The Accounting Onion

"Goodwill Impairment" Accounting Could Become Less Costly – and Earnings Management a Lot Easier

When you start with a bogus asset like goodwill (itself a misnomer), it's hard to find quality in the rules that govern its measurement. The root of the problem is that business combination accounting relies on a fantasy: that an utterly ineffable plug to balance the business combination journal entry must somehow be an asset. The initial measurement of that plug is, at its theoretical best (i.e., assuming that management did not overpay; and all other assets and liabilities are recognized and accurately measured), an amalgamation of assets and liabilities that (unlike every other asset on the balance sheet) can neither be separated from the rest of the entity nor measured directly. Among the broad panoply of misnomers in financial accounting, "goodwill" is among the least subtle.

If goodwill is a bogus asset, then testing it for impairment compounds the madness. But, there is at least some method to it. The propensity of management to overpay for corporate acquisitions is one of the most well-documented and uncontroverted phenomena in academic finance: with disturbing frequency, share price movements send a signal that investors believe that a corporate acquisition destroyed, rather than created, value. If goodwill were recognized as an asset, but not subject to either amortization or impairment, then management would be even more inclined to destroy shareholder value without having to worry that future earnings would suffer from its profligate spending. Since 2001 (see SFAS 142), impairment testing, albeit replete with opportunities for earnings management, has been seen as the less silly of two silly charades. As currently set forth in U.S. GAAP (ASC Topic 350), this is how it is choreographed:

• All of the goodwill currently on the balance sheet has to be assigned to "reporting units," another misnamed artifice, invented solely for the purpose of making the goodwill impairment test flexible and palatable.
• The fair value of each reporting unit has to be assessed at least once a year – a costly undertaking.
• The real mayhem begins if, heaven forbid, your estimate of the fair value of the reporting unit turns out to be less than its carrying amount. You must then estimate the "implied fair value" of goodwill, another artifice devised solely for goodwill impairment testing; and –
• Write goodwill down to the implied fair value, if its carrying amount is greater.

Recently, however, the FASB issued proposed Accounting Standards Update No. 2011-180, which if finalized would grant unfettered discretion to elect to avoid these processes altogether – so long as "qualitative factors" indicate that it is more likely than not (MLTN) that the fair value of a reporting unit is greater than its carrying amount. The option to consider qualitative factors applies to any year and any reporting unit.

The existing goodwill impairment ballet can be a very costly production. While I am sympathetic to cost issues, I have a number of concerns with the way the FASB is addressing them:

More management discretion means more earnings management – The Board ostensibly decided to give entities the discretion to skip any qualitative assessment at any time. The stated intention is to save entities the cost of performing the qualitative test in a period when breaching the MLTN threshold could be self-evident. Allow me to illustrate my skepticism for this explanation with a slight digression – an example of a similar impairment test that I got wind of just this week:

I was contacted by an analyst who was concerned about a foreign energy producer that prepares its financial statements under IFRS. In accounting for its oil and gas field development costs, the company capitalizes the development costs associated with both successful and unsuccessful wells. The question I was asked pertained to a very large reported impairment charge from writing off the capitalized costs allocated to one particular dry well, one component of a block of wells comprising a "cash generating unit." The write off took place even though every other well in the cash generating unit was performing well beyond original expectations.

I explained to the analyst that IFRS (see IAS 36) provides the option to assess impairment at the individual asset level at any time – ostensibly to allow an entity to provide more timely information regarding impairments. But, ulterior motives may have been at work: With oil prices (and revenues) so high lately, the entity likely had more than enough accounting profit to absorb impairment losses this year, and still meet earnings projections. Hence, the entity apparently elected to pave the way for higher future earnings by taking an impairment charge in the current year with the main effect being reduced future amortization of development costs.

One of my philosophies of financial reporting – and a major theme of this blog – is that discretion in financial reporting invariably leads to abuse. In criticizing the proposal that entities may use the qualitative assessment at their own discretion, am I being too cynical? Or, is the Board being disingenuous? As you decide these questions for yourself, consider that a qualitative assessment in no-brainer circumstances shouldn't cost that much in the first place. Also, while there are already plenty of opportunity for earnings management in the FASB's extant goodwill impairment standard, there will certainly be times when entities will want to write off goodwill before issuing its financials, but can't because the MLTN threshold prevents it.

The "more-likely-than-not" probability threshold is not auditable – I have conducted seminars on financial reporting to thousands of practicing accountants; and my own impression, which is corroborated by a colleague who has done even more of this kind of work than I, is that CPAs are not well-equipped, nor are they excited about, making probabilistic judgments. Yet, the latest fad among standards setters is to require them to do so (and in some cases – leases, loans, revenue recognition – to generate entire probability distributions). Again, my philosophy that discretion leads to abuse is directly applicable.

