On the revealed comparative advantages of Dutch cities

Davis and Dingel explain the distributions of skills, occupations, and sectors across cities. Their model predicts that larger cities will be relatively skill-abundant and specialize in skill-intensive activities. This relates the model to factor-driven comparative advantages. They also develop an elasticity test and pairwise comparison test for the spatial distributional implications of the model. What is not analyzed, however, is the associated structure of trade flows. This next step—the analysis of the structure of trade—is the main contribution of our article. We combine micro-economic data to analyze how the sorting process of factors of production across cities determines the revealed comparative advantage (RCA) distributions of Dutch cities. We find that (i) the sorting of factors of production across cities is consistent with Davis and Dingel, (ii) RCA patterns differ significantly across locations, (iii) RCA differences can be explained by the interaction of local skill-abundance and sector skill-intensity (in line with the factor abundance model), and (iv) the RCA analysis relative to the Netherlands mostly, but not always, coincides with that relative to the world.