Opening the "skill-bias technological change" black box: a look at the microfoundations of the technology-skill relationship

The present article is a review of the recent empirical literature developed around the issues of why technology is complementary to high levels \u2013 and substitute for low levels \u2013 of skill, and, in particular, of how the adoption of ICT and computer-based machines has changed the skill requirements of jobs. During the last two decades the discussion around the impact of technological diffusion on the skill mix of employees has been intense. On this purpose, different approaches have developed that provide different evidence to a common research question. The paper shows that traditional studies have just inferred the skill-biased technical change hypothesis by employing broad measures of technological change and human skills from sector- and firm-level data. While studies that rely on worker-level data depict a more puzzling phenomenon, a recent literature based on job-level analyses focuses the heterogeneity of both technology and skills and aims at determining the demand for labor by the tasks that occupations require. The main conclusion is that technology is a partial substitute for repetitive manual tasks, and a complement of nonroutine, non-manual skills, for which more educated workers enjoy a comparative advantage. However, some open questions still remain that claim a deeper and multidisciplinary investigation on the endogenous relationship between technology adoption and general versus specific skill accumulation