If that were the whole story - which I do not think it is, since one of the biggest predictor of income in industry in which the person is employed - one would not have to be a moralistic prick to draw an obvious conclusion that this calls for better parenting practices i.e. watching after your kids and making sure that they get what it takes to have a better start instead of, say, watching tee-vee, shooting drugs, and chasing after new girlfriends/boyfriends.
But again, I do not think this trite and hackneyed trapped-in-their-status trope tells us the whole story or even the most important part of the story.
Another point - most labor market studies worth the paper on which they are written show that - making a long story short - that social status of the employee accounts for a nontrivial share of variance in the income, so your revelations are hardly news. On the other hand, the share of variance attributable to social status (such as sex or ethnicity) is usually not that great as that attributable to, say, industry, work experience or education - albeit we all I agree that it should be zero. A far more interesting question is, imho, what actually determines wage.
First, a part of employee remuneration is not wage but rent (in economic sense - i.e. payment for the position in the market, so to speak, rather for the goods or services rendered). I would say that the rent part or employee remuneration increases with the power wielded in the organization and that may explain a significant part of income inequality on the top end of the spectrum. However, as we move down the power ladder, the rent part of remuneration becomes much les significant - it is mainly the function of market scarcity than power (i.e. positions experiencing shortages of qualified staff paying more). So that leaves with the question why, after accounting for the rent element - the hour of work of one person is priced higher than the hour of work of another person.
I am not a big fan of the human capital hypothesis (arguing that higher skills fetch higher prices), but I think it does account for some of the variance. It is just that share of variance is not as big as the human capital folk claim. My favored explanation, which is difficult to include in regression models of income is path dependence. That is to say, there is a certain "historical stickiness" in wages - those wages that historically started high, tend to remain high, while those that historically started low tend to remain low. You can thing of it as a form of inertia, if you will.
The original reasons why a position started high or low are not that important - it might have been supply/demand of the required skills, the power and prestige of the incumbents, redistribution (higher returns yield in that particular industry which were passed on employees). What matters is that these original wages created a set of expectations among both employers and employees what the "proper" wage should be, and those expectations stick for a long time, even if the original conditions determining the wage no longer obtain.
A similar idea was proposed by Barbara Reskin and Patricia Roos (my professor at Rutgers) in their book _Job queues, gender queues_, except that they end on social expectations as an explaining factor without explaining what accounts for these expectations. I think that the path dependency theory (which is a very cool thing in institutional economics) can travel some considerable length here.
What is the significance of all this? The income variance not accounted by the "human capital" variables (skills, experience, job responsibilities etc.) are not necessarily attributable to "social status" - i.e. discriminatory treatment due to one's gender or ethnicity (albeit some of it certainly is). Another powerful factor, often overlooked, is the institutional one - which relates to the institutional path dependencies and the systems of expectations that they create.
Wojtek