Commenter JKH has been doing some thinking about mark-to-market and its role in the current crisis, and WCI is delighted to report that he has been good enough to put those thoughts together and agree to have them posted here:
Lurking in the background to the credit crisis and the unfolding policy responses to it has been the issue of mark to market accounting.
Bankers distinguish between two types of accounting, known as accrual (ACC) and mark to market (MTM). Simplified, accrual accounting includes only ongoing interest or yield income in reported earnings. MTM accounting includes the same accrual plus an adjustment for changes in the value of the asset over the period. For example, annual ACC earnings on a 5 year bond would include interest plus any amortization of premium or accretion of discount, based on the original book price. Annual MTM earnings on the same bond would include the same accrual earnings, plus the change in the market value of the bond over the reporting period. Thus, accrual accounting, in addition to being a standalone method, is also a sub-component of MTM accounting. But when people talk about the issue of MTM, they generally refer to the accounting component that differentiates it from accrual accounting; i.e. the adjustment for changes in asset values. Finally, fair value (FV) accounting refers to the formal application of mark to market accounting under specified accounting rules, including the treatment of various types of assets according to their inherent liquidity and marketability characteristics. FV accounting rules are complex, and we won’t explore them in detail here.
Commercial banks use accrual accounting in accordance with the intended hold to maturity nature of their loan books. Broker dealers use MTM accounting due to the liquid nature of their securities sales and trading businesses. So called “universal banks” combine both types of banking and both types of accounting under one roof. Commercial bank deposit taking and lending is captured in the accrual book; broker dealer securities activity in the MTM book. The two segments are also known the “banking book” and the “trading book”, and the resulting accounting combination as the “mixed model”. Citigroup, JP Morgan, and Bank of America are universal banks. Goldman Sachs and Morgan Stanley are broker dealers (although they’ve recently obtained commercial banking licences). All five major Canadian banks are universal banks. (This overall description is generic and simplified.)
Regulators of universal banks have steadily expanded the scope of FV accounting over the past two decades. The motivation has been to increase the “transparency” of financial statements, so that changes in underlying asset values are open to investor scrutiny and evaluation. The result has been an incursion of MTM measures into what were previously accrual banking books, and an FV accounting regime of increasingly wide and complex scope within universal banks.
The issue of MTM accounting as it pertains to the credit crisis has much to do with the resulting ambiguous border-line between the two types of accounting within universal banks. When questioned at last week’s Congressional hearing, Goldman Sachs’ CEO suggested that the MTM problem lay primarily in the reluctance of universal banks to mark down assets properly to market (Goldman is not a universal bank). Citigroup’s CEO on the other hand claimed that assets already marked down to distressed levels on an MTM basis did not reflect true value, and therefore the bank was reluctant to sell them. (Citigroup is the prototype example of a universal bank.) These are different interpretations of the MTM issue. Such different viewpoints are widespread and contribute to the debate on how to solve the credit crisis.
The wisdom of MTM accounting has been debated before and since its original implementation. On the one hand, it “shines a light” on developing asset value problems. On the other, it has the potential to exaggerate volatility in reported earnings, in some cases unnecessarily. As a simple example, a 5 year fixed rate loan or bond that matures in tact may exhibit substantial MTM volatility throughout its lifetime. But such volatility accumulates to zero net effect by the time the bond matures. To the degree that MTM based risk measures such as “value at risk” (Var) and capital measures such as “risk adjusted return on capital” (Raroc) pay attention to this volatility, measured risk and attributed capital are higher than is the case where the bond is assumed to be held to maturity. This capital effect in terms of both risk measures and unrealized losses has been critical to the dynamics of the credit crisis. Unrealized MTM losses create capital shortages, motivating liquidations, creating realized losses, pushing prices lower, feeding back to further unrealized losses. Potential distortions spread from such plain vanilla examples to the wider scope of various credit risk instruments. Much of the credit crisis problem as it pertains to banking revolves around the pricing of so called “toxic” or “troubled” assets, and other assets that are at least guilty by association. The fact that both composition transparency and market liquidity are lacking means that the appropriate MTM treatment for these assets is problematic. Various credit crisis solutions have been proposed, such as MLEC, TARP I, bad bank, good bank, TARP II, and full nationalization. All of these depend on the problem of asset valuation and pricing and MTM, one way or another. Nevertheless, the wisdom of such aggressive MTM application is a question that lingers.
Willem Buiter (see below) has commented recently on one aspect of FV accounting, that being consistent treatment of assets and liabilities. E.g. an increase in market interest rates reduces the value of a fixed rate asset, which reduces FV earnings. The same effect on a similar fixed rate liability reduces its value, which increases FV earnings. The pathological case is that of deterioration in a bank’s credit quality, with a widening of credit spreads that reduces the value of its liabilities and increases its FV earnings. Buiter has noticed what bankers themselves complained about almost 20 years ago when FV accounting was first proposed. Given the broad transparency objective of FV accounting, there is no rationale for applying it to assets alone. This apparent “foolish consistency” is symptomatic of unintended consequences.
Events are rapidly unfolding in the global credit crisis. There is increasing momentum toward the view that governments should urgently nationalize banking systems on a broad scale. The tendency in that context is to assume the MTM issue is increasingly irrelevant. Before allowing this, we should highlight several fallacies about MTM, which are at the conceptual origin of the issue.
