“A sure-fire way to enable fraud detection”
It’s difficult to discover and prevent fraud. Fraudsters develop new schemes all the time and they often get more sophisticated than our methods of detection. A survey in 2008 conducted by the Association of Certified Fraud Examiners (ACFE) found that U.S. organizations lose seven percent of their annual revenue to fraud. Based on the 2008 GDP, this is approximately $994 billion in fraud losses and in the past 12 months, employees accounted for 48 percent of those cases.
Keeping with the unfortunate truth that in desperate times even trustworthy employees will go to desperate measures, today’s economic situation is giving way to employee theft and fraud. The most steadfast employees who would never think of committing fraud against their employers are more willing to take unlawful risks in order to have some extra money in their pockets. The core problem being several organisations simply does not conduct the appropriate testing to proactively detect fraudulent activity. By employing data mining techniques, however, organisations can significantly increase their detection of fraud and, as a result, deter fraudsters.
Why Data Mining?
Data mining is the process of extracting patterns from data. It uses refined data search options and statistical algorithms to unearth patterns and correlations and can be useful in a variety of applications, including fraud detection. Data mining can help your organisation find anomalies and spot internal control weaknesses, including conflicting data, abnormal transactions, duplicate payments, missing invoices, nonstandard transactions/vendors, and procurement and disbursement frauds.
Sheila Anderson, CPA, is Vice President,Accounting Consulting and Development for Owings Mills, MD–based Chesapeake System Solutions (www.chessys.com) that apply sophisticated technology to achieve financial governance, risk management and compliance. She feels that data mining, a methodology based on statistical theory, is a key way to uncover and combat corporate fraud. She recommends that data mining be a regular audit activity not just a procedure implemented when fraud is suspected. Further, she cautions that getting ‘matches’ is just the first step; it is important not to misinterpret preliminary data or jump to conclusions without doing further investigation. Anderson identifies three major areas where data-mining software is invaluable: payroll, vendor payments and expense reports.
Relative to payroll fraud, data-mining tools can perform the following analyses:
“It is important not to misinterpret preliminary data or jump to conclusions without doing further investigation.”
- Authorized Payments:
Run the payroll checks cut against a database of authorized employees.
Compute the annualized salary of all authorized employees and compare these amounts against the salary expense account. Anderson says it is possible to run the following analyses to uncover fraud in accounts payable relative to vendors:
- Payments to Vendors:
Look for repetitive/frequently appearing vendor names, dollar amounts and invoice dates.
- Vendor Address Locations:
Identify whether there are duplicate addresses for the same vendor. Also, run the master vendor address list against an employee address list.
- Vendor Activity Frequency:
Review exceptionally high activity levels for new and established vendors.
- Payments to Unauthorized Vendors:
Run vendor payment records against the authorised vendors’ list.
Compare addresses where refund payments have been sent against addresses on file. Check for extraordinary numbers of refunds going to the same addresses.
- Unbundling Purchases:
Check amounts that fall just below the threshold authorisation amount.
- Multiple Payments:
Evaluate whether multiple payments have been sent to the same vendor. Expense report fraud is another problem area. Here, data mining can evaluate:
- Threshold Analysis:
Look for payments that are just below the threshold where receipts are required.
- Excessive Filings:
Review whether an employee is filing excessively.
- Transaction Analysis:
See if an employee’s activity reflects numerous transactions falling just under the amount requiring receipt submission, such as travel fees.
It is often simple steps that go a long way in preventing corporate fraud. When you fail to set up a process to check fraud, you are increasing the possibility of fraud occurrence in your organisation. So, step up and take action.