Econometric Methods for Risk Assessment and Forecasting: SoFiE European Summer School (Brussels)

This intensive program offers participants the opportunity to engage with leading academics, develop practical skills through hands-on applications, and expand their network within the econometrics community. Whether you’re an academic, researcher, or industry professional, this summer school provides cutting-edge tools for advancing your expertise in time series forecasting and econometric modeling. 

The audience includes graduate students, academics, practitioners, and policymakers. 

Part 1: Time Series Density Forecasting and Combination
Francesco Ravazzolo
Density Forecasting
Reduced Form VARs: Introduction to Vector Autoregressions (VARs) and their application in modeling multivariate time series data.
Density Forecasting with VARs: Exploration of both Bayesian and frequentist approaches to generate density forecasts using VAR models.
Evaluation of Forecasts: Techniques for assessing the accuracy and reliability of density forecasts.
Applications: Participants will engage in practical applications, with a focus on sectors such as energy, to solidify their understanding.
Density Forecast Combinations
Bayesian Model Averaging: Study of methods to combine forecasts from multiple models using Bayesian principles.
Extension to Time-Varying Combination Weights and Learning: Investigation into dynamic weighting schemes that adapt over time for improved forecast combinations.
Combinations of Large Data Sets: Strategies for handling and integrating extensive datasets in forecast combinations.
Applications: Practical application exercises, potentially within the energy sector, to apply theoretical concepts.

Part 2: Structural Vector Autoregressions
Francesca Loria
Constant-parameter Structural VARs: Analysis of VAR models that incorporate structural information to identify causal relationships.
Regime-switching parameter VARs: Examination of VAR models that account for regime changes in time series data.
Proxy Structural VARs: VAR Models that use external instruments to identify and estimate the effects of structural shocks in economic time series data.
Applications:
Policy impacts: Identifying causal effects of monetary policy shocks on the macroeconomy using SVARs and proxy SVARs.
Growth at risk: Assessment of the potential downside risks to economic growth using  MS-VAR techniques.

Part 3: Factor Models and Other Extensions
Dimitris Korobilis
Quantile Factor Augmented Vector Autoregression (QFAVAR): Integration of factor models with quantile regression techniques to enhance forecasting accuracy.
Tail Index Regression: Methods for modeling and forecasting extreme values in financial and economic data.
Bayesian Density Regression: Application of Bayesian methods to estimate the entire distribution of a dependent variable, rather than just its mean.
Applications:
QFAVAR Structural Inference and Forecasting: Utilization of QFAVAR models for structural analysis and predictive purposes.
A Tail Risk Index: Development of an index to monitor tail risks in stock markets.