The mysterious statisticians shaping how we think about fertility

Humans have always tried to glimpse the future. Our methods have improved — from the stars and tarot cards to tremendously complex election forecasts. People love predictions, making them and reading about them. How else to explain the popularity of palm readers and legal gambling markets for sports, politics, and more? At Vox’s Future Perfect section, our annual predictions are some of our most-read stories.

There’s one kind of prediction that does not capture the public’s imagination as often but is nevertheless instrumental in shaping how politicians and other powerful people contemplate humanity’s future. Welcome to the world of population projections.

Estimates from highly respected demographers suggest that the United States of the future might be less populous than that of the past. Those predictions have galvanized the political right, combining a sense of cultural decline with xenophobia to persuade voters that a return to traditional American values — which means more child-rearing — is the solution.

Until recently, the fear had not been too few humans, but too many. A few generations ago, a widespread panic was set off by the dire projections published in the infamous book The Population Bomb, which warned of a wave of famines in the developing world because there would be too many people to feed. Other demographic experts warned around the same time that rapid overpopulation could lead to catastrophic outcomes and threaten humanity’s future. Countries like China and India thought they would have too many people in the future and sought to avert overpopulation — only to overcorrect and create a self-made demographic crisis in which they will have many elderly people and too few young people to care for them or power the economy.

National leaders are now confronting a question that would have been unthinkable a century ago: What if we don’t have enough people?

This dramatic reversal reflects the inherent uncertainties with population projections. These statistical estimates of the future of our species are imperfect but they are necessary. There are small groups of demographers, scattered around the world at some of our most revered institutions, asking some of the biggest questions of all: How many people will be alive 50 years from now? In 100 or, even 200? How many young people will there be, and how many old people will they have to support? And how will that future population be spread out across the planet, as different countries grow or shrink in wildly different ways?

These estimates have evolved from the days when overpopulation fears were rampant. Whereas studies once appeared to show such robust population growth that people feared mass starvation, they now anticipate stagnating or declining populations around the world. Already, more than half of the world’s nations are below the replacement fertility rate of 2.1 children per woman. The University of Washington’s Institute for Health Metrics and Evaluation projects that by 2100, nearly every country will have a fertility rate lower than what would be necessary to sustain the human population over time. Some longer-range forecasts, out to 2200 or 2300, paint a portrait of a dying species.

But the farther in the future you go, the more uncertain population predictions get — those that project centuries from now are little more than a guess. The United Nations’ current spread of projections for the global population in 2100 goes from less than 7 billion people on the low end all the way to more than 14 billion — representing two completely different futures for our species. The median guess is roughly 10 billion.

These sketches of humanity’s future are not carbon copies of each other, either. Population projections from the UN, the US Census Bureau, the University of Washington’s Institute for Health Metrics and Evaluation (IHME), and the International Institute for Applied Systems Analysis (IIASA), to name a few of the most prominent groups, often share a general trajectory but can significantly differ in specifics. When added up over the decades, the divergences are sometimes stark.

Everyone is reaching for answers because, as any good demographer might tell you, we have never seen a country rebound from low fertility rates. The trends have been down, down, down. Japan and South Korea, at the leading edge of this problem, are now in a genuine demographic crisis — the kind that awaits the US and Europe if the projections are to be believed, the kind that is already spurring so much societal angst.

”We’re all headed toward a smaller world. … The entire global economy has to adjust to an unprecedented reality,” Jennifer Sciubba, president and CEO of the Population Reference Bureau, told me. As she explained in a 2023 interview with WBUR: “All of our theories about the good life, our economic theories, our political theories — those were all developed under conditions of population growth and economic growth. … It’s really hard to get a paradigmatic shift and say, ‘What if we try to look at the world in a different way?’”

In other words, these projections aren’t just academic data computations. They matter. So it’s worth understanding how they work.

People have been methodically tracking their numbers for thousands of years. For the past 500 years, humans have been slowly sharpening their talent for projecting the populations of the future.

John Graunt, considered the first modern demographer, used London parish records to estimate population trends with crude statistical methods in the 1600s. A century later, Thomas Robert Malthus, another Englishman, predicted (incorrectly) that food production could not keep pace with exponential population growth, perhaps the first example of what would become a recurring theme of scholars and entire societies being struck by a population panic. In the 1920s, American biologist Raymond Pearl first began to publish population estimates that applied the kind of statistical modeling that is now the foundation of modern demography — with margins of error that can become enormous, given the many variables in play.

Today, the most authoritative projections come from agencies like the United Nations and the US Census Bureau, and from nongovernmental groups including IHME, the IIASA, and the Population Reference Bureau. Over the past six months, I’ve spoken with experts from most of them to better understand these all-important statistical estimates.

The foundation of all population projections is to take some basic data points — the current population, its breakdown by age and sex, fertility rates — and extrapolate them into the future. Right now, the IHME, IIASA and United Nations all project the global population in 2050 will be between 9.6 billion and 9.7 billion — not a lot of variation. Barring an unforeseen disruption, that makes sense, given the current world population of about 8 billion and the current global average fertility rate of 2.3 children per woman, a little bit above the human replacement fertility rate of 2.1.

