The decade that vanished in two years

The decade that vanished in two years

For the first time in over half a century, the world went backwards. Global life expectancy at birth fell from its all-time peak of 73.12 years in 2019 to 71.37 by 2021 — a drop of 1.75 years, with the slide accelerating in the second year. The 2021 figure is the level the world last passed around 2012. Roughly nine years of progress, gone in two.

What makes the reversal so jarring is the backdrop it interrupted. From 2000 to 2019, global life expectancy climbed 6.35 years, from 66.77 to 73.12, without a single down year anywhere in that 19-year stretch. The line only ever went up — until it didn't.

One caution before going further: this is period life expectancy, a snapshot of a single year's mortality conditions, not a forecast of how long a baby born in 2021 will actually live. The 2021 number does not predict a shorter life for today's newborns; it measures how deadly the world was in that one pandemic year. That distinction is what makes the whole story possible to tell honestly.

A long climb, powered by the poorest

To understand what COVID took away, you first have to understand whose progress it was. The pre-pandemic rise was not a story of rich countries getting marginally richer in years. It was a story of catch-up.

The fifteen largest life-expectancy gains from 2000 to 2019 were all sub-Saharan African. Rwanda gained 20.9 years, climbing from 46.9 to 67.8; Burundi added 20.0, Malawi 18.6, Ethiopia 18.1. These reflect recovery from catastrophe — the aftermath of conflict, falling child mortality, and above all the rollout of antiretroviral therapy after HIV/AIDS had gutted life expectancy across the region.

At the regional level the same pattern holds: WHO's Africa region surged 11.2 years over 2000–2019, more than double the gain of Europe (5.7) or the Western Pacific (5.5). The world's average rose because its floor rose far faster than its ceiling.

The income gradient tells it most cleanly. The gap between high- and low-income economies shrank from 23.92 years in 2000 to 16.69 in 2019, as low-income economies gained 10.6 years while high-income ones gained just 3.4. Progress was real, and it was disproportionately the progress of the poor.

One number, thirty-three years apart

However fast the floor was rising, it was rising from very far below. In 2021 the distance between the longest- and shortest-lived countries was 32.98 years — Japan at 84.46, Lesotho at 51.48. The "world average" of 71.37 is the midpoint of a chasm, not a description of any real place.

The top of the table is a cluster of high-income East Asian and Western European states: Japan first, then Singapore (83.86), Republic of Korea (83.80), Switzerland (83.33), Australia (83.10). The bottom is almost entirely sub-Saharan African and largely HIV-scarred: Lesotho last, just below the Central African Republic (52.31), Somalia (53.95), and Eswatini (54.59).

But a low number does not mean what it sounds like. Life expectancy at birth is pulled down sharply by deaths in the first years of life, so much of this gap historically reflected child survival rather than how long adults live — a low figure does not mean few people reach old age. The chasm is real; the simplest interpretation of it is not.

Japan

84.46

top

Switzerland

83.33

near top

Chad

59.09

near bottom

Lesotho

51.48

bottom

The chasm's endpoints. Flags: Wikimedia Commons, public domain.

The average that lied

Return now to the 1.75-year global drop — and watch it dissolve. Between 2019 and 2021, South-East Asia lost 3.01 years and the Americas 2.96, while the Western Pacific lost 0.07. The hardest-hit regions cratered roughly forty times more than the one that held steady. The single global figure is an average of experiences that have almost nothing in common.

At the country level the dispersion is starker still. Bolivia lost 7.51 years, more than four times the global drop; Peru lost 6.63, Mexico 5.00, the United States 2.37. Meanwhile 28 countries actually gained ground over the same window — Australia rose 0.46, New Zealand 0.39 — so the same two years that gutted the Andes barely touched the South Pacific.

The income pattern is the most counterintuitive of all: the loss was largest for lower-middle-income economies (–2.36 years) and smallest for the poorest, low-income economies (–0.55). The "great reversal" was concentrated in the middle of the world's income distribution, not at the bottom — a pattern that likely owes as much to younger populations and thinner surveillance as to any real resilience.

One advantage that barely moved

Against all this turbulence, one gap stayed almost perfectly still. In 2021 women outlived men in 182 of 183 countries; the lone exception is Israel, where men edge women by 0.29 years. The advantage is widest across the post-Soviet belt — Belarus at 10.07 years, with Latvia, Lithuania, Ukraine and Russia close behind — a pattern long tied to male alcohol, tobacco, and external-cause mortality.

Globally the gap has barely budged in two decades: 4.83 years in 2000, 5.07 in 2021, never straying outside a 4.82-to-5.41 band. Even COVID, which scrambled almost every other number in this dataset, left the female advantage essentially untouched. Some structures bend; this one didn't.

The gap was closing — and may be again

Step back to the longest view and the pandemic shrinks to a blip. The spread of life expectancy across countries narrowed steadily for two decades — the standard deviation across 183 nations fell from 9.84 years in 2000 to 7.29 in 2019 — as the poorest places gained fastest. COVID nudged it the other way, but only barely: dispersion ticked from 7.29 up to 7.15 by 2021. The long story is convergence; the pandemic was a small divergent interruption, not a reversal of the trend.

One last reminder of how to read every number here. In 2010 a single earthquake drove Haiti's life expectancy from 60.7 to 40.0 — and back to 61.1 the very next year. A 21-year plunge that vanished in twelve months is the clearest proof that period life expectancy is a snapshot of one year's mortality, not a verdict on a life. The same logic applies, more gently, to 2020 and 2021: a deep wound in the record, not necessarily a permanent loss of years for anyone alive.

So the honest closing question is not whether the world recovers its lost decade — early signs and the underlying convergence suggest it will — but whether we keep reading "global life expectancy" as a single shared fact, when the data has just spent two pandemic years proving it was never that at all.