Working Paper: NBER ID: w9020
Authors: Russell Hillberry; David Hummels
Abstract: We show that 'home bias' in trade patterns will arise endogenously due to the co-location decisions of intermediate and final goods producers. Our model identifies four implications of home bias arising out of specialized industrial demands. Regions absorb different bundles of goods. Buyers and sellers of intermediate goods co-locate. Intermediate input trade is highly localized. The effect of spatial frictions on trade are magnified. These implications are examined and confirmed using a unique data source that matches the detailed subnational geography of shipments to the characteristics of the shipping establishments. Our results broaden the measurement and interpretation of home bias, and provide new evidence on the role of intermediate inputs in concentrating production.
Keywords: home bias; intermediate goods; trade patterns
JEL Codes: F12; F15; R3
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| firm location choices (R30) | home bias in trade patterns (F14) |
| small trade costs (F19) | colocation of firms (R32) |
| colocation of firms (R32) | local availability of goods (R22) |
| industrial demands vary spatially (R22) | composition of trade (F10) |
| intermediate goods traded locally (F19) | final goods shipped over longer distances (L91) |
| trade barriers (F14) | magnified effect on trade (F69) |