A Critical Analysis of Standard Demography
by Wolfgang G. Gasser / June 2007


The prospects of the United Nations 2006 revision (RE2006) are quite astonishing. The demographic news of the last years have not only been characterized by population aging but also by always new records in fertility lows. A quote (from GUBHAJ) as an example:

Over the last 10 years, between 1990-1995 and 2000-2005, several countries exhibited rapid fertility declines. Islamic Republic of Iran exhibited a most spectacular decline in TFR [total fertility rate] by more than half, from 4.3 in 1990-1995 to 2.1 in 2000-2005. The TFR of Macao, China, which was already low at 1.6 in 1990-1995, further declined by almost half, reaching to a critically low level of 0.8 in 2000-2005. ... It is also to be noted that in the Republic of Korea, the TFR was further reduced by more than a quarter, from 1.7 to 1.2 during the last 10 years.

Nevertheless, the new 2006 revision predicts an even higher population for 2050 than the previous revisions of 2002 and 2004:

      Variant:           low   medium     high
      2002 revision    7.409    8.919   10.633
      2004 revision    7.680    9.076   10.646
      2006 revision    7.792    9.191   10.858

   Tab. 1: World population in billion in 2050 according to UN revisions

In a publication of the Australian government of 2004 (DUNWIL) we can read:

Of particular concern is that the UN appears to be underestimating the future decline in fertility rates, despite all evidence to the contrary.

A review of past UN population projections shows that there have been systematic biases in their assumptions about future trends in important factors. The most significant errors have been the consistent over-estimation of fertility rates in both developed and developing countries.

Such forecasts are often made to serve particular interests.

A similar criticism has been presented by Lutz, Skirbekk & Testa (LUSKTE):

In 1998, with a rapidly increasing number of countries falling much below the previously assumed magic target level of 2.1 children per woman, the UN (1999) finally abandoned their previous assumption that all countries of the world would converge to 2.1 and that no country that was still above 2.1 would ever fall below 2.1. This magic number is now assumed to be 1.85. All countries that are now already below 1.3 are assumed to recover rather quickly to 1.85; countries that are still above 1.85 are assumed to never fall below that level.

Later in the text they ask [1]:

For what reason do these projections assume such an unusual reversal in the trend? Typically in trend analysis, one would need to come up with a very strong and convincing reason to justify such a deviation from the pervasive trend of the past 50 years of cohort experience. Even more surprisingly, not even the low fertility scenarios produced by theses agencies assume a continuation of the trend of the past decades.

Have such reproaches been justified? And if yes, does the new 2006 revision take into account such criticism?

The most obvious anomalies of the 2006 revision

The revision provides detailed data for 195 countries with a minimum population of 0.1 million for the five-year periods from 1950-1955 to 2045-2050. The relevant scenario is the medium variant, which therefore is the subject of this analysis.

The authors of the 2006 revision have not taken into account much of the criticism cited above, maybe with the exception of the reproach that "all countries that are now already below 1.3 are assumed to recover rather quickly to 1.85". In the new revision, the fertility rates of 35 (out of 195) countries are assumed to recover so slowly that they remain still lower than the "target level" 1.85 in the period 2045-2050, then to range from 1.31 (Macao, 0.84 in 2000-2005) to 1.84 (Spain, 1.29 in 2000-2005) [2]. In order to nevertheless get a higher projection for the 2050 world population, the authors simply reduced the assumed speed of the future fertility decline of the countries still having high fertililty [3]. So the fertility rates of 69 high-fertility countries are assumed not yet to reach the endpoint value of 1.85 by 2045-2050, then to range from 1.86 (Qatar, 2.93 in 2000-2005) to 3.78 (Niger, 7.45 in 2000-2005) [2].

However, in the same way as previous revisions, the 2006 revision is once again based on the apriori rather improbable hypothesis that the current fertility minima are the eternal minima and that in the future no more spectacular declines in fertility will be seen. Low fertility rates are assumed to recover, whereas fertility rates of the countries still having high or transitional fertility are assumed to never fall to such low levels.

From the 195/3 = 65 countries with the lowest projected fertility rates in 2020-2025, not a single one is projected to experience fertility decline with respect to the previous 5-year period 2015-2020. If we look at the 65 countries with the lowest fertility rates in 1970-1975 and 1995-2000, however, we find that as many as 52 respectively 56 countries experienced a fertility decline with respect to the previous 5-year period [4].

