The authors could have just checked out the Atlas last year when I made and posted this graph.
I'm cross posting this from US General Politics since it seems relevant here, too. This chart should help guide one's attempts at a solution. The drift of the upper three quintiles away from the lower two quintiles has been generally slow and steady over the last 35 years. The decade before Reagan isn't appreciably different than the decade after, despite considerable differences in national policy.
It seems clear that wealth has migrated towards the skilled professions, even in the middle class, as the global information age has progressed. The greater the skills required, the more rapid the increase in wages. To me that suggests the most effective changes would direct more resources towards education for the skills needed in the current economy, not large scale wealth redistribution or an investment in jobs in less skilled sectors from economies of the past.
Here's a better chart in response to Ernest. This is also from the historical household income data at the US Census. I found that the top of the second quintile (40%) was the most stable in real dollars, only increasing 5% from 1969 to 2014, so I used that to compare the other quintiles. The bottom quintile remained almost unchanged compared to the second quintile during that span of years and is very close to half the second quintile.
The growth is in the upper three quintiles. The middle quintile grew about 17% compared to the bottom two quintiles. Since the bottom two quintiles had little growth in real dollars, that 17% is close to the growth in real dollars since 1967. The fourth quintile grew at 35% compared to the bottom two quintiles, or about double the rate of the middle. The limit for the upper 5% grew at 54% compared to the bottom two quintiles, or about triple the rate of the middle. My apologies for the year sequence which looked fine until the software rendered it to a bitmap.
