No Second Population Doubling
The 21st-century demographic story is peak, plateau, and divergence, not another doubling.
One of the easiest ways to get 2100 wrong is to extend 20th-century demographic experience into the rest of the 21st century. The world went from about 2.5 billion people in 1950 to more than 8 billion today, and that expansion shaped modern assumptions about demand, poverty, food, cities, migration, energy, and development. For much of the modern era, more people meant more schools, roads, hospitals, homes, factories, farms, power stations, ports, and transport systems. That history makes it tempting to treat population growth as a durable input to long-range demand projections. It is no longer that simple.
That intuition has deep roots. Thomas Malthus’s 1798 Essay on the Principle of Population argued that unchecked population would grow geometrically while food supply could grow only arithmetically. The result, in his view, was recurring pressure from famine, disease, and poverty unless birth rates fell or mortality rose. Malthus was wrong in the way he is usually remembered as being wrong: agricultural productivity, fertilizer, mechanization, logistics, trade, refrigeration, crop science, sanitation, vaccines, and public health bent the food and survival curves upward in ways he did not foresee. But he was also wrong in the second way that matters for this century. Human beings did not merely increase supply. They also bent the population curve itself.
That is one of the great development successes of the modern world. Fertility fell because more children survived, women and girls gained more education, contraception became more available, urban life changed household economics, women entered paid work in greater numbers, health systems improved, poverty fell in many regions, and parents no longer needed or wanted very large families as a default economic strategy. That did not happen evenly, and it did not happen without coercive mistakes in some countries and political fights in many others. But the broad pattern is real. The population curve bent because hundreds of millions of people, especially women, gained more control over their lives.
That is why the old Malthusian framing is the wrong way to read the current demographic moment. The main story is not an unavoidable population bomb. Nor is it that population decline has suddenly become the central catastrophe facing civilization, a claim that has become fashionable in some finance and technology circles. The more useful story is that demographic transition is working, unevenly but powerfully, and long-range projections now have to take that success seriously.
The conservative claim is already strong: the world is not heading into a second population doubling. The United Nations’ 2024 World Population Prospects has global population rising from about 8.2 billion in 2024 to roughly 10.3 billion in the mid-2080s, then easing slightly to about 10.2 billion by 2100. That is still a very large number of people, and it still implies substantial demand for food, housing, electricity, cooling, transport, health care, and infrastructure. But even the UN’s mainstream institutional projection is not another demographic explosion. It is a high-population plateau.
My working view is sharper than the UN medium case. I expect global population to peak between 2050 and 2070, with the UN serving as the conservative high-population reference rather than my central expectation. The Institute for Health Metrics and Evaluation’s 2020 Lancet work projected a global peak around 9.7 billion in 2064 and a decline to about 8.8 billion by 2100. Earth4All’s People and Planet scenarios are lower again, with one pathway peaking below 9 billion around mid-century and a more ambitious development pathway peaking earlier. These are competing curves, not three versions of the same model. They embed different assumptions about fertility, mortality, education, urbanization, health, development, migration, and policy.
Those curves matter because small changes in fertility assumptions compound into very large differences by 2100. A global average of 1.8 children per woman versus 2.1 is not a rounding error across generations. It changes age structure, school-age cohorts, workforce growth, dependency ratios, migration pressure, pension burdens, and long-range consumption patterns. Demographic projection is not a side note to 2100 strategy. It is one of the denominators.
The UN remains the institutional reference case because it is transparent, widely used, and comparable across countries. It is also cautious, as institutional projections tend to be. That is not carelessness. Caution is part of its role. The UN has to provide a stable common reference for governments, agencies, researchers, and policy systems. But a reference case is not the same thing as the best central case for every strategic purpose. When fertility is falling faster than expected across large parts of the world, the cautious institutional curve can become the high case.
The lower projections deserve attention because they put more weight on the mechanisms that have actually bent fertility downward: education, women’s agency, contraception access, urbanization, child survival, development, and changing expectations around family size. Those mechanisms are visible across rich countries, middle-income countries, and many lower-income countries. Fertility has fallen below replacement across much of Europe, East Asia, North America, Latin America, and parts of Southeast Asia. India, now the world’s largest country, is at or near replacement nationally, with major regional variation. China, once the default symbol of population pressure, is now a case study in rapid aging and sustained low fertility.
