impact and programme evaluation of plan and uhiki’s joint vsl … · 2017. 9. 25. · vsl member,...

80
1 Plan Tanzania Impact and Programme Evaluation of Plan and UHIKI’s joint VSL Programme in Tanzania Hugh Allen Solingen, Germany February 2009

Upload: others

Post on 17-Feb-2021

0 views

Category:

Documents


0 download

TRANSCRIPT

  • 1

    Plan Tanzania

    Impact and Programme Evaluation of Plan and UHIKI’s joint VSL Programme in Tanzania

    Hugh Allen Solingen, Germany February 2009

  • 2

    Table of Contents List of Acronyms ................................................................................................... 3 1 Executive Summary ............................................................................................. 4 2 Terms of reference ................................................................................................ 8 3 Evaluation design and approach ........................................................................ 10 4 Impact evaluation ............................................................................................... 14 5 Programme evaluation ....................................................................................... 34 6 Programme implementation ............................................................................... 47 7 Programme design ............................................................................................. 56 8 Annexes

    Annex 1: Full Terms of reference ................................................................... 58 Annex 2: Schedule of visits ............................................................................ 60 Annex 3: Impact evaluation questionnaire ...................................................... 61 Annex 4 Group questionnaire........................................................................ 72 Annex 5: Passbook-based record-keeping ..................................................... 78

  • 3

    List of Terms and Acronyms

    UHIKI ............................................................................... Uhamasishaji Hifadhi – Kisarawe CRP .......... Community Resource Person (a voluntary VSL trainer, living in the community) CD ............................................................................................................. Country Director FO .................................................................................................................... Field Officer IGA ............................................................................................. Income-generating activity TShs ................................................... Tanzania Shilling. Basic unit of Tanzanian currency UHIKI ............................................................................... Uhamasishaji Hifadhi – Kisarawe M&E ............................................................................................. Monitoring and evaluation MFI ................................................................................................. Micro-finance institution MIS ................................................................................... Management information system PC ................................................................................................. Programme Coordinator PM ............................................................................................................. Project Manager RESA .............................................................. Regional Office for East and Southern Africa RoI ..................................................................................................... Return on investment ROSCA ................................................................... Rotating savings and credit association SACCO ................................................................................ Savings and credit cooperative SPM ........................ Selection, planning and management (of income-generating activities) VSL .............................................................................................. Village Savings and Loan VSLA ........................................................................ Village Savings and Loan Association

  • 4

    1 Executive Summary

    1.1 Background and approach The evaluation of Plan’s Community-managed Microfinance programmes in Tanzania was undertaken in the last quarter of 2008. The intention was to conduct an evaluation of this entire initiative throughout Uganda and to analyse both impact and the quality and effectiveness of programme implementation. Owing to the short time-frame (12 days in the field) it was decided to focus on the closest of these programmes, around Kisarawe, about 60 Km to the south of Dar es Salaam, run by a local NGO, Uhamasishaji Hifadhi - Kisarawe (UHIKI)

    We conducted a livelihood survey of 104 households, 69 of whom comprised programme participants and 35 of whom were non participants. Table 1: Respondents’ profile

    This was to evaluate programme impact. To determine programme effectiveness we conducted in-depth interviews with 9 Village Savings and Loan Associations (VSLAs) and collected performance data from another 3. We also looked at Plan Kisarawe’s support to UHIKI and interviewed key UHIKI staff, as well as looking, in-depth, at UHIKI’s management information system, organisational structure and programme administrative systems. We were unable to talk directly to CARE staff who had undertaken the original training of UHIKI, except by telephone, since, at the time of the evaluation, they were working in Mwanza, some 700 Km to the North West. The following are the key findings.

    1.2 Impact. We did not have wealth ranking criteria with which to make a detailed assessment, but the evidence is clear that UHIKI has succeeded in providing VSL services to the poorest of the economically active. We found the following: • Participants tend to be drawn from households who own less than the average land

    holding and who are already engaged in income-generating activities. This suggests that they approximate to the definition of the ‘poorest of the economically active’

    • Their productive asset base (linked mainly to agriculture) was roughly the same as non-participants at the start of the programme and has not changed very much

    • There was a significant improvement in their household non-productive asset base and in the number of income-generating activities per participant household, much more so than non-participants. This places them now well ahead of the non-participants group in terms of economic security, despite owning less land.

    • There has been a dramatic shift in the use of VSL as an alternative to the home, friends, banks and ROSCAs for savings purposes. There has been a similar shift away from family and friends as a means of borrowing. There have been no changes in where non-participants borrow and only slight changes in where they save.

    • MFI and SACCO penetration is very low and most participants have abandoned formal institutions in favour of their VSLAs. Participant stores of maize are double that of non-participants, despite owning less land.

    • There has been a reduction in absence from school related to financial reasons, better than for the non-participants

  • 5

    • There has been a slight increase in the number of meals per day, and a significant increase in the variety of food consumed, but a market drift away from high-protein food.

    • Medical services are more accessible to all, owing to the construction of nearby medical facilities, but 15% more participants than non-participants perceive them to be more affordable.

    • There has been a greater than 33% increase in the number of household IGAs, representing just over an average of one IGA per participant household. There has been no change in the number of non-participant IGAs, which remain at 0.58 per household.

    • There has been a 79% increase in household labour allocated to IGAs, amongst participants compared to 45% for non-participants combined with a 39% increase in migration compared to 71% for non participants. This is accompanied by a small but definite increase in the use of child labour: a common finding in all three countries studied

    • There is a significant difference between participants and non-participants in awareness of child rights, but countervailing evidence that this affects school attendance rates.

    • Women’s decision-making in the household and in IGAs has increased both for participants and non-participants, but is far more pronounced amongst participants

    • There was a small decline in membership of financial and agricultural types of social grouping and an increase in membership of religious and politically-based associations (which recorded the biggest increase of 13% to 54%. There were negligible changes for the control group.

    It should be noted that all of these changes were in respect of pre-project indicators, and, most important, relative to changes in the livelihood status of the non-participant group. All in all the impact on livelihoods has been definite and positive and better than for the control group. The overall impression of the participant group and the control group is that of people whose livelihoods were under stress and with some negative trends for both: but VSLA members were faring better and appeared to be better placed to cope.

    1.3 Programme This looks at the manner in which the programme was implemented at three levels: VSLAs, UHIKI and Plan. The results were also positive, with the exception of difficulties in financial administration. We found the following:

    • VSLAs • Most VSLAs were organised and followed a standard procedure that was

    approximately consistent across groups: they had clearly been quite well trained and adequately supervised

    • Cash security was poor, with locks missing from boxes and cash often held by committee members.

    • Passbooks, when used, were accurate, but most VSLAs were not recording loans in the passbooks. The ledger-based loan tracking system was better than in other countries, but still quite inflexible and overly bureaucratic. In particular, the preparation of a formal business plan is inappropriate for this class of client - not many of us have to prepare a business plan when all we want is a small overdraft. Neither should the poor.

    • The ledgers and ledger cards used by the groups were good in respect of recording savings and loan activity. They were complicated and confusing in recording group-level financial information, none of which appeared to be used by the VSLAs. This should be eliminated. Use of passbooks only is recommended, but is not an essential reform.

    • Supervision by Community-based trainers (CRPs) appeared to be adequate, but CRPs have been mis-trained in the use of MIS monthly reporting requirements and need to be re-trained on a simpler version of the MIS. This training should focus on the meaning of the data and the means by which it is collected.

  • 6

    • Rates of return on member savings were low and, in the case of the groups directly studied were in negative territory, on average. This may simply be a data collection issue, or an issue of asking the question properly, but what we saw did not match what the Mis for those particular groups had said in the past. Why this is needs urgently to be examined in case there is a deeper problem of loss.

    Taken in general, the performance of the groups was good from procedural and participatory perspectives. There was a sense of order, application of standardised procedures and a generally positive spirit – but simplification of the loan record-keeping system and a move to passbooks will enhance this.

