Ask people to place their lives on a ladder — zero the worst life they can imagine, ten the best — and for most of the modern history of this question, the obvious story was geographic. Rich countries clustered high, poor countries clustered low, and the great gulf in human well-being ran between nations. Fourteen years of Gallup World Poll data, tracked through a fixed panel of 42 countries surveyed every single year from 2006 to 2019, show that gulf quietly changing its address.
Three things happened at once. The population-weighted world average climbed from 4.96 to 5.43 — the world, on this measure, got happier. The spread between countries' averages narrowed, from a population-weighted standard deviation of 0.89 to 0.70: nations converged. And the spread within countries — the typical distance of a citizen's answer from their own national average — widened from 1.92 to 2.37 ladder points, a rise of about 24%. Nor is that an artifact of a few big movers: on a robust three-year-average basis, comparing 2006–08 with 2017–19, within-country dispersion rose in 39 of the 42 panel countries. The salient divide in how life feels stopped being a matter of which country you live in and became, increasingly, a matter of who you are inside it.
Watch the middle thin and the edges fill
Averages hide this story completely; the shape of the distribution tells it. Below is the population-weighted distribution of ladder answers across the 42-country panel — what share of people stood on each rung, year by year. In 2006 humanity piled up in the middle rungs. By 2019 the pile is flatter: the middle has thinned and both edges have filled in, the very bottom rung and the very top rung most of all. The world's center of gravity drifted up the ladder even as its tails fattened.
Where the world stood on the ladder, 2006 → 2019
Share of people at each rung (0 = worst possible life, 10 = best), population-weighted across the 42-country panel. Tick on each baseline marks that year's distribution mean.
Two kinds of inequality, moving in opposite directions
Put the two dispersions on the same axis — both are in ladder points, so they are directly comparable — and the scissors open. Within-country dispersion climbs; between-country dispersion sinks. The series is honest-to-goodness survey data and it is noisy: 2006 itself sits a little high, and the line wobbles year to year. The robust signal is not any single step but the level shift — within-country dispersion sat around 1.8–1.9 through 2006–2010 and around 2.3–2.4 by 2018–2019. Treat the trend, not the jitter, as the finding.
The scissors: dispersion within vs. between countries
Standard deviation of ladder scores, 2006–2019, population-weighted across the 42-country panel.
Of all the variation in how people rate their lives, the share found inside countries rather than between them — 2019:
The decomposition makes the headline literal. Square the dispersions and you can ask: of the total variation in life evaluations across this panel, how much lives inside countries versus between them? Already in 2006 most of it sat inside borders — but by 2019 the within-country share had grown further while the between-country share shrank, in both absolute and relative terms. If you wanted to predict how someone rates their life, knowing their country told you less in 2019 than it did in 2006; knowing their position within their country told you more.
Two countries, one pattern
The aggregate ridgeline can feel abstract, so here are the two heaviest weights in the panel, each pooled over three survey years to strip out single-year noise. China widened as it rose: its average climbed from about 4.77 to 5.12, the upper rungs filled in fastest — the share at 8 and above more than doubled — and yet the very bottom rung grew too, stretching the distribution at both ends (SD 1.85 → 2.26). The United States took the opposite route to the same destination. Its average fell, from about 7.34 to 6.94, as people drained out of rungs 8 and above into the middle and lower-middle of the ladder; but the distribution did not tighten as it sank, it spread (SD 1.79 → 1.93). One country widened while rising, the other while falling: two different national trajectories, the same within-country widening found in 39 of the 42 panel countries.
Inside two countries: 2006–08 vs. 2017–19
Share of respondents at each ladder rung, survey-weighted within country and pooled over three-year windows. Both panels share the same vertical scale.
United States
China
What this can and cannot say
First, the panel. Holding the set of countries fixed — the same 42, present every year — is what makes 2006 and 2019 an apples-to-apples comparison rather than an artifact of Gallup adding and dropping countries. The price is coverage: this panel excludes much of the world, including most of Western Europe and several large countries that were not surveyed in every year. The panel does include the United States, China, Indonesia, Bangladesh and Vietnam, so it carries real demographic weight, and all "world" figures here are population-weighted — never a plain average of country means, which would let small countries swamp large ones. But "the world" in this piece means this panel's world.
Second, the clock stops in 2019, deliberately. The pandemic scrambled both survey operations and well-being itself; ending the series pre-pandemic keeps the trend clean, and means nothing here speaks to what happened after.
Third, the ladder is an evaluative measure — a judgment about one's life as a whole, not yesterday's mood — and cultures differ in how they use response scales. Some lean on the midpoint; others use the extremes freely. That mostly threatens comparisons of levels across countries. The within-country trend is more protected: each country is compared with its own past, and the dispersion rose inside countries across very different cultures at once — in 39 of the 42 panel countries on the pooled 2006–08 vs. 2017–19 comparison.
This finding also has a history worth being straight about. Stevenson and Wolfers (2008) documented the opposite movement in the postwar United States: happiness inequality there fell for decades even as income inequality rose. The widening shown here is a different era and a global panel — and it is corroborated: the World Happiness Report 2024 likewise reports within-country well-being inequality rising across most regions of the world. The reversal of direction, from Stevenson and Wolfers' converging America to this diverging world, is part of what makes the trend notable.
What the data describe, then, is a redistribution of difference. The world's average citizen climbed half a rung. The gap between nations closed by a fifth of a ladder point. And inside borders — in 39 of the 42 panel countries, rich and poor alike — the typical citizen's distance from their own national average grew by almost half a rung. Whatever is pulling well-being apart is operating less at the level of flags and more at the level of lives. The atlas of feeling still has borders drawn on it; they just explain less than they used to.