It's a very interesting hypothesis, but I hope that either the journalist misinterpreted the study, or I misinterpreted the journalist with > In fact, he found

It's a very interesting hypothesis, but I hope that either the journalist misinterpreted the study, or I misinterpreted the journalist with > In fact, he found that the more cherubic a black CEO’s face, the higher his salary—and even his company’s revenues. I hope the study controlled for firm revenues when evaluating the effect of facial features on salary. It's difficult to both simultaneously control for revenues and also attribute causation on revenues. I can imagine several plausible mechanisms (maybe a teddy bear face reduces power struggles within the C-suite) by which the face affects revenue and this in turn drives salary. I have no doubt that the salary differences remain after controlling for revenues, and there's some bidirectional causality, but the journalist's phrasing suggests to me a direction of causality that may not be warranted.