Since 1990, the World Bank developed and launched two major education loans in Peru, one (1995) aimed at improving urban primary education and the second (2002) at improving rural primary education. The first loan totaled US$146.5 million, and along with counterpart funds, invested nearly US$300 million in building urban primary schools, developing and distributing school textbooks, and improving classroom teaching. The second loan is still in progress, but it will end up investing roughly $170 million over four years (with counterpart funds nearly US$350 million) in improving rural primary teaching, testing incentive systems to
improve teacher and student attendance, and developing a secondary school distance education system.
The objective of this case study is to evaluate the relevance and effectiveness of World Bank efforts in supporting primary education in Peru. To carry out the task, the mission team interviewed 7 of the 16 current and former Ministers of Education from 1990 to 2005, a number of key past and present educational policymakers who have been involved in the negotiations and implementation of the two loans, the local World Bank education representative and representatives of other international agencies who lend for or provide technical assistance to education in Peru, including the Inter-American Development Bank and the German GTZ. The mission also visited a number of schools, where mission members interviewed administrators, teachers, and parents and observed classes.
During the period analyzed by this report, 1990–2005, the World Bank lent only for primary education (grades 1–6) in Peru, although the rural primary loan does include a component for secondary distance education. The Bank has been a major force in stimulating primary education improvement in Peru, largely because the Ministry—aside from counterpart funds for Bank loans—uses essentially its entire available primary education budget to pay salaries and to meet other usual current expenses. Further, Peru has changed education ministers annually, on average, over the past 15 years. Thus, the Bank has ended up being an important shaper (as well as institutional memory) of many, if not most, Peruvian primary educational improvement efforts during this period.
Primary Education in the National Economic and Political Context
The context for these efforts was an economy that suffered serious setbacks in the 1980s (GDP decline and rapid inflation), a political system threatened in the 1980s and 1990s by terrorists, assaults on the Constitution by the elected president in the late 1990s, and the undermining of the political system by drug cartels. In education, beginning in the 1970s, a series of
governments emphasized expanding access more than improving quality.
Educational attainment is relatively high in Peru but still very unequally distributed between urban and rural areas. The past 15 years of primary school expansion have produced near universal access to full primary education. The majority of urban youth are also likely to finish secondary education (64 percent of urban 16- to 18-year-olds have completed secondary school), but the vast majority of rural youth do not (only 24 percent of 16- to 18-year-olds have competed secondary). In urban areas, a relatively high percentage of youth also attends some years of post-secondary school.
Peru expanded education largely by making it less expensive—principally by reducing teacher salaries in real terms. Except for 1985–87 and an earlier spending jump in 1980–81, educational spending per student fell steadily since the early 1970s. Indeed, by 1990, spending per student had fallen about 60 percent from 1973-74 levels, whereas GDP had risen about 14 percent and GDP per capita had fallen about 23 percent. This necessarily meant steep declines in teachers’ real salaries.
Teachers earned about 25–30 percent more than per capita income in the early 1970s and earned about 23 percent less than per capita income in 1990, a drop of about 50 percent relative to the average Peruvian’s economic situation. Part of this fall in teachers’ relative position is due to an increase in average education in Peru’s labor force, but part is due to a fall in teachers’ wages relative to those of other professionals.
Quality of education, as measured by pupils’ scores on international tests, is at the low end in Latin America, much below Mexico, Chile, Argentina’s, and Colombia’s results on the same tests. This is not just an artifact of Peruvian students’ lower average socioeconomic background. The top 10 percent of achievers in Peru on Programme for International Student Assessment (PISA) scored at about the same level as the 60th percentile in Argentina. On UNESCO’s LLECE test, higher socioeconomic background Peruvian pupils also scored much lower than their counterparts in many other
Latin American countries, and rural Peruvian students scored among the lowest in Latin America.
