It looks like the FASB is in full retreat on fair value accounting. Below is an artifact from the period we published today in The IRA. Author is a very well respected analyst who worked for the Fed in the 2007 period and tried to sound a warning about the supervisory implications of a FVA regime. I am going to do a victory lap at AEI today when we convene the latest “Deflating Bubble” session. — Chris
March 17, 2009
The following article on fair value accounting (“FVA”) was authored in May 2007 by a researcher who at the time worked for the Federal Reserve System. The paper, which was not approved for publication by the Fed Board’s bank supervisory staff, outlines some of the issues in a transition to FVA, issues that have turned out to be critical. Many of these issues now may be obvious to students of financial institutions and the general public thanks to the financial crisis, yet two years ago the paper was dismissed by the Fed’s staff in Washington. To us, the story around this article provides yet another example of how the intellectual closed-mindedness of the Fed’s Washington staff results in bad public policy.
Some background for context. The Fed and other bank regulators historically did not push back on supervised firm accounting “choices” or otherwise second guess the external auditor on valuation issues for financial firms. That is, if a financial firm could get its paid accountant to sign off on a choice of valuation methodology — choices which in many cases are based purely on “stated intent” at the time to hold an asset for sale or to maturity, then the paid regulator at the OCC (figure roughly 50 examiners for each too-big-to fail bank) and its more poorly staffed step-brother, the Fed (roughly 7 examiners per TBTF bank), would simply accept the decision without question or review.
During the bubble years, the author and other members of the federal bank supervisory community fought internally and with the banks against apparent inconsistencies in accounting choices — for example, the same large bank would put a big chunk of liquid exchange traded equites that turned over frequently into AFS, while putting illiquid slow-to-turn distressed debt in trading. Most regulatory accountants, though knowledgeable and well meaning, were in some sense lazy, as they felt it much more important to maintain GAAP and regulatory accounting parity, then it was to have more correct reporting based on actual facts and behavior.
Though it is possible to understand this desire to avoid complications of mismatches — GAAP-RAP mismatches would have raised questions that could illuminate some interesting facts to the public — fair value or no, the increasing use of mark-to-market in the rising tide to get quick earnings was clear. See, for example, how much Deutsche Bank put in trading as a percentage of total assets (near 60%). Accounting or regulatory reporting examinations by regualtors essentially did not exist unless an obvious tying error or something of that sort persisted over time.
So to summarize the backdrop to the article:
1. The accounting rules were based on “stated intent” to hold or trade an asset. To optimize earnings in a rising tide, banking firms would put as much as possible in either trading (immediate earnings), or AFS (where they could sell the winners when EPS was needed and keep the dogs).
2. The regulators did not push back against suspect accounting and regulatory reporting classifications, especially for illiquid assets, as they rightly felt they were on thin ice politically and didn’t want to fight with public accounting firms.
3. Everyone now knows about structured finance and the shadow banking system, how big it was, and short-term based it was. A lot of structured finance vehicles including hedge funds are essentially agnostic as to whether their positions were/are levered longs or levered shorts. The interplay between FVA and RAP creates huge opportunities for regulatory arbitrage.
Given the above state of play, the author tried to describe, in non-expert terms, the FVA issue in such a way that people could understand what might happen, as we (as an entire economy) moved increasingly to a short-term, market price-based financial world. The purpose was and is not to argue against FVA, but rather to pose some questions and list some potential consequences with respect to the banking industry and the regulatory environment.
As the author wrote two years ago. “These consequences may amount to mild and transitional growing pains resulting from the accounting shift. On the other hand, certain unintended consequences may be more significant and enduring.”
Fair Value Increases Financial Statement Volatility; Removes a Solvency Buffer
Accrual accounting provides a balance sheet solvency buffer; Fair Value accounting takes the buffer away and can lead to increased volatility in earnings and capital.
