Fraud Alert: DYNEGY

The Situation: $300 million bank loan disguised to look like cash flow, through a series of complicated trades. Three Dynegy employees, including its vice president of taxation, have been charged with setting up a phony company in a scheme to buy and sell gas from Dynegy to turn the $300 million bank loan into sales, and generate $79 million in tax benefits in the process.

The Technical Answer: Fraud-detection software would have picked up on the sudden jump in gas sales and purchases even if they were disguised by a series of trades. Nonetheless, the complicated nature of the trades would have required substantial detective work on the part of the auditor.

The Potential: Knowing that fraud detection technology was in place and in active use might have deterred the Dynegy employees from trying this scheme. However, Bruce Nearon, a fraud and technology expert with the New York State Society of CPAs, says it shouldn’t even require a high-tech answer. Any number of parties—from the banks, to the auditors, to tax experts, to lawyers involved in setting up the phony firm and approving the $300 million loan—should have raised questions and stopped the scheme in its tracks.

Fraud Alert: RITE AID

The Situation: Earnings inflated by $1.6 billion through the improper recording of supplier rebates, expenses and calculation of the life of assets. A rich compensation plan would have awarded four top executives $220 million in stock if the pharmacy chain’s share price and earnings met certain targets.

The Technical Answer: Fraud-detection software likely would have isolated the sudden jump in supplier rebates. Data analysis could have identified changes in expense reporting, particularly at the end of a reporting period. Accounting software could have flagged changes to the life of assets and established an audit trail to lead to those who made the changes.

The Potential: Changes to financial records are difficult to discover if they are made in small increments over long periods of time, says Bruce Nearon, a fraud and technology expert with the New York State Society of CPAs. However, in a situation like this, where executives made abrupt changes to trigger their huge payday, software should have been able to easily uncover the scheme.


The Situation: $131 million in revenue booked prematurely, mainly by setting back computer clocks on the last day of a quarter.

The Technical Answer: Database queries by internal or external auditors to compare the average number of orders shipped per day during a given quarter versus those shipped on the quarter’s last day.

The Potential: Doing such a query, SEC investigators later found that on the last day of 1993’s third quarter, Sensormatic records show 1,692 orders shipped. But the daily average for the quarter was just 232. If fundamentals like that are too big or too small, says Ed Ketz, a professor at Pennsylvania State University who studies accounting scandals, “that would require investigation.”

Fraud Alert: WORLDCOM

The Situation: $11 billion worth of overstated sales and profits. Expenses disguised as assets.

The Technical Answer:

  • Database analysis to compare ratios of broad metrics, such as cash, accounts receivable and fixed assets, quarter-to-quarter and year-to-year.
  • Specialized fraud-detection software, such as IDEA from CaseWare Inc. in Toronto, Canada, or ACL from ACL Services in Vancouver, British Columbia, to analyze transactions in a given category.
  • Set triggers in accounting software that alert internal auditors when entries are made after the close of a quarter and when entries are too perfect, such as round, multimillion-dollar figures.

    The Potential: Late-and-perfect entries into the accounting system are glaring clues about misdeeds, at least in hindsight. Comparing ratios against those of telecommunication rivals would have turned up inconsistencies, says Bill Haslinger, assistant professor of economic crime investigation at Hilbert College in Hamburg, N.Y. “I can tell you that when fraud examiners are looking for financial-statement fraud, this is the first analysis they make,” says Haslinger, a former IRS criminal investigator.