The AAA Financial Accounting Standards Committee (FASC) recently submitted a comment letter in response to the FASB/IASB’s discussion paper on revenue recognition. This letter is unusual because it is one of few (if any others) that draw on existing academic research to support its arguments. In that regard, the FASC has provided a useful service to standard setters.

As you can imagine, there are more issues in the letter than I can possibly comment on in this post. So, let me draw your attention to just one thing (at least for now). In their response to Q1, the authors appear to support the boards’ decision to develop a model in which revenue is recognized when a performance obligation is satisfied by the transfer of goods and services to the customer. However, later in the letter, the authors state:

We advocate a model in which revenue is recognized in relation to the costs incurred to produce that revenue as long as there is a contract in place and the measurement of the revenue can be made with some threshold level of reliability. In situations where a contract does not exist or the revenue cannot be measured with sufficient reliability, then revenue would be based on the time at which an asset is transferred to the customer (at which point, it is usually much easier to measure the amount of revenue to be recognized).

The authors’ model is certainly a workable model and one that a number of staff and board members proposed and considered in early 2007. However, it is not the model the boards have tentatively chosen at this point. In fact, it is different in two significant ways. First, rather than recognizing revenue when a promised good or service is transferred to a customer, the authors essentially argue that revenue should be recognized as an asset (such as a building or ship) is created, as long as the price that will ultimately be received for that asset can be determined with some threshold level of reliability.

Second, the authors’ model does not focus on changes in a single asset or liability. Instead, it would recognize revenue from the creation of an asset as long as the price for that asset can be determined with some reliability, and in situations when that reliability doesn’t exist, they would recognize revenue from the satisfaction of a contractual promise to transfer a good to the customer. In this sense, the authors seem much more open to the idea that various changes in assets or liabilities can drive the recognition of revenue. In other words, they do not seem to share the boards’ view that revenue should be recognized solely from decreases in an entity’s net contract position with the customer.

So here are a few questions that occur to me.

  1. How important is it that a revenue recognition model focus on changes in a single asset or liability (such as an entity’s net contract position with the customer) rather than allowing revenue to arise from changes in many different assets of liabilities (such as the creation of a building or ship on the one hand and changes in an entity’s net contract position on the other hand)? In other words, do the boards need to limit revenue recognition to changes in an entity’s net contract position, or could it allow for multiple models based on changes in different assets or liabilities? Perhaps the asset or liability that drives revenue recognition should depend on the reliability with which changes in that asset or liability can be measured.
  2. How important is a contract? In the case of a building or ship under construction, the authors would recognize revenue only if the price can be determined with some reliability. The purpose of the contract in this situation seems to be that it includes explicit mention of the price that will ultimately be paid for that building or ship.
  3. If the authors are willing to recognize revenue for the creation of an asset (such as a building or ship) as long as the price to be received for that asset can be measured with some reliability, would they be comfortable recognizing revenue for a timber company’s growth in timber given that a spot prices for timber can be used to measure the change in the value of timber during the year? In this case, there doesn’t seem to be a need for a contract with a customer to provide an explicit price. So again, how important is a contract or what purpose does a contract serve in this situation? The same scenario would play out for the mining of precious metals where prices are easily observed even in the absence of a contract with a customer.

I’ll stop here with the questions for now. What are your reactions and thoughts?