Yeah, I think the author is drawing the wrong conclusions by comparing two variables. It looks like, from the graph, there's some staying power within the top 20%, but that still leaves a great amount of mobility for college graduates.