Mark-to-market's real role in the crisis: how accounting standards helped build the 'super bubble'.
Fair value accounting, as implemented, amplifies business cycles and seems to significantly contribute to bubbles and busts. Economists, regulators, and finance professionals overwhelmingly missed the unprecedented crisis because all were looking through their own perspective as if the concept of equity had not changed. Our metrics are based on the old concept of financial institution accounting equity, but mark-to-market changed equity in drastic ways. The unrecognized common denominator in the collective failure may have been the mark-to-market accounting data used by all market participants, increasingly full of unrealized profits and unstable gains during the boom. The theory is that risk and leverage became camouflaged, and the quality of equity compromised
amid unstable gains increasingly invading financial institution "equity" without adequate disclosure. Premature gains altered and disguised all financial and economic metric inputs and outputs used by regulators, economists, ratings agencies, investors, financial analysts, and the banks' themselves, at both the micro and macro levels.
The unrecognized distortions in risk, leverage, and equity from unrealized price changes perversely impacted investment decisions by insiders and outsiders, and synthetically affected supply and demand, risk perception and pricing, and asset price in feedback loops that forced balance sheets to expand at an unsustainable pace. This circularity is a result of a dynamic environment where all our financial metrics have become functions of mark-to-market accounting numbers. A mark-to-market accounting regime requiring the recognition of trillions in hypothetical gains falls victim to endogenous price changes much less potent in a non-mark-to-market environment through at least three channels: the Perceived Risk Effect, the Perceived Wealth Effect, and the Perceived Leverage Effect. The cyclical nature of asset prices and markets may be inevitable, but the potency is not. Do mark-to-market standards, as implemented, help amplify price changes that lead to "super bubbles" and severe crises that might be much less likely in a tempered-mark-to-market environment?
How We Got Here
For centuries, asset prices have deviated uncomfortably far from intrinsic value during euphoric periods and the subsequent downturns. Asset bubbles are inevitable and only concluded in hindsight, but they have proved economically dangerous. Intrinsic "value" often differs from fair value based on market price, but it is difficult to discern at what point the disparity occurs. As the American Enterprise Institute's (AEI) Alex Pollock says, "Price is what you pay for something. Value is what you get." The majority of the recent debate about mark-to-market's role in the crisis is either about the potential for manipulation because of the subjective nature of the asset valuation process or about contagion when the bubble burst. Virtually absent from the public debate is the necessary discussion about the increasing level of premature and unrealized gains over the past decade--especially from derivatives and derecognition--that could have potentially disguised risk and leverage at financial institutions enough to be a major factor in building the bubble. Disclosure about earnings quality, as well as equity quality, seems to have been compromised as unstable gains infiltrated our accounting concept of equity.
We must not fall victim to what Haresh Sapra of the University of Chicago calls "casual intuition" about mark-to-market's superiority. CPAs required the comingling of traditional real equity and profits with uncertain gains without segregation or adequate disclosure. Under Statement of Financial Accounting Standards (SFAS) 115, Accounting for Certain Investments in Debt and Equity Securities, assets are categorized according to function: trading, available for sale (AFS), and held to maturity (HTM). In the concurrent three-level valuation hierarchy under SFAS 157, Fair Value Measurements, increased disclosure about unrealized Level 3 value changes is now required. Unfortunately, it is the Level 2 asset class that lacks sufficient disclosure and houses the majority of the subjectively valued derivatives that far outpace Level 3. In the asset type categorization, available-for-sale gains must be segregated in equity, but such gains pale in comparison to the mysterious trading gains.
In an April 2010 CPA Journal article on derecognition gains, "The Subprime Lending Crisis and Reliable Reporting," Benjamin P. Foster and Trimbak Shastri highlight a study demonstrating how "the move to fair value accounting requires inclusion of more hypothetical transactions in the financial statements, which allowed subprime lenders to recognize income well before it was actually earned or received." This phenomenon is not limited to sub-prime lenders and toxic loans; it applies to all banks and all financial instruments, including derivatives. Misaligned compensation incentives are often mentioned as a cause of the recent crisis, but the financial institutions distributed what GAAP defined as profit and equity--so was a significant part of the problem created by the GAAP standards forcing all unstable trading gains to equity? That financial sector profit encompassed 40% of corporate profits prior to the collapse could be telling.
Does Mark-to-Market Increase Transparency?
A main goal of mark-to-market is increased transparency. The first inherent assumption about mark-to-market s superiority is that markets are generally efficient, but the efficient market theory has failed the latest test. The second assumption is that the remarking of asset prices is only a reflection of perceived value, but mark-to-market has real economic effects on asset pricing efficiency and financial stability. Mark-to-market significantly impacts investment decisions by banks, investors, and regulators as balance sheets and expected supply-and-demand dynamics are altered based on point-in-time valuation metrics. Marking to market every quarter actually leads to subsequent changes in asset prices, exacerbating and amplifying business cycles through behavioral phenomena significantly less potently than under a non-mark-to-market regime (Guillaume Plantin, Haresh Sapra, and Hyun Song Shin, "Marking to Market, Liquidity, and Financial Stability" working paper, 2005, www.imes.boj.or.jp/english/publication/mes/2005/me23-s1-7.pdf). The third assumption is that mark-to-market has increased disclosure about financial position, risk, and leverage. However, incorporating unrealized derivatives gains and other premature gains undistinguishable from real gains over the last decade appears to have increasingly distorted financial position, risk, and leverage as well as increased the need for off-balance sheet vehicles and unhealthy increases in derivatives to hedge the increased risks--all of which created more opacity than transparency.
