Original language: English

Santosh Mehrotra (India)

Studied at Nehru University, New School for Social Research and Cambridge University (Ph.D., 1985). Taught at Nehru University before joining UNICEF headquarters, New York, as an economic advisor and policy analyst. His most recent books include Development with a human face (1997) and India and the Soviet Union: trade and technology transfer (1990). At UNICEF he works on the links among macro-economic policy and health and education, and on the economics of basic education. Currently directing a major research project on public expenditure on basic social services in developing countries.

Primary school enrolment in sub-Saharan Africa increased rapidly after independence. However, between 1980 and 1990 the primary gross enrolment ratio fell from 81 to 71% under the impact of economic crisis and adjustment. Sub-Saharan Africa is the only region in the world to witness a declining gross primary enrolment ratio in recent history. While gross enrolment ratios on average declined on the continent, about half of the countries managed to increase their gross enrolment ratios after 1980. Eastern and Southern African countries seem to have done better than the countries in Western and Central Africa.

Achieving universal primary education (UPE) over the next ten years will be a major challenge for many sub-Saharan African countries. Given the high population growth rate, low enrolments and relatively high per pupil costs, the achievement of UPE in all developing countries over the next ten years will require about US$25 billion—an average of US$2.5 billion per annum in recurrent expenditure (UNICEF, 1995). This latter figure is equivalent to an additional recurrent allocation of approximately 0.7% of GNP to primary education.

The resources required in Africa are much more than those required in Asia and Latin America to achieve UPE. Part of the explanation lies in the large number of children still out of school, and part in the relatively high unit cost of primary schooling in Africa (see following section). Western and Central Africa will need to spend a total (the current level of expenditure plus the additional expenditure required to reach UPE) of about 2.5% of GNP each year between 1995 and 2005. Eastern and Southern Africa, on the other hand, will need to spend 2.1% of GNP. This would imply an increase in the real primary education budget of 5-6% every single year over the next decade—an unlikely scenario.

Given the substantial additional resources required, progress towards UPE will need major reforms which can be classified in two groups: more cost-effectiveness and additional financing.

Improved cost-effectiveness and greater diversification of financing are not the only reforms required to achieve UPE. Achieving UPE also requires pedagogical reforms: improving the preparation and motivation of teachers and strengthening institutional capabilities to manage the education systems; improving the curriculum; improved ways of learning; and sufficient time for learning. Since this paper is devoted to the cost and financing of education, pedagogical aspects will not be dwelt upon in any detail. However, they are critical in addressing the first question of how to use existing funds more efficiently. This is because lowering costs and raising funds for primary education are only means to the end of universal access with learning achievement. Pedagogical issues are closely related to many of the cost and financing reforms we discuss below, and an attempt is made to take those issues into account.

Options for improving cost-effectiveness

Although many sub-Saharan Africa countries will have to allocate more money to primary education, it will in most cases not be a sufficient condition to improve the situation. Efficient use of resources is a major issue. Unit costs are best expressed in relative terms because the principal input—teacher remuneration—is related to the level of development of the country. But even in relative terms, the unit cost of primary education is higher in sub-Saharan Africa than in Asia, for instance. The cost of primary education per pupil-year is equivalent about to 15% of per capita GNP, compared with 11% in South Asia and 8% in East Asia (UNESCO, 1995). Higher education is much more expensive in sub-Saharan Africa than in either South Asia or East Asia, with per student recurrent expenditure being 5.1% of per capita GNP in sub-Saharan Africa in 1992, as compared to 0.8% in South Asia and 0.9% in East Asia. Whilst economies of scale may account to some extent for the lower cost in Asia, there are other factors such as high boarding costs and student stipends which account for this discrepancy. In other words, different patterns of costs that have historically been practised, but not essential in pedagogical terms, account for a large part of the difference. In this section we explore six principal options for improving value for money.

IMPROVING TEACHER EFFICIENCY

Salaries represent the lion's share of the education budget. UNESCO reports that seventeen out of twenty-six African countries for which information is available spend more than 95% of their primary education budget on salaries (UNESCO, 1995). Although teachers in many African countries are poorly paid, they are also often poorly utilized.

Teachers' salaries are historically higher than per capita incomes in Africa. For example, in 1980 a primary teacher's salary in French-speaking Africa was 7.8 times the per capita income, whereas in English-speaking Africa it was 4.6 times more (Donors to African Education, 1994). High inflation and fiscal austerity have reduced real salaries in most countries. Nevertheless, they remain relatively high compared to either per capita incomes in their own countries or to those in Asia. In Asia the primary teacher's salary was only slightly higher than the per capita income (Tan & Mingat, 1992). At most it was double the per capita income. One reason for the high teachers' salaries in Africa relative to per capita income was the historical shortage of human capital, particularly in French-speaking countries.

As countries grow wealthier, teachers' salaries relative to per capita income fall because the education of an average income earner rises relative to teachers' education, and the proportion of the population represented by the labour force age also increases as countries grow richer. As such, Carnoy and Welmond (1996) rightly suggest that deviations of the ratio of average teachers' salaries relative to income per capita from an income per capita trend line is a more accurate measure than salaries relative to income per capita of the value countries give their teachers at different levels of development. Even by this measure, teachers' salaries in many African (mostly French-speaking) countries are relatively high: the Central African Republic, Rwanda, Mali, Ethiopia, Niger, Mauritania, Zimbabwe, Côte d'Ivoire, Burundi, Benin and Togo.² In the 1980s, teachers' salaries fell in real terms and, although it is virtually impossible to compress them further, it may be possible to decrease the relative size of the wage bill in the total budget for primary education. We discuss two methods of achieving this result.

Rationalizing the pupil-teacher ratio

The first prerequisite for containing the relative share of the wage bill is to prevent the number of teachers from rising when enrolments are falling. One management tool is to control the pupil-teacher ratio very carefully with a view to improving teacher utilization. In sub-Saharan Africa there was a dramatic increase in the number of teachers, which grew by 24% between 1985 and 1990. Over the same period the number of students increased by only 11% (Donors to African Education, 1994). In most parts of Africa, a ratio of one teacher for between forty and forty-five pupils, though far from ideal, has been found to be reasonable.³ The majority of African countries have teacher-pupil ratios well below these levels (UNESCO, 1995).⁴ For example, increasing the pupil-teacher ratio from thirty-five to forty may bring about a saving of more than US$6 per pupil per year (given that recurrent expenditure per primary pupil in sub-Saharan Africa is US$44), thus enabling the school to spend additional resources on learning materials.

