Abstract
Classification and prediction of bankruptcy of businesses is a widely-addressed topic at the international level, but there are few studies of this type applied to Chilean companies. In this context, the objective of this research is to identify which model classifies and predicts bankruptcy of businesses in Chile with the highest degree of reliability. To this end, we compare three commonly used models: multiple-discriminant analysis (MDA), logistic regression (LOGIT), and neural networks, which use different financial indexes, macroeconomic variables, and other control variables. The models were applied to a sample of 98 randomly selected companies with unrestricted lines of business, 49 bankrupt and 49 not bankrupt. The result of the research shows that while the neural networks model was superior, the MDA model as well as the LOGIT, with respect to classification and prediction, all require other tools to determine the optimal set of variables to use.
By submitting a contribution to consideration of the Editorial Board of Fórum Empresarial, the authors attest that it is an original, unpublished work, which has not been nor will be simultaneously submitted to another journal for consideration and publication; that they are responsible for the work carried out and the content of the article; and they have the corresponding copyrights.
The authors grant the right of first publication of their work to Fórum Empresarial, in any medium and physical and electronic format, including internet. Publication shall be subject to the Creative Commons Attribution-NonCommercial 4.0 International License, which allows third parties to share the work, provided that the author and Fórum Empresarial (as a first publication) are cited.
The journal allows authors to retain publishing rights without restrictions. Authors are able to enter into separate, additional arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in Fórum Empresarial.