The "unusual reversal in the trend" of the current revision becomes especially obvious if we determine the country ranking according to fertility for every 5-year period [5]. In 1950-1955, the number one country is Luxembourg with a fertility of 1.98. The number 63 country of 1960-1965 with a fertility of 5.1, Macao becomes number one in 2000-2005 with a fertility of 0.84. And Macao is projected to remain number one at least until 2045-2050! If we take the 20% * 195 = 39 countries with the lowest fertility rates in 2000-2005, then these same countries are the ones projected to have the lowest rates also in 2045-2050, with one single exception: Luxembourg (having left the top-20% group in 2000-2005) reenters in 2010-2015, whereas Trinidad and Tobago (having entered the top-20% only in 1995-2000) then leaves again [5].

Not only the rankings of the countries with the lowest fertility rates remain strangely unchanged from 2000-2005 on, but all country rankings. Whereas the ranking of each country changed on average by 7.3 positions from 1990-1995 to 1995-2000, this average ranking change is only 1.2 positions from 2010-2015 to 2015-2020 [5].

     Period    1955-1960   1975-1980   1995-2000   2015-2020   2035-2040
     Change       6.8         8.0         7.3         1.2         0.9

   Tab. 2: Average position changes in fertility ranking of the 195 countries
           with respect to the previous period (i.e. 5 years earlier)

The essence of demographic transition

Many representations of demographic transition agree with the following sequence: a fall in death rates (sometimes paralleled by an increase in fertility), a following period of substantial natural increase, a lagged decline of birth rates, and an eventual end of natural increase. Because fertility decline first appeared in the most developed countries of the world, one of its first explanations was the simple hypothesis that fertility is inversely correlated with prosperity. A huge number of other hypotheses followed.

As seen from standard demography, the variety of circumstances under which fertility started to decline all over the world has become more and more bewildering. In order to account for the ever-growing list of counter-examples and nuances, more and more ad-hoc-hypotheses had to be added to the original demographic transition theory. Finally, the following rhetorical question was formulated (Peter McDonald, 2001, see: HUGHED):

The search for a common set of conditions that will prevail in every society as it experiences the onset of fertility decline is necessarily fruitless because the extent of variation in economies, cultures, social and political structures and health conditions of different societies at the onset of fertility decline is clearly vast. Why would we expect that a single explanation could be found to this array of situations?

There is a counterquestion, however: why should a common outcome, the demographic transition occur everywhere unfailingly, without having a common cause? The only reasonable answer would be: a wide range of different conditions which everywhere cause fertility decline appear by chance. Yet this is highly unlikely. So there must be a common cause of the common effect. Astonishingly, this common cause is very obvious: substantial natural increase is enough for fertility to decline. In an analogous manner, high death rates leading to population decline tend to increase fertility rates [6].

Direct-replacement versus generation-replacement fertility

Many demographers had originally assumed that after demographic transition a homeostatic demographic equilibrium should be achieved. Interestingly, such an equilibrium actually has been achieved in more and more countries, regions and population subgroups. This has not been noticed, because demographic equilibrium has not been defined as a direct replacement of deaths by births, but as a fertility rate of slightly more than two children per woman. This 2.1 fertility rate entails generation replacement in low-mortality countries, i.e. 1000 mothers of one generation are replaced by 1000 mothers of the next generation and so on. The effect of this generation-replacement fertility on population size is quite complicated, depending on mean age at birth, life expectancy and age distribution.

Another reasonable definition of replacement fertility is: the fertility which depending on age-distribution and death rate entails a constant population. In order to delimit this concept from the standard generation-replacement fertility, let us call it direct-replacement fertility. The direct-replacement fertility leads to a constant population, by directly replacing the population loss due to deaths by a similar population increase due to births. At present, direct-replacement fertility is lower than generation-replacement fertility all over the world. In the most developed regions of China, such as Beijing, Hong Kong, Macao or Shanghai, direct-replacement fertility has fallen below one child per woman and actual fertility rates have quite significantly converged to direct-replacement fertility levels [7].

Because of population aging and decreasing fertile-age cohorts, direct-replacement fertility is already rising in many countries and eventually will surpass everywhere generation-replacement fertility. In the long run, the average of direct-replacement fertility is generation-replacement fertility. The lower actual fertility falls, the higher will be the rebound of direct-replacement fertility, because eventually the low-birthrate cohorts having reached fertile age will be confronted with more deaths of the baby boomer cohorts.

