ArticlePDF Available

Application of Beneish M-Score Models and Data Mining to Detect Financial Fraud

November 2015
Procedia - Social and Behavioral Sciences 211:924-930

DOI:10.1016/j.sbspro.2015.11.122

License
CC BY-NC-ND 4.0

Authors:

Universitas Trunojoyo Madura

This study aims to analyze the ability of m-score Beneish in detecting financial fraud. This research data refer to companies that commit fraud according to the fraud Database of Sanctions of Issuer Cases Public Companies that was released by the Financial Services Authority in the period of 2001-2014. The results showed that overall Beneish m-score model was capable to detect financial fraud. Gross margin index, depreciation index, index of sales and general administrative burden and total accruals were all significant in detecting financial fraud. Sales index, asset quality index, and leverage index was statistically not significant in detecting financial fraud.

. Logistic Regression Test Results

…

Figures - uploaded by Kashan Pirzada

Content may be subject to copyright.

Content uploaded by Kashan Pirzada

Content may be subject to copyright.

Available via license: CC BY-NC-ND 4.0

Content may be subject to copyright.

Procedia - Social and Behavioral Sciences 211 ( 2015 ) 924 – 930

Available online at www.sciencedirect.com

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the Organizing Committee of the 2nd GCBSS-2015

doi: 10.1016/j.sbspro.2015.11.122

ScienceDirect

2nd Global Conference on Business and Social Science-2015, GCBSS-2015, 17-18 September

2015, Bali, Indonesia

Application of Beneish M-Score Models and Data Mining to Detect

Financial Fraud

Tarjoa

, Nurul Herawatib

a,b University of Trunojoyo Madura, Raya Telang Street PO BOX 2, Bangkalan Madura 60192 - Indonesia

Abstract

This study aims to analyze the ability of m-score Beneish in detecting financial fraud. This research data refer to companies that

commit fraud according to the fraud Database of Sanctions of Issuer Cases Public Companies that was released by the Financial

Services Authority in the period of 2001-2014. The results showed that overall Beneish m-score model was capable to detect

financial fraud. Gross margin index, depreciation index, index of sales and general administrative burden and total accruals were

all significant in detecting financial fraud. Sales index, asset quality index, and leverage index was statistically not significant in

detecting financial fraud.

Peer-review under responsibility of the Organizing Committee of the 2nd GCBSS-2015.

Keywords: Ability; Detecting; financial; fraud, Beneish M-Score

1. Introduction

Based on research conducted by the Association of Certified Fraud Examiners (ACFE) in 2014, the most

disadvantage fraud in a row is statement of financial fraud by 73%, corruption by 18%, and asset misappropriation by

9%. Therefore, it can be stated that there are three (3) types of fraud. The financial statement fraud is the most harmful.

Various studies to detect financial fraud have often been committed by researchers with fraud triangle analysis. In

the last few years, some researchers have been using Beneish m-score model (Beneish, 1999, 2012), Dimitriois (2014),

data mining (Zaki & Theodoulidis, 2013), Ata & Seyrek (2009) on this problem.

* Corresponding author. Tel.: +6287853035508

E-mail address: tarjo@trunojoyo.ac.id

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the Organizing Committee of the 2nd GCBSS-2015

925

Tarjo and Nurul Herawati / Procedia - Social and Behavioral Sciences 211 ( 2015 ) 924 – 930

Research to detect financial fraud has been conducted by some researchers using financial ratios approach, data

mining, and Beneish m-score model, but the method has drawbacks. First, there is a misclassification. Second,

mistakes in data mining may lead to misclassification, this is due to the fact that different industries have different

financial reporting requirements. Therefore, the research problem is “can the application of m-score model detect

financial fraud?”

The results of this study will be very helpful for auditors in conducting audit anyone with an interest in company's

financial statements. M-score models and data mining can be used as an early indication (red flags) for detecting

financial fraud.

2. Detecting Financial Fraud

According to the ACFE (2014), financial statement fraud is a deliberate fraud performed by a manager or employee

with no reports on actual financial statement information, for example, fictitious revenues, too low expen ses report,

and so forth. In financial reporting context, an audit is designed to provide assurance to interested parties that the

company's financial statements are not influenced by a material misstatement and gives adequate confidence to the

management accountability of the assets (Koroy, 2008). Public companies and private sector have an opportunity for

profit manipulation. Tarjo (2008; 2010) confirms that the go public company, especially the manufacturing industry,

can manipulate profits. While Pradana, Tarjo & Kompyurini (2013) show that there are various factors which can be

used to detect fraud in the public sector organizations.

Who are in charge to perform financial fraud detection? Alleyne & Howard (2005) explores perceptions of auditors

and users that it is auditor's responsibility to uncover fraud, performance-related audit procedures, the nature and

extent of fraud in Barbados, as well as auditor’s response since Enron’s case came to light. The research findings of

Alleyne & Howard (2005) show that there has been a big gap in expectations auditors perceive fraud detection as the

responsibility of management, while users and management do not agree on this. Some research suggests that

prevention and detection of fraud is management responsibilities, such as Beasley (1996) finding that companies with

no fraud has board members from outside the company with a greater percentage than from inside the company. On

the other hand, subsequent studies showed that auditors are the subject of analysis of studies on fraud detection (such

as Owusu-Ansah, Moyes & Oyelere, 2002; Running (2009); Bishop 2004; Hutomo (2012); Mariana (2014); Rustiarini

& Novitasari, 2014; Gamar, & Djamhuri 2015; Lisic, Silveri, Song & Wang, 2015). The role and demographic factors

of auditors have become the subject in research on the detection of fraud.

