Christina and David Romer have a new working paper which examined the impact of the frequent, drastic, and heteregenous policy-induced changes in marginal tax rates on the incentive to work of those at the top 0.05% of the income distribution during the inter-war years. They write about four findings,
First, consistent with what one would expect given the tremendous identifying variation, they are very precise. Second, they show that taxes are indeed distortionary: the null hypothesis of no effect is overwhelmingly rejected. Third, they indicate that the distortions are small. Our baseline estimate of the elasticity of taxable income with respect to the after-tax share (that is, one minus the marginal tax rate) is approximately 0.2. This is considerably smaller than the findings of postwar studies (though generally within their confidence intervals). Finally, the estimates are extremely robust.
James Kwak interprets the Romer's finding that the elasticity of taxable income for these 0.05% (the super-rich) with respect to changes in the after-tax income share is 0.19.
"An elasticity of 0.19 implies that tax revenues would be maximized with a tax rate of 84 percent; that is, you could raise taxes up to 84 percent before people’s reduced incentives to make money would compensate for the higher tax rates."
He argues that instead of being dis-incentivized from working, these super-rich respond by trying to game the tax system,
Recent US history shows that when you raise taxes on the rich, they don’t stop trying to make money: they just pay their lawyers and accountants more to avoid paying taxes. The solution to that is a simpler tax code with fewer exclusions and deductions.
The interpretation of James Kwak is close to that estimated by a study on optimal taxes by Peter Diamond and Emmanuel Saez. Using parameters based on the literature, they suggest that the optimal tax rate on the highest earners is in the vicinity of 70%. See Paul Krugman's discussion here.
In this context Karl Smith points to the interaction between income and work to add another explanation for why the impact of higher marginal tax rates on the super-rich is likely to be minimal. As people grow rich, they slowly substitute their domestic and other non-core (other than their income earning activity) work and time/effort expenditures by hiring others to do that work (like household chores) or private charters (using their private jets instead of waiting at the airports) so that they can spend more time on their core-activity. This in turn increases incomes further. However, as they become super-rich, they would have more or less exhausted such substitution alternatives. They become indifferent to any marginal income changes.