The Accounting Onion

Financial Statement Presentation: Will Issuers or Investors Prevail?

The comment period has ended on the IASB and FASB's joint discussion paper (DP), Preliminary Views on Financial Statement Presentation, and it looks like there's going to be a rumble between investor advocates and issuer special interests. I have written three posts in the last few months on this topic (one, two, three), and I am drawing on two additional sources for this post: the CFA Institute's comment letter, and Tim Reason's CFO.com piece headlined "Critics Pan New Financial Statements."

The Statement of Cash Flows: Direct v. Indirect Method

The most basic of all proposals in the DP was to fix the loophole in SFAS 95 allowing the indirect method and mandating the direct method for the statement of cash flows. Among those who issuers claim to serve, however, it's a must have; according to the CFA Institute:

"We strongly support the adoption of the direct method cash flow statement as it would be a significant step towards improving financial reporting. The DM [direct method] offers insights into the quality of revenues, earnings and the characteristics of the cash conversion cycle, which are not available from an indirect method of reporting cash flows from operating activities. The DM is also consistent with and achieves greater relevance of disclosure of gross rather than net amounts. We strongly believe that the benefits outweigh the cost of adopting this method. In addition, we believe that the information provided by DM cash flows would be beneficial to management as well as investors. [Bolding in original]

But, according to CFO.com, issuers appear to sing a different tune in their comment letters. Here's CFO.com quoting from IBM's comment letter:

"IBM 'is aware that limited academic research supports the hypothesis that the direct method of cash flow provides better predictive value to future operating cash flows than either the indirect method or the income statement,' wrote Gregg L. Nelson, the company's vice president of accounting policy in financial reporting. But if that were true, he argued, companies themselves would use that information for internal reporting, something IBM and many other companies insist they do not do. 'To state the obvious,' added Nelson, 'future cash flow is largely driven by future transactions. Historical data is of limited predictive value.' IBM, he wrote, believes that revenue, days sales outstanding, changes in accrual balances and company information should be sufficient for investors to make their decisions." [Italics supplied.]

I love that part about 'if the direct method were useful, we'd be using it.'  It makes me feel like Voltaire's Candide, smugly assured by Pangloss, the pseudo-intellectual, that all is for the best; if there had been a better way, management would be all over it.

Right.  Are you really saying, Mr. Nelson, that IBM doesn't actually care to know where its precious cash came from, and where it went? How about cash planning?  If Nelson is actually saying that IBM management cannot answer those questions for itself, then I say "SELL!" Just to ratchet things up a little, ask yourself how credible Nelson's statements would be if they emanated from the ruby lips of his counterparts at GM, Ford, Chrysler or some other company flirting with bankruptcy. 

Implementation costs of switching to the direct method are also a factor noted by issuers. CFO.com cites Intel's letter, which claims that it will cost them $5 million in the first year to convert their accounting systems over to the direct method, plus $2 million every year thereafter. I'm not a software guru, but I have trouble taking such dire warnings at their face value. It can't be too hard to pull out each cash receipt and disbursement, sort them into about fifty different buckets, and add up the totals in each bucket; probably 99.9% of that be done automatically. 

On the other side of the cost equation, I'm guessing that the thousands of investors who actually want to understand Intel's financial statements spend a lot more than $5 million worth of time each year attempting to reverse engineer Intel's indirect method statements, along with trying to make heads or tails of the gobbledygook, boilerplate explanations of liquidity trends in their MD&A (I haven't actually read Intel's). For those of you accounting educators who I am fortunate to have as a reader, you will especially appreciate the millions of hours we spend trying to get students to think while they are standing on their head, as it were, in order to get their brains around preparing a cash flow statement; or just being able to read and understand the indirect format.

The Reconciliation of Cash Flows to the Comprehensive Income

The CFA Institute's comment:

"The requirement to present the schedule in the notes to financial statements that reconciles cash flows to comprehensive income and disaggregates comprehensive income into four components: (a) cash received or paid other than in transactions with owners, (b) accruals other then read measurements, (c) remeasurements that are recurring fair value changes or valuation adjustments, and (d) remeasurements that are not recurring fair value changes or valuation adjustments. We would not get information on operating changes versus investing decisions versus foreign exchange rate effects – currently provided (albeit poorly and/or incompletely) in the indirect cash flow statement reconciliation. We recommend and strongly prefer the statement of financial position reconciliation as it provides more comprehensive information and includes the statement of financial position and direct cash flow statement."

Just like one of my earlier posts, the CFA Institutes wants a line-by-line reconciliation of each balance sheet category. That's essentially what I have been pushing for (see this post), but the CFA is asking for comparative balance sheets in full reconciliation format. I imagine that the CFA Institute is working on the entirely valid assumption that disclosures are not taken as seriously by auditors and issuers as amounts on the financial statements. That's a great point, but I'm concerned that important information can become overly aggregated. For that reason, I had envisioned highly detailed reconciliations presented as tabular disclosures -- like what the IASB requires for some, but not all, accounts.  It's critical that it be required for all accounts.

CFO.com did not report that issuers squawked much about the cash flow reconciliation. Execs, no doubt, are aware that the Boards considered balance sheet reconciliations in their deliberations, but ultimately got enough pushback to contract a case of cold feet. So, issuers probably want to let sleeping dogs lie at this juncture. That's one reason why I would be willing to, albeit reluctantly at this point, settle for note disclosure, as a sort of compromise.

Classification According to Operations, Investing and Financing

CFA Institute supports the Boards' preliminary views to carry the same classification scheme (i.e., operations, investing and financing) across the three main financial statements, but issuers are objecting. Banks, in particular, are not sure how to reasonably go about separating their operating activities from their investing activities. There may also be a difference in preferences between the IASB and the FASB here; the former might be more comfortable with classification decisions made by management subject to some general 'principles' (a view I absolutely abhor), but U.S. issuers may prefer more FASB-like rules to back them up.

I am actually in the issuer camp on this one. I don't think it is possible to make a principled distinction between operations and investing for any industry. Try these:  Is the acquisition of inventory investing or operations? Is purchasing new equipment to replace old equipment an investing or an operating decision? Are planned expenditures on a 20-year-old oil field to extend proved reserves a part of normal operations, or are those expenditures an investment?

However, as I have stated here, I am very much for a principled distinction between financing transactions and everything else. I see no useful purpose in allowing management the discretion to fiddle with what goes where; but, it should be a straightforward exercise to write rules for classifying financing transactions apart from everything else.

Overall, though, I think organizations like the CFA Institute treat the FASB much too gently. They should be telling them also that, even though there are some satisfactory portions of the DP on financial statement presentation, it doesn't come close to going far enough. They should be telling them that convergence is not near as important as getting it right for U.S. investors. They should be SCREAMING that irrespective of the rules we have for recognition and measurement or the reporting of other comprehensive income,  financial reporting will never be what should be without a full balance sheet reconciliation and a direct method statement of cash flows.

Reason #2 to Dump on the IASB/FASB Leasing Proposal

In my previous post, I let 'er rip at the IASB/FASB "preliminary views" on measuring lease contract assets and liabilities. A close second on my list of investor-unfriendliness contained in their joint lease accounting Discussion Paper is the proposal that lease options (e.g., renewal, purchase, cancellation, etc.) and all other separately identifiable components of a contract should be balled up into one big ole asset and a corresponding liability. 

My unsubstantiated suspicion is that the leasing industry served this one up to the boards in exchange for lowering their resistance to the elimination of operating lease accounting.  It seems plausible, because, as I'm about to illustrate, that if we fail to require separate measurement for all of the material components of a lease contract, we will end up right back where we started. In other words, off-balance sheet lease accounting will continue to plague us.

Here's the example:

• The fair value of an airplane is $100; economic depreciation is expected to be $10 per year. Airline A enters into a one-year lease contract with Lessor B for $12, paid in advance. The lessee has an option to buy the airplane for $90.01 at the end of the one-year lease term.

• Separate and apart from its lease with Lessor B, Airline A purchases an option from a third party, Lessor C, to enter into another $12, one-year lease, commencing one year from now. The leased asset under this 'option to lease' would be a one-year-old airplane, with an option to buy for $80.01 at the end of the one-year lease term.

• And so on, and so forth. Airline A also enters into similar option contracts for subsequent years with new, or perhaps the same, lessors.

