5.2 The Money Question · More Than a Feeling

Philadelphia, 2021. Matt Killingsworth has spent the better part of a decade interrupting people. His research app, Track Your Happiness, pings its users at random moments and asks a question of almost rude simplicity: "How do you feel right now?" By the time he publishes, he has collected 1.7 million answers from more than 33,000 working American adults — and when he plots those feelings against income, the line rises, and keeps rising, straight through $75,000 a year without so much as a stumble.

Seventy-five thousand dollars is not a random address. It is the most famous number in the science of happiness — the income at which, according to a celebrated 2010 study, day-to-day emotional well-being stops improving. One of that study's authors is Daniel Kahneman: Nobel laureate, author of Thinking, Fast and Slow, probably the most famous living psychologist. He is eighty-six, and a younger researcher's data has just declared his signature number wrong.

What usually happens next is a feud — decades of dueling papers, footnotes sharpened into weapons, conference panels seated carefully apart. What happened instead: the two men agreed to join forces and settle the question together, with a referee empowered to keep them honest.

Two years later, the answer appeared under both their names. This lesson is the story of science's fifty-year argument about money — the rare academic brawl with a happy ending — and of what your next raise can and cannot buy.

Where This Takes You

You'll follow the money question through five decades and three reversals — ending with a Nobel laureate co-signing the correction of his own most famous finding.
You'll learn to read income the way the data does — in doublings, not dollars — and see why the same raise can mean everything to a barista and almost nothing to a banker.
You'll collect honest verdicts on two items from your orientation Self-Test: the $75,000 plateau (item 3) and experiences versus things (item 4).
You'll spend a few dollars in two directions this week and run one of the field's most charming experiments on a sample of one.

The Economist Who Noticed Nothing Happening

Start in 1974, at the University of Pennsylvania, where an economist named Richard Easterlin is doing something economists at the time considered slightly disreputable: reading happiness surveys.

The surveys had been running since just after the Second World War, asking Americans a plain question — taken all together, would you say you are "very happy," "pretty happy," or "not too happy"? Meanwhile the country underneath the surveys had transformed. Between the late 1940s and 1970, real income per person had climbed by more than half. Houses grew. Cars multiplied. Kitchens filled with machines that would have looked like sorcery a generation earlier.

And the happiness numbers did — nothing. The share of Americans calling themselves very happy in 1970 was about the same as in the lean years after the war. Easterlin found the same flatness abroad, in every country with data worth reading.

Here is what made it a paradox rather than a shrug: within any country, at any given moment, richer people reliably reported being happier than poorer people. Money clearly mattered when you compared neighbor to neighbor. It just seemed to stop mattering when a whole nation got richer together. That double finding — money matters across the street but not across the decades — became the Easterlin paradox, the opening move of the entire economics of happiness. It was also quietly subversive. Economics had long run on the assumption that more income means more well-being; that was the whole point of growth. Easterlin had done something almost nobody had bothered to do: checked.

Easterlin's explanation should sound familiar if you remember the treadmill. Part of what income buys, he argued, is position — and position is measured against other people. If everyone in the parade rises together, nobody feels taller. Add adaptation — the new kitchen becoming invisible within months — and you have two engines quietly eating economic growth before it can reach anyone's mood.

Now the honesty clause, because this course has one: the paradox is still contested, fifty years on. In 2008, economists Betsey Stevenson and Justin Wolfers reanalyzed the question with far more countries and far better data and found the opposite: life satisfaction rising with income everywhere they looked — richer people happier than poorer people, richer countries happier than poorer countries, and countries getting happier as they grew, with no ceiling in sight. Easterlin fired back that they were reading short-run wiggles, not long-run trends, and the two camps are trading papers to this day. When someone tells you "the science shows economic growth doesn't make people happier" — or that it definitely does — they are handing you one side of a live argument as if it were settled.

But notice what all those surveys were actually asking. "Taken all together, how is your life?" That's a judgment — a step-back, whole-life evaluation. It's only one of the dials on Diener's dashboard from Lesson 1.1. Nobody was asking how people felt — until somebody did, and the whole question split in two.

The Most Misquoted Number in the Field

In 2010, two Princeton professors — Daniel Kahneman and the economist Angus Deaton, each with a Nobel Prize in his present or future — got access to a remarkable dataset: a Gallup poll that had been interviewing a thousand Americans a day, around 450,000 responses in all.

