Automation, Learning, and Career Dynamics

This National Bureau of Economic Research working paper investigates the impact of automating technology on career dynamics, human capital accumulation, and overall welfare. The authors develop a continuous-time general equilibrium model where workers acquire skills through learning-by-doing while performing tasks. The study highlights that learning rates are jointly determined with the share of automated tasks, the frontier of tasks maintained by managers, and transitions from worker to manager roles. A key finding is that economies with high learning capacity can settle into one of two stationary equilibria. In a high-learning equilibrium, cheaper automation technology enhances welfare by boosting the learning channel. Conversely, in a low-learning equilibrium, the same technological cost reductions can trap the economy in a state of diminished human capital. To address these disparities, the authors propose a policy framework combining a tax on automation profits with a subsidy for frontier-maintenance expenditures at a common rate, aiming to achieve the planner's first-best outcome and mitigate potential negative impacts on workforce development.