The Skill-Specific Automatability of Aging Workers and Its Impact on Retirement Decisions

Abstract Research regarding the effects of automation on labor supply often assesses the labor force as a whole and disregards specific effects on aging workers. In light of rapid technological changes in the labor market, we assess the linkage between the automatability of aging workers and their retirement decisions. Based on the theoretical model of task-based technological changes and drawing data from the Health and Retirement Study and O*NET 2000–2018, we create an automatability index based on workers’ primary skills. Using the index as our main explanatory variable in Cox proportional hazards models and logit models, we find that skill-specific automatability increases the retirement likelihood, both in terms of their expected and actual timing of retirement. This work provides empirical evidence that individuals’ automatability renders the notion of “working at old age” less viable, despite the financial and health benefits of staying in the labor force for an extended period. Our findings offer important insights on how to better promote productive aging, for instance, by offering retraining programs for older workers to harness their soft skills to reduce automatability in the labor market.