In Cooking Their Books, Corporate Accountants Like a Good Recipe


steaming pot
John S. Dykes /

Every year, hundreds of US companies release documents called restatements, in which they acknowledge having made accounting errors in past earnings reports. These companies, which have typically overstated their earnings, often suggest that their bookkeepers made honest mistakes — an explanation that generally appears to satisfy the federal Securities and Exchange Commission, which punishes less than 15 percent of firms that issue such restatements.

But a new study by researchers from Columbia, Rutgers, and Nanyang Technological University in Singapore suggests that something more troubling is taking place. In analyzing 2,376 restatements that US companies issued between 1997 and 2008, the researchers discovered that when a prominent company admits to mismanaging its books, many of its competitors will begin mismanaging their own books in similar ways soon afterward. These firms will go on to overstate their earnings for about two and a half years, on average, before issuing their own restatements. The researchers say that these episodes of “restatement contagion,” as they call it, occur only if the company that initially admits to a particular accounting error doesn’t get disciplined by the SEC, sued by its investors, or criticized in the press; if the company is hit with any of these consequences, its peers rarely mirror the mistake.

“Many of the errors that get repeated, we believe, are just crafty means of inflating income,” says Shivaram Rajgopal, an accounting professor at Columbia Business School who is among the study’s authors. “A company might declare revenue from a multi-year contract a year early or neglect to devalue its old, unsold inventory.”

Rajgopal and his coauthors, whose paper appears in the Accounting Review, do not claim to be able to prove that any specific US corporations intentionally overstated their earnings over the twelve-year period covered by their study. But they say that their analysis leaves no doubt that US companies are routinely mimicking the purported mistakes that their competitors get away with — essentially using other firms’ financial restatements as crib sheets for their own bad behavior.

“The patterns that we’ve detected, when looked at statistically, could not possibly have resulted from pure chance,” Rajgopal says. “Our analysis shows that financial fraud is much more widespread than previously appreciated.”

According to Rajgopal, executives at publicly traded companies face powerful incentives to overstate their earnings and are often willing to do so even if it seems likely they will eventually need to correct their numbers. He says that his paper shows that corporate executives, in looking for ways to boost the bottom line, have become highly adept at anticipating the types of accounting errors that SEC investigators will believe are accidental. He suggests that the SEC, in order to keep companies off balance, should adopt analytic techniques similar to those used by his research team to periodically identify and go after copycat fraudsters.