The Number Everyone Knows
Ask a room full of people whether money buys happiness and someone will volunteer the figure: $75,000. Above that threshold, more income doesn't make you happier day to day. The claim comes from a 2010 PNAS paper by Nobel laureate Daniel Kahneman and economist Angus Deaton, who analyzed Gallup survey data from over 450,000 Americans and found that emotional well-being rose with income up to roughly $75,000, then flatlined, and the number entered TED talks, policy debates, and self-help books as settled truth.
It wasn't true for most people.
1.7 Million Moments of Happiness
Matthew Killingsworth, a senior fellow at Wharton, built a smartphone app called Track Your Happiness. It pinged users several times a day with one question: "How do you feel right now?" Participants answered on a continuous sliding scale from "Very bad" to "Very good," producing a granular, moment-by-moment record of emotional life as it happened, with no retrospective summaries and no yes/no checkboxes to flatten the signal.
Over several years, 33,391 employed US adults generated 1,725,994 of these real-time reports, and when Killingsworth plotted their average happiness against income, the relationship was clean: a straight line against the logarithm of income that extended well beyond $200,000 a year, with a slope above $80,000 that was statistically indistinguishable from the slope below it and no sign that the relationship weakened at any income level in the data.
Published in PNAS in January 2021, the finding directly contradicted Kahneman and Deaton: where they had found a plateau at $75,000 (the midpoint of their "$60K–$90K" income bin), Killingsworth found that each doubling of income produced a comparable increase in daily happiness whether someone earned $30,000 or $300,000, and the effect held for both experienced well-being (how people feel in the moment) and evaluative well-being (how they rate their life overall), with the two measures tracking each other closely across the entire income range.
Why Kahneman Was Wrong (and Partly Right)
Two scientists with contradictory findings in the same journal did something rare. They agreed to an adversarial collaboration, recruiting Penn psychologist Barbara Mellers as an independent arbiter, and their joint reanalysis (published in PNAS in March 2023 with Killingsworth, Kahneman, and Mellers as co-authors) produced one of the most elegant resolutions in recent social science.
Both were right, but about different people. When Killingsworth's data was split by baseline happiness, the plateau emerged only for the unhappiest 20% of the population, a minority for whom happiness flattened once household income crossed roughly $100,000 (the inflation-adjusted equivalent of Kahneman's $75,000), with sharp gains visible below that threshold but no further relief from additional money above it.
For the remaining 80%, the linear-log pattern held without interruption. And for the happiest 30%, the relationship didn't just continue — it accelerated, with the wealthiest, happiest people showing an even steeper link between income and well-being above $100,000 than below it.
The discrepancy traced to measurement. Kahneman and Deaton's Gallup survey used yes/no questions ("Did you experience happiness yesterday?"), and this binary format creates a ceiling effect: once someone reports "yes," the measure cannot distinguish between mildly content and deeply joyful. Among the majority who are generally happy, the instrument maxes out, making it look like happiness has plateaued when in reality the thermometer broke. Killingsworth's continuous scale captured gradations of well-being that dichotomous questions simply could not.
The Misinformation Audit
Here is a number neither study reported. The $75,000 threshold, adjusted for inflation to 2026, is roughly $100,000, and according to US Census data, approximately 40% of American households — around 52 million — earn above that figure, all of whom were told by Kahneman and Deaton's widely popularized finding that their next raise wouldn't improve their daily emotional experience. The 2023 adversarial collaboration showed this was wrong for about 80% of that group, meaning roughly 42 million households received inaccurate guidance about the relationship between their earnings and their daily well-being from the most famous happiness study ever published.
Did anyone actually turn down a promotion because of a PNAS abstract? That is unknowable, but the finding's influence on popular culture and economic policy debates is documented, and it shaped how a generation talked about work, ambition, and the point of earning more.
The Strongest Counterargument
Columbia statistician Andrew Gelman raised a serious caution: the income-happiness relationship in Killingsworth's data is correlational, not causal, and happier people might earn more because optimism, energy, and social fluency drive both happiness and professional success simultaneously, meaning the observational design cannot rule out reverse causation or confounding by stable personality traits. Lottery studies by Gardner and Oswald (2007) and Lindqvist, Östling, and Cesarini (2018), which use windfalls as natural experiments, do find causal happiness effects from exogenous income shocks, but those studies measure different populations and different magnitudes. The critique is real, and the linear-log relationship Killingsworth documents could reflect some mixture of income causing happiness and happiness causing income, with no way to decompose the two from his data alone.
What We Didn't Prove
The sample was self-selected: participants opted into Track Your Happiness, skewing toward smartphone-owning, English-speaking, employed Americans curious enough about their own emotions to install an app, and the lowest income bracket ($15,000) may underrepresent the severely disadvantaged, where different dynamics likely apply. Causation remains unresolved: the adversarial collaboration settled a measurement dispute but did not establish that giving people more money causes them to be happier. The finding concerns individual income within a society, not whether an entire country gets happier as GDP rises, which is a separate question (the Easterlin paradox) with much more ambiguous evidence. Finally, the "accelerating" happiness among the richest and happiest group could partly reflect personality or status effects that compound: the kind of person who earns $400,000 and reports peak happiness may differ systematically from someone earning $60,000 in ways that income alone cannot explain.
The Bottom Line
The most cited number in happiness research was an artifact of a yes/no question that couldn't distinguish happy from very happy. When a continuous measure was applied to 1.7 million real-time reports, daily happiness rose steadily with income across the entire range, and the scientist behind the original ceiling co-authored its correction. For most people, money does keep buying happiness, but with logarithmic diminishing returns: each additional dollar matters less, even as each additional doubling buys roughly the same increment of well-being.
What You Can Do
If you earn above average and have internalized the "$75K ceiling" as a reason not to prioritize income, revisit that assumption. The data suggests your next raise will make your daily experience measurably better, not because money solves problems, but because financial security and expanded choices continue to compound well-being at every level. When negotiating salary, frame it not as greed but as an investment in daily emotional experience with documented returns. And if you're already high-income but unhappy, the data carries a different message: you may belong to the 20% for whom additional income stops helping, and the returns on investing in relationships, meaning, or therapy likely exceed the returns on a higher paycheck.