from EVIDENCE to POLICY Learning what works for better programs and policies February 2018 PAKISTAN: Can private schools catering to the poor increase access and improve learning? Educating children is a priority across the world, but low-in- Development Association (IDA), the World Bank’s fund for come countries can face enormous challenges. Schools are often the poorest, the Sindh government received assistance to de- overcrowded and in disrepair. Teachers don’t always show up or velop and implement its Sindh Education Sector Reform Pro- may not be qualified or interested in teaching. Parents hesitate gram, which aimed to raise enrollment, improve learning, and to send children, especially girls, to schools that aren’t close by increase accountability and better governance, thereby reduc- or they may want to keep ing social disparities in education. This included a public-pri- them at home to help vate partnership program, in which the government provides with housework. The cash subsidies to private entrepreneurs who establish and oper- EDUCATION numbers tell the story: ate free, co-educational primary schools in villages in remote Worldwide, 58 million areas where government schools are not always available. children who should be An evaluation was integrated into this program to assess in primary school are its impacts. The impact evaluation found that boys and girls not, despite the push for in villages that were randomly assigned program schools were universal primary educa- more likely to be in school and they did better on tests than tion by national governments and international organizations. children in villages without such schools. Parents also had big- In the effort to boost enrollment, raise teaching standards and ger dreams: they were more likely to want their daughters to strengthen school accountability, policymakers and education become teachers rather than housewives and were more likely experts are exploring a variety of approaches, including lever- to want their sons to become doctors or engineers instead of aging the private and other non-governmental sectors to offer security personnel. The World Bank is supporting the Second quality education to disadvantaged children. Sindh Education Reform Program, which is helping contin- In Pakistan’s Sindh province, only about half of primary ue this public-private partnership program and support and school age children go to school. Through the International strengthen the quality of the schools. Context Private schools in Pakistan, historically catering to children of paid less and are often less educated than their counterparts in the country’s elite, have become popular among poorer popula- government schools. tions as lower-cost private options have expanded. More than Pakistani policymakers and education experts are turning to a third of all Pakistani students are now enrolled in private for-profit private education as one route to improve the quality schools, where tuition averages less than $5 a month in rural of education and expand access for the country’s primary school villages, a small fraction of average household income. Research age children, more than a quarter of whom aren’t in school. While studies of these schools have generally found that student learn- the models differ in details, these public-private partnership pro- ing is higher and teachers perform better, although teachers are grams generally require that schools waive tuition in exchange for receiving a per student subsidy that schools use for their costs. The working with a World Bank team. The program, launched in subsidy is set low, often lower than the cost to run a government 2009, seeks to encourage private entrepreneurs to build and school where teacher salaries tend to be much higher, and account- operate primary schools in educationally underserved villag- ability measures are built into the programs to promote quality. es, generally defined as having limited access to government In Pakistan’s Punjab province, for example, in a program initi- schools. Local private entrepreneurs are invited to apply to ated in 2005, existing low-cost private schools can qualify for the program through an open call in newspapers and to pro- a per student subsidy if they waive tuition and ensure that a pose underserved villages where they will establish and oper- minimum percentage of students pass a standardized academic ate schools. To qualify, schools must meet minimum facility test offered on a regular basis. This program, which has proven standards, be co-ed, waive tuition fees for all students, and successful at raising school enrollment and student test scores hire teachers with at least eight years of schooling. Two of the (see Evidence to Policy note, October 2012, “Can low-cost pri- teachers have to be female. In return, the private entrepreneurs vate schools improve learning?”), now covers more than 1.7 receive a per student subsidy. They also get additional non- million students in some 3,500 program schools. financial assistance, such as free textbooks and other materials, In Sindh province, education experts have also turned to teacher training, and regular visits by foundation staff to advise low-cost private schools to increase access to education and im- on how to improve teaching and learning. This policy note re- prove student learning, especially in villages that lacked local flects new and updated results since the previous Evidence to schools. The Promoting Low-Cost Private Schooling in Rural Policy note published in September 2013, “Using Low-Cost Sindh program was created by the Sindh Education Founda- Private Schools to Fill the Education Gap: An impact evalua- EDUCATION tion, a semi-autonomous agency of the provincial government, tion of a program in Pakistan.” Evaluation At the start, the program was implemented in eight districts, tiated subsidy treatment groups received training, textbooks chosen because they ranked very low on three school access and teaching materials. measures: the number of children in school, gender equality Because the schools opened in the summer of 2009, a few among children in school and the percentage of households less months after the start of the school year in the spring, students than 15 minutes walking time from the nearest primary school. in the program schools had a shorter first school year. The base- The evaluation sought to test not only the impact of pro- line survey was conducted in February 2009, with a follow-up gram schools on enrollment and student learning, but also survey in April-May 2011 after the second school year ended. whether giving some schools a bigger per student subsidy for The baseline survey interviewed 5,556 children aged five to female students had a larger impact on girls’ outcomes. A to- nine and