The Accounting Onion

Contingent Liabilities: A Troubling Signpost on the Winding Road to a Single Global Accounting Standard

By the logic of others, which I can’t explain, fuzzy lines in accounting standards have come to be exalted as “principles-based” and bright lines are disparaged as “rules-based.”  One of my favorite examples (actually a pet peeve) of this phenomenon is the difference in the accounting for leases between IFRS and U.S. GAAP.  The objective of the financial reporting game is to capture as much of the economic benefits of an asset as possible, while keeping the contractual liability for future lease payments off the balance sheet; a win is scored an “operating lease,” and a loss is scored a “capital lease.”  As in tennis, If the present value of the minimum lease payments turns out to be even a hair over the 90% line of the leased asset’s fair value, your shot is out and you lose the point.  

The counterpart to FAS 13 in IFRS is IAS 17, a putative principles-based standard.  It’s more a less a carbon copy of FAS 13 in its major provisions, except that bright lines are replaced with fuzzy lines: if the present value of the minimum lease payments is a “substantial portion” (whatever that means) of the leased asset’s fair value, you lose operating lease accounting.  If FAS 13 is tennis, then IAS 17 is tennis-without-lines.  Either way, the accounting game has another twist: the players call the balls landing on their side of the net; and the only job of the umpire—chosen and compensated by each player—is to opine on the reasonableness of their player's call.   So, one would confidently expect that the players of tennis-without- lines have a much lower risk of being overruled by their auditors… whoops, I meant umpires.

Although lease accounting is one example for which GAAP is bright-lined and IFRS is the fuzzy one, the opposite is sometimes the case, with accounting for contingencies under FAS 5 or IAS 37 being a prime exaple.  FAS 5 requires recognition of a contingent liability when it is “probable” that a future event will result in the occurrence of a liability.  What does “probable” mean?  According to FAS 5, it means “likely to occur.”  Wow, that sure clears things up.  With a recognition threshold as solid as Jell-o nailed to a tree and boilerplate footnote disclosures to keep up appearances, there should be little problem persuading one’s handpicked independent auditor of the “reasonableness” of any in or out call.

IAS 37 has a similar recognition threshold for a contingent liability (Note: I am adopting U.S. terminology throughout, even though "contingent liabilities" are referred to as "provisions" in IAS 37).  But in refreshing contrast to FAS 5, IAS 37 unambiguously nails down the definition of “probable” to be “more likely than not” —i.e., just a hair north of 50%.  Naively assuming that companies actually comply with the letter and spirit of IAS 37, then more liabilities should find their way onto the balance sheet under IFRS than GAAP.   And, IAS 37 also has more principled rules for measuring a liability, once recognized.   But, I won’t get into that here.  Just please take my word for it that IAS 37 is to FAS 5 as steak is to chopped liver. 

The Global Accounting Race to the Bottom

And so we have the IASB’s ineffable ongoing six-year project to make a hairball out of IAS 37.    If these two standards, IAS 37 and FAS 5, are to be brought closer together as the ballyhooed Memorandum of Understanding between IASB and FASB should portend, it would make much more sense for the FASB to revise FAS 5 to make it more like IAS 37.  After all, convergence isn’t supposed to take forever; even if you don’t think IAS 37 is perfect, there are a lot more serious problems IASB could be working harder on: leases, pensions, revenue recognition, securitizations, related party transactions, just to name a few off the top of my head.  But, the stakeholders in IFRS are evidently telling the IASB that they get their jollies from tennis without lines.  And, the IASB, dependent on the big boys for funding, is listening real close. 

Basically, the IASB has concluded that all present obligations – not just those that are more likely than not to result in an outflow of assets – should be recognized.  It sounds admirably principled and ambitious, but there’s a catch.  In place of the bright-line probability threshold in IAS 37, there would be the fuzziest line criteria one could possibly devise:  the liability must be capable of “reliable” measurement. We know that "probable" without further guidance must at least lie between 0 and 1, but what amount of measurement error is within range of “reliable”?  The answer, it seems, would be left to the whim of the issuer followed by the inevitable wave-your-hands-in-the-air rubber stamp of the auditor. 

It’s not as if the IASB doesn’t have history from which to learn.  Where the IASB is trying to go in revising IAS 37, we’ve already been in the U.S.  The result was all too often not a pretty sight as unrecognized liabilities suddenly slammed into balance sheets like freight trains.  As I discussed in an earlier post, retiree health care liabilities were kept off balance sheets until they were about to break unionized industrial companies.  Post-retirement benefits were doled out by earlier generations of management, long departed with their generous termination benefits, in order to persuade obstreperous unions to return to the assembly lines.  GM and Ford are now on the verge of settling faustian bargains of their forbearers with huge cash outlays: yet for decades the amount recognized on the balance sheet was precisely nil.  The accounting for these liabilities had been conveniently ignored, with only boilerplate disclosures in their stead, out of supposed concern for reliable measurement.  Yet, everyone knew that zero as the answer was as far from correct as Detroit is from Tokyo – where, as in most developed countries, health care costs of retirees are the responsibility of government. 

