John Downes and Niel Daniels of XL Re investigate the potential impact the smoking ban will have on policyholders
England will finally catch up with its neighbours when it introduces a smoking ban this summer. But, will it have an impact on protection business? Can we learn from countries that moved earlier? And could premiums for non-smokers increase due to the ban?
Chapter 28 of the Health Act 2006, which will become law on 1 July 2007, will introduce a smoking ban in England. After this date, it will be illegal to smoke in any place designated as an enclosed public place. Similar legislation has already been enacted in the Republic of Ireland (March 2004), Scotland (March 2006) and Wales and Northern Ireland (April 2007).
The prime aim of the ban is to reduce exposure to second-hand smoke in the workplace and in public venues. It also dovetails with the Government's target of a 3% reduction in overall smoker prevalence by 2010.
The immediate threat to the leisure industry, such as the loss of custom in pubs and restaurants, has grabbed the headlines. Less attention has been paid to the potential impact of the ban on individual health and lifestyles. So, will there be any lasting effects on the life and health insurance industry in years to come?
Smoker rating
Smoking is the only major rating factor used in life insurance that is behavioural - meaning the only factor the applicant can do something about. This makes smoker/non-smoker a powerful predictor of future experience. Other factors, such as alcohol consumption and body mass index are within the individual's control, but these rarely impact terms, except in extreme situations.
The UK market introduced differential pricing for smokers in the 1980s. Early differentials were modest. Over time, as actuaries tracked experience, differences widened. How much of this observed difference was caused by smoking, rather than being merely associated with smoking, is the vital question.
Remove smoking, and you remove a major cause of claim. But the extra risks associated with, but not due to, smoking will still remain. Assured lives that have given up smoking will now join the non-smokers, but they will still bring the associated risks previously identified and charged for in smoker rates. Counter-intuitively, with a reduction in smoker prevalence, non-smoker rates may increase as the more risky lives are no longer identified by their smoker status.
Passive smoking may also have an impact. Professor Konrad Jamrozik, formerly of Imperial College London, estimated that domestic exposure to second-hand smoke causes around 2,700 deaths in people aged 20 to 64 and a further 8,000 deaths in people aged 65 years or older in the UK. Exposure to second-hand smoke at work is estimated to cause more than two deaths a working day in the UK (617 per year, including 54 in the hospitality industry).
The smoking lobby believes the statistical evidence for illness caused by passive smoking is flawed and that the risks are greatly exaggerated, but, whatever the truth, the impact of banning smoky public and work places will be neutral to marginally positive on the non-smoking population. This could partially counter-balance the associated risks effect.
Underwriting
Separating applicants into smokers and non-smokers should be easy. In the early days of non-smoker definitions, there were attempts to allow certain types of smoker the benefit of non-smoker rates. An applicant may have been allowed non-smoker rates smoking "not more than 12 cigars a year". Today, the generally accepted underwriting definition of a non-smoker is someone who has "not used any form of tobacco or nicotine replacement products in the last 12 months".
For higher levels of cover, most life assurers test saliva or urine for cotinine - a by-product of nicotine - to verify the applicant's claim to be a non-smoker. If the test is positive, smoker rates are applied. For lower sums assured, however, non-smoker rates are granted on the basis of the declaration on the application form.
The 12-month qualification for non-smoker rates also raises the question of what happens when an applicant changes smoker status after policy issue. Statistics suggest that, of those who stopped smoking 6 to 12 months ago, roughly 50% are likely to relapse. In general, life offices do not ask ex-smokers or non-smokers to notify them should they recommence or take up smoking.
As we can see in the table on page 25, the differentials are significant, and what most people do not realise is that the differences are not totally due to smoking. Some of the extra cost is due to higher risks that are associated with smoking - such as general health awareness, socio-economic status and stress - which do not necessarily change on giving up.
For life offices, the prevalence of smoking in the population is important. With fewer smokers, an important marker of the risk an applicant brings is lost. As the ban may impact prevalence, it is worth examining the current trends, and asking what may happen. A good place to start is with the history of smoking trends in the UK over the years since smoking was identified as a serious threat to health.
According to the General Household Survey 2005, 45% of UK adults were smokers in 1974. This had gradually fallen to 24% by 2005.
More men than women smoke, but the gap has narrowed from 10% in 1974 to 2% in 2005. Adults aged 20 to 34 are the most likely smokers, while the proportion of heavy smokers (20 or more a day), fell to around 10% in 1998 and has remained virtually unchanged since.
Smoking prevalence also varies by social class in both sexes - for example, 35% of manual workers are smokers compared to just 19% among professionals. It also varies by marital status, with up to 37% of those who were single or co-habiting being smokers compared with only 22% of those who are married.
With the smoking ban approaching, the big question is whether there is going to be a significant shift in smoker prevalence that could lead to rate increases for non-smokers as companies rebalance their terms?
Early evidence for a rapid fall in smoking prevalence is mixed. The situation in the Republic of Ireland is probably closest to that in the UK and the graph below illustrates that experience.
As the ban came into place, smoking prevalence dropped, but it now appears to be edging upwards again.
Irish experience also suggests that the ban has displaced smoking from the workplace or from social gatherings to the home, and therefore the overall effect of the ban is likely to be muted.
Given that life underwriters already differentiate between smokers and non-smokers, it may seem surprising that the ban could actually increase costs for non-smokers, but it is indeed possible if there is a significant drop in smoker prevalence. This may be counter-balanced by reductions in claims due to passive smoking.
Looking to other territories, a rapid reduction in smoker prevalence does not seem likely, while other measures such as tobacco marketing restrictions and heavier tobacco duties will also have an effect over a longer time span. A 'One Year Review' by the Irish Population Health Observatory showed that 93% of the public thought that the ban was a good idea and 96% thought that it had been successful, including the vast majority of smokers.
So, despite the uncertain impact of anti-smoking legislation on protection assurers, one thing does seem certain - smoke does get in your eyes and the public, both smoker and non-smoker, would rather it did not.
John Downes is chief underwriter and Niel Daniels is UK pricing manager at XL Re, Life
