Abstract:

We used text-to-data analysis to build three new datasets from firm quarterly earnings call, burning glass job adverts and US daily newspaper stories. Armed with this and patenting data we identify 22 new technologies and their roll-out across firms and labor markets in the US. Three stylized facts jump out from our data. First, as technologies develop the number of workers using this technology grows in size but their education and wage level drops. This is a natural process in the move from early stage development of technologies which is heavy is PhD employees to the latter stage deployment which is typically heavier in college and high-school only educated employees. Second, as technologies develop their employment impact diffuses across the country. Initially, technologies appear to be concentrated in a few local hubs but over time with rising deployment their use diffuses across the country. Finally, the origin of technology hubs appears to be driven by the universities and high skilled labor pools. Technologies tend to originate in college towns and slowly diffuse across the country.