Project information
Forecasting with DSGE models: Optimal choice of model structure (DSGE_FORECAST)

This project focuses on forecasting performance of DSGE-DFM models, which can be seen as a combination of a regular DSGE model and a dynamic factor model. Unlike the regular approach, this method allows DSGE models to be estimated using a large data vector of time series. The DSGE-DFM framework can exploit the information from economic indicators that are not directly and unambiguously linked to a specific concept of the underlying DSGE model. Estimating a DSGE model in a data-rich environment seems to be useful for a proper identification and estimation of the state of the economy. A proper estimation of the state of the economy in turn appears to improve forecasts of important macroeconomic variables. Although the existing literature suggests that DSGE-DFM models might be very useful for forecasting macroeconomic aggregates, this research area remains quite narrow as only two studies have investigated the forecasting performance of DSGE-DFM approach. This research aims to thoroughly and systematically investigate the forecasting potential of DSGE-DFM setup, and it should answer the question, whether it is possible to achieve more accurate predictions of DSGE models when estimated in a data-rich environment. In order to obtain generally relevant results, the forecasting performance of DSGE-DFM approach will be comprehensively evaluated using a battery of DSGE models estimated for various economies.