I’ve heard some people express the opinion that the FASB’s recent emphasis on an asset-liability approach is just a thinly veiled attempt to move financial reporting more in the direction of fair values. Ignoring the fact that the primacy of assets and liabilities is as old as the conceptual framework itself, I think recent evidence from the revenue recognition project soundly rejects the notion that an asset-liability approach inevitably leads to fair value.

In that project, the boards decided early in their deliberations that they wanted revenue to be recognized based on a change in some asset or liability. They considered a number of assets or liabilities that could determine when revenue is recognized. For example,

  • Revenue could be recognized as an asset is created for eventual transfer to a customer. In this case, the asset driving revenue recognition is the good being created for the customer. As the value of that asset increases (based on some estimation process), revenue is recognized. The outcome of this approach would be similar to what we achieve today through the percentage completion method for construction contracts.
  • Revenue could be recognized based on when a company satisfies its promise to transfer a good or service to a customer. In this case, the liability driving revenue recognition is a company’s contractual promise to transfer a good or service to the customer. The outcome of this approach would be similar to most “delivery” models of revenue recognition that we have today.
  • Revenue could be recognized based on when cash is received from the customer. In this case, the asset driving revenue recognition is cash itself. Of course most people would object to this approach before any delivery has occurred, but we do use this approach when collectibility is not reasonably assured.

After considering these and other possibilities, the boards eventually decided that the general standard would recognize revenue when a company satisfies a contractual obligation to transfer a good or service to a customer. Having focused the model on a particular liability (i.e., a company’s performance obligation within a contract with the customer), some might have expected that the boards would quickly jump to the idea of measuring that liability at fair value. Indeed, the Boards spent a few years considering that possibility. However, in the summer of 2008, the Boards ultimately decided to measure the performance obligation not at fair value, but instead by allocating the measure of the rights in the contract to all of the identified performance obligations. The boards soundly rejected the possibility of measuring performance obligations at fair value, even though the recognition model itself was based on changes in a particular liability.

My conclusion—using changes in assets and liabilities to determine recognition of revenue (or expenses for that matter) does not inevitably lead to fair value. Now, dare we ask whether there are some circumstances in which the boards are perfectly willing to recognize revenue based on measures of fair value? Of course there are. Think of financial assets held by a bank as trading securities. Think of revenue recognized for biological assets harvested, but not yet sold. In both of these cases, revenue recognition is essentially driven by changes in the value of an asset (either the trading security or the biological asset). Interestingly, if the Boards decide that their general model on revenue recognition should apply to financial service entities and agricultural entities, their revenue recognition will likely become less informative to investors. That is perhaps a topic for a future post…