Nicholas Eberstadt, Ph.D.

Nicholas Eberstadt holds the Wendt Chair in Political Economy at the American Enterprise Institute. This essay is adapted from a forthcoming chapter in Earth Report 2000, edited by Ronald Bailey.

The year 2000 will conclude a century of utterly unprecedented population change—an abrupt, dramatic departure from previous experience.  Yet it is likely that developments in the 21^st century will unfold in ways that surprise even battle-tested demographers who think they are beyond surprise.

Although exact global numbers can never be known, population counters generally believe that the number of humans on the planet exceeded 6 billion sometime in 1999, and will total about 6.08 billion by the middle of this year. Data for the year 1900 are naturally even less precise, but a reasonable guess for the total population in that year is around 1.65 billion. Thus in a century, world population almost quadrupled in the process of adding almost four and a half billion souls.  No previous population boom comes even close, either in percentage terms or absolute numbers.

By historical yardsticks, moreover, the population explosion continues. The two premier demographic agencies -- US Census Bureau and the UN Population Division -- both project a rate of increase of 1.3 percent this year.  That is over twice the average rate in the 19^th century and something like four times the pace across the entire millennium. The total population is increasing by almost 80 million persons a year-- perhaps ten times the average annual number in the 19^th century.

Yet paradoxically, the trend appears to have reached a monumental turning point.  For as the 21^st century commences, the tempo of population growth is unmistakably in decline. 

As rapid as the current rate of natural increase appears against the backdrop of history, it is nonetheless considerably slower than in the immediate past.  In the early 1960s, for example, the pace exceeded 2.2 percent per year--a percentage point higher than the rate anticipated for the year 2000. The growth in absolute numbers apparently peaked at 85 million a year in the late 1980s, and has since been declining gradually.

The Longevity Explosion

To estimate population trends, it is necessary first to understand the forces that in turn generated the explosive acceleration, the peak and the decline in world population growth during this century.

The population explosion was entirely the result of health improvements and the expansion of life expectancy.  Between 1900 and 2000, life expectancy at birth at least doubled from something like 30 years to 63 years. Indeed, all other things equal, that health explosion would have resulted in an even greater growth of human numbers than has actually been witnessed. Rough calculations suggest that the world's population would be over fifty percent larger today if our century's recorded revolution in longevity had unfolded in the absence of other demographic changes.

Secular Fertility Decline

The world’s population currently totals about six billion rather than nine billion because fertility patterns were also changing over the century.  And of all the diverse changes in fertility trends registered over the past hundred years, the most significant has been "secular fertility decline"--that is to say, sustained and progressive reductions in family size due to deliberate birth control. 

In historic terms, this trend is a very new phenomenon: it apparently had not occurred in any human society until about two centuries ago.  France, where the trend began by the early 19^th century, was the first country to experience the sustained decline. Since that beginning, the decline has spread steadily if unevenly across the planet, embracing an ever rising fraction of the global population and depressing voluntary childbearing in the affected societies to successive record lows. 

A milestone in the process was passed during the era between the World Wars. Fertility rates in industrialized countries during peacetime dropped below the "net replacement" level-- below the level necessary for long-term population stability.

The inter-war fertility dips proved to be temporary, neither sufficiently prolonged nor sufficiently deep to bring on actual population decline.  But in the last quarter-century sub-replacement fertility has come back with a vengeance, slipping to levels previously unimaginable in prosperous societies at peace. Sub-replacement fertility has now been experienced over a generation or more in a growing number of countries, and it has come amazingly close to becoming the norm the world over.  Indeed, almost half of the world's population lives in societies that aren’t having enough babies to sustain their numbers.

In the early 1950s, the planet-wide "total fertility rate" -- the average number of children per woman per lifetime--is thought to have stood at about five.  Next year, the global TFR will likely be below 2.8.  That dramatic reduction in four decades has already curbed the relative and absolute pace of world population growth, in spite of increasing life expectancy. But even more dramatic changes – changes never contemplated by Malthusians – lie ahead.

The past generation demonstrated that it is possible for fertility levels to fall with startling speed, even in low-income societies.  And while we now know that countrywide fertility levels can plunge well below replacement and remain there for decade after decade, we do not yet know how low they can go. But it is entirely possible that, contrary even to quite recent expectations, sub-replacement fertility will soon typify the world as a whole.  Were that to occur, the 21^st century could turn out to be a time in which world population peaked and thereafter diminished.

