Combining micro- and macrosimulation models in the case of Luxembourg

Since the pioneering work of, a.o. François Bourguignon at the end of the 1980s, the field of linking micro-simulation and macro-economic models has generated a growing number of publications. Empirical applications foster notably in the areas of tax and trade policies, climate change, poverty reduction and structural reform. While CGE models are the overwhelmingly preferred choice for representing the macroeconomy, estimated macro‑econometric models are much less frequently used for that purpose. For a small country, Luxembourg exhibits a relative richness of large-scale models: recent work counts 3 microsimulation and at least 4 macro-economic models. This work links STATEC’s macroeconometric workhorse “Modux” with IGSS’s microsimulation tool “SPAFIL”. Both have been widely used in forecasting and/or policy analysis but never so far in combined simulations. The purpose of this work is hence to link both models and to generate a number of common shocks: reduction in household taxes, decrease in financial sector activity, etc. The shocks are simulated iteratively, with a set of pertinent variables exported sequentially in both directions. Convergence is assured but not full. Main results consist of a modified macroeconomic trajectory integrated into SPAFIL and the inclusion of distributional effects concerning households in Modux. The latter also has the potential to – sometimes substantially – change the macroeconomic outcome and, prominently so, the Keynesian multiplier.