Unconditional cash transfers in Gaza

Transcription

1 Unconditional cash transfers in Gaza An external review Provision of essential support to vulnerable communities in urban areas of the Gaza strip Result 5 Improved food security and livelihoods and reduced use of negative coping strategies of extremely vulnerable households in targeted areas in urban Gaza through the provision of unconditional cash grants commissioned by Oxfam GB Ben Mountfield 29 June 2012 Ben Mountfield relief recovery risk reduction benmountfield@gmail.com skype: benmountfield

2 Executive Summary This is a review of small pilot project, which targets extremely vulnerable households with monthly unconditional cash transfers for a limited period. The target population is made up of two groups, one group already receiving assistance through a voucher from WFP but still showing poor food diversity scores, the other initially receiving no other assistance and referred by the Ministry of Social Affairs (MoSA). All beneficiaries received a flat rate monthly amount of 260 NIS about 50, paid through the banking system. The Ma an Development Centre managed the project implementation as a partner of Oxfam GB. Project monitoring is comprehensive, and the resultant dataset provides a good basis for the determination of impact. This has been triangulated through interviews with key informants and focus groups with beneficiaries and staff. However, some revisions to monitoring tools and processes are necessary for the next round of the project, specifically relating to the measurement of expenditure and the better understanding of people s coping strategies and responses to the blockade and the resulting food insecurity. The relationship between Ma an and Oxfam should be reviewed to ensure it meets the expectations of both parties. Ma an is a mature organisation and the relationship is not typical, but neither is the context. The project is found to be appropriate and to have positive, quantifiable impact in terms of food consumption and dietary diversity; a quantifiable reduction in levels of short- term household debt and the use of negative coping strategies, and a reported but unquantifiable impact in reducing levels of domestic violence. This is true for both groups. The modality of an unconditional cash distribution is acknowledged to be more expensive than an in- kind distribution of food, but the added benefits reported include increased dignity, choice, and the ability to address a wider range of needs than can be addressed through the provision of staples. The intervention is complementary to the voucher for the sub- group that receive it. The impact of the project on gender dynamics in the household is seen as positive, and the choice of a cash modality appears to be a contributing factor. This needs to be understood in the context of the swiftly changing roles of men and women as a result of the blockade: where women have largely taken over responsibility for accessing resources to feed their household, yet those resources are in too short supply. The psychological impact of the blockade on men, especially associated with the loss of the role of breadwinner, might be explored in more depth. All indicators reviewed showed strong improvements over the baseline situation that can be attributed to the project. However, the proxy food security indicators have not improved to the point where household food security of beneficiaries could be described as good. The review found that the cash transfer is typically spent within 2-3 weeks of being received, leaving a shortfall towards the end of the month. It is found to be adequate for a household of 4 for a month, while the average household size within the project is around 7.6. The expenditure period is strongly (and inversely) correlated with the size of the household. The review therefore recommends that the amount of the cash transfer should be increased and scaled to household size, to ensure that it is sufficient for the targeted household and that the impact is maintained throughout the month. Further, since the project is of limited duration, the benefits will cease when the project ceases. For a sub- group of MoSA beneficiaries, this will be offset, at least in part, by their recent inclusion in a MoSA social protection scheme, which is longer- term. For most it presumably means a rapid return to their baseline situation. This effect could be quantified through a proposed final round of monitoring two months after the last payment. Detailed analysis of how the cash transfer is spent is complicated by the fact that it represents in many cases at least only a portion of the total household income. Recording the expenditure of the cash transfer in isolation invites a degree of selection on the part of the beneficiary as to what is reported under the transfer, and which expenditures are allocated elsewhere. Ben Mountfield Page 2 of June 2012

3 There are limited multiplier effects associated with these cash transfers. They are too small and too effectively targeted at those households that struggle to meet their daily needs: they provide little scope for promoting production or value addition. The review finds that too little is known about the processes by which families support each other, the networks through which they do so, and the health of those networks. Much is written about the resilience of the Gaza population, but we do not have the tools to measure that resilience, and we should not be too confident of its depth. A model to explore and understand these networks is proposed, together with some potential proxy indicators some of which are already collected through the SEFSec process. There is evidence that the resource base is being hollowed out: for example, some asset sales were identified during the implementation period amongst project beneficiaries, despite the fact the weakness of that particular monitoring tool would lead to under- reporting. As part of this understanding, future monitoring of impact should include a consideration of the total number of family members living in one residence, who share food frequently, as well as the standard measure of household size, typically used to determine the level of inputs. The widespread use of the Proxy Means Targeting Formula by the UN and the MoSA has been the subject of some debate, and this review highlights some concerns relating to its use and effectiveness. Oxfam could lobby to be included in an on- going review of the PMTF, and for the guarantee of a human over- ride in cases of inexplicable inclusion or exclusion errors. Acknowledgements This report would not have been possible without the excellent support provided by staff from Oxfam GB, the Ma an Development Centre and WFP. In addition to their consistent practical assistance and their timely response to my frequent requests for additional information and data, they were all reflective: focused on best possible outcomes for the targeted households, committed to the review process as a means of learning and improvement, and refreshingly open to consider criticism of current processes, challenges to existing ways of thinking and alternative perspectives. This report, then, embraces that willingness to reflect and presents a range of opinions: these have already generated responses through an early draft, and the report is expected to continue to evolve in the light of additional data and feedback. Of the opinions presented, some are strongly supported by substantial monitoring data, and others are much more tentative: some of these are presented as being opportunities for further research. The report will inevitably contain both bias and mistakes, and these remain the responsibility of the author. Comments, corrections and feedback on the report are welcome. Ben Mountfield Page 3 of June 2012

4 Contents Executive Summary... 2 Acknowledgements... 3 Contents... 4 List of figures... 4 List of tables... 5 Acronyms and abbreviations used... 5 Methodology... 6 Introduction... 7 Socio- economic context... 7 Programmatic context... 8 Targeting beneficiaries and the monitoring framework... 8 Observations and findings related to the objectives of the TOR Targeting Beneficiary expenditure Gender impacts Multiplier Effects Other cash programming Appropriateness Beneficiary profiles Additional observations and findings The impact on household food security Findings related to the monitoring tools The timing of monitoring visits Gifts, borrowing, lending, and networks Planning horizon Coping strategies Access to future support Ma an and partner relations Exit strategy The sustainability of the support network Recommendations Recommendations for Oxfam with regard to future cash transfer programming Recommendations for project stakeholders and other actors Annex 1: Terms of Reference Annex 2: the questions used to generate the PMTF Annex 3: key informants to the review process List of figures Figure 1: schedule of grants payments and monitoring... 9 Figure 2: distribution of household size Figure 3: balance of PMTF questions by category, adjusted for weighting Figure 4: expenditure breakdown for MoSA and CVP groups Figure 5: average short- term debt at the time of monitoring Figure 6: how long does the cash transfer last? Figure 7: distribution of Household Dietary Diversity scores Figure 8: changes in Household Food Insecurity Access category Figure 9: schematic of household support network Ben Mountfield Page 4 of June 2012

