Measuring the Economic Resilience of FinTech Firms under Financial Stress Using Panel Survival Analysis

This study quantifies the economic resilience of FinTech firms under financial stress using a panel survival framework with time-varying covariates and stress-regime interactions. The dataset contains 1,284 FinTech firms observed quarterly from 2014Q1 to 2025Q4, yielding 31,912 firm-quarter observations and 189 adverse events (distress or exit). Stress regimes are identified from a composite index, producing 21,436 firm-quarters in normal conditions, 7,982 in moderate stress, and 2,494 in high stress, with event rates of 0.39%, 0.61%, and 2.25%, respectively. A standardized resilience score declines across regimes, with median values of 0.18 (normal), 0.06 (moderate), and −0.21 (high), consistent with stress-driven deterioration and selection. Segment-stratified Cox models show resilience is strongly protective (hazard ratio, HR = 0.74; p < 0.001) and becomes more protective during high stress via interaction effects (HR for resilience × high stress = 0.61; p < 0.001). Control variables align with fragility mechanisms, including higher leverage increasing hazard (HR = 1.29; p < 0.001) and profitability and size reducing hazard (HR = 0.88; p = 0.003; HR = 0.92; p = 0.018), while funding concentration increases hazard (HR = 1.18; p = 0.011). Heterogeneity analysis indicates the largest resilience protection occurs in lending (HR = 0.69; interaction HR = 0.57) and seed-stage firms (HR = 0.66; interaction HR = 0.54), whereas later-stage firms exhibit smaller marginal effects (public: HR = 0.86; interaction HR = 0.74). Predictive validation improves monotonically across model variants: C-index rises from 0.71 (controls) to 0.77 (frailty Cox), and Integrated Brier Score over 0–12 quarters declines from 0.168 to 0.139, while calibration slope improves from 0.89 to 1.01. Robustness checks across event definitions, competing risks, country frailty, and alternative stress cutoffs preserve the protective direction and magnitude of resilience effects (main HR range: 0.73–0.81; interaction HR range: 0.60–0.69). The findings establish resilience as a measurable moderator of stress-induced failure risk and provide decision-relevant indicators for resilience governance in digitally mediated financial intermediation.