Offshoring and automation are sources of wage polarization. We reassess these two determinants of wage polarization in a single directed technical change setup that encompasses routine and nonroutine production. We empirically establish the conditional positive relationship between automation and relocations on one side and wage polarization on the other. Theoretically, we show that wage polarization increases with automation and offshoring. In particular, wage polarization in favor of domestic (nonroutine) high(low)-skilled workers is positively affected by an increase in domestic (nonroutine) high(low)-skilled labor quantity and/or absolute productivity. Additionally, it is also positively influenced by a rise in foreign (routine) medium-skilled labor quantity and/or absolute productivity while negatively impacted by an increase in domestic (routine) medium-skilled labor quantity and/or absolute productivity. We show that the effect of offshoring on wage polarization diminishes with the degree of substitutability between routine and nonroutine sectors in the economy, with the share of machines in the production of intermediate goods, and with the scale effect. We quantitatively assess the impact through a thorough data-based calibration exercise, where the numerical results confirmed the theoretical findings.

I develop an assignment model to quantify, in a unified framework, the causal effects of supply and demand forces on the evolution of the college wage premium in the US economy. Specifically, I quantify the relative contributions of four different forces: (i) a within-sector non-neutral technological change, (ii) the creation of new high-skill services/sectors, (iii) polarizing product demand shifts, and (iv) shifts in the relative supply of skilled labor. The model considers endogenous human capital accumulation. I find that positive supply shifts completely explain the fall of the skill premium during the period 1970–1980. Demand forces play a major role in the post-1980 period, when the skill premium rises. Among the demand forces, the results show an increasing contribution of polarizing product demand shifts over the decades. On the other hand, the effect of the within-sector non-neutral technological change is more important in the earlier decades of the post-1980 period.