Under traditional US GAAP accounting two major solvency buffers can be identified with respect to banking organizations during financial stress. The first is obviously the level of capital itself. During stressful markets it’s good to have a substantial amount of capital. It’s even better if capital is composed of low-risk, unencumbered assets easily convertible to cash to avoid illiquidity (which tends to be a faster and more prevalent killer of banks). Though banking institutions are traditionally highly levered, they tend to maintain sufficient amounts of capital to cover losses in difficult times.
A second solvency buffer has been the accrual method of accounting, which can dampen earnings and capital volatility under stressful market conditions. Most should be able to agree that with accrual accounting, many changes in fair value are effectively smoothed through accrual income streams over time. For example, loans booked as held-to- maturity and funded by retail deposits aren’t revalued through earnings at each reporting date. So unless a traditional banking firm was somehow able to maintain a perfect fair value hedge of loans against deposits , accrual accounting will result in less volatility of earnings and capital relative to a more immediate fair value approach.
Many argue that accrual accounting is a poor reflection of economic value at any point in time. But given that market risk factors fluctuate over time, just as today’s value can decline given a supply/demand-based risk factor move, tomorrow’s value can increase. In a sense, the true or realized value over the life or holding period of an instrument is never clear until after-the-fact. As we know, some, if not many risk factor moves, are not enduring and are attributable to imperfect markets (e.g. chunky trades, manipulation, or collusion) or could be characterized as coincidental or spurious. For example, if two big holders of a particular bond issue sell their holdings at the same time for unrelated reasons that have nothing to do with the credit risk of the issue, the “market”, with its imperfect information, may draw an erroneous conclusion from the sales volume that the bond issuer’s prospects are deteriorating. This could result in follow-on sales causing the fair value of the bonds to plummet, unless and until it becomes clear that the initial sell orders had nothing to do with actual repayment prospects of the subject bonds. It is essential to remember that there is neither perfect information nor perfect liquidity even in our best financial markets.
Another important point is that more than ever before, markets can be moved by those who “rent” assets (e.g. hedge funds) as opposed to those who buy or sell with more permanent intentions. On one hand, this rental society paradigm provides greater day-to-day price transparency and liquidity, at least under normal circumstances. But on the other hand, those who hold or control sizeable pools of assets (sometimes through the use of large amounts leverage) in efforts to derive short-term gains can contribute to extreme swings in asset prices, even within separate but correlated markets. What happens when the “renters” control a majority of the trading volume in a particular market, even for a relatively short period? Isn’t that the tail wagging the dog in terms of valuation?
Perhaps calculating a risk-adjusted discounted cashflows over the life of an asset is obsolete, and it’s the ability to sustain a short-term arbitrage that should result in fair value for financial accounting. An important question to ask is do we really understand the effects of putting highly-levered long-term-thinking banking firms (not to mention insurance companies, pension funds, etc.) into a paradigm where they are subject to the whims, fears, and possible manipulations of “renters”? Doesn’t that largely force the long term holders to act like “renters” too? It seems fairly clear that the volatility of asset prices and balance sheets could increase substantially in such a world.
Some note that FVA will encourage banks to better hedge the economics of their positions, since volatile results can be avoided through appropriate risk management. However, we should remember that hedging entails costs, and it may be quite expensive for the typical bank to hedge against short-term, moderate-size moves that could be driven by short-term market speculators, but nonetheless could wreak havoc on a given period’s earnings. In a full fair value world, to ensure stable period-to-period earnings, banks would likely need to incur substantial hedging expenses. In such a paradigm, we should ask ourselves if there will be enough remaining profit margin and management willingness left to hedge against the separate larger systemic stresses that could put a firm’s solvency in real jeopardy. That is, will a heightened focus on current fair values, given their potential impact on quarterly earnings, result in increased un-hedged tail-risk risk concentrations (i.e. moral hazard)?
Accrual accounting does not focus on quantifying current assumptions about future probabilities and expectations with respect to multiple risk factors; factors which sometimes move coincidentally. In a sense, accrual accounting takes things as they come, with only periodic adjustments for permanent-type changes. Beyond elimination of noise, the benefits of accrual accounting include its simplicity, transparency, and consistency. Everyone knows that book value makes no assumptions based on today’s forward view of risk factors, even if the current “consensus” views can be derived from market sources.