    • UHIKI • UHIKI is at, or slightly below target in terms of the number of VSLAs and VSLA

    members, and well below target in terms of CRPs. Because UHIKI needs to spend more time on strengthening itself internally and because the original caseloads proposed were very high these targets should in any case be reduced down to about 8,000 members and 400 or so groups. It should not seek to train more than a total of 40 CRPs, which will be a sufficient number for the proposed targets, computing to an annual caseload averaging about 4-5 each, which is realistic in the medium density terrain

    • UHIKI is among the most cost effective of Plan’s programmes. It spends $11.80 per VSL member, compared to $38 in Malawi and $11.63 in Uganda. But, unlike in Uganda, this may be at the expense of solid systems and adequate oversight. UHIKI has implemented VSL through a structure, that uses Field Officers (FOs) as supervisors of volunteer CRPs. The structure is appropriate but the addition of another full-time FO would leave the Supervisor free to exclusively supervise the FOs, without the need to carry a portfolio. Caseloads and are above the normal range for this type of programme, and within the normal range for CRPs

    • FO performance is tracked by an electronic management information system. The meaning of key parts of the MIS data are not understood correctly by 2 out of the 3 FOs, rendering the MIS highly inaccurate. In addition, MIS outputs are not regularly generated and transmitted to field staff but only passed on as reports. All of this needs to be changed

    • Programme records covering VSLA background information, financial performance, routine monitoring and progress through the training system do not exist. There is, in effect, no system for programme planning and management. UHIKI is simply a place where FOs meet and receive material and financial support and chat about the programme. There are no planning and management tools in place and no MIS data that can provide insights into programme problems and potential. This is an issue of institutional capacity, which has not been sufficiently considered in the design and execution of the project.

    • UHIKI’s organisational structure allows for dedicated technical supervision, because the VSL programme is its major activity. Unlike Malawi, where Plan was working through a multi-sectoral, large development organisation that was unwilling to invest significant management attention VSL, Plan has benefitted by working with an organisation that is focused on VSL. The chances that reforms will actually be implemented is therefore good.

    • Plan • Plan has ended up with a successful project in terms of member livelihoods, but it is

    less clear that VSLAs are profitable. It has proven that VSL makes a difference and that it is popular amongst targeted participants. It has created a partner of appropriate scale and structure but needs to invest much more in partner capacity development. This is basically an issue of planning and systems, married to a tighter and more comprehensive management effort

    • Having said this, Plan is structurally hamstrung because it does not have sufficient hands-on responsibility for programme outputs. If Plan is committed to partnership it

  • 7

    needs to be committed not only to providing technical support through additional sub-contracts, but also committed to developing internal expert, on-site technical oversight.

    • The gap between the National Coordinator and UHIKI is too large to ensure that Plan itself develops the sort of PU-based competence that is needed to facilitate expansion. PU-based Community Development Facilitators (CDFs) are responsible for programme design, fund disbursement, administration and reporting. But Plan’s CDFs have no technical training in VSL and are responsible for many other non-VSL activities, making it impossible to provide expert support. This means that Plan has had to depend on an external agency (CARE) to provide TA to the programme, without being able to assess the quality or currency of its knowledge and expertise, nor if the nature of the support is sufficiently comprehensive. The evaluators believe that the CDFs need either to be more actively engaged in the programmatic aspects of UHIKI’s work, or be replaced by a full-time CDF who does nothing but VSL support work, directed by the National VSL Coordinator.

    • Without dedicated technical oversight at the PU level Plan cannot responsibly expand the programme to other partners. CARE is a temporary participant in the process and, like all organisations, makes mistakes. Plan needs to have the capacity to recognise both opportunity and error; to challenge the way that work is done when it needs to be challenged, and to base this critique on a sound set of principles and sound understanding.

    • Plan, therefore, has to internalise the capacity to manage what it funds if it wants to expand VSL. Sub-contracting is the easy way out, but it robs the organisation of the capacity to maximise the benefits of the technology. Plan Tanzania’s dual level of dependency (UHIKI and CARE) means that it does not truly ‘own’ the programme and has neither understanding nor control over results. It cannot, then, think and act strategically with respect to VSL.

    • Without a more hands-on approach it is probable that: • VSL projects will remain isolated programmatically and fail to achieve their full

    potential for economic and social transformation • Plan will be forever dependent on technical support that (at the PUs) it does not

    fully understand; never internalise the knowledge of VSL and find it hard exploit its true potential

    • Things can go off track and become hard to fix.

  • 8

    2 Terms of reference The objectives and outputs of the evaluation were designed as a generic exercise, to cover Malawi, Uganda and Tanzania. Foe each country they are as follows:

    2.1 Purposes of the evaluation: The assessment fits into the framework of the Regional Microfinance Fund Acquisition Strategy agreed upon by RESA and Plan USA and serve a three-fold purpose: • To provide evidence in an attractive and readable format that participation in Village

    Savings and Loan Associations (VSLAs) has beneficial socio-economic effects on VSLA members, especially poor rural women. Plan hopes that the assessment reports can be used to attract the support of funding organizations for community-based microfinance programs in RESA and elsewhere in the Plan world.

    • To provide information on how well Plan country offices are managing the implementation of VS&L projects and how effectively and efficiently their partners are implementing them.

    • To provide a sense of the strategic vision of Plan country offices to scale up community-based microfinance and to provide suggestions to improve the vision and make it operationally effective

    The model for this assessment is the one done of the Kupfuma Ishungu program in Zimbabwe for CARE International by Hugh Allen and Pauline Hobane

    2.2 Tasks 2.2.1 Impact: • Adapt the basic survey questionnaire used in the Kupfuma Ishungu study to the context

    of the three countries. Categories of impact should include assets, health and nutrition, children’s education, habitat, the development of household enterprise, the social status of women and the effects of VSLAs on the community.

    • Choose a representative sample of the population and administer the questionnaire. Since there were no baseline surveys in the three countries, attempt to reconstruct a “before and after” situation through the use of recall.

    • Analyze the data and assess the results, especially in light of the age of the project. • Make recommendations as to how the project could initiate or improve a system of

    impact monitoring that would provide periodic data on the effects of VS&L on individuals (especially women), households and communities.

    2.2.2 Program effectiveness and efficiency: • Determine to what extent Plan’s partners were trained in the VS&L methodology and

    have mastered its basic principles and procedures. • Determine the effectiveness of the projects’ training and supervision of VSLAs. • Assess the profitability of VSLAs and determine the causes • Determine whether the project is using the VSL Portfolio Tracking System (MIS), and to

    what extent the system is being used properly (i.e. reports analyzed by supervisors and feedback given to FOs; reports used to identify and solve problems; reports being read and used by higher levels of management).

    • Assess the efficiency of the projects in terms of staff caseloads, drop-outs and cost per member

    • Assess the effectiveness of Plan in dedicating staff to VS&L program management and monitoring at the country and program unit levels.

  • 9

    • Determine to what extent Plan’s programme staff (country office and programme units) understand VS&L and the place and contribution of community-based microfinance in the country programme.

    • Make recommendations for improving the effectiveness and efficiency of each project

    2.2.3 Strategic vision and scaling up: • Determine to what extent Plan staff are aware of the content of the regional microfinance

    strategy and have used it for program planning. • Determine to what extent the Plan country programme is able to train multiple

    implementing partners to scale up • Assess the effectiveness of the use of Village Agents to extend outreach at low cost.

    Does the project have and use a methodology for identifying, training and supervising its Village Agents?

    • Determine the extent to which “spontaneous groups” figure in the outreach strategy of each project and whether the projects are able to identify and track such groups.

    • Do linkages with MFIs or banks figure in the scale up strategy? Assess to what extent such linkages are being systematically promoted so as to minimize the danger to VSLAs and increase the chances of success?

    • Make recommendations on strategy and operational measures to improve each country office’s ability to scale up VS&L services.

    2.2.4 Outputs expected from the Evaluation The major output of the final evaluation is a report of findings and recommendations on the way forward. The following process should be followed

    • Submission of draft report in soft copy within 5 days after the field data collection, for the CD, PM and the PC to review and approve.

