Revenue Recognition: Will a Single Model Fly?

Can U.S. and international accounting standard-setters realize their dream of fashioning a single revenue-recognition standard that would apply to all customer contracts? While the answer won't be known for some time, it's safe to say there are hurdles on the road ahead.

In a joint discussion paper issued last December in which the Financial Accounting Standards Board and the International Accounting Standards Board proposed a model for a lone standard, they acknowledged that an alternative approach could be needed for some contracts. The almost 200 letters they received in a comment period that ended June 19 did nothing to remove any doubts about whether having one standard will be viable.

Most of the letters agreed that the standards boards' goals are laudable. One main objective is to simplify and clarify FASB's revenue-recognition rules, which currently are scattered among more than 100 standards. Another is to offer more guidance than what's contained in IASB's broadly worded revenue-recognition principle.

In meeting those twin objectives, the boards would be advancing their overarching goal of converging U.S. and international standards. The major goals aside, however, many commenters registered alarm at specifics of the proposed model — especially concerning how revenue should be recognized under long-term contracts.

Today, entities typically recognize revenue when it's realized or realizable and the "earnings process" is substantially complete. The new model instead would direct the entity to record the gain when it performs an obligation under its contract, such as by delivering a promised good or service to the customer. (The contract need not be written; even a simple retail transaction involves an implicit contract in which the customer agrees to provide consideration in return for an item.)

In a simple example, if the entity had agreed to provide two products at different times, it would recognize revenue twice, even if the contract stipulated that payment would not be made until the second product was delivered. The discussion paper mentions several permissible bases on which revenue could be allocated to the different performance obligations. But the paper says the revenue should be in proportion to the stand-alone selling price of the good or service underlying a performance obligation. And for an item that's not sold separately, a stand-alone price should be estimated — something that the standards boards acknowledged could be hard to do.

A main purpose of the performance-obligation approach is to iron out many of the disparities in how businesses account for revenue, which the boards say make financial statements less useful than they should be. The discussion paper gave the example of cable television providers, which under FAS 51 account for connecting customers to the cable network and providing the cable signal over the subscription period as separate earnings processes. By contrast, under the Securities and Exchange Commission's SAB 104, telephone companies account for up-front activation fees and monthly fees for phone usage as part of the same earnings process.

"The fact that entities apply the earnings process approach differently to economically similar transactions calls into question the usefulness of that approach [and] reduces the comparability of revenue across entities and industries," the discussion paper stated.

Long Engagements
Perhaps the thorniest issue arising from the standards boards' proposal involves long-term construction or production contracts. Historically, under many such arrangements the company recognizes revenue using the "percentage-of-completion" method — if it's a three-year project with costs of $3 million, and $1 million of that is expended in the first year, one-third of the revenue is reflected for that year.

The single-model proposal, on the other hand, says that revenue should be recognized as an entity "transfers control" of goods and services to the customer. But many comment letters noted that the discussion paper did not clearly define what constitutes a transfer of control.

A company that is constructing a building for a customer may regard the materials and labor being provided as a continuous transfer of goods and services, which under the proposed model could be construed as allowing them to continue to recognize revenue over the duration of the contract. But if the standard setters hold that "transfer of control" occurs when the building is completed and turned over to the customer, all of the revenue would have to be recognized in the final year of the contract.

Lynne Triplett, a partner and revenue-recognition expert at Grant Thornton, told CFO.com that the way the discussion paper is written, "There could be questions as to whether there is continuous transfer of control, and to the extent there's not, there is going to be a significant difference between the way revenue is recognized today versus how it might be recognized in the future."

That would create misleading financial statements, according to some of the comment letters. "The most concerning area of the discussion paper is the potential change to the accounting for long-term contracts," wrote Financial Executives International Canada. "Creating a model which results in 'lumpy' revenue recognition ... with a waterfall effect in one accounting period at the very end, is not useful to the readers of financial statements."

• 1
• 2
• 3
• next »

LinkedIn Company Connections:

• FASB |
• IASB |
• Verizon |
• Grant Thornton |
• General Dymanics |
• FEI |
• Raytheon