Higher Earners Are Happier, But Not At Work

The relationship between money and happiness is a topic of enduring popular interest and a major focus of scholarly research. That research consistently finds that income and happiness are positively associated (1–12). In other words, people with more money tend, on average, to be happier.

This upward trajectory between larger incomes and greater happiness appears to continue well beyond $75,000/y (1, 2), a point at which some forms of happiness were once thought to plateau (7). More recent evidence shows this upward trajectory likely continues well beyond $500,000/y (3). Analysis of the shape of this upward trend finds that happiness rises almost precisely linearly, on average, with log(income). And while this linear-log association implies that extra dollars matter less to happiness when you have more of them, real-world incomes offset this by growing exponentially in size, yielding a steady and non-diminishing rise in happiness with larger real-world incomes (4). Studies of the causal relationship between money and happiness show that this is not merely a correlation: having more money does indeed cause people to be happier (13–16).

But this description leaves a major question unanswered. While much of research to date has focused on whether higher-income people are happier, comparatively little is known about when in daily life this extra happiness actually appears.

This distinction is crucial: averages tell us that money and happiness are related, but only by understanding the situations in which money does or does not predict greater happiness can we understand (a) why money and happiness are associated in the first place, and (b) how to navigate real-world decisions about money and happiness.

The experience sampling method, which has been called the “gold standard” for measuring happiness (17), provides an opportunity to answer this question. With experience sampling, people’s happiness is measured repeatedly in real-time as they go about their daily lives, along with other contextual details of their experience at that moment. In the case of money and happiness, experience sampling can reveal not just overall differences between richer and poorer individuals’ happiness, but also how their happiness varies by situation, such as what they are doing (e.g., eating, working, or watching TV), or where they are located (e.g., work vs. home).

A particular focus of this report is to analyze the relationship between happiness in different situations and an individual’s personal income, in contrast to most research which analyzes only household income. The results for what I will call “non-personal income”, household income the person does not themselves earn, will be briefly compared.

Method

Participants were 29,138 employed adults living in the U.S. from the large-scale experience-sampling project trackyourhappiness.org who collectively provided over 1.8 million real-time reports of their experience. Participants were signaled at random times during their waking hours and asked to report on their experience at the moment just before the signal, including how they felt, what they were doing, where they were located, and other details. Happiness was measured with the question “How do you feel right now?” on a continuous scale ranging from Very bad to Very good (coded as -50 to +50). See the Appendix for details of the measures analyzed in this report.

This report focuses mainly on situational variation in the context of different activities. For instance, when it refers to work it will typically refer moment’s in which work was the person’s primary activity, wherever it is performed, although work as a location is also discussed.

Like virtually all research on income and happiness, income was log transformed prior to analysis (for an in-depth analysis and discussion of how happiness relates to the logarithm of income, see (4)).

The effect estimates that follow are based on multilevel linear regressions that predict happiness as a function of log(income), account for repeated responses from each person, and control for other demographic variables, including age, gender, marriage, and education level, unless otherwise noted, and income refers to personal income, unless otherwise noted.

Results

We can start with a basic question: are people who personally earn more money happier? This is the first report I am aware of to assess the relationship between personal income and experienced happiness (i.e., happiness as measured with experience sampling), so it is a question worth addressing. The answer? People with larger personal incomes were indeed happier (P < 0.00001).

However, there were significant differences in when higher earners were happier. The most fundamental difference emerged at work. Specifically, the association between happiness and personal income depended significantly on whether people were working at the moment their happiness was measured (P < 0.00001). In a multilevel linear regression predicting happiness as a function of income, people with larger personal incomes were significantly happier when not working (b = 0.92, P < 0.00001), but no happier when working (b = -0.09, P = 0.40); see Fig. 1. Somewhat ironically, the higher a person’s income was, the more time they spent working (r = 0.37 when computed across all incomes, and r = 0.20 when excluding incomes below $20,000/y, both P’s < 0.00001), which was precisely when earning more money was not associated with greater happiness.

Fig. 1. Happiness estimated by a multilevel regression as a function of personal income when people are working (blue) and not working (black). Income axis is log-scaled. Happiness values control for other demographic variables (age, education, gender, and marriage).

A detailed breakdown by activity shows that out of 22 activities, almost every activity besides work exhibited increased happiness in people with larger incomes, see Fig. 2 (see Fig. S1 for an alternate figure with 95% confidence intervals and an activity frequency graph). When engaged in activities besides work, people with higher incomes tended to be happier, although how much happier they were varied by activity. Some of the activities that exhibited the most pronounced increase in happiness with larger incomes included leisure activities such as “listening to music”, “playing”, and “relaxing” as well as certain household activities including “taking care of your children” and “preparing food”.

