On the elasticity of substitution between clean and dirty energy: Reconciling empirical estimates and their implications for model calibration

Estimates of the elasticity of substitution between clean and dirty energy - and the values assumed in theoretical models - vary from less than 0.5 to as high as 10, creating substantial uncertainty about climate policy effectiveness and the feasibility of green growth. This paper develops an encompassing empirical specification that nests both low and high elasticity estimates and applies it to macroeconomic data from 13 OECD countries over 1980– 2020. Holding countries’ energy-related technology and infrastructure constant, the estimated elasticity of substitution remains well below 1, indicating limited short-run responsiveness of energy inputs to relative price changes. Allowing technology and infrastructure to respond to price changes, the estimated elasticity rises to between 2 and 3, a plausible long-run value. Without any controls, estimates reach as high as 6, greatly overstating true substitutability. We conclude that price-based policies alone, such as carbon taxes, are insufficient to trigger early-stage energy transitions. Investments in clean technology and infrastructure are essential, as they increase both the share of clean energy and the elasticity of substitution. Our findings can also guide the calibration of energy-augmented macroeconomic models. The elasticity values imposed in these models appear more often too high than too low.