Well, here's the hilarious followup. Brendan Nyhan, who clearly has too much time on his hands, took the raw data provided by Kevin Hassett and replotted it using standard tools. The result is on the right: if you do a linear regression you get the red line, while if you apply a quadratic model you get the blue curve.
Here's the funny part: For technical reasons Brendan thinks the linear regression is more likely to be correct, but still, the quadratic model is at least defensible. And if you look at its high point, it peaks at about.....29%. Pretty much the same place as the kindergarten curve the WSJ drew.
In other words, if the WSJ editorial page had even a smidgen of intellectual honesty, they could have plotted a curve that was at least colorably defensible and would have made their same point, namely that the United States might benefit from lower corporate tax rates. But they just couldn't bring themselves to treat the data honestly, and the result was to undermine their own cause. It seems a deserving fate, no?