Applied general and partial equilibrium models are widely used tools for ex ante analysis of trade policy changes. However, simulation results seem to exhibit significant variation across publications, and the often criticised ‘black box’ character of applied trade models makes meaningful comparisons of simulation results very difficult. As a potential remedy, this paper presents a meta-analysis of simulation-based Doha round publications. The meta-regression explains simulated welfare changes as a function of model characteristics, base-data and policy experiments. Regression results show that a major share of the variation within the dependent variable is explained by the covariates, and estimated coefficients show plausible signs and magnitudes. However, results also reveal that many model-based studies lack systematic documentation of their experimental settings.
meta-analysis partial equilibrium general equilibrium trade liberalisation WTO Doha