World Bank Support for Expanding and Improving Primary Education
The initial history of World Bank support for Peruvian education mirrors that of many other Latin American countries: loans for primary education in Peru started only in the mid-1980s, following the 1970s (technical and vocational) and 1960s (tertiary) project cycles. In 1984, a loan to improve and to expand primary education was signed, with the goal of supporting the first three-year phase of a ten-year education program designed to do the following: (a) provide sufficient and adequate student places for school-age children, (b) improve the quality of primary education, and (c) improve primary education management. The loan became effective in June 1985; less than two years later the Bank suspended disbursements to Peru. The project outcome was rated as unsatisfactory.
In 1993, the Fujimori government, with Bank support, developed an extensive diagnostic of Peruvian education and called for actions to improve educational quality, efficiency, and equity. That report led to the design of the Primary Education Quality Project (MECEP) project. The report pointed to key issues of instructional materials, teacher training, public school autonomy and accountability, school infrastructure, and bilingual-intercultural education.
Together, the first four issues became the basis for the broad 1994 MECEP loan in the amount of US$146.4 million (with a government contribution of US$107.5). Though not initially contemplated in the project design, school infrastructure became the project’s largest component (nearly half of project funds). This was the direct result of President Fujimori’s insistence on school construction as the project’s main goal. To ensure that school buildings did not take priority over “soft” investments, Bank staff set specific yearly targets for textbooks and training. Achievement of these targets triggered the release of funds for the construction component.
Beginning in 2001 the Bank signed a series of Programmatic Structural Adjustment Loans, designed to transfer funds directly to the Ministry of Finance in exchange for a broad array of social sector policy reforms (including health, education, and social protection). Each of the Programmatic Social Reform Loans I-IV (PSRLs) was signed in the amount of US$100 million (except for PSRL III, in the amount of US$150 million). Through the PSRLs, the Bank financed the publication of both international (UNESCO/LLECE) and national assessment results; establishment of monitoring and supervision systems, including creation of a payroll system to track the problem of ghost teachers and compare teaching responsibilities with payroll amounts; piloting of a program of local control in the distribution of salary incentives for rural teachers, guaranteeing budgetary allocations for counterpart funds for finalizing MECEP; and development of a monitoring and evaluation system designed to provide transparency of information during the decentralization process. In 2004, a Technical Assistance Loan in the amount of US$7.8 million was signed to support the development of an accountability system for decentralization in the social sectors, particularly with regard to improved monitoring and evaluation activities.
In 2003, the Bank and the Toledo government realized a long-in-gestation Rural Education Project. The first phase PEAR APL was signed in the amount of US$52.5 million (with a government contribution of US$29.5 million and an Inter-American Development Bank contribution of US$12.2 million). The total program amount of the 10-year, three-phase APL is expected to be US$347.2, of which US$172.5 is a World Bank loan. Project components include expanding access for rural children, improving quality in rural primary school, and reforming teacher policy and education management. Expansion of access under the first project component focuses on both preschool and secondary education.
These loans represent significant amounts of money in the context of Peruvian educational spending. The $300 million of the primary education loan in 1995-2000 represented about
5–6 percent of the total education budget for those 5 years and almost 20 percent of the total budget for primary education. The rural education loan now underway is much smaller but also represents a significant fraction of the money going to rural primary education.
The Bank’s Contribution to Sectoral Changes in the Past 15 Years
Each of the two Bank-supported projects and PSRLs implemented during this period (1990–2005) has generally been based on recommendations from detailed research-based diagnostics. These diagnostics were important to shaping the direction of the projects and helped to build consensus around the challenges and potential solutions for the MECEP project in the 1990s and the rural project. The activities outlined in the MECEP and PSRL projects seemed appropriate to education sector conditions in the country at the time and usually focused on areas where Bank could contribute with extensive experience and technical assistance, for example, in textbook production, teacher training, and teacher incentive pilots, and distance education in rural areas.
Although the design of the MECEP project was relevant to the needs identified in the diagnostic, as highlighted in the IEG review, the institutional development component was overly ambitious, especially given the volatile nature of the political context and the lack of specific project measures to help the Ministry develop and build consensus around proposed reforms of school governance (especially autonomy) and administration (for example, decentralization reforms).
