Tom Zylkin | Gravity with Persistence

The Moynihan Institute’s program for Trade, Development and Political Economy welcomes Tom Zylkin from the University of Richmond to discuss his working group's newest paper, “Gravity with Persistence.”

For a variety of reasons, trade flows may exhibit persistence. Exporting may require building destination-specific relationships and/or capital, and trade may generally need time to adjust to sudden trade cost shocks. Moreover, trade policies themselves may depend on recent changes in trade, feeding back into current trade dynamics.

We document how standard estimates of the effects of trade policies using static panel data gravity models are significantly biased under any of these types of dynamics. Yet accounting for them is no easy feat, since introducing endogenous regressors causes the three-way fixed effects PPML estimator to suffer from a severe asymptotic bias problem even if the number of countries and the number of time periods are both large.

To resolve these challenges, we propose a dynamic extension of the static gravity model that allows for zeroes in both present and past trade and demonstrate bias correction methods that remove most of the asymptotic bias. Even for static gravity model specifications, our bias corrections can serve to remove bias due to dynamic feedback between trade and trade policy variables.

Paper Authors: Amrei Stammann (University of Bayreuth, Amrei.Stammann@uni-bayreuth.de), Joschka Wanner (University of Würzburg, joschka.wanner@uni-wuerzburg.de) Tom Zylkin (University of Richmond, tzylkin@richmond.edu)