Bridging Econometrics and AI: VaR Estimation via Reinforcement Learning and GARCH Models
This addresses risk management challenges for financial institutions by providing a more adaptive approach, though it is incremental as it builds on existing econometric and AI methods.
The paper tackled the problem of accurately estimating Value-at-Risk (VaR) in volatile financial markets by proposing a hybrid framework that combines GARCH models with deep reinforcement learning, resulting in significant improvements in VaR accuracy, reduced breaches, and lower capital requirements on Eurostoxx 50 data.
In an environment of increasingly volatile financial markets, the accurate estimation of risk remains a major challenge. Traditional econometric models, such as GARCH and its variants, are based on assumptions that are often too rigid to adapt to the complexity of the current market dynamics. To overcome these limitations, we propose a hybrid framework for Value-at-Risk (VaR) estimation, combining GARCH volatility models with deep reinforcement learning. Our approach incorporates directional market forecasting using the Double Deep Q-Network (DDQN) model, treating the task as an imbalanced classification problem. This architecture enables the dynamic adjustment of risk-level forecasts according to market conditions. Empirical validation on daily Eurostoxx 50 data covering periods of crisis and high volatility shows a significant improvement in the accuracy of VaR estimates, as well as a reduction in the number of breaches and also in capital requirements, while respecting regulatory risk thresholds. The ability of the model to adjust risk levels in real time reinforces its relevance to modern and proactive risk management.