Age 47 has acquired an oddly precise reputation. In popular accounts of the “happiness U-curve”, it is the point where life satisfaction supposedly reaches the bottom before beginning a long ascent.
That number came from a large and serious body of survey analysis, but it was never a forecast for an individual life. More recent evidence makes the simple version harder to defend. In several major datasets, younger adults now report lower wellbeing or greater distress than people in midlife, flattening or reversing the left side of the old U.
The careful conclusion is not that people stop getting happier after middle age. It is that the U-curve described a broad statistical pattern in particular populations and periods. It was never a biological timetable, and it may now be changing.
Where the age-47 claim came from
In a 2020 paper in the Journal of Population Economics, economist David Blanchflower compared 109 datasets covering 145 countries. The measures varied, including happiness, life satisfaction and answers to the Cantril ladder question, which asks respondents to place their lives on a scale.
His models found a U-shaped relationship between age and subjective wellbeing across all 145 countries. The low points clustered near midlife. In an earlier working-paper version, the average minimum was 47.2 across estimates from advanced countries and 48.2 in developing countries. Across the expanded published analysis, the average of the country-level estimates was about 48.
That is the origin of the memorable claim. Scandinavia Standard has previously covered the classic U-curve evidence. The important update is that 47 or 48 was an average turning point produced by many fitted curves, not a birthday on which dissatisfaction peaks for everyone.
The curve depends on what the model holds constant
Blanchflower’s main estimates controlled for factors including education, marital status and employment. In plain terms, the model compared people of different ages while trying to hold those circumstances constant.
There is a useful question behind that approach: what is the association between age and wellbeing when two people otherwise have similar measured circumstances? But it is not the same as asking what people of different ages actually report in everyday life. Marriage, employment, health and family responsibilities are among the ways lives change with age. Statistically removing some of them can deepen the apparent curve.
Sociologist David Bartram tested this issue in a 2022 methodological analysis of European Social Survey data. He removed controls that could themselves be affected by age, included respondents older than 70 and avoided forcing every country into a quadratic curve.
Under those choices, the U-shape was much shallower and appeared clearly in only some countries. Norway retained the conventional pattern in one age-band analysis. Denmark did not: its estimated relationship pointed towards happiness increasing across the life course, although that trend was not statistically strong. Sweden was also not among the seven countries showing the conventional pattern in that model.
This disagreement does not make one analysis fraudulent and the other correct by definition. It shows that “happiness bottoms out at 47” is partly a description of a modelling specification.
Newer data has changed the left side of the U
The larger change is not statistical technique. It is the deterioration in reported wellbeing among younger people.
A 2025 peer-reviewed study in PLOS One, by Blanchflower, Alex Bryson and Xiaowei Xu, examined repeated surveys in the United States and United Kingdom alongside Global Minds data from 44 countries collected between 2020 and 2025.
The US data covered more than 30 years of reports about days of poor mental health. The UK sources included a household panel and national anxiety data. The international analysis included about 1.7 million online respondents aged 18 to 74 in countries with at least 10,000 observations.
The old mirror image of the happiness U was an “unhappiness hump”: distress tended to rise towards midlife and then decline. In the newer datasets, that hump was gone. Ill-being was highest among younger adults and generally declined with age. In the international sample, 13.4 per cent of respondents under 25 met the project’s threshold for distress, compared with 5.6 per cent overall.
This is one study, not settled consensus. Its global component was an online, non-probability survey, and its central international measures concerned mental health and distress rather than the same life-satisfaction question used in every earlier U-curve analysis. The finding cannot simply be substituted for a global measure of happiness.
Life-satisfaction surveys point in the same direction, with limits
A separate 2025 National Bureau of Economic Research working paper by Jean Twenge and Blanchflower brought together 11 studies. It reported declining life satisfaction and happiness among young adults in the United States and five other English-speaking countries: Australia, Canada, Ireland, New Zealand and the United Kingdom.
In data collected since 2020, happiness and life satisfaction generally rose with age across those six countries. That is closer to a direct challenge to the classic U-curve, although it is a working paper rather than a peer-reviewed journal article and its six-country comparison is concentrated in the English-speaking world.
The studies also cannot yet settle why younger wellbeing has fallen. The authors discuss employment prospects, housing, changing expectations and the pandemic among several possibilities. None is isolated as a single cause by these cross-sectional patterns.
Age, period and generation are difficult to separate
A survey taken today compares a 25-year-old and a 65-year-old at the same moment. The difference could reflect ageing, but it could also reflect the historical periods in which they grew up or conditions affecting everyone at the time of the survey.
Researchers call these age, cohort and period effects. They are notoriously difficult to separate because age is determined by birth year and survey year. A cross-sectional U could mean that individuals tend to dip and recover as they age. It could also partly mean that today’s middle-aged respondents belong to a less satisfied cohort, or that people with lower wellbeing are less likely to survive or remain in surveys at older ages.
Long-running panel studies, which follow the same people, help but introduce their own problem: participants drop out, and those who remain may not be representative. There is no single survey design that turns the curve into a universal law.
The old curve was a pattern, not a promise
The encouraging part of the original finding was never the decimal point attached to its nadir. It was the observation that lower wellbeing in midlife did not necessarily continue falling for the rest of life.
That remains true for many people and in many datasets. Older adults often report strengths that a simple story of decline misses. Yet population averages cannot promise that an individual will feel better at 50, 60 or 70, and the recent fall in younger wellbeing is not good news about ageing. The curve is changing because one end appears to have moved down.
The best current reading is therefore historical as much as psychological. For decades, many surveys produced a midlife low near the late forties. Some analyses still do. But newer evidence suggests the familiar U is neither everywhere nor permanent. Age 47 was never destiny, and the conditions shaping satisfaction can change faster than a life course.