Inequality in energy consumption: Statistical equilibrium or a question of accounting conventions?

Mitigating climate change requires information about the inequality in energy consumption. Recent contributions (Banerjee and Yakovenko, 2010; Lawrence et al., 2013; Yakovenko, 2010, 2013) have studied energy inequality through the lens of maximum entropy. They claim a weighted international distribution of total primary energy demand should approach a Boltzmann-Gibbs maximum entropy equilibrium distribution in the form of an exponential distribution, implying convergence to a Gini coefficient of 0.5 from above. The present paper challenges the validity of this claim and critically discusses the applicability of statistical equilibrium reasoning to economics from the viewpoint of social accounting. It is shown that the exponential distribution is only a robust candidate for a statistical equilibrium of energy inequality when employing one particular accounting convention for energy flows, the substitution method. But this method has become problematic with a higher renewable share in the international energy mix, and no other accounting method supports the claim of a convergence to a 0.5 Gini. We conclude that the findings based on maximum entropy reasoning are sensitive to accounting conventions and critically discuss the epistemological implications of this sensitivity for the use of maximum entropy approaches in social sciences.