Page 29 - Azerbaijan State University of Economics

THE JOURNAL OF ECONOMIC SCIENCES: THEORY AND PRACTICE, V.82, # 1, 2025, pp. 19-35

Practical application of the Beneish and Roxas models simultaneously to identify
signs of financial statement falsification among 50 marketing Ukrainian companies
for the period 2023-2024 revealed four distinct patterns of concordance and
divergence in the assessment results. The comparison of actual and threshold M-Score
values across both models enabled a more accurate confirmation of the reliability of
financial statements for 3 companies (6% of the sample), and identified potential
manipulations in 30 companies (60%). These findings indicate a high level of risk
related to financial statement manipulation within the marketing sector. Discrepancies
between the models were found in 17 companies; in particular, in 14 cases
manipulation was detected exclusively by the Beneish model, while in 3 cases only
the Roxas model indicated the presence of potential manipulations.
Based on the conducted research, the following recommendations can be formulated
for different groups of stakeholders. Audit entities should conduct an in-depth audit
of marketing companies with extreme Beneish M-Score values or a high probability
of manipulation, determined based on the Roxas M-Score. Since 60% of the studied
sample have a high level of risk, it is necessary to improve audit procedures by
detailed verification of contracts with customers for payment terms, analysis of the
correlation between revenue and cash flows, and assessment of the reasonableness of
marketing expenses. Capital borrowers are recommended to generally avoid
interaction with marketing companies that have a high risk of manipulation,
confirmed simultaneously by the Beneish M-Score and the Roxas M-Score.
Management of marketing companies with a high level of potential financial
statement manipulation should regularly monitor indicators that indicate
manipulation. In particular, avoid unjustified payment delays, reduce the share of
accruals in profit, and control a high level (more than 5% of the value of assets) of
cash flows from transactions.

REFERENCES
Achakzai, M. A. K., & Peng, J. (2023), “Detecting financial statement fraud using
dynamic ensemble machine learning”, International Review of Financial Analysis,
89, 102827.

Alkaraan, F., Albahloul, M., Abdoush, T., Elmarzouky, M., & Gulko, N. (2024), “Big
Four ‘rhetorical’ strategies: Carillion’s collapse”, Journal of Accounting and
Management Information Systems, 23(2).

Aqilah, N., Mohammed, N. F., & Kamaluddin, A. (2021), “Application of Beneish
M-Score model in detecting probable earnings manipulation in Malaysian public
listed companies”, IJBE (Integrated Journal of Business and Economics), 5(1), 86.
Beneish, M. D. (1999), “The detection of earnings manipulation”, Financial Analysts
Journal, 55(5), 24-36.

29