Moreover, who shall review the 'reasonableness' of these probabilistic judgments? In the most important cases, it will be the "independent" auditors of large public companies, who are paid millions of dollars via checks bearing the signature of the company's CEO. I can see it now: two weeks before the Form 10-K is due at the SEC, a newly-hired auditor who still takes his job too seriously will question the reasonableness of the MLTN determination. Should the auditor require a re-assessment, the company won't have enough time to conduct its goodwill impairment test. Guess who gets inserted between a rock and a hard place?

Non-existent cost-benefit analysis – Federal law requires the SEC to justify a new regulation with rigorous and transparent cost benefit analyses, generally with hard data and quantitative assessments. I find it strange that both the IASB and the FASB (essentially the SEC's agent when it comes to making accounting rules) can trumpet the rigor and extensiveness of their "due process" without the same. In ASU 2011-180, the FASB is proposing, without even acknowledging any potential negative consequences, to allow reporting entities to side-step a series of rules that clearly were intended to mitigate shareholder value destruction.

More convergence nonsense – The Board mentions that it considered a number of other alternatives for changing goodwill impairment accounting. I won't get into the details here, but some of those alternatives make a lot more sense that the proposed rule changes. Irritatingly, much of the Board's rationalizations are based on their concern that a revised goodwill impairment standard should move U.S. GAAP closer to IFRS. But in point of fact, the goodwill impairment rules under IFRS are already as different from U.S. GAAP (with no plans for convergence) as night is from day. Under U.S. GAAP, the last long-lived asset to be tested for impairment, and written down, is always goodwill; but under IFRS, all long-lived assets are tested together, and the first asset to be written down is generally goodwill (subject to earnings management options that are currently not in U.S. GAAP).

* * * * *

Financial accounting costs too much because it is too darn complex. I am sympathetic towards efforts to reduce the costs of financial reporting, but in most cases, improvements are only slightly incremental and they come with too much baggage. The latest proposal for modifying "goodwill impairment" accounting is a case in point.

 

Is GM's Equity For Real?

GM's recent filing of its registration statement with the SEC represents the collision of two topics I have written about. First, the fortuitous arrival of $30.2 billion of "goodwill" on GM's balance sheet so soon after the company crammed a bankruptcy organization down the throats of irate creditors and shareholders fits nicely into my collection of goodwill accounting curios. Second, GM's description of how that "goodwill" came to be will allow us to further consider the implications of my concern that the SEC's review of GM's IPO registration could be inadequate.

I became of aware of the goodwill issue from reading Jonathan Weil's article in Bloomberg last week. Out of all of the bad things that happened to GM, at least it became eligible for "fresh-start" accounting (see the Accounting Standards Codification starting at ASC 852-10-45-17). What it meant for GM is that measurement of all recognized assets and liabilities began anew on July 10, 2009, plus a fresh delivery of goodwill from being able to pretend that the reorganized shareholder group purchased the reorganized, "New GM," (which, of course, it didn't), and accounted for it by the rules for a business combination.

As set forth in the fresh-start accounting rules, the imaginary terms of the imaginary acquisition is that shareholders bought New GM for its "reorganization value" and assumed all of New GM's liabilities (of course, the shareholders of New GM don't actually pay anything). Any positive difference between the booked assets and liabilities is recognized as "goodwill"; and an asset shortfall would result in negative shareholders' equity.

Jon makes many interesting observations in his article, and I highly recommend that you read it, but the one point that I am going focus on is that GM's shareholders' equity at December 31, 2009 would have been a negative $6.2 billion if it were not able to book a whole bunch of goodwill. Even though everyone knows that goodwill is merely, as GM itself puts it, "a residual," it can be a pretty important residual. To say that few companies would be able to pull off a successful IPO with a negative number for shareholders' equity on its balance sheet would be an understatement. To say the same after applying fresh-start accounting would be a statement of fact.

Negative shareholders' equity is an accounting anomaly. If a balance sheet purports to portray financial position of the shareholders, then – just like a negative share price – negative shareholders' equity makes no sense. The only possible explanation is that assets and liabilities themselves are out of whack. Who wants to invest in a company, especially after fresh-start accounting is applied, whose assets and liabilities are so obviously out of whack?

Let's Play SEC

So, a couple of posts ago, I asked whether the SEC was capable of functioning as an effective gatekeeper of a GM IPO. GM's application of fresh-start accounting, especially the 'goodwill' part, could well be the acid test.

This could make pretty dry reading, but I'm going to play the role of an SEC staffer who is actually drafting comments to GM.  To avoid dry becoming turgid, I'm dispensing with a lot of the formalties to be found in a genuine comment letter, and virtually all of the standard form.  Anyway, here goes nothing.

