Community-Based Cash Transfer in fragile and conflict-affected communities

The objective of this research project is to understand the impact of cash transfers to beneficiary households through alternative targeting approaches for social assistance programs in conflict-affected settings. While the effectiveness of cash transfers has been widely discussed in the literature, in conflict settings, the choice of targeting methods brings additional challenges to the effectiveness of cash transfers. More specifically, the effectiveness of cash transfers may vary depending on how they are targeted. Over the last decade and half, with support from development partners, Ethiopia has implemented one of largest social assistance programs in Africa - known as the Productive Safety Net Programme (PSNP). The program reaches about 8 million rural people living in food insecure communities in the country. During much of this period, Ethiopia was characterized by relative stability and positive socioeconomic changes. The country has recently plunged into political unrest and recurrent conflicts, leaving millions in dire emergency and social assistance needs. This surge in the numbers of people in need of assistance, coupled with resource constraints faced by international aid agencies like the WFP and other development partners that necessitated a rethinking of existing targeting approaches, including community-based and other data-driven approaches such as PMT. The existing literature does not offer clear guidance as to which targeting methods are suitable in conflict-affected settings with limited data availability. Moreover, it is unclear whether and to what extent household level impacts of transfers are sensitive to the choice of targeting approaches. Our study aims to contribute to the broader discourse on impacts of alternative targeting strategies in conflict-affected and fragile settings.
We work with communities and community leaders in 180 Enumeration Areas (EAs) or villages across Ethiopia. An EA typically comprises 150 to 200 households within a Kebele, the lowest administration unit in Ethiopia. We build on a household survey conducted in 2019 comprised of a random sample of 20 households from each of the 180 sampled EAs. The community survey brought together six individuals composed of key Kebele leaders, including the Kebele chairman and others knowledgeable about the EA, such as community leaders, elders, priests, and teachers. Mimicking the actual targeting practices in Ethiopia and beyond, village leaders were assigned to use of one of different targeting approaches to select beneficiary households from their respective EAs and allocate cash to each beneficiary according to chosen targeting approach. Thus, the intervention in this study occurs at two levels: the community leaders and the community members. In this study, we focus on the household level impacts following the community level interventions.
The Intervention: allocation of real versus hypothetical cash transfers by community leaders
By way of mimicking traditional community-based targeting, we ask community leaders to rank households in their respective EAs from the most to the least needy based on their needs assessment for social assistance. We provide community leaders with a lump-sum of cash that will be transferred to their respective community members (20 households in each EA) ranked based on the targeting approach to which each community is randomly assigned to different budget categories and discretion levels built into the experimental design (see Experimental Design section). That is, we exogenously vary the nature of the transfers (hypothetical versus real), the amount of money available for transfers, and the level of discretion granted to community leaders. These would allow us to study the impact of different levels of transfers made through limited versus relaxed targeting discretion given to community leaders. Specifically, we aim to answer the following research questions.

Research Questions
This research project aims to evaluate the impact of cash transfers disbursed through alternative designs to CBT on the breadth and depth of social assistance transfers on several household level outcomes, including subjective wellbeing, mental health, and perceived stress level and depression, short-term food insecurity, and dietary diversity. Below is the list of specific research questions associated with these broad research objectives:
1) Do Community-Based cash transfers improve short-term food security, consumption of nutritious diets and dietary diversity?
2) Do Community-based cash transfers improve subjective well-being of beneficiaries in conflicted-affected settings?
3) Can relatively small community-based cash transfers improve mental health and stress in the face of recurrent conflicts? And to what extent does the size of transfer matter?
4) Do welfare impacts of community-based cash transfers vary across by the nature of the targeting (e.g., mandated or discretionary) approaches, or restricted or relaxed budgets?
5) Are community-based transfers more impactful for improving welfare outcomes in conflict affected communities?
6) What is the impact of community-based cash transfers on trust in local governance and social cohesion in the presence and absence of conflicts?
7) Do the impact of community-based cash transfers vary across gender and across areas with varying exposure to conflict?

