Innovation of Fraud Deterrence System in the Organization Using Forensic Accounting and Data Mining Techniques. Pornchai Naruedomkul, Ph.D.

Transcription

1 Innovation of Fraud Deterrence System in the Organization Using Forensic Accounting and Data Mining Techniques Pornchai Naruedomkul, Ph.D., CFE

2 Fraud: A Critical Issue for Business Fraud cases are increasing year on year. Loss due to fraudulent is about 5 7% of the revenues in the U.S.A.

3 Estimated Loss to Occupational Fraud and Abuse in U.S.A. (Trillion USD) Description Year Gross domestic products Loss to occupational fraud and abuse Loss to gross domestic products (%) Source: Adapted from the ACFE Report to the Nations 1996, 2002, 2004, 2006, 2008, 2010, 2012, and 2014

4 Fraud in Thailand From KPMG Survey in 2007: Fraud risk is a major issue. Fraud issues found in Thailand are similar to those in other countries. Fraud cases would increase in the next 2 years.

5 Estimated Financial Losses from Fraud Detected in Thailand Amount Baht 10 million and above 11% 16% Baht 5 million to less than 10 million 5% 8% Baht 1 million to less than 5 million 15% 18% Baht 100,000 to less than 1 million 31% 27% Less than Baht 100,000 38% 31% Source: Adapted from KMPG, 2007

6 Fraud Definition Albrecht et al. (2009: 7) defined fraud as: a generic term, and embraces all the multifarious means which human ingenuity can devise, which are resorted to by one individual, to get an advantage over another by false representations. No definite and invariable rule can be laid down as a general proposition in defining fraud, as it includes surprise, trickery, cunning and unfair ways by which another is cheated. The only boundaries defining it are those which limit human knavery.

7 Fraud Definition Wells (2010: 8 and 2011: 2) defined fraud as: encompass any crime for gain that uses deception as its principal modus operandi. Of the three ways to illegally relieve a victim of money force, trickery, or larceny all offenses that employ trickery are frauds.

8 Possible Causes for Fraud Selfactualization Esteem needs Love needs Safety needs Physiological needs Maslow s hierarchy of needs Source: Adapted from Maslow, 1943

9 Possible Causes for Fraud Theory of planned behavior Source: Adapted from Ajzen: 182

10 Fraud Triangle Source: Adapted from Singleton et al., 2006: 9

11 Fraud Management Lifecycle Theory Fraud deterrence will stop fraud before it happens. Fraud Deterrence Fraud Prevention Fraud Detection Fraud Mitigation Fraud Analysis Fraud Policy Fraud Investigation Fraud Prosecution Source: Adapted from Wilhelm, 2004

12 Risk Factors and Red Flags Indicators Living beyond means Financial difficulties Control issues, unwillingness to share duties Unusually close association with vendor/customer Wheeler-dealer attitude Divorce/family problems Irritability, suspiciousness, or defensiveness Addiction problems Source: Adapted from Association of Certified Fraud Examiners Report 1996, 2002, 2004, 2006, 2008 and 2010; Coenen, 2008

13 Risk Factors and Red Flags Indicators Refusal to take vacations Past employment-related problems Complained about inadequate pay Excessive pressure from within organization Past legal problems Instability in life circumstances Excessive family/peer pressure for success Complained about lack of authority Source: Adapted from Association of Certified Fraud Examiners Report 1996, 2002, 2004, 2006, 2008 and 2010; Coenen, 2008

14 Forensic Accounting Definition Bolgna and Linquist (1995) and The Accountant s Handbook on Fraud & Commercial Crime defined it as: the application of financial skills and an investigative mentality to unresolved issues, conducted within the context of the rules of evidence (cited in Digabriele J. A., 2008: 331; Mehta and Mathur, 2007: 1575). Hopwood, Leiner, and Young (2009: 3) defined it as: the application of investigative and analytical skills for the purpose of resolving financial issues in a manner that meets standards required by courts of law.

15 Data Mining Definition Han and Kamber (2001: 5) defined it as extracting or mining knowledge from large amount of data. Hand, Mannila, and Smyth (2001: 5) defined it as the analysis of (often large) observational data sets to find unsuspected relationships and to summarize the data in novel ways that are both understandable and useful to the data owner.

