Automation, Career Values, and Political Preferences (in cooperation with Maria Petrova, Gregor Schubert and Bledi Taska)

Career opportunities and expectations shape people's decisions and can diminish over time. In this paper, we study the career implications of automation and robotization using a novel data set of resumes from approximately 16 million individuals from the United States. We calculate the lifetime "career value" of various occupations, combining (1) the likelihood of future transitions to other occupations, and (2) the earning potential of these occupations. We first document a downward trend in the growth of career values in the U.S. between 2000 and 2016. While wage growth slows down over this period, the decline in the average career value growth is mainly due to reduced upward occupational mobility. We find that robotization contributes to the decline of average local labor market career values. One additional robot per 1000 workers decreased the average local market career value by $3.9K between 2004 and 2008 and by $2.48K between 2008 and 2016, corresponding to 1.7% and 1.1% of the average career values from the year 2000. In commuting zones that have been more exposed to robots, the average career value has declined further between 2000 and 2016. This decline was more pronounced for low-skilled individuals, with substantial part of the decline coming from their reduced upward mobility. We document that other sources of mobility mitigate the negative effects of automation on career values. We also show that the changes in career values are predictive of investment in long-term outcomes, such as investment into schooling and housing, and voting for a populist candidate, as proxied by the vote share of Trump in 2016. We also find further evidence that automation affected both the demand side and supply side of politics.