Estimation and Forecasting of Core Inflation in Poland using the Generalized Dynamic Factor Model
This research project constitutes the first attempt of applying the Generalized Dynamic Factor Model (GDFM) for the analysis of inflation in Poland. The family of dynamic factor models allows for analyzing sizeable sets of time series and their decomposition into a common (co-linear) component and an idiosyncratic component characteristic of individual processes. Core inflation is then estimated on the basis of the common component of individual price changes of elements of the CPI basket. Additionally these components are subjected to a spectral decomposition in order to remove variability associated with high frequencies - a process aimed at removing short-term fluctuation from the resultant core inflation series. Final core inflation index is calculated as a weighted average of common factors of individual CPI components which have previously been deprived of high-frequency variability. Research will also produce forecasts of core inflation according to a two-step procedure that will allow to generate several-year-ahead forecasts. The final part of research will compare alternative core inflation indices (including the one estimated by GDFM) and evaluate them using standard criteria used in literature, such as stability, unbiasedness or exogeneity vis-a-vis the CPI.
Polish Ministry of Learning and Advanced Studies
(Ministerstwo Nauki i Szkolnictwa Wyższego)