GDP Forecasting Performance, Monetary Policy Tools and Natural Resources: Mixed Frequency Bayesian VAR models for the Eurozone from 1999 to 2024

This study develops a Mixed-Frequency Bayesian Vector Autoregressive (MF-BVAR) model to predict GDP growth in the Eurozone from 1999 to 2024. The model integrates monetary policy tools, economic variables, and natural resources data to enhance prediction accuracy. The model captures the dynamic interaction between these factors by incorporating mixed-frequency data and Bayesian inference. In addition, employing a Bayesian framework, we consider prior distributions and update them with observed data, ensuring that our model remains robust and responsive to new information. Introducing natural resources data reduces forecasting errors measured by Root Mean Squared Errors (RMSE) and Theil inequality coefficient metrics. The results demonstrate the model's robustness in forecasting GDP growth, providing valuable insights for policymakers and economists. Moreover, it is also important to note that our findings highlight the significant and direct influence of natural resources on predicting GDP growth in the Eurozone. This conclusion has been overlooked in the current literature.