In accordance with the DP and the FASB's preliminary views, Airline A will conclude that the probability of exercising the purchase option is less than 50%. Therefore, Airline A will not add any portion of the expected future cash flows related to the option to the measurement of the rights acquired under the lease, or to the corresponding liability. Also, the numerous 'options to lease' do not meet the definition of a derivative per FAS 133 or IAS 39 (physical delivery of the leased asset will be required), so they won't be recognized until exercised.

Bottom line: $12 lease expense, no asset, no liability. Even though operating lease accounting will supposedly gone away, Airline A essentially achieves operating lease accounting in Year 1, and for subsequent years as well.

Is there a cost to constructing this crazy web of leases and options? You betcha, but I don't expect that would deter Mr. Schmo CEO from incurring that cost to get the accounting he 'needs.' Moreover, the result under the IASB's preliminary views would be even sweeter. (Are you surprised?) Let's say that the lease crosses over two accounting periods. The IASB approach would recognize $55 (=$10 + $90.01/2) of expense in the first fiscal year. This creates a $45 earnings bank that can be used in the second year. If Schmo CEO doesn't 'need' more earnings and the purchase option is in the money, then he will go ahead and exercise; otherwise he will let the option expire and withdraw $45 of earnings from the bank.

The Stated Rationale for Rejecting a Components Approach, and Why They're Wrong

Here's the boards' reasons—and my reasons for why theirs are bad reasons:

• A components approach may be difficult to apply.

Funny, but that hasn't stopped the boards from promulgating anti-abuse provisions before. I'm thinking about the FAS 133 requirement to bifurcate embedded derivatives. I'm also thinking about FIN 46(R) a complex an anti-abuse standard in its entirety. The point of my airline example is that, without some sort of components approach, the leasing industry is going to have a field day offering new and 'innovative' financial products designed solely to minimize the effect of lease capitalization on the balance sheets of lessees; and it's why I think all of this may be their idea.

• Components are often interrelated. For example, recognizing a liability in respect of a residual value guarantee may not provide useful information to users if the lessee is likely to exercise a purchase option.

That's like saying that the fair value of an option doesn't provide useful information if the probability of its exercise is low! It certainly has not deterred the boards from requiring fair value measurement for all options within the scope of FAS 133 or IAS 39, regardless of their probability of being exercised.

Having said that, however, any sensible and principles-based standard should provide that insignificant lease components need not be separately measured, and perhaps combined with the most basic rights and obligations associated with the lease contract.

• All components would have to be measured on the same basis, like fair value, in order to result in comparability and to curb abuse.

Amen. If you will be so kind as to read my previous post, I may be able to convince you that a principles-based approach to lease accounting would perforce require measurement of all rights and obligations at their replacement costs.

• The fair value [or replacement cost] of some components may be difficult to measure, because there is no market for them, and they are not normally priced separate from the lease contract.

Right again, but generating random numbers that are misleadingly labeled 'historic costs' is not at all a substitute for a good faith effort to measure a real financial attribute of an asset/liability.

But, I do understand the cost/benefit tradeoff, and would retain operating lease accounting for small, non-public companies. The IASB has done a good thing by creating a simpler version of their standards for use by smaller companies. If there is one area in which the FASB should follow the IASB's lead, it is in providing more accomodations to small companies, and leasing is an excellent place to start. But, while they are at it, they should also relax the requirements in FAS 133 for bifurcation of derivatives from certain host contracts, ease up on business combinations, goodwill impairment testing and FIN 48 (deferred taxes).

If you think I'm being paranoid about potental abuses of a new lease accounting standard that is not principles-based, then think about this:  if it took me 10 minutes to cook up my little example, then who knows what other schemes currently lurk in the minds of the 'financial engineers,' who actually get paid to think up new ways to steal from shareholders?

Lease Accounting: Replacement Cost is the Only Hope for a Principles-Based Solution

A couple of weeks ago, the FASB and IASB jointly published a discussion paper (DP) entitled Leases: Preliminary Views. I take no pleasure in saying that SFAS 13 is the oldest worst accounting standard currently living and breathing; and the DP is a wet blanket on my hopes that it (along with its politically tweaked clone, IAS 17) would ever be replaced with a principles-based standard.

