Abstract
This paper sheds new light on the disproportional wage decline of low-skilled workers in local labour markets with denser populations, an unsolved puzzle in the contemporary US economy. While technological change is often considered a potential underlying factor, its manifestation in local labour markets and the related creation of skill-biased factor demand changes is still not fully understood. Using USPTO patent data combined with detailed occupation data from the O*NET, I provide evidence that faster innovation across commuting zones creates cognitive-biased task shifts within occupations, whereby the population density of commuting zones intensifies this relationship. Based on this new finding, in the second part of the paper, I use a Bartik shift-share design to show that higher exposure to occupation-biased technological change increases the employment of high-skilled workers, causes a decline in low-skilled workers' wages, and increases the college wage premium. A deeper exploration unveils that the deskilling of low-skilled workers through sorting into low-wage occupations accounts only for a small proportion of the relative wage decline of low-skilled workers. Thus, a substantial role must be attributed to the rising return to cognitive skills within occupations. This finding is consistent with previous studies showing that task changes occur mainly within rather than between occupations.