Identifying key elasticities in a CGE model: a Monte Carlo approach

Abstract

This article presents a simple Monte Carlo (MC) procedure to improve sensitivity analysis in Computable general equilibrium (CGE) modelling. MC experiments provide the modeller with a population of randomly drawn exogenous parameters and corresponding endogenous outcomes. Standard econometrics can then shed light on the relationship between them and help to identify key parameters, particularly key elasticities.

Keywords:

• Monte Carlo
• elasticities
• CGE

JEL Classification:

• C15
• C10

Notes

¹Another strand of the literature uses maximum entropy techniques to set elasticities in CGE models when full estimation is not possible. See, for example, Arndt et al. (2001).

²How the modellers make use of this information to improve the CGE model has to do with standard issues of CGE modelling and is beyond the scope of this article. Interested readers may refer to Portier et al. (2008) for more information on how the Cameroon model was ultimately improved, in part thanks to the early identification of key parameters.