan officers lead to a higher repayment

Page 19 of 120

6. Methodology

6.1. Sampling Evaluation

6.1.1 Sampling method There are two methods which we can use, these are:

 Regression discontinuity A regression discontinuity design (RDD) is an impact evaluation method that can be used for programs that have a continuous eligibility index with a clearly defined cutoff score (e.g. the four kilometer boundary used by branches) to determine who is eligible and who is not (Gertler et al, 2012). The idea behind it is that if the boundary was not there, the people living on both sides of the boundary would be exactly the same. The only thing that makes them different is the boundary, and in this case, having a loan.   Pipeline approach Using the pipeline approach, we will compare ‘old’ with ‘new’ members, as a method to measure impact. We will also conduct surveys with people who are not a member (but who are interested) in order to control for selection effects. 

The regression we can use for both these methods is:

= a vector of outcome variables of household i in location j. (outcome variables have to be distinguished; e.g. income, expenditures, sales, profits, client retention, etc.)

= a vector of household characteristics

= a vector of location characteristics

= a client dummy variable; 1 if household is client of BRAC (including new clients who do not yet have borrowed); 0 if households do not have a BRAC loan.

= accounts the amount of months that somebody already has a loan

In order to use both or either one of the two approaches, we need to sample:

1. Household with access to BRAC loans; a. Households who have a loan and are old members b. Households who have a loan and are new members c. Households who do not have a loan 2. Household without access to BRAC loans (but who are interested)

Page 20 of 120

Depending on which method we will use, we can use the second group differently for RD and the pipeline approach;

 RD: This will be the group that does not have access (outside the four kilometers boundary).  Pipeline: We ask whether they are interested to become a member of BRAC (or interested in borrowing if BRAC would open a new office in this area). If interested, they will score a 1 on the M dummy variable.

We will use block randomization in order to block rural/urban, old/new branches and old/new members. This will be explained in the next section.

6.1.2 Block randomization of branches and members BRAC has in total 30 branches (Table 9) (Table 9, Appendix A), which are situated in urban and rural areas in Sierra Leone. In order to capture the differences between rural and urban members, we therefore block first between urban and rural (Block 1). The second block we use is between old and new branches, we use this block to compare branches that are operating for several years with branches that just started their group loans. In Table 1, we see the amount of branches set up on a certain month and year. If we want to make an equal division between old and new branches, we choose the month and year that makes this division as January 2010. Therefore we have 13 old branches and 17 new branches (Block 2).

Table 1 Set-up date branches

Date of set-up Amount of Division branches old/new June 2009 6 December 2009 5 13 OLD January 2010 2 November 2010 4 December 2010 6 17 NEW September 2011 6 June 2012 1

The last block we use will be within the branches itself. This block will separate ‘old members’ from ‘new members’. We define them as follows; Old members are members who became a member in the first six months that a branch was operating. New members are members who became a member in the last six months a branch was operating. Using this block, we can compare members who are older and therefore more influenced by the group loan with members who are new and therefore may not have any change in their own wellbeing and the wellbeing of their household yet. As a means to control for a selection effect we will use as a first control group households who have access to BRAC loans, but who do not have a loan (C1). A second control group will be households who do not have access to BRAC loans but who are interested (C2) (Block 3).

We randomize within every block with the help of Excel randomization. Excel randomization means that you list all the units of analysis in a column. Then you give every unit of analysis a random number using the RAND option of Excel. The units of analysis then all get a random number between 0 and 1. After giving all units of analysis a random number, you either choose all units above 0.5 or under 0.5 to be your ‘treatment’ group.

Page 21 of 120

Depicting these three levels in Table 2, you get;

Table 2 Block randomization

Block Division made by block 1 Rural Urban 2 Old branch New branch Old branch New branch Old New Old New Old New Old New member member member member member member member member Control Control Control Control Control Control Control Control 3 with without with without with without with without access access access access access access access access (inside (outside (inside (outside (inside (outside (inside (outside 4km) 4km) 4km) 4km) 4km) 4km) 4km) 4km)

6.2 Use of control groups Some of the research questions are descriptive research questions, therefore it is not necessary to use control groups for every one of the research question.

The first research question “What are the determinants of participation in microfinance groups?” is a descriptive research question. Therefore the use of a control group is not necessary, but can be used to check whether there are specific characteristics which differs members from non-members. In this case we would use the control group C1, because they differ from members due to their non- eligibility.

The second research question “What are the impacts of microfinance services on poverty and vulnerability?” consists of two sub questions. Both of them have different methodology approach. In order to assess the first, “Do BRAC’s microfinance services contribute to the income of their clients?” we will use a so called poverty outreach. Hereby we compare the income of BRAC’s members with the average income of the population of Sierra Leone, while making sure the population group we are using have the same characteristics (age, gender, etc.) as BRAC’s members. In this case, the control group used is the population of Sierra Leone. An overview of this data can be found in a national database. The second sub question “What is the impact of microfinance services on vulnerability of their clients?” assesses the vulnerability of BRAC’s members with non-BRAC members. Thus, here we make us of a control group(s) (C2 or C1+C2).

The third research question “What is the level of client satisfaction with the microfinance services of BRAC?” is also a descriptive question and only have to be asked at BRAC’s members. If we make use of a depression index, where you measure how depressed someone is, then we have to make use of a control group (C2).

The fourth research question “What is the impact of microfinance loan group composition on financial behavior?” is divided into two sub questions. The first question “What is the composition of groups?” describes the members of a group and the relation between them; therefore a control group is not needed. The second question “How does group composition affect financial behavior?” tests the composition of the group and the effect of the composition on repayment. A comparison between members and non-members is not needed, because relations in a group cannot be set-up as a means of control.

Page 22 of 120

To conclude this section, Table 3 is compiled to provide a clear overview.

Table 3 Treatment and control group per research question

Research question ‘Treatment’ group Control group Determinants of participation Members Non-members inside boundary C1 Impact on poverty Members Population Sierra Leone (national database) Impact on vulnerability Members Non-members outside (and inside) boundary C2 (and C1) Customer satisfaction Members Non-members outside boundary (C2) (only if we use a depression index) Group composition Members None Composition affect financial Members None behavior Composition affect impact Members None

6.3 Sampling size Phase I We divide this study into two phases: Phase I and Phase II. In Phase I, we conduct all four research question in one branch; branch Mile-91. In Phase II, we only conduct research question 1, 2 and 3. We only conduct the fourth research question in Phase I, because we make use of behavioural games and due to the fact that these are time consuming and costly, we decide to only conduct them in one branch. In order to grasp the effect of group composition, we need to conduct the survey and behavioural games with the entire microfinance group. Therefore, the sample size of Mile-91 will be bigger than in the other branches.

When sampling Phase I, we select 20 microfinance groups of Mile-91, which are around 600 members. The list with the 20 randomly selected groups can be found in Appendix C: Sample selection.

We also have to select control groups for assessing research question 1, 2 and 3. These control groups will each have the size of 50 respondents (Table 4).

Table 4 Sampling Phase I

Mile- BRAC members Control with Control Total per branch 91 access (inside without 4km) access (outside 4km) Sample 600 50 50 700

6.4 Sampling size Phase II In Phase II we only conduct research question 1, 2 and 3. The total amount of microfinance borrowers of BRAC is approximately 28000 members divided over 30 branches. For the first two blocks we will select in total eight branches. These are divided into four rural and four urban branches. The second block is then the division of the four rural/urban branches into two old

Page 23 of 120 branches and two new branches (Table 5). The list with the eight randomly selected branches can be found in Appendix C: Sample selection.

Table 5 Sampling branches Pahse II

Block Division made by block Rural Urban 1 4 4 Old branch New branch Old branch New branch 2 2 2 2 2

Regarding sampling of block 3, we select within and around each branch randomly four groups of respondents (Table 6).

1. Households who have access to BRAC loans and are old members 2. Households who have access to BRAC loans and are new members 3. Households who have access to BRAC loans but do not have a loan 4. Households who do not have access to BRAC loans but are interested

Table 6 Sampling Phase II

Block 3 Old BRAC New BRAC Control with Control without Total per Total members members access (inside access branch sample 4km) (outside 4km) size Sample 50 50 50 50 200 1600

6.5 Methods There are several methods which we will use for this research. These are:

1. Individual survey – Group members 2. Individual survey – Group leader 3. Individual survey – Loan officer 4. Individual survey – Non member 5. Network Analysis 6. Coordination Game 7. Public goods Game 8. Dictator Game

All instruments and protocols can be found in the Appendix D: Instruments and Appendix E: Protocols.

6.5.1 Individual survey – Group members & Group leader In order to answer RQ 1,2 and 3, we make use of an evaluation survey. The survey includes a range of questions capturing socio-demographics, income, housing, health, food security, consumption, client satisfaction and risk behaviour. The survey questions have been developed based on the outcome variables and determinants tables, which can be found in Appendix B: Linking research questions, outcomes and measurements.

Page 24 of 120

The second part of the research, which is about group composition, will be answered by means of module 8 and some behavioural games. This extra module will include questions on the bigger group level, such as how long a group exists, composition of group (leaders), questions on the smaller group level, also how long a group exists, composition of group (leader), questions about individual borrowers and their loan outcomes. This survey will therefore also include a network survey of the (big and small) groups. Next to the survey three games will be played; an coordination game (CG), a public goods game (PGG) and a dictator game (DG).

The survey will be in English, but the enumerators will conduct the survey in either Krio, Mende or Temne, depending on the area where they conduct the survey. A survey will take about 60 minutes/hours to conduct.

6.5.2 Individual survey – Loan officer Next to the group members and group leader, there will also be a short survey for the loan officer, and there will be some behavioural games which the loan officer needs to play. The survey includes a range of questions capturing socio-demographic characteristics, risk behaviour, how long a loan officer is working with BRAC, how long the loan officer is working with the branch and the MF group, and what the tasks of the loan officers are. Next to the survey three games will be played; a coordination game (CG), a public goods game (PGG) and a dictator game (DG).

The survey will be in English, but the enumerators will conduct the survey in either Krio, Mende or Temne, depending on the area where they conduct the survey. A survey will take about 40 minutes/hours to conduct.

6.5.3 Individual survey – Non member In order to control for RQ 1,2 and 3, we make use of an evaluation survey for non-members. The survey includes a range of questions capturing socio-demographics, income, housing, health, food security, consumption and risk behaviour. The survey questions have been developed based on the outcome variables and determinants tables, which can be found in Appendix B: Linking research questions, outcomes and measurements.

The survey will be in English, but the enumerators will conduct the survey in either Krio, Mende or Temne, depending on the area where they conduct the survey. A survey will take about 60 minutes/hours to conduct.

6.5.4 Network analysis Intra-group relationships, also known as relational social capital, in the form of personal trust between individuals and social homogeneity within groups have a positive effect on borrowing group performance (Cassar et al, 2007). Therefore, we expect that trust in groups positively influence the financial performance of a group. In order to test this trust in a group, between group members, group leaders and loan officers, we will conduct a network analysis with a trust component.

We will record the composition in a group in terms of relations by kin, friends, family, business, etc.

The network analysis is part of the group survey, and will therefore only be conducted in Phase I of the research. The instrument will be in English, but the enumerators will conduct the instrument in Temne. The network analysis will take about 20 minutes/hours to conduct.

Page 25 of 120

6.5.5 Coordination game Last, we do a coordination game. Here the borrowers have to think about what most of the people in the group would most likely answer, and base their answers on this. If they give the answer which most people of the group answer, they can win a small amount of 500 Leones. The maximum amount they can win will be therefore 2000 Leones.

The coordination game is part of the group survey, and will therefore only be conducted in Phase I of the research. The instrument will be in English, but the enumerators will conduct the instrument in Temne. The coordination game will take about 15 minutes/hours to conduct.

6.5.6 Public goods game A public goods game gives us an indication of ‘free-riding’ in a group. We will expect that the higher the cohesion in a group (which is measured with the coordination game and the network survey), the less people will free ride. We will play this game on the small group level and on the big group level.

The game will go as follows: Everyone gets 1000 Leones, we ask them to give a certain amount (can be all, can be nothing) to a public good (a pot which gets divided over the group members), this amount gets tripled and then goes to the public good. This public good (the pot) then gets divided over the amount of people playing the game. This amount can maximum be 3500 Leones in the small group and 4000 Leones in the big group. Everyone then receives: their share of the pot + amount of Leones they kept of the small group.

The public goods game is part of the group survey, and will therefore only be conducted in Phase I of the research. The instrument will be in English, but the enumerators will conduct the instrument in Temne. The PGG will take about 15 minutes/hours to conduct.

6.5.7 Dictator game A dictator game is designed to capture benevolence of big group leader/small group leader/loan officer to the group members. Therefore we will play the benevolence game with all 5 small groups with big leader and loan officer.

The idea is to ask every small group member how much of 2500 they want to give to their other group members, to the big group leader en to the loan officer. We will ask this for every member of every small group. Then we will ask the big group leader how much they want to give of 2500 Leones to every other group member (so 29 people) and the loan officer. Then we will ask the loan officer how much they want to give to all group members (30 people) and the big group leader. Everything is anonymous: no one knows who the person is who is giving them a certain amount.

The dictator game is part of the group survey, and will therefore only be conducted in Phase I of the research. The instrument will be in English, but the enumerators will conduct the instrument in Temne. The DG will take about 15 minutes/hours to conduct.

Page 26 of 120

6.5.8 Methods per research question The first research question “What are the determinants of participation in microfinance groups?” is a descriptive research question. In order to assess this question, we will conduct a survey regarding a wide range of certain characteristics. We then use a probability function explaining ‘membership’ with this wide range of individual characteristics. If we make use of the control group C2, we use a probability function explaining ‘non-membership’ and these certain characteristics.

The second research question “What are the impacts of microfinance services on poverty and vulnerability?” consists of two sub questions, for which both we will use a survey. In order to assess the first, “Do BRAC’s microfinance services contribute to the income of their clients?” we compare the income of the members with the average income of the population. We will do this by using a t-test on income. The second sub question “What is the impact of microfinance services on vulnerability of their clients?” assesses the vulnerability of BRAC’s members with non-BRAC members. In order to find a difference between the two groups, we will also make use of a t-test on the variables used.

In order to analyze the third research question “What is the level of client satisfaction with the microfinance services of BRAC?” we will use a survey and a so called depression index. In a depression index you ask questions about how someone’s sees their selves and how depressed someone is. To compare the depression index, you use a t-test on the outcomes of the members and the non-members (Control group 1).

The fourth research question “What is the impact of microfinance loan group composition on financial behavior?” is divided into two sub questions. In order to assess the first question “What is the composition of groups?” we will use a survey for group members, a survey for group leaders and a survey for loan officers. To give an answer to the second question “How does group composition affect financial behavior?” we use a network analysis, a coordination game, a public goods game, a dictator game and the loan history per borrower.

The methods used per research question can be found in Table 7.

Table 7 Methods per research question

Research question Research tool Analysis tool Determinants of 1. Individual survey of group probability function explaining participation members & group leader ‘(non-)membership’ and certain 2. Individual survey of non-members characteristics Impact on poverty 1. Individual survey of group t-test income between members & group leader members and population 2. Individual survey of non-members Impact on vulnerability 1. Individual survey of group t-test between members and members & group leader control group (C1 and C2) 2. Individual survey of non-members Customer satisfaction 1. Individual survey of group Descriptive analysis members & group leader 2. Depression index T-test depression index Group composition 1. Individual survey of group members 2. Individual survey of group leader 3. Individual survey of loan officer

Page 27 of 120

Composition affect 1. Network analysis financial behavior 2. Coordination game 3. Public goods game 4. Dictator game 5. Loan history of members Composition affect 1. Network analysis impact 2. Coordination game 3. Public goods game 4. Dictator game 5. Impact on povery/vulnerability

6.6 Limitations

6.6.1 Identification of causal effects The theoretical set-up of an impact evaluation is that it assesses the changes in the well-being of individuals that can be attributed to a particular intervention, which in the case of this study is microcredit by group loans. This focus on attribution is the key stone of impact evaluations and causal questions. Thus, the central challenge in carrying out effective impact evaluations is to identify the causal relationship between the project and the outcomes of interest (Gertler et al, 2011).

