Non-Linear Econometric Models for Assessing High Frequency Financial Market Stability Amid Economic Shocks and Regulatory Interventions

This study evaluates high-frequency financial market stability during major economic shocks and regulatory interventions using advanced nonlinear econometric models. Focusing on events like the Global Financial Crisis, Brexit, COVID-19 crisis, Quantitative Easing, and the Eurozone Debt Crisis, the analysis employs high-frequency datasets including enriched temporal and interaction-based variables. Comprehensive exploratory data analysis revealed significant variability in financial metrics showing the dynamic market responses to crises. Among the evaluated models—GARCH, ARIMA, LSTM, and Markov Switching—the GARCH model demonstrated superior performance with consistent error metrics (MAE: 0.72, RMSE: 0.68) across k-fold cross-validation. Results highlight heightened volatility and negative abnormal returns during geopolitical and economic shocks, with regulatory interventions like Quantitative Easing mitigating instability. The nonlinear patterns in stability indicators revealed persistent volatility clustering effects, particularly during crises. Methodological rigor ensured robust out-of-sample validation, confirming the reliability of GARCH in capturing market dynamics. The findings emphasize the importance of combining statistical and machine learning approaches for predictive modeling and informed decision-making in financial markets. This paper provides a foundational framework for understanding and forecasting market stability under varying economic conditions, contributing to the literature on nonlinear econometrics and financial risk analysis.