Typical street scene in Santa Ana, El Salvador. (Photo: iStock)
IMF Survey: Credit Market Turmoil Makes Valuation Key
January 15, 2008
- Valuation models broke down after crisis began
- Certain types of securities became relatively illiquid
- Rating downgrades further complicated valuation
Following the collapse of the U.S. subprime mortgage market in mid-2007, market concerns about the exposure of structured securities to subprime loans dramatically slashed liquidity.
Agencies were forced to revise assumptions on speed at which delinquent loans led to foreclosures (photo: Zuma Press)
Subprime Notes
As trading volumes declined, many market participants were forced to use pricing models that relied on historical data. However, the recent performance of some of these subprime loans has been much worse than the record would have suggested. This has caused valuation models to break down.
How is valuation done?
Market participants commonly use three types of valuation techniques, often described as
• mark-to-market,
• mark-to-matrix, and
• mark-to-model
Mark-to-market refers to the use of quoted prices for actively traded, identical assets. A second method, mark-to-matrix, is a technique used for less actively traded assets, such as emerging market securities, municipal bonds, and asset-backed securities (ABS). It involves estimating a credit spread of the asset relative to a more actively traded instrument that can be priced easily.
A third method of pricing is the mark-to-model technique that market participants are often forced to use for the least liquid assets including real estate, private equity investments, and complex structured securities such as certain tranches of collateralized debt obligations or CDOs. Mark-to-model assigns prices based on statistical inference.
Why is valuation an issue?
Valuation became a problem because certain types of structured securities became relatively illiquid following the subprime crisis. The absence of market quotes forced market participants to rely more heavily on mark-to-model as opposed to mark-to-matrix techniques.
Some types of structured securities were inherently illiquid at the time of issuance. This included most CDOs because each debt tranche had different levels of credit enhancement and the composition (and quality) of the underlying collateral varied from one deal to another. The size of the global asset-backed CDO market is no more than about $400 billion, whereas the subprime market is about $1 trillion.
Illiquidity of complex structured securities was compounded by a lack of transparency about the exposure to underlying nonprime mortgage loans and to uncertainty about ratings. The complexity of multiple derivations of securities with cash flows from mortgage loans to various tranches of ABS and then to tranches of CDOs according to deal-specific rules made it difficult and time consuming for many investors to model these securities independently. Some therefore increasingly relied on ratings as a measure of default risk and inappropriately compared them to those on plain vanilla corporate debt, which has different sensitivities to market conditions.
"Illiquidity of complex structured securities was compounded by a lack of transparency about the exposure to underlying nonprime mortgage loans and to uncertainty about ratings."
The uncertainty about ratings was a result of multiple-notch downgrades of mortgage-related securities (ABS and CDOs) that occurred en masse. On average, during 2007, the agencies downgraded structured securities by 3-4 notches. A four notch downgrade is, for example, from AAA to A+. More than 3,000 downgrades occurred in October 2007 alone.
The unexpectedly poor performance of the underlying collateral of recently originated mortgage loans required abrupt changes in the models For instance, the agencies were forced to revise assumptions concerning the speed with which loan delinquencies translated into foreclosures and lowered the recovery amounts from the process as U.S. home prices declined. The agencies were also forced to revise modeling assumptions concerning the correlation between assets making up the collateral for structured securities.
Modeling assumptions affected the expected performance of structured securities. The higher the assumed correlation on underlying assets, the more likely it is for a loss to appear on senior debt tranches. On the other hand, a relatively low assumption of correlation is likely to impact only the equity tranche. Separately, ratings uncertainty increases exponentially with a linear increase in the size of the underlying portfolio of assets (see Chart 1).
The impact of correlation on the equity tranche of CDOs may appear nonintuitive; thus the analogy to a ship going through a strait that contains hidden mines may help. When the mines are clumped together (high correlation), the ship will likely miss the mines (that is, avoid losses), compared with the situation when they are widely scattered (low correlation). In the higher correlation case, if the ship hits one mine then it will hit some of the others too, leading to a large loss compared with a low correlation case when it will likely just hit one mine.
Rating downgrades further complicated valuation. Agency downgrades qualified as credit events, which diverted cash flows from junior to senior tranches of CDOs. Also regulatory pressures forced many investors (such as insurers and pension funds) to sell at distressed prices.
Are valuations realistic?
Some market participants are seeking to trade at artificial prices. Anecdotal reports suggest that some holders of structured products may be marking their portfolios using favorable bilateral quotes. Other market participants have created off-balance-sheet vehicles that effectively allow them to trade with themselves; the prices paid for these distressed assets are likely to be higher than the fire sale prices that would be obtained if they were unwound in the market. A good comparison of such off-balance-sheet structures is the Reconstruction and Collection Corporation of Japan, which was a conduit for resolution of the nonperforming loans (see IMF Working Paper 04/ 86).
Recent moves to bring such off-balance-sheet structures on balance sheet are not at explicit "transfer prices," and thus the associated capital charge (for recapitalization) may not be a full reflection of potential losses. The market for such structures is similar to a monopsony (dominated by a few buyers of subprime products); thus it has been in the vested interest of the financial institutions to explicitly take on such liability and not jeopardize the relationship with these buyers.
The amount of capital injected simultaneously when off-balance-sheet structures are taken on the balance sheets may appear to fill the capital loss; however, some analytical work modeled on conservative assumptions suggests that potential losses may be higher and further capital injections are likely.
What is being done to help?
Central banks in mature markets conducted liquidity injections that temporarily improved trading in some structured securities. These liquidity operations helped stabilize markets for some structured products, such as asset-backed commercial paper, which was declining in outstanding stock at a rate of $50 billion per week during late August and early September.
Nonbanks (institutional investors, pension and insurance funds, hedge funds, and so on) do not benefit from such liquidity injections but regulations allow them "more room to maneuver" when pricing such assets on their books.
How can valuation difficulties be resolved?
In the short term, market participants should seek to transact a portion of their holdings of complex structured securities periodically in order to obtain valid market quotes. Some have come out clean, such as a few U.S. hedge funds that have written off the value of all junior notes issued by their structured vehicles.
However, most banks in the United States have not yet marked their assets to genuine transaction prices, or to proxy ABX indices (see Chart 2). Distressed or vulture funds are expecting large price declines this January and February, after the bonus season (see IMF Working paper 03/161 for a summary on returns from distressed debt). Although many market participants assume that resets will be over by 2008, there is an 18-month lag from end-2008 (or mid-2010) when such resets lead to delinquencies that result in cash flow problems.
Elsewhere, the valuation (and thus losses) may take longer to surface; for example, in Japan, it is a widely accepted practice that many institutions need not mark all their assets to markets, since they hold them to maturity (Euromoney, September 2005, page 216).
In the longer term, policymakers should encourage investors to reduce their reliance on ratings as measures of risk for structured securities. From a longer-term structural point of view, policymakers should create incentives for market participants to transact on exchanges rather than in over-the-counter operations.