In this study we will not be able to identify a causal relationship because the intervention has been implemented before the designing of this study. Therefore also randomization of treatment and control groups is not possible. A third limitation is that the baseline which has been conducted earlier in 2010 by BRAC cannot be used. This is due to the mobility of respondents; the merging of branches; the lack of keeping track of borrowers who completed or left their group, along with little identification data of borrowers. Due to these limitations, this study will not be identifying a causal relationship, but will be able to assess correlations.

6.6.2 Time Due to restricted availability of C.E.M. van Zaal until the 11th of April, the research has to be designed before this period and a part of the study has to be conducted.

Page 28 of 120

6.7 Planning The planning of this study is one of the most important aspects, due to the limited time available. An overview of the planning is being depicted in Table 8.

Table 8 Planning

Date What Where 20/02 Meeting Nasir FT 24/02-28/02 Exam enumerators FT Finish surveys/protocols Randomization branches Randomization groups Mile-91 Participant list Pay out sheets/record sheets 05/03 Pilot behavioural games and FT surveys 06/03 Edit surveys/protocols after FT Pilot 07/03 Presentation Tapan/Shohel FT 10/03-14/03 Training enumerators FT 16/03 Go to Mile-91 Mile-91 17/03-04/04 Conduct survey + games Mile-91 05/04-10/04 Edit surveys and logistics FT evaluation survey

Page 29 of 120

7. Data

7.1 Outcome indicators 1. What are the determinants of participation in microfinance groups? a. Outcome variable (Y): MF member

2. Do BRAC’s microfinance services contribute to the income of their clients? a. Outcome variable (Y1): Income MF members b. Outcome variable (Y2): Income population

3. What is the impact of microfinance services on vulnerability of their clients? a. Outcome variable (Y1): Education b. Outcome variable (Y2): Health c. Outcome variable (Y3): Housing d. Outcome variable (Y4): Financial security e. Outcome variable (Y5): Food security

4. What is the level of client satisfaction with the microfinance services of BRAC? a. Outcome variable (Y1): Client satisfaction b. Outcome variable (Y2): Depressiveness of clients c. Outcome variable (Y3): Aspirations d. Outcome variable (Y4): Locus of control

5. What is the composition of groups? a. Outcome variable (Y1): Homogeneity of group b. Outcome variable (Y2): Group formation c. Outcome variable (Y3): Composition of group leader in group d. Outcome variable (Y4): Composition of loan officer in group

6. How does group composition affect financial behavior? a. Outcome variable (Y1): Repayment behavior b. Outcome variable (Y2): Correlation between members c. Outcome variable (Y3): Strength of group d. Outcome variable (Y4): Group Leader in group e. Outcome variable (Y5): Loan officer in group

7. How does group composition affect the impact of microfinance services? a. Outcome variable (Y1): Income b. Outcome variable (Y2): Vulnerability c. Outcome variable (Y3): Correlation between members d. Outcome variable (Y4): Strength of group e. Outcome variable (Y5): Group Leader in group f. Outcome variable (Y6): Loan officer in group

Page 30 of 120

All determinants for the outcome variables can be found in Appendix B: Linking research questions, outcomes and measurements.

7.2 Control variables  control for other MFIs

7.3 Covariates - Age - Income - Ethnicity - Religion

Page 31 of 120

References Ackerly, B. (1997). What’s in a Design? The Effects of NGO Programme Delivery Choices on Women’s Empowerment in Bangladesh. In: Goetz, A.M. Getting the Institutions Right for Women in Development, pp. 140–60. Zed Books, New York.

African Development Bank (2006). Microfinance Policy and Strategy for the Bank Group. Operations Policies and Review Department.

Anyango, E., Sebstad, J. and Cohen, M. (2002). Assessment of the Use and Impact of MicroSave’s Market Research for MicroFinance Toolkit. MicroSave, Nairobi.

Armendáriz de Aghion, B. and Morduch, J. (2000). Microfinance beyond group lending. Economics of Transition. Vol 8, No. 2, pp. 401-420.

Banerjee, A.V. and Duflo, E. (2003). The economic lives of the poor. Journal of Economic Perspectives. Vol. 21, No. 1 ,pp. 141–167.

Bashir, M.S., Machali, M.M. and Mwinyi, A.M. (2012). The Effect of Service Quality and Government Role on Customer Satisfaction: Empirical Evidence of Microfinance in Kenya. International Journal of Business and Social Science. Vol. 3, No. 14.

Besley, T. and Coate, S. (1995). Group Lending, Repayment Incentives and Social Collateral. Journal of Development Economics. Vol. 46, No. 1, pp. 1–18.

Bowen, J. and Chen, S. (2001). The relationship between customer loyalty and customer satisfaction. International Journal of Contemporary Hospitality Management. Vol. 13, No. 5, pp. 213-217.

BRAC. (2013). Who We Are: Evolution. http://www.brac.net/content/who-we-are-evolution. 18-12- 2013.

Brau, J. C., Woller, G. M. (2004). Microfinance: A comprehensive review of the existing literature. Journal of Entrepreneurial Finance. Vol. 9, No. 1, pp. 1-27.

Cassar, A., Crowley, L. and Wydick, B. (2007). The effect of social capital on group loan repayment: Evidence from field experiments. The Economic Journal. Vol 117.

Chao-Béroff, R., Cao, T.H.H., Vandenbroucke, J.P., Musinga, M., Tiaro, E. and Mutesasira, L. (2000). A comparative analysis of member-based microfinance institutions in East and West Africa. Microsave, Kenya

Chaudhuri, S., Jalan, J. and Suryahadi, A. (2002). Assessing Household Vulnerability to Poverty from Cross-sectional Data: A Methodology and Estimates from Indonesia. Department of Economics, Columbia University, USA.

Churchill, C. (2000). Banking on Customer Loyalty. Journal of Microfinance. Vol. 2, No. 2.

Churchill, C. and Halpern, S. (2001). Building Customer Loyalty: A Practical Guide for Microfinance Institutions. Microfinance Network, Washington.

Page 32 of 120

Eakin, H. and Bojórquez-Tapia, L.A. (2008). Insights into the composition of household vulnerability from multicriteria decision analysis. Global Environmental Change. Vol. 18, pp. 112-127.

Dixon, R., Ritchie, J. and Siwale, J. (2007). Loan officers and loan ’delinquency’ in microfinance : a Zambian case. Accounting forum. Vol, 31, No. 1, pp. 47-71.

Fernando, J.L. (1997). Nongovernmental Organization, Mico-Credit, and Empowerment of Women. Annals of the American Academy of Political and Social Science. Vol 554, pp. 150-177.

Fischer, G. (2008). Contract Structure, Risk Sharing, and Investment Choice. Unpublished.

Gangopadhyay, S., Ghatak M. and Lensink, R. (2005). Joint Liability Lending and the Peer Selection Effect. Economic Journal. Vol. 115, No. 506, pp. 1005–1015.

Garikipati, S. (2008). The impact of lending to women on household vulnerability and women's empowerment: evidence from India. World Development.

Gertler, P.J., Martinez, S., Premand, P., Rawlings, L.B. and Vermeersch, C.M.J. (2011). Impact Evaluation in Practice. The World Bank.

Ghatak, M., (1999). Group lending, local information, and peer selection. Journal of Development Economics. Vol. 60, pp. 27–50.

Ghatak, M. and Guinnane, T.W. (1999). The economics of lending with joint liability: theory and practice. Journal of Development Economics. Vol 60, pp. 195-228.

Ghatak, M., (2000). Screening by the company you keep: joint liability lending and the peer selection effect. Economic Journal. Vol. 110, pp. 601–631.

Giné, X., Jakiela, P., Karlan, D.S and Morduch, J. (2006). Microfinance games, Center discussion paper. Economic Growth Center, No. 936

Giné, X., Jakiela, P., Karlan, D. and Morduch, J. (2010). Microfinance Games. American Economic Journal: Applied Economics. Vol. 2, pp. 60-95.

Gómez, C. and Rosen, B. (2001). The leader-member exchange as a link between managerial trust and employee empowerment. Group and Organization Management.

Gulli, H. (1998). Microfinance & Poverty: Questioning the Conventional Wisdom. Inter-American Development Bank.

Gulli, H. and Berger, M. (1999). Microfinance and poverty reduction – evidence from Latin America. Small Enterprise Development. Vol. 10, No. 3, pp. 16-28.

Guttman, J.M., (2008). Assortative matching, adverse selection, and group lending. Journal of Development Economics. Vol. 87, pp. 51-56.

Hentschel, J. and Waters, W.F. (2002). Rural poverty in Ecuador: Assessing local realities for the development of Anti-poverty programs, World Development. Vol. 30, No. 1, pp. 33-47.

Page 33 of 120

Henry, C., Sharma, M., Lapenu, C. and Zeller, M. (2003). Microfinance Poverty Assessment tool. Technical Tools Series No. 5. International Food Policy Research institute (IFPRI), commissioned by CGAP, The World Bank.

Hermes, N., Lensink, R. and Mehrteab, H.T. (2006). Does the Group Leader Matter? The Impact of Monitoring Activities and Social Ties of Group Leaders on the Repayment Performance of Group‐ based Lending in Eritrea. African Development.

Hermes, N., Lensink, R. and Meesters, A. (2008). Outreach and Efficiency of Microfinance Institutions. Centre for International Banking, Insurance and Finance, Faculty of Economics and Business, University of Groningen, the Netherlands.

Hulme, D. (2000). Impact Assessment Methodologies for Microfinance: Theory, Experience and Better Practice. World Development; Vol. 28, No. 1, pp. 79-98.

Huppi, M., Feder, G. (1990). The role of groups and credit cooperatives in rural lending. The World Bank Research Observer, Vol. 5.

Kandampully, J. and Suhartanto, D. (2000). Customer loyalty in the hotel industry: the role of customer satisfaction and image. International Journal of Contemporary Hospitality Management. Vol. 12, No.6, pp. 346-351.

Kono, H. (2014). Microcredit Games with Noisy Signals: Contagion and Free-Riding. Journal of the Japanese and International Economies.

Lensink, R. and Mehrteab, H. (2003). Risk behavior and group formation in microcredit groups in Eritrea. CREDIT Research Paper.

Mehra, A., Dixon, A.L., Brass, D.J. and Robertson, B. (2006). The Social Network Ties of Group Leaders: Implication for Group Performance and Leader Reputation. Organization Science. Vol. 17, No. 1, pp. 64-79.

Montgomery, H. (2004). Serving the poorest of the poor: the poverty impact of the Khushhali Bank’s microfinance lending in Pakistan. Asian Development Bank.

Oke, J.T.O., Adeyemo, R. and Agbonlahor, M.U. (2007). An Empirical Analysis of Microcredit Repayment in Southwestern Nigeria. Humanity and Social Sciences Journal. Vol 2, No. 1, pp. 63-74.

Olomola, A.S. (2002). Social capital, microfinance groups performance and poverty implications in Nigeria. Nigerian Institute of Social and Economic Research.

Paxton, J. (2003). A poverty outreach index and its application to microfinance. Economics Bulletin. Vol. 9, No. 2, pp. 1−10.

Rankin, K.N. (2000). Social Capital and the Foundations of Democratic Public Culture: Contributions from Critical Ethnography of Nepal. 6th Annual Conference of the University of Toronto - York University Joint Centre for Asia Pacific Studies. Democracy and Identity Conflicts in Asia: The Politics of Cultural Difference in an Age of Globalization. December 4–5, 2000.

Page 34 of 120

Rosenberg, R. (2006). Graduating the poorest into microfinance:Linking safety nets and financial services. CGAP Focus note 34.

Sadoulet, L. and Carpenter, S.B. (2001). Endogenous Matching and Risk Heterogeneity: Evidence on microcredit group formation in Guatemala. ECARES, Free University of Brussels.

Seibel, H. D., (2005). Does History Matter? The Old and the New World of Microfinance in Europe and Asia. Working paper. University of Cologne, Development Research Center, No. 2005, 10.

Schreiner, M. (1999). A scoring model of the risk of costly arrears at a microfinance lender in Bolivia. Unpublished manuscript. Available: http://www.microfinanace.com. Center for Social Development, Washington University in St Louis.

Siddiqi, K. O. (2011). Interrelations between Service Quality Attributes, Customer Satisfaction and Customer Loyalty in the Retail Banking Sector in Bangladesh. International Journal of Business and Management. Vol. 6, No. 3.

Srnec, K. and Svobodová, E. (2009). Microfinance in less developed countries: history, progress, present – charity or business? Institute of Tropics and Subtropics, Czech University of Life Science, Prague, Czech Republic.

Stiglitz, J. E. (1990). Peer Monitoring and Credit Markets. World Bank Economic Review. Vol. 4, No. 3, pp. 351–66.

Tassel, van, E., (1999). Group lending under asymmetric information. Journal of Development Economics. Vol. 60, pp. 3–25.

Westover, J. (2008). The record of microfinance: The effectiveness/ineffectiveness of microfinance programs as a means of alleviating poverty. Electronic Journal of Sociology.

Wilson, K. (2001). Microfinance Myopia. Journal of Microfinance. Vol. 3, No. 1.

Woller, G. (2002). From Market Failure to Marketing Failure: Market-Orientation As the Key to Deep Outreach in Microfinance. Journal of International Development. Vol. 14, No. 3.

Wright, G.A.N., Kasente, D., Ssemogerere, G. And Mutesasira, L. (1999). Vulnerability, risks, assets and empowerment – the impact of microfinance on poverty alleviation. World Development Report 2001.

Wright, G.A.N., Cracknell, D., Mutesasire, L. and Hudson, R. (2005). Strategic Marketing for MicroFinance Institutions. MicroSave – Market-led solutions for financial services.

Zaman, H. (2004). The scaling-up of microfinance in Bangladesh: determinants, impact, and lessons. World Bank Publications.

