from EVIDENCE to POLICY Learning what works for better programs and policies March 2016 INDONESIA: Can Community Preschools Improve Poor Children’s Development? Early childhood education is recognized as important for The World Bank and the international development children’s development, especially for those from poor or community have made early childhood development otherwise disadvantaged backgrounds. Increasingly, gov- programs a central part of plans to help improve people’s ernments and development lives. Knowing what makes a successful program is key to groups are grappling with how getting services to those who need them. In Indonesia, to create cost-effective services researchers evaluated a project to expand access to early that provide the stimulation childhood services in the country’s poorest areas by giving and support children need for communities grants for preschools and providing teacher healthy brain development training and facilitators to encourage use of services. The EDUCATION and growth. One problem is evaluation found that this project boosted enrollment and ensuring that poor children children’s development, especially for those from the most have the same access to early disadvantaged backgrounds. Now, with support from the Photo by Amer Hasan childhood services as those Strategic Impact Evaluation Fund, the research team from families who are relatively better off. Most evidence is going back to see whether the early childhood ser- of successful early childhood programs focuses on compre- vices created with the project’s support are still func- hensive programs that involve a combination of home vis- tioning and how they are sustaining themselves given its and preschools. In communities that have few resources, that project support has ended. As the Government of can increasing access to preschools lead to improvements Indonesia considers expanding compulsory education in enrollment and children’s development, especially among to include pre-primary school, the results can provide the poor? important evidence for how to do this. Context Indonesia offers a variety of early childhood education ser- vice, compared with an eight percent chance for those from vices. Playgroups, for example, target children ages three to the poorest households. four; kindergartens are aimed at children ages five and six. Pri- The Government of Indonesia has been taking active steps mary school, which is compulsory and free, starts at age seven. to redress this imbalance. Working with the World Bank and While enrollment in primary school is universal in Indonesia, the Government of the Kingdom of the Netherlands, it devel- access to preschool has long been unequal. Preschools often oped a project to expand access to early childhood services in charge a fee and the children who attend are usually from rural areas. The project, launched in 2009, provided selected wealthier families. For instance in 2007, children between the poor, rural communities with resources they could use to ei- ages of four and six from the wealthiest households had a 33 ther create new early childhood programs or improve exist- percent chance of being enrolled in an early childhood ser- ing ones. Communities received facilitators to raise awareness of the importance of early childhood education, training for to six. While 21 percent of communities used the grant to teachers and three-year grants for early childhood services. strengthen existing services, 79 percent of communities used The grants could be used to provide services for children the grant to start new services. The government envisioned from birth up to the age of six, including paying for teachers, that after the project closed, district governments would con- and most villages set up playgroups for children ages three tinue to support these services. Evaluation The project was implemented in 50 of the country’s 422 and 2013. This note focuses on the results from compar- districts. These districts were picked based on a ranking that ing villages that received the project to those that didn’t. incorporated, among other things, poverty rates and low The researchers looked at two groups of children. One levels of enrollment in early childhood programs. In each group was four years old when the project was launched in district, 60 villages were picked based on their poverty rates, the villages. They were immediately eligible to enroll in the the number of young children, and the village’s willingness playgroups established by the project. Children in the sec- to contribute financially to the project. In total, the project ond group were a year old when the project was launched. was introduced in 3,000 villages, or some four percent of They became eligible to enroll in playgroups two to three EDUCATION the country’s 69,000 villages. years after the project had been started. (It was decided to The researchers used the planned roll-out of the proj- include one year olds in the surveys because it was expected ect to design an impact evaluation that would allow the that some villages would create new services for young chil- effect of the project on children’s learning and develop- dren; in fact, almost all villages focused on playgroups for ment to be measured. Using a public lottery, villages were ages three and older.) By the second round of data collec- randomly assigned to receive the project in 2009 and tion, the children who had been four when the project was 2010 for the purposes of the impact evaluation. In total launched were generally old enough to move to a kinder- 218 villages were assigned to receive the project and be garten. By the third round of data collection, these chil- tracked for the impact evaluation. The team also collect- dren were all in primary school. In contrast, the children ed data on a comparison group of villages, which never who were one year olds when the project started were old received any services under the project. This group con- enough to enroll in playgroups by the time of the third sisted of 92 villages. Data were collected in 2009, 2010 round of data collection. Results Children in villages that received support through Enrollment in any kind of early childhood service was the project were more likely to be enrolled in 22.2 percentage points higher among children who were an early childhood program and were enrolled one year olds when the project was launched, compared for longer than children from villages that didn’t with one year olds in villages that didn’t get the project. receive the project’s support. Children who were four years old when the project started had a 7.5 percentage point increase in their use of early This police note summarizes World Bank Policy Research Working Paper 7372, “The Impact of Expanding Access to Early Childhood Services in Rural Indonesia: Evidence from two Cohorts of Children,” Sally Anne Brinkman, Amer Hasan, Haeil Jung, Angela Kinnell, Menno