Joint with Anna Salomons and Bryan Seegmiller
Recent theory stresses the role of new job types (‘new work’) in counterbalancing the erosive effect of task-displacing automation on labor demand. Drawing on a novel inventory of eight decades of new job titles linked to United States Census microdata, we estimate that the majority of contemporary employment is found in new job tasks added since 1940 but that the locus of new task creation has shifted—from middle-paid production and clerical occupations in the first four post-WWII decades, to high-paid professional and, secondarily, low-paid services since 1980. We hypothesize that new tasks emerge in occupations where new innovations complement their outputs (‘augmentation’) or market size expands, while conversely, employment contracts in occupations where innovations substitute for labor inputs (‘automation’) or market size contracts. Leveraging proxies for output-augmenting and task-automating innovations built from a century of patent data and harnessing occupational demand shifts stemming from trade and demographic shocks, we show that new occupational tasks emerge in response to both positive demand shifts and augmenting innovations, but not in response to negative demand shifts or automation innovations. We document that the flow of both augmentation and automation innovations is positively correlated across occupations, yet these two faces of innovation have strongly countervailing relationships with occupational labor demand.