Learning to Export from Your Neighbor by Miklós Koren and Ádám Szeidl
Wednesday, April 17, 2013 - 4:00pm to 6:00pm
Learning to Export from Your Neighbor
Miklos Koren and Adam Szeidl
We use the panel of all Hungarian firms 1992-2003 to document evidence for knowledge spillovers in firm exports. Our research design is to compare, in an event study, the export behavior of firms which did versus did not have a past exporter geographically close (in their street or city block). To deal with geographically correlated omitted variables, we exploit two identification strategies. (1) We use cross-destination variation, testing whether firms which had, in their street, a past exporter to country A but not to country B are more likely to export to country A. (2) We measure knowledge spillovers using a highly localized---street and block-level---definition of neighborhood, obtaining identification from the assumption that knowledge spillovers decay faster in distance than do omitted variables. Our estimates, which control for building-year fixed effects, imply that subsequent to a a firm entering an export market, knowledge spillovers generate 0.21 new exporters to that market over a five year time horizon. These results suggest that knowledge spillovers help explain three trade puzzles: why countries trade too little with each other, why trade flows are elastic to tariff reductions, and why the export behavior of seemingly similar firms is heterogenous.