Monday, 10 November 2008

Synthetic CDOs

On my list of things to do certain jobs get ticked off very quickly. Others, however, linger that little while longer. In the latter category is an entry that reads “write up synthetic CDOs”. By the end of this article, you might understand both CDOs and why it's been on the list for a while.

The article – in the New York Times – makes a stab of explaining the near impossible, giving a background to one of the most complex financial products around, as well as a human face or two. However, it falls a little short of providing a fuller explanation of the details. (For more on investment banks failing, here's the FT giving a good insight into the PwC team restructuring Lehman.)

“The synthetic CDO grew out of a structure that an elite team of JP Morgan bankers invented in 1997. Their goal was to reduce the risk that Morgan would lose money when it made loans to top-tier corporate borrowers like IBM, General Electric and Procter & Gamble.”

Normal (or cash) CDOs (collateralised debt obligations) are funds that invest in bonds and loans, and then are themselves parcelled and then sliced up and sold onto bond investors. A kind of bond squared, with a dose of leverage (borrowing) at each turn. Meanwhile synthetic CDOs are like normal CDOs but instead of being made up of lots of bonds and loans, they are packed with credit-default swaps, a form of derivative. (For more detail, here's one prescient essay on the subject, and here's another from 2004, but with the usual industry defence of the product.)

According to the NYT's jaundiced post-credit-crunch eye: “Synthetics could be slapped together faster, and they generated fatter fees than regular CDOs, making them especially attractive to Wall Street.”

The originator of the synthetic CDO would usually retain a stake in the product but sell on much of it to other investors, and it was originally used as a mechanism to offsetting the risk of high-grade corporate lending. But from around 2002 there was shift away from insurance (aka 'balance sheet') to speculation (aka 'arbitrage') CDOs, writes the NYT, when investors were looking for yields after they had sold out of equities but could expect little return from bonds. This led to more borrowing (much of which went into the property market) and more appetite for risky investments with ever-increasing complexity.

These developments were supported by the belief that by spreading risks throughout the system, and cutting these risky investments up into small pieces, would mean that a few bad apples would not, and could not, spoil the whole barrel. The above-linked article from Credit magazine notes the following:

"it is broadly accepted that the risk embedded in the super-senior tranche of a synthetic CLO referencing a pool of investment-grade assets is remote in the extreme. This is because even if a super-senior swap accounts for as much as 90% of a CDO’s capital structure (which is not uncommon), more than 10% of the assets within the reference pool would have to default for losses to be sustained by the most senior tranche – an improbably high default rate for investment-grade bonds when the historical norm has been for a default rate of about 0.3%."

The NYT focuses on Merrill Lynch’s land-grab for mortgage lending, explaining that the bank sought to emulate the profits made by Lehman Brothers. “The firm’s goal, according to people who met with Merrill executives about possible deals, was to generate in-house mortgages that it could package into CDOs. This allowed Merrill to avoid relying entirely on other companies for mortgages.” According to executives interviewed by the NYT, Merrill Lynch focused more on profits than evaluating risk.

And unlike the pioneers of synthetic CDOs, Merrill Lynch “seemed unafraid to stockpile CDOs to reap more fees”. By 2006 it was the world’s biggest underwriter of the products. But the risks began to spread after AIG suddenly stopped insuring the highest-quality portion of the company’s CDOs against default. Merrill Lynch couldn’t find a replacement but continued to create new CDOs, and the company’s unhedged exposure to mortgages continued to grow.

“In 2005, firms issued $178 billion in mortgage and other asset-backed CDOs, compared with just $4 billion worth of C.D.O.’s that used safer, high-grade corporate bonds as collateral. In 2006, issuance of mortgage and asset-backed CDOs totaled $316 billion, versus $40 billion backed by corporate bonds.

Firms underwriting the CDOs generated fees of 0.4 percent to 2.5 percent of the amount sold. So the fees generated on the $316 billion worth of mortgage- and asset-backed CDOs issued in 2006 alone, for example, would have been about $1.3 billion to $8 billion.”

But when the market turned, in 2007, “the impact was brutal”. The market collapsed, and Merrill Lynch was forced to take a USD 7.9bn write-down related to its exposure to mortgage CDOs. Chief Executive Stan O’Neal was also forced out, replaced with John Thain.

“None of the trading businesses should be taking risks, either single positions or single trades, that wipe out the entire year’s earnings of their own business,” Mr. Thain said in January. “And they certainly shouldn’t take a risk to wipe out the earnings of the entire firm.”

In August, Thain arranged the sale of the CDO assets for 22 cents on the dollar. For the first nine months of 2008, Merrill recorded a net loss of USD 14.7bn on its CDOs. Shortly afterwards, Merrill was taken over by Bank of America.

No comments: