DEA as a business failure prediction tool Application to the case of galician SMEs

Abstract

In the research group we are working to provide further empiricalevidence on the business failure forecast. Complex fittingmodelling; the study of variables such as the audit impacton business failure; the treatment of traditional variables andratios have led us to determine a starting point based on a referencemathematical model. In this regard, we have restrictedthe field of study to non-financial galician SMEs in order todevelop a model1 to diagnose and forecast business failure. Wehave developed models based on relevant financial variablesfrom the perspective of the financial logic, voltage and financialfailure, applying three methods of analysis: discriminant,logit and multivariate linear. Finally, we have closed the firstcycle using mathematical programming –DEA or Data EnvelopmentAnalysis– to support the failure forecast. The simultaneoususe of models was intended to compare their respectiveconclusions and to look for inter-relations. We can say that theresulting models are satisfactory on the basis of their capacityfor prediction. Nevertheless, DEA contains significant pointsof criticism regarding its applicability to business failure.