We can also toss in this paper by Raj Chetty et. al. showing how the Danish tax code creates sizable labor supply responses around the kinks in the tax brackets, thus creating a large macro elasticity.
What does it all mean?
How about that official estimates of the revenue consequences of tax changes will generally be incorrect in a systematic direction (overestimating the revenue gain of tax increases and underestimating the revenue loss of tax cuts).