Interfuel Substitution: A Meta-Analysis

This paper provides the first meta-analysis of the interfuel substitutability literature. The sample includes 47 studies of interfuel substitution in the industrial sector, in manufacturing industry or manufacturing sub-industries, or in the macro-economy of a variety of developed and developing countries. Meta-regressions are estimated for the six shadow elasticities of substitution between coal, oil, gas and electricity. I find that at the level of the industrial sector as a whole the oil-coal, oil-gas, oil-electricity and gas-electricity shadow elasticities of substitution are significantly greater than unity. The remaining elasticities - between coal and gas and coal and electricity - are insignificantly different from unity or zero due to their large standard errors. Substitution possibilities seem more constrained at the macro level and less constrained at lower levels of aggregation. Estimates of the elasticities of substitution from cross-section regressions are generally largest whereas fixed effects panel estimates are intermediate in magnitude, and time-series estimates are mostly much smaller. In order to obtain more precise and less biased estimates of long-run elasticities of substitution, there is a need for additional primary studies using large samples and/or the between estimator which econometric research suggests is likely to provide the best estimates of long-run elasticities.