Two of the most cited causes of the crisis--risk and leverage--were calculated and assessed using balance sheets increasingly subject to fair values. Therefore, it is ironic to realize that risk and leverage would have been assessed higher during the bubble period if we had used balance sheets less compromised by unstable, subjective gains. If early paper gains did not make leverage and risk appear to drop, then improper decisions and assessments by all parties could have been much reduced during the bubble. Banks could not have expanded with such disguised, but dangerous, levels of risk and leverage.
Mark-to-market's role in the crisis begins well before SFAS 157, with SFAS 133, Accounting for Derivative Instruments and Hedging Activities, and SFAS 140, Accounting for Transfers and Servicing of Financial Assets and Extinguishments of Liabilities, when unstable gains increasingly entered the balance sheet and distorted calculated leverage. Level 3 disclosures were increased because that category houses many of the subjectively priced toxic assets, but we neglected to envision the need for adequate disclosure about subjectively priced derecognition gains and derivative "trading" gains in Level 2. Because the value of Level 2 derivative gains is now measured in the trillions, knowing the amount of Level 2 fair value gains flowing to equity/income should also be paramount.
Derivative Gains Alone Dwarf Equity
Over the past few years, the word "procyclicality" (the amplification of business cycles) has entered our vocabulary, but minimal study exists on mark-to-market's effects on prices and the business cycle. Studying mark-to-market economic effects ex ante has proved problematic--especially because investment decisions would have been different along the way and disclosure of calculated gains has been lacking--but its perniciousness may be severe. Duncan Wood explains in a February 2010 Risk magazine article, "Out of Their Hands," how equity at the largest banks around the world is now dwarfed by paper derivatives gains.
Wood notes that in an exercise in early 2010, the 2008 balance sheets of the largest banks around the world were dissected and analyzed by a team attempting to reconcile banking regulators' and accounting standards setters' imperatives. The position that their imperatives compete is debatable, because capital adequacy is certainly akin to going concern. If GAAP numbers with minor adjustments do not foretell capital adequacy, then how can the numbers expose going-concern issues? After digging in the major banks' financial notes, Wood determined that the positive derivatives value (the in-the-money unrealized gain portion) was two to six times net equity--and this was after the major capita! infusions. The notional derivatives value is much larger, but the portion that the banks include as assets constitutes value changes run through profit as unrealized gains over the years.
Banking regulators were considering the exclusion of Level 2 and 3 paper gains from minimum capital requirement metrics, presumably because paper gains can evaporate, but that would have shown major banks around the world to be severely undercapitalized. Therefore, the under-capitalization and equity-lacking "quality" that became apparent in 2008 may partly stem from FASB and the International Accounting Standards Board (IASB) forcing all paper trading gains to equity, rather than banks making poor decisions of their own accord. Regardless, surely some value exists in these derivative instruments (and consideration should be given to the derivative liabilities in the multisided hedging equation), but must we include all point-in-time, subjective, and sometimes temporary price moves with the same consideration in equity, as if the full value changes are intrinsically real and stable? Unlike derivative positions, realized value changes cannot disappear or instantly turn into liabilities. In addition, valuing derivatives to effectively capture potential liabilities is inherently difficult, because contracts can fall in value very quickly when highly unexpected market circumstances materialize.
Looking at Goldman Sachs's 2009 annual report, derivative contracts after bilateral netting but before netting for cash collateral totaled $200 billion, compared to equity of only $71 billion. After reducing for cash collateral received, the positive derivatives value was $75 billion, still higher than equity. However, netting cash collateral can be deceptive if the cash is used to finance operations, because the risk exposure is not necessarily reduced if the cash collateral is not segregated. Collateral netting changes the visible net paper gain/loss because counterparties post anywhere from 0% to 225% in collateral. Netting for cash collateral distorts the true net asset/liability position, not only because it can be used, but because of "collateral arbitrage" opportunities due to varying collateral percentages. Banks can collect collateral at a much higher percentage than is paid out on liabilities (e.g., Goldman Sachs's 2009 cumulative net cash collateral balance was a favorable $110 billion by collecting 57% on its positive derivative assets but only paying 16% on its liabilities [Michael J. Moore and Christine Harper, "Goldman Sachs Demands Collateral It Won't Dish Out," Bloomberg, March 15, 2010, www.bloomberg.com/apps/news?pid=newsarchive&sid-af6uIAFTSorY&pos-10]).
Did we make the assumption that tallying and segregating trading gains (as was done with available--for-sale securities) was unnecessary because banks would run of out of cash before paper gains could take over equity? If so, did we forget how banks can subjectively re-price instruments in opaque and oligopolistic markets and use "observable" market proxies derived from thin or irrational markets to demand billions in cash collateral? Did we underestimate the amount of premature and overpriced derecognized assets "sold" to essentially related-party special-purpose entities financed on borrowed funds with ignored costly put options with contingent commitments?
The portion of cash collateral that is collected can be used, but what about the gain portion that is moved to equity and not collateralized by cash? Stock compensation does not require cash, hence the massive stock-based compensation bonuses we continuously witness. Distributing such unrealized gains in a non-cash manner is beneficial because it ultimately decreases the amount of remaining accounting equity requiring a return and raises the calculated return on equity (ROE). If this unstable new equity is not distributed, then the banks must lever up even more to earn an acceptable ROE. Unfortunately, we witnessed both premature distributions and unhealthy leverage levels regardless. It is no wonder that banks cannot stop delivering such massive stock option bonuses.
The banks also use non-cash collateral (e.g., securities pledged) on mark-to-market gains as collateral for themselves by rehypothecating (collateralizing the same collateral by multiple parties multiple times). Collateral assets pledged by counterparties can be re-pledged on new debt to finance operations and more assets. Gillian Tett of the Financial Times highlighted a recent International Monetary Fund (IMF) paper that estimated, "The seven largest US brokers were getting about $4.5 trillion of funding from rehypothecation activity" by 2007 ("Web of Shadow Banking Must be Unravelled," August 2010, www.ft.com/cms/s/0/112ff210-a62b-lldf-9cb9-00144 feabdc0.html).
