Energy efficiency and energy rebound across heterogeneous technologies: A global analysis

Cross-country studies of energy efficiency and rebound effects typically impose a common production frontier and therefore ignore technological heterogeneity across countries. This paper relaxes that assumption by combining latent class stochastic frontier analysis with a rebound-adjusted stochastic energy-demand frontier. Using an unbalanced panel of 85 countries over 1996–2021, we first identify latent technology regimes and then estimate energy efficiency and rebound effects for the pooled sample and within each class. The results show that technological heterogeneity is empirically important. The homogeneous model masks differences in the determinants of energy demand and rebound effects and produces estimates that are more sensitive to aggregation. By contrast, the latent class model identifies two broad regimes in frontier behavior. Class 1, which includes many emerging industrializers and lower- and middle-income economies, shows lower average energy efficiency and slightly weaker rebound effects, whereas Class 2, which includes many advanced and upper-middle-income economies, exhibits higher efficiency but also somewhat stronger rebound effects. We also find evidence supporting a Rebound Kuznets Curve: energy rebound is lowest at low-income levels, rises significantly during the middle of the income distribution, and declines again at higher income levels. These findings suggest that the design of energy-efficiency policy should account for both technological heterogeneity and differences in development stage.