Infant and child mortality as a determinant of fertility the policy implications

The availability of World Fertility Survey data on 25 developing countries has allowed us to estimate the effect of infant and child mortality on the number of births in two ways. First, it is possible to estimate the effect of an infant/child death on the period between births. For women who breastfeed their children, an infant death tends to shorten her period of infertility following a birth and thus, in the absence of contraception, reduce the interval between births. This has the effect of increasing the number of births over her reproductive life. While the results vary by the technique used, our best estimates show that in this way the prevention of one infant death averts between .016 births in Trinidad and .351 births in Lesotho with an average of 23 for neonatal and 17 for post-neonatal deaths. This represents the lower limit of the effect of an infant/child death.
The total effect of infant/child mortality on number of birthsdepends not only on biological effects through breastfeeding, but on behavioral adjustments through the use of contraception. Thus second method of estimating the effect is to measure the number of subsequent births a couple has depending on the survival or death of their first few births. This technique is likely to capture both the biological and behavioral effects of infant/child mortality. Our best estimates show that couples with an infant death have between 138 more births in Sri Lanka and 808 more births in Syria with an average of 484 more births than those who do not experience such a death. These represent the upper limit on the effect of an infant/child death.
These effects are larger on average than those reparted from earlier studies but far smaller than would be necessary for mortality declines to be fully compensated by averted births. There may, however, exist externalities or societal changes operating over the long run in response to community levels of mortality which cannot be estimated with data on individuals and their own mortality and fertility experiences.
In an illustrative analysis presented in this paper, the cost of averting a birth through a family planning program is compared with the cost of averting a birth indirectly by preventing an infant death. In the 16 countries for which estimates could be made, there was one case where mortality reduction was more "cost effective for reducing fertility than was family planning. This was the case of Kenya. If data were available for other high mortality African countries, we may well find the Kenya case replicated. Thus this illustrative analysis suggests that infant mortality may be an important component of a fertility reduction program in countries where mortality is high and few couples are able to have the number of surviving children they desire.