Comment #1 – Overstatement of reorganization value

On page 125 of your registration statement, you state that reorganization value is the sum of four components: (1) a discounted forecast of expected future cash flows; (2) "the fair value of operating liabilities"; (3) the fair value of non-consolidated investments, and (4) excess cash balances as of July 10, 2009. The following comment pertains to the fair value of operating liabilities (the second component), which you indicate on page 127 was $80,832 million.

We do not believe that you have determined reorganization value in accordance with GAAP. ASC 852-10-05-10 states as follows regarding the determination of reorganization value:

"…Reorganization value generally approximates fair value of the entity before considering liabilities…." [italics supplied]

PLEASE RESTATE to exclude the "fair value of operating liabilities" from the determination of reorganization value [and perhaps wipe out all of the goodwill in the process!].

Comment #2 – Definition of "operating liabilities"

The definition of "operating liabilities" is provided on page 127:

"Operating liabilities are our total liabilities excluding the liabilities listed in the reconciliation above …"

Nothwithstanding any effect of Comment #1 on your registration statement, the forgoing definition is deficient; it does not actually describe "operating liabilities," except in terms of an apparently arbitrary calculation. Should it become necessary to provide a substantive definition of "operating liabilities" after you have taken into consideration our Comment #1, you should provide a definition that is independent of total liabilities and its non-operating liability components. If necessary, restate the amount of "operating liabilities" to conform to the revised definition.

Comment #3 – Excess cash included in reorganization value

On page 126, you state that the excess cash component of the reorganization value was $33.8 billion, and that $21.2 of cash and marketable securities were restricted as to their use.

We note that if there were no excess cash balances, then reorganization value would be reduced such that goodwill recognized by fresh start accounting would be completely wiped out. In consideration of its materiality, PLEASE AMEND your registration statement to provide additional details as to how excess cash was determined. Provide an explanation of the cash flow forecasting model used, its key assumptions, how those assumptions adequately reflect the exceptionally high risks that New GM faces, and provide a sensitivity analysis on those assumptions in a tabular format. Also, provide a description of the negotiations with creditors that granted you the ability to retain excess cash; specifically, how it came to be that creditors did not successfully claim these amounts (if they were truly excess) for themselves. 

In summary, we are skeptical that, with the possible exception of your restricted cash, that New GM should consider itself to be in possession of excess cash balances.

Comment #4 – Effect of Taxes and Pensions on Goodwill

On page 136, you stated the following:

When applying fresh-start reporting, certain accounts, primarily employee benefit and income tax related, were recorded at amounts determined under specific U.S. GAAP rather than fair value, and the difference between the U.S. GAAP and fair value amounts gives rise to goodwill, which is a residual. … If all identifiable assets and liabilities had been recorded at fair value upon application of fresh-start reporting, no goodwill would have resulted." [italics supplied]

The last sentence, above, is incorrect because goodwill is also a function of reorganization value. Therefore, PLEASE AMEND your registration statement to correctly state the calculation of goodwill, and as follows:  

• Provide, preferably in tabular format, the following: the balance sheet categories for which fair value and recorded amounts differed as of July 10, 2009; their recorded amounts and respective estimated fair values; and the difference.
• In separate tables, disaggregate the forgoing differences by the variables that significantly contributed to the differences. State any assumptions that were required for the preparation of this table.

[Note: the deferred tax asset valuation allowance determination is a fertile source of earnings management. At the date of the fresh-start accounting, a "conservatively high" valuation allowance increases goodwill; and in subsequent periods, reductions in the allowance also reduce reported income tax expense. Therefore, it would behoove the SEC reviewer to examine the determination of the initial deferred tax valuation allowance very closely.]

What Will the Real Comment Letter Contain?

In summary, I reviewed only about ten pages of GM's registration statement, and generated the above comments. Among other things, I found three reasons why New GM probably doesn't have any goodwill, and as a consequence, negative shareholders' equity:

1. Overstated reorganization value due to inclusion of operating liabilities;
2. Overstated reorganization value due to overstated excess cash and marketable securities;
3. Potential overstatement of the deferred tax asset valuation allowance, with the added consequence of creating an earnings bank which can be drawn down by reducing the valuation  allowance.

Eventually, the SEC's comment letter on GM's registration statement will have to be made public. Here's hoping the Commission's staff is doing its job—even if it means that GM's IPO doesn't go forward.

Goodwill Impairment: I Love a Charade

I have written about initial recognition of goodwill on numerous occasions. I suppose it might be less bothersome if goodwill had the good grace to sit and stay like a good dog on the balance sheet at its opening 'value'; but alas, such is not the case. Not even close.

Goodwill impairment tests chew up good money, and as I am about to describe, they screw up the accounting for just about everything else. Even in the best of circumstances (by which I mean that goodwill could actually be something more than a garbage can for overpayments, mis-measurements and measurement exceptions) the goodwill impairment test is a ludicrous charade.