Page 35 of 120

Appendix A: BRAC

Figure 3 Map branches BRAC

Page 36 of 120

Table 9 Information branches

SL number Merged with Merged of the any other with branch branches Currently Sierra Branch established branch at that were operating List of current operating Leone Branch name District at month/year (yes/no) month/year merged (yes/no) branches at Jan, 14 1 Kambia Kambia sep'11 No Yes Kambia 2 Lunsar Portloko june'2009 No Yes Lunsar 3 Port Loko Portloko june '2009 No Yes Port Loko 4 Lungi 1 Portloko jan'2009 Closed August'13 4 No CLOSED 5 Lungi2 Portloko jan'09 Closed August'13 5 No CLOSED 6 Lokomasama Portloko oct'11 Closed August'13 6 No CLOSED 7 Masiaka Portloko june'2009 Closed August'13 7 No CLOSED 8 Rokupur Kambia sep'11 No Yes Rokupur 9 Kissy Bye Pass western urban june '2009 No Yes Kissy Bye Pass 10 Grass Field western urban jan'2010 No Yes Grass Field 11 Fullah Town western urban jan'2010 No Yes Fullah Town 12 Calaba Town western urban sep'11 No Yes Calaba Town 13 George Brook western urban jan'09 Merged March'13 13 No Merged with Krotown 17 14 Tengbeh Town western urban Dec'09 Merged March'13 14 No Merged with Krotown 17 15 Aberdeen western urban june '2009 No Yes Aberdeen 16 Goderich western urban june '2009 No Yes Goderich 17 Kroo Town western urban Dec'2009 No Yes Kroo Town 18 Lumly western urban Dec'09 Closed August'13 18 No CLOSED 19 Tagrin Portloko june '2012 No Yes Tagrin 20 Magburaka Tonkolili Dec'2009 No Yes Magburaka 21 Masuba Bombali Dec'2009 No Yes Masuba

Page 37 of 120

22 OIC Bombali Dec'2009 No Yes OIC 23 Wara Wara Koinadugu Dec'2010 No Yes Wara Wara 24 NP Manikala Bombali Dec'09 Merged March'13 24 No Merged with Masuba 21 25 Sengbeh Koinadugu Dec'10 Merged March'13 25 No Merged with Warawara 23 26 Kandeh Town Bo Nov'10 No Yes Kandeh Town 27 Mile-91 Tonkolili Dec'2009 No Yes Mile-91 28 Moriba Town Bo Nov'10 No Yes Moriba Town 29 Torkpoi Town Bo Dec'2010 No Yes Torkpoi Town 30 Kulandatown Bo Nov'10 Merged March'13 30 No Merged with Moriba town 28 31 Balama Road Kenema Nov'10 No Yes Balama Road 32 Small Bo Kenema Nov'10 No Yes Small Bo 33 Hangha Kenema Nov'10 Merged March'13 33 No Merged with Blama road 31 34 Dama Road Kenema Nov'10 Merged March'13 34 No Merged with Blama road 31 35 Tongo Field Kenema Dec'10 Closed August'13 35 No CLOSED 36 Masingbi Tonkolili Dec'10 No Yes Masingbi 37 Koquima Kono Dec'10 No Yes Koquima 38 So So Town Kono Dec'10 No Yes So So Town 39 Bongalow Kono Dec'10 Merged March'13 39 No Merged with SO SO town 38 40 Waterloo 1 western Rural june '2009 No Yes Waterloo 1 41 Waterloo 2 western Rural sep'11 No Yes Waterloo 2 42 Jui western Rural sep'11 No Yes Jui 43 Tombo western Rural sep'11 No Yes Tombo

Page 38 of 120

Appendix B: Linking research questions, outcomes and measurements

Research question 1:

What are the determinants of participation in microfinance groups?

Model: ,

= a vector of outcome variables of household i in location j. (outcome variables have to be distinguished; e.g. income, expenditures, sales, profits, client retention, etc.)

= a vector of household characteristics

= a vector of location characteristics

= a client dummy variable; 1 if household is client of BRAC (including new clients who do not yet have borrowed); 0 if households do not have a BRAC loan.

= accounts the amount of months that somebody already has a loan

Analysis: Probability function: ( ) ( )

(1) = MF member, thus probability that respondent is a MF member given certain characteristics.

Y X Survey Module Question Unit of analysis MF member 1. Age 1. 1 1. 1.1.2 1. New MF members 2. Education 2. 1 2. 1.1.3 2. Old MF members 3. Tribe 3. 1 3. 1.1.4 3. Non-member 4. Position in community 4. 1 4. 1.1.5 without access 5. Religion 5. 1 5. 1.1.6 6. Number of people in hh 6. 1 6. 1.1.7 Individual 7. Number of adults 7. 1 7. 1.1.8 Level 8. Number of children 8. 1 8. 1.1.9 survey 9. Occupation 9. 3 9. 3.1.1 10. Distance to branch 10. 8 10. 8.1.3 11. Distance to group meeting 11. 8 11. 8.1.2 12. Entrepreneurial skills 12. 3 12. 3.4.1 13. Risk type 13. 6 13. 6.1 3 3.3

Page 39 of 120

Research question 2:

What are the impacts of microfinance services on poverty and vulnerability?

2.1 Do BRAC’s microfinance services contribute to the income of their clients?

Model: ,

= income of MF members i

= a vector of income characteristics

= accounts the amount of months that somebody already has a loan

Analysis: T-test between “income MF members” and “income population”

Y X Survey Module Question Unit of analysis 1. Income MF 1. Weekly/monthly 1. 3 1. 3.1.2, 3.1.3, 1. New MF members income 2. 3 3.1.5, 3.1.6 members 2. Outstanding loans of 3. 2 2. 3.5.3 2. Old MF Individual hh 4. 5 members Level survey 3. Assets 3. 2.1 4. Expenditure 4. 5.1, 5.2

2. Income Income received from NA NA NA Population population national database

Page 40 of 120

2.2 What is the impact of microfinance services on vulnerability of their clients?

Model: ,

= education/health/housing/financial security/food security of respondent i

= a vector of (education/health/housing/financial security/food security) characteristics

= a client dummy variable; 1 if household is client of BRAC (including new clients who do not yet have borrowed); 0 if households do not have a BRAC loan.

= accounts the amount of months that somebody already has a loan

Analysis: t-test of every Y between the units of analysis

Y X Survey Module Question Unit of analysis 1. Education 1. Education level 1. 1 1. 1.1.3 member 2. Primary school 2. 1 2. 1.2.1, 1.2.2, Individual attendance 1.2.3 Level 3. Secondary school 3. 1 3. 1.2.4, 1.2.5, survey attendance 4. 1 1.2.6 4. Payment of school 4. 1.2.7 fees 2. Health 1. Perceived health 1. 4 1. 4.1.1, 4.1.2 status 2. Deaths in hh last 2. 4 2. 4.1.3 1. New MF year Individual member 3. Recent visits Level 3. 4 3. 4.1.4 2. Old MF hospital/doctor survey member 4. Access doctor 4. 4 4. 4.1.5 3. Non-member 5. Access hospital 5. 4 5. 4.1.6 with access 6. Diseases 6. 4 6. 4.2 4. Non-member 3. Housing 1. Assets 1. 2 1. 2.1 without access 2. Housing 2. 2 2. 2.2 Individual improvements Level 3. Housing 3. 2 3. 2.3 survey conditions 4. Ownership 4. 2 4. 2.4 4. Financial Savings Individual 3 3.6 security Level survey 5. Food security 1. 3 meals a day Individual 1. 5 1. 5.3.1 2. Experience hunger Level 2. 5 2. 5.3.2 3. Clean water survey 3. 5 3. 5.3.3

Page 41 of 120

Research question 3:

What is the level of client satisfaction with the microfinance services of BRAC?

Model client satisfaction: ,

= client satisfaction of MF member i = a vector of client satisfaction characteristics = accounts the amount of months that somebody already has a loan Model depressiveness of clients: ,

= depressiveness of respondent i

= a vector of depressiveness characteristics

= a client dummy variable; 1 if household is client of BRAC (including new clients who do not yet have borrowed); 0 if households do not have a BRAC loan.

= accounts the amount of months that somebody already has a loan

Analysis:

Client satisfaction: mean of (Y) Depressiveness of clients: t-test between MF member and non-member Y X Survey Module Question Unit of analysis 1. Client 1. Loan application system 1. 7.1.1 1. New MF member satisfaction 2. Waiting time to receive loan 2. 7.1.2 2. Old MF member 3. Maximum loan size 3. 7.1.3 4. Interest rate 4. 7.1.4 5. Loan application fee 5. 7.1.5 6. Repayment period 6. 7.1.6 7. Weekly instalment 7. 7.1.7 8. System of paying overdue Individual 8. 7.1.8 9. Increment in loan size Level 7 9. 7.1.9 10. Amount of meeting survey 10. 7.1.10 11. Duration of meeting 11. 7.1.11 12. Professional behavior of LO 12. 7.1.12 13. LO’s reaction to suggestions 13. 7.1.13 14. Accessibility of BRAC 14. 7.1.14 15. Distance to branch 15. 7.1.15 16. Distance to group meeting 16. 7.1.16 17. In general how satisfied? 17. 7.1.17 2. Depression index 1. New MF member Depressiveness 2. Old MF member Individual of clients 3. Non-member with Level 7 7.2 access survey 4. Non-member without access 3. Aspirations 4. Locus of control

Page 42 of 120

Research question 4:

What is the impact of microfinance loan group composition on financial behavior?

4.1 What is the composition of groups?

Model homogeneity of group:

= a vector of outcome variables of MF member i

= a vector of socio-demographic characteristics

= accounts the amount of months that somebody already has a loan

Model homogeneity of group leader:

= a vector of outcome variables of group leader i

= a vector of socio-demographic characteristics

= accounts the amount of months that the group leader is a group leader

Model homogeneity of loan officer:

= a vector of outcome variables of loan officer i

= a vector of socio-demographic characteristics

= accounts the amount of months that a loan officer has been working with the group

Analysis:

Homogeneity of group: F-test for X’s amongst members of small group, testing = 0

Position group leader: F-test for socio-demographics amongst members of small group and group leader, testing = 0. Other X’s are descriptive.

Position loan officer F-test for socio-demographics amongst members of small group and group leader, testing = 0. Other X’s are descriptive.

Y X Survey Module Unit of analysis 1. 1. Age 1. 1 1. 1.1.2 1. Lg. group Homogeneity 2. Education 2. 1 2. 1.1.3 members of group 3. Tribe Individual 3. 1 3. 1.1.4 2. Sm. Group 4. Position in community Level 4. 1 4. 1.1.5 members 5. Religion survey 5. 1 5. 1.1.6 6. Number of people in hh (Group 6. 1 6. 1.1.7 7. Number of adults members 7. 1 7. 1.1.8 8. Number of children and Group 8. 1 8. 1.1.9 9. Occupation Leader) 9. 3 9. 3.1.1 10. Distance to branch 10. 8 10. 8.1.3 11. Distance to group meeting 11. 8 11. 8.1.2

Page 43 of 120

12. Entrepreneurial skills 12. 3 12. 3.4.1 13. Risk type 13. 6 13. 6.1 3 3.2 2. Group 1. When group formed 1. 8 1. 8.1.4 1. Group formation 2. Why do members join group Individual 2. 8 2. 8.1.5 members 3. Group formation based on Level 3. 8 3. 8.1.6 2. Group Leader 4. Refusal of membership Survey – 4. 8 4. 8.1.14 5. Decision making in meeting Group 5. 8 5. 8.1.15 6. Inability to pay weekly members 6. 8 6. 8.1.19 repayments and Group 7. 8 7. 8.1.20 7. Members pay for other leader members 3. Group 1. Socio-dem characteristics 1. 1 1. 1.1- 1. Group leader leader Individual 1.9 2. Risk type Level 2. 6 2. 6.1 Survey – 3 3.2 3. How long leader in group Group 3. 8 3. 8.1.9 4. How chosen Leader 4. 8 4. 8.1.8 5. What tasks 5. 8 5. 8.1.10 4. Loan officer 1. Socio-dem characteristics 1. 1 1. 1.1 – 1. Loan officer 1.6 2. Risk type 2. 6 2. 6.1 Individual 3 3.2 Level 3. How long LO at BRAC 3. 8 3. 8.1.1 Survey – 4. How long LO at branch 4. 8 4. 8.1.2 Loan officer 5. How long LO for group 5. 8 5. 8.1.5 6. Rate working with group 6. 8 6. 8.1.7 7. What tasks 7. 8 7. 8.1.12

Page 44 of 120

4.2 How does group composition affect financial behavior?

Model repayment rate:

= repayment behavior of MF member i

= a vector of repayment characteristics

= accounts the amount of months that somebody already has a loan

Model strength of group:

= strength of group i

= social cohesion of group i using coordination game

= social cohesion of group i using network analysis

= freeriding of group i using public goods game

= accounts the amount of months that somebody already has a loan

Analysis:

Repayment rate: take mean (Y) of group members

Strength of group: take (Y) of group

Group composition  financial behavior:

Probability function: ( ) ( )

(1) = ‘Perfect’ repayment, thus probability that group has perfect repayment behavior given the strength of the group

Or correlation?!?

Y X Survey Module Question Unit of analysis 1. Repayment 1. Number of previous NA NA NA 1. Lg. group members behavior loans 2. Sm. Group members 2. Number of current loans 3. Amount of previous loans 4. Amount of current loans 5. Are current loans paid off? 2. Correlation 1. Best income of past 5 Survey 1. 3 1. 3.2 1. Are answers of between years members similar members 2. Worst income of past 2. 3 2. 3.2 (some bad and 5 years worst years) 3. Fractions of revenue 3. 5 3. 5.2 2. Check for

Page 45 of 120

spend on different correlations on the categories percentage spending of group members 3. Strength of 1. The amount of what a 1. Social 1. Lg. Group members group small group wins; so cohesion in 2. Sm. Group members the amount of ‘right’ group 3. Group leader answers (Coordinatio 4. Loan officer 2. The amount of links in n game) a small group 2. Network 3. The amount of links in analysis a big group 3. Network 4. The total amount that analysis members of a small 4. Freeriding in group contribute to group the Public good (Public 5. The total amount of goods game) the small group 5. Benevolence members willing to in group share to: (Dictator - the other small group game) members 6. Benevolence - the big group leader in group - the loan officer (Dictator 6. The total amount of game) the big group leader 7. Benevolence willing to share to: in group - the group members (Dictator of the big group game) - the loan officer 7. The total amount of the loan officer willing to share to: - the group members of the big group - the big group leader 4. Group 1. The amount of links 1. Network Leader between group and analysis group member 2. Freeriding in 2. The total amount that group the group contributes (Public to the public good goods game) when group leader is 3. Freeriding in in the group group 3. The total amount that (Public the big group leader goods game) contributes to the 4. Benevolence Public good in group 4. The total amount of (Dictator the small group game) members willing to 5. Benevolence

Page 46 of 120

share to: in group - the other small group (Dictator members game) - the big group leader 6. Benevolence - the loan officer in group 5. The total amount of (Dictator the big group leader game) willing to share to: - the group members of the big group - the loan officer 6. The total amount of the loan officer willing to share to: - the group members 5. Loan officer 1. The amount of links 1. Network between group and analysis loan officer 2. Freeriding in 2. The total amount that group the group contributes (Public to the public good goods game) when loan officer is in 3. Freeriding in the group group 3. The total amount that (Public the loan officer goods game) contributes to the 4. Benevolence Public good in group 4. The total amount of (Dictator the small group game) members willing to 5. Benevolence share to: in group - the other small group (Dictator members game) - the big group leader 6. Benevolence - the loan officer in group 5. The total amount of (Dictator the big group leader game) willing to share to: - the group members of the big group - the loan officer 6. The total amount of the loan officer willing to share to: - the group members

Page 47 of 120

4.3 How does group composition affect the impact of microfinance services?

Model income: : ,

= income of respondent i

= a vector of income characteristics

= accounts the amount of months that somebody already has a loan

Model vulnerability: : ,

= education/health/housing/financial security/food security of respondent i

= a vector of (education/health/housing/financial security/food security) characteristics

= accounts the amount of months that somebody already has a loan

Model strength of group:

= strength of group i

= social cohesion of group i using coordination game

= social cohesion of group i using network analysis

= freeriding of group i using public goods game

= accounts the amount of months that somebody already has a loan

Analysis:

Group composition  impact on income and vulnerability of MF members:

Correlation between impact on income and vulnerability and strength of group

Y X Unit of analysis Income Same as RQ 2.1 Group members Vulnerability Same as RQ 2.2 Group members Strength of group Same as RQ 4.2

Correlation members Same as RQ 4.2 Position of group Same as RQ 4.2 Leader in group Position of loan officer Same as RQ 4.2 in group

Page 48 of 120

Appendix C: Sample selection

PHASE 1

In the first Phase of the research we will implement the survey in the branch Mile-91. The 20 groups which are randomly selected are:

1. 2030 2. 2028 3. 2026 4. 2024 5. 2023 6. 2022 7. 2021 8. 2018 9. 2017 10. 2016 11. 2014 12. 2013 13. 2011 14. 2009 15. 2008 16. 2007 17. 2006 18. 2005 19. 2004 20. 2002

PHASE 2

In the second Phase of the research, we will implement the survey in eight randomly selected branches. These branches are:

1. Waterloo 2 2. Small Bo 3. Kandeh Town 4. Kroo Town 5. Aberdeen 6. OIC 7. Magburaka 8. Tombo

Page 49 of 120

Appendix D: Instruments

Page 50 of 120

Evaluation MF BRAC Sierra Leone 2014

Group Member Survey

Consent of Respondent

(Interviewers: Clearly read out the following to the respondent and take her consent before taking interview)

I am …………………………………………….. from BRAC – an international NGO operating in microfinance services. We are currently conducting a survey in your community which broadly aims at generating information that will augment our services to people in this community. You have been selected as one of the respondents for this survey in this community. Please be aware that your participation is completely voluntary and you are free to withdraw at any time. However, because we want this survey to be as representative as possible of the views of people in this community, your participation is vital to enhance the success of this research. No financial token will be provided nor do we require any token from you for your participation. Do you agree to participate?