Though proposed in the original project design, which did not adequately take into account issues of political will, neither school autonomy nor regional decentralization was implemented under the MECEP design, though some aspects of the original design would appear both in later projects (such as the rural project) and through independent Ministry actions, such as the new teacher-hiring process implemented at the beginning of the Toledo presidency.
There are three important caveats to the overall positive assessment of the relevance of Bank-supported project activities. The first is the inclusion of the construction component in the MECEP project, which was not originally seen as a priority in the sector diagnostic. The Fujimori government, however, had threatened not to have a project at all unless the construction component was included; in exchange for guaranteeing advances in other areas, Bank staff included school infrastructure. In hindsight, there was considerable need for physical school improvements, though, as discussed below, these likely would have occurred even without an MECEP project.
The second caveat relates to the low level of institutional capacity building in project activities. The Bank did help modernize the Ministry through financing the technical assistance, hardware, and software to install information systems for payroll and record keeping. The Bank financed the technical assistance to make the Ministry more cost efficient through the elimination of many superfluous payroll positions. The Bank also supported the Ministry to develop and sustain the Quality Measurement Unit, which has done excellent work in achievement measurement and analysis over the past 10 years. Yet at the same time, the Bank-created and -financed Project Management Unit in the Ministry has had little impact on training people in the rest of the Ministry or in departmental offices or installing management systems that permanently become part of the Ministry’s mode of operation.
One of the main problems in this regard has not been under the Bank’s control—the almost constant change of ministers in the past 15 years. It is telling that the implementation of the MECEP project and PLANCAD (the National Plan for Teacher Training) is largely due to the fact that one minister, Domingo Palermo, lasted three years during the Fujimori regime.
The third caveat concerns the absence of evaluation and monitoring of project activities and impact. There have been no ex post evaluations of project impact even though data are or could have been available for assessing the effect of textbooks and teacher training on student achievement over the five-year period 1996–2001. There is some indication that test scores for primary school children have remained relatively constant throughout the period, and this at a very low level compared to other large Latin American countries. However, this indication is not based on strict comparisons of like items on tests at the fourth-grade level, for example, which would have been possible if Bank or Bank-financed Ministry staff had built project evaluation into the Bank-supported project. In the absence of further evaluation, we do not even know if teachers changed their practice. We do know that thousands of teachers received training of varying quality from a variety of agencies contracted by the Ministry of Education.
Each of the PSRLs was highly relevant in establishing key administrative and legislative benchmarks for improvements within the education sector as well as protecting key social sector antipoverty measures from budgetary cuts during the transition period. Highly relevant measures include laying the administrative groundwork for the Rural Education Project, reforming the payroll system, and creating additional transparency within the Ministry of Education budget system.
Regarding the Rural Education Project, project activities are relevant to the sector, especially given the advances achieved under the PSRLs in terms of the creation of school councils, more autonomous regions, and schools. There is, however, a question that the rural project may be doing too much (that is, it is spread too thin across a variety of activities). Some of the elements of the REP are being evaluated carefully, using comparison groups. But it does not appear that an evaluation component using learning outcomes data has been built into the overall effort to improve the quality of rural education. A recent progress report on this project shows some major problems, especially the lack of an implementation strategy, an overall monitoring, and an evaluation plan; and a communications strategy aimed mostly at parents, teachers, and administrative personnel linked with the project.
Lessons
Peru’s history of progress in primary education is typical of developing countries in some ways and very atypical in others. Peru has reached high levels of incorporating its population into primary education, even in poor rural areas, and rather high completion rates for primary schooling (and secondary school attendance) for marginalized urban and rural youth. This makes it somewhat atypical for a lower-middle income country. It is also atypical in the financial effort it has expended to accomplish these goals. Peru spends relatively little on its primary education system. Its costs per pupil are among the lowest in Latin America, and its teachers are paid among the lowest in the region relative to per capita income and compared to other public servants.