    • Workshop presentation of the interim findings to wider stakeholders for discussion and refinement.

    • Submission of the final report both in hard and soft copies within 3 days after incorporating comments.

  • 10

    3 Evaluation design and approach

    3.1 Background The evaluation was undertaken by an international consultant, leading a team of 6 enumerators and supported by a Canadian-based data analyst. Field visits were accompanied by partner staff and designated community-based FOs, who undertook translation services.

    The team leader took major responsibility for design of the evaluation framework, which was pre-tested on 12 households and revised. A survey of

    The international consultant was a microfinance specialist with extensive experience in community-managed microfinance: he concentrated more on the financial and institutional quality/performance of the women’s groups and cooperatives, while the enumerators took care of the household level interviews. Debriefings and quality control on the interview results were held daily. The evaluation team members were engaged in constant consultations and feedback on aspects of their observations and findings.

    3.2 Design

    Using the PO as the point of departure for the design of the evaluation, the following general framework was used.

    Figure 1: General framework

    It was decided to conduct a simultaneous impact and programme evaluation.

    The impact evaluation focussed on the final goal of the project and looked principally at the lives of programme participants. This required the evaluation team to adopt a two-pronged approach • interviews with at least 69 programme clients in their households and interviews with a

    control group of 35 non-participants drawn from the same communities • Observation of 12 VSLA meetings (196 members present in total) and interviews with

    management committees.

    The following framework for the impact evaluation was followed:

    Area of Inquiry Type of Evaluation Level of Inquiry

    Permanent changes in the lives of the target group

    Impact • Clients: • Household • Individual

    Community-based service provider (VSLA) capacity, performance and efficiency

    Programme • Community-based service provider

    • Clients

    Implementation: effectiveness and efficiency of the implementing institution and capacity to sustain the programme

    Programme • Implementing institutions • Plan Malawi • UHIKI

    Programme Design: relevance of the programme to the needs of clients and method of delivery

    Programme • Community-based service provider

    • Clients

  • 11

    Figure 2: Impact evaluation framework

    The programme evaluation looked at three different levels:

    • The efficiency and effectiveness of VSLAs in terms of: • Their organisational capacity, • The degree to which they satisfy their members needs • Their level of financial sustainability and; • T heir efficiency

    • The efficiency of UHIKI in the delivery of its services • Staff efficiency • Cost per member served • Strategy for sustainability

    • The quality and appropriateness of the programme’s design in terms of: • Interventions: were the interventions appropriate • Methodology: was the methodology used to deliver the interventions appropriate • Was the way in which the services were delivered by the project and through its

    partnership appropriate and effective

    The following table outlines the issues that were considered in looking at women’s groups and cooperatives:

    Area of Inquiry Indicators

    Means of Measurement Approach

    Individual • Control of resources by women, including enterprise resources, business and loan decision making

    • Household decision making • Level of self-esteem on the part of women clients • Female labour allocation • Number of working children • HIV/Aids awareness • Capacity to define ‘empowerment’ • Awareness of legal rights • Participation in group & public activities, at what level

    • Individual participant interview

    • Random sample of households and interview with group member. 60 interviews with members

    • 28 interviews with non members.

    • Household interviews using structured questionnaire Daily review of findings

    Household • Level of household assets: productive and non-productive

    • Consumption of goods & services: housing, education, food, health

    • Number of income-generating activities • Allocation of HH labour to income-generating activities

    • Household survey

    Community • Changes in attitudes towards women’s rights • Changes in attitudes towards women’s status and

    participation in decision-making

    • FGD and household survey

  • 12

    Figure 3: VSLA performance

    Area of Inquiry Indicators Means of Measurement Approach

    Institutional capacity

    • Goals • Staff/leadership capacity • Level of financial

    resources/sustainability • Plans exist and are followed • Appropriate organisational structure • Systems (financial, HR, M&E) • Quality and nature of linkages to other

    organisations

    • FGD of members • Interview with

    staff/leaders • Review of records

    • FGD • Interview with

    staff/leaders • Review of records • FGD

    • 11 groups • Evaluators

    first witnessed VSLA meetings and assessed the VSLAs quality in terms of leadership and participant performance and behaviour.

    .

    Client/member satisfaction proxies

    • Client transaction costs (time and ease of participation) and tradeoffs

    • Participation in group-level decision-making

    • Attendance rates • Retention rate (%) • Membership growth rate

    • FGD

    • FGD

    • Review of records • Review of records • Review of records

    Financial performance

    • Average savings per member mobilised to date

    • Annualised return on savings • Average member investment • Average outstanding loan size

    • Review of records

    • Review of records • Review of records • Review of records

    Operating efficiency

    • % of members with active loans • Loan fund utilisation rate

    • Review of records • Review of records

    Figure 4 outlines the issues that were considered in looking at the manner of UHIKI’s ’s implementation of the project.

    Figure 4: UHIKI/Plan implementation strategy and efficiency

    Area of Inquiry Indicators Means of Measurement

    Strategy • Service delivery channel (groups and cooperatives) • Partnership strategy • Strategy for sustainability of service delivery • Strategy for long-term programme growth

    • Review of project proposal

    Operating efficiency

    • Caseload: Associations per field staff • Caseload: members per field staff • Ratio of field staff to total staff • Staff turnover • Cost per member assisted: Outputs 1-4

    • Review of records

    Institutional capacity

    • Goals • Staff/leadership capacity • Level of financial resources/sustainability • Plans exist and are followed • Appropriate organisational structure • Systems (financial, HR, M&E) • Quality and nature of linkages to other organisations

    • Interview with staff/leaders

    • Review of records

  • 13

    Figure 5 below outlines the issues that were considered in evaluating the programme’s design.

    Figure 5: Programme design

    Area of Inquiry Indicators Means of Measurement

    Interventions and Methodology

    • Credit, savings and insurance services are appropriate to the needs and characteristics of clients and the environment

    • Service delivery methodology is appropriate to clients and environment

    • Interventions and methodology are delivered cost-effectively (Cost/member/intervention)

    • A realistic strategy for sustainability/exit exists

    • FGD and household survey and staff enquiry

    • FGD • Review of budgets and

    expenditure • Project plan and

    management interviews

    Delivery Channel (Plan-UHIKI)

    • Strategic plans exist that are compatible with the organisation’s goals and objectives and are regularly revised

    • Operational plans exist that are compatible with the strategic plan and are regularly updated

    • Individual work plans exist and are regularly updated. These focus on results in conformity with outputs and goals

    • Organisational structure is adequate and appropriate • Systems: • HR • M&E: a system of performance measurement exists,

    which focuses on results and ensures close supervision of field activities

    • Financial management • Relationships with partners

    • Review of programme proposal or business plan

    • Operational or program reports

    • Review of Work plans • Review of organigram • Review of HR records • Review of M&E/MIS

    records • Review of financial

    records • Partner interviews

  • 14

    4 Impact evaluation 4.1 Goals of the project Plan’s strategy in Kisarawe is to implement VSL through local partners. As part of this strategy, UHIKI was formed and registered as a local NGO in October 2007, growing out of a three-year food security project . The first VSLA groups were formed prior to the establishment of UHIKI, in January 2007. Thus, the results shown here are for 22 months out of 48. As such UHIKI has reached 51% of the number of members targeted in slightly more than half the number of members required under the log frame in 46% of the time and has formed 48% of the targeted number of groups targeted. Thus, the project is on track in terms of its outputs. It has done this with only 20% of the proposed number of Community Resource Persons (CRPs).

