54851 March 2010 Note Number 32 Output-Based Aid in Education: A Solution for Quality Education | Lars Johannes Inga Murariu A number of approaches have been tried to improve school attendance and educational attainment, including the use of Output-Based Aid (OBA). The challenge of introducing OBA in education has often been finding an appropriate definition of "output" that balances achievement of results with reasonable transfer of perfor- mance risk. This review examines several OBA projects in education and describes how they compare to other results- based schemes. It also discusses where the OBA approach fits in the larger spectrum of results-based mechanisms in education, and why it can be successful. OBA in the Education Sector Output-based aid (OBA) in education is used to bridge million in Chile, and to over $130 million for the two the gap between the cost of providing quality education phases of the Bangladesh FSSAP. Three of the schemes and the funds available. As a rule, it involves payments to aimed at improving student enrollment, attendance, and schools based on predefined outputs such as enrollment quality of education, and one provided learning op- and attendance of specified school-age children, and may portunities for adults. All schemes employed extensive include school achievement as a performance indicator government co-funding. (see Box 1 for other results-based schemes). The funding for the identified OBA schemes comes Few OBA schemes have been tried in the educa- from a variety of sources: the World Bank Group's In- tion sector. A new review (Mumssen, Johannes, and ternational Development Association (IDA) and In- Kumar, 2010) identified four OBA schemes funded by ternational Bank for Reconstruction and Development the World Bank that provide performance-based grants (IBRD), government revenues, parental contributions, for the actual delivery of education services, and this and private investments. review adds another, government-funded one. The five schemes are the Female Secondary Assistance Programs Targeting in Bangladesh I and II (FSSAP),1 the Lifelong Learning and Training Project in Chile,2 the Balochistan Edu- The OBA educational schemes in developing countries cation Support Project (ESP),3 the Punjab Education aim at targeting services mostly to low-income house- Support Project in Pakistan,4 and a Concession Schools holds. Geographic targeting is common, particularly Program in Bogota, Colombia.5 The scale of subsidy where projects are small and thus confined to a specific disbursements ranged from $2.1 million for the scheme geographic area. In countries where the literacy gap in Balochistan, to $77.5 million in Punjab, to over $100 between male and female students is significant, the OBA schemes can target girls, in particular. For example, the districts in Bangladesh for FSSAP were identified based on their economic level of development, low female Inga Murariu is a Consultant and Lars Johannes is an Infrastructure Specialist with the Global Partnership on Output-Based Aid literacy rates, and low female attendance levels. New pri- vate schools in Balochistan, Pakistan were opened if no Supporting the delivery of basic services in developing countries Box 1. Other Results-Based Schemes in the in need of school spaces. The Lifelong Learning and Education Sector Training Project in Chile relies on self-selection and targeted marketing to provide vocational training to A variety of other result-based schemes exist in the adults between 15 and 65 years of age, since vocation- education sector. al training is more attractive to the poorer strata of the population. Conditional Cash Transfers CCT schemes focus entirely on the demand-side of education services by providing incentives for Performance Risk families to send their children to school. Thus these schemes encourage families to choose to invest in The challenge of introducing OBA in education has education, which may have an opportunity cost for often been to define an appropriate "output" that is the family, such as preventing children from work- closely linked to the desired outcomes without putting ing on the family farm. CCT schemes in education efficient service providers at too much risk. For exam- differ from OBA in that there is no performance ple, paying only for enrollment may be considered in- risk for the supplier of education services. While they differ, OBA and CCT schemes can be comple- sufficient incentive to an institution to provide quality ments: CCT provides an incentive for the demand education, while there is a limit to the extent to which for education, while OBA provides an incentive for disbursement can be tied to academic achievement, as the supply. Both are needed, particularly in social achievement is influenced by a number of factors that services sectors, where the benefits of services are not fully under the control of the service provid- such as education are not always apparent to ers. As a result, a project could make disbursements poor households, or where they may carry a high through a combination of attendance and measures of opportunity cost in terms of perceived income forgone. the quality of education. For example: Cash on Delivery · The Punjab Education Support project offers a COD is a results-based approach proposed by monthly per-student subsidy to low-cost private the Center for Global Development (CGDEV) as schools, with free rein on how the school spends a hands-off way to improve results in the educa- the subsidy. In return, schools must offer educa- tion sector. CGDEV proposes to pay the recipient tion tuition free and achieve a minimum level of government a specific amount for a specific mea- student performance in an externally adminis- sure of progress, without prescribing the means or policy to achieve it. The scheme does not address tered, semi-annual, independent test. The pro- a specific funding gap; rather it is an incentive gram also offers large group bonuses to teachers payment to government for a specific output. For and competitive bonuses to schools for high levels example, donors could promise to pay a certain of student test performance. amount for every additional child who completes · Under the Concession Schools Program in primary education and takes a test. The further use Bogota, private school operators manage public of funds is at the government's discretion. How schools and must meet pre-established targets much COD and OBA schemes compare in practice remains to be seen once the first COD schemes are for standardized tests and drop-out rates for two implemented. consecutive years in order to qualify for contin- ued state funding. · Under the Lifelong Learning and Training Project in Chile, the performance risk is born by the pri- vate service provider. Subsidies are paid based on students' demonstrated completion of a learning other schools operated within a two-kilometer radius; module. enrollment of girls was required to reach at least 40 percent. Private Sector Capital and Expertise Some schemes make use of existing means testing or proxy means testing systems or social stratification In most countries, education is largely provided by