ABSTRACT : This study applies machine learning classification models to predict bankruptcy risk among real estate firms in Vietnam, particularly in the context of heightened market volatility and elevated credit risk during the period 2019–2023. Using firm-level financial data and macroeconomic variables obtained from the National Statistics Office, the authors develop and compare the predictive performance of several algorithms, including LightGBM, XGBoost, Random Forest, and Decision Tree, against the traditional Logistic Regression model. The empirical results indicate that boosting techniques, specifically LightGBM and XGBoost, outperform other approaches in terms of predictive accuracy and overall performance metrics. SHAP analysis reveals that the ratio of shareholders’ equity to total liabilities is the most influential predictor of bankruptcy risk. In addition, asset utilization efficiency and capital structure exert significant effects on financial distress probability. Profitability indicators and cash flow generation capacity further confirm their essential role in maintaining financial stability. While institutional and macroeconomic factors exhibit certain effects, their influence is not dominant relative to firm-specific financial characteristics. Overall, the findings suggest that bankruptcy risk in the real estate sector primarily stems from firms’ internal financial structure and operational efficiency rather than external macroeconomic conditions.
Chi et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: