Below-replacement fertility is defined as a combination of fertility and mortality levels that leads to a negative population growth rate, hence a declining population size, in a closed stable population. Equivalent definitions of the term, still with reference to a closed stable population, include: the number of deaths exceeds the number of births; the absolute number of births declines over time; the life-expectancy is below the inverse of the crude birth rate; and the net reproduction rate (NRR) is below one. As the reference to the NRR indicates, replacement is most conveniently interpreted for a single-sex (female) population: replacement then means that one female generation replaces itself in the next generation. This condition is satisfied if 1,000 newborn female babies–that is to say, their survivors–give birth to 1,000 female babies over their lifetime; or, equivalently, if 1,000 women at age 15 have female births over their lifetime in numbers that yield 1,000 women survivors at age 15 in the next generation.
The most common measure of fertility, the period total fertility rate (TFR), refers, however, to a two-sex population: it includes both male and female births. TFR indicates the number of births women would have, on average, if during their reproductive lifetime they experienced the age-specific fertility rates observed in a given period (such as a year) in the absence of mortality. It is evident from the definitions above that the TFR, by itself, does not provide an unambiguous definition of below-replacement fertility. Yet a TFR of 2.1 is often referred to as the replacement level fertility. It is in fact a good approximation to replacement level under conditions of very low mortality. The value 2.1 reflects the fact that the sex ratio at birth (the ratio of male births to female births) is about 1.05 in most human populations. Thus replacement of the population would require that, in the complete absence of mortality, women on average have 2.05 children (i.e., 1,000 women would have to have 1,000 female births and 1,050 male births). The small difference between 2.1 and 2.05 allows for the effect of mortality–a good enough approximation, as noted above, in very low mortality populations. In such populations, then, a TFR below 2.1 is below-replacement fertility. But when mortality is higher, below-replacement fertility may be present even though TFR are levels considerably higher than 2.1. How much higher will depend primarily on the overall level of mortality and, to a lesser but not negligible extent, also on the precise age pattern of mortality and fertility. The following are the replacement levels for TFRs associated with various levels of e゠ (life expectancy at birth)–a good summary index of the overall level of mortality–and an average age of maternity of 29 years. Each of these combinations yields
stable populations with a zero rate of population growth (see Table 1).
TFR values lower than those indicated imply stable populations with negative growth rates. If the shortfall is substantial, the rate of population decline in the stable state will be rapid. For instance, a TFR of 1.3 implies an annual decline of the population size by approximately 1.5 percent in a stable population with very low mortality levels. Such a decline implies a reduction of the birth cohort by 50 percent and a halving of the population size every 45 years. Small TFR differences become increasingly important when fertility is lower: a difference of 0.3 between a total fertility rate of 1.0 and 1.3 is approximately equivalent, in terms of stable population growth rates, to the more than threefold greater difference between a total fertility rate of 3.2 and 4.2 in high fertility contexts (assuming low mortality in both instances).
If a population is not stable but has an age distribution shaped by fluctuating past fertility and/or mortality levels, the above definitions of below-replacement fertility are no longer equivalent. Replacement in the period sense–that is, a zero difference between births and deaths occurring in a given calendar year–depends strongly on the prevailing age structure of the population. A period TFR of 2.1 (assuming low mortality) may then be associated with a surplus of births over deaths if women of childbearing age represent an unusually high share of the population (compared to the share in the stable population implied by the prevailing fertility and mortality) and a surplus of deaths over births if women of childbearing age are underrepresented. Similarly, replacement in a cohort sense depends on the long-term trend in TFR levels. Thus, temporary shortfalls of period fertility from replacement level can be consistent with full cohort replacement if there is sufficient subsequent recovery in period fertility. The common, if casual, use of TFR interprets the eventual average fertility a cohort would experience if subject to the current period fertility pattern during its life-course as not just an indicator of the current level of fertility but as implying also its long-term continuation. Hence familiar statements are often used such as "Italian women in the late 1990s have 1.2 children." Such interpretations of the period TFR may, however, be misleading as the ultimate cohort experiences could turn out to be substantially different. In particular, the characteristic postponement of childbearing in many low fertility countries leads to tempo effects that reduce measures of period fertility below the level that would have been observed in the absence of timing changes. In addition, rapid fertility declines and/or delays imply that the period parity distribution is out of equilibrium, with an over-representation of women at higher parities; this condition further depresses period measures such as the TFR due to compositional effects.