China is useful because it breaks the older intuition cleanly. For decades it was treated as shorthand for overpopulation. Now its population is declining, its fertility is far below replacement, and its age structure is moving quickly. South Korea and Japan are more extreme versions of the same broad pattern. Europe is older and slower-moving, but the direction is similar. Large parts of Latin America have moved through the fertility transition faster than many older models expected. Once fertility falls and stays below replacement, it is hard to move back up at scale. Governments have tried cash payments, tax breaks, housing supports, parental leave, and patriotic appeals. The results have generally been modest. Demography is not very responsive to press conferences.
The countercase is real, and it is mostly about Africa and momentum. Sub-Saharan Africa remains young and growing. Nigeria, the Democratic Republic of Congo, Ethiopia, Tanzania, and other countries still have substantial demographic momentum. If fertility declines more slowly than expected, if girls’ education and reproductive health access stall, if child mortality improvements do not translate into smaller desired family sizes as quickly, or if governance and economic development lag, the global peak could arrive later and higher. Migration can also change national population trajectories even when global totals are heading toward peak. Treating early peak as certainty would be as weak as treating the UN medium case as destiny.
That is why my view is a range, not a point forecast. Peak population between 2050 and 2070 is the practical working assumption. It gives weight to continued growth and demographic momentum in younger regions, especially Africa, while also taking seriously the speed and persistence of fertility decline across much of the rest of the world. It treats the UN as the conservative high institutional case, IHME/Lancet as a plausible lower central case, and Earth4All as a useful stress test of what happens if education, development, reproductive autonomy, and policy bend the curve harder. The useful work is to understand what each projection assumes and which evidence would move the range.
The evidence to watch is straightforward. Fertility trends in sub-Saharan Africa matter most for the high-side case. The persistence of ultra-low fertility in East Asia and Europe matters for the low-side case. India matters because of its scale and regional diversity. Urban housing costs, women’s education, contraception access, child mortality, youth unemployment, migration policy, and the fiscal cost of aging all matter because they are the mechanisms through which demographic projections become reality. These variables move slowly, but they are measurable.
There is also a political trap in the demographic discussion. Population can become a lazy proxy for anxiety about migration, poverty, consumption, culture, or national power. That is not useful. The important issue for 2100 analysis is not whether more people are good or bad. It is whether the denominator being used in long-range forecasts is plausible. A world peaking near 9.5 billion in the 2060s and declining by 2100 is materially different from a world continuing toward 10.3 billion in the 2080s and staying near that level. Both are high-population worlds. Neither is a repeat of the 20th century.
There is a trap in the opposite direction as well: treating lower population growth as an excuse for complacency. The world can avoid a second population doubling and still face enormous demands for housing, food, clean water, electricity, health care, education, cooling, infrastructure, adaptation, and better lives. Lower peak population reduces some long-range pressure, but it does not solve climate change, poverty, biodiversity loss, air pollution, or bad governance. It is a success story, not a solution.
This is why I have used population peak as a supporting assumption in my 2100 work, not as a stand-alone forecast. In my cement displacement work, I noted that current demographic projections have peak global population between 2050 and 2070, reducing long-range demand pressure alongside the end of China’s infrastructure boom and the spread of already-working displacement levers. In my aviation demand work, slowing population growth and aging demographics were part of the reason historic demand growth curves did not carry cleanly through 2100. In my maritime decarbonization work, population is one denominator among others, with cargo mix, fuel costs, and the decline of fossil cargo often mattering more than headcount. The demographic assumption matters precisely because it changes the scale of the other questions without answering them by itself.
For investors, policymakers, and infrastructure planners, population should not be a vague growth amplifier inside the model. It should be a tested assumption with a range, drivers, and update signals. A model that uses the UN medium projection as central should say why. A model using a lower peak should say what fertility, education, urbanization, health, and policy assumptions justify it. A model that effectively assumes another century of demographic compounding needs a very strong defense.
The better starting point is simple: there will be no second population doubling, and the global peak is likely between 2050 and 2070. That is not a claim that demand collapses or that the world becomes small. It is a claim that one of the dominant curves of the 20th century has changed shape. That change was not accidental. It was the result of survival, education, health, urbanization, women’s agency, and development bending the curve that Malthus expected to run away from us.
I do not claim to be right. I claim to be less wrong than most. In this case, being less wrong starts by treating the UN as the high institutional case, not the only serious demographic future, and by recognizing that the population curve has already bent because development changed the underlying drivers.
Good 2100 strategy starts with better denominators. Subscribe to TFIE Strategy Briefing for the deeper pathway reviews, workbooks, assumptions, and evidence tests behind the public articles.