# 2017/2018  KAN-CMECV1702U  Cross Section and Panel Econometrics

 English Title Cross Section and Panel Econometrics

# Course information

Language English
Course ECTS 7.5 ECTS
Type Elective
Level Full Degree Master
Duration One Semester
Start time of the course Spring
Timetable Course schedule will be posted at calendar.cbs.dk
Max. participants 80
Study board
Study Board for HA/cand.merc. i erhvervsøkonomi og matematik, MSc
Course coordinator
• Ralf Andreas Wilke - Department of Economics (ECON)
• Mathematics
• Statistics and quantitative methods
• Economics
Last updated on 22-08-2017

Learning objectives
To achieve the grade 12, students should meet the following learning objectives with no or only minor mistakes or errors:
• Detect situations in which the ordinary least squares estimator is not adequate and be able to explain why.
• Understand econometric methods of estimation and inference for cross section data and panel data.
• Understand proofs and derivations in matrix notation.
• Choosing an econometric model, form those introduced in the course, and explaining why it is the suitable model for the specific situation
• Interpret estimation results in R/STATA output correctly and comment on appropriateness of their presentation.
• Relate R/STATA code and R/STATA output to the econometric models introduced in the course.
• Competence in R or STATA to do econometric analysis with the introduced models of the course.
Course prerequisites
The course has a high technical level. Students are expected to have knowledge of the statistical properties of ordinary least squares estimation and maximum likelihood estimation, as well as hypotheses tests about parameters in regression analysis.
Knowledge of matrix algebra, fundamentals of probability and mathematical statistics are required.
Basic knowledge of either R or STATA.
Examination
 Cross Section and Panel Econometrics: Exam ECTS 7,5 Examination form Home assignment - written product Individual or group exam Individual exam Size of written product Please see text below Maximum 10 pages, excluding references and appendices. The main textbody must not exceed 10 pages and should contain main result tables and figures which are required for the understanding of the main text. References and appendices with additional material (additional results or figures, code or proofs/derivations) do not count for the page limit but should not be too excessive. Assignment type Project Duration Written product to be submitted on specified date and time. Grading scale 7-step scale Examiner(s) One internal examiner Exam period Summer Make-up exam/re-exam Same examination form as the ordinary exam Description of the exam procedure Students do an unsupervised project work and submit a project report that is assessed. The analysis in the project should use one or more methods that have been covered in the course. The project can be methodological (by for example modyfying or extending an approach and/or by doing simulations to investigate properties) or it can be empirical (by applying the methods to an empirical problem) if the student has interesting data at hand. The topic of the project is to be suggested by the student and has to be approved upfront by the assessor.
Course content and structure

Topics in Cross Section Econometrics:

We consider various violations of common model assumptions (such as Gauss-Markov assumptions), including heteroskedasticity, auto/serial correlation, omitted variables, functional form misspecification, measurement error and simultaneity. We see how this can be tested for and what solutions exist (e.g. robust inference statistics, GLS, IV/2SLS methods). We then consider crafting and estimation of a system of equations including simultaneous equations and seemingly unrelated regression by 2SLS/3SLS and System GMM.

Topics in Panel Models:

We start with the main static linear panel models: Pooled OLS, FE,RE,FD. We then combine common static linear panel models with IV methods (FD-IV, FE-IV and RE- IV, Hausman-Taylor type models). We consider dynamic panel models (Anderson-Hsiao/ System 2SLS, Arrelano-Bond/ SGMM) before we move to nonlinear panel models (binary dependent variable only). Here we extend ML estimation to cope with dependent observations (using Kullback-Leibler Information criterion) and consider methods to correct estimated standard errors and statistics. The final point will be to consider panel attrition/unbalanced panels, if time permits.

Teaching methods
The course comprises of 25 hours of lectures and 8 hours of computer classes. The first computer class is an introductory class. The following classes cover problem sets.
Feedback during the teaching period
office hours