Expertise-Leveling Technology and the Division of Creative Labor

Abstract

What happens when an expertise-leveling technology diffuses through an industry? I study digital sampling, which makes recorded performances reusable and reduces the expertise required for music production. I assemble novel data on production credits, sample use, and content for 70,712 US records released between 1980 and 2000. A difference-in-differences design identifies the causal effects of geographic exposure to a new sampler’s 1988 promotional campaign attached to Michael Jackson’s tour. Exposure increased sampling. Main artists on sampled records were more likely to be self-taught and less likely to report formal musical training. Participation in music production, measured as participants per 100,000 working-age residents, showed no detectable change. On exposed records, the number of people working exclusively in writing, composition, arrangement, and production grew by 0.5, roughly 15 percent of the pre-period mean, with no detectable broadening of individual task bundles. Records became more acoustically distinct from recent music before the organizational response emerged, while lyrics showed no significant change. These findings suggest that making existing performances easier to reuse can support specialized creative contributions and music that departs from the recent past.