ClaudiaFest reflection: Warnings about economic history and education

This weekend is a special conference held by the National Bureau of Economic Research, Human Capital and History: The American Record, or as the Twitter hashtag is suggesting, ClaudiaFest, in part a celebration of Claudia Goldin’s accomplishments as an economic historian. Oh, yes, and an economist, too. One of the papers is by Larry Katz and Bob Margo, and it provides an update to the optimistic view of Goldin and Katz in their 2008 book, The Race Between Education and Technology. In 2008, Goldin and Katz wove a story around what economists call technology-skills complementarity: briefly, that more education among U.S. adults could address the growing wage gap between those currently with high levels of education and those with either just high school diplomas or who are high school dropouts. The root of this was in what they called technology-skills complementarity, which is the tendency across a large economy for greater reliance on technology where there was anticipated growth in skills that would match the technology.1

The new paper by Katz and Margo put a subtle twist on this, noting that the deskilling of workers in 19th century manufacturing was only in a part of the economy, leading to a hollowing-out of demand for artisanal skills at the same time that the early-19th century factories hired lesser-skilled workers they could train to specific tasks as well as the early progenitors of white-collar workers. But most of the early 19th century economy was not manufacturing, and as a result the hollowing-out effect in manufacturing was swamped by other shifts in labor demands. They broaden this to a general explanation (on pp. 8-9):

In equilibrium the labor market assigns workers to tasks at a point in time.   Over time, technical change alters the assignment of workers to tasks, thereby feeding back on the demand for the underlying skills.   In recent years, for example, there has been dramatic erosion in demand for workers in middle‐skill white collar work, as these tasks can now be more cheaply undertaken by computer‐based technologies which also facilitate international outsourcing.  However, while the demand for middle‐skill jobs has eroded, the demand for those with higher levels of skills – for example, those who can design and market new software applications or invent more powerful algorithms or design faster computer chips – has increased.  Task‐based models demonstrate that technical change need not be uniformly skill‐biased but rather can be complementary with  skills in some tasks while substituting for skills in other tasks.

In the context of the 19th century, they argue, there was sufficient movement out of farm labor to consistently raise the skill level of the labor force. Thus, Katz and Margo argue that the consistently increased demand for skilled labor began well before the 20th century. The story they tell about the last 30-40 years is different in result if not mechanism, another hollowing-out of the skill distribution, except this time more broadly in the labor market rather than just in manufacturing. And yet, they argue, it comes from the same technical-change phenomenon, reflecting what they see as “the continuity of the effects of technical change on the relative demand for skill” (p. 39).

Their view is different from Paul Krugman, who posited a fairly straightforward capital-labor substitution as the explanation for the decreasing proportion of income going to labor. Who is correct, and why does that matter? If Goldin and Katz and Margo are correct about a monotonic increase in demand for skilled labor, and a mechanism that is persistent, then we should focus on that big picture rather than shorter-term consequences such as the recent hollowing-out of the middle of the skills distribution. That means we push for improved education and higher educational achievement and fight inequality more outside the labor market than inside it: the first improves the conditions for individuals, and the second addresses broader concerns abou inequality. If Krugman is correct, I think, then the hollowing-out may signal an important discontinuity in labor-market mechanisms, and increased education is not nearly as salutary as Goldin, Katz, and Margo imply. Or at least there are much more worrisome signs under the Krugman gloss.

Update 12/9: More from Krugman, with abstract graph, even.

If you enjoyed this post, please consider subscribing to the RSS feed to have future articles delivered to your feed reader, and sign up for my irregular newsletter below!


  1. This is not really the story of growing wealth inequality in the U.S., since a very large portion of increasing inequality happens within the top 1%, not the difference between the top 2-20% of wage earners and the rest of the scale. []