Average pupil-teacher ratios hide substantial heterogeneity within countries. It is not uncommon to find situations where some schools, particularly in urban areas, are seriously overcrowded. This must necessarily mean that some schools have less than twenty pupils per teacher. Therefore, any attempt to modify the class size must pay attention to the spatial distribution of schools and teachers.

A careful monitoring of the pupil-teacher ratio in the country by district and by school is a useful means of controlling the salary component of current expenditure. How may this be achieved? Three ways of rationalizing the pupil-teacher ratio without jeopardizing the quality of education and demoralizing teachers could be as follows.

The most obvious method is to relocate teachers (and to build additional classes where needed so that every child will have access) in a way which decreases the overcrowding of some schools (e.g. in low-income urban areas) by increasing the class size in others (e.g. affluent urban areas). Even if this relocation increases the average pupil-teacher ratio, it could improve quality by reducing overcrowding.

An alternative option is to introduce multi-grade teaching where the pupil-teacher ratio is low, which requires special training for teachers, as well as special educational materials. The implementation of such measures requires adequate 'school mapping'. However, deploying and utilizing teachers efficiently is no easy task. In many countries, the population is so dispersed that a primary school within a reasonable distance cannot have thirty to forty students in each class. Multi-grade teaching then is essential to make efficient use of teachers. Multi-grade schools should not be regarded negatively as a second-class solution. In France, the United Kingdom and the Netherlands, multi-grade schools have been established in small communities in response to falling pupil intakes and staff reductions, and in small towns and villages as an alternative to closing uneconomical single-grade schools (Thomas & Shaw, 1992). To succeed, multi-grade schools require modified pedagogical techniques—self-directed learning, peer tutoring, careful lesson planning and appropriate texts.

Another way of increasing pupil-teacher ratios is to increase the average teaching load. For this measure to succeed, the number of classroom hours must be increased only for teachers and not for pupils. This can be done by increasing the number of hours taught per day, days per week or weeks per year—or any combination thereof. This means some system of double shifting—i.e. splitting the school day, week or year—with each session catering to half of the pupils enrolled. Teachers' salaries can be increased on account of the increased workload, but per pupil costs will fall if the increase in pupil-teacher ratios is greater than the increase in teachers' salaries. Variations of this approach have been tried in many countries. For example, when primary education in Sweden was first made compulsory a century ago, teachers worked a six-day week, whereas pupils attended only three days of school per week. In many parts of the United States today, Grade 1 pupils are taught for only three hours per day in classes of twenty pupils, but Grade 1 teachers normally take two such classes per day.

A variant is to reduce the average attendance of pupils—either by decreasing the number of hours per day, days per week or weeks per year—without reducing the workload of the teachers (i.e. introduce double-shifting). There may be a tradeoff with learning achievement, although research indicates that—beyond a minimum threshold—the length of instruction in the initial years is not a strong determinant of academic performance in subsequent years. Countries in which all primary grades require the same number of hours per week could reduce the length of the school day for the lower grades to permit a second shift (Lockheed & Verspoor, 1991). Parents may welcome a shorter school day since it may reduce the opportunity cost of the child's schooling. In Bangladesh, the primary education system is based on double-shifting: Grades 1-3 meet for two hours in the morning, and 4-5 for three hours in the afternoon. It is essential that limiting the number of hours a learner is at school is done judiciously. There is a trade-off between reaching large numbers with limited resources and offering high-quality education to a small number. A balance has to be struck to protect an essential degree of relevance and quality while reaching the largest possible number of children.

The choice of any of these options or a combination of them as the most appropriate one will depend on the national and sub-national conditions. Flexibility and a high degree of decentralization on the part of the local school and community are of big importance in policy design.

Using teacher assistants and community volunteers

The hiring of teacher assistants and community helpers, who are provided with adequate but low-cost training, could further lower the relative size of the salary component. The introduction of teacher assistants could play an important role in those situations where teachers are working longer hours in order to accommodate double-shifts or splitting of the school calendar. They could also play a significant role when the coverage of the system and enrolments are expanding rapidly. Teacher assistants are well integrated into the American system of primary education today, allowing teachers to concentrate their specialist skills on smaller groups of children at a time, whilst teacher assistants supervise the practice of knowledge and skills with other groups. In Guinea, the introduction of pairing between a government-employed teacher and a community-employed para-professional has been pivotal in persuading communities to participate in schooling. Incidentally, this strategy also lowered unit costs. State grants to parental and community groups to allow employment of community members within the school system can be more efficient than the State directly employing more teachers. Education, particularly at the primary level, requires a great deal of practice, and this practice can be done with teacher assistants, as well as with classmates, parents and community volunteers. Employing the most qualified youths from the village in rural schools will address another problem—that qualified teachers of urban origin often resist working in rural areas. In order to restore teacher morale and motivation and reduce teacher absenteeism (partly caused by second jobs), teachers' salaries will have to be increased over time (Adedeji, Green & Janha, 1995). Large increases in salaries may not be affordable in slow-growing economies, especially in economies where teachers' salaries are linked to other public service pay scales. Competing claims on expenditure (e.g. for educational materials, teacher training) may also constrain increases in teachers' salaries in the short run—but in the medium run teachers' salaries will have to increase to improve incentives.⁵

If the objective of increasing teachers' salaries is to be realized, it is absolutely critical to monitor pupil-teacher ratios in the short run because fiscal constraints cannot permit the simultaneous increase of the total teacher force and an increase in salaries. Teachers have some political clout, and with the recent trend of democratic politics in the region, there is a tendency to protect teachers (ILO, 1996).

Adjustment programmes during the early 1990s emphasized protecting or increasing expenditures on non-salary inputs. The examination of incentive systems as a means of improving performance in education is more recent. The overwhelming emphasis on cutting civil servant salaries (to which teachers' salaries were tied) and public sector employment had the effect that the impact on teachers, and in turn on their performance and productivity, remained largely ignored. That will need to change if the recovery in primary education in Africa is the ultimate goal of the reform programmes.

