by investors. What is front-loading of revenues, you ask? Complex software programs, such as i2’s, usually require significant updating and servicing over an extended time after sale. Recording the total value of the software package, including payments for future services, as revenue at the time of sale is front-loading. But is front-loading a lie? No, if the relevant fact is a finalized software sale, and i2’s front-loading indeed related to finalized sales. However, GAAP—the accounting “law” in the United States—requires the allocation of revenue and income from multiperiod projects over the respective delivery and subsequent service periods (quarters, years) to properly match periodic revenues and expenses. Recording the total revenue of the software in the delivery period and the cost of servicing and updating it in subsequent periods distorts such matching and the consequent income number. Thus, according to GAAP, i2 Technologies’ front-loading of revenues violated truthful reporting and was accordingly sanctioned by the SEC.11 That’s how GAAP substitutes for the missing or hazy facts, in the establishment of truth or falsehood in financial reporting. A GAAP violation is manipulation.
This may seem clear-cut, but nothing in accounting is. There are no clear criteria in GAAP for making reliable estimates, except for the general dictum—use the best information available—which is as helpful as “buy low and sell high.” This is serious, because most financial information items are based on managerial estimates. Not surprisingly, studies have shown that many manipulations of financial information, particularly the “fine tuning” of EPS to beat the consensus analysts’ estimate by a penny or two, are done through “massaging” accounting estimates and projections.12 The upshot is that GAAP is often an insufficient benchmark for investors and regulators to establish truth or trickery in financial reporting, particularly for emerging industries, such as software in the 1980s, biotech in the 1990s, and the Internet services in the 2000s, and for new technologies, such as risk hedging or securitization of intellectual property, where GAAP is lagging considerably behind business developments. For example, in 2002, the FASB started its “revenue recognition” project to update the measurement rules for this key income statement item. As of early 2011 (yes, nine years later), work on this project continues.
GAAP’s inadequacy exposes managers and directors to considerable risk: what may seem truthful reporting to them, the SEC or class-action lawyers might claim later on to be fraudulent.13 AOL is an early case in point. Starting in 1995, the company capitalized, justifiably in my opinion, certain advertising and promotion expenses, claiming that they create an asset—customer franchise—generating future benefits. Trial lawyers and the SEC challenged this practice, which reduced substantially AOL’s reported losses, as manipulative, finally forcing the company in 1998 to reverse the capitalization and write off (charge to expenses) close to $400 million of customer acquisition costs. This early experience with the hazards of intangibles’ capitalization had detrimental effects even when GAAP was clear. The software industry provides an example. GAAP in FASB Statement 87, 1986, calls for the capitalization of software development costs—recognizing them as an asset, rather than an expense—when a project successfully passes the technological feasibility stage. Many software companies, however, particularly the large ones like Microsoft, expense all their software development costs, lest analysts accuse them of earnings manipulation. And yet the evidence shows that software capitalization, when justified, yields earnings that better predict future performance than software expensing.14
Thus, except for clear-cut frauds, such as Parmalat’s claim of having billions of dollars in an empty bank account, or obvious GAAP violations, such as WorldCom’s multibillion-dollar capitalization of regular expenses in the late 1990s, many cases of managed or manipulated information are in a gray area, difficult to ascertain and prosecute. This vagueness poses a serious challenge to managers wishing to make sure that the information they disclose and certify, a Sarbanes-Oxley requirement, is truthful, and to financial information users who rely on the presumed integrity of the information. A thorough analysis of the motives to manipulate financial information and the means of manipulation—both vividly demonstrated by the following cases—goes a long way in drawing a bright line between truth and falsity in financial reporting.
Why and How They Do It
The closest one gets to a proven manipulation (defendants rarely admit wrongdoings) are the SEC’s Accounting and Auditing Enforcement Releases (AAERs), which report on cease-and-desist orders concerning accounting and financial reporting issues imposed on companies, their managers, and auditors for violations of the 1933 and 1934 Securities Acts. These actions are generally the culmination of elaborate and extended investigations by the SEC. Here is a varied selection of such actions from the early 2000s, a period ripe with manipulations, each followed by my message for managers.
GE Brings Good Things to Life, Yet Not to Accounting
In 2009, the SEC filed a suit against General Electric, alleging that it had misreported revenues and earnings during 2002 and 2003 in order to beat analysts’ consensus earnings estimates, partially explaining a “world record.” From 1995 through 2004, GE met or exceeded the consensus estimate in practically every quarter: forty consecutive hits! The SEC alleged several manipulations, some of which, while involving millions of dollars, require an advanced accounting degree to fully comprehend. One was a plain-vanilla accelerated revenue recognition scheme. In the fourth quarters of 2002 and 2003, GE purportedly sold locomotives to financial institutions (heavy users of locomotives) to the tune of $223 million and $158 million, respectively. These transactions, the SEC alleged, were not real sales and shouldn’t have been recorded as revenue, because they were structured in a way that didn’t transfer the risk of ownership to the institutions, a GAAP requirement for revenue recognition. The financial institutions were, of course, expected to subsequently deliver the locomotives to real customers, but GE needed immediate delivery to boost fourth-quarter sales and earnings to beat the consensus estimates. GE settled the SEC’s lawsuit by paying a $50 million penalty without, as is customary in such cases, admitting to wrongdoing.15
The major lesson learned is that many manipulative schemes cannibalize future revenues and earnings, necessitating an ever-increasing intensity of trickery, soon to spiral out of control. This was evident in 2007 when GE restated its 2002 and 2003 reports for the locomotives “sales,” stating that the 2002 fourth-quarter segment revenues and profits were overstated by 8.8 percent and 14.6 percent, respectively, while the corresponding 2003 overstatements were substantially higher: 22.6 percent and 16.7 percent, respectively. This cannibalization renders most reporting manipulations unsustainable.
How Gateway “Immunized Itself from the Vagaries of the Market”
In 1999, Gateway Inc., a direct marketer of personal computers and related products, embarked on a diversification strategy—dubbed venturing beyond the box—offering software, Internet access services, and training and support programs to customers. By the end of 1999, Gateway’s beyond-the-box income reached 20 percent of total earnings, seemingly demonstrating the success of the diversification program. Soon, however, things turned ugly. The tech bubble burst in 2000, the economy fell into a recession, and demand for computers and related products plummeted. Gateway, however, continued to present a happy face. While competitors reported a significant softening of demand and the consequent reduction in revenues and earnings, Gateway managers consistently claimed they were bucking the trend and reported increasing revenues and earnings to boot. How did Gateway manage, in the words of an analyst, to immunize itself from the vagaries of the market?
The SEC provides the answer.16 In late 1999, Gateway initiated a program to sell computers on credit to persons who were previously rejected because of poor-credit status. At first, management characterized the program “a test” and limited it to $10 million. Soon, however, as Gateway’s revenues declined and the specter of missing analyst estimates loomed large, it aggressively accelerated the program. In the second quarter of 2000, this high-risk sales drive—for
