Quasi-Monte Carlo methods in stochastic simulations: An application to policy simulations using a disequilibrium model of the West German economy 1960-1994
Klaus Göggelmann (),
Peter Winker (),
Martin Schellhorn () and
Wolfgang Franz ()
Additional contact information Klaus Göggelmann: Credit Suisse Asset Management, Uetlibergstrasse 231, CH-8070 Zürich, Switzerland
Martin Schellhorn: Volkswirtschaftliches Institut, Universität Bern, Gesellschaftsstr. 49, CH-3012 Bern, Switzerland
Wolfgang Franz: Zentrum für Europäische Wirtschaftsforschung, P.O. Box 10 34 43, D-68034 Mannheim, Germany
Abstract:
Different stochastic simulation methods are used in order to check the robustness of the outcome of policy simulations. The application of a macroeconometric disequilibrium model of the West German economy to a fiscal policy simulation is taken as an example. Due to nonlinearities arising from regime specific reactions inside the model, confidence intervals for the simulation results have to be obtained by means of stochastic simulations. The robustness of the results is assessed using different methodologies. In particular, different methods for the generation of uniform error terms and their conversion to normal variates are applied. These methods include standard approaches as well as quasi-Monte Carlo methods.