16 Data Mining Technique Data mining combines many fields into its technique; machine learning, pattern recognition, statistics, databases and visualization to execute the unknown information by extraction from large database (Cabena et al., 1999; Han and Kamber, 2001). Data Mining Techniques 1. Supervised method Classification Prediction 2. Unsupervised method Association rule Clustering Outlier

17 Data Mining Technique Classification The unknown data will be classified into groups based on the similar data known by the classification to develop the patterns (Han and Kamber, 2001; Shmueli, Patel, and Bruce, 2007). Prediction Prediction is similar to classification except that prediction tries to predict the value of a numerical variable rather than a class (Han and Kamber, 2001; Shmueli, Patel, and Bruce, 2007). Association rule This technique will be used with patterns which occur frequently in data (Han and Kamber, 2001). Frequent patterns can be item sets, subsequences, and substructures; item sets refers to a set of items normally appeared together (i.e., a customer normally buys shampoo and conditioner); subsequent refers to the pattern that a customer has an intention to buy a camera will follow by buying a memory card. Mining this pattern will lead to a discovery of useful association of data (Han and Kamber, 2001).

18 Data Mining Technique Cluster analysis The data will be clustered or grouped without consulting a known class label. The objects are clustered or grouped based on the principle of maximizing the intraclass similarity and minimizing the interclass similarity (Han and Kamber, 2001: 26). Outlier analysis The data objects that are totally different from other groups of data are considered as outliers. This method can be used in fraud detection (Han and Kamber, 2001: 451) by detecting the unusual transactions.

19 Fraud Deterrence Model Attitude and behavior survey was conducted to collect the information from 2,000 participants. Cluster Analysis Cluster algorithms in WEKA* were used to generate the patterns of fraud risk behaviors which would be used as fraud deterrence model. 1. K-means 2. Expectation-Maximization (EM) 3. Filtered cluster 4. Make density-based cluster *Waikato Environment for Knowledge Analysis (WEKA) implemented by University of Waikato in New Zealand is the leading open-source project in machine learning.

20 Fraud Deterrence Model Patterns from cluster algorithms are appropriate for fraud detection because they classify the similar cases in the same cluster. Prediction of potential frauds varies depending on the behaviors of fraudsters. Fraud risk behaviors may not be suited with all clusters generated. Cluster analysis is not appropriate to be used to generate the patterns or rules for prediction of fraud risk behaviors.

21 Fraud Deterrence Model Association Rule Association rule is more appropriate to be used to find out the important rules for fraud potential to occur in association with each fraud risk factor. Apriori algorithm in association rule of WEKA program is selected to be used to find out the best rules. Association rule is generally expressed in the form of X Y or if X then Y

22 Fraud Deterrence Model Minimum support of X Y means the percentage of transactions in the database that contain both X and Y. Example Total participants of attitude and behavior survey = 2,000 Participants who had living beyond means And financial difficulties (X and Y) = 225 Minimum support (X Y) = 225/2000 = 0.11 Copyright by Pornchai Naruedomkul, Ph.D., CFE

23 Fraud Deterrence Model Minimum confidence of X Y means the percentage of transaction that contain items in X and also contain items in Y. Example Total participants of attitude and behavior survey = 2,000 Participants who had living beyond means (X) = 275 Participants who had financial difficulties (Y) = 260 Participants who had living beyond means and financial difficulties (X and Y) = 225 Minimum confidence (X Y) = 225/275 Copyright by Pornchai Naruedomkul, Ph.D., CFE = 0.82

24 Fraud Deterrence Model The minimum support and the minimum metric (confidence) were set up in 4 sets as follows: 1. Minimum support = 0.01 and minimum metric (confidence) = Minimum support = 0.01 and minimum metric (confidence) = Minimum support = 0.01 and minimum metric (confidence) = Minimum support = 0.10 and minimum metric (confidence) = 0.90 Copyright by Pornchai Naruedomkul, Ph.D., CFE

