AbstractThe present paper discusses methods originally proposed by Adams and Miovic in 1968 (then refined and used later by many other authors) for calculating the output elasticity of useful energy consumption (ϵ). We first show that this methodological approach is quite dubious. Better alternative methods of estimation are then proposed. We also stress that, since the marginal rate of interfuel substitution depends on the GDP functional form, the simulataneous use of several functional forms of GDP in energy studies leads obviously to misleading interpretations. Using thermal efficiency coefficients and OECD countries figures for the 1959–73 period, we finally found that ϵ did steadily fall from high values to values which are still higher than