Logit business failure prediction in V4 countries

Abstract

The paper presents the creation of the model that predicts the business failure of companies operating in V4 countries. Based on logistic regression analysis, significant predictors are identified to forecast potential business failure one year in advance. The research is based on the data set of financial indicators of more than 173 000 companies operating in V4 countries for the years 2016 and 2017. A stepwise binary logistic regression approach was used to create a prediction model. Using a classification table and ROC curve, the prediction ability of the final model was analysed. The main result is a model for business failure prediction of companies operating under the economic conditions of V4 countries. Statistically significant financial parameters were identified that reflect the impending failure situation. The developed model achieves a high prediction ability of more than 88%. The research confirms the applicability of the logistic regression approach in business failure prediction. The high predictive ability of the created model is comparable to models created by especially sophisticated artificial intelligence approaches. The created model can be applied in the economies of V4 countries for business failure prediction one year in advance, which is important for companies as well as all stakeholders.

Corporate Failure Prediction of Construction Companies in Poland: Evidence from Logit Model

“…The relevance is measured with precision (positive predictive value) and recall (also called sensitivity). Based on that, we can say how many selected items are relevant and how many relevant items are selected (Durica, Valaskova, and Janoskova, 2019). Both measures are presented in figure 1, where the precision is marked in red, and the recall measure is marked in blue.…”

“…On the other hand, the recall measure is equal to 99%, which means that in 99% of cases, the model (or classifier) correctly identifies companies with a bad financial condition. A detailed analysis will be carried out based on the error matrices presented in Tables 4A-4C. Error matrix examines the classification model's ability to predict failure among a new set of companies (Durica, Valaskova, and Janoskova, 2019). Modeling companies' financial situation is an important factor for recognition of the early signs of deterioration of the financial condition.…”

Business risks and their impact on business future concerning the entrepreneur’s experience with business bankruptcy: case of Czech Republic

“…Business owners often focus attention and initial attempts on developing and selling their products and services without tracking expenses (Durica, Valaskova, & Janoskova, 2019). Accounting often receives secondary attention, although it should be the basis for running an efficient business (Brozyna, Mentel, & Pisula, 2016).…”

Zastosowanie modeli logitowych do zdiagnozowania zagrożenia bankructwem przedsiębiorstw

“…Ponadto wskazali, że konstruując model należy uwzględnić istniejące warunki makroekonomiczne 9 . Z kolei Słowaccy naukowcy stwierdzili, że model regresji logistycznej powinien obejmować informację o wielkości firmy, ponieważ jest to istotny predyktor prawdopodobieństwa trudności finansowych przedsiębiorstw 10 . Warto zauważyć, że E. Belyaeva skonstruowała model regresji logistycznej do prognozowania upadłości w branży IT 11 .…”