What’s Your Fraud Score?

Receive the wrong grade on this test, and your auditor may be scouring your company's financial statements for evidence of earnings management.
Marie LeoneAugust 17, 2007

A new research report contains formulas to help auditors and investors predict the likelihood that a public company is playing earnings management games. Sponsored by the Big Four accounting firms, the report, “Predicting Material Accounting Manipulations,” explains how to calculate a “fraud score” for public companies, thereby identifying which corporate books deserve further scrutiny.

A fraud score that exceeds 1.00 is a “red flag” indicating that a company may be toying with how it accounts for cash and accruals so that it can ultimately boost stock prices, says study co-author Weili Ge, an accounting professor at the University of Washington. The mathematical model, presented in the paper released in late June, hunts for abnormal patterns in five key areas where manipulation likely takes place: accrual quality (in terms of the number of accruals being booked), financial performance (including earnings growth, cash margins, and transaction management), nonfinancial performance (order backlog and employee head count), off-balance-sheet activities (operating leases and pension assumptions), and market-based measures (valuations and price-to-earnings ratio).

Essentially, the model detects mismatched metrics in each key area. For example, the math may flag a company that has recorded an atypical increase in the number of accruals even though it has declining cash margins, a situation that can occur when receivables and inventory numbers are manipulated. Similarly, the model draws attention to unusual increases in the number of operating leases a company carries. Since leases allow companies to record lower expenses early on in the life of the contracts, an abnormal spike may indicate that a manager is focused on “financial statement window-dressing,” says the report.

4 Powerful Communication Strategies for Your Next Board Meeting

4 Powerful Communication Strategies for Your Next Board Meeting

This whitepaper outlines four powerful strategies to amplify board meeting conversations during a time of economic volatility. 

The model also notes changes in employee head count and total assets. Managers attempting to mask deteriorating financial performance often reduce the number of workers to boost the bottom line. But if they also overstate assets, the drop in workforce numbers will seem out of sync with economic reality.

To be sure, the more a measure deviates from the model’s prescribed norm, the more likely the company is manipulating that variable. If enough variables are skewed, then the chance is greater that something fishy is going on, say the authors, who along with Ge include Patricia M. Dechow of the University of California-Berkeley’s Haas School of Business, Chad R. Larson of the University of Michigan’s Ross School of Business, and Richard G. Sloan of Barclays Global Investors. According to Ge, the calculations are meant to predict the likelihood of earnings manipulations, in much the same way bankruptcy risk models predict a company’s chance of falling into insolvency.

The researchers developed the model by examining 2,191 Accounting and Auditing Enforcement Releases issued by the Securities and Exchange Commission between 1982 and 2005. The releases are public documents that identify corporate accounting and auditing misconduct allegedly committed by executives and auditors. From that batch of releases, the study authors identified 680 companies with alleged manipulations in at least one of their quarterly or annual financial statements. From those companies, which were not named in the report, Ge and her colleagues concluded that:

• Revenue, which was overstated by 55 percent of the sample companies, is by far the most commonly misstated account.

• Manipulating reserves — including allowances for doubtful debts — is also common, occurring in 10 percent of the sample.

• Manipulating inventory and cost of goods sold occurred in 25 percent of the sample.

• In general, manipulations are clustered in certain industries, most commonly computers and computer services, retail, and general services — which include telecommunications and health care. This is likely because companies in those industries tend to be current or former high-growth organizations that need to maintain the status quo.

• Large companies, defined by market capitalization, are more likely to manipulate earnings. This conclusion is probably due to the SEC’s focus on investigating the most material and visible misconduct involving the largest losses and highest number of affected investors.

The authors uncovered some surprising findings, too. For instance, Ge says that, in general, when the companies misstated earnings, their rates of return were dropping, earnings growth rates were negative, and cash margins were shrinking. Yet cash sales were unexpectedly rising during the manipulation periods. The authors were puzzled because they expected managers to boost sales by manipulating accruals; that is, credit sales. A little more database digging revealed why cash sales were up.

Indeed, the companies that massaged earnings also tended to be expanding their capital bases and increasing the scale of their business operations — which led to an increase of both cash and credit sales. Also, many of the companies in question manipulated sales via transaction management — for example, by selling goods to related parties, forcing goods onto customers at the end of a quarter, and encouraging sales to customers with return provisions that violated the true definition of a “sale.”

In its current iteration, the model would probably miss this type of misconduct, say the researchers, mainly because it was set up to probe accrual-based measures, not manipulation of techniques designed to boost cash sales. However, the authors agree that their unexpected discovery is a “useful area for future research” into developing metrics to measure earnings quality that captures cash-based earnings manipulation.