Saturated populations as the endpoint of demographic transition

In the absence of confounding factors, the endpoint of demographic transition is a state, where a concerned population does not grow further, due to some kind of homeostatic equilibrium between births and deaths. So there is a substantial difference between a (rather) constant population after demographic transition and a more or less constant population before or during demographic transition. After demographic transition, a population already has fully exploited its potential population growth and cannot grow further from within. We can call it a demographically saturated population, i.e. a population with a saturation value of roughly 100%.

The population figures published in the Eurostat yearbook 2006-07 (EURSTA) for Europe as a whole represent an ideal scenario expected by the demographic saturation model:

     Year         1960 1965 1970 1975 1980 1985 1990 1995 2000 2005
     Population    604  634  656  676  692  706  721  728  728  728
     Change              30   22   20   16   14   15    7    0    0
     Saturation %   83   87   90   93   95   97   99  100  100  100

   Tab. 3: Population in million and population change over the 
           previous 5 years in Europe (Source: EURSTA, Table 1.4)

But even if these estimated population figures should correspond to reality, the actual saturation value of the 604 million population of 1960 would be higher than the 83% of Tab. 3, because Europe's population would not have reached 728 million in the absence of immigration. It is true that this extraordinarily precise agreement with the demographic saturation model is partly by chance. Nevertheless, Europe is the continent which always has been leading in the demographic transition. If we focus on fertility (Tab. 6) and assume in accordance with the 2006 revision an endpoint fertility of 1.85, then Europe's 2000-2005 value of 1.41 is completely arbitrary, whereas Thailand seems to be the country most advanced in the demographic transition, because its estimated 2000-2005 fertility of 1.83 is projected to reach the endpoint value of 1.85 already in 2005-2010 [2].

Before the onset of demographic transition, the saturation values of populations are generally much lower than 100%. The more difficult survival is and the higher mortality risks are, the lower are saturation values. Thus the main cause of demographic transition is neither urbanisation, nor prosperity, nor education, nor contraception and abortion, nor family-size policy imposed by authorities, nor the perceived costs and benefits associated with children [8], but increasing demographic saturation caused by decreasing mortality due to

Increasing demographic saturation leads directly to lower fecundability. At least under not too exceptional circumstances, this lower fecundability entails lower fertility, irrespective of other causes (e.g. at the individual choice-decision level), and actual fertility of a fully saturated population cannot significantly exceed direct-replacement fertility. Thus infertility of some couples is an unavoidable outcome, if more children are desired than direct-replacement fertility allows. A substantial number of couples does not seek infertility treatment, despite wishing for years for a first or a further child.

Classification in subpopulations and evolutionary relatedness

A main problem of demography in general consists in classifying the world population into useful subpopulations, normally countries, geographic regions or socio-economic groups, for which reasonable projections can be made. Brazil, for instance, is considered as a unit in the 2006 revision, with a fertility slowly converging to the assumed endpoint of demographic transition, namely the 1.85-fertility. Nevertheless, in 2000, Brazil's fertility ranged from 1.02 in the highest income-group of the south to 6.3 in the lowest income-group of the north (CENBRA, ‘Tabela 15’).

According to the saturation model, the endpoint of demographic transition is a fertility oscillating near direct-replacement fertility, resulting in a rather constant population. In reality, however, the effect of direct-replacement fertility after demographic transition can interfere with other effects. Despite being already saturated, populations of child-oriented countries or groups, having lower prevalence of contraception and abortion, can still increase at the expense of evolutionarily related less child-oriented countries or groups. The reason is simple: evolutionarily related countries or groups can be seen as subpopulations of a unit, having as a whole a maximum potential population. So if something hinders one subpopulation from replacing its deaths by births then another subpopulation can further grow at the expense of the first. The more evolutionarily-related subpopulations are, the easier they can grow at the expense of each other.

Any reproach that the demographic saturation model is based on arbitrarily defined populations can be raised in same way against the hypotheses of standard demography. Also, the concept of direct-replacement fertility as the key figure of a demographic model is apriori no less arbitrary respectively scientific than the concept of generation-replacement fertility. And because direct-replacement fertility as the endpoint of demographic transition entails a constant population size, we can use this endpoint population as a benchmark in order to rate the previous, lower population sizes. This directly leads to the concept of demographic saturation, which is nothing more than the ratio of the population size at a given point in time to the same population after demographic transition.

