Lessons Learned from Ongoing Social Cash Transfer Programmes in Zimbabwe

Transcription

1 Lessons Learned from Ongoing Social Cash Transfer Programmes in Zimbabwe Ministry of Labour and Social Services March 2011 Ministry of Labour and Social Services

2

3 Lessons Learned from Ongoing Social Cash Transfer Programmes in Zimbabwe March 2011 Prepared on behalf of Team Consult for UNICEF by: Bernd Schubert In cooperation with: Officers of the Ministry of Labour and Social Services on national and district level Officers of UNICEF Zimbabwe Harare, November 2010 Ministry of Labour and Social Services

4 ii Acknowledgments The authors acknowledge that information was generously provided by all organizations contacted. Special thanks go to Mrs. Kerina Zvobgo, the Chair of the Cash Transfer Working Group. The report as also based on the following recent publications on cash transfer programmes in Zimbabwe: Justine Smith (2009), Cash transfers as a humanitarian and development programming tool in Zimbabwe. Harare Fabienne Wydler, and Ebbie Dengu, (2009), Assessment of appropriate social transfer mechanisms in Zimbabwe, Harare Acknowledgments

5 1 Contents Acknowledgements ii Background to the Report Introduction Types of ongoing cash transfer programmes Emergency aid - drought relief - public works cash transfer programmes Livelihood support programmes Contents 2.3 Categorical social cash transfer programmes Social cash transfers to labour constrained food poor households Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe Target groups and targeting criteria Targeting mechanisms Volume and frequency of transfers Delivery mechanisms Costs effectiveness Impact Concerns, risks and challenges Tentative conclusions on implications for cash transfer programming References

6 2 Annexes 1. Social Cash Transfer Programmes in Zimbabwe (Source; Kerina Zvobgo) Review of the Pilot Project to Support OVCs in Zimbabwe Review of the Zimbabwe Emergency Cash Transfer Pilot Programme (ZECT) Definitions of OVC Application form for DSS Welfare Programmes Contents 6. Effectiveness of Targeting of PRP Programmes (3 selected slides from a PRP Power Point Presentation) Some Advantages and Disadvantages of Different Forms of Delivery Mechanism Concerns about using Cash as a Programming Tool

7 3 Background to the Report The first phase of the Government of Zimbabwe's National Action Plan for Orphans and Vulnerable Children (NAP I) was approved in It aimed to address the needs of the large numbers of orphans and other vulnerable children in Zimbabwe and was operational up to the end of In 2006 a Programme of Support (PoS) was established to coordinate and scale-up international support for implementation of NAP I. During 2010 an independent review of PoS against OECD evaluation criteria found the programme to be relevant, efficient and effective however noted that there was potential for mistargeting since poverty criteria had been excluded from beneficiary selection. The key lessons learned from PoS included the following: Targeting vulnerable children and their households together is likely to be more effective than focusing on children alone A comprehensive approach with effective follow-up is needed to target the range of deprivations faced by orphans and vulnerable children Household poverty is a major cause of child vulnerability in Zimbabwe Investments in government capacity building are essential to deliver national quality results Background to the Report Within the changing context of widespread vulnerability in Zimbabwe a shift was recorded from targeting on the basis of HIV and AIDS, which was the foundation for NAP I and PoS towards scale up of social protection systems which are child and HIV sensitive, rather than HIV focused. This meant including poverty as a determinant of child vulnerability. Thus the revised NAP II describes a child sensitive social protection framework for Zimbabwe including strengthening household economy as one of its key pillars. The Child Protection Fund in support of NAP II (the successor to PoS) reflects this by aiming to reduce household poverty of 55,000 extremely poor households by implementing a national cash transfer mechanism, as well as enhancing all vulnerable children's access to social welfare and justice and other basic social services. The design of NAP II and CPF was informed by a critical research paper, "Child Sensitive Social Protection Thought Paper", commissioned by UNICEF in 2010 in order to: Clarify the concept and the terminology related to poverty, vulnerability and child-sensitive social protection Briefly analyse the social protection needs of children and other vulnerable groups and households in Zimbabwe Provide an overview of ongoing social protection interventions in Zimbabwe in order to identify the main gaps between needs and services provided Provide recommendations on how the upcoming social protection strategy, the future Programme of Support and other programmes managed by UNICEF and by others could contribute to effective child-sensitive social protection

8 4 Elaborate guidelines for targeting the social protection interventions of the Programme of Support to the National Action Plan for OVC (the guidelines are provided in a separate report) This report is part of a series of eight design reports and manuals which collectively provide background detail and operational guidance for the Government of Zimbabwe Harmonised Social Cash Transfer Programme. The first design report "Lessons Learned from Ongoing Social Cash Transfer Programmes in Zimbabwe", summarises lessons learned from past and ongoing cash transfer programmes implemented by Government and by Non-State-Actors, and draws conclusions for the design of a national Government owned social cash transfer program as part of the revised National Action Plan for Orphans and Vulnerable Children (NAP for OVC) as well as broader DSS social protection programming. Background to the Report The second design report "Process and Results of Baseline Survey of Labour Constrained, Extremely Poor Households in Goromonzi, Ward 5", details the results of a baseline household survey simultaneously conducted in Ward 5 of Goromonzi District in Zimbabwe to inform the design of a targeting mechanism for the cash transfer scheme. This provides detail on (a) the test of potential targeting methods, and (b) the identification of target groups for the cash transfer scheme, in order to determine a way forward in design. The third design report "Review of Cash Transfer Databases and Management Information Systems (MIS) In Zimbabwe" assesses Management Information Systems (MIS) and databases of ongoing cash transfer programs and focuses on the capacity and potential of existing database systems and the possibility of using them as basis for refining the requirements of a scalable and robust MIS for cash transfer. The fourth design report provides a detailed "Manual of Operations for the Zimbabwe Harmonised Social Cash Transfer Programme", based on outcomes from the test phase January-March The fifth design report on the "Test Run of the Zimbabwe Harmonized Social Cash Transfer Programme in Ward 15 Goromonzi District", describes the full implementation test of the whole targeting, approval, registration and payment cycle as specified in the draft Manual of Operations was implemented in Ward 15 of Goromonzi District, including lessons learned for revising the Manual of Operations. The sixth design report provides detail on the parameters for design of the social cash transfer programme including harmonisation with other social protection programmes. The seventh design report is the guidance for internal monitoring. The Internal Monitoring System activities are designed to provide timely information that is tailored to the requirements of decision makers. The main users of this information are all stakeholders involved in implementing, assisting and co-funding the programme, namely officers of the Department of Social Services (DSS) at national, provincial and district level, UNICEF and cooperating partners funding the Child Protection Fund. The eighth and final design report is the "Operational Guide for the Management Information System of the Zimbabwe Harmonised Social Cash Transfer".

9 5 1 Introduction

10 6 1. Introduction The Ministry of Labour and Social Services (MoLSS), Department of Social Services (DSS), is in the process of designing a social cash transfer programme for labour constrained extremely poor households. The design will be partly based on lessons learned from past and ongoing cash transfer programmes implemented by Government and by Non-State-Actors. The objective of this report is to summarise lessons learned and to draw conclusions for the design of a national Government owned social cash transfer programme as part of the revised National Action Plan for Orphans and Vulnerable Children (NAP for OVC) as well as broader DSS social protection programming. Introduction The work is being technically and financially supported by UNICEF. During UNICEF managed a pooled donor fund in support of implementation of the NAP for OVC. Known as the Programme of Support (PoS) to the NAP for OVC, this mechanism was independently reviewed during 2010 and a new proposal designed for a national child sensitive social protection and child protection programme. A key programme pillar of the NAP for OVC II and its accompanying second phase of implementation, the Child Protection Fund (CPF) will be cash transfers to the poorest and most vulnerable families. Team Consult was recruited by the MoLSS and UNICEF to design the cash transfer component of the NAP II implementation. The primary purpose of the assignment is to support the Ministry of Labour and Social Services' (MoLSS) Department of Social Services (DSS) to design a national government owned and coordinated cash transfer programme which targets ultra- poor and labour constrained households with clear costing, targeting, monitoring and evaluation and implementation strategy. The report is based on reports of and interviews with DSS as well as with some of the more than 20 NGOs implementing cash transfer programmes in Zimbabwe (see Annex 1). Most of the programmes are small and have started only recently. Altogether DSS welfare programmes reach 4,104 households while Non-State-Actors reach 37,613. The report starts out with a classification of the different types of ongoing cash transfer programmes. Each type is briefly outlined (Chapter 2). A more detailed analysis of two of the bigger programmes is given in Annexes 2 and 3. Chapter 3 summarises lessons learned with regard to: Target groups, targeting criteria and targeting mechanisms Volume and frequency of transfers Delivery mechanisms Cost effectiveness Concerns, risks and challenges Chapter 4 draws tentative conclusions with regard to implications of the lessons learned for the design of a national social cash transfer programme for labour constrained extremely poor households.

11 7 2 Types of ongoing cash transfer programmes

12 8 2. Types of ongoing cash transfer programmes 2.1 Emergency aid - drought relief - public works cash transfer programmes Simultaneous with the economic decline of the last decade emergency assistance, predominantly in the form of food aid, has increased. In response to the drought in 2008 the two main emergency food aid organizations (WFP and C-safe) together provided aid to more than 5 million people. Types of ongoing cash transfer programmes Food aid is based on annual ZimVAC vulnerability assessments of the food deficits of each Ward in the country. Food insecure Wards are allocated a quota indicating the percentage of people that can be provided with food aid. Administration and delivery is organised through NGOs and private sector transporters. Household targeting is done on village level by mechanisms like poverty mapping and wealth ranking involving the whole community. A more detailed analysis of one of the World Food Programme (WFP) pilot cash transfer programme implemented by Concern Worldwide is given in Annex 3. The food rations are provided on a monthly basis from October to March. They consist of 10 kg maize, 1 kg beans and 0.6 litres vegetable oil. This ration provides 1,700 Kcal food energy per person per day. In March 2010 prices, the market value of a ration was approximately USD 6. It is estimates that the market value of the ration represents about half of the total costs of the programmes. Some of the emergency assistance is provided in the form of Food for Work or Cash for Work again mainly channelled through NGOs. A recent study commissioned by the MDTF concludes: Today the use of public works approaches within the humanitarian sector is not coordinated and erratic and the states' own public works initiative has been grounded by loss of manpower, skills, transparency and adequate financial and logistical resources. (Wallace-Karenga 2009) Some of these programmes provide free food distribution for households with no adult labour capacity and food for work for households with adult labour capacity. In view of the high transfer costs of food aid, a number of pilot projects are experimenting with different forms of cash transfers. Most of these programmes simply replace the food rations by an amount of cash equivalent to the value of the ration at current market prices which are recalculated every month. In practice that means that beneficiaries of the pilot scheme in Gokwe (WFP financed and implemented by Concern Worldwide) received USD 5.21 per person in November, USD 5.22 in December and USD 6.42 in January.

13 9 As pilot activities in the area of cash in emergencies have started only recently, the following monitoring and evaluation results are still preliminary: Transfer costs for cash transfers are significantly lower compared to transfers in kind Markets respond effectively: Traders welcome the additional purchasing power resulting in increased business turnover and stock up to meet the demand of cash recipients Pilots use different delivery mechanism like direct delivery by the NGOs or hiring specialised private companies. So far no security problems have been experienced Given the background of hyperinflation and initial problems experienced during the transition to US Dollars, beneficiaries initially seemed to prefer food rations or a mixture of food and cash but are now increasingly realising the advantage of cash There are concerns that cash transfers being more attractive compared to food may lead to increased interference into the targeting. Up to now there is no documented evidence with regard to interferences. This risk will have to be closely monitored. In summary the preliminary monitoring and evaluation results of the cash in emergency pilots are positive. Types of ongoing cash transfer programmes In terms of social protection the emergency food aid programmes ensured the survival of millions of extremely needy people and will still be required - hopefully at decreasing levels to combat nutritional deprivations of many children. If progressively transformed from food transfers to cash transfers they could in future reduce transaction costs, avoid the risk of disrupting markets and stimulate local economic activities. By broadening their objectives from mainly smoothing food consumption to supporting livelihoods they could reduce the danger of creating a dependency syndrome. However, for labour constrained ultra poor households and for the children living in these households, emergency aid in food or cash is not an appropriate tool for providing social protection. While these households require continuous, reliable and predictable transfers, emergency aid is unpredictable (will their Ward be declared food insecure or not) and is only available from October to March. 2.2 Livelihood support programmes Livelihood support programmes provide inputs, voucher or cash to poor households in order to empower them to work themselves out of poverty. While emergency aid is short term, livelihood assistance is given for one

14 10 or two years after which beneficiary households are supposed to graduate. The following paragraph summarises livelihood assistance provided by the Protracted Relief Programme II: PRP II is targeting households both in rural and urban settings and aims to reach more than 2 million people by improving their livelihoods. Components of PRP II include increased food production through provision of agricultural inputs, promotion of approaches including conservation farming, improved farmer extension methodologies, provision of small livestock, increased access to potable water and sanitation, improved hygiene practice, promotion of improved nutrition through small vegetable gardens, home based care, and provision of social transfers (PRP II, 2009). Types of ongoing cash transfer programmes PRP II is a multi-donor, multi-sector fund coordinating 21 NGO implementing partners who work with 30 local partners. PRP is managed by MTLC (Managing, Technical Learning coordination) a project of the private company GRM International. Many of the implementing NGOs are also implementing partners of PoS. Livelihood programmes are not suitable for labour constrained households. NGOs implementing livelihood programmes either do not register labour constrained households or - in case they accept them as beneficiaries - are faced with the fact that these households are unable to graduate. They would like to refer them to Government welfare programmes as soon as these programmes have the capacity to absorb them. 2.3 Categorical social cash transfer programmes Categorical social cash transfer programmes target individuals like those targeted at orphans and vulnerable children (currently defined differently in different programmes in Zimbabwe), elderly, or disabled persons. Some of these programmes just target the individual beneficiary disregarding the household in which he or she lives (e.g. the DSS elderly maintenance reaches 3,070 elderly while the disabled maintenance programme reaches 2,000 beneficiaries). Others target households in which elderly or OVC or disabled persons live. All these programmes target individuals or households that are extremely poor using different tools for poverty targeting. Help Age Zimbabwe (HAZ) runs a small pilot for 180 beneficiaries that are older than 80 years. The Christian Relief Service (CRS) and the Joint Initiative (JI) implement pilots that do not exclusively target extremely poor labour constrained households but include them in their programmes. The dilemma of these pilots is that: Due to the limitations of programme

15 11 funding cycles, NGOs are not able to offer long-term assistance to those households among the chronically poor that may continue to need direct assistance, for example because of high dependency ratios, chronic illness or disability. This is the role of Government, who are ultimately responsible for providing long-term, reliable and predictable social welfare assistance to the chronically poor. As such the exit strategy for NGO assistance for this category is referrals to the Department of Social Services, under the government's wider Social Protection Framework. However, it is also important to acknowledge that in the short and medium term, government systems in Zimbabwe are unlikely to have the capacity to absorb all these additional households (Smith 2009). In the context of categorical programmes it is worthwhile to also analyze the Basic Education Assistance Module (BEAM). BEAM does not transfer cash to individuals or households but pays school fees and school levies for primary school children living in very poor households. School fees are paid for "vulnerable children" (as defined in the BEAM Manual 2009) directly to more than 5,000 primary schools by UNICEF with donor funding and to secondary schools through complementary Government funding. Donor funding is channeled through the current PoS. School Development Committees organise a participatory targeting process and control school attendance of beneficiaries. Types of ongoing cash transfer programmes The revised NAP II restates the importance of BEAM to support children's access to school and includes support to education (BEAM) and health (Assisted Medical Treatment Orders AMTOs). Donor funding for BEAM, however, is unlikely to continue beyond UNICEF also coordinates, together with the Ministry of Education, an Education Transition Fund which has provided a full set of textbooks to every child in the country as a complementary initiative. BEAM aims to reach 560,000 primary school children which otherwise would not have been able to attend school and a total of 243,000 secondary school children. Compared to a total of approximately 3.6 million school age children of whom 3,277,000 were enrolled in 2006 (Gandura 2009), BEAM will cover 22%. Considering that more than half of the population in Zimbabwe is living in extreme poverty (MDTF 2010), the planned coverage of 16% is still low. BEAM is regarded as an innovative transitional funding model. Complemented by UNICEF programmes that provide textbooks, stationary and learning material or by block grants to schools, BEAM could significantly reduce educational deprivations. Problematic is the fact that school fees are a standalone intervention and that conditionality is imposed which could lead to situations where children from the poorest households are excluded because they are unable to meet the condition of regular attendance as their caregivers e.g. are unable to afford school uniforms or even shoes and soap.

16 12 The process of BEAM's revitalization and potential success can be interpreted as a movement towards Government - led social transfer schemes at district and ward level. While initially educational funding was channeled through NGOs, the funds for the 3rd term 2009 and for all terms 2010 (costed at USD 20 million in the PoS budget) were delivered to BEAM because this Government run scheme was evaluated as both, more cost-effective and able to target more children. The BEAM will be evaluated towards the end of Social cash transfers to labour constrained food poor households Types of ongoing cash transfer programmes It is estimated that 250,000 households in Zimbabwe are labour constrained and at the same time food poor (Schubert 2010, MDTF 2010). Labour constrained means that they have no household member in the working age (19 to 60) who is fit for productive work or that they Graph 1: Members of 111 labour constrained and at the same time food poor households in Ward 5 of Goromonzi District Source: Team Consult baseline survey November 2010 Adults not fit for work (19 t0 64) 67.0 Adults fit for work (19 to 64) Chronocally ill persons Disabled Adults Double Orphans Paternal Orphans Maternal Orphans 31.0 Orphans (all categories) Children (below 19) Elderly (65+) 72.0 Female HH Members Male HH Members Average Size of HHs 4.6 Total Number of HH members Percentages given for the share of different population groups in the total population of Zimbabwe are quoted from MDTF (2010)

17 13 have a dependency ratio of more than 3. Members of labour constrained households are either too young to work (below 18 years) or too old (60+) or severely disabled or chronically sick. Graph 1 shows the composition of 111 labour constrained and at the same time food poor households in Ward 5 of Goromonzi District. Members of these households are predominantly children (67% of all household members) of whom the majority are orphans (66%) while the average incidence of orphans among all children in Zimbabwe is 22%. 72 (14%) are elderly while the average percentage of elderly in the population of Zimbabwe is 3.9%. 46 (9%) are chronically ill. Nearly all household members belong to the so called vulnerable groups. Only 32 (6%) are working age adults fit for productive work while the share of the population aged 15 to 64 in the total population is A further analysis of the labour constrained food poor households in Ward 5 of Goromonzi District is given in Team Consult Report 2. Except for the DSS welfare programme for 'destitute households' there is no other programme targeting these households. Extremely poor labour constrained households are neither able to access labour based social protection programmes (like public works or livelihood programmes) nor can their needs be met by the short term transfer programmes offered by NGOs, while government social welfare programmes do not have the capacity to absorb them. The DSS 'destitute households' programme is presently reaching only 4,140 households with monthly payments of USD 20. Households are self identified and only those able to report can access benefits (if funding is available). This means that many labour constrained extremely poor households, the neediest of all households, fall through all safety nets. This is one of the most pressing unsolved problems of social protection in Zimbabwe. The hardest to reach households are arguably the most important to invest in for Government and human rights organizations in equity terms as well as for their potential to contribute to the alleviation of extreme poverty. Types of ongoing cash transfer programmes

18 14 3 Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe

19 15 3. Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe 3.1 Target groups and targeting criteria While the target groups of cash of emergency programmes are 'food insecure households' in drought prone Wards that are annually identified by the Zimbabwe Vulnerability Assessment Committee (ZimVac), nearly all other programmes target 'vulnerable households' as such or households that contain a child or an orphan or an "OVC" and are at the same time "vulnerable". Often the fact that a household cares for one or more orphans is used as a criterion for vulnerability of the household. For example the Pilot Project to Support Orphans and Vulnerable Children (OVCs) in Zimbabwe implemented by CRS defines its target group as follows: Households will be defined as vulnerable and thereby eligible for inclusion in the programme if there are children under the age of 18 years resident in the household and the household satisfies one or more of the following conditions: a. The household is in the poorest quintile of households (20%) at baseline. b. The household contains one or more orphans at baseline c. The household head is under the age of 18 years at project inception d. The household contains a chronically ill member e. The household contains a disabled member. (Catholic Relief Service 2009) Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe Using these criteria 4,116 households (24% of the 16,887 households living in the programme area) have been identified as eligible at baseline. Household vulnerability criteria of ZimVac are: a. Households with orphans b. Households with chronically ill persons c. Households with physically or mentally challenged persons Interestingly, the 2010 National Nutrition Survey does not link orphanhood with poor nutrition status (Food and Nutritional Council, 2010). Annex 4 gives the definitions of OVC used by the National Plan of Action

20 16 for OVC (NAP), the Programme of Support for the NAP (PoS) the Multiple Indicator Monitoring Survey (MIMS) and by BEAM. The list of definitions shows that all the programmes and surveys struggle with the task to define the terms 'vulnerable' or 'OVC'. The definitions are just long lists of examples and do - when used for analysis, programming and targeting - result in substantial inclusion and exclusion errors. Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe In addition to fuzzy definitions of vulnerability discussed above, there is also the problem that caring for orphans is by no means a valid indicator for poverty or food insecurity (see Table 1). A study based on 60 national household surveys in 36 countries concludes: Results indicate that orphanhood and co-residence with a chronically ill or HIV-positive adult are not universally robust measures of child vulnerability across national and epidemic contexts. For wasting, early sexual debut, and to a lesser extent, school attendance, in the majority of surveys analyzed, there were few significant differences between orphans and non-orphans or children living with chronically ill or HIV-positive adults and children not living with chronically ill or HIV-positive adults. (Akwara et.al. 2010) Table 1: Share of households with orphans that are in the lowest income, expenditure, assets, and food security quartile calculated from ZIMVAC 2006 Source: Miller, C. Zimbabwe child and family vulnerability analysis 2007 Households with orphans (percent in category) Lowest income quartile 27.1% Lowest expenditure quartile 26.0% Lowest asset quartile 15.7% Food insecure 23.5% In summary: Most cash transfer programmes suffer from vaguely defined target groups and use eligibility criteria which may not necessarily be appropriate for identifying the neediest households 3.2 Targeting mechanisms Basically three mechanisms are used for the actual targeting: a) Government welfare programmes for the elderly, disabled and destitute use an administrative process where needy individuals or households have to file applications to the District Social Welfare officers.

21 17 b) Nearly all other programmes use community based targeting mechanisms c) The CRS pilot programme in Manicaland is using a census of all households in their programme areas as the basis for computer aided targeting a) Administrative targeting by the welfare administration To apply for one of the three Government Welfare Programmes an applicant has to fill in an application (see Annex 6) accompanied a copy of the Identity Card and by a letter of the Counselor of his or her Ward confirming the residence. The Welfare Officer is supposed to verify the information provided by the applicant. This is usually done in the district office because household visits for verification are mostly not possible due to lack of transport. Once the verification is completed, the applications are forwarded to the provincial office for approval and then passed on to the national level. Graph 2: DSS Social Welfare Schemes Operations Flow Chart Start Applicants apply for welfare support at the District. Application vetted by DSSO Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe Means test criteria met? Verify the accuracy and consistency of the data details at the community level End Completed application forms including banking details signed by DSSO and sent to provincial level Payroll of beneficiaries POSB pays into beneficiaries accounts Details checked and approved by PSSO and sent to national level Details verified at DSS and complete forms sent to Central Computing for data entry Pay sheet printed at Central Computing at Ministry of Finance and forwarded to DSS Payroll of beneficiaries approved and forwarded to POSB

22 18 In fact this targeting mechanism is similar to the procedures used in the different welfare programmes in South Africa. Evaluations of the quality of this targeting mechanism are not available. The fact that all welfare programmes together reach only 9,210 beneficiaries indicates that there is an enormous 'exclusion error'. However, this is not due to the targeting mechanism but to a lack of Government funds allocated for social welfare. b) Community based targeting Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe Nearly all the other cash transfer programmes use different forms of community based targeting using a variety of tools like community meetings for awareness raising, community committees, poverty mapping on village level, and wealth ranking exercises. They also verify the lists received as a result of community targeting by organizing household visits and finally set up help desks to respond to complaints of community members who are not satisfied with the results of the targeting process. In summary: Most programmes go out of their way to ensure that the principles of transparency, participation and community ownership complemented by verification exercises are used in every possible manner in order to ensure that inclusion errors and exclusion errors are as low as possible (high effectiveness of targeting). Unfortunately nearly all available evaluations and appraisals of community based targeting in Zimbabwe indicate that the results of community based targeting are far from satisfactory. For example the assessment of inclusion errors and exclusion errors given in the July 2010 evaluation report of the ZECT pilot by Oxford Policy Management states (for more information on ZECT targeting see Annex 3): Fieldwork indicated that inclusion and exclusion errors are reasonably common in current targeting. As an indication of this, the headman of every village visited was a recipient, and while some may have been food insecure, not all were. Inclusion errors are not unexpected in the present community based targeting system which relies on community members to nominate and confirm vulnerable households, using a series of categories suggested by Concern. The major weakness in this system is that in a communal setting, it is very hard for marginalised households to speak freely and relatively easy for powerful households to secure nominations and validations from those in their social network (Kardan et.al. 2010). This is not an exception. Annex 5 provides graphs of the effectiveness of targeting of a number of schemes under the Protracted Relief Programme II. The graphs show that beneficiary households are more or less evenly spread over all wealth groups. If poverty targeting has been

23 19 intended it has only partly been achieved. Similar opinions have been voiced by nearly all respondents interviewed. In summary: Poverty targeting is an important concern of programme officers in all organizations. At the same time there seems to be a widely shared experience that the results of community based poverty targeting are not satisfactory. It seems that shifting the responsibility for targeting to the community not only leads to high exclusion and inclusion errors but also creates conflicts in the communities. In interviews with NGO personnel a community member was quoted as stating: "You make us hate each other!" If this already happens in the context of food transfers it may be worse in the context of cash transfers because cash is more attractive and will not easily be shared between beneficiaries and nonbeneficiaries. c) Computer aided targeting based on census data The CRS implemented Pilot Project to Support OVCs in Zimbabwe (for details see Annex 2) conducted a baseline survey of all the 16,887 households in the programme area. They practically did a census. Enumerators were hired and trained for the purpose of household verification. This process was conducted with active participation of the community and all the partners. The findings from this process were handed to the Biomedical Research and Training Institute (BRTI) who then designed a computer system to enter them, a process which made the selection process systematic with less interference and bias. These lists of beneficiaries were taken back to the community for final verification. This has helped in coming up with the most deserving beneficiaries lists for all the sites. Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe In fact the computerization made it possible to analyze how many beneficiary households would result from using different combinations of targeting criteria (see graph below). As a result of this analysis the programme decided to use the targeting criteria listed above which resulted in having 4,116 beneficiary households (including the control group). In this way the programme managed to match the number of beneficiaries to the amount of funds available to the programme. The fact that only 28% of the eligible households belonged to the poorest quintile indicates that poverty targeting was not effective (if at all intended). It also shows that using the criteria given above (containing an orphan or a chronically ill member or a disabled member) is not appropriate for poverty targeting. In summary, computer targeting based on baseline data from all households in the programme area is an interesting alternative to other targeting mechanisms. However, if the targeting criteria used are not appropriate, the results will still not be convincing.

