L 01 What does life expectancy mean?
L 02 Longevity is increasing
L 03 Healthy life expectancy
L 11 One particular longevity table
L 12 What if you have a better estimate of your own longevity?
Thinking about longevity risk
L 21 How long should you plan to make your money last?
L 31 What longevity insurance is – and isn’t
STAGE L 01: WHAT DOES LIFE EXPECTANCY MEAN?
Where the route takes us
Why do so many people misunderstand life expectancy? Is it the math or the concept? Let’s take a look. (Spoiler alert: the math is simple.)
I thought I’d start by interviewing a few people on what the phrase “life expectancy” brings to their mind.
Tour Guide: The treatment of the topic of longevity is short and sweet. A friend asked me to start with that thought! Seriously, now that we’ve looked at psychology and happiness, and we know we need to save, and we know we need to invest our savings, the next natural question is: how long will our savings have to last? That’s where your future life expectancy is relevant. In my experience, most people misunderstand and underestimate it. I have three panelists with me today, as we start this topic, who I know have given the subject much more thought than most people.
J, you deal with financial matters all day long, in your job. You’ve probably given all kinds of matters a lot more thought than the average person does. So, what do you think about when someone mentions “your future life expectancy” to you?
Panelist J: That’s how long I’ll live. And you know, that keeps me awake at night!
TG: Why is that?
Panelist J: I’m sitting in a genetic pool with great activity. My great-grandmother went to 112, and at 99 was cross-country skiing to get around. So I’m going to have to cope with probably a very long life. I’m amazed at people who say: plan as if you’ll live to age 90. That’s not nearly long enough for me! I notice long-lived people all around me. A family friend, still living in her own home, is 107. I’m very, very cautious about this – and afraid.
TG: Why afraid? Is a really long life not something you want?
Panelist J: Of course, it’s what we all want – until we think about it. But it has a couple of bad aspects, potentially. What if my body outlives my mind? That scares me. And then there’s the financial consequences. What if things become much more expensive? You know, all in all these thoughts make an annuity almost an automatic decision for me.
TG: Gosh, you’ve thought about this very deeply! Let me turn to K next. How about you? Same sort of family history, or different?
Panelist K: Same sort, kind of. Not as extreme – yet. My parents are 84 and 74. I’m a health and fitness person, so I ought to have a long life. The first ice storm of winter makes me think about how I’ll have to walk on ice at 85 – that’s when I think about longevity. My friends joke and say: no need to worry, you’ll be in a home! I compare the amount of stress across the generations.
TG: Meaning what, exactly?
Panelist K: Do I have more stress than my parents? Certainly I have a different kind of stress – things are physically laborious for my parents, and for me it’s mental stress, whereas they’ve stayed mentally innocent. So if I end up with physical stress too, that’s bad.
Life expectancy to me is not so much about length of life as about quality of life. It’s only a healthy lifespan I’m interested in. This year my husband and I visited friends in Iceland, which is surely unusual – but we did it because one day it’ll be too late.
Expectancy is full living, enjoying things while we’re healthy, not just extending life regardless of its quality.
TG: And L, how about you? What are your thoughts on the subject?
Panelist L: Life expectancy to me implies life balance. How many years will I live? My parents went at 77 and 84. My mom had dementia. My older brother died at 70, of a rare form of dementia. My older sister is developing dementia at 76. Fortunately, no sign in my other siblings. But I don’t want to plan beyond 80. It’s not likely to be a healthy life. I don’t want to be in a retirement home. It’s only a healthy lifespan I’m interested in. Quality of life is all-important – without it life is nothing.
TG: I’m not sure what I expected. But my takeaways are these (and yours may be different). First, while the numbers are important, it’s the quality of life that really counts. Second, the numbers could be big ones – and the longer we live, the tougher things get, potentially, both financially and psychologically.
The question of our future health is a scary one, and there are few population statistics that tell us much about that issue if we’re currently healthy, as we’ll see in Stage L 03.
There are good statistics available on the length of life expectancy, so let’s understand what life expectancy means, and how you can get a starting idea of what yours might be – particularly as ignorance on that subject (as we’ll see in stage F43) generates the biggest fear that retirees have
I’ve found that many people have a vague idea about how long life expectancy is, and that typically they underestimate it. It’s an important subject, because if you’re going to plan to make your assets last a lifetime, you need to make some estimate about how long that lifetime may be.
This is an attempt to explain it. I know that it’s too simple for techies. I had no idea if it’s the right level or too complicated for the non-expert reader, but my website readers said that they found it useful.
In this stage I want to make four points about life expectancy.
The first is that most people misunderstand it. Even if they have heard that in some fictitious country life expectancy at birth is 80 years, they don’t understand that, at age 65, the average person there can expect to live another 15 years (in fact, probably more like 20 years). Really? How does the arithmetic work?
Here’s a way to understand it.
Suppose I were to ask you what’s the average of the numbers from 0 to 100. It’s not a trick question. It’s simple arithmetic. You know the answer: it’s 50.
Now suppose we leave out the lower numbers. What’s the average of the numbers from 40 to 100? Obviously, it’ll be higher; in fact, the average now rises to (yes, you know this answer too) 70.
It’s similar with life expectancy.
