Multiple classification schemes for signalling corporate collapse

Abstract

Purpose – The purpose of this paper is to undertake an analysis of two recent classification schemes in the literature for ratio-based modelling of corporate collapse; namely the dual-classification scheme (DCS) and the multi-classification scheme (MCS). Its contribution to the literature lies in investigating whether a trade-off exists between the structural efficiency and the practical adeptness of these two schemes. Design/methodology/approach – The methodological approach for the DCS relies on a combination of multiple discriminant analysis and multi-level modelling, whereas that for the MCS utilizes multi-classification constrained-covariance regression analysis. Findings – Based on a unified data set of 112 collapsed companies and 341 non-collapsed companies utilised across both the DCS and the MCS, the results indicate that whilst both classification schemes are comparable in their predictive accuracy with respect to signalling collapse, they exhibit a trade-off between their structural efficiency and their practical adeptness. Originality/value – Whilst novel classification schemes such as the DCS and the MCS have been successful in addressing the inherent problem of identifying unclassifiable companies in the literature for ratio-based modelling of corporate collapse, thus far no attempt has been made to investigate the trade-off between their structural efficiency and their practical adeptness. Moreover, by utilising a unified data set, the robustness of this investigation is enhanced. Accordingly, this paper provides economic insight into more stable financial modelling.