FIRM FINANCIAL DISTRESS PREDICTION WITH STATISTICAL METHODS: PREDICTION ACCURACY IMPROVEMENTS BASED ON THE FINANCIAL DATA RESTATEMENTS

Firm failure phenomenon has been in the focus of academic research for many years. Pioneering work of Beaver and Altman triggered many papers, which were trying to explain and predict firm failure. However, developed failure models entirely relay on original financial data and do not take into account potential data problems resulting from accounting manipulations. However, empirical research and anecdotal evidence (Enron, WorldCom, Parmalat, etc.) shows that accounting data can be manipulated. In this paper authors proposed the model for restatement of financial statements and tested it on the sample of 345 firms from Croatia. Empirical testing has shown that usage of restated financial data increases overall failure prediction accuracy by 5.3 percentage points. In the segment of non-distressed firms prediction accuracy was increased by 10.4 percentage points, while in the segment of distressed firms prediction accuracy was increased by 1.5 percentage points. Such findings indicate that accounting manipulations can affect failure prediction accuracy and that proposed model can be useful for prediction accuracy improvements.