SFAS 13 was issued almost 35 years ago, irreparably broken right out of the box. For decades, it has been Exhibit A for critics of bright line rules, while IAS 17 has been similarly excoriated from the outset for its fuzzy rules. There has been no better example than these two standards of the proposition that rules without principles cannot result in high quality financial reporting—unless you're a big fan of off-balance sheet financing and earnings management. The rules are so bad, that a flurry of articles and textbook chapters appeared almost immediately after SFAS 13 was issued to guide users in their restatements of 'operating lease' accounting.

Financial reporting lived with SFAS 13 for 20 years until some quasi authority, the G4+1 group of standard setters, finally published a discussion paper advocating the capitalization of all leases. Duh.  Ten more years had to pass for the FASB and IASB to commence its joint project on lease accounting. It has taken almost another three years for these accounting whizzes to come out with this watery "preliminary views" document.

Why so long? First, issuers are in no hurry to give up the chicken salad they ordered, and were served on a silver platter in FAS 13; and second, as we are finally coming to recognize, investors have not pushed hard enough on their own for the changes in accounting they are entitled to. If there is any good that might come out of the mark-to-market debate following the financial meltdown, it is the realization that investors can't count on any other 'stakeholder' group to help them unless they themselves become a much stronger voice.

The Good News First

By far, the best thing to come out of the PV document is the proposal to treat all leases as giving rise to an asset (termed the "right to use the leased item" in the PV), and a liability for the obligation to pay rentals. In other words, 'operating lease accounting' and the gift of off-balance sheet accounting that went with it, may finally be out the window! Unfortunately, the rest of the PV follows a long-practiced FASB gambit of doing just enough, and no more, to cross the project off the list -- and to keep those pesky investors at bay for another 15 – 30 years.

Now for the bad news. Actually, there's lots of it, but I don't have time to write a book. So, in this post I'm just going to focus on the one issue that relates to some of my previous posts on 'fair value' versus 'replacement cost' measurements.

Let's Pretend These Numbers Actually Mean Something

To illustrate the point I want to make, let's take a relatively straightforward variation of a plain vanilla 8-year lease on retail store space: the lease term is divided into an initial 5-year period with an option to renew after 3 years at an inflation-adjusted price. Monthly rents contain a fixed component plus a variable component depending on store sales.

The FASB and IASB have slightly different preferences at this point, but let's start with the IASB's only because it's easier to explain: they propose that initial measurement take place as follows:

Estimate the expected lease term, and the expected (i.e., probability weighted) cash payments to be made by the lessee; and discount the expected cash payments at the lessee's incremental borrowing rate.

That's rules-based accounting for you. The DP does not state what the objective of this calculation is, because there isn't one, other than to come up with some number, any number, which they think enough people will say that 'it looks about right.' Stated plainly, the calculation results produces a hash—a number that can only be described by the chicken salad recipe used to cook it up.

There actually is a better way to measure a lease based on discounted cash flows, but it only applies in very restrictive circumstances. Let's set aside the choice of discount rate for the moment, which is a big problem in and of itself, and focus on measuring the numerators, if you will, of the terms in an expected net present value calculation. Finance 101 would say that there should be one term for each possible year (i.e., the maximum lease term), and the amounts for each year should be their expected cash flows.

Too see why that's the only rational approach to present value, think about how you would calculate the present value of an annuity for the rest of your life. Choice 'a' is to calculate your life expectancy (say, 80 years) and to discount the cash payments you would receive only up to the year in which you attain the age of 80. Choice 'b' would be to discount the expected cash flows for each year you could possibly live, where the expected cash flow is the contractual amount multiplied by the probability you should live so long. Choice 'b' is obviously the correct one, and should also be the approach for measuring the present value of a lease obligation of an uncertain term.

For some reason I fail to grasp, the flawed choice 'a' is essentially what the IASB is advocating. The DP describes some other, even more obtuse approaches they purportedly considered, but Choice 'b' is not even given lip service; probably because they knew of no way to put it back in the bag once they let it out. In this, and other aspects, the DP's reasoning is disingenuous at best, but more likely, intentionally deceptive.

