35 Years of studies on business failure: an overview of the classic statistical methodologies and their related problems

Abstract

Over the last 35 years, the topic of business failure prediction has developed to a major research domain in corporate finance. A gigantic number of academic researchers from all over the world have been developing corporate failure prediction models, based on various modelling techniques. The ‘classic cross-sectional statistical' methods have appeared to be most popular. Numerous ‘single-period' or ‘static' models have been developed, especially multivariate discriminant models and logit models. As to date, a clear overview and discussion of the application of the classic cross-sectional statistical methods in corporate failure prediction is still lacking, this paper extensively elaborates on the application of (1) univariate analysis, (2) risk index models, (3) multivariate discriminant analysis, and (4) conditional probability models, such as logit, probit and linear probability models. It discusses the main features of these methods and their specific assumptions, advantages and disadvantages and it gives an overview of a large number of academically developed corporate failure prediction models. Despite the popularity of the classic statistical methods, there have appeared to be several problems related to the application of these methods to the topic of corporate failure prediction. However, in the existing literature there is no clear and comprehensive analysis of the diverse problems. Therefore, this paper brings together all criticisms and problems and extensively enlarges upon each of these issues. So as to give a clear overview, the diverse problems are categorized into a number of broad topics: problems related to (1) the dichotomous dependent variable, (2) the sampling method, (3) non-stationarity and data instability, (4) the use of annual account information, (5) the selection of the independent variables, and (6) the time dimension. This paper contributes towards a thorough understanding of the features of the classic statistical business failure prediction models and their related problems.