Holding the recognition of a liability hostage to “reliable” measurement is bad accounting.  There is just no other way I can put it.  If this is the way the IASB is going to spend its time as we are supposed to be moving to a single global standard, then let the race to the bottom begin.  
 

FAS 157: The FASB's Prelude and Fugue on Fair Value of Liabilities

FAS 157 on fair value measurements was supposed to provide comprehensive guidance for determining the fair value of pretty much any asset or liability.  Yet, almost two years after its initial publication, and well after companies have had to apply the standard to certain accounts, CFO.com reports that the FASB is still making up some of its rules on the fly, and having a tough slog to boot.  The problem described in the article has immediate consequences for derivative financial instruments that are classified as liabilities, but it could eventually affect the measurement of many other liability accounts as fair value measurement becomes more broadly applied:

"At an unusually heated FASB meeting last week [no minutes published on the FASB's website yet], for instance, the members debated how companies should estimate the market value of liabilities when there's no actual market on which to base the estimate.

During one point in the discussion, which concerned a proposed guidance by FASB's staff on how to mark liabilities to market under 157, chairman Robert Herz seemed, to member Leslie Seidman, to be contemplating an overhaul of the brand-new standard itself. Matters got so confusing that the board ordered its staff to go back and summarize the members' positions so that they could understand what they themselves had said.

At issue was the question of how to measure the fair value of a liability for "which there is little, if any, market activity," according to 157. The standard defines fair value as "the price that would be received ... to transfer a liability in an orderly transaction between market participants at the measurement date." The question that FASB struggled with was: How do you determine the fair value of a liability that can only be settled, rather than sold?

...Often, for instance, when a company borrows money, it can't transfer its obligation to another party without an agreement from the bank. Or a market may not exist for transferring such liabilities."

It's a mess that the FASB has gotten itself into for two related reasons.  The first is that the problems now being addressed are significant, and they were known long before FAS 157 was let out the door. The second is that FAS 157 is fundamentally flawed in its approach to fair value measurement of liabilities.  The solution, as I am about to describe, seems to me to be surprisingly simple. 

This particular flaw in FAS 157 (see my previous post on many others) occurs in paragraph 5:

"Fair value is the price that would be received to sell an asset or paid to transfer a liability [italics supplied] in an orderly transaction between market participants at the measurement date." 

For every liability there is a counter party that holds an asset, and the economic value of the liability must be equal to the economic value of the asset. These are basic economic principles, which are not acknowledged in FAS 157.    If they were acknowledged, there would be no need for the phrase "or paid to transfer a liability."  That's because the value of any liability -- even one that cannot be transferred --must equal the value of the counter party's asset, which, perforce, can always be transferred.  Even though the evidence directly available to value the liability may be scant, the asset value might even be quoted in the newspaper; the non-transferability restriction on the debtor is just one more valuation parameter from the viewpoint of the creditor. 

If you need further convincing that the solution to the problem of valuing any liability is to value the counter party's asset, let's consider an even thornier non-transferable liability that the FASB briefly considered and then dropped like a hot potato:  contingent environmental liabilities.  My understanding of federal environmental law is that the cleanup liability of a "potentially responsible party" is joint and several.  No other party can assume the liability, so the only way out from under it is to settle with the government.  Although I am not aware that the government has done this, it is theoretically possible for the government to transfer its contingent receivable to a third party.  Is the contingent receivable difficult to value?  Yes, but certainly no harder than many of the complex, illiquid derivatives that are roiling the global economy.  (And by the way, I recall seeing the issue of the fair value of contingent environmental liabilities posted on the FASB's website during the project phase of FAS 157.  The Board expressed a tentative conclusion, but it soon disappeared mysteriously, and without explanation.  I have searched Board minutes, and have come up with nothing.  If anyone has any further information on this that they would like to share, please contact me!)

Because my solution to liability valuation is so simple (attention: CIFiR - SEC Advisory Committee on Improvements to Financial Reporting) and obvious, I can't help but fear I have overlooked something.  If that is indeed the case, I hope a reader of this post will take the time to point it out, and I will gladly issue a mea culpa forthwith.  Yet, I derive some measure of comfort (and optimism) by an entry in the minutes of an FASB meeting (11/14/07) where Bob Herz stated that he disagrees with the measurement principles for liabilities in SFAS 157. 

Who knows, maybe Bob and I are thinking along the same lines?  That gives me hope for the future.  But, I have to express my disappointment that liabilities were not dealt with in a comprehensive way before SFAS 157 was issued. There is much to be said for getting it right the first time.