The real effective exchange rate is an aggregation of several bilateral real exchange rates with respect to other countries. The aggregation is usually done under the assumption of constant elasticity of substitution (CES) between products from different countries. We investigate the validity of this assumption by estimating manufacturing export equations for 56 countries over 26 years. We find that the hypothesis of CES is rejected and that the export equations that contain two real effective exchange rates (one in relation to OECD countries and one in relation to non-OECD countries) perform on average considerably better than the traditional ones.