Linking financial and economic agent-based models: An econometric approach

The project addresses the disconnection issue of the current financial and economic agent-based research. A complex agent-based economic model with an integrated financial sector will be suggested and a focus will be devoted to its econometric estimation. This interconnection will allow an overall study of the economic system under realistic assumptions about the behavior of economic agents. The estimated model can be utilized as a new financial regulatory policy tool or to support economic decisions of policymakers. As merging the two frameworks increases the model complexity, methodological advancements designed and solved in this project will significantly support estimation procedures and sound empirical results. The project aims at a novel introduction of machine learning methods to agent-based econometrics. For instance, the setup of the simulated method of moments will be extended with an optimized selection of the moment set. To further deepen the knowledge of the empirical validation of agent-based models, a comparison study of feasible estimation methods will be conducted.