Suppose we encountered a peculiar population of 100 people in which one person dies before the first birthday, one dies between ages 1 and 2, and so on, the last one dying between 99 and 100.
What would be the average age at death? Again, not a trick question: it’s 50.
Now suppose we leave out all of those who die before age 40. What’s the average age at death of this smaller group? Again, not a trick question: it’s 70.
So, what does this tell us about the life expectancy of the population?
It tells us two things. First, at birth, if we don’t know which person we’re talking about, all we can talk about is the average, and for the average person, then, the life expectancy is 50 years. Second, if we consider only those who have survived until age 40, and again we don’t know which individual we’re talking about, their average age at death is 70. Their future life expectancy, once they’ve reached 40, is another 30 years, because that’s what “life expectancy” means: it’s the average number of future years to be lived by the average member of a well-defined group.
Notice that the people in the second group (those who have survived to age 40) are also members of the first group (the entire population). But the two groups are not the same, even though they contain some identical members. The second group excludes those who have already died before 40; that makes it a different group, and a longer-lived group. So, if we are to define life expectancy, it’s important to define the group we’re talking about very clearly.
OK, now let’s go back to the fictitious country in the first paragraph, and interpret those numbers.
What the numbers tell us is two things. First, if you include the entire population, the average age at death is expected to be 80. Second, if you exclude those who have already died before age 65, and include only those who survive past that age, their average age at death is of course higher than 80; for the fictitious country it’s 85. And that’s why the future life expectancy of someone in this country who has already survived to age 65 is a further 20 years, not the 15 years that people often misunderstand it to be.
Life expectancy tables vary by gender (typically, the life expectancy of a female is longer than that of a male), by country, by race – all kinds of factors, in addition to age.
OK, that’s my first point. Now to my second point. It comes from that example of the peculiar population of 100 people.
Remember that their average life expectancy at birth is 50 years. Now here’s a point that most people miss. Notice that 50 people (half the population) live longer than the average and half live shorter than the average.
Similarly, for the 60 of them who survive to age 40, 30 people (half of them) live longer than the new average life expectancy of 30 more years, and half live shorter than this average.
That’s what life expectancy means. It’s not the limit of life, it’s the average future life. And half of the people to whom it applies will outlive their life expectancy.
Talking about the average conceals the fact that, for any individual, the actual date of death is uncertain. If there’s something in your family history, or your personal health record, that makes you wonder if you may be longer-lived or shorter-lived than the average group of people who have survived to your age, see if you can find a way to estimate what this might mean for your future life expectancy.
Your doctor may be able to help. Also, there are websites with calculators that purport to help you make adjustments.
Whatever they tell you, however, it’s still not a prediction. For most people, until they’re near death, their specific future life expectancy is still reasonably unpredictable.
And so, here’s my third point: when you make financial plans about the future, it’s important to take this unpredictability into account.
My fourth and final point relates to the life expectancy of a couple. More specifically, this relates to how long before the second death of the couple. Techies call this the “joint and last survivor” life expectancy. It’s important because it’s necessary to provide for the longer-lived member of a couple, whichever one that may turn out to be.
Suppose there’s a couple whose individual future life expectancies, at some point in time, are roughly 15 years and 20 years. How long until the second death?
Most people say: well, after 15 years you expect one to die, and after 20 years the second one will die; so it’s 20 years to the second death, right?
It actually turns out to be a little more complicated than that. So I won’t go into the arithmetic, I’ll just try to give you a general explanation of why the expected time to the second death is longer than 20 years.
Consider the person with the longer (20 year) expectancy. Remember that 20 years is an average. For this person, it’s more or less a 50/50 chance of living longer than, or less than, 20 years. (That’s what the average means.) Now, what about the other partner in the couple? That partner has a 50/50 chance of living longer than 15 years. And a smaller chance than 50/50, obviously, of living longer than 20 years. A smaller chance, yes, but some chance nevertheless. And that some chance of living longer than 20 years, combined with the first partner’s 50% chance of living longer than 20 years, means that the couple together has more than a 50% chance of having one of them live more than 20 years.
And that’s why, for the couple together, the average expectancy to the second death is longer than the longer of the two individual life expectancies.
That’s a tough one to understand, first time round, I’ve found. So don’t worry if it still isn’t clear to you. For now you may simply prefer to accept it.
(One website reader with a sense of humor chose to deliberately misunderstand my point. He said: “Given that the average expectancy to the second death is longer than the longer of the two individual life expectancies, I have decided to go second.”)
Life expectancy is not only uncertain, it’s also typically underestimated, particularly for a couple.
STAGE L 02: LONGEVITY IS INCREASING
Where the route takes us
Longevity is just an estimated average, right? Yes. The strange thing is that the average itself keeps moving.
In Stage L 01 I explained what life expectancy means. In this stage we’ll look at some interesting numbers. In particular, I’ll stress the fact that, no matter what published number you see for life expectancy, the odds are that it’s out of date, and that actual life expectancy is a little bit higher than that number. In short, longevity (the expected length of life) is increasing.
It’s not a big deal. But I mention it simply to suggest that it might make sense to plan for a future lifetime that’s longer than your future expectancy.
I’ll start with a couple of quotes that I hope you’ll enjoy.