To Illustrate

Let's start with a simple set of made-up facts, albeit loosely based on a real situation as described to me. The teller of the tale attended a business combinations workshop that I had recently led:

Acquiror Company purchased 100% of the outstanding shares of Target, a consulting company for $1,000. None of Target's liabilities were assumed; the only asset eligible for recognition other than goodwill was a customer-related intangible, with a fair value of $400 million and an expected economic life of 8 years. Therefore, goodwill was initially "measured" (the FASB's misleading term, not mine) at $600 million (=$1,000-$400). Why such a high amount for goodwill? Acquiror viewed Target's assembled workforce to be its most valuable asset, which may not be recognized separately under GAAP (or IFRS).

One-year later, negative events associated with the recession significantly diminished Target's prospects. Principally, the value of the previously recognized customer related intangibles declined significantly; however, their remaining expected economic life still has seven years to run. (The remaining expected economic life of the asset is not particularly relevant to my analysis to follow, but there it is, anyway.)

U.S. GAAP requires that the customer-related intangible be tested for impairment before the goodwill is tested. To make a long story mercifully short, one first assigns the current carrying amount to the smallest cash-generating unit (CGU) for which cash flows can be reasonably attributed. Acquiror determines the CGU to be Target itself, and then determines the expected undiscounted cash flows to be just slightly greater than $350 million (which is the carrying amount of the CGU. Thus, no impairment of the customer-related intangible is recognized, even though its value is surely far below the carrying amount.

Now, it's on to the goodwill; and for the FAS 142 cognoscenti among you, we shall stipulate that Target constitutes a "reporting unit." Cutting to the chase again, GAAP requires that the current fair value of Target be compared to its carrying amount. Let's say that the fair value of Target is determined to be $300 million. Since that's less than Target's carrying amount of $950 million (=$350 + $600), Acquiror must launch itself through the gauntlet of the goodwill impairment test, known as "Step 2."

Step 2 requires Acquiror to pretend that it purchased Target today for its fair value, and to figure out what goodwill would be recorded at today, should that impossible fantasy somehow be the reality. So, let's see: assuming a purchase price of $300 million and a fair value for the customer related intangible of $200 million, the "implied fair value of goodwill" (another fabricated and misleading term to add to one's collection) comes to $100 million.

Thus, a "goodwill impairment" of $500 million (=$600 - $100) must be recorded, even though everyone and their brothers and sisters all know that it is the customer-related intangible that is deep underwater. Acquiror's management knows that the customer-related intangible is worth $200 million less than the amount reported on the consolidated financial statements, but investors don't know that. All they see is a writedown to goodwill; which everyone and their brothers and sisters dismiss as merely the result of an arbitrary recalculation of an arbitrary calculation.

So now you know why issuers don't complain too much about goodwill impairment accounting, even though the charade by which it is calculated can be a gigantic pain in the tuchas. It's just one more line of defense for hiding information about real impairments on any kind of long-lived asset other than goodwill that you can imagine. Nothing actually re-measured, only goodwill actually lost.

Is IFRS Any Better?

No. It's worse and with no prospects of improvement anywhere close to being on the horizon. Unless management elects to separately estimate the "recoverable amount" (higher of net fair value and "value in use"), then all of the goodwill and customer-related intangibles carrying amounts are thrown into the CGU bucket. If the CGU's fair value is less than its total carrying amount, then you will always write down goodwill before you touch any of the other long-lived assets.

Although in this case the GAAP and IFRS answer would be identical, the difference is that under GAAP there is at least some chance that non-goodwill assets would be stated at a more realistic value for them. Moreover, the SEC has demonstrated its awareness of the anomaly and willingness to hold a registrant's feet to the fire. For example, here is one case of an SEC comment letter to a company expressing its incredulity that goodwill was written down while miraculously preserving the carrying amounts of its non-goodwill assets:

Taking into consideration the circumstances that caused you to recognize an impairment charge on the Birmingham market goodwill, tell us whether you first tested your long-lived assets … If you did test your long-lived assets for impairment, explain to us in why an impairment charge was not recognized. If you have not tested your long-lived assets for impairment explain to us why not. Please also tell us how you group your long-lived assets for purposes of testing your long-lived assets for impairment …. [Letter from the SEC to Cox Radio, Inc., dated July 17, 2006]

Don't Worry, Be Happy

Small wonder that goodwill and long-lived asset impairment is not on the rush-rush 2011 convergence agenda, or even anytime thereafter. The financial crisis has clearly demonstrated, asset impairment accounting is a sacred cow that may only be approached in circumstances involving extreme unction. Not very long ago, it was hard enough for the FASB to push through any sort of consistent impairment standard for long-lived assets. Now, with the EU already threatening to jump ship on financial instrument impairment, the only choice the Boards have is to pretend that the shortcomings of impairment standards are not a high priority, not to mention the gaping inconsistencies within and between IFRS and GAAP.

I love a charade.

