Education-gender earnings crossover

It has been remarked that, in the olden days, men with only a high school education earned more than women with a college education. That’s true, as I reported to Stephanie Coontz for today’s New York Times column. And the olden days ended about 20 years ago.

Until the early 1990s, men who were high school graduates – but not college graduates – earned more than women who were college graduates. And men who were high school dropouts earned more than women who were high school graduates.

cps-educ-gender-earnings-62-2012Source: My analysis of March Current Population Survey data from IPUMS.

I don’t know what’s going on with the big 1992 drops. It could have to do with CPS survey design changes (I used last year’s earnings for people who worked 35+ hours last week and 50+ weeks last year).

You could describe these crossovers as a modernization of the labor force, with education rising in importance relative to gender. Or not.

4 thoughts on “Education-gender earnings crossover

  1. Be very wary of the coding change in the education variable. I have charts like this with crossovers for veterans and non-veterans at the same time. If I recall, some college is wonky. You might break that out and see what it shows.


    1. Yeah, I would not mess around with “some college” in the older data. For one thing it was less common as a category, and for another it was not well measured. But high school graduate and college graduate are pretty stable I think. My guess is this has more to do with the estimation of hours and weeks worked. Plus, it was a recession so some decline in 1992 is to be expected.


  2. Emailed comment from Kim Weeden, posted with permission (thanks):

    Before 1992, the CPS used highest grade completed to measure education. After 1992, it used highest degree attained. Unfortunately, this means that even the HS and College grad categories aren’t strictly comparable over the series. David Jaeger has come up with an algorithm to reconcile the “new” and “old” coding, and also estimated the effect of the switch on the observed returns to education (Jaeger, Economic Letters 2003). I don’t remember if the earnings effect he found was large enough to explain the drop you’re seeing in your graphs. and obviously there could be other stuff going on as well, both in the data and in the economy.


Comments welcome (may be moderated)

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s