With my actuarial background, it’s not surprising that I’m interested in the advance of longevity. But I admit it took me aback when I saw this quote from Woody Allen: “I’m not seeking immortality through my work … I want to achieve it by not dying.” Lovely! It left me with a big grin on my face.
So too did this headline many years ago in the satirical weekly paper The Onion (and it would still hold true today, obviously): “World death rate holding steady at 100 percent.” The article includes a bar chart (a statistical diagram), showing a constant 100% every year. And the text included this gem: “Death, a metabolic affliction causing a total shutdown of all life functions, has long been considered humanity’s number one health concern. Responsible for 100% of all recorded fatalities worldwide, the condition has no cure.” It’s great when even a grim topic can generate clever humor.
Yes, death is certain. But we don’t know when it will occur for each of us, and that’s why averages are useful, because they give us a ballpark figure, a rough first estimate, to start with.
As a very rough and over-simplified version of human longevity’s history, I have the impression that human life expectancy at birth stayed in the range of 30-40 years from the earliest civilizations all the way through to the early 1800s. Since then it has been increasing, and a child born today can expect to live beyond age 80 in many countries. Mostly, that’s because death rates involving childbirth and violence have declined rapidly, and more recently modern medicine has dramatically reduced death rates at older ages. People are reaching old ages in greater numbers and in better health than ever before. And so the average lifespan has continued to increase.
While it’s still an unsolved mystery as to whether or not there is some way to extend the ultimate age beyond which nobody seems to survive (the “maximal lifespan” – which historically seems to be somewhere around age 120), I thought it would be interesting to give you a couple of facts and one unusual interpretive comparison.
The first fact is that life expectancy at birth has been increasing by roughly 3 months every year. If this continues, then 10 years from now today’s estimates of life expectancy at birth will be 2.5 years longer than they are today. (Fancifully – because it’s obviously not true – I fondly imagine that the last quarter of every year, from the beginning of October onwards, is like free time for me, because with the increase in longevity, it’s as if I have only aged 9 months!)
I ought to mention that every now and then there’s a year in which life expectancy doesn’t seem to increase, in some country. Who knows, that might just be a random fluctuation, or it might be the start of a trend in which life expectancy no longer increases, or even decreases. Only time will tell.
Another fact is the effect of how recent the improvement in longevity (since the early 1800s) has been.
The world’s population today is, in very round numbers, about 7 billion. The total number of people who have ever lived is about 107 billion. So today we have on earth about 7% of all the people who have ever lived. But how many people have lived beyond age 65? Today we have about 550 million people alive over age 65. What proportion is that of all the people who have ever lived and reached 65? Answer: perhaps one-third. That shows how rare it has been historically for people to reach 65.
The unusual interpretation comes from a conversation I had with Dr John Shoven at a conference held at Stanford University in 2012. I mentioned that our retirement systems aren’t terribly dynamic when it comes to defining what they call the “normal” retirement age; and since these formal structures influence so much of our thinking, we tend to think of people as becoming old when they reach retirement age – even though they may still be as vigorous and vibrant and healthy as people 20 years younger, a couple of generations ago. Dr Shoven had an even more vivid interpretation. He said that, based on a comparison of current longevity tables with those that prevailed generations ago (by the way, you’ll have noticed that at my age I prefer to call them “longevity tables,” rather than the more common “mortality tables”), people should be considered young until they reach the age at which their mortality rate reaches 1%, and middle-aged until it reaches 4%.
Now, I’m aware that, other than actuaries, few people know what a mortality rate is or where they would find one, so let me convert Shoven’s formula into something easier for the layman to understand. We’re young until the age at which our future life expectancy declines to 20 more years, and middle-aged until it declines to 10 more years.
And what does that imply? Based on a table published by the US Social Security Administration, and relevant for the year 2007, that means that American men should be considered young until they reach, in round numbers, roughly age 60 (65 for an American woman), and middle-aged until age 75 (80 for a woman). In many countries the ages would be even higher.
There! Isn’t that cheery news for you young people!
Seriously, there really is a “so what?” to add to these fun facts. And here’s what it is. Perhaps every five years or so, check the latest longevity statistics, via any friend who knows where to find stuff like this, and see if you want to change the time horizon for which you’re planning to make your money last.
Longevity has been increasing, almost every year. It might be useful for you to check the latest estimates every five years or so.
STAGE L 03: HEALTHY LIFE EXPECTANCY
Where the route takes us
It isn’t just how long we might live that’s of interest. It also matters a lot how healthy our future years are. We’ll see in this stage that health is a tough thing to measure.
Life expectancy measures a quantity of time. It says nothing about the quality of that time. We saw from the panel before Stage L 01 that people naturally think in terms of that quality, even before they think of a quantity of time. And therefore a natural question that arises is: how much of one’s future life expectancy is likely to be spent in good health? That’s what “healthy life expectancy” measures.
You can guess one thing about such a measure before you ever look at the actual numbers. There’s not a lot of ambiguity about whether someone is alive or dead. In contrast, there’s a lot of ambiguity about whether someone is healthy or not, and if not, how unhealthy that person is. And so it’s inevitable that measures of healthy life expectancy are less reliable than measures of life expectancy; that measures will vary depending on how health is perceived; and that from time to time there will be changes in the classification of health perception and therefore in the measures themselves even when nothing has actually changed. In short, we know there’ll be some subjectivity rather than pure objectivity in the numbers.