Patterns of Below-Replacement Fertility
Fertility at or above replacement levels has prevailed for most of human history: by definition, it was necessary for human survival. As a long-term average, fertility must have been slightly above replacement. Above, since human numbers were growing; slightly, because average long-term growth was very slow. Nevertheless, below-replacement fertility is not a new phenomenon. Many human populations did become extinct in the remote and not so remote past. In almost every such case the main driving force was a high mortality rate that raised the level of fertility that would have been required for population replacement much above the actually prevailing fertility level. However, as mortality (up to the end of the childbearing years) increasingly came under control, fertility behavior became the key determinant of population growth. Low fertility strategies were initially adopted by some subpopulations, such as the bourgeoisie in nineteenth-century Europe, but the practice gradually spread also to other social classes. Despite this spread, overall fertility levels, reflecting still high mortality risks, remained relatively high. For instance, the lowest national fertility level at the beginning of the twentieth century was in France, which had a TFR of 2.79.
Total fertility levels below two, unambiguously below replacement, became common in many countries of the West in the period between World War I and II. This raised considerable concern about depopulation, even though population growth still continued, as the age distribution, reflecting past demographic conditions, temporarily delayed the onset of a decline in population numbers. After World War II, fertility levels rose significantly, creating the unexpected baby boom. But by the late 1950s in the United States and in the early 1960s in much of Europe, the trend reversed and fertility fell rapidly. By the end of the twentieth century, virtually all developed countries and a few developing countries exhibited below-replacement fertility and fertility levels in an increasing number of other developing countries were approaching a TFR of 2.1. The formerly quite distinct fertility regimes of the developed and the developing worlds have become increasingly similar.
Several features of this situation are particularly striking. First, the spread of below-replacement fertility to formerly high-fertility countries occurred at a remarkably rapid pace: The global convergence of fertility indicators has been quicker than the convergence of many other socioeconomic characteristics. Second, earlier notions that fertility levels may naturally stabilize close to replacement level have proven incorrect. In the early 1990s, for instance, fertility levels in Italy and Spain sank below a TFR of 1.3–an unprecedentedly low level for a national population. At the end of the 1990s there were 14 countries in Southern, Central, and Eastern Europe with TFRs of1.3 or less. Several other countries, such as Germany, Japan, and South Korea, had TFRs not much above1.3. Third, there has been a remarkable divergence in the fertility levels of developed countries. For instance, the TFR in the United States rose from a trough of 1.74 in 1976 to levels slightly above 2.05 in the late 1990s. Similarly, TFR levels in the Netherlands, Denmark, France, and several other countries have recovered and stabilized at between 1.7 and 1.9. This divergence of fertility levels in developed countries has been accompanied by a shift or even a reversal of many formerly observed associations between fertility and other demographic and social behaviors. For instance, the cross-sectional correlations in OECD countries (member countries of the Organisation for Economic Co-operation and Development) of fertility levels with the first marriage rate, the proportion of births outside marriage, and the female labor force participation rate have reversed during the period from 1975 to 1999. At the end of the 1990s, divorce levels seemed no longer to be negatively associated with fertility levels in Europe. Hence, there have been crucial changes in the relationships of fertility to its traditional determinants–such as marriage, divorce, leav ing home, and female labor force participation. A high prevalence of marriage and long-term partnerships is no longer associated with higher fertility in cross-sectional comparisons among European countries.
