Labour-saving technologies and occupational exposure
Dr. Maria Enrica Virgillito , Sant’Anna School of Advanced Studies
The increasing diffusion of artificial intelligence (hereafter, AI) and robotic technology in the last decade has became a renewed object of analysis in both economics and technology studies. Robots, and intelligent robots more so, represent, among the wider spectrum of the recent Industry 4.0 wave, the technological artefacts ‘naturally’ apt in substituting human labour. However, the actual implementation of these artefacts may well be labour-friendly, as in the case of collaborative robots. At the current stage, the economic literature tends to rely on experts judgement (so-called Delphi method) when constructing automation probability measures of occupations (see Arntz et al., 2016; Frey and Osborne, 2017; Nedelkoska and Quintini, 2018). However, a direct measure of human substitutability and occupational exposure, ideally based on the effective functions and operations which labour-saving technology aims at executing, is still missing.
In this paper we intend to fill the existing vacuum in the literature and to propose a direct measure of the actual penetration of labour-saving technologies within the occupational structure. To accomplish this objective, we develop a multistep strategy. First, leveraging on the identification of labour-saving technologies by means of natural language processing on robotic patents (Montobbio et al., 2020), we perform a task-based textual match between the descriptions of elicited CPC codes attributed to labour-saving patents and the O*NET dictionary of occupations. The match is constructed by means of a cosine-similarity matrix that informs us about the “closeness” of the two dictionaries of words.
After recovering a CPC-task matching, we weight each entry of the matrix by the frequency of the respective CPC in LS patents. In this respect we attribute a LS trait to each pair. We then aggregate tasks into occupations by assigning the cosine-similarity measure to each task weighted according to being core or supplementary as defined by the “Task Statements” of O*NET. In this way we recover a measure of exposure of each task and related occupations to LS technologies. According to our results, most affected occupations are “Material MovingWorker”, “Vehicle and Mobile Equipment Mechanics, Installers, and Repairers” (Logistic), “Food Processing Workers”. In order to externally validate our measure, we match with Occupational Employment Statistics (OES) from US Bureau of Labor Statistics and with median wage data for 6-digit SOC occupations (1999-2019). Lowess estimates present a monotonically negative relationship between occupational exposure and (i) wage levels, and (ii) employment growth. Remarkably, no expected U-shaped pattern is recovered. Cutting edge-innovative efforts look to be directed towards the weakest and cheapest segment of the labour market.
To view the full reference list of the abstract please click on the 'More Information' button below.
About the speaker
Maria Enrica Virgillito is Assistant Professor in Economics at the Institute of Economics and EMbeDS Department, Sant’Anna School of Advanced Studies. Formerly she was Assistant Professor at the Department of Economic Policy, Università Cattolica del Sacro Cuore, where she is currently research fellow. She was fellow at the Labor and Worklife Program at Harvard University and she is currently GLO fellow. She is actively engaged in EU H2020 projects as task coordinator (GROWINPRO and ISIGROWTH) and JRC tenders. She undertook research collaborations with the ILO and invited as expert by EU-OSHA.
Her research interests range from technological change, industrial dynamics, labour market organization and institutions, macroeconomic dynamics, agent-based modelling, technology and labour relations to evolutionary economics. Her research outputs have been published in international peer-reviewed journals, including Industrial and Corporate Change, Journal of Economic Behaviour and Organization, Journal of Economic Dynamics and Control, Journal of Evolutionary Economics, Journal of Economic Interaction and Coordination, Journal of Economic Surveys, Socio-economic Review, Structural Change and Economic Dynamics, World Development, International Labour Review.
She acts as Editor for Industrial and Corporate Change (Macro and Development), she is member of the editorial board of SINAPPSI, and Associate Editor for Structural Change and Economic Dynamics and for the Review of Evolutionary Political Economy.
Venue: via Zoom (please contact us at seminars@merit.unu.edu for the Zoom link)
Date: 25 February 2021
Time: 12:00 - 13:00 CEST