25 Fraud Deterrence Model 7 attributes could not find the best rules at minimum support of 0.10 and minimum metric (confidence) of Only one attribute could not find the best rules at minimum support set up of 0.01 and minimum metric (confidence) set up of 0.95 and All contributes could find the best rules at minimum support set up of 0.01 and minimum metric (confidence) set up of 0.90 and the rules would be used to develop fraud deterrence model. Copyright by Pornchai Naruedomkul, Ph.D., CFE

26 Best rules, min sup 0.1, min metric (confidence) 0.9 The meaning of the rules can be explained as follows: Example: Rule No.1 X Y Addiction problems and borrow money from coworkers Living beyond means (Fraud risk behavior) (Fraud risk factor) Probability to occur fraud = 0.92 (92%) Copyright by Pornchai Naruedomkul, Ph.D., CFE

27 Fraud Risk Ranking Model Fraud survey was conducted with 280 companies from the research of Risk Factors for Fraudulent in the Organization in Thailand that indicated that they experienced the corporate fraud in the past 3 years. Frequency and consequences for those fraudulent activities were examined based on risk management Principles and Guidelines (ISO 31000) of the International Organization for Standardization. Type and method of corporate fraud used in the questionnaire were complied with the type and fraud method of Association of Certified Fraud Examiners. Copyright by Pornchai Naruedomkul, Ph.D., CFE

28 Fraud Deterrence System Work Flow Process Copyright by Pornchai Naruedomkul, Ph.D., CFE

29 Report of Prediction Potential Fraud in the Organization To: Board of Directors of ABC Company Limited Scope We have performed the procedure agreed with you as detailed in the written instructions on March 1, 2011, and described below with respect to the fraudulent activities in the organization of ABC Company Limited as of December 31, The procedure was performed solely to assist you in evaluating potential employee frauds and is summarized as follows: We obtained the data of attitude and behavior survey that was filled out by employees of ABC Company Limited as of December 31, We used the data to evaluate potential employee frauds. Findings We report as follows: With respect to the above we found that the significant risk factors due to fraudulent activities in your organization are lack of authority, past legal, past employment, wheeler-dealer attitude, and inadequate pay. Furthermore, the analysis indicates that 16 out of 17 persons could be potential fraudsters. For more details, please refer to the documents attached. Our report is solely for the purpose set forth in the first paragraph of this report and for your further action on risk management as well as risk minimization due to fraudulent activities. This report relates only to the items specified above. Date Address Signature Copyright by Pornchai Naruedomkul, Ph.D., CFE

