AbstractThis study examines the ability of exchange rate changes to aid in the explanation and prediction of the geographic segment earnings of U.S.‐based multinational corporations. In order to evaluate the potential usefulness of geographic segment earnings disclosures, it is important to understand the effect that currency changes should be expected to have on these earnings. Two types of exchange rate effects are examined. These include the mechanical translation effects of an exchange rate change, as well as the operating effects. A sample of geographic segment earnings disclosures is developed for two geographic locations (Canada and Europe) and six industries in which adequate country‐industry specific subsamples can be identified. Because the factors that impact exchange rate exposure tend to depend on country, industry, or both, provision is made for country, industry, and country‐industry specific sensitivity to currency changes. Regressions and nonparametric analyses are performed to determine whether exchange rate changes can help explain the geographic segment earnings within these samples. Results indicate that the accounting‐based translation effect that is typically modeled as the impact of exchange rate changes on earnings does not adequately capture the effect of exchange rate changes on segment earnings. Operating effects of past and current exchange rate changes do help explain earnings for geographic segments located in Europe and in four industries. Prediction models which are country or country‐industry specific are able to outperform a random‐walk prediction of geographic segment earnings in some ci