Testing the Automation Revolution Hypothesis (original) (raw)

We test basic theory, two expert-derived vulnerability metrics, and 251 O*NET job features as predictors of 1505 expert reports regarding automation levels in 832 U.S. job types from 1999 to 2019. Pay, employment, and vulnerability metrics are predictive (R 2~0 .15), but add little to the top 25 O*NET job features, which together predict far better (R 2~0 .55), seem understandable in terms of traditional types of automation, and did not change over this period. Instead, job features changed to be more suitable for automation. Over this period, job automation increases have predicted neither changes in pay nor employment.