Investment Plan Detail

Investment Plan summary 1For further detail on the costing model and financing model, see Education Commission Analytical Background Paper available at http://report.educationcommission.org/resources.

The investment plan for the Learning Generation proposed by the Commission includes ambitious but achievable targets for domestic and international, public and private financing:

• The Commission’s investment plan calls for low and middle-income countries to increase domestic public expenditures on education from an estimated $1 trillion in 2015 to $2.7 trillion by 2030, or from 4 to 5.8 percent of GDP.2The Education Commission Secretariat (2016) analyzed government expenditures on education as a function of GDP per capita, region, fragility and population, and projected an ambitious path of government education expenditure based on the historical experience of countries with relatively higher spending relative to the average prediction. More details can be found in the Education Commission Analytical Background Paper at http://report.educationcommission.org/resources.  It projects that public expenditures expand to increase quality as well as coverage and are allocated according to the principle of progressive universalism. Governments allocate the bulk of public financing to pre-primary, primary, and secondary education, with a focus on what works to increase access and learning for the poor or disadvantaged. By gradually expanding coverage, they achieve quality, free education from pre-primary through to secondary school. That financing should also include progressively eliminating in-school incidental costs for households such as textbooks and learning materials. Based on country-by-country projections in the costing model, financing will be overwhelmingly – about 90 percent— for recurrent costs with the remaining 10 percent for constructing classrooms and other capital costs.3(2016) projects investment costs as the construction of classrooms (a simplifying assumption). Data or estimates of school construction costs from the literature and assumptions on the lifetime of schools and related maintenance costs are taken from the UNESCO costing model. According to the costing model, construction costs accounted for 9 percent of public expenditures in primary and secondary in LIC and MIC countries in 2015. Another estimate, from Development Finance International (2015), is 15 percent. Our investigation of the difference found that DFI includes six countries with very high capital spending. Of these countries, DFI says: “Many of the countries which spend higher proportions of their education budgets on ‘capital’ spending are countries receiving large amounts of project aid, so it may well be that a major explanation for the variations is that project aid (whether spent on recurrent or capital items) is generally classified in recipient country budgets as ‘capital’ spending. This skews the amount reflected in capital versus recurrent, especially in high- ly donor dependent countries, and thus does not reflect the normal ratios of capital versus current spend in education.” If these countries are excluded, then the average capital spending in LIC and MIC countries in the DFI report is 11 percent. Given the uncertainty of capital investment data in general, this can be said to be consistent with the Education Commission estimate.
• Households continue to bear a share of the cost, declining in volume from 1.5 to 1 percent of GDP as governments increase their share of financing. The pathway reduces household expenditure on preschool to secondary education very substantially, especially in low-income countries. Household financing will need to shift up the education ladder to post-secondary education, where even after taking account of reforms to reduce costs and of available government financing, there will still be a need for substantial household participation in cost-sharing mechanisms, particularly student loans and fees. However, in the case
  of upper-middle income countries, growth in public revenues would enable governments to assume more of the costs of post-secondary education, permitting household expenditures to decline.
• International sources including official development assistance (ODA), emerging donors, official non-concessional loans, and private development assistance — such as philanthropies, civil-society organizations, and corporate giving – would be available for all countries that need it, but would need to be prioritized according to where needs are greatest and where commitment to reform is demonstrated. The pathway allocates each of these flows to the three country income groups, with emphasis in allocations of concessional finance to the low-income group, which is made up primarily of fragile states. With increased domestic financing and efficiency, only 3 percent of the total financing package will be needed from international financing, primarily in low-income countries. This small share of financing still requires total international finance for education to rise by an average of 11 percent per year, from today’s estimated $16 billion per year to $89 billion per year by 2030. ODA would need to rise by 9 percent per year, from today’s $13 billion per year to $49 billion per year. This is feasible if ODA increases to at least 0.5 percent of donor GDP, a wider range of actors engage in financing education, and education and health are prioritized equally by international funders, each at 15 percent of global finance. These funds will remain critical for low-income countries, covering nearly half of their education costs. These countries will be home to almost 20 percent of the world’s school-age children (three to 18 years) by 2030, and without this support they will fall irretrievably behind.

