In our paper, Oligopoly, Disclosure, and Earnings Management, which is forthcoming in The Accounting Review, we theoretically examine whether firms bias their disclosures (manage earnings) to gain a competitive advantage in their product market. Our specific motivation comes from the claims of C. Michael Armstrong who was the CEO of AT&T from 1997 to 2002. In statements made by Armstrong, he argues that accounting fraud at Worldcom was the cause of AT&T’s perceived strategic failures, its inability to compete with Worldcom and, in the end, the decision to break up the company. He specifically suggests that Worldcom’s fraudulently reported revenues, margins and costs drove AT&T’s layoffs, cost cutting and a very unprofitable price war that left AT&T unable to service the debt he incurred to revive the company. This view is supported by William Esrey who was Sprint’s CEO during that time. We interpret Armstrong and Esrey’s argument as suggesting that a rival’s earnings management affected competition in the product market, led to a misevaluation of their relative performance and caused dramatic, potentially non-optimal changes in their business strategy.
We employ an incomplete information Cournot duopoly model in which each firm knows its own production costs but not its rival’s to evaluate this argument. In our model, each firm provides a disclosure through, for example, its income statement. The firm’s rival can use the disclosure to update its beliefs about the disclosing firm’s production costs prior to competing in the product market. Our model differs from prior work on disclosure in incomplete information Cournot models because we assume that firms can provide biased reports. However, if they do so, they incur a cost of misreporting.
We find that Cournot competitors bias their financial reports so as to create the impression that their costs of production are lower than they actually are. This bias leads to lower total production, a higher price and each competitor earning greater product market profits. These results obtain even though no firm is fooled by its rival’s disclosure. We also find that the magnitude of the bias (the amount of earnings management) is larger when firms compete in more profitable product markets but smaller when they can extract more information about their rival’s costs from their own. When the costs of misreporting are asymmetric, the lower cost firm engages in more earnings management than its rival, and it produces more and earns greater profits than it would in a full-information environment.
Our analysis also offers new, testable implications for the relationship between earnings management, reported and actual earnings and industry structure.
The full paper is available for download here.