Michael Perelman puts forward evidence of racial discrimination in the NBA:

Even in sports, one of the few venues where society associates Blacks with excellence, Blacks still face discrimination. For example, in cities where the population is more White, professional basketball teams hire fewer Black players (Brown, Spiro, and Keenan 1991).

In an unpublished paper, Dan Rascher and Ha Hoang found that after adjusting for a number of factors, Black basketball players have a 36 percent higher chance of being cut than Whites of comparable ability. This statistic reinforces the widely held impression that while teams will want to employ the Black superstar for a better chance of winning, they will prefer a higher mix of White players on the bench to please their predominately white audience.

This might well be true. Here's a 2005 paper entitled, "Are NBA Fans Becoming Indifferent to Race," published in the Journal of Sports Economics, that charts a few trends. During the 1990s, teams that played in whiter cities really did tend to have a greater number of white players. The correlation is higher for starters than it is for benchwarmer, as you'd expect—"star" players who are white, in fact, are especially likely to end up in cities with relatively higher white populations.

Meanwhile, white players are less likely to be traded away from teams with a larger population of whites in their market area (and black players more likely to be traded away). And the kicker is that, controlling for other factors, teams that better "matched" the racial composition of their market area boosted their home-game attendance revenue. That dovetails somewhat with this 2001 paper, which, judging from the abstract, finds that the size of television audiences for NBA games increases with increasing participation by white players.

Now granted, none of this is conclusive. I don't know much about the NBA, but it seems like there could be all sorts of perfectly benign reasons why the Utah Jazz, playing in a lily-white city, just happened to be 44 percent white during the 1990s (luck of the draft, existing contractual obligations, the way the free agent pool happens to shake out). But it also doesn't seem so unrealistic to think that a black fan base would rather see black players (and a white fan base white players), and that teams respond to this preference.

On the brighter side, racial discrimination in pay seems to have largely vanished during the 1990s, unlike the previous decade, when white players were paid more than blacks for the same level of performance. Here's a long-ish paper on the topic if anyone's interested.