Revenue Recognition: What is a Sale, and When Do You Book It?
It’s worth repeating: Flexibility, the great virtue of most accounting systems as practiced in most of the developed world, is also those systems' weak point. Thanks to their flexibility, most accounting systems can measure the financial impact of a wide variety of transactions, from the simplest retail exchange of cash for physical goods to complex trades of intangible assets whose value and useful life are uncertain. But where there is flexibility, there is also ambiguity, and where there is ambiguity, there is the opportunity for managers to abuse the accounting system.
In honest hands, most accounting systems are up to the job of providing a reasonably accurate rendering of a business’s financial performance. Accuracy, however, is not always the aim of those who manage a corporation or stand to gain from cooperating with management, be they board members or outside accountants, analysts or investment bankers. All may have an interest in manipulating the accounting system to produce a misleading, distorted, or downright fictional portrayal of a business’s health and prospects. And when the manipulation starts, it usually starts with the recognition of revenue (also called turnover or sales). Of the 720 restatements of earnings from 1997 to 2000 in the United States, 376—more than half—involved methods of revenue recognition. Restatements to correct improper revenue-recognition methods were reported by 183 non-computer-manufacturing companies, 116 software businesses, and 77 computer-manufacturing firms. As cheaters like to say, everybody does it.
The restatements were necessary because the original financial reports fell afoul of at least one of the two basic requirements of revenue recognition. First, the revenue must be “earned,” which is generally understood to mean that the sales process is complete. In the case of goods, legal title has been, or is on the verge of being, transferred to a willing buyer. In the case of services, the service in question has been rendered. The second requirement is that the seller has to be reasonably certain of collecting the money publicly claimed (or, in accounting parlance, recognized) as revenue. The history of accounting chicanery is in large part a history of many attempts—some of them more ingenious than others—to evade those two requirements.
One of the simplest revenue recognition games is also one of the most common. Essentially, it’s an attempt to dodge a rule known to every first-year accounting student: Goods shipped on consignment cannot be booked as revenue. Neither condition of revenue recognition has been fulfilled—ownership and its attendant risks have not been transferred, and since the goods might not even be sold, there can be no certainty of getting paid. But those strictures haven’t stopped some managers from using consigned goods to fatten the top line—that is, the revenue line—of the corporate income statement.
Sunbeam, the appliance maker run into the ground by Albert “Chainsaw Al” Dunlap, shipped barbecue grills to retailers during the fall and winter of 1996–97, even though consumer demand for outdoor cooking equipment is a strictly warm-weather phenomenon. To induce retailers to go along with the scheme, Sunbeam allowed them to defer payment until they had sold the grills. They were also assured they could return any unsold goods. As a final fillip, Sunbeam picked up the tab for the warehouses where the grills were stored for the winter. In short, the grills were sold on consignment. Nonetheless, Sunbeam accounted for them as if they had been sold in the last quarter of 1997.
Sunbeam’s exertions, known generically as “channel stuffing” (in this game, a company stuffs its distribution channels with merchandise, which it then claims to have sold) added $71 million to the company’s 1997 profits. The company later had to restate those amounts after Dunlap resigned from Sunbeam under pressure from the board of directors. In September 2002, Dunlap agreed with the Securities and Exchange Commission that he would never again serve as officer or director of another public company and paid $500,000 to settle the government’s civil case against him. He also paid $15 million to settle a shareholder class action. As is usual in such cases, Dunlap settled without admitting or denying guilt, but had the charges against Chainsaw Al gone to trial, he might have faced much stiffer penalties, given the current climate of revulsion from CEO excesses.
Financially ailing, suffering from years of mismanagement, Sunbeam fits the stereotype of the sort of company that would resort to a crude scheme like channel stuffing. But even apparently healthy, well-managed companies such as BristolMyers Squibb have been caught at the same game. In 2000 and 2001, the pharmaceuticals giant boosted sales by offering discounts to wholesalers to purchase more drugs than patients needed. In the first quarter of 2002, when those wholesalers stopped building up inventory, BristolMyers Squibb’s sales plummeted. In October 2002, the company restated its 2000 and 2001 results, a move that affected more than $2 billion in revenue.
Other revenue-recognition games are far more sophisticated than those played by Sunbeam and BristolMyers Squibb. Some are downright elegant. But the difference between channel stuffing and fancier methods of prematurely recognizing revenue is one of degree, not of kind. At bottom, most revenue-recognition games are variations on a few simple themes.
Truth or Consequences: Why Companies Cheat
Companies often attempt to pass off revenue-recognition games as mere differences of opinion. Accounting, they say, is full of judgment calls, and management, far from trying to inflate revenue, is merely making a judgment about which reasonable people can disagree. It is true that, in some transactions, the moment when revenue has been earned cannot be fixed with precision. Because the consequences of management’s decisions about revenue recognition can be so enormous, though, managerial protestations of pure intentions sound more than a little disingenuous. One set of managerial decisions can result in revenue reports that suggest high growth and strong customer demand—the sort of reports that encourage bankers to continue lending and investors to bid up the shares of the company doing the reporting. Another set of decisions yields a revenue figure that suggests slow growth, grudging customer acceptance, dubious future prospects. The latter set may send investors a more accurate signal about the future of the business, but how many managers are going to opt for accuracy over optimism, especially when their compensation consists mainly of stock and stock options, which grow more valuable as the stock price increases?
