World Development Report 2018: Learning to Realize Education’s Promise

Since 1995, England has substantially improved the literacy and numeracy skills of primary schoolchildren using good political strategy and sound technical solutions. As a result, the proportion of students in grade 4 reaching the intermediate benchmark in the Trends in International Mathematics and Science Study (TIMSS) assessment of mathematics shot up from 54 percent in 1995 to 80 percent in 2015 (figure 11.1)—an achievement matched by few other countries. Poor education outcomes had become an important issue in the 1997 national elections, and the new government responded with a national strategy at the start of its term in 1998.

At the heart of the reforms was a redesign of how teachers taught. The new strategy set clear targets for the country, as well as for individual schools, based on regular, publicly available data on student achievement. The targets provided incentives for local education authorities, teachers, and principals. The government adjusted school inspections to reflect the new curriculum; it also strengthened the links between teacher performance and pay. A revamped professional development program, supported by local literacy consultants, helped teachers implement the new strategy. Local governments received substantial new funding for implementation. Literacy and numeracy “hours,” introduced as part of the new strategy, significantly improved early learning outcomes. The program has continued to evolve, with more support focused on disadvantaged learners.

Reforms that improve learning rely on good strategies—both political and technical. This chapter draws lessons from various experiences to identify how opportunities for reform emerge and how politicians, bureaucrats, parents, and students can seize them. It focuses on three entry points for addressing systemic political and technical challenges: improving information, building coalitions and strengthening incentives, and encouraging innovation and agility. Most countries need all three.

Improving information

Addressing weaknesses in education systems is difficult when accurate, usable information on learning is lacking. Without it, stakeholders cannot hold politicians and bureaucrats accountable, assess system performance, or design effective policies to improve learning. Though it might not be enough on its own, better information on learning can provide the substance needed for better political strategies and the evidence base needed for effective policies.

The absence of information on learning can weaken the political incentives to provide good public services. Targeted programs or even direct vote buying are sometimes exchanged for political support, resulting in poor service delivery. Better information increased competition between schools. As a result, can encourage voters to elect politicians who deliver learning outcomes improved in both public and priresults. For example, using a metric that combines vate schools, and private school fees were cut. Parstudent passing rates with test scores, the federal ents can also use information to pressure schools to government in Brazil sets credible education targets raise standards. For example, the provision of report that are widely scrutinized (box 11.1). Meeting these cards has strengthened accountability in some countargets increases the chances of an incumbent politries.Interventions of this kind work best where tician being reelected and of bureaucrats keeping power relations between actors in an education systheir jobs. This example also highlights the value tem are not highly unequal or organized to support of providing information on learning for areas that patronage networks, and where frontline service correspond with political jurisdictions; because of providers have autonomy to respond to community the overlap, citizens can hold politicians accountable demands. When these factors prevent parents’ voices for progress on education targets. But whether inforfrom being heard, it can encourage some, especially mation can shift incentives toward a greater focus on middle-class parents, to opt out of the public educalearning depends on the broader context. For examtion system, weakening pressure on governments to ple, better information in just one sector is unlikely improve learning across the system.

to disrupt patronage networks in countries where

Information can also help ensure that resources clientelism is entrenched across the political system.

go where they are intended. In the mid-1990s, schools in Uganda received only around a quarter of their intended per student grant allocations. The government began to publish information on the timing

Information on school performance can make local and amount of transfers made to districts for school education systems work better. In many developing capitation grants so that schools could monitor local countries, parents have limited information on the administrators. This move increased the share of grant quality of their local schools. In Pakistan, providing funding reaching schools by reducing capture of fundparents with information on learning outcomes ing by district offices. Consistent with the feedback loop described in the next section, schools in areas with better access to newspapers benefited the most.

System managers need information to monitor and analyze system performance. School supervisors need information on student learning outcomes to identify and address poorly performing schools. Good research and evaluation on programs and policies aimed at improving learning can support better implementation by enabling feedback loops. In the early 2000s, Cambodia’s scholarship program sought to improve learning outcomes for disadvantaged students. An early evaluation of the program found that it improved attainment and narrowed gender gaps in enrollment, but it failed to reach the poorest children or improve learning. In 2006, as a result of these findings, the government improved the targeting of poorer children. It then experimented with using the scholarships to encourage learning. Introducing merit-based criteria into student selection increased enrollment and improved learning, raising mathematics test scores by about 0.17 standard deviations.

Research and evaluation can also build support for effective programs across political cycles.

Oportunidades, Mexico’s conditional cash transfer program, has endured since 1997 despite political and economic changes. Because they provided solid evidence of how the program improved the lives of children, impact evaluations were key to the decision to continue the program after a new government was elected in 2000.

Information needed to improve learning is lacking in many countries. An assessment of capacity to monitor progress toward the Sustainable Development Goals found that, of 121 countries, a third lacked data on learning outcomes at the end of primary school, and half had insufficient information on learning at the end of lower secondary school. Even fewer have the data to track these learning outcomes over time. Information systems in the education sector, which are often weak, are rarely used for decision making, planning, or implementation.

There are many barriers to using information to improve learning outcomes. In Tanzania, widely publicized results from citizen-led learning assessments influenced public perceptions of education and shifted the government’s focus toward learning (box 11.2). Yet such direct links between evidence and policy making are often missing. Some evaluations take too long to inform decision making; others fail to track key drivers of low system performance. Even where usable information exists, government agencies may lack the incentives or capacity to use it well. Independence also matters: reliable, salient information can provide incentives for better performance, but biased media may protect the interests of particular groups at the expense of better public services. In Argentina between 1998 and 2007, newspapers that received government funding published fewer reports on corruption than did others.

What are the characteristics of an information system that promotes learning (table 11.1)? First, information needs to be credible, politically salient, and publicly available. Second, clear targets for progress on learning can strengthen incentives by providing measures of system performance. Third, meaningful information on learning needs to be aligned with political or decision-making power, so that the public can hold education decision makers more accountable. Finally, information needs to be usable by policy makers, administrators, and other system actors— that is, it must be timely, accurate, policy relevant, and sensitive to the policy cycle.