However, whereas it is not clear how the completely different fertility rates of the above-mentioned Brazilian sub-populations ranging from 1.02 to 6.3 should result together from 2035-2040 on in the endpoint fertility of 1.85, it is obvious that saturation as an endpoint of the whole Brazilian population simply requires saturation of all its subpopulations [9].

The effect of migration on direct-replacement fertility

The most important long-ranging factor confounding the demographic saturation model is migration. Economic immigrants, especially if they come from high-fertility countries tend to have higher fertility than non-migrants. But they affect fertility rates far less than natural increase rates [10]. Because most economic migrants are young, their impact on absolute death figures in both the countries of origin and those of destination is small for many years. The children and grandchildren of migrants, however, are counted in the countries of destination, whereas they are missing in the countries of origin [11].

Let us look at the breakdown of the Soviet Union which triggered substantial migration, not only because of an aggravation of the socio-economic situation but also because inhabitants coming from other republics became foreigners and often felt discriminated against in the new nations. The lingua franca Russian has been replaced more and more by the national languages, and inhabitants may not have been able or willing to learn these languages. In many cases the result has been emigration. But whereas older emigrants have tended to return to the nations where they or their parents had come from, younger persons often have preferred to emigrate to richer countries outside the former Soviet Bloc (Warsaw Pact). But if a substantial part of the migrants to e.g. Russia have been older people seeking a less unsecure life, and young people from Russia have emigrated to richer countries, then this obviously has negative effects on natural increase in Russia, even if the net migration rate is zero or positive. The reason is simple: immigration of older persons increases the number of deaths, whereas emigration of young persons reduces the number of births.

So if we deal with deviations from zero-natural-increase respectively from direct-replacement fertility in order to judge the validity of the demographic saturation model, then the age structure of both the emigrants and the immigrants must be taken into consideration. From the 195 countries (with detailed data in the 2006 revision) having a natural increase above 0.2% in 2000-2005 [12], the countries of Tab. 4 can be considered to have started demographic transition long ago.

     Country  Nat.incr.    Country  Nat.incr.    Country  Nat.incr.
     Malta       0.23%     France      0.36%     Australia   0.59%
     Norway      0.28%     Netherlands 0.37%     New Zealand 0.71%
     Luxembourg  0.31%     Cyprus      0.51%     Ireland     0.76% 
     Canada      0.35%     U.S.A.      0.58%     Iceland     0.83%

   Tab. 4: Countries in an advanced stage of demographic 
   transition with natural increase >0.2% in 2000-2005 

Because these countries are in an advanced stage of demographic transition, within the demographic saturation model, their high natural increase rates lead to the conclusion that either not yet saturated subpopulations subsist, or there has been substantial immigration of rather young persons.

The emergence of natural decrease

According to the 2006 revision, natural population increase (respectively decrease) of 2000-2005 ranged from –0.74% per year in Ukraine to +3.56% in Niger. All countries the natural increase of which in one of the five-year periods once fell below –0.2% per year in the absence of war, belong to the former Soviet Bloc after breakdown of communism [12]. These countries are considered by the 2006 revision as forerunners in the demographic transition insofar as they are the first which seem to confirm a prediction of standard demography, namely that low fertility eventually leads to a substantial population decline by natural decrease. However, such predictions of negative natural increase were not made for these Eastern countries [13] but for countries in the West which were more advanced in the demographic transition.

From the countries not having pertained to the Soviet Bloc, Germany's 1975-1980 value of –0.19% is the biggest natural decrease. Germany's average over the next 25 years from 1980 to 2005 of –0.12% is obviously even closer to the direct-replacement level, and there is no indication of a downward trend [14]. It is true that the non-occuring of such predicted population-declines outside the former Soviet Bloc results rather from higher than projected life expectancies [15], than from higher than projected fertility rates. Nevertheless, natural increase in Europe has remained quite close to replacement for many years now.

     Period        Total      East   North   South    West
     1950-1955     1.07%     1.55%   0.58%   1.06%   0.61%
     1955-1960     1.08%     1.49%   0.60%   1.09%   0.62%
     1960-1965     0.94%     1.08%   0.69%   1.12%   0.69%
     1965-1970     0.70%     0.71%   0.59%   1.01%   0.51%
     1970-1975     0.56%     0.68%   0.36%   0.85%   0.21%
     1975-1980     0.44%     0.63%   0.16%   0.67%   0.08%
     1980-1985     0.36%     0.52%   0.22%   0.42%   0.12%
     1985-1990     0.31%     0.43%   0.26%   0.26%   0.17%
     1990-1995     0.03%    -0.12%   0.20%   0.14%   0.12%
     1995-2000    -0.13%    -0.41%   0.12%   0.04%   0.09%
     2000-2005    -0.14%    -0.48%   0.13%   0.05%   0.08%

   Tab. 5: Natural increase in Europe according to the 2006 revision

The question arises whether Eastern Europe with a continuous fall in natural increase to –0.48% or Western Europe with a rather constant natural increase close to zero since the mid-seventies does exemplify an advanced stage of demographic transition. In order to clear up this question we can look at the corresponding fertility data.