Noted Crisis 'Causes'
While historical cost accounting (HCA) could have helped prevent the bubble from expanding as far as it did, it has other issues that make it suboptimal. Therefore, just ignoring fair value gains/losses is not advocated, but analyzing HCA versus mark-to-market is necessary to understand the unintended consequences and economic impacts of mark-to-market as implemented. The "causes" of the financial crisis, such as failed regulation, misaligned incentives, greed, risk under pricing, and high leverage, could have been severely affected and camouflaged by early recognition of a material amount of premature gains under mark-to-market accounting. Is it a coincidence the latest super bubble coincided with the implementation of SFAS 133 and SFAS 140? Is it a coincidence that mark-to-market's use was on the rise during loose accounting standards prevalent in the 1920s, another speculative super bubble that led to the Great Depression? No super bubble occurred in the six decades between, when mark-to-market was predominately not in effect. China did not experience a deflating property bubble, but, interestingly, it did not implement mark-to-market until 2007. Now, almost four years later, China is believed to be experiencing a major property bubble. One can suggest yield-hungry investors injected capital into China, causing the increases, but capital also follows a rapidly rising economy subject to mark-to-market accounting. Maybe it was mark-to-market premature gains that made investing in China seem more attractive in the first place.
To bank regulators, net equity equals common equity (e.g., generally defined as common stock and retained earnings), minus goodwill, with some other adjustments. There is no distinction between most realized versus unrealized gains transferred to equity. Some argue that paper gains segregated in OCI are prohibited from certain capital metrics, but the segregated AFS category only applies to non-trading instruments and is miniscule compared to derivatives in the trading category. Therefore, the majority of the paper derivatives gains are essentially "capital" in minimum capital requirements. Some consideration, although inadequate, is given to the risk of such gains in other risk-adjusted metrics like risk-weighted capital, but these paper gains are not excluded from capital or distribution per se. For example, Goldman Sachs's 2009 "equity" was $70.7 billion, whereas fair value gains in the AFS category only totaled $97 million. Paper derivative gains at $200 billion, or 2,000-plus times the AFS category, are almost triple equity. For decades, bank net equity generally meant capital and real earnings without vast adjustments for paper gains. After SFAS 133 and SFAS 140, we now taint our concept of equity by comingling real profits with huge amounts of subjectively calculated gains as of today, but not necessarily tomorrow. Paper gains essentially become paper equity. For financial institutions, the quality of equity is especially paramount, because equity is the foundation upon which leverage is built, and it must be strong, not unstable.
Banking regulation could be modified to restrict distributions of paper equity because the gains might be unreliable, overvalued, disappear, or morph into liabilities, but accounting does not segregate such gains in equity, and it is treated as equivalent to cash profits or paid-in capital. Level 2 and 3 derivative values are subject to information asymmetry and are often priced using proprietary methods rather than subject to open market scrutiny. In addition, if the pricing proxies used in calculations such as the "observable" ABX index stray from reality or the chosen proxy seems reasonable but is questionably used, then calculated gains can become far from reality as well. A recent study by Fitch suggests that derivative prices can be poor indicators of bond defaults, not least because such markets are often thin. Regardless, aggressive margin calls become difficult to fend off. A need for disclosure regarding potential losses became apparent, hence SFAS 133. Yet does the needed disclosure about potential losses necessitate the need for trillions in hypothetical gains to be unsegregated in the body of the financial statements? Some values are likely still inflated and others could evaporate with counterparty risk or after unexpected market changes. Worse, these unstable gains housed in equity facilitated the house of cards and dangerous leverage. The more illusory gains are forced to equity, the more financial institutions experience the necessity--and have justification--to increase their balance sheet and inherent leverage, as demonstrated later in this article.
What's Different After SFAS 133 and SFAS 140?
The forced comingling of paper gains with real equity creates a false sense of security and the ripe opportunity for broad suboptimal decision making across the board. Financial institutions distribute equity as dividends and bonuses, and have always done so. What is different after SFAS 133 and SFAS 140 is that equity is different. Prior mark-to-market standards changed equity but not to the scale of the newer standards. We are tempted to chastise the banks for taking too much risk, but mark-to-market may have severely disguised the level of risk the banks were taking. The rising level of unstable gains increased the fungible pool of equity, so it seems that early gains filtrating into equity allowed the premature distribution of traditional real equity. Slowly, the real equity was distributed as it was replaced by unstable equity. The scenario is akin to a teenager indulging in his parents' vodka bottle and replacing the spirits with water, diluting it further and further; the bottle looks to be the same, but after careful observation and testing we can determine the liquid is quite different. GAAP and IFRS said these gains were real enough to flow to equity, so can we fully blame the banks for their naivete? Many mark-to-market opponents claim using fair values is problematic because it brings increased artificial volatility. However true this may be, the potential bigger issue is that real stable equity unknowingly became replaced with unstable equity.
Will New Capital Requirements Be Sufficient?
Enhanced capital requirements can certainly help prohibit future dire scenarios, but will these be enough? FASB leaders suggest that because banking regulators have concurrent imperatives that overlap but are not necessarily the same, banking regulators need to adjust GAAP numbers for their purposes. Besides the shared going-concern issue, there are two other problems with this outlook.
