Automation and Occupational Network

Objective

While most research has analysed automation as a within-occupation substitution process, little is known about how technological shocks diffuse through networks of interdependent tasks and occupations. AUTON proposes a novel framework that integrates task-level data, automation technologies, and employer–employee microdata to quantify both direct and systemic effects of robotics and AI adoption. The first contribution is the construction of a task-based occupational input–output network, using O*NET and Large Language Models (LLMs) to capture productive interdependencies between occupations at an unprecedented level of detail. The second contribution is the identification of tasks that can be automated within the occupational input–output network, to analyse how automation alters the input–output relationships between occupations that depend on those tasks. The third contribution is an occupational model that embeds these shocks into a general equilibrium setting, disciplined with linked employer–employee data from Australia, allowing us to trace equilibrium feedbacks on wages, employment, and occupational mobility. By combining micro-level task structures with macroeconomic modelling, the project moves beyond aggregate measures of automation risk, offering a systemic perspective on technological diffusion. The expected outcomes include new empirical measures of automation spillovers, simulation-based evidence on systemic vulnerabilities, and a calibrated model suitable for counterfactual policy analysis. This framework will provide both academic and policy communities with novel tools to evaluate how automation affects the resilience of labour markets and the distribution of work.