2023-11-10

2024-01-31

1. Household Dietary Diversity Score (HDDS)
2. Women’s dietary diversity score (WDDS)
3. Consumption of nutritious diets
4. Per capita consumption expenditure
5. Household Food Insecurity Experience Scale (FIES)
6. Subjective well-being
7. Perceived stress level
8. Depression scale

Primary outcomes explanation
1. Household Dietary Diversity Score, which will be constructed based on 7-day recall following the guidelines by Food and Agriculture Organization (FAO). The HDDS is a qualitative measure of household-level food security and hence reflects the economic ability of the household to access a variety of foods (FAO, 2023). Previous studies have shown that an increase in dietary diversity is strongly associated with household food security (e.g., Hoddinott and Yohannes, 2002; Hatloy et al., 2000). Respondents are first asked if household members have consumed one or more of the food groups over the preceding week. Then, the items are re-categorized into 12 groups to arrive at a score which consists of a simple count of food groups consumed, ranging from 0 (no consumption of any group) to 12 (consumption of all groups).
2. The women’s dietary diversity score (WDDS) comes from a report of foods consumed in the last 24-hours. The WDDS captures the number of food groups consumed in the previous day by women of reproductive age. The food groups in reference are grouped into 10 groups. The 10 food groups include: (1) grains, roots, and tubers; (2) legumes and beans; (3) nuts and seeds; (4) dairy products; (5) eggs; (6) flesh foods, including organ meat and miscellaneous small animal protein; (7) vitamin A-rich dark green leafy vegetables; (8) other vitamin A-rich vegetables and fruits; (9) other fruits; and (10) other vegetables. The value of DDS ranges from 0 to 10.
3. Consumption of nutritious diets, which is constructed from the set of questions on weather women of reproductive age consumed high-value (nutritious) foods in the 24 hours preceding the survey. For each of the high value consumption groups comprising dairy products, eggs, and meats, the outcome variable is constructed as a dummy indicator that takes value 1 if a woman consumes a high value item (e.g., eggs), and 0 otherwise. Because these measures are related to the WDDS, they will be analyzed together.
4. Consumption expenditure is measured and elicited using household consumption expenditure module, which collects information on households’ consumption of various food and non-food items. Household consumption expenditure is widely used as a proxy for household well-being and incomes, based on the assumption that a household’s consumption and income are strongly correlated.
5. Food Insecurity Experience Scale (FIES) is a self-reported metric which captures households’ access to adequate food and associated difficulties due to financial or other resource constraints. The FIES is an experience-based food insecurity metric developed by the FAO of the United Nations and is widely applied to measure perception and prevalence of food insecurity (FAO, 2014; FAO, 2020). The FIES builds on an eight-question module related to respondents’ experiences and associated difficulties to access sufficient and nutritious food in the last 30 days. The aggregate FIES is constructed by summing the responses to the eight questions. Its value ranges from zero to eight, zero standing for those households reporting no experience of food insecurity across all eight dimensions of food insecurity. Based on the various indicators and questions used to measure Food Insecurity Experience Scale (FIES), we also aim to generate an indicator variable assuming a value of 1 if the household experiences one or more types of food insecurity and 0 otherwise.
6. Subjective Well-being: this is measured using an ordered indicator of overall life satisfaction. This scale ranges from 1 (“completely dissatisfied”) to 10 (“completely satisfied”).
7. Perceived stress level. We use two measures of perceived stress level. One that captures the overall perceived stress of respondents related to everything in their life, like work, family, health, and so on. This is a scale ranging from 1 (not stressed at all) to 10 (extremely stressed). A second and more comprehensive stress measure is constructed from 10 questions in the standard stress assessment instrument Perceived Stress Scale (PSS). Respondents are asked 10 questions on their feelings and thoughts over the last month.
8. Depression scale. We measure depression scale using responses to the Patient Health Questionnaire (PHQ-9). The PHQ-9 consists of 9 questions that ask respondents how often they have been bothered by a case-finding problem item (e.g., little interest or pleasure in doing things) in the past two weeks. Each item is scored on as 0 – not at all, 1 – several days, 2 – more than half days, and 3 – nearly every day. These scores are then summed to generate a depression scale.

1. Trust in local governance, trust in community leaders and other stakeholders in the village
2. Social cohesion among households

Experimental design
The intervention follows community level clustered randomization in which the 180 communities are randomly assigned into control group or one of three treatment arms. The treatment assignment is based on (i) whether communities receive actual transfer or hypothetical (control); (ii) the size of the transfer pool available to community leaders to be distributed among households within the community (constrained budget involving 10,000 Birr versus relaxed budget of 20,000 Birr); and (iii) the level of discretion given to community leaders (rule-based or discretionary). Effectively, the first stage randomization involves assigning the 180 communities into a control and cash transfer group. We assign about 30 percent of the communities to control group and the remaining 70 percent into treatment group. We further split the treatment group communities (enumeration areas) into one of three treatment arms: Relaxed budget with rule-based approach for distributing resources (i.e., ETB 20,000), relaxed budget with some discretion on the criteria to distributing cash transfer (i.e., ETB 20,000); constrained budget with rule-based approach for targeting (i.e., ETB 10,000). Figure 1 shows the random assignment and associated control and treatment groups.