REDUCING REPETITION AND DROP OUT

Even where the per pupil unit cost is low, the cost per primary graduate can still be very high because many children repeat grades or drop out before completing the primary cycle. Repetition rates tended to go up during the 1980s. In French-speaking Africa the median repetition rate was 22% in 1980 and 25% in 1990; in English-speaking Africa it was 11% and 15% respectively (Donors to African Education, 1994). In at least seventeen countries—most of them French-speaking—repeaters occupy more than a quarter of the available school places—an inefficient way of utilizing scarce resources. Pupils in low-income countries generally take four additional years to complete the primary cycle as compared to almost two (1.8) years in low-middle income countries and just over one (1.2) year in upper-middle income countries (Lockheed & Verspoor, 1991). It is paradoxical that countries where the waste of education resources is the highest are the ones that can least afford it. Lowering the repetition rate necessarily entails improving learning achievement.

Less than half of all the children in the region complete five years of primary education. In addition to low enrolment, about one-third of those who start primary education drop out before completing fifth grade (UNICEF, 1995). This partial investment yields a very low return for society, because four to five years of primary education is considered a minimum to retain permanent literacy and numeracy.

Studies show that the drop-out rate is highest among the repeaters. Therefore, a policy of automatic promotion may be more cost-effective. Many of the countries which achieved UPE in the early years of their development (such as Sri Lanka, Kerala state in India, Zimbabwe, the Republic of Korea, Malaysia and Barbados) adopted automatic promotion in the early grades (Mehrotra, 1997). Automatic promotion may be a simple way of reducing repetition and saving on costs, because lowering the mean time required to complete primary education implies that more children can be catered for. However, appropriate measures are necessary to address students who have difficulties—otherwise the problem is merely deferred to the end of the primary cycle (IBE & UNICEF, 1996). A policy of automatic promotion does not mean the removal of standards or performance criteria. It merely recognizes that not all children learn at the same speed in all subjects. Automatic promotion helps children to learn to the best of their potential within the time-frame of the primary cycle without leading to unnecessary repetition. Instead of repetition, extra help and attention are given to those who fall behind in some areas.

Such remedial classes require special courses for teachers on how to handle slow learners and manage the heterogeneity of students, as well as adequate testing and diagnostic systems within the primary education cycle. Establishing such units within the primary education system may be much more cost-effective than permitting repetitions.

Repetition is highest in the early grades as well as at the end of the primary cycle as students try to improve their chances of passing the exam for entry to the secondary level. The majority of sub-Saharan African countries have to restrict entry to secondary education to a small percentage of primary school leavers because of the limited places available. A possible cost-effective strategy could be to separate the primary cycle leaving exam from the secondary cycle entrance exam. The opening up of secondary education opportunities, using the savings which would accrue from stopping primary school repetitions, is an option taken by countries such as Zimbabwe—which increased secondary school enrolments from about 4% at independence in 1980 to about 50% ten years later.

Increased opportunities for secondary education generally enhance the motivation to complete primary education. Secondary education is also likely to increase the pace of development, including economic development, as secondary graduates are more likely to develop independent entrepreneurial skills than those with only primary education (Crouch, 1993). Secondary school unit costs can be reduced sharply through the adoption of lower cost models, such as reducing the number of boarding schools. In Zimbabwe the capital costs of a boarding school are seven times that of a similar size day school, and recurrent unit costs of boarders are three to four times those of day students. Hence Zimbabwe preferred to expand day schools. Developing schools closer to the children's homes allowed more girls to attend and lowered costs for all (Chung, 1993).

INCREASING ACCESS TO EDUCATIONAL MATERIALS

If repetition and drop-out rates are to be reduced by improving the quality of education, increasing access to educational materials will be a key strategy. While the latter is likely to increase the total costs of education, its cost-effectiveness is beyond doubt. If improved quality increases completion rates, the additional costs will be offset and unit costs per graduate will fall. Apart from improving teachers' motivation and skills, teaching materials may be the most cost-effective input for learning. UNESCO data indicate that non-salary expenditure per primary pupil in sub-Saharan Africa amounts to only US$3 per year, against a recommended minimum of US$5 per pupil per annum (Colclough & Lewin, 1993).

One major reason for worsening quality is that education materials have unduly suffered from years of fiscal austerity. A 1995 UNICEF and UNESCO survey of primary schools in fourteen of the world's poorest countries showed that even chalkboards were missing from 79% of the classrooms in Burkina Faso, 54% in Tanzania, 51% in Madagascar, 48% in Equatorial Guinea, 44% in Zambia, 36% in Ethiopia and 35% in Uganda (Schleicher, Siniscalo & Postlethwaite, 1995). In most African countries, notebooks, pens and chalkboards have to be provided by the household. Moreover, in many countries books are imported and expensive, partly because of agreements among governments, former colonial country publishers and donor agencies.

Given that few countries have the expertise to write good textbooks, they are often written by committees of international experts or are imported rather than stimulating and strengthening national capacities and institutions to develop and produce quality textbooks (Torres, 1995). Donors, in particular, need to guard against such practices. It may be particularly useful for donors to support a group of countries to publish textbooks—which can yield very substantial economies of scale and hugely lower the cost of textbooks.⁶

Learning materials that require teachers to follow a step-by-step pedagogical routine and ensure that children are engaged in an active learning process have been found to be a great help in improving pupil achievement, as shown by the experience of escuela nueva in Colombia and BRAC in Bangladesh (Anderson, 1992).

The importance of teaching materials, however, should not overshadow the ultimate role of the teacher. For years, the World Bank has emphasized resource allocation to educational materials, almost at the expense of teachers' living and working conditions. Given the weight of World Bank lending to education, this advice has influenced decision-making. However, the surest and most direct manner of improving the quality of education is by enhancing the motivation, preparation and working conditions of teachers. Textbooks are educational tools; teachers are the educational agents (Torres, 1995).

REDUCING TEACHER-TRAINING COSTS

In the previous sections, the importance of better trained teachers has been stressed. As training is costly, ways to reduce these costs should be looked into. In many countries, a large proportion of the intake in teacher-training colleges has not completed secondary education. Empirical evidence indicates that teachers' behaviour in the classroom tends to be based more on what they experienced as students themselves than on what they learned in their formal preparation as teachers. Pre-service and in-service training are to a large extent compensatory and remedial strategies for a deficient education provided by a low-quality school system.

One way of reducing teacher-training costs is to shift the secondary education component of pre-service training to general secondary schools, and reduce the length of pre-service training.⁷ However, such a measure is contingent upon ensuring that an increasing proportion of all newly recruited teachers have completed secondary education. One example of this is in Zimbabwe, where secondary teachers' colleges recruit students who have completed 'A' or Advanced levels. They undertake one year of pre-service training, followed by one year of in-service training, in contrast to the past when students were recruited with only 'O' or Ordinary level passes, and spent three years in pre-service training.

Another way of reducing costs is to increase school-based in-service training, which could yield savings in both boarding and tuition costs as well as in teacher salary costs, because part of the training would consist of on-site in-school teaching. The problem, however, is that while the proposal to increase in-service training as a substitute for lower quality pre-service qualifications of new teachers is reasonable, the practical difficulty is that it is not implemented on account of concurrent shortages in qualified teacher trainers and limited public funds. Meanwhile, because unemployment among the educated is often quite high, there is still an incentive to join the teaching profession, even though salaries may be declining in relative and absolute terms. But many countries in the region undergoing adjustment have not followed recommendations to recruit less-educated teachers even when teachers' salaries have fallen sharply—thus putting pressure on the total salary bill (in addition to squeezing out non-salary expenditures).

The limited funds to provide in-service teacher training has resulted in innovative solutions. Efforts need to be made to recognize the innovations of teachers in teaching methodology, and allow for experiences to be shared. In Zambia each school could be called a teacher-training centre—a teacher can become a resource teacher who has the duty of improving the quality of teaching of colleagues (UNESCO & UNICEF, 1996). The emphasis is on the fostering and nurturing of replicable innovative techniques.

COST SAVINGS IN DONOR ASSISTANCE

Although donor assistance to primary education has been limited (an issue we discuss later), the modalities of donor assistance also need modification. Donors have normally adopted a project approach, and a multiplicity of donors in any given country can lead to a large number of projects that are not well integrated into a primary education sector strategy. In a project approach (as opposed to a sectoral approach), costs can be very high because of the duplication of management structures.⁸ Thus donors could contribute to cost savings in projects if they adopted a sectoral approach, and ensured better co-ordination among themselves and the ministry of education (MOE). However, this requires that the recipient government takes the lead in formulating a sectoral strategy and actively takes on the task of coordination between the multiplicity of donor projects. The sectoral approach encouraged by the United Nations Special Initiative on Africa and the Sectoral Investment Programmes of the World Bank can facilitate the realization of such cost savings.

COST SAVINGS ON SCHOOL CONSTRUCTION AND FURNITURE

Some cost reduction is usually possible in respect of furniture. Desks and chairs can be made locally at comparatively lower cost. In a context of community management of schools, there should be an opportunity of getting parents and teachers to decide on priority and essential items for their school. This will encourage schools to provide simpler furniture, such as locally produced mats, thereby cutting furniture costs in general.

Expensive design standards and imported construction materials for primary schools are not uncommon in sub-Saharan Africa. There is scope to make more use of local materials, which will reduce costs and facilitate the transfer of responsibility for primary school construction and maintenance from the central government to local authorities and communities. Imported designs and materials often imply poor maintenance on account of high costs. Local materials can also improve the quality of construction, while decreasing costs. A survey in six sub-Saharan Africa countries indicates that brick and mortar buildings are, on average, more than twice as costly as the construction of schools with local materials. The cost range varies between 1.8 in Mali to 3.4 in Senegal (Lockheed & Verspoor, 1991). In Malawi it is reported that the construction cost of a primary school has been reduced by a factor of five by using local materials and community contributions. State grants to communities for construction and maintenance could be much more cost-effective than construction by the State.

Aid-financed school building projects can also result in high unit costs. In several countries, there is evidence that buildings financed from donor resources are more costly than locally financed schools.⁹

Finally, a cautionary note: there are clear risks of introducing a series of cost-cutting measures simultaneously, without giving serious thought to their implications for the quality of teaching and learning. In Tanzania, the World Bank and other donors have been calling for a reduction in per pupil expenditures—in a country that experienced a decline in enrolment over the 1980s. This has resulted in the implementation of a number of cost-cutting measures: use of volunteer teachers, use of distance and in-service rather than pre-service training and cost-sharing with the population. While these measures have reduced costs, they have also adversely affected equity and teachers' conditions (ILO, 1996).

Options for diversifying the sources of financing

We noted in the first section that the additional costs of reaching UPE in sub-Saharan Africa are very substantial, amounting to 0.7% of GNP, much more than in any other region. While efficiency of resource use is important, more important is that the total resources available for primary education must increase if UPE is to be achieved. Sub-Saharan Africa spends more on education as a proportion of GNP than Latin America, South and East Asia, the Middle East and North Africa (UNESCO, 1995). However, sub-Saharan Africa's low per capita GNP level means that absolute expenditure is much lower than in the other regions. Moreover, the region also suffers from the most serious shortage of human capital. So while there is scope for better utilization of existing resources, there is an overwhelming case for increasing resources for primary education. This case is strengthened by the fact that education expenditure declined throughout the 1980s, along with the primary enrolment ratio.

Seven principal sources of funding for primary schooling are discussed here.

GOVERNMENTS

Given the widespread policy of fiscal conservatism on account of adjustment and stabilization programmes, public expenditure as a share of GNP cannot be expected to increase in most countries. Economic growth, which was below population growth in the 1980s, has remained sluggish in the 1990s. The growth scenarios for sub-Saharan Africa indicate that, at best, per capita income will grow by between 0.5% and 1% per year over the next decade. Growth is unlikely to lead to much increase in government revenue and education expenditures. Hence, the potential for raising additional taxes earmarked for primary education needs examination.

