The $75,000 happiness threshold became one of those numbers people remember even after the study behind it has disappeared from view. It appeared in career advice, salary negotiations and conversations about how much is enough. Earn less, the story went, and more money will make life feel better. Earn more, and the emotional return drops to zero.
The research never supported such a universal rule. The original paper separated day-to-day feelings from reflective life evaluation, and only one of those measures flattened. Later researchers used a more sensitive tool and found no average plateau. When the disagreeing scientists examined the data together, they discovered that different parts of the happiness distribution followed different paths.
The result is less quotable than $75,000, but more useful: income is associated with higher happiness well beyond six figures for most people, while a plateau appears mainly among the least-happy minority. Even then, the studies show association, not a guaranteed emotional return from the next raise.
The original paper measured two different kinds of well-being
Daniel Kahneman and Angus Deaton analysed more than 450,000 responses from the Gallup-Healthways Well-Being Index, collected from US residents in 2008 and 2009. Kahneman had received the Nobel Memorial Prize in Economic Sciences in 2002; Deaton would receive it in 2015. Their 2010 paper in PNAS asked whether household income related differently to two forms of subjective well-being.
Life evaluation was reflective. Respondents placed their lives on a ladder whose bottom represented the worst possible life and whose top represented the best. That score continued to rise with log income across the measured range, including well above $75,000.
Emotional well-being concerned the previous day. Respondents answered questions about happiness, enjoyment, frequent smiling or laughter, worry, sadness and stress. These daily-emotion measures improved as income rose through the lower brackets, then showed no further statistically detectable improvement beyond roughly $75,000 a year.
The popular retelling collapsed those separate findings into “happiness stops rising.” But the authors themselves found that richer households continued to rate their lives more favourably. The plateau applied to their measure of yesterday’s emotional experience, not to every meaning of happiness.
The $75,000 figure was never a universal salary line
The number referred to annual household income in the United States, not one person’s salary. It came from 2008–2009 dollars, before more than a decade of inflation, and it was not a cost-of-living threshold adjusted for San Francisco, rural Iowa or Stockholm. Household size also changes what the same income can provide.
The Gallup survey recorded income in bands. At the upper end, respondents selected categories such as $60,000 to $90,000, $90,000 to $120,000 and more than $120,000. Kahneman and Deaton inferred saturation because emotional well-being improved between the lower two categories but not between the top two. The familiar $75,000 was therefore an approximate point inside a bracketed pattern, not a precise switch.
Income was analysed logarithmically. On a log scale, proportional changes matter: moving from $30,000 to $60,000 is comparable to moving from $100,000 to $200,000 because both are doublings. This captures diminishing marginal value. Happiness can continue rising with income even while each additional dollar is associated with a smaller change than the dollar before it.
A yes-or-no scale ran into a ceiling
The most important measurement detail was easy to miss. The original daily-emotion questions were dichotomous. Someone either reported experiencing happiness yesterday or did not. The positive-affect score combined yes-or-no answers about happiness, enjoyment and smiling or laughter; negative states were captured in similarly coarse form.
This design can identify movement from distress toward feeling okay. It struggles to distinguish someone who had a moderately pleasant day from someone who felt intensely joyful, because both simply answer yes. As more respondents reach the positive answer, the measure approaches its ceiling and has less room to register further improvement.
That is why the later collaboration argued that the original emotional measure was better understood as tracking the disappearance of unhappiness. Below the threshold, rising income was associated with fewer people reporting emotional pain or the absence of positive feeling. Above it, the binary questions could not reliably separate degrees of happiness among people already answering positively.
This does not make the 2010 study careless. Its enormous sample and distinction between experienced and evaluative well-being were major contributions. It shows how a conclusion can be statistically correct for the instrument used yet acquire a broader meaning the instrument was not designed to support.
Smartphone sampling found happiness still rising
Matthew Killingsworth took a different approach. Through the Track Your Happiness project, participants received prompts on their smartphones at randomly timed moments and rated how they felt immediately before the signal on a continuous scale from very bad to very good.
His 2021 PNAS study analysed 1,725,994 experience-sampling reports from 33,391 employed US adults. Both momentary well-being and evaluative life satisfaction rose approximately linearly with log income. The slope above $75,000 was about as steep as the slope below it, with no obvious break at the famous threshold.
The continuous scale mattered because it retained differences at the positive end. Two people who would both have answered yes to “Were you happy yesterday?” could give different momentary ratings. The smartphone method also reduced the burden of reconstructing a whole day from memory.
It introduced other limitations. Participants volunteered for an app-based study, were employed, lived in the United States and were not a nationally representative probability sample. Household income and well-being were self-reported. The design remained observational, so it could not show that income itself caused the higher ratings.
The adversarial collaboration found different curves for different people
Kahneman and Killingsworth had produced apparently contradictory answers. Rather than trading critiques, they worked with University of Pennsylvania statistician Barbara Mellers in an adversarial collaboration, a format designed to let researchers with opposing expectations agree on tests and interpretations.
Their 2023 reanalysis returned to Killingsworth’s experience-sampling data and examined different percentiles of person-level happiness. Averaging everyone together had hidden important differences in the shape of the distribution.
At roughly the 15th percentile, representing the least-happy group, experienced well-being rose with log income up to around $100,000 and then flattened. This was the part of the population for which the plateau pattern survived. The $100,000 dividing point was also roughly an inflation-adjusted counterpart to the earlier $75,000 figure.
For people at the 30th, 50th and 70th percentiles, happiness continued rising above $100,000. At the 85th percentile, the association became steeper beyond that point. The overall near-straight line was therefore an average of different nonlinear curves: flattening at the unhappy end and acceleration at the happy end.
The collaboration also explained why both earlier papers missed part of the story. Kahneman and Deaton’s binary measure compressed degrees of happiness near the top. Killingsworth’s 2021 average obscured a plateau confined to the lower tail. Neither initial analysis anticipated that the distribution itself would change shape with income.
Money matters, but the studies do not turn it into destiny
It would be equally misleading to replace “money stops mattering at $75,000” with “more money always makes everyone happier.” These studies compare people with different incomes. Health, employment, family, location, personality and opportunity can influence both earnings and well-being. A later methodological commentary likewise cautioned against treating the reported curves as causal effects.
The average associations were modest and logarithmic. Doubling income was linked to a similar increment at different income levels, which means reaching the next emotional increment requires increasingly large dollar gains. Group patterns cannot predict how a particular raise, promotion or job with longer hours will affect one person.
Nor should an American threshold be exported unchanged. A 2018 analysis of 1.7 million people across 164 countries found that apparent satiation points varied by world region and by whether researchers measured emotional well-being or life evaluation. Prices, public services, inequality, expectations and social comparison all alter what income means.
Income also sits beside other strong correlates of well-being. Scandinavia Standard has covered World Happiness Report evidence that sharing meals is associated with life satisfaction as strongly as income or employment status. Financial security can reduce daily pain and expand choice, while relationships, health, time and purpose shape what those choices feel like.
The famous plateau was not fabricated. It emerged from a real pattern in a huge dataset, measured with a scale that was much better at identifying who was no longer unhappy than how happy already-positive people had become. Newer data did not simply overturn it; they located it.
For the least-happy minority, the curve flattened above roughly $100,000. For most people, it kept climbing, and for the happiest it grew steeper. The honest conclusion has no magical salary: more income is generally associated with more happiness, with diminishing value per dollar, substantial variation between people and no promise that money alone can build a good life.