Accounting-based early warning models for financial distress are sensitive to sectoral and macroeconomic conditions, leading to non-comparability of indicators and temporal instability in prediction. We propose a projection-based representation of accounting data that removes external structure prior to modeling. Using an external analysis framework, firm-level accounting indicators are orthogonally decomposed into a component explained by sector-level and macroeconomic information and a residual component capturing firm-specific deviations from context-dependent average behavior. Unlike standard approaches that include macroeconomic variables directly in predictive models or rely on demeaning or fixed effects, the proposed method operates at the represen tation level, yielding context-adjusted indicators that are orthogonal to external conditions by construction. Predictive performance is evaluated in an out-of-time early-warning setting using Italian small and medium sized enterprises observed from 2016 to 2024. Results show stable discriminatory performance across test years, supporting the use of context adjusted representations for temporally robust early-warning analysis.

Integrating External Information in Accounting-Based Early Warning Models / Ariante, I., Cavicchia, C.. - (2026), pp. 290-295. (53rd Scientific Meeting of the Italian Statistical Society and 1st Scientific Meeting of the European Statistical Societies ) [10.1007/978-3-032-30873-3_50].

Integrating External Information in Accounting-Based Early Warning Models

Ariante, Irene
Primo
;
Cavicchia, Carlo
Ultimo
2026

Abstract

Accounting-based early warning models for financial distress are sensitive to sectoral and macroeconomic conditions, leading to non-comparability of indicators and temporal instability in prediction. We propose a projection-based representation of accounting data that removes external structure prior to modeling. Using an external analysis framework, firm-level accounting indicators are orthogonally decomposed into a component explained by sector-level and macroeconomic information and a residual component capturing firm-specific deviations from context-dependent average behavior. Unlike standard approaches that include macroeconomic variables directly in predictive models or rely on demeaning or fixed effects, the proposed method operates at the represen tation level, yielding context-adjusted indicators that are orthogonal to external conditions by construction. Predictive performance is evaluated in an out-of-time early-warning setting using Italian small and medium sized enterprises observed from 2016 to 2024. Results show stable discriminatory performance across test years, supporting the use of context adjusted representations for temporally robust early-warning analysis.
2026
9783032308726
9783032308733
Integrating External Information in Accounting-Based Early Warning Models / Ariante, I., Cavicchia, C.. - (2026), pp. 290-295. (53rd Scientific Meeting of the Italian Statistical Society and 1st Scientific Meeting of the European Statistical Societies ) [10.1007/978-3-032-30873-3_50].
File in questo prodotto:
File Dimensione Formato  
978-3-032-30873-3 (1)-314-319.pdf

solo utenti autorizzati

Tipologia: Versione Editoriale (PDF)
Licenza: Accesso privato/ristretto
Dimensione 223.5 kB
Formato Adobe PDF
223.5 kB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11588/1074982
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact