DOI: 10.30518/jav.1884951 ISSN: 2587-1676

Nonlinear Dynamics and Chaotic Microstructure of European Airline Stocks: Evidence from Recurrence Quantification Analysis

Eyyüp Ensari Şahin
This study investigates the nonlinear dynamics and chaotic microstructure of ten major European airline stocks—comprising eight European carriers and two Turkish airlines—over the period from January 2015 to January 2026, a timeframe that encompasses five structurally distinct crisis regimes: pre-pandemic stability, COVID-19 shock, recovery phase, Ukraine war impact, and post-war normalization. Employing a multi-method chaos framework that integrates Lyapunov exponents, Hurst exponents, BDS tests, and Recurrence Quantification Analysis (RQA)—including rolling-window and cross-recurrence extensions—the study addresses three core research questions: whether chaotic dynamics differ by business model, whether global crises induce permanent structural breaks, and whether Turkish airlines exhibit systematically distinct chaotic profiles from their European counterparts. The empirical results reveal heterogeneous chaos patterns: four carriers exhibit deterministic chaos while six display borderline or stable dynamics, challenging uniform business model categorizations. Crisis periods produce statistically significant and persistent structural transformations, with laminarity increasing 33% and entropy rising 28% between pre-COVID and post-war sub-periods, indicating regime shifts that do not revert to pre-pandemic baselines. Cross-recurrence analysis documents intensifying synchronization during crises, with post-war network density reaching complete integration and thereby eliminating diversification benefits. Turkish airlines demonstrate unexpected stability relative to their European peers, a pattern potentially attributable to domestic monetary policy buffering effects. Rolling-window analysis further reveals that determinism declines 30–40% prior to major market shocks, establishing RQA metrics as early warning indicators complementary to conventional volatility measures.

More from our Archive