From GARCH to Artificial Intelligence: Bibliometric and Thematic Review (1995-2025)

Abstract

This bibliometric study focuses on the changing nature of factors affecting the stock market volatility between the year 1995-2025. Using a data source of 89 reviewed research articles in Scopus and analyzed using VOSviewer through visualizations. The analysis also reveals the main clusters of influence, which include the macroeconomic indicators, political stability, and oil-price shocks, as well as behavioural drivers, including investor sentiment. The results show that the previous papers focused on the GARCH-based modeling and seasonality, whereas recent literature predicts the incorporation of artificial intelligence, deep learning methods, and ESG-associated volatility spill-overs. Bibliographic coupling highlights the fact that network theory is becoming more and more relevant to understand spatial correlations and dynamics of regional contagion. The study reveals that volatility is no longer necessarily understood as an economic variable but instead as a dynamic of complicated world systems. These results serve vital information to policy makers who are working hard to reduce systemic risk as well as retail investors who deal with information asymmetry or imbalance. Future studies should address the gap of methodological fragmentation by combining explainable artificial intelligence with theory-consistent financial models.