On January 14, 2020, economists Brian Phelan of DePaul University and Daniel Aaronson of the Federal Reserve Bank of Chicago released a paper that tracks how low-wage automation has evolved during the first two decades of the twenty-first century. Their analysis focuses on the period after the 2008 Financial Crisis, a turning point that reshaped many aspects of the U.S. labor market.
The authors define technological substitution as the process by which machines or software replace human labor in specific tasks. By examining occupational data across the nation, they quantify how vulnerable different low-wage jobs are to this substitution and identify patterns that reveal both winners and losers in the evolving economy.
Acceleration of low-wage automation after the crisis
The study finds that the speed of low-wage automation has markedly increased since 2008. Jobs that involve cognitively routine tasks—such as data entry, basic bookkeeping, or simple customer-service scripting—experienced notable declines in employment. At the same time, occupations relying on manually routine activities, like assembly-line work or basic food-preparation roles, also saw a steady erosion of positions.
These trends are not confined to a narrow set of sectors. The authors report that automation has spread to a broader spectrum of occupations, meaning that the pool of vulnerable jobs is expanding. In quantitative terms, the employment decline in routine low-wage roles outpaces the
Growth in interpersonal-task jobs, but not enough to balance losses
While automation chips away at routine positions, the paper documents a parallel rise in occupations that demand strong interpersonal skills. Roles such as personal care aides, community outreach coordinators, and certain service-industry positions have seen increased hiring. The authors attribute this shift to the fact that machines are still less capable of handling nuanced human interaction, empathy, and relationship-building.
However, the surge in people-focused jobs does not fully counteract the The study notes that, before the crisis, the growth in interpersonal-task occupations was sufficient to offset the negative impact of automation on low-wage workers. Post-crisis, the rate of new interpersonal-task job creation lags behind the accelerating automation of routine tasks, leaving a net employment deficit for many low-skill workers.
Geographic and demographic disparities
One of the more striking findings is that the intensity of low-wage automation varies sharply across the United States. The phenomenon is most pronounced outside the nation’s largest metropolitan areas, where firms often have fewer resources to invest in worker retraining programs. Consequently, workers in smaller cities and rural regions are more likely to experience job displacement.
Demographically, the research highlights that younger, less-educated males—particularly those belonging to minority groups—bear the brunt of automation-driven job losses. These workers typically have limited access to advanced education or vocational training, making it harder for them to transition into the growing interpersonal-task segment of the economy.
Policymakers and educators face the challenge of designing interventions that accelerate retraining, especially in regions and communities most affected by the rapid advance of automation.
