Title
Financial Consequences of Fraud in Amman Stock Exchange Firms
Authors
Abstract
Purpose: This study investigates the impact of financial fraud risk, proxied by the Beneish M-Score, and key macroeconomic variables on the financial performance (Return on Assets – ROA) of firms listed on the Amman Stock Exchange (ASE).
Methodology: Employing panel data from 140 ASE-listed firms between 2015 and 2020, the research utilizes Ordinary Least Squares (OLS) regression and several machine learning regression models (Support Vector Machines, Random Forest, Gradient Boosting). The analysis examines the influence of the Beneish M-Score, GDP growth, inflation, and company size on return on assets (ROA).
Findings: The results reveal a significant positive impact of GDP growth and firm size on ROA. While inflation’s linear effect was insignificant, we uncovered a compelling non-linear, inverted U-shaped relationship between the Beneish M-Score and ROA. This suggests that while moderate levels of earnings management risk may coincide with performance-enhancing activities, higher levels are unequivocally detrimental. Notably, machine learning models, particularly Random Forest, demonstrated superior predictive accuracy over traditional OLS regression, underscoring the importance of capturing these non-linear dynamics.
Recommendations: Jordanian firms are advised to strengthen internal controls and foster transparent financial reporting. Regulators should enhance oversight and consider advanced analytical tools, including machine learning, for risk assessment. Investors should critically evaluate fraud risk indicators, recognizing their complex impact on performance.
Originality: This study offers novel insights into the nonlinear performance implications of financial fraud risk in an emerging market context (Jordan). It distinctively integrates macroeconomic factors and compares traditional econometric techniques with machine learning approaches, contributing to the financial fraud literature in developing economies by highlighting the complex dynamics between earnings manipulation risk and firm performance. This study contributes to the field of Decision Sciences by demonstrating how hybrid econometric-ML models can enhance fraud risk assessment and corporate decision-making in developing economies.
Keywords
Financial fraud, financial performance, Beneish M-Score, Regression analysis, Amman Stock Exchange.
Classification-JEL
G33, G34 , M41
Pages
83-111
How to Cite
AlQudah, A., Jaradat, S., & Al-Haddad, L. (2025). Financial Consequences of Fraud in Amman Stock Exchange Firms. Advances in Decision Sciences, 29(1), 83-111.