The Office for National Statistics’ recently-published national population projections expect continued population growth for the United Kingdom at a relatively strong pace until 2047. This forecast might come as something of a surprise given the constant drumbeat of news about falling fertility rates in the UK and abroad. As this brief review of British population dynamics will suggest, the ONS’ forecast includes several unfounded assumptions which lead to considerable excess optimism about the future of British population growth.
Projection vs. Forecasting: a distinction without difference
Before addressing the data the ONS provides, it should be noted that the ONS adopts an unreasonable distinction: that “projections” are different from “forecasts.” The ONS places a warning on their published document saying, “National population projections are not forecasts and do not attempt to predict potential changes in international migration. Migration assumptions do not directly account for recent and future policy or economic changes. Demographic assumptions for future fertility, mortality and migration are based on observed demographic trends.” This disclaimer is strange, since “projection” is simply, etymologically, derived from the Latin “to launch forward,” while “forecast” is simply, etymologically, derived from the Old High German “to launch forward.” They are literally synonyms, and in fact identically derived, it’s just that Alfred the Great made forecasts, while William the Conqueror made projections.
The ONS insists on this kind of logomachy in order to avoid serious scrutiny. “Projections” presumably based simply on walking past trends forward sound safe. If they are wrong, nobody can be blamed: it’s only a model. “Forecasts” however might be wrong, they are statements about what will be. Outside observers might track and check the accuracy of forecasts. In reality, of course, the ONS is perfectly aware that virtually every single reader of their reports will immediately interpret their “it’s only a model” as an actual prediction of the future, because that is, in fact, what it is. Because this is the inevitable and nearly exclusive popular interpretation of their data, they owe it to the public not to hem and haw about “projections,” but in fact to produce a credible and defensible forecast, or prediction.
The projections produced are, unfortunately, not credible or defensible. Below, I will lay out reasons to believe that not only are the projected immigration values dubious, but the projections for fertility and mortality are perhaps even more so. We’ll start with immigration.
Immigration
The ONS does anticipate that deaths will outnumber births in the very near future. How, then, does population growth continue apace? The answer is international migration. The ONS anticipates that international migration will decline over the next few years, then remain stable at around a net gain of 350,000 people per year. The ONS says they “do not attempt to predict potential changes in international migration,” but this is not true: they predict international migration will fall quite a bit in the mid-to-late 2020s, but then it will remain at levels greater than those observed any time in the 20th century. The ONS anticipates that net migration into the U.K. will be about the same in 2030-2039 as it was for the entire cumulative period from 1940-2009.
Figure 1: Net migration by decade
This could be correct, or it might not be, but it seems clear that this is neither a simple forecast of “current rates continue into the future,” nor “future trends return to historic baselines.” Rather, the ONS is assuming that the U.K. will maintain an immigration stance which is more open to immigration than it was at any point in the 20th Century – but not quite as liberal as it has been in the last few years. This may or may not be true, but the ONS should forthrightly state that their view of population presupposes a specific set of political outcomes, namely, persistent failure of political efforts to reduce migration for many decades to come. If that political forecast implied by these projections does not come to pass, British population growth will come crashing down. Moreover, the 21st century thus far saw a wave of quite unique events: European Union expansion allowed a dramatic increase in immigration in the 2000s and 2010s, while major refugee crises (especially in the Middle East and Ukraine) drove huge increases in the 2010s and 2020s. If the ONS expects the U.K. to rejoin the E.U. or expects that the unusual pace of civil and interstate conflict observed 2010-2025 will continue through the 2040s, they should say so, and perhaps defense budgets will need adjusting too.
Mortality
Life expectancy is stagnant in the UK. According to the ONS’ data release, in 2012, English baby boys had a life expectancy of 79.3 years. In 2023, English baby boys still had a life expectancy of 79.3 years. This pattern of stagnation was also generally true for baby girls, and in Northern Ireland, Scotland, and Wales.
And yet, rather than recognize this stagnation and take seriously the possibility that the health of British people may not inexorably improve as it did during the 20th Century, the ONS insists on assuming that “long run rates of mortality improvement” will continue. These long-run rates assume that gains to health in the next 20 years will be better predicted by the gains made in the 1970s, 1980s, and 1990s, than the stagnation experienced over the last decade.
