Targeting in a National Social Safety Net Programme WFP Turkey
Emergency Social Safety Net Background EU funded nationwide assistance programme to refugees in Turkey Registration: Ministry of Interior Implementation: Ministry of Family & Social Policies Aligns with national social welfare system Objective: Meeting basic needs Targeting 1,000,000 vulnerable refugees ATM card uploaded monthly Unconditional unrestricted cash transfer of 100 TL (~$30) per person
Agenda 1. Establishing Criteria 2. Implementation & Practical Constraints 3. Analysis: Inclusion/Exclusion Errors 4. Process: Revising targeting methodology 5. Q&A
1. Establishing Criteria: Systems National Systems MoFSP provides assistance by specific eligibility criteria e.g. disability, pregnancy, education Applications submitted at local Social Assistance Solidarity Foundation (SASF) All applicants receive a visit by SASF officer. Complete 12 page form, determine eligibility based on responses and visual assessment of household ESSN Systems Use sub-set of essential questions within MoFSP application form All applications assessed immediately for eligibility Followed by household visit within one year: addresses inclusion error Select questions from application form that are verifiable, and allow reliable identification of vulnerable households Select thresholds that allow selection of one million refugees (of ~three million)
1. Establishing Criteria: Analysis Only available data for eligible populations: DGMM demographic data (MoI) VAM mission in March 2016: analysis of available data Working Groups in April 2016 (WFP, UNHCR, UNICEF, TRC) Agreed upon criteria using: Regression analysis (statistically significant predictors of HH welfare, defined as per capita expenditure) WB UNHCR Welfare of Syrian Refugees Internationally used criteria
1. Establishing Criteria Criteria Definition Dependency ratio >1.5 ((0-17 yrs + 60+ yrs) / (18-59 yrs)) > 1.5 Large Families Any household with four or more children (<18 yrs) Two disabled members Single Parent Household Elderly Headed household (>60 years) Any two members of household are disabled (requires medical certificate showing at least 40 percent disability to align with the national system) Not accompanied by other adults (18-59yrs) and with at least 1 child (<18 yrs) Not accompanied by any other adult (18-59 yrs) Single female: Not accompanied by other household members (1 person HH) Plan to have additional referral systems to capture exceptions to these rules
1. Establishing Criteria: Eligibility 26 December 25 January: Eligible Households (19,145) 26 Dec - 25 January: Eligible Households(19,145) by Critiera Dependency Ratio 81.27% 4+ Children 77.38% Single Parents 20.97% Single Female 2.92% Elderly Headed 2.37% Two Disabled 0.15% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90%
2. Implementation & Practical Considerations (1) 1. DGMM data (Ministry of Interior): Intended use: verify HH demographics declared on application Requires: Integration of MoI system with MoFSP system Challenge: Significant delays in integration Data quality issues Unverified. Must have IDs for all declared family members. Can choose not to declare family members in application.
2. Implementation & Practical Considerations (2) 2. Lack of transparency/accountability No publication of criteria due to lack of verification Lack of understanding around eligibility confusion, anger, re-applications, crowding
2. Implementation & Practical Considerations (3) 3. Household vs. Family Turkish national system: assessed by address Refugees: many families living within one address SASF staff unfamiliar with splitting families within address; many not doing this 4+ children criteria - unreliable 4. Nationwide Implementation: Applications accepted at 980 SASF offices Each office with own practice, application procedures vary 5. Referrals System: No referrals system established in or out In: challenges accepting NGOs, establishing reliable system Out: no practical system in place. Referring to what? By whom? How? Results in 100% reliance on demographic criteria = exclusion error 6. Lack of data for analysis!
