As of 2017:
" I think most people don't realize just how much money the richest people have. People generally think of normal(ish) distributions like height, where if you're 10% taller or shorter than average, you're a tall or short person, and 40% taller makes you the tallest person in the world. In comparison, wealth has a very, very long tail, making it hard to comprehend.
Here's what I've come up with to visualize wealth in the United States. Suppose you start counting, going up by 1 million dollars every second, and people sit down when you reach their net worth. Most people will sit down immediately. After about 9 seconds, people in the "1%" will start sitting down. Near the 17 minute mark, billionaires would start sitting down. Donald Trump would sit down just before the hour mark. An hour and 10 minutes into the second day of billionaires sitting down, Bill Gates would sit, followed by Jeff Bezos just three minutes later.
The point of this is there's a huge range of billionaires (analogous to comparing 17 minutes to a day). The 1% hardly even registers on this scale (a few seconds). (I should also mention that there should be huge error bars on reported net worth numbers.) " -- Kens
https://www.forbes.com/billionaires/list/ lists about 2000 billionaries (corroborated by [1], which estimates about 2300).
There are thought to be about 200000 "Ultra high-net-worth individuals", or people with over $30 million in 2013 dollars [2].
There are estimated to be about 13 million millionaires in the word [3].
Note that the word 'millionaire' was first used in print in America in 1843 [4]. $1 million in 1843 would be about $32 million (inflation-adjusted) in 2017, according to [5], or $26 million in 2016, according to [6]; $1 million in 1913 would be about $25 million in 2017, according to [7]. $1 million in 1950 would be about $10 million in 2017, according to [8]. $1 million in 1975 would be about $5 million in 2017, according to [9]. So, for much of its life, the word 'millionaire' (as applied to Americans/USD, at least) meant much more than it did today.
" Behavioral economists Michael Norton and Dan Ariely recently showed sample distributions of wealth to Americans, in which the people in the bottom fifth have X percentage of the wealth, those in the next fifth have Y percentage of the wealth, and so on. They found that Americans are very wrong about just how unequal their country is—they think that the bottom 40 percent has 9 percent of the wealth and the top 20 percent has 59 percent, while the actual proportions are 0.3 percent and 84 percent.
They also find that when asked about what distribution would be ideal, Americans, regardless of political party, want a far more equal society than they actually live in or believe that they live in. In an article published in The Atlantic, Ariely writes, “the vast majority of Americans prefer a distribution of wealth more equal than what exists in Sweden, which is often placed rhetorically at the extreme far left in terms of political ideology—embraced by liberals as an ideal society and disparaged by conservatives as an overreaching socialist nanny state.”
These are important findings, but Frankfurt’s analysis motivates us to question what they really mean. Ariely emphasizes that Americans want a far more equal society than they have, but it’s worth noting that they don’t actually want equality. The study finds that when asked to create a perfect society, respondents choose one in which those in the top fifth have about three times more wealth than those in the bottom fifth. "
some facts from "Factfulness: Ten Reasons We're Wrong About the World—and Why Things Are Better Than You Think", via [10]:
4 levels of income " One billion people live on level 1. This is what we think of as extreme poverty. If you’re on level 1, you survive on less than $2 a day and get around by walking barefoot. Your food is cooked over an open fire, and you spend most of your day traveling to fetch water. At night, you and your children sleep on a dirt floor.
Three billion people live on level 2, between $2 and $8 a day. Level 2 means that you can buy shoes and maybe a bike, so it doesn’t take so long to get water. Your kids go to school instead of working all day. Dinner is made over a gas stove, and your family sleeps on mattresses instead of the floor.
Two billion people live on level 3, between $8 and $32 a day. You have running water and a fridge in your home. You can also afford a motorbike to make getting around easier. Some of your kids start (and even finish) high school.
One billion people live on level 4. If you spend more than $32 a day, you’re on level 4. You have at least a high school education and can probably afford to buy a car and take a vacation once in a while. "
Banks will always fail from time to time (unless bailed out), and many banks will always fail at once. People speak as if this is avoidable, but it's not. Think about it; banks borrow more than their net worth from their customers, and then lend most of this out. This means that if more than a certain percentage of their loans go bad, they go broke. When the economy goes south, lots of loans tend to fail at the same time, and this same thing will be happening to all of the banks at the same time, which means that there is some small probability that all of the n-largest banks will go broke at roughly the same time. And the associated probability distributions are not Gaussian distributions, but heavy-tailed ones, which means that the probability of extremely bad outcomes is not lifetime-of-the-universe-so-dont-worry-about-it-small but just infrequent-but-it-happens-small.
So for a given amount of caution (eg regulator-imposed capital ratios), one could derive a predicted mean-time-between-failure duration (or perhaps a survival time) for banks. For political reasons, regulators pretend like their target is for it to NEVER occur that lots of banks all fail at once, but in fact failures are correlated, and a survival time metric is implied, and personally i suspect that it's on the order of tens or hundreds of years, not eg hundreds of thousands of years. For example, [11] find a survival time in a Nigerian bank dataset to be on the order of 50 years (i don't have data on correlated failures of the top-N banks but this could be calculated).