That’s why it’s so important to know a company’s revenue-recognition policies. For example, how does a company recognize revenue when a customer takes delivery of a product but makes payments on it over several years? One approach is to consider all of the revenue as earned when the product is delivered. But a more prudent approach is to consider such things as the costs of servicing and supporting the product. Are they incurred all at once, or are they spread out over the life of the product, so that they should be matched against the stream of customer payments? And what of the customer’s ability to meet its long-term commitments? How should a software company treat a ten-year agreement to provide e-business support to a client that is burning through millions of dollars a quarter and has no product on the market, much less revenue on its income statement? In the late 1990s, many suppliers treated such agreements as sure things and booked the revenue as if ten years’ contract payments had already been collected. To a large extent, the collapse in stock prices after March 2000 represented the belated recognition by corporations and their shareholders that much of the revenue so confidently booked during the boom times would never materialize.
In addition to deciding when to recognize revenue, managers have discretion to define what to designate as revenue. Suppose an auction business sells an item for $100. Of that amount, $5 goes to the auctioneer as commission. On its financial statements, should the auctioneer include the total amount of the sale as revenue and call the $95 payment to the item’s original owner an expense? Or should it count only the commission as revenue and show no expense? Most accountants would prefer to treat only the commission as revenue. The auctioneer is merely a conduit for the payment from buyer to seller; the money it passes to the seller doesn’t come from its own pocket. All the same, some Internet companies, recognizing the importance investors place on sales growth, took advantage of ambiguities in U.S. accounting rules to treat the gross proceeds of auctions as revenue, a procedure known as grossing up.
Companies get away with grossing up because there are instances where claiming the gross proceeds of a sale as revenue is perfectly reasonable and justified. The existence of a legitimate precedent creates a penumbra of “reasonable doubt” around more aggressive accounting moves. Suppose Dell Corporation sells, as a component of one of its PCs, a computer monitor it purchased from an independent manufacturer. Does it recognize as revenue only its gross profit margin—the difference between what it pays for the monitor and what it charges its customer? Or does it recognize the price of the complete PC as revenue, treating the cost of the monitor as an expense, just as Ford recognizes the full price of a car as revenue and counts as expenses the costs of the parts it purchases from outside suppliers? In such a case, Dell recognizes the full price of the monitor, and rightly so. But what if Dell were to arrange—as it often does—for the monitor to be shipped directly from the manufacturer to the customer? Should Dell include the monitor’s selling price in its revenue, or only Dell’s cut? In other words, should Dell’s sales figures suffer just because of an efficient logistics arrangement? Or should the decision hinge on a legal question, such as who would be responsible if the goods were damaged in shipping? Dell has resolved the ambiguity by recognizing the gross proceeds of the transaction as revenue, since its money and reputation are at risk if the product is damaged in shipping or fails to perform as advertised. That’s a far cry from an auctioneer laying claim to money that it merely funnels, risk free, from one side of a transaction to another.
Grossing up was one of the dubious accounting practices that became almost routine during the Internet boom, and no company indulged in it more enthusiastically—or controversially—than Priceline.com, the company that invited customers to “name your own price” for airline tickets, hotel rooms, and rental cars. In a 2000 quarterly filing, the company reported revenue of $152 million. Priceline arrived at that figure by grossing up—summing the full amount customers paid for those tickets, rooms, and cars. Like any travel agency, though, the company kept only a small portion of gross bookings—the spread between the customers’ accepted bids and the price it paid to travel and lodging providers. In Priceline’s case, that spread was $18 million, meaning that it claimed $134 million in revenue that it actually passed on to various providers, booking the payments as expenses.
Priceline’s accounting policy prompted skeptical questioning from the SEC and some stock analysts. Yet Priceline persisted. Why? Perhaps Priceline’s senior managers actually believed their public pronouncements that as “merchant of record” the company assumed all the risks of ownership. But there is another possible explanation: the rewards that were lavished on companies reporting rapid revenue growth during the Internet boom. Investors paid stratospheric prices for the shares of companies that had never recorded a profit and weren’t likely to do so in the foreseeable future. To justify this evident folly, many investors took to valuing Internet companies on the basis of their sales, theorizing that “critical mass”—a large customer base—was more important than profit. Grossing up was a cheap and easy way of showing sales growth. The few skeptics who complained that Priceline’s grossed-up revenue reports wildly overstated the company’s performance were told they “didn’t get” the New Economy.
Grossing up is not a practice exclusive to Web merchants. In 1999, Professional Detailing, a New Jersey–based firm that recruits and manages sales personnel for drug companies, started counting as revenue the reimbursements it received from clients for placing help-wanted ads. In serving as a conduit between its clients and the newspapers and other media where the ads were placed, Professional Detailing was no different from an advertising agency that uses client funds to buy television ad time. No advertising firm would count the client funds as revenue—the agency is only a way station for the money, not its destination. Similarly, Professional Detailing had no legitimate rationale for treating as revenue the want-ad reimbursements from its clients. Its auditors demanded that the company restate its 1999 results and report 5 percent less revenue than originally claimed. In less than a month after the restatement, Professional Detailing’s share price fell 31 percent.
Edison Schools, the ostensibly for-profit education company that has never made a profit, played a similar game with a contract to run the Philadelphia public school system. Edison booked $30 million in revenue from that contract, even though the company paid $21.3 million of the total straight out to teachers and other school personnel. To potential sources of finance to the money-losing company, revenue of $30 million gives a far more favorable impression of the company’s health and prospects than the more reasonable and accurate revenue figure of $8.7 million ($30 million less $21.3 million).
Of course, no discussion of grossing up would be complete without mentioning Enron, which treated as revenue the entire amount of the energy contracts it traded. As a result, Enron slotted in as fifth largest U.S. corporation on the 2001 Fortune 500 (published in 2002), with $139 billion in revenue, even though it had recently filed for Chapter 11 bankruptcy protection. If Merrill Lynch, number 36 on the Fortune ranking, had grossed up all the securities transactions it handled for customers in 2001, it would have reported revenue in the trillions of dollars—enough, obviously, to move Merrill to the number one spot.