The prediction of bankruptcy is the art of predicting bankruptcy and various measures of financial distress of public enterprises. It is an extensive area of finance and accounting research. The importance of the area is due in part to the relevance to creditors and investors in evaluating the likelihood that a company may go bankrupt. The amount of research also depends on the availability of the data: for public companies that have declared bankruptcy or not, numerous accounting ratios can be calculated that can indicate danger and numerous other possible explanatory variables. Consequently, the area is suitable for testing increasingly sophisticated and data intensive forecasting approaches.
The history of bankruptcy prediction includes the application of numerous statistical tools that gradually became available, and involves deepening the appreciation of several dangers in the first analysis. Interestingly, research is still published that suffers the traps that have been understood for many years.
Bankruptcy prediction has been the subject of formal analysis since at least 1932, when FitzPatrick published a study of 20 pairs of signatures, one failed and one survived, matched by date, size and industry in The Certified Public Accountant. He did not perform statistical analysis as is common today, but he carefully interpreted the ratios and trends of the ratios. His interpretation was indeed a complex, multi-variable analysis.
In 1967, William Beaver applied t-tests to assess the importance of individual accounting ratios within a similar paired-like pairing. In 1968, in the first formal analysis of multiple variables, Edward I. Altman applied multiple discriminant analysis within a paired sample. One of the most prominent early models of bankruptcy prediction is the Altman Z-score, which still applies today. In 1980, James Ohlson applied logit regression on a much larger sample that did not imply matching-matching.