This graph really doesn't seem to support their hypothesis AT ALL - the correlation looks really weak, and they give no strength statistics for it at all. Of the municipalities with the lowest revenue from fines, all of those with the highest percentages (from 80% up to around 100%) of African Americans are represented. On the other hand, in the scattering of municipalities with high revenue from fines (say, over 20%) the percentage of African Americans is anything from less than 5% to 80% with no clear pattern. I'm not convinced...