First, it is not just the regulators who sometimes need to manipulate GAAP numbers, for inevitable irrational periods create a need for all financial statement users to adjust. Markets will always have periods of inefficiency with unstable or abrupt changes in value, so we cannot assume the same naive market will know when and how to adjust their metrics accordingly. It was the poor investment decisions by bankers, regulators, and investors, whose reliance on increasingly irrational fair value mark-lo-market numbers contributed to their decisions. Because banks are extremely sensitive to metrics like ROE based on GAAP numbers, they do not make investment decisions based on bank regulatory measures and internal models alone. In fact, the intense market pressure on quarterly accounting numbers can be as potent as regulatory pressure, if not more so. Bond ratings agencies use accounting numbers, and banks fear declines in ratings because the drops have severe consequences on margin calls and cost of capital.
Second, this outlook ignores the potential of the economic real impacts of our rules. CPAs often like to think of ourselves as impartial referees, and "don't shoot the messenger" is a common refrain. Although it is true CPAs are referees of sorts, we are also the line judges, and our rules can sometimes move the chains in our evolving fair value regime. Referees usually move the chains only after the players make a legitimate first down, but they sometimes move the chains when they shouldn't, by making bad calls. Did we make a bad call by forcing the premature gains into equity that led to further increases in price and facilitated early distributions?
Specifically, how do our accounting rules affect market prices and have real economic effects? When we force market values on the financial statements--especially during increasingly irrationally exuberant or irrationally fearful time frames--in the name of relevance, reliability, and transparency, we are blessing the irrational market and making it even more irrational. Our requirement that unrealized trading positions must flow to equity will always lead to distortions of true financial position and perverse decision making as a bubble is forming, but most investors won't recognize it. A distorting bubble reallocates economic resources across sectors in an ineffective manner, but one that seems effective at the time. Mark-to-market, with its comingling of realized and unrealized trading gains in equity/profit, obscures the situation and actually contributes to asset price changes--rather than the conventional wisdom that mark-to-market standards only reflect changes in asset prices. The theory is mark-to-market fair value changes can affect prices through at least three channels: the Perceived Risk Effect, the Perceived Wealth Effect, and the Perceived Leverage Effect.
The Perceived Risk Effect
The perceived risk level affects prices as the perception of risk changes, which leads to further price changes. As described by Plantin, Sapra, and Shin (2005), as unrealized gains increase, risk perception declines. Investments appear safer, so risk is assessed lower and the market lowers the required risk premium because of the positive effect on loan-to-value ratios and default probabilities. Since financial markets use discounted cash flow metrics for pricing and because risk premiums and asset prices are inversely related, asset prices for derivatives and securitized assets rise accordingly when investments appear to have become less risky (or fall when investments appear more risky). A circular feedback loop is then set in motion. The authors explain this "synchronization" of asset price moves under mark-to-market and the subsequent effects on investment decisions.
The Perceived Wealth Effect
Perceived illusory wealth adversely affects investment decision making. The perceived wealth effect is different from the traditional economic wealth effect. The wealth effect in economics attempts to measure wealth impact on consumption and assumes all wealth is inherently real, but the perceived wealth effect (especially active in a mark-to-market world) could be the behavioral phenomenon that perceived wealth has on investment decision making. Abundant, cheap credit from the Federal Reserve may have fueled the crisis, but mark-to-market gains fueled the abundant, cheap credit decisions made by banks that lent to thousands without verifiable income.
Perceived wealth leads individuals to be more confident in risk and leverage taking. Just as homeowners felt wealthier and used their homes as ATMs, banks felt wealthier with healthier looking accounting ratios and took on more risk and loaned out more and more money in their quest for yield, amplified under mark-to-market rules. Banks can essentially create leverage out of thin air with premature derecognized gains and other paper gains that make the banks feel more confident to create the leverage and take on the increased risk. For example, $1 trillion in paper gains across the world can be used to expand to $10 trillion in new assets with a leverage of 10, or $30 trillion with a leverage of 30--with hypothecating collateral on that $1 trillion could lead to an even further expansion. Banks' trading books and accounting books became full of unstable equity showing that perverse decisions could (and should) be made in order to earn an acceptable ROE--if only the bank's trading book showed such paper equity, as was the case for decades, maybe decision making would not have occurred on the same scale in the boom years, maybe risk would have been better priced, and maybe the super bubble might not have been able to inflate (hence, mark-to-market might explain at least part of the $30 trillion-plus in new wealth created from 2004 to early 2008 that mostly evaporated by early 2009).
Finance metrics using GAAP numbers worked relatively well for decades in a non-mark-to-market world but became less effective with an increasing amount of unrealized gains. A portion of the increase in wealth that evaporated was only perceived paper wealth from demand created under false pretenses that helped drive asset prices well beyond their intrinsic value. Cheap money and low interests rates are also often cited as a large factor in the crisis, but regulators' biggest mistake might have been the failure to recognize the liquidity injection from perceived increases in wealth under mark-to-market rules on all the relied-upon metrics that worked relatively well in a non-mark-to-market world. Who needs quantitative easing or money presses when perceived equity gains can be used to justify and increase leverage? Unstable gains pass muster as long as the perceived increase in value can be documented by verifying the value (or verifying an input to the valuation metric), but things are not always as they seem. In hindsight, we know a large portion of the increases in wealth proved artificial. Certainly some recognized the bubble, but wise men cannot override irrational human behavior in gray areas. Rationality cannot win until it is very clear a valuation assumption is wrong, but incorrect assumptions rampant during a bubble are difficult to discern in real time.
The parabolic jump in securitized asset and derivatives gains greatly inflated the perceived wealth of financial institutions. Will forcing more OTC derivatives through clearinghouses be transparent enough to make the market efficient and make fair values reflect intrinsic value? Despite open stock markets, stocks still often deviate from fundamental value. Therefore, why would we expect exchanges to most often reflect rational and efficient prices? Clearly, having more derivative trades conducted through clearinghouses with structured collateral requirements would be a step in the right direction, but a significant portion of OTC derivatives are likely to remain off exchanges and stay inside a "black box" (literally the black filing cabinets in bankers' offices).
How exactly does mark-to-market distort actual leverage and lead to further increases in leverage? Incorporating paper gains eventually leads to distortions by camouflaging leverage and equity quality. Perceived increases in wealth from paper gains was infused into equity and used to increase debt and leverage; it actually forced the need for more and more securitized asset purchases. Unrecognized gains do not affect leverage metrics, but gains flowing to equity, whether realized or unrealized, have the same effect on leverage: premature gains, which 1) make leverage appear to drop with debt unchanged; and 2) create false demand for the creation of new assets and corresponding debt, as holes on bank balance sheets need to be filled. The more unrealized gains a bank experiences, the more assets it needs to add to its balance sheet to earn a return on the unstable equity. (We could even call this the ROE effect.) In contrast, when a bank experiences a realized gain, the balance sheet shrinks in overall size as the asset and any corresponding liability are removed from the balance sheet and net cash is received--whereas paper gains not only do not decrease the balance sheet, they increase it. Our economy was on "escape velocity" because bank balance sheets were not waxing and waning, as happened for decades when mostly only realized gains plump with cash flowed to equity. The business model failed because the growth required under mark-to-market was unsustainable.
The Perceived Leverage Effect
Mark-to-market produced holes in the balance sheet even during the good times. We all saw how unrealized losses caused deep holes on balance sheet equity in 2007 and 2008, but mark-to-market gains also affected the balance sheet during the bubble by creating different holes. Mark-to-market losses caused direct holes in equity, while mark-to-market gains on securitized assets and derivatives caused hills in equity, indirectly causing holes in leverage and ROE metrics. Lacking scientific economic studies on mark-to-market's procyclicality during bubbles, we can at least analyze the effect of mark-to-market gains/losses on leverage and bank decisions with some simple scenarios. According to Tobias Adrian and Hyun Song Shin in their working papers "Liquidity and Leverage" (www.imf.org/external/np/seminars/eng/2008/fincycl/pdf/adsh.pdf, April 2008) and "Liquidity and Financial Contagion" (www.princeton.edu/~hsshin/www/BdPFSR.pdf, February 2008), banks target a certain leverage as they target favorable credit ratings and ROE that is highly sensitive to mark-to-market gains/losses. The resulting phenomenon helps explain how the unstable gains caused leverage holes, why banks took on more debt than they could handle (and why they didn't recognize the danger), how leverage and risk got disguised, and why the banks prematurely distributed profits. In "Fair Value Accounting and Financial Stability" (www.princeton.edu/~hsshin/www/BdPPSRmtm.pdf, October 2008), Plantin, Sapra, and Shin explain how mark-to-market can give a "misleading indicator of financial position" and further state:
Leverage targeting entails upward-sloping demands and downward-sloping supplies ... When the securities price goes up, the upward adjustment of leverage entails purchases of securities that are even larger than that for the case of constant leverage. If, in addition, there is the possibility of feedback, then the adjustment of leverage and price changes will reinforce each other in an amplification of the financial system.
In other words, maintaining leverage in the presence of fair value gains essentially creates fabricated endogenous demand for the creation or purchase of more assets (endogenous as in demand stemming from the mark-to-market process rather than real investment demand). Economics 101 explains how increases in demand increase price, but this demand was artificial. There are various bank leverage metrics, but the following demonstration, taken and expanded from Adrian and Shin, uses the metric of assets/equity and shows how mark-to-market gains essentially created demand and disguised leverage. As the gains made leverage appear to drop, intrinsic leverage actually grows because calculated gains in equity increasingly become unstable and temporary.
If a bank possesses a seemingly harmless leverage ratio of 10, equity would rise 10% for every 1 % increase in asset value (1% x leverage of 10):
Assets 100 Liabilities 90 Equity 10 Assets 101 Liabilities 90 Equity 11
A 1% gain leads to a 10% increase in equity value; leverage appears to drop from 10 to 9.18. To get leverage back to 10, it takes $9 in new debt and $9 in new assets, or $1 in equity distributions (or some combination). A 10% gain leads to a 100% increase or doubling of equity value; leverage appears to drop from 10 to 5.5. To get leverage back to 10, it takes $90 in new debt and assets or $10 in equity distributions (or some combination) so that the balance sheet expansion is not as severe as to double in size.
Was it insatiable investor demand for risky loan assets or rather the banks' insatiable appetite to till mark-to-market leverage holes? With a leverage of 10, every $1 in mark-to-market gains requires new assets and debt nine times the gain (or an equity distribution of 100% of the gain, or some combination) just to maintain constant leverage. Much higher leverage ratios of 30% or more were condoned when financial positions became distorted because of premature gains, but these higher leverage levels are much more problematic and sensitive to fair value changes than the 10% example above. Was it insatiable investor demand or demand created by the mark-to-market process? Maybe it was really insatiable bank appetite for more and more loan assets as paper gains increasingly entered equity.
Maintaining and even increasing leverage is necessary to earn an acceptable ROE because banks cannot let paper gains sit idle in equity not earning a return. Because balance sheets expand rapidly under mark-to-market during consecutive periods of gains and because balance sheets become starved of cash, an unsustainable path to higher and higher leverage results. In the 10% example above, equity doubles, so the new required return doubles as well. With regard to the major investment banks, Adrian and Shin state in "Liquidity and Financial Contagion" (2008): "Leverage is procyclical. Leverage increases when balance sheets expand. ... Although 'procyclical leverage' is not a term that the banks themselves would use in describing how they behave, this is in fact what they are doing." Interestingly, this phenomenon was especially particular to investment banks, and a major difference between investment banks and commercial banks is the level of subjective gains from derivative transactions and prematurely derecognized assets. Can this really be a coincidence? Furthermore, the leverage ratio does not visibly portray the inherent dangerous leverage actually absorbed, revealed only later when the illusory asset values vanish. An explanation as to why commercial banks without high levels of such activities suffered as well during the collapse is that investment bank activities were material enough to drastically affect asset prices for the whole system.
Plantin, Sapra, and Shin explain, "The increased mark-to-market equity that results from a boom in asset prices leads to a feedback effect as they attempt to expand lending in order to keep leverage high enough to sustain an acceptable return on equity." Therefore, banks experience intense pressure to obtain a return on that unrealized "gain' no matter what the consequences. If premature gains were not recognized in equity, investors would not require a return on such gains, and banks would not lever up to dangerous levels.
A large part of the demand for new securitized loan assets and synthetics was essentially fabricated on the supply side rather than being real market demand. The upward-sloping demand curve as described by Plantin, Sapra, and Shin was the result of banks needing to fill holes on their balance sheets rather than investor demand or borrower demand. When there were not enough borrowers, we witnessed the lowering of loan standards to one basic requirement: having a heartbeat. Therefore, mark-to-market indirectly fed into to the need for relaxed lending standards and the need for synthetic instruments, because banks became addicted to new loan assets to fill their mark-to-market leverage and ROE holes that necessitated new asset additions. The affordable housing mandate is only a convenient excuse for the dangerous relaxation of lending standards.
The Leverage Effect on the Way Down
The leverage effect on the way up is problematic but nothing compared to the accelerated economic effect during the inevitable downturn. As asset prices drop and paper equity evaporates, banks' calculated leverage will rise with debt levels unchanged. The leverage was really there, but it was camouflaged by unstable mark-to-market gains. With a leverage ratio at 10, each 1% decline in asset value is multiplied by 10 to get the comparable drop in equity value. Banks are inevitably forced to delever after only a small drop in asset value or face increasing leverage levels. Both realized and unrealized losses increase calculated leverage. Realized losses are worse initially because of the cash required, but unrealized losses are still very problematic, especially when margin and collateral calls are triggered. With debt unchanged, the following occurs:
* A 5% decline in overall asset value leads to a 50% decrease in equity value; leverage appears to increase from 10 to 19.
* A 9% decline in overall asset value leads to a 90% decrease in equity value; leverage appears to increase from 10 to 91.
A 10% decline in overall asset value leads to a 100% decrease in equity value. Equity is wiped out--leverage cannot even be calculated since one cannot divide by zero. No wonder Lehman Brothers (and other banks) increasingly entered into large repo transactions to affect calculated leverage. This phenomenon could explain one of the reasons why repos became part of the necessary "financial innovations." To see how difficult it is for banks to de-lever, consider the 5% decline scenario again:
* A 5% decline in asset value leads to a 50% decrease in equity value; leverage jumps from 10 to 19.
* A sale of 45% of assets and a reduction of the corresponding liability is necessary to bring leverage back to 10.
If overall asset values drop 9%--hardly an unthinkable amount anymore--then leverage appears to leap to 91. It would take a sale of 85% of assets to get back to a leverage of 10, The above scenarios assume the banks can delever by selling their assets at current balance sheet value. Inevitable asset price declines will occur because mark-to-market previously allowed inflated values that were not fully market-or liquidity-tested and because all banks will be attempting to sell at the same time when there are no buyers. Before the decline occurred, effective hedging and responsible risk management required more derivatives that became severely mispriced partly due to mark-to-market distortions, so the casino expands, increasing embedded risk and counterparty risk that unfolded during a crisis. Risk is said to "migrate down to those less apt to handle it." Risk does not disappear, but remains in the system even though our widely accepted models collectively showed it could be distributed off the Earth. How could all the risk models provide overly optimistic results? Risk metrics failed to fully account for tail risk and relied on data from an inadequate historical time frame, but the failure is partly because the models used fair value GAAP numbers in their metrics tainted with only perceived wealth, which drastically distorted their outputs.
While disclosure rules have not required explicit details about the amount of mark-to-market versus real gains for all asset classes, we can speculate that the overall mark-to-market percentage was high enough to cause distortions. For net positive and negative derivative values to outpace traditional "equity" would be impossible under HCA. A bank's temporary paper gains would have never entered the metrics to produce distorted outputs erroneously showing leverage was dropping and reasonable, an additional return on equity was required, and premature equity could be distributed. The same conscious increases in debt (and therefore risk and leverage) would not be required, nor would they appear reasonable and safe. Ironically, if the same debt-level decisions were made under HCA and derivative positions were mostly off-balance sheet, only the denominator would be affected in the leverage ratio, so banks would have shown extremely high levels of leverage and accounting insolvency well before such imprudent decisions could be made. Knowing current market values is important, especially for risk assessment, but if fair value gains are included on the balance sheet, doing so requires adjusted metrics and better disclosures.
If we rearrange the leverage equation of leverage = assets/equity by considering that equity = assets - liabilities, the leverage equation becomes leverage = assets/(assets - liabilities). Here one can see how increasing debt without increasing assets by the unrealized portion (as is the case under historical cost), makes the denominator smaller and leverage higher. Therefore, leverage would be higher in the presence of gains not marked in assets or equity. However, under mark-to-market, assets are increased in the numerator and the denominator, so leverage is calculated lower than under HCA. Isn't it ironic that under the goals of transparency and disclosure we created more opacity?
The CPA's Role at the Center
There are several ways that accounting can significantly contribute to a goal of financial stability, but we first need a comprehensive and fair debate about mark-to-market's procyclicality and its consequences as outlined above. We need a new SEC mark-to-market study on the bubble period that analyzes the potential impact of mark-to-market gains, since current disclosures were not required until 2007 and are still inadequate. Mark-to-market could prove to be one of root causes during the upswing (and therefore the bust) if it contributed to disguising leverage and risk, helped to drive asset prices beyond intrinsic value, and allowed and encouraged premature equity distributions while easy money policies acted as the fertilizer. The interactive economic environment is extremely dynamic, but we sometimes lose perspective as if our main audit role is relatively static with similar but different imperatives than other participants.
The dynamic environment includes businesses, financial institutions, investors, economists, pension funds, hedge funds, state and sovereign wealth funds, regulators, economists, ratings agencies, and policymakers. Not only are CPAs important participants in this environment, but we are also at the center of the data flow. We set the rules in order for financial institutions to provide relevant and reliable information, but our mark-to-market rules affect metrics used in investment and planning decisions, which affect market supply and demand and the reallocation of capital, which affects asset prices. Did mark-to-market gains contribute to the unrecognized extortion of real wealth from the fixed-income investors to "build a bubble with fake assets funded by real debt" (phrase borrowed from Andrew Hodge of the Bureau of Economic Analysis).
Many investors claim that they favor full fair value on the balance sheet because its superiority is intuitive, but this is inadequate, or even fallacious, reasoning. Investors--as well as most accounting, economic, and finance professionals--clearly do not understand the full implications and suboptimal decisions mark-to-market brings.
Market prices are functions of balance sheet and income statement values. Investors will make different investment decisions in a mark-to-market world as compared to a historical cost or tempered mark-to-market world.
We do not operate in frictionless markets where accounting numbers do not affect values. After deliberation and good-faith intentions, FASB determines the fair value rules for the game, assuming that the rules have an insignificant impact on values. Then CPAs ensure institutions follow these standards and essentially bless the audited numbers that will be used in thousands of decision support systems by thousands of other participants. The balance sheet and income statement numbers are thought of by CPAs as outputs but are used as inputs in decision support systems at the micro and macro levels by economists, regulators, and investors. GAAP numbers are some of the inputs into relatively simple metrics like ROE and leverage, as well as complex metrics such as risk pricing. This makes market prices and metrics functions of calculated balance sheet and income statement values. Looking at it this way helps us understand why investors will act differently in a mark-to-market world versus a historical cost world. Market metrics will present different answers with inputs based on fair value financial statement numbers full of subjective gains versus historical cost numbers without such gains.
The Need for Control Mechanisms and Stabilizers
Over a decade ago, a Canadian study by Norman Macintosh and colleagues at Queen's University claimed accounting and markets, now tainted with fair value standards, "circulates in a hyperreality of self-referential models" ("Accounting as Simulacrum and Hypeareality," Accounting, Organizations and Society, January 2000).
A recent article in the journal of the Canadian Institute of Chartered Accountants, "FVA: Smoke and Mirrors," by Michel Magnan and Dan Thornton, reminded us of this study and expanded on this circularity claim with the analogy "accounting and the market are like two mirrors reflecting each other, each depending on the other for its information, in an endless endogeneity-loop where it is not clear what determines what" (CA Magazine, March 2010, www.camagazine.com/archives/print-edition/2010/march/features/camagazine34815.aspx).
The argument of Macintosh et al (2000) is that "companies" earnings determine security prices, which determines derivative prices, which determines companies' earnings. In short, neither the accounting sign nor the financial market sign appear to be grounded in any external reality. Instead each model appeals to the other for the only 'reality check' available." The authors further explain that "This has the potential to create market bubbles, where asset prices are decoupled from underlying cash flows, and to create systemic risk, where companies (especially financial institutions) depend on others' assessment of fair values of derivatives without reference to underlying property values."
Accounting numbers are used in all the financial metrics, so these circular references created in finance metrics are proving to be like feedback loops in engineering that will always become unstable. We acknowledge that mark-to-market feedback loops occur during contagion presumably because prices drop so fast, so it is easier to envision and comprehend the direct causal effect. However, mark-to-market feedback loops are not exclusive to declining periods--feedback loops also occur during upswings, although not quite as fast. The amplification during the upswing occurs over the period of quarters or years rather than a few months (or even days), so the causal link is harder to perceive. If it is inevitable that the market becomes irrational at times, we must ask the deep question of whether we should contribute to the irrationality by forcing certain hypothetical values to common equity that may distort market metrics further.
The Scientific Concepts We Forgot to Incorporate
The fields of finance and economics have increasingly relied upon scientific concepts and have arguably become true scientific disciplines. Wall Street, along with academia and their quantitative analysts, developed financial metrics like net present value, modern portfolio theory, and complicated copulas to price risk. We want to treat each field as a deterministic science, when each is more akin to quantum physics wherein some things are not so determinable. Both physicists and economists are striving to develop their disciplines' unified field theory. Other sciences like biology are extremely sensitive to behavioral factors. As in biology, finance and economics are subject to behavioral aspects unable to fit neatly in models (if at all). Traditional scientific fields are tainted by behavioral factors in the nature-versus-nurture debate--finance and economics are no different.
We have incorporated the mathematical concept of equilibrium in our metrics, but we neglected to realize the importance of other scientific phenomena, such as momentum, amplification, feedback loops, control theory, and the uncertainty principle. We have relied on our metric outputs as if efficient markets do not need stabilizers and behavioral finance phenomena are inconsequential and strategic opportunities for wise investors. Economic and financial models assume investors are rational and will pursue utility maximization, but behavioral factors often alter expected outcomes, sometimes to a severe degree.
