Education and Health

Smaller birth cohorts should reduce the pressure on schools, allowing improvements in education. Across developing countries, expenditures per pupil rise as the proportion of the population that is of school age declines. But education spending in total does not necessarily increase, and enrollment rates do not always rise (Schultz, 1987). The "demographic bonus" from declining fertility, which could be spent to increase enrollment or improve the quality of education, is sometimes spent instead on ineffective educational systems or on other things besides education.

Where the "demographic bonus" does go to improve education, the results can be salutary. For instance, South Korea raised net secondary enrollment from 38 to 84 percent between 1970 and 1990 and more than tripled expenditure per secondary pupil (Ahlburg and Jensen, 1997).¹¹ Demographic factors were not the only ones at work, but for East Asia as a whole they are estimated to have contributed 3 - 4 percent of the rise in enrollment and 10-13 percent of the increased expenditure (Williamson and Higgins, 1997).

¹¹In the 1960-1985 period, high enrollment levels (especially at the primary level but also at the secondary level) accounted for more than twice as much economic growth in Korea as the level of investment (Page et al., 1993).

Lower fertility presumably is a boon not only for the educational system as a whole but also for individual parents in educating their children. With fewer children, parents should have more resources, time, and energy to spend on each child. Empirical studies generally confirm that fewer children means each gets more education, although the effect is neither universal nor necessarily sizable (see, for example, Rosenzweig and Wolpin, 1980; Knodel et al., 1990; Foster and Roy, 1993; Lloyd, 1994). One of the larger effects is for the Dominican Republic: In households that had avoided any unwanted births, 56 percent of children completed primary school, as opposed to only 39 percent of children in households with two or more unwanted births. This effect of "unwantedness" is similar, although weaker, in the Philippines but does not appear in Egypt or Kenya (Montgomery and Lloyd, 1996).¹² As with the demographic bonus at the societal level, the effect of fertility on education at the household level may depend on context. When parents see improved futures for their children with education, consider it an affordable priority, bear at least some costs of sending children to school, and have begun to plan their families with children's education in mind, lower fertility can contribute to more educated children (Lloyd, 1994).

¹²Montgomery et al. (forthcoming) confirm the effect of unwantedness in further analysis but suggest it may be smaller, less than a year of schooling completed on average in the countries where it appears.

Lower fertility also produces healthier children. Closely spaced children, large numbers of children in each family, and children born to younger mothers are all more common before fertility declines, and such children all face higher mortality risks. Even up to the age of five years, the risk of death is greater if the interval since the preceding birth is shorter. The risk of death in the first month of life (the neonatal period) is 35-60 percent higher after a birth interval under 2 years than after an interval of 2-4 years (as Figure 8 shows for three regions). On the other hand, if the interval is longer than four years, the risk of neonatal death is reduced by 10 percent. Deaths in the postneonatal period (one month to one year) and in early childhood (1-4 years) show even more striking effects (Figure 8). Some of these effects of child survival may reflect limits on household resources for nutrition or health care, but physiological factors are also at work.

Figure 8 - Percentage Change in Risk of Death When Preceding Birth Interval Is Shorter or Longer than 24-48 Months

NOTE: Medians for 12 sub-Saharan, 10 Latin America, and 6 Asian - North African countries, excluding cases where the preceding sibling died before the following child's second birthday (Sullivan et al., 1994).

Close spacing interferes with breastfeeding, which has an important role in child nutrition and in protecting the child from infection.

Child survival is also affected by birth order, although this is not evident in sub-Saharan Africa. In several countries in that region, the risk of mortality under age 5 is still as high as one in five, and the risk in the region as a whole is equally high for all children (Figure 9). But in other regions, first, second, and third children clearly benefit from lower mortality, while children beyond the sixth continue to bear substantially greater risk. The effect of the mother's age is similar: Children of older mothers (35 and older) do not seem to be at greater risk in sub-Saharan Africa because every child is at high risk, but elsewhere children of older mothers bear a 10-25 percent greater risk. Even in sub-Saharan Africa, however, children of mothers under 20 have a 20-30 percent higher risk of death than do children of older mothers.

Figure 9 - Under-Five Mortality (per thousand), by Birth Order

The mother's health also benefits from lower fertility. The clearest benefit is a reduced risk of maternal death. Death in childbirth is about 20 times as likely for each birth in developing countries as in industrial countries. Having many successive pregnancies puts a mother at even greater risk. At the total fertility rate for sub-Saharan Africa of 5.6 children, the average woman has a lifetime risk of dying in childbirth of about 1 in 18. If total fertility could be approximately halved, this lifetime risk would also be halved, to 1 in 35.¹³

¹³ Maternal mortality ratios are taken from Tsui et al. (1997, p. 115) and total fertility from World Bank (1997a), adjusted upward by 20 percent to allow for pregnancies terminated without a live birth. An important assumption is that the women whose fertility would fall to bring down average fertility bear the same underlying risk per birth as those whose fertility would not fall. If those women whose fertility would fall bear a lowerrisk per birth perhaps related to higher socioeconomic status, which could produce greater propensity to use contraception - the reduction in average life-time risk would be less. On the other hand, these women might bear higher risk per birth - perhaps because of a greater likelihood their pregnancies are unintended and could lead to potentially unsafe abortions - in which case the reduction in average lifetime risk would be greater.