D552 Task 2 Fraud Analysis Report Example

Twenty-Three Repeats, One Shared Address and $41,000 to Strangers: A Fraud Analysis of a Year of Invoice Data at a Composite Industrial Parts Manufacturer

Student Name

School of Business, Western Governors University

D552: Data Analytics for Accountants I, Task 2

Course Instructor

Month Day, Year

Twenty-Three Repeats, One Shared Address and $41,000 to Strangers: A Fraud Analysis of a Year of Invoice Data at a Composite Industrial Parts Manufacturer

Purpose and Scope

Lakeshore Precision Components, a composite manufacturer of machined parts in Wisconsin with annual revenue of $86 million, asked the internal audit team to test its accounts payable data for signs of fraud after an anonymous tip reported that a purchasing employee was living beyond his means. Tips are the most common way occupational fraud is first detected, and billing schemes, in which an employee causes the organization to pay false or inflated invoices, rank high among the ways employees steal from employers (Association of Certified Fraud Examiners [ACFE], 2024). The analysis covered all 9,840 invoices paid in the last fiscal year, totaling $31.2 million, together with the approved vendor master file and the employee master file from human resources.

Data Preparation

The invoice file, vendor master and employee file were extracted from the enterprise system, and control totals were agreed to the general ledger. Vendor names and addresses were standardized, removing punctuation, suite numbers and abbreviations such as St and Street, so that matching would not fail on formatting. Employee addresses were loaded to a protected table with access limited to the audit team. Because the tip named no one and a fraud analysis can damage reputations, the team agreed in advance that results would be reported only to the chief audit executive and legal counsel, that no employee would be contacted during the analysis, and that every test would run on the full population rather than on the purchasing employee alone, so the findings could not be shaped by the tip.

Test 1: Duplicate Payments

Method: invoices were matched on vendor, amount and invoice number, and then on vendor and amount within 30 days where invoice numbers differed only by a character or suffix, such as 10442 and 10442A.

Results: 23 pairs were flagged, totaling $118,600 in second payments. Seventeen pairs, $71,300, involve large established suppliers and appear to result from the same invoice being entered once from the mailed copy and once from the emailed copy; three have already been credited back. Six pairs, $47,300, involve a single vendor, Midstate Industrial Supply, and in each case the second invoice number has an added suffix and was approved by the same purchasing agent.

Assessment: the 17 large-supplier pairs are most likely processing errors that indicate a control weakness. The six Midstate pairs follow a pattern consistent with deliberate duplication and need investigation.

Test 2: Vendor Addresses Matching Employees

Method: standardized vendor addresses and bank account numbers were compared with employee home addresses and direct deposit accounts.

Results: one vendor, Midstate Industrial Supply, shares a mailing address with a purchasing agent employed for six years. Midstate was added to the vendor master two years ago, was paid $212,000 last year, and supplies general items such as abrasives and gloves that Lakeshore also buys from two large distributors. No bank account matches were found.

Assessment: a vendor sharing an employee's address is a classic indicator of a shell company billing scheme. Combined with the duplicate invoices and the fact that the same agent approved Midstate's purchases, this is the highest-priority finding.

Test 3: Payments to Unapproved Vendors

Method: every vendor paid during the year was compared with the approved vendor list.

Results: $41,000 was paid to 12 vendors that never went through the approval process. Eleven, $17,400, are one-time payments for items such as emergency repairs, entered as one-time vendors by accounts payable. One, Northgate Tooling Services, received $23,600 in nine payments, each just under the $2,700 threshold above which a second approval is required.

Assessment: the one-time vendors reflect a process gap. Northgate's pattern of payments just below an approval limit is a red flag for splitting payments to avoid review and needs follow-up.

Additional Analysis

A first-digit test of invoice amounts against Benford's Law, which predicts that in many naturally occurring financial data sets lower first digits appear more often than higher ones, was run on the full population (Nigrini, 2012). The overall distribution conformed closely, but Midstate's invoices showed far more amounts beginning with 4 than expected, consistent with invoices sized to stay under a purchasing agent's $5,000 approval limit. Weekend and holiday invoice dates were also reviewed; none were unusual.

Limits of the Analysis

Data analysis identifies where to look; it does not prove intent. The Midstate findings could have an innocent explanation, such as an employee's relative operating a legitimate business from a shared home, though that would still be an undisclosed conflict of interest. Physical verification is needed: whether Midstate's goods were received, whether Midstate exists at the address and who owns it. The tests covered one year, so earlier losses are unknown. Richardson and colleagues stress that analytics results should be refined and confirmed before they are communicated as conclusions (Richardson et al., 2023).

Recommendations

Immediate: refer the Midstate and Northgate findings to legal counsel and the audit committee for investigation, preserve the purchasing agent's email and system records, and suspend payments to both vendors pending review. Do not confront the employee until the investigation plan is in place.

Controls: require new vendors to be approved by someone outside purchasing, with verification of tax identification number, address and ownership; run the address and bank account match monthly; configure the system to block invoice numbers that duplicate an existing number with a suffix; separate the ability to add vendors from the ability to approve purchases; review payments just below approval thresholds quarterly; and limit the use of one-time vendors to a defined list of emergencies with after-the-fact review.

Conclusion

A year of invoice data points to one vendor, Midstate Industrial Supply, where a shared address with a purchasing agent, suffixed duplicate invoices and amounts clustered below an approval limit form a consistent pattern, and to a second vendor whose payments appear split to avoid review. Most other flags look like processing errors, but they reveal the same weaknesses in vendor setup and duplicate controls. Investigating the two vendors and closing those gaps should come next.

References

Association of Certified Fraud Examiners. (2024). Occupational fraud 2024: A report to the nations. https://legacy.acfe.com/report-to-the-nations/2024/

Nigrini, M. J. (2012). Benford's law: Applications for forensic accounting, auditing, and fraud detection. Wiley.

Richardson, V. J., Teeter, R. A., & Terrell, K. L. (2023). Data analytics for accounting (3rd ed.). McGraw Hill.