The surprising global variation in replacement fertility

It is frequently assumed by the general public and also by some population experts that the value of replacement-level fertility is everywhere an average of 2.1 lifetime births per woman. Nothing could be farther from the truth. The global variation in replacement fertility is substantial, ranging by almost 1.4 live births from less than 2.1 to nearly 3.5. This range is due almost entirely to cross-country differences in mortality, concentrated in the less developed world. Policy makers need to be sensitive to own-country replacement rates. Failure to do so could result in fertility levels that are below replacement and lead to long-run population decline. For example, the current replacement total fertility rate for the East Africa region is 2.94. Lowering fertility to 2.10 would, under current mortality conditions, result in a regional birthrate 29 percent below replacement. It is frequently assumed by the general public, the media, and even by some demographers that the value of replacement-level fertility is universally 2.1. For example, in an article about recent dramatic declines in fertility in some developing countries, Crossette (2002: D8) notes, “Today, village women and slum families in some of the poorest countries are beginning to prove [the experts] wrong, as fertility rates drop faster than predicted toward the replacement level – 2.1 children for the average mother”. The 2000 Revision of the United Nations (2001) world population projections refers to “the” replacement level as a total fertility rate of 2.1. Finally, after assessing trends in total fertility rates for 143 developing countries from 1950 to 2000, John Bongaarts (2002: 2) concludes, “It is highly unlikely that developing countries will converge on replacement fertility of 2.1 children per woman as is often assumed in population projections”. The global variation in replacement-level fertility is shown in Figure 1. It is substantial, ranging from a low of 2.05 for Réunion to a high of 3.43 in Sierra Leone. A majority of the world’s countries – and all of the more developed ones – have replacement-level total fertility rates within 0.1 of 2.1. But roughly 80 developing countries display replacement fertility rates higher than 2.2. This range of behaviors shows that there is not just one universal constant for replacement-level total fertility; instead, the values for replacement fertility are highly country-specific. 576 THOMAS J. ESPENSHADE ET AL. F ig ur e 1. D is tr ib ut io n of re pl ac em en tle ve l to ta l fe rt ili ty ra te s by co un tr y, 19 95 -2 00 0 (N = 18 7) . So ur ce : A ut ho rs ’ ca lc ul at io ns fr om da ta in U ni te d N at io ns (2 00 1) . THE SURPRISING GLOBAL VARIATION 577 The numerical values for replacement fertility in Figure 1 are calculated using the approximation in Preston et al. (2001: 115): TFRR ≈ (1 + SRB)/p(AM), (1) where TFRR is the replacement value for the total fertility rate (TFR), SRB is the sex ratio at birth and equal to the ratio of the number of male to female births, and p(AM) is the probability of surviving to the mean age of the fertility schedule. The major determinant of cross-national variations in TFRR is differences in mortality. TFRR will be close to 2.1 when practically all women survive to the mean age of the fertility schedule (usually to age 25 or 30). However, when this survival proportion falls closer to 0.60 – as it does in Afghanistan, Burundi, and Sierra Leone – then replacement total fertility rates above 3.3 are implied. A comparison between current total fertility rates and associated levels of replacement fertility highlights the significance of replacement fertility. In the absence of migration, a continuation of current fertility and mortality conditions leads to long-run population increase if fertility is above replacement and to eventual population decline if fertility is below replacement. The relation between actual and replacement fertility for the world’s less developed countries is shown in Figure 2. Current total fertility rates are shown along the vertical axis, and replacement TFRs are measured horizontally. If replacement fertility were universally 2.1, then countries that are located above (below) the horizontal line corresponding to TFR = 2.1 would have fertility above (below) replacement. The 45-degree line traces out points where the values for current and replacement TFR are equal. Countries that lie above (below) this line have fertility above (below) replacement. A majority of the 143 less developed countries, and especially those that the United Nations (2001) terms the least developed, have fertility rates that are substantially above replacement. This is so whether we include countries that lie above the horizontal line corresponding to TFR = 2.1 or, more appropriately, count those above the 45-degree line. The variation in current total fertility is striking, ranging from just above 1.0 for Hong Kong to 8.0 in Niger. But the cluster of data points for countries with total fertility rates above 4 or 5 also shows that the value for replacement fertility is not always 2.1 and can reach nearly 3.5. The measurement of replacement has obvious consequences for how far current fertility is from replacement. For example, in the case of Botswana in Figure 2, the current total fertility rate (4.35) is 107 percent above 2.1 but only 45 percent greater than its own replacement value of 3.01. The corresponding situation for the world’s more developed countries is shown in Figure 3. For the period 1995–2000, with the exception of Albania, 578 THOMAS J. ESPENSHADE ET AL. Figure 2. Total fertility rate and total fertility rate at replacement for the less and least developed countries, 1995–2000 (N = 143). Source: Authors’ calculations from data in United Nations (2001). fertility is below replacement in all 44 more developed countries. Figure 3 also confirms the observation from Figure 1 that all more developed countries exhibit values for replacement-level fertility that are very close to 2.1. Failure to appreciate the significant global variation in replacement fertility is more than a demographic peccadillo. It can have important policy consequences. Table 1 shows the gap between current fertility and replacement fertility for the world and its major regions. Current levels of total fertility are shown in column 2; column 3 presents values for replacement fertility, calculated using Equation (1). Column 4 shows the deviation of actual from replacement fertility using 2.1 as the standard for replacement. World fertility by this measure is 34 percent above replacement. By contrast, column 5 derives deviations between actual and replacement fertility from own-region replacement values. Seen from this new perspective, world ferTHE SURPRISING GLOBAL VARIATION 579 Figure 3. Total fertility rate and total fertility rate at replacement for the more developed countries, 1995-2000 (N = 44). Source: Authors’ calculations from data in United Nations (2001). tility is 21 percent above replacement – still high by many standards, but not as high as when judged by the conventional, though misleading, rule of thumb of 2.1. European fertility is 33 percent below replacement, regardless of which measure of replacement is used. On the other hand, fertility in Africa is 151 percent above replacement as measured by 2.1 but 95 percent above its own-region replacement level. In many policy circles the implicit goal of population policy in highfertility countries is a reduction to a total fertility rate of 2.1. The data in column 6 indicate that, for most regions of the world, this is a recipe for population decline. On a global scale, a TFR of 2.1 would produce a fertility level 10 percent below replacement. For the more developed regions generally, as well as for North America and Europe individually, total fertility of 2.1 is roughly equivalent to fertility at replacement. But for the least developed regions, and especially for Africa, reducing fertility to 2.1 lifetime 580 THOMAS J. ESPENSHADE ET AL. Table 1. The gap between actual fertility and replacement fertility for the world and major regions, 1995–2000 Region Total Total TFR/2.1 TFR/TFRR 2.1/TFRR fertility fertility at