In 2002, the Sarbanes-Oxley Act (SOX) set stringent new accounting and board standards for publicly held US companies in the wake of widespread fraud at Enron, Tyco, WorldCom, and a number of other large corporations. Legislators, business leaders, and the press, as well as academics, have argued that SOX has resulted in higher quality audit reports and increased the cost of earnings management, which creates stronger incentives for companies to avoid the detection of earnings management. Academic researchers have generally concluded that SOX has led companies to switch from earnings management through accrual manipulation to (arguably) costlier earnings management through real activities, because the latter is more difficult to detect than accrual manipulation.
There are significant differences between the two types of earnings management. Accrual-based earnings management exploits the subjectivity inherent in the estimation of certain balance sheet and/or P&L items — inherently subjective because this means estimating tomorrow’s uncollectible receivables today. In contrast, real earnings management involves the use of real transactions, for example, pulling in sales to the current period from the next period or cutting discretionary expenses such as advertising and R&D.
Professor Nahum Melumad and his colleague Itay Kama of Tel Aviv University were skeptical about whether SOX has really resulted in a shift toward real earnings management. “Real earnings management is costly to employ,” says Melumad, who frequently investigates aspects of earnings management and other forms of corporate reporting. “If a company wants to manage earnings in a manner that is difficult to detect, it is more likely to resort first to another available, often less costly, method of disguising earnings management: managing cash to replicate the impact of a true improvement in performance and thereby masking earnings management.”
Two common red flags that signal potential earnings management are (1) cash that does not follow earnings and/or revenues and (2) accruals (defined as the differences between earnings and cash) that do not correspond to earnings and/or revenues. If indeed the incentives to avoid detection of earnings management have intensified post-SOX, then one option available to firms is masking earnings management by assuring that changes in cash closely follow changes in earnings/revenues and that accruals are closely associated with changes in earnings/revenues. One means of achieving this is by converting accruals into cash to mimic impact of a true sales increase or expense decrease on cash and accruals. Kama and Melumad label such activity accruals conversion cash management.
Kama and Melumad employed an analytical framework to illustrate the identifiable traits in financial reporting of attempts to camouflage earnings management by using accruals conversion cash management. They introduce new proxies for accruals conversion cash management that result in camouflaged earnings management. Specifically, they show that the forward variation of key financial variables (a measure of how much the variable fluctuates over subsequent quarters), like the forward variation of cash flow from operating activities relative to sales, increases significantly after companies attempt to camouflage earnings management. They analyzed over 120,000 quarterly financial disclosures of about 6,000 publicly traded companies for several years before and after the passage of SOX, to determine changes in key financial variables. The results support their assertion that companies have increased the use of accruals conversion cash management to disguise their earnings management efforts since the enactment of SOX.
While Kama and Melumad’s work was prompted by questions raised by the passage of SOX, the research has broader implications. Analyzing the delayed effects on the key financial variables the researchers identify here could help analysts, regulators, and other market participants detect suspect earnings management. Additionally, the work suggests that relying exclusively on the commonly analyzed indicators that ignore the option of camouflaging earnings management could result in misleading inferences about the scope of earnings management activity.
Nahum Melumad is the James L. Dohr Professor of Accounting and Business Law in the Accounting Division at Columbia Business School.