How can we detect financial fraud? It has been a lot of research on fraud detection. The results of investigations by

Bishop (2004) to 450 companies, which are victims of fraud in 2004 years since the ACFE report, on how fraud was

initially detected, indicate that there are several ways of doing so, such as internal audit, tip form employees, by

accident, internal control, external audit, tip from customers, tip from vendors, anonymous tip, and notified by police.

Bishop (2004) states that fraud was most frequently detected by internal audit (23.8%) while the information of

workers is the second most common detection method (23.6%). However, if all categories of information (employees,

vendors, consumers, and anonymous) are combined, it becomes far and the most common detection method, 39.6%.

These results indicate that the facilities / services to report violations provide the opportunity and encouragement to

report the potential violation.

Other studies on reliability of financial ratios in detecting financial fraud are by Persons (1999); Spathis (2002);

Kaminski, Wetzel & Guan (2004). Owusu-Ansah, Moyes & Oyelere (2002) does a study on fraud detection through

effectiveness of audit procedures. The detection of fraud through control is conducted by Alleyne & Howard (2005);

Bierstaker, Brody & Pacini (2006); Gottschalk (2011); Akbar, Tarjo & Carolina (2012) & Kummer, Sing & Best

(2015). There are also financial fraud detection model approaches, such as by Spathis (2002), developing reliable

models in the detection of False Financial Statements or FFS for the Greek company with ten financial variables

chosen for testing as a predictor of potential FFS. Skousen & Twedt (2009) use a fraud score model, as determined

by Dechow, to determine the possibility of manipulation on financial reports and found that the F-Score is an indicator

of the risk of fraud, but not a direct signal that there is no fraud. Glancy & Yadav (2011) proposed fraud detection

computational models / CFDM to detect fraud in financial reporting. CFDM uses a quantitative approach on textual

data.

926 Tarjo and Nurul Herawati / Procedia - Social and Behavioral Sciences 211 ( 2015 ) 924 – 930

In its development, data mining can be used to detect financial fraud of companies. According to Turban, Aronson

& Liang (2005), data mining is a semi-automatic process that uses statistical techniques, mathematics, artificial

intelligence, and machine learning to extract and identify potential knowledge and useful information stored in a large

database. There are several definitions of data mining.

3. Research Methodology

3.1 Data

This research data was company’s fraud financial statements issued by the Capital Market Supervisory Agency or

Services Authority Finance (FSA) in the period of 2001-2014. For comparison, the financial data of companies that

committed no fraud of the period of the year and of the same industries was used.

3.2 Beneish M-Score

Beneish M-Score is a method that can be used to detect companies with a tendency to commit fraud on their

financial statements (Beneish, 2012). Empirically, companies with higher M-Score have higher tendency to commit

fraud. Beneish M-Score is a probabilistic model, so that one of the limitations is that the ability to detect fraud is not

with 100% accuracy. The formula of Beneish M-Score is as follows:

M = -4.840 + 0,920XDSRI + 0,528XGMI + 0,0404XAQI + 0,892XSGI + 0,115XDEPI-0,172XSGAI + 4,679X \

TATA- 0,327XLVGI

If the M-score> -2.22, it shows indications of financial fraud within companies.

3.3 Data Mining

Beneish Model M-Score is used for data mining of fraud-committed companies. According to Turban et al. (2005),

data mining is the semi-automatic process that uses statistical techniques, mathematics, artificial intelligence, and

machine learning to extract and identify potential knowledge and useful information stored in a large database. Data

mining technique used in this research is Logit Regression. Data mining begins with testing the Principle Component

Analysis (PCA) to variable of Beneish M-Score model. PCA is used to determine what variables of M-Score Beneish

model that is predictive of financial fraud. The Logit Regression model of this study are:

FRAUD = β0 + β1DSRI + β2GMI + β3AQI + β4SGI + β5DEPI + β6SGAI + β7TATA + β8LVGI + εi

Where:

FRAUD= Dummy Variable (1 for fraud-committed companies and 0 for non-fraud-committed companies)

DSRI = Sales Index

GMI = Gross Margin Index

AQI = Asset Quality Index

SGI = Sales Growth Index

DEPI = Depreciation Index

SGAI = Sales and General Administration Expenses Index

TATA = Total Accrual

LVGI = Leverage Index

εi = Residual

927

Tarjo and Nurul Herawati / Procedia - Social and Behavioral Sciences 211 ( 2015 ) 924 – 930

4. Research Result

This research data was company’s fraud financial statements issued by the Capital Market Supervisory Agency or

Services Authority Finance (FSA) in the period of 2001-2014. After tabulation, we obtained 35 companies that

committed fraud. For comparison, 35 companies that did not commit fraud were selected based on the equality of

assets and same category of industry. Table 1 below is descriptive statistics of the Beneish M-Score.

