Anders Bredahl Kock

Staff portrait: Anders Bredahl Kock

Anders Bredahl Kock is a postdoc at the Department and is affiliated with CREATES.

How long have you been employed at the Department of Economics and Business?
I have been a Postdoc since September 2011, but I started on my PhD in September 2007, which I guess is also some sort of employment, so technically speaking I have been employed here since 2007.

What is your background?
I was born in Copenhagen but moved to Hamburg, Germany, when I was very little. When I was around 11 years old my family moved to Vejle, and so it was (geographically) natural for me to study in Aarhus when I graduated from high school in 2003.

What are your main areas of research?
I guess I am working somewhere in the (non-empty) intersection between econometrics and statistics. I am mainly interested in theory, but I have also worked on some more applied projects together with my former supervisor Timo Teräsvirta which I also enjoyed a lot. The applied projects with Timo mainly revolved around using tool from machine learning and approximation theory to forecast economic time series. This is not entirely non-trivial in the sense that it can often be difficult to beat even quite simple time series models.

On the more theoretical side I am interested in high-dimensional econometrics/statistics. That is the study of models where the number of parameters to be estimated might be of a much larger order of magnitude than the sample size. How does one select the right variables if only a few are relevant and can one establish so-called oracle inequalities? These are questions which I am quite interested in. In particular, I have tried to take the sample size as given and take a non-asymptotic approach which seems to be popular these days (and it also reflects reality better since in practice one only has a given finite sample...).

What research projects are you working on at the moment?
- Modeling and predicting retirement ages (joint with Johannes Tang Kristensen and Malene Kallestrup Lamb)
- Forecasting vast realized covariance matrices (joint with Laurent Callot and Marcelo Medeiros)
- Oracle inequalities for convex loss functions for the elastic net (joint with Mehmet Caner)
- Oracle inequalities for high-dimensional panel data models

What are you involved in teaching-wise at the moment?
Nothing! But I am teaching the Mathematical Economics 1 course in the fall. It is a first semester math course. I actually enjoy teaching this course quite a lot - it is usually good fun (for me). 

Are you planning to take part in any conferences or the like abroad in the near future?
I am going to Budapest to attend the European Meeting of Statisticians in July.

Are you expecting any visiting researchers from abroad in the near future (or have you had any visitors from abroad recently)?
I was hosting a guest from Vienna last week. Perhaps I will meet with a co-author in August.

Are you cooperating with researchers from other departments at AU at the moment?
No, but I speak regularly to Jens Ledet Jensen from the Department of Mathematical Sciences who has similar interests as me.

What do you do when you are not at work?
I liked Mikkel's answer to this question from the previous interview... but I guess I try to spend some time with my friends. Since I am getting old I try to take at least one day of every weekend.

Supplementary question from Rune Vejlin: Did you ever complete Civilization 3 on the hardest level with 7 opponents?
I didn't know that you could choose less than seven opponents??? (on a serious note, I actually never did - but I never tried either since I was struggling at much simpler levels. Also I was more of a Civ1 and Civ2 guy. Those were my hay-days.)

Supplementary question from Mikkel Bennedsen: What is it like to be a university rock star?!?
It is very nice and as you might have noticed I have also let my hair grow! Don't you think I have the full package now?

Please give us your suggestion for the next department staff member to be presented:
Eric Hillebrand

What specific supplementary question would you like him or her to answer?
How low can you go? (when you dive)

18490 / i32