     Period        Total      East   North   South    West
     1950-1955      2.66      2.91    2.32    2.65    2.39
     1955-1960      2.66      2.79    2.52    2.65    2.49
     1960-1965      2.58      2.43    2.73    2.71    2.67
     1965-1970      2.36      2.15    2.47    2.67    2.46
     1970-1975      2.16      2.15    2.08    2.54    1.93
     1975-1980      1.97      2.07    1.81    2.24    1.65
     1980-1985      1.89      2.09    1.81    1.83    1.61
     1985-1990      1.83      2.10    1.84    1.56    1.57
     1990-1995      1.57      1.63    1.81    1.41    1.50
     1995-2000      1.40      1.29    1.68    1.32    1.52
     2000-2005      1.41      1.26    1.69    1.36    1.58

   Tab. 6: Fertility in Europe according to the 2006 revision

Western Europe with a fertility minimum of 1.50 in 1990-1995 is already in the stage of a fertility rebound (in parallel with an increasing direct-replacement fertility), whereas Eastern Europe has been at least until 2000-2005 in the stage of a decreasing fertility. So we may conclude, that the substantial natural decrease in some countries of Eastern Europe is rather an exception caused by migration and other special causes (e.g. high abortion rates) than a new stage in demographic transition.

Tab. 6 is also consistent with this prediction of the demographic saturation model: the faster fertility declines, the lower is its final minimum. The speed of demographic transition primarily depends on the speed of mortality decline, but also on the homogeneity of the corresponding population. In a homogenous population, all subpopulations have the same saturation and the same fertility rates at the same time. In a heterogenous population, fertility of one subpopulation can already increase because of aging, whereas a second subpopulation is at the peak and a third still at the beginning of fertility decline. Because the subpopulations reach their respective fertility lows at different times, fertility of the whole population as an average is always higher than the fertility of the subpopulation with the currently lowest fertility [16].

The biggest anomaly in Tab. 6 is the increase in fertility of Northern Europe from 2.32 in 1950-1955 to 2.73 in 1960-1965. Part of the explanation is the decreasing number of persons in fertile age, especially in the 25-29 age group, because fecundability is inversely correlated to the number of persons in fertile age.

     Year   10-14 15-19 20-24   25-29   30-34 35-39 40-44 45-49    20-34
     1950    5.47  5.31  5.60    5.91    5.36  5.87  5.82  5.33    16.88
     1960    6.86  5.93  5.38    5.20    5.38  5.73  5.17  5.62    15.97
     1970    6.62  6.31  6.82    5.82    5.30  5.11  5.33  5.61    17.94

   Tab. 7: Population in million by 5-year-age-group of Northern Europe

The effect of wars on fecundability and on sex ratio at birth

It seems to be an ancient wisdom that wars do not only increase fertility but also the proportion of male births. The existence of post-war baby booms is so obvious, that a positive effect of wars with high death tolls on fertility is beyond reasonable doubt [17]. In Japan the age-groups born just after World War II are still the age-groups with the biggest numbers. In this case, the high fertility might be explained by the lack of contraceptives in those days, and the following fertility decline by the increasing availability of contraceptives. At least in the case of Timor-Leste, however, higher fecundability is a much more reasonable explanation than unavailability of contraceptives. The fertility rate of Timor-Leste is estimated as high as 7 births per woman during 1995-2005, due to a lengthy conflict which ended in 1999 with a UN-supervised referendum leading to independence.

