Fraudulent Financial Reporting in the Banking Sector of Bangladesh: A Prediction

The purpose of this study is to predict the areas in financial statements susceptive to fraud in the banking sector of Bangladesh. Data of ۱۳ years ranging from ۲۰۰۶ to ۲۰۱۸ of ۲۹ listed banks in Bangladesh were examined for the purpose of this study. Financial data suggested by International Standard on Auditing (ISA) ۲۴۰ as fraud risk indicators were used as the independent variables and banks identified by Centre for Policy Dialogue (CPD) to be engaged in fraud, scam and heists were taken as dependent variable. Multilayer Perceptron Network (MLP), a class of feedforward Artificial Neural Network (ANN) model was used as the analytical tool. It is found that loan disbursement, assets, profit, operating expenses and tax are the areas that can signal the probable fraud in financial statements of the listed banks of Bangladesh. The findings of this study will have policy implications for auditors and the regulators of money market in Bangladesh.

Keywords:

Fraud , fraudulent financial reporting , Artificial neural network (ANN)