Good Primer on CDOs

The Financial Times’ Paul Davies has written a good short piece on the basics of CDOs, which is useful if you are ever in the unfortunate position of having to explain them to someone new to the concept. He also suggests that subprime-related CDOs going pear shaped is not an indictment of the technology. Nevertheless, he also points out that a related similar financial structure, collateralized loan obligations, most commonly used for LBO debt, may be on the cusp of getting a bad name due to similar overly optimistic assumptions about the likely performance of the underlying assets.

One quibble: Davies is talking about “passive” CDOs. CDOs can also be “active” or “managed.” In that case, the debt is issued on a blind pool basis (at most, only some of the assets will have been purchased at the time of underwriting) and a manager trades the assets over the life of the CDO. It is a continuing source of frustration that I have never seen any information on the relative magnitude of active versus passive CDOs.

From the FT:

With the crisis at the riskier end of the US mortgage market gathering pace and investor jitters spreading to much broader segments of the debt markets, attention is increasingly being focused on the complex structured bonds that helped to fund this lending boom.

The intensity of this focus will only sharpen following news that two of Bear Stearns Asset Management’s hedge funds which were heavily exposed to something called collateralised debt obligations (CDOs) have essentially lost everything.

So what is a collateralised debt obligation?

The first thing to understand is that CDOs are a process, a kind of financial technology, rather than an asset class. The roots of this technology are in securitisation, which is the trick of borrowing money against cash you expect to receive in the future.

The reason sub-prime mortgages – lending to people who have patchy credit records, who cannot prove their incomes, or who want to borrow many times their incomes – took off in the first place was the discovery that such lending could be financed through wrapping the loans up into large pools and “securitising” the expected repayments.

Banks weren’t keen to lend to such risky borrowers and instead new companies such as Kensington in the UK were created to exploit this new business model.

Why would anybody else be keen to lend to these risky borrowers?

Herein lies the wonderful alchemy of securitisation and CDOs. The pool of debts created for these vehicles is like one side of a company’s balance sheet – its assets. They are an asset because they will provide income from repayments in the future.

The “company” that is created here must fund the purchase of these assets somehow and it does this by issuing debt and equity like any other company.

Through the magic of mathematical probability, the credit rating agencies will look at this “company” and say that the chances of, for example, 70 per cent of its assets all going bad and not providing income are very, very low. Thus it will give debt representing that share of the pool its safest AAA rating.

Many kinds of investors from banks to pension funds and life insurers love this triple-A debt, especially when it pays a higher interest coupon than other kinds of debt with a similar rating.

What about the rest of this pool of debts?

The next, say, 10 per cent of the pool faces a higher likelihood of turning bad so it gets a lower rating and so on down to the first 5 per cent of the pool, which is quite likely to see problems and thus gets no rating. This last bit is the “company’s” equity – it receives the highest returns, but bears the highest risk of suffering losses.

Mortgage-backed bonds are the debts that fund pools of mortgages directly, while debt issued by CDOs is used to fund pools of other kinds of bonds or loans that are issued by public and private companies, property investors and, crucially, mortgage securitisations.

So who buys this stuff?

Are you sitting down? Your pension fund is likely to have holdings of most parts of these kinds of deals from the safest triple-A paper down to the equity. Recent research from Citigroup suggests in fact that traditional asset managers – pensions and bond funds – have more than 40 per cent of their exposure to CDOs in the equity and less than 20 per cent in the safest triple-A bonds.

Are they mad? Am I going to lose everything?

Relax. There are a number of reasons why your money is not going to be entirely sucked down some strange financial vortex. First, equity tranches are very small in nominal terms and any pension fund’s money that is actually at risk will be far, far smaller than the amounts they have in other areas such as equities and ordinary bonds.

Second, so far in most markets the theories around ratings and mathematical probabilities have held up fairly well – and the kind of debts that these vehicles finance is not all off the scale of prudence, in the same way that US sub-prime lending became in recent years.

However, some hard questions are being asked about the aggressiveness of lending to leveraged buy-outs of companies such as Boots and much of this lending is financed by a special type of CDO known as collateralised loan obligations, which your pension may well have exposure to.