Yes No

Signature/Fingerprint ------(optional)

[Take the interview after having obtained the consent]

Branch Code

Group Code

Loan officer Code

MF member Code

Non MF member Code

DATA COLLECTION Enumerator Code Date _ _/_ _/_ _

Supervisor Code Date _ _/_ _/_ _

DATA ENTRY Enumerator 1 Code Date _ _/_ _/_ _

Enumerator 2 Code Date _ _/_ _/_ _

Supervisor Code Date _ _/_ _/_ _

Page 52 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

MODULE 1: SOCIO-DEMOGRAPHIC PROFILE

1.1 Household

1. Name 2. Age 3. Education Level 4. Tribe 5. Position in 6. Religion 7. No. of 8. No. of 9. No. of Community people in Adults Children HH 1. None 1. Mende Wife of the 1. Christian 2. Muslim 2. Some primary 2. Time chief school 3. Fulla 3. Other, specify 3.Completed primary 4. Gola Women’s school 5. Limba leader 4. Some JSS 6. Gbandi Youth Leader 5.Completed JSS 7. Kissi Union board 6. Some SS 8. Vai 7.Completed SS 9. Madingo 8. Some college 10. Other, 9.Completed college specify 10. Arabic education 11. Other, specify

1.2 Education of children

1. How many 2. How many of 3. If any do not attend 4. How many 5. How many of them 6. Of any do not attend 7. Who pays the children are primary them attend primary school, why not? children are attend secondary secondary school, why school fees? school aged? primary school? secondary school school? not?? aged? Amount Amount 1. No money for school fee Amount Amount 1. No money for school fee 1. We do ourselves 2. No money for transport 2. No money for transport 2. Family 3. No school 3. No school 3. Friends 4. Need children to work in 4. Need children to work in 4. Scholarship business business 5. Donor 5. Other, specify 5. Other, specify 6. Other, specify

Page 53 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

MODULE 2: HOUSING

2.1 Assets

Type of asset 1. Do you 2. How 3. If you 2.2 Housing improvements own this many do would buy a 1. type of 2. Costs of 3. Source of funding item? you own? similar item, improvement of improvement how much house would it cost? 1. Walls In Leones 1. Earnings/savings 2. Doors and windows 2. Relatives 1. Yes In Leones 3. Roof 3. Friends/neighbours 2. No 4. Kitchen 4. BRAC 1. Bed 5. Fences 5. Other NGO 2. Table 6. Tubewell 6. Government 3. Chair 7. Sanitary latrine 7. Other, specify 4. Light/lantern 8. Other, specify 5. Big pot 1. 6. Bucket 2. 7. Bible/Koran 3. 8. Homestead land

(town lot) 2.3 Housing conditions 9. Shop premises 10. Radio/cassette player Does your home have; 1. Yes, new; 2. Yes, old; 3. No 11. Television 12. Fan 1. Pan roof 13. Cellular phone 2. Mud and sticks walls 14. Generator 3. Cements walls 15. Bicycle 4. Earth floor 16. Motorbike 5. Wood floor 17. Refrigerator 6. Concrete floor

18. Sewing machine 7. Electricity 19. Computer 8. Private toilet 20. Other valuable assets

Page 54 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

2.4 Ownership

1. For how long have you been living in your house? Mention in years 2. What is the ownership status of your house? 1. Owned and built 5. Rented 2. Owned and purchased 6. Using without paying rent 3. Owned and inherited 7. Temporary dwelling 4. Given as a favour 8. Other, specify 3. If owned, what is the value of your house? In Leones 4. If rented, how much do pay per month? In Leones

MODULE 3: ECONOMIC AND INCOME RELATED ACTIVITIES

3.1 Occupation

1. What is your first 2. Daily income first 3. Weekly income first 4. What is your second 5. Daily income second 6. Weekly income occupation? occupation occupation occupation? occupation second occupation 1. Farming In Leones In Leones 1. Farming In Leones In Leones 2. Petty trading 2. Petty trading 3. Salary employee 3. Salary employee 4. Wage employee 4. Wage employee 5. Shop keeping 5. Shop keeping 6. Carpeting 6. Carpeting 7. Tailoring 7. Tailoring 8. Fishing 8. Fishing 9. Mining 9. Mining 10. Hair dressing 10. Hair dressing 11. Blacksmithing 11. Blacksmithing 12. Gardening 12. Gardening 13. Other, specify 13. Other, specify

3.2 Income past five years

1. Which of the past five years was BEST in terms of household income? 2009/2010/2011/2012/2013 2. Which of the past five years was WORST in terms of household income? 2009/2010/2011/2012/2013

Page 55 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

3.3 Expected income

3. What would your income be in the coming year if it were a GOOD year? In Leones 4. What would your income be in the coming year if it were a BAD year? In Leones 5. What do you expect your income to be this year (2014)? In Leones 6. What do you expect your income to be next year (2015)? In Leones

3.4 Business plans

3.4.1 Do you have any new business plan you want to start? 1. Yes; 2. No If NO, skip 3.2.2

3.4.2 New business plan

Activity 1. Farming 8. Fishing 2. Petty trading 9. Mining 1. 3. Salary employee 10. Hair dressing 4. Wage employee 11. Blacksmithing 2. 5. Shop keeping 12. Gardening 6. Carpeting 13. Other, specify 3. 7. Tailoring

3.5 Outstanding loans

3.5.1 Do you have an outstanding loan? 1. Yes; 2. No If NO, skip 3.3.3

3.5.2 Does anyone in your household have an outstanding loan? 1. Yes; 2. No If NO, skip 3.3.3

Page 56 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

3.5.3 Outstanding loans

1. Who 2. From 3. Amount of loan 4. Interest rate 1. Yourself 1. Formal bank 6. Relative In Leones In percentages 2. Husband 2. BRAC 7. Friend/neighbour 3. Other hh 3. Other MFI 8. Local money lender member 4. Landlord 9. Other, specify 5. Employer 1. 2. 3.

3.6 Savings

How much (cash) savings do you have?

1. Savings at home 2. Savings with 3. Bank 4. NGO 5. If NGO, which one someone you trust Amount in Leones

3.7 Involvement NGOs

Which other development programs (NGOs) do you know in your community?

Name of development program/NGO Activity Are you involved with this program/NGO? 1. Financial aid 5. Microfinance lending 1. Yes 2. Food aid 6. Human rights training 2. No 3. Medical aid 7. Women’s empowerment 4. Education aid 8. Other, specify 1. 2. 3. 4. 5. Page 57 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

MODULE 4: HEALTH

4.1 Health and health access

4.1.1 How do you perceive your current health status? 1. Good; 2. Average; 3. Bad 4.1.2 How do you consider your health compared to last year? 1. Better than last year; 2. Same as last year; 3. Worse than last year 4.1.3 How many death in household last year Amount 4.1.4 How many recent visits to doctor/hospital last year Amount 4.1.5 How long will it take you to get to a doctor? 1. Less than 15 min; 2. 15 - 30 min; 3. 30 - 1 hour; 4. 1 hour; 5. 2 hours or more 4.1.6 How long will it take you to get to a hospital? 1. Less than 15 min; 2. 15 - 30 min; 3. 30 - 1 hour; 4. 1 hour; 5. 2 hours or more

4.2 Diseases

For the following diseases, how many people in the household were sick in the last six weeks

1. Number of adults 2. Number of children 1. Malaria 2. Diarrhea 3. Blood in stool 4. Infection 5. Yellow fever 6. Typhoid 7. Cholera 8. Common cold 9. Other, specify

Page 58 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

MODULE 5: CONSUMPTION

5.1 Food consumption

1. Consumption 2. Did you 3. Did you buy this 4. How much did you 5. Did you use your stock 6. If you would buy a similar item, good consume good? spend on this good? for consumption of this how much would it cost? this good? good? 1. Yes 1. Yes In Leones 1. Yes In Leones 2. No 2. No  skip to 5. 2. No 1. Rice (white) 2. Rice (native) 3. Cassava 4. Yams 5. Groundnut 6. Beans 7. Plantain 8. Fruits 9. Fish 10. Chicken 11. Beef 12. Palm oil 13. Maggi 14. Other, specify

5.2 Fractions of revenue spend on different categories

How much is the fraction (%) of your income spend on In percentages 1. Consumption 2. Transport 3. Education (uniforms/fees) 4. Health (medicines/doctor) 5. Clothing 6. Building/improvement costs of house BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘ Page 59 of 120

5.3 Non-food consumption

EDUCATION Amount in Leones CLOTHING Amount in Leones 1. School fees 9. Wearings 2. Books, pencils, etc. 10. Other, specify 3. School uniforms 4. Other, specify CEREMONIES Amount in Leones 11. Religious ceremonies HEALTH Amount in Leones 12. Memorial ceremonies 5. Purchase of medicines 13. Burials 6. Stay in hospital 14. Marriage 7. Herbalist 15. Secret society 8. Other, specify 16. Other, specify

5.4 Food security

5.3.1 How many meals do you eat a day Amount 5.3.2 Do you ever experience hunger? 1. Yes; 2. Sometimes; 3. No 5.3.3 Do you drink welled or bottled water? 1. Yes; 2. Sometimes; 3. No

MODULE 6: RISK

6.1 Would you like to receive A with certainty or B with a chance of 50 % of receiving it and a chance of 50% of receiving nothing?

A B A or B 5.000 Leones with certainty (100%) 4.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 3.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 2.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 1.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 0.500 Leones with uncertainty (50%)

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘ Page 60 of 120

MODULE 7: CLIENT SATISFACTION (ONLY FOR MFI MEMBERS)

7.1 Client satisfaction Please rank the following

Please rank the following 1 = Rarely or none (less than 1 day a week) 2 = Sometimes (1 to 2 days a week) 1 = Highly satisfied 3 = Occasionally (3 to 4 days a week) 2 = Moderately satisfied 4 = Most of the time (5 to 7 days a week) 3 = Moderately unsatisfied 4 = Highly unsatisfied 1. I was bothered by things that usually don’t bother me (bothered = angry/irritated) 1. Loan application system 2. I did not feel like eating; my appetite was poor 2. Waiting time to receive loan 3. I felt that I could not shake off the blues even with help 3. Maximum loan size from my family or friends 4. Interest rate 4. I felt I was just as good as other people 5. Loan application fee 5. I had trouble keeping my mind on what I was doing 6. Repayment period 6. I felt depressed 7. Weekly instalment 7. I felt that everything I did was an effort 8. System of paying overdue 8. I felt hopeful about the future 9. Increment in loan size 9. I thought my life had been a failure 10. Amount of meeting 10. I felt fearful 11. Duration of meeting 11. My sleep was restless 12. Professional behaviour of LO 12. I was happy 13. LO’s reaction to suggestions 13. I talked less than usual 14. Accessibility of BRAC 14. I felt lonely 15. Distance to branch 15. People were unfriendly 16. Distance to group meeting 16. I enjoyed life 17. In general, how satisfied are you 17. I had crying spells with BRAC’s services 18. I felt sad 19. I felt that people dislike me 20. I could not get going

7.2 Depression index Page 61 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

7.3 Aspirations

Have a look at the picture The first level (1) represents the best possible life for you, the lowest level (10) the worst 1. On what level are you now? Level 1-10 2. On what level do you want to be? Level 1-10 3. On what level do you think you are Level 1-10 in 5 years? In 5 years [question] will be 4. Your income A = much better B = better 5. Your health C = the same D = bad 6. Opportunities for you children E = much worse X = I do not know During the next five years I will: 7. Prove my education 1. Yes; 2. No 8. Find a better job 1. Yes; 2. No 9. Improve my education 1. Yes; 2. No 10. Marry and have children 1. Yes; 2. No

11. I strongly pursue my goals 1 = Definitely true 12. There are a lot of ways to solve a problem 2 = Somewhat true 13. I can think several ways to get things in life that are most important to me 3 = Slightly true 14. Even when others get discouraged, I know that I can find a way to resolve a problem 4 = Slightly false 15. My past experiences made me well prepared for my future 5 = Somewhat false 16. I had some success in life 6 = Definitely false 17. I attain the goals I set for myself

7.4 Locus of control 18. With which statement do you agree? A: everyone is primarily responsible for his success or failure in his life. B: success or failure in life is a question for each of the person. 19. With which statement do you agree? A: To be successful, above all should work very hard. B: To be successful, above all must have the luck

Page 62 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

MODULE 8: GROUP (ONLY FOR MFI MEMBERS)

8.1 Group information

1. When did you become a member of BRAC? Month – Year 2. When did you become a borrower of BRAC Month - Year 3. When did you become a member of the group? Month – Year 4. When was the group formed? Month – Year 5. Have you been a member of this group since it started? 1. Yes; 2. No 6. Why did you join the group? 1. Needed loans 2. Peer pressure 3. No access to bank 4. No access to other sources of savings 5. Other, specify 7. What is the group formation based on? 1. Similar income 5. Friends 2. Similar occupation 6. Relatives 3. Neighbouring land 7. Other, specify 4. Neighbouring house 8. What is your position in the group 1. Big group leader 4. Normal member 2. Small group leader 5. Loan officer 3. Big and small group leader 9. If you have are [1], [2] or [3], what are your specific tasks? 1. Opening group meeting 5. Punish if group member 2. Collecting weekly did not paid weekly repayments repayments 3. Decide if new borrower is 6. Kick group members out allowed in group of program 4. Punish if group member is 7. Other, specify too late 10. If you are a big group leader, when did you become the big group leader? Month – Year 11. How were you selected? 1. Anonymous election by group 2. Non-anonymous election by group 3. Elected by loan officer 4. Elected by previous group leader 6. Other, specify Page 63 of 120

12. What are the most important qualifications that a group leader must 1. Wealthy have? 2. Position in community 3. Good speaker Multiple answers possible 4. Trustworthy 5. Oldest of the group 6. Other, specify 13. If a woman you didn’t know requested to join your group, and hasn’t been 1. Yes approved by BRAC, would you allow her? 2. No 14. If no, what’s the main reason? 1. There is a limit to group size 2. We don’t know her 3. It’s a hassle to get her approved by BRAC 4. Other, specify 15. Has anyone been refused membership? 1. Yes 2. No 16. If yes, what’s the main reason? 1. Did not need any more 4. We did not know her members 5. Other, specify 2. Not trustworthy 3. Not similar income 17. How many members have joined in the last five years? Amount 18. How many people who would like to be member of this group but cannot? Amount 19. How many people have left this group? Amount 20. Who takes the decisions in the meetings for the following activities? 1. Voting of majority of members a. ____ a) Group positions 2. Big group leader only b. ____ b) Loan to members 3. Small group leader only c. ____ c) Penalty for non-payment 4. Big and small group leaders d. ____ d) Activities of meeting 5. Loan officer of BRAC 6. Other, specify 21. Do you attend the weekly meetings? 1. Yes; 2. Often; 3. Sometimes; 4. No 22. How long do you have to walk from your living place to the weekly In minutes meeting place? 23. How long do you have to walk from your living place to the branch? In minutes 24. How many times (in terms of weeks) have you been unable to repay your weekly instalments in the last three months? 25. Has one of your group members ever been unable to repay the weekly 1. Yes repayment? 2. No 26. If Yes, did you ever paid for her weekly repayment? 1. Yes Page 64 of 120

2. No 27. In the history of this group, have group members ever repaid loans for 1. Yes other members when they were unable to? 2. No 28. How many times did this happen in your group? Amount 29. Can you remember which group members were unable to repay their ID loans? 30. How many weekly meetings have you missed in the last three months? 31. Is there a fee for a group member/leader for not attending the meetings? 1. Yes 2. No 32. If Yes, how much is the fee? In Leones 33. Is the group leader aware of the debt and savings that members have with 1. Yes BRAC 2. No 34. Are other members aware of the debt and savings that members have 1. Yes with BRAC? 2. No 35. Is the group leader aware of the debt and savings that members have with 1. Yes institutions other than BRAC? 2. No 36. Are other members aware of the debt and savings that members have 1. Yes with institutions other than BRAC? 2. No 37. Is the group leader aware of debts that members have to individuals? 1. Yes 2. No 38. Are other members aware of debts that members have to individuals? 1. Yes 2. No

8.2 Coordination game

If you had to guess what most people in the big group think, guess which answer most of the people in the big group would give.

