ETD

• Commodity markets
• CVaR
• Extreme Value Theory
• GARCH
• Tail risk.
• VaR

This master’s thesis explores how Extreme Value Theory (EVT) can improve the measurement of extreme risk in commodity prices, focusing on Gold and Brent crude oil between 2000–2023. It integrates AR–GARCH models with Peaks-Over-Threshold (POT) methods using the Generalized Pareto Distribution (GPD) to capture volatility clustering and fat-tailed behavior typical of financial returns during crises such as the 2008 financial crash, COVID-19 pandemic, and Russia–Ukraine conflict.
The study applies Monte Carlo backtesting to assess the accuracy of Value-at-Risk (VaR) and Conditional Value-at-Risk (CVaR) estimates under systemic stress. Results show that EVT significantly enhances tail-risk estimation, producing VaR exceedances closely matching theoretical levels (within 5%). Gold shows moderate tail risk, confirming its limited safe-haven function, while Brent displays higher tail sensitivity during shocks. The research demonstrates that the AR–GARCH–EVT framework yields robust, policy-relevant insights for investors and regulators facing volatile commodity markets.