The Role of Apparent Signs of Financial Distress in Test Samples and Verification Samples of Bankrupt Models

Abstract

Financial bankrupt models are characterized as quite accurate and above all very fast tools for quantitative evaluation of financial health of company. The creators report the accuracy of the predicted bankruptcy usually in the range of 70 to 90%. But the problem of bankruptcy models is the test sample on which the models were created. The sample affects the predictive power of these models. Usually indicated accuracy rate differs from the real predictive power of these models. The financial distress of certain businesses may be obvious even without the use of bankruptcy models. Apparent signs of financial distress may be insolvency, negative equity, VAT unreliability, negative economic result for several years in a row. Survey conducted by more than 270 companies has shown that more businesses with apparent signs of financial distress in the sample increase the reported accuracy of the bankruptcy model. The research carried out also has determined the real accuracy of selected bankruptcy models on the standard sample of Czech firms and also on a sample of companies where companies with obvious signs of financial distress were eliminated. Due to the modification of the test sample subsequently the accuracy of the selected models changed radically.