Nevertheless, it’s interesting to at least get some order of magnitude for healthy life expectancies, to compare them with life expectancies, even if we should interpret the numbers in a sort of touchy-feely way.
Of all the many papers on the subject, there’s an early one by Michael Wolfson, then the Director General of Statistics Canada, that sets out the concepts (and one particular approach to measurement) very well, so I’ll rely on it for my explanation here.
A sample survey of Canada’s population was taken for two purposes. One was to establish the prevalence of six broad kinds of ill health: sensory problems (vision, hearing and speech); mobility; emotional state; thinking and memory; dexterity; and level of pain and discomfort. The other was to assign a score to how bad the ill health was, by asking the respondents to rank preferences for various health conditions (yes, very subjective). From that a proportion and a degree of ill health were estimated for each age and gender, and then these numbers were combined with life expectancy tables.
For example, if for a particular combination of age and gender 30% of the population had some form of ill health with an average severity of 40%, the aggregate ill health is as if the product of the two numbers (meaning 12%) are totally ill and the remainder (meaning 88%) are perfectly healthy; so this group is assessed as having spent 88% of its year in good health (and therefore 12% of its year in a state of perfectly ill health, or total disability). And so on. Eventually it becomes possible to estimate how many years of perfectly good health and how many years of perfectly ill health are contained in the group’s life expectancy.
Many other approaches are feasible. For example, in 2001 the UK census for the first time asked the question: “Over the last 12 months would you say your health has on the whole been: Good? Fairly good? Not good?” And “good” and “fairly good” are counted as healthy and “not good” as unhealthy. Or there’s a “global burden of disease” study to estimate severity-adjusted prevalence by age and gender. For example, dementia is assessed as 66.6% disabled, AIDS 54.7%, low back pain 32.2%, blindness 19.5%. That was in 2010. But 6 years earlier, low back pain was 6.1% and blindness 59.4%. Assessments can change substantially.
You get the idea that these numbers should not be treated as gospel.
Now for some results, and then some helpful overall conclusions from Wolfson.
Table L 03.1 shows a small and arbitrary selection of countries for which The World Health Organization published these international comparisons in 2016, for males and females combined, for life expectancy and healthy life expectancy at birth. And Table L 03.2 shows the corresponding numbers for those who survive to age 60.
Table L 03.1: Comparison of life expectancy and healthy life expectancy at birth
Country LE at birth (yrs) HALE at birth (yrs) Ratio HALE/LE
Japan 83.7 74.9 89%
Australia 82.8 71.9 87%
France 82.4 72.0 87%
Canada 82.2 72.3 88%
Netherlands 81.9 72.2 88%
United Kingdom 81.2 71.4 88%
Germany 81.0 71.3 88%
USA 79.3 69.1 87%
Poland 77.5 68.7 89%
China 76.1 68.5 90%
Russian Federation 70.5 63.3 90%
India 68.3 59.5 87%
Table L 03.2: Comparison of life expectancy and healthy life expectancy at age 60
Country LE at 60 (yrs) HALE at 60 (yrs) Ratio HALE/LE
Japan 26.1 21.1 81%
Australia 25.5 19.6 77%
France 25.7 20.3 79%
Canada 25.0 19.7 79%
Netherlands 24.2 19.3 80%
United Kingdom 24.1 18.8 78%
Germany 23.7 18.6 78%
USA 23.6 18.1 77%
Poland 21.8 17.0 78%
China 19.7 15.9 81%
Russian Federation 18.6 15.1 81%
India 17.9 13.3 74%
Notice that all the healthy life expectancies at birth are almost 90% of the respective life expectancies. And since young and middle-aged years are typically much healthier than post-work years, much of the gap between the total and healthy life expectancies must occur after age 60. In other words, the ratio of healthy to total remaining life expectancy at 60 must be noticeably lower than 90%.
Sure enough, the ratio at age 60 has dropped to around 80%.
Remember that the measuring convention for HALE is to split the count between those who are notionally completely healthy and those who are notionally completely disabled. So the 80% is not as clean as saying that we can expect 80% of our future lifespan at 60 to be completely free from disability, unless we also accept that the other 20% will be spent completely disabled. In practice, therefore, what this means is that we are likely, on average, to spend something less than 80% completely healthy and therefore something more than 20% at least partially disabled.
I searched extensively for statistics that I would have found insightful, without success. For example, in any given country, for each gender/age combination, show me the prevalence of disability, that is, the proportion of the population of that gender and age with some form of disability. Show me also, for those without disability at each gender/age combination, the incidence of disability, that is, what proportion of the healthy change from healthy to disabled at that age. If I’m healthy, this would give me some initial idea of how likely it is that I will develop some disability in the next year. Some prevalence statistics exist. But incidence statistics were beyond my ability to locate.
On the whole, while the actual numbers are indications at best, I found Wolfson’s four overall conclusions very helpful in interpreting them.
- First, he says that the societal burden of ill health is higher for women than for men. “Since the prevalence of chronic conditions increases with age and women live longer, they spend a longer period with chronic conditions. Also, at age 65 and older, women tend to be in notably poorer health than men the same age.”
