In policy, age is more than a number
Unlock Editorial Digest for Free
FT editor Roula Khalaf chooses her favorite stories in this weekly newsletter.
I think there's an unwritten rule that every article or policy discussion about population aging must frame the debate with some scary statistics. Here is some information from the United Nations. Between 2015 and 2050, the proportion of the world's population over 60 years old will nearly double, from 12% to 22%. In 2021, for every 100 people aged 20 to 64, there will be 17 people aged 65 and over (this is the so-called “old age dependency ratio”); by 2050, there will be 29 for every 100 people.
So far, it's still so familiar. But what if these statistics don’t serve as a useful framework for debate? What if “65 and older” is the wrong definition of “elderly”? In fact, what if chronological age is not a good measure of aging at all?
The only thing a person's actual age can really tell you is how many years they have lived. Policymakers are concerned about the above statistics because they use chronological age as a proxy for other things they worry about, such as the number of frail or sick people who will need health or social care in the future, or economic and social conditions. The fiscal implications of fewer workers and more pensioners, among other things.
If chronological age was a reasonable proxy for all of these things, that would be fair enough, but is it? A paper published last month by economists Rainer Kotschy, David Bloom and Andrew Scott argued that relying on chronological age is “at best is incomplete and, at worst, misleading” because it “provides only limited information about the aging process.”
Most obviously, people of the same age can vary greatly in frailty or health. Kotschy, Bloom and Scott used US and UK data on the physical capabilities of people over 50 and found that the healthiest 10% of the population at age 90 were nearly as frail as the median at age 50.
Average health and fitness levels by chronological age also change over time. For example, according to the Office for National Statistics, women aged 70 in the UK had roughly the same level of poor overall health in 2017 as women aged 60 in 1981.
If you use chronological age as a proxy for when people stopped working, then age also varies significantly across countries and over time (and is, of course, particularly sensitive to changes in state pension ages). In a country like the UK, where the proportion of people over 65 in employment has risen from 27% in 2014 to 40%, what is the significance of the “old age dependency ratio” that classifies people over 65 as “dependants”? 2024?
As Warren Sanderson and Sergei Scherbov, the leading researchers in the field, put it: “Should a 60-year-old Russian in 1950 be considered as old as a 60-year-old Swede in 2050? If not, is there a better alternative?”
The alternative proposed by Sanderson and Scherbov is to define the onset of “old age” as the point at which 15 years of life expectancy remain. Through this lens, the past, present and future look very different.
In the UK, for example, life expectancy has increased significantly until the past decade, with the number of people over 65 increasing by 8.3 million between 1981 and 2017, but the number of people with a life expectancy below 15 falling by 7.4 million. If old-age dependency ratios are recalculated using the definition of “old age,” old-age dependency ratios are lower in all regions of the world except sub-Saharan Africa and Oceania (excluding Australia and New Zealand), and are projected to rise at a much smaller rate.
Of course, it's also possible that this isn't the right lens – it depends on the specific issue you're worried about. Asking the question of when people should be able to receive their state pension. In recent years, new “clocks” have emerged that aim to measure a person's “biological age” based on indicators such as proteins in the blood. Given that any system using chronological age or average life expectancy is unfair to poor people who live shorter lives, could they one day be used to determine everyone's state pension age?
Scott told me that even if clocks became scientifically powerful enough, he wasn't sure people would accept it. “Can you imagine two people of the same age, working the same job? … but people can claim (the state pension) three years in advance?”
There is no perfect indicator that can replace chronological age as a measure of population aging. But once you understand the definition of “old” as something other than the number of years people live, it starts to look more malleable than inevitable, and those scary statistics about how quickly we age look more like a challenge Than fate.
This article has been revised to clarify a sentence about the employment rate of those aged 65