DETECTION AND PREDICTION OF MANAGERIAL FRAUD IN THE FINANCIAL STATEMENTS OF TUNISIAN BANKS

This article models the detection and prediction of managerial fraud in the financial statements of Tunisian banks. The methodology used consist of examining a battery of financial ratios used by the Federal Deposit Insurance Corporation (FDIC) as indicators of the financial situation of a bank. We test the predictive power of these ratios using logistic regression. The results show that we can detect managerial fraud in the financial statements of Tunisian banks using performance ratios three years before its occurrence with a classification rate of 71.1%.