Their move was to keep two dials separate — the same two you've been tracking since Module 1. Dial one: life evaluation, measured with the Cantril ladder you met in Lesson 4.4 — picture a ladder with rungs from 0 (worst possible life) to 10 (best possible), and place yourself on it. Dial two: emotional well-being — not a verdict on your life, but yesterday's actual weather. Did you smile or laugh a lot yesterday? Experience enjoyment? Stress? Worry? Sadness? Yes or no.

The two dials told two different stories. Life evaluation rose with income and kept rising to the top of the data — no plateau, no ceiling, more money, higher rung, all the way up. But the emotional dial behaved differently. Reports of yesterday's enjoyment and stress improved with income up to roughly $75,000 a year — and then flattened. Above that line, richer households climbed the ladder but their Tuesdays didn't seem to feel any better. The authors also found the reverse edge sharper: below it, low income didn't just mean fewer comforts — it amplified the emotional pain of everything else, divorce, illness, loneliness. Their conclusion, in the paper's own words: "high income buys life satisfaction but not happiness, and low income is associated both with low life evaluation and low emotional well-being."

You can guess what happened next, because you live in the world where it happened. The finding escaped the lab, and the caveats did not make it over the wall. "Money stops buying happiness at $75,000" became a headline, then a factoid, then furniture — quoted in salary negotiations, graduation speeches, and a thousand personal-finance columns. In 2015 a Seattle CEO named Dan Price made the news by citing the study as he set a $70,000 minimum salary at his company. The factoid was comforting in a very particular way: it told everyone below the line that relief had a street address, and everyone above it that the rich had nothing on them. It was also strangely frozen in time — quoted for over a decade at exactly $75,000, though the figure was in 2010 dollars and inflation alone would push it well past $100,000 today. A number that never needs updating is usually a number that has stopped being a measurement and become a proverb.

Notice what the paper never said. It never said money stops mattering — in the very same dataset, life evaluation climbed without limit. It said one dial flattens. And even that claim came with a technical detail almost nobody quoted: the income axis on those charts wasn't marked off in equal dollar steps. It doubled at every tick. Hold that thought; it turns out to be half the story.

The other half arrived by iPhone.

An App, a Rival, and a Referee

Killingsworth's method — experience sampling, the descendant of the beeper studies you met in the last lesson — has one great advantage over asking about yesterday: there is no yesterday. The app catches you inside the moment itself, no memory required. And instead of yes-or-no questions, it offers a continuous slider from "very bad" to "very good" — a dial with hundreds of positions rather than two.

In 2021, Killingsworth published what all those interrupted moments added up to: experienced well-being rose with the logarithm of income — smoothly, steadily, and straight past $75,000 with no bend, no kink, no plateau. Measured in the moment, on a fine-grained scale, more money went with better-feeling days as far up as his data could see.

Two famous findings, flatly contradictory. And here the story does its most unusual thing. Kahneman had spent twenty years arguing that when scientists disagree, dueling publications are a terrible way to find truth: each side writes for its own supporters, nobody concedes, and the argument outlives the arguers. He'd helped pioneer an alternative called adversarial collaboration, in which rivals design a joint analysis, agree in advance on what would settle the question, and publish together whatever comes out — a format that requires each party to admit, in writing, before the data arrives, what would prove them wrong. Now, at the end of his career, he took his own medicine. Killingsworth and Kahneman teamed up, with psychologist Barbara Mellers — who had refereed one of Kahneman's first adversarial collaborations two decades earlier — as arbiter.

Their joint paper appeared in March 2023, title unusually plainspoken for a scientific journal: "Income and emotional well-being: a conflict resolved."

The resolution had two parts. First: for most people, Killingsworth was right. Happiness rises with income on that log scale, with no plateau anywhere in the data — and for the happiest third, the rise actually accelerates above $100,000. But second, hiding in the same data, a group Kahneman's instincts had half-detected: the unhappiest 15 to 20 percent. For them, money helps briskly up to about $100,000 — and then stops. Whatever is causing their misery above that point — grief, depression, loneliness, heartbreak — is not for sale, and further income cannot touch it. There was a plateau after all. It just belonged to an unhappy minority, not to everyone.

So why had the 2010 study seen a universal plateau? The autopsy turned up a ceiling effect — a flaw not in the people but in the ruler. Those yes/no questions were superb at detecting misery, but nearly everyone above modest incomes answered them the same way: yes I smiled, no I wasn't sad. It was like measuring basketball players with a ruler that ends at six feet and concluding that human height plateaus at six feet. The instrument had run out of scale, not the happiness.

Kahneman was eighty-eight when the correction appeared, with his name on it — a man who spent his career cataloguing overconfidence, closing it with a public demonstration of the alternative. He died a year later. It is hard to think of a better final lesson.

