It seems perfectly reasonable to me to respond to this by pointing out that upper-band tax rates have been distinctly higher in the past, and that innovation and wealth creation have done just fine when they were.
Incidentally, it seems pretty odd to me to say "more or less flat" and give a number as precise as 18.1% when the actual variations have been from 15% to over 20%. I wouldn't call that "more or less flat" (though I would agree there's no long-term trend, which is an entirely different matter) -- but, regardless of that, I suggest an experiment. Ask someone to plot a graph of something that's "more or less flat at 18.1%", a graph of something that's "more or less flat at 18%", or a graph of something that's "more or less flat, between 15% and 20%". I predict that the first will be much, much flatter than the actual revenue graph; the second will merely be much flatter; and the third might be about right (though probably still too flat; "varying between 15% and 20%" would most likely get you a better result).
In other words, your choice of wording gives a very false impression of how the graph actually behaves.