To be fair, under normal times, the noise and volatility created by technical distortions and imperfect information applied to fair valuation may cancel-out for firms with numerous diversified positions. And perhaps, when all parties in the markets operate on a full fair value basis, we will get closer to perfect markets. But are all stakeholders psychologically ready for full fair valuation under a broadly stressed environment? One could argue that under stress, when the estimation of probabilities and expectations becomes extremely unclear, something closer to accrual-based accounting is actually a more transparent and an easier to understand reflection of a position’s value, verses a point-in-time estimate based on today’s best-available information. Let’s not forget that a firm’s best information available during a market stress may come from an individual who is inherently conflicted (e.g. the trader who bought the position in the first place).
Do We Need A Different Valuation Technique or Just Better Information About Risk Positions?
To move to FVA really amounts to a change in accounting convention, which should be largely unimportant to professional analysts and sophisticated stakeholders. What the sophisticated really want is more objective information on positions and exposures.
Those who want to the estimate fair value of a bank’s accrual assets at any reporting date can already make a reasonably sound estimate for most positions, and have been able to do so for some time. An average yield is reported on various position types. Analysts can compare those yields with current market rates and forward curves. Given this information, re-valuations can be performed in a way roughly analogous to how FASB would compute a “fair value” today. This general technique can be used for many assets and liabilities. The more granular and transparent the reported information on positions and risk factors within the positions, the “better” estimate of fair value can be made by an external analyst. So maybe we should ask ourselves, “Do analysts need a new accounting convention, or do they really want to see more detailed objective position and risk factor information regarding a company’s holdings?” With more information financial statement analysts can arrive at their own summary statistic of value and perform sensitivity analysis around it. A firm’s own estimate of the current fair value of its positions is certainly convenient, but given the valuation of the complex and illiquid items is often more art than science, isn’t the underlying position information what analysts really need to understand the past, current, and potential future circumstances of the firm?
To reiterate, the exercise of computing a “fair value” at any point in time is really just a convention. Just as accrual accounting is a convention. One can argue that accountants should change the convention so that there is more “direct” and possibly better estimates of current value. But in a way, changing accounting valuation convention is analogous to transforming the current position data, just as a statistician might transform data in different ways to get to a desired fit. If financial statement reviewers don’t have the details of what accounting transformation is taking place, will they really have more information? Assuming the same position information is disclosed, there may not be compelling reasons why analysts would argue for a change in accounting convention.
Bank Supervision, Liquidity, and Behavioral Considerations
If less informed stakeholders perceive increased volatility and risk given accounting changes, their subsequent behaviors may exacerbate volatility and actual risk within banking firms. Supervisors must be prepared for this paradigm in the individual firm context, as well as in broader stressed environments.
In terms of supervision, a strong argument can be made that the traditional bank accounting method, with its “accrual buffer”, should result in less-frequent intervention by bank regulators relative to a “fair value” convention. Why? Obviously, if there is increased volatility, fair value exposes banks to more downside risk in earnings and capital. With more downward spikes in earnings and capital, banking firms are not just more likely to become completely insolvent (e.g. 0 financial capital), but are more likely to hit Prompt Corrective Action capital (PCA) levels which may exacerbate related liquidity issues. Furthermore initial downward spikes may lead various bank stakeholders to reduce exposures leading to a downward spiral.
The first-order issue of reporting more volatile results may not alone lead to significantly more bank closures. However the follow-on effects should not be underestimated. An important and unanswered question is, “Will changing the accounting convention change behavior or trigger points of depositors and other stakeholders?” If it does, under fair value accounting, illiquidity could become a more frequent killer of banks.
