1.7 The education of girls and women

The preceding analysis has drawn attention to several respects in which access to education amongst girls and women is inferior to that of boys and men in many, but not all, developing countries. There is a strong case, now widely recognised, that investing in the education of girls is a critical input for development and has a cluster of interrelating benefits. The general case that girls and women are relatively educationally deprived is easy to demonstrate. Female enrolments lag behind boys in most developing countries.

Table 5: Male and Female Gross Enrolment Ratios by Level of Education 1990 and Adult Literacy (percentages)

	First		Second		Third		Adult
	Level		Level		Level		Literacy
	M	F	M	F	M	F	M	F
Developing Countries	105.5	90.4	50.3	37.5	10.1	6.5	74.9	55.0
Sub Saharan Africa	73.5	59.9	1.2	13.8	2.8	1.0	59.0	36.1
Arab States	92.3	74.2	60.2	44.9	15.6	9.5	64.3	38.0
Latin America	111.4	107.2	55.7	59.6	19.3	18.2	86.4	83.0
Caribbean East Asia	124.6	114.9	58.7	47.7	7.3	4.9	85.7	66.4
Oceania South Asia	100.8	75.1	47.8	28.2	12.1	5.3	59.1	32.2

Source: UNESCO 1991:53 Table 3.2;:26, Table 2.2)

Though disparities in enrolments have been reducing at the first level they remain high in Sub-Saharan Africa, the Arab States, and South Asia. At the second and third levels progress towards more equal enrolment has been slower. Low enrolments are associated with lower GNP per capita and the association is stronger for countries with lower female enrolments (Behrman 1991:20).

There is a strong relationship between the incidence of low GERs at primary level and relative under-enrolment of girls. Countries where female enrolment is low are also countries where overall gross enrolment rates are low (UNESCO 1991:54). The bigger the disparity in enrolments between boys and girls the more likely it is that a smaller proportion of primary school children will be in school (Colclough with Lewin 1 993: Chapter 2).

Drop out is higher for girls than for boys in the majority of African and Asian countries (UNESCO 1991:122 Table 5). Persistence rates to grade four are consistently lower for females in countries in Asia and the Near East (ANE) though the gender gap is less than that associated with enrolments. Enrolment rates for girls have been increasing faster than for boys in the ANE region. However there is some evidence that improvements in persistence have favoured boys rather than girls (Behrman 1991:8)

The proportion of teachers who are female has been increasing though in two regions they remain a minority at the primary level. In Sub Saharan Africa and South Asia females account for about 30% of all teachers at the first and second level. In all other regions half or more of teachers are women at the first level (UNESCO 1991:81).

Adult illiteracy is substantially higher for women than for men. In low human development countries, excluding India, the adult literacy rate has a mean of 59% for males and 34% for females. In the least developed countries the figures are 47 and 24% respectively and in Sub Saharan Africa the same pattern is apparent (56% to 34%). Of 59 low human development countries, fully 54 have male female disparities for literacy of 10% or greater. In only one, Lesotho, does female literacy exceed male (UNDP 1991:129).

Data from the IEA studies show that boys outperform girls in mathematics achievement at all ages in most countries. This is also the case in science subjects. The IEA Second International Science study demonstrates that sex differences greatly favoured boys in the countries with the lowest overall scores, in general and in terms of the performance of both the bottom 20% and the top 20%. Though in some high scoring countries sex differences were minimal (e.g. Hungary), in others (e.g. Japan), boys outperfommed girls consistently at all levels of ability. Typically sex differences in performance are greatest in physics and least in chemistry. Patterns do vary. It is not difficult to find studies where male scores are as much as half a standard deviation greater than for girls in lower secondary science (MSU 1986-1990). Analyses of O-level examination results for Kenya, Zambia and Botswana demonstrate that girls perform less well than boys in almost all subjects, but particularly in the physical sciences. In a small number of countries (e.g. Trinidad and Tobago) there seems some evidence that girls perform consistently better than boys in all types of schools and at all grade levels in science (Kutnick and Jules, 1988).

The reasons why female participation is characteristically lower than that for males have been enumerated in a recent research report (Brock 1991). They include factors, with differing emphases in different countries, that are geographical (the physical access to schools which often favours urban boys and discriminates against rural girls), cultural (the result of marriage practices, the gender stereotyping in economic roles, patriarchal social organisation), health related (preferential treatment of male children), economic (the opportunity costs of schooling girls), religious (disposition inherent in some religious practices), legal (asymmetric legal rights and ability to use the legal protection that is available), political and administrative (interest groups and implementation problems), and educational (especially where female teachers are scarce, and where investment in boys education is favoured).

The case in favour of increasing female participation has many dimensions, most obviously the simple observation that there is no equitable reason for the continued existence of under enrolment compared to boys. There are many other aspects to the case that illustrate that developmental benefits are available and that these probably accrue at a greater marginal rate than for boys where enrolment disparities are large. Early studies of the effects of female education on child health suggest that the impact of women's schooling is about twice that for men (Cochrane, Leslie, O'Hara 1980,1982). A complex picture is provided by later studies which take a wide range of approaches to controlling for possibly significant variables. In these there is no simple consensus about the size of the effects but some agreement about the direction. Mensch, Lentzner and Preston (1985) in a fifteen country study attribute considerable significance to mother's education in relation to child mortality. Rosenzweig and Schultz (1982) explored the determinants of male female survival in India in relation to expected rates of return in the labour market. No significant effects of parental schooling emerge in this study except that the difference in survival rates is less if fathers have matriculated (there is no effect for mothers). Pitt and Rosenzweig (1985) report greater incidence of illness in Indonesian households the higher the head of household's education perhaps reflecting greater propensity to self report. The education of wives has no effect on reported frequency of illness. Boulier and Paqueo (1988) suggest that infant and child mortality in Shri Lanka is only reduced for mothers who have ten years or more of education and there is no effect for less education. There are many other studies (reviewed in Behrman 1991:67) and simple generalisations across countries and without specification of controls are elusive. Such controls include individual endowments (e.g. mother's height), community factors (good health care associated with higher levels of schooling), interactions between mothers and fathers schooling, and response biases in reported morbidity.

