Which Percent are You? — The Actual Income Distribution in the United States

May 22, 2016 · 7 min read

There’s been a lot said and written about income and wealth inequality in the United States recently, but most of it refers to he-said-she-said quotes and data and statistics. I thought it was time to actually dive into the primary source data to see what the real story is. And the real story is that even the good graphs are deceptive and the skew is way worse than most people comprehend.

The Tool and Visualization

I built a tool which you can see here (http://thekeesh.com/whichpercent/) which is based on Social Security Administration wage data, as well as Census income data. You can type in an individual annual income and it shows you what “percent” you are in the US. You can also dive into the source data yourself.

However, what the rest of the page tries to do is help you actually visualize what this means. The reason this is important is because the variance in incomes is so wide that it is hard to actually understand it. The incomes follow a power law, both in where the money is going and numbers of people in each bucket.

Some Benchmarks

Start with \$28,851, the median income or the 50%. Thats \$13/hour and it takes a standard work year of 260 days to earn that income.

Now type in \$60,000. That’s the 21%. That puts you in the top 34 million out of 158 million wage earners.

Now type in \$250,000. That’s the 1%. That puts you in the top 1.85 million with 156 million people below. That’s a rate of \$120/hour. So you only need to work 30 days of the whole year to make the median wage.

Now type in \$11,770. That’s the poverty line or the 77%. At that rate it would take 637 days a year to earn the median wage.

The Inequality Issue

Now type in \$1,000,000. That’s the 0.0825%. That’s the top 130,000. That’s \$480/hour. It’s 34.66084 times the median and would take just 60 hours (7.5013 work days) to make the median annual income.

Now type in \$1,000,000,000. That’s the 0.00001%. That’s \$480,769.23 per hour. It takes just 3 minutes to earn the median wage.

How does that make any sense? No one is arguing the straw-man that everyone should make the exact same income. What is being argued is that the distribution is too skewed.

Now if we graph a histogram of the SSA income data it looks like this, which at first glance doesn’t seem that bad.

Unfortunately this is a very deceptive graph because the x-axis lacks uniform scaling so paints a very incorrect picture of what the income skew is. The initial buckets go up by \$5000, then the later buckets go up by …. millions. The graph is literally skewed by thousands of times.

Now when you normalize the graph with 10,000 buckets it looks like this, which also doesn’t seem so bad here. However, it stops at the largest bucket of \$250,000 and above, which doesn’t make it seem that bad.

Now if we group by \$50,000 we get a more accurate picture. Unfortunately we are not even close to capturing the range.

So then we create a more accurate picture grouping by \$100,000. And you’ll notice that almost everyone earns under \$100,000 so the graph makes sense. But there is something very misleading about this graph as well. It is that we’ve cut off the graph at \$600k and above. We’re actually not capturing the range by even a tiny fraction.

That’s because what the actual graph of the US income distribution is is this one.

Yes, the little bar on the left is everyone. Actually this graph isn’t right either because the SSA truncates at \$50m. The real graph is this one because it needs to go to \$3.5 billion.

This. This is the actual graph. Everyone — the 99.9999 percent is the bottom sliver. Actually the sliver needs to be so much smaller because it should be 1/18,000 of that first tick mark at \$0.5 billion or \$500 million. You’d need to zoom this graph in 18,000 times to see it. So the graph we think is accurate, that shows that income inequality is bad, but not that bad — is off by a zoom factor of 18,000.

If this graph is taking up 6 inches on your display then the full graph should be 1.70 miles and the small bar showing everyone in the United States would be one pixel. That is actually the skew.

If this graph is taking up 6 inches on your display then the full graph should be 1.70 miles and the small bar showing everyone in the United States would be one pixel. That is actually the skew. You’d draw a dot, then walk for 35 minutes and that would be how far away the graph would be at scale.

So when Bernie talks about the people at the top, he is actually not exaggerating. For them to buy off a politician for \$1 million dollars is like you buying a coffee. Actually it’s not that either. It’s like you dropping a penny on the sidewalk accidentally.

I’ve often posed this hypothetical: What if we could choose 20 people who had to make 5 million dollars and then we could double the wage of millions of people, would you do it? And people say an unequivocal yes. But then they say we should not pay people that much to double everyone, a much smaller amount would do. They don’t realize how it works at the top. This hypothetical is ultimately an embedded tax policy around income distribution that so greatly lifts up the bottom and is hardly even noticeable at the top.

At the bottom, we’re arguing over peanuts. We’re mad that someone we know makes 10% more when someone makes 10,000 times more than you and is continually tilting the system in their favor — unnoticed.

What Bernie is saying about income inequality is both accurate and not an exaggeration.

What Bernie is saying about income inequality is both accurate and not an exaggeration. You’re likely somewhere on the left sliver with 90% of people under \$85k. When one argues that we can’t afford to lift up the poorest or it would ruin the economy, the math doesn’t support it. You could tax the income of ten people, they could still be billionaires, and we’d lift up a million people. The effects of power law distributions with these orders of magnitude are very difficult to comprehend.

Hillary Clinton makes \$16,000,000 a year. Type that in. That is the 0.002%, and she needs to work just 3 hours (on a continuous basis) to make the median income. She is literally the 1%, the 0.1% and more. By definition half the people are under the median income and we saw 90% are under \$85k. She only has to give about 10 minutes of that Wall St. speech to make what you make in a whole year. She also is saying that someone making \$30,000 a year is too high. She makes that in the first 6 minutes of her speech. Hm. Having Clinton and Trump try to talk about income inequality is disingenuous at best. I’d prefer Bernie over Hillary, and Hillary over Trump, but that says very little because Trump is a legitimate national threat and fascist.

So, those are the numbers. I hope and encourage you to dive in and analyze them to learn more. If you are in California, tomorrow (May 23rd) is the last day to register to vote in the primary, and I encourage you to vote for Bernie Sanders if you understand that the fundamentals around income inequality and subsequently the role that money is playing in politics are critical to actually having a functional government that represents the people. His message is not an overreaction. And if you are in the top 10% voting for Hillary, I’d encourage you to consider that you may have more in common with the bottom 90% than the top 0.001%.

Try the tool: http://thekeesh.com/whichpercent/ (view code)

Register to Vote in California Primaries: http://registertovote.ca.gov/

Enjoy this post? Hit recommend.

Written by

Jeremy Keeshin

Blog Posts by Jeremy Keeshin (also at thekeesh.com)

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just \$5/month. Upgrade