It is axiomatic that, once death rates in a population have fallen steeply and irreversibly, birth rates must eventually follow. The alternative is rapid population growth that is unsustainable in the long term. However, the characterization of the link between mortality and fertility in classic statements of demographic transition theory has differed. As described in the pioneering 1953 study of the population scientist Frank Notestein, societal modernization was the common cause of declines in both mortality and fertility, though falls in death rates preceded falls in birth rates. In his view, improved survival was a contributory cause of fertility reduction, among many others. By contrast, in 1963, Kingsley Davis the eminent demographer, saw improved survival as the central cause of such decline–both necessary and sufficient. To Davis, societal modernization was largely irrelevant; the effect of improved survival was as strong in subsistence, agrarian societies as in more urban, industrialized settings.
Since these early theoretical contributions, scientists have accumulated a great deal of empirical evidence on the mortality—fertility link at the societal and family level. It has become clear that the relationship is not a simple mechanical one. Though it remains true that mortality decline has always preceded secular declines in fertility, the degree of prior mortality improvement, the absolute level and age pattern of mortality at the onset of fertility decline, and the time lags involved vary widely among societies.
In Europe, improvement in life expectancy typically started in the eighteenth century with the gradual elimination of mortality crises. In the latter half of that century and the early decades of the nineteenth century, gains in life expectancy appear to have been widespread; but this period was followed in some countries by an era of stabilization before further and more pronounced improvements started at the end of the nineteenth century, when the onset of fertility decline also occurred. In much of Europe, declines in childhood mortality preceded falls in infant mortality: Indeed, large improvements in infant survival coincided with fertility decline in many provinces. In some Northern European countries, the crude death rate had fallen to 15 per 1,000 population and infant mortality to about 100 per 1,000 births by the start of fertility transition. Conversely, the crude death rate was about 30 and infant mortality over 250 in much of Eastern Europe and Germany when their fertility transitions started.
In the developing regions, the prior imprint of mortality decline was much more substantial than in Europe. In Asia and Latin America, by 1960, which broadly marks the onset of fertility transition in these regions, mortality at all ages, including infancy and childhood, had been falling sharply for several decades, mainly in response to public health initiatives rather than improved living standards. As a consequence, and also because the pre-transition levels of the crude birth rate were higher than were those in Europe, rates of natural increase were much higher than in Europe at the end of the nineteenth century. In sub-Saharan Africa, large gains in life expectancy occurred a decade or so later, as did the onset of fertility decline. But even in the developing world, societal levels of mortality varied widely at the onset of fertility decline–for instance, ranging from life expectancies of 67 and 60 years in Costa Rica and China to 47 and 45 years in Bolivia and Bangladesh.
These irregularities in the mortality—fertility relationship do not disprove the existence of a strong underlying causal connection between the two. Rather, they may reflect the influence on reproductive behavior of many intervening factors–economic, cultural, and political–that mediate its response to improved survival. Opinion on the causal centrality of the relationship is divided and the difference in emphasis between Notestein and Davis is not yet reconciled. The main reason for this continued uncertainty is the failure to identify strong and convincing linkages between improved survival and fertility.
Possible links between mortality decline and fertility are depicted in Figure 1. Four main links at the family level have been identified. One of these is physiological. Early death of a child necessarily stops breastfeeding; thus lactational protection against conception is lost and the interval to next birth shortened. Moreover, in societies where the custom of prolonged sexual abstinence during lactation is observed, the death of a child may trigger the resumption of intercourse, also contributing to a shortening of the next birth interval. Ample evidence confirms that, in societies characterized by prolonged breastfeeding, reductions in infant mortality will act to lengthen birth intervals and hence reduce the frequency of childbearing. However the effect on fertility of even major reductions in infant mortality in these societies is minor.
Improved survival of children reduces any tendency to replace children who have died and thus represents a second possible pathway of influence. Intentional replacement of children who have died
can only be an important consideration in societies where birth control is widely practiced. But, of course in most of these societies, child mortality is already low and any need to replace dead children is uncommon. In a synthesis of evidence from a variety of transitional and low fertility societies, the demographer Samuel Preston concluded that parents who had lost a child were only 20 to 30 percent more likely to proceed to the next birth than other parents. This effect is also too small to constitute an important part of any overall explanation.
The theory that parents insure against possible future deaths of children by having more children than they would otherwise want is highly plausible. When life expectancy is 30 years (typical of many pre-transitional societies), 4.5 births are required to ensure that two children, on average, survive to the next generation. When life expectancy has risen to 65 years, only a fraction over two births are required to achieve the same outcome. All that is required to imbue this insurance hypothesis with causal force is for the adult generation to perceive that survival has improved and to respond by bearing fewer children. Surprisingly, the actualization of this hypothesis has little empirical support. An individual's perception of mortality change corresponds weakly with actual change, and perceptions correlate poorly with reproductive behavior. Moreover, improved child survival is rarely proffered by parents as a reason for having fewer births.
The last of the four family-level links between mortality decline and fertility concerns rising costs of childbearing. To understand the potential explanatory power of this factor, some characteristics of pre-transitional societies need to be outlined. In such societies, only two children per family survived to adulthood per family on average but there was wide inter-family variability. Parents with many children alleviated the costs of nurture by transferring surplus offspring to those with few or no children: by fostering, adoption, or offering them as servants or apprentices, for instance. Such redistributive mechanisms appear to have been a common feature in most traditional societies.
