The alternative version of J-curve hypothesis testing: Evidence between the USA and Canada

Abstract

This study aims to eliminate an embedded data bias in testing the J-curve hypothesis in relevant literature. In all previous trade models about the J-curve, the bilateral trade balance (BTB) ratio, as the dependent variable, is generated and used based on total exports data, i.e. as total exports over total import. However, total exports are the sum of domestic exports and re-exports, and some countries also re-export to their partners. Furthermore, the dynamics and impacts of changing exchange rates on export volumes of domestically produced goods may differ from those on re-exported goods. Therefore, in this study, in testing the asymmetric J-curve hypothesis, we, for the first time, re-define two new proposed forms of BTBs, namely, domestic-export-based J-curve hypothesis BTB and re-export-based J-curve hypothesis BTB, based on domestic exports and re-exports, respectively, for USA–Canada trade. The main finding shows that the numbers of industries that support the asymmetric J-curve concerning these two forms of BTBs are entirely different. While the J-curve is supported by domestic-export-based J-curve hypothesis BTB for 27 industries, it is supported by re-export-based J-curve hypothesis BTB for 38. This support is 22 for the total-export-based J-curve hypothesis BTB.