Third, the reason the Bear Stearns’ funds lost everything was because they borrowed money to invest in CDOs, which your pension fund will certainly not be doing – or not directly at least.

So I’m all right then . . .?

Not exactly. Such investments remain highly risky, although they can also be equally highly rewarding. Some kinds of structures have held up well in previous economic downturns, although others have been blown apart. But one thing’s for certain: they have never been used so widely to fund so many different kinds of debt as now.

Print Friendly, PDF & Email


  1. Anonymous

    Please provide feedback on this timeline.


    Prelude: securitization of mortgages spurs lenders to issue as many loans as possible because investors keep buying RMBS. Lenders make more and more loans and pay less and less attention to creditworthiness and ability to pay off borrowers, resulting in a mortgage and home price bubble. Eventually the bubble implodes in gradual phases:

    Phase 1
    Events: ARMs reset, and home prices stagnate because of oversupply and prices so high that buyers are priced out of the market.
    Effects: Subprime borrowers start defaulting on loans at a higher than anticipated rate. Prime borrowers start defaulting on investment property that they are now unable to sell as expected.

    Phase 2
    Events: RMBS investors realize what is happening and start to pull out of the market or at least stop pouring more money into it.
    Effects: Wall Street is a little worried, but not much because of hopes the problems are contained to the subprime market.

    Phase 3
    Events: Lending becomes more restricted in reaction to the defaults.
    Effects: Consumers cannot refinance their home or take cash out of their home as easily as before, and cannot as easily take out a second mortgage to buy investment property for speculation.

    Phase 4
    Events: Home prices start dropping because less mortgages means less buyers, and the oversupply is still there.
    Effects: The collateral to the RMBS loses value, which means the RMBS loses value not just because of defaults but also because of vaporized collateral value.

    Phase 5
    Events: Housing speculation comes to a complete halt because of dropping prices. Churning of housing stops.
    Effects: Less new loans to feed the pipeline of RMBS. Existing RMBS lose value consistently because of defaults and dropping value of collateral. Wall Street starts to really get worried about not just the whole RMBS sector, but debt securities in general because financial institutions and rating agencies got the RMBS situation so wrong.

    Phase 6
    Events: Consumers start spending less both because their homes are worth less and because it is harder to get a mortgage.
    Effects: Consumers start defaulting on credit cards, and start spending less on retail, cars, luxury items, etc. Home prices drop further. Wall Street now has to deal with a drag on the whole economy.

    Phase 7
    Events: Wall Street processes the losses, demands better quality RMBS and more realistic credit ratings in general, and borrowers, lenders, and rating agencies become more responsible.
    Effects: The economy and home prices stabilize.

  2. Yves Smith

    At a conceptual level, this looks like a good description of the arc, but it would be helpful if someone closer to the action could comment on the timing of distress in subprime related paper vs. non-subprime RMBS.

    Also, a couple of minor points: in Phase 2, “Wall Street is worried.” In the initial phases, at least some people on Wall Street weren’t worried. Merrill, Morgan Stanley, and Lehman were interested in buying failed mortgage brokers as late as mid-February, seeing the decline as a buying opportunity (

    Also, on Phase 6, although it seems obvious to me that declining housing prices will have a negative wealth effect and lead to lower consumer spending (even the Economist has commented that housing market declines have a bigger economic impact than stock market declines), some economists dispute that notion.

  3. david

    I am not close to the action, so this is just a guess. Traditionally housing prices tend to be sticky ever under stress. Given a deteriorating situation prices have to come down but is a good bet that this process will be going on at least until late 2008 early 2009.

    Precisely because housing has very slow time scales some people (the fed?) can assert there is no evidence of contagion. If it is going to spread (which it very possibly will), it’s gonna take months to a year or so at least.

  4. Anonymous

    Re actively managed or static pools in CDOs: my experience drafting them from 2001-2006 was that when the bond mkts and interest rate environments were relatively stable, active mgmt was favored and predominant. It seemed that when things were turbulent (making it hard to price deals because hard to forecast with confidence), the CDOs tended toward static or very lightly managed pools. One shouldn’t assume that a static pool is necesarily safer. Indeed, a good manager with latitude written into his Portfolio Mgmt Agreement can rescue a deal by selling stuff he senses is about to take a fall.

Comments are closed.