Faster mortality decline – slower population ageing. How could this be?

Posted on by

Professor Alistair Woodward and Professor Tony Blakely

SNZ predictions of number of New Zealanders aged 65 plus

SNZ predictions of number of New Zealanders aged 65 plus

Have you sat in a meeting recently, or listened to the radio, where someone is invoking the aging population as a harbringer of doom and gloom due to the tsunami of older people with poor health? A tsunami that will overload health and social services, etc. Have you then asked yourself “Hang on a minute, if people are living longer are they not also healthier?”. Or even “Why on earth do government agencies and some academics keep talking about the number of people aged greater than 65 as a marker of some dependency on the state?”. Well, this Blog is for you. We use a paper just published in PLoS ONE that demonstrates how with falling mortality rates, population aging may actually slow due to a rapid increase in the age at which your remaining life expectancy is 15 years or less. And we draw on analyses from our own book, The Healthy Country? A History of Life and Death in New Zealand, to critique naïve assumptions about population aging. Which begs many questions, such as “Why is the age of entitlement for New Zealand superannuation still 65?” Read on.

Life expectancy at birth in New Zealand is rocketing upwards, recently gaining 2-3 years every decade (and see just released 2012-14 lifetables). But is this altogether a good thing? Many worry that we face a future that will be overwhelmed by old people. They ask: will we be weighed down by an aged population, threatened by a top-heavy age pyramid due to an ever-growing fraction that is elderly?

You might think that the faster mortality falls, the more quickly the population ages (assuming no changes in fertility and migration). But this isn’t necessarily the case, according to a recent paper in PLoS ONE.

At issue here is what defines “old age”.

When Dick Seddon introduced the Old Age Pension in 1898, he said you reached old age on your 65th birthday. And that is still the measure of eligibility for New Zealand Super, and is widely used as a definition of old age (1). Elsewhere, old age starts earlier. For instance, in its World Ageing Report, the United Nations treats 60 years as the beginning of old age (2).

Taking this approach, it is plain that the lower mortality rates fall, the greater the proportion who reach their 60th or 65th birthdays, and therefore the faster the population ages.

But is “years lived” an appropriate measure, when “years yet to live” are increasing so fast, and at every age? In New Zealand, life expectancy for a female once she turned 65 was 19 years in 1981; in 2011 a non-Māori female aged 65 can expect to live another 24 years.

Sanderson, ScherbovIn the PLOS paper Sanderson & Scherbov argue that instead of a fixed chronological age, being old should be defined by a dynamic, prospective measure, such as remaining life expectancy, one that takes account of changing circumstances. For instance, one might use the age at which one has less than 15 years life expectancy remaining, since that corresponds fairly well with the onset of significant age-related disabilities.

From this angle, it is apparent circumstances have changed greatly, and human prospects have improved to a remarkable extent. Using the “15 years yet to live” standard, French women would have been defined as old beginning at age 58.4 years in 1900, at age 64.8 in 1956 and at 74.6 in 2012.

Looking ahead, Sanderson & Scherbov ask: what proportion of the population will be “old”, using the conventional definition (65 years+) and a prospective measure (less than 15 years to live)? They apply these scenarios to the German population, rolling forward from 2013 to 2030 and 2050 and, and this is the innovative bit, they test the effect of varying the rate of mortality decline.

With no change in mortality rates, the population ages at a similar speed, whichever definition is applied. Assuming mortality declines in the future, the proportion over 65 increases, and the faster that mortality rates fall, the greater the increase in the aged segment of the population.

The picture is quite different when the prospective measure of “being old” is used. In this German dataset, the proportion with less than 15 years to live increases from 2013 to 2050, but is less, at any point in time, if mortality is falling than if it is static. And the faster mortality rates fall, the slower the increase in the proportion defined as “old”.

A Healthy Country - Image of bookThis is the answer to what seems a paradox. As life expectancy stretches out into the future, the threshold of old age also advances. Depending on the rate of change in life expectancy, and the existing age structure of the population, it is possible for mortality decline to co-exist with a very modest increase in the proportion defined as elderly. All else being equal, the faster the increase in life expectancy, the slower the rate of population ageing – using a more rational measure than a ‘static’ age of 65.

