In a recent article on food deserts, Richard Florida had this quick throwaway line that caught my eye, with my emphasis added:

There are deeper reasons, again tied to class, that enable affluent and educated households to put this nutritional information to use. For one, they simply have more time and resources to devote to their health and well-being.

This struck me as an interesting and unlikely to be true claim. Higher income and more educated households work more, and therefore you would think they have less free time for things like cooking dinner and healthy behaviors. Indeed, this struck me as a case of a bias I identified in a recent post: we look for the most complementary explanations for the behaviors of lower socioeconomic status people. Let's look at the data.

Shutterstock

In Mark Aguiar and Erik Hurst's paper "Measuring Trends in Leisure: the Allocation of Time Over Five Decades", they use American Time Use Survey data from 1965 to 2003. They found that for both men and women, more education is associated with less time devoted to leisure. A man with less than 12 years of schooling has 116.34 hours of leisure a week and a man with 16 or more hours has 101.44. This difference isn't due to non-market work activities, like chores around the house, which more educated men do a little bit more of. The difference is due to greater work hours for more educated men. Similar patterns exist for women, with less educated women having more leisure time.

Interestingly, this gap has grown over time. The gap in leisure time by education used to be much smaller, resulting in what Aguiar and Hurst call "a growing inequality in leisure that is the mirror image of the growing inequality of wages and expenditures".

You can see a similar story about time availability when it comes to spending time with children. In their paper "Parental Education and Parental Time with Children", Jonathan Guryan, Erik Hurst, and Melissa Kearney find that more educated and higher earning men and women spend more time with their children. This is true when you control for a variety of demographic characteristics, and when you control for employment status. Consistent with the other research, they find that the greater time spent on child care comes out of less leisure and somewhat less home production.

This is a lot of information for a quick throwaway sentence from Richard Florida, but I think his sentiment is a popular one. In fact I think it reflects a very fundamental bias that I discussed in my past previous post, which I call the "complimentary bias". People search for and favor theories of behavior and outcomes which are complimentary to groups they want to raise in status. For liberals, this is low socioeconomic status households. Unfortunately, this bias reduces our understanding of the actual reasons for different outcomes and behaviors.