Many studies of business failure prediction have been carried out in the last thirty years or so. Prediction is based on an economic and financial analysis of failing and non-failing firms with the aim of determining the variables, mainly of an accounting nature, that best distinguish the two categories. The purpose of this review is to assess the relative effectiveness of the different classification methods employed. To that end we identify the common elements while highlighting the different ways in which they are applied empirically. We describe the principle of the available techniques and compare their performance, focusing non-exhaustively on both the founding studies and on more recent research.