The Polanyi Paradox

In 2014 the labor economist David Autor gave a name to something that had been sitting in the employment statistics for three decades without one. The pattern was this: automation had devastated middle-skill, routine work—bookkeeping, assembly, data entry—while leaving largely untouched both the high-skilled professional work at the top of the wage distribution and the manual, physical, perceptual work at the bottom. The naive story was that machines take the unskilled work and leave the high-skilled. The data refused that story: what machines had taken was not the unskilled work but the explicit work, the tasks whose procedures could be fully specified, regardless of whether the specification was high-status or low. What resisted automation was the tacit: the tasks performed by people who could not say, in complete rules, how they did them. Autor called this the Polanyi paradox, attributing it to Michael Polanyi’s 1966 principle that