Fig. 2. The association between happiness and personal income during different activities. Point size indicates the frequency of each activity. Values indicate the slope of the relationship between happiness and Log(personal income) during each activity, controlling for other demographic variables. Work was the most frequent activity and happiness during work did not increase with income.

Income also predicted differences in happiness by location (At Work, At Home, In a Car/Vehicle, Other). Higher earners were significantly happier than lower earners in every location except work, see Fig. 3. The largest gains in happiness was observed in “Other” locations, followed closely by “At Home”. Perhaps “Other” locations reflect the kinds of places and activities where money can be most directly converted into happiness (e.g., restaurants, entertainment, recreation). Time spent “In a Car/Vehicle” barely improved with income, and was much closer to zero than to “At Home” or “Other” locations in its association between income and happiness.

Fig. 3. The association between happiness and personal income in different locations. Values indicate the slope of the relationship between happiness and Log(personal income) in each location, controlling for other demographic variables. Error bars represent 95% confidence intervals.

Why does happiness rise in most situations but not when working? There are at least two potential explanations.

One potential explanation is that money simply doesn’t matter for happiness at work. For instance, you can straightforwardly buy a nicer home or a nicer vacation, but not a nicer boss.

Another potential explanation is that offsetting positive and negative effects of a higher-paying job average out to approximately zero at work.

For instance, all else equal, less pleasant jobs must pay a higher salary to attract workers, a concept discussed by Adam Smith (18) and referred to in modern economics as a “compensating wage differential”. This explains why a job like collecting trash might pay surprisingly high wages. Perhaps lagging happiness at work results from the fact that higher income jobs, all else equal, tend to be more unpleasant, while that unpleasantness is offset by the benefits of a higher salary.

On the other hand, higher-income jobs could potentially convey various advantages, such as more autonomy, prestige, creativity, impact, or other desirable characteristics. Perhaps such advantages exist and genuinely benefit happiness at work when they are present, but are offset in aggregate by various disadvantages that also sometimes accompany higher-paying jobs, leading to the (approximately) net zero relationship overall between income and happiness at work overall?

A way to gain some insight into these competing explanations of the work-nonwork divergence is to examine its relationship to “non-personal income”, which is computed here as total household income minus personal income for the roughly half of the participants who reported household income in excess of personal income. In other words, how much money does the household earn that is not earned by the person whose happiness is being measured?

If money simply doesn’t matter for happiness at work, then we might expect non-personal income to show the same pattern as personal income. On the other hand, if earning a higher income entails a cost to happiness while working, then we might expect money you don’t earn (non-personal income) to show a more positive association with happiness while working.

Results show that, unlike personal income, larger non-personal income was positively related to both happiness at work (P < 0.00001) and happiness outside of work (P = 0.02). Interestingly, the slope was steeper for happiness at work (P < 0.00001).

A related way to understand this is to hold household income constant, and analyze the association between happiness at work and the fraction of household income an individual personally earns (i.e., personal income divided by household income). Sure enough, the more of one’s household income an individual personally earns, the lower their happiness is when working (P = 0.0016). Interestingly, when summed across all situations (i.e., work and nonwork), personally earning a larger fraction of household income did not reliably predict higher or lower happiness overall.

Money a person does not themselves earn can in principle provide the benefits of money without the costs of a higher paying job. The fact that larger non-personal income was associated with increased happiness at work while personally earning a larger fraction of one’s household income was associated with decreased happiness at work are both consistent with the possibility that higher income jobs come at a cost to happiness during work, which is then offset by the benefits associated with the income they provide, leading to the average “effect” of personal income of approximately zero. It is also possible that larger non-personal incomes allow people more freedom to prioritize their happiness at work, perhaps by taking jobs they personally like (even if the salary is lower), quitting jobs they dislike, and so on, which one could argue is just a more circuitous way to for a higher income job to impose a cost on happiness at work, by comparison.

The fact that non-personal income had such a robust relationship to happiness when working seems to suggest that the second explanation is operative to at least some degree, although both explanations (and others) could simultaneously be true.