A Sampling of What Lurks at the Bottom of the Goodwill Garbage Heap

I have already reported stumbling upon a fascinating interview of Clarence Sampson, SEC Chief Accountant for more than a decade starting in the mid-1970s. Of his many tales of peculiar interactions with special interests, this one struck me right in one of my biggest pet peeves:

"In the process of recording ... [a business combination transaction] ... they discovered, by golly, that in a $300,000,000 acquisition, $100,000,000 of assets they thought they had didn't exist. And so the company tromped in with their auditors and said, the rules say the difference between what we got and what we paid is goodwill. I simply wasn't able to accept the fact that there should be $100,000,000 goodwill on their books, which didn't exist, and we told them to write it off."

I have explained in a previous post many months ago why I think the process of measuring goodwill and periodically testing it for impairment is a shameful waste of time and money. I would be hard pressed to think of a better example than Clarence's story to back that up. But, I also want to explain why Clarence's story is more than merely an interesting anomaly.

Goodwill (I despise the term, but will use it here for the sake of clarity and with the understanding that it's meaning as a term of art bears no relation whatsoever to what regular folks think it means) arises from two sources. One source is genuine assets that have been acquired, but for various and sundry good reasons those assets are never separately recognized under GAAP. Even the management that bought those assets probably can't adequately explain to you what those assets actually are in anything but very general and vague terms. Yet, in a business combination, we recognize them all together (and mixing them in with liabilities of a similar ilk as part of the process) as 'goodwill.'

The second source of goodwill are 'mistakes.' In other words, paying a price to acquire a company greater than its value. Although the amounts of money in Clarence's story are extreme, the fact of the matter is that mistakes happen all the time. There are business school academics who spend virtually their entire careers trying to explain why it is so often the case that an acquiror's stock price goes down after they have proudly announced their plans to acquire another company. During my part-time career as litigation consultant, I can recall at least four cases where acquirors have claimed that assets they purportedly purchased either didn't exist, or those assets were worth less than they were represented to be worth by acquirees. In all of those cases I was involved in, how did a mistake get accounted for? Capitalized as goodwill, of course! No Clarence Sampson or auditor suggested they do otherwise.

I suppose that one could justify initial capitalization of mistakes as goodwill, because they are impossible to detect at the time a transaction takes place; if they could have been detected, then the purchase price presumably would have been adjusted. But, don't business combination accounting rules give one a full year to adjust the values of assets acquired and liabilities assumed? Sometimes they do, but the rules don't mention that mistakes aren't supposed to go to goodwill; so that's where they go.

But, won't impairment testing eventually catch the mistakes and chase them out of goodwill? Not usually. If it ever should happen that a mistake pops out as an impairment charge, it's usually years after the mistake has become known to management. The goodwill impairment tests allow companies to aggregate subsidiaries into 'reporting units,' which are usually large enough to allow any mistakes to be offset by goodwill from other acquisitions that have accumulated a successful enough track record over time to protect their own goodwill, plus the goodwill generated by any recent mistakes.

At least the big mistakes will get caught by the Chief Accountant, right? Ironically, I doubt whether the current chief accountant or his predecessor would have the gumption Clarence did to stand up to a registrant and its auditor like that. Unlike Clarence, who spent decades coming up through the ranks of the SEC, these guys spent their distinguished careers chest bumping their fellow Big Four partners. When an erstwhile comrade-in-arms "tromps" into the SEC as his client's Doberman Pincer, will he be welcomed with the secret Big Four handshake? But to be fair, today's SEC staff may not have the technical ammunition Clarence did; the FASB's sausage factory has created a new line of business combinations rules; their literal application has come to be the generally accepted method for leveling the M&A playing field…

… as opposed to Clarence Sampson's application of common sense principles:

"And that's the kind of thing that the Commission can say - look that's just too far; you can't look at the written words and try to apply them to a situation where it just doesn't make sense. And as a matter of fact there's some language, and I'll bet you can tell me where it is, which says if it doesn't make sense, you can't do it."

Those "written words" (principles-based rules?) Clarence couldn't specifically recall are still in the cupboard (see Exchange Act Rule 12b-20, and AICPA Ethics Rule 203-1), but they haven't been taken off the shelf in a real long time.

Anyway, I hope you enjoyed Clarence's story as much as I did.

FAS 52: Another Goodwill Charade, and IFRS Convergence To Boot

In a recent post, I argued that goodwill arising from a business combination was just a random number; therefore, any attempt to measure impairment amounted to nothing more than a costly charade.  By coincidence, I recently had an inquiry from a client about the accounting for goodwill at foreign subsidiaries per FAS 52.  Thinking about it for my client also reminded me of yet another sordid tale of accounting convergence for its own sake. 