30 Result from Fraud Deterrence Module Testing Potential Fraud Report ID Fraud Risk Factor Probability Fraud Risk Behaviors 1 - Mr. Charles Dolye Wheeler_Dealer Past_employement 1 - Mr. Charles Dolye Past_legal Close_relationship + Past_employement 1 - Mr. Charles Dolye Financial Life_circumstances + Past_employement 1 - Mr. Charles Dolye Borrow_money Close_relationship + Past_employement 1 - Mr. Charles Dolye Living Close_relationship + Past_employement 2 - Mr. Sura Chan Wheeler_Dealer Financial 2 - Mr. Sura Chan Living Financial + Inadequate_pay 3 - Mr. Tony Gana Wheeler_Dealer Financial + Past_employement 3 - Mr. Tony Gana Inadequate_pay Family_problems + Past_employement 4 - Mr. Somsak Chanmeeboon Lack_authority Life_circumstances + Pressure_organization 5 - Mrs. Soontaree Chaiudom Wheeler_Dealer Living + Financial + Borrow_money 5 - Mrs. Soontaree Chaiudom Inadequate_pay Family_problems + Wheeler_Dealer + Borrow_money 5 - Mrs. Soontaree Chaiudom Addiction_problems Living + Financial + Family_problems + Wheeler_Dealer + Borrow_money 6 - Mr. Sakchai Lertmongkol Unwilling Life_circumstances + Pressure_organization + Lack_authority 6 - Mr. Sakchai Lertmongkol Lack_authority Life_circumstances + Pressure_organization 7 - Mrs. Wan Wanna Close_relationship Family_problems + Inadequate_pay 7 - Mrs. Wan Wanna Wheeler_Dealer Financial 7 - Mrs. Wan Wanna Living Financial + Inadequate_pay 9 - Mr. Sakka Changsook Close_relationship Wheeler_Dealer + Irriabilities 9 - Mr. Sakka Changsook Wheeler_Dealer Financial + Refusal 9 - Mr. Sakka Changsook Past_legal Financial + Refusal + Wheeler_Dealer 9 - Mr. Sakka Changsook Inadequate_pay Wheeler_Dealer + Irriabilities 9 - Mr. Sakka Changsook Unwilling Refusal + Wheeler_Dealer + Irriabilities 9 - Mr. Sakka Changsook Pressure_family Gambling + Refusal + Wheeler_Dealer 10 - Mr. Nares Meesook Wheeler_Dealer Past_legal + Past_employement 10 - Mr. Nares Meesook Past_legal Close_relationship + Past_employement 10 - Mr. Nares Meesook Borrow_money Close_relationship + Past_legal + Past_employement 10 - Mr. Nares Meesook Living Close_relationship + Past_legal + Past_employement 11 - Mr. Tawin Chaisri Close_relationship Family_problems + Pressure_family 11 - Mr. Tawin Chaisri Wheeler_Dealer Financial + Addiction_problems 11 - Mr. Tawin Chaisri Past_legal Financial + Close_relationship 11 - Mr. Tawin Chaisri Inadequate_pay Family_problems + Pressure_family + Close_relationship 11 - Mr. Tawin Chaisri Living Financial + Addiction_problems 11 - Mr. Tawin Chaisri Financial Pressure_family + Gambling 11 - Mr. Tawin Chaisri Addiction_problems Living + Financial + Family_problems + Pressure_family + Close_relationship 11 - Mr. Tawin Chaisri Refusal Pressure_family + Close_relationship 12 - Mr. Srichan Wanpen Past_legal Financial + Close_relationship 12 - Mr. Srichan Wanpen Wheeler_Dealer Financial 12 - Mr. Srichan Wanpen Living Financial + Close_relationship 13 - Ms. Netdao Sopa Close_relationship Family_problems + Inadequate_pay 13 - Ms. Netdao Sopa Wheeler_Dealer Living + Close_relationship 13 - Ms. Netdao Sopa Past_legal Financial + Close_relationship 13 - Ms. Netdao Sopa Living Financial + Life_circumstances 13 - Ms. Netdao Sopa Inadequate_pay Family_problems + Close_relationship 13 - Ms. Netdao Sopa Addiction_problems Living + Financial + Family_problems + Inadequate_pay + Close_relationship 14 - Mr. Sak Rodboon Wheeler_Dealer Living + Close_relationship 14 - Mr. Sak Rodboon Past_legal Financial + Close_relationship 14 - Mr. Sak Rodboon Living Financial + Inadequate_pay 15 - Mr. Keng Chaikla Inadequate_pay Close_relationship + Irriabilities 16 - Mr. Songsak Hanta Unwilling Life_circumstances + Pressure_organization + Lack_authority 16 - Mr. Songsak Hanta Lack_authority Life_circumstances + Pressure_organization 17 - Mr. Thai Rakdee Close_relationship Life_circumstances + Irriabilities 17 - Mr. Thai Rakdee Inadequate_pay Life_circumstances + Irriabilities 17 - Mr. Thai Rakdee Lack_authority Life_circumstances + Pressure_organization Copyright by Pornchai Naruedomkul, Ph.D., CFE

31 Result from Fraud Deterrence Module Testing Non-Potential Fraud Report Name Fraud Risk Factor Probability Fraud Risk Behaviors 8. Mr.Ake Kajonkul No Risk Factor 0 No Risk Behaviors Copyright by Pornchai Naruedomkul, Ph.D., CFE