3. Analysis: Inclusion & Exclusion Errors Exclusion Error: poor, vulnerable households who do not meet the demographic criteria Inclusion error: wealthier households who meet the criteria. Addressed through SASF household visits Quantifying errors: per capita expenditure, FCS
Baseline: Exclusion Error Expenditure Exclusion/Inclusion: Baseline Data (1562 HHs) Wealthiest 4 3 2 Poorest 63.3% 54.3% 45.5% 33.1% 23.7% 36.7% 45.7% 54.5% 66.9% 76.3% 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0% Excluded Included Exclusion Error: 23.7% of poorest households excluded = 4.74% of total population Bottom two quintiles = 11.37% of total population
Baseline Expenditure & Food Consumption Baseline Data Percentile Group of per capita expenditure Eligibility FCG Poorer 2 3 4 Wealthier ESSN ineligible Poor 7 8 14 11 4 Borderline 32 40 43 44 24 Acceptable 47 69 105 138 187 ESSN Eligible Poor 18 11 7 6 5 Borderline 57 44 32 23 10 Acceptable 146 135 113 89 80 PAB # % of total pop Exclusion Error 275 17.75% Inclusion Error 282 18.21%
Summary of Results Three datasets Exclusion of the very poorest (0-20%) households: 1.76 5.33% of total population Exclusion of poor (0-40%) households + poor, borderline FCS: 9.16-17.7% of total population Likely higher these quintiles are relative within a dataset of mostly poor households
Addressing Exclusion Error: Socioeconomic Criteria Based on WB/HCR Welfare of Syrian Refugees 1. High Crowding 2. House without proper latrine and/or kitchen 3. Substandard housing (defined as not concrete in WB report) 4. Not paying rent or owning accommodation (almost all pay rent here ) 5. Education level of HH head (not available in our datasets)
1a. Crowding: Evidence Number of individuals in residence/ Number of (sleeping) rooms in residence Per capita living space - significant predictor in regression analysis Clear link between HH welfare and crowding Verification: Crowding (people per sleeping room) PAB: Crowding (people per meter squared) 4.50 4.00 3.50 3.00 2.50 2.00 1.50 1.00 0.50 4.08 3.18 2.60 1.98 1.48.14.12.10.08.06.04.02.13.12.10.09.08 0.00 Poorest 2 3 4 Wealthiest.00 Poorest 2 3 4 Wealthiest
2a. Housing Facilities Running water and electricity in almost every HH (only 1 without) Internal toilet (83.7%); Kitchen (85%) linked to HH welfare Housing Facilities by wealth 100.0% 90.0% 80.0% 70.0% 60.0% 50.0% 40.0% 30.0% 20.0% 10.0% 93.20% 69.7% 69.60% 95.40% 0.0% Internal Toilet Kitchen Poorest: bottom two quintiles Better off (40-100%)
2b. Facilities in Residence: Families not meeting demographic criteria Electrical Infrastructure in Residence Yes 99.7% No 0.3% Toilet Facilities Acceptable Toilet (inside) 94.4% Unacceptable Toilet (outside) 5.6% Application Form: What facilities are available in your residence? Bathroom or shower Toilet (inside the house) Kitchen Central Heating Electricity System Water tank Piped water system (tap water) Natural gas Hot water Cable TV Elevator Parking Garage
3a. Housing Type Residence Type Frequency Percent Apartment 1469 51.1 Barn 1.0 Unfinished building 25.9 Basement 56 1.9 Tent 10.3 Store - unused shop 198 6.9 Slum 534 18.6 Detached house 568 19.8 Abandoned construction 8.3 Other 4.1 Total 2873 100.0 Substandard Housing: All Assessed Households 51.22% Sub-standard housing 48.78% Acceptable housing
3b. Housing Type Did not track well with other variables categories is too subjective/ open to different interpretations by field staff 80.0% Substandard housing 75.9% 70.0% 63.3% 60.0% 50.0% 55.2% 51.7% 40.0% 40.0% 30.0% 20.0% 10.0% 0.0% Poorest 2 3 4 Wealthiest
Defining Additional Criteria: Thresholds & Challenges Application data is required in order to understand the percentage of households who meet each criteria To set threshold, we need to establish a combination of criteria i.e. a household who meets 3 of 5 of the socioeconomic criteria will be eligible. Ability to do this depends on the quality of the data have all applications reliably included these variables? I.e. does 0 facilities really indicate that the household has no facilities, or simply they did not ask that question?? Key Challenge: No source of verification. Requires household visit prior to determining eligibility? Operationally feasible? Implications on publicising criteria
Addressing Exclusion Error: Other 1. Burden Index Used in Lebanon Different weightings for different HH members: gender, age, ability 2. Referrals Any system relying exclusively on statistics will have some exclusion error Human element to address exceptions to rules Current system disadvantages less educated, less connected Balance: standardisation vs. subjectivity
4. Process: Targeting Revisions 1. Targeting Working Group with all stakeholders to discuss options 2. Application data analysis: determine reliability, thresholds 3. Consultations with SASFs: Possibilities for new criteria, implications on process and household visits Political acceptability of referrals, and operational implications 4. Targeting Working Group to endorse proposal for changes 5. Presentation to ESSN Governing Board
Questions?
World Food Programme Via C.G. Viola, 68/70-00148 Rome, Italy