This study analyses the comparison between a traditional statistical methodology for bankruptcy classification and prediction, i.e. linear discriminant analysis (LDA), and an artificial intelligence algorithm known as Genetic Algorithm (GA). The study was carried out at Centrale dei Bilanci, in Turin, Italy, analysing 1920 unsound and 1920 sound industrial Italian companies from 1982–1995. This paper follows our earlier examination of neural networks (NN) (see Altman et al., 1994. Corporate distress diagnosis: Comparisons using discriminant analysis and neural network. Journal of Banking and Finance XVIII, 505–529). The experiments on GA were oriented along two different lines: the genetic generation of linear functions and the genetic generation of scores based on rules. The two types of experiments showed GA to be a very effective instrument for insolvency diagnosis, even if the results obtained with LDA analysis perhaps proved to be superior to those obtained from GA. Of particular interest, it should be noted that the results of GA were obtained in less time and with more limited contributions from the financial analyst than the LDA. Of additional interest is the relevance for credit risk management of financial institutions.