    The following table lists the goals, Table 1: Programme logic and summary of achievements to date

  • 15

    4.2 Approach to the evaluation 4.2.1 Stratification of Associations First we selected 12 VSL Associations, listed in Annex 2. The Associations were spread across the entire project area and were selected from amongst the oldest: the oldest was established in December 2006 and the youngest in December 2007. This was done to determine the financial viability of the oldest groups and also to contact members who had participated in the programme for the longest time of any member. A three-year time frame is normally required to measure significant change when VSL programmes are implemented. Since even the oldest groups were less than two years old, it was important to try and contact only the older groups, whose average age was 56 weeks. 100 of the VSLAs were in their second cycle (average age 75 weeks) while 140 groups were in their first cycle, with an average age of 30 weeks). The relative youth of the groups made it likely that while group profitability could be optimal, household level impact was likely to be less evident than in Malawi, where Plan has been implementing VSL for 3 years. This proved to be the case.

    4.2.2 Choice of respondents and survey methodology

    The project did not have a baseline survey, nor control group, making it hard to establish livelihood benchmarks at the time of programme inception. Thus, we used a method that has been applied in the other VSL impact studies carried out under this contract, using recall techniques. First each enumerator randomly selected two respondents from the VSLA. Then each respondent identified a non-member, from the same community, whom they considered to be of similar social and economic status to the VSLA member at the time they became a member. Both the members and the non-members were then administered a questionnaire that required them to recall their situation at the time of the programme’s inception and to report on the same set of data at the present time. The questionnaire is appended in Annex 3. The recall methodology depends for its accuracy on identification of livelihood indicators that are easily remembered and can be counted.

    The breakdown of respondent is shown in table 2 below:

    Table 2: Composition of respondents

    The main areas of enquiry were:

    Household • Level of household assets: both productive and non-productive. Quantitative and

    usually reliably recalled • Consumption of goods & services: housing, education, food, health. This is measured

    more in terms of positive or negative changes, except for investment in housing, which is quantitative. Highly reliable

    • Number of income-generating activities. Quantitative and reliable • Allocation of HH labour to income-generating activities. Qualitative and reliable Community • Changes in attitudes towards women’s rights. Qualitative but frequently tied to specific

    events.

  • 16

    • Changes in attitudes towards women’s status and participation in decision-making. Status indicators are qualitative, but concrete, while participation in household decision-making is only qualitative.

    Individual • Control of resources by women, including enterprise resources, business and loan

    decision making. Qualitative • Household decision making. Qualitative • Level of self-esteem on the part of women clients. Qualitative • Female labour allocation. Qualitative • Number of working children. Reliable recall. • Awareness of children’s legal rights. Reliable recall • Participation (at different levels) in group & public activities. Measurable

    4.3 Household survey findings 4.3.1 Changes in asset ownership The following tables show changes in productive and non-productive asset levels. No wealth ranking had been carried out by the project prior to the evaluation, so it was not possible to determine the socio-economic level from these data. Nevertheless it is possible to draw some basic conclusions. Both the participant groups and the non-participants were moderately secure by Tanzanian standards, with relatively large land area at their disposal, but with participants owning 32% less land (3.75 acres compared to 5.49 for the non-participant group. Both groups had a low incidence of productive asset ownership, except for hoes and pangas. This was less true of non-productive assets, with most households having significant investments in furnishings, tableware and glassware. There was a high incidence of mosquito-net ownership, which is important in this coastal region. Of all the studies conducted in this 3-country evaluation exercise, the participant group and non-participant group were the most closely matched.

    The level of productive asset ownership was small and changes in asset ownership were also small, except for land. In this case the participant group had increased its land holding by 12% to an average of 4.18 acres, while the non participants had increased by 8% to 5.93 acres. In terms of acreage both groups had increased land holding by the same amount: .44 of an acre. Thus, we concluded that there was no conclusive evidence of differences in the levels of productive asset ownership or rate of acquisition.

    The difference between the two groups is clearer when looking at non-productive assets. Taken in general the participant group started out with a higher level of ownership in most categories and made bigger gains, except for the upgrading from straw to foam mattresses, where the non-participant group had made significant gains (but still had a lower incidence of ownership). There is a definite difference in the gains made by the participant group relative to the non-participant group and this may be ascribed to higher levels of participation in off-farm IGAs. The participant group started out with 2.6 IGAs per household and ended up with 3.03, while the non-participant group started out with 1.03 and ended up with 1.31: a lower average increase (0.42 vs. 0.28) although larger in percentage terms (16% to 27%). Thus, although ownership of the most nominally important productive asset (land) favours non-participants, it is clear that non-productive asset ownership probably accrues more to off-farm IGAs. We can therefore conclude that the difference in non-productive asset ownership, when considered in the light of the number of household IGAs is a proxy for higher levels of economic security. We can also conclude that running an IGA is likely to lead to a higher level of asset ownership than land ownership. Land ownership does not necessarily translate into extra production or productivity. The type of tools that are owned will not permit the average family of 5-6 people to cultivate much more than they own: but IGA participation may be decisive.

  • 17

    Table 3: Productive assets – participants and non-participants

  • 18

    Table 4: Non-productive assets – participants and non-participants

  • 19

    4.3.2 Housing investment Table 5 shows a small but clear difference between the participant group and the control group.

    Table 5: Housing investment

    A small but indicative percentage of the participant group has made major investment in improved housing, while non-participants seem to have made less, except in terms of roofing sheets. 10.1% of participants have built cement block houses and, starting off with a 14% higher incidence of mud-house ownership have reduced this to 65.2%, slightly less than the non-participant group – more of whom are now living in mud houses (65.7% up from 57.1%). 42.0% of the participant group now live in houses with cement floors, up from 31.3%, while none of the non-participant group has done so. Similarly significant improvements in roofing material (13.2%) have accrued to the participant group, while only 5.7% of the non-participants have done so. The evidence is clear. At the start of the programme most participants lived in lower standard housing than non-participants (particularly in terms of the material used in wall construction), while now the situation is reversed. The non-participant group’s housing quality has diminished and the participant’s housing quality has improved. The gains apply only to a maximum of 10% of the participant group but are an important indicator of changing levels of purchasing power.

    4.3.3 How people save and how people borrow It is clear from table 6 that VSL members were more active savers before joining the programme than non participants and were more likely to use a wider and more formal range of savings instruments. While participants reduced their membership of savings institutions after joining VSLAs this was by small percentages: they did not abandon them wholesale, as happened in Malawi and Uganda. Strikingly, a very high percentage were account holders in local banks (25%) and this has only declined by 2%. Less than half this

  • 20

    number of the non-participants had bank accounts, (11%) and this has remained unchanged. It is clear that VSL is not so much a replacement for local savings opportunities as it is an addition to an already rich menu. Food stocks of participants are higher than for non-participants across all the major staples.

    Table 6: How people save and borrow

    When it comes to access to credit, the results are fairly similar. While credit access through banks and shopkeepers has remain quite high, it seems that VSLAs have reduced dependence on family and friends for loans, while for non-participants this dependence has increased.

    These results are in line with the early inference in 4.3.1 (changes in asset ownership) that VSLA participants are probably more active in business and the formal economy than non-participants is confirmed by these results.

    4.3.4 Welfare: access to education The results here are of some concern. Tanzania has universal primary education, so one would expect to find high enrolment rates. That seems to be the case. It is, however, of concern that while the already low rate of male absence from school has dropped close to Zero, it has more than doubled in the case of the female children of participants while declining to zero for the girls of non-participants. A finding that is constant across all three country studies is that VSL membership seems to have a negative affect on school attendance relative to a non-participant control-group. But this is the first case where it has an emphatic gender dimension: there is an actual increase in female absence while male absence has decreased. This needs to be taken in proportion: it is only 3 individuals affected in over 100 households: but it is a consistent picture and may get worse as investment in IGAs continues to increase.

  • 21

    Table 7: Children per household unable to attend school

    4.3.5 Welfare: nutrition - meals eaten daily and food variety

    Table 8: Number of meals eaten daily per household

    This is, universally, a well-nourished society, but VSL participants ate more frequently than non-participants three years ago, but not by much (2.85 meals a day average compared to 2.74). Thus, neither group is food insecure. Having said that, the participant group is, on average, eating slightly more frequently every day (2.93 times), while the non-participant group’s food intake has declined slightly from 2.74 to 2.71 meals per day. This correlates with the levels of per-household savings in basic staples, which, for maize and millet is 30% greater for participants. The reasons for increase are headed by easier access to cash, followed by VSLA membership. Since local production is a long way third as a cause of increased consumption, it appears, again, that IGA participation is the most likely cause of increased income. Not much should be made of these changes because they still describe a food secure population, both participant and non-participant. But the trend favours the participants.

