The article shows a logarithmic trend (i.e. doubling your income increases your happiness by a fixed amount), which is a straight line on the log plot. This insinuates (if the trend continues, which is assumed from the results) that there is no point at which increased income won't make someone happier.

The traditionally-held view is that there is an income (I've heard something like 70k in the USA) where the plot should become a horizontal line, i.e. increasing your income increases your happiness by zero.

These are very different conclusions. The plot isn't really falling "significantly towards the top of the scale". Log is the correct way of looking at income, because people consider changes to their income proportionately to their current income (for example, a 5% raise to someone making 1M dollars a year is about as significant to his life as a 5% raise for someone making 100k, even though the actual additive amounts are very different).

This might seem like a strange view, but the notion is not that money ceases to increase happiness, but rather increased happiness costs associated with earning that increased income start to negate the gains.

Anyway, this graph is self-reported life satisfaction. We would need to examine the metric used in the studies that came up with the ~70k number to see if we can even compare the two studies.

And that's why the cap for Nigeria is at $8k...

In fact, as someone else has commented: https://news.ycombinator.com/item?id=7047548

"According to the paper (page 13), they only look at incomes between the 10th and 90th percentiles. That's why they cut off at $128,000 for US residents"

In my opinion, ignoring data from the top 10% earners is the perfect way to get plots that seem to disprove a paradox that states "Become rich enough, and a bigger paycheque no longer leads to more happiness".

They are not; in the real world, logarithmic growth and bounded growth are essentially the same because the possibilities for income growth are severely limited, and you reach a point where any realistic increase in income cannot yield a measurable increase in happiness.

Firstly, you're generalizing towards the origin from a relatively high section of the graph. Leaving aside the relative happiness, a 5% increase in income is actually much more significant in someone's life when they make $10,000 a year than when they make $100,000. There are thresholds established by absolute resource requirements. This makes it dangerous to reason from the 'flat' part of the curve downwards. Relevant analogy: transistor response curve.

Secondly, insofar as income represents a portion of an absolute resource [ie to some extent], some information is obscured that would otherwise be intuitively and immediately available. It's remarkable that the money required to increase a person's happiness by one unit is, in every case, enough money that another person could share their initial happiness. A graph with a linear relation would demonstrate this as an (immediately obvious) diminishing return on a finite resource. A graph with a log relation presents a diminishing return as natural and efficient. In this way, the choice of scaling method significantly biases the meaning attributed to the data.

Are you asserting that as fact? Where I live, someone making 100K is just getting by and not saving very much, even with the 5% raise (he certainly couldn't buy a house), whereas for someone making 1M, the extra $50K is "free" money.

I've seen many people complain about not making enough to live where they live, when I clearly see lifestyle changes that would fix the income problem, and I see people living alongside the first who make less and save more.

I don't know you. Or your life circumstances. It might not be true for you. But "$100k/year isn't enough to survive" is the kind of claim that I've learned to be automatically suspicious of.

There's a huge difference between 100K in Tennessee vs 100k in NYC or SF.

Let's take NYC. The fact is that in New York City, the median household income is about $50k/year. (Last year it was $50,895.) So if we define a "middle class lifestyle" as that which those in the middle can afford, well, by definition you're going to get by very well following that lifestyle while making $100k/year! Even if your household has 3 people in it.

But if you define "middle class lifestyle" according to the way you think that you peer group is living, well, I've known people making $200k+ who experience "trouble getting by".

Now I can understand making more, choosing to live better, and running into financial trouble because of it. But if you do that and blame anyone but yourself, well, I've got nothing in the way of sympathy for you.

If you can't afford decent housing, adequate healthcare, and a good education for your kids, of course more money will buy you happiness. And in the US, the income at which you can afford those three ... is high and increasing fast.

There is just a vast gap between the point where you stop spending money for consumption and when you can start spending money for influence.

Maybe they don't have data for people making more than that? There are relatively few people with incomes that high, and their time is probably valuable enough that they're disinclined to waste it filling out happiness inventories.

The chart is of household income, not individual income. There are plenty of households who break that mark.

It doesn't - the title of the article and the interpretation provided with in is almost as misleading. They start off quickly by telling you about the 'Easterlin paradox' - that is what they have specifically found evidence against. i.e. 'more money makes you happier' is backed by evidence.

The fact that more money makes you happier does not mean that any amount of money can make you as happy as you want though...

This is not surprising - its probably common sense.

And "more money" is not always the easiest or most available "life hack."

With that said, I do think it'd be interesting to see a graph that continues past the 90th percentile.

The paper: http://www.nber.org/papers/w18992?utm_campaign=ntw&utm_mediu...