- Second, the burden is highest among those in early old age, not among the most elderly. By this he means that there are so many more people in early old age that their ill health aggregates to a more serious problem for society than the fewer (though less healthy) really elderly. Note that that’s a societal rather than an individual burden he’s talking about. From an individual perspective, I have no doubt the burden increases with age.
- Third, sensory problems and pain comprise the largest components of the burden of ill health. (And that was before low back pain was assessed more severely by the WHO.)
- Finally, higher socioeconomic status confers a dual advantage, longer life expectancy as well as a lower burden of ill health.
I’m not sure how to interpret all this. But it certainly adds to the relevance of the issues discussed in Stage H 81, on aging with dignity.
In most countries, the average person can expect to spend something up to 80% of our expected future lifespan after age 60 in reasonably good health.
STAGE L 11: ONE PARTICULAR LONGEVITY TABLE
Where the route takes us
There’s a particularly useful table available online. Here I’ll show you how to use it.
In my first draft of this book, I planned a stage telling you about longevity tables. Since then I discovered a very useful one, at www.longevityillustrator.org, and so I’ll use it as the standard one I refer to, and I’ll explain it in this stage.
But I also recognize that it may not apply to many readers. So in Stage L 12 I’ll explain how to adapt it to fit your own circumstances. In that way it should be useful to everyone.
The simplest and clearest way to start is to quote from the Welcome message on the website:
“Developed by the American Academy of Actuaries and the Society of Actuaries [in 2016], the Longevity Illustrator is designed to provide you with perspectives on your longevity risk—the uncertainty of how long you and your spouse/partner might live. It does not address your finances, your investments, your earning potential or your anticipated expenses; consult with a financial professional about those aspects of your retirement planning. We invite you to use the Longevity Illustrator to enhance your understanding of the potential risk for outliving your financial resources.”
Beautifully expressed, and exactly what we need here.
“The projections shown in the Actuaries Longevity Illustrator are based on mortality tables used by the Social Security Administration in the annual Trustees’ Report. There are separate rates for males and females.
“Because mortality rates show a long-standing trend of improving, the Longevity Illustrator is based on the assumption they will continue to improve according to the MP-2015 rates published by the Society of Actuaries. These improvement rates are applied to the 2010 Social Security mortality tables to project mortality rates to future years.
“Additional adjustments are made for health status ranging from 80% to 125% depending on age and for smoker status ranging from 77% to 211% depending on age…
“There are many factors that affect longevity. These include: income, family history, geography, life style, occupation, current medical conditions, and ethnicity. While these all affect longevity, the four factors (age, gender, smoking and health) chosen for this calculator have been shown to account for a significant amount of the individual variations in longevity.”
As you’ll find, the website asks you a (very) few questions about yourself, so that you can get the output you seek.
And then you see the projections. You can print them, change the inputs, and do other useful things. (You can see why I like the website so much.)
The numbers I’m going to refer to are those called “Planning Horizon” – those are the ones I believe are the most useful, though of course the other numbers can be fascinating in their own right. You can get numbers for yourself alone or for you and your partner. I’ll assume you’re using the ones for you and your partner. A nice feature is that you can be a male/female partnership, or male/male or female/female – that saves me having to explain how to adapt the projections for same-sex couples, as my original book draft did.
Let’s start with the 50% numbers. These show four things:
- Your life expectancy, in years. Remember that in Stage L 01 we said that half the population outlives their life expectancy. So you have a 50/50 chance of living at least this number of years.
- Your partner’s life expectancy. Same comment.
- Your “joint and last survivor” (or JLS, for short) expectancy, as explained in Stage L 01. That’s the average length of time until the second “estate event” (a term some American insurance agents use, as a euphemism for – how should I say this? – passing away). So this is the best estimate of the length of time that at least one of you is likely to survive.
- Your “joint” life expectancy, which is the period for which both of you should survive together.
Notice that the numbers are round numbers – no fractions, no decimals. I like this feature. Of course it’s possible for techies to go back to the actuaries’ original tables and calculate the outputs to as many decimals as one likes. But frankly, who cares? These are all rough estimates, for initial planning purposes, and the decimals are irrelevant.
I produced one set of numbers (the inputs don’t matter), for which these numbers turned out to be 13, 15, 18 and 9. Let’s interpret them.
13 and 15 are the individual life expectancies of the two partners. Their JLS expectancy is 18 years, meaning that that’s the best estimate of how long at least one partner of the pair will survive. Their joint expectancy is 9 years, meaning the best estimate is that both will be alive together for 9 years. Put the last two estimates together, and it means that the best estimate is that they’ll be alive together for the first 9 years, and then the survivor will be alive for a further 9 years.
By the way, the Actuaries website allows you to click on the following question: “Why is the number of years that either one or both of us will live greater than the number of years one of us will live?”
This is the same question I attempted to answer briefly in Stage L 01, in my fourth point there. The Actuaries website goes into some of the numbers. If you like playing numbers and probabilities, you might enjoy looking at them!
I’m going to ignore the 90% and 75% numbers, which I suspect are for curiosity rather than for planning purposes. Let’s go to the 25% and 10% numbers.