The uncertainty principle in physics applies to accounting and finance too. Most of our current metrics (mark-to-market accounting rules, value-at-risk models, risk pricing metrics, derivative/option pricing models like Black-Scholes, modern portfolio theory, and even Basel banking rules) all fall victim to an phenomenon analogous to one physicists recognized nearly a century ago--the uncertainty principle, also known as the Heisenberg principle. The analogy mentioned by AEI's Pollock and others is very fitting, because the uncertainty principle states that measurement disturbs momentum. Alexander Bleck and Pingyang Gao open their mark-to-market paper with the 1929 Heisenberg quote, "The measurement of position necessarily disturbs momentum, and vice versa." ("Where Does the Information in Mark-to-Market Come From?" May 2010, w4.stern.nyu.edu/accounting/docs/Summer_Camp/Bleck.Gao.Where.Does.the.Information.in.M2M.pdf). In other words, attempting to measure position (price) makes the price move. Therefore, we affect asset price in our chosen accounting regime with our rules about how to measure values. In a mark-to-market world, asset price is a moving target--market metrics continuously move asset prices as we attempt to measure income and equity also using fair value metrics.
Our metrics are all dependent on some other metric output, which is dependent on another metric (which is dependent on the first metric), thus creating circular references everywhere. CPAs who are avid Excel users know that Excel does not allow a cell to reference back to itself and forces us to resolve our circular references. Financial, accounting and risk metrics, however, are designed and based on such loops.
Accounting numbers like profits, gains, and losses are used to determine risk premiums, but accounting numbers are often derived using risk premiums in their calculations. Which came first--the chicken or the egg? A recent e-mail from the Global Association of Risk Professionals recently quoted JBGlobal's James Berman. Beta is a key concept in modern portfolio theory and attempts to measure how volatile an asset price is relative to the market, but "attention to beta as a de rigeur risk measurement tool has been a complete failure. ... Volatility is a poor measure of risk, since it is the market's own measure. If the market's price is not right, how could the risk be?" Our markets have these omnipresent circular references and feedback loops throughout, which exacerbate asset price movements, sometimes to the point of speculation. When irrational exuberance can no longer continue because of a shock to the system (or because the markets can climb no more), irrational fear and contagion kick in.
Current finance models seem to work for a time absent an extreme "tail" event, but our current metrics have been systemically insufficient in the new mark-to-market-world coinciding with increased complex financial innovation and more global and interconnected markets. The overreliance on our metrics did not lead to a major crisis for many decades, because there were various stopgaps in place to keep leverage, risk, and even greed in check. One by one we lost all our irrational market stabilizers, which were not foolproof but had prevented the catastrophic and unsustainable buildup of risk, paper equity, and leverage we recently witnessed.
Each control mechanism (e.g., the gold standard, transparent exchanges coordinating most derivatives, historical cost accounting, and the Glass-Steagall Act) had their own issues and unintended consequences, but we let those stabilizers go without replacing them with sufficient substitutes, revised metrics, and effective countercyclical measures. Efficient market theory under an untamed mark-to-market environment only works for a time until the leverage effect and the behavioral aspects of perceived wealth, greed, group-think, speculation, and contagion take over.
What's the Solution?
Can FASB continue to blame others without looking in the mirror if its rules allowed equity to become distorted? We have fallen victim to the "logic of auditability" when we can "confirm" a value by checking the "market" price or testing a model, but verifying that a number matches an irrational market value is not relevant or useful but rather irrelevant and harmful. Many economists and finance experts are recognizing that the belief in rational markets was naive, so should accountants continue to develop standards with such blind faith and disregard for the inevitable irrational market behavior that will always be present? Or is it that markets act in a relatively efficient manner but lack revised metrics and knowledge of the increasing percentage of unstable equity? It is time to acknowledge that accounting based on market prices requires major enhancements, including revisiting how unrealized positions should be handled and consideration for countercyclical measures such as dynamic loan provisioning where loan loss reserves can be built up during the euphoric times. The conventional wisdom that forward-looking loss measures are earnings management must be changed.
If we could historically calculate and add up mark-to-market gains moved to equity from 1) securitized assets on balance sheet and off, 2) derivatives gains, and 3) premature realized gains from overpriced assets because of inaccurate mark-to-market proxies that led to the systemwide misjudgment of risk, we could begin to understand the scale of mark-to-market's potential role. (Even realized gains become based on duped market prices derived from pricing metrics utilizing information tainted by mark-to-market.) The complete picture would add in subsequent premature gains from 1) the increased capital flowing in to chase returns amplified under mark-to-market, 2) rehypothecation activity arising from collateral pledged on mark-to-market gains, and 3) commodity overpricing gains because of the artificial housing material demand. Only an agency like the SEC or the Federal Reserve could force disclosure of the needed historical data, but neither agency has acknowledged that fair value accounting is too procyclical for banking regulation to effectively handle.
If we require adequate disclosures about all mark-to-market gains, then the markets will hopefully do a better job of pricing and be more efficient. If we don't temper mark-to-market by valuing with multiyear moving averages, as suggested by Sapra, then we definitely need to adjust our relied upon metrics to better account for uncertain gains. Let's begin the new mark-to-market discussion today in order to facilitate improved accounting and better finance and economic metrics. Maybe we can help prevent the next normal business cycle from inflating into a super bubble and minimize the need for many of the other planned regulatory measures. Otherwise, the next super bubble and meltdown might end capitalism as we know it.
Gina McMahon, CPA, is an auditor with the federal government and a recent graduate of the University of Maryland, University College, Adelphi, Md.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||markets & investments|
|Publication:||The CPA Journal|
|Date:||Feb 1, 2011|
|Previous Article:||The anatomy of an excise tax: indoor tanning services.|
|Next Article:||Accounting for unclaimed property.|