New taxes that are specifically earmarked for primary education have been levied in countries at different levels of income and can take the form of taxes on property, business, selected luxury commodities, payroll, imports and interest income.¹⁰ Such tax receipts could be used for quality-related inputs such as a national textbook fund or a community fund with matching grants from the government. It is important that such taxes supplement (not substitute for) existing budget allocations to education. However, the revenue potential of such taxes should not be overestimated.

Therefore, inter-sectoral budget restructuring must be considered as a way of releasing more resources for education. A larger share of the current package to the social sectors can be realized by redirecting spending away from the military, debt servicing and inefficient parastatals. Since a very substantial part of Africa's debt is external, the capacity of many governments to restructure public expenditure is severely constrained. Within external debt, multilateral debt accounts for an increasing proportion of the total debt service, rising from 8% in 1980 to 20% in 1990 and 41% in 1994 (United Nations, 1996). The multilateral debt relief for Uganda in 1997 specifies that resources released will be allocated to basic education. 'Debt for social services' swaps as promoted by UNICEF have helped to redirect interest payments to be re-invested into human resource development in the debtor country.

Controlling defence expenditure is equally important in a number of countries as a way of increasing the total package of funds for education. The scope for further reductions remains considerable, with 2.8% of the GNP being spent on defence and 2.3% on interest payments as compared to only 1.8% on primary education (UNDP, 1995).

Another area where there has been rather limited action is the privatization and commercialization of parastatals. But privatization has not taken place on any significant scale, and has been limited to a few countries, with six countries—Benin, Ghana, Guinea, Mozambique, Nigeria and Senegal—accounting for two-thirds of divestitures (World Bank, 1994).

Within the education sector, there is greater scope for intra-sectoral budget reallocation in favour of primary education. There is a strong argument for doing so on grounds of both efficiency and equity. Public expenditure per student at the post-primary level is considerably higher than spending per primary school pupil in all regions and countries of the world, but the discrepancy is the largest in sub-Saharan Africa. In Latin America and South Asia, expenditure per secondary pupil was about twice as high as spending per primary pupil and a little over two times in East Asia between 1980 and 1992, whereas in sub-Saharan Africa it is three and a half times as much. The discrepancy is much greater for higher education. In Africa, public expenditure per university student is equivalent to government spending on about thirty-five primary school students. The corresponding ratio for Asia and Latin America varies between eight and fifteen for 1980-92. This is not to underestimate the importance of higher education for national capacity building or to call for a further reduction in public expenditure on post-primary education. But the wide discrepancies highlight the potential for a more effective and efficient use of available resources.

Indeed, as the World Bank's education policy review (1995a) shows, one domain within the education sector that could release resources includes student amenities such as food, accommodation, transport costs and so on. The budget share devoted to student welfare in secondary and particularly in higher education in sub-Saharan Africa is very high. In Asia, less than 10% of the budget is allocated to the welfare of university students, whereas the equivalent share is about one-quarter in English-speaking countries in Africa and one-half in French-speaking countries.

The state of African universities is such that they deserve attention and additional resources for essential areas—which makes inter-sectoral restructuring of public expenditure all the more necessary. A larger slice of the education pie for primary schools might be more acceptable to politically powerful groups if the size of the education pie was growing, thus allowing public resources to be allocated to these essential areas. Meanwhile, the view that the costs of student housing and meals should be recovered from students is increasingly accepted in Africa, mainly among English-speaking countries (Saint, 1992). But where students are crowded into dormitory rooms, lack sanitary facilities and live on poor food, the introduction of cost recovery would be unfair and may lead to student riots. Increased fees will have to be justified by improved quality of meals and accommodations. Meanwhile, the release of resources from inter-sectoral restructuring will enable essential expenditure to be undertaken on educational inputs, research and maintenance.

The World Bank has been calling for protecting health and education expenditures during adjustment, and public expenditure reviews normally recommend intra-sectoral reallocation of resources within the education sector (Toye & Jackson, 1996). However, the fact is that in the 1990s, education expenditures have risen in relation to GNP in only a few countries. Even in countries where macro-economic improvements have occurred, there is no clear evidence that primary schooling has benefited (Jayarajah, Branson & Sen, 1996).

DONORS

Negotiations with donors to provide official development assistance to basic education and establishing a compact between donors and governments are other areas for resource mobilization for primary education. In 1990-1991, a study of fourteen Western and Central African countries revealed that less than 3% of all Overseas Development Agency (ODA) grants went to basic education. In twelve Eastern and Southern African countries, the share was less than 2% of ODA (Mehrotra, 1994). Most international aid to the education sector in the 1980s went to secondary and higher education. The 20/20 initiative calls on governments and donors to spend 20% of the national budgets and aid flows on basic social services, including primary education.

For donor funding to become more effective, two changes will be necessary. First, the capital-output ratio of aid, which is extremely high (Mosley, 1996), must be lowered. In the education sector, one reason for the low effectiveness of aid is that a substantial part of it has been focused on the hardware of education (buildings, furniture and equipment) and technical assistance and rather less on capacity development, books, instructional materials and budgetary assistance for teachers' pay. Second, general recurrent support has been provided mainly to middle-income countries (Colclough & Lewin, 1993). Donors may consider concentrating future support more on recurrent costs and less on capital expenditure within a framework of a national commitment, plan and timetable for gradually phasing out reliance on external assistance for recurrent expenditure. There should be a special emphasis on the provision of books and instructional materials, especially the domestic production of books rather than those imported from donor countries.

At present a large proportion of donor aid is utilized for technical assistance from the North, generally from the donor country itself. If such aid were instead utilized to build up institutional capacities within recipient countries and to strengthen human resource capacities, this would have a more sustainable impact than outside technical assistance.

Most structural adjustment programmes have underestimated the impact of institutional capacity constraints in Africa on programme implementation. Planning capacity, personnel management and supervision capacity are weak or have been weakened, especially in the education sector (Makau & Coombe, 1994). Similarly, while adjustment measures have often called for decentralization of decision-making (in education and many other sectors), these requirements often do not take into account that personnel and financial management skills are deficient at the lower levels of government.

PARENTS

There is evidence that as a result of falling government support for primary education, households have been shouldering a growing burden of the costs, particularly at the primary level. There has been a proliferation of different kinds of fees. A 1992 household budget survey in Kenya showed that households' direct contributions covered 34% of the total cost of primary education. A recent survey in Zambia found that households spend five times more on education than on health care. Ghana's education sector reform has emphasized community-based primary education. Not surprisingly, statistical data from the latest two rounds of the Ghana Living Standards Survey show a substantial rise in household expenditure on education, but recent studies have found that both educational officials and parents reported a decline in the quality of education. These private contributions have seldom enhanced the quality of public schooling but merely substituted for government expenditure. An internal review at the World Bank on cost-sharing in education and health in sub-Saharan Africa concludes that user fees for primary education should be avoided (World Bank, 1995b).

UNICEF studies in two African (Burkina Faso and Uganda) and three Asian countries (Bhutan, Myanmar and Viet Nam) confirm the importance of private costs in discouraging school attendance. They found that private costs for uniforms, textbooks, building funds and parent-teacher association contributions ranged between 10 and 20% of per capita income. The figure is even more alarming if private tuition and in-kind contributions are added (Mehrotra & Delamonica, forthcoming).

Uniforms, which often constitute the single most important cost item for parents, have been subject to extensive policy debates in many countries. The main arguments in their favour, especially in the African context, are that they improve discipline, eliminate the difference between rich and poor children, and protect children in conditions of civil conflict. However, the recent experience of Malawi is relevant, where enrolment sharply increased in 1994 in the wake of the elimination of school uniforms and school fees. One possible policy is for the authorities to indicate that uniforms are not mandatory, and leave it to school administrations and parent-teacher associations to decide at the local level whether the uniform should be retained or dropped.

Primary education should be free and should not involve tuition costs. Regrettably, school fees at the primary level have emerged as a new source of financing, although educational fees as a share of unit costs of higher education are mostly low in sub-Saharan Africa. A survey of fifteen African universities found that only half of them charged fees—which generated on the average only about 10% of their recurrent budgets (World Bank, 1995a).

The historical experience from the early stages of development of both middle-and low-income countries such as Malaysia, Mauritius, Botswana, Costa Rica, Sri Lanka, Kerala state in India and Zimbabwe, is that they managed to eliminate or minimize tuition fees. Not all of the reduction in costs came at once, but were sequenced depending upon the fiscal capacity of the State. Thus Mauritius introduced free primary education in the 1950s, but free textbooks at primary level were introduced only in 1988. Tuition fees were eliminated in Sri Lanka in 1945; a free textbook scheme and a free mid-day meal began operating in the 1950s; a free school uniform came in 1991. In Botswana, enrolment received a major boost from the decision to cut fees by half in 1973, and were totally eliminated in 1980 (Mehrotra, 1997).

COMMUNITIES

Communities have traditionally provided support for school construction in sub-Saharan Africa to a greater extent than in other regions. Harambee in Kenya is probably the best known example of this practice (Lillis & Ayot, 1988). However, communities cannot be expected to increase their support under conditions when government support for primary education is declining, and when the out-of-pocket costs of their members are already very high. Their response is already clear in the form of falling enrolment rates in many countries.

Community participation normally occurs through parent-teacher associations and parent-school community organizations. Enhanced community participation will only be forthcoming if communities gain a stronger say in the organization and management of primary schools, and if the perceived quality of education increases in the early stages of cost sharing. Decision-making by the communities can comprise: hiring of school teachers or assistants (preferably from within the community); providing inputs into the school curricula; governing the finances of the school; participating in setting time-tables and calendars; and monitoring achievement.

There is little possibility that with the decline in incomes, villages will be willing and able to build houses for teachers or additional classrooms, without the community having a greater say than before in the running of the school. For communities to participate effectively in decision-making, there may be need for some training in management and participation—of both school heads and of members of the parent-teacher association and/or other citizens who represent the community at large. In this context, special attention should be paid to the training of women, as they play an important role in child education in the family. In addition, an effective mechanism has to be established to ensure that direct central government contributions will compensate for the different levels of income and development among communities (Bray with Lillis, 1988). Community participation should not serve as a foil for reducing the responsibility of the State. Community participation cannot be accepted as a euphemism for additional extraction and taxation. 'Decentralization' is sometimes utilized as a way for the State to divest itself of responsibility for education. Instead, decentralization should be a community-school partnership.

SCHOOLS AND PUPILS

Pupils can also contribute to school finances through income-generating activities, mainly by growing crops or producing goods with wood, metal and other locally available materials. The proceeds can be reinvested in the school to purchase teaching materials and textbooks or to supplement the salary of the teacher. However, such activities are often carried out at the expense of the time available for learning, thereby further undermining quality. In addition, the sums they generate are usually small, and they are not always reinvested in education materials. The management of the funds by headmasters and teachers has often been problematic. However, it is important that the lessons from past experience are learnt by evaluating school-based income-generating activities. The objective of these activities is not merely to support the school—its educational aspect is equally, if not more, important; the activity needs to be an integral part of the school curriculum.

The Self-Help Action Plan of Education (SHAPE) in Zambia has placed appropriate emphasis on self-help through production—and its efforts have reached almost all primary schools in the country, and hence deserve special attention. Education with production has three main frameworks—industrial arts, home economics and agriculture—but the scope of each activity changes from school to school because some skills (e.g. agricultural techniques) are easier to practice in some areas than others. Education with production has four main results: a) it facilitates learning based on practical knowledge, problem-solving, planning and initiative; b) it changes pupils' views of production and manual labour; c) it has financial benefits for the pupil and school through the sale of finished products; and d) its involvement of the surrounding population and the parents removes many of the barriers between the school and community (UNESCO & UNICEF, 1996).

PRIVATE SCHOOLS

Private schools are of three major types: community schools, mission and mosque schools, and for-profit schools. This section is mainly about the last category. Public resources can be released by giving the right incentives to the private sector to increase the supply of school places, especially in urban areas. Currently, cases of multiple providers of primary education are limited in many African countries. The share of private schools in primary enrolment is very low—under 10% in most sub-Saharan Africa countries. In the mid-1980s, private education played a larger role in English-speaking Africa (22% of enrolment) than in French-speaking Africa (3%) (World Bank, 1988).

