> If the US GDP increases faster than population, then the per-capita income will always increase as well, even if the employment-population ratio plummets. Ye

> If the US GDP increases faster than population, then the per-capita income will always increase as well, even if the employment-population ratio plummets. Yes, but irrelevant because GDP is not produced by non-workers, it's produced by workers. GDP is #workers * output per worker. If the former decreases more than the latter increases, GDP goes down. Even if GDP goes up, GDP per capita can go down. Suppose we start with 100% workers and then double the population while keeping the number of workers constant AND making them 25% more productive. The result is 25% more GDP but less GDP per person because there are twice as many people. > Regarding state and local taxes in 2050, they are included in the $19 trillion figure. In other words, they're part of the 19%. Which means that the current average rate is no where near 38%. You're claiming that nearly doubling the average tax burden will have no effect. I'm pointing out that you're actually proposing something closer to 3.5x. Even if we assume that you're correct, why do you think that Americans will pay those taxes? Note that your "the numbers work for the economy as a whole" argument is irrelevant to what an individual does. The Clinton welfare reforms as well as current experience with unemployment benefits shows that legal employment is sensitive to fairly small amounts of money. How likely is it that it's insensitive to large amounts of money? Yes, an interesting fraction of folks will go off book. They'll contribute to GDP but they won't be paying taxes. So, you'll need even higher rates on the "taxed GDP" which will push more folks off the book. Yes, I realize that you're claiming that the remaining workers will have more after-tax than they have today but that's not how people think. High tax rates make evasion more profitable and generous payments to non-workers make not working more attractive.