The rich data set of cohort life tables provided by StatsNZ shows how much population longevity has improved and how variation in lifespans has reduced.
StatsNZ publishes life tables for each cohort of New Zealand population. The latest update, published in April 2021, has data for cohorts born each year from 1876 to 2019.
The average lifespans for these cohorts – achieved lifespans for completed cohorts or estimated for those with some people still alive – have increased over the nearly 150 years. The increase has been steady, with noticeable dips only because of major events such as wars.
The lower two lines on the chart show cohort life expectancy at birth (see here for a definition). For the other lines, the cohort life expectancy at age 65 or 80 has been added to 65 or 80 to calculate the average lifespans expected for people of the birth cohort who reached (or are expected to reach) those ages. The chart shows 3 key trends:
- All the lines are increasing and flattening over time. This means that average lifespans are getting longer, but the pace of improvement is slowing.
- The steepest improvements are in the “at birth” lines between 1876 and the 1940s. This shows that the most important driver for longer lifespans over this period has been health improvements under age 65.
- The lines are now close together. This means that most people are living to age 65 and many to age 80, rather than only a ‘select’ group surviving to those ages. Deaths used to occur more evenly across the entire age range, but now are concentrated at older ages.
Looking at 100 years more closely
100 years on from the 1876 cohort is the cohort born in 1976, reaching their 45th birthday in 2021. This cohort is a prime target for retirement income policy makers and retirement product providers. Members of the cohort may be thinking about retirement planning themselves. The cohort provides a neat illustration of how much mortality has changed, and how differently we need to think about longevity, compared to 100 years ago.
More likely to reach retirement
The first change is the likelihood of reaching an age when retirement may be possible. New Zealand Superannuation is available from age 65, and people are now unlucky not to live to that age. Only 10 out of every 100 babies born in 1976 are expected not to do so. For the cohort born in 1876, the chances were fifty:fifty.
| Probability of living to age 65 from birth | ||||
| Cohort birth year | Female | Male | ||
| 1876 | 53% | 47% | ||
| 1976 | 93% | 89% | ||
Secondly, for those reaching age 65, there is now a reasonable chance of reaching age 100, especially for women. This would have been exceptional in the 1876 cohort.
| Probability of living to age 100 from age 65 | ||||
| Cohort birth year | Female | Male | ||
| 1876 | 0.4% | 0.1% | ||
| 1976 | 8.6% | 4.4% | ||
The reasons for these longevity improvements across the world and in New Zealand are well understood, and broadly come down to public health initiatives and medical technology keeping ahead of diseases.
We are living longer, differently
As shown in my last post, another good way of understanding longevity is to look at charts of the distribution of lifespans achieved in that cohort, and calculate indicators such as average lifespan, the median age at death, the most common age at death, and the age to which at least one in five people lived to.
All these indicators have improved markedly over 100 years. The chart below shows this for men; and here is the full set of data. The picture is similar for women.
Between the two cohorts a century apart, life expectancy at birth increased by 32 years, and life expectancy at age 65 by 10 years. The difference between these two figures confirms the importance of mortality improvements below age 65 over this period.
For those reaching age 65, the median lifespan increased by 12 years, the most common lifespan by 18 years, and the age to which one in five live by 10 years. The changing shape of the distribution of lifespans over the century is confirmed by the various indicators changing by different amounts.
For the 1876 cohort, the chart of lifespans (ages at death) shows a fairly flat distribution. Variation in age at death was across the entire age range. For the 1976 cohort, the distribution has a more noticeable peak, meaning that variation is now concentrated in older ages. Most people are now living to older ages and uncertainty about age at death runs within a smaller range of ages which falls within the usual retirement period.
Useful indicators for retirement planning
Another implication of the previously flat distribution becoming a peak is that the usefulness of different indicators has changed. Life expectancy, the average, was a reasonable indicator of population lifespan, close to both median and most common age at death. Now, life expectancy falls well below median and mode as the distribution has extended to the right.
A key issue for retirement planning is to guard against longevity risk by using a range of realistic estimates for expected lifespan. Life expectancy is a poor indicator in this context – it will suggest too low an estimate.
This means that for the 1976 male cohort shown above, retirement planning should not focus only on 88 years as the expected average lifespan. Given the modal lifespan of 91 years and the ‘one in five’ indicator of 94 years suggests planning for a lifespan in the range of 90 to 95 years.