A brief, high-level explanation of the dispute: Under mark-to-market accounting, assets on your balance sheet have to be valued at their current market values. So if you have $10 million worth of stock in Microsoft, but that stock falls to $5 million, you have to write it down on your balance sheet and take a $5 million loss on your income statement. The criticism was that mark-to-market was forcing financial institutions to take severe writedowns on assets whose market values had fallen precipitously, not because of their inherent value, but because nobody was buying these assets – think CDOs – and that banks were becoming insolvent because of an accounting technicality. Under this view, banks should be able to keep these assets at their “true” long-term values, instead of having to take writedowns due to short-term market fluctuations.
I am instinctively skeptical of this view, and in favor of mark-to-market accounting, because I believe that while market valuations may not be perfect, they are generally better than the alternative, which is allowing companies to estimate the values themselves, subject only to their auditors and regulators. But the issue is considerably more complicated than either the simple criticism or my simple defense would imply.
Earlier this week, the SEC released its study of mark-to-market accounting as required by the bailout bill. Their conclusions are simple:
fair value [mark-to-market, as will be explained] accounting did not appear to play a meaningful role in bank failures occurring during 2008. Rather, bank failures in the U.S. appeared to be the result of growing probable credit losses, concerns about asset quality, and, in certain cases, eroding lender and investor confidence.
This should not be surprising. At a high level, accounting conventions are artificial constructs designed to ensure some measure of uniformity in financial reporting. Whatever the rules are for calculating certain numbers, savvy investors know those rules, and can make adjustments as they feel appropriate. (A good example of this is accounting for stock options as expenses – even before this became a mandatory part of the income statement, it was in the footnotes, so analysts knew what was going on; as a result, when it did become mandatory, it had little or no impact on stock prices.) In this case, even if banks did not have account for certain assets at market values, the investors still knew exactly what was going on in the markets for those assets, and could draw their own conclusions.
At 259 pages, I doubt many of you will read it, so I will provide a bit of a summary and commentary. Still, many parts of it are both educational and interesting. The executive summary is only 10 pages long. If you aren’t familiar with the basics of financial accounting, sections I.B-D make a good introduction.
What Is Mark-to-Market Accounting?
The first thing to understand is that the world is not neatly divided into “mark-to-market” accounting and some single other form. The broad concept is “fair value” accounting; assets subject to this treatment must be valued at the price they would receive in an arms-length market transaction. Fair value accounting may apply to assets that are not traded on visible, liquid markets (like exchange-traded stocks), so in itself in can involve estimates. And there are a number of alternatives to fair value accounting, of which the most familiar is probably historical cost accounting (assets are carried on the balance sheet at whatever you paid for them).
Companies have a fair amount of latitude in deciding how they account for different assets. In some cases, the accounting treatment depends not on the nature of the asset itself, but on what the institution plans to do with it. for example, the same security can be designated as part of a trading account, available for sale (AFS), or to be held to maturity (HTM). Trading assets are accounted for at fair value, and changes in their value affect the income statement (profits and losses) directly; AFS assets are accounted for a fair value, but changes in value do not show up on the income statement (only in a line of adjustments to equity, and these adjustments to equity do not affect regulatory capital requirements); and HTM assets are not accounted for at fair value. In addition, there are also assets that only become subject to fair value accounting if they are subject to other-than-temporary impairment (OTTI); the idea here is precisely to ignore short-term fluctuations, but only write them down if they lose long-term value.
In short, the system is already designed to protect financial institutions from having to take writedowns in their asset portfolios due to short-term market movements, which is what fair value accounting stands accused of.
What Impact Did Fair Value Accounting Have During the Crisis?
The first thing to note is that a majority of financial institution assets (55%) are not accounted for at fair value, and only half of those that are at fair value are of the type that affect the income statement (and therefore regulatory capital).
The second thing to note is that changes in fair-value assets during the first three quarters of 2008 were relatively small as a percentage of overall equity. Across a broad sample of the financial industry:
Items reported at fair value on a recurring basis, . . . resulted in . . . [a] 3% and 4% increase (on a comparable nine-month basis) for the first quarter and the first three quarters of 2008, respectively. . . .
impairment charges . . . represented 3% and 8% of equity (on a comparable nine-month basis) for the first quarter of 2008 and the first three quarters of 2008, respectively.
