Forensic Analytics, 2nd Edition

CHAPTER 8Identifying Abnormal Duplications

THE DUPLICATION-BASED TESTS DISCUSSED in this chapter are based on the premise that duplications or excessive duplications within subsets are indicators of fraud and errors. Since there is usually some level of normal duplication, and some level of abnormal duplication, we will have to review our results very carefully to find those duplications that are indicators of errors or fraud. One way to review long lists of false positives is to only review duplications above some dollar threshold. One way to avoid lots of false positives is to run tests where abnormal duplication are reasonably reliably indicators of fraud or errors. The case study discussed in the chapter makes it clear that fraudulent duplications can be subtle, as in an employee claiming an air ticket as a reimbursable expense as well as claiming mileage for having driven to the specified location.

The first test discussed is a straightforward test to find duplicate records. Duplicate payments with all fields being equal are rare in accounts payable systems because the payments system should issue an alert if an exact duplicate of a transaction is entered for payment. Duplicates can, however, occur in other data sets. The second test discussed is a little more complex in that we are looking for partial duplicates. The biggest paybacks from these partial duplicates test has been from identifying accounts payable cases of the same dollar amounts, the same date, the same ...