Peeling away financial reporting issues one layer at a time

Gross versus Net Presentation: The First of Many Revenue Recognition Debacles to Come?

Pick any class of upper-level accounting students, and give them a few cases of supplier/customer relationships for which “gross versus net presentation” of revenues is an issue, and I’ll wager that the class will come to a consensus on the presentation most consistent with the objectives of financial accounting 90% of the time.

Yet, somehow, this issue has become the first battleground over the new joint revenue recognition standard with the IASB.

How Did This Get So Hard?

The question is nothing more than this: if two parties are involved in selling goods and services to a customer, how should each of the parties involved in the sale present their revenues?  Not all of the fact patterns are basic, but let’s start with one that is:

Facts—An online retailer receives an order from a customer.  The customer uses a credit card to pay the retailer $100. The retailer notifies its supplier, who ships the merchandise directly to the customer.  Soon thereafter, the retailer pays the supplier $85. The retailer’s gross margin is $14, which is straightforwardly calculated as $100 minus $85, and minus $1 withheld by the credit card company.

Question—How should the retailer present revenue and cost of sales: (a) $100 revenues and $86 cost of sales; or (b) $15 and $1 cost of sales?


It would seem that an accounting standard would not be necessary to get the online retailer to the correct answer, which is that it may only record $15 in revenue.  But since 1999, a long list of subjective criteria have been added to the authoritative literature to cover this mundane situation.  IMHO, this was  owing to widespread auditor reluctance to push back against a “creative” client without an accounting rule for a crutch.

A Bit of a History Lesson 

During the .com boom era, “profitable internet IPO” was an oxymoron.  Consequently, the sales pitch to investors was that nothing mattered more than capturing market share.  In other words, revenue growth — and not conventional measures of profitability — should have been the key performance metric. It’s not surprising, therefore, that it would have been common for an online retailer to recognize $100 in revenue and $86 in cost of sales — even though the merchandise never passed through its warehouse.  EITF 99-19 (ASC 605-45) put a stop to most of these abuses by requiring an issuer to determine whether it was acting as a “principal” to its customer, or merely as an “agent” of … somebody.  If the latter, then the appropriate amount of revenue to recognize would only be the net $15 retained after paying the supplier minus the credit card processing fee.

The New Order

Now, let’s fast forward to the present.  The FASB and IASB jointly issued a new revenue recognition standard a few months ago, which established common criteria for determining whether gross or net presentation would be appropriate.  Considering the many many other changes made by the new standard, gross versus net presentation would appear to the casual observer to be no more important than a flea on the hind quarter of of a Holstein.  So whaddya know, it became the very first topic to be considered by the boards’ Transition Resource Group.

And I’m sure that the leadership of the TRG now regrets it. By all unbiased accounts, the TRG’s maiden deliberations were a fiasco.  Perhaps it was because the principal v. agent framework would be applied within a new overarching concept of transfer of control.  This might be problematic when selling services instead of tangible goods as in our first example. During the discussion, it also came out that some TRG members on the IFRS side of the table were more than a little concerned that current practice would be affected. (Whaddya know.)

Making Common Sense Out of Gross versus Net Presentation

As I see the problem, it is that the principal/agent dichotomy is being misapplied to model relationships that vary on a continuum, as opposed to a strict principal/agent dichotomy.  An either/or solution might work well for the extreme cases, but many other fact patterns fall through the cracks.  The defining characteristic of arrangements having multiple purveyors of goods/services and one final customer is not the mere presentation of revenue, but recognition of expenses.

Let me explain.  An implicit assumption of these fact patterns is that the issuer knows what it’s gross margin is from a transaction.   The remainder of the accounting question boils down to an algebraic equation with two unknowns: revenues and cost of sales.  Instead of letting the amount of revenue recognized determine the amount of cost of sales to plug, an issuer should have to determine whether the amount of expense recognized is appropriate.

That would require a conceptual definition of “expense,” so let’s start with this one:

Expenses, except for  interest, taxes, and other obligations to deliver assets,  are outflows or other using up of real economic resources that the entity controls from delivering or producing goods, rendering services, or carrying out other activities that constitute the entity’s ongoing major or central operations.

My definition is very similar to the FASB’s definition as set forth in CON 6, and should produce the same results, but I made my own variation so as to be clear that the general case of an expense is the use of an economic resource, i.e., a real asset, which is independent of how acquisition of the resource will be paid for.  Moreover, I want to take this opportunity to point out that we make an exception to the criterion of real asset consumption for interest and taxes because these are distributions to other stakeholders (i.e., lenders and governments), which are reasonably regarded as “expenses” from the perspective of common shareholders.

Applying the definition of “expense” to the retailer example is straightforward.  Since the retailer never had any control how the asset would be used, it can’t have an expense; and since gross margin of $14 is known, the only possible measure of revenue is $15.

Okay, let’s try another example.  The TRG spent a long time discussing one similar to this:

A “bricks and mortar” retail store offers gift cards for sale.  For $X, a purchaser of the gift card is entitled to spend $Y at a local restaurant.  The grocery story remits $Z to the restaurant, and (ignoring the cost of the card) earns a gross margin of $X minus $Z.

If you don’t think that it’s obvious how much revenue the grocery store should recognize, then I nominate you for the TRG.  You wouldn’t believe the esoterica that was discussed at the meeting; and I won’t bore you with any of it here.  I merely want to point out how easy it is under my suggested approach to determine how much revenue the grocery store would report from the transaction.  Since none of the costs of serving the customer do not meet the conceptual definition of an expense for the retail store, its revenue is limited to the amount it earned, $X minus $Z.

