Abstract
In a general sense, bankruptcy prediction modelling based on accounting numbers is quite straightforward. The main ingredients in a statistical approach to estimate such models consist of a sample of bankrupt and non-bankrupt companies, accounting numbers from financial statements prior to the bankruptcy event, and some statistical model allowing for a stochastic relationship between the dependent variable (‘bankruptcy’ versus ‘non-bankruptcy’) and the accounting numbers. However, as most researchers in the area have experienced, several more or less open-ended questions soon tend to surface. One of these questions, concerning the choice and specification of the independent variables, is re-addressed in the article by Peat (2007). This is clearly a worthwhile and legitimate purpose, in particular as the author has chosen a new ‘managerial decision based approach’ in order to provide a well-founded rationale for the suggested variables. In principle, the chosen methodology holds promise for the derivation of something as unique as a theory-based probabilistic bankruptcy prediction model.