The first fallacy is that MTM accounting is necessary in order to produce balance sheet transparency. This is not the case. Given the nature of the information it conveys, we can say more accurately that MTM accounting is a sufficient, but not a necessary condition for transparency. Insofar as asset values are concerned, the necessary condition is MTM disclosure, not MTM accounting. Exactly the same information can be made public through disclosure, without necessarily affecting reported earnings and capital positions. This would still allow the market to incorporate internal MTM information in the stock price. Shareholders could evaluate the impact of illiquid asset markets, uncertain pricing, and the longer term relevance of immediate MTM asset values, and impound all of this in the current stock value. This avoids a compound and to some degree duplicative and arguably inefficient pricing process in which internal assets and external stock value are both subject to an MTM valuation process.
The second fallacy is the contention that choice of accounting has no bearing on the economics of value. This would be conventionally correct if accounting had no bearing on the actual requirement for bank capital. But it is wrong for precisely this reason. And this is of particular importance to the credit crisis.
A relevant and connected analogy exists in the household sector. The housing boom was based on easy credit and an implied belief in MTM accounting for houses, otherwise known as the ATM effect. Current house prices, however inflated, were the basis for many hundreds of billions of dollars of mortgage equity withdrawals (MEW). Those who bet on housing MTM as an indicator of sustainable price appreciation behaved accordingly. So did mortgage lenders. Those who restrained from using MTM as a mental accounting of a sustained trajectory of personal wealth, and who considered such MTM information with more restraint, probably checked their behaviour more wisely in terms of allowing for risk. In order words, behaviour was a function of how households viewed the effect of housing MTM on their personal balance sheets and longer term capital positions. Those who resisted incorporating full MTM into their own capital evaluations were probably more restrained in MEW transactions and related spending. Those who acted aggressively on their housing MTM profile overextended their balance sheets and spent the money from their MEW proceeds. The economist Steven Roach referred to the corresponding macroeconomic effect in terms of the “asset economy”. The trajectory of this economy is tracked in the Federal Reserve’s quarterly computation of US household wealth. Trillions of dollars of wealth have now been eroded from household balance sheets since the credit crisis began.
A number of reputable economists and analysts have judged the US banking system to be insolvent. They conclude this in part by comparing the probable value of bank assets and liabilities with system capital. Forecast loan losses apart from MTM losses are also an important factor. These forecasters may be right. Who would bet strongly against Nouriel Roubini at this stage? President Obama has invoked “catastrophe” as the worst case scenario for the US economy, and positioned the importance of fiscal action accordingly. One could describe the banking problem similarly, and the importance of policy action there. Prescriptions for resolving the banking crisis include permutations and combinations of government intervention and asset reconfigurations such as “bad bank” (MLEC, TARP I, TARP II), “good bank”, and full nationalization. More commentators are pronouncing both US and European banking systems as insolvent, and urging quick nationalization, in order to avoid a Japan style protracted bank balance sheet deflation. Martin Wolf (see below) alleges the new US administration has already failed to recognize that the US banking system is insolvent. In his view the current policy response has already failed; nationalization is required. He makes the point that aggressive action is required to prevent a calamitous consequence that is asymmetric in its severity. All of this may be right. The MTM issue has been overtaken by this momentum. So where does it fit now?
The government assumes asset risk in any of these bail-out mechanisms. But unlike private sector banks, it incurs no additional capital requirements due to unrealized MTM losses on assets that it purchases (not that it guarantees). Neither the Treasury nor the Fed requires more capital funding because of MTM changes in the prices of assets that they hold. (The issue of Fed capital adequacy is really part of the same question for the government as a whole.) In other words, Treasury and the Fed can hold assets to maturity, but the private sector may be inhibited from doing so, given its required response to MTM unrealized losses and capital constraints.
The Treasury and Fed balance sheets offer considerable potential leverage to a financial system that is undergoing massive deleveraging. This is the essence of the requirement for their balance sheet intervention. But they also offer MTM immunization by virtue of their own balance sheet process. Therefore, why not provide MTM immunization (i.e. “relief”) directly to the banks, as part of a broader package of intervention? Indeed, this is what a number of commentators have urged for some time. Among them are Paul Volcker (former Fed chairman), William Isaac (former FDIC chairman), and Bob McTeer (former Fed governor; see below). Opposing this general proposition is the purist MTM view, which happens to note correctly that Japan’s banking system failed to recover due to half-way measures. But this doesn’t deny the potential dysfunctional effect of MTM on bank capital requirements in any environment.
TARP II as outlined by US Treasury secretary Geithner envisages a public-private partnership for the purchase of bank assets. The idea is that government will provide the leverage necessary to make these assets attractive to private equity capital. This would presumably also lift current MTM pricing, simply due to financial engineering. The irony is that private capital because of this financing may discover assets worth holding on the basis of underlying cash flow, i.e. on an accrual basis. Thus, the government may move assets out of the banking system that are disabled on an MTM basis to an investment platform where they make sense on an accrual basis. This would include the decomposition of complex assets such as CDOs in order to identity cash flow arising from the underlying assets. So long as these individual assets generate cash flow, they are “worth” something rather than nothing.
In summary, the banking crisis reflects the problem of dealing with so-called “toxic assets” and associated write-downs forced by MTM and FV accounting. Various solutions have been proposed involving different ways of severing toxic and non-toxic portions of a bank balance sheet, including full nationalization. These mechanisms will depend on the price at which these assets can be sold or transferred away, and how MTM accounting will factor into the arrangement. There may also be some scope for meaningful relief on MTM accounting in universal banks. The credit crisis itself reflected an asset bubble that in good times was enabled by aggressive MTM accounting, explicit or implicit, across all sectors of the economy, including banks and households. Universal banks combined and integrated MTM oriented investment bank operations with more staid, accrual accounted commercial operations. Perhaps the financial architecture of the future will feature a return to commercial banks that resemble deposit taking and lending utility functions, with minimal exposure to MTM businesses, and investment banks that are still MTM driven but small enough to fail.
Various opinions on mark to market: