Logo
Munich Personal RePEc Archive

Die Umsetzung der Annual-Overlap-Methode in ökonometrischen Modellen – eine Analyse der programmtechnischen Möglichkeiten von E-Views

Quaas, Georg (2009): Die Umsetzung der Annual-Overlap-Methode in ökonometrischen Modellen – eine Analyse der programmtechnischen Möglichkeiten von E-Views.

[thumbnail of MPRA_paper_19435.pdf]
Preview
PDF
MPRA_paper_19435.pdf

Download (265kB) | Preview

Abstract

The study addresses (i) the problem of non-additivity of some time series expressed in real terms that is one of the features of Germany’s new system of national accounts since 2005, and (ii) methods of the computation of quarterly data in econometric forecasting models under this condition. The problem of non-additivity is demonstrated with the architecture of the RWI-business cycle models how they were constructed by 2005. One of the theses of this study is that the unchained volumes measured by constant prices of the previous year (which can be ordered at Germany’s Office for National Statistics) are solving the problem (i) not only partly (Nierhaus 2004), but totally when embedded in the environment of an econometric model in an appropriate way. This study shows (ii) how the problem of computing quarterly data in econometric forecasting models on basis of E-Views appears. Six trials to solve this problem are discussed with the conclusion that only three of them are successful. The solutions of the problems consist in (i) a restriction of the explaining variables in the regressions to unchained volumes, or, alternatively, (ii) in the implementation of a 2-step-procedure that is operating with approximations and iterations (in this connection the convergence of the model-solutions is shortly discussed), and, last but not least, (iii) the embedding of the econometric model into a framework of forward- and backward dated variables and implementing a special algorithm that does not hark back to iterative procedures, but procures the exact values of all dimensions of an macroeconomic aggregate directly.

Atom RSS 1.0 RSS 2.0

Contact us: mpra@ub.uni-muenchen.de

This repository has been built using EPrints software.

MPRA is a RePEc service hosted by Logo of the University Library LMU Munich.