Study: Automation's Ripple Effects on Careers and Voting Patterns
A new NBER working paper ("Automation, Career Values, and Political Preferences" by Maria Petrova, Gregor Schubert, Bledi Taska, and Pinar Yildirim) examines how automation and robotization affected career prospects and political preferences in the U.S. from 2000 to 2016.
The big picture: Increased robot adoption led to declining career values and upward mobility, especially for low-skilled workers. This correlated with shifts in voting behavior and support for populist candidates.
By the numbers:
One additional robot per 1,000 workers decreased the average local market career value by $3,820 (1.7% of the 2000 average)
Low-skilled workers saw a $1,520 reduction in lifetime earnings from upward mobility vs. $1,040 for high-skilled workers.
One standard deviation increase in career values led to a 0.68 percentage point decrease in Trump's 2016 vote share.
Average local market career value grew by $2,100 (0.9%) between 2000-2008 but declined by $200 (0.09%) from 2008-2016
Key findings:
Career values (expected lifetime earnings) declined, especially after 2008
Automation reduced both wages and upward job mobility
Effects spilled over to non-manufacturing industries like retail and services
Areas with more education opportunities saw fewer negative impacts
Lower career values correlated with less investment in housing and education
Automation exposure increased support for Trump in 2016 but not previous GOP candidates
Younger voters' income was more strongly associated with Trump support, consistent with career value concerns
Republican candidates made more campaign visits to areas with higher robot exposure
Methodology:
Used resume data from 16 million individuals to calculate "career values" based on job transition probabilities and wages
Instrumented U.S. robot adoption using industry-level adoption in EU countries
Controlled for demographic characteristics, manufacturing share, and other factors
Between the lines: The study suggests automation's effects on future career prospects, not just immediate job losses, shaped economic and political outcomes. It also found evidence of the political impacts of supply-side (e.g., campaign visits) and demand-side (voting intentions).
What's next: The authors hypothesize AI could exacerbate inequality and erode the middle class by replacing white-collar jobs and reducing demand for middle managers.
The bottom line: Policymakers should consider automation's broader impacts on career mobility and inequality when crafting responses. The study highlights the importance of forward-looking measures like career values in understanding labor market dynamics.