A girl born in Japan in 1947 had a two percent chance of reaching her ninetieth birthday. A girl born there today gets a coin flip.

Both numbers sit in the same government spreadsheet. Japan’s Ministry of Health, Labour and Welfare publishes abridged life tables every year, and the 2024 edition puts female life expectancy at birth at 87.13 years. It prints the 1947 figure alongside it for comparison: 53.96. That is thirty-three years added in under eight decades, or close to five months of extra life expectancy for every year that passed.

What the number is actually counting

Life expectancy at birth sounds like a prediction about a baby. It behaves more like a photograph of a single year. Statisticians take every death rate recorded in 1947, freeze them in place, and work out how long a hypothetical newborn would last if nothing ever improved. So 53.96 describes how dangerous it was to be alive in 1947, not the fate of anyone actually born that year.

Which matters, because averages hide bodies. In 1947, half of Japanese women were still alive at 64.45. The published expectancy sat more than ten years below that midpoint, hauled down by babies dying in enormous numbers. By 2024 the midpoint had climbed to 90.04, less than three years above the headline figure. That gap narrowed because the deaths distorting it had stopped happening.

A cleaner way to see the shift is survival odds. Of the women in the 1947 table, 49.1 percent made it to sixty-five. In the 2024 table, 94.4 percent do. Pension age stopped being something a woman had to be lucky to see.

Men gained nearly as much, from 50.06 years to 81.09, though the gap between the sexes widened along the way, from 3.90 years to 6.03.

The babies who stopped dying

Japan lost 76.7 infants for every thousand born in 1947, or about one child in thirteen.

Where did that go? Takashi Yorifuji and colleagues traced the collapse prefecture by prefecture in Paediatric and Perinatal Epidemiology, finding that infant deaths began falling before the economic boom and before national insurance arrived. What predicted a prefecture’s improvement was medical supply: doctors and nurses per head, and the share of babies delivered in hospital. By 2009 the rate was 2.38 per thousand. One paper, one method, so treat it as a strong clue rather than a closed case.

Tuberculosis first, then blood pressure

More than 100,000 deaths a year, every year, until 1950. That was tuberculosis, and it held the top spot on Japan’s cause-of-death list for the whole stretch, per a public health page maintained by the city of Hiroshima.

Nayu Ikeda and a large team of Japanese and international researchers mapped what came next in a 2011 review in The Lancet. They split the gain into two acts: a fast collapse in deaths from infectious disease through the 1950s and into the early 1960s, then a long grind downward in stroke deaths from the mid-1960s on. They call the second phase one of the biggest drivers of everything that followed.

Stroke was Japan’s outlier, and the lever was blood pressure. Ikeda’s team credits population-wide salt reduction campaigns alongside cheap, widely prescribed blood pressure drugs. Soy sauce, miso, pickles and dried fish carry serious sodium loads, and a review in Hypertension Research tracks decades of national campaigning to bring intake down. It still sits above the World Health Organization’s recommended ceiling.

The 1961 decision

Cheap drugs help nobody who cannot get to a doctor. Naoki Ikegami and ten co-authors, also writing in The Lancet, date near-universal insurance coverage in Japan to 1961, enforced through a single national fee schedule binding almost every plan and provider. Same price for the same treatment, whoever you happened to be. Whether the insurance system alone bought those extra years is hard to prove, given how much else in Japan was changing at the same time, but the timing lines up awfully well.

Where the curve is now

Flattened, and slightly dented. Female life expectancy peaked at 87.71 in 2020, fell through the pandemic years, and landed at 87.13 for 2024, a hair below 2023. Japanese women are no longer the longest-lived population in the ministry’s own report, which lists Hong Kong at 88.41.

The cause-of-death breakdown explains that ceiling better than any policy argument does. For a Japanese girl born today, the likeliest eventual cause of death on the official list is senility, at 20.75 percent, edging out cancer at 19.06.

The same tables run a useful thought experiment. Wipe out cancer entirely, cure every case, and the female figure rises by 2.68 years. Heart disease is worth another 1.12 years. The great killers of the twentieth century have been reduced to one to three years apiece, which is what happens when the easy wins are already banked.

Infection has been beaten so comprehensively that the biggest killer left has no pathogen, no vaccine and no screening programme. It is just time, showing up on schedule.