This is a fundamentally important insight. Continuing improvements in life expectancy will certainly add years to the lives of individuals. But service planners and others concerned about population ageing must pay attention not only to the number of future 65 year olds, but also the characteristics of this group. And steps to drive mortality lower are not necessarily at odds with efforts to slow population ageing.

One very important characteristic of older cohorts is that they are living longer. But there are other significant differences as well. These include cognitive improvements (the “Flynn effect” applies at all ages (3)), better education, less disability than was experienced by age-equivalents previously, and as a consequence, increased healthy life expectancy. Whilst it is very hard to find ‘good studies’ testing whether the percentage of one’s life in disability is changing as we live longer, the broad brush consensus is that the proportion lived in good health is holding up (4). Put another way, people are healthier at older ages.  A review of the results from the New York Marathon found that the average finishing time of the 50 fastest 70-79 year olds was falling by 2 minutes a year, the most rapid improvement of any age group (5)!

It does not take too much reflection to realise that as the population ages, it is both healthier and the age at which you have 15 years of expected remaining life expectancy rises. The Sanderson & Scherbov paper elegantly brings this home, and makes it explicit. The implications for society are profound, most obviously that demographic and other social projections cannot naïvely assume that the expected health status of a 75 year old in 20 years is going to be the same as today. Critically, there is little logic in keeping the age of eligibility for New Zealand superannuation at 65.


1. Roebuck J. When does old age begin? The evolution of the English definition. J Social History 1979;28:416-426

2. United Nations. World Population Ageing 2013 [Internet]. New York, USA: Department of Economic and Social Affairs, Population Division; 2013. Report No.: ST/ESA/SER.A/348. Available: http://www.

3. Skirbekk V, Stonawski M, Bonsang E, Staudinger UM. The Flynn effect and population aging. Intelligence 2013;41:169-177

4. Christensen, K., G. Doblhammer, R. Rau and J. Vaupel (2009). “Ageing populations: the challenges ahead.” Lancet 374: 1196-1208.

5. Jokl P, Sethi PM, Cooper AJ. Master’s performance in the New York City Marathon 1983-1999. Br J Sports Med 2004;38:408-412

This entry was posted in Uncategorized and tagged by Kate Sloane. Bookmark the permalink.

3 thoughts on “Faster mortality decline – slower population ageing. How could this be?

  1. Fascinating piece. Thank you.

    A question: Is there a sense of what any decompression of morbidity may do? Mortality is declining, and age at which we enter our last 15 years is increasing, but what happens when we overlay morbidities such as T2DM over this? Is the age of onset of T2DM decreasing? Does this mean we will have a population which is living longer, but living with morbitities that don’t lead to mortality at perhaps the same rate as smoking-related diseases and other environmental exposures once did?

    • An important question Jamie. I have two answers. First, it is actually very hard to undertake ‘good’ studies on compression or expansion or morbidity over time. You need the same measures asked repeatedly, and assumptions that any subjectivity in how people answer does not change over time (e.g. that expectations do not increase over time of what is ‘good’, but this can be overcome by asking ‘objective’ questions like ‘walking up flights of stairs’). That all said, one of the most authoritative reviews of this areas (Christensen, K., G. Doblhammer, R. Rau and J. Vaupel (2009). “Ageing populations: the challenges ahead.” Lancet 374: 1196-1208.) concluded that, on balance, we are living longer, living longer in good health, and that the percentage of one’s life in poor health is probably not increasing – in the past at least. There are variants on this, such as the idea that we live less of our life in severe morbidity but more in mild morbidity – the so-called dynamic equilibrium hypothesis, for which there is some NZ evidence (Graham, P., T. Blakely, P. Davis, A. Sporle and N. Pearce (2004). “Compression, expansion, or dynamic equilibrium? The evolution of health expectancy in New Zealand.” Journal of Epidemiology & Community Health 58(8): 659-666.)
      Second, though, the future may not be like the past. And without wanting to over-sell the obesity epidemic, it does tend to generate morbidity – without necessarily killing you as successfully as (say) smoking. Think increased diabetes, osteoarthritis, and the morbidity per se of obesity. So, it might be that we will see some expansion of morbidity (i.e. increase in the percentage of lives lived in poor health) in the next few decades. Watch this space, or apply the precautionary principle to act on the food and physical activity environments now.

      Tony Blakely

Leave a Reply

Your email address will not be published. Required fields are marked *

* *