But by 2100, the organizations’ estimates range from 8.8 billion people (IHME) to more than 10 billion (the UN). The Census Bureau doesn’t do global projections, expects the population to peak at 369.4 million in 2080 (from today’s 340 million), and drop to 365.6 million in 2100. The UN, on the other hand, projects that the US population will exceed 400 million in 2100 — a difference of some 35 million, or almost as many people as now live in California.

Generally speaking, population estimates tend to be fairly accurate in the short term (up to 20 years out), but less so once you reach 50 and 100 years.

The Census Bureau, for example, made some of its most accurate projections about the short-term baby boom of the post-World War II era, when fertility rates briefly rebounded after a Great Depression-era decline. But their longer-range forecasts that expected that boom to continue projected that the US would exceed 300 million people by the year 2000. Instead, the advent of the birth control pill and other cultural and economic trends pushed the fertility rate down — evidence of how factors outside a demographer’s field of study can have enormous unanticipated effects. By the turn of the millennium, the US was home to 281 million people.

It is challenging to predict advancements in medicine or shifts in migration or the outbreak of war or pestilence or an economic depression that might affect how many people are alive in the future and where they will live. Still, we have to try. Population projections are necessary to plan for the future.

Critical government programs such as Social Security — one I think about a lot as a millennial worker — depend on our ability to predict the demographic makeup of the country: How many working-age people will there be to pay into the program? How many elderly people will need benefits? Policymakers are constantly monitoring the program’s solvency date, adjusting it forward or backward by a few years based on projected population trends, and considering new policies — usually either tax hikes or benefit cuts — to try to make the program more sustainable. Right now, Social Security is projected to be solvent for 10 more years — if the projections are right.

Population figures are also a core component of any climate change forecast; the volume of future emissions depends in large part on how many people are alive to consume energy. “If we peeled away all the layers of the onion, what’s at the center? It’s population dynamics,” Sciubba told me.

According to one estimate, carbon emissions would be more than 40 percent lower by 2100 if the world’s population fell to 7 billion than if it continued to grow to 15 billion — roughly the full range of the UN’s global projections.

Here’s one way to visualize the imprecision: Any time you’re looking at a chart of future population projections, there is a wide cone with a bolded line inside of it. We typically pay the most attention to that line, which is the median outcome.

But because they are making probabilistic projections, statisticians would never say “the population in year X will definitely be Y.” So while there is a median outcome in which projectors are most confident, there are extreme, if less likely, scenarios at the edges of the cone. The set of dotted red lines represents a 95 percent confidence interval, meaning that the UN believes there’s a 95 percent chance that world population will fall between 9 billion and 11.4 billion in 2100. The narrower band bordered by the red dashed lines, inside the red dotted lines, shows a smaller range — 9.4 billion to 11 billion — that reflects an 80 percent confidence interval. The more precise the projection, the less confident forecasters are in its accuracy.

Why the uncertainty? Because fertility rates, the foundation of population projections, can be fickle, and even a small shift, only a decimal point or two, could have huge ramifications. Look at the chart above again: The widest ranges — the dashed blue lines — reflect the arbitrary addition or subtraction of 0.5 children to the fertility rate. That would “only” require one out of every two couples to decide to have one additional or one fewer child. It doesn’t seem like a big change at the interpersonal level, but apply it to the entire human race and it makes a massive difference. In recent years, the UN has overestimated its predictions for fertility rates in South Korea, Colombia, and other nations. That could happen again, moving up the timeline for a demographic crisis.

You can technically read projections for the year 2300, which paint a dire picture of humanity’s future. These are fascinating documents, and clearly carefully considered. Yet their authors are transparent about how many assumptions they had to make to come up with a cogent estimate — and any one of those assumptions could lead them ending up horribly wrong when extrapolating over 200 years.

One paper that projected there would be between 2 billion and 26 billion people in the year 2300, for example, made a number of necessary but nonetheless debatable assumptions, namely that global fertility rates would never drop below 1.2 births per woman, given the societal consequences. At such a low fertility rate, the authors write, “the human species would then be on a path to extinction. It seems at least plausible that humanity would act collectively or individually well before that point to avoid such an outcome.”

And yet, in places like Japan and South Korea, where the demographic crisis has already arrived, national leaders have struggled to find effective policies to reverse fertility decline. While these decisions have national and global implications, they are also intensely personal. It’s not clear what the government can do to persuade people to have more children.

Declines in the fertility rates of wealthy countries have driven much of the recent population anxiety discourse — more educated women with higher incomes tend to have fewer babies. These trends might soon impact the rest of the world: Africa’s current birth rates are high, but as the continent continues to develop economically over the coming decades, fertility rates are expected to drop accordingly. But there is a range of possibilities: In 2021, the UN’s median estimate was 3.8 billion people in sub-Saharan Africa by 2100, up from about 1.1 billion today, while IIASA anticipated a billion fewer (2.6 billion) and IHME falls somewhere in between.

