The Moderating Effect of Professional Skepticism on the Relationship Between Tone at the top & Management’s Accountant Fraud Prevention and Detection

Purpose – This study examines the moderating role of professional skepticism in the relationship between tone at the top and fraud prevention and detection among Management Accountants (MAs) in Malaysia. Additionaly, it explores whether behavior that may help MAs detect fraud more successfully are linked to higher degrees of skepticism with working experience and holding professional licence as moderating effect. Design/methodology/approach – A total of 237 Management Accountants Malaysia participated in an audit judgement experiment, which required them to indicate their beliefs about the tone of the top at their company (poor or good) and evaluate the ability of Managements Accountant’s to detect fraud. This study is grounded in Attribution Theory as the underpinning theory and this quantitative study employed Partial Least Squares (PLS) analysis to analyze the data. PLS-SEM is selected over Covariance-Based SEM (CB-SEM) due to its superior suitability for predictive modeling and its ability to effectively manage moderation effects in complex structural models. Findings – The study reveals that professional skepticism significantly enhances fraud prevention and detection among MAs, particularly for less experienced accountants and those holding professional qualifications. However, professional skepticism does not moderate the relationship between tone at the top and fraud prevention and detection. Originality/value – This research contributes to the limited literature on Management Accountants in Malaysia, specifically addressing how professional skepticism influences the impact of tone at the top on fraud prevention and detection. By incorporating a theoretical framework and a robust SEM analysis, this research provides novel insights into the behavioral mechanisms underlying fraud prevention. Notably, this study is the first to employ PLS-SEM in this context, providing a more comprehensive and robust analysis of the factors affecting fraud detection. Future studies could expand this investigation to include MAs from diverse geographical contexts and explore additional moderating variables for broader applicability.