Automation, Job Polarisation, and Structural Change

Abstract

The increasing automation of tasks traditionally performed by labour is reshaping the relationship between skills and tasks of workers, unevenly affecting labour demand for low, middle, and high-skill occupations. To investigate the economy-wide response to automation, we designed a multisector Agent-Based Macroeconomic model accounting for workers’ heterogeneity in skills and tasks. The model features endogenous skill-biased technical change, and heterogeneous consumption preferences for goods and personal services across workers of different skill types. Following available empirical evidence, we model automation as a manufacturing-specific, productivity-enhancing, and skill-biased technological process. We show how automation can trigger a structural change process from manufactory to personal services, which eventually increases the share of high and low-skilled occupations, while reducing the share of middle-skilled ones. Following the literature, we label this dynamics as job polarisation throughout the paper. Finally, we study how labour market policies can feedback in the model dynamics. In our framework, a minimum wage policy (i) slows down the structural change process, (ii) boosts aggregate productivity, and (iii) accelerates the automation process, strengthening productivity growth within the manufacturing sector.