Other common characteristics of firms who manipulate financial results include unusually high growth in cash sales but declines in cash profit margins and earnings growth; declines in order backlog and employee headcount; and abnormally high increases in financing and related off-balance sheet activities such as operating leases.

Those were the findings from a comprehensive analysis of of Securities and Exchange Commission accounting and auditing enforcement releases, which the agency issues to document enforcement actions against companies, auditors, and officers for alleged accounting misconduct. Dechow and her coauthors examined more than 2,000 SEC releases from 1982 to 2005, which resulted in a final sample of 680 firms alleged to have manipulated financial statements.

The analysis was conducted by Dechow; Weili Ge of the University of Washington Business School; and Chad Larson and Richard Sloan, both from the University of Michigan's Stephen Ross School of Business.

Dechow and her coauthors outlined the results of their analysis in a recent working paper, Predicting Material Accounting Manipulations. Their research was sponsored by the Research Advisory Board established by the Big Four accounting firms to develop a model to help identify firms that manipulate earnings or commit fraud.

"A consistent theme among manipulating firms is that they have shown strong performance prior to the manipulations," the researchers noted in their paper. "Manipulations appear to be motivated by managements' desire to disguise a moderating financial performance."

Managers may want to disguise such performance to ensure their stock-based compensation remains valuable or to raise capital at better prices, Dechow notes.

Based on the research, Dechow and her co-authors devised a so-called Fraud-Score, or F-Score, to be used by investors, auditors, and regulators as a preliminary assessment of "earnings quality" to determine whether further investigation into possible fraud is warranted. Enron, for instance, received an F-score almost twice a high as the average firm.

Overall, alleged manipulations are more common in large firms. About 15 percent of the manipulations occur in the largest 10 percent of firms, most likely because of the SEC's incentive to identify only the most material and visible manipulations involving large losses to numerous investors.

In addition, more than 20 percent of manipulating firms were in the computer industry, but the computer industry only comprised 11.9 percent of public companies. Retail firms made up 13 percent of manipulating firms, compared with 9.7 percent of public companies. And service firms such as telecommunications and health care made up 12.4 percent of manipulating firms, compared with 10.4 percent of public companies.

The most manipulations occurred in 1999 and 2000, perhaps because slowing growth during this time gave managers incentive to manipulate earnings.

Other findings of the research include:

• Investors have abnormally high expectations about the future growth opportunities of manipulating firms, as evidenced by unusually high price-earnings and market-to-book ratios prior to manipulations.
• Manipulating firms tended to have abnormally low free cash flows. Many firms were actively seeking new financing to cover negative operating and investing cash flows. 
• Cash sales, surprisingly, increased during manipulations. That's because many firms (Coca Cola, Sunbeam, Computer Associates) allegedly front-loaded their sales and engaged in unusual transactions at the end of the quarter. 
• More firms issued either debt or equity in years in which they manipulated financials compared with other years. And cash from financing more than doubled during manipulating years compared with other years.
• Companies engaged in abnormally high leasing activities during manipulation periods, consistent with managements' increased use of the flexibility granted by lease accounting rules to manipulate their firms' financial statements. 
• Accruals increase in manipulating years. Accruals are the difference between reported earnings and actual cash flows, so high accruals indicate more accounting adjustments being made to boost earnings. 
• Revenue is by far the most commonly manipulated line item on income statements, with 55 percent of sample firms allegedly manipulating revenue. Types of revenue manipulations include front-loading sales from future quarters (Coca Cola and Computer Associates), creating fictitious sales (ZZZZ Best), incorrect recognition of barter arrangements (Qwest), and shipping goods without customer authorization (Florafax International).
• Manipulations of inventory and cost of goods sold occurred in 25 percent of sample firms. Manipulations of allowances, including the allowance for doubtful debts (an estimate of how many customers who purchased goods on credit will not pay), occurred in 10 percent of sample firms.

Dechow's latest research builds on her previous work, which examined how corporate governance is correlated to accounting manipulations. In that work -- first published in 1996, years before the most recent round of accounting scandals --– she found that manipulating firms have a higher number of insiders on the board and a CEO who is more powerful and entrenched.