Financial stress shocks and digital asset performance across market conditions

Abstract

This study examines the predictive capacity of the Financial Stress Index (FSI) for major digital asset categories - Bitcoin, DeFi tokens, and NFT tokens - across bearish, normal, and bullish market regimes. We deploy a rolling-window wavelet quantile Granger causality (RWWQGC) framework capturing time variation, frequency heterogeneity, and distributional asymmetries, supported by quantile-on-quantile regressions (QQR) for robustness. FSI predictability is state- and horizon-dependent. Under normal market conditions, predictive power stabilizes at lower frequencies. Conversely, in bearish and bullish regimes, predictability shifts toward higher frequencies. QQR estimates confirm nonlinear, asymmetric, and tail-dependent systemic risk transmission. Systemic financial stress drives long-horizon portfolio rebalancing during normal market states, but induces short-term speculative trading in extreme regimes, offering direct implications for cross-asset hedging and macroprudential oversight. This study extends the systemic risk literature across heterogeneous crypto segments through a unified time–frequency–quantile framework, uncovering state-contingent risk propagation overlooked by conventional linear models.