Key Risk Indicator
Forecast Model
November 2025
Key Risk Indicator Forecasting Model
Confidentiality
Note that due to the confidentiality of this project I cannot disclose any critical information as of now.
During my internship at Commerzbank, I am developing a forecasting model designed to predict one of the bank's Key Risk Indicators in a stress scenario.
Summary
- Developing a Machine Learning forecasting model to predict a major Key Risk Indicator
- End-to-end responsibility for data processing, modeling, testing, and documentation
- Delivering forecasts and analytical insights to support analytical work within the risk domain
Motivation
Stress-testing requires banks to simulate how their capital structure reacts to adverse economic environments. To forecast how economic stress imapcts the bank models need to be developed.
I was tasked with designing such a model and ensuring that it produces stable and interpretable results.
I hold sole responsibility for the technical implementation.
I work under the supervision of a senior colleague who provides project management guidance
to ensure we meet our goals and deadlines. Additionally, I collaborate closely with another colleague
who serves as a mental sparring partner and validates the statistical correctness of my approach.
Methodology
The modeling process consists of several stages. While specific implementation details remain confidential, the general workflow includes:
Modeling Steps
- Data Integration: Collecting and processing relevant parameters and aligning them with historical values of the Key Risk Indicator
- Feature Engineering: Applying appropriate transformations and structures based on exploratory analysis and domain knowledge
- Model Design: Developing an econometric approach to capture relationships between economic drivers and changes in key financial metrics
- Validation: Testing the model for stability, sensitivity, and predictive accuracy using industry-standard methods
- Model usage: Generating forecasts under various economic scenarios
What I Learned
This project is strengthening my ability to combine quantitative modeling with communication tailored to non-technical stakeholders. I'm gaining deep insight into the intersection of macroeconomics and banking regulation, and learning how to deliver models that balance statistical rigor with regulatory requirements and practical business needs.
Skills Demonstrated
- Econometric modeling & statistical analysis
- Data preprocessing and feature engineering
- Model validation & scenario simulation
- Technical documentation for risk and regulatory stakeholders
- Independent problem-solving and project ownership
- Cross-functional communication with risk management teams