Title: Forecasting statistical methods in business: a comparative study of discriminant and logit analysis in predicting business failure

Authors: Ana García-Gallego; María-Jesús Mures-Quintana; M. Eva Vallejo-Pascual

Addresses: Faculty of Economics and Business Studies, Department of Economy and Statistics, Universidad de León, Campus de Vegazana s/n, 24071 León, Spain ' Faculty of Economics and Business Studies, Department of Economy and Statistics, Universidad de León, Campus de Vegazana s/n, 24071 León, Spain ' Faculty of Economics and Business Studies, Department of Economy and Statistics, Universidad de León, Campus de Vegazana s/n, 24071 León, Spain

Abstract: The application of statistics in business is essential in order to make decisions in a rigorous and reliable way. One of the fields where forecasting methods are important focuses on business failure. In a comparative study, discriminant analysis and logistic regression are applied on a sample of small and medium-sized firms with head offices in Castilla y León (Spain) in order to predict business failure using a set of financial ratios as independent variables to enter the corresponding models. The achieved results show that there are some differences in the variables becoming significant in each method, but factors related to resources generation are common to both. The classification results reveal that the two methods are appropriate to predict business failure, but logistic regression turns out to be somewhat better, since the percentages of correctly classified firms are higher.

Keywords: forecasting; statistical methods; comparison; discriminant analysis; logit analysis; logistic regression; Spain; financial ratios; business failure; bankruptcy; firm classification; failure prediction; statistics; small and medium-sized enterprises; SMEs; SME failure; resource generation.

DOI: 10.1504/GBER.2015.066534

Global Business and Economics Review, 2015 Vol.17 No.1, pp.76 - 92

Received: 03 Apr 2013
Accepted: 27 Sep 2013
Published online: 24 Dec 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article