In trying to argue that very high tax rates lead to lower total government revenues, the Wall Street Journal used the following graph. They think the Laffer Curve is the line of best fit, or something close.
As the mildest attention to the graph reveals, however, it's just a random line that goes through an outlier and doesn't bear any interesting relation to the collection of points on the graph.
We've all encountered many cases of people wantonly imposing bad theories on recalcitrant data. But I've never seen such a graphic example before.