>That is, taxes can in practice be increased
a lot until the rich start fantasizing about going Galt.
The problem is not that they stop participating in the economy, the problem is that higher taxes induce more economically inefficient behavior solely for the purpose of tax avoidance. Corporations will engage in whatever contortions necessary and waste enormous resources and engage in otherwise-irrational activities with billions of dollars behind them solely for the purposes of reducing taxable income. If you can spend a billion dollars to save $1.2B in taxes then corporations will do it. So corporations hoard assets in offshore subsidiaries because repatriating it causes it to be taxed and issuing it as a dividend causes it to be taxed again, but doing neither allows the shareholders to take the profit as share price appreciation which can be tax deferred indefinitely and even then comes in at a lower rate. Result is that corporations like Apple end up as de facto mutual fund managers even though their business expertise has absolutely nothing to do with that. The amount of inefficiency caused by making poor (or just excessively conservative) investment decisions with billions of dollars per corporation is difficult to even imagine. Then there is the matter of jurisdiction shopping, so again, the problem is not that they stop engaging in economic activity, the problem is they move across borders, and each instance of that causes you to get 30% of nothing instead of 20% of something, meanwhile the businesses that stay are put at an economic disadvantage (or reduced economic advantage) against their foreign competitors.
>You also don't seem to take into account elasticity, the decreasing utility of money, or any notion of optimizing social utility.
That's because we're discussing business growth. And D&S have the opposite problem: They try to account for utility to the taxpayer (though they go on to assign that value to zero for high income earners), but they never account for the investment value of money. When a business owner pays more in taxes and in so doing has only enough earnings to reinvest into opening three new facilities instead of four, the owner may not have any change whatsoever in standard of living, but the employees who would have worked at the fourth facility might have a different opinion.
D&S also calculate the "optimal" tax rate as the one that generates the most revenue from high income earners, which ignores the essential question of whether the government can make more economically productive use of the money than the private sector. They punt on the question by assuming that because high earners have a lower marginal utility for the money than low income earners, shifting the tax burden at any given spending level toward high income earners will always result in a net gain in social utility, but that falls apart pretty quickly because the lowest income earners already have a negative effective tax rate. So that assumption devolves into the idea that redistribution of wealth has positive social utility and we should always take from someone who has more money and give to someone who has less money to take advantage of the social utility gain, i.e. naked Marxist rhetoric with no economic basis.
>If I'm Bill Gates, I could not give enough of a shit that you taxed from me $5k dollars, or even $500k. But, use that to pay for nine or ten heart surgeries, and suddenly social utility is increased by, I would guesstimate, 100x (total lifetime earnings of those ten people, starting after surgery).
This is a perfect example because of what Bill Gates actually does with his money. If you take $500k that could have gone to the Bill and Melinda Gates Foundation, Bill Gates may not give one damn about the money, but what about the people he was going to buy mosquito nets for? And what if the government spends the money building F35s that don't fly rather than on heart surgery?
>Also, there's not really any empirical support for a inverse relation between business taxes and economic growth. It's just one data point, granted, but we could take as an example Bush's years. Massive tax cuts, but very poor growth.
Why are you assuming the result of low tax rates would be seen immediately? A business doesn't always decide to build a new facility when taxes change, they just decide where to build the new facility based on what prevailing tax rates are. You might also want to look into Chamley and Judd (as discussed in D&S at 14), regarding the consequences of taxing investment earnings that would have been reinvested. The short version is, because investments collect interest with exponential growth, taxing would-be investment capital has an exponential cost to the economy if future earnings would always be similarly reinvested.
D&S go on to discount this because most families or individuals will eventually choose to consume their savings rather than continually reinvesting their earnings forever, but that doesn't really apply when the entity collecting the earnings is a publicly traded corporation, and even for individuals the original reasoning still holds to extent that earnings are reinvested.