Exploration of the Parameter Space in Macroeconomic Agent-Based Models

Agent-Based Models (ABM) are computational scenario-generators, which can be used to predict the possible future outcomes of complex system they represent. To better understand robustness these predictions, it is necessary full scope phenomena model generate. Most often, due high-dimensional parameter spaces, this a computationally expensive task. Inspired by ideas coming from systems biology, we show that for multiple macroeconomic models, including an agent-based and several Dynamic Stochastic General Equilibrium (DSGE) there only few stiff combinations have strong effects, while other sloppy directions irrelevant. This suggest algorithm efficiently explores space parameters primarily moving along directions. We apply our medium-sized model, recovers all dynamics unemployment rate. The application method Agent-based may lead more thorough robust understanding their features, provide enhanced sensitivity analyses. Several promising paths research discussed.