This course covers advanced subjects in econometrics, focusing on causal inference and model building. We will cover common problems of regression analysis and potential remedies. Applied examples and assignments will be laid out to use the R language.

The following topics will be covered in this course:

• Causal inference (endogeneity, instrumental variables, matching)
• Maximum likelihood estimation (limited dependent variables)
• Prediction (bias-variance trade-off, regularisation, averaging)

Prior knowledge of the following topics is expected:

• Multivariate regression (application, interpretation)
• Regression properties (least squares estimation, classical assumptions, estimator properties, Gauss-Markov theorem)
• Regression inference (hypothesis testing, confidence intervals, model selection)
• Assumption failures and remedies (heteroskedasticity, serial correlation, multicollinearity)
• Functional forms (dummy variables, interaction terms, log-transformations)

These are covered in Econometrics I -- it is assumed that you have a solid understanding of them. In addition, it would be beneficial to have working knowledge of R, e.g. from the Statistics with R course.

After this course you will be equipped to conduct advanced econometric analyses. You will be aware of common pitfalls and how they may be dealt with. That includes a solid understanding of causal inference. You will be able to independently apply your knowledge using R and critically review applied research.

Attendance is compulsory. Students are allowed to miss up to two units.

The course consists of lectures with focus on econometric theory. Assignments focus on  practical issues are designed to use R and will be worked on in groups.

Assessments are based on three components that must all be positive

• 40% assignments
• 20% midterm exam
• 40% final exam

The grading scheme is

1. [90, 100]
2. [78, 89]
3. [65, 77]
4. [51, 64]
5. [0, 50]

Sound knowledge of basic statistics, mathematics and matrix algebra. Successful completion of Econometrics I and basic knowledge in R is highly recommended.

nikolas.kuschnig@wu.ac.at