Validating GTAP-E using the Shale Revolution

The GTAP-E computable general equilibrium model is widely used to forecast the economic impacts of energy and environmental policies (McDougall and Golub 2007; Burniaux and Truong 2002). It has been used to look at topics as varied as EU biofuel policy, to the effects of climate change on tourism (Banse et al. 2008; Berrittella et al. 2006). However, the GTAP-E literature has mostly focused on applying the model to predict the impacts of policies, with much less emphasis on proving the predictive power of the model in the first place (Hong et al. 2014; Dixon and Jorgenson 2013). To validate the effectiveness of the GTAP-E model, I use GTAP-E to forecast the impact of the Shale Revolution, which drastically reduced the price of natural gas in the US from 2005 to 2016. I find that given the observed price change, the GTAP-E model predicts changes in production and trade that are orders of magnitude larger than what actually occurred. But GTAP-E’s predictions become much closer to reality if smaller Armington elasticities of substitution are used instead.