alt Hacker NewsPeterson before he went nuts covered the research on this quite bluntly:
Lifetime earnings is most strongly predicted by IQ, then conscientiousness. So those with both are considerably more likely. Social connectedness is entirely dependent on personality as well - extraversion predicts this, and extraversion also predicts ones interest in financial incentives rather curiously
Super uncertain causality. "Earnings is most strongly predicted by IQ" suggests mechanistic and innate causality, but SES acts on IQ and on outcomes (as IQ does on SES). Education powerfully influences lifetime earnings, but education is obviously and potently impacted by the SES of your family.
This is like that hoary chestnut about the Pentagon and pizza delivery orders. Military strikes, to hear the story go, are most powerfully predicted by pizza orders. But pizza itself does not cause wars (except in Chicago and New York City).
A nit, but the point is you can't distill these problems down to a single statistic and then do policy based on them (not that you were proposing policy).