Now, let's talk about the discount rate. Using the lessee's incremental borrowing rate for a fixed term lease could make sense, because the result of the calculation would yield the 'settlement amount': i.e., the amount that should discharge the liability in a single payment at the inception of the lease. But, in the presence of the renewal option, or any other type of option, the incremental borrowing rate is just a number pulled out of the air; kind of like adding red food coloring to artificially flavored jell-o. Where the FASB and IASB differ is only with respect to the shade of red: in place of expected cash flows, the FASB would use "most likely" cash flows. Big Hairy Deal. Nothing about either calculation method will produce historic cost, fair value, replacement cost, settlement amount, or any other attribute of assets and liabilities. Neither calculation will result in a "fair presentation" or a "true and fair presentation," or any other paliative weasel phrase the boards want to associate with the financial reporting solutions they are pushing.

A Principles-Based Approach to Lease Accounting

There are two basic principles of lease accounting that should guide the project. First, I agree with the FASB that lease contracts give rise to assets and liabilities; operating leases should be history. Second, and this is the payoff from reading this overly long post, the only relevant measurement attribute of rights and obligations arising from lease contracts is replacement cost. In fact, if you believe that non-transferable rights should be recognized as assets, I can think of no better example to show why replacement cost is preferable to fair value as a comprehensive basis of measurement.

Both boards take the position that contractual rights meet the conceptual definition of an 'asset,' including non-transferable contractual rights. But, if you are contractually prohibited from selling something, the only logical answer to the fair value question is 'zero'! Imagine the fair value balance sheet of an airline that finances its entire fleet of planes with non-transferable leases. It's not just ugly, it's of absolutely no use investors.

However, the replacement cost of a leased asset, regardless of transferability, can be substantial. And, unlike the FASB or IASB's improvised approaches to measurement, any rules that the boards would deign to impose with regard to how one should go about estimating replacement cost can be based on sound principles.

For stating that operating lease accounting should go away, I give the boards a 10. But for the rest of their preliminary views, even though I haven't told you about everything that bothers me, I give them a big fat zero. When I ponder the measurement rules they are proposing, I can't help thinking of the game we used to play as little tykes: "Let's Pretend."

Replacement Cost Rebound

In the public FASB meeting on Tuesday, one of the topics was revenue recognition under circumstances where the consideration received is other than cash.  The Board discussed how to value the consideration received, and I am told that a majority of the Board favored fair (exit) value.

HOWEVER, Chair Bob Herz made a "strong case" for replacement cost. I don't know anything yet about the substance of Bob's comments, but following on the publication of the SEC's study of mark-to-market accounting, this is the first ray of hope that a full-blown reconsideration of fair value could occur.

Another reason I am writing so soon again is to provide those of you who happened to read yesterday's epistle via email with a link to Walter Schuetze's speech, which I inadvertently omitted.  I also want to take this opportunity to address some of the concerns that some readers have already expressed about the positions I have taken. An investor named Andy and an experienced appraiser, Alfred King, have provided representative comments.

The Preferences of "Investor Andy"

Andy wrote that he prefers historic cost "about 80% of the time" because of its objectivity; for the other 20%, mainly marketable securities and loans, he prefers fair value. He is "…most concerned with what an asset cost, because that represents the initial investment, and will allow me to calculate a return on investment."

As to the claim that historic cost is objective, in yesterday's post I spoke primarily about the balance sheet, yet nothing makes my point better than an historic cost income statement.  There is nothing objective about allocations to the income statement for depreciation, bad debt expense, warranty costs, and much more. 

But, the comment I really want to respond to concerns how to go about measuring return on investment.

Consider the oil and gas industry. The value of a producing field has nothing to do with what it cost to develop perhaps twenty years ago, or even how much it has produced to date.  Valuation has everything to do with all sorts of things ignored by historic cost accounting: current oil prices, the current amount of reserves, and the current capability of the current owner to operate and manage the field to maximize its remaining economic returns.

If the current operator cannot generate an adequate return based on the current value of the field, then management should consider selling it to an operator that can achieve a return commensurate with that current value. Historic cost cannot provide me with the information I need to judge whether the current operator is making an adequate return. I have consulted in the oil and gas industry, and the company I worked for adjusted its goals for field managers every year; those goals had absolutely nothing to do with the sunk costs that were reported on the balance sheet by historic cost accounting; and I am positive that investors take the same approach as management.

