Too Old to Adjust. Aging and the Speed of Automation Adoption

We study how demographic composition changes the transition costs of rapid automation adoption in a speed–capacity model where displaced workers differ by age and the young share of the displacement pool declines over time. Older workers face higher discouragement hazards and lower retraining completion rates, so demographic aging deteriorates aggregate transition outcomes even when institutional capacity is unchanged—a demographic composition effect. This effect interacts with adoption speed through congestion: fast adoption in an older displacement pool is worse than the sum of its parts (supermodularity). We derive a demographic gradient in optimal adoption speed: the planner adopts faster when the displacement pool is younger. Numerically, in the calibrated region, this demographic gradient becomes steeper when aging is faster. We also identify a demographic buffer: a timing boundary beyond which earlier adoption may dominate delay, although the sufficient condition is quantitatively demanding under OECD-style calibrations. We derive six cross-country predictions and an explicit measurement strategy.