Tabel 1. Descriptive Statistics

Variable

Minimum

Maximum

Mean

Std. Deviation

Fraud

Non Fraud

Fraud

Non Fraud

Fraud

Non Fraud

Fraud

Non Fraud

DSRI

.000

.103

15.845

3.036

1.367

1.111

2.576

.470

GMI

-81.180

-18.928

48.335

26.811

.601

.902

16.349

6.517

AQI

.057

.027

14.432

4.478

1.533

1.035

2.445

.819

SGI

.320

.515

4.286

11.431

1.249

1.379

.722

1.771

DEPI

-728.537

.002

1.732

1.434

-21.281

.923

123.117

.232

SGAI

.158

.079

5.570

1.595

1.071

1.070

.834

.287

LVGI

.317

.100

12.737

1242.494

1.376

36.453

2.004

209.854

TATA

-4.363

-.975

.544

.568

-.365

.003

1.018

.270

TOTAL

0.008

-21.360

270.288

465.848

10.660

12.313

45.871

79.066

Table 1 shows that DSRI value for companies committing fraud had the lowest value of 0.000, the highest of

15.845, and the average value of 1.367, and a standard deviation of 2.576. While the non-fraud companies had a lowest

value of 0.103, the highest of 3.036, the average value of 1.111 with a standard deviation of 0.470.

Table 1 also shows that GMI value for companies committing fraud had the lowest value of -81.180, the highest

of 48.335, the average value of 0.601, and a standard deviation of 16.349. While the non-fraud companies had the

lowest value of -18.928, the highest of 26.811, the average of 0.902 with a standard deviation of 6.517.

The test results also showed that DEPI for companies committing fraud had the lowest value of -728.537, the

highest of 1.732, the average of -21.281, and a standard deviation of 123.117. While the non-fraud companies had a

lowest value low of 0.002, the highest of 1.434, the average value of 0.923, and a standard deviation of 0.232.

The next step was to test the Principle Component Analysis (PCA) to variables of Beneish M-Score model. Based

on the Principle Component Analysis (attachment 1), the SGI component was not used for explanatory factor because

it had a negative value, so the variables on Beneish M-Score Model used to detect fraud were DSRI (Sales Index),

GMI (Gross Margin Index), AQI (Asset Quality Index), DEPI (Depreciation Index), SGAI (Sales and General

Administration Expenses Index), TATA (Total Accrual), and LVGI (Leverage Index).

In assessing the suitability of this research model, we used the statistic value of Chi-Square Homer and Lemeshow

Goodness of Fit. Based on logistic regression test, as in Table 3, the value of Chi-Square Hosmer and Lemeshow

Goodness of Fit was 5.459 with a significance value of 0.708 (sig> 0.05), this means that the model developed fit the

data used in this study. Therefore, it can be said that the binary models in this study is acceptable and feasible to be

used on further analysis.

Test on the overall model (overall model fit) is determined based on the value of -2 Log Likehood. Based on tests

carried out, the initial value of -2 Log Likehood (Block Number: 0) was 97.041, while the final value of -2 Log

Lihehood (Block Number: 1) was 60.193. These tests showed a decrease in the value of Log Likehood as much as to

36.848, therefore it can be stated that these figures illustrate the suitability of the model used in this study. Nagelkerke

R Square value was 0.546. It shows that the variability of the fraud data can be explained by the independent

variable by 54.6%, while the rest was explained by other variables not used in this study.

928 Tarjo and Nurul Herawati / Procedia - Social and Behavioral Sciences 211 ( 2015 ) 924 – 930

Table 2. Logistic Regression Test Results

Variable

Coefficients

Wald

Sig.

DSRI

0,614

1,945

0,163

GMI

0,984

4,141

0,042

AQI

0,830

2,521

0,112

DEPI

-1,254

3,266

0,071

SGAI

-3,420

5,333

0,021

LVGI

-0,029

1,716

0,190

TATA

-2,592

5,582

0,075

Constant

1,356

1,278

Chi-Square (Hosmer and Lemeshow Test)

5,459

0,708

-2 Likelihood (Block Number: 0)

97,041

-2 Likelihood (Block Number: 1)

60,193

Omnibus Test of Model Coefficients

36,848

0,000

Nagelkerke R Square

0,546

Overall Percentage

78,6%

Where:

* : Significant level 5%

** : Significant level 10%

The data used in the study were 35 companies that committed fraud and 35 non-fraud companies. Based on data

mining test using regression logistic, classification accuracy to detect fraud was 77.1% (27 of 35 companies that

committed fraud). From the 35 non-fraud companies, as many as 28 (80%) were accurately classified as not

committing fraud. However, the interesting finding in this study is that from the 35 companies categorized as non-

fraud, it turned out that 7 of them were then classified as committing fraud.

Table 2 shows that GMI (Gross Margin Index), DEPI (Depreciation Index), SGAI (Sales and General

Administration Expenses Index) and TATA (Total Accrual) influence the detection of financial fraud. This shows that

the GMI, DEPI, SGAI, and TATA can be used to detect financial fraud. DSRI (Sales Index), AQI (Asset Quality

Index), and LVGI (Leverage Index) statistically have no significant effect on the detection of financial fraud. This

shows that DSRI, AQI, and LVGI are not able to detect financial fraud.

5. Conclusion

From the discussion, Beneish M-Score model can be used to detect financial fraud. The better the prospects of the

company, the better it can be used to detect financial fraud. On the other hand, depreciation, cost of sales, and

discretionary accrual on accounting policy may reduce the ability to detect financial fraud. Indication of recognition

on increasing revenue and profit, suspension on cost of the asset, as well as the amount of debt to creditors cannot be

used for detecting financial fraud.

References

Association of Certified Fraud Examiners (ACFE), 2014, Report to The Nation, Austin, Texas.

Akbar, F., 2013. Fraud Pengadaan Barang/Jasa, Proceeding of Conference in Business, Accounting and Management (CBAM), CBAM Vol. 1, No.

929

Tarjo and Nurul Herawati / Procedia - Social and Behavioral Sciences 211 ( 2015 ) 924 – 930

Alleyne, Philmore and Howard, Michael. 2005. An Exploratory Study of Auditors’ Responsibility for Fraud Detection in Barbados. Managerial

Auditing Journal Vol. 20 No. 3, pp. 284-303.

Ata, Ali. And Seyrek, Ibrahim, 2009. The Use of Data Mining Techniques in Detecting Fraudulent Financial Statements: An Application on

Manufacturing Firms, The Journal of Faculty of Economics and Administrative Sciences, Vol. 14, No. 2, pp. 157 -170.

Beneish, Messod D., 1999. The Detection of Earnings Manipulation, Financial Analysts Journal, September/October, pp. 24 -36.

Beneish, Messod D., 2012. Fraud Detection and Expected Return, http://papers.ssrn.com/sol3/papers.cfm? abstract_id=1998387

Beasley, Mark S., 1996. An Empirical Analysis of the Relation Between the Board of Director Composition and Financial Statement Fraud. The

Accounting Review Vol. 71, No. 4, pp. 443-465.

Bierstaker, James L.; Brody, Richard G.; & Pacini, Carl. 2006. Accountants’ Perceptions Regarding Fraud Detection and Prevention Methods.

Managerial Auditing Journal Vol. 21 No. 5, pp. 520-535.

Bishop, Toby J.F., (2004)."Preventing, Deterring, and Detecting Fraud: What Works and What Doesn’t ", Journal of Investment Compliance, Vol.

5 Iss 2 pp. 120 – 127.

Debreceny, Roger S. & Gray, Glen L. 2010. Data Mining Journal Entries for Fraud Detection: An Exploratory Study. International Journal of

Accounting Information Systems 11, pp. 157–181.

Gamar, Nur; dan Djamhuri, Ali. 2015. Auditor Internal Sebagai “Dokter” Fraud Di Pemerintah Daerah. Jurnal Akuntansi Multiparadigma Vol 6,

No 1.

Glancy, Fletcher H. & Yadav, Surya B. .2011. A Computational Model for ﬁnancial Reporting Fraud Detection. Decision Support Systems 50, pp.

595–601.

Gottschalk, Petter. 2011. Prevention of White-Collar Crime: The Role of Accounting. Journal of Forensic & Investigative Accounting, Vol. 3,

Issue 1.

Hutomo, Oki Suryo. 2012. Cara Mendeteksi Fraudulent Financial Reporting Dengan Menggunakan Rasio-Rasio Finansial. Fakultas Ekonomika

Dan Bisnis. Universitas Diponegoro. Semarang.

Kaminski, Kathleen A.; Wetzel, T. Sterling; & Guan, Liming. 2004. Can Financial Ratios Detect Fraudulent Financial Reporting? Managerial

Auditing Journal, Vol. 19 No. 1, pp. 15-28.

Kathleen Kaminski, A. 2004. Can Financial Ratios Detect Fraudulent Financial Reporting. Journal Of Accounting Research. Vol. 19 No. 1, Pp.15-

28.

Kirkos, E., Spathis, C., Manolopoulos, Y., 2007. Data Mining techniques for the detection of fraudulent financial statement, Expert Systems with

Applications 32, pp. 995-1003.

Koroy, T.R. 2008. Pendeteksian Kecurangan (fraud) Laporan Keuangan Oleh Auditor Eksternal. Jurnal Akuntansi dan Keuangan, Vol. 10, No.1,

Mei 2008:22-33

Kummer, Tyge-F.; Singh, Kishore; & Best, Peter, 2015. "The Effectiveness of Fraud Detection Instruments in Not-For-Profit Organizations",

Managerial Auditing Journal, Vol. 30 Iss 4/5 pp. 435 – 455.

Lari, Leonard Rang’Ala. 2009. The Power Of Financial Ratios In Detecting Fraudelant Financial reporting: The Case Of Savings And Credit Co-

Operative Sosieties In Kenya. Strathmore University.

Lisic , Lin g Lei ; Silveri, Sabatino (Dino); Song, Yanheng; & Wang, Kun. 2015. Accounting Fraud, Auditing, and The Role Of Government

Sanctions in China. Journal of Business Research 68, pp. 1186–1195.

Mariana, Lia, 2014. Analisis Kecurangan Laporan Keuangan: Studi kasus pada PT Bumi Resources Tbk dan PT Berau Coal Energy Tbk.

Owusu-Ansah, Stephen; Moyes, Glen David; Oyelere, Peter Babangida; & Hay, David, 2002."An empirical analysis of the likelihood of detecting

fraud in New Zealand", Managerial Auditing Journal, Vol. 17 Iss 4 pp. 192 – 204

Persons, Obeua S. 1999. “Using Financial Information To Defferentiate Failed Vs Surviving Finance Companies In Thailand: An Implication For

Emerging Economies”, Multinational Finance Journal, Vol. 3 No.2, Pp. 127 -145.

Person s, Obeua S. 1999. “Using Financial Statement Data To Identify Factors Associated With Fraudulent Financial Reporting. Journal Of Applied

Business Research. Vol. 11 (3): Pp.131-146

Pradana, Y., Tarjo & Kompyurini, N., 2013, Detection of Fraud in Public Sector Organization : Case in Indonesia, The 5th International Conference

on Financial Criminology (ICFC) 2013 “Global Trends in Financial Crime in the New Economics”, 28-29 May 2013 Holiday Inn Glenmarie,

Shah Alam, Malaysia.

Rezaee, Hogan, E.Chris, Zabihollah, Richard A. Riley, Jr., & Uma K. Velury. 2008. Financial Statement Fraud: Insights From The Academic

Literature. Journal Of Auditing. Vol. 27 No.2, Pp.231-252.

Skousen, C. J., & C. J. Wright. 2006. Contemporaneous risk factors and the prediction of financial statement fraud. Working paper, University of

Texas at Arlington.

Skousen, Christopher J. dan Brady James Twedt. 2009. “Fraud Score Analysis in Emerging Markets”. Journal of Accounting and Economics, Vol.

16, No. 3, h. 301-315

Summers, S. L., Dan John T. Sweeney. 1998. Fraudulently Misstated Financial Statement And Insider Trading: An Emphirical Analysis The

Accounting Review. (January): Pp. 131-146.

Spathis, Charalambos T. 2002. Detecting False Financial Statements Using Published Data: Some Evidence from Greece. Managerial Auditing

Journal 17/4, p.179-191.

Tarjo, 2008. Effect of Institutional Ownership Concentration and Free Cash Flow to Leverage, Profit Management, Shareholder Value and Cost of

Equity Capital, Dissertation, Graduate Program, University of Airlangga, Surabaya, Unplished.

Tarjo, 2010, Relevance concentration of ownership and the Debt Policy Affecting Earnings Management, Shareholder Value and Cost of Equity

Capital.

Turban, Aron son, & Liang, 2005, Decision Support Systems and Intelligent Systems, 7th Edition, Prentice Hall, New Jersey.

930 Tarjo and Nurul Herawati / Procedia - Social and Behavioral Sciences 211 ( 2015 ) 924 – 930

Zaki, M., & Theodoulidis, Analyzing Financial Fraud Cases Using a Linguistics-Based Text Mining Approach, Thus paper was part of PhD research

investigated at Manchester Business Schoo, University of Manchester.

Effects of Financial Literacy and Financial Management Teaching on Financial Behavior (Student Study of Hotel & Tourism Business Study Program)

Article

May 2022

Financial literacy is an important factor in improving people's economic welfare. When a person's financial literacy is high, it is certain that the quality of life and welfare will also increase. This can happen because of the higher one's financial literacy, the better one's financial behavior. Students are the driving force of the nation's economy because students have the largest population, character, and literacy level. However, students are also faced with various very complex financial problems because some students still do not have their own source of income while the level of student needs is high. For this reason, students need to gain more knowledge about financial literacy through financial management learning models in their courses. This was done by the Hotel & Tourism Business study program at Ciputra University, Surabaya batch 2019 and 2020. However, in the process, it is necessary to evaluate the effectiveness of whether teaching changes the financial behavior of students. We used linear regression analysis to process the data in this study. Based on the description above, this study aims to find the effect of financial literacy on financial behavior with financial management teaching as a moderation. The results of this study indicate that there are significant differences in perceptions related to financial behavior between someone who has high and low financial literacy, while the teaching of financial management has a positive effect in moderating the relationship between financial literacy and financial behavior but is not significant.

Predicting fraudulent financial statement using cash flow shenanigans

Article

Jan 2023

Detection of fraudulent financial stewardship in the cash flow section is an exciting thing and is rarely studied. This research empirically tests the discovery of fraudulent financial statements based on basic cash flow shenanigans. The sample of this study amounted to 470 data mining companies in Indonesia, Malaysia, China, and Japan. The analysis method used is a positive approach. The results show that all ratios used can predict fraudulent financial statements. Three ratios of cash flow shenanigans, namely change in receivable to cash flow operations, days payable outstanding, and change in inventory to cash flow operations, significantly affect the F-Score. Meanwhile, the six cash flow shenanigans ratios, namely cash flow operations to current liability, operating cash flow ratio, free cash flow, cash flow operations to total liability, days payable outstanding, and change in inventory to cash flow operations, have a significant effect on the M-Score.

Deteksi Financial Statement Fraud : Model Beneish M-Score, dan Model F-Score

Article

Jan 2023

The research aims to find out which model is most effective in detecting financial statement fraud by using the Beneish M-Score model and the F-Score model. The research population of the banking sector is listed on the Indonesia Stock Exchange for the period 2018 - 2020 with a total of 48 banks. The research sample totaled 40 banks using a purposive sampling method. The results of the study stated that by calculating the index using the beneficial M-Score model, it showed that there were indications of banks committing financial statement fraud in 2018 at 2.5%, in 2019 at 95% and in 2020 at 97.5%. Meanwhile, using the F-Score model in 2018 it was 5%, in 2019 it was 7.5% and in 2020 it was 5%. Based on these two models, an effective model for detecting financial statement fraud is the Beneish M-Score model. Keywords: Financial Statement Fraud; Model Beneish M-Score; Model F-Score.

