Network-driven differences in mobility and optimal transitions among automatable jobs

Abstract

The potential for widespread job automation has become an important topic of discussion in recent years, and it is thought that many American workers may need to learn new skills or transition to new jobs to maintain stable positions in the workforce. Because workers’ existing skills may make such transitions more or less difficult, the likelihood of a given job being automated only tells part of the story. As such, this study uses network science and statistics to investigate the links between jobs that arise from their necessary skills, knowledge and abilities. The resulting network structure is found to enhance the burden of automation within some sectors while lessening the burden in others. Additionally, a model is proposed for quantifying the expected benefit of specific job transitions. Its optimization reveals that the consideration of shared skills yields better transition recommendations than automatability and job growth alone. Finally, the potential benefit of increasing individual skills is quantified, with respect to facilitating both job transitions and within-occupation skill redefinition. Broadly, this study presents a framework for measuring the links between jobs and demonstrates the importance of these links for understanding the complex effects of automation.