Some secondary statistics from the results of LW Survey
| Global LW (N=643) vs USA LW (N=403) vs. Average US Household (Comparable Income) | |||||||||||||
| Income Bracket | LW Mean Contributions | USA LW Mean Contribution | US Mean Contributions** [1] | LW Mean Income | USA LW Mean Income | US Mean*** Income [1] | LW Contributions /Income | USA LW Contributions/Income | US Contributions/Income [1] | ||||
| $0 - $25000 (41% of LW) | $1,395.11 | $935.47 | $1,177.52 | $11,241.14 | $11,326.18 | $15,109.85 | 12.41% | 8.26% | 7.79% | ||||
| $25000-$50000 (17% of LW) | $438.25 | $571.00 | $1,748.08 | $34,147.14 | $32,758.06 | $38,203.79 | 1.28% | 1.74% | 4.58% | ||||
| $50000-$75000 (12% of LW) | $1,757.77 | $1,638.59 | $2,191.58 | $60,387.69 | $61,489.30 | $62,342.05 | 2.91% | 2.66% | 3.52% | ||||
| $75000-$100000 (9% of LW) | $1,883.36 | $2,211.81 | $2,624.81 | $84,204.09 | $83,049.54 | $87,182.68 | 2.24% | 2.66% | 3.01% | ||||
| $100000-$200000 (16% of LW) | $3,645.73 | $3,372.84 | $3,555.02 | $123,581.28 | $124,577.88 | $137,397.03 | 2.95% | 2.71% | 2.59% | ||||
| >$200000 (5% of LW) | $14,162.35 | $15,970.67 | $15,843.97 | $296,884.63 | $299,444.44 | $569,447.35 | 4.77% | 5.33% | 2.78% | ||||
| Total | $2,265.56 | $2,669.85 | $3,949.26 | $62,285.72 | $75,130.37 | $133,734.60 | 3.64% | 3.55% | 2.95% | ||||
| All < $200000 | $1,689.36 | $1,649.32 | $2,515.29 | $51,254.43 | $58,306.81 | $81,207.03 | 3.30% | 2.83% | 3.10% | ||||
| Global LW (N=643) vs USA LW (N=403) vs. Average US Citizen (Comparable Age) | |||
| Age Bracket* | LW Median | US LW Median | US Median*** [2] |
| 15-24 | $17,000.00 | $20,000.00 | $26,999.13 |
| 25-34 | $50,000.00 | $60,504.00 | $45,328.70 |
| All <35 | $40,000.00 | $58,000.00 | $40,889.57 |
| Global LW (N=407) vs USA LW (N=243) vs. Average US Citizen (Comparable IQ) | |||
| Average LW** | US LW | US Between 125-155 IQ [3] | |
| Median Income | $40,000.00 | $58,000.00 | $60,528.70 |
| Mean Contributions | $2,265.56 | $2,669.85 | $2,016.00 |
Note: Three data points were removed from the sample due to my subjective opinion that they were fake. Any self-reported IQs of 0 were removed. Any self-reported income of 0 was removed.
*89% of the LW population is between the age of 15 and 34.
**88% of the LW population has an IQ between 125 and 155, with an average IQ of 138.
****Median numbers were adjusted down by a factor of 1.15 to account for the fact that the source data was calculating household median income rather than individual median income.
[1] Internal Revenue Service, Charitable Giving by Households that Itemize Deductions (AGI and Itemized Contributions Summary by Zip, 2012), The Urban Institute, National Center for Charitable Statistics
[2] U.S. Census Bureau, Current Population Survey, 2013 and 2014 Annual Social and Economic Supplements.
[3] Do you have to be smart to be rich? The impact of IQ on wealth, income and financial distress Intelligence, Vol. 35, No. 5. (September 2007), by Jay L. Zagorsky
Update 1: Updated chart 1&2 to account for the fact that the source data was calculating household median income rather than individual income.
Update 2: Reverted Chart 1 back to original because I realized that the purpose was to compare LWers to those in similar income brackets. So in that situation, whether it’s a household or an individual is not as relevant. It does penalize households to an extent because they have less money available to donate to charity because they’re splitting their money three ways.
Update 3: Updated all charts to include data that is filtered for US only.
How have you treated students? LW has a large student population. I notice that LW incomes are below US median in the 15-24 bracket but above US median in the 25-34 bracket. I suspect LW also has a large population of people who have spent longer at university than most (e.g., PhDs) which will tend to depress earnings at the young end but maybe boost them later on.
It seems (at least according to self-reports) that the LW population gives a little more, as a fraction of income, than the US population at large. I wonder what the distributions look like.
I updated the chart to be a bit more useful: I broke the <$100000 bracket down into $25000 increments.
Now students are accounted for indirectly: there is an extremely strong correlation between age and average income for that age among LWers. (r^2 of .83). So it’s safe to assume that there are a large number of 15-24 year olds in that “<$25000” income bracket.
Let’s pretend that each LWer represents a snapshot in the life of one big meta-uber-LWer. Here is the story of hisher life (with my theory in parentheses):
It starts off its life making less money than average, but donating more of their overall salary. (Perhaps an idealistic student who spends a large chunk of its early 20s in school. Since most of school ends up being paid for by Uber-Parents, it has more disposable income to donate.)
