The U.S. Congress heard rumblings of corrupt business behavior by Enron, Tyco, and other large corporations, the U.S. Congress in 2002, and in response, it passed the Sarbanes-Oxley Act (SOX). The act gave the federal government powers over corporations that had traditionally been the province of individual states. It also thrust significant amounts of extra responsibility for corporate governance, including criminal liability, on CEOs.
But new research suggests that the legislation has fallen short of its goal of controlling illegal business practices. Two papers coauthored by Thomas Lys (Professor of Accounting Information and Management at the Kellogg School of Management) cast doubts on the efficacy of the legislation. “Our data tend to support the fact that SOX was necessary, because people were really cooking the books,” Lys says. “But the questions are: Should it have been this act? And was the cure worse than the disease? There’s some evidence that the legislation was detrimental to the U.S. economy and the competitiveness of U.S. markets.”
Lys and two colleagues—Daniel Cohen (Assistant Professor of Accounting at the Stern School of Business, New York University) and Aiyesha Dey (Assistant Professor of Accounting at the Booth School of Business, University of Chicago)—examined the impact of SOX on two aspects of corporate behavior: management of earnings and the effect of CEOs’ compensation on their willingness to take business risks. The study of earnings management convinced the researchers of the need for some sort of control.
“The most significant thing is that everybody was cheating during this scandal period; it wasn’t just the problem of a few bad apples,” Lys says. “If you wanted to make a case for regulation, it was an easy case to make.” However, they found that companies can still use earnings management strategies to mislead investors, and that SOX reduces corporations’ willingness to take risks, outcomes that raise serious questions about the value of the legislation. “We suggest,” Lys says, “that a more considered version of SOX may have been better.”
Earnings Management Pre- and Post-SOX
Lys and his colleagues outline the aim of their first study: “It was unclear whether there really was a widespread breakdown of the reliability of financial reporting prior to the passage of SOX or whether the highly publicized scandals were isolated instances of individuals engaging in blatant financial manipulations,” they write. “And if it were the former, how did the passage of SOX affect firms’ reporting practices?” To answer those questions, they examined the prevalence of accrual-based and real earnings management before and after SOX became law.
Accrual accounting registers sales when they are made, regardless of when they are actually collected. Real earnings management, by contrast, records only actual payments. Why should a firm choose one over the other? “You change the accounting method to get the desired earnings,” Lys explains. “Real earnings management means that you make actual transactions. You might sell something to record a gain that you might not otherwise have considered selling to book a profit, or if you want a lower net income you might sell an asset that has declined in value to realize a loss.”
“But even more problematic is cutting R&D (research and development) expenditures,” Lys says. “Because R&D is expensed under GAAP (U.S. Generally Accepted Accounting Principles), a reduction in R&D increases current earnings but has the potential of reducing long-term profitability. So from the perspective of shareholders, real earnings management is a much more expensive way of managing earnings.” SOX has also made accrual accounting “more expensive,” Lys says. “As a result, SOX makes real earnings management seemingly cheaper to management.”
Using publicly available financial reports and related data, the team examined earnings management overall before the scandals emerged, between that time and the passage of SOX, and after the legislation. “Rather than focusing on offenders, we focused on the market,” Lys explains. “We found that was much more widespread than most people believe it was.”
The team also studied some individual cases. They examined the pre- and post-SOX earnings management of a subset of firms that they regarded as more likely than others to have managed their earnings to meet or beat previous years’ earnings or analysts’ forecasts or simply to avoid reporting losses. “We looked at ‘suspect’ firms that were either indicted or had allegations of impropriety against them,” Lys says. When Fortune magazine published a list of thirty disgraced firms, the metric that the team had developed correctly identified twenty-nine of them. The team’s metric missed only one firm: Enron. “Enron worked in a very sophisticated manner,” Lys explains. “They were really smart. They were good crooks—the smartest people in the room.”
The data for firms overall and the suspect group yielded a clear conclusion. Accrual-based earnings increased before passage of SOX, but then declined thereafter. Firms were prone to switch to real earnings management that disadvantaged shareholders but are more difficult to detect. The change in suspect firms’ earnings behavior was particularly noticeable—they tended to rely more on real earnings management post-SOX than they previously had under similar conditions.
The data also revealed what Lys describes as “a very large correlation between the percentage of compensation from stock options and the amount of earnings management.” He sees two possible reasons for that. “When you have lots of stock options, managements have a big incentive to manipulate, as a small change in the stock has a large impact on their wealth. Also, companies that are very risky have more stock options and are also more likely to have bigger accounting effects.” Whatever the cause, Lys draws one firm conclusion from the data. “If I were on the board of a company that offered a lot of stock options,” he asserts, “I would want to make sure that my audit committee is working diligently.”
In its second study, the team set out specifically to find SOX’s impact on CEOs’ compensation contracts, as well as CEOs’ responses to any changes by reducing their firms’ investments in risky projects. Again using public sources, the researchers examined several variables: changes in total compensation; individual components of that compensation such as salaries, bonuses, and option grants; and the ratio of incentives to fixed salaries.
The data showed a clear pattern. While overall compensation levels did not increase after SOX passed Congress, Lys and his colleagues did find that salaries and bonuses grew fatter while option grant levels dwindled. The ratio of incentive compensation to fixed salary also dropped. The team reasons that those changes stem from the heavy responsibilities that SOX places on CEO. “This shift is likely to represent firms’ response for shielding executives from some of the risks imposed by SOX,” the paper states. In fact, the team computed the change in CEO wealth when CEOs increase the risk of their firms. The data clearly show that boards of directors reduced the incentives to invest in risky projects following SOX.
The team also calculated corporations’ risky investments, defined as the sum of R&D expenditures, acquisitions, and net capital expenditures. After all, the team notes, “One potential deleterious effect of governance reforms such as SOX is reduced risk-taking activities by managers—incentives to take risks are reduced when managers face penalties for bad outcomes.” The result: CEOs made significantly fewer risky investments after SOX was passed. Thus, managers reduced investments beyond what was expected from the changes in compensation contracts.
The results, of course, raise the possibility that other events, such as the stock market crash of 2000-2001 or the passage of SFAS 123R (Statement of Financial Accounting Standards 123R, which mandated the expensing of stock options) are responsible for the results in the study. However, the results clearly point to SOX—through a change in incentives, the act induced boards to encourage their CEOs to take less risk and also created incentives for managers to take less risk. As a result, firms reduced their investments in risky endeavors for two reasons: changes in compensation and changes in personal liability.
Beneficial or Detrimental?
Doubt also remains about SOX’s broad role in changing corporate behavior. For at the same time the legislation passed, the FBI started to arrest corporate malefactors in a very public way. “Was the reason for the changes SOX or seeing your neighbor schlepped away in handcuffs on the evening news?” Lys asks. “The arrests emphasized the personal cost of cooking the books.”
What overall conclusions does Lys’s team reach about the legislation as a result of its two projects? “We have a very strong suspicion that SOX was detrimental,” he says. “Regulation sometimes goes overboard.” But he does not hold much hope that Congress will remedy the weaknesses exposed in the SOX legislation. “Politicians never make a mistake,” he says. “To compensate for a piece of bad legislation, they produce more legislation.”