| Περίληψη: | This thesis presents and summarizes a versatile and lately developed statisticalmethod for evaluating and exploiting large datasets for time series analysis. Lately,Factor Models are widely used in macroeconomics to unveil interrelationshipsamong variables and exploit the information which is scattered across thousandsof economic variables that are available today. VAR models are widely used inmacroeconomic analysis and the estimated factors can be inserted in the VARsystem in order to achieve dimensionality reduction and avoid missing variablebias.The first part of this dissertation deals with an extensive literature review of factor models and their development through the variations that have been presentedin the international bibliography. As such, factor models have been developed inthe time domain as well as in the frequency domain, based on the assumptions onthe behaviour of the idiosyncratic component (exact and approximate) and on thebehaviour of the factors which can be static or dynamic. Moreover, an extensivereview of the empirical literature that uses Factor models on the estimation ofbusiness cycle behaviour, monetary transmission, economic condition indicatorsand forecasting topics.In the next section special emphasis is given on the estimation techniques onthe time and frequency domain along with the estimation algorithms provided. Inthe last section a static and dynamic factor model is employed on Greek inflationseries to decompose Greek inflation into common components. Static factoranalysis reveals goods prices that exhibit common variability in the Greek economyand suggests the need of estimating comprehensive inflation measures that willbe used for structural analysis and forecasting. Additionally, a dynamic factoranalysis is employed in order to decompose inflation into several price indexes thatrepresent key macroeconomic shocks in the economy. A careful examination ofthe correlation of inflation variability with the estimated indexes is provided anda thorough examination of the relative weight of anticipated monetary shocks onseveral measures of monetary policy.
|
|---|
