Topic: Contributed on computational methods for complex data Title: Estimation of a DSGE model with heterogeneous agents using an indirect inference estimator Authors:  Artem Duplinskiy - VU University Amsterdam (Netherlands) [presenting]
Alexey Gorn - Bocconi University (Italy)
Abstract: Policy makers use macro models with individuals being the same in terms of their level of income and skill. Models with heterogeneous agents allow individuals to have different starting endowments (wealth) as well as different income (not everyone is employed). These models are more difficult so researchers use simulation based solution methods to solve them and calibrate the parameters. Without estimation, it is hard to tell which of the new models represents the data better. Calibration takes away the uncertainty regarding parameter values necessary for meaningful forecasts. We asked whether an indirect inference estimator could be used to estimate a simple dynamic stochastic general equilibrium model with heterogeneous agents. This estimator naturally fits to the task since it requires only that one can simulate data from the economic model for different values of its parameters. We use macro variables together with the dynamics of the income distribution to estimate the model. A Monte Carlo study reveals the importance of the income distribution data to identify the parameters. The data application highlights that the model picks up precautionary savings effect observed in the data, suggesting that II estimator can be used to fit new models to the data and advance macroeconomic research.