Don’t Kill CDO’s

I’ll be on the Kudlow Report at 7 p.m. talking about collateralized debt obligations (CDO’s). My crib sheet:

There is very clear case for CDOs: the most obvious is that they are now an important vehicle for loans to medium-sized business at a time when the banks have reduced their loan book by 20 percent over the past 12 months. A CDO simply puts credits into a pool and sells different risks to different investors. You can cheat investors with a CDO, but you can also cheat them with common stock any other financial instrument you might name. The financial industry and the ratings agencies have lost investor confidence by misusing the structure and have to straighten up and fly right—but killing CDOs would damage precisely the sectors of the economy that need capital the most.

Roger Lowenstein in the New York Times today claims that synthetic CDOs are just gambling:

Wall Street’s purpose, you will recall, is to raise money for industry: to finance steel mills and technology companies and, yes, even mortgages. But the collateralized debt obligations involved in the Goldman trades, like billions of dollars of similar trades sponsored by most every Wall Street firm, raised nothing for nobody. In essence, they were simply a side bet—like those in a casino—that allowed speculators to increase society’s mortgage wager without financing a single house.

The mortgage investment that is the focus of the S.E.C.’s civil lawsuit against Goldman, Abacus 2007-AC1, didn’t contain any actual mortgage bonds. Rather, it was made up of credit default swaps that “referenced” such bonds. Thus the investors weren’t truly “investing”—they were gambling on the success or failure of the bonds that actually did own mortgages. Some parties bet that the mortgage bonds would pay off; others (notably the hedge fund manager John Paulson) bet that they would fail. But no actual bonds—and no actual mortgages—were created or owned by the parties involved.

He misses the obvious: If investors demand to own a certain kind of risk in the synthetic market, it makes it easier for borrowers to issue cash debt (among other reasons, by making it easy for investors to hedge cash debt if they choose to).

Synthetic CDO’s work the same way as cash CDO’s. The industry used synthetics rather than cash simply because the transaction costs are much lower. You can write credit default swaps against a name at any given moment you want; cash CDO’s require you to go into the market and buy the actual bonds. Bonds aren’t fungible, like common stock; you have to The Twilight Saga Eclipse 2010 find specific issues all with matching maturities. This makes it necessary to “ramp up” the CDO while you scour the market for the bonds and warehouse them for two or three months, which means that the investment bank has to take the market risk of holding the bond portfolio until the deal closes. The paper work is costlier, registration takes longer, and so on. He doesn’t know a thing about the industry.

And I might add that repackaging the risk of already-issued cash bonds doesn’t raise money for anyone either, because the bonds have to have been issued already. Of course, it may create demand for additional issuance because it enables investors who can’t own the bonds as they are issued (for example, because their ratings are too low, as in case of middle-market obligors) to participate.

Now, the synthetic structure can be subject to chicanery. In a cash CDO, you get money and buy bonds and swap the cash flows into floating rate instruments. In a synthetic, you get money, put it in escrow, and sell protection on reference names. The income from selling protection is quite close to that of owning a floating-rate cash credit (spread to LIBOR). There sometimes is a difference (the CDS “basis”) but that’s usually minor. Investment banks got into trouble on the escrow side: some of them took the cash they held in escrow (in case credits defaulted and the CDO had to actually pay out on the protection it sold), and invested it in other CDOs. That’s a sort of pyramiding. But you can do that with cash bonds, too. Some investment banks at the dawn of the high-yield market sold junk bonds for companies on condition that they invest some of the proceeds into other junk bond deals. So there is nothing unique to the potential for chicanery with synthetic CDOs.

Risks should be transparent. Banks should make their internal default models available to investors so that everyone is looking at the same information—and investors should be able to compare the models offered by various banks.