If you say “I don’t know” no earnings can be made!

Who would most people in the group think of: Name of person ID of person 1. If a new group leader is needed, who would be the most likely to be chosen 2. Who is most likely to show up early in a meeting 3. Who is the best dressed of the group 4. Who is the best singer of the group

Page 65 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

8.3 Network

ID 1. Are 2. Are 3. Have 4. How 5. Do 6. Do you go ID 1. Are 2. Are 3. Have 4. How 5. Do 6. Do you go you you you lent many you do to the same you you you lent many you do to the same friends blood money to times per business church/mosq friends blood money to times per business church/mosq with family this week do with this ue as this with family this week do with this ue as this this with this person? you meet person? person? this with this person? you meet person? person? person person? outside of person person? outside of ? the BRAC ? the BRAC meeting? meeting? 1. 21. 2. 22. 3. 23. 4. 24. 5. 25. 6. 26. 7. 27. 8. 28. 9. 29. 10. 30. 11. 31. 12. 32. 13. 33. 14. 34. 15. 35. 16. 36. 17. 37. 18. 38. 19. 39. 20. 40.

Page 66 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

8.4. Public good game

Say to the respondent:

We are going to play a game. The first round of the game will be played with the members of your small group. I am going to give you and your group members 2000 Leones. Then we will ask every member individually how much of the 2000 Leones they want to put in a public pot. There is no obligation to put something in the public pot, it is only what you want to put in the public pot. The total amount of the public pot will then get tripled and equally divided over the members of the group, no matter how much someone has put in the pot. The amount you will receive at the end of the game is how much you kept, so the amount you did not put in the game, plus whatever you get returned from the public pot.

How much do you want to put into the public pot?

ID Amount put in the pot Amount returned from the pot Total amount received

The second round of the game will be played with the members of the big group. You are getting the same amount of Leones and are asked again how much you want to put in the public pot. The amount you will receive at the end of the game is the amount you kept and the amount you get returned from the public pot.

How much do you want to put into the public pot?

ID Amount put in the pot Amount returned from the pot Total amount received

8.5 Dictator game

Say to the respondent: I am giving you 2500 Leones. For each person that I ask you about, it is a new 2500 Leones. The money is yours; you can share some, share none, or share all. There is no obligation to share, it is only what you want to do. The receiver will not know who the money was sent by.

ID 1 2 3 4 5 Loan Officer Group Leader How much would you like to share with

Evaluation MF BRAC Sierra Leone 2014

Group Leader Survey

Consent of Respondent

(Interviewers: Clearly read out the following to the respondent and take her consent before taking interview)

I am …………………………………………….. from BRAC – an international NGO operating in microfinance services. We are currently conducting a survey in your community which broadly aims at generating information that will augment our services to people in this community. You have been selected as one of the respondents for this survey in this community. Please be aware that your participation is completely voluntary and you are free to withdraw at any time. However, because we want this survey to be as representative as possible of the views of people in this community, your participation is vital to enhance the success of this research. No financial token will be provided nor do we require any token from you for your participation. Do you agree to participate?

Yes No

Signature/Fingerprint ------(optional)

[Take the interview after having obtained the consent]

Branch Code

Group Code

Loan officer Code

MF member Code

Non MF member Code

DATA COLLECTION Enumerator Code Date _ _/_ _/_ _

Supervisor Code Date _ _/_ _/_ _

DATA ENTRY Enumerator 1 Code Date _ _/_ _/_ _

Enumerator 2 Code Date _ _/_ _/_ _

Supervisor Code Date _ _/_ _/_ _

Page 69 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

MODULE 1: SOCIO-DEMOGRAPHIC PROFILE

1.1 Household

1. Name 2. Age 3. Education Level 4. Tribe 5. Position in 6. Religion 7. No. of 8. No. of 9. No. of Community people in Adults Children HH 1. None 1. Mende Wife of the 1. Christian 2. Muslim 2. Some primary 2. Time chief school 3. Fulla 3. Other, specify 3.Completed primary 4. Gola Women’s school 5. Limba leader 4. Some JSS 6. Gbandi Youth Leader 5.Completed JSS 7. Kissi Union board 6. Some SS 8. Vai 7.Completed SS 9. Madingo 8. Some college 10. Other, 9.Completed college specify 10. Arabic education 11. Other, specify

1.2 Education of children

1. How many 2. How many of 3. If any do not attend 4. How many 5. How many of them 6. Of any do not attend 7. Who pays the children are primary them attend primary school, why not? children are attend secondary secondary school, why school fees? school aged? primary school? secondary school school? not?? aged? Amount Amount 1. No money for school fee Amount Amount 1. No money for school fee 1. We do ourselves 2. No money for transport 2. No money for transport 2. Family 3. No school 3. No school 3. Friends 4. Need children to work in 4. Need children to work in 4. Scholarship business business 5. Donor 5. Other, specify 5. Other, specify 6. Other, specify

Page 70 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

MODULE 2: HOUSING

2.1 Assets

Type of asset 1. Do you 2. How 3. If you 2.2 Housing improvements own this many do would buy a 1. type of 2. Costs of 3. Source of funding item? you own? similar item, improvement of improvement how much house would it cost? 1. Walls In Leones 1. Earnings/savings 2. Doors and windows 2. Relatives 1. Yes In Leones 3. Roof 3. Friends/neighbours 2. No 4. Kitchen 4. BRAC 1. Bed 5. Fences 5. Other NGO 2. Table 6. Tubewell 6. Government 3. Chair 7. Sanitary latrine 7. Other, specify 4. Light/lantern 8. Other, specify 5. Big pot 1. 6. Bucket 2. 7. Bible/Koran 3. 8. Homestead land

(town lot) 2.3 Housing conditions 9. Shop premises 10. Radio/cassette player Does your home have; 1. Yes, new; 2. Yes, old; 3. No 11. Television 12. Fan 1. Pan roof 13. Cellular phone 2. Mud and sticks walls 14. Generator 3. Cements walls 15. Bicycle 4. Earth floor 16. Motorbike 5. Wood floor 17. Refrigerator 6. Concrete floor

18. Sewing machine 7. Electricity 19. Computer 8. Private toilet 20. Other valuable assets

Page 71 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

2.4 Ownership

1. For how long have you been living in your house? Mention in years 2. What is the ownership status of your house? 1. Owned and built 5. Rented 2. Owned and purchased 6. Using without paying rent 3. Owned and inherited 7. Temporary dwelling 4. Given as a favour 8. Other, specify 3. If owned, what is the value of your house? In Leones 4. If rented, how much do pay per month? In Leones

MODULE 3: ECONOMIC AND INCOME RELATED ACTIVITIES

3.1 Occupation

1. What is your first 2. Daily income first 3. Weekly income first 4. What is your second 5. Daily income second 6. Weekly income occupation? occupation occupation occupation? occupation second occupation 1. Farming In Leones In Leones 1. Farming In Leones In Leones 2. Petty trading 2. Petty trading 3. Salary employee 3. Salary employee 4. Wage employee 4. Wage employee 5. Shop keeping 5. Shop keeping 6. Carpeting 6. Carpeting 7. Tailoring 7. Tailoring 8. Fishing 8. Fishing 9. Mining 9. Mining 10. Hair dressing 10. Hair dressing 11. Blacksmithing 11. Blacksmithing 12. Gardening 12. Gardening 13. Other, specify 13. Other, specify

3.2 Income past five years

7. Which of the past five years was BEST in terms of household income? 2009/2010/2011/2012/2013 8. Which of the past five years was WORST in terms of household income? 2009/2010/2011/2012/2013

Page 72 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

3.3 Expected income

9. What would your income be in the coming year if it were a GOOD year? In Leones 10. What would your income be in the coming year if it were a BAD year? In Leones 11. What do you expect your income to be this year (2014)? In Leones 12. What do you expect your income to be next year (2015)? In Leones

3.4 Business plans

3.4.1 Do you have any new business plan you want to start? 1. Yes; 2. No If NO, skip 3.2.2

3.4.2 New business plan

Activity 1. Farming 8. Fishing 2. Petty trading 9. Mining 1. 3. Salary employee 10. Hair dressing 4. Wage employee 11. Blacksmithing 2. 5. Shop keeping 12. Gardening 6. Carpeting 13. Other, specify 3. 7. Tailoring

3.5 Outstanding loans

3.5.1 Do you have an outstanding loan? 1. Yes; 2. No If NO, skip 3.3.3

3.5.2 Does anyone in your household have an outstanding loan? 1. Yes; 2. No If NO, skip 3.3.3

Page 73 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

3.5.3 Outstanding loans

1. Who 2. From 3. Amount of loan 4. Interest rate 1. Yourself 1. Formal bank 6. Relative In Leones In percentages 2. Husband 2. BRAC 7. Friend/neighbour 3. Other hh 3. Other MFI 8. Local money lender member 4. Landlord 9. Other, specify 5. Employer 1. 2. 3.

3.6 Savings

How much (cash) savings do you have?

1. Savings at home 2. Savings with 3. Bank 4. NGO 5. If NGO, which one someone you trust Amount in Leones

3.7 Involvement NGOs

Which other development programs (NGOs) do you know in your community?

Name of development program/NGO Activity Are you involved with this program/NGO? 1. Financial aid 5. Microfinance lending 1. Yes 2. Food aid 6. Human rights training 2. No 3. Medical aid 7. Women’s empowerment 4. Education aid 8. Other, specify 1. 2. 3. 4. 5. Page 74 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

MODULE 4: HEALTH

4.1 Health and health access

4.1.1 How do you perceive your current health status? 1. Good; 2. Average; 3. Bad 4.1.2 How do you consider your health compared to last year? 1. Better than last year; 2. Same as last year; 3. Worse than last year 4.1.3 How many death in household last year Amount 4.1.4 How many recent visits to doctor/hospital last year Amount 4.1.5 How long will it take you to get to a doctor? 1. Less than 15 min; 2. 15 - 30 min; 3. 30 - 1 hour; 4. 1 hour; 5. 2 hours or more 4.1.6 How long will it take you to get to a hospital? 1. Less than 15 min; 2. 15 - 30 min; 3. 30 - 1 hour; 4. 1 hour; 5. 2 hours or more

4.2 Diseases

For the following diseases, how many people in the household were sick in the last six weeks

1. Number of adults 2. Number of children 1. Malaria 2. Diarrhea 3. Blood in stool 4. Infection 5. Yellow fever 6. Typhoid 7. Cholera 8. Common cold 9. Other, specify

Page 75 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

MODULE 5: CONSUMPTION

5.1 Food consumption

1. Consumption 2. Did you 3. Did you buy this 4. How much did you 5. Did you use your stock 6. If you would buy a similar item, good consume good? spend on this good? for consumption of this how much would it cost? this good? good? 1. Yes 1. Yes In Leones 1. Yes In Leones 2. No 2. No  skip to 5. 2. No 1. Rice (white) 2. Rice (native) 3. Cassava 4. Yams 5. Groundnut 6. Beans 7. Plantain 8. Fruits 9. Fish 10. Chicken 11. Beef 12. Palm oil 13. Maggi 14. Other, specify

5.2 Fractions of revenue spend on different categories

How much is the fraction (%) of your income spend on In percentages 20. Consumption 21. Transport 22. Education (uniforms/fees) 23. Health (medicines/doctor) 24. Clothing 25. Building/improvement costs of house

Page 76 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

5.3 Non-food consumption

EDUCATION Amount in Leones CLOTHING Amount in Leones 1. School fees 9. Wearings 2. Books, pencils, etc. 10. Other, specify 3. School uniforms 4. Other, specify CEREMONIES Amount in Leones 11. Religious ceremonies HEALTH Amount in Leones 12. Memorial ceremonies 5. Purchase of medicines 13. Burials 6. Stay in hospital 14. Marriage 7. Herbalist 15. Secret society 8. Other, specify 16. Other, specify

5.4 Food security

5.3.1 How many meals do you eat a day Amount 5.3.2 Do you ever experience hunger? 1. Yes; 2. Sometimes; 3. No 5.3.3 Do you drink welled or bottled water? 1. Yes; 2. Sometimes; 3. No

MODULE 6: RISK

6.1 Would you like to receive A with certainty or B with a chance of 50 % of receiving it and a chance of 50% of receiving nothing?

A B A or B 5.000 Leones with certainty (100%) 4.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 3.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 2.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 1.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 0.500 Leones with uncertainty (50%)

Page 77 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

MODULE 7: CLIENT SATISFACTION (ONLY FOR MFI MEMBERS)

7.1 Client satisfaction 7.2 Depression index

Please rank the following Please rank the following

1 = Highly satisfied 1 = Rarely or none (less than 1 day a week) 2 = Moderately satisfied 2 = Sometimes (1 to 2 days a week) 3 = Moderately unsatisfied 3 = Occasionally (3 to 4 days a week) 4 = Highly unsatisfied 4 = Most of the time (5 to 7 days a week)

1. Loan application system 1. I was bothered by things that usually don’t bother me 2. Waiting time to receive loan (bothered = angry/irritated) 3. Maximum loan size 2. I did not feel like eating; my appetite was poor 4. Interest rate 3. I felt that I could not shake off the blues even with help 5. Loan application fee from my family or friends 6. Repayment period 4. I felt I was just as good as other people 7. Weekly instalment 5. I had trouble keeping my mind on what I was doing 8. System of paying overdue 6. I felt depressed 9. Increment in loan size 7. I felt that everything I did was an effort 10. Amount of meeting 8. I felt hopeful about the future 11. Duration of meeting 9. I thought my life had been a failure 12. Professional behaviour of LO 10. I felt fearful 13. LO’s reaction to suggestions 11. My sleep was restless 14. Accessibility of BRAC 12. I was happy 15. Distance to branch 13. I talked less than usual 16. Distance to group meeting 14. I felt lonely 17. In general, how satisfied are you 15. People were unfriendly with BRAC’s services 16. I enjoyed life 17. I had crying spells 18. I felt sad 19. I felt that people dislike me 20. I could not get going

Page 78 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

7.3 Aspirations

Have a look at the picture The first level (1) represents the best possible life for you, the lowest level (10) the worst 1. On what level are you now? Level 1-10 2. On what level do you want to be? Level 1-10 3. On what level do you think you are Level 1-10 in 5 years? In 5 years [question] will be 4. Your income A = much better B = better 5. Your health C = the same D = bad 6. Opportunities for you children E = much worse X = I do not know During the next five years I will: 7. Prove my education 1. Yes; 2. No 8. Find a better job 1. Yes; 2. No 9. Improve my education 1. Yes; 2. No 10. Marry and have children 1. Yes; 2. No