The General Problem of Goodwill Arising from Foreign Subs under FAS 52

By way of background, FAS 52 requires that assets of a foreign subsidiary with carrying amounts based on historic costs (e.g., plant and equipment) be translated into the reporting currency (e.g., dollars) at the exchange rate existing on the balance sheet date (e.g., the 'current rate).  As I have noted in an earlier post, multiplying historic costs by current exchange rates is the equivalent of multiplying apples and rocks -- with the inevitable result being a random number. 

What about goodwill arising from the acquisition of a foreign subsidiary?  As I am about to show, the choices arising out of a slavish application of the general approach of FAS 52 is so absurd that the FASB buried its own decision deep in an Appendix.  Be that as it may, the specific question is whether goodwill should be measured in dollars, or measured in the currency that is used to measure all of the other assets and liabilities of the subsidiary and translated into dollars.  The implications are the following:

1. If the foreign currency is the answer, then goodwill will be translated to dollars at each balance sheet using the current exchange rate.  Exchange rate movements will affect the carrying amount of goodwill expressed in dollars, which will in turn affect consolidated shareholders' equity through other comprehensive income (the so-called 'translation adjustment'). 
2. If goodwill is measured in dollars from the outset it will remain at a constant dollar amount over time, unless it becomes impaired per the FAS 142 charade.   

From an investor's viewpoint, which treatment of goodwill is preferable? Measuring in a foreign currency and translating with current exchange rates (the first choice, above) adds another information-less element to goodwill measurement, confounding even further any potential to glean something relevant out of the translation adjustment.  Therefore, measurement in dollars (the second choice) would be the lesser of two evils. 

Which answer did the FASB pick?  Hint: look for the answer that creates a consistent set of rules, irrespective of whether the rules themselves make any sense.  Yep, FAS 52, para. 101 requires goodwill measurement in the foreign currency.  The only conceivable reason must be that it is consistent with the treatment of other historic-cost-based assets of the foreign subsidiary.   

Unintended Consequences

That brings me to the question my client asked, which goes something like this:

The functional currency of Company A is the U.S. dollar.  Company A acquires 100% of the outstanding shares of Company B, for which the dollar is also its functional currency.  However, Company B has two foreign subsidiaries, C and D, whose functional currencies are their respective local currencies.  Given the requirement of paragraph 101 of FAS 52, must a portion of the goodwill recognized from A's purchase of B be attributed to C and D? 

It would seem that the fact pattern described above is not uncommon.  Yet, I could not find even a glimmer of insight in the official GAAP literature that would acknowledge that the problem even exists. So this is how my own thinking goes:

• Given that the only purpose of paragraph 101 would appear to be consistent with the accounting for other assets, one would think that a reasonable allocation of goodwill would be intended.  Take the extreme example where the U.S. operations of the acquired company might be insignificant: failure to allocate goodwill among the subsidiaries would result in a significant inconsistency, and perhaps even, an opportunity for manipulation.
• FAS 142, charade that it is, does require a rigorous apportionment of goodwill among reporting units for the purposes of determining whether goodwill is impaired.  A reasonable approach to allocating goodwill to foreign subsidiaries might be similar to the reporting unit approach of FAS 142.   

However, to my surprise, the few inquiries I have made of practitioners with experience in the area are unanimous in the view that, in practice, goodwill would not allocated below the intermediate parent -- in this case, Company B.  Why might practitioners take that view?  Because no rule prevents them from doing otherwise; and because it benefits them to reduce the volatility caused by translating goodwill from a foreign currency to the reporting currency.  In fairness, and as I have noted above, even though what appears to be broad practice may result in inconsistencies, it is arguably better accounting -- or rather less bad -- than results from a rote extrapolation of the extant rules (which is the most I can say of my own reasoning).

The International Convergence Connection

As I have pointed out elsewhere (interest cost capitalization), some changes to IFRS are made for the sake of convergence, with well-reasoned standards thrown on the junk heap for the sake of convergence with U.S. GAAP and the prospect of IASB financial reporting hegemony. The goodwill provisions of IAS 21, the IFRS counterpart to FAS 52, is another case in point.  At one time, IAS 21 allowed either of the approaches we have discussed for goodwill measurement.  But as part of an omnibus project to eliminate alternatives in numerous standards (brought on by the 2005 adoption of IFRS by the EU), the IASB eliminated the option to measure goodwill in the reporting currency, making it the same as FAS 52.  Again, the only conceivable motivation appears to be convergence for convergence's sake. 

I imagine that the progenitors of the original IAS 21 were thinking that at least part of goodwill arises from synergies between a parent that 'thinks' in its reporting currency and a subsidiary that 'thinks' in a foreign currency.  Thus, the source of the goodwill is not strictly foreign or domestic; so, who is to say in which currency goodwill should be measured?  Evidently a race for convergence trumps reason and reasonableness. 

What Good Comes from Goodwill Accounting?

In an earlier post, I described how SFAS 141R resulted in some incremental improvements to the accounting for business combinations.  However, warts remain, and the purposes of this post is describe the ugliest and most painful of them all: the accounting for so-called 'goodwill.'