The Likelihood of Fraud from The Fraud Hexagon Perspective: Evidence from Indonesia

Article

Full-text available

Jan 2023

Fraudulence can cause financial loss and investor mistrust. Fraud is not only unethical but also a punishable sin. As the impact is enormous, it is crucial to examine what factors motivate or impact a company to commit fraud. Literature has archived several models to explain elements of fraudulent activities, such as the Fraud Triangle, Fraud Diamond, and Fraud Pentagon. Georgios L. Vousinas introduced the fraud hexagon in 2019, the latest model exhibiting six factors that motivate companies to commit fraud. This model consists of stimulus(pressure), capability, collusion, opportunity, rationalization, and ego (arrogance). This research aims to examine the effect of the fraud hexagon elements on the likelihood of fraud. Seventy-six manufacturing firms listed on the Indonesia Stock Exchange during 2015-2019 were chosen to be included in the sample. This study uses the Beneish M-Score model to separate companies likely to commit fraud. Logistic regression analysis was then used to test the hypothesis. The findings indicate that stimulus impacts the likelihood of fraud.

Analysis on The Effectiveness of Augmented Artificial Intelligence Implementation in Preventing Fraudulent Financial Statement by Utilizing Beneish M-Score Model

Conference Paper

Mar 2023

Comparison of the Prediction Accuracy of Earning Manipulation in Development of the Beneish Model by combination of Artificial Neural Network and Imperialist Competition and Particle Swarm Optimization Algorithms

Article

Jun 2022

Hosein Asgari

Predicting financial statement manipulation in South Africa: A comparison of the Beneish and Dechow models

Article

Mar 2023

Recently, South Africa has suffered from several large financial statement frauds. To assist stakeholders in identifying fraud, this study investigated the ability of the Beneish M-score and the Dechow et al. F-score to identify fraud in South Africa. The study also explored similarities in earnings management characteristics between false positives and fraudulent companies. Finally, the study re-estimated the models’ coefficients based on current South African data to determine if this improved their predictive capabilities. The study used a sample of 23 manipulated and 2 320 non-manipulated observations from 2006 to 2018 and found that both scores showed low sensitivity and precision. The false positives share similar, or higher, earnings management characteristics to the manipulators. Re-estimating the coefficients reduced the M-scores’ sensitivity by, on average, 6.52% but improved precision by, on average, 4.21%. Conversely, re-estimation increased the F-scores’ sensitivity by, on average, 58.70% but increased the type II error by, on average, 48.09%. These findings suggested that either the M- and F-scores are unsuitable in the South African context or that regulators have failed to identify manipulators adequately. Therefore, investors and other stakeholders should use caution when applying these models in South Africa.

Impact of Earnings Management on Working Capital Management Efficiency

Article

Mar 2023
Finance Res Lett

Peranan Sistem Pengendalian Internal dan Interaksinya dengan Proporsi Dewan Komisaris Independen Terhadap Praktik Manajemen Laba

Article

Jan 2023

This study analyzes the effect of the internal control system on earnings management with the proportion of independent commissioners as a moderating variable. The data analysis used includes descriptive statistical analysis and logistic regression using PROCESS macro v4.1 in SPSS version 25 software. Based on the determination of the research criteria, there were 33 companies consisting of 16 health sector companies with 64 sample data and 17 transportation sector companies. with 67 sample data. Through hypothesis testing, it was found that the internal control system decreases earnings management and the independent board of commissioners proportion variable moderated the relationship between the internal control system and earnings management. Keywords: Internal Control System; Earnings Management; Independent Commisioners;

Deteksi Kecurangan Laporan Keuangan dan Prediksi Kebangkrutan Perusahaan Sebelum dan Saat Pandemi Covid-19 dengan Menggunakan Perbandingan Pengukuran Model Altman Z –Score, Grover, Springate dan Zmijewski

Article

Jan 2023

Penelitian ini dilakukan untuk mengetahui: 1) Perusahaan yang tergolong manipulator atau non manipulator dalam mendeteksi kecurangan laporan keuangan, 2) membandingkan prediksi kebangkrutan perusahaan dengan menggunakan model Altman, Grover, Springate, dan Zmijewski, dan mengidentifikasi perusahaan yang diklasifikasikan sebagai zona aman, zona abu-abu dan zona marabahaya. Populasi penelitian ini adalah 43 perusahaan yang terdaftar di BEI pada perusahaan manufaktur subsektor makanan dan minuman tahun 2018-2021. Penelitian ini menggunakan metode purposive sampling dan diperoleh sampel penelitian sebanyak 25 perusahaan. Penelitian ini menggunakan metode deskriptif kuantitatif dengan data sekunder dari website www.idx.co.id. teknik analisis data menggunakan model Beneish M-Score dan Altman, Grover, Springate, dan Zmijewski. Hasil dari penelitian ini adalah 1) terdapat 6 perusahaan yang tergolong manipulator dengan persentase 24% dan 19 perusahaan tergolong non manipulator dengan persentase 76%; 2) model Zmijewski memiliki tingkat akurasi paling tinggi dari ketiga model sebelumnya yaitu tingkat akurasi sebelum pandemi 96 persen dan saat pandemi 96 persen.

