Local Customs: Industry-Specific Revenue Games
Each industry has its own special ways of accelerating revenue recognition. In the software business, companies have been known to book revenue from upgrades that have not even been produced yet, much less provided to the customer. While one might concede that some judgment is required to establish revenue-recognition policies in such businesses, it should be clear to managers and auditors alike that revenue can never be recognized until it is earned. And how can revenue be earned when the product supposed to generate revenue does not even exist?
The soft-drink industry has a few games of its own. One of the authors heard about one from a former chief financial officer of a large Coca-Cola bottler in South America. Whenever Coca-Cola wanted to boost sales in the country, it would ship massive quantities of concentrate to the bottler. It even kept refrigerated trailers on the bottler’s premises for just such an occasion. The bottler incurred no obligation to Coca-Cola until it was ready to use the concentrate. Thus, while the concentrate was physically transferred and invoiced to the bottler, none of the unordered quantities were included in its inventory, and it was not expected to pay any interest to Coca-Cola on the unpaid invoice. The sole purpose of the practice was to allow Coca-Cola to artificially boost revenue, practically at will.
Why haven’t Coca-Cola’s auditors pointed out this transparent charade and demanded that the company put a halt to it? The answer is unclear, but it may be worth pointing out that the same firm that audits Coca-Cola worldwide, Ernst & Young, also audits the South American bottler.
Another common ruse, found in many industries, is to misclassify the gains that arise from selling off valuable real estate and other corporate assets. Such one-time gains should clearly be segregated from regularly recurring revenue and, if large enough, reported as separate line items on the income statement. But some companies try to classify one-time gains as operating income. This maneuver is not reserved for marginal, fly-by-night operations. In 2002, IBM received embarrassing press for trying to pass off as operating income a $300 million gain from the sale of a business. To cynical minds, it appeared that IBM was trying to fool investors and bankers into believing that the company’s underlying, day-to-day business operations were more profitable than they really were.
Back when the stock market was throwing money at anything Internet related, Web site operators played a game they called “concurrent transactions.” For centuries, the word “barter” sufficed to express the same concept. In a concurrent transaction, two companies would trade banner spots on each other’s Web site, recording the deal as both revenue and expense. The net effect on profit was zero, but at least the companies could report revenue to convince investors that there was some substance to their business. GAAP forbids this practice except under highly restrictive conditions, but many Internet businesses did it anyway, and few analysts, investors, or regulators called them on it. Nor has the practiced died out with the dot-coms. Siebel Systems took a big hit in its share price in 2002 when investors realized that nearly one-fifth of the software developer’s second-quarter revenue came from concurrent transactions, a big number in itself and a sharp increase from the second quarter of 2001.
There are real costs associated with adopting revenue-recognition policies so aggressive they must eventually be reversed. In October 2000, the UK’s Cedar Group, which was trying to bring a stock offering to market, received unfavorable media attention for its revenue-recognition policies. With investors uncertain about the company’s true value, the company’s shares lost a third of their value. The underwriters that were handling Cedar Group’s stock offering found themselves holding shares for which they’d paid a 28 percent premium above market price. Cedar subsequently adopted revenue-recognition policies consistent with U.S. standards, transforming an 8.5-million-pound profit into a 24.4-million-pound loss.
A small amount of revenue manipulation can have a substantial impact. In 2000 and 2001 AOL recorded $190 million in revenue that was really nothing more than advertising fees paid by some of AOL’s own divisions—garden-variety intercompany transfers of funds, which are, of course, not revenue. The revenue was minuscule relative to the company’s overall revenue, but by recognizing it when it did, AOL was able to report that ad revenue was growing just as it was preparing to acquire Time Warner. By misrepresenting the state of its business, AOL was able to extract more favorable terms from Time Warner than it could have otherwise. Indeed, had Time Warner known the true state of AOL’s business, the deal might not have gone through at all.