Novel Market Temperature Definition Through Fluctuation Theorem: A Statistical Physics Framework for Financial Crisis Prediction

This paper introduces a novel approach to financial crisis prediction by establishing a thermodynamic-like framework derived from the fluctuation theorem of statistical physics. We define market temperature through the probability ratio of positive to negative returns and demonstrate its effectiveness in identifying market states and predicting potential crises. Our empirical analysis spans nine major global indices from 2005 to 2025, revealing statistically significant differences in temperature dynamics between crisis and non-crisis periods. Most notably, we discover a counterintuitive relationship between market temperature stability and crisis occurrence: crises tend to emerge more frequently during periods of apparent temperature stability rather than instability. This finding suggests that unusually stable periods in market temperature might signal the accumulation of systemic risks, similar to the calm before a storm in physical systems.