You’ll remember the concept from Stage L 01. We start by considering a large group of people of the same age and gender. As we follow them through life, the number remaining will gradually decline. At some point (the 50% numbers above), only half of the original group will be left. Some time later, only one-quarter of the original group will be left. Later still, only one in every 10 original members will be left. And so on.
The 25% and 10% numbers show the points in time at which 25% and 10% will be left, respectively.
Now you understand and know how to find the 50%, 25% and 10% longevity estimates for you and your partner.
STAGE L 12: WHAT IF YOU HAVE A BETTER ESTIMATE OF YOUR OWN LONGEVITY?
Where the route takes us
If you have an independent way of getting your own longevity estimate, you can adapt that online table to fit your own circumstances.
In Stage L11 I quoted from the Actuaries website, with their explanation of the many factors that affect longevity. Remember? “These include: income, family history, geography, life style, occupation, current medical conditions, and ethnicity.”
What if, for example, you’re a reader in a country other than US? You might be familiar with, or prefer to use, a longevity table published there. Or you may like a particular app on the internet that asks you lifestyle and health questions that get to the level of detail you prefer. Or perhaps, if you have a medical condition, your doctor has given you a personal life expectancy estimate. Whatever the reason, you have your own independent starting point, your own 50% life expectancy estimate, and it’s different from the Actuaries 50% estimate.
Now what? Where do you get JLS estimates, 25% and 10% estimates, and whatever other longevity estimates you wish you had for planning purposes?
The trick is to tie together your personal independent estimate and the estimates contained in the Actuaries table. And it turns out to be quite simple to make that adjustment.
In essence, you take your own independent estimate, and fit it to the Actuaries estimates. By that I mean that you see what age your estimate corresponds to in the Actuaries estimate. Then continue as if you are an American of that “fitted” age.
To save you some tedious work, I compiled a list of 50% life expectancies from the Actuaries website, for a non-smoker in average health. They’re shown in Table L 12.1.
Table L 12.1 Some life expectancy numbers from the Actuaries
Let me give you a couple of examples to show you how to use it.
Example 1: You’re a male and you have an independent life expectancy estimate of 14 years. It’s as if you are a male American non-smoker in average health aged 72 or 73.
Which age should you use, 72 or 73?
It really doesn’t matter. Remember that the numbers generated by this approach aren’t predictions, they’re approximations for you to use for planning purposes. Pick either 72 or 73, and the resulting numbers are close enough for planning purposes.
Example 2: You’re a female and you have an independent life expectancy estimate of 12.5 years. It’s as if you are a female American non-smoker in average health aged 76, 77 or 78.
Again, it doesn’t matter, for planning purposes, which of those ages you use to get the Actuaries 50%, 25% and 10% longevity estimates.
All you need to do is find the age at which your independent longevity estimate matches one in the table.
STAGE L 21: HOW LONG SHOULD YOU PLAN TO MAKE YOUR MONEY LAST?
Where the route takes us
Since you don’t know how long you’ll live, what is a sensible planning horizon for the length of your retirement?
As you plan for the future, there’s a fundamental question. How long do your savings need to last, as you draw down from them for spending? The answer is simple, but not straightforward. It’s simple, in the sense that it needs to last as long as you live. But that’s not terribly helpful, because (assuming your health is at least average) you have no idea how long that is.
So now, what do you do? Because if you don’t know, then you also have no idea how much you can afford to spend each year from your savings. (This amount, by the way, is often called your “sustainable drawdown,” in the jargon used by financial professionals. You’re drawing down an amount from your saved assets, and that level of drawdown should be capable of being sustained for as long as necessary, so you can go on living the lifestyle implied by that rate of spending.)
If you want the best safety that money can buy, you’ll seriously consider buying a lifetime income annuity (see Stages F 61 and F 62).
If you aren’t going down that path, then you’re self-insuring, and I’ll assume that you want a safety margin. That’s where the relevant life expectancies (the 50%, 25% and 10% points) explained in Stage L 11 become relevant.
Suppose you choose the halfway (50%) point as your planning horizon. You may live longer or shorter than that halfway point. They’re equally likely. So, if that’s your planning horizon, you’re as likely to fail to make your money last, as to succeed.
For most people, that’s too chancy. Running out of money before you run out of life turns out to be the biggest fear that retirees have. So you’ll probably want to make the money last longer than the halfway point.
How about the point at which three-quarters of a group like you have gone, and only one quarter still alive? Well, now you’re giving yourself three chances out of four, of succeeding to make the money last. That’s better. Or, if you’re very cautious, you’ll choose the point at which only one in 10 of a group like you are likely to be alive.
Wouldn’t you naturally want to be as cautious as possible, with something like this? Not necessarily. Why is that? It’s because the longer you plan to make your money last, the less you can draw down each year. That’s just common sense, not higher mathematics. If you have $1,000 and want it to last for 20 years, you’ll be able to withdraw less than if you want the same $1,000 to last for only 10 years.
So there’s a trade-off. What you might do, in practice, is two things.
The first is to ask someone to calculate (or maybe do it yourself) how much you can sustainably draw down from each $1,000 if you plan for the one-quarter or one-tenth points. See how that compares with your desired lifestyle. That could help you decide.
The other is to resolve to adjust, as time advances. For example, this is what my wife and I are doing. We’re using the one-quarter (25%) JLS point for planning purposes.
When the first of us goes, that’ll require a new calculation, with only the future expectancy of the survivor then being relevant.
If, as time goes by, we’re both still around and in good shape, we’ll re-examine our future JLS expectancy at some point, and adjust our planning horizon accordingly. Our hope is that by then we’ll have gone from the so-called “go-go” years to the “slow-go” years, and won’t need as much money to support our slow-go lifestyle.
Or we’ll buy a lifetime income annuity with the balance of our assets, to ensure that we don’t outlive our assets.
Depending on how safe you want your plan to be, you might buy a lifetime income annuity, or use the 25% or 10% longevity point as your initial planning horizon, with a potential reassessment later.
STAGE L 31: WHAT LONGEVITY INSURANCE IS – AND ISN’T
Where the route takes us
Just as we consider buying insurance if an early death hurts our family financially, so too we might consider buying insurance against living far longer than our money can support.
Wikipedia says: “Insurance is a means of protection from financial loss.” A good starting point. It is particularly appropriate when the contingency giving rise to the loss has a low probability, but if it occurs, its financial impact is large. In those circumstances, a relatively small premium secures the insurance, and the financial risk has been hedged via the insurance company’s guarantee of payment.
Most of the time, we don’t want the contingency to occur, for example a car accident or a house fire. And if it doesn’t occur we lose the insurance premium, but it’s a small price to pay for the financial security and the peace of mind that the premium secures for us.
Term insurance is similar. We don’t buy it just for the sake of it. We buy it in the circumstances cited above. For example, in the case of a young family, if the breadwinner dies early (unlikely though the event is), the future work earnings that the family counts on will be lost, and by taking out term insurance covering the breadwinner’s death before retirement, that financial loss can be hedged. Again, we’re happy not to collect.
There is, however, one form of insurance which is odd, in the sense that we actually do want to collect on it. It’s longevity insurance. Here’s how it works.
How long we’ll live is unknown. Most people’s biggest fear after retirement, according to consistent surveys, is outliving our assets by living “too long” – I put that in quotes because typically we actually really do want to live a long time. It’s just that the financial cost of supporting a long life could be very great, far more than we can afford. If we think of life well beyond our average life expectancy, we have the two conditions for which insurance is a logical solution: low probability, but high cost if it occurs.
The relevant insurance is called “longevity insurance.” We pay a premium. If we live beyond the date we select, the insurance company pays us an income for as long thereafter as we live. If the contingency doesn’t occur (we don’t reach the selected date) we lose the premium.
That’s longevity insurance.
Now let’s start to think about what isn’t longevity insurance.
One ambiguous feature is what is meant by saying that the event is unlikely to occur. With a car accident or a house fire, it’s reasonably clear whether or not the event has occurred. But not with longevity. What exactly is living “too long”? What should be the selected date after which one is deemed to have lived “too long”?
It really varies from one person to another. It depends on how long we can afford to sustain the lifestyle we’d like, given the assets and sources of post-retirement income that we have. (Stages F 21 through F 28, in Route 4, expand on this notion.)
But in general, to really be thought of as insurance comparable with car and fire and term insurance, we can say that the selected date for longevity insurance should be one that we have only a small chance of living beyond – even though “small” isn’t a precisely defined term. For example, if we choose the 10% age in Walk 13 in Life Two, then we have a 10% chance of living beyond that age. That’s a higher probability than collecting on car and fire insurance, but in fact that 10% age is far older than the insurance offered in practice.
For example, for a 65-year-old non-smoking American male in average health in 2019, that 10% “living beyond” age would be 97. Yes, that feels like insurance: it pays off only if we live past age 97. But typically the “living beyond” age offered by an insurance company is 85, which is roughly equal to a 65-year-old male’s average future life expectancy.
So the longevity insurance offered in practice gives us a very high chance of collecting on it, maybe 50/50. And therefore it feels as much like an investment as insurance. And therefore the notion of not collecting makes it feel like money thrown away – we feel we’ve been cheated.
This is interesting! If we use age 97, the premium is relatively small, because there’s a relatively small (10%) chance that the insurance company will have to pay us. So, as with other small insurance premiums, we may not mind if we don’t collect. But if we use age 85, the premium is large, because there’s a high chance (roughly 50%) that the insurance company will have to pay us. We may not mind losing a small premium. We definitely do mind losing a large premium!
And so, because 85 has (unfortunately) become a sort of standard age for longevity insurance, insurance companies cater to our psychology. They add an element to what I’ve described as longevity insurance, promising us not only to pay as promised if we survive past the selected date, but also to make a payment if we don’t survive to that date. That don’t-survive payment is sometimes equal to the premium paid, and sometimes it’s the premium plus interest from the date the premium was paid.
That is a wonderful addition psychologically. In fact, I understand that focus groups suggest that most people won’t buy longevity insurance unless it has that return-of-premium feature. But I just want to point out that bundling a return-of-premium feature (with or without interest) makes it more than pure longevity insurance, and can be unnecessarily costly.
Let me demonstrate that.