The pattern of financing is very different for the three income groups:

• Low-income countries would need to increase domestic public expenditures for education from $13 billion to $50 billion between 2015 and 2030, or from 3.2 to 4.9 percent of GDP. Overall domestic public expenditures would grow from 20 to 25 percent of GDP and the share of education in total expenditure would grow from 16 to 20 percent. International finance will remain critical for low-income countries, covering nearly half of their education costs. The combination of higher public and international financing allows household financing for education in these poorest countries to decline from 2.5 percent of GDP – double that of middle-income countries — to just under 1 percent by 2030.
• Lower-middle income countries would increase domestic public expenditures for education from $214 billion to $712 billion, or from 4.1 to 6 percent of GDP. Domestic public expenditures would increase from 27 to 32 percent of GDP, still lower than all but one OECD country, and the share of education would increase from 15 to 19 percent.
• Upper-middle income countries would increase domestic public expenditures for education from $779 billion to $1.93 trillion, or from 4.5 to 6.3 percent of GDP. Public expenditures would go from 32 to 37 percent, and the share of education from 15 to 18 percent.4Although some UMIC countries could raise more domestic funding for education, in those countries domestic finance is capped when all of the education costs are met. Thus, those countries could in fact invest and achieve more in access and learning than called for in the Education Commission projections.
• In lower and upper-middle income countries, government budgets are able to finance an increasing share of the growing costs of education. International finance plays a very minor role (under 1 percent of GDP in 2015) and will be declining over time. In these countries, international finance is important, however, for its catalytic effect in encouraging greater domestic resource mobilization and reform.

The Commission costing model

The Education Commission financing model is built on the UNESCO 2015 costing model which was the model developed to estimate the costs of reaching the SDGs and the associated finance gaps.5UNESCO. 2015. “Pricing the right to education.” Education for All Global Monitoring Report. Policy Paper 18. UNESCO: Paris. The Commission substantially added to this model. In addition to the inclusion of more countries (adding the upper-middle income group), the Commission model includes projections of learning, post-secondary, and the top 25 percent or country-specific trends for future pathways rather than the fixed targets of the SDGs. It also includes an option to project the impacts of specific interventions on education costs and outcomes. The model uses a detailed bottom-up approach and projects education progress of students by grade over time from pre-school to secondary.6With a few exceptions, the data were obtained from international UN databases, such as the World Bank Development Indicators, the UNESCO Institute for Statistics, the IMF, and the UN Population Division. Data on learning were obtained from international assessments (PIRLS, TIMSS, PISA, SACMEQ, LLECE), and data on intervention impacts from a background study (Conn, Katherine. 2016. “The Effectiveness of Education Programs Worldwide: Evidence from a Meta-Analytic Dataset.” Background Paper for the Education Commission). Where data were not available, estimates were made based either on regional averages, or on predictions using coefficients from multi-variable OLS regression models. Costs are the sum of teacher salaries, other recurrent costs, capital investments, and support for marginalized students or specific interventions (depending on the scenario).

Assumptions Used
The table below presents a summary of the assumptions that were used for the Learning Generation “vision scenario” discussed in the report. The assumptions regarding access to education – from preschool to post-secondary – and the assumptions regarding improvements in learning are taken from the Learning Generation aims. Assumptions regarding resources – teachers, salaries, other recurrent costs, construction, and support for marginalized students – are the same as those used in the UNESCO model. These include, by 2030, a convergence of teacher salaries toward the average levels of the top-paying half of countries (controlling for average incomes); and added spending for poor children ranging from 20 to 40 percent of base costs. The assumed costs cover the envisioned access and quality improvements (if programs are effectively implemented) based on calculations using evidence on the costs and impacts of various practices.

The Commission’s projections were developed on the basis of detailed analysis regarding maximum achievable expansion and improvement rates and the most cost-effective ways these could be achieved. Specific growth and spending paths included are illustrative and not prescriptive – some countries will be able to go further or will choose to prioritize spending differently. For the costing estimates, the Commission made some ambitious but feasible general assumptions, but recognizes that each country’s strategy for achieving higher quality will be unique.

