Figuring Out When You Are Going To Die: Life Expectancy

One of the biggest challenges of retirement planning and figuring out how much money will be needed, how much you need to save, is deciding for how many years you will need income, in other words, estimating when you will die.

This somewhat macabre exercise has become necessary with rising life expectancy and the decline in defined benefit pensions. The success of medicine, public health and improved nutrition in the 20th century is reflected in the vastly increased life expectancy of people in all developed countries like Canada, the USA and the UK. As Sherry Cooper outlines in her new book, The New Retirement (which I will review soon), when the retirement age of 65 was first introduced few people died much older than that. Even as recently as 1951 (for people my age that's recent!), this Maclean's article What's the Magic in 65? says that average life expectancy was 67 for men and 71 for women.


Those figures are far higher today and they are still climbing as the table below from the Statistics Canada website shows. Baby boys born in Canada in 2005 could expect to live to age 78, while baby girls would get to 82.7 on average. It is amazing to me that in a mere five years from the year 2000 to 2005, life expectancy would rise by 1.4 years for males. More pertinent for retirement planning, is that those men age 65 in 2005 could expect to live another 17.9 years, i.e. to age 82.9, while women would get to 86.1.


Three important points come out of this for retirement planning:

1. you can live many years in retirement - 20 or more years; and remember that's only the average so about half will live longer (? not sure how close the average age is to the median, which is the true dividing line in terms of half the number of people above and half below)
   
2. if you get to 65, there are more years left than you might have thought (most of the weak, the foolish and the unlucky have disappeared from the statistics)
   
3. the number of years you will live is still creeping steadily upwards, though there must be a limit somewhere we don't seem to have reached it yet

With respect to point three, this UK report from Which says that by 2050, men reaching 65 could expect to live to 91. Wow! ... update Feb.19 ... This is not a joke. In this BBC report Warning on Pensioners' Longevity, the Pensions Regulator is quoted as saying that pension administrators should assume men aged 65 today will live to age 89. Further on, the article quotes a pension consultant to the effect that many companies are not factoring in the extra longevity into their pension calculations. In other words, they are being too generous to pensioners today; the question will inevitably be "at what cost tomorrow?"

Point two surprised me. The older you are, the longer you can expect to live. It goes up steadily as this other table from Stats Can shows.

Since I am around 55 and if I retire I would need to plan for at least 25 years of income as a start. When I first did my retirement planning at my birth around about 1951, per the Macleans figures, I only needed to consider living to 67, or 12 years. This suggests a need to regularly re-evaluate retirement income plans upwards as one gets older. If you have died along the way that isn't necessary!

The story doesn't end there, however. Those averages can be refined. If you take care of yourself in various ways (being fit, not fat, drinking less, not smoking, eating food not processed junk, having a pet, mental stimulation, a neutral worldview, friends, cuddles for the women and sex for the guys) you will live even longer. The Sherry Cooper book recounts that taking care of yourself is especially effective in prolonging life expectancy when you get older.

How much longer you ask? What I could find on-line is sketchy and incomplete. Here is a UK report from Channel 4 that discusses the factors and provides some estimates of their impact. Then there is this life expectancy calculator from the University of Pennsylvania where you select your own circumstances from drop down boxes to get a prediction of your own life expectancy. You can use it to do trade-off analysis - if you like driving fast, how much is it worth in years to keep to speed limits? Should you bother sleeping more if that means missing that great late night poker show on TV? Trade-off pleasure for years ... shades of Dorian Gray! Another simpler calculator is at Death Timer. The Northwestern Mutual longevity calculator, based on US data, has a 12 question amusing quiz that shows how each factor adds or subtracts years as you answer each question. The Foundation for Infinite Survival (ambitious goal, huh?) has a lengthy and seemingly actuarily-founded life expectancy calculator too. Canadian Business' MoneySense offers this calculator for Canadians. Finally, you could simply predict your own death; if I understand this academic gobledygook by Siegel, Bradley and Kassel at Gerontology, it seems to say that people are pretty accurate in predicting when they will die.

What is needed:
Since the only way to know for absolute certain when one is to die is not palatable, something better is needed. As Moshe Milevsky pointed out in his book Insurance Logic, the financial challenge is to not run out of money, to mitigate the so-called longevity risk. I wish the discount brokerages or maybe our governments would put up a proper death calculator on the web to develop good individual estimates. This should include what is missing from all the above - confidence intervals, such as, "if your are 65 and you have the following charactersitics and lifestyle, your life expectancy is 81 years; there is 99% probability you will die by 87 years, a 95% probability of dying by age 85" and so on. Then we could plan effectively.

Alas, we cannot follow W. Somerset Maughan's admonition on WisdomQuotes: "Dying is a very dull, dreary affair. And my advice to you is to have nothing whatever to do with it."