Time-Series Foundation AI Model for Value-at-Risk Forecasting
This addresses the problem of financial risk management for institutions by demonstrating improved VaR forecasting with a foundation model, though it is incremental as it builds on existing foundation model concepts applied to a specific domain.
This study tackled the problem of forecasting Value-at-Risk (VaR) by analyzing a time-series foundation AI model, showing that the fine-tuned model consistently outperformed traditional methods like GARCH and GAS in actual-over-expected ratios over 8.5 years of out-of-sample data, ranking as the best or among the top performers across multiple quantiles.
This study is the first to analyze the performance of a time-series foundation AI model for Value-at-Risk (VaR), which essentially forecasts the left-tail quantiles of returns. Foundation models, pre-trained on diverse datasets, can be applied in a zero-shot setting with minimal data or further improved through finetuning. We compare Google's TimesFM model to conventional parametric and non-parametric models, including GARCH and Generalized Autoregressive Score (GAS), using 19 years of daily returns from the SP 100 index and its constituents. Backtesting with over 8.5 years of out-of-sample data shows that the fine-tuned foundation model consistently outperforms traditional methods in actual-over-expected ratios. For the quantile score loss function, it performs comparably to the best econometric model, GAS. Overall, the foundation model ranks as the best or among the top performers across the 0.01, 0.025, 0.05, and 0.1 quantile forecasting. Fine-tuning significantly improves accuracy, showing that zero-shot use is not optimal for VaR.