By allowing more private schools, the supply of primary school places will be increased without adding to the financial burden of government. Private schools, because they usually charge higher fees, are targeting the better-off families. That means that the leakage of public subsidies to the non-poor will be reduced. The oversight function of private schools should remain vested in the MOE, but this should be one of facilitator, and not that of principal financier. This oversight function becomes particularly important in a situation where the private school may be the only one available in a poor area, offering relatively expensive education of poor quality.

PRIVATE ENTERPRISES

Private enterprises could be given greater incentives to support public schools in the area where they operate and beyond. This would be a non-profit endeavour— with returns in the form of public goodwill—although it could be argued that the productivity of employees will increase in return for the enhanced quality of education of their children. Such contributions to public schools could be made tax deductible.

The room for manoeuvre by local schools could be enlarged to undertake fundraising activities with the private sector. Unilever and Press Corporation in Malawi, for example, provide support to public primary schools.

But this is not the only role for the private sector in the public school system. Apart from being a service provider, the private sector can contribute in other ways: supplies, materials and transport. In other words, there is need for more innovative arrangements and partnerships. The State need not be the only producer/distributor of textbooks, supplies or transport services to the school system.

Conclusion

This paper has presented a menu of options in respect to reforms necessary to address the reasons for the decline in the enrolment ratio witnessed in Africa over the 1980s. Given slow economic growth and the need to keep fiscal deficits down, education budgets will remain constrained in most countries. Hence, more effective utilization of existing resources will be essential. Reducing cycle costs (through reduced repetition and drop outs), rationalizing teacher deployment and (where possible) increasing the pupil-teacher ratio offer the greatest scope for cost savings in many countries. Improving the quality of education through increased teacher effectiveness and greater investment in useful learning materials offer additional cost savings—essentially through reduced repetition and drop out. Investing in the teacher is the best guarantee of improved learning—by ensuring recruitment of only full secondary school graduates (assisted by lower paid assistants with lower qualifications), upgrading skills through in-service training (on-site in school), and providing the scope for teachers' innovations, so that each school is also a teacher-training centre where emphasis is put on teachers' initiative, resourcefulness and ability to train colleagues in new methodologies. Lowering school construction costs by using more appropriate standards and local materials could further reduce costs.

However, given that the total costs of achieving UPE were seen to be quite substantial in Africa, much more attention needs to be given to mobilizing additional resources for primary education. If teacher motivation is to be improved, making investments in teacher training alone will not suffice. Teachers' salaries, which have been sharply eroded, must rise in the medium term—and greater cost-effectiveness alone cannot yield sufficient resources to allow that to occur.

In terms of diversifying the sources of funding of primary education, the greatest potential lies with inter-sectoral and intra-sectoral budget restructuring, targeted taxation, innovative fundraising initiatives by schools with private enterprises, and earmarking ODA and debt cancellations, particularly the redirection of interest repayments to primary education through direct budget support.

The capacity of parents, communities and pupils to further increase their contributions to primary education is limited—though school effectiveness would certainly be improved by the participation of the community in certain aspects of school governance, which in turn may bring forth greater community contribution. In some instances, out-of-pocket costs will have to be cut because they provide strong disincentives for parents to send their children to school. The financing of primary education is a core government responsibility, which is too important to the society as a whole to be left to the variability of parental or local community financing.

Notes

1. For important inputs and valuable discussions, thanks are due to Jan Vandemoortele. Fay Chung, Rosa María Torres, Mary Joy Pigozzi, Anna Obura, Manzoor Ahmed and Enrique Delamonica helped with comments on an earlier draft. The usual disclaimer applies. An earlier version of this paper was presented at the Mid-Decade Review of Progress Towards Education for All in Africa in Yaounde (for Western and Central Africa) and Johannesburg (for Eastern and Southern Africa) in February 1996.

2. It is striking, however, that some African countries (e.g. Tanzania, Malawi, Madagascar, Kenya and Guinea) that pay teachers relatively well compared to the average income per capita pay much less when adjusted for the GNP trend line (Carnoy & Welmond, 1996).

3. In comparison, in the Republic of Korea the teacher-pupil ratio was 1:68 in 1950 and decreased to 1:52 as late as 1975. However, it is critical that in South Korea, primary education was well funded by both the government and parents, and the quality achieved was high despite large class sizes (Mason et al, 1980).

4. The following countries have pupil-teacher ratios in primary education below forty: Angola (32), Benin (35), Botswana (29), Cape Verde (33), Côte d'Ivoire (37), Madagascar (38), Nigeria (39), Sierra Leone (34), Sudan (34), Swaziland (33), Tanzania (36), Uganda (35), and Sao Tome and Principe (35). In thirteen countries the ratio is over 50, and an effort to reduce it is needed (UNESCO, 1995).

5. For an analysis of options on improving teacher motivation and incentives, especially in low-income countries, see Mehrotra, forthcoming.

6. The Minister of Education of Kenya made this proposal at the Johannesburg meeting in February 1996 of African education ministers to review progress towards the Jomtien goal of achieving UPE by the year 2000.

7. In a sample of twenty-seven countries, the ratio of recurrent cost to train one teacher in a teacher-training institute compared to putting the same candidate through general secondary education was 1:7.9 (Lockheed & Verspoor, 1991).

8. For example, the Self-Help Action Plan of Education (SHAPE), conceived as a support system for basic education in Zambia and supported by the Swedish International Development Agency (SIDA), has developed within the framework of an existing ministry and not as an independent entity. Launched in 1987, it has now reached every primary school in the country. Moreover, unlike many other donor-supported projects in many countries, no SHAPE member receives a salary from SIDA or from the SHAPE structure (UNESCO & UNICEF, 1996).

9. Thus in Bhutan, school buildings that were part of World Bank project cost US$826 per school place at 1992 prices; United Nations Capital Development Fund schools were also relatively high cost at US$293 per place. Government-built schools cost much less, and community-built schools even more so (Bray, 1995). Similarly, in low-income Guyana the per pupil cost of an Inter-American Development Bank-assisted school building project was initially US$3,750, and was later brought down to US$1,875 for a primary/secondary school (IDB, 1994).