mechanisms to avoid subsidizing wealthier students public schools and is financed publicly. Private sec- already enrolled in private schools. For example, the tor capital is mainly mobilized to build, operate, and Concession Schools Program in Colombia specifi- maintain education infrastructure. Projects can also cally targets marginalized low-income areas of Bogota rely on existing infrastructure to some extent: for March 2010 Note Number 32 example, through vouchers that pay for additional Box 2. The Impact of Public-Private students to attend existing private schools. Partnerships on the Concession Schools Evidence shows that private sector involvement Program in Bogota can effectively address issues related to quality of edu- cation and school management (see Box 2). In devel- The public-private partnership in Bogota illustrates oping countries, where scarce government resources the positive impact private sector involvement can cannot provide education for all, partnering with the have on attendance and school attainment. Cre- private sector may be a solution.6 ated to provide education to 45,000 students from For instance, in Pakistan, the inclusion of low-fee low-income neighborhoods, the program is based on a bidding process in which the applicant must private schools in the Balochistan Educational Sup- demonstrate previous experience in the educa- port Project was based on the successful implementa- tion sector. The winning private school operators tion of a pilot phase in which private schools sup- qualify for continued state funding only if they meet plied low-cost, high-quality education for girls from pre-established targets on standardized tests and very poor urban areas of the province. The private drop-out rates for two consecutive years in the schools selected by the project receive annual per- public schools they manage. The program has had student subsidies for facilities and material costs, and the following impact: a monthly subsidy linked to student enrollment and · It made the concession schools financially attendance. stable, since the state paid a fixed sum per stu- dent, and focused more on academic achieve- Monitoring ment, as the selected private operators had a history of high test scores. To make payments based on outputs feasible, outputs · The private participation model applied a proven pedagogic model. Moreover, the free- need to be defined in such a way that they are measur- dom to choose the teaching and administrative able and discrete. Such outputs can include enroll- staff led to improvements in the quality of edu- ment, attendance, or educational attainment, which cation. On average, 55 percent of the subsidy may be measured by standardized tests. Monitoring amount was allocated to human resources-- can be undertaken by schools themselves, govern- well below the 90 percent in the public school ment entities, consultancy firms, or nongovernmental system--freeing up 33 percent for nutritional organizations. Projects based on school choice rely on support and education materials. · The eligibility of private operators for continued parents to judge the quality of schools in which they support depends on meeting pre-established enroll their children. standardized test scores, which emphasizes Monitoring systems for OBA projects in education the importance of higher quality of education need to be carefully designed. Self-reporting of enroll- rather than mere operation of the schools. ment or attendance may provide an incentive to inflate · Private sector providers had an incentive to output figures, and standardized tests have a risk of partner with parents and the community, which providing an incentive for "teaching to the test" and helped reduce dropout rates and improve for cheating in order to secure funding. Independent educational attainment. verification can help ensure that subsidies are paid Source: Barrera-Osorio (2007). only for outputs that have been achieved. If outputs are verified by government institutions, it is important that such institutions are sufficiently independent and have sufficient capacity. To mitigate the risk of teach- ing to the test and cheating, the Cash on Delivery ap- have generally chosen a more balanced set of outputs proach proposes to tie incentive payments to partici- that aim at addressing not only the issues of enrol- pation in standardized tests and publishing results. ment and attendance, but also of quality. Evidence shows that private sector expertise and discipline can Conclusion improve both operational management and learn- ing outcomes without an increase in cost. However, A focus on higher attendance and school completion given the limited use of OBA in education and the rates alone does not necessarily translate into im- small number of projects studied by this review, more provements in knowledge and skills--which should experience and evidence will be required to formulate be the ultimate policy objective. Thus, OBA projects better grounded conclusions. References 1 For information on FSSAP I, see: http://web.worldbank.org/exter- nal/projects/main?pagePK=64283627&piPK=73230&theSitePK =40941&menuPK=228424&Projectid=P009555 Barrera-Osorio, Felipe. 2007. "The Impact of Private For information on FSSAP II, see: http://web.worldbank.org/ Provision of Public Education: Empirical Evidence external/projects/main?pagePK=64283627&piPK=73230&theS from Bogota's Concession Schools." World Bank, itePK=40941&menuPK=228424&Projectid=P044876 Washington, DC. 2 http://web.worldbank.org/external/projects/main?pagePK=642 Mumssen, Yogita, L. Johannes, and G. Kumar. 2010. 83627&piPK=73230&theSitePK=40941&menuPK=228424&Pr ojectid=P068271 Output-Based Aid: Lessons Learned and Best Prac- 3 http://web.worldbank.org/external/projects/main?pagePK=642 tices. Washington, DC: World Bank. 83627&piPK=73230&theSitePK=40941&menuPK=228424&Pr Patrinos, Harry Anthony, F. Barrera-Osorio, and J. ojectid=P094086 Guaqueta. 2009. The Role and Impact of Public- 4 http://web.worldbank.org/external/projects/main?pagePK=642 Private Partnerships in Education. Washington, 83627&piPK=73230&theSitePK=40941&menuPK=228424&Pr ojectid=P102608 DC: World Bank. 5 Barrera-Osorio (2007). 6 For an extensive analysis of private-public partnerships in educa- tion, see Patrinos, Barrera-Osorio, and Guaqueta (2009). About OBApproaches OBApproaches is a forum for discussing and disseminating The case studies have been chosen and presented by the au- recent experiences and innovations in supporting the delivery thors in agreement with the GPOBA management team and are of basic services to the poor. The series focuses on the provi- not to be attributed to GPOBA's donors, the World Bank, or any sion of water, energy, telecommunications, transport, health, other affiliated organizations. Nor do any of the conclusions rep- and education in developing countries, in particular through resent official policy of GPOBA, the World Bank, or the countries output-, or performance-, based approaches. they represent. To find out more, visit www.gpoba.org e Global Partnership on Output-Based Aid e Global Partnership on Output-Based Aid Supporting the delivery of basic services in developing countries