Will global population peak sooner than expected?
On forecasting peak population and the population bust
It’s my book’s 1st birthday! And the world is still one of 8 billion and counting. The question we tackle in this newsletter is: Will we soon reach peak population, or do I get to release a 9 Billion and Counting edition in a decade or so? It’s a peek into the peak. Thanks for reading, and if you haven’t done so yet, please consider buying a copy of my latest for you or a friend to support my work.
The Club of Rome, famous for commissioning the 1972 Limits to Growth study, has now commissioned a new study of global population that predicts the population bomb they warned about 50 years ago will never actually detonate. If current trends continue, they expect a peak of 8.8 billion before the middle of the century, then rapid decline.
Several times a week I’m asked at what number and when I think world population will peak so I thought I’d give a glimpse into the logic I use when I answer. My answer depends on (1) how fast I think fertility will fall where it is currently high, and (2) whether I think fertility will rise where it is currently super low. When you answer, you’re making predictions about the same variables. My answer has varied over the years, and that’s a good thing. The quality of our data changes, new political leaders come into office, funds get slashed. It’s important to consider that new information and revise. That’s something Philip Tetlock emphasizes in his approach to forecasting. Let’s use this approach to think about whether fertility will come down in places where it’s high.
To make predictions, Tetlock tells us to go outside-in versus inside-out. Here’s what he means:
He asks us to imagine we’re at a wedding and surmising the newly married couple’s likelihood of divorce. The inside-out prediction would be to think of the vows the couple just made, their professions of love for each other, and the closeness of their bodies on the dance floor. In that moment, the likelihood of divorce looks low, and saying it’s higher would be in poor taste.
The outside-in prediction would first consider the baseline divorce rates for the couple’s socioeconomic group, then move to inside factors, like how long they’ve known each other, how often and how well they fight, etc. You’d move your prediction of the likelihood of divorce higher or lower from there.
So, how could we use these lessons to think about peak population? I start with the scenarios offered by the various organizations doing demographic projections. The UN predicts world population to peak at 10.4 billion in the 2080s. The Institute for Health Metrics and Evaluation (IMHE) predicts a peak of 9.7 billion by 2064.
Those predictions fall within a range of projections made by each organization. From those projections, organizations generally choose a few scenarios to illustrate how tweaking underlying assumptions about fertility and mortality affect the ultimate outcome. The UN actually offers 9 different scenarios (or variants) and their low fertility scenario has population peaking at 8.7 billion in the 2050s. Various organizations give a range of projections and then typically offer their prediction as to which is most likely—they call them probabilistic projections and they pick the most probable (I say that’s a prediction). I’ll try to use those terms intentionally in this post.
To get to peak population, things have to change NOW. Rather than the whole century, I thought a more digestible task would be to look just 17 years out, to the 20 most populous countries in 2040, and only consider fertility declines. Most journalists and think tanks use the UN’s medium fertility variant but might not know what’s behind it. By comparing it with the constant fertility variant—which just holds fertility where it is today (in 2022, when the data came out)—we can see what has to happen for that future to come to fruition.
I’ve highlighted the most populous countries with the highest fertility to show how much the average number of children per woman would have to fall for the medium scenario to materialize.
How likely is Tanzania to go from its current fertility of 4.66 children per woman to 3.49 children per woman over the next 17 years? To look at the outside factors, our baseline, we could see how long it took other countries in the region to see similar fertility declines. Three African countries have fertility rates close to 3.49 today (where the medium variant puts Tanzania in 2040): Gabon, Ghana, and Zimbabwe. How long did it take each of them to go from about 4.66 (where Tanzania is now) to where they are today? The following chart shows their varied fertility over time.
Zimbabwe had the longest and most meandering journey, at 32 years. Gabon took 26 years and Ghana only 21 years. Given that none of those countries saw declines faster than 18 years, we should revise down our confidence for Tanzania. This is also the time to bring in the larger literature on African fertility declines. That would make this newsletter too long, but I’ll refer you to the discussion in my book on p. 38 and Bruno Schoumaker’s work.
Now we turn to inside factors. What kind of leadership commitments or investments have been made in Tanzania that would lead us to think fertility will decline to the UN’s medium scenario level by 2040?
Tanzania’s previous president, John Magufuli, said in 2018 that Tanzanians should stop using birth control because the country needed more people and he was afraid of population aging. He also insinuated that foreigners peddling contraception campaigns had sinister motives. Samia Suluhu Hassan took office in 2021, and has the opposite view of family planning, seeing it as essential to the country’s success. Leadership commitment, plus resources towards education and family planning, are key inside factors to consider.
As for the other countries on our top 20 list, for the DRC to reach the medium scenario fertility rate it would need to go from 6.11 to 4.62 children per woman, basically where Tanzania is today, or Gambia or Burkina Faso. It took Tanzania 32 years and Gambia 31 years to make similar declines. Burkina Faso did it in about half the time, just 16 years. Is DRC more like Burkina Faso, or Tanzania?
Working through these steps, which is just a starting point for a comprehensive evaluation, may feel like a lot of work, but companies and governments use population projections all the time in their work and not all carefully consider the assumptions behind them. When a non-profit, like Charity Water, builds a well in Pakistan or Madagascar, they have to think about how many people will use the resource today and ten years from now. When MTN is setting up infrastructure for cell service in Nigeria, they have to consider future usage as well. The same holds true for a host of other advocacy, non-profit, and for-profit outfits. I’d love to know what inside factors you see as relevant for fertility declines in the countries we’ve discussed. How fast do you think it will decline, and why?
If you want to understand more about what’s behind the various projections here are a few resources. First, my colleague Richard Cincotta has written about the projections in more detail in a handy 3-page document published by the Population Institute. Check out the third page especially. The UN discusses its methodology in this document starting around p. 27. And noted demographers Stuart Gietel-Basten and Tomas Sobotka offer a scathing review of the IHME projections here.
The Incredible Challenge of Counting Every Global Birth and Death - This haunting article by Jeneen Interlandi for the NYT illustrates the importance of good data. For millions, it’s a matter of life and death. I highly recommend taking the time to read it. Key takeaways:
Nonprofits Vital Strategies and Bloomberg Philanthropies estimate that some two billion people do not have birth certificates, and only half of the 60 million or so deaths that occur each year are recorded in any meaningful way. Mostly these are in low-income countries.
This is in sharp contrast to the abundance of data in wealthy countries. Discussions of data here are about privacy and use—a totally different set of issues. I’m always writing about the demographic divide. There’s a data divide too, of course.
Without a good baseline, how can we measure progress on key indicators, like health, education, and poverty? As I write about in my book, we do an especially poor job of collecting data disaggregated by gender, even though we claim to have specific goals around lifting up women.
And some news…I’m giving a TED Talk in a couple of weeks, which is why I’ve been quiet on the newsletter front. Preparing is hard. I hope it goes well.