Here's a simple example to contemplate:

• Company P determines that Company S has a value of $1,100, and negotiates an acquisition for 100% of its outstanding shares for $1,000.
• S has the following assets:
  • Plant and equipment with a fair value of $200.
  • An assembled workforce with a fair value of $100.
• S has no liabilities eligible for accounting recognition.
• Company S will be run independently from Company P; thus, any synergies created by the acquisition are negligible. 

The root of the problem is literally that debits (the assets acquired) do not equal credits (the purchase price).  Business combination accounting is a collision of fantasy and reality: the fantasy is that accounting can fully reflect the economic impact of past events on an enterprise, and the reality is that it cannot be so. A balance sheet produced by even the most principled of accounting systems imaginable cannot possibly comprehend the entire set of economic assets and liabilities.  One example on the asset side would be that S has been put together in such a way as to rapidly and inexpensively expand or contract capacity as market conditions change.  In other words, S holds 'real options', and the shareholders of S would want P to pay for them. On the liability side, not all obligations are legal, amounts are highly uncertain, and the probability of payment may be low. 

The FASB's solution to the debit and credit problem is to plug the shortfall in debits and to weave a fantasy around it.  The plug is euphoniously dubbed 'goodwill' -- to be classified on the balance sheet as an asset and tested for impairment at least once each year. In the above example, the amount reported as goodwill would be $800 (=$1,000 - $200). 

Whipped Cream on the Balance Sheet   

As described above, the amount reported as goodwill is, at its best, a conglomeration of assets offset by liabilities.  Nowhere else in accounting would there be permitted such a hodgepodge, and by no other means other than a narrowly defined 'business combination' may it -- whatever it is -- be recognized.  But even granting that offsetting assets with unrelated liabilities may be permissible, of what use to investors is the assignment of a number to something that, by definition, is beyond description?  (Ironically, even though the value of S's assembled workforce may be measurable and significant, separate recognition of this asset is streng verboten and kept a dark secret from investors.) 

As I have reported in my earlier post, Walter Schuetze (former SEC Chief Accountant and FASB member) derisively characterizes reported goodwill as "the lump left over." Actually, I think he was being generous.  FAS 141R contains some significant exceptions to fair valuation of assets acquired and liabilities assumed.  Thus, the unknown difference between recorded amounts and their fair values are  shoveled into the goodwill muddle.  As if that weren't enough, the math of the goodwill calculation blithely compares apples with oranges: prices paid with values received.  "Lump", "goodwill" or whatever name you can think of implies that the number is associated with actual attributes, but what we are dealing with here is nothing more than just a number--an arbitrary number.

So, dear readers, I hope you are not disillusioned to realize that 'goodwill' is invariably anything but.  If it must be recognized at all, let's drop the obvious pretension and call it what it is: in this example, "excess of purchase price over recognized amounts of identified assets acquired and liabilities assumed."  However, dropping the pretension is not as easy as it seems.  If a muddle is to be reported as an asset, it must be subject to an impairment test; and without a dressy name that belies the muddle that is 'goodwill', there can be no pretense for the charade of an impairment test that is FAS 142. 

The Goodwill Impairment Mess

Recognition of goodwill may seem but a curious anomaly until you get to the impairment test specified in FAS 142.  It's a real money pit:  goodwill has to be assigned to "reporting units" (a new concept rife with opportunities for manipulation); the fair value of each reporting unit has to be assessed at least once a year (another opportunity for manipulation); and the real mayhem begins if, heaven forbid, you are required to estimate the "implied fair value of goodwill" (another new concept rife with opportunities for manipulation).  The only good that comes out of goodwill impairment testing are the jobs created for valuation consultants, accountants and attorneys. 

A Proposed Solution

In olden days, the British permitted a charge to contributed capital for the amount that would otherwise have been recognized as goodwill.  While imperfect, it may well be the only reasonable solution to the problem; for as I have shown above, there can be no perfect solution.  If you can't describe what something is, than what possible good can come from purporting to measure it?

By the way, even though business combinations rules have been somewhat converged by the issuance of FAS 141R and a revised IFRS 3, goodwill impairment remains one of the most significant differences between IFRS and U.S. GAAP. The two approaches are fundamentally at odds, but it should be said that IFRS's impairment rules are much less worse.  But that's not the most important point I want to make.  Whatever the merits of the two approaches, by eliminating goodwill and the inevitably screwball impairment tests, standard setters would not only be improving financial reporting, they could also say that they have resolved one of the thorniest convergence issues.   

Peeling the Onion on the New Business Combination Standards: FAS 141R and FAS 160

This post examines the onion skin, if you will, of the new business combination standards. I'm going to explain the differences between the so-called 'purchase' method of accounting and the new 'acquisition' method. As is my habit, let's begin with a simple example.

