Spin glass models and market volatility: A multiple threshold nonlinear autoregressive distributed lag approach

This paper investigates how market volatility influences systemic financial tension across different regimes. Drawing inspiration from the Sherrington–Kirkpatrick spin glass model, we construct a disordered interaction matrix among major technology stocks and simulate Hamiltonian energy to represent systemic tension. We then employ a Multiple Threshold Nonlinear Autoregressive Distributed Lag Approach (MT-NARDL) model to explore how volatility, segmented into three distinct regimes, affects the energy dynamics. Our findings reveal that both prolonged low-volatility and high-volatility regimes significantly reduce systemic disorder over time, while moderate volatility has no statistically significant effect. These asymmetric results underscore the regime-specific nature of financial dynamics and prove the utility of interdisciplinary approaches for capturing complex market behaviors. The spin glass model, when adapted to financial time series, offers a novel and theoretically grounded framework for understanding disorder in evolving market structures. The Gini coefficient results reveal that financial volatility and asset interaction networks exhibit markedly lower structural inequality than group-based social systems, though both are governed by spin-glass-inspired disorder.