There is a famous shape in the science of happiness. Ask people across the world to rate their lives from 0 to 10 — the worst possible life to the best — and the average has long bent like a smile: highest among the young, lowest somewhere in the middle of life, climbing back up among the old. The midlife dip became one of the most reliable findings in social science, replicated across dozens of countries. Evaluative This is the evaluative kind of well-being — a considered judgment of one’s whole life, not a snapshot of yesterday’s mood. Everything in this piece is that one ruler.
In the high-income world, the smile has stopped smiling. Pool every wealthy country surveyed by Gallup, weight each by population, and the young (ages 15–29) have dropped from a ladder average of 6.98 in 2010–2014 to 6.64 in 2021–2025 — a fall of 0.34 points. Over the same stretch the old (60+) rose, from 6.51 to 6.70, a gain of 0.18. The young used to sit well above the old; now they sit just below. The young-minus-old gap swung from +0.46 to −0.06 — an inversion of more than half a ladder point.
That single sign change — a positive gap going negative — is the story. The line that always ran from young-and-content down to midlife and back up is being pulled by both ends: the youthful peak collapsing inward, the elderly tail lifting up. To see it whole, look at the age profile itself, then watch it move.
The smile flattens, then tips over
Average life-evaluation ladder by five-year age band, high-income countries, population-weighted. The faint line is the early period (2010–2014); the bold line is recent (2021–2025). In the early profile the youngest tower over everyone; in the recent profile the curve has flattened — the youthful peak has fallen to sit level with the oldest band.
The fine-grained profile makes the reversal vivid. In 2010–2014 the rich world’s 15–19-year-olds averaged 7.33 on the ladder, the highest of any age; by 2021–2025 that band had fallen to 6.80. At the other end, the oldest band climbed: those 70 and over went from 6.52 to 6.80, rising from the middle of the pack to sit level with the youngest. The once-commanding youth peak has collapsed: the two ends of the curve, once a ladder point apart, are now essentially tied. The classic U has not just deepened or shifted — it has flattened, with its high youthful arm pulled down to meet the rising old.
The young end fell. The old end rose. Where the curve once peaked, it now bottoms out.
It did not happen everywhere the same way
Zoom out from the rich world to the entire planet and the pattern changes character. Globally, life evaluation rose at almost every age — but unevenly, and that unevenness is itself the point. The world’s young (15–29) were essentially flat, slipping 0.01 points; the world’s old (60+) rose the most, gaining 0.24. In between, the gains grew with age: 30–44 up 0.04, 45–59 up 0.12. So the age gap widened on every continent — but by opposite mechanisms. In poorer countries a rising tide lifted the old fastest while leaving the young behind. In the rich world the tide went out from under the young.
Two worlds, one direction
Change in average ladder from 2010–2014 to 2021–2025, by age band. Left, all high-income countries; right, the whole world. Bars above the line are gains, below are losses. In both, age and fortune now run together — but the rich-world young are the only group that fell hard.
The crossover, year by year
The gap did not lurch to zero in a single bad survey. Tracked wave by wave across the high-income world, the young (15–29) and the old (60+) ladder lines start the 2010s a comfortable distance apart, with the young on top. Through the late 2010s the young line drifts down and the old line drifts up; around 2022 they touch and trade places. The two lines have not pulled back apart since — the post-2020 era is one of persistence, not rebound.
The lines that crossed
Average ladder for the rich-world young (15–29) and old (60+), each Gallup wave. The young begin the decade clearly above the old; the lines converge through the late 2010s and cross in the early 2020s. Shaded between them is the gap that flipped sign.
A reversal, country by country
The aggregate hides a near-universal local pattern. In country after wealthy country the young-minus-old gap shrank, and in many it crossed below zero. The United States is the starkest case: in 2010–2014 its young (15–29) and old were nearly level, a gap of −0.07; by 2021–2025 the young trailed the old by 0.76 ladder points. Canada moved almost identically, from −0.02 to −0.89. Even places where the young still lead — Italy, Greece, Portugal, Poland — saw their youth advantage shrink by half or more. The collapse of the young’s edge is the common thread; in the largest rich economies it has become an outright deficit.
