Community CCT in Indonesia The Generasi Project

Transcription

1 Community CCT in Indonesia The Generasi Project November 12 th, 2008 Junko Onishi

2 Two Pilot Projects In 2007 GoI started two pilot projects: Household CCT the traditional model Quarterly tranches of cash transfers Statistically identified 500,000 poor households with children 500 sub-districts in 68 districts Generasi Community Block Grants Addresses the same health and education indicators 178 sub-districts covering approx. 300,000 poor households with children

3 Why two approaches? HH-CCT has proven to be an effective tool However, Indonesia is very different from LAC countries: Patterns of vulnerability - a large cluster of the near poor, Poor targeting in previous cash transfer and other programs, Poor provision of social services in poor regions, Lack of national leadership to bring supply-side and demandside efforts together, Weak institutional capacities to administer complex programs, Lack of interest for quality monitoring required for an effective CCT implementation Indonesia s 10 years of experience with CDD programs has proven to be a powerful tool for resource transfer to poor communities

4 Why Generasi Project? Takes full advantage of the existing effective machinery and social capital built through 10 years of CDD Transferring funds to communities allows solutions to common and shared hurdles, not individual or household problems Communities self-diagnosis coupled with flexibility in funds allows investments in locally appropriate solutions Communities address demand-side problems as well as small-scale supply-side problems through collective action Participatory planning allows flexible local targeting without relying on central statistical systems Strengthens CDD project implementation by placing financial incentives to provide motivation and focus on key indicators

5 The Generasi Project Objectives: Accelerate the achievements of MDGs Reduce maternal mortality Reduce child mortality, and Ensure universal coverage of basic education Conditionalities: Places incentives on communities to identify problems and seek solutions to improving the 12 health and education indicators. Implementation: Implemented through the national CDD program.

6 Generasi Project Design 12 indicators: communities are required to work on improving these indicators Health: 1. Four prenatal care visits during pregnancy 2. Taking iron folate tablets during pregnancy 3. Delivery assisted by trained professional 4. Two postnatal care visits 5. Complete childhood immunization 6. Ensuring monthly weight increases for infants 7. Regular weighing for under-fives 8. Taking Vitamin A twice a year for under-fives Education: 9. Primary school enrolment 10. Regular primary school attendance 11. Junior secondary school enrolment 12. Regular secondary school attendance

7 Generasi Project Design Geographical coverage: 178 sub-districts in 21 districts, five provinces About 450,000 poor households Approx 300,000 poor households with children Block grant amounts: 2007 average per village amount 8,400 USD 2008 average per village amount 11,600 USD Timeframe: First block grant disbursed to villages in Nov 2007 Second year disbursement to villages in Nov 2008

8 In 2007 villagers used their funds for 56% of block grants on education: School materials, equipment and uniforms (59%) Financial assistance and school fees (30%) Infrastructure (satellite classrooms and access roads) (13%) Financial incentives for part time teachers (3%) Training and behavior change communication (3%) 44% of block grants on health activities: Supplementary feeding activities (40%) Financial assistance for pregnant mothers to use services (30%) Infrastructure (13%) Facilities and equipment (11%) Training and behavior change communication (3%) Incentives for health workers (3%)

9 Design challenges and innovations Objective: to design a CDD project directly comparable to HH-CCT Conditionalities = 12 indicators Adapted from HH-CCT conditionalities Easy to understand for communities Measurable by communities Follow national health & education protocols Village monitoring of 12 indicators Coupon books for pregnant mothers & infants Stamps for health providers Copying attendance books from classrooms

10 Design challenges and innovations Financial incentives = inter-village yardstick competition Fixed amounts of block grants to subdistricts Year 1 block grants: distributed to villages according to population in target age groups Year 2 block grants: 80% according to population of target age groups 20% according to year 1 performance on the 12 indicators

11 Design challenges and innovations Key advantages of having fixed bonus pool per subdistrict: Ensures budget predictability for the government Common shocks across subdistricts cancel out Villages within subdistricts tend to be similar Ensures each subdistrict gets same amount of funds i.e, prevents funds from being concentrated in richer subdistricts Provides small incentive for cross-village monitoring to make sure villages accurately report indicators

12 Challenges in implementation Health-supply shortages particularly in poor districts: Lack of health providers Vaccine stockouts, supplied centrally Guiding communities to target their investments on improving the indicators rather than spending on general health and education activities Monitoring of indicators is labor-intensive and requires strong facilitation Are the indicators/targets the right ones? In particular, education indicators need rethinking: Near universal primary school enrollment in Indonesia School attendance is hard to monitor and to verify

13 Impact Evaluation Uses a randomized evaluation Randomizes subdistricts into three groups: with performance incentives, without performance incentives, and controls Subdistrict randomization addresses spillovers and crowding out Design structure allows comparison with HH-CCT Common survey instruments for HH-CCT and Generasi evaluation

14 Impact Evaluation Three rounds of surveys: Baseline (2007) One-year followup (in progress in Generasi locations) Two-year followup (2009) Survey design: 26,000 households per wave spread over 660 subdistricts including Anthropometric measurements of < 3 Math and Indonesian tests administered to school-aged children School and health provider interviews to track supplyside effects Qualitative studies to understand bottlenecks in use and provision of services

15 Early Findings In general: Generasi startup was smooth & continues to run well, Communities understand the indicators and the performance incentives, Institutional machinery is entirely within the grasp of the executing agency, Anecdotal evidence indicate that a there has been a big leap in the number of mothers and their children receiving health services regularly and on time, Strong support from local health & education departments who appreciate village-level data and community mobilization.

16 Early Findings On performance incentives: Generasi s performance incentives affect communities block grant allocation. Subdistricts with performance incentives: Spent less on school uniforms (5.3 percentage points) Spent more on health (4 percentage points) specifically on: Supplementary feeding (3.7 percentage points) Infrastructure (2.5 percentage points) Equipments and supplies for weighing posts (0.7 percentage points)

17 Food for thought What are the trade-offs and synergies between a HH-CCT and a community-based approach such as Generasi? How replicable might Generasi be in Africa or in other SE Asian countries? Generasi s preliminary one-year impact evaluation results will be available June-July 2009.

18 How competition works Specifics of yardstick competition: Set relative weights/prices for each indicator (wi) For each village, set minimum achievement level based on number of children and access to services (mvi) Set at 70% of predicted average achievement level Measure achievement on each indicator each year (yvi) Total village bonus points: Pv=Σ[ wi (yvi - mvi)] Village s allocation depends on their bonus points, relative to total points achieved in subdistrict, i.e., Pv / (Σ Pj)