    Table 9 presents a very clear picture. Participants report a higher incidence of increased food consumption across all food groups and have, in terms of the percentage of participants, suffered the least reduction, except for dairy products. Both groups report a dramatic drop in meat consumption, but the non-participants are the worse affected.

  • 22

    Table 9: Variety of food consumed – most common food groups

    4.3.6 Welfare: Access to affordable health services Table 10 confirms a difference in the ability of the participant group and the non-participant group to afford medical services. This question was framed in terms of affordability and not simply access. Alarmingly, both groups report a decline in access, although VSLA members are less affected.

    Table 10: Affordability of medical services

  • 23

    There is evidently some confusion in the responses, because, although the question is framed in terms of affordability, the clear implication is that there is reduced access, indicating that there may have been fewer service providers in the local area than before. We did not confirm this hypothesis. But the second most important reason given was decreased income. We cannot draw definitive conclusions, but because it is clear that access (however defined) has reduced, Plan Tanzania would do well to discover if this is a more widely held view and, if so, to determine causes. When it is evident that VSL members (if not the non-participants) have more money in their pockets than heretofore, this result is counter-intuitive, unless there are exogenous factors at work.

    4.3.7 Income-generating activities Table 11: Incidence of off-farm income-generating activities per household

    There is a small increase in the number of IGAs in participant group households but a bigger one in the case of non-participants (16% vs. 27%). Translated into actual numbers, however, and noting that participant households start out with about a 2.5:1 superiority in the incidence of IGAs per household, the per-household increase in IGAs is much greater in participant households than non-participant households This gives participant households a massive – and increasing – advantage in securing basic livelihood necessities. This correlates closely to land holding. When land-holdings are small, people are driven to invest in IGAs because their basic needs may not be met. In Tanzania, this disadvantage on the part of the participant group (they own less land) has been turned into an advantage. One may reasonably infer that IGA activity is turning out to be more profitable than farming and the proximity of Dar es salaam and the availability of low-cost transport over good roads seems to mean that the disadvantage of small landholding can be converted into an advantage where IGA investment opportunities – and financial services – abound. This explains the high use rate of formal financial services amongst the participant group. The evident importance of IGAs and the opportunities provided by the market for IGA participation is an important indicator of the potential relevance of promoting IGA training. Table 12: Adult labour allocated to income-generating activities

    Table 12 indicates that adults are able to take advantage of IGA investment, but there is a striking superiority amongst the participant group. Starting from a slightly less favourable position men from participant households have significantly increased the amount of time they spend in IGAs, which now exceeds that of men from non-participant households. Amongst women the percentage change is almost the same, but while only one household

  • 24

    in 2 is likely to have a woman participate in IGA activity, this rises to 3 households out of 4 for participants. This has led to an almost identical labour burden between men and women in participant households, while women in non-participant households are much less likely than their male family members to be engaged in IGAs. While this shows an increase in female labour burden, it is viewed positively by respondents, who view their investment in IGAs as empowering and enabling them to achieve status in the community and equality in the home.

    Where the findings are distinctly negative is when we examined its impact on children, with a disproportionate burden felt by girls. Table 13 below indicates the level of child labour participation in IGAs. Table 13: Child labour participation in IGAs

    The change in direction for male children is small and the incidence per household is also low (roughly about 1 household in 20). For girls the story is different. Prior to the project the household incidence of female child labour was about 1 household in 8. Now it is approximately 1 household in 3. In both cases (male and female) there appears to be no change in child labour levels in non-participant households, but a large increase in participant households that cannot be due to chance. Research may be needed into what this ‘child labour’ actually entails. It may be as little as helping the mother to shell peas at night before going to the market, but it might be more significant. What cannot be disputed, however, is that VSL is probably responsible for increasing the burden of female child labour. And it is much more noticeable in Kisarawe than in Tororo in Uganda or Mzuzu in Malawi. This is not good news and Plan Tanzania needs first to understand the issue more comprehensively and then, if it turns out to be significant, work with VSLAs to highlight the issue and see what solutions the VSLAs themselves may work out.

    Table 14: Length of annual operation of IGAs

    The findings here are counter-intuitive. They indicate that the length of time that IGAs stay in operation for the participant group households is actually reducing by a significant amount, with non-participant household IGAs affected even more. This is inconsistent with other findings, except to the extent that increased numbers of IGAs in households may allow for more agile shifts of labour and capital to alternative investments. Further examination of this issue is needed.

  • 25

    Table 15: Wage labour

    The findings on wage labour indicate that the participant group was slightly less likely to have a household member working for a wage than the non-participant group, prior to joining a VSLA (0.44 per household vs. 0.49). This ratio has changed significantly, with non-participants more likely to be working outside the home, or home-based IGAs than participants, who are much less likely to be doing so now than prior to their joining a VSLA (0.44 down to 0.35). This finding is value neutral, without deeper enquiry. It is not clear what type of work is involved (high or low value/status) and whether or not it is undertaken as a matter of necessity or a matter of opportunity, but a plausible hypothesis is that increased income from IGAs reduces the need to seek work (which is fairly easily available in nearby Dar es salaam).

    Table 16: Migration

    The hypothesis suggested above is reinforced by this finding. Participant family members are approximately 3 times less likely to be seeking work today compared to non-participant family members. While the incidence in VSLA member households has remained static at 6%, the non-participant households are increasingly likely to have a migrant member, with an incidence of 17%, up from 14% prior to the project. It seems probable that participant families have other income-earning opportunities in the community than non-participants and are therefore less likely to seek wage-earning work or to migrate. These tendencies existed prior to the advent of VSL, but have become more pronounced.

    4.3.8 Change in social status

    Table 17: Perceptions of change in social status

    A significant majority of both participants and non-participants believe that their social standing is better than it was at the time the project started. but the belief is strongly pronounced in the case of participants. Tanzania has enjoyed steady economic growth over the last 20 years and has achieved an annual increase in GDP of about 6% for the last 10 years. This may account for the general prevalence of optimism. But the emphatic optimism of VSL members may be linked to the fact that this group is more in touch with markets and is therefore more likely to be prosperous and therefore optimistic. It cannot be concluded that VSL membership is the cause: rather it may be the result of a fundamentally different level of economic engagement and entrepreneurial capacity.

  • 26

    4.3.9 Attribution for positive change in social status The findings suggest that those whose livelihoods have improved believe that their individual positive attitude and social manners were, more than any other factor, the reason that they are better regarded in their communities. Access to financial services (of all types) was only third in line. This is the first such result from the three studies. In Uganda and Malawi, there was an iron relationship between access to finance, business investment and social status. Here, it seems to be cultural, with social status linked to non-material attributes. Kisarawe is a Muslim culture and social bonds are strong. It can be speculated that the importance placed on intangible, value-based interaction may be inherent to the culture, displacing more tangible indicators. Table 18: Why people think their social status has improved

    Conversely, those who had suffered a perceived loss of status did attribute this to purely material or exogenous conditions. It is worthy of note that a decrease in social status is most strongly linked to reduced access or control over material conditions (including IGAs). This correlates with what we know about the lower incidence of IGA ownership amongst the control group, whose non-productive asset base and housing status was markedly inferior to the participant group. Thus, while status may be linked to matters of positive personality, it could be speculated that a positive personility is correlated with IGA ownership and growth. Put simply, a prosperous business person tends to be more cheerful (and therefore better liked) than someone whose material condition is deteriorating. Table 19: Why people think their social status has deteriorated

  • 27

    4.3.10 Gender: control over family resources There seems to be a steady, population-wide shift towards significant female control over material and financial resources. At least 2/3 of women (and up to ¾) are reported to control family material resources and both groups have achieved similar percentage gains. But participant families report significant gains in control over financial resources compared to non-participant families (9% vs. 0%), with participants reporting that nearly 60% of women are the chief decision-makers regarding finance in the family, compared to less than half of the control group.

