A Gentle Introduction to Factor Models
Factor models became the dominant methodology for forecasting economic time series with high dimensional data. Several empirical studies suggest that a factor structure is present in large panels of macroeconomic time series. In my opinion there is a big deal of confusion about the different types of factor models that have been proposed and how they relate to each other: Dynamic vs. static, infinite vs finite dimensional factor spaces, approximate vs exact, etc.
The goal of this talk is to give a gentle introduction to factor models from a geometric point of view and to answer the following questions:
- Which types of factor models are out there?
- How do they relate to each other?
- Why are factormodels so successful in forecasting?
- How are they used for forecasting macroeconomic variables (GDP-growth, inflation, unemployment, etc.)?
Please register at event(at)ihs.ac.at.