Who knows the impact of Brexit? Why ONS projections are not predictions

The ONS’ latest population projections suggest an extra 3 million people will be living in the UK by 2028. But it’s important to understand why projections, which are based on past trends, are not predictions.   As Andrew Nash explains, the hugely uncertain impact of Brexit is just one reason why the ONS does not attempt to guess the future.        

There are many reasons for wanting to know the future population of the UK. How many school places will we need? How many hospitals? How many people will claim a State Pension? These are big questions, and ONS’s latest national population projections, published today, will help answer them.

Today’s figures project the UK population will continue to grow. In our latest (mid-2018) estimates it was 66.4 million, but we project a rise of 3 million to 69.4 million by mid-2028.

To reach these figures, we consider how birth rates, life expectancy and international migration will change. For example, we assume net international migration will average 190,000 per year in the long-term. The ultimate accuracy of the projections is dependent on how these factors change in reality.

Setting the assumptions

We can’t know for certain what will happen in future, but choose assumptions that are representative of what has happened in the UK to date. There is no single correct method of doing this, but we adhere to the standard good practice of considering trends over time. If we just looked at the last few years, this could be unrepresentative.

Taking net international migration, 190,000 is the average annual figure for the 25 years up to mid-2018. It is true that many recent years have seen higher figures, but as recently as the early 1990s some years experienced a net outflow of migrants from the UK. Given how much change can occur in a single generation, taking a longer-term average is preferred. Likewise, our assumptions for birth rates and life expectancy are not simply based on the very latest data.

Why don’t we factor in Brexit?

Our users may reasonably ask “Brexit is important. How do you factor it in?”. The simple answer is that we don’t. This is because we base our assumptions on past trends, and do not attempt to predict the possible effects of any future political or economic developments. An important reason for this is that any assumptions would inevitably be speculative. There are many different views on what impact Brexit may have, but no way of knowing for certain.

On that basis, we stick to the convention of producing a projection rather than a prediction. In addition, our use of longer-term trends ensures that a range of different circumstances is reflected in our assumptions.

Influence of births, deaths and migration on growth

In recent years, women have been having fewer children, so we assume fewer new babies than in our last set of projections. In addition, the rate of increase in life expectancy has slowed, reducing the size of our projected increases. Combined, this means the numbers of births and deaths are much more similar. Meanwhile, our long-term net annual migration assumption of 190,000 is 25,000 higher than last time. This is because we have retained the approach of taking a 25-year average.

The overall effect of these changes is that net migration accounts for 73% of growth over the first ten years, compared with 54% previously. The increased migration assumption contributes to this increase, but it is the changed life expectancy and birth rate assumptions which have the biggest impact. In other words, the increased similarity between the numbers of births and deaths means that the proportion of growth attributable to migration grows by default.

Alternative futures

What if migration is higher or lower than our main (principal) projection? What if women have more or fewer babies? What if life expectancy grows more quickly or more slowly?

All of these outcomes are certainly plausible. To reflect this, we produce variant projections with alternative values for birth rates, life expectancies and migration. Although our users will generally use the main projection, this range of variants allows them to explore the effect on the population of these alternative futures. If they believe that one of the variants is a more likely outcome, they can then use it for their analysis.

You can see our 2018-based national projections in detail by reading our bulletin here.

Andrew Nash is Head of Population Projections at the ONS