    Table 20: Female control over important household livelihood resources

  • 28

    Table 21: Income and IGA decision-making

    Table 21 shows only a slight shift away from exclusive male decision-making related to cash income, with women, reaping this small gain. It seems that unlike the studies in Uganda and Malawi, there is a pre-existing culture of mutual decision-making in respect of the disposition of cash income.

    With respect to control over IGAs a similar phenomenon can be observed. This is a culture in which IGAs are long-established and, while VSL appears to have accelerated the growth in numbers it does not appear unduly to have influenced pre-existing decision-making powers. There is a slight shift in favour of female decision-making amongst participant households, but the bulk of households, approaching 2/3 make these decisions mutually, with non-participant households

    more likely to do so. It seems that VSLAs are service providers that fit into already dynamic patterns of economic activity. Their influence on decision-making appears to be additive to what are already established, fundamentally healthy ways of families working together and no special claims can be made.

    4.3.11 Child rights Table 22 on the following page compares the awareness of both sets of respondents about the legally enshrined set of children’s rights. In every category there was a higher awareness of these rights amongst participants than non-participants, excepting non-participants’ awareness of right to play (2%). There is very little that is surprising here, except to note an average 6% higher level of rights awareness. This does not reflect on the VSLA as such, except to the extent that a VSLA is a forum through which Plan is able to make its messages heard.

  • 29

    Table 22: Awareness of children’s rights

    Simple awareness of rights has no meaning unless there are actual changes in behaviour, leading to concrete changes in childhood conditions. The following were cited.

    Table 23: Evidence of impact of the awareness of children’s rights

    The responses cited here refer to the frequency with which examples of change, given by all respondents, correlated with a particular right. There was a generally greater frequency of response amongst the participant group, except for improved care of children.

    4.3.12 Respondent labour

    Table 24: Participant time/labour allocation to areas of major responsibility

    The findings here are unsurprising: greater time spent on IGAs by participants (55%) compared to 40% for the control group, while much more time (64%) spent on group based activities, compared to 23% for the control group. Compared to the non-participant

  • 30

    group, a much larger percentage of the participant group had reduced the time they spent on housework and child care (49% vs. 37%). The differences between the two groups are smaller than in the Malawian and Ugandan studies, but still clear.

    Earlier in this study it emerged that the more economically active participants are tending, slightly, to increase the burden of child labour. Having less time to spend on household chores may mean that this is tending to fall on the shoulders of children. Since this is also showing up in terms of school attendance, Plan needs to be aware of the issue and to understand it better. .

    4.3.13 Social capital: group participation

    Table 25: Incidence of group membership

    This table is interesting. It shows a generally low rate of participation in associations that are related to social services (women’s guild, education, health) and agriculture. It shows a highest rate of general participation in political associations amongst all the countries studied. But there was also a higher rate of participation in financial services, right from the start, which correlates with what we learned about the pre-project numbers of participant IGAs. VSL participation seems to have had a negative impact on membership in other types of association.

    Table 26 below shows that household membership in community groups was unchanged amongst non-participants, it increased by 0.58 members per

    household amongst the participant group. Since these were people who, by definition were VSL members it implies a drop in participation in other types of social grouping of one person for every two participant households.

    Table 26: Per-household average incidence of group membership (all types9

  • 31

    4.3.14 Attribution

    Table 27: Causation – principal causes of change in the lives of participants

    59 participants provided 80 responses to this question and indicate a strong consensus by 56% that access to financial services (principally VSLA savings and credit access) is the main reason for changes in their lives and this correlated quite strongly to increased participation in IGAs. It is particularly interesting that increased income is so far down the list. It can be inferred from these findings that social status is less linked to actual income than participation in IGA activity. This correlates well with what we have found distinguishing the household assets of participants and non-participants (if we presume that higher levels of visible household asset ownership confers status). The frequency with which hard work was mentioned also correlates with what we have found concerning increases in household labour allocation to off-farm economic activity. There is a clear picture here. Because they have access to financial resources, this leads to increased levels of economic activity which, in and of itself, confers status.

    4.3.15 Inability to participate UHIKI seems to be offering VSL services to people in the communities studied who were already quite economically active. Table 28 explains why non-participants have not joined VSLAs. Table 28: Reasons who non-participants have not joined a VSLA

    About 1/3 of the 32 participants responding to this question felt that they did not have the ability to save the minimum need to make regular contributions to a VSLA. This is an important finding. An equally important finding is that 22% of respondents were simply unaware of VSL, which correlates strongly with the low general rate of participation in other types of group-based activities. Plan and UHIKI should take more active measures in any follow-on activity to target people who are poorer than the average participant at this time and should also be more active in promoting awareness of VSL. This can be done through ‘open days’ (probably at the share-out) when testimonials from some of the poorest members can attest to the social and economic benefits and the low comparative level of risk. Community-based facilitators can stress on such occasions the ability of a

  • 32

    group to set its own share value or even suspend contributions for the whole group at the harshest time of the year. UHIKI and Plan should build this rhetoric and approach into the next phase, but introduce it only in areas where VSL has already taken hold.

    Table 29 below shows that only 43% of non-participants are interested in joining VSLAs and a similar number have a positive view of what VSLAs are able to achieve. This is a surprisingly low percentage. In Uganda 67% of people thought well of VSL. The findings here and the preceding table may simply reflect that fact that the programme is relatively young and most groups have been in operation for less than a year. But since the impact findings are quite positive it is probably wise to initiate a deliberate process, suggested above, of organised observation/involvement in group meetings by outside observers, to create a deeper appreciation of what is happening. In Zimbabwe (before the current economic meltdown) CARE’s VSL groups had an open day every three months. This was a meeting on a sports field, attended by as many as 50 groups. Local leaders and important government officials were invited and members brought along their products for sale. The centrepiece of the day-long meeting was songs, theatre and personal testimonials of members who had improved their living standard. Other members of the community were also invited, to witness the event and ask questions. No-one in the project area was unaware of VSL. A similar approach can be considered in Kisarawe. Table 29: What non-participants think of VSLAs

    4.3.16 Attitudes towards the future Table 30: Attitudes towards the future

    4.3.17 Interpretation of results The results of this survey are moderately positive. The use of a control group significantly strengthens the findings. The following broad conclusions can be drawn: • Participants tend to be drawn from households who own less than the average land

    holding and who are already engaged in income-generating activities. This suggests that they approximate to the definition of the ‘poorest of the economically active’

    • Their productive asset base (linked mainly to agriculture) was roughly the same as non-participants at the start of the programme and has not changed very much

  • 33

    • There was a significant improvement in their household non-productive asset base and in the number of income-generating activities per participant household, much more so than non-participants. This places them now well ahead of the non-participants group in terms of economic security, despite owning less land.

    • There has been a dramatic shift in the use of VSL as an alternative to the home, friends, banks and ROSCAs for savings purposes. There has been a similar shift away from family and friends as a means of borrowing. There have been no changes in where non-participants borrow and only slight changes in where they save.

    • MFI and SACCO penetration is very low and most participants have abandoned formal institutions in favour of their VSLAs. Participant stores of maize are double that of non-participants, despite owning less land.

    • There has been a reduction in absence from school related to financial reasons, better than for the non-participants

    • There has been a slight increase in the number of meals per day, and a significant increase in the variety of food consumed, but a market drift away from high-protein food.

    • Medical services are more accessible to all, owing to the construction of nearby medical facilities, but 15% more participants than non-participants perceive them to be more affordable.

    • There has been a greater than 33% increase in the number of household IGAs, representing just over an average of one IGA per participant household. There has been no change in the number of non-participant IGAs, which remain at 0.58 per household.