10. In 1982, the Republic of Korea's education budget could not meet the total costs of its education system. It levied a national tax for five years on liquor, tobacco, interest on dividend income, and the banking and insurance industry. The tax provided 15% of MOE's budget. In Guinea, the sous-prefecture collects a US$4 poll tax from all persons age 15 or older—which finances education and other social expenditures. Indonesia recently introduced a 2% profit tax on big businesses to fund the nation's poverty alleviation programme, including an expanded programme of primary education. Lockheed and Verspoor (1991) discuss some of the pros and cons of earmarking.

References

Adedeji, A.; Green, R.; Janha, A. 1995. Pay, productivity and public service: priorities for recovery in sub-Saharan Africa. New York, UNICEF and UNDP.

Ahmed, M.. 1975. The economics of non-formal education: resources, costs and benefit. New York, Prager.

Anderson, Mary B. 1992. Education for all: what are we waiting for? New York, UNICEF.

Bray, M. 1995. The costs and financing of primary schooling in Bhutan. Thimpu, UNICEF. (Mimeo.)

Bray, M.; with Lillis, K. 1988. Community financing of education: issues and policy implications in less developed countries. Oxford, Pergamon Press.

Carnoy, M.; Welmond, M. 1996. Are teachers paid too much? Stanford, CA, Stanford University. (Mimeo.)

Chung, F. 1993. Educational expansion, cost considerations, and curriculum development in Zimbabwe. In: Conway; Bourque, eds. The politics of women's education. Perspectives from Asia, Africa, and Latin America. Ann Arbor, University of Michigan Press.

Colclough, C.; Lewin, K. 1993. Educating all the children: strategies for primary schooling in the South. Oxford, Clarendon Press.

Crouch, L. 1993. Policy dialogue and reform in the education sector: necessary steps and conditions. Washington, DC, Academy of Educational Development.

Donors to African Education. 1994. A statistical profile of education in sub-Saharan Africa in the 1980s. Paris.

Inter-American Development Bank. Social Agenda Policy Group. 1994. Building consensus for social and economic reconstruction. Washington, DC, IDB. (Report of the IDB Pilot Mission on Socio-Economic Reform in the Cooperative Republic of Guyana.)

International Bureau of Education; UNICEF. 1996. Primary school repetition: a global perspective. Geneva, IBE.

International Labour Organization. 1996. Impact of structural adjustment on the employment and training of teachers. Geneva, ILO.

Jayarajah, C.; Branson, W.; Sen, B. 1996. Social dimensions of adjustment: World Bank experience 1980-93. Washington, DC, World Bank. (A World Bank operations evaluations study.)

Lillis, K.; Ayot, H. 1988. Community financing of education in Kenya. In: Bray with Lillis, op cit.

Lockheed, M.; Verspoor, A. 1991. Improving primary education in developing countries. London, Oxford University Press.

Makau, B.; Coombe, C. 1994. Teacher morale and motivation in sub-Saharan Africa: making practical improvements. Paper prepared for the DAE Working Group on the Teaching Profession, Commonwealth Secretariat, London.

Mason, E.R., et al. 1980. The economic and social modernization of the Republic of Korea. Cambridge, MA, Harvard University Press.

Mehrotra, S. 1994. An analysis of donor assistance to the basic social services. New York, UNICEF. (Mimeo.)

Mehrotra, S. 1997. Health and education policies in high-achieving countries: some lessons. In: Mehrotra, S.; Jolly, R., eds. Development with a human face. Experiences in social achievement and economic growth. Oxford, Clarendon Press.

Mehrotra, S. (Forthcoming.) Teacher motivation, compensation and incentives in low-income countries: the role of the State and the community. (UNICEF Staff Working Paper.)

Mehrotra, S.; Delamonica, E. (Forthcoming.) Private costs and public expenditure on primary education: a comparative analysis of five low-income countries. International journal of educational development (Oxford, U.K.).

Mehrotra, S.; Thet, A.T. 1995. Public expenditure on basic social services. The scope for budget restructuring in selected Asian and African economies. New York, UNICEF. (UNICEF Staff Working Paper no. 14)

Mosley, P. 1996. Globalization, economic policy and poverty in sub-Saharan Africa 1970-95. In: UNCTAD, Globalization and liberalization: the effects of international economic relations on poverty. Geneva, UNCTAD.

Saint, W.S. 1992. Universities in Africa. Strategies for stabilization and revitalization. Washington, DC, World Bank. (World Bank Technical Paper no. 194.)

Schleicher, A.; Siniscalo, M.T.; Postlethwaite, N. 1995. The conditions of public schools: a pilot study in the least developed countries. Report to UNESCO and UNICEF.

Tan, Jee-Peng; Mingat, A. 1992. Education in Asia: a comparative study of cost and financing. Washington, DC, World Bank.

Thomas, C.; Shaw, C. 1992. Issues in the development of multigrade schools. Washington, DC, World Bank. (World Bank Technical Paper no. 172)

Torres, R.M. 1995. ¿Mejorar la Calidad de la Educación Básica? Las Estrategias del Banco Mundial [How to improve the quality of education: World Bank strategies]. New York, UNICEF. (Mimeo.)

Toye, J.; Jackson, C. 1996. Public expenditure policy and poverty reduction: has the World Bank got it right? IDS bulletin (Brighton, U.K.), vol. 27, no. 1, January.

United Nations. 1996. The United Nations system-wide special initiative on Africa. New York, United Nations.

United Nations Development Programme. 1995. Human development report 1994. New York, Oxford University Press.

UNESCO. 1995. Statistical yearbook, 1994. Paris, UNESCO.

UNESCO; UNICEF. 1996. The Self-Help Action Plan for Education (SHAPE) in Zambia: of copper and fire. Paris, UNESCO.

UNICEF. Office of Social Policy and Economic Analysis. 1995. The global cost of universal primary education. New York, UNICEF. (Mimeo.)

World Bank. 1988. Education in sub-Saharan Africa. Policies for adjustment, revitalization, and expansion. Washington, DC, World Bank.

World Bank. 1994. Adjustment in Africa. Reforms, results and the road ahead. Washington, DC, World Bank.

World Bank. 1995a. Strategies and priorities for education: a World Bank review. Washington, DC, World Bank.

World Bank. 1995b. The poor and cost sharing in the social sectors of sub-Saharan Africa. Washington, DC, World Bank.