The most commonly used theoretical framework describing why prices in some countries are higher than in others (i.e. explaining deviations from purchasing power parity) is the Balassa-Samuelson model. The Balassa-Samuelson model implies that stronger productivity growth should tend to cause a country's real exchange rate to appreciate. This paper develops measures of productivity and real exchange rate levels across industries and countries to allow the Balassa-Samuelson hypothesis to be tested.

We show that the model finds empirical support in 17 OECD economies: there is a link between real exchanges and sectoral productivity levels both across countries and over time. We then show theoretically and empirically that relaxing the model's assumptions about wage determination and the role of labour market differences across sectors and countries helps improve the performance of the model. However, there remains large unexplained deviations in real exchange rates across countries that the model cannot account for.