    • There has been a 79% increase in household labour allocated to IGAs, amongst participants compared to 45% for non-participants combined with a 39% increase in migration compared to 71% for non participants. This is accompanied by a small but definite increase in the use of child labour: a common finding in all three countries studied

    • There is a significant difference between participants and non-participants in awareness of child rights, but countervailing evidence that this affects school attendance rates.

    • Women’s decision-making in the household and in IGAs has increased both for participants and non-participants, but is far more pronounced amongst participants

    • There was a small decline in membership of financial and agricultural types of social grouping and an increase in membership of religious and politically-based associations (which recorded the biggest increase of 13% to 54%. There were negligible changes for the control group.

    The study is weakened by the absence of a historical control-group and the inevitably subjective methods used to identify non-participant controls during the research. Nevertheless, the results should remove most doubts about the efficacy of VSL as a tool for securing and improving social and economic conditions – at least for the middle classes. The major challenges for UHIKI and Plan Tanzania are:

    • to make the programme more inclusive, especially down-market • to ensure that child labour practices are ameliorated • to identify collateral interventions, such as training support to the establishment,

    financing and management of household IGAs, that are related to the major activities of the VSLA and from which synergies can be expected to accrue.

  • 34

    5 Programme evaluation

    5.1 Service provider performance: institutional capacity In talking about service providers we are referring to the VSLAs, 12 of which we visited. We first of all observed savings and loan meetings and then interviewed the management committee about the group’s financial performance. VSLAs trained by UHIKI are providing a valuable service to their members and are well supported and attended, but the rates of profitability cannot easily be determined because the results from the MIS do not match the results we found in the field amongst our sample of 12 groups and much of the data that has been fed into the management information system (MIS) is clearly inaccurate. For the one FO whose MIS data seems more accurate than the others the return on member savings appears to average 18.4%. This is lower than for Malawi (about 44%) and Uganda (about 35%). This may have to do with a lower average interest-rate charged compared to the other two countries (5% per month vs. 10%). Nevertheless from the visits made to 12 VSLAs there appears to have been a loss of equity for about half the groups, which may have to do with poor security procedures.

    5.1.1 Procedures and transparency All of the groups that we visited were older groups that had been operating more or less independently for a long time. As such they had developed different ways of conducting meetings, with some sticking to a formal manner of meetings and some less so, as these pictures indicate. All of the meetings that we attended were well-organised and quite formal, following a strict procedure. In most cases the money-counting bowls were in use and procedures were transparent and generally consistent.

    The use of written ledgers (addressed below) tended to put the meeting firmly in the hands of a single individual (the Record-keeper)

    with the Chairperson playing a somewhat vestigial role. Sometimes this tended to exclude participation, although, when it mattered (such as when loan disbursements, repayments and savings were mobilised) this impression diminished. It is clear that the CRPs employed by UHIKI have properly emphasised the importance of orderliness, transparency and the consistent application of standard procedures.

    5.1.2 Cash security In general cash security is not as good as it should be, although we noted cases in which the balance of cash was held outside the box and starting balances were usually not counted at the start of the meeting. In all cases the metal cash boxes were well made and in use, but there was a definite reduction in security as many of the boxes only had two or even only 1 lock. In some cases passbooks were kept in the box and sometimes they were carried home. This should be consistent and we normally recommend that passbooks are held in the box to guard against loss and to prevent the unauthorised alteration of entries. Frequently, also, money was not kept in the box but at the home of

  • 35

    the Treasurer, which is a dangerous development. The improper use of member savings for private purposes is very common where cash is held outside the box. Cash handling in the meetings was good and most groups used the money-counting bowls to manage cash movement.

    5.1.3 Record-keeping

    Savings The general standard of record-keeping in this programme is good. Records of savings and records of social fund contributions (Mfuko wa Elimu na Jamii – education and family

    fund) are excellent, carefully maintained on yellow ledger-cards, with matching records entered (also carefully) in members passbooks. The forms that are used to record member savings and the passbooks are well-designed and appropriate. By using share stamps the members’ passbooks visually represent the totality of a member’s savings (especially useful for non-literate members, while the ledger does so using numerals. The ledger is a duplication (and probably not necessary – passbooks alone would be fine), but it is simple and well thought out. We did, however, note that for many members savings is irregular, with many interruptions. This may simply reflect the reality of living in an economy where income is irregular.

    We also noted that a duplicate set of savings records (seen in the last photograph of this page) are created in detail (and also carefully) for Plan but question the necessity for this. VSLA members should not have to do work that is not directly related to the operation of the VSLA. If called upon to do so, and seeing no value-added at the level of the VSLA, this is likely to decline in accuracy over time.

    Loans Where the record-keeping system is least inadequate is in recording loan liabilities. We noted the following:

    • Loans have to be applied for in writing and are not approved until the next meeting. The forms ask why a loan is needed, ho much is needed and for how long and requires the signature of a guarantor. It then needs the signature of the VSLA’s Secretary and the attachment of a detailed business plan, written on a standard form. The idea behind all this is good, but it makes for a cumbersome and bureaucratic process that is, in practice, neither carefully followed nor necessary. There is no reason to believe, for example, that a borrower wants money for business, or has the capacity to think through and articulate these issues. There is also no reason to believe that the average group member is in any position to analyse this type of application. It should be eliminated

  • 36

    • Members do not have their loans recorded in passbooks (although the passbooks permit it) but in a central ledger. The Secretary has to fill out a form in which the borrower states the plan for the reimbursement of a loan. This shows principal owing, the amount of interest to be paid monthly for three months and the total amount that has to be reimbursed.

    The second form is a record of what actually happened and mirrors the form on which the plan was developed. This is illustrated in Figure 6. The loan book has the virtue of simplicity. The initial loan plan lays out quite clearly what is borrowed, what will be charged in interest, the time period of the loan and the total to be repaid. It is limited because it does not allow for loans that are longer than 3 months (and many loans have been given that last longer) but is a clear indication of intention. Figure 6 below is a record of repayment and shows how the loan was repaid – exactly in accordance Figure 6: The loan book

    with the plan. The problem arises when the borrower deviates from the plan. If, for example, the loan lasts for longer than 3 months (a common occurrence) then the Secretary will be unsure of how to carry on. If a fine is levied more than once (if a loan interest repayment is missed a few times again the Secretary will be uncertain as to how to fill it out and we did see some individual adaptations of the model. The main problems with so many loan record-keeping system are three:

    • they try to allow for too many eventualities (such as fines) that can easily be recorded elsewhere

    • they only allow for three loan repayments: it is very often the case with VSLA that people take longer to repay than they plan and;

    • they use lateral records in which the arithmetic has to be done horizontally. Most of us learn to calculate laying out the figures in columns, not rows, to facilitate adding and subtraction.

    Having said this, the loan record book appears to work, with some occasional modification. Of all the systems that we saw in Malawi, Uganda and Tanzania, this was the simplest and most practical. We have appended to this report in Annex 5 an alternative system that can be used either in a ledger or a passbook and which addresses the problems listed above.

    .

  • 37

    Figure 7: VSLA financial report

    VSLA group-level financial report Figure 7 is a representation of half a page in a VSLA’s monthly report to UHIKI and was originally developed by CARE in its VSL work in Zanzibar. It is a summary of the group’s financial position. (Taarifa ya Fedha ya Kikundi cha Akiba na Mikopo – Financial report for a savings and credit association) and, on a single page, is intended to act as a monthly presentation of every week’s financial transactions. Taken by and large this is confusing, rarely accurate and, crucially not much used by VSLAs to manage their affairs. Rather, it is used by UHIKI, a small part of which feeds in to the MIS.

