[Symmetry]

Your referring to an idea called the Laffer Curve. It's true that the government would bring in essentially $0 revenue with an income tax rate of 0% and it's, well, more or less true for a tax rate of 100% too. And so logically there is a point between those two where revenue is maximized. But in the US at least we're no where near that point. For people payed in proportion to their labor like people working hourly jobs it seems to peak over 50% and for salaried professionals the peak seems to be between 70% and 80%. So if you're talking about lowering the top tax bracket from 90% to 70% you can raise more money. But you can't for any of the sorts of tax rates we have today.

[kcorbitt]

But government revenue is almost certainly the wrong thing to optimize for. You probably want to optimize for something closer "wellbeing of the median citizen," although there's of course a lot of disagreement on what the actual right metric is.

This is a very practical difference. For the sake of easy math, let's assume that the maximum-revenue point on the curve is a 50% tax rate. It then follows that for every % you charge someone over 50%, tax revenue decreases. Which means that at this rate, every additional X% of tax is actually decreasing total economic output by 2X%! And assuming that we're on a smooth curve, that means that every X% you DECREASE your tax from the Laffer point is creating 2X% of economic value. Now, based on how fairly the created value is distributed and how efficiently you believe the government can spend its tax revenue you may have a different opinion on how much value you're comfortable destroying to increase tax revenue, but I'd submit that 2X is quite a high number.

I believe we need much-higher-than-0 taxes, but "we're still below the Laffer point" is a poor argument for raising them.