Network of Affiliates and Canada-U.S. Border Effect

Abstract

The objective of this thesis is to evaluate the influence that the pattern of affiliate linkages- establishments associated with companies as affiliates, subsidiaries and divisions- between U.S. states and Canadian provinces has on the effect of the border on trade between the two countries. The gravity model is used to estimate the border effect. Two hypotheses are tested – that the border effect is greater in the presence of affiliate linkages and that the strength of the border effect varies between industrial sectors. The results support the first hypothesis indicating that when all sectors combined, the presence of affiliates has a positive impact and it significantly strengthens the Canada-U.S. border effect. However, for the second hypothesis, nine sectors analyzed in this study present mixed results. For six sectors of agricultural, mineral, chemical, plastic, machinery and motor vehicle the border effect is not significantly different in the presence and absence of affiliates, while for the other three sectors of wood, textile, and base metals, the impact of affiliates has been significant. However, for wood and base metals sectors border effect has become stronger and for textile it has become weaker in the presence of affiliates. This research shows how the complicated relationship between trade and foreign direct investment determines firms operating in various sectors make decisions between export and establishing affiliates. The implication of the results can shed light on the study of the border effect and trade policy.