Effectiveness of bankruptcy prediction models constructed for differently selected diagnostic variables

Bernard Kokczyński; Dorota Witkowska

doi:https://doi.org/10.59139/stattrans-2026-021

Effectiveness of bankruptcy prediction models constructed for differently selected diagnostic variables

Bernard Kokczyński Department of Corporate Financial Management, Faculty of Management, University of Lodz, Lodz, Poland ORCID:https://orcid.org/0000-0002-9379-0376 , Dorota Witkowska Department of Corporate Financial Management, Faculty of Management, University of Lodz, Lodz, Poland ORCID:https://orcid.org/0000-0001-9538-9589 Statistics in Transition new series, vol. 27, 2026, 2, pages: 145-160 Published online: 9 June 2026 https://doi.org/10.59139/stattrans-2026-021 Citation: Kokczyński B., Witkowska D., 2026. Effectiveness of bankruptcy prediction models constructed for differently selected diagnostic variables. Statistics in Transition new series, 27(2), pp. 145-160. https://doi.org/10.59139/stattrans-2026-021

24 Views 2 Downloads

ARTICLE

(English) PDF

ABSTRACT

The purpose of the article is to compare the effectiveness of discriminant models for predicting corporate bankruptcy, constructed using different methods of diagnostic variables selection. We compared several methods, such as arbitrary selection, the forward stepwise method applied after the initial selection of variables by means of a significance test of differences between group averages (further called the two-step method), the Hellwig method, the tstatistics and the backward stepwise method. We also assessed the models’ accuracy in terms of the synthetic measure. It was constructed by applying eight measurements of the classification effectiveness, such as the values of Wilks' lambda statistic and AUC together with the percentage of correctly identified companies, i.e. total, bankrupts and non-bankrupts in training and testing sets. The results show that the backward stepwise method and the two-step method generate models with the highest accuracy of classification. In addition, the study found that Wilks' lambda statistic is not a good approximation of the classification abilities of bankruptcy models. The contribution of our paper is a comparative methodological study, focusing on the impact of alternative diagnostic variable selection techniques on the linear discriminant function accuracy used to bankruptcy prediction.

KEYWORDS

diagnostic variable selection, linear discriminant function, bankruptcy prediction, Euclidean distance.

REFERENCES

Aczel, A., (2000). Statystyka w zarządzaniu [Statistics for management]. Wydawnictwo Naukowe PWN, Warsaw.

Altman, E., (1968). Financial Ratios, Discriminant Analysis and the Prediction of the Corporate Bankruptcy. The Journal of Finance, Vol. 23, pp. 589–609.

Chudik, A., Pesaran, H., Sharifvaghef, M., (2024). Variable selection in high dimensional linear regressions with parameter instability. Journal of Econometrics, Vol. 246, pp. 1–25.

Fan, J., Li, R., (2001). Variable Selection via Nonconcave Penalized Likelihood and Its Oracle Properties. Journal of the American Statistical Association, Vol. 96/456, pp. 1348–1360.

Gajdka, J., Stos, D., (1996). Wykorzystanie analizy dyskryminacyjnej do badania podatności przedsiębiorstwa na bankructwo [Application of discriminant analysis to examine corporate bankruptcy risk], In: Duraj, J. (ed.), Przedsiębiorstwo na rynku kapitałowym [Enterprise on the capital market]. Wydawnictwo Uniwersytetu Łódzkiego, Łódź, pp. 138–148.

Gruszczyński, M., (2012). Mikroekonometria, Modele i metody analizy danych indywidualnych [Microeconometrics: Models and methods of individual data analysis], 2nd ed., Wolters Kluwer Polska, Warsaw.

Grzybowska, U., Karwański, M., (2023). Selekcja zmiennych metodami statystycznymi i uczenia maszynowego. Porównanie podejść na przykładzie danych finansowych, [Variable selection using statistical and machine learning methods: Comparison of approaches based on financial data], Metody Ilościowe w Badaniach Ekonomicznych [Quantitative Methods in Economics], Vol. XXIV/4, pp. 229–241.

Harańczyk G., (2010). Krzywe ROC, czyli ocena jakości klasyfikatora i poszukiwanie optymalnego punktu odcięcia [ROC curves: Evaluation of classifier quality and search for the optimal cut-off point], StatSoft Polska, available at: 25.06.2024, https://media.statsoft.pl/_old_dnn/downloads/krzywe_roc_czyli_o cena_jakosci.pdf

Herman, S., (2018). Analiza porównawcza zdolności predykcyjnej wybranych metod prognozowania upadłości przedsiębiorstw [Comparative analysis of predictive ability of selected corporate bankruptcy forecasting methods]. Ruch Prawniczy, Ekonomiczny i Socjologiczny [Legal, Economic and Sociological Movement], Vol. LXXX/3, pp. 199–216.

