Representative versus real households in the macro-economic modeling of inequality

To analyze issues of income distribution, most disaggregated macroeconomic models of the Computable General Equilibrium (CGE) type specify a few representative household groups (RHG) differentiated by their endowments of factors of production. To capture “within-group” inequality, it is often assumed, in addition, that each RHG represents an aggregation of households in which the distribution of relative income within each group follows an exogenously fixed statistical law. Analysis of changes in economic inequality in these models focuses on changes in inequality between RHGs. Empirically, however, analysis of household surveys indicates that changes in overall inequality are usually due at least as much to changes in within-group inequality as to changes in the between-group component. One way to overcome this weakness in the RHG specification is to use real households, as they are observed in standard household surveys, in CGE models designed to analyze distributional issues. In this integrated approach, the full heterogeneity of households, reflecting differences in factor endowments, labor supply, and consumption behavior, can be taken into account. With such a model, one could explore how household heterogeneity combines with market equilibrium mechanisms to produce more or less inequality in economic welfare as a consequence of shocks or policy changes. An integrated microsimulation-CGE model must be quite large and raises many issues of model specification and data reconciliation. This paper presents an alternative, top-down method for integrating micro-economic data on real households into modelling. It relies on a set of assumptions that yield a degree of separability between the macro, or CGE, part of the model and the micro-econometric modelling of income generation at the household level. This method is used to analyze the impact of a change in the foreign trade balance, and the resulting change in the equilibrium real exchange rate, in Indonesia (before the Asian financial crisis). A comparison with the standard RHG approach is provided. _________________________________ Ce papier présente une méthodologie qui permet de s’affranchir de l’hypothèse de l’agent représentatif couramment utilisée dans les modèles d’Equilibre Général Calculable (EGC). Il s’agit de remplacer les traditionnels agrégats correspondant à d’hypothétiques « ménages représentatifs » par un échantillon de ménages réels, tirés d’une enquête budget-consommation. Cette approche permet de prendre en compte toute l’hétérogénéité des ménages étudiés, non seulement économique mais également démographique, sociologique, etc. Ce type d’approche présente néanmoins l’inconvénient de la taille puisqu’il s’agit de manipuler des bases de données de plusieurs milliers d’individus. Ce papier présente un modèle appliqué à l’Indonésie pour étudier l’impact social d’un choc d’épargne extérieure et l’évolution du taux de change reel d’équilibre qui résulte de ce choc (avant la crise financière asiatique). Les résultats obtenus sont comparés à ceux obtenus avec un modèle plus “standard” à ménages représentatifs.