The two key international institutions that make long-range population forecasts are the World Bank (Bos et al, 1994) and the Population Division of the UN Department for Economic and Social Information and Policy Analysis (1992 and 1993). The UN makes forecasts biannually. In its 1992 assessment the UN medium forecast was for global population of 10.0 billion in 2050, 11.2 billion in 2100, and 11.5 billion in 2150. The UN provides, however, an extreme range of forecasts, ranging from 7.8 to 12.5 billion in 2050, from 6.0 to 17.6 billion in 2100, and from 4.3 to 28.0 in 2150.
The United Nations has revised recent forecasts downward. Its medium forecast for 2050 was 9.4 billion in 1996, compared to 9.8 billion in 1994. This is a fairly substantial downward revision and reflects new information about declines in fertility rates.
The World Bank projects that world population will reach 8 billion in 2023, 10 billion in 2060, and 11 billion in 2103. This forecast is higher than that made in 1987-88, but lower than those made in intervening years.
The Population Reference Bureau (1996) anticates a global population in 2010 of 6.97 billion people and a population in 2025 of 8.19 billion. It also provides forecasts by countrry. Like many forecasting agencies, it has steadily revised its forecasts downward over recent years. For instance, in 1991 it projected global populations of 7.19 and 8.65 for 2010 and 2025. In contrast, the UN1980 projection for 2025 was 8.19 billion, while its 1990 forecast was 8.5 billion.
The U.S. Bureau of the Census also makes global forecasts (1991).
The keys to all forecasts are their assumptions about the future of fertility and mortality rates, especially the former. Most long-range forecasts of recent years produce global equilibrium totals of 10-12 billion many years after fertility reaches replacement rates. Most forecasters assume (for reasons that are not at all clear) movement to replacement fertility rates over time.
Herwig Birg (1995) has undertaken extensive forecasts of population through 2150, with varying equation forms (hyperbolic, S-shaped or logistic, and linear) and representing fertility decline and variable speeds of decline (the key unknowns). Birg believes the UN projections of an ultimate global population of 10.7 billion in 2150 to be based on overly optimistic assumptions about global fertility decline (to 2.5 in 2020-25 and 2.3 in 2025-30). He is also less optimistic, however, about increases in life expectancy than is the World Bank (which posits 82.3 years global average by 2100) or the UN (which sees life expectancy of 84.7 in 2150). Mortality assumptions make much less difference in forecasts than do fertility assumptions.
Carchetti, Meyer, and Ausubel (1996) use logistic curves to model fertility declines and do not assume that they lead to replacement fertility rates. On the contrary, their analysis suggests that fertility in developed countries well stay well below replacement levels. They find, however, that logistic curves suggest a much more predictable movement towards an apparently biological limit on life expectancy of about 80 for men and 85 for women.