Firms’ Default – from Prediction Accuracy to Informational Capacity of Predictors

Research background: Bankruptcy literature is populated with scores of (econometric) models ranging from Altman’s Z-score, Ohlson’s O-score, Zmijewski’s probit model to k-nearest neighbors, classification trees, support vector machines, mathematical programming, evolutionary algorithms or neural networks, all designed to predict financial distress with highest precision. Purpose of the article: We believe corporate default is too an important research topic to be identified with the prediction accuracy only. Despite the wealth of modelling effort, a unified theory of default is yet to be proposed. Due to the disagreement, both on the definition and hence the timing of default as well as on the measurement of prediction accuracy, the comparison (of predictive power) of various models can be seriously misleading. The purpose of the article is to argue for the shift in research focus from maximizing accuracy to the analysis of the information capacity of predictors. By doing this, we may yet come closer to understand default itself. Methodology/methods: We have critically appraised the bankruptcy research literature for its methodological variety and empirical findings. Default definitions, sampling procedures, in and out-of-sample testing and accuracy measurement have all been scrutinized. We believe the bankruptcy models currently used are, using the language of Feyerabend, incommensurable. Findings: Instead of what we call the population of models paradigm (the comparison of predictive power of different models) prevailing today, we propose the model of population paradigm, consisting in the estimation a single unified default forecasting platform for both listed and non-listed firms, and analyze the marginal contribution of the different information sources. In addition to classical corporate financial data, information on both firm's strategic position and its macroeconomic environment should be studied.