Publication Details
Abstract
Against the background of further market transformations and strengthening of economic systems, the financial sustainability of an enterprise has become one of the important household factors of the macroeconomic stability and the national economic security. Although a variety of bankruptcy prediction methods exist, such as discriminant analysis and logistic regression, panel econometrics, and fuzzy logic, their straightforward application in transition economies like Uzbekistan is rarely visible in practice due to institutional specificities, accounting differences, and structural volatility. A holistic early warning framework able to be empirically validated is currently unavailable for Uzbek joint-stock companies, and traditional static models do not show adequate predictive adaptability in emerging market environments. The objective of this research is to design and test Altman Z-score, Taffler model, logistic regression, ROC analysis, panel data techniques and fuzzy-logit hybrid modelling based integrated bankruptcy prediction system from balanced panel dataset for 2022–2024. The performance of the logistic model was 87.5% accuracy and AUC = 0.89; the integrated hybrid framework performed best (accuracy 0.91; AUC = 0.93), and the Fixed Effects estimation was favored based on the Hausman test (p=0.013). The study presents a pragmatic logit-panel-fuzzy framework based on the institutional and structural features of Uzbekistan. The proposed framework increases capacity for early detection, methodological rigour and practical applicability to banks, auditors and financial managers in emerging markets.