Computerization and Skill Bifurcation: The Role of Task Complexity in Creating Skill Gains and Losses

Abstract

Does computerization increase or reduce the extent of skills that workers are required to have? The authors argue that the effects of computer-based technologies (CBT) are neither universal nor uniform. Rather, effects depend on the level of skill required by occupations prior to the introduction of CBT, and as such, a bifurcation emerges: occupations that historically (pre-computerization) could be accomplished with low skills and that entailed low-complexity tasks do not experience significant CBT, and they remain low skill or become less skilled occupations, whereas historically high-skill occupations that entailed high-complexity tasks see an increase in CBT and the skills they require. The authors test this proposition using a unique data set that includes measures of the degree of computerization and changes attendant to computerization in the level of seven skills of core employees for a sample of 819 firms during 2000. This data set is linked by core employees’ occupation to U.S. occupation-level data on three dimensions of task complexity during 1971 (pre-CBT). The authors find that occupations with higher pre-CBT task complexity are associated with subsequent adoption and intensity of CBT, and that CBT affects most skills positively. For simple tasks, however, CBT does not affect skills or affects them negatively. Results shed light on the skill-based technological change and skilling-deskilling debates and suggest that the relationships are contingent in more nuanced ways than the literature has suggested.

Keywords: skills, CBT, computerization, complexity