Pradhan, Education Global Practice Group, July 2015. http://documents.worldbank.org/curated/en/2015/07/24808116/impact-expanding-access-early-childhood-services-rural-indonesia-evidence-two-cohorts-children childhood services, compared with similarly-aged children ing). In fact, children from less poor families scored worse in villages that didn’t get the extra support. on measures of social competence and language and cog- Children in villages that received the project were not nitive development than similar children in villages that only more likely to be in an early childhood program, but didn’t receive the project. they went for longer periods of time. Those who were one The research team hypothesizes that this is because when the project started spent an extra six months in early when the new playgroups were started, they didn’t charge childhood programs and those who were four years old fees. As a result, families that might otherwise have sent when it started spent one month more than children in villages that didn’t get the project. Children in villages that received support for early childhood services showed improvements in cognitive and socio-emotional skills. Children who were four years old when the project was launched showed higher emotional maturity when com- pared with those in villages that didn’t get the project. For the children who were one when the project was launched, the evidence suggests that the project led to improved physical health and wellbeing, and language and cognitive development. their five and six year olds to kindergarten—which gener- When the project was launched, the poorer ally does charge a fee—may have kept them in the play- children in the communities used the services as groups set up through the project. For children from the much as children from families that weren’t as less poor families, not going to kindergarten at the right poor… age meant they missed on the age-appropriate stimulation they’d otherwise have received. This affected their devel- Everyone in these villages was poor, but it was possible opment. For children from the poorest families, being in to separate families into less poor and more poor. Both playgroups helped make up for other disadvantages. groups were as likely to be enrolled in preschool services. These results are confirmed when being disadvan- …but the more disadvantaged children benefited taged is defined differently. The researchers looked more from these services than less disadvantaged at the cognitive and other development outcomes children. of children whose parents reported lower parent- ing skills and children who had never been to a One year after the project was launched, poorer children preschool before the project started. in these villages who were age four at baseline were doing better on measures of language and cognitive development There were also gains for children whose parents had be- and social competence when compared with similarly poor low average scores on measures of parenting skills. These children in villages that didn’t receive the project. Children children also did better in language and cognitive devel- in villages where the project was carried out also exhibited opment, and social competence after just one year of the fewer behavioral problems. At the end of the project, these project being up and running, when compared to children children were still performing better on these measures from similar families in villages that didn’t receive the proj- of development. The same distinction wasn’t seen among ect. These children also had less problems with their peers. children from families that were less poor (relatively speak- Three years later, when these children were about eight years old, they were still scoring higher on these measures When the project was launched, the majority of the proj- of development. ect playgroups didn’t charge any fees. Those that did only Children who hadn’t been enrolled in any early child- charged a minimal amount. By 2013, more than two- hood program before these playgroups were established thirds of these playgroups had begun to charge fees. The scored better in the first year on language and cognitive de- fees were still less than what was charged by comparable velopment, and social competence. Three years later, they early childhood education services in villages that never re- were doing better in terms of physical health and emotional ceived the project, but they weren’t as low as had been at maturity. the start. As a result, the type of children enrolled changed, with more children from less poor families enrolling than When the project ended in 2013, almost all of the children from the poorest families. playgroups that had been established were still functioning, but many had started to charge fees and the children who now enrolled tended to be Fifty percent of mothers had no more than a primary school education, and sometimes less. less poor than before. Conclusion EDUCATION The Government of Indonesia is currently considering poli- (all amounts in 2014 dollars). Under the most conservative cies to expand compulsory education to include preschool assumptions, estimates suggest that every dollar spent will education. Quantifying the impact of increasing access to pre- generate 1.3 dollars in benefits. Less conservative assump- school services on children’s school readiness and later educa- tions would suggest that the benefit could be as high as 4.3 tional performance, in low-resource settings in particular, will dollars for every dollar spent. While this is an underesti- aid the government in making an informed decision. mate that only considers private returns to individuals and Was this project a worthwhile investment? Comparable does not account for social returns, it does suggest that the interventions range in cost from US$37 per child a year in project was a worthwhile investment and that community India to US$289 in Colombia. The project in Indonesia, preschools that are given support can help improve poor on the other hand, cost approximately US$30 per child children’s development. The Strategic Impact Evaluation Fund, part of the World Bank Group, supports and disseminates research evaluating the impact of development projects to help alleviate poverty. The goal is to collect and build empirical evidence that can help governments and development organizations design and implement the most appropriate and effective policies for better educational, health and job opportunities for people in developing countries. For more information about who we are and what we do, go to: http://www.worldbank.org/sief. The Evidence to Policy note series is produced by SIEF with generous support from the British government’s Department for International Development. THE WORLD BANK, STRATEGIC IMPACT EVALUATION FUND 1818 H STREET, NW WASHINGTON, DC 20433 Produced by the Strategic Impact Evaluation Fund Series Editor and Writer: Aliza Marcus