Volume 28 - Article 39 | Pages 1145–1166
Most recent first
28 March 2014 | Response Letter
What makes a demographic scenario ‘unrealistic’? Response to Le Bourg
Le Bourg makes critical references to the work of many other scholars in his letter, but in this response we focus on the two core bones of contention with ours, namely a possible life expectancy of 120 and a possible African population of 12 billion.
We acknowledge that we could have made further reference in the paper to the ongoing debate over both the recent history of life expectancy gains and, in particular, their possible future trajectories. This is no place for a recounting of the ongoing discussion between Vaupel, Christensen and others who argue for forecasting very long life expectancies as against Olshansky, Carnes and Le Bourg himself who argue for a much tighter ceiling to maximal lifespan increase. However, given the lack of consensus around the issue, and drawing on Oeppen and Vaupel (2002) and their observation of life expectancy gains in developed countries, we felt that a hypothesis of a future of life expectancy at 120 was worth exploring. Furthermore, we also note that nearly one-third of the 60 ‘demographers, geronotologists, and aging researchers’ who were asked in Richel (2010) to give their opinions concerning the average life expectancy for a newborn in the year 2100 suggested 120 years or greater. In this same survey Shripad Tuljapurkar - one of the world’s leading experts on aging – suggested a life expectancy of 130. Other experts who suggested life expectancies in excess of ours include Michael R. Rose (Director of Ecology & Evolutionary Biology, UC Irvine); Claudio Franceschi (former Scientific Director of the Italian National Research Centres on Aging); and Alexandre Sidorenko (former UN Focal Point on Ageing). Besides, what our paper was actually trying to show was that even under what might be termed ‘extreme’ (another subjective term) increases in life expectancy, ultimately the key driver of future total population size will be fertility. As such, even if one does not agree with a life expectancy of 120 at some point in the future, precisely the same message holds.
With agricultural technologies currently used in Europe, Africa would not theoretically have a problem feeding 12 billion. The old IIASA/UNFPA land carrying capacity study showed that Sudan alone with only intermediate agricultural inputs (much less than is used in Europe and North America today) could easily feed more than a billion. Indeed Joel Cohen’s (1996) book "How Many People can the Earth Support" includes many estimates of much higher populations, each of which is dependent on future changes in technology. Despite this, we clearly state on page 1155 that there are a number of feedback mechanisms which would mean that some of these scenarios are highly unlikely to transpire.
To zoom out, though, it seems that the main thrust of the letter is about not publishing things which are deemed to be ‘unrealistic’. This, indeed, is a deeply substantive question and one to which an easy answer cannot be given. If we turn to the literature on scenario building, this speaks about "not impossible" futures which can be very unlikely but cannot be ruled out due to internal inconsistency or obvious (bio-physical) constraints. Such scenarios can be informative and illustrative even though they are not very likely/realistic; especially when they are consistent with already available technologies. If there are plausible stories about future technological trends they can also be included. The title of our paper, Very long range global population scenarios as well as numerous references throughout made clear our intention to study ‘not impossible scenarios’.
What, however, we will concede is that the relationship between demographers and biologists is far from where it should be. Le Bourg’s (2001) ‘Mini-review of the evolutionary theories of aging’ was a critical step in fostering this relationship and remains an important reference piece for demographers. Building on this, though, a deeper understanding of the inter-related processes of population and the environment in both a theoretical and empirical setting is clearly required.
Oeppen, J. & Vaupel, J.W. (2002). Broken limits to life expectancy. Science 296(5570): 1029-1031. doi: 10.1126/science.1069675
Richel, T. (2003). Will human life expectancy quadruple in the next hundred years? Sixty gerontologists say public debate on life extension is necessary. Journal of Anti-Aging Medicine 6(4): 2003
20 March 2014 | Response Letter
Unrealistic demographic scenarios should not be published
Demographers often wonder “what would happen if …”. For instance, they try to forecast the world population under various scenarios, as done by Basten et al. (2013) in a recent article in Demographic Research. Such forecasts are useful, provided they rely on realistic and relevant hypotheses. Basten et al. (2013) wondered what could happen if the global total fertility rate would vary in the 0.75-2.50 range, but they rightly did not consider 3, 3.50, or higher global fertility levels, because there is not any chance to observe such values during a long period.
However, from time to time, articles make predictions that have not any chance to be ever observed, simply because they are totally unrealistic from a biological viewpoint. While it could be understood that an unrealistic scenario is used as the extreme boundary of the forecast, i.e. the scenario that will never be observed, some demographers however do not state that point, leaving the reader with the unpleasant feeling that the authors do not discard extreme scenarios. The present article, written by a biogerontologist, is a plea for discarding unrealistic scenarios from demographic forecasts, because they are a source of confusion.
2. Some examples of unrealistic scenarios
Bourgeois-Pichat (1988) made the hypothesis that, in the future, lifespan without incapacity could be 140 years and women could reproduce up to 100 years. Vallin and Caselli (1997) relied on this hypothesis and published a graph showing, in 2100, women with first and second children at modal ages of 35 and 70 years, the last births being observed beyond 80 years of age. For any biologist, making the hypothesis that most women could have a child at a modal age of 70 years and beyond has simply nothing to do with biology. Forecasting extreme lifespans also simply ignores basic knowledge in biology (for a criticism of extreme longevity forecasts, see Le Bourg, 2012).
