For those of you who are relatively new to the complexities involved in the pricing of collateralized debt obligations (CDOs), this Financial Times article, “Worries grow about the true value of repackaged debt,” gives a good overview. Since the article is lengthy, and the first part covers largely familiar ground, I’ve excerpted the second half.
One thing nags at me as I read this article. The assumption, which is articulated by a consultant, Christopher Whalen is that more liquidity would help pricing (in financial markets, that is close to a tautology):
The lack of a publicly quoted market for CDOs and like assets is exacerbating the liquidity problems for these assets beyond the underlying economics, for example, in subprime real estate
Now it is undeniable that more trading of CDOs would give more pricing benchmarks. Something is obviously better than nothing. But this market is a harder nut to crack than most people imagine. My belief is that a very considerable range of CDOs would need to be traded to provide enough data points to be useful. And lacking natural buyers and sellers (no one in the normal course of events sells this stuff, unless there is a problem, which means buyers are likely to be chary), it’s hard to imagine how market participants are going to get around this issue.
The reason I stress this issue is that readers might easily imagine that if a few “benchmark” CDOs were to be traded with some frequency, that would give this sector the anchors it needed for more realistic price marking for the CDOs that don’t trade. But these instruments are so arcane, so complex, and so highly differentiated on so many axes that one is likely to need to have a large number of CDOs trading to capture enough permutations to allow for realistic pricing.
Let’s consider the variations: underlying assets (they can contain any tranche of asset backed securities, other CDOs or even CDOs of CDOs, whole loans, mortgages), degree of credit enhancement (whether via overcollateralization or the use of guarantees), leverage, use of synthetics (I may have managed to miss an attribute or two, but you get the picture). As a result, the maturity of the deals vary, and the structures used to achieve the desired credit ratings are all over the map.
So even if a few of these puppies traded, I’m not sure what it tells you. You could try to infer what that means for illiquid issues, but unless you have a statistically reasonable sample trading, it’s hard to decompose why the liquid issues are priced the way they are priced. Or else you make an educated guess as to why they trade the way they trade and use a very complicated model to relate the price to the untraded paper you own. I’m sure it would be an improvement on what we have now, but my sense is that the public at large is overestimating the benefit of more active trading of a few issues. “A few issues” won’t scratch the surface of the variety and hairiness of the paper out there.
And there are a few other barriers: you can’t get the deal documents. No kidding. The Fed can’t even get them because it isn’t a “qualified investor.” (Should the Fed start a hedge fund so it can study this problem?). From “Where Did the Risk Go? How Misapplied Bond Ratings Cause Mortgage Backed Securities and Collateralized Debt Obligation Market Disruptions,” by Joshua Rosner and Joseph Mason (pages 83-4):
At present, even financial regulators are hampered by the opacity of over-the-counter CDO and MBS markets, where only “qualified investors” may peruse the deal documents and performance reports. Currently none of the bank regulatory agencies (OCC, Federal Reserve, or FDIC) are deemed “qualified investors.” Even after that designation, however, those regulators must receive permission from each issuer to view their deal performance data and prospectus in order to monitor the sector.
So if regulators can’t get the description of the securities, market participants certainly won’t. So what good is a price if you aren’t really certain what is being traded?
In addition, the discussion in the FT article presupposes the CDOs are passive CDOs, meaning the assets are assembled and the CDO is structured before it is sold to investors. Yet many CDOs are “active” or “managed” CDOs, meaning blind pools. Blind pools that are tranched, often with leverage and often buying other CDOs or “CDO squared” (CDOs of CDOs). That means the investors pony up money before the fund (it is like a convoluted mutual fund) is formed, and the managed gets to trade it over its three to five year life. No CDO manager is going to disclose his holdings (it would put him at a competitive disadvantage) but how can you value it otherwise?
I don’t understand how anyone with an operating brain cell could have bought this stuff, and the supervising grown-ups (the regulators) seem powerless to do anything to ameliorate the situation.
From the Financial Times:
To an extent, the valuation problem for CDOs reflects the fact that the frenetic pace of innovation seen in the financial industry this decade has outpaced the development of its infrastructure. It has often been the case that when new instruments emerge in the banking world, the market is initially quite illiquid, meaning that the level of trading is low. But the murky nature of new products has rarely had broad systemic implications, because they have typically occupied a small niche.
What makes the CDO sector unusual is that it has exploded at such a breakneck pace with bankers packaging bonds, loans and other debts into ever more complex structures. Last year alone, about $1,000bn (£500bn, €745bn) in cash and derivatives CDOs was issued in Europe and the US, according to data from the Bank for International Settlements. More than one-third was composed of asset-backed securities, often including low-grade mortgages.
As this explosion has occurred, some corners of this universe have already become relatively widely traded and transparent. Every day in the London and New York markets, for example, billions of dollars worth of deals are struck involving indices of derivatives on well-known corporate bonds – making it easy to obtain prices.
However, many other such products are created by bankers directly with their clients and then simply left to sit on the books of an investor. Since such instruments typically last three to five years – and the CDO boom is so recent – many have not come to the end of their life. Nor have they been traded. Christopher Whalen of Institutional Risk Analytics, a consultancy, says: “The lack of a publicly quoted market for CDOs and like assets is exacerbating the liquidity problems for these assets beyond the underlying economics, for example, in subprime real estate.”
To compensate, investment institutions and banks use a variety of techniques to assign a value to these instruments in their accounts. In some areas, third-party data groups exist that can offer price estimates. However, the pace of innovation is so intense that it is hard for these providers to keep up with all corners of the market. So in many cases, investors are turning to alternative techniques to create prices. One tactic used by hedge funds entails asking several brokers for price quotes and taking an average. Results vary – not least because dealer banks may hold positions in these instruments themselves.
“It is very easy for hedge funds to shop around to find valuations that suit them best and then book their assets at that,” says one banker who advises hedge funds. “Going back to the bank that sold you a CDO and asking for a price is rarely likely to produce an accurate picture.”
Another approach is to estimate valuations based on the ratings the instruments receive from credit rating agencies. Yet this does not offer a fail-safe valuation method either. The rating agencies have been downgrading bonds backed by subprime mortgages in recent weeks but critics say they have been slow to act and face difficulties in analysing the market.
Christian Stracke, analyst at CreditSights, a research company, says: “With so little truly relevant historical data on the behaviour of subprime mortgages, and with such massive structural changes having occurred in the mortgage landscape in recent years, any time-series analysis approach is little more than a not-so-educated guess.”
Moreover, while ratings attempt guidance on the chance of default, they offer no indication of how market prices could behave – as the rating agencies stress. As the BIS noted in its annual report this week, ratings reflect expected credit losses rather than the “unusually high probability” of events that “could have large effects on market values”.
That means that on the rare occasions that instruments are traded, a large gap can suddenly emerge between the market price and its book value. This week Queen’s Walk Fund, a London hedge fund, admitted it had been forced to write down the value of its US subprime securities by almost 50 per cent in just a few months. That was because when it was forced to sell them, the price achieved was far lower than the value created with the models the fund had previously used – which had been supplemented with brokers’ quotes.
But unless circumstances arise that force a market trade, valuations often remain at the investment managers’ discretion. While managers say they strive to assign honest values, these are often difficult for an outside accountant to verify, since the techniques used are invariably highly complex.
Moreover, incentives do not always encourage fair valuations: hedge fund managers, for example, are typically paid a percentage of the profits they book, giving them a vested interest in reporting a high asset valuation. At best, this means that the valuations of CDOs, for example, may often lag behind any swings in broader asset classes; at worst, this ambiguity may enable hedge fund managers or investment bankers to keep posting profits – even when markets fall.
But Amitabh Arora, head of interest rate strategies at Lehman Brothers, points to a further potential impact from the Bear Stearns upheaval. “The bigger risk now is that it calls into question CDOs as a financing vehicle in the corporate credit market – I think in the next six to 12 months we will see a significant reassessment of CDOs as a financial vehicle not just in the subprime world but the corporate world too.”
Adil Abdulali, a risk manager at Protégé Partners, a fund of funds, recently studied the performance of hedge funds and discovered clear statistical indications that they tend to stage-manage their earnings [known in the industry as “smoothing” them] when they trade illiquid instruments. “Conservatively, 30 per cent of funds trading illiquid securities smooth their returns,” says Mr Abdulali.
Some bankers and policymakers argue that this is simply a teething problem that will fade as structured finance becomes more mature. History suggests that most opaque, illiquid markets eventually become more transparent when they grow large enough – and behind the scenes, the Bear Stearns hedge fund problems are prompting bankers and investment managers to re-examine their valuation techniques. “We are getting a lot of calls from worried people,” says one third-party data provider.
However, history also shows that large-scale structural dislocations – such as a serious mispricing of assets – are rarely corrected in an orderly manner. Thus the big risk now is that if thousands of banks and investment groups suddenly have to slash the value of the securities they hold, the wave of accounting losses might at best leave investors wary of purchasing all manner of complex financial instruments. At worst, it could trigger more distressed sales and a broader repricing of financial assets, not just in the subprime sector but in other illiquid markets too.
“If every CDO [manager] was forced to mark to market their subprime holdings, it would be – well, I can’t think of a strong enough word to describe what it would be,” confesses a US policymaker.