Finance: Financial Econometrics and Machine Learning

Our journey will begin with the Gauss-Markov assumptions, which establish that linear regression is the Best Linear Unbiased Estimator (BLUE). In other words, if certain assumptions are met, you can be certain that the linear regression is the best approach. However, when theory meets reality: “Everybody has a plan until they get punched in the mouth” (Mike Tyson).

You quickly realise that you need tools to address the messy reality. Hence, we will utilise the theoretical underpinnings of linear regression to determine when other econometric tools are necessary.

To simplify, when you run into those problems, there are tools that address the problems:

• Autocorrelation: you use an Autoregressive Integrated Moving Average (ARIMA) model
• Heteroscedasticity: you use Heteroscedasticity and Autocorrelation Consistent (HAC) estimators
• Endogeneity: you use panel data, which also opens the door to causal inference (correlation is not the same thing as causality)
• Non-linear and complex relationships: you use machine learning (universal function approximators)

The goal is that you gain a robust foundation for your future understanding of statistics and econometrics.

Course literature

The course literature listed may be updated up to eight weeks before the course begins.

The course will consist of in-person lectures and computer sessions to practice the concepts. The examination consists of two parts:

• a written exam at the end of the exam
• homework assignments to put what you have learnt into practice.

Additionally, there are a couple of online-quizzes throughout the course, which yields bonus points that counts in addition to your score on the written exam.

Prerequisites

Students admitted to the Master Programmes in Finance are qualified for this course. For other students, at least 90 ECTS-credits in economics or business administration, which must include a course in finance, and at least 15 ECTS-credits in statistics are required.