5 List of tables Table 1: numbers of households passing through selection process Table 2: stages in the screening process for the CVP group Table 3: simple vulnerability analysis of applicants Table 4: how the time taken for the transfer to be spent varies with household size Table 5: the expenditure period of the grant, in people- days Table 6: scaling a sufficient transfer by household size Table 7: coping strategies and their weighting within the CSI tool Table 8: classifying coping strategies by income, expenditure, production and consumption Table 9: additional coping strategies identified by social workers and through other discussions Table 10: six levels of assets Table 11: the social acceptability of different coping strategies Table 12: frequency of children being sent out to work Acronyms and abbreviations used CFW CSI CVP ECHO HDDS HFIAS MoSA NIS opt PMT (PMTF) SEFSec TOR UNRWA WFP Cash for Work Coping Strategy Index Cash Voucher Project (previous Urban Voucher Project) of WFP / Oxfam / Ma an European Commission, Humanitarian Aid and Civil Protection Household Dietary Diversity Score Household Food Insecurity Access Scale Ministry of Social Affairs, Gaza (also MoSA group: beneficiaries referred by MoSA) Israeli New Shekel currency common to Israel and opt. (also ILS) occupied Palestinian territories Proxy Means Targeting (Formula) Socio- Economic and Food Security Survey Terms of Reference (for this evaluation) United Nations Relief and Works Agency for Palestine Refugees in the Near East World Food Programme Ben Mountfield Page 5 of June 2012

6 Methodology A balance of quantitative and qualitative methods was used to gain a thorough understanding of the project, its impact, and the context in which it operates. Secondary data A range of project- specific secondary data was reviewed, including the project proposal, the TOR for the review, the mid- term review of the CVP project and some monitoring reports produced by Ma an using the data collected through the project under review. Other secondary data was reviewed relating to the food security situation in Gaza, the socio- economic context, and the broader situation with regards to the current state of the blockade. Various papers analysing gender dynamics in the Gaza strip were also consulted. Quantitative analysis of monitoring data The project monitoring is regular and extensive. The dataset generated required little cleaning and was easily accessible. Although data was collected in Arabic, the field headings were all translated into English and the much of the data is numerical. Standardised and common answers were also translated. Some of the qualitative data, where answers were non- standard, was not analysed. The sample sizes for the monitoring were a compromise between ensuring cost- effectiveness and providing truly representative information. This is particularly true for the baseline. Monthly monitoring could usually be combined to form a fully representative sample of the target population. Qualitative analysis through participatory methods The main approaches used were semi- structured interviews, with key informants and stakeholders to the project, and with focus groups of project beneficiaries and staff. Focus groups were primarily used to triangulate and validate hypotheses suggested by the data analysis, place findings in context, and explore weaknesses and gaps in the data. A small number of remote stakeholders were consulted by and telephone. Focus groups of male and female beneficiaries were interviewed separately. In some focus groups, ranking and other exercises were undertaken to clarify findings and ensure consensus. Seven household visits were also made, four to randomly selected beneficiary households and one to follow up specific issues that arose during focus group discussions. Two visits were to observe the Oxfam verification processes. One supermarket participating in the CVP was also visited. A list of key informants and focus groups arranged is included as Annex 3 Ben Mountfield Page 6 of June 2012

7 Introduction Socio-economic context It is not necessary here to rehearse at length the complex context of the Gaza situation, nor its severe and growing effects on the poorest segments of society. The following summary is incomplete, but relevant to the analysis below. Three interconnected groups of factors inform the socio- economic context and critically, all are underpinned by the blockade: Restrictions on movement: subsequent loss of human, financial and social capital; limited imports, no exports. At the moment this is most strongly felt in the fuel crisis; at other times it is expressed in liquidity shortages of shekels or dollars, or in other ways. The shortage of employment opportunities is a constant result of the closure. Large families and rapid population growth, reduced land and property holdings with each generation. Put bluntly, despite the current building boom, the next generation of poor urban Gaza residents has nowhere to live. Fixed or reducing natural asset base, with increasing environmental contamination this is particularly obvious in the water sector, but also true in farming and fisheries. With land being shared out at each generation, household level productive units are getting smaller and will eventually become non- viable, pushing people off the land. The combined impact of these three sets of factors is a reduction in resilience and an increase in vulnerability as the resource and asset base of many households and businesses is inexorably depleted. No convincing scenario mapping exists that might explain how this situation will develop: at what point the growing demands will completely outstrip the dwindling supply, and what the nature of the resulting crisis will be. The network that provides support to vulnerable households in Gaza is intricate, and humanitarian actors are just one part of it. The absence of tools to understand this support network or indicators to measure its health is a weakness implicit in the humanitarian system. The joint SEFSec survey presents the best macro perspective on food security, and it shows a general improvement since 2009 when WFP programming started after Operation Cast Lead. However, while this is the baseline for the WFP programme, it is a poor reference point against which to compare the current situation: a low point in Gaza s recent history from a number of perspectives. Indeed, even if continuous data were available, it would not be possible to identify a normal time to use as a reference no such point exists. All that is possible is to monitor trends and the trends do not show an increase in the proportion of people who are considered to be food secure in the Gaza strip since 2009, despite the start point, all the humanitarian aid and the so- called easing of the blockade. People have moved out of the category of food insecure into the intermediate categories: there has been no corresponding move into the category food secure. And whatever the trends, the overall numbers of people who remain food insecure are intolerable. The second concern with the SEFSec approach is that the published summaries focus largely on food consumption: an area where good results might be expected, given the amount of food aid provided. Other data is collected and analysed: for example food insecurity by land holding type for 2011, but trend data over the last three years is not presented. There is no information on the size of landholdings, which must be of a certain size to be viable. In particular, information on the continued sale of assets is absent; despite the fact that this practice clearly indicates that the resources available to a household are inadequate for the household needs, even while food access levels remain acceptable. Ben Mountfield Page 7 of June 2012

