Estimation of unknown parameters in system dynamics models using the Method of Simulated Moments

In principle the Method of Simulated Moments (MSM) combines simulation-based methods (e.g. Monte Carlo methods) with non-parametric statistical estimations techniques such as General Method of Moments (GMM). The MSM is useful when there are empirical data related to the behavior of different entities. Different statistical moments (e.g. mean, variance, correlation, etc.) of empirical data can be matched against the moments of model-generated data in order to estimate some structural parameters of the model. In this paper, we introduce the MSM as a non-parametric method of estimating the parameters of dynamic models. The major value of the MSM for estimating dynamic models is in its flexibility to be used with any type of data, including cross-sectional data and time series data.