Abstract (eng)
In recent years leading economic indicators have become more and more important. Nowadays, the entire media landscape follows the development of these measures and their information content is often used by economists whenever future predictions are needed. Despite the fact that leading economic indicators are heavily utilized, their real predictive power is almost uninvestigated. This provides the motivation for analysing whether leading economic indicators are really useful tools for forecasting the upcoming economic development.
The present diploma thesis covers and analyses some of the most important leading indicators in Europe (Ifo Index, Consumer Confidence and German Stock Index). Various statistical methods are used in order to test if there is a link between these measures and the upcoming production within the economy.
In the first part of the analysis the well-known time series models are implemented. In the process the concept of Granger Causality is applied for investigating the existence of a causal relation between leading economic indicators and production in the time domain.
Furthermore, in the second part spectral methods are utilized to get a more insightful view on the topic. These methods allow the formal decomposition of the data in their basic components, which enables me to investigate the relation of leading indicators and production over different frequencies. In this way it is possible to determine the exact frequency bands in which leading indicators contain useful information for the explanation of future economic developments.