Using Financial Statement Data To Identify Factors Associated With Fraudulent Financial Reporting

Article

Full-text available

Sep 2011

Obeua S. Persons

Based on stepwise-logistic models, this study finds that financial leverage, capital turnover, asset composition and firm size are significant factors associated with fraudulent financial reporting Prediction results suggest that these models outperform a nae strategy of classifying all firms as nonfraud firms for all levels of relative costs of type I and type II errors. The models also correctly identify a large percentage of fraud firms and misclassify a relatively small percentage of nonfraud firms when realistic relative error costs are assumed.

Auditor Internal sebagai "Dokter" Fraud di Pemerintah Daerah

Article

Full-text available

Apr 2015

Abstrak: Auditor Internal sebagai “Dokter” Fraud di Pemerintah Daerah. Penelitian ini bertujuan memahami peran auditor internal pemerintah daerah dalam upaya meminimalisir fraud. Metode riset yang digunakan adalah etnometodologi pada lingkungan pemerintah kabupaten Songulara. Melalui 9 (sembilan) orang pengawas/auditor sebagai informan kunci, ditemukan bahwa auditor internal pemerintah daerah adalah “Dokter” fraud. Namun demikian, “Dokter” fraud belum dapat melaksanakan perannya secara maksimal, karena minimnya kompetensi auditor internal, dan belum adanya komitmen dari manajemen puncak serta unsur-unsur terkait dalam pemerintahan di daerah. Temuan lain mengindikasikan bahwa informan sekadar “melunturkan” sifat wajib dari penugasan yang diberikan. Abstract: Internal Auditor as Fraud “Doctor” in Local Government". This research aims to understand how internal auditors of local governments understand their role in their efforts to minimize fraud. Ethnometodology was employed as method in a local government namely Sogulara. Through 9 (nine) key informants comprising supervisors/auditors, it is found that the internal auditors of local government is the fraud "Doctor", However, fraud "Doctors" has not been able to carry out their role to the maximum, because of the lack of competence of internal auditors, and lack of commitment from top management and related elements in the district. Other findings indicate that informants only do their jobs in order to lessen their obligation.

Using Financial Information to Differentiate Failed vs. Surviving Finance Companies in Thailand: An Implication for Emerging Economies

Article

Full-text available

Jun 1999

Obeua S. Persons

This article combines qualitative and quantitative information from financial statements and auditors' reports with logistic models to differentiate failed from surviving finance companies in Thailand. Failed companies are those that were forced to suspend their operations in mid-1997. The results indicate that auditors' reports from the 1996 financial statements did not differentiate failed from surviving finance companies. On the other hand, the logistic regression models indicate that failed finance companies had lower profitability, lower foreign borrowing possibly due to their poorer credit rating, lower management quality, and smaller size. These models have relatively high predictive ability for failed finance companies and low expected costs of misclassification (JEL G21, M41).

The Detection of Earnings Manipulation

Article

Full-text available

Sep 1999
FAJ

Messod D. Beneish

Abstract The Detection of Earnings Manipulation The paper profiles a sample of earnings manipulators, identifies their distinguishing characteristics,and estimates a model for detecting manipulation. The model’s variables are designed to capture either the effects of manipulation or preconditionsthat may prompt firms to engage in such activity. The results suggest a systematic relation between the probability of manipulation and financial statement variables. The evidence is consistent with accounting data being useful in detecting manipulation and assessing the reliability of accounting earnings. In holdout sample tests, the model identifies approximately half of the companies involved

An Exploratory Study of Auditors’ Responsibility for Fraud Detection

Article

Full-text available

Apr 2005
Manag Audit J

Purpose Recently, fraud has been brought to the forefront with the scandals of Enron and Worldcom. Fraudulent financial reporting and misappropriation of assets served to undermine investors’ confidence in audited financial statements. This study investigates how auditors and users perceive the auditors’ responsibility for uncovering fraud, the nature and extent of fraud in Barbados, and audit procedures utilised in Barbados since Enron. Design/methodology/approach A total of 43 respondents (19 auditors and 24 users) were surveyed regarding their perceptions and experiences on fraud, using qualitative and quantitative approaches. Findings Indicates that the expectation gap is wide, as auditors felt that the detection of fraud is management's responsibility, while users and management disagreed. Also finds that fraud is not a major issue in Barbados and that companies who have internal auditors, sound internal controls and effective audit committees are better equipped to deal with fraud prevention and detection. Research limitations/implications The sample size is relatively small and it is not intended nor claimed that those interviewed comprise a representative sample. Practical implications This research fills a void in research in this area in a small country like Barbados. These findings have important implications for users of Barbadian accounts, especially investors, auditors and regulators. Originality/value This paper fulfils a resource need for academics and practitioners, and makes an interesting contribution to our understanding of fraud in Barbados.

