April 8, 2023
By Oussama A. Nasr*
Background
For many observers, the catastrophe that befell Silicon Valley Bank (“SVB”) in March can be attributed in large part to accounting principles that tolerate alternatives to fair value accounting. It will be remembered, for example, that bonds classified in the held-to-maturity category (“HTM”) need not be marked to market either on the balance sheet or in the income statement when their price falls in response to interest rate increases. The rationale for this rule is that fluctuations in value prior to the maturity date will prove temporary in most cases and should therefore be ignored if the owner has the “ability and intent” to hold the bonds until maturity. It is this very principle, however, that permitted SVB to meander slowly towards economic ruin, without recognizing a fraction even of the billions in unrealized losses that afflicted its HTM portfolio – whether on the balance sheet or in the income statement.
To many of these observers, the solution (or deus ex machina, as the more pompous might say) lies in fair value accounting. Under such a regime the bonds would be recognized on the balance sheet at the price, approximately, at which they could be sold on the date of measurement. Increases in price would additionally be reflected in the income statement as gains while declines in price would be reflected as losses. Greater transparency would prevail and the balance sheet would, among other consequences, provide an accurate snapshot of what would happen if the bank were liquidated immediately. If in liquidation the fair value of the assets exceeds the aggregate amount of liabilities, the bank’s net worth is positive and the shareholders can expect to recover an amount equal to this excess; while if in liquidation the fair value of the assets falls short of aggregate liabilities, shareholders will be wiped out and liabilities will endure a partial haircut equal to the excess of liabilities over assets.
Much of this seems eminently desirable, and in many ways an improvement on current rules. We argue in this paper, however, that fair value accounting produces a number of bizarre results that arguably overwhelm its benefits in aggregate. We leave out of this discussion the important question of how fair value accounting might deal with assets for which no market exists.
The art dealer.
We begin by imagining a jurisdiction in which art dealers are required by law to have a net worth of $8 million at least. Art dealer AD starts off with $10 million in cash and no liabilities, implying a net worth of $10 million; he is thus able to obtain a license to practice his trade in full compliance with the law. AD now identifies an attractive painting that is for sale for $12 million. He obtains a long-term loan for $2 million, adds to this his original $10 million, and purchases the painting. He now has assets of $12 million and liabilities of $2 million, implying an unchanged net worth of $10 million, all as expected. He retains his license to deal in artwork.
Then one day it is revealed that the painter of AD’s new acquisition has a criminal record involving acts of moral turpitude. The price at which the painter’s various artworks clear the market drops uniformly by 20%. The painting owned by AD will most probably fetch only $9.6 million if it is sold immediately, but AD is adamant that he will not sell it until its price recovers to $12 million at least. The loan being a long-term loan, he can afford to wait. Would/should AD keep his license?
Most of us would feel, I doubt not, that AD’s assets are now worth $9.6 million only, while his liabilities still equal $2 million so that his net worth is $7.6 million – a violation of capital requirements. At this point AD’s license is probably better used to light up the last of his Cuban cigars. In reasoning in this manner we would be deploying fair value accounting in its most virtuous incarnation: the affected party is prevented from hiding behind some archaic principle of historical cost accounting – specifically one that would allow AD to continue valuing the artwork at $12 million, on the basis that he has the ability and intent to hold onto it until it returns majestically to its original value or higher.
An important difference between AD’s situation and SVB’s is that SVB’s HTM portfolio is contractually required to be redeemed at par after some specified number of years. Nothing similar exists in the case of the artwork, it will be argued. But AD would insist that (i) the artwork can be expected to rise in value over the long term as memories of the painter’s moral turpitude recede and buyers increasingly choose to separate the man’s works from his repugnant deeds, and also (ii) the artwork can be expected to rise in value over the long term by the cumulative rate of inflation. In short, after a sufficiently long passage of time, the painting is bound to reprice at or above $12 million, just as the bonds in SVB’s HTM portfolio, after a sufficiently long passage of time, are bound to reprice at par. Round one of the dispute appears to favor fair value accounting.
The bank with the ALM gap.
But now consider a bank that begins operations with $10 million in equity. This bank attracts $90 million in new 1-year deposits on which it pays 1% interest annually. With $100 million in hand the bank purchases a $100 million 30-year treasury bond that pays 5% interest annually. Everybody agrees that in Year 1 the bank earns $5 million in interest income and incurs $0.9 million in interest expense. Ignoring operating expenses and taxes, the bank has Year 1 net income of $4.1 million and a return on equity (“ROE”) of 41%. Stellar by any measure, although the inclusion of operating expenses and taxes would have reduced this figure by an amount that is not trivial. We assume that the bank distributes in dividends its entire net income for Year 1 and prepares to replicate its success in Year 2.
Just as Year 2 is about to begin, however, interest rates rise by 1% along the entire yield curve. Of course this does no good for the bank on the left-hand side of its balance sheet: the bond has a fixed coupon that continues to generate $5 million exactly during the course of Year 2, but not a penny more. On the right-hand side, however, the liabilities reprice at 2% annually, so interest expense for the year now climbs from the previous $0.9 million to $1.8 million. Net income drops to $3.2 million and ROE follows it down to 32%. Fair-minded people would describe Year 2’s performance as quite excellent still, although not as excellent as that of Year 1. The bank remains a solidly profitable institution with a healthy net worth, very clean balance sheet, robust interest margins and appetizing prospects. Shareholders and depositors should be ululating, and should be very eager to continue doing business with this goldmine that gives and gives.
