Translated into simple english, this means that if we know the actual death rates for people of all ages in a given time period for a subgroup of the population (e.g., a country or county), we can estimate the expected remaining duration of life for an average person in that population at any age. The metrics of life expectancy and survival probabilities are drawn from what’s known as a life table.
The Origin of The Life Table
Life tables were first invented by John Graunt and Edmond Halley in the 17th century. In the first edition of Graunt’s famous publication ‘Natural and Political Observations mentioned in a following Index and made upon the Bills of Mortality’, Graunt included a table showing how many survivors there were out of 100 people beginning at age 0 and every 10 years beginning at age 6 through age 80. While Graunt didn’t record ages at death, it was possible to assemble a very crude life table from these numbers. Life tables today reflect the complete mortality experience of people at all ages where they show the complete population alive at the beginning of the time period, the population at risk of dying, and the observed total number of deaths. From this it becomes possible to calculate death rates and survival probabilities. Both sets of numbers are provided by Lapetus in our final reports.
What is the Difference Between Period Life Expectancy and Cohort Life Expectancy?
When we estimate life expectancy for the U.S. population today, you might wonder how we know how long these folks are likely to live – given that they’re still alive. We do this by assuming that the death rates that prevail today will hold constant for the entire duration of life of the cohort. This is called a Period Life Table – and it is the basis for almost all life expectancy reports that appear in the media and in publications.
The reported ‘period’ life expectancy for a 70-year old man in the U.S. in the year 2020 (the latest published data) was 13.74 years or 165 months. This assumes death rates at ages 70+ observed in 2020 do not change for the remaining life of men currently aged 70. If death rates decline at ages 70+ in the future, then the life expectancy provided will underestimate LE; if death rates increase at ages 70+ in the future, then the period life expectancy provided will overestimate LE. As you might imagine, period estimates of life expectancy in the U.S. made during the Covid pandemic are likely to underestimate survival in the future (e.g., people will live longer than the tables suggest).
To address this caveat in the period estimate of LE, scientists have created a Cohort Life Table, which in one version is basically the same life table as before, but with assumptions included about how much death rates observed today will change in the future. This is where it gets tricky because now assumptions enter into the equation, and we should all know by now that assumptions are the mother of all “something or other”. I am personally not a fan of using cohort life tables because I just don’t trust the assumptions that have been used or are being used to anticipate future changes in death rates.
However, historical cohort life tables can also be created and they can be quite interesting. For example, a cohort life table for babies born in 1900 reveals how long these children actually lived compared with how long they were projected to live based on death rates present in 1900 (because everyone born in 1900 is already dead). As it turns out, historical cohort life tables have no practical use in the world of life settlements.
Has Life Expectancy Increased Linearly Since 1900?
Contrary to the view held by some, the answer is a definitive NO! In the 20th century in the United States, life expectancy at birth actually declined during 24 of those 100 years for men and in 18 of those 100 years for women. Overall, the 20th century led to the most dramatic increase in life expectancy ever experienced by humans, but it was most definitely not a linear increase. Some of my fellow scientists in aging have used linear extrapolation from past trends to conclude that period life expectancy at birth will rise to 100 in this century in the United States, but as I’ve demonstrated, this conclusion is based more on wishful thinking rather than solid science.
Is Life Expectancy a Gauge of The Health Status of A Population?
If you look up the definition of life expectancy, you’ll often find the conclusion that this metric is a gauge of the health status of the population. I cringe when I read this. Sure, if you think health is defined 100% by when death occurs, then you’re onto something. And yes, when life expectancy goes up or down, you’re getting a very generic signal from vital statistics that the risk of death is going up or down. But you won’t know why things are changing, you would need to look at specific causes of death to determine what specifically is leading to the change; and most important of all, you would have no clue whatsoever about the health status of people still alive. So no, life expectancy is an extremely crude and very poor metric when it comes to gauging health status.
Having said that, when a report from an LE provider like Lapetus indicates that the expected duration of life is well below average for a person of a given age, that does signal a health challenge is present. The direction and magnitude of projected variation in LE from average should always be defended in LE reports by specific reference to the medical and scientific literature.
Life Expectancy vs Lifespan
Technically, the term ‘life expectancy’ should only be applied to populations since the metric itself is drawn from a population-based life table. The term ‘lifespan’ refers to the observed duration of life of an individual and would actually be more appropriate in the context of what LE providers do (LS providers would actually make more sense). I’m not sure why or how ‘life expectancy’ became the term of art in the life settlement industry, but as long as we all interpret its meaning in the exact same way, I guess it doesn’t matter.