8 Programmatic context This is a review of a small pilot project supported by short- term funding from the ECHO, the European Commission s Directorate General for Humanitarian Aid and Civil Protection. Agencies in Gaza operate in a protracted limbo, which is neither strictly humanitarian nor developmental. The principal approach is the maintenance of an unsustainable status quo, holding the growing numbers of very poor just above destitution while trying to stimulate an economy rendered irretrievably moribund by the blockade. This depressing situation is punctuated by periods of real humanitarian activity prompted by escalations, incursions or perhaps shortages of critical commodities. There is an uncomfortable parallel between the hand- to- mouth, year- by- year existence of many agencies, and the equivalent hand- to- mouth, week- by- week or even day- by- day existence of their beneficiaries. This may go some way to explain the apparent lack of a strategic approach to the problems on the part of the agencies. Stymied by the impossibility of promoting any real development, constrained by short- term perspectives, and drained by the many challenges of delivering their projects, vision and strategy have given way to determination and tactics. Denied the opportunity to export, the Gaza economy is almost entirely contingent on external support, much of it in the form of PA, UN and NGO salaries, programme funds spent in Gaza and resources transferred to beneficiaries, presumably supplemented by external support to the Gaza authorities from supporters elsewhere. Imports from Israel and Egypt differ enormously in their price, quality, acceptability and regulatory frameworks, and this complicates the marketplace. Local productive capacity has in many cases been undermined by cheap imports. Regulations (both external and internal) placed on the expenditure of humanitarian actors limit their flexibility, and further increase the dependence on imports. The project supports beneficiaries recruited from two separate sources, and this report examines them separately where this is appropriate. In common with all families in Gaza, beneficiary households have access to a support network that may include family, mosque, zakat, and neighbours to varying degrees. The scale of support they receive in this way relates to their circumstances, their willingness to ask for help, and the capacity of their particular network to provide support. In this way, the two groups are similar. In terms of external support, the two groups of beneficiaries are different. Of the 240 beneficiary households, 100 were referred by the Ministry of Social Affairs (MoSA) and at the start of the project were receiving no external support. The other 140 were already beneficiaries of a WFP/Oxfam conditional voucher, to be spent in a designated supermarket on food commodities selected from a range of 10 essential items. In this report, these households are described as the Cash Voucher Project (CVP) group. The value of the weekly voucher varies in steps with household size from 14 NIS to 96 NIS. Through the unconditional cash grant project, both groups received a monthly amount of 260 NIS regardless of family size or circumstance. Targeting beneficiaries and the monitoring framework A thorough process was undertaken by the project partner Ma an Development Centre to select beneficiaries from the lists provided by MoSA and WFP. Oxfam staff validated a proportion of selected households and found some inclusion errors: action was taken in these cases. Monitoring was undertaken at the same time after each payment round, to maintain comparability between the various rounds. The specific time chosen was three weeks after the monthly payments. The target population were divided into two parts. A fixed group of approximately 10% of the beneficiaries was randomly selected: 10 households from the MoSA group and 15 from the CVP group. These households were monitored at each round. The remainder were monitored in rotation, with the households selected at each round randomly selected from those who had not yet been studied. This process allows for longitudinal study to be undertaken on the fixed group, lowers the monitoring burden on most of the households, and ensures that all beneficiaries receive at least one Ben Mountfield Page 8 of June 2012

9 follow- up visit from the field workers they were all visited during the targeting process. There are no standard mechanisms for onward referrals should these prove to be appropriate. The baseline survey and subsequent monitoring were done with a detailed questionnaire, which includes a wide range of questions including three standard proxy indicators for food insecurity: Household Dietary Diversity Score (HDDS), Household Food Insecurity Access Score (HFIAS) and the Coping Strategy Index (CSI). This apparent redundancy was deliberate on the part of Ma an to develop their understanding and learning about the complexities of food insecurity in Gaza, and it provides an opportunity to compare the various tools in this context. Opportunity exists to do a more thorough longitudinal analysis of the situation of the small fixed group, and Ma an are already looking specifically at this subset of the data. Figure 1: schedule of grants payments and monitoring Sample size Despite the significant effort required to monitor around 75 households after each payment, the sample is not large enough to be considered representative of the beneficiary population of 240. A truly representative sample would have incurred a monitoring burden that would not have been cost effective. A 5% margin of error, at a 95% confidence interval for a population of 240 would require a sample size of 148 at each round of monitoring. To do detailed analysis between the two groups, the sample sizes required are even larger (and less cost- effective). For the sub- population of 100 MoSA beneficiaries, it would necessary to monitor 80 households each round in order to achieve the same levels of confidence. For the 140 CVP households, 103 households would need to be visited. Detailed monthly monitoring at this level is not realistic. Fortunately, once the monitoring rounds are combined the confidence can be much higher: the combined data includes visits to all households at different times and is therefore representative although averaged across the project period. In most cases in this report, the combined data will be used unless the monthly numbers show something that looks significant. In addition, the fixed group provides a further opportunity to identify trends over time and isolate noise in the data. In addition, where there is a high level of consistency between the answers in a group, confidence increases and a smaller sample will also be representative. This characteristic may well apply in this case, based on the consistency of responses to many questions in the focus groups. However, this may not apply to all the questions in the monitoring process. As is normal in the sector, the unit of data collection, analysis and allocation of resources is the household, which is consistently described as that group of people who cook and eat together on a regular basis. It should be noted that there are two other groupings that are relevant in this context: the (extended) family, and (in many cases) the cluster of households from one family that share a property. Where two related households live together in one property, we can generally expect the level of sharing to be much higher than a situation where they live separately. In essence, the unit of analysis we use the household is in many cases different from the unit within which resources are largely pooled the family. This theme will be further explored in the report. Ben Mountfield Page 9 of June 2012

10 Observations and findings related to the objectives of the TOR Targeting TOR Objectives: To assess whether appropriate targeting criteria have been identified to select the most appropriate beneficiaries to receive unconditional cash transfers (in both CVP and non- CVP caseload); To assess the extent to which these criteria have been correctly applied to beneficiary selection in the current project. Processes The processes used for beneficiary targeting and selection were thorough although not totally rigorous. The start point for the process were two lists provided by WFP (from the CVP programme) and MoSA. A two stage selection process was adopted, as shown in the table below. Table 1: numbers of households passing through selection process Short interview Passed short interview MoSA CVP Long interview (baseline) The interview processes were fairly comprehensive: the questionnaires are quite long and detailed. Some of the field workers (called Social Workers in the MoSA project and Data Monitors in the CVP component) managed to complete a large number of surveys a day, which is surprising. However the data provided appears to be clean and of high quality, and project partner Ma an have high level of experience and skills in data capture and analysis. The reduction from the CVP cohort from 291 to 140 was done on the basis of the information collected and held on the database, using a ranking process. Criteria included HH size, ability or inability to work, numbers of women and children, numbers of disabled, other sources of income, and similar. In general the WFP candidates ranked higher than the MoSA candidates, as would be expected from the lack of inputs received by the MoSA group. A sample of the selected candidates were then verified by Oxfam field staff: a process that produced evidence that the desk exercise was not entirely satisfactory. From the 140 CVP candidates, 35 were visited and of these, 7 were found not to meet the criteria. Likewise, from the 100 MoSA candidates, 15 were visited of whom 3 were found to be unsuitable. In each case this is a 20% error, in the light of which it may be wise to revisit the criteria used, or the ranking system. Appropriate, qualifying alternative households were found from the remainder to take their places. The breakdown of the successful applicants by beneficiary group and household size is shown below, disaggregated by the gender of the household head. Ben Mountfield Page 10 of June 2012

11 Figure 2: distribution of household size Outputs The demand for support of this type appears to outstrip the supply available. It is not surprising that no examples of inclusion errors were observed during the field visits: all those met and interviewed were found to be extremely vulnerable and appropriately selected. These observations are supported by analysis of the detailed baseline and monitoring data. The selection of households visited was by Ma an. To fully investigate errors of exclusion would require field visits to non- beneficiaries, which presents ethical challenges. However, the data relating to the selection of the eventual beneficiaries from the lists provided offers some insights. Much of the initial application process is in narrative form and in Arabic: it does not readily lend itself to analysis. Thus the selection of the fields below is somewhat arbitrary they are relatively easy to extract from the larger dataset and relevant to the question. Table 2: stages in the screening process for the CVP group. has work owns business land / dunums # of disabled in HH yes no yes no 0 <=1 > Excluded at first stage Excluded at second stage Included in programme At first glance, this looks satisfactory people with their own resources are being excluded from entry in favour of those without. But if we turn this analysis around and look at people without specific assets, and in particular without the whole range of these assets, then the picture looks less rosy. The following table is based on the same data (slightly cleaned) and uses four simple tests for vulnerability: no job, no business, no land, one or more disabled people. It scores each category 1 or zero: a household with no work, no business, no land and disabled members would score 4. The results look like this: Table 3: simple vulnerability analysis of applicants 0 factors 1 factor 2 factors 3 factors 4 factors Excluded at first stage Excluded at second stage Included in programme Ben Mountfield Page 11 of June 2012

12 Some people may have been excluded for a variety of reasons that are not immediately apparent. Some households did not want to participate in the programme; others were not available. But it is clear that the process of beneficiary selection warrants some further analysis, if more people with all four of these factors were excluded than included. The Proxy Means Targeting Formula The main tool used to target assistance in Gaza and throughout the occupied Palestinian territories (opt) is the Proxy Means Targeting [Formula] (PMTF, PMT), which is used by UNRWA, WFP and MoSA. It is tested through a common questionnaire, some (but not all) of the answers to which are used to produce a household score. The PMT score is usually described as a measure of poverty or income levels the means of the title. The PMTF has been the subject of some debate, which is driven by reports of inconsistent results, and fuelled by the slightly odd reluctance to widely share the formula by which the result is generated from the questions. The usual challenge raised against the PMTF is that it is strongly asset- based, and thus discriminates against the new poor : those who had jobs and assets before the closure, and now have a good house and a TV set but no income and little food. The challenge to the PMTF is worth considering. The following chart groups the PMT questions into types, and combines their individual weightings to provide a score for the group. The allocation of questions into groups is shown within Annex 2. Figure 3: balance of PMTF questions by category, adjusted for weighting It must be acknowledged that measuring household income is tricky in any circumstance, and much more complicated in situations where informal transfers of cash, food and assets between households are so common. Much effort has gone into the development of the PMTF: it works in most cases, it is cheap to administer and it is objective. But as the chart demonstrates, there is still some scope for development 1 : this balance seems intuitively wrong for the Gaza context. Most of the questions used to collect the data relate to a range of typical household assets or to household demographics. No attempt is made to actually measure income or expenditure, or to look at support networks beyond the household. The questions examine the existence of assets, but not changes or asset sales, and the only question about productive assets relates to a car. There are many more common productive assets in Gaza that might reward observation. Over a third of the scoring relates to household and fixed assets, which supports the common charge made against the weighting of the tool; less than 20% relates to income, expenditure and production, and then only indirectly. Human capital might be under- represented in the context of the West Bank, but it appears to have too high a weighting for the Gaza situation, where master s degrees are so common, and jobs so rare. The MoSA PMTF is common to the West Bank and Gaza, but that of UNRWA is slightly different and has a variant for each location. It is understood that work is planned or underway to review the tool, and that harmonisation is a longer- term objective. 1 Three recent EU documents by different authors describe the PMT formula as weak and not appropriate for a more complex social assistance system, according to a CTP background document related to an ongoing review. Ben Mountfield Page 12 of June 2012

13 Until that time and for some time afterwards it might be sensible to ensure that there remains the possibility of a human override of the calculation, at each end of the scale. Once the answers to the questions have been checked and an experienced field officer still raises questions about the inclusion or exclusion of a household, the limitations of the tool should be recognised and human wisdom allowed the final say with appropriate checks and balances applied. Each such case should be thoroughly investigated against a broader set of questions, to identify weaknesses in the tool as it stands and to inform the next process of review. Beneficiary expenditure TOR Objective: To analyse the expenditure patterns of both groups of beneficiaries and how this has varied between them Expenditure The project monitoring looks closely at the expenditure of the cash grant. The data entered in the database is generally very clean, and the field workers described a process that was thorough. There was no opportunity to cross check the validity of the data capture process as no monitoring took place during the review period. As noted above, the sample size for each round of monitoring is too small to have high levels of confidence in the results. However, the full picture has been provided, as it shows broad trends that are valid. Small fluctuations between the rounds are likely to be caused by the sample size. Figure 4: expenditure breakdown for MoSA and CVP groups The chart shows a significant expenditure on clothing at the start of the project which coincides with the al- Adha feast which reduces to zero towards the end of the period. Food costs were low in the first round also, which might again relate to the feast period, when many poorer people are given gifts of food. Expenditure on non- food items gradually increases through the project period. Expenditure on production is effectively zero. Several of the MoSA group had saved quite a proportion of the cash grant at the time of the final monitoring, which is interesting. A decision on monitoring the 7 th round has yet to be taken. Relative expenditure on food is interesting: given that the CVP group all receive a conditional voucher designed to cover food needs, we might expect them to spend less of the cash transfer on food. Put Ben Mountfield Page 13 of June 2012

14 another way, we might expect the MSA group to spend proportionally more of their transfer on food, as they may not have access to the essentials. This is not evident in the data. It was reported that between 70 of the MoSA beneficiary group were included onto the MoSA National Programme for Social Protection caseload, from January They were previously on the waiting list. They will now receive between 750 NIS and 1,800 NIS every three months, depending on household size. The cash transfers are not the only source of income for many families. Some will have inputs from MoSA, others will have small amounts of income from casual work, for example. But the monitoring examines only the expenditure from the cash transfer. This is unrealistic, as money is fungible: even if the beneficiaries are being completely honest about the expenditure, they could allocate food expenditure to the cash transfer and tobacco to their other income, if they chose. The monitoring tool would be strengthened if total expenditure was measured, and expenditure allocated to the cash grant proportionally. This process should take place after recording total income, including cash and in- kind gifts. Once the total short- term debt at the start and end of the process is included (see below), this would provide a more representative picture of household finances. The usual problems of collecting financial information at the household level will still remain, of course. Monitoring The chart above is drawn to suggest a longer gap between the monitoring of rounds 4 and 6, as shown in the timeline in Figure 1: schedule of grants payments and monitoring. Internal issues delayed payment 5, and two payments were made within a few days. A decision was made to delay the monitoring until 22 days after the second payment, and monitor both payments together. With hindsight it would have been useful to monitor the non- payment of round 5 at the regular time, around the 21 st of March. Expenditure monitoring would not have been possible the money from the previous round was all spent. But a consideration of changes in the levels of coping strategies, for example, would have provided valuable insight into the effects stopping the payments a situation that is now being replicated at the end of the programme and may have helped to prepare an exit strategy. Servicing short-term debt Expenditure on debt is a major feature of both groups, and this is mostly short- term debt. The chart above shows that especially in the first couple of months, debt payments were quite high. The survey asked specific questions about accumulated short- term debt, and the Figure 5 below summarises this information for the two groups. Figure 5: average short- term debt at the time of monitoring. Ben Mountfield Page 14 of June 2012

15 This chart suggests that project households seem to be comfortable with between 200 and 300 NIS of short- term debt. Interestingly, this is very similar to the size of the cash transfer, shown in orange in the centre. Average levels of debt at the baseline were much higher for both groups, more than double in the case of the MoSA group. The expenditure chart in Figure 4 shows average monthly payments on debt in the order of 36 shekels. However, levels of debt after the first round are not decreasing: the debt is rotating, and the money released is being spent on something else. Expenditure recorded as debt should only relate to the decrease of the absolute amount of short- term debt: debt being paid off. The format of the monitoring form could be amended to provide an easy check that the figures balance at the household level. The expenditure data shows that all reported expenditure is appropriate, although it might have been wise to include a tobacco field in the monitoring form. The debt levels suggest that the size of the transfer is in the correct order of magnitude. How long does the cash transfer last? We know from the focus groups that some families spent the cash transfer quickly on a durable commodity, some on debt payments (especially in the early rounds) or on medical expenses. Others made an effort to eke it out, so that it appears to be sufficient for the whole month. However, the data presents a clear and consistent suggestion that it lasts between 2 and 3 weeks, as Figure 6 below demonstrates. As the data for the first month was significantly different from the data for the following rounds, it is presented separately in each case. Figure 6: how long does the cash transfer last? When is debt expenditure? If the total short- term debt of a household changes from 310 NIS to 275 NIS over the reporting period, this corresponds to an expenditure of 35 NIS, regardless of the amount of old debt paid off and new debt taken on. If the amount of debt increases from 310 NIS to 340 NIS this corresponds to an increase in the cash available to the household that month by 30 NIS (and a corresponding increase in the debt burden). If interest is ever payable on debt, (and it is thought that this is not normally the case) this should be considered as expenditure. Ben Mountfield Page 15 of June 2012

16 There is also a strong inverse correlation between the size of the household and the time taken for the transfer to be spent, as the table below illustrates. We can call this time the expenditure period. Table 4: how the time taken for the transfer to be spent varies with household size HH size Expenditure period (days) The table above demonstrates that on average, even very small households consume the grant faster than the monthly cycle. Note that a household of average size around 7.5 members will consume the grant in a little over around 2 weeks. This suggests that the size of the cash transfer is too small for a typical family, and that consideration should be given to moving away from a flat rate. Short term vs longer term expenditure Focus groups and household visits asked about people s ability to put some of their cash transfer towards longer term expenditure, in the form of durable assets or investment into projects to generate an income in the future. The usual answer was no : all the expenditure was for daily consumables. But there were a few interesting exceptions, for example: One man in a focus group, undergoing cancer treatment, out of work for a long time and receiving no additional support, had been able to provide basic furnishings for his house, where there was he said previously no furniture at all. A divorced woman living with her mother, children, siblings and several disabled relatives had been able to buy some ducklings cheaply and was feeding them up to sell. Another man bred rabbits from one male and two females purchased using the transfer. A woman set aside 50 NIS per month to purchase a fridge and cooker she gave the money to her sister for safe keeping so that she would not be tempted to spend it. Working through a rotating savings group, a woman was able to tile her bare floor with second- hand tiles. A man from the MSA group bought a vender s trolley and began to sell vegetables. Since most of this list was generated through the two focus groups with the field workers, who between them knew all the beneficiaries, it is possible that it is complete or almost complete. The large majority of beneficiaries have spent the money entirely on immediate needs. It may be possible to set a more appropriate value of the grant over the next couple of cycles of the project, by setting a target for expenditure on durable items. For example, if the target set were: 50% of households are able to put some expenditure into durable items at some time during the six- month duration of the project then the current size of the cash transfer would be too small. The size of the grant As noted above, the current grant is a flat rate, and there is an inverse relation between household size and the expenditure period. This could be simply addressed by scaling the grant depending on household size, perhaps with an upper limit. This approach would then be consistent with much other aid provided in Gaza. If we multiply the HH size by the number of days the grant was reported to have lasted (7, 14, 21 or 28) we can find the number of person- days for each HH interviewed, and an average for each group for each monitoring round. Ben Mountfield Page 16 of June 2012

17 Table 5: the expenditure period of the grant, in people- days Monitoring round MSA group Average number of people- days UVP group MSA & UVP combined & Rounds As noted before, expenditure in the first round was different, and the conditions of the last round were unusual. In normal rounds, the cash transfer lasted an average of 119 people- days, and this figure is used for the calculations below. Let us call a cash transfer that lasts 30 days sufficient. We can calculate the size of a sufficient cash transfer for households of a range of sizes. Table 6: scaling a sufficient transfer by household size HH size Proportion of current transfer Amount of transfer required 25% 101% 126% 152% 177% 202% 227% 253% 379% 505% This approach makes a number of assumptions, including: Expenditure patterns would remain constant as the amount of income increases. Expenditure patterns are proportional to household size. It seems unlikely that the first of these is true: a larger transfer may open possibilities for expenditure on durable items that a smaller transfer may not, for example. However, this would be readily monitored through existing systems. Households and families In this context, all those inputs that are provided scaled against household size should perhaps also be monitored against family size. As noted in the introduction, the framework used by humanitarian actors appears to be different from that used by the people themselves, when it comes to the movement of food and resources amongst family sub- units, especially when they all live within a single property. The concept of family size would need to be clearly defined, but in principle it is the total number of people living together who regularly share food, regardless of whether they cook together on a normal basis. Cross- referencing indicators like HDDS scores and cash transfer expenditure periods against family size, once the inputs have been scaled to size of the household, should provide additional information on the levels and nature of sharing. Note that it is not being recommended that organisations switch to family size as an alternative to households just that this be monitored and analysed in order to better understand the sources of vulnerability and the complexity of social and family networks. Health expenditure Feedback from focus groups and household visits showed that for many families, it is essential medical expenditure that impacts most heavily on the household economy. This is not clear from the expenditure chart, which shows average health expenditure to be around 20% of food expenditure. But while every household must buy food, only certain households need pay for expensive procedures, medication and accessories. For these households, the expenditure on these items is perceived as essential - who can think of food when your child is suffering? and it has a significant impact on the ability of these households to meet their food needs. The averages produced by the monitoring are not a good reflection of the situation of these households. Ben Mountfield Page 17 of June 2012

18 Larger families are more likely to have one or more member with a disability or a serious medical or mental health condition that makes frequent demands on household finances. There are 12 occurrences in the monitoring data when health expenditure accounted for half or more of the total, and 2 occasions when it consumed the entire amount. In 9% of individual monitoring (30 occasions) expenditure on health was equal to or greater than expenditure on food. Gender impacts TOR Objective: To assess the extent to which this targeting has impacted women and men s access to and control over resources and/or impacted women s role and decision- making power within the home. According to the focus groups arranged for this review, women manage the cash transfers in most households, whereas men generally managed previous food distributions. This seems to be true both of the CVP voucher and the cash transfer. Some of the CVP beneficiaries were transferred quite recently from a CHF list and their modality changed from food to cash: they reported a change in household roles at this point. The following generalisation was heard in several focus groups: If we receive a food distribution, the men are needed to carry the heavy load. They will sell some of the foodstuffs, and use the money to buy cigarettes. If the aid comes in the form of a cash transfer, it generally comes to the women, who retain control and use the money to feed the family, and on essentials. This is probably true in many cases, at least within Gaza City, but it is not a finding that is replicated reliably in other research, and it is certainly worth looking beyond this. Role inversion What appears to have happened over a period of some years is an inversion of some of the traditional gender roles. As the supply of resources, especially jobs, has gradually diminished, men have increasingly felt disempowered. This was already happening over a period of time, but it became absolute with the closure. Many men have lost their capacity to feed their families, and with it a measure of their self respect and dignity. Men did not, usually, actively control the resources, but they did provide them. Women have largely stepped into the breach. They have taken on the responsibility to feed the family, and this has been encouraged by aid agencies actively targeting women to receive resources. As a result, women reported becoming more active, being more mobile, and learning new skills. There is a flip side to this newfound responsibility. The resources to meet their obligations are insufficient. Many women cannot meet these new expectations, and as a result they are under enormous stress. Gender-based violence Much has been written about gender- based violence in Gaza, but there is little quantitative data available to measure trends. This is a shame, as focus groups with women and men were unambiguous about the link between food security and domestic harmony. They told us: when food supplies are both adequate and predictable, levels of tension in the household are reduced. When food is absent or there is little confidence in its availability in the near future, tensions increase. Men s loss of dignity is a compounding long- term factor, which presumably lowers the threshold at which tension or violence is likely. Groups of women and men described the impact of these raised tensions in different but overlapping ways. The men spoke of tension, anger and shouting, of leaving the house for a while or sending the wife away to her parents for a few days until things calmed down. The women added to this list verbal and physical violence against themselves or the children. The prevalence of this kind of violence appears to be extremely high. Every woman in each focus group reported it herself or acknowledged the experiences of the women who spoke up. During one Ben Mountfield Page 18 of June 2012

19 household visit, a woman spoke about it openly, in front of the neighbours children, as if it were perfectly normal. During another, a woman discussed it in front of her mother and brother. Considerations for programming Food security is more than food consumption. For a household to consider itself food secure, they must have confidence that their supply is reliable. If the link between food insecurity and domestic violence is as strong as was suggested during the focus groups, then it may be possible to infer the level or frequency of one from the other, if indicators for the level of gender- based violence could be identified and measured, and the correlation established. More immediately, the provision of cash or food appears to reduce violence against women. The end of such provision in the absence of any other change in circumstances will presumably lead to an increase in such violence. This has serious consequences for the exit phase of this project, and of others like it. In this context, the missing round of monitoring when cash transfer 5 was delayed takes on a new significance. At the end of one focus group, the women were explicitly asked if they liked their new empowerment. The question came immediately after the discussion about violence in the household. They were quite clear that they did not: they would like to go back inside their houses. This surprising answer may have been the result of the conflation of the two ideas, generated by the previous discussion. Programming that implicitly places the responsibility for feeding the family in the hands of women, yet does not provide them with sufficient resources at sufficient levels of confidence, actually disempowers women, and places them at higher risk. Projects that actively create meaningful employment for men, while maintaining the gains made by women, could go some way to reducing levels of domestic violence. To achieve this at scale, of course, requires the lifting of the blockade. Considerations for monitoring and research Different researchers, working in different parts of Gaza, have reported quite different findings about preferences for food, vouchers and cash. These variances may be due to the location, the individuals being asked, their social and livelihood situation, the external context, previous topics loading the dialogue, the framing of the question, the bias of the researcher or their organisation, and many other factors. To try to reduce this bias, the questions could perhaps be framed in a neutral manner: Aid can be delivered as food, as a voucher, or as cash. In what circumstances would you prefer which option? Which of these descriptions most closely reflects your current situation? What might cause your preferences to change? These might work best in individual household interviews rather than in focus groups unless those groups are highly homogenous. It will be important to distinguish between unconditional cash transfers and cash- for- work in such discussions. Multiplier Effects TOR Objective: To identify and attempt to quantify any actual or potential multiplier effects on the local economy that can be attributed to the cash transfers. The multiplier effects of the project are probably limited to the impact of the monthly injection of an additional 62,000 NIS into the economy of Gaza city, plus the creation or maintenance of a few jobs for staff of the project partner. At best, this probably increases the turnover over some traders and shops by a modest amount. The low numbers of households perhaps as few as 2 investing in productive projects suggests that this activity is not going to have a multiplier effect. If the cash transfer were designed to support the Ben Mountfield Page 19 of June 2012

20 development of livelihood strategies, it would need to be larger, and then multiplier effects per household would be larger, and could perhaps be observed and measured. If the size of the transfer remains as it is, and the scale of the project were increased, then the modest multiplier effect would simply scale up proportionally. It is implicit in the nature of unconditional transfers that they can be spent anywhere or indeed saved, which makes both quantification and attribution very difficult. In contrast the funds transferred through the CVP project are spent through specific outlets, whose financial records and physical properties are open to analysis. In addition, the ambition of non- participating outlets to join the scheme provides another indicator of multiplication. Other cash programming TOR Objective: Identify other cash transfer programmes in Gaza, highlighting the scope, scale, frequency and beneficiary profile of recipients of these transfers. This objective was not addressed in much detail, as it is more relevant to a follow- up review of the CVP planned for June. The use of cash as a modality in Gaza is overwhelmingly in the form of Cash For Work (CFW), or to be more accurate, job creation projects. Few of the projects have the characteristics associated with good practice in CFW. Cash transfers are also used extensively by MoSA, as a form of social protection for the poorest and most vulnerable households, using the PMTF as the basis for targeting. Appropriateness TOR Objective: Using the experience of this pilot, evaluate the appropriateness of the cash grant modality as a humanitarian intervention in the Gaza context, including an assessment of the capacity of the Gaza economy to absorb a larger injection of cash as an intervention modality without having other negative impacts (e.g. inflationary pressures). The programme is a humanitarian intervention that conforms with the EU s definition of humanitarian assistance, designed to ensure that the basic food needs of extremely vulnerable households are met by providing sufficient unconditional cash transfers intended to meet basic household needs. The programme meets an acute need for the target beneficiaries, although only a small (and unknown) proportion of the potential target population is reached. The approach used is completely appropriate for these households: the market system is strong, the commodities are available, but their access to these commodities is limited. The modality of an unconditional cash transfer meets these access needs effectively. For the subgroup of targeted CVP beneficiaries, the evidence of the baseline suggests that the voucher alone was not sufficient to ensure food security. The MoSA beneficiaries were on average in an even worse food security situation, not even having access to a voucher equivalent to a general ration. The unconditional nature of the cash transfer is in interesting contrast to the e- voucher approach of the CVP project and the larger background of long- term food distributions in the Gaza strip. As shown in Figure 4, the money is spent on a wide range of items beyond just food. All of the recorded expenditure is appropriate for households in this vulnerability group, although the monitoring tool does not capture the whole of household expenditure. The key impacts of the expenditure of the cash transfer are twofold. The first is a direct effect on food security: increasing the amount of food available in the household and increasing dietary Ben Mountfield Page 20 of June 2012

