Does smoking affect mortality improvements?

Smoking kills. No surprises there: it says so on every packet of cigarettes. It has also been recognised in the pricing of term insurance in the UK since the early 1980s and more recently in annuity pricing. But what do we know about the impact of smoking on mortality improvements?

Actuaries have been criticised for consistently underestimating future gains in life expectancy compared with emerging experience. In retrospect, such underestimation was hardly surprising.  Levels of mortality improvements experienced in recent decades have been unprecedented, particularly among over-65s in the last 30 years. Without a crystal ball, who could predict that?

By improving our understanding of the underlying drivers of mortality improvements we can be better equipped to assess what might happen in future. Consider the effect of smoking. The life expectancy of long-term smokers has been shown to be, on average, between 10 and 20 years less than that of non-smokers. Mortality rates from certain causes of death, notably circulatory diseases and cancers, are massively increased by smoking, in some cases up to 20 times. (More information on the impact of smoking – and tobacco consumption generally – is available here)

It stands to reason, therefore, that lower levels of smoking should have a positive effect on mortality. Our recent analysis shows that reduced smoking prevalence in the UK has made a significant contribution to observed mortality improvements in recent decades. We have quantified this by comparing overall improvements with those experienced separately by smokers, ex-smokers and lifelong non-smokers, derived by combining cause-of-death analysis with smoking prevalence rates and the results of published research into the relative mortality risk from various causes of death for each of the three smoking categories.

We found that – for the ages of most interest to annuitant and pensioner portfolios – around 20% of the observed improvements in UK mortality in recent decades (equivalent to about ½% a year) can be attributed to reductions in smoking. As prevalence rates now appear to be levelling off at historically low levels, and past reductions cannot be repeated, we can therefore expect lower mortality improvements in future. Such analysis can be used by actuaries as part of their argument to demonstrate why – despite being proved wrong in the past – they do not expect future gains in life expectancy to be as large as those we have seen recently.

  Subscribe to RSS feed