The higher proportion of male births is not as evident as the increased fertility. Nevertheless, statements like the following can be found:

There can be no reasonable doubt that sex ratios (proportions male at birth) have risen during and just after major wars. This happend to the sex ratios of all the belligerent countries in both World Wars. During the same period, the sex ratios of the non-belligerent countries rose to a lesser extent or remained stable. [JAMESW]

Explanations, ranging from physiology to biological evolution have been brought forward:

Sperm motility or hormone levels, as well many other explanations proposed in the literature, are ‘proximate’ causes: physiological changes that would act to produce biases in the sex ratios. But from what is called an ‘ultimate’ point of view, changes in the sex ratio can be seen as an adaptive equilibrium after a decimation of males after a war. ... The Paraguayan war at the end of the 19th century, for example, destroyed most of the male population and was followed by an spontaneous increase in male births. [BISIOL]

According to the saturation model, higher mortality of men leads to a lower saturation value of the male population with repect to the female, and this lower saturation entails a higher fecundability with a male child. Thus, insofar as wars increase death rates of men at higher levels than death rates of women, increases in the sex ratio at birth are a simple logical consequence.

China's missing girls

In a recent UN publication (UNCHIN) we read:

While the overall life expectancy has been increasing in China from 67 years to 73 years, the disparity of it between male and female has become greater from less then 3 years in the early 1980s to more than 4.6 years in the early 2000s. It can be expected more the case in the future. Because women tend to live longer than men, it will end up with more women than men surviving in the high age.

On the next page, the same UN publication proceeds:

Another issue related with rapid fertility decline is arguably the abnormal sex ratio at birth (SRB) as well as infant mortality. China's sex ratio among new-born babies has risen since the mid-1980s. The 1990 population census reports China's SRB at 111, much above the acceptable normal level observed internationally. The abnormality of SRB reported from census received much concern but also created a great debate about the trueness of the reported abnormality. The debate has lasted for ten years until the release of the results of the 2000 population census, which reports China's SRB up to 117, way above the normal level. The abnormality of SRB has been observed in China for two decades. While China is not the only population with abnormal SRB, it is the population with the most serious abnormal SRB for a longest period in the world.

According to the 2006 revision data, the sex ratio at death (SRD) of China has been continuously increasing from 104 (males per 100 females) in 1975-1980 to 123 in 2000-2005.

      1975-1980:  104.0     1990-1995:  115.1
      1980-1985:  108.8     1995-2000:  120.0
      1985-1990:  111.6     2000-2005:  123.1

   Tab. 8: Sex ratio at death in China according to the 2006 revision

According to the same source, China's SRB was 115 in 2000-2005, thus substantially lower than the SRD of 123. In light of the homeostatic link between deaths and births, even this question arises: why is the proportion of girls at birth higher than the proportion of females at death? The answer lies primarily in the fact that China's population as a whole is not yet fully saturated. For 2000-2005, the crude birth rate (per 1000) is given as 13.6 and the crude death rate as 6.6. Thus only around half (6.6 out of 13.6) of the births are directly influenced by the SR at death whereas the rest (7.0 out of 13.6) depends on the sex ratio of China's population as a whole, which is estimated to have been around 107 since 1950. If we calculate the weighted average, we get a prediction of 6.6/13.6 * 123.1 + 7.0/13.6 * 107 = 114.8 for SRB, coming close to the actual value given by the same data.

There are two other big countries where death rates of men have been substantially higher than death rates of women in recent times: Russia and Ukraine. Life expectancies of 2000-2005 are respectively 58.5 and 62 years for men and 71.8 and 73.4 years for women. The reason why in these two countries with even fully saturated populations no abnormal SRB shows up is obvious: the majority of their populations are women (only around 86 men per 100 women).

Consequently, in order to explain high male to female ratios at birth in China and other regions, it is not necessary to invoke widespread sex-selective abortion and infanticide. It is enough to assume that the sex ratio of a country tends to remain unchanged. Then ‘the missing girls’ are simply the counterpart of the ‘feminization of the elderly population’, and the fact that globally less girls than boys are born, primarily results from the higher life expectancy of women.


Standard demography, focusing on generation-replacement fertility, is completely in the dark [18]. The only model of demographic transition having predicted the appearence of extremely low fertility is demographic saturation. According to this model, the end state of demographic transition is direct-replacement fertility of saturated populations. A saturated population already has exploited its full potential population growth and cannot grow further from within. Low fertility after demographic transition is determined primarily not by voluntary regulation but simply by low fecundability (probability of becoming pregnant) [19].

In the end let us ask how probable it is that the ‘most probable’ medium-variant projection of the 2006 revision, a population of 9.2 billion in 2050, comes true? If we extrapolate from the past to the future instead of assuming from now on a completely new pattern never observed in the past (a slow convergence of the fertility rates of all countries from their current values to 1.85 [2]), then we must conclude that this probability is virtually zero.


[1] It is true that the just cited paragraph does not follow the previously cited paragraph but a discussion of European fertility, where such "a reversal in the trend", from decreasing to rather increasing fertility, actually is on the way.
2] See: excel table (fertility rates; ‘target level’ of 1.85 highlighted)
3] This reduction in the speed of the estimated fertility decline is justified by alleged fertility stalls in several countries. However, the experience of the past shows that fertility stalls are temporary and can be followed by substantial declines.
4] See: excel table (fertility rate changes over five years in percent; negative changes highlighted
5] See: excel table (country rankings according to fertility and ‘ranking_changes’)
6] According to HEUPOC, the relationship between mortality and fertility changes has puzzled demographers for decades, and leading demographers have pointed out that mortality change must have played a central role in fertility declines: "Hirschmann, for instance, suggests that we study fertility declines more generally in terms of how ‘demographic equilibria’ are ‘reestablished after mortality declines’... Cleland more directly calls for a ‘a reinstatement of mortality decline at the center of fertility transition theory’ , arguing that mortality decline is the only common feature of theses fertility changes that have otherwise occured under vastly different conditions."
7] According to RE2006, the 2000-2005 fertility was 0.94 in Hong Kong and 0.84 in Macao. According to UNCHIN, Table 4, the 2000 fertility was around 0.7 both in Shanghai and Beijing. Because birth rates are close to the death rates and equality of birth and death rate is the definition of direct-replacement fertility, actual fertility is close to direct-replacement fertility.
8] Arguments against almost every line of explanation can easily be found in the demographic literature, especially in publications of those advocating other lines of explantion. A few examples: "Paul Demeny of the Population Council points out that the drop to replacement-level fertility in much of the U.S. urban population by the late nineteenth century is not explained by ‘...people's access to some superior contraceptive technology - modern methods were yet to be invented - but was the result of individual motivation to keep fertility low...’" (ABERNE). "The purposiveness of reproductive behavior can also be inferred from the ineffectiveness of authorities in imposing family-size policy. ... Low fertility in the face of official pronatalist policy occurs repeatedly" (ABERNE). "Sudan's fertility rate dropped by 17 percent (to 5.0 births per woman) during the late 1980s. This trend could have many explanations, but prosperity and modern family planning are nonstarters" (ABERNE). "An intriguing - indeed very puzzling - feature of the Ghanaian decline is that modern contraception usage has not kept up pace with the decline in fertility" (AGYEIM). "... Also in Brazil, the worsening conditions of the 1980s coincided with a 50 percent drop in the fertility rate" (ABERNE). "In low fertility countries, ideal family sizes as measured in surveys [are] typically higher than actual fertility rates" (LUSKTE).
9] The mathematics of how the saturation values of subpopulations contribute to the saturation of the whole population is very simple. Let us assume a country with two clearly distinguishable subpopulations, one with 27 million and a saturation value of 90% and another with 15 million and a saturation of 30%. The current population obviously is 27+15 = 42 million. The first subpopulation, having exploited already 90% of its potential maximum (by definition 100%), can only grow from 27 to 27 * (100%/90%) = 30 million. The potential maximum of the smaller subpopulation is 15 * (100%/30%) = 50 million. Thus the potential maximum population of the two subpopulations taken together is 30 + 50 = 80 million. A current population of 42 million in the case of a potential maximum of 80 million results in a saturation of 42/80 = 52.5%.
10] "Analysis of data collected by Census Bureau in 2002 shows that women from the top-10 immigrant-sending countries living in the United States collectively tend to have higher fertility than women in their home countries. As a group, immigrants from these countries have 23 percent more children than women in their home countries" (CAMARO). Analogous data for France can be found in TOULEM. According to HERPIS, in France foreign women were responsible for around 12% of the births in 2005. The fertility rate of the foreigners was 3.3 against 1.8 of the Frenchmen (in 2004). Nevertheless, the fertility rate in France increased only from 1.8 to 1.9 due to the foreigners. In order to show that the fertility rate of France would also be (relatively) high without immigrants, the authors also use the concept ‘immigrated which combines foreigners and immigrants having received French citizenship. The immigrated are responsible for 15% of the births. The impact of the second generation of immigrants, born in France and having French citizenship, is not addressed.
11] This can best be seen if we assume that all young adults start to leave a given region. After some time, there won't be any more births, whereas absolute death figures are not strongly affected by the emigration of the young. (Because of a decreasing population size as the denominator, death rates will increase until reaching 100% in the year, the last remaining inhabitants of the region die.)
12] See: excel table (natural increase; values below –0.2% highlighted)
13] On the contrary, still in 1992 the UN estimated that the population of the former Soviet Union would grow from 289 million in 1990, to 308 million in 2000, and further to 416 million in 2150 (Long-range World Population Projections: Two Centuries of Population Growth, 1950-2150. Sales No. E.92.XIII.3.).
14] A similar situation with many years of natural decrease has emerged in Shanghai. "In 2005, Shanghai had 123,900 new births among the permanent population, a birth rate of 0.7 percent, and 107,000 deaths, a rate of 0.6 percent, ... The city was the first in China to see a 0.12 percent decline in the permanent population in 1994. The decline lasted until 2004 when it saw a zero percent natural growth rate, a 0.14 percent increase from 2003." (source)
15] "In the past decades, population projections were based on the expectation that after the end of demographic transition, life expectancy would reach a certain maximum level... In fact, the trend in the countries with the world's highest life expectancy at any point in time shows an almost perfectly linear trend for more than a century with no sign of leveling off" (LUSKTE). "Maximum life expectancy has risen steadily by more than two years per decade (Oeppen and Vaupel, 2002)" (HUGHED).
16] Different saturation values for different subpopulations can also be the cause of fertility stalls. While the proportion of a subpopulation with a fertilty substantially above the average is increasing, average fertilty can remain constant or even increase despite the fact that fertility rates of both the high-fertility subpopulation and the rest-population decrease.
17] Mortality crises in general lead to post-crisis baby booms. "After falling during the 1958-1961 famine in China, fertility rates rose sharply in 1962-1963, above their pre-famine levels (Ashton et al. 1984)" (HEUPOC).
18] "With respect to actual fertility, for decades population forecasters have assumed that it would not fall below replacement only to find that already more than half of the world's population today is below replacement" (LUSKTE).
19] One reason why fertility declines often have coincided with the availability of contraceptives, is the following: As long as countries or populations are living in an equilibrium of high death and high birth rates, there is no need for contraceptives, but need for protection against violent death, need for better food and water, for better sanitation and health care. Only in the case that an infrastructure has been built up in order to meet the latter needs, population can start to increase steadily. The same infrastructure is then prepared to promote contraception as soon as population growth seems to become a problem.

