Gravity Model Specification and the Effects of the Canada-U.S. Border

There is a well-established literature finding that the Canada-U.S. border has a large dampening effect on trade, is asymmetric, and differs across provinces. In this paper, I demonstrate that the standard gravity model used to obtain these results provides biased estimates of the volume of trade. I attribute this to heterogeneity bias and reestimate the effects of the border using a gravity model that allows for heterogeneous gravity equations. Doing so does not alter the general results of existing studies, although it does yield a border effect that is 40 percent larger, reverses the border's asymmetry, and indicates very different provincial effects.