Where the young lead, and where they fell behind
Each row is a high-income country’s young-minus-old ladder gap: the open dot is 2010–2014, the filled dot is 2021–2025. Dots to the right of zero mean the young rate their lives higher than the old; to the left, the old lead. Almost every arrow points left — toward the old.
A second instrument agrees
Could this be a quirk of one survey’s ladder question? The United States offers a check. The General Social Survey has asked Americans the same plain question — taken all together, how happy would you say you are? — since 1972, on a three-point scale where higher is happier. Evaluative It is a different instrument, a different mode, a different scale, but the same kind of evaluative judgment. And it tells the same story. Among Americans 18–29, mean happiness fell from 2.19 in 2010–2014 to 1.90 in 2021–2024 — a drop of 0.29, much steeper than the 60+ decline of 0.15. The young-minus-old happiness gap moved from essentially zero (−0.001) to −0.14: the young, once even with their elders, fell behind.
What this does and doesn’t show
It is a real reversal, and it supersedes the old picture. The midlife U was always an evaluative finding, and so is its inversion. Earlier work in this series mapped the U-curve across countries as a fixed fact of the lifespan; this is the correction. The shape was never permanent — it was a feature of a particular cohort of young people, and that cohort’s successors are not enjoying the same head start.
Two weightings, one answer. The headline figures population-weight countries, so the large rich economies count for more. Weighting every country equally instead, the same inversion appears in muted form — the young still slip and the old still rise — so the finding is not an artifact of any single big country dominating the average.
Small differences are soft. Cross-period and cross-country ladder gaps under about 0.2 points should not be pressed; response styles and translation differ. The recent rich-world young-minus-old gap of −0.06 is itself inside that soft zone — the news is not that the old have decisively overtaken the young, but that a once-dependable half-point youth advantage has vanished. The 0.34-point fall among the young, and the 0.52-point swing in the gap, are well outside it.
This is description, not diagnosis. The data show that the young rate their lives lower, not why. The usual suspects — smartphones, housing costs, labor markets, the pandemic’s social toll, a thinning of in-person life — are not adjudicated here. What the ladder measures, it measures cleanly: the generation that used to be happiest is no longer, and the curve that defined a lifetime of well-being has turned upside down.
Notes & data
- Sources. Gallup World Poll, 2005–2025 (Cantril ladder, current life evaluation, 0–10); General Social Survey, 1972–2024 (
happy, 3-point, recoded so higher = happier) for the US echo. Analysis code:analysis/the-u-curve-inverts/build_data.py; every figure in the text is recomputed there and traced inclaims.json. - Periods & geography. Early = 2010–2014, recent = 2021–2025. “Rich world” = countries labeled High income in the World Bank income grouping (about 50 countries early, 49 recent). “World” pools all income groups with population data.
- Weights. Within each country, estimates use Gallup’s within-country weight (
wgt). Cross-country aggregates weight country means by population (ctry_pop_millions, a static ~2018 vintage); countries missing the population figure are excluded from population-weighted aggregates only. An equal-country robustness check shows the same inversion. GSS useswtssps(post-stratified, full 1972–2024 coverage). - Construct. Evaluative throughout — the Cantril ladder and GSS
happyare both considered judgments of life as a whole. No experiential (daily-emotion) items appear here, and none are mixed onto these axes. - Suppression & samples. Gallup country×age×period cells with fewer than 100 ladder responses are suppressed; GSS cells with Kish effective N below 50 are suppressed (none triggered). Rich-world period totals run to tens of thousands of interviews per age band; world bands to the hundreds of thousands.
- Known limits. The GSS 2021–2022 push-to-web mode change can shift levels and is flagged in the text; the mode-stable Gallup US ladder agrees, so the direction is robust. Cross-period and cross-country ladder gaps under 0.2 points are not definitive. This article supersedes the earlier across-countries U-curve treatment.
- Prior art. Stone, Schwartz, Broderick & Deaton 2010 (the U-shape of well-being over the life course); Blanchflower’s work on the midlife low; and recent World Happiness Report findings on falling youth well-being in North America and Western Europe, which this fixed-period, population-weighted re-computation traces as a full inversion of the curve.