PhD Courses in Denmark

Faculty
Lasse Heje Pedersen (LHP) and Christian Stolborg (CST), videos made by Theis Ingerslev Jensen

Prerequisites
The course is designed as a first-year Ph.D. course. The prerequisites are knowledge of asset pricing theory and econometrics at a M.Sc. level and an ability to work independently with data using a programmatic computer language such as Matlab, R, or Python. Students must participate in the whole course and do all problem sets. 

Aim
The aim of the class is to introduce PhD students in finance and related fields to empirical asset pricing research methods using big data. 

Content
The course provides students with empirical asset pricing tools to use big data to analyze modern topics in financial economics. The course starts with a quick overview of asset pricing, empirical asset pricing, and how to work with big financial data. The course then covers the factor zoo, multiple testing adjustments, replication, machine learning in asset pricing, and asset pricing with frictions. In addition to the theoretical discussion, the students will gain access to a large data set of global equity returns and use this data to solve several mandatory exercises, which constitute an essential part of the course. Each student must make their own solution to each exercise and be able to explain this solution and present it.  Students are allowed to discuss the exercises and solution methods, but students are not allowed to copy each other. Students must disclose in their solutions if code has been copied from public sources (using public code is perfectly fine, but should be disclosed), and should disclose any other material used.

Slides, exercises, and other material: Available on Canvas. Preliminary slides and exercises available here

Lecture plan 
(Preliminary, 1h means 1 hour consisting of 45 minutes lecture and 15 minutes break)

Lecture 1 - 9 January, 9-12: A primer on asset pricing (hybrid, 3h) LHP 
Stochastic discount factors, tradable and non-tradeable factors, factor models

Lecture 2 - 16 January, 9-12: A primer on empirical asset pricing (hybrid, 3h) LHP 
How to make and use factors, time series and cross-sectional regressions, predictability in the time series and the cross section, further methods

Discussion of Exercise 1 (Beta-dollar neutral portfolios)

Lecture 3 - 23 January (video): Working with big asset pricing data (video, 3h) 
WRDS, CRSP, Compustat, JKPfactors, global data 

Lecture 4 - 30 January, 9-12: The factor zoo and replication (hybrid,3h) LHP and CST 
Replication crisis, frequentist and Bayesian multiple testing adjustments

Discussion of Exercise 2: Construct Value Factors

Lecture 5 - 6 February, 9-16: Machine learning in asset pricing (on campus, 6h) LHP and CST
Validation, hyper-parameters, penalized regressions, trees, neural networks, feature importance, asset pricing applications 

Discussion of Exercise 3: Factor replication analysis

Work on Exercise 4 

Lecture 6 - 7 February, 9-13: Asset pricing with frictions (on campus, 4h) LHP
Transaction costs, market liquidity risk, funding liquidity risk, frictions meet machine learning

Lecture 7 - 20 February, 9-10: Discussion of Exercise 4: High-dimensional return prediction (hybrid, 1h) CST
Participation on campus in lectures 5 and 6 (6-7 February) is mandatory.

Learning objectives 
The course objectives are to:

• Work with big financial data, including making factors

• Apply factor models to estimate risk and expected return

• Estimate stock return predictability via regressions and portfolio sorts 

• Evaluate potential replication crisis and the factor zoo

• Implement multiple testing adjustments using frequentist and Bayesian methods

• Apply machine learning to asset pricing data

• Analyze financial market frictions

For further information and registration please see www.cbs.dk