Summary of assumptions for the scenarios

The key assumptions include:

Expansion rates:
The expansion rates envisioned by the Commission, particularly in low-income countries, are rapid but achievable for almost all countries. They are based upon the rates achieved by the 25 percent of countries whose rates of growth most outperformed that of countries with a similar starting point on a given measure.

These rates mean that, for example, the number of secondary students in low-income countries is projected to increase from 36 million in 2015 to 94 million in 2030. The average annual growth rate in secondary pupils would be 6.9 percent – this is more than one and a half time faster than the average rates achieved since 2000 by countries who had at that time levels of secondary enrollment comparable to those in low income countries today.7Because growth rates decline steeply as enrollment rates increase (slowing to near zero as 100 percent is approached) one needs to compare the projected growth rate of LIC countries 2015-30 to a group of countries that is similar to where the LICs are in 2015. In 2000, the LMICs that were not former Soviet states had an average lower secondary GER of 58 and upper secondary was 38 (UIS data via Edstats). This is similar to the LIC average enrollment rates of 2015, estimated at 50 for lower secondary and 31 for upper secondary, making these two broadly comparable groups. The average historical growth rate of the LMIC group from 2000-2015 was 4.3 percent (computed from UIS data via EdStats); while the average projected growth rate of LIC group is 6.9 percent annually.
Even faster growth is projected for preschool and post-secondary in low-income countries: preschool pupils are projected to rise from 5 million to 21 million by 2030, and post-secondary students from 5 million to 17 million by 2030.

To support this expansion rate, the number of teachers needed in low-income countries will double in 15 years, from 3.6 million to 6.6 million. This would require an average of more than 60 percent of tertiary graduates from 2015-30 to go into teaching.8Even if non-tertiary post-secondary graduates are included, 30 percent of all post-secondary graduates would need to go into teaching. The number of classrooms would need to grow at roughly the same rate. Overall, the expansion as proposed requires huge effort by all; achieving it within a realistic funding envelope constrains some options for how quality can be improved, for example reducing class size more rapidly.

Teacher salaries:
The Commission’s costings assume that teacher salaries will, on average, be improved. Assumptions are based upon the historical trend relationship between teacher salaries and GDP per capita, and the assumption that countries will gradually converge at a global average. Relative to GDP per capita, teacher salaries tend to be higher in poorer countries because relevant skills are scarce. In the Commission’s projections, average salaries rise to meet those of the top 50 percent of teacher salaries globally (controlling for GDP per capita). This is to ensure that that salaries are sufficient to attract qualified candidates into the profession. In the very poorest countries, teacher salaries would be seven to eight times the average GDP per capita. There is evidence that raising overall teacher salaries has an impact on learning, more so in the long-run as better candidates choose to become teachers and in some cases when those increases are linked to student learning improvements.9See for example Dolton, Peter, and Oscar D. Marcenaro-Gutierrez. 2011. “If you pay peanuts do you get monkeys? A cross-country analysis of teacher pay and pupil performance.” Economic Policy 26, No. 65 (2011): 5- 55; Bruns, Barbara, and Javier Luque. 2014. Great teachers: How to raise student learning in Latin America and the Caribbean. World Bank: Washington, DC. However, the impact of salary increases needs to be weighed against the impact of other investments in teachers. The Commission’s salary assumptions reflect evidence on the impact of improved pay alongside further evidence that teacher quality and engagement can also be improved by other, complementary, and often lower-cost measures, as discussed in this report. In the costing assumptions, the salaries for teachers in lower-secondary are assumed to be 50 percent higher than those of primary teachers, and upper-secondary teachers another 20 percent higher. This is in line with observed salary scales in Africa and Asia today.

