Baumol, Goldin, and benchmark choices

A bit of an extension on Saturday’s entry: As an historian of education, I find Claudia Goldin to be a much more interesting economist than William Baumol, because she does not rely on arbitrary choices of starting points for her economic stories. Like many economists (and historians), Baumol tells a story that depends on a starting point, in his case in the 20th century. Baumol’s story is one of differential productivity increases, and specifically lower productivity increases in labor-intensive fields than in capital-intensive fields (or fields that become more capital-intensive). The classic story Baumol tells is of musicians who still play the same Mozart quartet in the same time. Voila! No productivity increases, except for all that messy stuff about recording and the relevant measure being listening time and not playing time and some other things Tyler Cowen describes succinctly.

Baumol is certainly not the only person who has tried to apply this concept to education — for recent examples, see Matt Reed and Matthew Ladner (and also people not named either William or Matthew). I’ve explained before why I am not entirely persuaded by Baumol, and explored some issues with thinking about educational productivity, but let me briefly explain here the fifth argument of Tyler Cowen: high- and low-productivity-increase sectors change over time. And that means that the starting point of one’s analysis matters. If you were to look at American education in the early 1970s, you would have to say that over the previous 50 years, it had seen remarkable “productivity increases” by objective measures, with huge increases in the proportion of young adults graduating from high school and college. It is only in the past 40 years or so that one could argue K-12 schools have become less productive in some ways. (For the purposes of argument, I am going to ignore the question of which measures might show K-12 schooling to be more or less productive since 1970. The point is that the first two-thirds of the twentieth century show remarkable growth in high school and thus college attainment.)

That shift is the setup for Goldin and Katz’s 2008 book, The Race Between Education and Technology. Baumol and many others ignore the shift and assume that low productivity increases for various measures are inherent in either the labor-intensive nature of schooling or the domination of K-12 education by public schooling. Neither is supportable if you move the starting point to the beginning of the twentieth century. I am not saying I can explain that shift, but that explanations ignoring the shift do not hold much water.