Assume that ParentCo acquires 70% of the outstanding shares of SubCo for $1,000. Additional facts are as follows:

• ParentCo estimates that the fair value of 100% of SubCo is $1,405:
  • You should note that the fair value of SubCo may not ordinarily be calculated by extrapolating the purchase price paid to the remaining shares outstanding (i.e., $1,000/70% = $1,429 is not ordinarily the fair value). The reason is that a portion of the purchase price contains a payment for the ability to exercise control.
  • In this case, the control premium would be $55, calculated as follows:

($1000 - .7($1405))/(1-.7) = $55

• It may be difficult to estimate the control premium, because it may have to be derived from an estimate of the full fair value of the acquired company, as above.
• The book value of SubCo's assets and liabilities approximate their fair value, except for one asset with a remaining useful life of 10 years. For that asset, the fair value exceeds the book value by $100.

The table below displays the following: (1) respective balance sheets of ParentCo and Subco at the date of acquisition, (2) consolidated results under the purchase and acquisition methods, and (3) the goodwill calculations under each method.

The purchase method had a lot of warts, but it took the FASB decades to replace it. In essence, it was nothing more than a slavish application of the historic cost principle to mute the future effect of an acquisition on operating expenses. The idea was that if you purchased 70 percent of the outstanding shares of another company, then revaluation of assets would take place only to the extent of the shares purchased. Among other things, this meant that the basis of the assets of a subsidiary acquired in business combination transaction would be the sum of the fair value of the portion acquired and the historic cost of the portion not acquired. This is illustrated above by the calculation of consolidated assets: $2,000 + $800 + .7($900 - 800) = $2,870. The mix of fair value and historic cost to measure one asset reminds me of something my father would say when someone (mainly me) was less than fully committed to a principle: "you can't be half pregnant."

The acquisition method represents a full commitment to fair value, yet ironically, the FASB still doesn't require fair value for all assets and liabilities assumed (more on that in another post). In other words, if the transaction results in the acquisition of control of an entity, assets acquired and liabilities assumed will be initially measured at 100% of their fair value--even if less than 100% of the outstanding shares are purchased. That's why consolidated assets are $30 higher in the illustration under the acquisition method.

The example also illustrates two other important differences between the methods:

• The consolidated amount attributed in consolidation to the non-selling shareholders was� termed 'minority interests' under the purchase method. It was reported on the balance sheet between liabilities and shareholders' equity'; and following the same reasoning allowing half-pregnant measures of consolidated assets and liabilities, was measured based on the historic costs of the subsidiary. 'Noncontrolling interests' (NCI), has replaced minority interests, and are now measured on the same basis as the 'controlling interests', and firmly categorized as part of shareholders equity.
  • Well, maybe not so firm. See my post on the FASB's recently issued Preliminary Views document entitled Financial Instruments with Characteristics of Equity, which proposes to move NCI to liabilities. In fact, one intrepid FASB member objected to the issuance of FAS 141R until the FASB could actually enunciate clear principles for distinguishing between liabilities and equity. Until then, it's just another rule that's subject to change.
• "Goodwill" is higher under the acquisition method, because as can be seen by comparing the calculations, it now includes amounts attributable to NCI.
  • What hasn't changed, though, is that goodwill is still a euphonious sobriquet for a random number masquerading on the balance sheet as an asset measurement. (I'll write more about this in a post soon to come.) Under the purchase method, it's the difference between the purchase price and amounts for the other assets and liabilities recorded. Under the acquisition method, it's fair value of SubCo (equivalently, purchase price plus amount recognized for noncontrolling interests) minus the amounts for the other assets and liabilities recorded. SFAS 141R may describe goodwill in more dignified terms, but the way I just did is more accurate. At least I'm not as blunt as Walter Schuetze, former SEC Chief Accountant, who is fond of calling goodwill "the lump left over." I smile to myself every time I think of that.

Since only balance sheet differences are illustrated above, I have only told part of the story so far. The table below extends the illustration to the end of the first year subsequent to the acquisition of SubCo.

The foregoing illustrates that there is an additional step on the income statement. It may be hard for some people to get used to, but net income is no longer the bottom line. In fact, the bottom line doesn't change in this case, because $3 of additional depreciation expense is recorded, but it is taken out of the earnings attributed to NCI. If you are familiar with the SEC's required presentation of 'net income applicable to common stock (SAB Topic 6.B.1), you will know that the idea is not new.

Well, that's all for now. If you want to look at my spreadsheet, you can get it by clicking here. I'll have a lot more to say about goodwill and accounting for business combinations in the near future, and it won't be pretty.

Note: The original version of this post contained an error in the calculation of the control premium. Thanks to Wallace Enman of Moody's for pointing out the error, and for helping me to get it right. Thanks also to Barb Foerster of Western Union for finding another technical error that I have now corrected.