A systematic review of the evidence of the impact of school voucher programmes in developing countries

A systematic review of the evidence of the impact of school voucher programmes in developing countries


Demand-side financing is based on the principle of governments either channelling education resources through students and their parents, or basing school funding on enrolments or attendance. Regardless of the direct path of funding, demandside financing is viewed as a way of addressing inequities that prevent poor children from continuing their education, as well as a means of introducing school choice. Education vouchers, a demand-side financing intervention involving the public subsidy of private schooling based on the number of eligible voucher students per school, generally aim to expand parental school choice, which is often promoted to increase competition in the school system. Opponents of vouchers argue that private schools do not necessarily provide a higher-quality education; affluent families with more social capital and access to voucher programme information are more likely to find the best schools; and it is very difficult to set up effective systems of accountability to guard against 'cream skimming' and sorting. In this paper, we report on a systematic review of evaluations of education voucher programmes in developing countries.

Through extensive searching, including electronic keyword searches of bibliographic databases, handsearches of relevant journals, examinations of online holdings of international development organisations and research firms, citation chasing, examining grey literature, and contacting experts in the field, we identified studies that responded to the following question:

What is the evidence of the impact of school vouchers in developing countries?

Eligible studies had to meet the following criteria: The evaluation took place in a low-income developing nation as defined by the World Bank at the time of the intervention; and the evaluation directly assessed the impact of a school voucher programme on participants' educational outcomes. With the intent to conduct meta-analysis, we focused on identifying randomised controlled trials (RCTs) or quasi-experimental evaluations (QEDs) with some evidence that the groups being compared were equivalent. In addition, we sought for contextual information - but not for inclusion in effect size estimates - i.e., quasi-experiments without pre-test group equivalency, and other quantitative and qualitative studies that shed light on implementation and context issues.

Each RCT or QED located in the search that appeared to be a possibility for inclusion was carefully reviewed by two authors and a structured abstract was prepared for each study, detailing the context, methodology and findings. For each study deemed eligible for inclusion following this screening process, a coding instrument was completed that included items in the following areas: researcher and study characteristics, study methods and methodological quality, intervention and control conditions data, participants in the study and outcome data. To evaluate study quality, we recorded details on three key implementation issues: how the groups were equated and whether any problems with equating were reported, information on attrition, and whether the programme experienced significant implementation or fidelity problems.

We identified two studies that met our inclusion criteria - one examining the Colombia PACES programme and the other evaluating the Quetta, Pakistan Urban Fellowship programme. We also identified four quantitative studies on the Chile voucher system that did not meet our criteria for inclusion in effect size estimates but were examined to shed light on possible theory, implementation and context issues. Given the very small number of studies that met our inclusion criteria, we provide the results in a narrative fashion, rather than through meta-analysis. Both the Colombia and Pakistan programmes increased private school enrolment amongst the countries' poorest income groups, thus probably improving equity. The Pakistan programme resulted in girls being educated for less than it would have cost for the government to create public school spaces, while the Colombia programme cost rather more, but will most likely prove cost-effective in terms of long-term economic gains.

Clearly, more rigorous research in developing country contexts is necessary to determine whether the gains from these two programmes can be replicated and enhanced and to elucidate the many issues surrounding vouchers. Pilot programmes employing random assignment or lotteries should be accompanied by rigorous impact evaluation. This approach would enable governments to design innovative initiatives and target resources most efficiently and equitably.


Morgan, C.; Petrosino, A.; Fronius, T. A systematic review of the evidence of the impact of school voucher programmes in developing countries. EPPICentre, Social Science Research Unit, Institute of Education, University of London, London, UK (2013) iii + 92 pp. ISBN 978-1-907345-45-6

A systematic review of the evidence of the impact of school voucher programmes in developing countries

Help us improve GOV.UK

Don’t include personal or financial information like your National Insurance number or credit card details.