Doublings, Not Dollars

Now the part that applies to your paycheck: what does "rises with log income" actually mean?

It means the straight, steady line in all these studies is only straight because of how the income axis is drawn. The ticks don't go $30,000, $60,000, $90,000, $120,000. They go $30,000, $60,000, $120,000, $240,000 — each tick doubles. Happiness rises by roughly the same amount for each doubling of income, not for each additional dollar. Money buys happiness at a price that doubles every time you buy.

Make it concrete. A barista earning $30,000 gets a $30,000 raise: her income has doubled — one full step up the happiness curve. A banker earning $300,000 gets the same $30,000: that's a tenth of a doubling, a nudge he may not feel. For the banker to gain what the barista gained, he needs another $300,000. Same dollars, wildly different yield — which is why "does money buy happiness?" was always the wrong question. The right one is whose money, and which dollars.

Two morals fall out of the arithmetic. First, money matters most exactly where it is scarcest — the same dollar produces more well-being at the bottom of the income ladder than at the top, a fact with consequences for how a society arranges itself, not just how you negotiate a raise. (Module 6 will pick that thread up.) Second, there is no finish line — no magic number where money stops working — only steps that get exponentially more expensive. The treadmill from Lesson 1.2 didn't disappear; it got measured.

And now keep the whole thing in proportion, because this is where headlines fail in the other direction. The income effect is real, durable, and modest. Across all these enormous datasets, differences in income explain only a small slice of the happiness differences between people; you will meet factors in Lesson 5.4 — the quality of your closest relationships, above all — that dwarf it. Killingsworth himself has been careful to say that money is one contributor among many, not the secret. If money is a happiness lever, it is a genuine lever with a short arm.

There is one more variable, though, and it's the one you control this week: not how much, but how spent. Three findings, with honest sizes attached.

Experiences over things. Since Leaf Van Boven and Thomas Gilovich's 2003 studies ("To Do or to Have?"), research has repeatedly found that experiential purchases — trips, meals, concerts, lessons — tend to yield more happiness per dollar than material ones. The proposed reasons fit everything you've learned: possessions sit in the corner and become wallpaper (adaptation again), while experiences end before you can adapt to them, turn into stories, improve with retelling, resist side-by-side comparison, and usually include other people. Caveats, because you were promised caveats: the effect is modest, not magical; the boundary blurs (is a guitar a thing or ten years of Saturday afternoons?); the advantage is strongest for social experiences and for people whose basics are already covered; and larger, stricter replications find the gap smaller than the early studies did. That settles Self-Test item 4 from orientation: true-ish — mark yourself right, with an asterisk.

Others over yourself. In a 2008 experiment, Elizabeth Dunn and colleagues handed people $5 or $20 and an instruction: spend it by tonight, on yourself or on someone else. That evening, the prosocial spenders were happier — and the amount made no difference. Prosocial spending has since been correlated with happiness in Gallup data across more than a hundred countries, rich and poor; large preregistered replications have shrunk the effect, but a small, stubborn signal survives.

Time over stuff. People who spend money to buy back time — paying to skip the chore, shorten the commute, outsource what they dread — report higher life satisfaction than equally rich people who don't, and in experiments a $40 time-saving purchase beat a $40 material one for that day's mood. Researchers call the chronic rushed feeling it treats time famine. The strange part: even people who can easily afford it rarely spend this way.

None of this transforms a life; every effect here is a lean toward the margin. But if the money question has a practical answer, it's this: the exchange rate between dollars and happiness is not fixed — and you're the one setting it. A king in Jerusalem once ran the unlimited-budget version of this experiment, you may recall, and filed his report: vapor. The modern data is gentler than Qohelet — but not by as much as the ads assume.

Spotlight

The Lab: Killingsworth, Kahneman & Mellers, "Income and Emotional Well-Being: A Conflict Resolved" (2023)

The question. Two rigorous studies said opposite things: does day-to-day happiness plateau around $75,000 (Kahneman & Deaton, 2010), or rise with log income without limit (Killingsworth, 2021)?

The setup. Instead of dueling publications, the two principals ran an adversarial collaboration — a joint reanalysis of Killingsworth's experience-sampling dataset (over 33,000 employed U.S. adults, roughly 1.7 million real-time mood reports), with Barbara Mellers as agreed arbiter and the analyses specified before anyone knew the results. Published in PNAS in March 2023.

