Thanks for sharing! This paper, like the American correlational ones, seems to gloss over the smallness of the effect. They find that a $100,000 USD payout increases happiness by 0.016 standard deviations, and life satisfaction by 0.037 SDs.
In the American data, 1 SD is 12.4 points on a 100-point scale.
So we're talking about (roughly) an increase in one-fifth of a happiness point, or one half of a "life satisfaction" point, on a 100-point scale. Not much!
To really compare to the American estimates, it'd be good to convert the $100k into a proportion of income gained. So:
The average annual income in the Swedish study was ~$34k USD. They say "a 100,000 prize corresponds to an increase in net annual income of 5,996." So $100k can be thought of as an accross-time income boost of 17.5% ($6k / $34k.)
So if we scale that, and assume it holds up at higher levels as the authors suggest it does ... let's say the payout were $600k, then the person would basically be getting a doubling of the lifetime income since the payout. In that case, we'd also multiply the SDs by about 6.
So, the new SDs, implied by the Swedish study, regarding a doubling of income: Happiness: .096. Life satisfaction: .222.
Converted to the American 100-point scale data: Happiness increases by *1.2 points* from a doubling, while life satisfaction increases by *2.75 points* from a doubling.
That's actually really, really close to the American correlational estimates! ...With the biggest difference being that the impact on life satisfaction is half of what Greenberg estimates for life satisfaction.
The Swedish data's lower estimate on life satisfaction may not necessarily be contradictory, as the Swedish data are based on lottery winnings, and it's possible that earned income increases "life satisfaction" more.
Nice post. One may also note that those living in destitute third world countries generally appear *much* happier than Americans on average, for whatever reason (more community? more free time? a non-materialist culture?)
I think it varies by third world country, from what I've noticed. But it's certainly true that the difference is less than one would expect. Worth researching further.
I believe there is a fairly strong cross-country relationship between income and happiness, which I believe is also there in the time series data though may be weaker. Obviously, it is just correlation, and there are also cultural differences. But, the fit of the data is not perfect and there of course are poor countries that are happier than rich countries.
The research I've seen, filtered through the popular scientific press, is that once a country gets above "dire poverty", the average reported happiness doesn't increase with average income. But some studies have shown a significant effect of an individual's income compared to the *national* income. That doesn't surprise me at all.
OTOH, neither of those mesh well with the results described in the column, so it seems to be up in the air.
The Greenberg post has more on that, but basically the international comparisons seem to use questionable measures of happiness, so I’ll need to dig before before believing much about them again.
I think this is one where lived experience probably trumps scientific studies on the subject.
I have two views. Basically that it is and isn't important.
I had a two year period where I 2x'ed my income with entrepreneurship and trading. And then in year 3, I 5x'ed my income. I was a single guy who moved to a nicer apartment at the time. It was really great. And then, partly from the FTX collapse, I had to reign in my spending subsequently. And that was a really rough experience. I think I didn't sleep at all for four days after the FTX thing happened, even though it was only a fraction of my net worth. Certainly, changes in income/wealth matter as well as the level.
Another take: in my old building there were these Guatemalan janitors who would be cleaning the toilets and so forth every day, while I was doing my white collar work, academic research and so forth. I was mostly miserable post-FTX, and these ladies were laughing and telling jokes all day long, having the time of their lives while mopping the floor for what was likely minimum wage. Their frame was much more mature than mine. A lot of life's pleasures aren't necessarily that expensive, and many are free. I also buy the "rapidly diminishing marginal returns" to income argument, even if I also believe it's also always positive at the margin.
Interesting, thanks for sharing! My own experience also has a bunch major shifts in that, and, introspecting, I’d say it’s in line with the study data in the post.
Definitely agreed that the housekeepers there show an example of how one can be happy at different levels, which is also what the data show.
Though the Guatemalan janitors might be showing the effects of "what do you compare your life with". If you used to be a peasant in Guatemala subject to political and criminal violence, doing backbreaking agricultural work, cleaning toilets in the US might seem like winning the lottery. There are reports of immigrant/guest workers into rich countries who are treated quite shabbily by the standards of their new countries, but who evaluate their situation based on where they left and are quite happy about it. But e.g. their children raised in the rich country compare their lives to the average of the new country and they're a lot less happy.
Have often questioned this narrative, since there is definitely some arbitrary point where people will give up a higher salary for more leisure time / less stress but that amount seems fairly low and does not make sense for anyone living in a big city, multiple children etc.
True. And maybe their methods don’t pick up on what happens when those worries become real — I doubt many broke cancer patients were using a happiness tracking app.