Basten et al. (2013) made population projections up to 2300, under various total fertility rates (from 0.75 to 2.50) and life expectancy hypotheses (90, 100, 120 years: but this last hypothesis is irrelevant, see Le Bourg, 2012). In this article, the authors calculated the population size of sub-saharian Africa in 2100 if the total fertility rate would remain around 5 in this region. It would be 12 billion, i.e. more than the usually expected total population in 2100 and the authors stressed that “the results are only presented to 2200 because they would quickly reach impossibly high levels”. My opinion is that even the 2100 level is highly improbable, to the very least, and that there is no ground for a scenario with such a high total fertility rate at a continent level.
Unrealistic forecasts of the global population size have been made in the past. Jacquard (1978) stressed that if the yearly growth rate of the population would remain 1.9%, as it was during the 1970s, the population would be 45 billion in 2100 and 150 billion in 2200. However, such forecasts simply ignore that taking into account the environment and its constraints is mandatory for species to thrive. Bacteria in petri dishes show an exponential growth only before reaching the walls of the dish. Flies can lay hundreds of eggs a week but the planet is not overwhelmed by flies because, on the whole, one fly is replaced by one fly in the next generation, no matter the number of eggs laid. Like bacteria and flies, the size of human populations does not evolve freely and stating that it could increase without limit is a mere ignorance of ecological constraints and of basic population biology.
3. Blatantly irrelevant hypotheses should be avoided
In biology and demography, not all hypotheses are probable to the same extent. Some hypotheses are clearly so unrealistic that there is not any heuristic interest in studying them. For instance, human mean body size is increasing for decades and one could make the hypothesis that it could increase in a linear way with no limit, despite it is not actually the case (Komlos and Lauderdale, 2007). Then, one could wonder, for instance, what would be the consequences of 3-meters tall humans for real estate market, because buildings in the world could no longer be used. However, biological variables show asymptotic limits, as observed for world records in sports for which the progression rate is not linear but exponentially decaying (Berthelot et al., 2008), and there is thus not any ground to make the hypothesis that, one day, humans will be 3-meters tall.
The same could be said for demographic forecasts. There is no ground for forecasts of unlimited population size or lifespan, not to mention childbearing at 80 years of age and beyond. Thus, publishing such forecasts is simply a source of confusion for readers. In 2000, Leridon severely criticised “useless computations” of the United Nations forecasts on replacement migration as a solution to population ageing. For instance, Leridon (2000) emphasised that these computations showed that France should welcome 100 millions immigrants before 2050 to keep constant the 15-64/65+ ratio, i.e. nearly twice its current population. Regarding Korea, the UN report stated that “it would be necessary to have a total of 5.1 billion immigrants from 1995 through 2050” (UN, 2000) to keep this ratio constant. One can agree with Leridon (2000): what is the interest of computations showing that ca half the world population should emigrate to Korea?
Mass-media often publish demographic forecasts, as emphasised by Leridon (2000), but biologists also rely on them, and these are two reasons for avoiding unrealistic forecasts in demography articles. These forecasts are not only useless and meaningless, but they are also harmful.
Basten, S., Lutz, W., Scherbov, S. (2013). Very long range global population scenarios to 2300 and the implications of sustained low fertility. Demographic Research 28: 1145-1166.
Berthelot, G., Thibault, V., Tafflet, M., Escolano, S., El Helou, N., Jouven, X., Hermine, O., Toussaint, J.F. (2008). The Citius End: world records progression announces the completion of a brief ultra-physiological quest. PLoS ONE 3, e1552.
Bourgeois-Pichat, J. (1988). Du XXe au XXIe siècle : l’Europe et sa population après l’an 2000 (From XXth to XXIth century: the population of Europe after 2000) Population 43: 9-44.
Jacquard, A. (1978). Éloge de la différence (Praise the difference). Éditions du Seuil, Paris.
Komlos, J., Lauderdale, J.E. (2007). Underperformance in affluence: the remarkable relative decline in U.S. heights in the second half of the 20th century. Social Science Quarterly 88: 283-305.
Le Bourg, E. (2012). Forecasting continuously increasing life expectancy: what implications ? Ageing Research Reviews 11: 325-328.
Leridon, H. (2000). Vieillissement démographique et migrations : quand les Nations unies veulent remplir le tonneau des Danaïdes... (Ageing and migrations: when the United Nations want to fill in the tub of Danaïdes…). Population et Sociétés 358.
United Nations, Population Division, (2000). Replacement migration: is it a solution to declining and ageing populations?, New York, (ESA/P/WP.160).
Vallin, J., Caselli, G. (1997). Towards a new horizon in demographic trends: the combined effects of 150 years life expectancy and new fertility models. In: Robine, J.M., Vaupel, L.W., Jeune, B., Allard, M. (eds.). Longevity: to the Limits and Beyond. Springer, Berlin: 29-68.
Cited References: 53
Download to Citation Manager