Engel & Rogers (1996) find that crossing the US-Canada border can considerably raise relative price volatility and that exchange rate fluctuations explain about one-third of the volatility increase. Using a decomposition method, this study re-evaluates the border effect. It is shown that cross-country heterogeneity in price volatility can induce a bias in measuring the border effect unless proper adjustment is made to correct it. We further examine the implication of symmetric sampling for the border effect estimation under the decomposition approach. Two conditions governing the strength of the border effect are identified. In particular, the more dissimilar the price shocks are across countries, the greater the border effect will be. Decomposition estimates also suggest that exchange rate fluctuations actually account for a large majority of the border effect.