DSpace Repository

A model to predict corporate failure with particular reference to the Gambler's ruin approach : an empirical study

Show simple item record

dc.contributor.author Silvis J en
dc.date.accessioned 2016-09-22T09:16:39Z
dc.date.available 2016-09-22T09:16:39Z
dc.date.submitted 1978 en
dc.identifier.uri http://hdl.handle.net/20.500.11892/52579
dc.description.abstract Bankruptcy, insolvency and failure are subjects of interest to managers, investors and auditors alike. Prediction of impending failure is obviously of great value, yet little attention is given to the subject in the financial literature available. In this paper an attempt is made to enlarge upon the subject. The emphasis is on the prediction of corporate failure and the developments in the construction of models are discussed to put the topic in perspective. The model advocated in this paper is split into two sections. The non-accounting data is used to develop a corporate personality or image, whilst the accounting data is used as input in a gambler's ruin model. The non-accounting data centres around identification of mismanagement which is regarded as the starting point of the process of failure. Mismanagement is identified by unbalanced decision making, lack of management information and inability of management to respond to changes in the environment. The symptoms of failure are discussed and the total of non-accounting data is used to develop a failure or non-failure company image. This failure or non-failure image is supplemented by accounting data in constructing the model to predict failure. The gambler's ruin model to predict failure was selected as the most appropriate tool to represent the accounting data in this model. The model is a statistical model based on the Markovian chain process where P (R) = probability q = probability of failure p = probability of success y = the number of states away from failure The above parameters are approximated by accounting data available in the annual financial statement and a sample of companies quoted on the Johannesburg Stock Exchange was selected to test the model empirically. Two types of mistakes were identified. The first mistake (type 1) is the instance where companies were classified as "failure" companies by the model but which were still in existence at the time of writing this report. Type 2 mistake is where companies were classified as "solvent" by the model but which went bankrupt subsequently. Despite the misclassification the proportion correctly classified as solvent companies for years 1-5 before failure is .92, .69, .85, .64 and .77 whilst the proportion is .77, .69, .69, .64 and .89 for the failure companies. The companies misclassified were further studied and the possible reasons for misclassification sought. Some of these reasons gave insight and justification for the misclassification which makes the results of the empirical test even more remarkable. The practical implications of the results for the financial manager, investor and auditor are discussed and a few suggestions are made as to how the model can be improved. In conclusion it can be said that the model proved that accounting data has definite predictive value in distinguishing between failure and non-failure companies. This could be done up to five years prior to failure but success was lower as time between data and failure increased. en
dc.language English en
dc.subject Business administration / Business leadership en
dc.subject Models for decision making and operational research en
dc.title A model to predict corporate failure with particular reference to the Gambler's ruin approach : an empirical study en
dc.type Masters degree en
dc.description.degree MBA en


Files in this item

Files Size Format View

There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record