We could not care less about Armington elasticities – but should we? : a meta-sensitivity analysis of the influence of Armington elasticity misspecification on simulation results

This paper investigates the robustness of CGE models with respect to the elasticities of substitution in demand between domestically produced goods and foreign goods – the so-called Armington elasticities. The Armington-type modeling of trade is still one of the most extensively used specifications in CGE modeling. For a long time the choice of the respective elasticities of substitution has not been given much attention. We resimulate 50 published CGE policy simulations each a 1000 times and randomize the elasticities. The results of this experiment clearly indicate that a change in the elasticities has noteworthy quantitative and qualitative effects on the results in more than half of the models. We thus conclude that the choice of the elasticities should get more attention and robustness of models with respect to elasticities should be tested by modellers.