Working Paper: CEPR ID: DP16278
Authors: Philipp Herkenhoff; Sebastian Krautheim; Philip Saur
Abstract: The recent literature on firm-to-firm trade has documented salient empirical regularities of the buyer-seller network. We propose a simplistic re-interpretation of the classical Krugman (1980) model that accounts for surprisingly many of the empirical regularities. This re-interpretation relies on randomized bundling of Krugman-varieties into heterogeneous firms, economically neutral `sales units' that import foreign varieties but belong to local firms, and a statistical reporting threshold that applies to firm-to-firm transactions. We argue that our model provides an important benchmark for the assessment of theoretical models that aim to identify the determinants of firm-to-firm networks.
Keywords: firm-to-firm; buyer-seller; trade network; random matching
JEL Codes: F10; F12; F14
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| gravity forces (F20) | buyer margin (G21) |
| gravity forces (F20) | seller margin (L81) |
| number of partners (J12) | aggregate trade volume (E10) |
| number of sellers or buyers (D44) | share of local firms with more foreign sellers or buyers (F23) |
| size and connectedness of a firm (L25) | sales to median buyers (D19) |
| importer's set of exporters (F10) | larger importer's set of exporters (F10) |
| connected firms (L14) | less well-connected partners (L14) |