In practice it is hard to exhaustively validate a global CGE
model. Therefore, we focus our discussion of model validation on
the central dimensions of the problem at hand – namely the
model’s characterization of international trade and the robustness
of the results. One critical characterization of international trade is
the ‘‘Armington assumption’’, which assumes differentiated
products by country of origin. Villoria and Hertel (2011) have
empirically examined this assumption in their study of global land
use change due to regional shortfalls in the supply-demand
balance for coarse grains. They contrast this approach to the
commonly employed assumption of integrated world markets and
reject the latter in favor of the Armington specification. Those
authors also show that results under the two competing
assumptions could be quite different – thereby reinforcing the
strength of a model which takes into account the geography of
trade as does GTAP-BIO-W.