Dynamic Model Averaging in Large Model Spaces Using Dynamic Occam's Window.

Area of Science:

• Econometrics
• Macroeconomic Modeling
• Time Series Analysis

Background:

• Bayesian model averaging (BMA) addresses uncertainty in model structure.
• Dynamic Model Averaging (DMA) extends BMA for sequential data and changing models.
• Standard DMA is computationally infeasible with numerous variables in macroeconomics.

Purpose of the Study:

• To develop a computationally efficient method for Dynamic Model Averaging (DMA) in large-scale macroeconomic models.
• To address the challenge of model space explosion in time-varying parameter models.
• To introduce a dynamic Occam's window approach for model selection.

Main Methods:

• Proposes a novel method to perform DMA on a dynamic subset of models, avoiding exhaustive search.
• Implements a dynamic selection of models at each time point.
• Applies the method to nowcasting Gross Domestic Product (GDP) in the Euro area.

Main Results:

• The proposed method effectively manages large model spaces in DMA.
• Dynamic Occam's window selection proves efficient.
• The method achieves competitive forecasting performance for Euro area GDP nowcasting.

Conclusions:

• The new DMA approach offers a practical solution for complex macroeconomic modeling.
• Dynamic model subset selection enhances the applicability of DMA.
• The method demonstrates strong performance in real-world economic forecasting applications.