Non-salary spending:
Recurrent spending on items such as high quality learning materials, in-service training, teaching support programs, special education programs, management improvement, and other programs is projected to rise substantially, reflecting evidence that schools need to be much better resourced, and that teachers and students need far greater support. As discussed throughout this report, leveraging existing resources (teachers and classrooms) better through more effective practices and reforms can improve outcomes for relatively low cost. In low-income countries, this category of spending is projected to grow from 10 percent of total education costs in primary in 2015 to 30 percent by 2030 (in lower-middle income countries, from 15 to 35 percent). At secondary level, this category is projected to grow to more than one quarter of total education costs. In addition, resources to help marginalized students start and stay in school are added to the cost assumptions.10These supportive measures are assumed to cost 20 percent of base unit costs for marginalized primary students; 30 percent for lower secondary; and 40 percent for upper secondary. The marginalized students are those who are categorized as living in extreme poverty. The subsidies are assumed for additional poor students (because poor students who are already completing primary or secondary already have, through existing programs, support to get them through school). Because the measures are provided only to a portion of students, typically this cost item adds less than 10 percent to overall costs. This adds an average of 6 percent of primary costs in low-income countries; 9 percent at lower-secondary; and 10 percent at upper-secondary by 2030.

Class sizes:
Evidence on the negative impact of very large class sizes is strong, so costings assume that average class size in low and middle-income countries will be reduced – in preschool to an average of 20 children per teacher; in primary to an average of 40; and in secondary to an average of 35. While many countries opt to invest in smaller class sizes than this, evidence on the effectiveness of this policy is mixed and the marginal benefits drop once a class size of around 40 is reached.11These estimates are based on a variety of studies which include: Lockheed, Marlaine E., Adriaan M. Verspoor, and associates. 1991. “Improving Primary Education in Developing Countries.” Oxford University Press for the World Bank: Oxford; Bruns, Barbara, Alain Mingat, and Ramahatra Rakotomalala. 2003. “Achieving Universal Primary Education by 2015: A Chance for Every Child.” World Bank: Washington, DC; EQUIP 1. 2007. “Large class sizes in the developing world: What do we know and what can we do?” USAID: Washington, DC; and UNESCO. 2006. Global Monitoring Report: Education for All – Strong Foundations. Early childhood care and education. UNESCO: Paris. A teacher:pupil ratio of 40 is widely used as a benchmark and, given the great demands on teacher supply and overall resource constraints, reducing much further than this, on average, is unlikely to be achievable or cost-effective relative to other, lower-cost measures for improving teaching and learning.12Organizations such as UNESCO or UIS use 40 as a benchmark – for example: UIS. 2015. “Sustainable Development Goal for Education Cannot Advance Without More Teachers.” Cost implications of further reductions would be considerable – for example, reducing the primary school pupil:teacher ratio to 20 would result in the total education costs in low-income countries rising to $130 billion in 2030 (rather than $102 billion) and the external finance needs would be $71 billion by 2030 (instead of $45 billion).

Construction costs:
The cost of classroom construction can vary widely, depending on the procurement method and who builds the classrooms. The model assumes that there is one classroom for every teacher and that these need to be furnished adequately, maintained annually, and incur recurrent utility costs.13Leathes, Bill, Roger Bonner, P.K. Das, Ripin Kalra, and Nigel Wakeham. 2004. “Delivering Cost Effective and Sustainable School Infrastructure.” The TI-UP Resource Centre and DFID: London. This brief provides a discussion of different costs of constructing classrooms depending on procurement, and estimates of the costs for furniture and maintenance. The benchmark cost of classroom construction was taken from the UNESCO model, which obtained it from various country studies (no international database of classroom construction costs exists).14A brief and clear discussion of classroom construction and estimates of costs in developing countries can be found in Theyndyck, Serge. 2003. “Education for All: Building Schools.” Policy Notes. World Bank: Washington, DC. All of the dollar values found in this report as well as other sources were converted to multiples of GDP per capita to be used in the projections, so construction costs rise with general incomes as one would expect. The total costs for construction are higher when education is expanding, because of the need to add new classrooms. In low-income countries in 2030, construction costs are assumed as 18 percent of the total for preschools; 14-15 percent in secondary; but only 9 percent in primary.