    Figure 7 shows data filled out up to the start of the second week of the month. It is not clear what purpose is served by this report. To fill it out is quite difficult. First it requires that income from fines, sale of forms, loans repaid, loans provided by UHIKI, social fund contributions, income from insurance payments, loan interest and donations are added together and loans disbursed deducted to obtain a cash balance. Yet there is no provision for expenses, no provision for other assets that the group may have purchased and no provision for debts (or receivables). Next, to fill this form the group has to keep track of fines and interest income, disaggregated from the total amount paid in loan reimbursement. This requires

  • 38

    that cash movement is recorded in a cash book and that interest is disaggregated from any loans repaid. All of this is possible (though complex) but it is also pointless because it is insufficient to prepare a balance sheet and insufficient to prepare a profit and loss account. No-one in the groups we met was able to explain what this form was for and were not able to say what specific information that it contained had actually helped them to manage their funds. It was something that Plan/CRP wanted. Without question, there is no value-added for anyone in creating this cumulative data, while adding a lot of complexity. It should be dropped.

    If a summary report is needed (and it very rarely is requested by VSLA members) a group should not need to calculate the inputs. When groups are asked to take primary information (such as individual savings), aggregate it (to total savings), transfer it somewhere else and use it as input to another table (cash book) it is common for errors to arise. They are best avoided by eliminating the need for transaction record-keeping and depending instead on a process of witness of what has happened. If this is done a group (or, better still, a FO) can develop a balance sheet comprised only of the following information:

    • Cash on hand in loan fund ...................................................................................... plus • Cash on hand in insurance and other funds ........................................................... plus • Total value of loan balances outstanding ................................................................ plus • Other assets (i.e. livestock, furnishings etc.) ........................................................ minus • Debts ................................................................................................................. equals • Equity .............................................. (made up of member savings and interest income)

    If this formula is used a FO can quickly calculate equity and, if he or she is interested, profit can be calculated simply by subtracting total savings from equity.

    Record-keeping - Conclusion Of all the systems reviewed in this series of evaluations, this one is the most practical, apart from the financial report, which serves no purpose at all. The loan record-keeping system could be improved to make it able to deal with contingencies that are quite common and it is questionable whether any of this needs to be done in a ledger. But if nothing is done the system will continue to work quite well.

    Having made due allowance for the complexity of the tools the record-keeping in UHIKI’s programme is consistent between groups and, critically, the savings records in the passbooks and ledgers are very well maintained. The single biggest difficulty over the years has been to identify a record-keeping system that is adapted to the skill levels of VSLAs (and CRPs), without sacrificing safety and, crucially, savings and loan flexibility. The consensus view in all Plan countries in West Africa, is that the standard VSL approach needs only to make use of simplified passbooks and recording ending balances, while doing away with complex transaction record-keeping and, critically, efforts to produce financial statements. There is usually a lot of resistance to this point of view.

    Natural prudence and professional training seems to indicate that ledger-based written records are an automatic pre-requisite. But they are not. Because groups are small and because groups observe all transactions, there is no need for laborious recording. By witnessing transactions, deeming that they are fair and that the ending cash result is reached logically, legally and in accordance with procedure groups don’t feel the need to write it all down – except in terms of personal savings and personal loan liabilities, recorded in passbooks or, if a backup system is felt to be essential, recorded in exactly the same form in a notebook. This is what is done in all 9 Plan countries in West Africa. In addition to the passbook entries, a separate recording of ending loan fund and social fund cash balances is made in a small note book. The following advantages of this system are: • Meetings take less time

  • 39

    • The system is well adapted to the skill levels of most members and Record-keepers • Training time is significantly shortened • Supervision is simpler • Everyone gets to see what they have saved and what they owe at every meeting • Transparency is enhanced (records are in each member’s hands, and not confined to

    a technical elite who may, or may not (usually the latter) regularly inform the members about their financial state of affairs.

    • Written record-keeping may militate against regular committee turnover, because the skills needed to maintain written records may be scarce

    Thus, it is recommended that Plan Tanzania, at the least, simplifies the ledger-based record-keeping system and adopts a passbook-based system, using the reformed loan record-keeping system shown in Annex 4. No changes are needed to the share section of the passbook.

  • 40

    5.2 Financial performance To determine the financial performance of the groups we visited, and the performance of the entire portfolio, we turned to the MIS. and reviewed the results from Kisarawe. Figure 8: Programme performance reported in Kisarawe December 2008

  • 41

    It is clear from reviewing the results that the return on savings is, at 1.7%, extremely low. When we looked closely at the FO reports we discovered that two of the three FOs were mis-reporting savings data. It appeared that they had derived the value of savings from a calculation that added together cash and loans outstanding instead of counting actual savings. This over-counts savings because it incorporates interest income and means that the MIS financial results do not have validity, if only because they do not measure profitability or return on savings. This can clearly be seen in Figure 9 below, which shows part of a FO’s MIS results. This indicates that all but two VSLAs have made neither a profit nor a loss (in addition, Figure 9 shows that no-one has dropped out, which is highly implausible). Figure 9: Part of one FO’s portfolio

    Clearly, there is a problem relating to the FO’s understanding of the data going in to the MIS. The same problem was noted in Plan Malawi. Subsequent work with CARE Tanzania in 2009 has shown that they are committing the same error and have evidently passed it on to Plan

  • 42

    Tanzania. It is remarkable, however, that one of the three FOs (Shadrack) seems to have entered correct data as per figure 10 and achieved a return on savings of 18%. Since one Field Offi8cer seems to have done it correctly and two have not done so, it would appear that it is left to FOs to define the data for themselves and no clear guidance given. Figure 10: Part of Shadrack’s portfolio

    As a result, we selected groups for interview that seemed to be profitable, and carried out a review of their financial situation. Annex 2 lists 39 groups from which we selected 12 for visits. The initial 39 were chosen because they were groups having a meeting on the days of our proposed visit and the 12 we finally selected were those who were carrying out loan transactions (repayment and disbursement) on the day in question.

    We gathered data from the groups visited and carried out in-depth enquiries about the current state of group assets and liabilities. What we found bore very little relationship to the data entered in the MIS. Figure 11 on the following page compares results derived from the MIS with results measured by the evaluators. The first part of Figure 11 is a comparison of performance of the random sample of 12 VSLAs. Bearing in mind that

  • 43

    most MIS data was 3-4 months old, discrepancies between the results are likely to be the result of the passage of time. The increase in savings per member and the small shift from a growth in membership to a retention rate of only 95% are cases in point. The one disturbing difference, however, is the very sharp drop in member equity, which imputes a loss of $25.3 per member rather than a profit of $21.1, as reported by the MIS. Figure 11: Performance December 2008

  • 44

    We were careful to question the groups about the possibility that surplus funds may have been invested in other activities, or that an informal interim share-out had taken place, but were assured that nothing of these sorts had happened. The results, then. speak for themselves. Figure 12 is a detailed listing of assets and liabilities, disaggregated by group and show that only 4 of the 12 groups selected may be making a profit, and 7 a loss. Figure 12: Detailed listing of assets, liabilities and equity, by group interviewed.

    It is entirely possible that surplus funds have been legitimately invested elsewhere or distributed. But it is also possible that the money has been misused. The point is that no-one knows and, more to the point, data collected to date by the FOs and community-based trainers does not show this.

    The MIS is also being implemented in its most complex form. This was originally designed for programmes that wanted to maintain information on a standard financial measure called portfolio at risk (PAR). This requires staff to collect additional data to the standard 19 data sets that apply to the basic MIS version. These additional pieces of data are:

  • 45

    • Cumulative number of loans since start of cycle • Cumulative value of loans since start of cycle • Total value of loans past due • Value of loans written off at share-out

    All of these are difficult data to gather, but Total value of loans past due is particularly hard and, with VSL hardly relevant, since loans are not expected to be paid on a fixed schedule, but only within a given period of time. Thus, we strongly recommend that the basic version is used, which requires a checkmark on SETUP page of MIS 2.13 (the current version) to be removed as follows.

    This will render the data easier to collect and more accurate. The MIS, then, is neither accurate nor properly understood and it does not provide a reliable picture of the reality in the field. All it has been able to do to date is to approximately indicate programme scale and, in this evaluation, to suggest that there may be a problem of loss of equity, which urgently needs to be addressed. . Retraining and simplification is, then, an urgent necessity if the MISD is