The Great (Male) Stagnation

You have probably seen something like the following graph which shows real GDP per capita and median male income since 1947. Typically, the graph is shown with family or household income but to avoid family-size effects I use male income. It’s evident that real gdp per capita and median male income became disconnected in the early 1970s. Why?  Explanations include rising inequality (mean male income does track real gdp per capita somewhat more closely), Tyler speculates that the nature of technological advances has changed, other people have speculated about rising corporate profits. Definitive answers are hard to come by.

Here is another set of data that most people have not incorporated into their analysis:

Median female income tracks real GDP per capita much more closely than does median male income. It’s unclear which, if any, of the above explanations are consistent with this finding. Increasing inequality, for example, predicts an increasing divergence in real GDP per capita and female median income but we don’t see this in the graph (there is a slight increase in the absolute difference but the ratios don’t increase). Similarly, we would expect changes in technology and corporate profits to affect both male and female median income equally but in fact the trends are very different.

One can, of course, do the Ptolemaic move and add an epicycle for differences in male and female inequality and so forth. Not necessarily wrong but not that satisfying either.

The big difference between female and males as far as jobs, of course, has been labor force participation rates, increasing strongly for the former and decreasing somewhat for the latter. Most of the female change, however, was over by the mid to late 1980s, and the (structural) male change has been gradual. Other differences are that female education levels have increased dramatically and male levels have been relatively flat.  Females are also more predominant in services and males in manufacturing: plumbers, car mechanics, carpenters, construction workers,  electricians,  and firefighters, for example are still 95%+ male.  Putting these together points to a skills and sectoral story, probably amplified by follow-on changes in labor force participation rates.

Thinking about the story this way also reminds us that the median male or female is not a person but a place in a distribution. The median male in 1970 can get rich by 1990 even though median male income is flat.

Again, no definitive answers, but the raw patterns are striking.

Note: An extra high tip of the hat to Scott Winship who whipped up all of the data during a discussion.


Comments for this post are closed