ELITE NETZWERK BAYERN

English  Sprachen Icon  |  Gebärdensprache  |  Leichte Sprache  |  Kontakt


Forschungsarbeit

A Comparison of traditional and stochastic forecasting in economics 

Von Luis-Andre Zitzmann (14.09.2010)

The development of stochastic models for making reliable and accurate predictions will probably turn out to be the key technology for society and industry of the 21st century. So far, there is no ISO standard for producing prediction procedures safeguarding the usefulness and reproducibility of predictions. As a consequence many different procedures are used yielding divers predictions for the same situation leading almost necessarily to confusion and to wrong decisions. The described circumstances can be observed for instance in the case of predictions of the gross domestic product (GDP), which play an important role in Political Economy. Many institutions, organizations and governmental offices produce and circulate continuously different numbers as predictions that practically never occur.

These dissertation aims at developing specifications for prediction procedures that could be used for a prediction standard similar to standards that exist for almost every product and process on the market. If such a standard would exist, a prediction not meeting the specifications could be classified as nonconforming and discarded from any decision-making process.

As an illustration substantive prediction procedures of economic practice like prediction procedures for the GDP or Credit-Ratings will be examined whether they are conforming to the anticipated specifications or not. Figure 1 shows a prediction of a certain future event X with a given reliability β (denoted A(β)X (D) ). This prediction is based on a stochastic model for a given process of reality. The stochastic model is named Bernoulli-Space and denoted BX;D = (D;X;P).

Fig. 1: Prediction. Source: XiaoMin Zhai: Graphical Representation of Stochastic Models and Procedures, The 7th Workshop of Stochastic Models and Their Application, Würzburg 2005.[Bildunterschrift / Subline]: Figure 1: Prediction. Source: XiaoMin Zhai: Graphical Representation of Stochastic Models and Procedures, The 7th Workshop of Stochastic Models and Their Application, Würzburg 2005.

Luis Zitzmann
* 1981

Stationen
  • since April 2009
  • PhD student and research assistant of Prof. Dr. Peter Bofnger, Chair for Monetary Policy and International Economics, University of Würzburg
  • April 2007 - March 2009
  • PhD student and research assistant of Prof. Dr. Elart von Collani, Chair of Econometrics, University of Würzburg
  • Dec. 2006
  • Diploma in Economics (Diplom-Volkswirt)
  • Aug. 2004 - March 2007
  • Student Assistant of Prof. Dr. Elart von Collani at the University of Würzburg
  • Oct. 2004 - Dec. 2006
  • Studies in Economics, University of Würzburg
  • Oct. 2001 - Sep. 2004
  • Studies in Business Administration, University of Würzburg

Stipendien
  • since May 2008
  • Research Scholarship in the Elite Network of Bavaria