What they found. For the large majority, experienced well-being rises with log income with no plateau — and for the happiest respondents the slope steepens above $100,000. For the unhappiest 15–20 percent, however, happiness rises with income only up to about $100,000 and then goes flat: beyond that point their distress stops responding to money. The 2010 plateau was diagnosed largely as a ceiling effect — its yes/no emotion items couldn't distinguish degrees of feeling good, so nearly everyone above middle incomes maxed out the scale.

The fine print. One country, and a tilted sample: employed American adults who owned iPhones and volunteered — younger, better educated, and better paid than average, with the very rich barely represented (this study says nothing about billionaires). Income was self-reported in bands. The design is correlational: it cannot show that money causes happiness — happier people may earn more, and health or temperament could drive both. And the effect, though real across a huge sample, is modest: income differences account for a small share of the happiness differences between people. The paper's method may matter as much as its finding — two rivals agreeing in advance on what would count as an answer, then publishing it together, is what self-correcting science looks like when it works.

Myth check

“Money stops improving happiness after about $75,000 a year.”

False

And you can now settle item 3 on the Self-Test. The 2010 study behind the factoid found a plateau in one dial only — yesterday's yes/no emotions — while life evaluation climbed without limit in the same dataset; the 2023 adversarial collaboration then showed even that emotional plateau was mostly a measurement ceiling. The current best evidence: average happiness rises with each doubling of income, with no known stopping point — except for an unhappy minority, roughly one person in six, whose remaining misery stops responding to money around $100,000, because what hurts them isn't for sale. What survives of the factoid is its modest kernel: returns per dollar really do shrink as you climb. "The rich are no happier past $75k" is folklore. "Money's help is real, modest, and logarithmic" is the finding.

Echoes Today

The $75,000 factoid still circulates — in salary-negotiation columns, personal-finance videos, and dinner-party wisdom — usually unadjusted for inflation and unaware it was corrected by its own author. Meanwhile, every "treat yourself" ad is a wager that you'll keep spending in the low-yield column: on things rather than experiences, on yourself rather than anyone else, on objects that save you no time. The data's spending advice and the algorithm's spending advice point in opposite directions, which is worth knowing about the algorithm.

Key Takeaways

The Easterlin paradox (1974): richer people report more happiness within a country, yet national income growth long appeared not to raise national happiness — a founding puzzle that remains genuinely contested today.
Kahneman & Deaton (2010) split happiness into two dials: life evaluation rose with income indefinitely, while day-to-day emotional well-being appeared to plateau near $75,000. Headlines kept the number and dropped the dials.
Killingsworth (2021) found no plateau with finer instruments, and the 2023 adversarial collaboration resolved the conflict: happiness rises with log income for most people — except an unhappy minority (~15–20%) whose misery stops responding around $100,000. The old plateau was largely a ceiling effect in the measures.
Money's returns arrive per doubling, not per dollar — the same raise means far more to a barista than a banker — and the overall effect is modest next to factors like relationship quality.
How you spend shifts the exchange rate at the margin: experiences modestly beat things, spending on others modestly beats spending on yourself, and buying time helps more than people predict.

The Ledger

This lesson pushes back on the treadmill's bleakest reading: circumstances do move the needle — modestly, logarithmically, and most powerfully near the bottom. If your working definition has no room for material conditions, it now needs some; if it quietly assumed a salary at which happiness arrives, strike the number. And note what the unhappy minority adds: part of happiness can be bought at diminishing rates, and part of it is conspicuously not for sale.

Try this week

The Two-Purchases Experiment

Two small purchases you'd probably make anyway, plus about fifteen minutes of journaling across the week

Hypothesis. Before spending anything, write your prediction: which purchase will leave the longer afterglow — the one aimed at you, or the one aimed at someone else? By a lot or a little? Commit now, so the data can argue with you.

Protocol. Pick a small amount you can spend twice without wincing — five to fifteen dollars, or your local equivalent. Once this week, spend it on yourself: something genuinely enjoyable, not toothpaste. Once, the same amount on someone else: cover a coworker's coffee, send a small unexpected gift, feed a stranger's parking meter, give it to a cause you actually care about. It must be freely chosen — an obligatory birthday present doesn't count. The evening after each purchase, rate the afterglow from 1 to 10 in your journal, plus three words. The next morning, rate it again.

What to notice. Whether the other-directed purchase felt awkward beforehand — it often does, and the original study found people mispredict exactly this. Whether the two afterglows fade at different speeds. Whether the recipient's reaction mattered, or the act itself was enough.

Debrief. Compare the results with your hypothesis. Then be a good scientist about your own experiment: you knew the hypothesis going in, so expectation could be doing some of the work, and two purchases prove nothing. In the 2008 study, the amount made no difference — only the direction. Did it for you? That's not a finding. It's a datum about the only subject in your trial.