21 diversity. The second is to protect the household from the need to dispose of assets, and thus prevent (or at least fend off) destitution. This can be inferred from the increased spending power, and was confirmed during focus groups. To quantify this effect we would need to know about the history of asset depletion in the months prior to the programme, and make a comparison. The appropriateness of the short- term nature of the project is questionable, particularly for the reasons outlined in the section on gender relations above. For some of the MOSA beneficiaries the project has acted as valuable bridging support, and they have recently been able to access longer term support from the Ministry of Social Affairs. This cash support is provided every three months, and is between 750 NIS and 1,800 NIS depending on circumstances. (Compare to 780 NIS for this project in the same period). For those who remain on the waiting list, repeat access to this project would be appropriate in the second round. The economy of Gaza may be entirely dependent on imported commodities and aid on the one hand, but it is also vibrant and quite robust. It will not struggle to absorb a much larger injection of cash than it currently does: indeed, such injections of cash could be said to be the backbone of the economy. Beneficiary profiles TOR Objective: Identify other beneficiary profiles that require economic support to access basic needs for which unconditional cash would be the most appropriate mechanism. The evidence collected and examined suggests that project targeting is good despite some challenges. The many surveys undertaken in Gaza for a wide range of actors notwithstanding, it seems that the social safety net has not yet managed to sweep up all those it would hope to reach. The list provided by MoSA was of people who were, until recently, undiscovered, unregistered and in dire need. There is a word used to describe at least some of these people: Masturin, which implies being hidden, not pushing oneself forward to gain resources, too dignified to beg, being modest. Unfortunately for identification purposes, there is no evidence that there is a profile as such. In these circumstances the best approach will be to continue to get names from MoSA and from WFP, and act as a bridging support until there is capacity within MoSA to provide more long- term assistance. With WFP, the plan is to provide a complementary voucher and food distribution from 2013, for those in the bottom left section of Error! Reference source not found.. The monitoring form developed by project partner Ma an asks a question of the beneficiaries if they know of other people in similar situations. This could have two potential uses: to help to identify those who might be considered for programme support, and as a potential indicator of the depth of extreme poverty and food insecurity more widely. There is a risk of a gap around understanding the context and specific needs of the Bedouin communities in Gaza. There are Bedouin amongst the beneficiary groups, but not, it appears, amongst the project staff. It is not possible to be sure how many, as this is not data that is collected. They are under- represented in the Oxfam office. Ben Mountfield Page 21 of June 2012

22 Additional observations and findings The impact on household food security As noted in the introduction, three separate standard proxy indicators are included in the monitoring questionnaire. Results for two of these, which consider the diversity of foods eaten and the level of access to food generally, will be discussed under this heading; the third, which looks at coping strategies, will be tackled later in this section. Household Dietary Diversity Score The HDDS 2 is a tool that measures the number of different food groups consumed in the household. Dietary diversity is an attractive proxy indicator for food security for a range of reasons, and it is easy to collect. It uses a 24 hour recall period and asks about twelve food groups: Cereals Fish and seafood Root and tubers Pulses/legumes/nuts Vegetables Milk and milk products Fruits Oil/fats Meat, poultry, offal Sugar/honey Eggs Miscellaneous The HDDS works as a proxy measure of the socio- economic level of the household. The output of the tool is an integer between 0 and 12. Figure 7 below shows the results for the two sub- groups. Figure 7: distribution of Household Dietary Diversity scores 2 Swindale, Anne, and Paula Bilinsky. Household Dietary Diversity Score (HDDS) for Measurement of Household Food Access: Indicator Guide (v.2). Washington, D.C.: Food and Nutrition Technical Assistance Project, Academy for Educational Development, Ben Mountfield Page 22 of June 2012

23 This chart demonstrates that at the baseline, both groups had poor dietary diversity scores and the scores for the MoSA group were especially weak, with many households reporting scores of 3, 4 and 5 only. The baseline scores for the CVP group are better, which is encouraging as the CVP aims to improve diversity and choice. The impact of the project has increased household dietary diversity by a significant amount, especially in the lowest- scoring households. The effect is more noticeable for the MoSA group than for the CVP, with an average increase of 1.4 points. One of the key targets for the project was an increase in Food Consumption Score (which is effectively HDDS) for 90% of the targeted population. If we consider the 10% of the population that were fixed and monitored in each round (and can therefore be tracked individually) we see that every household in this subgroup showed an increase in HDDS, and the average increase was 1.3 points. Household Food Insecurity Access Scale The HFIAS 3 tool is a broader approach to measuring the levels of access to food at the household level. It is based on the evidence that the experience of food insecurity (access) causes predictable reactions and responses that can be captured and quantified through a survey and summarized in a scale. Qualitative research provided insight into the following ways that households experience food insecurity (access) which can be reduced to three main domains: Anxiety and uncertainty about the household food supply Insufficient quality which includes variety and preferences Insufficient food intake and its physical consequences The tool has been widely tested and found to be effective in a range of contexts. It takes the form of 9 questions, related to the domains above, each asked with a recall period of 4 weeks. Each question has two parts: an occurrence question, and a frequency question: In the past four weeks, did you? How often? The questions are organised in order of severity, and the answers are used to classify households into four categories. The results are not simply added up and presented as a score, as this misses some of nuances. For example, any positive response to the last three questions places the household into the seriously food insecure category. In the specific case of our project, this has made the tool a rather blunt instrument, pushing most beneficiaries into the same group. It does, however, demonstrate the severity of food insecurity within the target population. Figure 8: changes in Household Food Insecurity Access category Figure 8 illustrates an improvement for both subgroups, but perhaps not as substantial an improvement as might have been anticipated. This is due in part to the nature of the tool, and in part because of the frequency of positive answers to questions 7 and 8: 3 Coates, Jennifer, Anne Swindale and Paula Bilinsky. Household Food Insecurity Access Scale (HFIAS) for Measurement of Household Food Access: Indicator Guide (v. 3). Washington, D.C.: Food and Nutrition Technical Assistance Project, Academy for Educational Development, August Ben Mountfield Page 23 of June 2012

