Assessing Predictive Power and Earnings Manipulations. Applied Study on Listed Consumer Goods and Service Companies in Ghana Using 3 Z-Score Models

This study uses data from Ghana’s public listed consumer goods and service companies for the period of 2014-2018 to test the predictive power of Altman’s (2000), Taffler’s (1983), and Beneish’s (1999) models in detecting bankruptcy and Earnings Manipulation. Prediction power (accuracy) was tested for two Z-Score models: Altman’s (2000) revised model and Taffler’s (1983) model. All two models were found to have significant predictive power. Altman’s revised model was found to be accurate for listed consumer firms in Ghana at a predictive power rate of 66%. Taffler’s (1983) Z-Score model was equally found to be accurate for prediction at a higher predictive power of 88%. The Taffler (1983) model has a higher predictive power compared with Altman’s (2000) revised model. The Beneish (1999) model also revealed that the financial statements of the industry were manipulated at a different degree. The study recommends that stakeholders would be better protected when the three models are deployed simultaneously as an important part of an Audit engagement. Also, Altman’s (2000), Taffler (1983), and Beneish’s (1999) model should be applied in predicting bankruptcy and financial statement fraud evaluation in the banking and mining sectors in Ghana taking into account the frequent collapse, mergers, and acquisitions that do occur.