This course deals with multivariate time series analysis from an applied perspective. After briefly refreshing the knowledge about univariate time series models (ARMA), we will continue with a short look at error correction models that take cointegration relationships explicitly into account. The main part of the lectures concerns the analysis of the multivariate time series models (VARs) and their extensions to incorporate structural characteristics (SVAR). After discussing the estimation routines in more depth, we will tackle the problem of identifying the nature of the structural shocks and the possible econometric strategies (short- vs. long-run restrictions and sign restrictions). Finally, we will derive some recommendations for policymakers based on an impulse response analysis. Moreover, this course provides a brief introduction to the Bayesian paradigm in econometrics and its advantages compared to the frequentist approach.

The course will be helpful for students interested in working at research institutions or financial institutions. Rather than focus narrowly on the application of econometric tools in macroeconomics, we will try to convey a deeper understanding of the most important tools used in applied time series analysis, their proper use and their limitations, illustrated by applications to questions considered in macroeconomics. The discussed methods are used heavily in Central Banks and policy institutions and will be covered with a special emphasis on their applications and interpretations. Finally, the students will be enabled to conduct own small research projects applying time series analysis.

Attendance is mandatory (however, three missed units are tolerated)

This lecture consists of two main blocks. While in the first block we are discussing the topics mentioned in the syllabus (slides, literature, and papers will be provided), the second block is dedicated to students' presentations of famous examples in the literature.  The group (max. 5 students) is expected to scrutinize the paper in depth (objective, relevant assumptions, model framework, and results) and provide (a) a detailed discussion as well as (b) potential comments/questions/suggestions to the authors. There will be sufficient time for a deeper discussion.

Please make sure that you read the assigned literature PRIOR to the lecture.

Final Exam (40Points), Paper presentation (30Points), and Exercises (30Points)

A positive final test (50% threshold of total exam points) is required for passing the course.

Grading Key: <60Points: fail; >61Points: sufficient; >71Points: satisfactory; >81points: good; >91Points: very good.

The final exam will be a mix of multiple-choice and open questions (90 minutes).

Currently, the final exam is planned to take place offline at the campus. However, in case of another surge of the pandemic, it will be switched towards an online exam.

Students should have a sound knowledge of statistics (probability, random variables, expectations, joint/conditional distributions), mathematics (linear algebra, differential/integral calculus, algebra), and basic econometrics (OLS/ML estimation). Moreover, it is expected that the students are familiar with univariate time series econometrics (if not, it is expected that students refresh their knowledge with "Applied Econometric Time Series" by Enders (CH1-4). However, we will refresh the essentials in the second session.)