The Insurer's Dilemma: Future Cause of Death UnknownPublished: April 11, 2011 in Knowledge@Australian School of Business
Population longevity has big implications for individuals, businesses and society. Sooner or later everyone has to do the sums and divide their retirement savings by the number of years they think they have left. How long is that? One way of looking at it is by considering what's most likely to cause an individual to die. But with advances in medical science and social change, that's becoming increasingly difficult.
Michael Sherris, a professor of actuarial studies at the Australian School of Business, says understanding past mortality trends by cause provides useful insights into what's driving increasingly longer lives worldwide. But there's a catch. Sherris warns that projections from the headline figures work on the assumption that past patterns of what causes people to die will continue. Drilling down from the aggregate level to the underlying causes of death, the trends are more complicated.
Premature death used to start with infant mortality. Then came accidents in youth, exacerbated by not wearing seat belts and drink driving. Smoking and the consumption of fatty and salty foods brought on an epidemic of heart disease and stroke. These health risks are receding. Cancer remains the greatest threat. But what if breakthroughs in medical science reversed the trends with this sometimes life-threatening disease as well? An unexpected increase in life span is a risk to retirement income and Australians are now facing up to the reality that they are under-insured against old age.
Longevity is a challenge with no easy fixes, according to John Piggott, professor of economics at the Australian School of Business and director of the ARC Centre of Excellence in Population Ageing Research. "Only in the last 20 or 30 years have people who lived to their 50s been expected to live a lot longer," Piggott says. "People will have to delay retirement. Between the 1970s and the turn of the century, the average retirement age was going down. Now it's going up again. It is 59 to 60 in Australia – and in 20 years' time, I predict it will be 65."
The federal government is raising the age at which Australians can qualify for a pension – currently it's 65 for males – it will increase on a sliding scale between 2017 and 2023 to 67. Tax-free withdrawal from superannuation schemes is allowed from age 60. "If you cut off access to super from 60 to 65 in a hurry you would get a lot of pushback," says Piggott. "We do not have any longevity insurance market in Australia and the aged pension has been carrying a lot of the weight. That has to be fixed up. We need an environment for innovative longevity insurance products."
Research into the causes of mortality can assist private sector life insurers, who already provide longevity protection products such as lifetime annuities, to improve pricing by being able to better model longevity risk, notes Challenger Life chief executive Richard Howes. "There is not yet a fully underwritten product offering individual profiling and reflecting that in the price," he says. "There may well be a time in the future that will happen. When annuities increase and the market for longevity insurance products develops, one would expect the underwriting and profiling could become more viable. Globally, there is not individual underwriting for annuities – you do not have to go and have a medical to get one. There is no test of whether you are pre-disposed to a heart condition or are overweight.
"In some countries, including the UK, there is annuity pricing profiling by postcode – so depending on where you live, you get different pricing, similar to car insurance. If you are in a certain demographic, you will have a different life expectancy. White-collar workers are expected to live longer than blue-collar workers and so could pay more for annuities," Howes suggests.
Better Rates for Smokers
A market for impaired annuities already exists in the UK. For example, a smoker can get a better-priced annuity than someone who does not smoke. Curiously, an unhealthy profile can be a financial advantage. For life insurance companies, the more information they have about how long a person is going to live will be reflected in the price of the annuity. If insurers dealing annuities are able to more specifically reflect life expectancy, then industry fund members in sectors where people may not be likely to live as long may be able to get better pricing than, say, wealthy investors who have longer life expectancies.
Separating causes of death from mortality statistics is one way of calculating future mortality improvements, Howes says. "We believe that there is a cohort effect under which people born in certain years experience different rates of mortality improvement."
So what are the trends in the causes of mortality? There are different trends for ages, sexes, diseases and countries. The causes of mortality are not independent. A decrease in one cause may result in an increase – or a decrease – in other causes. Sedentary living in many industrialised countries is set to contribute to an increase of deaths due to chronic diseases later in life. On the other hand, the anti-smoking campaign showed how recent interventions against the risk factors can have a positive impact on future mortality. Even if cancer deaths are still increasing at older ages, there is a declining trend from the 1990s in the US, Japan and the UK, indicating that smoking control measures are starting to affect older ages. While the death rate from diseases of the circulatory system have fallen over the last 50 years, thanks to prevention campaigns, it is unlikely that the trends will continue, Sherris believes.
Obesity Overtaking Cancer?
