Multivariate techniques in corporate failure prediction

Abstract

This thesis introduce Multidimensional Scaling (MDS) to explain and predict the corporate failure event.  After critically reviewing the relevant literature, it is argued that refinements of existing models, which are mostly mechanical, and the construction of new ones using similar techniques, are likely to yield nothing more than marginal benefits to the users.  However, users will benefit from the development of statistically sound but easy-to-comprehend models, which will assist them in making an informed decision regarding the financial health of the companies in question.  Scaling techniques enable the construction of robust and user-friendly bankruptcy models as they offer pictorial representations of the results, via which the user can visualise the underlying relationships in the data. 

The research begins with the application of logit analysis and neural networks (NNs), and the investigation of the usefulness of operating cash flows in bankruptcy models, reflecting the recent popularity of NNs, the increased interest in cash flow reporting in the UK and the scarcity of recent UK failure studies.  The NN model is not found to be superior to the logit model and empirical evidence shows that operating cash flow variables can add significantly to the explanatory power of accrual-based models.

The second study of the thesis proposes classic MDS as an alternative to conventional techniques.  The MDS approach has the advantage of visualising the main features of the data in the form of statistical maps, which lend themselves to intuitive interpretation, while keeping a strong theoretical basis.  The methodology is demonstrated using a recent sample of UK companies, while a future-dated holdout sample is employed to illustrate how MDS can aid practitioners when assessing the financial health of a company.

Three-way scaling analysis is introduced in the third study to examine the differences in the financial structure of a large sample of UK companies over a number of years.  The results indicate that the most important determinants of a company’s financial health are its capital structure, the profitability/operating cash flow position and the ratios that relate to the shareholders’ return.

More information

Identifiers

Catalogue record

Export record

ASCII CitationAtomBibTeXData Cite XMLDublin CoreDublin CoreEP3 XMLEndNoteHTML CitationHTML CitationHTML ListJSONMETSMODSMPEG-21 DIDLOpenURL ContextObjectOpenURL ContextObject in SpanRDF+N-TriplesRDF+N3RDF+XMLRIOXX2 XMLReferReference ManagerSimple Metadata