Working Paper: NBER ID: w8518
Authors: Donald R. Davis; David E. Weinstein
Abstract: One account of spatial concentration focuses on productivity advantages arising from market size. We investigate this for forty regions of Japan. Our results identify important effects of a region's own size, as well as cost linkages between producers and suppliers of inputs. Productivity links to a more general form of 'market potential' or Marshall-Arrow-Romer externalities do not appear to be robust in our data. Landlocked status does not matter for productivity of regions in Japan. The effects we identify are economically quite important, accounting for a substantial portion of cross-regional productivity differences. A simple counterfactual shows that if economic activity were spread evenly over the forty regions of Japan, aggregate output would fall by nearly twenty percent.
Keywords: No keywords provided
JEL Codes: F1; R1
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Region's own aggregate size (R10) | Productivity (O49) |
| Access to suppliers of inputs (L14) | Productivity (O49) |
| Good access to consumers (L81) | Productivity (O49) |
| Own region size (R12) | Market access (L17) |
| Neighboring regions (R11) | Market access (L17) |
| Market size (L25) | Productivity (O49) |
| Supplier access (L81) | Marshallian externalities (D62) |
| Economic activity distribution (D39) | Aggregate output (E23) |