The results so far have been based on assessing the slope of a linear relation between log(income) and happiness, which is an extremely good fit for data on happiness overall (e.g., for household income, see (4)). Next, let us more closely inspect the trend, and we will see something interesting (see Fig. 4). As the figure shows, happiness when not working trends linearly upward with income. Happiness during work, however, appears approximately flat or trends slightly downward for almost all of the income distribution. As already reported, a linear slope fit to the entire income range was not significantly different from zero (i.e., “flat”). But there appears to be a potential exception: people who earned very high incomes (the top three income points in Fig. 4, which display the top 3% of personal incomes and correspond to personal incomes above $200k) were all directionally happier during work than any lower income level.

Fig. 4. Happiness plotted as a function of personal income when people are working (blue) and not working (black). Point size indicates income frequency. Income axis is log-scaled. Happiness values control for other demographic variables (age, education, gender, and marriage).

To test whether this represents a statistically significant rise in happiness, we can compute a regression as before, predicting happiness during work from log(income) alongside demographic covariates, but add a binary indicator variable that distinguishes incomes above vs. below $200k. The result? The indicator variable is positive and statistically significant during work (P < 0.0001). This shows that people who earned very high incomes were indeed disproportionately happier at work, representing a deviation from the trend estimated by a single linear slope (during non-work activities, the same indicator was not significant, P = 0.6). In fact, once the indicator variable is included, the remaining linear slope in happiness with log(income) at work was modestly but significantly negative, b = -0.25, P = 0.02. Similarly, if the top 3% of incomes are excluded from analysis, a linear model finds that larger log(income) is associated with reduced happiness at work (P = 0.01).

To summarize, it appears that, for the small group of people who earned the highest 3% of incomes, they were both happier outside of work (consistent with the overall upward trend outside of work) and happier at work as well (deviating from the otherwise flat to slightly negative association between income and happiness at work). For the other 97% of the income distribution, larger incomes were associated with steadily higher happiness outside of work but with flat or perhaps reduced happiness while working.

What might explain why this small group of high earners was happier at work? It is impossible to know for sure, but one factor which showed a congruent pattern was people’s sense of control. In previous research, I found that people’s sense of control over their lives was the strongest explanation for why people with larger incomes are happier; it could account for 74% of the association between larger incomes and greater happiness (1).

To evaluate a similar possibility here, I turn to a measure of people’s momentary sense of control, since control in this case would need to be measured specifically while working. Just like happiness, how in control people felt while working did not rise with income when estimated by a linear slope fit to the entire income range (P = 0.65). But adding a binary indicator variable above $200k shows that high earners significantly increase in their sense of control compared to people at lower incomes (P = 0.016). So one possibility is that people with very high incomes experience an increased sense of control while working, and that this may explain why this exceptional group is happier at work, contrary to the otherwise flat-to-downward trend.

Discussion

People who earn more money are, on average, happier. But that extra happiness does not appear everywhere in daily life. These results, based on over 1.8 million real-time reports from more than 29,000 employed adults, find that higher personal income predicted greater happiness in nearly every activity except work, which was the single activity that occupied the largest share of their waking lives.

This pattern helps clarify both why money and happiness are related and perhaps why they are not more strongly so. A larger personal income was associated with greater happiness when people were doing things that money can directly improve such as leisure or family activities, but not when they were working to earn it. For most people, then, higher income appears to enhance life outside of work, but brings little net benefit to experience at work itself.

Two complementary explanations may account for this pattern. First, money may simply matter less for happiness at work: one can purchase a better home or vacation, but not a better boss or fewer meetings. Second, offsetting forces may cancel each other out. High-paying jobs can offer more autonomy, influence, or prestige, yet also more stress, responsibility, or longer hours. Money itself might offer benefits at work, for instance, by providing more financial flexibility in choosing one’s job or when navigating job demands, yet, all else equal, unpleasant jobs may tend to offer greater pay to attract workers. Opposing factors like these may yield the flat or potentially negative association between personal income and happiness at work observed here, especially when summed across people who might individually experience an unequal mix of costs and benefits of a larger income.

Evidence from “non-personal” income reinforces the latter interpretation. Household income that participants did not personally earn was positively associated with happiness during both work and nonwork, suggesting that income itself can provide benefits at work when it does not come bundled with the costs of earning it. Moreover, when total household income was held constant, individuals who personally earned a larger fraction of it were significantly less happy while working, consistent with the idea that higher-paying jobs involve trade-offs between income and happiness during work.

There was one notable exception. Compared to lower earners, the highest-earning 3% of participants (>$200,000 per year) were modestly but significantly happier while working. These individuals also reported a greater sense of control while working, echoing earlier evidence from my research that perceived control can explain much of the overall income:happiness link. For the general population, a larger income is associated with greater freedom and autonomy in life overall; for the very top of the income distribution, it appears to be associated with greater sense of autonomy at work, as well.

