One of the issues that adds greatly to the complexity of accounting is the concept known as “unit of account.” The question is whether to make accounting decisions item by item or to make those decisions based upon some grouping of items.

We have come up with many different answers to the unit of account issue, each seemingly logical to the situation it addresses.  Unfortunately, the result of all these individually logical answers is somewhat baffling to a lot of accountants and, much like the measurement attribute issue I wrote about earlier this year, we don't have any guiding principles for how to select the unit of account, meaning that each time we address the issue, we're largely starting from scratch.

A simple example of the issue is determining at what level to account for something that is acquired in pieces but used as a whole , such as a building that an entity constructs. Of course, different parts of the building are purchased separately, and different parts wear out at different times. International Financial Reporting Standards require that each major component of a building be separately depreciated, so that when that component is replaced, the new component can be capitalized without seeming to “double count” anything.

In U.S. Generally Accepted Accounting Principles, however, we record a building as a single asset, with a single depreciable life, making the initial accounting much simpler. An unfortunate consequence of that, though, is that we capitalize the cost of a replacement roof even though the original roof may not be fully depreciated.

In another example of the different ways to think about an asset, consider a company that provides an incentive to new customers in the hopes that they will make more purchases over time. While each individual new customer may not make sufficient purchases to cover the cost of the incentive, the company might well be able to show that, as a group, the customers receiving the incentive will provide sufficient business to cover the total cost of providing the incentives. In this situation, if each customer's incentive is analyzed separately, capitalization of the incentive costs would not seem justified, while consideration of the incentive costs in total would make capitalization seem appropriate.

Unit of account also comes into play when expenditures produce a large number of individual failures in order to achieve a small number of successes. Oil exploration companies, for example, spend a ton of money on failed exploration activities, but the small number of successful efforts is expected to more than cover the costs of all exploration, including the costs of dry holes.

The question for accountants, then, is whether the costs of the dry holes must be written off, as those costs did not specifically lead to any economic benefits, or whether the costs of all exploration should be pooled together. While the Financial Accounting Standards Board issued a standard mandating the “successful efforts” approach, in which costs of unsuccessful efforts would be expensed, the Securities and Exchange Commission overrode FASB and let public oil and gas exploration companies choose the “full cost” method instead, capitalizing all exploration costs as essentially a single asset.

While successful efforts is generally mandated in other areas, there are sufficient uses of full-cost accounting, such as direct response advertising costs, that the issue keeps raising its head over and over.

The Law of Large Numbers

Sometimes we group items in order to recognize the effect of things that we can't measure on an item-by-item basis. When a retailer sells goods to consumers with a right of return, for example, it is impossible to know exactly which items will be returned. However, if there are large number of homogeneous sales, history may show what percentage of sales are typically returned, allowing us to account for the right of return for the whole group of transactions when we can't do it for individual transactions.

This strategy is repeated in other areas of accounting, such as bad debts, in which reserves are booked based on the analysis of a portfolio as opposed to individual accounts or loans. Warranty costs, stock option forfeitures, and inventory write-downs are all similarly accounted for at a group level, as accounting for them item-by-item would be incredibly difficult.

My advice to accountants is to bone up on the grouping issues relevant to your business, because I believe the complexity involved in the unit-of-account issue is here to stay.

Other groupings are applied elsewhere. Impairment testing for many assets is intended to reflect the way an asset is used. A machine used in manufacturing, for example, is bundled for impairment testing with other assets (and possibly liabilities) to form a group that generates cash flows that are largely independent of other asset groups. Goodwill, on the other hand, is tested for impairment at a reporting-unit level, which is something like a separately operating component of a company.

While it might be very difficult to test these assets for impairment individually, the results sometimes seem illogical. For example, when a business is acquired and placed into a larger reporting unit, the goodwill from the acquisition becomes associated with the unit as a whole, and therefore might not be written off even if the acquired business completely fails. Similarly, an impairment of real estate might not be recognized even though property values drop precipitously, if the real estate in question is used in an asset group that continues to be profitable.

Complexity on Top of Complexity

The complexity brought about by all these different units of account is exacerbated by the fact that, in most cases, a particular item is only grouped for certain topics, while it is analyzed alone or in a different group for other issues. As a result, a single asset may be part of five or more different units of account for different accounting analyses. It's no wonder that accountants are often confused about what level to analyze an issue at.

All of the major current projects being worked on by FASB and the International Accounting Standards Board have unit-of-account issues running through them. The revenue recognition standard is likely to require revenue to be recognized at a “performance obligation” level, which is often a portion of a contract. Contract origination costs, however, will be analyzed at a contract level. Rights of return will continue to be analyzed on a portfolio level, as will price variability if there is a large volume of similar transactions.

A major issue of the leasing project was whether optional lease extensions would be treated as part of the overall lease obligation, or whether the option would be evaluated separately. FASB and IASB originally concluded that the unit of account is the lease, thereby requiring all options to be included in the measure of the lease liability. Eventually, the boards changed the unit of accounting to the option, so that the liability to make lease payments during an optional extension will only be recognized if that option is highly likely to be exercised.

With the financial instruments project, loan impairment has been looked at almost exclusively in regard to portfolios to the point where the proposals are almost unintelligible when applied to an individual financial asset.

So What Do We Do?

The answers to the unit-of-account questions in these projects do not exhibit consistent thinking, nor are there common principles being applied to reach the answers. Indeed, part of the reason that we have so much difficulty resolving unit-of-account issues is that there hasn't been a real attempt to consider what factors should lead to grouping and at what levels. I would like FASB and IASB to do some conceptual thinking in this area, if for no other reason than to help accountants deal with situations that aren't covered in the literature.

I'm not optimistic, however, that a significant reduction in the many different units of account employed in U.S. GAAP can be achieved. There are a great many different economic phenomena that accounting needs to reflect if transparency is to be achieved, and some of those phenomena can only be captured adequately and efficiently by looking at things in a group, while others are more easily reflected by considering each item on its own. As a colleague of mine used to say, “Everything is a unit-of-account issue.” She's right, and there is no way to make that go away.

So unfortunately, while the standard setters can certainly help us to some extent by providing some principles for when grouping should occur and at what level, I believe that this issue is actually an example of complexity that is necessary to make financial reporting relevant. Einstein has been quoted as saying “Everything should be made as simple as possible, but no simpler.” In this case, I'm afraid the issue just isn't a simple one, so my advice to accountants is to bone up on the grouping issues relevant to your business, because I believe the complexity involved in the unit-of-account issue is here to stay.