Bayesian Modeling and Forecasting of Matrix-Valued Time Series

 

• \(\{ Y_t \in R^{n \times k} : t = 1, \ldots, T \}\): Variables x Regions
• New Challenge in Empirical Macroeconomics:
  • Many macroeconomic variables show extreme movements after the COVID-19 pandemic
  • Recent studies suggest addressing COVID-19 outliers in Bayesian VARs
  • This project extends the issue into matrix-valued time series
• Central question: Can we forecast global economy during COVID-19 period?