The J-curve revisited: An empirical analysis for Canada

This paper examines exchange rate and trade data to determine whether J-curve effects have occurred for Canada in recent history. With a unique set of data on import and export volumes, the response of the trade balance and its components to exchange rate changes are determined empirically over 1981:1-2005:12 through impulse response functions. The responses have taken into account dominant long-run feedback effects, where the long-run parameters have been estimated by the Johansen cointegration technique. The results show the J-curve phenomenon does not exist for Canada. Robustness checks show the results do not change when looking at the responses over the pre-NAFTA and NAFTA periods. © International Atlantic Economic Society 2008.