Title: Predicting bankruptcy of micro-enterprises by industry: Integrating financial and web-based indicators Authors:  Carlo Bottai - University of Milano-Bicocca (Italy)
Lisa Crosato - Ca Foscari University of Venice (Italy)
Caterina Liberati - University of Milano-Bicocca (Italy) [presenting]
Abstract: The aim is to study bankruptcy prediction for micro-sized enterprises, often underrepresented in credit risk modeling due to their limited financial data quality. Building on previous research advocating for sector-specific approaches, we develop separate prediction models for selected industries using a dataset of 84,019 Italian micro-enterprises, of which only 1,308 (1.15\%) defaulted. The low default rate makes the classification problem particularly complex, especially when analyzed by sector. To overcome the limitations of models based solely on balance sheets, we integrate an innovative non-financial information source: features extracted from the HTML structure of company websites. These web-based indicators are combined with traditional financial variables to enhance model performance. A cross-validation scheme ensures the robustness and generalizability of results. Findings show that website data add significant predictive power, particularly in industries where digital presence is actively maintained. The relevance of these features varies across sectors, underlining the presence of sector-specific heterogeneity not only in financial patterns but also in web behavior. We demonstrate that website information represents a valuable and innovative signal for early-warning systems, especially in data-poor environments. This approach offers new perspectives for more accurate and industry-aware credit risk models for micro-enterprises.