This is not a credible forecast. Every single ONS projection since 2012 has forecast rising life expectancy, and every single forecast since 2012 has had to start that forecast rise at a lower point, because the rise in lifespans continues to not happen. The ONS would be better off just assuming that life expectancy is going to remain stable where it is, that British lifestyles and the British healthcare system are not going to improve considerably in the near future. To their credit, the ONS has reduced the rosiness of their projections: while in 2012 they expected annual life expectancy gains of 0.18 years between 2012 and 2047, in these latest figures they expect just annual life expectancy gains of just 0.12 years between 2022 and 2047. Nonetheless, that is still far greater than recent experience suggests may be credible.
Fertility
But perhaps the strangest error the ONS makes is on fertility. Since 2010, the ONS has been persistently forecasting that fertility would remain stable. In fact, fertility has fallen dramatically, as the figure below shows.
Figure 2: ONS fertility projections, against actual
The 2016-based forecast, for example, anticipated a 2023 total fertility rate of 1.79. The true value was likely nearer 1.41-1.45 based on data reported for 2023 thus far (the statistical agencies in the United Kingdom report birth data at a ponderous speed compared to other countries). That miss of about 0.35 children per woman is enormous: birth rates were 20% lower than expected. The entire COVID pandemic only reduced life expectancy by about 1.2% at its height, and the large forecasting error for mortality I noted above nonetheless amounted to just a 3.3% error. The ONS’ misses on fertility in their previous forecasts have seismic implications for planning in the U.K. Schools planning their entering cohorts of 5 or 6 year olds based on these data would have found fully one-in-five of expected incoming children to be missing.
So, what about the newest fertility forecast? Is it also excessively rosy?
Bizarrely, no. It’s actually too pessimistic – at least in the short term. Northern Ireland and Scotland both publish relatively reliable provisional birth data and, while England and Wales do not, England’s health system publishes monthly data on the number of newborns registered with pediatricians and other doctors, which turns out to be extremely strong predictive of eventual birth reports. As such, although we do not know for sure what birth rates were in the U.K. in 2024, we know that births in Scotland through October were pretty much unchanged vs. 2023 after years of decline, births in Northern Ireland were down just 1.6% (vs. an average annual decline of 2.7% over the prior 5 years), and that infant patient counts in England were up nearly 2% year-on-year in September (vs. an average annual decline of 2% over the prior 5 years). In other words, unless Welsh women went on an extraordinary birth strike in 2024, UK births in 2024 almost certainly rose vs. 2023. And since the pace of that increase actually grew bigger over the course of 2024, it’s reasonable to think 2025’s birth rate may not be as low as expected either.
Figure 3: Demographic Intelligence vs ONS projections, when compared to actual
Using these data to calibrate nearby years, as well as a demographic cohort model comparable to that used by ONS, my consulting firm notified clients this previous December of our expectations for British fertility, shown in the figure above as the “Demographic Intelligence” forecast. Our forecast is almost certainly more accurate than the ONS’, not least since we incorporated 2023 and 2024 provisional data, which the ONS elected not to do. We find that the ONS has probably underestimated fertility for the mid-2020s. The gap is not enormous; the ONS expects fertility rates around 1.41-1.43, whereas we expect them nearer 1.46-1.49. Nonetheless, that’s a 4% difference in the number of children who will be starting school around 2028-2032, and a 4% difference in how much space will be needed in maternity wards. It may sound small, but if pharmaceutical companies use these forecasts to plan production of vaccines (which can be expensive to store in large quantities), it won’t feel small to the parents whose child’s jabs have to be rescheduled because the local clinic ran out.
Furthermore, while the ONS expects a gradual long term increase in fertility, my firm projects a decline. In my view, the fundamental dynamics driving low fertility (namely, poor economic outcomes for young adults and especially young adult men, housing costs at levels which are ruinously unaffordable for young adults, and rapidly falling marriage rates) are not about to disappear. While there may be year-to-year volatility, without a major course correction, there is only one reasonable direction for a projection of British fertility in the long run: down.
It should be noted, however, that course corrections are possible. Migration is obviously highly sensitive to politics, but fertility is too – as the SMF has argued. Pro-family policies in most countries are relatively small: even in unabashedly pronatal Hungary, the reality is that the benefits a family receives for investing time in raising their children are extremely modest compared to the wages they receive from investing time in market work. But even these modest policies are known to be effective: recent academic reviews strongly support the notion that more generous family policies modestly boost births. The British Government should not allow the ONS projections to lull it into a false sense of security: policies to tackle the root causes of falling fertility and directly support families with childrearing are needed to avert a worsening demographic picture.
