This repository contains three applied projects from the Quantitative Financial Risk Management course at Vrije Universiteit Amsterdam.
Across the assignments, we worked on market risk measurement, model validation, dependence modelling, and tail-risk analysis using real financial time series data.
The first assignment focused on Value at Risk and Expected Shortfall for a multi-asset portfolio. We constructed and cleaned the data, defined the portfolio, estimated risk measures under several models, and compared their performance through backtesting and stress testing. The main models included the Normal model, Student-t model, Historical Simulation, Filtered Historical Simulation, and volatility-based approaches.
The second assignment was a model validation exercise. We reviewed another team’s risk model, checked their documentation, assessed their assumptions, and independently reproduced parts of their analysis. The goal was not only to verify the numerical results, but also to identify where modelling choices, data handling, or missing documentation could affect the reliability of the risk estimates.
The third assignment extended the analysis to PCA, Factor Analysis, Copulas, and Extreme Value Theory. We studied the common drivers of asset returns using both statistical and economically motivated factor methods. We then analysed dependence between selected asset pairs using Gaussian and Student-t copulas, with particular attention to tail dependence. Finally, we applied EVT methods to model extreme losses and estimate tail-risk measures such as VaR and Expected Shortfall.
- Value at Risk
- Expected Shortfall
- Market risk modelling
- Backtesting
- Stress testing
- Model validation
- Historical Simulation
- Filtered Historical Simulation
- Volatility modelling
- Principal Component Analysis
- Factor Analysis
- Copulas
- Tail dependence
- Extreme Value Theory
- Peak-over-Threshold modelling
- Financial time series analysis
The purpose of this repository is to collect the full workflow of the course assignments in one place: from basic portfolio risk measurement, through independent model validation, to more advanced dependence and tail-risk modelling.
The work is academic and intended to demonstrate practical implementation of quantitative risk management methods.