In the early 1990s more than 160 million children of primary school age were not attending school. Most had either never attended or had dropped out at an earlier stage. Most were illiterate and the majority were girls. Most of them lived in two developing regions, sub-Saharan Africa and South Asia.

The period 1960-1980 saw unprecedented expansion of enrolments at all levels of education, but after 1980, when recessionary pressures had a major impact on growth, things changed, particularly in sub-Saharan Africa (SSA). Over the decade to 1990 the gross enrolment ratio at primary level in SSA fell from 78 to 68 percent of the age group (see chart, below). By 1995 it had recovered somewhat but still remained less than 1980 levels. The impact of recession, rapid population growth and weak government commitment to the adoption of effective education policies were the main causes of this stagnation. Budgetary cuts and high levels of domestic inflation resulted in reductions in real spending of more than 50 percent over the decade, leading to decline in school quality.

During this period the proportion of girls relative to boys increased from around 77 to 83 percent - but this was caused by male enrolments falling faster than those of girls, rather than by any significant improvements in female enrolments. In South Asia enrolment growth was maintained at a steady pace over the whole period, but here too, girls remain a minority and the gender gap in enrolments is currently greater than in sub-Saharan Africa, and indeed is worse than in any other region.

Although resources available for education have not kept pace with population growth, particularly in sub-Saharan Africa, the proportion of GNP allocated to education spending in these two regions has risen by more than 20 percent since 1985. The macroeconomic difficulties facing countries in these regions have thus had an important impact upon these outcomes.

Nevertheless, it is becoming increasingly clear that even countries which are amongst the poorest in the world, can - given high commitment and careful policy choice - achieve schooling for all with gender equity. There is, of course, no single package of measures which can be applied from place to place. Effective policy responses rely first, on careful national diagnosis of the extent and causes of under-enrolment amongst boys and girls. They also depend upon the design and selection of appropriate policy options and on a commitment to implement the required changes. This often raises tough political challenges, where required policy change is controversial or where changes in resource allocations offend the interests of powerful groups.

Several projects featured in this issue indicate the power and complexity of the constraints that need to be targeted by policy reforms. All show that, whilst the problem of resources is crucial to the supply side of educational provision, the achievement of gender equity in schooling is a much more complex issue, that cannot be solved merely by increasing expenditures.

The studies in Guinea and Ethiopia, for example, indicate that poverty at the levels of the household and of the State is a prime cause of under-enrolment. Yet the persistence of gender inequities is more profoundly related to questions of cultural practice, which may prove intransigent even with a shift to quite rapid income growth.

In Ghana, life histories reveal that it is the interaction between the culture of the home, the school and the society, which causes early drop-out and under-enrolment of girls. In India, discriminatory practice leads to lower labour market returns to women's than to men's education. But even if returns were equal, cultural practice would still lead to a male bias in schooling. This happens because early marriage, the dowry system and traditional obligations for male - but not for female - children, to support parents in old age, provides strong incentives for parents to favour the education of their sons.

Although the problems are deep-seated, solutions are not out of reach. Studies under way in nine African countries under the auspices of the Forum for African Women Educationalists (FAWE) reveal that careful diagnosis of constraints and policy choices can lead to practical programmes of policy reform that promise to win popular support. The studies in Ethiopia and Guinea reported overleaf form part of the first phase of the FAWE work. Typical constraints on the demand side such as direct costs (for fees, textbooks, transport, and uniforms) may often be equally prohibitive for boys and girls. But the opportunity costs of school attendance, caused by the loss of domestic or agricultural labour, are as a rule disproportionately borne by girls. They must be reduced or subsidised if enrolment patterns are to change. Examples of policy measures that could help achieve gender equity in enrolments are:

• more flexibility in scheduling the school year to match it to seasonal farming needs
• more initiatives to enrol children early at age 5-6
• efforts to increase the average age of first marriage
• gender awareness training for teachers and community leaders.