Who is Afraid of Technological Change? A Bayesian IRT Modeling Approach

Author – Tobias Tober

Abstract :

The labor market implications of rapid technological change are vast and — given recent quantum leaps in areas like artificial intelligence — uncertain, and so are its political consequences. A potential way to deal with this uncertainty is to rely on subjective perceptions of technology-related unemployment risks. However, in order for these subjective risks to exert any explanatory power, we need to have a thorough understanding of what explains their formation. Thus, this paper asks the following question: What are the determinants of subjectively perceived employment risks from technological change? Drawing on two original survey sources, I attempt to answer this question using Bayesian Item Response Theory modeling as a way to infer the latent concept — i.e., fear of technology-induced job loss — from the data and identify the factors that cause these subjective fears. The results show that general exposure to modern technologies at the workplace is a strong predictor of subjective risk. In contrast to the view that individuals with high skills and cognitively demanding job tasks are less vulnerable, I find that perceived employment risks increase with the work-related use of complex technologies like programming languages and statistical software. Moreover, the empirical analysis suggests that the welfare-state context as measured by the generosity of unemployment insurance has a dampening effect on subjective technological risk.