7.4 Locus of control 11. I strongly pursue my goals 1 = Definitely true 12. There are a lot of ways to solve a problem 2 = Somewhat true 13. I can think several ways to get things in life that are most important to me 3 = Slightly true 14. Even when others get discouraged, I know that I can find a way to resolve a problem 4 = Slightly false 15. My past experiences made me well prepared for my future 5 = Somewhat false 16. I had some success in life 6 = Definitely false 17. I attain the goals I set for myself

18. With which statement do you agree? A: everyone is primarily responsible for his success or failure in his life. B: success or failure in life is a question for each of the person. 19. With which statement do you agree? A: To be successful, above all should work very hard. B: To be successful, above all must have the luck

Page 79 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

MODULE 8: GROUP (ONLY FOR MFI MEMBERS)

8.1 Group information

1. When did you become a member of BRAC? Month – Year 2. When did you become a borrower of BRAC Month - Year 3. When did you become a member of the group? Month – Year 4. When was the group formed? Month – Year 5. Have you been a member of this group since it started? 1. Yes 2. No 6. Why did you join the group? 1. Needed loans 2. Peer pressure 3. No access to bank 4. No access to other sources of savings 5. Other, specify 7. What is the group formation based on? 1. Similar income 5. Friends 2. Similar occupation 6. Relatives 3. Neighbouring land 7. Other, specify 4. Neighbouring house 8. What is your position in the group 1. Big group leader 4. Normal member 2. Small group leader 5. Loan officer 3. Big and small group leader 9. If you have are [1], [2] or [3], what are your specific tasks? 1. Opening group meeting 5. Punish if group member 2. Collecting weekly did not paid weekly repayments repayments 3. Decide if new borrower is 6. Kick group members out allowed in group of program 4. Punish if group member is 7. Other, specify too late 10. If you are a big group leader, when did you become the big group leader? Month – Year 11. How were you selected? 1. Anonymous election by group 2. Non-anonymous election by group 3. Elected by loan officer 4. Elected by previous group leader Page 80 of 120

6. Other, specify 12. What are the most important qualifications that a group leader must 1. Wealthy have? 2. Position in community 3. Good speaker Multiple answers possible 4. Trustworthy 5. Oldest of the group 6. Other, specify 13. If a woman you didn’t know requested to join your group, and hasn’t been 1. Yes approved by BRAC, would you allow her? 2. No 14. If no, what’s the main reason? 1. There is a limit to group size 2. We don’t know her 3. It’s a hassle to get her approved by BRAC 4. Other, specify 15. Has anyone been refused membership? 1. Yes 2. No 16. If yes, what’s the main reason? 1. Did not need any more 4. We did not know her members 5. Other, specify 2. Not trustworthy 3. Not similar income 17. How many members have joined in the last five years? Amount 18. How many people who would like to be member of this group but cannot? Amount 19. How many people have left this group? Amount 20. Who takes the decisions in the meetings for the following activities? 1. Voting of majority of members a. ____ a) Group positions 2. Big group leader only b. ____ b) Loan to members 3. Small group leader only c. ____ c) Penalty for non-payment 4. Big and small group leaders d. ____ d) Activities of meeting 5. Loan officer of BRAC 6. Other, specify 21. Do you attend the weekly meetings? 1. Yes; 2. Often; 3. Sometimes; 4. No 22. How long do you have to walk from your living place to the weekly In minutes meeting place? 23. How long do you have to walk from your living place to the branch? In minutes 24. How many times (in terms of weeks) have you been unable to repay your weekly instalments in the last three months? 25. Has one of your group members ever been unable to repay the weekly 1. Yes repayment? 2. No Page 81 of 120

26. If Yes, did you ever paid for her weekly repayment? 1. Yes 2. No 27. In the history of this group, have group members ever repaid loans for 1. Yes other members when they were unable to? 2. No 28. How many times did this happen in your group? Amount 29. Can you remember which group members were unable to repay their ID loans? 30. How many weekly meetings have you missed in the last three months? 31. Is there a fee for a group member/leader for not attending the meetings? 1. Yes 2. No 32. If Yes, how much is the fee? In Leones 33. Is the group leader aware of the debt and savings that members have with 1. Yes BRAC 2. No 34. Are other members aware of the debt and savings that members have 1. Yes with BRAC? 2. No 35. Is the group leader aware of the debt and savings that members have with 1. Yes institutions other than BRAC? 2. No 36. Are other members aware of the debt and savings that members have 1. Yes with institutions other than BRAC? 2. No 37. Is the group leader aware of debts that members have to individuals? 1. Yes 2. No 38. Are other members aware of debts that members have to individuals? 1. Yes 2. No

8.2 Coordination game

If you had to guess what most people in the big group think, guess which answer most of the people in the big group would give.

If you say “I don’t know” no earnings can be made!

Who would most people in the group think of: Name of person ID of person 1. If a new group leader is needed, who would be the most likely to be chosen 2. Who is most likely to show up early in a meeting 3. Who is the best dressed of the group 4. Who is the best singer of the group Page 82 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

8.3 Network

ID 1. Are 2. Are 3. Have 4. How 6. Do 7. Do you go ID 1. Are 2. Are 3. Have 4. How 6. Do 7. Do you go you you you lent many you do to the same you you you lent many you do to the same friends blood money to times per business church/mosq friends blood money to times per business church/mosq with family this week do with this ue as this with family this week do with this ue as this this with this person? you meet person? person? this with this person? you meet person? person? person person? outside of person person? outside of ? the BRAC ? the BRAC meeting? meeting? 1. 21. 2. 22. 3. 23. 4. 24. 5. 25. 6. 26. 7. 27. 8. 28. 9. 29. 10. 30. 11. 31. 12. 32. 13. 33. 14. 34. 15. 35. 16. 36. 17. 37. 18. 38. 19. 39. 20. 40.

Page 83 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

8.4. Public good game

Say to the respondent: We are going to play a game. The first round of the game will be played with the members of your small group. I am going to give you and your group members 2000 Leones. Then we will ask every member individually how much of the 2000 Leones they want to put in a public pot. There is no obligation to put something in the public pot, it is only what you want to put in the public pot. The total amount of the public pot will then get tripled and equally divided over the members of the group, no matter how much someone has put in the pot. The amount you will receive at the end of the game is how much you kept, so the amount you did not put in the game, plus whatever you get returned from the public pot. How much do you want to put into the public pot?

ID Amount put in the pot Amount returned from the pot Total amount received

The second round of the game will be played with the members of the big group. You are getting the same amount of Leones and are asked again how much you want to put in the public pot. The amount you will receive at the end of the game is the amount you kept and the amount you get returned from the public pot. How much do you want to put into the public pot?

ID Amount put in the pot Amount returned from the pot Total amount received

8.5 Dictator game

Say to the respondent: I am giving you 2500 Leones. For each person that I ask you about, it is a new 2500 Leones. The money is yours; you can share some, share none, or share all. There is no obligation to share, it is only what you want to do. The receiver will not know who the money was sent by.

ID 1 2 3 4 5 6 7 8 9 10 11 How much share with ID 12 13 14 15 16 17 18 19 20 21 22 How much share with ID 23 24 25 26 27 28 29 30 Loan officer How much share with

Page 84 of 120

Evaluation MF BRAC Sierra Leone 2014

Loan Officer Survey

Consent of Respondent

(Interviewers: Clearly read out the following to the respondent and take her consent before taking interview)

I am …………………………………………….. from BRAC – an international NGO operating in microfinance services. We are currently conducting a survey in your community which broadly aims at generating information that will augment our services to people in this community. You have been selected as one of the respondents for this survey in this community. Please be aware that your participation is completely voluntary and you are free to withdraw at any time. However, because we want this survey to be as representative as possible of the views of people in this community, your participation is vital to enhance the success of this research. No financial token will be provided nor do we require any token from you for your participation. Do you agree to participate?

Yes No

Signature/Fingerprint ------(optional)

[Take the interview after having obtained the consent]

Branch Code

Group Code

Loan officer Code

DATA COLLECTION Enumerator Code Date _ _/_ _/_ _

Supervisor Code Date _ _/_ _/_ _

DATA ENTRY Enumerator 1 Code Date _ _/_ _/_ _

Enumerator 2 Code Date _ _/_ _/_ _

Supervisor Code Date _ _/_ _/_ _

Page 86 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ LOAN OFFICER CODE └──┴──┴──┘

MODULE 1: SOCIO-DEMOGRAPHIC PROFILE

1.1 Household

1. Name 2. Age 3. Education Level 4. Tribe 5. Position in 6. Religion 7. No. of 8. No. of 9. No. of Community people in Adults Children HH 1. None 1. Mende Wife of the 1. Christian 2. Muslim 2. Some primary 2. Time chief school 3. Fulla 3. Other, specify 3.Completed primary 4. Gola Women’s school 5. Limba leader 4. Some JSS 6. Gbandi Youth Leader 5.Completed JSS 7. Kissi Union board 6. Some SS 8. Vai 7.Completed SS 9. Madingo 8. Some college 10. Other, 9.Completed college specify 10. Arabic education 11. Other, specify

1.2 Education of children

1. How many 2. How many of 3. If any do not attend 4. How many 5. How many of them 6. Of any do not attend 7. Who pays the children are primary them attend primary school, why not? children are attend secondary secondary school, why school fees? school aged? primary school? secondary school school? not?? aged? Amount Amount 1. No money for school fee Amount Amount 1. No money for school fee 1. We do ourselves 2. No money for transport 2. No money for transport 2. Family 3. No school 3. No school 3. Friends 4. Need children to work in 4. Need children to work in 4. Scholarship business business 5. Donor 5. Other, specify 5. Other, specify 6. Other, specify

Page 87 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ LOAN OFFICER CODE └──┴──┴──┘

MODULE 3: ECONOMIC AND INCOME RELATED ACTIVITIES

3.2 Income past five years

1. Which of the past five years was BEST in terms of household income? 2009/2010/2011/2012/2013 2. Which of the past five years was WORST in terms of household income? 2009/2010/2011/2012/2013

3.3 Expected income

3. What would your income be in the coming year if it were a GOOD year? In Leones 4. What would your income be in the coming year if it were a BAD year? In Leones 5. What do you expect your income to be this year (2014)? In Leones 6. What do you expect your income to be next year (2015)? In Leones

MODULE 6: RISK

6.1 Would you like to receive A with certainty or B with a chance of 50 % of receiving it and a chance of 50% of receiving nothing?

A B A or B 5.000 Leones with certainty (100%) 4.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 3.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 2.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 1.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 0.500 Leones with uncertainty (50%)

Page 88 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ LOAN OFFICER CODE └──┴──┴──┘

MODULE 8: GROUP

8.1 General information

1. Since when have you been working for BRAC Month - Year 2. Since when have you been working for this branch? Month - Year 3. Have you been working at other branches before this branch? 1. Yes; 2. No 4. If yes, at which branches have worked before? 5. How long have you been working with this MF group? Month - Year 6. With how many other MF groups do you work? 7. How do you rate working with this group 1. No problems since the start 2. Some group members are not (Multiple answers possible) satisfied with BRAC 3. Some of group members do not attend the meeting 4. Some of group members do not pay their repayments 5. Majority of group members do not attend meeting and do not pay their repayment 8. What is the group formation based on? 1. Similar income 2. Similar occupation 3. Neighbouring land 4. Neighbouring house 5. Friends 6. Relatives 7. Other, specify 9. Has anyone been refused membership? 1. Yes; 2. No 10. If yes, what’s the main reason? 1. Did not need any more members 2. Not trustworthy 3. Not similar income 4. We did not know her 5. Other, specify 11. Who takes the decisions in the meetings for the following 1. Voting of majority of members a. ____ Page 89 of 120

activities? 2. Big group leader only b. ____ a) Group positions 3. Small group leader only c. ____ b) Loan to members 4. Big and small group leaders d. ____ c) Penalty for non-payment 5. Loan officer of BRAC d) Activities of meeting 6. Other, specify 12. What are your tasks in the group? 1. Opening group meeting 2. Collecting weekly repayments 3. Decide if new borrower is allowed in group 4. Punish if group member is too late 5. Punish if group member did not paid weekly repayments 6. Kick group members out of program 7. Other, specify

8.2 Coordination game

If you had to guess what most people in the big group think, guess which answer most of the people in the big group would give.

If you say “I don’t know” no earnings can be made!

Who would most people in the group think of: Name of person ID of person 1. If a new group leader is needed, who would be the most likely to be chosen 2. Who is most likely to show up early in a meeting 3. Who is the best dressed of the group 4. Who is the best singer of the group

Page 90 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

8.3 Network

ID 1. Are 2. Are 3. Have 4. How 6. Do 7. Do you go ID 1. Are 2. Are 3. Have 4. How 6. Do 7. Do you go you you you lent many you do to the same you you you lent many you do to the same friends blood money to times per business church/mosq friends blood money to times per business church/mosq with family this week do with this ue as this with family this week do with this ue as this this with this person? you meet person? person? this with this person? you meet person? person? person person? outside of person person? outside of ? the BRAC ? the BRAC meeting? meeting? 1. 21. 2. 22. 3. 23. 4. 24. 5. 25. 6. 26. 7. 27. 8. 28. 9. 29. 10. 30. 11. 31. 12. 32. 13. 33. 14. 34. 15. 35. 16. 36. 17. 37. 18. 38. 19. 39. 20. 40.

Page 91 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ MEMBER CODE └──┴──┴──┘

8.4. Public good game

Say to the respondent: We are going to play a game. The game will be played with the members of the big group. I am going to give you and your group members 2000 Leones. Then we will ask every member individually how much of the 2000 Leones they want to put in a public pot. There is no obligation to put something in the public pot, it is only what you want to put in the public pot. The total amount of the public pot will then get tripled and equally divided over the members of the group, no matter how much someone has put in the pot. The amount you will receive at the end of the game is how much you kept, so the amount you did not put in the game, plus whatever you get returned from the public pot. .How much do you want to put into the public pot?

ID Amount put in the pot Amount returned from the pot Total amount received

8.5 Dictator game

Say to the respondent: I am giving you 2500 Leones. For each person that I ask you about, it is a new 2500 Leones. The money is yours; you can share some, share none, or share all. There is no obligation to share, it is only what you want to do. The receiver will not know who the money was sent by. How much would you like to share with

ID 1 2 3 4 5 6 7 8 9 10 11 How much share with ID 12 13 14 15 16 17 18 19 20 21 22 How much share with ID 23 24 25 26 27 28 29 30 Group Leader How much share with

Page 92 of 120

Evaluation MF BRAC Sierra Leone 2014

Non Member Survey

Consent of Respondent

(Interviewers: Clearly read out the following to the respondent and take her consent before taking interview)

I am …………………………………………….. from BRAC – an international NGO operating in microfinance services. We are currently conducting a survey in your community which broadly aims at generating information that will augment our services to people in this community. You have been selected as one of the respondents for this survey in this community. Please be aware that your participation is completely voluntary and you are free to withdraw at any time. However, because we want this survey to be as representative as possible of the views of people in this community, your participation is vital to enhance the success of this research. No financial token will be provided nor do we require any token from you for your participation. Do you agree to participate?