In summary, Andy, I want the managers of my investment to maximize profit, not simply to generate a satisfactory return based on the amount invested some arbitrary time ago.

Andy also accused me of having a "double standard" when comparing fair value to replacement cost:

"… You eviscerate fair-value accounting, not theoretically, but how it was put into application. Alternatively, you do not consider how replacement value could be taken advantage of preferring instead to think about it theoretically."

I plead "not guilty." I stated that fair value has problems in both principle and its application. But, to be clearer, I should have stated that SFAS 157 is a failed application because it begins with a flawed principle: the flaw being that wealth is command over goods and services. There are two parts to this. First, goods and services are not cash, and fair value measures the potential for cash. Second, and much more important, you can't command anyone to buy your asset. However, every seller does have a 'reservation' price; therefore, you can command goods and services by bidding high enough. In that sense, Andy, replacement cost, as its name implies is actually nearer in concept to historic cost than fair value.

The Appraiser, Alfred King

I have had a series of interesting exchanges over the last few months with Alfred King, author of Executive's Guide to Fair Value. I haven't read the entire book, but I generally agree with his critique of FAS 157.

After today's post, Alfred King wrote, in part:

"I spent about two years of my life working with replacement cost and price level accounting [when that information was required by GAAP].  With the best will in the world, and with some clients no limit on the resources that could be applied, we still could not come up with satisfactory information. …. Nobody in their right mind would replace an inefficient factory building that had been added to five times over 25 years at the same site.  You could build a new modern building with 20% less square feet, and much lower utility costs if you really would replace it.  But of course buildings go on for many many years." [emphasis in original]

"This was one, but not the only, reason the supplemental information died a natural death within three years.  I remember it so well because for about six months we in the valuation business thought it was going to be our equivalent of what the 1933 and 1934 Acts did for Public Accountants.  It wasn't, and we were soon back to our old type of valuations."

I will admit that I haven't read FAS 33 in years, and don't have much recollection as to the details of the requirements. Although Mr. King's critique may have been an important consideration when FAS 33 was in effect, I don't relate it to my vision of what replacement cost accounting should now be.

The most straightforward way to determine replacement cost to meet the wealth measurement objective is to ask oneself what would be the least amount one would have to pay for an asset (or a similar asset that provided the same utility), if one did not actually already own it. It seems to me that real estate appraisers make estimates for specific properties on that basis as a matter of course. Often, their best estimate is the result of making somewhat objective adjustments to 'comparables' for age, floor space and even location.

Having said that, I would allow for any number of approaches to approximating replacement cost, so long as they adequately answered the question I posed in the previous paragraph.  Like FAS 157, the greater the subjectivity in the estimates, the more detailed would be the disclosures.  However, in all cases, I would require reconciliations of the changes in balance sheet accounts in sufficient detail to make all assumptions, and changes in assumptions, transparent.

To be sure, there must be many implementation hurdles to overcome in broadly applying replacement cost accounting, but for goodness sake, consider the alternatives. Sorry, Andy, but I think historic cost has become a bucket of offal, and fair value is falling flat on its face.

Beyond the "Inactive Markets" Problem: The Crisis in Accounting

"It's painful now to read a lecture that Mr. [Lawrence] Summers [who is currently President Obama's principle economic adviser as chair of the NEC] gave in early 2000, as the economic crisis of the 1990s was winding down. Discussing the causes of that crisis, Mr. Summers pointed to things that the crisis countries lacked — and that, by implication, the United States had. These things included "well-capitalized and supervised banks" and reliable, transparent corporate accounting. Oh well."

Paul Krugman, New York Times, March 29, 2009 [emphasis supplied}

"Oh well," indeed. It seems that NYU economist Nouriel Roubini has been given the lion's share of the credit for prognosticating the economic calamity we now find ourselves in. But many independent-minded accountants have warned, decades before Roubini, that our financial reporting standards – not just those rules having to do with mark-to-market – are as thin and flimsy as the paper they are printed on. That nobody would heed the warnings can only be due to the collective weight of the business and banking lobbies pressing on standard setters, with the paid support of the Big 8, 7, 6, 5, 4.   