This reversal in cross-sectional associations between fertility and related behaviors is in part due to the different demographic factors driving fertility change. Initially, the decline toward low fertility has been strongly related to stopping behavior–that is, to a reduction in higher parity births. More recently, the postponement of fertility–particularly the postponement of first births–has emerged as a crucial determinant of differences in fertility levels among developed countries. For instance, during the period 1980 to 1999 the period mean age at first birth increased from 25.0 to 29.0 years in Spain and from 25.7 to 28.7 years in the Netherlands; in the United States it has increased from 22.0 years in 1972 to 24.9 in 2000. This postponement affects fertility levels through two distinct mechanisms. First, the tempo-distortions described earlier contribute to reduced period fertility levels. (These reductions may be absent or substantially muted in cohort fertility.) Second, delays in childbearing also affect parity progression probabilities because women start being at risk of higher parity births only at later ages.
Determinants of Below-Replacement Fertility
In light of these demographic factors leading to low fertility, explanations must differentiate between contexts where the primary response of individuals to changing socioeconomic circumstances is stopping behavior and contexts in which the primary response is postponement of births. In the former case, the key issue is what determines the demand for children and hence the quantum of fertility (the lifetime number of births per woman or its period equivalent calculated for synthetic cohorts). The frameworks used for explaining fertility decline during the demographic transitions are largely applicable in answering that question. For instance, theories of fertility variously relate reductions in the quantum of fertility to increased child costs, a reversal of intergenerational wealth flows, increased levels of education (especially for females), higher opportunity costs of time due to increased opportunities of female labor force participation, fertility-friendly population policies, and other factors affecting incentives for women or couples to have children. In general, it can be said that countries with below-replacement fertility share an institutional and socioeconomic context that favors an overall low quantum of fertility. In addition, there is evidence that the emergence and persistence of low fertility is also due to the diffusion of low fertility norms and value orientations. This explanation has been particularly emphasized in theorizing about the so-called second demographic transition, in which demographic change in developed countries since the 1970s has been closely linked to ideational shifts toward more postmodern, individualistic, and post-materialistic value orientations. As a consequence, both acceptability and practice of cohabitation, out-of-wedlock childbearing, and divorce have increasingly spread through young cohorts along with desires for low fertility; these desires were achieved by the availability of effective contraception.
The factors explaining the emergence of below-replacement fertility over time can differ significantly from country to country. For example, as shown by Kohler and his colleagues in 2002, a rise in childlessness (as calculated in period terms) is not a primary driving force leading to very low fertility levels in Southern, Central, and Eastern European countries. Childlessness, however, does constitute an important factor in Germany and Austria. This suggests that even in situations characterized by fertility well below replacement level, biological, social, and economic incentives are generally strong enough to make most women (or couples) want at least one child, and that high levels of childlessness in some countries are likely due to special institutional factors that favor a polarization of fertility behavior toward either childlessness or relatively high fertility.
The reasons for the postponement of childbearing in many developed countries seem to be twofold. First, several factors make late childbearing a rational response to socioeconomic changes. These factors include increased incentives to invest in higher education and labor market experience, and economic uncertainty that may be particularly acute in early adulthood. Second, social interaction effects are likely to reinforce individuals' desire to delay childbearing in response to socioeconomic changes. These interaction effects are a result of social learning and social influence in the decision processes about the timing of fertility, and can also be caused by feedback in the labor and marriage market that make late fertility individually more rational the later the population age-pattern of fertility is. As a consequence of these interaction effects, a delay of childbearing follows what may be called a postponement transition. This is a behavioral shift that shares many characteristics with the earlier fertility transition in Europe and contemporary developing countries: It occurs across a wide range of socioeconomic conditions; once initiated, it results in a rapid and persistent delay in the timing of childbearing; and it is likely to continue even if the socioeconomic changes that initiated the transition are reversed.