IBM., (2023). Tabela badań Levene'a, available at: 26.12.24, https://www.ibm.com/docs/pl/spss-statistics/29.0.0?topic=variances-levene-test-table .

Jagiełło, R., (2013). Analiza dyskryminacyjna i regresja logistyczna w procesie oceny zdolności kredytowej przedsiębiorstw [Discriminant analysis and logistic regression in the process of assessing corporate creditworthiness]. Materiały i Studia [Materials and Studies], No. 286, pp. 1–116.

Kopczyński, P., (2022). Prognozowanie upadłości przedsiębiorstw [Corporate bankruptcy forecasting]. Wydawnictwo Uniwersytetu Łódzkiego, Łódź.

Korzeniewski, J., (2016). New method of variable selection for binary data cluster analysis. Statistics in Transition new series, Vol. 17/2, pp. 1–10.

Mączyńska, E., Zawadzki, M., (2006). Dyskryminacyjne modele predykcji upadłości przedsiębiorstw [Discriminant models for corporate bankruptcy prediction]. Ekonomista [Economist], No. 2, pp. 205–235.

Malarska, A., (2005). Statystyczna analiza danych wspomagana programem SPSS [Statistical data analysis supported by the SPSS software], SPSS Polska, Cracow.

Misztal, M., (2014). Wybrane metody oceny jakości klasyfikatorów – przegląd i przykłady zastosowań [Selected methods for classifier quality assessment – review and application examples], In: Jajuga, K., Walesiak, M. (eds.), Taksonomia 23. Klasyfikacja i analiza danych – teoria i zastosowania [Taxonomy 23. Classification and data analysis – theory and applications], Prace Naukowe Uniwersytetu Ekonomicznego we Wrocławiu [Research Papers of Wrocław University of Economics]. Wydawnictwo Uniwersytetu Ekonomicznego we Wrocławiu, Wrocław, pp. 156–166.

Nowakowski, M., (2019). The ANOVA method as a popular research tool. Studia i Prace WNEiZ US, 55, pp. 67–77.

Pociecha, J., Pawełek, B., Baryła, M., Augustyn, S., (2014). Statystyczne metody prognozowania bankructwa w zmieniającej się koniunkturze gospodarczej [Statistical methods of bankruptcy forecasting in changing economic conditions], Fundacja Uniwersytetu Ekonomicznego w Krakowie, Cracow.

Ptak-Chmielewska, A., (2012). Wykorzystanie modeli przeżycia i analizy dyskryminacyjnej do oceny ryzyka upadłości przedsiębiorstw [Application of survival models and discriminant analysis in assessing corporate bankruptcy risk]. Ekonometria [Econometrics], Vol. 4(38), pp. 157–172.

Raftery, E., Dean, N., (2006). Variable selection for model-based clustering. Journal of the American Statistical Association, 101/473, pp. 168–178.

Shi, Y., Li, X., (2019). An overview of bankruptcy prediction models for corporate firms: A Systematic literature review. Intangible Capital, 15/2, pp. 114–127.

Tibshirani, R., (1996). Regression Shrinkage and Selection via the Lasso, Journal of the Royal Statistical Society, Ser. B, Vol. 58/1, pp. 267–288.

Valaskova, K., Gajdosikova, D., Belas, J., (2023). Bankruptcy Prediction in the Post- Pandemic Period: A Case Study of Visegrad Group Countries. Oeconomia Copernicana, 14(1), pp. 253–293.

Wang, Z., Zhu, Z., Yu, C., (2023). Variable Selection in Macroeconomic Forecasting with Many Predictors. Econometrics and Statistics, pp. 1–18.

Wiktorowicz, J., Grzelak, M., Grzeszkiewicz-Radulska, K., (2020). Analiza Statystyczna z IBM SPSS Statistics [Statistical analysis with IBM SPSS Statistics]. Wydawnictwo Uniwersytetu Łódzkiego, Łódź.

Witkowska, D., (2023). Wybrane metody ilościowe w finansach [Selected quantitative methods in finance]. Wydawnictwo Uniwersytetu Łódzkiego, Łódź.

Woo Ahn, K., (2016). Modern variable selection techniques. Datum Newsletter Division of Biostatistics, Vol. 22/1.