People with high incomes sometimes live paycheck to paycheck with massive debts. Retirees often have lower incomes than when they were working but much more free time. Did the studies control for anything like this? Income alone doesn't seem granular enough.
>>Looking at US data, every doubling in income is associated with … 6 points on a 100-point scale for life satisfaction, and just 1 point for hedonic wellbeing…
>So for example, if someone’s income goes from $62,500 to $250,000 (two doublings) someone’s life satisfaction would tend to go from, say, 60 to 72.
How do those scales work, what happens at the fringes? At 8M/yr (7 doublings) we'd expect 102 on the 100-point scale (which is not selectable).
So in some sense, the scale itself is exponential: There's probably a bigger difference between people scoring 98 and 99 than people scoring 50 and 51 respectively.
This is difficult to get rid of, since it's hard to compare values if people give scores on an unbounded scale.
The data don't differentiate beyond "above $600k" -- that's the highest bucket. If someone is making $8M/yr, they can take pride in being well beyond the ability of happiness science to evaluate.
Basically, the numbers in my post just don't apply there. Of course someone could theoretically study it, and see what happens at that level, but we don't have data for it currently. Clearly, happiness does not infinitely rise linearly with income, for the reason you give.
Btw, a causal Swedish study linked in the comments, which I just went through, suggests more like a 3-point increase on life satisfaction (though it's based on lottery winnings. It's possible that "earned income" increases satisfaction more, though.) https://www.maximumtruth.org/p/data-dive-money-does-not-buy-much/comment/57415730
I was talking more about the mathematical / statistical properties of the scale - the effect should start long before hitting the boundary. Has it been measured, e.g. how marriage impacts people based on how "satisfied" they already are? Just from the mathematical properties, I'd expect people lower on the scale (let's say 40) get a bigger increase than people already high (70?).
Marriage (or other interventions) becoming less impactful on satisfied people, would be a sign that the focus on money here is overly narrow.
Footnote 1 somehow plays into this, but I can't wrap my head around it.
Shocking "take-aways": 1) Academicians are often incompetent and too narrowly focused. 2) Studies are often poorly designed and gather a lot of inconsequential data. 3) Always remember and apply the genius statement of Samuel Langhorne Clemens regarding statistics.
Statistics are a great tool if the data is carefully gathered and analyzed. Unfortunately this is often not the case. Kind of analogous to: "those that can, do; those that can't, teach" : "those that know explain , those that do not know, publish statistics" When dealing with statistics, the first thing to analyze is the statisticians.
I've long thought something about life satisfaction measures, but I'm not sure I can explain it well. Basically, my thoughts are this: how is life satisfaction beneficial? How does one 'feel' satisfied? In the present moment. The only way in which it can be beneficial is by making you happier in the present moment, as that's all we ever have. The only way that your life not being satisfying, being a total loser, can negatively affect you is by making you feel sad in the present moment. Therefore, it seems like a nonsense measurement to me... it all just comes back to the present moment positive affect measurements. Someone who scores higher than another on the happiness measure will necessarily always be better off than someone who has a lower score, even if they have a billion times the life satisfaction score. It's like saying there's 'happiness' vs 'happiness because of x'. Interestingly, Kahneman thought life satisfaction was more important than happiness, and he's said this is one of the reasons he stopped focusing on happiness. It's not a perspective I understand.
If I remember correctly, life satisfaction in these studies is measured by asking people to sort of place their success in life, or something like that, on a ladder or steps. It's inviting them to compare themselves, in that moment, to the rest of society and consider societal expectations of success. It all seems a bit abstract and perhaps irrelevant compared to the much clearer and measurable 'how do you feel now (or yesterday)?' questions, to me.
Causal data from Swedish lotteries.
https://academic.oup.com/restud/article/87/6/2703/5734654
Thanks for sharing! This paper, like the American correlational ones, seems to gloss over the smallness of the effect. They find that a $100,000 USD payout increases happiness by 0.016 standard deviations, and life satisfaction by 0.037 SDs.
In the American data, 1 SD is 12.4 points on a 100-point scale.
So we're talking about (roughly) an increase in one-fifth of a happiness point, or one half of a "life satisfaction" point, on a 100-point scale. Not much!
To really compare to the American estimates, it'd be good to convert the $100k into a proportion of income gained. So:
The average annual income in the Swedish study was ~$34k USD. They say "a 100,000 prize corresponds to an increase in net annual income of 5,996." So $100k can be thought of as an accross-time income boost of 17.5% ($6k / $34k.)
