Several months ago, Jeb Bush gaffed that, for the US economy to flourish, Americans should work more hours. Pile-on ensued; critics exclaimed modern Americans are already over-worked.

Well, not exactly, at least historically speaking.

Several years ago, economists Mark Aguiar and Erik Hurst studied work and leisure trends over the last 40 years. They noticed that, in the 1960s, most men—regardless of their education, which serves as a proxy for income—worked the same amount of hours, about 50 per week, and spent about 105 hours dedicated to leisure activities. By 2003, a divergence emerged that mirrored growing income inequality: Men with less than 12 years of education worked, on average, 37.5 hours a week, while more educated (higher earning) men worked 43.4 hours. Both groups gained more leisure time (socializing, watching TV, playing sports), though the less educated group spent about 6 to 7 more hours a week engaged in leisure activities than their more educated (and presumably higher earning) peers.

About a decade later, it’s still the case. The figure below plots hours worked over the last decade by education for employed Americans ages 18 to 65, from the US Census American Time Use Survey.

It’s still true that the more education you have, the more hours you work each week. But Americans of all education levels are working less than in 2007, just as the US headed into the recession.

A similar pattern emerges when you break up work hours by earnings. The figure below shows hours worked by income quintile.

In some ways, it’s not surprising. The more you work, the more money you make. But it wasn’t always this way. Research by economists Peter Kuhn and Fernando Lozano estimate that in 1979 the bottom 20% of earners were more likely to put in long work hours (more than 50 hours a week) than the top 20%. But by 2006, it was mainly the top earners putting in the long hours.

The fact that trends in hours worked mirrors trends in income inequality raises the question of how work hours is contributing to the growing distance between the haves and the have-nots. It could be employers are offering low wage, hourly workers fewer hours and no overtime, dooming them to lower earnings. Or it could be stagnating or declining real earnings reduce people’s incentives to work more hours.

Unfortunately, there are not many real answers. Kuhn and Lozano argue that salaried, skilled workers face greater incentives to put in longer hours than they used to because long hours lead to bigger bonuses and job promotions. But it is not clear why hourly low wage Americans work less.