Consider the effects of a sustained decline in mortality. The number of surviving children per family doubles, ratcheting up expenditures on food, clothing, education, and so on. At the same time opportunities for children to be adopted by the childless decrease. Moreover, parents themselves survive longer, thus delaying inter-generational transfer of assets. In short, improved survival places a strain on families. Emigration represents a possible, but typically short-term, solution. Sooner or later, reproduction itself is modified, perhaps by postponement of marriage but ultimately by birth control within marriage. This characterization is the kernel of Kingsley Davis's theory.
Several features of this proposed link, between higher survival and lower reproduction, need to be stressed. Unlike the insurance mechanism, it does not depend on accurate perceptions of changes in mortality. The pressure on families does not suddenly appear. Rather it increases gradually but relentlessly. The increase in numbers of surviving children is experienced first by relatively privileged urban sectors of society, the very groups most concerned with providing children with sufficient means to maintain the social position of the family.
This theory is consistent with evidence that, even in subsistence economies, children consume more than they produce until at least the early teenage years. It is also consistent with the most commonly cited reason for wishing to reduce childbearing, namely that childrearing costs are too high. The demographer John Cleland claims that the theory provides coherence to an otherwise baffling body of evidence. Improved survival is the most plausible common underlying cause of the fertility declines that have arisen over a relatively short period (1950–2000) in most developing countries despite their wide economic, cultural, and political diversity. This interpretation, however, is not amenable to empirical appraisal and offers a less compelling explanation of the earlier transition in Europe, when fertility fell in the context of more modest mortality improvements.
Identification of the main societal link between mortality and fertility decline is attributable to English economist T. R. Malthus who claimed that population growth, resulting from an imbalance between death and birth rates, would drive down living standards and thereby produce a resurgence of mortality (the positive check) or a lowering of fertility (the preventive check). The role of such homeostatic forces has been a major theme of historical demography and assumed a new relevance in the context of rapid population growth in the latter half of the twentieth century. In the 1960s, a broad–but by no means universal–consensus arose that rapid population growth posed a serious threat to development goals in Asia, Latin America, and Africa. The main policy response took the form of state-sponsored family planning programs. Such policies have facilitated reproductive change in many developing countries (particularly in Asia), played a minor role in others (particularly in Latin America), and of course played no role at all in the European transition.
The fact that government policies, driven by neo-Malthusian concerns, do not constitute a central part of the mortality—fertility relationship, does not necessarily imply that a Malthusian perspective is invalid. Yet the thesis that negative feedback from population growth, at a societal level, can trigger a fertility response has little support. Fertility has declined in buoyant economies (e.g., Republic of Korea) and stagnant ones (e.g., Kenya), in densely populated countries (e.g., Bangladesh) and in ones with abundant under-utilized land (e.g., Colombia). Nor, despite intensive empirical investigation, is there decisive evidence that rapid population growth has had negative effects on improvements in living standards. Whatever threats further population growth in the twenty-first century may bring, the twentieth century has seen unprecedented improvement in living standards in parallel with unprecended population growth, the exact opposite of Malthusian predictions.
Relative, rather than absolute, deprivation resulting from the effect of mortality decline on age structure is a further possible link. The central idea, derived from the economist Richard Easterlin, is that the economic and social fortunes of a cohort tend to be inversely proportional to its size relative to other cohorts. Mortality decline at younger ages results, some 20 years later, in an increase of young adults relative to older generations. Compared with their parents, these young adults are vulnerable to relative deprivation: The labor market becomes saturated leading to stagnation in wage increases and family assets have to be divided among more survivors. In short, the natural desire of the younger generation to maintain or improve upon the living standards of its parents is jeopardized and this pressure stimulates a reduction in childbearing. In a simple statistical test using data from 184 countries, economist Diane Macunovich found that increases in the ratio of males aged 15 to 24 years to males aged 25 to 59 were more strongly predictive of declines in fertility than were declines in infant mortality. Pending further assessment of such linkages, this thesis remains an interesting possibility.
Mortality decline must remain at the center of attempts to understand the fertility transition of the past 120 years. Steep declines in childbearing from over five births to around two births per woman were only possible in the context of vastly improved survival. Beyond this obvious truth, few other generalizations can be stated with confidence. Because fertility decline occurs under widely differing mortality conditions, it is clear that improved survival, while it is probably the underlying cause, is not the sole nor, in the short term, necessarily the dominant influence.
Cleland, John. 2001. "The Effects of Improved Survival on Fertility: a Reassessment." Population and Development Review 27(Supplement):60–92.
Davis, Kingsley. 1963. "The Theory of Change and Response in Modern Demographic History." Population Index 29: 345–366.
Macunovich, Diane H. 2000. "Relative Cohort Size: Source of a Unifying Theory of Global Fertility Transition." Population and Development Review 26(2): 235–261.
Notestein, Frank W. 1953. "Economic Problems of Population Change." In Proceedings of the Eighth International Conference of Agricultural Economists. Oxford: Oxford University Press.
Preston, Samuel H. 1978. "Introduction." The Effects of Infant and Child Mortality on Fertility, ed. Samuel H. Preston. New York: Academic Press.
van de Walle, Francine. 1986. "Infant Mortality and the European Demographic Transition." In The Decline of Fertility in Europe, ed. Ansley J. Coale and Susan Cotts Watkins. Princeton, NJ: Princeton University Press.
John G. Cleland