"We have had a lot of success with cardiovascular disease to cut heart attacks and strokes, with the prospects of surviving improving – thanks to drugs, surgery and prevention – since the 1970s," Sherris says. But there's a limit to future improvements. "It's important to understand where that improvement is coming from – otherwise you do not understand what's ahead," explains Sherris. "Statistics show middle-aged deaths from heart attacks and strokes have been replaced by people dying from cancer, although the number of cancer cases has remained fairly steady. In 1955, 16% of 55-year-old deaths were from cancer, while in 2000, it tripled to 45%."
But there has not been a lot of success with curing cancer, notes Sherris. "The importance and uncertainty of future changes in cancer death rates are critical to any longevity risk model," he says, "The future trend affects the price of retirement products. If the trend is going to improve, then the cost of the longevity products is going to increase. If there is an obesity epidemic, it may reduce – unless a prevention plan overcomes that. The main thing to recognise is that there is a huge amount of uncertainty and that has to reflect in pricing and risk management."
In research with Séverine Gaille at the University of Lausanne in Switzerland, Sherris analysed mortality rate age patterns for the main causes of death – circulatory system, cancer, respiratory system, accidents, infectious diseases – for nine major countries over 50 years to understand the reasons for recent improvements in mortality rates and the extent to which these will continue in the future. "Everyone is going to die but the question is how long can we live," says Sherris. "Some people argue there is potentially no limit to life – but not everyone agrees. It may be possible to improve our lifespan a lot. We need to consider this in context. Currently the population is set up to retire at 65 when half of it may now live beyond 85."
There have been significant changes in the causes of death in the past century. Take infant mortality. Infectious and parasitic diseases caused 10% of child deaths (aged 5) in Australia in 1955, but only 2% in 2000. These diseases affect mostly children and people aged above 60 across all countries. Mortality from childbirth for women aged 20-30 is thought to have caused a statistical "hump" that is noticed from the 1950s but disappears from the 1960s to the 1970s. HIV appeared as a killer of males aged 30 to 40 from the 1980s or 1990s, depending on the country. The disease spread easily in different parts of the world as travel and commerce grew during the 20th century.
Globalisation and Dying
There are some big differences between countries on causes of death although these are narrowing with globalisation. For example, overall Japan had five times the rate of infectious and parasitic deaths in 1955 than Australia, but only a quarter of the mortality rate from circulatory (heart and stroke) disease. Japanese eat less fat and food from animal sources but more salt than those in Western countries. Over the past 50 years, the consumption of animal products and fats increased considerably in Japan, while the salt intake declined as Japanese started to have more similar culinary habits to other developed countries. Mortality due to stroke drastically decreased over the last decades. Now the researchers say the growth in obesity suggests that mortality due to several heart diseases may increase in the near future.
Trends tend to run internationally, with common developments in healthcare and other campaigns targeting specific diseases. Many national healthcare systems in advanced countries were launched in the 1950s or 1960s, along with institutions dealing with diseases of the circulatory system, such as the Heart Foundation and the Australian Cancer Society (now called Cancer Council Australia), which helped mortality rates decline from the 1970s.
Research into why we are living longer shows that the causes of death vary with age. For instance, mortality rates in regard to diseases of the circulatory system have generally decreased at each age, although at different rates, but it still remains the most important mortality for ages above 75. Cancer is becoming the major cause of death for middle ages. Respiratory system deaths have reduced substantially at younger ages.
With external causes of death, such as accidents, the age pattern of mortality has been very different for males and females, particularly in the 1950s. The male accident "hump" is not visible for females and only appears later, reflecting the more general use of cars and other social behaviours including alcohol consumption by females. Accident levels vary across nations but with a general decline.
Respiratory system diseases have decreased for most countries and both males and females aged under 60 or 70 over time. For older ages, mortality increased from the 1950s up to the 1980s, followed by a decrease – a trend again linked to changes in tobacco consumption. Another factor is the increase in air pollution post-World War 2 in Europe and the US.
Gender is also relevant with male and female mortality rates showing different trends for cancer. Females are more affected by breast cancer and prostate cancer impacts males. Also the smoking prevalence among males shows up in statistics for lung cancer mortality. Male cancer death rates across advanced countries have shown similar evolution over time and across ages, with a general decrease in younger age groups. For middle ages, cancer mortality rates started to increase in the 1950s, and then decrease from the 1970s or 1980s, again a major factor being the decline in smoking. The mortality increase at older ages also reflects that with medical advances and treatment of cancer extending life, people diagnosed with cancer are living longer.
Internationally, more people are reaching older ages, due to declines in the main causes of death. Yet the age pattern for each cause of death has remained relatively the same over the last 50 years and across countries. What changes is the relative importance of the five main causes of death over time.
"The analysis points to the need for more careful modelling of trends in mortality improvement. By not taking into account the potential effects of trends in different causes across different ages will reduce the value of forecasts of future mortality rate trends," says Sherris.