Instead of keeping the actual proportion of black residents, the data set contains a presumed model B of how housing prices change as black resident population

Instead of keeping the actual proportion of black residents, the data set contains a presumed model B of how housing prices change as black resident population moves away from 63%. The problem is B gives you the same answer for 62% and 64% of black population, so you can recover the actual proportion.