Purchasing Power Parity: A Canada/U.S. Exploration

Abstract

The paper examines the possible explanations for deviations from purchasing power parity (PPP) between Canada and the United States in the 1980s and 1990s. The Balassa-Samuelson (B-S) model is used as the basis for the empirical exercise. In the B-S model, where PPP is assumed to hold for tradable goods, the real exchange rate (corrected by the exchange rate between two countries) reflects the bilateral differences in the relative productivity of the tradable and nontradable sectors. We investigate both the productivity effect and the underlying PPP assumption for tradable goods.

Using a Canada/U.S. micro data set on four benchmark years (1985, 1990, 1993 and 1996), we apply univariate and nonparametric analysis and obtain several results. First, purchasing power parity is rejected for both tradables and nontradables during the time period examined. Within tradables, however, PPP is rejected for differentiated but not for homogeneous commodities. Second, price differences for tradables are decreasing in the 1990s, a period of increasing free trade between the two countries. Third, we find little support for a simple version of the Balassa-Samuelson productivity explanation of the diverging average prices (adjusted by exchange rate) between the two countries in the 1990s. We do find a relationship between these variables but it involves a lag structure that requires further study.

Keywords: purchasing power parity, Canada and United States