Teaching/Classes: EC220, Introduction to Econometrics.
Office hour: Monday S684 16.30-17.30 or by appointment
Course webpage: EC220
|- Teaching comments - Comments on problem sets - Past exams - Personal page|
|Teaching comments This form is here to allow you to make anonymous comments and suggestions in order to improve the classes. I will be happy about any comments. If you have got any questions concerning the presented material, please do not hesitate to ask them.|
Comments on past problem-sets
The first source of information is the course webpage. The files below contain the first slides of the past classes with a brief comment on your problem sets and additional remarks on some topics. They only point out some general problems and are not intended as a substitute to the lectures/classes or the course webpage in any way. You can find answers to the similar starred exercises from the book in the study guide.
|Michaelmas Term||Lent and Summer Terms|
|Pdf file||Updated on:||Pdf file||Updated on:|
|PS2 Stata file||10/27||PS11||02/02|
|PS3 Gauss Markov||11/10||PS12||02/02|
|PS5||11/18||PS14 not available|
|PS6 Dummy trap||11/26||PS15 not available|
|PS8 not available||Note on ADF test||03/17|
|PS9 not available||PS18 not available|
|Questions Here are a few answers to some of the questions I have had at this point:|
|- Why is the formula "TSS=ESS+RSS" not valid when the model has no intercept? Answer.|
|- Multi-collinearity and confusion between MSD and standard-errors. Answer.|
|- The dummy variable trap and exact multicollinearity. Answer.|
|- Heckman selection model and binary dependent variables. Answer.|
|- Data sets for PS14 to PS18 (time series). Answer.|
Past exams and solutions
The past exams and their solutions and pretty much everything you need for this course (besides lectures and classes) are available on the course webpage. When looking at the past exams (before 2007/2008) you should pay a particular attention to the notations which are slightly different from those we use this year. I advise you to have a look at some of the questions of the past exams during the winter break.
In particular, these questions are interesting at this stage (01/12):
2004 (Q1a Q1bi Q1bii Q3a Q4), 2005 (Q1 Q4b), 2007 (Q1 Q2)
This is just to illustrate the Gauss-Markov theorem on efficieny of OLS among linear unbiased estimators for homoskedastic linear models.
"The Gauss-Markov theorem is an efficiency justication for the least-squares estimator, but it is quite limited in scope. Not only has the class of models been restricted to homoskedastic linear regressions, the class of potential estimators has been restricted to linear unbiased estimators. This latter restriction is particularly unsatisfactory, as the theorem leaves open the possibility that a non-linear or biased estimator could have lower mean squared error than the least-squares estimator". (B. Hansen, 2007).
Simulations 2, 3 and 4 illustrate this statement for biased estimators (beta1 is different from 0).