Autori
Magrini, Alessandro
Balzano, Marco

Titolo
No easy way out: dissecting firm heterogeneity to enhance default risk prediction
Periodico
Sinergie
Anno: 2025 - Volume: 43 - Fascicolo: 128 - Pagina iniziale: 161 - Pagina finale: 183

Purpose of the paper: Effective risk assessment is central to managerial decision-making in financial institutions, corporate finance, and strategic planning. The paper investigates how default risk prediction varies depending on firm characteristics. It examines whether financial indicators differ in predictive relevance based on industry, technology level, size, and age. Methodology: The analysis draws on data from the AIDA database, covering 121,809 Italian firms operating across different sectors and technological contexts. The study uses logistic regression and random forests to test whether financial indicators—grouped into liquidity, efficiency, profitability, and growth—vary in predictive strength depending on firm-specific factors. Findings: Results support that (a) service-oriented firms are more affected by liquidity indicators, (b) high-tech firms are more responsive to efficiency metrics, (c) smaller firms are more influenced by profitability measures, and (d) younger firms are more sensitive to growth indicators. These findings support the use of tailored prediction models rather than uniform approaches. Research limits: The study is based on firms operating in Italy and does not account for possible institutional or macroeconomic differences in other national contexts. It also focuses on financial indicators, without including qualitative or behavioral variables. Practical implications: The study suggests that risk assessment models can be refined by incorporating firm-level contingencies. This has potential implications for analysts, risk officers, and institutions involved in SME financing or credit scoring. Originality of the paper: The paper contributes to research on default prediction by combining a contingency perspective with both statistical and machine learning techniques.



SICI: 0393-5108(2025)43:128<161:NEWODF>2.0.ZU;2-E
Testo completo: https://ojs.sijm.it/index.php/sinergie/article/view/1905/1032

Esportazione dati in Refworks (solo per utenti abilitati)

Record salvabile in Zotero

Biblioteche ACNP che possiedono il periodico