The population has fallen before—in the 14^th century, for example. Those earlier reductions, however, were the consequence of catastrophes: bubonic plague decimated societies across Asia, Europe and North Africa between 1333 and 1355. The 21^st century population implosion, by contrast, would take place under conditions of steadily improving life expectancy and living standards. 

Even so, such a de-population would confront the world with its own socio-economic challenges —- quite possibly formidable ones.  Though negative population growth is not yet imagined to be a problem, we must now recognize that it is possible that this, rather than Malthusian over-population will emerge as the great demographic issue of the coming century.

Fertility Trends At Century's End

Many analysts do not appreciate just how far low-fertility regimens have already progressed.  To understand the current situation, divide the world into three categories: (1) countries where fertility levels are currently believed to be below replacement (2) countries where fertility is above replacement but rapidly declining (3) countries where fertility levels remain high and seemingly immune to secular fertility declines.

1) Sub-replacement Fertility.Table 1 catalogues the countries in this first category. Since the numbers are extrapolations based on recent history, they may exaggerate the dimensions of fertility decline. On the other hand, the population data for most places listed here is both relatively up-to-date and reasonably accurate. So it is equally the numbers understate the trend. Indeed, this list, based on Census Bureau estimates, is more conservative than the UN Population Division's figures.

In all, 79 countries and territories, with 44 percent of the world’s population, fit the below-replacement category. And the countries themselves are strikingly diverse in geography, culture and level of economic development.

Virtually every advanced industrial democracy is on the list.  In fact, 27 of the Organization for Economic Cooperation and Development’s 29 members have total fertility rates of less than 2.1 —- more or less the level required for long-term population replacement. The two exceptions, by the way, are Mexico and Turkey, countries at the low end of the OECD as measured by income and education.  Within the regular OECD grouping, the highest total fertility rates are the United States (2.07) and Iceland (2.04)--levels just shy of replacement. At the other end, Germany and Spain's current TFRs are just over 1.2 -- and Italy's is even lower. 

Most OECD members are in Western Europe, which had a collective TFR of 1.4 in 1998.  But overall fertility levels appear to be even lower in Eastern Europe--by Census Bureau reckoning, about 1.3.  Bulgaria, in fact, has the lowest fertility level ever witnessed in modern nation not at war, with women averaging only 1.14 births in a lifetime. Were that pattern maintained indefinitely, each new generation would be half the size of the one before.  In all of Europe, only remote Albania and the tiny outposts of Gibraltar and the Faeroe Islands are thought to be above-replacement enclaves--and in those cases, only barely so.

Within the former USSR, fertility has fallen far below replacement since the collapse of the Soviet empire.  While fertility rates in the six former Islamic Soviet republics all appear to be above the net replacement level (from Kazakhstan's projected TFR of 2.1 to Tajikistan's 3.5), the other nine states are far below replacement.  In the Russian Federation, by far the most populous of the former Soviet republics, fertility is a shade over 1.3.  In the next largest, Ukraine, the TFR is just over 1.3--as it is in Belarus and in the three Baltic states. With a projected TFR of 1.88, Moldova would rank as distinctly the most fertile European enclave within the former USSR today.

Secular fertility decline, as already noted, originated in Europe. And today, virtually every population group in the world of European origin has a fertility rate below replacement. But these account for only about a billion of the 2-1/2 billion people living in sub-replacement regions. Below-replacement fertility thus is no longer an exclusively European – or even predominantly European phenomenon.

Much of the Caribbean is now bearing children at below the replacement rate. Island societies as culturally and economically diverse as Barbados, Cuba and Guadeloupe have fertility rates below that of the United States.  A few tiny territories off the coast of Africa and in Micronesia have apparently also slipped into the sub-replacement category.  But the largest concentration of sub-replacement populations is in East Asia.

Japan was the first non-European society to report sub-replacement fertility in peacetime. The rate fell below replacement in the late 1950s and has remained there for the past four decades. The U.S. Census Bureau puts Japan's TFR at 1.46 –- and Japanese authorities say it is even lower.

All four East Asian tigers -- Hong Kong, South Korea, Singapore and Taiwan -- have been below replacement fertility since the early 1980s. South Korea has the highest fertility level of that group, with a TFR estimated at just under 1.8; Hong Kong, with a TFR of 1.36, is lowest.  Thailand's fertility rate also fell below replacement in the last decade; its current estimated TFR is a bit over 1.8. 