The direction of causality is also noted as problematic in some of the studies. There is evidence from Indonesia (Pitt and Rosenzweig (1990) that a one standard deviation increase in morbidity of siblings under four results in a 15% reduction in the number of teenage daughters whose primary activity is attending school. This suggests that raising attendance rates can be achieved by lowering child morbidity. In some countries, (e.g. Pakistan) gender gaps in enrolments, and achievement, seem most closely related to the nature of the public provision of schooling which provides more difficult access for girls (Alderman, Behrman, Ross and Sabot (1991).

Rates of return for females are at least as high as those for males, especially when controls are applied related to female participation in the labour force and participation in waged labour. In a number of studies it appears that rates of return for females are higher than for males at particular levels. Thus in Thailand (Schultz 1991) suggests a rate of return to secondary schooling of 25% compared to 13% for males. Gender gaps in paid employment may also contribute to reductions in the perceived benefits of schooling for girls despite the fact that rates of return for the education of girls may actually be higher for those who get access to the labour market. To the extent that this is true reductions in the gaps in both participation and wage rates should encourage more female schooling for longer. Schultz (1991) argues that increased schooling for women is warranted because it is generally associated with relatively high private rates of return, and relatively high social rates of return with significant externalities which include decreased child mortality and unwanted fertility, and more equitable social policy. The latter is the case because it increases the productivity of a marginalised group that is relatively poor, improves the inter-generational distribution of health and educational opportunities for children, and because it slows the population growth rate. Thus even if basic rates of return are comparable between boys and girls, the effects of these externalities will be to increase adjusted rates of return for females.

Behrman (1991:126) notes an element of caution that enters into Shultz's advocacy of the justification of special public subsidies to female education, but does not contest the basic thrust of the analysis offered. The analysis Behrman offers argues that there is a leap between recognising that there are externalities of the kind Schultz draws attention to and accepting the need for public subsidies. It is argued that it may be that this would be a second best solution - if the problem is contagious illness more effective immunisation may be a better strategy, if there are negative externalities to population growth over grazing, deforestation appropriate pricing of such resources may be more effective. This seems to over look what may be strong interaction effects - effective immunisation programmes depend on the understanding and cooperation of mothers, family size is not simply determined by costs in the present and in the future.

Despite many technical reservations Behrman still concludes that allocation patterns for female education probably do justify a shift in resources (Behrman 1991:129). The analysis offered further notes that though it is sometimes claimed that the gender gap in schooling is less important than it appears because girls receive equivalent but informal educational inputs in the home and/or from other community based or religious organisations This is difficult to demonstrate and there seem several contra-indicators. Available evidence supports the view that in most cases female formal education is associated with increased household production and increased non-market productivity. It also appears unlikely that home based education prepares girls better for the changing demands of developing societies than would a judicious mix of school and home based education. Much educational experience is likely to be unavailable in the home environment - if there are few books in the schools there are likely to be even less in the home; mothers, the main source of educational experience in homes in many cultures are likely to be less educated than fathers (and much less so than teachers) and more likely to be illiterate. Many skills can only be acquired by systematic practice which is not characteristic of much informal learning (numeracy, literacy etc.). If it were true that other forms of education apart from schooling were as effective it is difficult to understand why this should not be equally true for boys, a proposition that is rarely argued in debate of this kind.

King (1990) highlights some successful strategies that have been used in increasing female participation. School location has proved important in Bangladesh, Bhutan and Liberia in attracting and retaining more girls in school. Girls are less likely to travel beyond local catchment areas than boys so local provision is critical. Dormitories for girls have also been introduced where travel distances cannot be reduced. Providing a- secure environment for girls in school has also been influential in reassuring parents of the safety of girl children. Efforts to increase the proportion of female teachers have focused on local training and posting teachers in schools close to teachers' homes. This has boosted the supply of female teachers in parts of Pakistan and Nepal. Financial incentives have been used in Guatemala to encourage continued enrolment and discourage early pregnancy. Bangladesh has also employed female scholarship programmes with a positive impact on enrolments - in the Sharastri Upazali area more than 20,000 girls benefited during the 1 980's and female enrolments reached twice the national average. Community care schemes have been implemented in Colombia to free daughters of caring for siblings and domestic work and enable them to attend school. Improvements in technology can reduce the burden of domestic work that falls on girls and women - such that for example less fire wood needs to be collected. Flexible scheduling of classes can also provide opportunities to attend school outside the hours needed for basic economic activity. Adult education can also be used to increase awareness of the benefits of educating female children and can provide opportunities to those who missed the opportunity to attend normal schools. The removal of gender bias from curriculum materials and courses may also assist.

It can be concluded that on grounds of equitable access to education, and from indicators of participation and educational outcomes there is an overwhelming case for assistance to improve the attendance levels and achievement of girls in many developing countries. The evidence on rates of return suggests that these are generally no less than those for boys and may be greater. The significance of the widely cited externalities - improved nutrition, reduced child mortality, smaller family sizes, improved school; attendance, greater equity, etc. - varies from case to case. However, if any of these externalities are substantial, as it seems that often they are, this adds further weight to what is already a very compelling case.