24 20 Manicaland Cash Transfer Trial Households Enumerated & Found Eligible* Target 4269 All HH HH with children Eligibleoriginal criteria Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe Eligiblerevised criteria Contains a disabled member Contains a chronically ill member HH head is under 18 years Conatins one or more orphans Poorest quintile of HH % Number of Households * In both the BRTI census & the DOM CCP community verification exercise, the greater than expected number of households eligible meeting the original criteria was due to more than expected households overall - following earlier refusals due to Satanist fears Computerization also facilitated the following breakdown of the 4116 eligible households: 5% 28% 48% 69% 0% 10% 20% 30% 40% 50% 60% 70% 80% 3.3 Volume and frequency of transfers While the Government welfare programmes pay USD 20 per beneficiary or beneficiary household. Government public works programmes also pay USD 20 per month for 15 days of work. Programmes of Non-State- Actors pay between USD 15 and USD 25 per month (see Annex 1). Some differentiate according to household size. Others pay a flat rate. The rationae for the Government to pay USD 20 is that this is consistent with the current Food Poverty Line and is also the amount paid to the lowest category of pensioners. For the ZECT pilot, transfers were

25 21 calculated per person and were based on the costs of a food ration (10 kg of maize grain, 1kg of beans, and 0.6kg of oil. As prices varied by region and by month, ZECT paid different amounts in different districts and in different months (see Table 2). While most programmes including the Government programmes pay monthly (if funds are available) other programmes like CRS pay every second month. Bi-monthly payments are more cost effective. Table 2: ZECT transfers per person (in USD) Location Nov 2009 Dec 2009 Jan 2010 Feb 2010 Mar 2010 Nyanga Gokwe Delivery mechanisms While Government is using the Post Office Savings Bank (POSB) and the Central African Building Society (all beneficiaries are required to open POSB or CAPS accounts), most of the other programmes pay directly to beneficiaries (see Annex 1). Other transfer mechanisms are: Cash in Transit (CIT) implemented by private companies like Safeguard or FAWCET. These companies transport the money in armored cars from the bank to the pay points. While CRS uses FAWCET just for transport and uses their own staff to pay directly to the beneficiaries, Concern has outsourced the whole range of activities from packing envelopes for each beneficiary, transporting to pay points and paying to beneficiaries, to Safeguard. Swipe cards issued by large retail chains like SPAR, OK and TIM are used by number of urban cash transfer programmes. Some of these organizations insist that beneficiaries use at least 50 % of the transfers for purchases at their outlets. E-vouchers in the form of electronic scratch cards are also offered by chain stores like SPAR and OK. WFP uses this system managed by Redan Mobile Transfers (RMT) Few small programmes use bank accounts (TBD and CABS) In addition an increasing number of banks (like TN Bank offering cash cards), cell phone companies (like ECONET offering cell phone based transfers) and other IT companies like Redan Mobile (offering e- vouchers) are developing or already offering IT based cash transfer services. Annex 6 gives a systematic review of the pros and cons of all the different transfer mechanisms. In summary it seems that Cash in Transfer (CIT) is still the most cost effective (about 1 Dollar per transfer), client friendly (pay points close to beneficiaries residence) and developmental (bringing cash into the rural economy where beneficiaries spend it in their communities) delivery option for rural area especially as two thirds of the rural areas still do not Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe

26 22 have internet access. Those programmes that use CIT are satisfied with their performance and are not reporting any serious problems. E- vouchers like the ones used by WFP/RMT might be an alternative for regions with satellite coverage, if pay points with sufficient liquidity can be established in each Ward. For urban areas the number of options for cash transfer delivery mechanisms is increasing and it is still an open question which of them is the most appropriate. 3.5 Costs effectiveness Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe One of the most surprising experiences of this rapid survey is the fact that it is difficult to find data on operational costs defined as the percentage of total programme funds that are not passed on to the beneficiaries but are required to organise all the activities of the implementing agencies from targeting to delivery including monitoring and all overheads. One would expect that pilot programmes that have the objectives to identify costs-effective ways of organizing social cash transfers would have cost data at their finger tips - but most have not. In spite of elaborate M+E systems, data bases, MIS and backstopping by scientific institutes - costs remain unknown. Even the two reviews of cash transfer mechanisms in Zimbabwe (Wydler and Dengu, 2009 and Smith, 2009) remain silent on costs. An exception is the Concern ZECT pilot which reports the following cost structure: Under the cash only aid modality 75% of total costs were due to the value of the cash transfer itself (USD 7.23 per person per month). Operational costs (USD 2.43 per beneficiary per month) include WFP management fees (6.5%), administrative and operational costs (6.2%), Concern management fees (4.2%) and Concern staff costs (3.9%) (Kardan et.al.2010). ZECT also compared the operational costs per transfer per person for three different types of transfers: For food in kind it was USD 4.98, for half food/half cash USD 4.13, for all cash it was USD In percent of total costs operational costs were 52.65% for food aid, 42.61% for food plus cash and 25.15% for cash. A more detailed analysis is given in Table 3. The message of this analysis is loud and clear: Operational costs for cash transfer are significantly lower compared to transfers in kind. 3.5 Impact The context of utilization of this review is not to inform policy decisions if social cash transfers should be introduced or not. This decision has already been taken. The review therefore concentrates on information that facilitates decisions on how to design a social cash transfer programme. For this reason the report will not cover issues related to the impact of social cash transfer programmes. The following quote from the report of a 2009 conference on social cash transfers in Zimbabwe (Smith

