This book was accepted in 1999 as doctoral thesis (Dr. oec. publ. ) by the faculty of economics of the Ludwig-Maximilians-Universitat in Munich. It won the dissertation price of the Alumni-Club of the economics faculty in the academic year 2000. Financial assistance by the German Research Asso ciation (DFG) for printing costs is gratefully acknowledged. I have to thank the people behind the B'IF,X-project, in particular Donald Knuth and Leslie Lamport, for their wonderful typesetting program. The supervisors of my thesis were Prof. Dr. G. Flaig and Prof. Dr. D. Marin. I would like thank both for their thorough reading of the book and their ideas and critical remarks. lowe much gratitude to my academic teacher Prof. Dr. G. Flaig for many insightful conversations about modern time series econometrics, asymptotic theory and data problems. His empha sis on clear theory combined with sound econometric methods formed my personal ideal of how to do applied economics. After he left the university in 1998I had the opportunity to continue my work at the Center for Economic Studies. Prof. Dr. H. -W. Sinn and my new colleaguesat CES provided a stim ulating environment for the completion of my dissertation. Finally, I want to thank my parents, my sister and my friends for being there and sharing my cheers and worries. Munich, October 2000 Michael Reutter Our main concern in philosophy and in science should be the search fortruth. Justification is not an aim; and brilliance and cleverness as such are boring.