Has job polarization squeezed the middle class? Evidence from the allocation of talents

Over the last two decades, earnings in the United States increased at the top and at the bottom of the wage distribution but not in the middle - the intensely debated middle class squeeze. At the same time there was a substantial decline of employment in middle-skill production and clerical occupations - so-called job polarization. I study whether job polarization has caused the middle class squeeze. So far little evidence exists about this because the endogenous selection of skills into occupations prevents credible identification of polarization’s effect on wages. I solve the selection-bias problem by studying the changes in returns to occupation-specific skills instead of the changes in occupational wages using data over the two cohorts of the National Longitudinal Study of Youth (NLSY). This data features multidimensional and pre-determined test scores, which predict occupational sorting and thus measure relative occupation-specific skills. My estimation equations are derived from the Roy (1951) model over two cross-sections with job polarization amounting to a shift in the occupation-specific skill prices. In line with polarization, I find that a one percentage point higher propensity to enter high- (low-) as opposed to middle-skill occupations is associated with a .29 (.70) percent increase in expected wages over time. I then compute a counterfactual wage distribution using my estimates of the shifts in occupation-specific skill prices and show that it matches the increase at the top of the wage distribution but fails to explain the increase at the bottom. Thus, despite the strong association of job polarization with changes in the returns to occupation-specific skills, there remains room for alternative (e.g. policy related) explanations about the increase in the lower part of the wage distribution.