BAN427 Insurance Analytics (E)

Autumn 2026

Topics
Across all industries, the ability to utilize data and data science methods is essential for gaining a competitive advantage. Insurance is of particular interest due to the abundance of data and the long tradition for advanced risk modelling. This course gives you an introduction to insurance economics and how data science typically is applied in order to solve real business problems. We will work with data from insurance and work hands on using Python and Jupyter Notebook. The topics we focus on will be:
- Introduction to Non-Life Insurance. Adverse selection and moral hazard - and how to measure this empirically. We will also touch upon EU Taxonomy regarding "Leadership in modelling and pricing of climate risks".
- Big data in insurance/finance. How to establish a Customer TimeLine. The importance of event data and how to utilize event data in real life predictions.
- Prediction methods with applications to insurance. Introduction to the standard "ML tool set" (logit regression, regression trees, random forest, ensemble methods and more).
- Prediction versus causation. Causal models, combining ML and causal methods.
- How to use randomized experiments in order to improve business processes.
- From predictive modelling to production. How to deploy and maintain many prediction models in a business environment? Keywords: Microservices, Streaming data, on-the-fly scoring.
Learning outcome
After completing the course students:
Knowledge
- Know how big data and machine learning techniques is used in the insurance industry.
- Know how to build, deploy and test models and treatments using randomized experiments.
- Have brief knowledge about EU taxonomy and the modelling of climate risks.
Skills
- Can bring insurance problem into a statistical model.
- Can analyze and predict important insurance outcomes using machine learning techniques.
General Competence
- Have general knowledge about measuring adverse selection and moral hazard from insurance data.
- Know how domain knowledge can be used to extract "causal" knowledge from observational data.
- Have knowledge about correlation vs causality - and how to empirically address causality using domain knowledge.
Teaching

7 lectures of 2 x 45 minutes. Anonymous data will be provided for applications of ML methods in the insurance business.
Recommended prerequisites

Econometrics - for example ECN402, BUS444 or BAN431.
Knowledge with Python and Jupyter Notebook is useful, but not a requirement. The assignment given at the end of the course will involve data and prediction modelling. Feel free to use whatever code language you prefer for the assignment (R, Python, Stata, SAS).
Credit reduction due to overlap

None.
Compulsory Activity

Hand in of a small assignment given at the first day of the course.
The old compulsory activity from previous years is still valid.
Assessment

Written assignment (group work, 2-3 students in each group).
The students will work on the assignment for approximately 10 days.
Grading Scale

Pass-Fail.
Computer tools

Python combined Jupyter Notebook, R is optional.
Literature

Suggested general background/supportive literature (more details given during the course):
Einav & Finkelstein "Selection in Insurance Markets", https://doi.org/10.1257/jep.25.1.115
Aarbu (2015) - "Asymmetric Information in the Home Insurance Market". https://doi.org/10.1111/jori.12084
Zhang, Bradlow & Small (2013) : "New measures of clumpiness for incidence data https://doi.org/10.1080/02664763.2013.818627
Varian (2014): "Big Data: New Tricks for Econometrics" https://doi.org/10.1257/jep.28.2.3
Mullaainathan & Spiess (2017): "Machine Learning: An Applied Econometric Approach" https://doi.org/10.1257/jep.31.2.87
Breiman (2001), "Statistical Modelling": The Two Cultures https://www.jstor.org/stable/2676681
Sutton & Barto, "Reinforcement Learning: An Introduction (chapter 1 and 2) http://incompleteideas.net/book/RLbook2018.pdf
Loss Data Analytics (https://openacttexts.github.io/Loss-Data-Analytics/)