I am skeptical, because almost all of the effect was due to white populations in republican counties (as they point out in the article). Other ethnicities apparently performed similarly across counties (again, as they point out). Seems a lot more likely that the decline of racism over the recent decades has just made poor white people look like poor other ethnicities (which we do see in the data). This also correlates with switching to vote Republican because you tend to vote populist when you are getting screwed and the Republicans have been the party of populism for at least the last decade. Anyways, the point is this study should be viewed with extreme skepticism because it is doubtful they captured all the external confounding factors.