Once it reaches its mid-20s, it starts making more money than the average. But, on the downside, it starts donating a lot less both overall and as a percentage of its income. (That college education pays off with a decent job, but the reality of growing up sinks in… Bills must be paid, rent is due, groceries must be bought, and less money is available to give away.)
As it reaches its mid-30s, it advances its career sufficiently, and starts making pretty good money, but not millionaire money. (Now that it has a lot more money to throw around, it’s safe to start being idealistic again. After all, it finally made it!)
Despite its success, the Uber-LWer is still making less money than the Uber-HighIQer. (Perhaps people are overreporting their IQ. Or maybe there’s some social consequences to rationalism. Or maybe the career paths of LWers tend to be lower paying. Not sure what to make of this one.)
[edit] Incidentally, the US Mean Income shows us why we typically use Median income instead: the mean is dramatically thrown off by the small percentage of very rich individuals.
IQ doesn’t change much over the lifetime, being an LW’ler does. Decisions that you made ten years ago might determine your current salary.
The data that you provided doesn’t prove that claim. To make that claim you would need to filter for LW’ler who live in the US. Living in the US is correlated with higher income.
I’d have to do a lot more than just filter for USA. There’s a strong correlation between age and income, and the LW population is much younger than the average US citizen. There’s a pretty decent chance that the gap in IQ-income is more a function of age. I’d have to divide high IQ individuals into age buckets and show that their average income is higher. I don’t think that data currently exists, or at least I wasn’t able to find it.
Filtering for USA is done in a line of R code. It’s not much additional effort if you already have set up the rest.
Oh yeah, it will be easy to do. I’m just mobile right now and don’t have access to the computer with the data on it. I’ll update tonight though.
I think so. Don’t we have quite a lot of people in academia, for instance?
Oh, I guess the other thing is where one lives. Salaries are high in the US, and LW has plenty of people from elsewhere. (I think the difference is particularly large in the software industry, where lots of LW people and lots of high-IQ people work.) Do you fancy taking a look at what happens if you restrict the LW sample to people who live in the US?
[EDITED to add: Oops, I see ChristianKI has already made this observation.]
Updated to include US-only data.
It’s interesting how much smaller the difference in salary between all-LW and US-LW looks when you look at data broken down by salary brackets than when you don’t. Not very surprising in retrospect, but still interesting.
There are some interesting effects in the charitable contribution figures. For instance, in the first row we see that US-LW contributions in this bracket are way lower than US-general, US-LW income is about the same as US-general, but US-LW contributions/income are bigger than US-general contributions/income. The 100k-200k bracket shows a similar oddity: very similar contributions, very similar incomes, but much larger contributions/income.
My guess is that the contributions/income figures are averages of ratios rather than ratios of averages, so that if you have one person with an income of $1 giving $100 away then that’s going to pull your average way up. I guess there are quite a lot of low-income high-relative-giving LW folks and that they explain the low-end figures, and that at the high end there are maybe one or two high-relative-giving LW folks to explain the high-end figures.
Agggh! I’m glad you pointed out that incongruency. When I was reformatting the graph, I only copied in half of the US Mean Contribuions so the first 4 rows of that were incorrect. The graph has been updated. The numbers are in fact a ratio of averages. LW mean contribution is calculated as follows: (Total Contributions of Subgroup)/(Total Income of Subgroup).
However, I am not seeing where you’re getting that US-LW income is about the same as US-General; for 0-25k the USA-LW mean income is $11k, the USA-General is $15k.
Sheer hallucination, I think. Sorry.
Why did you remove zero income individuals? Students or unemployed people might plausibly list this.
Because I was comparing it against US income data which excludes the unemployed.
Ah, that makes sense. Thanks.
On the first chart, you’re comparing household figures with individuals which isn’t fair to the households. It’s better to have an amount of money split one way than three ways.
Looks like ~29% of households are dual-income. Meaning we would need to adjust the mean income and contributions for the “US Household” columns by a factor of 1.29. The contributions/income stay the same though.
There’s really not a good on-the-fly adjustment one could make for the median measurements. The best approximation I could come up with was 15%, (which is the national median househould income divided by the national median individual income).
I have updated the original post with these adjusted numbers. It looks like my hypothesis was incorrect. An average LWer makes less and donate less than the average American.
Note: I know this is not a perfect method for transforming household income into individual income. But after two iterations, the end result is much the same, and I really don’t like the idea of creating ad hoc methodologies just because I don’t like the results.
Good catch; that axtually applies to both graphs. I will need to drag up some numbers on how many households have two income earners (as that is the only case where it would differ from the current setup). Of course, in absence of individual earner data broken out by age and income bracket (which I don’t think is out there), any attempt to adjust will be just a guess.
Before I go out and try to find data, I’ll say in advance the method and hypothesis: take the % of overall tax returns that involve two income owners. Increass the sample by that much: if 20% of 100 households have dual earners, that accounts for 120 total individual income earners. I predict that it will bring the national average down to be much more comparable with LW’s stats.
(Forgive poor style, typing this from my phone)
Contributions to what—all charities?
On the LW Survey: “How much money, in number of dollars, have you donated to charity over the past year?” On the IRS Report: A list of what they consider a charity is here: http://nccsweb.urban.org/nonprofit-overview.php
Ah, now I also see the footnotes in the chart headings.