So if we scale that, and assume it holds up at higher levels as the authors suggest it does ... let's say the payout were $600k, then the person would basically be getting a doubling of the lifetime income since the payout. In that case, we'd also multiply the SDs by about 6.
So, the new SDs, implied by the Swedish study, regarding a doubling of income: Happiness: .096. Life satisfaction: .222.
Converted to the American 100-point scale data: Happiness increases by *1.2 points* from a doubling, while life satisfaction increases by *2.75 points* from a doubling.
That's actually really, really close to the American correlational estimates! ...With the biggest difference being that the impact on life satisfaction is half of what Greenberg estimates for life satisfaction.
The Swedish data's lower estimate on life satisfaction may not necessarily be contradictory, as the Swedish data are based on lottery winnings, and it's possible that earned income increases "life satisfaction" more.
Nice post. One may also note that those living in destitute third world countries generally appear *much* happier than Americans on average, for whatever reason (more community? more free time? a non-materialist culture?)
I think it varies by third world country, from what I've noticed. But it's certainly true that the difference is less than one would expect. Worth researching further.
Relative satisfaction plays a huge role.
A third world country citizen may get happier by getting ahead of his peers (where the bar is lower)
A first world citizen has bigger Joneses to keep up with.
"If you want to make progress, surround yourself with people more successful than you.
If you want to be happy, surround yourself with people less successful than you." -
Comparison is not only the thief of joy, comparison also brings joy.
I believe there is a fairly strong cross-country relationship between income and happiness, which I believe is also there in the time series data though may be weaker. Obviously, it is just correlation, and there are also cultural differences. But, the fit of the data is not perfect and there of course are poor countries that are happier than rich countries.
The research I've seen, filtered through the popular scientific press, is that once a country gets above "dire poverty", the average reported happiness doesn't increase with average income. But some studies have shown a significant effect of an individual's income compared to the *national* income. That doesn't surprise me at all.
OTOH, neither of those mesh well with the results described in the column, so it seems to be up in the air.
If you go to Greenberg’s article, I believe he has charts on that, and they show there isn’t such a cutoff regarding national income and happiness
The Greenberg post has more on that, but basically the international comparisons seem to use questionable measures of happiness, so I’ll need to dig before before believing much about them again.
I think this is one where lived experience probably trumps scientific studies on the subject.
I have two views. Basically that it is and isn't important.
I had a two year period where I 2x'ed my income with entrepreneurship and trading. And then in year 3, I 5x'ed my income. I was a single guy who moved to a nicer apartment at the time. It was really great. And then, partly from the FTX collapse, I had to reign in my spending subsequently. And that was a really rough experience. I think I didn't sleep at all for four days after the FTX thing happened, even though it was only a fraction of my net worth. Certainly, changes in income/wealth matter as well as the level.
Another take: in my old building there were these Guatemalan janitors who would be cleaning the toilets and so forth every day, while I was doing my white collar work, academic research and so forth. I was mostly miserable post-FTX, and these ladies were laughing and telling jokes all day long, having the time of their lives while mopping the floor for what was likely minimum wage. Their frame was much more mature than mine. A lot of life's pleasures aren't necessarily that expensive, and many are free. I also buy the "rapidly diminishing marginal returns" to income argument, even if I also believe it's also always positive at the margin.
Interesting, thanks for sharing! My own experience also has a bunch major shifts in that, and, introspecting, I’d say it’s in line with the study data in the post.
Definitely agreed that the housekeepers there show an example of how one can be happy at different levels, which is also what the data show.
Though the Guatemalan janitors might be showing the effects of "what do you compare your life with". If you used to be a peasant in Guatemala subject to political and criminal violence, doing backbreaking agricultural work, cleaning toilets in the US might seem like winning the lottery. There are reports of immigrant/guest workers into rich countries who are treated quite shabbily by the standards of their new countries, but who evaluate their situation based on where they left and are quite happy about it. But e.g. their children raised in the rich country compare their lives to the average of the new country and they're a lot less happy.
Great article!
Have often questioned this narrative, since there is definitely some arbitrary point where people will give up a higher salary for more leisure time / less stress but that amount seems fairly low and does not make sense for anyone living in a big city, multiple children etc.
Thank you!
It can buy relief from certain worries, which is important
True. And maybe their methods don’t pick up on what happens when those worries become real — I doubt many broke cancer patients were using a happiness tracking app.
People with high incomes sometimes live paycheck to paycheck with massive debts. Retirees often have lower incomes than when they were working but much more free time. Did the studies control for anything like this? Income alone doesn't seem granular enough.
