Abstract. In the field of credit risk management, the calculation of the probability of default of companies plays a key role. For that reason, bankruptcy prediction of companies has generated extensive research in the past decades. This paper applies one of the most popular techniques, the logistic regression. This technique is extensively used both by professionals and academics and is employed in many studies as a benchmark. Here we will apply it on a vast data base of the Spanish companies and a statistical analysis of the robustness of the model will be undertaken, with very satisfactory results.JEL classification: G32, G33.