Check yourself

1. What did Kahneman & Deaton (2010) actually find?

A. Money stops affecting happiness entirely once income passes $75,000, after which extra dollars add nothing.
B. Life evaluation and day-to-day emotion both plateau at around $75,000.
C. Life evaluation rose with income indefinitely, while their measures of day-to-day emotion leveled off around $75,000.
D. Money makes no measurable difference to happiness at any income, rich or poor.

2. A barista earning $30,000 and a banker earning $300,000 each receive a $30,000 raise. What does the log-income finding predict?

A. Equal happiness boosts — a dollar is a dollar, whoever receives it.
B. A much bigger boost for the barista: the raise doubles her income but adds only a tenth of a doubling to his.
C. A bigger boost for the banker, who has far more experience converting money into enjoyment and more ways to spend it well.
D. No boost for either — income and happiness are unrelated, so no raise of any size changes how anyone feels in daily life.

3. What did the 2023 adversarial collaboration conclude?

A. Kahneman was right all along: happiness plateaus at $75,000 for everyone, and extra income adds nothing to anyone's well-being beyond that.
B. Killingsworth was right all along: money helps everyone equally, without limit, and the supposed plateau was only an illusion of the data.
C. Neither was right: once the data were pooled, money turned out to have no reliable link to happiness at any level of income, for anyone.
D. Happiness rises with log income for most people — but for an unhappy minority, money stops helping around $100,000.

4. (Callback to Module 1) Lesson 1.2 corrected the famous Brickman lottery study with a much larger Swedish lottery study (Lindqvist, 2020). What did that study find, and which side of this lesson's story does it fit?

A. Winners fully returned to baseline within a year — supporting a universal happiness plateau at moderate incomes, with no lasting effect from the money at all.
B. Winners ended up less happy than non-winners — money backfires and ruins the lives of those who get it.
C. Winners showed modest but durable gains in life satisfaction lasting over a decade — consistent with money's real, lasting, and modest effect.
D. Winners were dramatically transformed — evidence that income is the main driver of happiness.

Show answers & explanations

1 — C. Option A is the factoid that escaped the lab — see The Most Misquoted Number in the Field. The study's own data showed life evaluation rising to the top of the income range; only the yes/no emotion measures flattened, and the 2023 reanalysis traced most of that flattening to a ceiling effect in the instrument. Option D is the overcorrection: every dataset in this lesson shows a real income–happiness link.

2 — B. The income axis in all of these studies doubles at each tick, and happiness climbs by roughly equal steps per doubling — see Doublings, Not Dollars. A $30,000 raise is a full doubling for the barista and a tenth of one for the banker; to gain what she gained, he would need roughly another $300,000. Option A is the intuition the log scale replaces; option C has no support in the data; option D overreads even the Easterlin paradox, which always granted that richer individuals report more happiness than poorer ones.

3 — D. Options A and B are what adversarial collaboration exists to prevent: total victory declared by one side. The joint reanalysis found Killingsworth right for the majority — no plateau, and acceleration above $100,000 for the happiest — while vindicating a sliver of Kahneman's finding in the unhappiest 15–20 percent, whose distress stops responding to income. See An App, a Rival, and a Referee and the Spotlight, fine print included.

4 — C. The pop version of Brickman (option A) was already corrected in Lesson 1.2: adaptation is real but incomplete, and the tiny 1978 sample never showed full return to baseline. Lindqvist's large Swedish study found winners' life satisfaction modestly but durably elevated more than a decade after the win — which slots neatly into this lesson: money moves the dial, the movement lasts, and it is smaller than daydreams predict. Options B and D are the two fairy tales — money as curse, money as cure — that the actual data politely declines to tell.

Journal Prompt

Write down your number — the annual income at which, you suspect, you'd finally stop worrying and be fine. Look at it next to what you earn now. Is it, by any chance, somewhere between one-and-a-half and two times your current income? Now apply this lesson's scale: if you got there, you'd be one doubling up the curve — a real but modest step — and standing at a new income from which a new number would likely suggest itself. Write a paragraph on what, specifically, you imagine that money buying. How much of the list is things? How much is time, security, or other people? And how much of it could tonight's version of you start buying at the current exchange rate?

Coming Up

The money debate quietly assumed we know who you are. But there are two of you — one living your life minute by minute, the other telling its story afterward — and they disagree about almost everything: vacations, endings, even how bad a colonoscopy was. Next: which one of you is supposed to be happy?