24 In the past four weeks, was there ever no food to eat of any kind in your household because of lack of resources to get food? In the past four weeks, did you or any member of your household go to sleep at night hungry because there was not enough food? This is highly consistent with the findings that for most households, the cash transfer was only sufficient for 2-3 weeks. A cash transfer sufficient for 4 weeks might produce very different results. The SEFSec survey contains a very similar tool: the WFP Food Consumption Score (FCS) which has a 7 day recall period and 20 food groups, and the complete HFIAS as well as the CARE version of the CSI (see below). This might provide a valuable community- wide baseline for comparison, and also provide the basis to compare the PMTF with recognised food security proxy indicators. Findings related to the monitoring tools The monitoring tool is large, and it takes some time and patience to administer. As noted above, it contains three largely complementary tools and a number of additional questions. Monitoring such a large amount of data is difficult in real time: much of the analysis is done retrospectively and is not available for decision making during implementation. Experiencing the three food security tools as a beneficiary may be confusing: they ask a number of overlapping questions but use three different recall periods: 24 hours for the HDDS, 7 days for the Coping Strategy Index and a month for the HFIAS. Cash transfer programmes tend to attract more scrutiny than others, perhaps unfairly. However, it is appropriate to monitor such a pilot project, in such a context very closely, and the generous amount of information produced provides an excellent database to better understand both the context and the impact of the programming. It may still be possible to review the monitoring tool to make it more focused. The result might be able to be used across a larger, more representative sample after each distribution. The timing of monitoring visits. It is apparent that the timing of the monitoring visit could have a significant impact on the nature of the results. Focus group discussions suggested that at normal levels of expenditure, the cash grants were depleted in around two weeks although some families chose to stretch them out over the whole period. In the former case, a monitoring visit in the first half of the month will provide quite different results from one in the second half especially for the HDDS and CIS results. With the exception of the first round, when the monitoring was late, and the absence of round 5, the timing of the monitoring was appropriate. Monitoring visits took place as planned around 21 days after the transfer became available. The final monitoring visit There remains the possibility of a final monitoring visit: no follow- up was made after the additional distribution round 7. There would be value in making such a follow- up 2-3 months after the final payments. The sample size should be increased: to cover a good number from each group and the MoSA group should include both those who have been taken into the main MoSA support programme, and those still on the waiting list. Gifts, borrowing, lending, and networks Standard tools for measuring or estimating household economy (which is difficult in any context) will run up against two additional challenges in Gaza. The first is that by focusing on income, expenditure and consumption they may miss the significant contributions that come in- kind as gifts. This can be tackled by careful design of the tools themselves. The second challenge relates to the distinction Ben Mountfield Page 24 of June 2012

25 between the concepts of family and household, and the different perspectives of people and agencies. Gifts and sharing Ad- hoc gifts within the family are difficult to measure but form a major part of the household economy in the poor households we visited. The monitoring systems are quite strong on debt, but not on gifts: nor on properly linking debt and expenditure. This may be because lending is usually currency based, while gifts are often in kind and harder to quantify. In the most part we can expect the beneficiaries of this programme to be the recipients of gifts, although there may be occasions when the reverse is true. There are also traditional expectations that people will pay Sunna on gifts and large purchases although this is not thought to apply to the cash transfers and zakat of 2.5% of cash amounts. Payments and receipts through these mechanisms are not monitored, and while they probably don t often apply to these beneficiaries, this should be validated. The perception of family in Gaza is much broader than the notion of household used in humanitarian sector. Flows of resources - money, food, clothing, furniture, household items between family members may not be seen as gifts but simply as sharing. A net flow of goods, especially food, from the household to the family would imply that the grant and inputs are being shared more widely than intended and their impact may be reduced; also that there are others in the family in a position of significant need. A net flow of gifts from the family to the household may imply that the inputs alone are not sufficient but that other family resources are complementing them. Borrowing and lending Borrowing and lending are much more formalised, and are normally currency- based transactions. There is an expectation of repayment, although some debt within the family is probably nominal. Debts are a normal part of the household economy and there does not appear to be any substantial stigma attached to modest debts. Borrowing requires either collateral or credit- worthiness. Most of the families visited would not have been able to get goods on credit from a local shop prior to the programme; some still cannot. However, shops are not the only source of credit and as seen in Figure 5, levels of debt prior to the programme were quite significant and higher than the apparent comfort level. Family support networks Households and families do not exist in isolation. In Gaza, the ties within families are strong, and this contributes in no small part to the resilience of the Palestinian people. This section will discuss resources (primarily meaning money, food, clothing and household items) and the way they move about within the family network in response to various stimuli. It introduces the concept of a support network, meaning the sum of ways in which a family receives resources that it does not earn or produce itself, it presents a model of the network that might be tested through the next phase of the project, and it offers a hypothesis of what factors might influence the flow. While there remain sufficient resources within the family support network, a portion of them they will flow out to the point where they are needed. Rather like water flowing downhill, resources can flow down a socio- economic gradient within a family structure. Indeed, they also flow beyond the family, to neighbours and others, or through donations to mosque or zakat. The outcome of this is to compensate to a degree at least for any deficits in people s ability to provide for themselves and for deficits in aid provision. If we provide inputs ourselves directly to a household, then a portion of those resources may become available for others. This may allow onward flow of resources to an even lower part of the family network we had not identified, or may even (in the case of a substantial transfer, the creation of a job or a good production project) reverse the flow. We are not neutral observers of this network: we become a part of it the moment we inject resources. The flow of such resources will probably depend on three main factors: Ben Mountfield Page 25 of June 2012

26 The proximity of the recipient household to the wealthier parts of the family, in terms of birth, marriage, etc. The difference in socio- economic status between these two points and the absolute amount of resources available The quality of the relationships between the key individuals, and along the chain between them. Figure 9 below attempts to illustrate this hypothesis. Figure 9: schematic of household support network Discussions with individuals and focus groups suggest that resources normally move through the network one step at a time, except perhaps at times like funerals and medical emergencies. This model may help to explain, for example, why loans within the family are sometimes described in unclear terms. There is perhaps an expectation for recipients to pay them back, but not unless and until their situation improves sufficiently to make that realistic. As the overall level of resources in the network drops, so does the reach of any possible surplus. There are a number of reasons to think this might be the case at the current time. The data within the project shows that assets are still being sold. Land holdings are divided at each generation and plot size (and thus yields per family) are reduced. Availability of affordable housing is outstripping supply. If we can identify an indicator for the health of this broader support network (the sum of the support networks at household level) this could be very valuable. We don t know how to measure the health of such a network, let alone the many intersecting networks that connect different households in Gaza. We would need to use a proxy for the network as a whole, and we would need to measure it across a representative sample of poor households, and this would require an investment. However, it may prove that data already being collected through the PMT questionnaire, for example, is adequate, or that a small addition could provide a suitable indicator. It seems likely that a single well- chosen proxy indicator might be more valuable in this regard than a complex analysis of many inter- related factors. Ben Mountfield Page 26 of June 2012