ABERNE: The Demographic Transition Revisited: Lessons for Foreign Aid and U.S. Immigration Policy; Virginia Abernethy; 1994
AGYEIM: The Fertility Transition in Ghana Revisited; Samuel Agyei-Mensah; University of Ghana, 2005
BISIOL: Sex ratio of births conceived during wartime; Claudio Bisioli; Human Reproduction, Vol. 19, No. 1, 218-219, January 2004
CAMARO: Birth Rates Among Immigrants in America - Comparing Fertility in the U.S. and Home Countries; Steven A. Camarota; October 2005
CENBRA: Censo Demográfico 2000; Instituto Brasileiro de Geografia e Estatística
DUNWIL: Global Population Projections – Is the UN getting it wrong?; Ron Duncan & Chris Wilson; February 2004
EURSTA: Eurostat yearbook 2006-07 – Population
GUBHAJ: Fertility Decline in Asia: Opportunities and Challenges; Bhakta Gubhaju; The Japanese Journal of Population, Vol 5, No. 1 (March 2007)
HERPIS: Deux enfants par femme dans la France de 2006: la faute aux immigrées?; Francois Héran et Gilles Pison; Population & Société; Mars 2007
HEUPOC: Mortality Influences on Fertility: Cambodia During and After the Khmer Rouges; Patrick Heuveline and Bunnak Poch
HUGHED: Rethinking the Demographic Transition; Edward Hugh; Barcelona, January 2006; First draft only
JAMESW: Sex ratios of births conceived during wartime; William H. James; Human Reproduction, Vol. 18, No. 5, 1133-1134, May 2003
LUSKTE: The low fertility trap hypothesis; Wolfgang Lutz, Vegard Skirbekk & Maria Rita Testa
RE2006: World Population Prospects: The 2006 Revision
TOULEM: La fécondité des immigrées: nouvelles données, nouvelle approche; Laurent Toulemon; Population & Société; Avril 2004
UNCHIN: Low Fertility in China: Trends, Policy and Impact; UN, December 2006

© No rights reserved, 2007