Consider the lead/crime correlation, in which the increase in crime and incarceration rates starting in 1960 and peaking in 1991 mirrors a rise and fall of atmospheric lead offset 20 years prior.
The idea is that exposure to lead during childhood makes someone more likely to commit crime in adulthood. Those same factors would arguably make a person less employable, or at least less productive in whatever employment they did have.
But whereas crime rates can just as easily slide one way as the other over time, income distribution cannot.
(Disclaimer: This is purely speculative on my part. It seems plausible, but I'm not claiming it's true.)
The idea is that exposure to lead during childhood makes someone more likely to commit crime in adulthood. Those same factors would arguably make a person less employable, or at least less productive in whatever employment they did have.
But whereas crime rates can just as easily slide one way as the other over time, income distribution cannot.
(Disclaimer: This is purely speculative on my part. It seems plausible, but I'm not claiming it's true.)