Sorting through these different projections for sub-Saharan Africa, one of the most volatile regions for demographers because it has both high fertility rates and the potential for rapid economic development, can illuminate how these experts arrive at their different conclusions. Demographers differ in their assumptions about how quickly fertility rates will decline across the continent: In Angola, for example, the UN expects that each woman will bear more than 2.5 children over their lifetime by 2100, while IHME and IIASA anticipate fewer than 1.75 babies. The latter groups include women’s educational attainment in their projections, and IHME also accounts for other family planning trends, while the UN doesn’t account for those variables. That’s because, while there is certainly evidence that more economic progress leads women to having fewer babies, there is also countervailing research identifying other mitigating factors that may soften the downward trend.

That could lead IHME and IIASA to underestimate Africa’s future population compared to the UN. On the other hand, given UN has overestimated the fertility rates of numerous countries over the past 25 years, there is a case for giving more weight to educational and economic trends.

Migration patterns, meanwhile, are not as important when projecting the global population, but they are crucial to how that population might be distributed. Consider the United States: In a high immigration scenario, supported by permissive immigration policy, the Census Bureau projects the US could reach more than 435 million people by 2100. But in a low-immigration scenario, with more restrictions on people entering the country, the population could fall below 320 million in the same time frame. The fertility rate among Americans is already below the 2.1 replacement level.

Demographers struggle to account for migration, because it’s so heavily shaped by unpredictable developments in politics, and many decide to hold current immigration rates constant instead of trying to predict it. But that could affect their accuracy, because there can be big swings in migration patterns, which could then alter the host country’s fertility rate — see, for example, the influx of migrants from Africa and the Middle East to Europe over the past decade-plus. Likewise, population experts know that climate change could affect the population of the future, but it’s hard to know how, so they tend to leave it out of their estimates.

Population projections feel so relevant now because of the surge in American natalist discourse. American thought leaders look at projections of a plateauing or perhaps even dropping US native-born population and see different crises: one of declining traditional family values, or a failure of the social welfare state to support families, or a looming economic catastrophe, or the risks of a regressive immigration policy. These figures underscore many of the major political debates in the US and the world today.

The question will be what we do about them. As we’ve seen in the past and the present, the anxieties spurred by population projections — whatever their accuracy — can result in terrible social damage or inspire aspirational political agendas.

India and China are the clearest examples of population fears spiraling out of control. In the 1970s, Indian leaders worried that their population would explode so quickly that the country would be unable to feed all of its people — so they instituted a forced sterilization program. Millions of men were subjected to an unwanted and irreversible medical procedure in what is now widely regarded as a serious violation of human rights. And the program was not particularly effective in slowing population growth. Instead, it merely eroded public trust in the government.

China, meanwhile, implemented its one-child policy in 1979 for similar reasons, which faced its own criticisms over coercion and unintended consequences. The country formally ended the rule in 2015, and it is now contending with the reciprocal problem to the one it was trying to prevent: It has too few people of a working age, born during the one-child period, and too many people from older generations to support. Many of its cities are emptying out, and advertisements for cemeteries are oddly commonplace.

For most of human history, this would have been unthinkable. Humans have long had more children than they expected to survive into adulthood because it was economically necessary. We needed as many children as we could get simply to produce enough food to eat. But the economic and medical improvements of modernity created a possibility that had never existed before: We could have too many people, many in the 20th century feared, because fewer kids were dying in childhood.

At this point, it seems safe to say that fears of overpopulation were overblown and we overreacted. Now we’re back to asking the age-old question: How do we get more people?

In the United States, we are seeing diametrically opposed responses to that issue. Conservatives, best represented by Vice President JD Vance, blame the loss of “traditional” family values and religiosity for the dramatic drops in marriage rates over the last 50 years. Those on the left blame the unbearable costs of raising children in the US, our weak welfare state, and assorted world crises like climate change for discouraging people from having kids. Solve those problems, their thinking goes, and you might see future populations rebound.

The stakes are starting to appear existential, threatening the high and rising living standards that much of the world has enjoyed over the last century. If current fertility trends continue, Japan is projected to have only one child total 695 years from now. The country has resorted to desperate attempts to reverse their course, such as clumsily trying to incentivize dating through cash handouts.

Extinction fears can even leak into the practice of population projections. Demographers sometimes make assumptions that fertility rates can’t fall below a certain level.

“It’s almost like a reluctance to predict the future where people just aren’t having kids anymore,” Mark Mather, associate vice president at the Population Reference Bureau, told me. “But based on what’s been happening over the past generation, the way things have been going, that’s exactly what we would predict in 20 or 30 years. That’s a challenge for people trying to make these projections.”

It is, in a way, a reminder that population forecasters are human beings, contemplating their own species’s future. Their projections may tell us the path we’re headed on now — but they cannot tell us how humanity will respond. That will ultimately be up to us.