Relatedness diffuses impacts from automation: a novel approach to estimate job risk in US cities

Modern Robotics, artificial intelligence and other new technologies are changing the workforce landscape worldwide. They threaten some jobs and improve some other jobs, directly and indirectly. When workers get automated, some of their co-workers might become more productive and higher in demand. Some other co-workers might instead become more prone to getting automated too, as the new technology develops further. This paper empirically investigates how automatable jobs have affected their neighbouring jobs in North American cities between 2007 and 2016. Results indicate that impacts of new technologies can be positive or negative depending on how occupations relate to each other. More concretely, occupations that share similarities with neighbouring high-risk occupations grew less, even when controlling for their own technical risk of automation. Conversely, occupations that share complementarities with neighbouring high-risk occupations grew more, possibly indicating productivity gains from working with recently automated jobs.