An Uncertain Financial Fragility Modeling Framework Based on Uncertainty Theory

Financial fragility is a forward-looking indicator that characterizes the potential financial deterioration risks of economic entities under adverse shocks. Household financial fragility assessment is often subject to epistemic uncertainty induced by limited information and subjective expert judgement, which cannot be properly captured by classical probabilistic frameworks. This paper develops an uncertain logistic regression framework for household financial fragility modeling rooted in uncertainty theory. We first construct two measurement indicators, namely financial margin and liquidity capacity, to quantify household financial fragility. Then we derive the uncertain maximum likelihood estimator and design a numerical algorithm to implement parameter estimation. Using micro household data from the China Family Panel Studies (CFPS), we carry out empirical analysis for rural labor-transfer households. The empirical results show that the proposed uncertain logistic regression model delivers stable and economically interpretable estimates under epistemic uncertainty with unstable event frequencies. Compared with classical probabilistic logistic regression, the uncertain framework exhibits better applicability for such small-sample scenarios dominated by cognitive uncertainty.