TFR falls up to 300k in household income, then starts to rise pretty sharply.
That's where I would peg the diminishing returns to start kicking in. If you live in a high COL area its at least 500k.
1) Does your wife have to work (at a real job)?
2) Can you afford a decent place in a good neighborhood?
3) Do you kids go to schools you feel good about?
4) Do you have an Au Pair or whatever other "help" smooths out the edges?
5) Do you have "fuck you money" if things at work ever became unbearable?
6) Does any of that change if you have 3+ kids?
Past these limits I could buy that income has sharply demising returns. But 75k won't buy you that.
>>Looking at US data, every doubling in income is associated with … 6 points on a 100-point scale for life satisfaction, and just 1 point for hedonic wellbeing…
>So for example, if someone’s income goes from $62,500 to $250,000 (two doublings) someone’s life satisfaction would tend to go from, say, 60 to 72.
How do those scales work, what happens at the fringes? At 8M/yr (7 doublings) we'd expect 102 on the 100-point scale (which is not selectable).
So in some sense, the scale itself is exponential: There's probably a bigger difference between people scoring 98 and 99 than people scoring 50 and 51 respectively.
This is difficult to get rid of, since it's hard to compare values if people give scores on an unbounded scale.
The data don't differentiate beyond "above $600k" -- that's the highest bucket. If someone is making $8M/yr, they can take pride in being well beyond the ability of happiness science to evaluate.
Basically, the numbers in my post just don't apply there. Of course someone could theoretically study it, and see what happens at that level, but we don't have data for it currently. Clearly, happiness does not infinitely rise linearly with income, for the reason you give.
Btw, a causal Swedish study linked in the comments, which I just went through, suggests more like a 3-point increase on life satisfaction (though it's based on lottery winnings. It's possible that "earned income" increases satisfaction more, though.) https://www.maximumtruth.org/p/data-dive-money-does-not-buy-much/comment/57415730
I was talking more about the mathematical / statistical properties of the scale - the effect should start long before hitting the boundary. Has it been measured, e.g. how marriage impacts people based on how "satisfied" they already are? Just from the mathematical properties, I'd expect people lower on the scale (let's say 40) get a bigger increase than people already high (70?).
Marriage (or other interventions) becoming less impactful on satisfied people, would be a sign that the focus on money here is overly narrow.
Footnote 1 somehow plays into this, but I can't wrap my head around it.
>To solve the conflicting headlines/studies above, there was an adversarial collaboration between Khaneman, [...]
Typo "Khan"
Shocking "take-aways": 1) Academicians are often incompetent and too narrowly focused. 2) Studies are often poorly designed and gather a lot of inconsequential data. 3) Always remember and apply the genius statement of Samuel Langhorne Clemens regarding statistics.
True! Though the anti-stats quotes can get a bit nihilistic. We need to start somewhere to figure things out.
Statistics are a great tool if the data is carefully gathered and analyzed. Unfortunately this is often not the case. Kind of analogous to: "those that can, do; those that can't, teach" : "those that know explain , those that do not know, publish statistics" When dealing with statistics, the first thing to analyze is the statisticians.
I've long thought something about life satisfaction measures, but I'm not sure I can explain it well. Basically, my thoughts are this: how is life satisfaction beneficial? How does one 'feel' satisfied? In the present moment. The only way in which it can be beneficial is by making you happier in the present moment, as that's all we ever have. The only way that your life not being satisfying, being a total loser, can negatively affect you is by making you feel sad in the present moment. Therefore, it seems like a nonsense measurement to me... it all just comes back to the present moment positive affect measurements. Someone who scores higher than another on the happiness measure will necessarily always be better off than someone who has a lower score, even if they have a billion times the life satisfaction score. It's like saying there's 'happiness' vs 'happiness because of x'. Interestingly, Kahneman thought life satisfaction was more important than happiness, and he's said this is one of the reasons he stopped focusing on happiness. It's not a perspective I understand.
If I remember correctly, life satisfaction in these studies is measured by asking people to sort of place their success in life, or something like that, on a ladder or steps. It's inviting them to compare themselves, in that moment, to the rest of society and consider societal expectations of success. It all seems a bit abstract and perhaps irrelevant compared to the much clearer and measurable 'how do you feel now (or yesterday)?' questions, to me.
> So happiness at any given moment does rise with happens, but the effect is small.
Typo
Fixed, thanks.
It now reads:
"So happiness at any given moment does rise with happiness, but the effect is small."
The second "happiness" should be"income".
Fixed, thanks.