Are we over-stating COVID-19 mortality?

Data and Metrics Data Literacy

A: No. It’s bad.

➡️ The Covid-19 pandemic has resulted in the largest loss of life expectancy seen in many countries since World War II.

⬇️ U.S. males saw some of the worst losses—2.2 years of lost life expectancy in 2020. The drop was 1.65 years for U.S. females.

We’ve talked about “excess mortality” in previous posts– the idea that regardless of whether people die “with” or “of” COVID, we can tally the number of people dying ABOVE AND BEYOND what would be expected from previous years.

By this estimate, actual COVID deaths in the US may be 30% HIGHER than official COVID-19 deaths. (More here)

*Life expectancy* is another measure we use to summarize the mortality patterns of different countries over time. You’ve probably heard this statistic in the news your whole life—”life expectancy at birth for American males is 79 years,” for example. Like excess mortality, life expectancy is calculated using deaths from ALL causes, so it doesn’t depend on the accuracy of recording COVID deaths.

❓But what does life expectancy mean? The number of years someone born today can expect to live….right?

Almost. Hold on for some nerdy demography 🤓:

💥Life expectancy is the number of years someone born today could expect to live… IF THEY WERE TO LIVE THROUGH THE AGE-SPECIFIC DEATH RATES OF *THIS* YEAR.💥

✅ Life expectancy is thus a snapshot of *current* mortality conditions.

❌ It’s not a forecast of the actual experience of newborns.

You read that right–life expectancy is calculated based on the pretend lives of an imaginary (or “synthetic”) group of people living through the risk of mortality at each age that exists today, without any improvements (or declines).

While this may seem counter to how life expectancy is commonly perceived, demographers nonetheless find this a very useful summary measure of population mortality.

❓WHY is life expectancy useful if it’s not actually how long I can expect to live?!

❓Why not JUST use excess mortality to summarize the COVID death toll?

❓COVID isn’t going to happen *every* year (though 2021 not looking so good TBH…) so isn’t this “fall” in life expectancy deceiving?

Excellent questions!

Life expectancy has some advantages over excess mortality or raw mortality rates. The total number of deaths depends a lot on population size, as well how many young versus old people there are in a country. This makes it more difficult to compare across countries or over time.

In addition, we might care more about deaths at younger ages when someone has more remaining years to live. Life expectancy accounts for all these issues, allowing direct comparisons across countries of different sizes and age structures, and for the same country over time. Deaths that happen at younger ages contribute more to losses in life expectancy than deaths that happen at older ages. This is why countries that reduce infant and child mortality see rapid gains in life expectancy.

Nerdy Girl Jenn Dowd and her colleagues at the Leverhulme Centre for Demographic Science, Oxford University just published the first estimates of 2020 life expectancy for 29 countries for whom complete data was available.

Key takeaways:

➡️ Life expectancy fell in 2020 for 27 out of 29 countries, and the losses in the U.S were among the worst.
➡️ Men suffered larger life expectancy declines than women across most countries.
➡️ Losses this large had not been seen since World War II for Western European countries such as Spain, England and Wales, Italy and Belgium.
➡️ The size of the fall in US life expectancy is unprecedented going back the Great Depression.
➡️ In the US, increases in mortality in the under-60 age group contributed most to life expectancy declines, whereas across most of Europe increases in mortality above age 60 contributed more.


Life expectancy, despite the name, does not predict the future. It provides a useful snapshot of *current* mortality in a population.

This helps us quantify the losses due to COVID relative to other mortality “shocks” – in this case losses not seen since World War II. Even if these losses are transient (they won’t happen every year🤞), life expectancy helps us compare shocks across time and place.

By any measure, these losses are real and devastating. And as we often remind ourselves, these deaths are not just numbers. Each one represents a friend, a co-worker, a family member.

COVID-19 is now a vaccine preventable illness, so let’s all do our part to limit the impact COVID has on life expectancy in 2022 and beyond. #VaxUp

Those Nerdy Girls

Further Reading:


New life expectancy paper by Jenn Dowd & colleagues

“Covid-19: Life expectancy is down but what does this mean?”

“Covid has wiped out years of progress on life expectancy, finds study”

“Covid-19 isn’t the reason that US life expectancy is stagnating”

Twitter thread primer on life expectancy by demographer Dr. Ilya Kashnitsky

Link to Original FB Post