The Balassa-Samuelson hypothesis suggests that countries with a weak relative productivity performance should, over time, see a low or falling real exchange rate. This note uses detailed sectoral data to test the hypothesis over the period 1978-2006 and also fails to find any evidence of the expected effect.