Estimation of Growth Elasticities in the General Equilibrium Framework under Model Uncertainty

In recent years the concept of uncertainty analysis is being employed more frequently in order to test the validity and robustness of results of deterministic Computable General Equilibrium (CGE) models. This paper suggests the formal method to evaluate the Total Factor Productivity (TFP) growth elasticities under the uncertainty of all exogenous parameters of the CGE model developed by the International Food Policy Research Institute (IFPRI). First, we use the historical data and our assumptions to parametrize the uncertainty of the model parameters and present it as a covariance matrix. Then we employ the Latin Hypercube Sampling method to draw a sample of 1000 simulation scenarios from the multivariate normal distribution characterized by this covariance matrix and the vector of the baseline parameters. Then for each generated scenario we run the IFPRI CGE model and get the endogenous response variables. Finally, we employ a metamodeling concept to formalize and evaluate the relationship between the policy-induced sector-specific TFP growth and variables of interest given the uncertainty of other model parameters. The approach is applied empirically to a highly aggregated CGE model of the Senegalese economy.