You might be right about an attempt at misdirection (I wouldn't speculate one way or another, but obviously EPI agrees with you), but I don't think that invalidates the case about income inequality being partially (among many other sources) caused by AM. It seems fairly well-supported by the evidence from NBER, which Cowen wasn't involved in: http://www.economist.com/news/united-states/21595972-how-sex...

The (also liberal) Brookings institution also seems to support it, so I don't think the opposition case is exactly ironclad: http://www.brookings.edu/blogs/social-mobility-memos/posts/2...

Another thing to consider is that income inequality is increasing via second-order AM effects, via propagation across generations, rather than directly due to the marriages (http://www.economist.com/node/21640316)