Trade Shocks and the Shifting Landscape of U.S. Manufacturing

"...: Using data over more than a century, we show that shifts in the location of manufacturing industries are a domestic reflection of what the international trade literature refers to as the product cycle in a cross-country context, with industries spawning in high-wage areas with larger pools of educated workers and moving to lower-wage areas with less education as they age or become “standardized.” We exploit the China shock industries as a set of industries that were in the late-stage product cycle by 1990 and show how the activity in those industries shifted from high-innovation areas to low-education areas over the 20th century.... the analysis in this paper suggests that the resilience of local labor markets to manufacturing shocks depends on local industries’ phase in the product cycle, on their education levels, and their manufacturing wages, with the greatest risk of unemployment and detachment from the labor force occurring when a shock hits areas where an industry already is phasing out, wages are high, or education levels are low. The results are consistent with the belief that there are long-term, secular trends in U.S. industrial structure driving the movement of industries, which shocks may mitigate or accelerate....The analysis also suggests that the resilience of local labor markets to manufacturing shocks depends on local industries’ phase in the product cycle, on local education levels, and on local manufacturing wages....Areas where the shock-affected industries had been reducing employment the most between 1960 and 1980— which we interpret as a characteristic of local industries being in a late stage of the product cycle—experienced a roughly 0.25 basis point additional increase in the unemployment rate above that attributable to the shock by itself. Locations with high manufacturing wages also felt more pain conditional on their level of exposure. This could occur if firms had been holding on in a location due to sunk costs as in Fillat and Garreto (2015) or adjustment costs as in Rodriguez- Lopez (2014) amidst increasing pressures to move or exit, then a large shock jolted them out of the range of inaction. Areas with a greater fraction of workers with college degrees had somewhat less adverse labor-market outcomes conditional on the shock, suggesting that a highly educated workforce may be better able to achieve resilience through innovation or expansion in different sectors along the lines of Feenstra, Ma, and Xu (2017), for instance.....The risk of unemployment and detachment from the labor force rises most when a shock hits in areas where an industry already has begun phasing out, wages are high, or education levels are low. The results are consistent with the belief that there are long-term, secular trends in U.S. industrial structure driving the movement of industries, which shocks may mitigate or accelerate.... As points of comparison, we compute an analogous “Japan shock" for the period from 1975- 1985 and “Tiger shock” from 1975-1988 and find that the regions most directly affected were quite different from those hit by the China shock. In particular, exposure to the China shock is correlated with lower educational attainment, lower incomes, and higher unemployment rates. In contrast, exposure to the Japan shock shows no such pattern, nor do areas with high concentrations in China-Shock industries in 1910 through 1960. Thus, the China shock was concentrated in areas that look distinctly different from regions hit by earlier shocks to manufacturing. We argue that the China shock hit industries late in their product cycle, while the Japan shock did not...."

Katherine Eriksson, Katheryn Russ, Jay C. Shambaugh, Minfei Xu, "Trade Shocks and the Shifting Landscape of U.S. Manufacturing," National Bureau of Economic Research, March 8, 2019, https://katherynruss.weebly.com/uploads/4/5/1/7/45179109/mfgshifts_ersx.pdf