A virtue of this research worth reiterating is the use of experience sampling. When these results refer to happiness during work, for example, it is based on repeated, real-time reports of happiness in the moments in which a person is actually working. At the same time, it is worth keeping in mind that these results reflect real-world associations, not the outcome of a controlled experiment. Direct tests of causation (such as very large-scale random assignment to jobs or income) would be needed to draw definitive conclusions about the causal relationship between income and happiness. Additionally, these results reflect averages across many people. There could be people or occupations or particular circumstances that trend in ways that deviate from the patterns described here.

Many people would probably like to have a high income, and there are at least two possible reasons why: because life outside of work is better when you have more to spend, or because higher income jobs are themselves more enjoyable. These results offer robust support for the first reason but, for all but the highest incomes, a lack of support for the second. Earning more money was associated with greater happiness in contexts where money is spent, but not where it is earned.

Cite this

Killingsworth, M. A. (2024). Higher earners are happier, but not at work. Happiness Science. https://happiness-science.org/income-happiness-situation

References

1. M. A. Killingsworth, Experienced well-being rises with income, even above $75,000 per year. Proceedings of the National Academy of Sciences 118, e2016976118 (2021).
2. M. A. Killingsworth, D. Kahneman, B. Mellers, Income and emotional well-being: A conflict resolved. Proceedings of the National Academy of Sciences 120 (2023).
3. M. A. Killingsworth, Money and Happiness: Extended Evidence Against Satiation. Happiness Science (2024). https://happiness-science.org/money-happiness-satiation
4. M. A. Killingsworth, The Price of Happiness. Happiness Science (2024). https://happiness-science.org/price-of-happiness
5. B. Stevenson, J. Wolfers, Economic growth and subjective well-being: Reassessing the Easterlin Paradox. Brookings Papers on Economic Activity 2008, 1–87 (2008).
6. B. Stevenson, J. Wolfers, Subjective Well-Being and Income: Is There Any Evidence of Satiation? American Economic Review 103, 598–604 (2013).
7. D. Kahneman, A. Deaton, High income improves evaluation of life but not emotional well-being. Proceedings of the National Academy of Sciences of the United States of America 107, 16489–93 (2010).
8. A. Deaton, Income, Health, and Well-Being around the World: Evidence from the Gallup World Poll. Journal of Economic Perspectives 22, 53–72 (2008).
9. M. Argyle, “Causes and correlates of happiness” in Wellbeing: The Foundations of Hedonic Psychology, D. Kahneman, E. Diener, N. Schwarz, Eds. (Russell Sage Foundation, 1999), pp. 353–373.
10. R. Veenhoven, M. Hagerty, Rising Happiness in Nations 1946–2004: A Reply to Easterlin. Social Indicators Research 79, 421–436 (2006).
11. R. Layard, S. Nickell, G. Mayraz, The marginal utility of income. Journal of Public Economics 92, 1846–1857 (2008).
12. E. Diener, S. Oishi, Money and happiness: Income and subjective well-being across nations. Culture and subjective well-being 185–218 (2000).
13. E. Lindqvist, R. Östling, D. Cesarini, Long-Run Effects of Lottery Wealth on Psychological Well-Being. The Review of Economic Studies 87, 2703–2726 (2020).
14. J. Gardner, A. J. Oswald, W. W. Worldwide, Money and Mental Wellbeing: A Longitudinal Study of Medium-Sized Lottery Wins. 1–26 (2006).
15. S. Kim, A. J. Oswald, Happy Lottery Winners and Lottery-Ticket Bias. Review of Income and Wealth 67 (2021).
16. R. J. Dwyer, E. W. Dunn, Wealth redistribution promotes happiness. Proceedings of the National Academy of Sciences 119 (2022).
17. D. Kahneman, A. B. Krueger, D. Schkade, N. Schwarz, A. A. Stone, A Survey Method for Characterizing Daily Life Experience: The Day Reconstruction Method. Science 306, 1776–1780 (2004).
18. A. Smith, An Inquiry Into the Nature and Causes of the Wealth of Nations (Strahan, 1776).

Appendix

Fig. S1. The association between happiness during an activity and personal income. Values indicate the slope of the relationship between happiness and Log(personal income) during each activity, controlling for other demographic variables. Work was the most frequent activity and happiness during work did not increase with income. Error bars represent 95% confidence intervals, and bar length indicates relative frequencies of different activities.

