Border Effect or Country Effect? Seattle May Not Be so Far from Vancouver After All

This paper reexamines the evidence on the border effect. We argue that if there is cross-country heterogeneity in the distribution of within-country price differentials, there is no clear benchmark from which to gauge the effect of a border. In the absence of a structural model or a (natural) experiment, it is impossible to separate the "border" effect from the effect of trading with a country with a different distribution of prices. We show that the border effect identified by Engel and Rogers (1996) is entirely driven by the difference in the distribution of prices within the United States and Canada. (JEL F11, F14)