By far the largest sub-replacement population, though, is China's: total fertility on the mainland dropped below net replacement in the early 1990s and was thought to be 1.8 for 1998. China's fertility, of course, has been affected by Beijing's two decades-old population control campaign.  It is only reasonable to suppose that the Chinese State Family Planning Commission's strident and muscular intervention in the name of a "One Child Norm" has had an impact.  But it is not possible to say how great that impact has been--or what would happen if that program were abandoned.

2) Above-replacement, but Rapidly Declining, Fertility.  Another large portion of humanity lives in countries where fertility is still above the net replacement level, but is declining rapidly. Just how large a fraction of humanity this grouping encompasses, of course, depends upon one's definition of "rapid." 

According to the UN Population Division's most recent estimates, total fertility rates for "less developed regions" dropped by half between the late 1960s and the late 1990s, from an average of six births per woman per lifetime down to three. Of course, those averages were strongly affected by the trend in China, and China's circumstances were anomalous.  But the Chinese experience should not divert attention from the breadth and scale of fertility declines in other low-income settings.

Table 2 presents UN Population Division estimates of fertility change between the late 1960s and the late 1990s in the fifteen most populous developing countries -- home to three-quarters of the population of the less-developed regions.  During those three decades, fertility levels fell by over half in seven of the 15 -- which is to say, in five countries besides China and Thailand.  Two of those five (Brazil and Mexico) are traditional Catholic societies, while the other three (Bangladesh, Iran and Turkey) are Islamic. 

Note that 12 of the 15 countries in Table 2 experienced fertility declines of two-fifths or more while, in absolute terms, 10 of the 15 registered TFR declines of over three births per woman!  Only one of the fifteen (Ethiopia) has yet to reduce fertility significantly. Note, too, that in spite of the religion-based pro-natal traditions of Mexico and Iran, fertility has dropped by four births per woman in just 30 years.

Table 2 underscores a more general feature of the fertility decline process now underway in low-income countries: there is no single obvious cause. We think of fertility as being inversely associated with living standards, and in a very general way it is. Yet one of the world’s poorest countries, Vietnam, has a fertility level in the same range as the United States of the Vietnam War era. The rate of income growth is an equally problematic indicator of fertility decline. For while Indonesia's dramatic fertility reductions coincided with relatively rapid GDP growth, per capita GDP actually shrank in Iran over those same years--and yet Iran's fertility declines were even steeper than Indonesia's.

We often hear that education, particularly female education, has a big influence on fertility.  But Table 2 shows no clearcut international pattern.  According to the World Bank Nigeria's 1995 adult literacy rate was higher than Egypt's--but its current TFR is also higher, and by almost two births per woman.  Conversely, the World Bank estimates that adult literacy is currently 50 percentage points lower in Bangladesh than in the Philippines, and the gap between male and female literacy rates is closer to 70 percentage points. Yet fertility is lower in Bangladesh than the Philippines, and Bangladeshi fertility declines over the past three decades have been distinctly greater in magnitude.

What about family planning?  It has become a matter of faith that government population control initiatives have been instrumental, if not indispensable, in fertility decline in low-income settings.  And it is true that most of the developing countries reporting major fertility declines in recent decades did have national birth control programs in place -- some of which were overtly coercive. But even though Brazil has never sponsored family planning, its fertility decline in percentage terms has been almost identical to Mexico's, where a strong national program operates. And although Brazil is both poorer and more poorly educated than Mexico, its fertility level is lower as well.

It is all too clear, then, that the standard socioeconomic predictors of fertility differences are of distinctly limited utility. Indeed, one is tempted to conclude that the strongest predictor of a country’s fertility level is the calendar: the later the date, the lower the fertility level.

3) High and "Resistant" Fertility.  The third share of humanity lives in countries where TFRs are very high --say, six or more -- and for whom convincing evidence of fertility decline has yet to be reported.

People often assume that such countries typify the developing regions.  But since fertility levels in the less developed regions as a whole are thought to have fallen by half over the past thirty years, this clearly cannot be the case.  In the entire Asian land mass, TFRs were down by half between the late 1960s and the late 1990s, from about 5.7 to 2.8.  In Latin America and the Caribbean, TFRs have fallen from roughly 5.6 to 2.7. 

The regions where fertility remains highest (and where fertility declines have been most modest) are sub-Saharan Africa and the Islamic expanse to its north and east--more specifically, the Arab Middle East. These areas had a population of about 930 million in 1998 -- a little less than a fifth of the total for less developed regions, and a bit under a sixth of the world total.

Fertility levels are much higher within this agglomeration than in any other comparably sized group of nations. However, the image of uniformly high, persistently "traditional" fertility regimens is already badly outdated.  A revolution in family formation patterns has begun in the region.