In summary, therefore, the emergence and persistence of below-replacement fertility is related to three distinct transition processes: The (first) demographic transition leading to parity-specific stopping behavior within marriage; the second demographic transition resulting in ideational changes and in the rise of non-marital family forms; and, most recently, the postponement transition toward late childbearing regimes. As a consequence of the still ongoing postponement transition, the extent to which specific socioeconomic and institutional contexts accommodate late childbearing has emerged as an essential determinant of cross-country variation in fertility levels. In particular, the delay of childbearing is usually associated with substantially increased investments in female education and labor market experience prior to parenthood–investments that increase the opportunity costs of childbearing in terms of wages foregone. The extent to which these increased opportunity costs affect the quantum of fertility appears to be strongly influenced by the degree of compatibility between childbearing and female labor force participation. Countries with below replacement fertility show marked differences in this regard and these differences are reflected in the degree to which fertility falls short of replacement level. Countries with low compatibility between female labor force participation and childbearing, such as Italy and Spain, exhibit substantially delayed childbearing and especially large reductions in completed fertility.
The Future of Below-Replacement Fertility
Given the socioeconomic and institutional conditions that favor generally low fertility, it is difficult to foresee any widespread tendency for fertility levels in Europe or other developed countries to return to levels persistently above a TFR of 2.1. Many additional countries are likely to experience below-replacement fertility in the near future, and a TFR of 2.1 does not constitute a natural endpoint to the fertility decline. The feasibility of widespread, safe, and reliable childbearing above age 35 that could counteract some of the effects of late-starting motherhood on total fertility, is at best weakly supported by the medical literature, and there are no signs that the process of postponement of childbearing to later ages will come to a halt in the near future.
There are some mechanisms that could potentially lead to a reversal of below-replacement fertility. The quantum and desired level of fertility could be increased by improvements in the economic situation, especially for young adults, and by social policies that provide increased incentives for having children–for example, improved child-care provision, better access to labor markets for women with children, and increased income transfers to families with children. Homeostatic forces may emerge that increase the quantum of fertility as rapid fertility declines lead to substantially reduced relative cohort sizes. When these small cohorts begin higher education, or begin to enter the labor and housing markets, they are likely to encounter substantially more favorable conditions than their older predecessors in large cohorts experienced, and this could lead to an earlier onset and higher level of fertility. This fertility-enhancing effect of small cohort sizes, first proposed by American economist and demographer Richard Easterlin in the context of the U.S. baby boom may be particularly potent in countries in which fertility has fallen far below replacement level.
Bongaarts, John, and Griffith Feeney. 1998. "On the Quantum and Tempo of Fertility." Population and Development Review 24(2): 271–291.
Brewster, Karin L., and Ronald R. Rindfuss. 2000. "Fertility and Women's Employment in Industrialized Nations." Annual Review of Sociology 26: 271–296.
Caldwell, John C. 1982. Theory of Fertility Decline. New York: Academic Press.
Coale, Ansley J., and Paul Demeny. 1983. Regional Model Life Tables and Stable Populations, 2nd ed. New York: Academic Press.
Easterlin, Richard A. 1980. Birth and Fortune: The Impact of Numbers on Personal Welfare. Chicago: University of Chicago Press.
Foster, Caroline. 2000. "The Limits to Low Fertility: A Biosocial Approach." Population and Development Review 26(2): 209–234.
Kohler, Hans-Peter, Francesco C. Billari, and José Antonio Ortega. 2002. "The Emergence of Lowest-low Fertility in Europe during the 1990s." Population and Development Review. 28(4): 641–680.
Morgan, Philip S., and Rosalind B. King. 2001. "Why Have Children in the 21st Century? Biological Predispositions, Social Coercion, Rational Choice." European Journal of Population 17(1): 3–20.
Teitelbaum, Michael S., and Jay Winter. 1985. The Fear of Population Decline. London: Academic Press.
Vallin, Jacques, and Graziella Caselli. 2001. "Le remplacement de la population." In Démographie: Analyse et synthèse, ed. G. Caselli, J. Vallin and G. Wunsch, pp. 403–420. Paris: PUF–INED.
van de Kaa, Dirk J. 1987. "Europe's Second Demographic Transition." Population Bulletin 42(1):1–59.
Wilson, Chris. 2001. "On the Scale of Global Demographic Convergence 1950–2000." Population and Development Review 27(1): 155–172.
JosÉ Antonio Ortega