Empirical Estimation of the Elasticity of Substitution: A Review

This review of time series and cross-sectional studies indicates that there is little agreement as to the "true" value of the elasticity of substitution. There appears to be no clear-cut explanation for the diversity of the estimates. Not only do the various studies attribute widely different absolute levels to the elasticity of substitution in the industries considered, but there is also no apparent consistency between the ordinal rankings of the industries within each study. There is however a general pattern of differences between the cross-sectional and time series estimates -- namely, the former are usually larger (averaging about 1) than the latter (averaging perhaps about 0.5). Please note: The PDF download of this fairly old paper is an optical scan.