the follow-up survey interviewed 17,720 children tal of 199 villages in the districts were randomly assigned to aged five to 17. During the follow-up household survey, chil- one of three groups: 82 villages in which program schools dren aged five to nine were given a language and math test received a subsidy of 350 rupees per student ($5 in 2008 and queried about work they did apart from school, their en- US dollars), regardless of gender; 79 villages where program rollment status, and their desired future occupations. Schools schools received the 350-rupee subsidy for each student, with were also surveyed at follow-up to gather information on their an additional 100 rupees ($1.50 in 2008 US dollars) for each operations, facility, and staff. Teachers were interviewed and female student, and 38 control villages where no program attendance lists were checked through unannounced visits to support was offered. Both the gender neutral and differen- cross verify reported enrollment and attendance. This policy note is based on “Delivering Education to the Underserved through a Public-Private Partnership Program in Pakistan,” Felipe Barrera-Osorio, David S. Blakeslee, Matthew Hoover, Leigh Linden, Dhushyanth Raju, and Stephen P. Ryan, Policy Research Working Paper 8177, World Bank Group, South Asia Region, August 2017. Findings Enrollment among children in villages with program disadvantaged backgrounds: their fathers were less educated schools rose sharply, and children in these villages and more likely to be farmers, and their families lived in continued in school longer. more poorly-constructed homes. The primary school enrollment rate in rural Sindh for children aged six to 10 was about 56 percent at the start of the program—among the lowest in the country—with a significant gender gap: Enrollment was 65 percent for boys, compared to 46 percent for girls. In the areas where the program was implemented, enrollment was even lower, about 30 percent for primary-school aged boys and girls. Two years after the program started, enrollment of boys and girls aged six through 10 had increased by 30 percentage points relative to the control group. There was no differ- ence in the increase for villages where schools also got an additional subsidy for enrolling girls. Researchers also found that children in program villages completed an extra half a grade on average as compared to children in control villages. And despite the fact that the program didn’t target them, enrollment for children aged 11 to 17 also increased by 12 percentage points. Children in program villages did much better on standardized tests. Program impacts on school enrollment or student test scores didn’t differ by gender or by the gender- Math and language tests were administered to children age based structure of the subsidy. five to 10 years old in household samples in both program and control villages as part of the follow-up measurement. The impact of the program on enrollment and test scores was Overall, when looking at all children in villages where the similar for both girls and boys. The additional subsidy to encour- program was rolled out—regardless of whether or not the age program schools to attract girls to school didn’t improve girls’ child had enrolled in school—test scores in both math and enrollment or test scores beyond the subsidy provided to other language rose by 0.63 standard deviations, which is equiva- programs schools for all children. lent to moving a child from the average score to the 74th percentile. When just looking at the children who enrolled Program schools not only did a better job in school because of the program, the impact was even larg- teaching— despite having less experienced teachers er: test scores for those children increased by two standard than government schools—but they provided their deviations, equivalent to moving a child from the average students with better facilities. score to the 98th percentile. These higher test scores are par- ticularly significant because children enrolled in program Similar to findings from studies of low-cost private schools in schools were more likely to come from socioeconomically other parts of Pakistan, the program school operators in Sindh province did a better job of keeping the schools staffed and run- The evaluation found that going to a program school ning. And the type of subsidy a school was offered – whether didn’t just give children the chance to learn, but it gender neutral or with an additional payment for girls – didn’t gave them the chance to dream bigger. affect how schools were structured or operated. The schools usu- ally had more classrooms than government schools and were Families in program villages were more likely to hope that their more likely to have sufficient desks for students, drinking water daughters became teachers, and less likely to hope that they be- facilities, an electrical connection, and toilets. Researchers also come housewives. They were more likely to want their sons to found that program schools were open an extra half day a week, be doctors or engineers, and less likely to want them to become on average. The program required that these schools hire at least security personnel, as compared with families in the control two female teachers, and the total number of female teachers villages. Families also raised their expectation of the ideal edu- was higher than in government schools. Program school female cation level by one and a half years. teachers generally had less teaching experience than their coun- terparts in government schools, and they were paid less. Conclusion EDUCATION Private schools catering to poorer populations in Pakistan counterparts in government schools in the general vicinity. are proving very successful at attracting students—boys and There is still more to learn about ways to create and support girls—and teaching them effectively for less money than it programs that expand educational access and improve school costs to run a government school. Some of the lower costs quality. But as policymakers explore solutions to the global come from hiring teachers who receive lower salaries than learning crisis, it’s important to keep in mind that with mini- government school teachers, but this doesn’t appear to be mal government support, local actors and entrepreneurs have hurting the quality of education. On the contrary, program the potential to create cost-effective solutions to address local school students did substantially better on tests than their challenges in education. 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 and the London-based Children’s Investment Fund Foundation (CIFF). THE WORLD BANK, STRATEGIC IMPACT EVALUATION FUND 1818 H STREET, NW WASHINGTON, DC 20433 Series editor: Aliza Marcus; Writer: Daphna Berman