Table 3 suggests that it is no longer accurate to describe the "Arab world" as a place isolated from fertility decline.  Over the past thirty years, in fact, fertility in "Northern Africa"--the territory stretching from Western Sahara to Egypt--is believed to have dropped by almost half, to an average of 3.3 births per woman.  There are still outposts of awesomely high and nearly unyielding fertility: the Gaza Strip (with a TFR of 7.3), Oman (5.9), Saudi Arabia (5.8) and Yemen (7.6).  But these four places account for barely 40 million of the Arab Middle East's quarter of a billion people.  Elsewhere relatively rapid fertility decline is underway, and in some Arab countries the decline is rapid.  In Algeria, Libya, and Morocco and Syria, TFRs have dropped by nearly half in just the past two decades.   And there are now Arab countries – notably Tunisia and Lebanon -- where fertility levels are falling close to the net-replacement level.

Even less generally appreciated than the fertility reductions throughout the Arab world is the fact that fertility decline appears to be underway in parts of sub-Saharan Africa too. Particularly noteworthy is the recent Kenyan experience: over the past two decades, the country's TFR fell by two-fifths, to an average of 3.7 births per woman.

Kenya speaks to another aspect of fertility decline in low-income countries worth bearing in mind: the long lag between its onset and the moment demographers can assay its magnitude. Like all sub-Saharan countries, Kenya lacks reliable vital statistics. Its population trends must therefore be divined from occasional censuses and episodic surveys. 

Kenya collects more demographic information than most. Even so, the pace of the country's recent fertility declines has surprised the experts. In 1994, the UN Population Division's very lowest projection for Kenya for the early 1990s was a TFR of 6.2. In fact, the TFR was 5.4--or by almost one birth per woman lower than the lowest contemporaneous UN projections had imagined. By the same token, contemporaneous fertility estimates for some other sub-Saharan locales may be way off base.

What Happens Next?

Barring utter cataclysm, population trends in the next century will be dominated – indeed, determined -- by trends in fertility. That’s just a matter of arithmetic as long as death rates remain low and stable. But what course will world fertility trends take?

There is no way for predicting with much certainty. The UN Population Division's Expert Group Meeting on Below Replacement Fertility summarized the state of knowledge succinctly: "There exists no compelling and quantifiable theory of reproductive behavior in low fertility societies." The same might be said about reproductive behavior in contemporary societies with higher levels of fertility.  As a practical matter, fertility prediction is a matter of educated guesswork.

Some eminent demographers, notably Charles Westoff, have advanced a "homeostatic" hypothesis suggesting that sub-replacement fertility rates will be drawn upward to replacement as if by "a magnetic force." But that is pure speculation.  By the same token, demographers including Antonio Golini have recently suggested that TFRs in stable industrial societies are unlikely to dip below 0.7 -- the level already observed in eastern Germany after unification.  The discomfiting reality is that today's sub-replacement fertility patterns are terra incognita--and that demographers don’t even have much to say about what happens once a population enters it.

The situation is not much more certain for countries still above the replacement rate. The demographer Dudley Kirk has argued that "no country has been modernized without going through the demographic transition [that leads to both low levels of mortality and fertility]." But the definition of "modernization" must now be sufficiently elastic to stretch around cases like Bangladesh and Iran--countries where low average incomes, high incidences of extreme poverty, slow economic growth, mass illiteracy, high levels of mortality and other ostensibly "non-modern" characteristics have proved compatible with massive voluntary reductions in fertility. 

We now know, moreover, that fertility decline can kick in swiftly in low-income settings. As Table 4 attests, there are now examples of countries in which fertility levels have declined by 1.5 births per woman per decade for a full quarter of a century.         In sub-Saharan Africa total fertility has been falling by 20 percent per decade, while in Latin America and the Middle East the pace is around 30 percent per decade. The comparable figure for East Asia is nearly 40 percent, thus far sustained for two and a half decades. 

Forget theory. It is a fact that fertility levels have fallen by three-fifths in just twenty-five years in one Arab country (Tunisia) where upwards of half of all women of childbearing ages have had no formal schooling, and by 45 percent in a sub-Saharan country (Kenya) with a dismal incidence of poverty. For whatever reasons, the constraints against fertility decline appear to be receding remarkably in our own time—and may possibly continue to recede in the coming century.

The UN’s “medium variant” fertility projections envisions a world thirty years hence with an overall TFR of 2.23 -- 1.75 in "more developed regions" and 2.31 in the "less developed regions." That would mean a global population of 7.8 billion in 2025 and 8.9 billion in 2050. 

