The included graph is an excellent example of graph distortion. One of the fun things about graphs is that they can be incredibly deceptive. This is one of thos

The included graph is an excellent example of graph distortion. One of the fun things about graphs is that they can be incredibly deceptive. This is one of those examples -- "Hey, let's make the bottom of the graph the normal percent, whatever that is, and then stretch it out so the top of the graph is the percent we're trying to highlight as abnormal" Furthermore, mc32 nails it on the head by mentioning the disproportion of the self-employed in the 1%. I think a more revealing statistic would be, among those who run companies, which percentage of their children work with them at some point. With that filter in place, effect of wealth on that would be far more interesting, and I think the effect will be much much less, and perhaps even negatively correlated.