A growing concern for using large scale applied general equilibrium models to analyze energy and environmental policies has been whether these models produce reliable projections. Based on the latest MIT Economic Projection and Policy Analysis model we developed, this study aims to tackle this question in several ways, including enriching the representation of consumer preferences to generate changes in consumption pattern consistent to those observed in different stages of economic development, comparing results of historical simulations against actual data, and conducting sensitivity analyses of future projections to key parameters under various policy scenarios. We find that: 1) as the economies grow, the empirically observed income elasticities of demand are better represented by our setting than by a pure Stone–Geary approach, 2) historical simulations in general perform better in developed regions than in developing regions, and 3) simulation results are more sensitive to GDP growth than energy and non-energy substitution elasticities and autonomous energy efficiency improvement.

© 2016 Elsevier B.V.