Measuring the possibilities of interfuel substitution

What are the costs of making consumption of production activities use less-polluting fuels? The author reviews how the fuel mix used by different industries has changed over time and examines 2 techniques for estimating the responsiveness of fuel demand to fuel prices: econometric models and the engineering approach. With econometric models, the elasticity of substitution between energy and other inputs determines the costs of making activities less energy-intensive, while the elasticity of substitution between sources of energy (interfuel substitutability) determines the marginal costs of replacing one energy source with another. The engineering approach uses more detailed technical information and can draw a more complete picture, but with less ability to inform about activities with a vast number of different economic agents. Among the author's main conclusions: There are surprisingly large variations in energy and fuel use over time and between countries. Industrial output increased 62 percent in OECD countries between 1971 and 1988, for example, while energy use stayed unchanged. Also, shares of energy sources for industry and electricity vary greatly with local availablity, indicating that these sectors have some flexibility in choice of energy source. A judgment on whether this variability indicates that an economy responds cheaply if energy prices are changed selectively depends on how one reads the more detailed studies in the econometric and engineering literature. Lack of data is the biggest problem in estimating fuel and energy substitutability in non-OECD countries. Engineering studies of fuel switching in industry are rarely available. They exist, however, for the power industry and could be used to estimate the costs of alternative fuel-mixes for particular greenfield sites. The technique could not be used for assessment of economywide policies. Econometric studies are useful inasmuch as they take a sector- or economywide perspective. Econometric techniques are challenging, but often represent the state of the art in providing reliable estimates for elasticies of substitution - particularly when data are scarce and the level of aggregation is high. The issue of whether econometrically estimated structural parameters can be transferred across borders has not been thoroughly investigated.