Think of that hypothetical 65-year-old male friend who wants pure longevity insurance that kicks in if he survives past age 85. He gets a quote from an insurance company, telling him how many dollars a year they’ll pay him once he survives to 85, for an initial premium of $100. That’s the end of the matter. We’ll call it Example A. (It doesn’t matter what the quoted income payment is.)
Now let’s consider a variation. This is Example B. The friend says: “How much will I get after 85, if in addition you also guarantee to refund my $100 if I die before 85?”
Aha! What our friend has asked for, in addition to the longevity insurance, is a 20-year term insurance policy that pays $100 if he dies before 85, and nothing if he survives to 85. As I said, that’s in addition to the pure longevity insurance, which kicks in on survival to age 85.
Very roughly (again, the exact amount doesn’t matter), the cost of that term insurance policy is the same as the cost of the pure longevity insurance.
That means that, of the aggregate $100 that your friend is willing to pay, roughly half will go for the term policy, and the remaining half will buy him pure longevity insurance. Whatever quote he received in Example A, he will now be quoted an income after age 85 of roughly half of that amount, in Example B.
For the psychological satisfaction of “I can’t lose, whether I live or die,” he has sacrificed half of his longevity insurance coverage.
For whose benefit? For whoever benefits from his estate. His personal retirement prospects after 85 have been greatly (50%) sacrificed, to create a bequest.
That may be what he wants, but then he is admitting that he has two financial goals, a bequest as well as longevity insurance, and the bequest goal is significant enough that he’ll significantly (50%) compromise his own prospects after age 85. (Not something that most of us want to do, according to surveys.)
Note also that it’s an odd use of term insurance. I explained earlier the standard use of term insurance, to make good a loss of earnings on early death while in the workforce. Term insurance after retirement is much more difficult to justify.
Or, of course, this may not be what your friend really wants. All of this may be far too complex for him. He may not realise that he has been sold two bundled contracts, because nobody ever told him you could split them apart and buy just the one you want. So he now has an unwanted and unnecessary and costly 20-year term insurance policy. Too bad, because it’s not an informed choice.
OK, finally, Example C.
Here your friend wants pure longevity insurance, plus a return of premium with interest if he dies before 85.
This time, what is he actually doing? He’s not actually buying longevity insurance at all, right now. What he has done is deposited $100 with the insurance company, and it has a great similarity to a locked-in bank deposit earning interest. If he dies before 85, his estate will get the deposit plus accrued interest. If he survives to 85, the accrued amount will, in effect, buy him an immediate lifetime income annuity at that point.
Why would he want to do that? After all, he’s locking in the money, thus losing the flexibility to use it for other purposes before age 85, and it’s invested purely in a fixed-rate deposit account.
He could just as easily invest the $100 in his own way, whether in a bank deposit or in fixed income or in growth-seeking assets – in other words, with total flexibility and control – and approach the insurance company if he survives to 85 and buy his lifetime income then. In fact, for the first 20 years there is no insurance element at all. Not unless your friend wants a pure fixed-income investment and believes that the insurance company is best qualified to deliver it, would this be a sensible choice.
Insurance companies in most countries typically don’t offer Example A. Typically they only offer Examples B and C. They call B and C longevity insurance, even though the pure longevity insurance element (which is A) may be much smaller than perceived, or non-existent.
It’s not as if insurance companies don’t underwrite A. They do, routinely, because both B and C include A.
People typically don’t understand what they’re buying, as this is not only a complex field, but one that we don’t want to think about.
If people are offered a choice between Examples A, B and C, and reject A for psychological or other reasons, fine. That’s an informed choice.
But surely it makes sense to offer Example A (pure longevity insurance) with a high age beyond which payments commence.
Otherwise it’s not an informed choice, it’s “Take this bundle or nothing at all.”
Longevity insurance is a way to hedge the financial impact of survival to an unexpectedly high age. It ought to kick in at an age that we’re very unlikely to live past. But typically it’s only available from a much earlier age, in a bundle with other contracts that may be unnecessary, expensive or involve no insurance at all.
 For the techies, I acknowledge that I’m not being strictly accurate here, since 50/50 relates to the median, not to the mean, which is what life expectancy is.
 Lax (1975).
 January 22, 1997.
 Blagosklonny (2010), Oeppen et al (2002), IHME (2015).
 See www.prb.org/Publications/Articles/2002/HowManyPeopleHaveEverLivedonEarth.aspx. Retrieved on May 23, 2016.
 See poetryofphysics.blogspot.ca, December 12, 2011. Retrieved on May 23, 2016.
 He is the Wallace R. Hawley Director of the Stanford Institute for Economic Policy Research and the Charles R. Schwab Professor of Economics at Stanford.
 It goes by other names too, such as “health-adjusted life expectancy (HALE)” or “healthy life years” or “disability-adjusted life years”.
 Wolfson (1996).
 Cited in wikipedia, “World Health Statistics 2016: Monitoring health for the SDGs Annex B: tables of health statistics by country, WHO region and globally”. World Health Organization. 2016. Retrieved 27 June 2016.
 As at March 17, 2018.
 Also as at March 17, 2018.
 Compiled on March 16, 2018
 OK, for the techies, the purchase will be at a guaranteed conversion rate.