Contrary to those with whom I usually stand on financial reporting issues, I am not going to excoriate the FASB for its latest kluges to GAAP, which they made at the behest of Congress.  Methinks that a few little tweaks to make banks look sounder was better than at least one other alternative: doing nothing. Who can argue that a 20% run-up in bank stocks isn't something that the economy needed RIGHT NOW?

My own frustration and sense of gloom is from two sources.  First, next-to-know-nothing-about-accounting congressmen may be further emboldened to hunt for messengers to shoot instead of focusing on real problems.  Second, as Krugman points out, financial reporting has been exposed as totally kaput, and it ain't getting fixed any time soon.  The question policy makers should be addressing is not whether we should keep GAAP or adopt IFRS, but rather how long it will take for both the FASB and IASB to become completely discredited as agents of investor protection.

Accounting Needs to Start with a Clean Sheet of Paper

Everyone agrees that there is no silver bullet that can fix the economy. Whatever the capitalist or socialist ideologues will preach, we can't simply scrap the current system and make a new one; every policy decision has to be made at the margin.

But, when it comes to financial accounting and regulation of financial institutions, I believe that 76 years after the enactment of the first federal securities laws, establishing the accounting and auditing disciplines as we know them today, it's high time to start over.

First, we have to start by completing de-linking those banking regulations from the rules of financial accounting. Second, we need to recognize that neither historic cost accounting nor fair value accounting are providing the answers we need. The second point is where I am going for the rest of this post.

The Choices: Fiction-, Fantasy- or Fact-Based Accounting?

The three candidates competing to be the foundation of accounting standards are historic cost, fair value, or replacement cost: stated another way, the choices are fiction, fantasy, or fact. By that characterization, I suppose you can already guess which of the three I would choose.

Historic Cost Accounting as Fiction

The conventional wisdom is that historic cost accounting is fact-based, objective and verifiable. The sad truth is that it is a fiction, however it has been practiced.

To account for historic costs, we subjectively choose which portion of a particular cost should be capitalized; and consequently, which portion should be expensed immediately.  Of the capitalized costs, we have to decide on a plan for transferring them to the income statement.  There can be no right answers, and there is hardly a pretense of impartiality in the numbers that management presents for the auditors to bless as "reasonable."

But even setting aside concerns of bias, so-called "historic cost accounting" only pretends to measure the historic cost of, say PP&E. Instead, fragments of historic costs are dumped in a parking place, bearing the fictitious name PP&E on a financial statement, fictitiously and pompously titled the "statement of financial position."  So, cutting through the pretense, historic cost accounting is simply a rules-based protocol for describing past cash flows and selected 'expected' future cash flows. Big deal.   

Proponents of historic cost, defend it as "verifiable," but as the example of PP&E illustrates, net book values are not verifiable; neither is the allowance for doubtful accounts verifiable, or any other estimate management makes.  The term "verify" stems from "truth", and there is nothing true about these numbers other than the unavoidable fact that management made them up.  

And the auditors?  We derive little value, if any, from an audit report that opines on "reasonableness" of the estimates that make up those fictitious book values.   Walter Schuetze, former FASB board member and SEC Chief Accountant (where he was my boss) gave a zinger of a speech in 2003 on this topic a few years ago to the New York State Society of CPAs. Among other things, he proposed a much more limited role for auditors (and an expanded role for other independent appraisers).  To my mind, the speech is Roubini-esque – so disquieting that the accounting profession and its regulators could respond only by acting as if one of the most visible and prominent accountants in the world did not make it.  If I were teaching an accounting theory class or auditing today, I would spend at least two sessions discussing its implications.  You can access it here.

Fair Value as Fantasy

Fair value, sometimes referred to as mark-to-market or exit value accounting, consists of a series of predictions of the outcomes of liquidating transactions that probably won't occur.  Some people like exit values, especially for gauging the health of financial institutions, because the values assigned to assets provide some indication of margin of safety in the reporting entity's financial structure, should liquidations be required to pay off creditors.

But the sturm und drang of the past few years in implemting SFAS 157 should be some indication that fair value has some real problems -- if not as a principle, then at least in its implementation.  I think it's both.

