Exploring the use of dynamic linear panel data models for evaluating energy/economy/environment models — an application for the transportation sector

Abstract

This paper uses the RoSE transportation sector scenarios of the GCAM and REMIND energy-economy-models for the U.S. region to derive and compare these models’ intrinsic elasticities with those resulting from historical trends, estimates from the literature, and across each other. To estimate the model-intrinsic elasticities, we explore the use of dynamic linear panel data models. On the basis of 26 scenarios (panels) between 2010 and 2050, our analysis suggests that nearly all model-intrinsic elasticities with respect to final energy use are roughly comparable to each other, to those observed historically, and to those from other studies. The key difference is these models’ comparatively low intrinsic income elasticity of final energy use. This and other minor differences are interpreted through key assumptions underlying both energy-economy-models.