The effectiveness of fraud detection instruments in not-for-profit organizations

Article

May 2015
Manag Audit J

Purpose – The purpose of this study is to investigate the effectiveness of fraud detection instruments in not-for-profit (NFP) organizations. Not-for-profit organizations rely on trust and volunteer support. They are often small in size and do not have relevant expertise to prevent fraud. Such organizations are more vulnerable to fraud and, consequently, require effective fraud detection instruments. The existing literature on fraud detection is primarily descriptive and does not measure instrument performance. The authors address this research gap and provide a detailed overview of the impact of nine common fraud detection instruments. Design/methodology/approach – Data were obtained from an NFP fraud survey conducted in Australia and New Zealand. A set of contingency tables is produced to explore the relationship between the existence of a specific fraud detection instrument and actual detection of fraud. We also investigate the relationship between organization size and fraud detection strategy. Findings – The findings provide valuable insights into understanding fraud detection mechanisms. Although most fraud detection measures may not lead to more fraud detection, three highly effective instruments emerge, namely, fraud control policies, whistle-blower policies and fraud risk registers. The results also reveal that commonly used fraud detection instruments are not necessarily the most effective. This is true in a significant number of small organizations that appear to be focusing on ineffective fraud detection instruments. Practical implications – Implementation of more effective fraud detection measures will reduce the damage caused to an organization and is highly relevant for practitioners. Originality/value – The results show that differences in the effectiveness of fraud detection instruments in the NFP sector exist. This knowledge is directly applicable by related organizations to reduce fraud damage.

Accounting Fraud, Auditing and the Role of Government Sanctions in China

Article

Jun 2015
J BUS RES

We use the unique economic, legal, and political landscape of China to examine the impact of auditors on the incidence of accounting fraud. In particular, we examine whether large audit firms reduce the incidence of financial statement fraud in China, an emerging market in which auditors face strong government sanctions but low litigation risk associated with audit failures. We find that companies audited by large audit firms are less likely to commit financial statement fraud. This effect is stronger for regulated industries and for revenue-related frauds. Our results are robust to considering the severity of fraud, excluding firms cross-listing in other jurisdictions, using alternative measures of fraud, accounting for the self-selection of auditors, and controlling for other corporate governance mechanisms. Our results highlight the role of government sanctions in assuring audit quality and have important practical implications to help international audit firms – and businesses more generally – successfully compete in China.

Data mining journal entries for fraud detection: An exploratory study

Article

Sep 2010
Int J Account Inform Syst

Fraud detection has become a critical component of financial audits and audit standards have heightened emphasis on journal entries as part of fraud detection. This paper canvasses perspectives on applying data mining techniques to journal entries. In the past, the impediment to researching journal entry data mining is getting access to journal entry data sets, which may explain why the published research in this area is a null set. For this project, we had access to journal entry data sets for 29 different organizations. Our initial exploratory test of the data sets had interesting preliminary findings. (1) For all 29 entities, the distribution of first digits of journal dollar amounts differed from that expected by Benford's Law. (2) Regarding last digits, unlike first digits, which are expected to have a logarithmic distribution, the last digits would be expected to have a uniform distribution. Our test found that the distribution was not uniform for many of the entities. In fact, eight entities had one number whose frequency was three times more than expected. (3) We compared the number of accounts related to the top five most frequently occurring three last digit combinations. Four entities had a very high occurrences of the most frequent three digit combinations that involved only a small set of accounts, one entity had a low occurrences of the most frequent three digit combination that involved a large set of accounts and 24 had a low occurrences of the most frequent three digit combinations that involved a small set of accounts. In general, the first four entities would probably pose the highest risk of fraud because it could indicate that the fraudster is covering up or falsifying a particular class of transactions. In the future, we will apply more data mining techniques to discover other patterns and relationships in the data sets. We also want to seed the dataset with fraud indicators (e.g., pairs of accounts that would not be expected in a journal entry) and compare the sensitivity of the different data mining techniques to find these seeded indicators.

Can financial ratios detect fraudulent financial reporting?

Article

Jan 2004
Manag Audit J

Fraudulent financial reporting is a matter of grave social and economic concern. The Treadway Commission recommended that the Auditing Standards Board require the use of analytical procedures to improve the detection of fraudulent financial reporting. This is an exploratory study to determine if financial ratios of fraudulent companies differ from those of nonfraudulent companies. Fraudulent firms were identified by examining the SEC's Accounting and Auditing Enforcement Releases issued between 1982 and 1999. The fraudulent firms (n=79) were then matched with nonfraudulent firms on the basis of firm size, time period, and industry. Using this matched-pairs design, ratio analysis for a seven-year period (i.e. the fraud year -/+ 3 years) was conducted on 21 ratios. Overall, 16 ratios were found to be significant. Of these, only three ratios were significant for three time periods. Of the 16 statistically significant ratios, only five were significant during the period prior to the fraud year. Using discriminant analysis, misclassifications for fraud firms ranged from 58 percent to 98 percent. These results provide empirical evidence of the limited ability of financial ratios to detect and/or predict fraudulent financial reporting.

Fraud score analysis in emerging markets

Article

Jul 2009
Cross Cult Manag

Purpose – The purpose of this research is to determine the likelihood of financial statement manipulations in companies throughout a variety of emerging market countries and compare this potential wirh that of firms within the USA. Design/methodology/approach – The authors utilize the Fraud Score Model, as set forth by Dechow et al., to determine the likelihood of financial statement manipulations. By adjusting their model to work in an international setting, the authors are able to study nine industries across 23 countries, including the USA. Findings – The results vary from industry to industry, with some countries performing extremely well in one industry, only to prove remarkably risky in the next. Originality/value – The findings may be used by a variety of market participants, especially investors, to determine the risk levels of potential foreign investments. Therefore, this research can help lead to a more overall efficient placement of global capital.