|
|
|
|
EDITORS: Media copies of Professor Beneish's article in the Journal of Accounting and Public Policy are available from George Vlahakis, Office of Communications and Marketing, 812-855-0846 or 812-855-3911, gvlahaki@indiana.edu
BLOOMINGTON, Ind. -- In a recent issue, Forbes magazine posed a burning question. It noted, "Wall Street swarms with highly paid security analysts, loaded with advanced degrees and surrounded by hype. So how come there are so many unpleasant earnings surprises?"
The magazine went on to describe the case of Oxford Health Plans, a $2 billion health maintenance organization located in Norwalk, Conn., that announced a third-quarter 1997 loss of at least 83 cents per share, compared with per-share earnings of 33 cents the year before. The next Monday, its stock plunged from $76 to around $23 in just one day of trading, wiping out $4.2 billion in market value.
Another story of woe for investors was Sulcus Computer. Sulcus Computer disclosed that it was being investigated by the Securities and Exchange Commission and saw its stock price fall by 46 percent. The SEC subsequently reported that Sulcus Computer's financial statements in 1991 and 1992 materially overstated earnings by capitalizing expenses and by failing to record both expenses and liabilities.
Under a model developed by an accounting professor in the Kelley School of Business at Indiana University, investors involved in both scenarios could have avoided heavy stock losses by learning as much as two years early that the companies had been manipulating earnings reports.
These are two examples of how a model developed by Daniel Beneish, IU professor of accounting, can improve the speed and ability with which investors are able to detect companies that are manipulating their earnings figures. The results of his research appear in the fall issue of the Journal of Accounting and Public Policy.
Beneish's findings are important because, when companies are discovered to be manipulating earnings, their stocks plummet and it opens a Pandora's Box for anyone associated with the firm.
The forensic accounting model can also be used by auditors, lenders, securities analysts and regulators to quickly screen a large number of firms and identify those with the greatest potential for manipulation of their earnings reports. Beneish bases his findings on firms that were either charged by the Securities and Exchange Commission with violating Generally Accepted Accounting Principles (GAAP) or publicly admitted to violating GAAP.
"When the market learns that something is wrong with their accounting practices, the stock price drop exceeds the market movement by 20 percent on that day," he said of the average price response in his sample. "The amount of the surprise is perhaps indicative that nobody's watching in a systematic manner.
"The evaluation of the model suggests a systematic relation between financial statement disclosures, stock market variables and the likelihood of earnings manipulation," said Beneish. And it's simple. "Most of the data in the model can be extracted from an annual report.
"This model predicts manipulation a little bit ahead -- a year to a year and a half ahead of when the public discovers it. Those caught have typically manipulated earnings for two to three years before it is discovered," he said. "Indeed, the variables in the model also tell you about a slowdown in future prospects."
Beneish said that firms caught manipulating their earnings are generally in, or facing, tough times; they tend to be growth firms that probably manipulated the figures to dispel investors' perceptions that growth was decelerating; they tend to be younger firms and many infractions occur in the process of, or after, issuing securities.
The model contains variables intended to capture distortions in financial data produced by earnings manipulation. "The basic premise underlying these variables is that accounting is a double-entry system. That is, firms cannot inflate revenues or deflate expenses without simultaneously bloating an asset account," Beneish said.
For example, an influential variable in the model is the days sales in receivables index, which is computed as the ratio of days sales in receivables in two consecutive years. A large increase in days sales in receivables raises the likelihood that sales are overstated.
"While it could be the result of a change in credit policy to spur sales in the face of increased competition, disproportionate increases in receivables relative to sales are also suggestive of front-loading or other forms of sales inflation," Beneish observed.
The model includes similar variables that are intended to capture deterioration in gross margins and in asset quality, reductions in depreciation, and increases in general expenses. The analysis leads to a model somewhat akin to bankruptcy prediction models.
One limitation of the model is that public financial information must be available for the companies. Therefore, it cannot be reliably used in studying privately-held firms.
But with public companies, "it is very easy to implement."
Return to the Office of Communications Home Page