Yes No

Signature/Fingerprint ------(optional)

[Take the interview after having obtained the consent]

Page 93 of 120

Branch Code

Non MF member Code

Inside boundary Code

Outside boundary Code

DATA COLLECTION Enumerator Code Date _ _/_ _/_ _

Supervisor Code Date _ _/_ _/_ _

DATA ENTRY Enumerator 1 Code Date _ _/_ _/_ _

Enumerator 2 Code Date _ _/_ _/_ _

Supervisor Code Date _ _/_ _/_ _

Page 94 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ NON-MEMBER CODE └──┴──┴──┘

MODULE 1: SOCIO-DEMOGRAPHIC PROFILE

1.1 Household

1. Name 2. Age 3. Education Level 4. Tribe 5. Position in 6. Religion 7. No. of 8. No. of 9. No. of Community people in Adults Children HH 1. None 1. Mende Wife of the 1. Christian 2. Muslim 2. Some primary 2. Time chief school 3. Fulla 3. Other, specify 3.Completed primary 4. Gola Women’s school 5. Limba leader 4. Some JSS 6. Gbandi Youth Leader 5.Completed JSS 7. Kissi Union board 6. Some SS 8. Vai 7.Completed SS 9. Madingo 8. Some college 10. Other, 9.Completed college specify 10. Arabic education 11. Other, specify

1.2 Education of children

1. How many 2. How many of 3. If any do not attend 4. How many 5. How many of them 6. Of any do not attend 7. Who pays the children are primary them attend primary school, why not? children are attend secondary secondary school, why school fees? school aged? primary school? secondary school school? not?? aged? Amount Amount 1. No money for school fee Amount Amount 1. No money for school fee 1. We do ourselves 2. No money for transport 2. No money for transport 2. Family 3. No school 3. No school 3. Friends 4. Need children to work in 4. Need children to work in 4. Scholarship business business 5. Donor 5. Other, specify 5. Other, specify 6. Other, specify

Page 95 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ NON-MEMBER CODE └──┴──┴──┘

MODULE 2: HOUSING

2.1 Assets

Type of asset 1. Do you 2. How 3. If you 2.2 Housing improvements own this many do would buy a 1. type of 2. Costs of 3. Source of funding item? you own? similar item, improvement of improvement how much house would it cost? 1. Walls In Leones 1. Earnings/savings 2. Doors and windows 2. Relatives 1. Yes In Leones 3. Roof 3. Friends/neighbours 2. No 4. Kitchen 4. BRAC 1. Bed 5. Fences 5. Other NGO 2. Table 6. Tubewell 6. Government 3. Chair 7. Sanitary latrine 7. Other, specify 4. Light/lantern 8. Other, specify 5. Big pot 1. 6. Bucket 2. 7. Bible/Koran 3. 8. Homestead land

(town lot) 2.3 Housing conditions 9. Shop premises 10. Radio/cassette player Does your home have; 1. Yes, new; 2. Yes, old; 3. No 11. Television 12. Fan 1. Pan roof 13. Cellular phone 2. Mud and sticks walls 14. Generator 3. Cements walls 15. Bicycle 4. Earth floor 16. Motorbike 5. Wood floor 17. Refrigerator 6. Concrete floor

18. Sewing machine 7. Electricity 19. Computer 8. Private toilet 20. Other valuable assets

Page 96 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ NON-MEMBER CODE └──┴──┴──┘

2.4 Ownership

1. For how long have you been living in your house? Mention in years 2. What is the ownership status of your house? 1. Owned and built 5. Rented 2. Owned and purchased 6. Using without paying rent 3. Owned and inherited 7. Temporary dwelling 4. Given as a favour 8. Other, specify 3. If owned, what is the value of your house? In Leones 4. If rented, how much do pay per month? In Leones

MODULE 3: ECONOMIC AND INCOME RELATED ACTIVITIES

3.1 Occupation

1. What is your first 2. Daily income first 3. Weekly income first 4. What is your second 5. Daily income second 6. Weekly income occupation? occupation occupation occupation? occupation second occupation 1. Farming In Leones In Leones 1. Farming In Leones In Leones 2. Petty trading 2. Petty trading 3. Salary employee 3. Salary employee 4. Wage employee 4. Wage employee 5. Shop keeping 5. Shop keeping 6. Carpeting 6. Carpeting 7. Tailoring 7. Tailoring 8. Fishing 8. Fishing 9. Mining 9. Mining 10. Hair dressing 10. Hair dressing 11. Blacksmithing 11. Blacksmithing 12. Gardening 12. Gardening 13. Other, specify 13. Other, specify

3.2 Income past five years

7. Which of the past five years was BEST in terms of household income? 2009/2010/2011/2012/2013 8. Which of the past five years was WORST in terms of household income? 2009/2010/2011/2012/2013

Page 97 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ NON-MEMBER CODE └──┴──┴──┘

3.3 Expected income

9. What would your income be in the coming year if it were a GOOD year? In Leones 10. What would your income be in the coming year if it were a BAD year? In Leones 11. What do you expect your income to be this year (2014)? In Leones 12. What do you expect your income to be next year (2015)? In Leones

3.4 Business plans

3.4.1 Do you have any new business plan you want to start? 1. Yes; 2. No If NO, skip 3.2.2

3.4.2 New business plan

Activity 1. Farming 8. Fishing 2. Petty trading 9. Mining 1. 3. Salary employee 10. Hair dressing 4. Wage employee 11. Blacksmithing 2. 5. Shop keeping 12. Gardening 6. Carpeting 13. Other, specify 3. 7. Tailoring

3.5 Outstanding loans

3.5.1 Do you have an outstanding loan? 1. Yes; 2. No If NO, skip 3.3.3

3.5.2 Does anyone in your household have an outstanding loan? 1. Yes; 2. No If NO, skip 3.3.3

Page 98 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ NON-MEMBER CODE └──┴──┴──┘

3.5.3 Outstanding loans

1. Who 2. From 3. Amount of loan 4. Interest rate 1. Yourself 1. Formal bank 6. Relative In Leones In percentages 2. Husband 2. BRAC 7. Friend/neighbour 3. Other hh 3. Other MFI 8. Local money lender member 4. Landlord 9. Other, specify 5. Employer 1. 2. 3.

3.6 Savings

How much (cash) savings do you have?

1. Savings at home 2. Savings with 3. Bank 4. NGO 5. If NGO, which one someone you trust Amount in Leones

3.7 Involvement NGOs

Which other development programs (NGOs) do you know in your community?

Name of development program/NGO Activity Are you involved with this program/NGO? 1. Financial aid 5. Microfinance lending 1. Yes 2. Food aid 6. Human rights training 2. No 3. Medical aid 7. Women’s empowerment 4. Education aid 8. Other, specify 1. 2. 3. 4. 5. Page 99 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ NON-MEMBER CODE └──┴──┴──┘

MODULE 4: HEALTH

4.1 Health and health access

4.1.1 How do you perceive your current health status? 1. Good; 2. Average; 3. Bad 4.1.2 How do you consider your health compared to last year? 1. Better than last year; 2. Same as last year; 3. Worse than last year 4.1.3 How many death in household last year Amount 4.1.4 How many recent visits to doctor/hospital last year Amount 4.1.5 How long will it take you to get to a doctor? 1. Less than 15 min; 2. 15 - 30 min; 3. 30 - 1 hour; 4. 1 hour; 5. 2 hours or more 4.1.6 How long will it take you to get to a hospital? 1. Less than 15 min; 2. 15 - 30 min; 3. 30 - 1 hour; 4. 1 hour; 5. 2 hours or more

4.2 Diseases

For the following diseases, how many people in the household were sick in the last six weeks

1. Number of adults 2. Number of children 1. Malaria 2. Diarrhea 3. Blood in stool 4. Infection 5. Yellow fever 6. Typhoid 7. Cholera 8. Common cold 9. Other, specify

Page 100 of 120

BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ NON-MEMBER CODE └──┴──┴──┘

MODULE 5: CONSUMPTION

5.1 Food consumption

1. Consumption 2. Did you 3. Did you buy this 4. How much did you 5. Did you use your stock 6. If you would buy a similar item, good consume good? spend on this good? for consumption of this how much would it cost? this good? good? 1. Yes 1. Yes In Leones 1. Yes In Leones 2. No 2. No  skip to 5. 2. No 1. Rice (white) 2. Rice (native) 3. Cassava 4. Yams 5. Groundnut 6. Beans 7. Plantain 8. Fruits 9. Fish 10. Chicken 11. Beef 12. Palm oil 13. Maggi 14. Other, specify

5.2 Fractions of revenue spend on different categories

How much is the fraction (%) of your income spend on In percentages 26. Consumption 27. Transport 28. Education (uniforms/fees) 29. Health (medicines/doctor) 30. Clothing 31. Building/improvement costs of house BRANCH CODE └──┴──┴──┘ GROUP CODE └──┴──┴──┘ NON-MEMBER CODE └──┴──┴──┘ Page 101 of 120

5.3 Non-food consumption

EDUCATION Amount in Leones CLOTHING Amount in Leones 1. School fees 9. Wearings 2. Books, pencils, etc. 10. Other, specify 3. School uniforms 4. Other, specify CEREMONIES Amount in Leones 11. Religious ceremonies HEALTH Amount in Leones 12. Memorial ceremonies 5. Purchase of medicines 13. Burials 6. Stay in hospital 14. Marriage 7. Herbalist 15. Secret society 8. Other, specify 16. Other, specify

5.3 Food security

5.3.1 How many meals do you eat a day Amount 5.3.2 Do you ever experience hunger? 1. Yes; 2. Sometimes; 3. No 5.3.3 Do you drink welled or bottled water? 1. Yes; 2. Sometimes; 3. No

MODULE 6: RISK

6.1 Would you like to receive A with certainty or B with a chance of 50 % of receiving it and a chance of 50% of receiving nothing?

A B A or B 5.000 Leones with certainty (100%) 4.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 3.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 2.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 1.500 Leones with uncertainty (50%) 5.000 Leones with certainty (100%) 0.500 Leones with uncertainty (50%)

Page 102 of 120

Control Group Census Form: Inside boundary

Enumerator Code RA TEAM Date _ _ / _ _ / _ _

1. Full name Write full name 2. Age Amount 3. AAdress Street, number, town/village 4. Telephone number 1 Number 5. Telephone number 2 Number 6. Interested in BRAC loan? 1. Yes; 2. No 7. Type of business 1. Farming 8. Fishing 2. Petty trading 9. Mining 3. Salary employee 10. Hair dressing 4. Wage employee 11. 5. Shop keeping Blacksmithing 6. Carpentry 12. Gardening 7. Tailoring 13. Other, specify 8. Willing to conduct survey 1. Yes; 2. No 9. Have time on Friday the 28th of March 1. Yes; 2. No  go to Q10 10. If no, when does the respondent have 1. ______time? 1. Day 2. Time 2. ______

Control Group Census Form: Outside boundary

Enumerator Code RA TEAM Date _ _ / _ _ / _ _

11. Full name Write full name 12. Age Amount 13. AAdress Street, number, town/village 14. Telephone number 1 Number 15. Telephone number 2 Number 16. Interested in BRAC loan? 1. Yes; 2. No 17. Type of business 1. Farming 8. Fishing 2. Petty trading 9. Mining 3. Salary employee 10. Hair dressing 4. Wage employee 11. 5. Shop keeping Blacksmithing 6. Carpentry 12. Gardening 7. Tailoring 13. Other, specify 18. Willing to conduct survey 1. Yes; 2. No 19. Have time on Thursday the 27th of 1. Yes;

March 2. No  go to Q10 20. If no, when does the respondent have 1. ______time? 1. Day 2. Time 2. ______

Page 103 of 120

Branch Code Group Payments Sheet Group Code

Enumerator 1 Code 1 Enumerator 2 Code 2 Team Code Date (DD MM YYYY) Silence Participant Activity #3 Participant Name Activity #2 Activity #4 Token Total Signature/RTP ID (Big group) (1.000)

001 =

002 =

003

004

005

006

007

008

009

010

011

012

013

014

015

016

017

018

019

020

Page 104 of 120

Branch Code Group Payments Sheet Group Code Enumerator 1 Code 1 Enumerator 2 Code 2 Team Code Date (DD MM YYYY) Participant Silence Participant Name Activity #2 Activity #3 Activity #4 Token Total Signature/RTP ID (1.000)

021 =

022 =

023

024

025

026

027

028

029

020

031

032

033

034

035

036 037 038

039

040 Page 105 of 120

Coordination Game Record Sheet Branch Code Group Code Loan Officer Code Enumerator Code RA Team code Date _ _/_ _/_ _

Question 1: New Group Leader Name ID Tally

Question 2: Show up early in meeting Name ID Tally

Page 106 of 120

Question 3: Best Dressed Name ID Tally

Question 4: Best Singer Name ID Tally

Page 107 of 120

Appendix E: Protocols

Page 108 of 120

Evaluation MF SL BRAC – Protocols Wageningen University/BRAC Sierra Leone

Version: 08-03-2014

Contents Overview of activities ...... 110 Introduction ...... 110 Phase 1: Group composition study ...... 110 Phase 2: Evaluation study ...... 111 Sample ...... 111 Phase 1: Group composition study ...... 112 Pre-visit activities ...... 112 Schedule of activities during visit ...... 112 Group Meeting ...... 112 General comments ...... 113 Phase 2: Evaluation study ...... 114 Pre-visit activities ...... 114 Schedule of activities during visit ...... 114 Meeting with participants ...... 114 General comments ...... 115

Page 109 of 120

Overview of activities

Introduction In this document we outline the main activities to be undertaken in Sierra Leone starting from January 2014. The objective of this survey is to assess (i) the determinants of clients, (ii) whether clients improve their wellbeing, and are less vulnerable and (iii) whether clients are satisfied with the services provided by BRAC and (iv) to understand the processes underlying the composition of the groups and the influence of group composition on financial behaviour. This document includes the instrument used when conducting the survey.

Phase 1: Group composition study The group composition study consists of several parts (PART 1 and PART 2 OF SURVEY);

1. Group member (& group leader) survey This survey will include questions on the bigger group level, such as how long a group exists, composition of group, questions on the smaller group level, also how long a group exists, composition of group, questions about individual borrowers and their loan outcomes. But also questions about the characteristics of a group leader, how long a group leader is positioned, what her tasks are. 2. Loan officer survey The survey will include questions about the characteristics of the loan officer, what her tasks are, how long she is positioned. 3. Coordination game The coordination game is a measure of social closeness and coordination within a group. This game is meant to understand how much the group members know the other group members. The enumerator explains that when answering this question person A has to think about what most people in the group would answer. So for example; “Which group member dresses the best?” Then person A has to give the name of the person she thinks most people would answer. When she guesses right, she can win 500 Leones, when she guesses wrong, she wins nothing. This game will be played in the big group, and later when analysing the data we will check for any outcomes on the small group level. 4. Network analysis To understand the possible intertwined relationships in groups, we do a network analysis where we ask the borrowers for example who of the other borrowers are family, or are friends. We conduct this network analysis with group members, group leaders and the loan officers. 5. Public good game A public good game gives us an indication of ‘free-riding’ in a group. We will expect that the higher the cohesion in a group (which is measured with the coordination game and the network survey), the less people will free ride. We will play this game on the small group level and on the big group level. The game will go as follows: Everyone gets 1000 Leones, we ask them to give a certain amount (can be all, can be nothing) to a public good (a pot which gets divided over the group members), this amount gets tripled and then goes to the public good. This public good (the

Page 110 of 120

pot) then gets divided over the amount of people playing the game. This amount can maximum be 3500 Leones in the small group and 4000 Leones in the big group. Everyone then receives: their share of the pot + amount of Leones they kept of the small group. 6. Dictator game A dictator game is designed to capture benevolence of big group leader/small group leader/loan officer to the group members. Therefore we will play the benevolence game with all 5 small groups with big leader and loan officer. The idea is to ask every small group member how much of 2500 they want to give to their other group members, to the big group leader and to the loan officer. We will ask this for every member of every small group. Then we will ask the big group leader how much they want to give of 2500 Leones to every other group member (so 29 people) and the loan officer. Then we will ask the loan officer how much they want to give to all group members (30 people) and the big group leader. Everything is anonymous: no one knows who the person is who is giving them a certain amount.

Phase 2: Evaluation study An individual survey will be conducted in eight branches and will be divided in four groups per branch. The survey includes a range of questions capturing household identification, socio- demographics, income, housing, health, food security, consumption, client satisfaction and risk behaviour (PART 1 OF SURVEY).