Finally, let’s try a “hard” problem.  This is one that the TRG might not consider, but is nonetheless very prevalent; and to my knowledge, practice varies considerably:

A real estate investor owns a hotel property, and engages a management company to operate its hotel.  All of the operating cash flows from the hotel are collected/paid by the management company. They are subsequently remitted to the real estate investor, minus the management fee.

It would be a ludicrous exercise for the management company to determine whether it is acting as a principal or an agent for the real estate investor.  Reference to criteria for expense recognition, however, would be a relative piece of cake.  For example, depreciation and interest are not expenses of the management company, but most other operating costs probably are.  And since the management company would know: (1) what those other operating costs are; and (2) its gross margin from providing management services, that would dictate the amount of revenue to recognize.

* * * * * *

Revenue presentation should not be a hard accounting problem.  Yet, the TRG couldn’t even agree on how a grocery store should present its revenues from sales of gift cards!  Compared to the other questions that issuers will want the TRG to address before the new standard becomes effective, that one should have been low hanging fruit.  Instead, it has become a litmus test for the TRG and maybe even for the new revenue recognition standard itself.

But to be fair, the TRG didn’t cause the problem.  Fifteen years ago, the FASB punted the revenue presentation inquiries from auditors to the EITF, which proceeded to introduce a brand new concept to financial reporting, and it did so without any real semblance of “due process” that should accompany such weighty and consequential deliberations.

It may have worked for a while, but it has finally blown up in their faces.


  1. Reply Martha Nelson October 17, 2014

    Thanks for the article, Professor Selling. It raises some really interesting points, and I very much enjoyed reading it.

    I am curious to know whether the philosophy of “economic realities” that is often used in tax accounting for recognizing income would be a good standard for revenue recognition in cases like these. I don’t know if there is a better term for the principle, but I would assume that the best yardstick for determining revenue recognition would be the overall economic effects of revenue and cost of sales (i.e., the final economic effects of the first example you gave in the article are $15 in revenue and $1 in cost of sales, not $100 and $86). My guess is that using the economic realities approach as a basis for revenue recognition would focus more on the meat of the issue, which is the amount of revenue and expense that an entity can legitimately claim when it is acting as a middle-man in a transaction (basically, a sort of commission on the sale). I think it would also be helpful to emphasize the principle of conservative revenue recognition when creating regulations to address this issue.

    Am I on the right track? What are your thoughts?

    • Reply Tom Selling October 17, 2014

      Hi, Martha:
      You raise an important question that I chose not to directly address in my post, which was already getting too long. I think “economic reality” is a crucial consideration. Take my first example: what if the online retailer took legal title to the goods just before they were shipped to the customer? So, any risk of loss/damage would be borne by the retailer? To use a term from financial reporting, that aspect of the arrangement would lack sufficient “commercial substance” and that it should be disregarded when determining whether the goods are part of the cost of sales of the online retailer.

  2. Reply Brad October 18, 2014

    Speculation here, but I am guessing the biggest issue that the TRG may be having with “gross v. net” presentation in the gift card scenario is the concept of credit risk. Now, even EITF 99-19 indicates this is weaker evidence of gross reporting, but I could see an example set-forth below that could cause individuals on the TRG some difficulty in reaching a conclusion:

    In the above example (grocery store), I could see a situation where, the store sells a gift card to a customer via a stolen credit card. Said CC entity then rejects the transactions since the credit card is fraudulent. But the gift card has been used before the store can cancel the card (a common way for thieves to use stolen credit cards).

    Assume the value is $100.00 gift card, in which $85 gets sent to the restaurant. A situation could exist where in one quarter, the store reported $15 in revenues with $1 in cost of sales, and then in the next quarter, is forced to report a $100 loss when the transaction gets rejected by the CC company. (Unless somehow the store is able to recoup the $85 is sent to the restaurant, and thereby forcing the restaurant to eat the loss, but I am not well versed enough in these arrangements to understand who ultimately is responsible for losses on fraudulent gift card purchases).

    Hence why, one of the factors in determining gross vs. net presentation under EITF 99-19 is credit risk. In the gift card scenario, if the risk of loss is invariably borne upon the retailer, not the restaurant, then net presentation could result in a significant difference between revenues reported in one period, and the loss recognized in a future period.

    • Reply Tom Selling October 18, 2014

      Hi, Brad:

      I understand that implementing the EITF 99-19 criteria would be challenging, but credit risk is not a factor that I would consider. My perspective is that the seller of the gift card is providing a form of insurance to the restaurant, and that the compensation the seller of the gift card receives contains implicit compensation for bearing that risk. In other words, that implicit compensation is nothing more than prepaid insurance (for a very short duration insurance contract), and should be accounted for in that way. For me, this is no justification for grossing up revenues by amounts that don’t meet the conceptual definition of an expense.

  3. Reply Tim October 30, 2014

    Great article, Tom. Great response to Martha also- I too was wondering about economic realties

  4. Reply Donald April 29, 2015

    I buy used auto parts from various vendors. I set the price and take the credit risk. I pay the supplier and the customer pays me. I do not get a commission from the supplier and I arrange transportation. If something is damaged, I am liable. The customer is one I developed and normally the supplier does not know who it is until I give them my order to ship. I can go to any supplier to buy my products. I offer the warranty, the customer pays me and I in turn pay the vendor. Can I report what the customer pays me as gross revenue, show cost of auto part as cost of goods sold less freight as my cost to establish gross margins.

Leave a Comment