"The subject of longevity has unduly taxed the actuarial profession," says RisCura's Andrew Slater as he asks why forecasting mortality has been so difficult.
Last year the BBC website carried an article titled 'where are the missing 90-year-olds?' available at. It focused on Europe and the US, where the latest censuses are showing fewer people at extreme ages than expected. For example, in the UK we expected 457,000 people in their 90s, whereas there were only 429,000, a shortfall of 30,000. Sadly, the longevity improvements have not been as good as we hoped and more people have died sooner than expected.
In my opinion the subject of longevity has unduly taxed the actuarial profession. Perhaps, during the boom years of high equity returns and surplus to spend, the eye was taken off the ball. Then, once the mistake was realised, there was over-reaction. We have had nonsense like the first person to live to 1,000 has already been born. These forecasting errors are costly: early retirements in the 1980s and 1990s were allowed which could not be afforded; more recently monies have been pumped into pension schemes that have in my view unnecessarily interfered with the business growth of sponsors.
The BBC article reproduced a chart from the World Health Organisation. It shows the average life expectancy at birth between 1990 and 2011. There are several upward sloping lines, each a different colour to represent a different region of the world. And some lines had a kink around the mid-point year 2000. Some lines kinked more than others, and some kinked one way and some kinked the other way.
The question as I see it, is why has forecasting mortality been so difficult? Computers may have replaced slide rules as the calculating aid for actuaries, but the chart is surprisingly linear: a ruler would have sufficed. Sure, there has been a slight acceleration in the most recent years in certain regions, and a slight deceleration in others. In plain English, we see a kink upwards from 2000 for Europe for example, a kink downwards for the Americas, while South-East Asia is without deviation.
You can extend the chart backwards and see that the kinked linearity is not just a function of the time period selected. Longevity is increasing and we know why: across the population improved living standards and access to medical care, which itself advances rapidly for those able to afford it.
What really interests me is the bottom line which shows life expectancy for the Africa continent. I won't dwell on the fact that it is the bottom line and the inequality (lifespan of 50 not 70 in 1990) it represents. What is interesting is the fact that there is a kink, and the strongest kink. Life expectancy is increasing faster in Africa – a continent that has just passed one billion people, a seventh of the world's population – than anywhere else. My ruler predicts that African life expectancy will have caught up with Europe and America in a generation.
Written by Andrew Slater, managing director, RisCura
aslater@uk.riscura.com
