Looking around at different ways of measuring income inequality (which I'm not convinced is a problem in and of itself: "people start startups in the hope of becoming much richer than they were before. And if your society tries to prevent anyone from being much richer than anyone else, it will also prevent one person from being much richer at t2 than t1") and I see that "scale independence" is valued in an income inequality metric. I'm not sure it should be. If you have a society and make everyone 2x richer, then a scale independent metric would say inequality hasn't changed. But what if you make everyone 1000x richer? Because the value of money is not linear in the amount you have (I would rather have a 50% chance of gaining $1M than a 25% chance of gaining $2M. Heck, I'd even prefer it to a 25% chance of gaining $3M.), as you scale incomes up inequality decreases. What I want is an income inequality metric that is explicitly scale dependent, considering along with each income the decreasing value of large amounts of money.