Managing a Polarized Structural Change

We build a multi-sector occupation-choice model where routinization leads to job and wage polarization across occupations, and to a structural change across industrial sectors. Our model also incorporates inequality between managers and workers, which we call vertical polarization (to differentiate from polarization among workers in horizontally differentiated occupations). Individuals are heterogeneous in managerial talent and worker human capital, and choose from four occupations: a manager or a high/middle/low-skill task worker. Their choices determine sector-level/aggregate TFPs. Consistent with the data, an increase in the productivity of the middle-skill/routine task leads to job and wage polarization across workers. A novelty is that routinization leads to vertical polarization---increases in the employment share and wages of managers relative to workers. The model predicts that the speed of polarization is faster in sectors that use the routine task more intensively. An empirical contribution is the documentation of new facts that support these predictions. The differential impact of routinization across sectors leads to structural change by shifting capital and labor toward sectors that are less dependent on the routine task. In the limiting balanced growth path, the routine task/occupation vanishes, but all sectors coexist and grow at the same rate.