Or should they? After all the bond’s value would have declined from par to 86.8 (please trust me on this, or verify it by means of the PV function in Excel), generating a loss to the bank of $13.2 million under fair value accounting principles. This turns our solidly profitable institution into one that is drowning in oceans of red; it also eviscerates the institution’s net worth, which approaches zero. Our previous description of the Year 2 performance as “quite excellent still, although not as excellent as that of Year 1” suddenly seems like fantasy.
And yet the bank began life with a net worth of $10 million and has remained meaningfully cash-flow-positive since then. Is there not something deeply unsatisfying about an accounting regime that declares insolvent a well-capitalized institution whose cash flow has been significantly positive in every year of its existence? To whom should we award the next prize in our accounting dispute? Read on!
The financial trainer.
A final thought-experiment begins with a financial trainer who lectures on fixed income and derivatives. A very liquid lunch with an important client concludes with a contract that requires the trainer to teach one course annually in return for a $1,000 fee for the next 10 years. The trainer is elated by this development and chivalrously treats his spouse to a beautiful bouquet of 24 red roses and a juicy steak dinner (or, more likely, kale salad with fresh balsamic dressing), even though he knows that accounting principles preclude him from recognizing a penny of revenue until he begins delivering the courses. Still, life appears to be good as it gets . . .
. . . until one day GAAP is amended to require the application of fair value accounting to service contracts. Meanwhile the appetite for training courses in fixed income and derivatives explodes, and trainers with comparable competence to our own find themselves able to charge $2,000 per course. Does this development affect the manner in which our trainer should account for his cash-flow-positive contract? Perhaps not under traditional accounting, but yes if we apply fair value accounting. The contract now embodies an opportunity cost of $1,000 annually, and therefore its fair value from the perspective of the trainer is negative $10,000 (assuming interest rates are zero for all maturities).
To understand fully the basis for this assertion (all while eschewing the nebulous concept of opportunity cost), we assume the trainer becomes paralyzed and unable to deliver the courses in question. Eager to protect his reputation with his clients he searches for a substitute trainer, of comparable competence, to step into his shoes. This second trainer will demand payment of $2,000 per course while the original trainer will receive no more than $1,000 per course from his client. This translates into an annual loss for the original trainer of $1,000 per course and an aggregate loss, over 10 years, of $10,000, whose present value is also (negative) $10,000 since interest rates are zero. Q.E.D.
It follows that from the moment fair value accounting starts to apply, the trainer must immediately recognize a liability of $10,000 on his balance sheet and a loss of $10,000 in his income statement, even if not a single penny in cash has entered or left his pocket just yet. He now bitterly regrets the chivalrous bouquet of red roses and the juicy steak dinner (or, more likely, the kale salad with fresh balsamic dressing). Assuming the trainer has no other assets, he is bankrupt, despite every expectation that he will generate positive cash flows throughout each year of the contract’s life. All very strange.
Now assume that fees for competent financial trainers stabilize at $2,000 per course. During Year 1 our trainer delivers his first course and generates $1,000 of cash flow. This works its way into the trainer’s income statement for Year 1. At the same time however the $10,000 liability has now shrunk to $9,000 (assuming interest rates are still zero). This reduction the liability also works its way into the income statement, producing a fair value gain of an additional $1,000. So for Year 1 the income statement will reflect $2,000 in revenue/profit, even if from a cash flow perspective only $1,000 arrives in that year. A bit strange as well.
The same will recur in every year thereafter, for as long as fees for competent trainers are stable at $2,000 per course. This outcome is at least as strange as the outcome in the preceding paragraphs. Fortunately a minimum amount of sanity comes to prevail by the end of Year 10, at last: the initial loss of $10,000 has been followed by ten annual profits of $2,000 each, producing an aggregate profit, over the life of the contract, of (positive) $10,000 – exactly the amount of cash flow generated in aggregate. While cash flow and net income diverged meaningfully year after year, they converged seamlessly over the 10 years, cumulatively. So perhaps not so strange after all.
Conclusion.
In summary, expanded use of fair value accounting often enhances the transparency of financial reports and reflects more accurately the respective positions of competing stakeholders in liquidation; but it gives rise to a variety of bizarre outcomes that are difficult to reconcile with financial common sense. In other words the price paid arguably overwhelms the benefits achieved.
For us, to be truthful, the jury is still out. Each of the two accounting alternatives carries plusses and minuses. Comparing them with some objectivity is not easy. We will continue to think about these issues and will publish another article if our thoughts produce tangible insights.
*Oussama A. Nasr has worked as a lawyer, banker, trainer and financial consultant in New York City and Beirut for 37 years. He began his career with a 4-year stint at the corporate law firm of Shearman & Sterling in NYC, followed by 7 years at Citigroup, before establishing the consultancy DNA Training & Consulting more than 25 years ago. He holds BA and MA degrees in Mathematics and Philosophy from the University of Cambridge and a Juris Doctor degree from Cornell Law School. His company URL is dnatrainingconsulting.com and he can be reached at onasr@dnatrainingconsulting.com.