Inflation forecasting and business cycle analysis in a data-rich environment: an application of dynamic factor models to the euro area.

Summary We implement modeling strategies consistent with the problems raised by the study of the euro zone and the concerns of a central bank. We seek to verify the operational usefulness of exploiting a heterogeneous and large body of information to forecast inflation, and to extract the economic fluctuations responsible for inter-economic developments. To this end, we have made use of the latest developments in statistical theory regarding factor models, which allow us to apply these models to a large amount of non-stationary data. The main methodological aim of the thesis is to distinguish between common and country-specific effects, and between short- and medium-term phenomena and long-term phenomena. The predictive power of this synthesis of inflation information is also examined and compared in particular with that of a model based on economic theory.