CGE Analysis of Mega Events: A Timely Issue

CGE models are more and more frequently used to evaluate the real impact of mega events. However, the implementation of CGE models reveals a number of methodological features that may provide misguiding policy recommendations. A first series of issues relates to how time is represented in CGE models, where most applications make simplifying assumptions that can profoundly alter the realism of results. The proposed contribution aims at examining these various assumptions and evaluate their impact based on a reference CGE model. We review questions such as : 1- dynamicity. Many models still are based on a single period. In some applications, the multiperiodicity of the model is accounted for through a set of independent simulations, losing the interlinkages between the different periods. It seems however useful to produce quantitative results which show how much the model results are impacted by dynamic linkage, for this particular class of applications. 2- time concentration of visitors impacts. Time concentration is a striking aspect of events, and one can wonder whether the general annual periodicity used by typical CGE are suitable to represent phenomenon with such a time pattern. Using a model that neglects the time concentration of the phenomenon certainly risks provoking errors of large magnitude. 3- how infrastructures are often accelerated rather than generated by the event and how infrastructure legacy impacts span on an even longer period. Interestingly, a more accurate consideration for time issues makes other questions apparent. This relates to issues linked with the productivity of infrastructure, to the likely overcosts of these infrastructures compared with normal infrastructure provision, and others. We provide results on these issues and argument for the need to properly account for a whole set of more realistic assumptions in the models.