This is the same survey used in Phase 1.

Sample Phase 1 will only be conducted in Mile-91, and will have a total of ± 700 respondents.

 600 group members (old and new members)  50 non-members inside the boundary of Mile-91  50 non-members outside the boundary of Mile-91

Phase 2 will be conducted in 8 branches and will have 1600 respondents in total. Per branch we will have 200 respondents:

 50 old group members  50 new group members  50 non-members inside the boundary of the branch  50 non-members outside the boundary of the branch

Page 111 of 120

Phase 1: Group composition study

Pre-visit activities Before the research activities can start, go to the branch and announce:

 We would like to conduct a survey in your branch. Hand the Branch Manager the Letter of Participation- Branch.  Several groups are randomly selected. Ask the Branch Manager and Loan Officers to check whether the Participants List is complete, or whether people are now out of the program, or maybe new people have entered the program.  Tell the Loan Officer that it is very important that everyone on the Participants List is present on the day of the visit.  Have the Loan Officer read the Letter of Participation - Groups to the selected groups one week before the research team visits.

Schedule of activities during visit After arrival in the branch, the team leader will present his team to the branch manager. After the acquaintance, start the schedule.

1. Group meeting 2. Research Activities a. Loan Officer survey (done by the team leader) b. Group Leader survey (done by the team leader) c. Group Member survey (all research assistants and the team leader when he/she is done with the Loan Officer/Group Leader survey) d. Coordination Game e. Network Analysis f. Public Goods Game g. Dictator Game 3. Calculation of earnings 4. Handing out of pay-slips. After the schedule has been completed, the night can be spent in Mile- 91.

Group Meeting 1. Convene a meeting and announce:  Introduce yourself; first as representative of Wageningen University, BRAC Sierra Leone and Njala University, second as yourself, hence your name. Tell them you are Wageningen University researchers, working for BRAC Sierra Leone.  We would like to conduct some survey activities with all group members, the group leader and the loan officer. BRAC Sierra Leone will not know the names, all is anonymous.  Tell them you would like to ask them questions and engage in some activities in which they can win some money based on their choices in the activities.  This money will paid out, after we visited all the groups which are scheduled on that day, by our research supervisor.  We would like to talk to you each individually. It may take us some time to talk to all of you. We ask you to please be patient.  Since we can’t talk to everyone at once, we ask that you wait before you have taken your turn.

Page 112 of 120

 After you have had your turn, you are welcome to go do other work until all participants have been surveyed. Then we will calculate the amounts earned by all participants. After this is calculated you will receive your pay-slip, which you can cash in with our research supervisor that same day.  While you wait for your turn in this area, we ask you not to talk to those who have already talked to us. If we overhear you discussing the survey questions or games then we will take away your silence token and you will not earn the additional Le 1,000.

After announcements;  Welcome all participants again  Make sure the Participants List is complete  Give all participants their ID numbers (Tape+markers)  Divide up the participants on the Participants List among all the research assistants  Each research assistant should find a private place and call their first participant to interview (for the team leader this will be the Loan Officer and the Group Leader)  For the Loan Officer: ask the participant the survey questions on the Loan Officer Survey and Record Sheets. See protocol for survey and activity #1 - #4 below.  For the Group Leader: ask the participant the survey questions on the Group Leader Survey and Record Sheets. See protocol for survey and activity #1 - #4 below.  For the Group Member: ask the participant the survey questions on the Group Member Survey and Record Sheets. See protocol for survey and activity #1 - #4 below.  Conduct the survey and activities and use the form to record their answer and choices. Write neat and clear.  At the end, tell the participant: o Thank you for your time today o We will add up everything you won and write a pay-slip for you. Keep this pay-slip at a safe place. After all participants have been surveyed, our research supervisor will come to make the pay-out. o Please do not talk to those who have not yet had their interview. If you do, and we catch you, you will not get your 1.000 Leones silence token. o Do you have any questions?  Make sure the participants Survey and Record Sheets are completely filled out.  Put it in a safe place where all of the forms are stored during the day.  Call your next participant and repeat the interview procedures.

General comments 1. Make sure that you divide the tasks clearly among the different members of the group, so everybody knows where he or she is responsible for. Not only the bigger tasks (paperwork, material, money), but also tasks as who is responsible for the final payment, for writing the receipts etcetera. But if for example the material/stationery person is not around/still busy, make sure that you take care of the stationery for him/her. 2. Think ahead: when you don’t have anything to do, already think about what you can do next. For example, writing names and villages on sheets.

Page 113 of 120

Phase 2: Evaluation study

Pre-visit activities Before the research activities can start, go to the branch and announce:

 We would like to conduct a survey in your branch. Hand the Branch Manager the Letter of Participation - Branch.  Several groups are randomly selected. Ask the Branch Manager and Loan Officers to check whether the Participants List is complete, or whether people are now out of the program, or maybe new people have entered the program.  Tell the Loan Officer that it is very important that everyone on the Participants List is present on the day of the visit.  Have the Loan Officer read the Letter of Participation – Members to the selected members one week before the research team visits. To have them gather the next week at the stated time and place.

Schedule of activities during visit After arrival in the branch, the team leader will present his team to the branch manager. After the acquaintance, start the schedule.

1. Meeting with participant 2. Research Activities a. Group Member survey b. Non-member survey inside boundary c. Non-member survey outside boundary 3. After the schedule has been completed, the night can be spent in the city or village where the branch is located.

Meeting with participants 2. Convene a meeting and announce:  Introduce yourself; first as representative of Wageningen University, BRAC Sierra Leone and Njala University, second as yourself, hence your name. Tell them you are Wageningen University researchers, working for BRAC Sierra Leone.  We would like to conduct some survey activities with all participants, i.e. you would like to ask them some questions. BRAC Sierra Leone will not know the names, all is anonymous.  We would like to talk to you each individually. It may take us some time to talk to all of you. We ask you to please be patient.  Since we can’t talk to everyone at once, we ask that you wait before you have taken your turn.  After you have had your turn, you are welcome to go.  While you wait for your turn in this area, we ask you not to talk to those who have already talked to us.

After announcements;  Welcome all participants again  Make sure the Participants List is complete  Give all participants their ID numbers (Tape+markers)  Divide up the participants on the Participants List among all the research assistants  Each research assistant should find a private place and call their first participant to interview (for the team leader this will be the Loan Officer and the Group Leader)

Page 114 of 120

 For the Group Member: ask the participant the survey questions on the Group Member Survey. See protocol for survey below.  For the Non-Member: ask the participant the survey questions on the Non-Member Survey. See protocol for survey below.  Conduct the survey and use the form to record their answer and choices. Write neat and clear.  At the end, tell the participant: o Thank you for your time today o Please do not talk to those who have not yet had their interview o Do you have any questions?  Make sure the participants Survey are completely filled out.  Put it in a safe place where all of the forms are stored during the day.  Call your next participant and repeat the interview procedures.

General comments 3. Make sure that you divide the tasks clearly among the different members of the group, so everybody knows where he or she is responsible for. Not only the bigger tasks (paperwork, material, money), but also tasks as who is responsible for the final payment, for writing the receipts etcetera. But if for example the material/stationery person is not around/still busy, make sure that you take care of the stationery for him/her. 4. Think ahead: when you don’t have anything to do, already think about what you can do next. For example, writing names and villages on sheets.

Page 115 of 120

Respondent Survey PART 1 1. Tell the participant:  My name is______. We are here to collect information about BRAC’s microfinance services and their clients, for a study of BRAC Sierra Leone, Njala University and a university in the Netherlands. Your branch was selected to be part of this survey.  The researchers will keep your responses confidential. Your full name will never be used anywhere to ensure confidentiality.  You are not obliged to answer questions if you do not want to and you are free to stop the interview at all times.  We hope that the research will benefit Sierra Leone by assisting us to understand BRAC’s microfinance services and maybe even improve these in the future.  Do you have any questions for me? You may ask questions about this study at any time  The survey will take approximately half an hour. 2. Hand over the informed consent sheet and review the Informed Consent form. Emphasize that:  No one has to participate.  There will be no negative consequences to anyone if you do not want to participate.  Everything you do with us will be confidential. We will not tell anyone in your group what you tell us. When we record your information, we will not include your name with your file so that anyone can come back and say anything to you about what you tell us. Everything we record today will be kept in a safe location in Europe. 3. Ask each question and record their answers in the appropriate boxes. 4. If you are part of the PHASE I study, you continue with the protocol 5. If you are part of the PHASE II study, you stop here.  Tell the participant: o Thank you for your great patience and cooperation. We have enjoyed working with you. o Look up your next participant and start all over again.

Respondent Survey PART 2 1. Tell the participant that the first part of the survey has been completed, and that we are now starting with the second part. 2. This part includes questions and activities where a small amount of Leones can be earned, based on their decisions in the activity. 3. Tell the participant that this will be explained later on in the survey. 4. Ask each question and perform each activity and record their answers in the appropriate boxes.

Network Analysis 1. Tell the participant:  We will be asking everyone that we survey the same questions.  We will say the name of your group member and then ask 6 questions regarding the relationship of this group member with you.  Give the name of the first ID.  Then ask the participant: o Are you friends with this person? o Are you blood family with this person? o Have you lent money to this person? o How many times per week do you meet outside of the BRAC meeting?

Page 116 of 120

o Do you do business with this person? o Do you go to the same church/mosque as this person?  Repeat for all IDs (including group leader and loan officer)

Activity #1: Coordination Game 1. Tell the participant:  We will be asking everyone that we survey these same questions.  When thinking of the answer you want to give, try to guess the name of the person you think most people in your big group will say for each of these questions.  You will win 500 Le for every time you give the same answer that is most often given.  For example, if we asked you the question, “What is the best Sierra Leonean dish?” and 10 people said groundnut soup and 5 people said potato leaves, then if you were one of those who said groundnut soup you would win 500 Le but if you were one of those that said potato leaves you would not win anything.  You may answer “No one” or “None”, but those types of answers cannot win, so you should do your best to pick a specific person’s name (e.g. slightly agree or neither agree or disagree). 1. Make sure that they give you the right name and spelling. Names are often quite similar. Also fill the person’s ID in the sheet. 2. For questions 1 through 4, read the below prompt to the participant:  If you had to guess what most people in your big group think, who do you think they would say is: o If a new group leader is needed, who would be the most likely to be chosen o Who is most likely to show up early in a meeting o Who is the best dressed of the group o Who is the best singer of the group 3. Make sure that you do NOT ask what the interviewee thinks is the best cook in the big group, but what MOST OF THE GROUP MEMBERS think is the best cook in the big group. 4. Record the full name of their response or the appropriate code for their response.

Activity #2: Public Good Game 1. Tell the participant:  We are going to play a game, we will play this game with all the participants.  There are two rounds in this game; the first round is with the members of your small group. The second round is with the members of the big group. 2. Tell the participant: We will begin with the first round;  I am going to give you and your group members 2.000 Le. Then we will ask every member individually how much of the 2.000 Le they want to put in a public pot.  There is no obligation to put something in the public pot, it is only what you want to put in the public pot.  The total amount of the public pot will then get tripled and equally divided over the members of the group, no matter how much someone has put in the pot.  The amount you will receive at the end of the game is how much you kept, so the amount you did not put in the public pot, plus whatever you get returned from the public pot. 3. Ask the participant: How much do you want to put in the public pot?

Page 117 of 120

4. Tell the participant: Now we will play the second round of the game.  This time we will play the game with the members of the big group.  You are getting the same amount of Leones; 2.000 Le.  And you are asked again how much you want to put in the public pot.  The amount you will receive at the end of the game is the amount you kept and the amount you get returned from the public pot. 5. Ask the participant: How much do you want to put in the public pot?

Activity #3: Dictator Game 1. Tell the partipant:  I am giving you 2.500 Le  For each person that I will ask you about, it is a new 2.500 Le  The money is yours; you can share some, share none, or share it all  There is no obligation to share; it is only what you want to share  The receiver will not know who the money was sent by

Pay-slip payment 1. When everyone has finished the survey and activities, tell the participants:  Thank you for your great patience and cooperation. We have enjoyed working with you. 2. Calculate the winning answers for the coordination game on the Coordination Game Record Sheet and mark which questions people got correct on their Loan Officer Survey and Record Sheets, Group Leader Survey and Record Sheets and Group Member Survey and Record Sheets. 3. Use the Loan Officer Survey and Record Sheets, Group Leader Survey and Record Sheets and Group Member Survey and Record Sheets to calculate the total amount of money owed to each participant in each of the games. 4. Record these amounts on the Group Payment Sheet. (This should be done by the same person always!) 5. Record on the Group Payment Sheet how much the two public pots in Activity #3 is, triple the amount of the public pot of the big group and divide it over the amount of people in the group. Write the amount on the Group Payment Sheet. Remember: the first public pot is played with the small group and the second public pot is played with the big group. (The Loan Officer only plays Activity #3 with the big group). 6. Record on the Group Payment Sheet how much each participant is owed based on his or her matched participant in Activity #4. Remember: For all households you just add up one to come to the matched participants. So participant #2 shares with participant #3, participant #3 shared with participant #4, etc. The last participant shares with participant #1. And the group leader shares with the loan officer, and the loan officer with the group leader. 7. Add a bonus of 1.000 Le on the Group Payment Sheet if the participant kept their silence token. 8. Calculate the total amount owed to the participant. 9. Wait for Karen to come to the group to pay out the participants.

Page 118 of 120

10. Ask the first person in the line to come into the private place to get their payment 11. Have them give a thumbprint in the ‘received’ column, certifying they received their payment. 12. Send the participant out of the private room, and repeat this process with all remaining participants. 13. Only in very special occasions: If a participant cannot wait for Karen, make a pay-slip, with the date and location and A STAMP.

Page 119 of 120

Protocol - Control Group Census In this document we outline the activities which have to be undertaken in Sierra Leone in March 2014. The objective of this document is to find a control group which can be surveyed in the BRAC MF SL Evaluation study. Due to the fact that BRAC uses boundaries, which determine who can be a member and who cannot be a member, we will also make use of these boundaries to find our two control groups.

Sample The control group will consist of two different groups;

1. 50 women inside the boundary of the branch 2. 50 women outside the boundary of the branch

Control Group Census 1. Go to the assigned area. 2. Walk around the area and look for women who are between 18 and 50 years old. 3. When you see a woman who looks between 18 and 50 years old, approach her. 4. Ask her: a. Can I ask you something? b. If no  thank her and look for your next female participant c. If yes  ask her; i. Are you between 18 and 50 years old? 1. If no  thank her and look for your next female participant 2. If yes  continue ii. Do you have a loan with BRAC? 1. If no  thank her and look for your next female participant 2. If yes  continue iii. Do you have a (small) business? 1. If no  thank her and look for your next female participant 2. If yes  continue d. If she is between 18 and 50, does not have a loan with BRAC and owns a business, tell her; i. I am a representative of Wageningen University and Njala University working for BRAC Sierra Leone. BRAC is an international microfinance organisation which provides small loans for women with a business. The process goes as follows: women form small groups with other women, and they are jointly liable for each other. This means; when someone cannot repay their loan, the other will pay for her. Of course, the women who couldn’t repay pay the other woman back. These groups form, together with other smaller groups, a bigger group which meet every week for a half an hour. Women can lend up to 500.000 Leones for their first loan, which they have to repay within either 10 of 20 weeks. e. Tell them; i. We have two questions for you 1. Would you be interested in a loan from BRAC? a. If no  thank her and look for your next female participant b. If yes  continue 2. Would you have time to help us answering some questions? This will take a half an hour. We are not able to provide the actual loan for you, but we are interested in you and your household. a. For outside the boundary: this will be on Thursday b. For inside the boundary: this will be on Friday 3. If yes; a. Write down her information on the Control Group Census Form and thank her for her time. 5. Look for the next female participant and go through the process all over again.

Page 120 of 120