27 ) seems to reflect the general feeling that social cash transfers are a powerful tool to combat poverty and that their impact is significant. "The existing documentation of cash- and voucher-based responses shows that they are overwhelmingly successful in terms of their impact. People spend the money they are given sensibly, cash projects have not generally resulted in sustained price rises and women have been able to participate, and have a say in how cash is spent. Cash responses have usually been found to be more cost-effective than commodity-based alternatives." (Harvey, 2005) 3.6 Concerns, risks and challenges Market response Before starting cash transfer programmes there were concerns about inadequate market elasticity to respond to the additional demand and that cash transfers would result in food price inflation. Some programmes therefore analyzed the markets and informed traders before starting the programme. Partly because of these precautions markets have responded effectively. Traders welcomed the additional purchasing power resulting in increased business turnover and stocked up to meet the demand of cash recipients. Evidence shows that price increases following cash distributions are short lived. Following Concern's first distribution in Gokwe, for example, prices of cooking oil had doubled by the end of the day; three days later prices had reverted to normal (Smith, Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe ZECT reports: No complaints were recorded through Post Distribution Monitoring (PDM) on the incapacity of local markets to respond adequately to household demands. Monitoring price data also shows prices of basic commodities remaining somewhat stable. Monthly monitoring data collected on basic commodity prices found no evidence that the ZECT programme had a negative impact on those prices but a positive one. Prices in cash only markets were found lower than in food only markets where unfair bartering practices were still common. In spite of the relative short duration of the programme (5 months) there has been a reported increase in the number of traders, volume of commodities and number of commodity traded iin local markets while the cash transfer programme was implemented. (Roman 2010). Security concerns In total, Concern together with Safeguard -the Cash in Transit Company (CIT)- distributed on average USD107, in cash and 180mT of food every month to an average of 29,300 beneficiaries in almost 6,000

28 24 households. In spite of the potential security problems the programme could have had, one of the main successes of the ZECT was the secure atmosphere that the CIT, Concern staff and benefiting communities managed to ensure in all distributions, resulting in few complaints from beneficiaries and no incidents of corruption or theft. (Roman, 2010) Misuse of transfers by beneficiaries Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe Comparing the use of cash transfer between male headed households and female headed households during the whole duration of the programme, we observed that they were no significant differences between these two groups when looking into how they spend the transfer. Both groups spent most of the cash allocation purchasing and milling maize, and purchasing other food items and non-food items. And although women seem to be spending more money on health and education this could be because women headed households tend to have more dependents than male headed households. Therefore the "irresponsible spending" attached to men when distributing cash transfer was not observed (Roman 2010). While some cash may be used to buy things like cigarettes and alcohol, evidence shows that poor people prioritise basic needs - food, shelter, healthcare and education - for the majority of their expenditure (Smith, 2009). Intra household conflicts The ZECT programme paid the transfers predominantly to women. The evaluation report states: Giving cash to women would appear to be contrary to cultural practices but by doing a strong sensitization during cash distributions no problems were reported. The possibility of the cash increasing intra-household violence was discussed separately with men and women in 14 villages when carrying out focus group discussions. There were no reported incidents of gender based violence; on the contrary some women in Gokwe South reported that tensions had decreased at household level thanks to the cash transfer programme, as there is less pressure in the household to look for money to buy basic commodities (Kardan et.al.2010). While some cash may be used to buy things like cigarettes and alcohol, evidence shows that poor people prioritise basic needs - food, shelter, healthcare and education - for the majority of their expenditure (Smith, 2009). Stigma The ZECT evaluation reports: Recipients, including women and the elderly, felt that they were accorded additional respect within and outside

29 25 the household from their ability to meet household food needs (Kardan et.al. 2010). With regard to her social status since she is receiving cash transfers a 65 year old widow from Nyanga North is quoted as saying: It's strange what money can do, I am suddenly like everyone else in my community' (Roman, 2010). Dependency syndrome, graduation and exit strategies Labour endowed households (the transitory poor) will have to be assisted in such a way that their livelihood and food security is reestablished after which assistance can be phased out (graduation). Labour constrained households (the chronically poor) depend on long term social assistance. The concern is that Government welfare programmes are presently not able to absorb a significant number of the chronically poor. At the same time Non-State-Actors are not well placed to meet their needs for long term assistance. This important concern has been phrased by Non-State-Actors as follows: Exit strategies for the chronically poor: Although some of these households may also participate in livelihoods activities, many have no or limited capacity to lift themselves out of poverty, even with assistance, for example those with very high dependency ratios, or chronic illness or disability. For this small group of beneficiaries it is ultimately the responsibility of government to provide ongoing, long term social transfers. As such the exit strategy for NGO assistance for this category is referrals to the Department of Social Services, under the government's wider Social Protection Framework. However, it is also important to acknowledge that in the short to medium term, government systems in Zimbabwe are unlikely to have the capacity to absorb all of these additional households. Without additional assistance these households risk slipping back into destitution; where possible NGOs need to try to find additional sources of funding to continue assisting these households until such a time that the government can take over. If this problem persists or becomes wide-scale NGOs will need to engage other NSAs and government in seeking longer term solutions (Smith, 2009). Analysis of parameters and procedures used by cash transfer programmes in Zimbabwe Political interference Political interference is a concern of donors and implementing organizations alike. However, reports and interviews revealed no incidents that have seriously disrupted or derailed ongoing cash transfer programmes. This is maybe due to the fact that implementing agencies ensured that all stakeholders on all levels were consulted, informed and integrated in the preparation and implementation of the programmes

