I don't know what the copyright issues are and so, rather than assuming I have the right, I'll simply give this link to an excellent exposition of the Laffer Curve by Tim Groseclose. It's titled "Do High Taxes Raise More Money?" but, of course, as you'll see from the exposition, it should really be titled "Do High Tax Rates Raise More Money?"
Near the end, Tim makes a controversial statement. He says, at about the 4:30 point, that the Romer and Romer study implies a revenue-maximizing rate of 33%. He doesn't say that they say that. Rather he says, "if you do the math, the results imply that the hump on the Laffer curve occurs where the tax rate is around 33%." Yet, if James Kwak has it right, some later work by the Romers says that the revenue-maximizing rate is well over twice that. Would the Romers disagree with what Groseclose says is the implication of their earlier study? I must admit that when I look at the study, I can't tease out of it what Groseclose does. I don't even know how to do it based on their study and it seems to me as if one can't. But that could be my failing. Groseclose is very good and so it's quite conceivable that he found something I didn't.
The answer matters a lot. My own gut feel is that the revenue-maximizing tax rate, which would clearly be above the welfare-maximizing rate, unless high-income people's welfare is weighted by zero, would be somewhere in the mid-40s. But, as I admit, that's a gut feel. Even if he's wrong, by the way, his exposition up to the controversial part of the video is beautifully done.
HT to Jeff Hummel.
UPDATE: MatthewH, in a comment below, gives a link to the cite where Groseclose gives the proof. The proof makes sense to me. I told you that it's quite conceivable that he found something I didn't.