The fundamental problem with an exit value approach is the need to imagine the market conditions that would exist at the time of the hypothetical liquidation. Therein lies the fantasy; the Pollyanna FASB wants preparers to assume that liquidation won't happen under duress, but the reality is that unplanned liquidations happen under duress more often than not.  Not only that, whatever the market conditions that are to be imagined do nothing more or less than replace the market conditions that actually did exist around the time of the balance sheet date. 

To clarify the implications of fair value, let's momentarily step away from financial assets, and consider a very simple scenario (inspired by an unnamed source, who claims that it is based on fact):

A Ford dealership in rural Vermont has an inventory of 300 model F150 pickup trucks. The dealer sold about 50 - 60 trucks a month in 2006 and 2007.  In 2008, the dealer sold 30 trucks all year, with only 10 from July through December.  In January 2009, the dealer had one sale.  In February 2009, the dealer had no sales.  In March 2009, the dealer has had no sales.  What is the fair value of the 300 trucks in the dealer's inventory? 

Zero, or something close, would be a simple answer that everyone can understand, but most would consider useless as information relevant to assessing the present financial position and future prospects of the dealership. But, under these circumstances, no higher amount could be legitimately described as an 'exit value,'  simply because anything other than a recent price is a made-up number.   

Seeking fair value is just as problematic for subprime mortgages as it is for those pickup trucks. That's because, as I just stated, the question of 'fair value' or 'exit value' is nonsensical when there are no currently active markets for either trucks or subprime mortgages. Valuing a subprime mortgage by the discounted cash flows to be derived from holding it does not magically change a 'value in use' to an 'exit value,' even if the parameters to the DCF model are based on so-called 'marketplace assumptions.'  When we rely on management to estimate the sales price that would clear out the inventory of pickup trucks, or to estimate default rates and discount rates for subprime mortgages, that's fantasy and not fair value. Even worse, we are back to where we started with historic costs: biased estimates buttressed by audits of, to be kind, questionable value.

Replacement Cost as Fact

Every western economist, from Paul Krugman to Gary Becker, accepts as axiomatic the notion that wealth is manifested by the command over goods and services. For a non-cash asset, the appropriate measure of its wealth effect is the amount it would cost to replace it (let's leave aside the question of obsolete assets for the moment).

In a nutshell, if you think accounting is about reporting financial margin of safety, then fair value is for you. And by "you," I mean primarily the bank regulators. But if financial accounting is supposed to be about reporting wealth and changes in wealth to investors, then you gotta go with replacement costs.

Replacement cost accounting is fact-based. It accounts for the wealth you now have. It's not about what you hypothetically could have if, in some imagined scenario, you suddenly  sold an asset for its 'fair' price, whatever that means. Replacement cost accounting is most certainly not about fantasizing that you can command a nice price for an asset from selling it in a market that may not, and in many cases does not, actually exist. Replacement cost accounting is much simpler and, dare I say, saner: it measures actual wealth and actual changes in wealth – isn't that what we were taught accounting is supposed to be about?

Of course, measuring the replacement cost of what you have is subject to error, just as is any other measurement. Getting back to trucks, how does replacement cost accounting handle the absence of a market in which to sell an asset? Perhaps not easily, but a heckuva lot easier than FAS 157, or any other implementation of exit value accounting could. The Ford dealer still knows almost exactly how much he would have to pay to buy 300 trucks, even if every single Ford dealer in New England were feeling the heavy breathing of lenders on the backs of their necks. Nevertheless, the price would be a lot higher than zero; and, that's because the trucks are still worth a substantial amount of money. As for the subprime mortgages, even though there haven't been a lot of recent transactions, dealers and other holders should have a pretty good idea (indepdent of management) of how much it would take others to part with their holdings.

I prefer fact-based reporting of wealth to fantasies of exit value. In both of the new proposed FSPs, the FASB has invoked the SEC's recent study on fair value accounting. That study has also recommended that replacement cost should be given a closer look.

When the FASB takes time from responding to the congressional command to chase its own tail, perhaps we can take out that clean sheet of paper and write accounting rules that are based on fact, instead of fiction or fantasy. I'll be honest, I don't know if it can be done, but I think it can; and we won't know until someone tries.