30 26 4 Tentative conclusions on implications for cash transfer programming

31 27 4. Tentative conclusions on implications for cash transfer programming Most ongoing cash transfer programmes invest a substantial amount of human and financial resources in their targeting systems. The effectiveness of these targeting systems has so far not been systematically evaluated. The limited evidence available households are significant. Reasons for low effectiveness of targeting seem to be: Target groups are not clearly defined like "support to OVC" or "support to vulnerable households" Targeting criteria like "households with orphans" or "households with a disabled person" are not necessarily indicating extreme poverty Targeting mechanisms like community based targeting are prone to abuse and manipulation by community leaders The analysis of targeting seems to suggest that a national social cash transfer programme implemented by DSS has to use a targeting system that is less open to abuse, is less costly in terms of human and financial resources and is at the same time more effective in avoiding inclusion and exclusion errors. Tentative conclusions on implications for cash transfer programming In order to harmonise with ongoing Government programmes (welfare programmes as well as public works) a volume of USD 20 per household per month seems to be appropriate. However, this should be seen as an average amount while the individual households should receive transfers in accordance with the size of the household. Taking the specific needs of the target group into account, delivery of transfers needs to be done at points that are not far from where they live. Cash in Transit (CIT) seems to be a client friendly, costeffective and well tested option. An alternative is a scratch card system provided that at least one pay point (private business) per ward will always have sufficient cash available to pay to beneficiaries. Cost-effectiveness could be increased by streamlining targeting, delivery and administrative procedures. To achieve this, the programme design has to be as simple and undemanding as possible. The M+E system has to provide up to date information on cost-effectiveness.

32 28 The analysis of concerns and challenges experienced by ongoing programmes does not indicate any risks that cannot be managed or mitigated by appropriate programme design and effective programme management. To achieve effective programme management, well equipped management units will be required on national and district level. Tentative conclusions on implications for cash transfer programming However, most ongoing programmes have been implemented by organizations that are equipped with appropriate human and financial resources and equipment (offices, vehicles, computers, copiers, etc.). DSS on national and district level has hardly any of these resources. The main challenge of the new programme is to strengthen DSS in all these dimensions. At the same time the programme has to be designed in such a way that it can function under poor frame conditions and scarce management resources.

33 29 References

34 30 References Akwara, Priscilla Atwani, Noubary, Behzad, Lim Ah Ken, Patricia, Johnson, Kiersten, Yates, Rachel,Winfrey, William, Chandan, Upjeet Kaur, Mulenga, Doreen, Kolker, Jimmy and Luo, Chewe (2010) 'Who is the vulnerable child? Using survey data to identify children at risk in the era of HIV and AIDS', AIDS Care, 22: 9, Concern Worldwide/ WFP (2009/10), Zimbabwe Emergency Cash Transfer Pilot Reports November, December 2009 and January Harare Catholic Relief Service (2009), Manicaland Cash Transfer Trial. Preliminary results on the impact of cash transfers in Eastern Zimbabwe. Harare References Food and Nutitional Council (2010), Zimbabwe National Nutrition Survey Preliminary Findings. Harare Gandure, Sithabiso (2009), Baseline study of social protection in Zimbabwe. Harare Gregson, Simon (no date), Manicaland Cash Transfer Trial. Scientific Evaluation: Design & Baseline Findings (Power point presentation) Harvey, Paul (2005), Cash and vouchers in emergencies, London 2005 Kardan, Andrew, MacAuslan, Ian and Marimo, Ngoni (2010), Evaluation of Zimbabwe's emergency cash Transfer Programme (Oxford Policy Management), Harare MDTF (2010), Challenges in Financing Education, Health, and Social Protection Expenditures in Zimbabwe. Harare Ministry of Public Service, Labour and Social Welfare (2001), Enhanced Social Protection Project - Basic Education Assistance Module (BEAM). Harare Moyo, Busisive (2008), The design of a pilot project for cash transfers to support OVCs in Zimbabwe. Harare Protracted Relief Programme (2009), PRP II Profile, Harare Roman, Elena Ruiz (2010a), ZECT monitoring consolidated report, November 2009 to March 2010) Harare Roman, Elena Ruiz (2010b), Programme evolution, planning and implementation Zimbabwe Emergency Cash Transfer Pilot Programme November 2009 to March Harare Schubert, Bernd (2010), Child-sensitive social protection in Zimbabwe. Harare Smith, Justine (2009), Cash transfers as a humanitarian and development programming tool in Zimbabwe. Harare Wallace-Karenga, Katrine (2009), An analysis of Government and NGOs Public Works/Food for Work approaches in Zimbabwe, Harare Wydler, Fabienne and Dengu, Ebbie (2009), Assessment of appropriate social transfer mechanisms in Zimbabwe, Harare

35 31 Annexes