Last November, the FASB issued Proposed Accounting Standards Update 1900-100, Transfers and Servicing (Topic 860): Reconsideration of Effective Control for Repurchase Agreements:
"During the global economic crisis, capital market participants questioned the necessity and usefulness of the collateral maintenance guidance when determining whether a repo should be accounted for as a sale or as a secured borrowing. …
[The determination of whether sales accounting is permitted] is based, in part, on whether the entity has maintained effective control over the transferred financial assets. One of the relevant considerations for assessing effective control is … whether there is an exchange of collateral in sufficient amount so as to reasonably assure the arrangement's completion on substantially the agreed terms, even in the event of the transferee's default.
The Board determined that the criterion pertaining to an exchange of collateral should not be a determining factor of effective control. … The Board believes that the remaining criteria are sufficient to determine effective control."
Please allow me to translate and remove the varnish:
We were burned by Lehman Brothers, who on multiple occasions dressed up its balance sheet by exploiting a loophole in our repo accounting rules. We allowed Lehman to presume that, because they received 'only' 5 percent less cash collateral than the value of the assets they transferred to an accommodating counterparty, it would be able to account for certain repo contracts as 'sales.'
Our rationale for including in GAAP a size-of-collateral loophole was that even if the counterparty defaults on its obligation to return the securities, the transferor (e.g., Lehman) would be holding sufficient cash collateral so that the transferor could buy replacement securities in the market.1 We now realize that our rationale completely ignored the basic fact that cash collateral is fully fungible with other sources of cash. The Lehmans of the world would have all sorts of cash hanging around for purchasing equivalent assets in the market, if that's what they wanted to do. So, let's be frank: we totally screwed up.
This proposed ASU is really our admission that collateral considerations had no basis in logic or reality. However, we really, really now believe that the remaining loopholes are sufficient for determining which party to a repo agreement has 'effective control'—whatever that means—over the transferred assets.
Will the Fix Work?
Under the remaining loopholes, a transferor would have retained 'effective control' of the non-cash assets (and have to account for the repo as a collateralized borrowing instead of a sale) unless it can wriggle through at least one:
- The financial assets to be repurchased are the same, or substantially the same, as the ones transferred;
- The repurchase is for a fixed price;
- The repo agreement is entered into at the same time, or in contemplation of, the transfer.
The FASB states in the proposed ASU that "…the assessment of effective control should focus on a transferor's contractual rights and obligations with respect to transferred financial assets," but I can't see how these three surviving loopholes have somehow fundamentally transformed the focus of repo accounting. I am also not sage enough to foretell what new devices issuers will invent, but wriggling through only one loophole is still all that it will take to achieve sale accounting.
Moreover, the logic of an overarching 'effective control' criteria remains, let's say, interesting. Has there ever been a coin flip that turned out to be 'effectively heads'? Sorry, but I can't resist this one: has there ever been an 'effective pregnancy'? Like 'control,' either you are, or you aren't;
A Principles-Based Solution
I described my approach to repo accounting in a previous post, and I'll briefly summarize it in enough detail here so as to add a clarification in response to a question I received.
The following procedures would apply to securities for which legal ownership (as opposed to 'effective control') has transferred:
- At the inception of the agreement, a transferor would recognize an asset for the cash received and derecognize the securities transferred. Any difference between these two amounts would preferably be reflected on the balance sheet as a change in shareholders' equity through net income.
- To recognize the commitment to repurchase the transferred securities, a receivable for the future return of the transferred securities would be recognized at its fair value; and a payable would be recognized at its fair value for the price to be paid for the return of the transferred securities. In contrast to existing accounting rules, these two amounts must be presented "gross"— i.e., they may not be offset. And again, any difference between these two amounts would be reflected on the balance sheet as a change in shareholders' equity through net income.
One reader of the post in which I described my proposed repo accounting asked whether I would also show a receivable on the balance sheet of a company that simply has committed to buy an asset under any sort of agreement (i.e., not just a repo agreement). My response is: no – except for financial instruments.
First, I would amend GAAP to require that all commitments or options to purchase or sell a financial instrument (perhaps with limited exceptions such as regular-way security trades) would be accounted for as any other derivative. I will readily admit that I am not an expert on determining whether a particular contract is accounted for as a derivative, but my understanding of the derivatives rules in Codification Topic 815 is that the 'forward' or 'option' aspect of a repo arrangement is scoped out. My proposal would fix that.
Second, I would require gross presentation of all derivatives. In other words, the receivable leg may not be netted with the payable leg – even if the terms of the contract were to permit net settlement.
Taken together, these two changes to GAAP would effectively eliminate opportunities for window dressing with repos. Even more significant, and as I have described in a post dealing with the credit default swap mess, these enhancements to GAAP would also result in more conservative balance sheet indicators of a bank's capital adequacy. By grossing up an interest rate swap, for example, a bank would recognize a receivable for the expected interest inflows and a payable for expected interest outflows. Under the current rules, capital adequacy is unaffected by the swap, whether the notional amount is $1 million or $100 billion.
Let's forget and forgive the FASB for the loophole that Lehman wriggled through, even though there was never any good reason for its existence. But, let's not allow the FASB to pretend that all the repo accounting rules need is a slight trim around the edges.
1As explained by former FASB chair, Robert Herz, in his letter to the House Financial Services Committee dated April 19, 2010.