Figure 1. Random assignment of communities across treatment and control arms
(1) Control: Rule-based targeting using hypothetical transfer of 20,000 Birr (C): This group serves as a control cluster where community leaders are not given any actual funds but are instructed to act as if they have a hypothetical budget of 20,000 Birr to distribute among households in their community. Community leaders are first asked to rank households based on their need for social assistance. They are then asked to allocate this notional budget among the 20 households included in our sample. During this ranking process, leaders are required to strictly adhere to pre-defined rules provided by the research team. These rules are carefully selected to mimic the targeting criteria used in actual social assistance programs in Ethiopia. More specifically, community leaders are asked to prioritize those households who: (i) had difficulty satisfying their food needs; (ii) own no or little asset (e.g., livestock, land); (iii) have limited income-generating activities or capacity; (iv) have lost productive assets due to shocks (e.g., conflict, drought); and (v) have lost family members recently. Effectively, for the current purpose of the analysis, this group receives no actual transfer and hence serves as our control group.
(2) Rule-based targeting with relaxed budget (T1): Another group of communities are randomly assigned to a cluster that receives real transfer funds with a budget of 20,000 Birr (about 360 USD). In this cluster (i) Community leaders are required to rank households based on five pre-determined targeting criteria and allocate the transfers. These criteria are similar to those in the control group and mimic the targeting criteria used by the national safety net program in Ethiopia (e.g., Gilligan et al., 2009; Hoddinott et al., 2012; Abay et al., 2022).
(3) Rule-based incentivized targeting with constrained budget (T2): This group of communities follows similar rules as those in control group, but they receive a constrained budget of 10,000 Birr (about 180 USD). Community leaders are required to rank households based on the five criteria outlined above and allocate the 10, 000 Birr to the community members in our sample. These criteria are designed to mimic the targeting criteria used by the national safety net program in Ethiopia, the PSNP. This treatment arm serves to test the implication of budget constraint.
(4) Discretionary targeting (T3): The fourth group of communities are provided a budget of 20,000 Birr to distribute as social assistance to households identified as in need. Here, community leaders rank households based on their own criteria they collectively agree upon. The establishment of these ranking criteria is entirely left to the discretion of the community leaders. It is up to the leaders to determine who among the ranked households gets how much of the 20,000 Birr transfer assigned to the community.

The randomization was done at the village level using the baseline list of villages. We initially selected those villages that are accessible for a survey, and we randomly assigned these villages into four groups. A reserve list was also prepared in case some of the villages are not accessible because of conflict.

Village or community level

Yes

180 communities s

About 3,060 households

C (Control): 53 villages
T2 (Rule-based, 20, 000 ETB): 41 villages
T3 (Rule-based, 10, 000 ETB): 42 villages
T4 (Discretionary, 20, 000 ETB): 44 villages

We compute the number of clusters needed for the primary outcomes described above, assuming that there are a known and fixed number of households in each cluster (village). The baseline sample includes an average of 20 households in each village. We assume an attrition rate of 15 percent and hence we expect to revisit an average of 17 households per village. Our power calculations aim to achieve 80 percent power at a significance level of 5 percent. Power calculations are performed only for the primary outcomes described above. Given that we have primary outcomes, we computed the number of clusters and associated sample size needed for each primary outcome separately, and then selected the maximum sample needed to detect impacts across these outcomes. We focus on quantifying the impact of the community-based cash transfers on short-term welfare outcomes. Our outcomes of interest include: households’ dietary diversity score (HDDS), women’s dietary diversity score (WDDS), consumption of nutritious foods, food security, consumption expenditure, subjective well-being, and perceived stress level. We compiled mean and standard deviation of these primary outcomes as well as minimum detectable effects (MDEs) for each outcome using the baseline sample as well as other external data and evidence from previous studies. The mean HDDS in our baseline stands at a very low level of 4.5 food groups. Previous studies that evaluated comparable cash transfers programs report average impacts ranging from 0.12 to 0.33 or a 6-12 percent increase (Hidrobo et al., 2014; Savy et al., 2020; Leight et al., 2023). In our study, an assumed 11 percent increase in HDDS requires 51 control clusters and 120 treatment communities. As we are comparing the hypothetical arm, where households receive no actual transfers, with the rest of the treatment arms which involve actual transfers to households, this comparison is sufficiently powered since we can combine the three treatment arms with actual transfers. The 180 clusters (53 control and 127 treatment group) allow us to detect the assumed MDEs for 6 of the 7 primary outcome variables listed in Table 1. Table 1 summarizes our power calculations involving several primary outcomes meant to test alternative hypotheses.