Dynamic Volatility Adjustment Models for Financial Time Series under News-Induced Shocks

Volatility forecasting is a cornerstone of financial analysis, but standard models often fail to account for abrupt market changes caused by news shocks. This study proposes a dynamic adjustment framework for volatility forecasting, enhancing traditional GARCH models to incorporate the transient and persistent effects of news-induced shocks. By leveraging a novel similarity-based method, the approach identifies historical periods with comparable market conditions to adjust forecasts dynamically in response to shock intensity. The framework is validated using financial market data from major geopolitical events, including the 2016 U.S. presidential election. Results demonstrate significant improvements in forecast accuracy, providing robust tools for risk management and investment decision-making in volatile financial environments. This work contributes to the growing intersection of econometrics and real-time financial decision-making, offering practical insights into managing uncertainty in global markets.