Comparative Advantage in (Non-)Routine Production

We illustrate a new source of comparative advantage that stems from countries’ different ability to adjust to technological change. In our model, workers in codifiable (routine) tasks can be substituted for more efficient machines, a process extensively documented in the labor literature. Our key hypothesis is that automating routine tasks, and the corresponding labor reallocation, is subject to frictions, the importance of which varies by country. The arrival of capitalaugmenting innovations triggers the movement of workers out of routine tasks. More flexible countries become relatively abundant in non-routine labor and increasingly specialize in producing goods that are intensive in non-routine tasks. We document empirically that the ranking of countries with respect to the routine intensity of their exports is strongly related to labor market regulations and the organization of the workplace. This mechanism has the best explanatory power for intra-EU trade flows of all five mechanisms that we compare