32 Fraud Deterrence System Verification Comparison of 1 st and 2 nd train test for fraud deterrence module Name 1st Test 2st Test 1. M r.charles Dolye P P 2. M r.sura Chan P P 3. M r.tony Gana P P 4. Mr.Somsak Chanmeeboon X P 5. M rs.sonntaree Chaiudom P P 6. M r.sakchai Lertmongkol X P 7. Mrs.Wan Wanna P P 8. M r.ake Kajonkul P P 9. M r.sakka Changsook P P 10. M r.nares M eesook P P 11. Mr.Tawin Chaisri P P 12. Mr.Srichan Wanpen P P 13. Ms.Netdao Sopa P P 14. High Ranking Officer P P 15. M r.keng Chaikla P P 16. Mr.Songsak Hanta X P 17. M r.thai Rakdee P P P Predict potential fraud correctly X Predict poential fraud incorrecly Copyright by Pornchai Naruedomkul, Ph.D., CFE

33 Report of Level of Risk Due to Fraudulent Activities in the Organization To: Board of Directors of ABC Company Limited Scope We have performed the procedure agreed with you as detailed in the written instructions on March 1, 2011, and described below with respect to the fraudulent activities in the organization of ABC Company Limited as of December 31, The procedure was performed solely to assist you in evaluating the level of risk due to fraudulent activities in the firm and is summarized as follows: We obtained the frequency, consequences, and details of risk issues due to fraudulent activities of ABC Company Limited and used them to generate the level of risk for the organization. Findings We report as follows: With respect to the above we found that asset misappropriation is the most popular for fraud committed: 72.72% (24 out of 33 cases) while corruption is 18.18% (6 out 33 cases) and fraudulent statement is 9.1% (3 out of 33 cases). The most important assets that the company should protect from fraudulent activities are cash on hand, inventory, expense reimbursement, and fictitious of revenues and expenses. For more details, please refer to the documents attached. Our report is solely for the purpose set forth in the first paragraph of this report and for your further action on risk management as well as risk minimization due to fraudulent activities. This report relates only to the items specified above. Date Address Copyright by Pornchai Naruedomkul, Ph.D., CFE Signature

34 Results from Fraud Risk Ranking Module Testing Fraud risk ranking summary report Company Name: A Accounting and Finance Likelihood Low Asset Misappropriation Impact Medium Agree Risk Low Corruption Likelihood Impact Agree Risk Fraudulent Statements Likelihood Impact Agree Risk Likelihood Low Overall Impact Low Agree Risk Low Copyright by Pornchai Naruedomkul, Ph.D., CFE

35 Fraud Risk Ranking Report Company A Department Name Risk No. Type of Fraud Fraud Method Fraud Risk Issue Likelihood Estimated % Loss Risk Level Likelihood Risk Level Impact Agreed Risk Status Accounting and Finance 1 Asset Misappropriation cash-larceny-cashonhand Cashier steals money for personal use 6 3 High Medium High New Accounting and Finance 2 Asset Misappropriation cash-fraudulent disbursementsoverstated expense Reimburse exceeding expenses without receipts Very low Very low Very low New Accounting and Finance 3 Asset Misappropriation cash-fraudulent disbursementsmultiple reimbursements expense many times Use copy of the same receipt to reimburse the same High Very low Low New Accounting and Finance 4 Asset Misappropriation cash-skimming-unrecorded Not recording some sales transactions and receive cash from customers Medium Very low Low New Accounting and Finance 5 Asset Misappropriation cash-skimming-lappingschemes Receive cash from clients and use it for personal use and return the money to the company with the next clients money Medium Medium High New Copyright by Pornchai Naruedomkul, Ph.D., CFE

36 Fraud Risk Ranking Report Example No. 1 Department Name = Accounting and Finance Risk No. = 1 Type of Fraud = Asset Misappropriation Fraud Method = Cash Larceny, Cash on hand Fraud Risk Issue = Cashier steals money for personal use Likelihood = 6 Estimated % Loss = 3 Risk Level Likelihood = High Risk Level Impact = Medium Agreed Risk = High Status = New Copyright by Pornchai Naruedomkul, Ph.D., CFE

37 Fraud Deterrence System A tool to indicate potential frauds in the organization. Forensic accountant is needed for further investigation for the staff who fits the potential fraud or is lower than what it should be (i.e., reject all questions in the form). Copyright by Pornchai Naruedomkul, Ph.D., CFE

38 Thank you for your attention!