Peru is typical of countries investing so little per pupil in public primary education (Central American countries, for example) that its students score very low on international achievement tests, both at the primary level (LLECE) and in middle school (PISA), even when adjusted for socioeconomic class differences. Peru is also typical of most developing countries in that the teaching supervision system and teacher and school accountability systems are essentially nonexistent. Finally, Peru shares with most countries a fundamental lack of capacity for managing a massive and highly spread out primary education system. That is one more reason why the quality of these services is so low.
These underlying conditions suggest that improving teacher capacity and the governance of primary (and secondary) education are crucial to improving quality and to increasing efficiently the amount of schooling taken by each student. The Peruvian case suggests that management capacity building, from Ministry to school to classroom, should be a priority for governments and for agencies lending for primary education in developing countries.
Development Effectiveness of Bank Support
The Bank strategy under such conditions seems to have been to invest in projects that emphasized successful delivery of educational inputs rather than the delivery of educational outcomes. In the 1996 urban primary education loan, MECEP focused on two inputs—textbook distribution and improved classroom pedagogy. In theory, the delivery of these inputs should produce higher student outcomes, but this is not what the Bank emphasized.
Under programs that emphasize input delivery, managers are considered successful if they deliver so many repaired buildings and textbooks or train so many teachers. In a Latin American country it should be expected that projects could go to the level of outcomes: delivering textbooks that are used in instruction and changing teacher and management behaviors. It appears that the project took the less-demanding road and focused on the inputs but not on actual textbook use, teacher behavior in the classroom, or learning outcomes.
The Bank’s strategy implicitly assumed that if textbooks arrive at the school, teachers and students would use them effectively, and that if teachers learned better teaching techniques, they would utilize them effectively. Although there was slippage in textbook distribution and some teacher corruption in taking commissions from competing publishers to not use the free textbooks, the presence of textbooks and exercise books probably did contribute positively to pupils’ learning. But a greater emphasis on the effectiveness of textbook use would have had to include considerable investment in management capacity. Teachers did apparently use at least some of what they learned in the in service training courses, and, based on teacher interviews, teachers who took the courses considered them valuable. Contract teachers who were not eligible for the courses also wanted very much to take them.
But even though investing in such inputs is a correct strategy, the question is whether without supportive investments in supervision and content knowledge their yield is high enough to justify spending considerable sums on them (particularly the much more expensive pedagogical training part). It does not appear that the yield on pedagogical improvement was very high in the context of teachers’ low content knowledge, but an emphasis on outcomes may have forced a more effective investment strategy.
In the Rural Education Loan (2003), the emphasis is also on improvements that emphasize input delivery, such as expanding access to rural education, nonformal preschools run by community implementation agents, and providing direct access to distance secondary education based on programming from a centralized location. Other improvements do, however, emphasize outcomes, such as pilot community incentive programs to improve rural teacher attendance and teacher accountability to local rural communities.
Probably the Bank’s most input delivery-oriented project was the questionably large amount it spent on school construction in the 1990s under pressure from President Fujimori himself. As long as school construction itself is monitored (the Bank claims that the 450 schools’ construction that it controlled met specifications), the finished school is the product itself, and it does provide additional or at least improved classroom space for educational activities.
Thus, the Bank seems to invest in input improvements whose delivery alone signals project success, regardless of whether student outcomes increase as a result of such investments. This is a prudent choice for the Bank in a low management-capacity environment, although there exist some serious questions regarding whether because of poor management at the school level, the investments turned out to have relatively low yield in terms of improving student learning significantly. In the long term, educational improvement will depend on the ability of projects to influence management and teacher behaviors and to improve educational outcomes.
With constant changes in ministers, it is admittedly difficult to maintain continuity in reform efforts. The Bank has been fairly successful in Peru despite this difficulty because of the personnel in the local office of the Bank and the fact that the Bank’s education sector specialist has been in place for 10 years.
Thus, the Bank has been an important part of the institutional memory for reform, and has, by being firm in not changing the shape of its loan agreements once signed, been able to get most of what it wants in the agreement and implementation of the loan. This is not always a good thing, but for the most part, keeping implementation on course has worked reasonably well. All in all, the Bank should be more aware of the longer-term nature of successful educational reforms, particularly in a country in which the educational system requires long-term improvements in quality.