OTTI on securities comprised the largest component of total impairment charges, at $62 billion or 5.1% of equity.
In English: Changes in fair-value assets that affect the income statement actually increased equity by 4%; changes that do not affect the income statement reduced equity by 8% (remember, that’s a percentage of equity, not assets); and most of that was other-than-temporary impairment, meaning that the institutions themselves thought these were permanent changes, not short-term fluctuations. Instead of fair-value assets, it was good old-fashioned loan losses that hurt the financial industry’s income statement:
net income for banking, credit institutions, and GSEs was most significantly impacted by the increase in the charge for provision for loan losses, which is a historical cost concept, as the provision for loan losses is primarily based on “incurred” losses.
The SEC also specifically studied those banks that failed during the crisis:
For most of the failed banks studied, fair value accounting was applied in limited circumstances, and fair value losses recognized did not have a significant impact on the bank’s capital. For the failed banks that did recognize sizable fair value losses, it does not appear that the reporting of these losses was the reason the bank failed. Market concerns about these companies, as evidenced by their share price, appear to indicate that the marketplace factored in losses for these banks that had not been recognized in U.S. GAAP reported income.
For small banks (<$30 billion in assets), declines in capital were overwhelmingly (~90%) due to increased loan loss provisions for the loans they held on their books. The same was true of Washington Mutual. The only exception was IndyMac, for which increased loan loss provisions only accounted for about 50% of capital declines. Even for IndyMac, though writedowns on fair-value assets were not made at fire-sale prices:
While IndyMac stated that it believed that a portion of the fair value losses it recognized during 2008 would recover over time, IndyMac also stated that it used its judgment to arrive at a fair value estimate for these securities that it believed did not represent a fire-sale valuation.
For the three largest banks, the SEC compared bank stock prices to book values (which reflect writedowns), and found that “market concerns regarding these companies pre-dated any significant fair value losses that these companies recognized.” In other words, investors were concerned because they knew that the banks had large mortgage portfolios, and they could see what was happening to the values of houses, mortgages, and mortgage-backed securities, and they drew their own conclusions independent of writedowns in quarterly statements. And the deathblow to Washington Mutual was caused not by a new accounting statement: “Instead of reduced capital, the proximate cause for the failure of WaMu appears to have been dramatic increase in deposit outflows sparked by concerns about the quality of the bank’s mortgage loan assets.”
The report draws a similar conclusion regarding non-banks, such as Bear Stearns:
Instead of accounting and reporting being the crisis’ primary driver, the observations indicate that the liquidity positions of some financial institutions, concerns about asset quality, lending practices, risk management practice, and a failure of other financial institutions to extend credit appear to be the primary drivers. . . .
liquidity pressures brought on by risk management practices, and concerns about asset quality precipitated by a rapid decline in confidence in these financial institutions, appears to be the primary cause of their financial distress and in some cases bankruptcy.
This goes back to the basic point that whether or not a financial institution is solvent – and I have heard it said by people who should know that Bear Stearns was solvent at the time of its collapse – it can still suffer a liquidity run.
Why Is Fair Value Accounting Good?
Ultimately, the point of fair value accounting is to provide accurate information to investors. The basic principle is that where possible, companies should account for their assets at their real values, not at some other value that they can make up. The SEC study cites one example of where not using fair value accounting caused a problem:
…in the Savings and Loan Crisis in the U.S., historic cost accounting masked the [extent of the] problem by allowing losses to show up gradually through negative net interest income. It can be argued that a mark-to-market approach would have helped to reveal to regulators and investors that these institutions had problems.
(Citing Franklin Allen & Elena Carletti, Mark-to-Market Accounting and Liquidity Pricing, 45 Journal of Accounting and Economics, at 358-378.) In the S&L crisis, thrifts did not have to account for the fact that their loan portfolios had plummeted in value because the interest rates they were receiving were lower than the interest rates they were paying depositors (due to a surge in inflation).
There is no chance that one report from a largely discredited agency will settle this question once and for all. But hopefully it will at least teach people that the issue is a lot more complicated than you would think from reading newspaper opinion pages.
Originally published at The Baseline Scenario and reproduced here with the author’s permission.