Principal components at work: the empirical analysis of monetary policy with large datasets

The empirical analysis of monetary policy requires the construction of instruments for future expected inflation. Dynamic factor models have been applied rather successfully to inflation forecasting. In fact, two competing methods have recently been developed to estimate large-scale dynamic factor models based, respectively, on static and dynamic principal components. This paper combines the econometric literature on dynamic principal components and the empirical analysis of monetary policy. We assess the two competing methods for extracting factors on the basis of their success in instrumenting future expected inflation in the empirical analysis of monetary policy. We use two large data sets of macroeconomic variables for the USA and for the Euro area. Our results show that estimated factors do provide a useful parsimonious summary of the information used in designing monetary policy.