Life expectancy was 35. Almost nobody died at 35.

The finding

This box is the article. The rest is evidence.

TL;DR

When life expectancy at birth was under 40, people already 65 still expected about 75.

75.39median expected age if alive at 65
211 country-years, 11 countries

0.27share of births that reached 65

195.9infant deaths per 1,000 births

75.07same age-at-65 figure when birth LE was 30 to 35

The myth “Life expectancy was 30 to 35, so adults died at 30 to 35.”

What the tables say Early deaths were mostly babies and children. People who reached old age were not timed to die at 35.

Sweden, 1800: life expectancy at birth 32.2. About 227 infant deaths per 1,000. About 21% of births reached 65. Those who did expected age about 73.7.

Second myth, also wrong: adults improved too. Remaining life at 65 rose about +8.45 years from before 1900 to after 2000 (12 countries; about 7.7 to 9.1).

Scope: countries with strong vital registration in open life tables (Human Mortality Database style). One year’s death rates, not every past society, not one person’s full life story.

Plain English

If 75 and 27% feel like they fight each other, start here.

Simple version

Picture 1,000 babies in a hard year. Hundreds die before age one. More die as small children. Average age at death for the whole group can sit in the thirties. That average is real. It is a bad summary of how long a grown-up lived.

The meme number

~32

Life expectancy at birth: average years for a newborn if this year’s death rates never changed. Dead babies pull it down hard.

The forgotten number

~75

If you already reached 65 under those rates, expected age at death is around the mid-seventies. Not 35.

The catch

~1 in 4

Only about a quarter of births reached 65 when birth life expectancy was under 40. Most never got the old-age number.

Two questions, same year

Life expectancy at birth: for a baby born into this year’s death rates, how many years on average?
Age if already 65: you already survived; under this year’s rates, how old do people like you die?
Same rates. Different questions. The myth swaps them.

So did people die young?

Yes. Childhood killed a huge share of births. That is why the newborn average was low.
No. Adults were not on a 35-year clock. By 15, expected age is already the mid-fifties. By 65, about 75.

Two myths

Both fail for the same reason: one number, wrong question.

Myth A

“Life expectancy was 30 to 35, so people died at 30 to 35.” Confuses the newborn average with adult death age.

FALSE

Myth B

“Only baby survival got better. Adults always lived modern lengths.” Remaining life at 65 still rose hard.

FALSE

The rule

Always pair survivor age with infant death or survival to 65. Never post 75 without the 27%. Never post 32 without asking who died early.

DO THIS

Proof

One long series first. Then many countries. Then the exact meme window.

Sweden

What to look at

Longest clean series. As baby death falls, life expectancy at birth rises. The “already 65” line sits high the whole time, then rises further.

Sweden: life expectancy, infant mortality, survival to 65, expected age at 65 — **Sweden over time.** Low newborn average and high old-age expectation in the same country. Infant death and survival to 65 explain the gap.

Eleven countries, same pattern

When life expectancy at birth < 40	Value
Country-years (both sexes, cleaned)	211
Countries	11
Median life expectancy at birth	37.0
Median infant deaths per 1,000	195.9
Median share reaching 65	0.270
Median expected age if alive at 65	75.39
Median of country medians (one vote each)	75.37
Range of country medians	74.78 to 76.08

Sweden has many years in the sample. One vote per country still lands near 75.

Multi-country low birth LE with high age at 65 — **Not just Sweden.** In low birth-LE years, expected age if alive at 65 clusters near the mid-seventies.

Scatter: birth LE vs expected age at 65 — **The gap on a scatter.** Low birth LE, high age-if-65.

Distributions when birth LE is low — **When birth LE is low.** High infant death, low survival to 65, high age if already 65.

Year-weighted vs equal-country medians — **Weighting check.** Count every year, or one median per country. Both answers are about 75.4.

Deaths pile up early

Where deaths concentrate early in life — **Where deaths sit.** Early life takes the big bite. That is why the newborn average looks like the thirties.

Birth LE versus infant mortality — **Birth LE tracks infant death.** Median within-country correlation about -0.94. When babies die less, the newborn average rises.

Narrow to the meme window: birth LE 30 to 35

Only years where life expectancy at birth was 30 to 35:

45 country-years, 8 countries
Median birth LE about 33.5
Median infant deaths per 1,000 about 213
Median share reaching 65 about 0.22
Median expected age if alive at 65 still about 75.07

Strict 30-35 birth LE band still near 75 at age 65 — **The meme window.** Early death gets worse. Age if already 65 does not fall to the thirties.

Age at 65 by country when birth LE is low — **Country by country.** Every country median for age if alive at 65 sits near 75 when birth LE is under 40.

The myth fails long before 65

The myth already fails at 15. Each rung is expected age if you are already alive there.

Simple version

At birth the average is low because of baby death. If you are still alive at 15, expected age jumps into the fifties. At 30, about 60. At 65, about 75.

Already this age	Expected age (single-table)	Expected age (median of tables)
0 (birth)	~35	~37
5	~52	~53
15	~56	~58
20	~57	~59
30	~60	~62
50	~68	~69
65	~75	~75

Source: Human Life-Table Database, years when birth LE was under 40. Single-table cells and multi-table medians both show the same climb.

Age ladder of expected ages when birth LE under 40 — **Age ladder.** Expected age climbs once childhood is cleared. The myth is broken by the teen rung.

Single-table and multi-table ladders — **Two ways to build the ladder.** Coverage differs. The shape does not.

Adults improved too

Fixing Myth A does not mean adult longevity never moved.

Remaining life at age 65	Value
Mean change (after 2000 minus before 1900)	+8.45 years
95% interval (bootstrap over countries)	7.74 to 9.14
Median country change	+8.44
Range across countries	+6.4 to +10.7
Countries	12

Remaining life at 65 before 1900 vs after 2000 — **Before 1900 vs after 2000.** Remaining life at 65 rose about 8 to 9 years. Baby survival was the big early story. Old-age remaining life still moved.

Data and checks

Headline numbers: Human Mortality Database public summaries (birth LE, LE at 65, infant mortality, survival to 65). Age ladder: Human Life-Table Database. Context: OWID, Clio, Eurostat.

Piece	What we used
Main panel	4,277 country-years, 44 regions, 1751 to 2023
Myth A sample	Birth LE under 40: 211 country-years, 11 countries
Long series	Sweden from 1751; France from 1816; England and Wales from 1841; USA from 1933
Cleanup	Drop double-counted population variants; do not let multi-table life tables flood the average

Coverage by country and year — **Who is in the data.** Low birth-LE years are mostly historical vital-registration Europe plus early Sweden.

Birth LE and infant death move together

Simple regressions and within-country correlations show it. High R-squared is expected: both come from the same year’s death rates. Description, not a causal claim about one policy lever.

How tightly birth LE tracks infant death — **Fit of the association.** Useful description only.

Year-to-year changes in birth LE and infant death — **Year-to-year changes.** When infant death falls, birth LE usually rises with it.

Men, women, worst year

Sex-specific results including Iceland 1843 — **By sex.** Medians still near 75. Worst female year in the low birth-LE sample: Iceland 1843, about 69.9, with tiny survival to 65. Still about 70, not 30.

python -m src.analysis.run_final_agrade
python -m pytest -q

# outputs/bulletproof/final_claims.json
# outputs/reports/FINAL_A_GRADE_REPORT.md

Built with Grok 4.5 as a coding partner. Numbers are checked against open tables and repo claim gates.

Limits

Strong vital-registration countries in open databases. Not every past society.
Period rates for a year. Not one person’s full biography from birth to death.
“Age if already 65” only applies to people who reached 65. Pair it with survival or infant death.
Mid-age ladder tables vary by method. We show two builds on purpose.
Iceland 1843 is a crisis floor near 70 for women, not proof of the 35 myth.

One line to share

Life expectancy of 30 meant childhood was deadly, not that adults died at 30. Survivors to 65 still expected about 75. Adults later gained about 8 more years at that age.

Post 75 with the 27%. Post 32 with the infant deaths. Keep both numbers together.