As an example, assume a person is told how cold it is outside in numerical terms, but isn’t told the scale from which the quote is derived, or doesn’t have experience with the scale. If the person is told it is 32 degrees, they may think, “at least the number is positive”. But if they are told it is 0 degrees, without quoting a scale, will they think it’s colder and behave differently? Of course, well-informed individuals will realize there is no difference if one quote is in Fahrenheit and one is Celsius. However, just as analyst or buy-out group can understand that an insolvent company (i.e. zero financial capital) may have real long-term economic value based on potential future value of cashflows, the uninformed person may be more hesitant to place his funds in such a company. So if XYZ bank has $32 in accounting capital or $0, will it make a difference to the non-analyst depositor or investor? If changing the convention will impact behavior of key, but less informed constituents, it seems we must be wary of unintended consequences.
When earnings and reported asset values at a bank decline sharply, one could reasonably imagine a subsequent outflow of deposits. Even prior to withdrawals, in terms of option value, deposit fair values decline as depositors get closer to exercising their withdrawal option. In a sense, banking firms are naturally engaged a structural form of “wrong way risk”. With declining asset values, fearful depositors may pull deposits, causing declines in the value of liabilities simultaneously. The whole scenario might sound similar to a hedge fund that has had a bad quarter where investors suddenly want their money back. Though that is an oversimplification, under “fair value” accounting, financial statements may not only more quickly indicate that a banking firm is having problems on the asset side but the convention may exacerbate value losses on the liability side. In the trader’s mind this is like being short Gamma – when an underlying value starts moving against you, each incremental move downward results in an increasing level of loss.
Readers may be thinking, “Banks can’t be compared with hedge funds. Since traditional banks have deposit insurance, depositors won’t withdrawal their money just because of a bad quarter’s earnings”. It should be uncontested that there is value in the “stickiness” of deposits (i.e. “elasticity” or “value of the depositor relationship”) under normal conditions. However, when a firm is under stress, do we believe that FDIC insurance is a perfect backstop to stave deposit outflow under stress? Clearly some will exercise their withdrawal option rather than sit tight and wait for the FDIC. But before we even consider insured retail deposits, let’s not forget about the large uninsured depositors, and lenders. This “smart money” (or “hot money”, if you prefer) will likely have no good reason to undertake any extra real or perceived risk by maintaining funds in an institution that reports poor results.
If problems last over more than a single reporting period and/or if problems are reported in the press, history shows even the smaller insured deposits may evaporate, as Southeast Bank of Miami learned in the early 1990’s. Southeast, the second largest banking company in Florida at the time was closed by the FDIC primarily due to illiquidity. Press reports of the firm’s real estate problems and poor earnings performance led to large and rapid withdrawals by small depositors, accelerating illiquidity. The Federal Reserve provided a discount window loan temporarily, but when that loan was called, the firm had inadequate liquid funds to continue, and the institution was taken over by the FDIC. Incidentally, the FDIC ultimately made a profit of $200 million on the disposition of assets it took over from Southeast.
It is not unreasonable to conclude that fair value’s more rapid quantification of problems, potential problems, noise, and market stress within bank financial statements can lead to stakeholder behavior that exacerbates real problems or perceived problems. As a result regulators like the Federal Reserve (as lender of last resort through Discount Window operations) and the FDIC may need to make more explicit “live or die” decisions, and make those decisions rapidly. Is this a bad thing? Not necessarily. But regulators should certainly be prepared for this potential new paradigm which entails its own hazards. Maintaining a continual understanding of specific markets, ongoing analysis of firm portfolios, and access to key communication channels within firms, will likely become even more important aspects of supervision under a fair value world than today. In the case of broader market stresses in a fair value world where many firms are effected, there is the possibility that the Federal Reserve or others will need to be prepared to call for bank holidays, market closures, or other forms of intervention to allow time for uncertainty to pass.
Don’t get me wrong, I’m not arguing that regulators, analysts, or anyone else should believe in accrual accounting as a true sense of value, or shouldn’t already be assessing the viability of firms under stress. Thinking long-term, migration to full fair value accounting may increase discipline, transparency, and comparability of financial reporting. However, inevitably some will not be prepared to manage at fair value, as the fair value option is ultimately replaced by a fair value requirement. And let’s not forget to consider potential stakeholder behaviors as traditional accounting standards are replaced by methods that may lead to less stable results.
Questions? Comments? firstname.lastname@example.org