But... the Laffer curve napkin!
The Laffer curve is correct, its just that most American jurisdictions are on the left side of the curve.
I've never seen explanation of the Laffer curve that acknowledges the income effect. That's a glaring oversight.
Like many basic economic models the Laffer curve does not take into account multivariate features of the economy. That doesn't invalidate it, it just limits its predictive power. The curve itself long predates Laffer and he attributed it to the medieval Islamic scholar Ibn Khaldun (who incidentally was a student of the works of Ibn Rushd).
The curve is based on two theorems and one economic assumption. The first theorem is that if the tax rate is zero then revenue will be zero. The second is that if a function is continuous and positive, then it reaches a maximum between two points where the function is zero. The economic assumption is that as taxation rises, people will change their economic behavior to reduce their activity that generates the tax (This is the principle argument given in favor of raising cigarette taxes - it encourages people to cut back on smoking.) The corollary to this assumption is that if the tax rate is 100% then people will completely avoid that activity since they would get no value from it. The logical conclusion from the assumption is that there is a tax rate less than 100% that generates maximum revenue.
The problem is that when there are multiple methods of taxation then calculation of that point of maximum revenue becomes difficult. It is almost certainly the case that the revenue rate curve is not symmetric and may have more than one peak. Another problem is that consumption taxes (like the sales tax) don't confiscate the good or service so there is always some value and the 100% point isn't reached for that tax.