How Long Has this Been Going on? A Discussion of “Recent Flattening in the Higher Education Wage Premium: Polarization, Skill

Robert Valletta’s paper “Recent Flattening in the Higher Education Wage Premium...” illuminates one of the leading puzzles for contemporary U.S. labor economics: the unexpected ‘flattening’ of the premium to higher education in the United States in the 2000’s. This single metric—the college/high-school wage premium—has been the North Star guiding neoclassical analysis of the evolution of wage inequality during a period of rapidly shifting wage structures. Two impactful papers by Beaudry, Green and Sand (2014, 2016, BGS hereafter) argue that since approximately the year 2000, this North Star has become an increasingly dubious point of navigation. Specifically, BGS highlight the failure of the college premium to rise in the 2000s following two decades of steep increases. They interpret this deceleration as reflecting the maturation of the information technology revolution, which in turn has spurred a slackening in the pace of workplace IT investments and a consequent slowdown in the trend of rising demand for highly educated labor. A key piece of evidence favoring BGS’s narrative is the precipitous fall in U.S. investment information processing equipment and software in the U.S. after 1999 (Figure 1), which seems to have precisely the right timing to explain a falloff in IT-augmentation of skilled labor demand. Valletta’s careful analysis extends and probes the BGS findings, verifies their robustness, and considers their interpretation in the light of both the BGS conceptual framework and an alternative framing offered by Acemoglu and Autor (2011). There are many things to admire about Valletta’s paper: it is empirically rigorous, intellectually ecumenical, and commendably ambitious in synthesizing and adjudicating between two conceptual models that are not, to a first approximation, speaking the same language. My remarks focus exclusively on one question that is core to both Valletta’s BGS’s work: when did rising demand for college-educated labor decelerate? I argue below that (1) the recent flattening of the skill premium in the 2000s is not surprising in light of the canonical supply-demand model; and (2) what is surprising is that the underlying demand for college labor decelerated sharply and (to date) inexplicably almost a decade beforehand. These observations render the phenomenon that Valletta tackles no less consequential. But they may suggest a different set of explanations for the slowdown than those focusing on discontinuous changes in economic trends