Wrong Number: Telecom Tricks
The telecommunications industry had its own bizarre take on revenue recognition during the boom. From 1997 to 2000, Global Crossing took on over $7 billion of debt to lay 1.7 million miles of fiber-optic cable to transport data via the Internet. When completed in summer 2001, the network spanned 27 countries and 200 major cities around the globe. The company’s debt load didn’t seem to faze investors—Global Crossing’s market capitalization reached $40 billion in 1999. But then other carriers entered the market, worldwide economic growth began to slow, and Internet usage, while growing fast, was not taking off quite as fast as company management had expected.
As a result, demand for Global Crossing’s fiber-optic capacity began to wane. Fearing that deteriorating financial performance would cause its share price to collapse and call into question the company’s ability to service its debts, management began concocting revenue from capacity swaps with other carriers. In one such swap, executed in the first half of 2001, Global Crossing “sold” $100 million of capacity to Qwest Communications, which was also suffering a demand slowdown, while “buying” an equal amount of capacity from the same firm. The $100 million price tag was an essentially arbitrary value placed on the transaction by executives of both companies. The asset they were trading was unused fiber-optic capacity (known in the industry as “dark fiber”), for which there was no demand and for which there might be no demand for years to come. Nearly 20 percent of Global Crossing’s $3.2 billion in revenue in the second quarter of 2001 came from capacity swaps. For the first nine months of 2001, such swaps accounted for $600 million of Qwest’s $15 billion in revenue. While the amount of the swaps appears modest as a percent of total revenue, it accounted for most of the company’s sales growth in that period.
When the accounting treatment was questioned, Global Crossing defended itself by asserting that each leg of the trade was priced and contracted independently. But any reasonable interpretation of the deal would conclude that no real sale or purchase had taken place. Compounding the deception, Global Crossing, Qwest, and other telecom companies would record the sale of capacity as revenue without recording the offsetting purchase of capacity as an expense. Instead, the purchase was capitalized. Rather than being posted on the income statement as an expense, deductible immediately from earnings, it was listed on the company’s balance sheet as a capital investment, its value being gradually reduced, or amortized, over several years. The maneuver increased reported profits by completely mischaracterizing the transaction, treating the payment of cash for capacity as if it were an investment instead of an expense.
The telecoms can no longer disguise the fact that they already have more than enough fiber-optic capacity, and so do all their competitors. They can no longer pretend that swaps of this useless, unwanted commodity have any real value, any more than they can pretend that the fiber-optic capacity they own is still worth what they paid for it. They must acknowledge the reduced value—or impaired value, as accountants say—by reducing their earnings by an amount equal to the decline in value of their fiber-optic assets. And since the SEC ruled fiber-optic swaps invalid in 2002, the telecoms must also restate previous revenue and earnings reports pumped up by swap transactions. For example, Qwest was required to reverse $950 million of revenue from capacity swaps.
Another game the telecoms played was to generate revenue unrelated to the basic business and claim it as revenue generated in the ordinary course of business. For example, Qwest would purchase networking equipment from Cisco, which it would then resell, at a profit, to KMC, a company that built and maintained such networks. Then, over the course of several years, Qwest would return most of that money to KMC in the form of payments for servicing the network supported by the equipment. Such transactions, which the company justified as a means of speeding network development, contributed to Qwest’s sales growth, but it was quite a stretch to record their proceeds as revenue from the company’s core business of selling telecommunications capacity.
Qwest had still more revenue-recognition tricks up its corporate sleeve, such as adjusting the publication date of its Yellow Pages directories to shift revenue from one quarter to another. In fact, as a general rule, it’s safe to say that companies rarely play just one accounting game. Games tend to come in clusters, so when you spot one, remember the cockroach theory: If you see one, you can be pretty sure there are many more hiding in the dark.
