This phenomenon complicates the interpretation of tax return data. For example, when one looks at the growth of the 1 percent, or the 0.1 percent, in the Piketty-Saez data, that growth is likely exaggerated because some income is merely being shifted from corporate returns. I don't know how much. If someone has already quantified the magnitude of this effect, please email me the answer. If not, someone should write that paper.
Click on graphic to enlarge.