While the die seems to be cast for the advanced industrial countries, it is easy to imagine factors that significantly raised the pace of population growth above the medium variant elsewhere. For example, a slower than expected fertility decline in the Indian sub-continent. Or a later-than-postulated advent of fertility decline in sub-Saharan Africa. Or the resumption of above-replacement fertility in China. But it is also possible to imagine alternatives that lead to a cessation of population growth and a subsequent depopulation, despite assumed steady improvements in life expectancy.

One Scenario for Global De-population

The UN Population Division's "low variant" population projections, in fact, offer just such a vision. In this scenario, global life expectancies--and life expectancies in every region of the world--rise steadily over the first half of the century.  But population crests in 2040 at 7.47 billion (growing by about 25%, or about a billion and a half people, between now and then), drops by about 120 million between 2040 and 2050, and then continues to decline by over 25 percent per generation.

Of course, assumptions about fertility drive this model.  The "low variant" scenario imagines net-replacement fertility for the world as a whole by the years 2000-2005 —- that is to say, more or less now -- and sub-replacement fertility thereafter.  We must therefore ask whether such presumptions are plausible.

One way of assessing these projections is to compare them to the recent historical past. Table 5 examines the "low variant" model's assumed fertility declines for the three decades between the early 1990s and the early 2020s and compares them with the actually recorded declines over the three previous decades. 

For every area of the world, these "low variant" projections imagine vastly lower fertility levels by the early 2020s than ever before recorded. Some may regard that in itself as a sign of inherent implausibility. But looked at a bit differently, the low variant projections simply suggest a continuation of already existing trends--and in fact, a slight deceleration from the pace of decline between the early 1960s and the early 1990s.

The only region of the world for which the low variant projections imply a faster fertility decline is the African continent. Here, fertility would have to fall as rapidly in absolute terms in the coming decades as it did in Asia from the early 1960s through the early 1990s. That likelihood is, of course, debatable. But in light of the Asian and Latin American experiences, one can hardly maintain that the low variant projections for Africa are out of the realm of possibility.           And the assumptions embodied in the scenario projecting a peaking, and subsequent decline, in global population in the coming century do not seem especially heroic.

Population Issues For A Depopulating World

If the pace of global fertility decline continues for another generation--and the world consequently heads toward negative population growth--the population issues of the future won’t resemble those of the recent past. In a world of long life expectancies, small families and negative population growth, the Malthusian specter will cease to be relevant to public policy.  On the other hand, pervasive and prolonged sub-replacement fertility in the context of generally improving health would likely pose social, political and economic challenges of its own.    Rising life expectancy guarantees that the world's population will be aging in the coming century. If fertility is low and falling, it will age all the more rapidly.  At the moment, the median age of the world's population is about 26 years.  If the trajectory anticipated by the current UN low variant projections were to come to pass, this figure would rise to 44 years --higher than those for even the “elderly" populations like Japan in our contemporary world.

Rapid global aging would have a number of ineluctable implications.  For one thing, it would increase the salience of addressing the health care and income security needs of the elderly.  In Western countries, current public programs for these purposes are coming under increasing demographic pressure, and require far-reaching overhauls to maintain financial soundness. In low-income countries, where coverage by public pension and health systems is limited, the issue of how to take care of the elderly could be all the more pressing.

Rapid global aging would also likely beg the question of how to educate and train the work force of the future. It is not difficult to imagine circumstances in which a majority of a country's workers were over the age of 50.  And such a scenario would exacerbate the already existing tensions between an educational system designed to train the young and the desire of workers to enjoy long, worthwhile careers in an increasingly complex economy.

Finally, prolonged sub-replacement fertility in a world of long life expectancy would presage a radical change in family structure along the lines of one-child China. For the first time in the human experience, there could be societies in which the only biological relatives for many people would be their ancestors.  With sufficiently low fertility for just two generations, people with blood siblings and cousins would become the exception. Exactly how a society would operate under such conditions – how, for example, children would be socialized -- is difficult to imagine. 

All this is merely constrained speculation. But if it does indeed turn out that today’s sub-replacement fertility societies are only a foretaste of what lies in store, such speculation is hardly an indulgence.  Long-term population decline under conditions of steady health improvement is not a disastrous demographic portent—any more than was the Twentieth Century’s dual explosions of population and health. But an orderly global depopulation would require strange new adjustments—some of them wrenching—in both economies and societies, and we are more likely to cope with these changes successfully if we have contemplated them in advance.