The ZEW Combined Microsimulation-CGE Model : Innovative Tool for Applied Policy Analysis

Ex ante evaluation of policy reform proposals in general relies either on microsimulation models or computable general equilibrium (CGE) models. This paper describes how the Centre of European Economic Research (ZEW) has merged two established models - the so-called STSM and PACE-L models – in order to obtain a linked microsimulationcomputable general equilibrium model. This state of the art approach for applied policy analysis combines the advantages of the two model types. Concerning microsimulation models, their main advantage is the foundation on individual household level data. This allows a detailed analysis of who gains and who loses from a reform proposal. The main disadvantage of microsimulation models is that they neglect general equilibrium effects. In particular, these models assume constant wages and interest rates. i.e. that the labour market equilibrium is the same before and after the analyzed policy change. CGE models, on the other hand, incorporate these effects. Their disadvantage ist that they are based on aggregated household types rather than individual households. This brings a loss of information as regards heterogeneity of household which impedes in particular thorough distributional analysis. Merging the microsimulation and the CGE-approach to policy evaluation is a means to combine the advantages of both model types. The approach delivers detailed results as regards distributional outcomes whilst taking into account the possible employment and wage impacts of policy reforms. This paper provides an introduction to the fully integrated CGE-microsimulation model STSM/PACE-L. This model, developed and maintained at the ZEW, is the most advanced model of its kind in Germany. It integrates all individual households that are included the microsimulation model with flexible labour supply in the CGE model, whereby it achieves a much more precise representation of the household sector on the macroeconomic level than typical models with at best few representative households. The advantages of fully integrating simulation at the micro and macro level have shown in several cases of practical ex-ante evaluation studies