So the income curve is becoming skewed instead of remaining Gaussian or unimodal in appearance. If we realize that the wierdness comes from the superimposition of two normal distributions--that the population (aside . . .
[mbs] the income distribution is skewed to begin with; always has been. I don't think anybody has shown the 'new' component is normally distributed within itself. Whatever the answer to that, chances are there are not enough immigrants to alter the fact of skewness in the distribution -- only the degree. Nor is there any sign of more than one modal point, far as I know, though I don't live with this sort of data.
Obviously the lower percentiles are disproportionately non-white, so a racial subtext to such a discussion is conceivable. That could go two ways: towards a malign view of ability or genes, or towards a look at discrimination, social welfare spending, etc.
The drift of Cox is libertarian. Distribution may have gotten worse for racial reasons, but everybody's doing better on the whole. I'd be surprised to know of any racial themes in Cox, tho I haven't read him so I could be wrong.
mbs