OK, you want a truly surprising result? Two Japanese scholars constructed an unbalanced panel of NBA players from the 1985-86 season to the 2015-16 season and found this:
Contrary to the results of previous studies, we find that non-white players are paid equally to white players with similar characteristics in the 1980s and 1990s, but that white players started to be paid 20 percent more than non-white players in the last 10 years. Our results are robust in all specification checks such as the quantile regressions, controlling the sample selection and controlling different contract types. Non-parametrically estimated density of the counter-factual salary of non-white players confirms our results. In addition, we find that neither the employers preference nor income gap of white and black fans explain this increasing salary gap.
You want to know how much we are talking about? Well, with the caveat that these estimates are very imprecise, you can go to page 3: “Consistent with the previous literature, which shows that racial salary discrimination was disappearing in the 1990s, we find that during the 1980s and 1990s, there was no white premium. However, in the 2000s, we find that the white premium becomes about 9 percent (p < 0.05) and in the 2010s, it reached 26 percent.”
Italics mine. And warranted, I think. Is the result convincing, and if so, what is going on?