Once again, don’t make the mistake of thinking that revenue games are a purely American invention. The accounting firm Ernst & Young analyzed 41 UK software firms and characterized the revenue-recognition practices of more than half of them as poor or very poor. And in 2002, the UK-based Vodafone, the world’s largest mobile phone operator, admitted to playing games similar to those played by Edison Schools and Professional Detailing. It booked all the revenue generated by wireless Internet services, even when it redirected a sizable portion of that revenue to third parties that actually provided the Internet content. Vodaphone’s practice stands in sharp contrast to that of rivals MMO2, the mobile business spun off by BT, and Orange, the mobile arm of France Telecom, both of which strip out of reported revenues any amount owed to third-party content providers. In its defense, Vodaphone says that the payments made to third parties are treated as cost of sales—that is, an expense—leaving gross profit and other profit measures unaffected by the practice. Besides, says the company, the amounts involved represent only a small percentage of overall group revenues. Then why bother with the practice at all? One answer suggests itself: By inflating revenue relative to major competitors, Vodaphone can report higher average revenue per user, a statistic (or metric, in business-speak) closely watched by telecom analysts.
In response to the sort of gaming described in this chapter, the U.S. Securities and Exchange Commission in 1999 issued Staff Accounting Bulletin 101. The agency issues such bulletins from time to time to address accounting questions thrown up by a new sort of transaction or a new sort of business—or, as in this case, to rein in marginal practices before they become mainstream. In this document, the SEC staff declared that revenue should not be recognized until it is “realized or realizable and earned.” To achieve this standard, all of the following criteria must be met:
Persuasive evidence of a sales arrangement exists. If a particular company relies mainly on written sales contracts, such a contract must be in place and signed by the seller and the buyer before revenue can be recognized. Not only must customers signal intent to buy the product or service before revenue can be recognized, they must do so in a manner consistent with company or industry practice.
Delivery has occurred or services have been rendered. In the case of goods, the risk and rewards of product ownership have been transferred and the buyer no longer has any right of return. In the case of services, the service has already been performed. Contractual obligations on the part of a customer to pay for services rendered or goods delivered in the future do not qualify.
The selling price is fixed or determinable. Without such a price, there is no reasonable basis upon which to measure the amount of revenue to be recognized.
Collection is reasonably assured. It should go without saying, but selling to a customer who might not have the financial means to pay is not a proper sale.
The SEC’s bulletin didn’t just come out of the blue, of course. The agency’s staff prepared it specifically in response to the gross abuses of revenue-recognition rules that proliferated during the Internet boom.
One of the most dangerous accounting games is played with a method known as “percentage of completion.” Popular in the construction industry, though used elsewhere, too, it allows companies to recognize revenue gradually over the life of a long-term project, instead of waiting, sometimes for years, until the project is complete. For example, if the total revenue to be earned from the project is expected to be $20 million, and 30 percent of the work was done this year, $6 million of revenue would be recognized in this year’s income statement. While this approach has important conceptual merits, it is widely subject to abuse.
Percentage-of-completion accounting found its way into U.S. news in 2002, when Halliburton was accused of abusing the rules during the period when the Vice President of the United States, Richard Cheney, was CEO of the oilfield services firm. Like other firms with large, multiyear contracts, Halliburton recognizes revenue and profit on a percentage-of-completion basis. Since projects that run for several years often encounter unexpected costs, most contracts have cost-overrun provisions, which allow firms like Halliburton to increase their fees, as long as they can document the increased costs and win client approval. Until Cheney’s tenure as CEO, Halliburton accounted for possible overruns in the most cautious and conservative manner possible, recording the increased fees only when the customer had explicitly agreed to pay them. But under Cheney, Halliburton adopted a new policy, recognizing fee increases as long as customer consent was merely anticipated. Most customers eventually approved the increases, but Halliburton’s haste to recognize revenue and profits in advance of customer approval raised questions that the company was using the accounting change to dress up its income statement. Its merger with Dresser Industries was pending, and improved financial results would help Halliburton extract the best possible terms for the deal.