Materials and Methods

Sample information. Participants were 29,138 employed adults living in the U.S. Mean age was 32, mean personal income was $65,100/year, 64% were female, and 35% were married. To better understand the association between personal income and happiness across the situations of daily life, participants were included if they were employed adults living in the U.S. of working age (18-65), if they reported their personal income, age, and marital status, and provided at least 10 reports of happiness, including at least one report while working and at least one report while not working.

Procedure. After providing informed consent, participants completed an intake survey, which included demographic questions, as detailed below. Participants were next asked to indicate the times at which they typically woke up and went to sleep, and how many times during the day they wished to report on their experiences (default = 3). A computer algorithm then divided each participant’s day into a number of intervals equal to the number of desired reports, and a random time was chosen within each interval. New random times were generated each day, and the times were independently randomized for each participant. At each of these times, participants were signaled via a notification on their smartphone, asking them to respond to a variety of questions about their experiences at the moment just before the signal. The primary happiness question, location question, and activity questions were asked in every survey, while other measures, including the situational sense of control measure were assessed in independently randomized subsets of surveys. Other questions unrelated to the present investigation were also asked. Participants received notifications requesting a report until they chose to discontinue participation. If 50 samples had been collected, reporting stopped for 6 months or until the participant requested that it be restarted.

Measures. Happiness was measured with the question “How do you feel right now?” on a continuous response scale with endpoints labeled “Very bad” and “Very good” (coded -50 to +50).

Primary activity was measured with the question, “What are you doing?” with a list of activity choices. If the user endorsed more than one activity, they were required to identify their primary activity. For results reported here, a moment was categorized as work only if work was the primary activity at that moment.

Location was measured with the question, “Where are you?” with response options “At Home”, “At Work”, “In a Car / Vehicle”, and “Other.”

Hours worked per week was measured with the question, “On average, how many hours do you work each week?” with response options: Less than 5 hours, 5-10 hours, 10-20 hours, 20-30 hours, 30-40 hours, 40-50 hours, 50-60 hours, 60-70 hours, 70-80 hours, 80-90 hours, 90-100 hours, 100+ hours, and This doesn’t apply.

Situational sense of control was measured with the question, “To what extent do you feel in control of your current situation?” with responses collected on a continuous scale ranging from Not at all to Extremely.

Income was measured on an intake survey that occurred prior to and on a different occasion from happiness or other momentary measures such as activity, location or perceived control, which were all via experience sampling reports. Thus, income was not made salient by the study design to participants when they were reporting happiness.

Personal income was measured by asking people, “What is your total annual personal income before taxes?” with response options in $10,000 increments up to $100,000/year, followed by “$100,001-$125,000, $125,001-$150,000, 150,001-$200,000, and over $200,000. The same question was used to measure household income, but replacing the word “personal” with “household” in the question.

If a person selected “over $200,000” then an expanded income range was offered including $200,001 - $300,000, $300,001 - $500,000, $500,001 - $750,000, $750,001 - $1,000,000, $1,000,001 - $2,000,000, $2,000,001 - $4,000,000, $4,000,001 - $7,000,000, $7,000,001 - $10,000,000, $10,000,001 - $20,000,000, $20,000,001 - $50,000,000, $50,000,001 - $100,000,000, and More than $100,000,000.

For analysis and visualization, income values were set to the midpoint of the income range selected, e.g., the income value for the income band $100,001 - $125,000 was set to $112,500. In practice, only 3% of people reported personal incomes above $200,000/year. Personal incomes over $500,000 were very rare, collectively comprising less than 0.5% of the sample, and were pooled together and set to a value of $625,000/year for visualization and analysis (the midpoint of the income band above $500,000/year).

Analysis. Multilevel regression with a random person-level intercept was used for regression analyses using the lmer function from the lme4 package in R. Results described are based on treating different values of demographic covariates as factors, to allow the shape of their association with happiness to freely vary. For those analyses, age was coded into age groups (up to 29, 30-39, 40-49, and 55-65). However, if demographic covariates are treated as numeric variables (which imposes a linear relationship between variables like age or education level and happiness), the basic conclusions were the same. To visualize the appropriate happiness level by income in each context (e.g., as in Fig. 1) while accommodating multilevel regression with covariates (including factor covariates which cannot simply be centered), the plotted estimates were anchored to model-based context means (based on a random-intercept mixed model of happiness, i.e., the model-based grand mean). The intercept was defined so that the weighted mean of the plotted values equals the model-based context mean. This preserves all relative differences across income as predicted by the full multilevel model, while aligning the visualization to the correct grand mean per context.