Does the average income in the US drop by \$9,500 if you exclude the ten richest Americans?

A leftist friend of mine shared this meme:

Averages are tricky like that

• The average income in the United States is 74,500\$.
• Excluding the top 10 Americans, it's only 65,000.
• Excluding the top 50, it drops to 48,000\$.
• Excluding the top 1,000, it drops to just 35,500\$

It alleges that the average income in the United States drops by \$9,500 if you exclude the ten richest Americans, \$25,000 if you exclude the 50 richest, and \$39,000 if you exclude the 1,000 richest.

Is this true?

• Is this meme widely shared, or is its claim widely believed? It's unclear presently whether this meets the site's notability guidelines. Can you find the original post (or at the very least, a copy of the meme that's been widely viewed)? Commented Sep 5 at 11:08
• @F1Krazy I tried to find the source for the image, because I saw it somewhere on Reddit in the last few days, but I wasn't able to find that post anymore. A Google Lens search for exact matches find some Thread posts from June, so I guess it originated there. Commented Sep 6 at 5:56

This is false because the "starting" value used by the meme is the median value, not the mean, of household income in the United States (as seen in this chart Real Median Household Income in the United States, showing \$74,580 for 2022)

The reductions are made as though that value was an average, not a median. While excluding outliers can change an average substantially, it affects the median only very slightly.

It also appears that the ostensible reductions from excluding the "top" earners are based on counting total wealth, not yearly income-- which is itself at least an order-of-magnitude error. The 10 highest income Americans do not average \$95,000,000,000 in annual income.

From the chart Mean Family Income in the United States you can see that mean family income (as opposed to median) is actually 126,500. This is income and not wealth, but even if we are as generous as possible to the original graphic and assume we can just deduct total wealth of the top X, we see that excluding the top 10 reduces the mean to \$116,000, and reducing the top 1,000 reduces to \$87,000. As percentage reductions, these are much less dramatic-- and the resulting mean is still higher than the (all but unchanged) median. But remember, even these values are comparing apples (income) to oranges (total wealth)-- the total share of annual income going to the very wealthiest is much smaller.

• Interestingly the answer is almost “yes”, for the absolute amount the incomes drop by, except that the starting values are completely wrong in the original meme (since it used median instead of mean). Commented Sep 5 at 18:38
• @fyrepenguin - But to consider the absolute amounts accurate, you would be subtracting total wealth from yearly income, which is itself highly dishonest. Commented Sep 5 at 18:56
• Also CBP estimates 271 million as the noninstitutionalized adult population; \$24 trillion/271 million is \$88,561. Note the answer states family income not individual. @DavePhD Commented Sep 5 at 20:24
• "even if we are as generous as possible to the original graphic and assume we can just deduct total wealth of the top X" - this is way too generous, and evidently misled at least @fyrepenguin. As you note yourself, net worth and annual income are wildly different. Even if we take a broad definition of "income" which includes unrealised gains, rich businessmen's net worths took them a probably-decades-long career to accrue, so you should expect this approach to overestimate the impact of removing them, probably by a factor of 10x or more (though it depends on how bursty the gains are). Commented Sep 5 at 21:40
• @Relaxed the original meme confuses income with wealth (in order to make the math works, as the answer points out) so unfortunately no useful information can be extracted here. Commented Sep 5 at 22:49

The income of individuals is confidential; however, Propublica says they obtained secret IRS information and published the income of the top 400 individuals for 2013-2018. This is gross income before tax deductions. Generally, the list includes spousal income unless tax returns were separately filed. Propublica stated in 2022 "All numbers were adjusted for inflation".

Bill Gates \$2.85B
Michael Bloomberg \$2.05B
Jan Koum \$1.93B
Ken Griffin \$1.68B
Laurene Powell Jobs \$1.57B
Jeffrey Yass \$1.30B
David Siegel \$1.17B
John Overdeck \$1.17B
Larry Ellison \$1.07B
Steve Ballmer \$1.05B

Total: \$15.84 Billion

The US population was 331,449,281 in 2020, so removing the top 10 would reduce the average income by \$48.

Adult population was 258.3 million, so removing the top 10 would reduce the average adult income by \$61.

Number of households was 126.8 million, so removing the top 10 would reduce the average household income by \$125.

• Those "income" values are nonsense; they represent "how much did they choose to get taxed this year" not an actual measure of how much money they earned.
– Yakk
Commented Sep 6 at 14:06
• @Yakk If anything, that makes things even worse, since their actual incomes will be higher and removing them causes the averages to drop even more. Commented Sep 6 at 14:19
• @Yakk the Propublica website goes into annual wealth increase also, but if you apply that standard to the top 10, you'd need to apply it to everyone. Increase in home value is a big factor. Also, apparent wealth increase includes lots of inflation. Commented Sep 6 at 14:54
• @DavePhD You can use constant dollars for that. The nice thing about doing total wealth increase is that you can get "net US assets" from one year, subtract the previous year, and divide by population.
– Yakk
Commented Sep 6 at 19:02
• @Yakk should use net worth of households, fred.stlouisfed.org/series/BOGZ1FL192090005Q 152 trillion as of Q1 2024. Only about 1% is that of the top 10 households. Commented Sep 6 at 20:10