I thought I would revisit a post I wrote in November on two types of credit derivatives, because it may prove to be a useful way of thinking about how to regulate CDOs, synthetic CDOs, and credit derivatives generally.
We shouldn’t lump all of these complex financial products together. They can have very different values to society and pose radically different kinds of costs. The key distinction is whether at least one of the parties is using the transaction to hedge and existing credit risk – or whether neither party has a pre-existing risk and the financial instrument is a pure bet.
Pure bets have no net social value but no net social cost with a very big EXCEPT
Pure bets have no social value because they are zero sum games. Except one party may default on its obligations. One reason it may default is because it made too many bad bets and may go insolvent. If it defaults, it may cause a chain reaction of defaults by other parties. Fanciful theory? Well domino risk from credit derivatives this is the reason the government gave for bailing out AIG.
If these bets have no social benefit and possible risks, the obvious solution is to do away with them. Lynn Stout (UCLA) who has been analyzing this type of pure bet derivative for years, has argued for a return to the old common law rule that these bets are not enforceable contracts because neither party has an insurable interest. In other words, courts would refuse to enforce pure gambles. She recently reframed this proposal as “regulate derivatives by deregulating them.”
Hedging risk is great…
For hedges, when one of the party is using a derivative to hedge a pre-existing risk, the financial instrument may indeed have great social value in allowing risk to be reallocated and spread efficiently. Mortgage-backed securities and traditional “cash” CDOs perform this same risk spreading function. Credit derivatives, mortgage backed securities and CDOs also allow credit to be funneled from investors back to borrowers. If a party holding bonds or loans can either hedge the credit risk with a credit derivative or sell part of the bonds or loans in a securitization, it can take the money and re-invest or re-lend. That extra money can make its way back to borrowers like home buyers.
…except when it’s not: manufacturing leverage and printing money
So if these derivatives may have social value, what's the cost? Well they may have the same counterparty and mispricing risk of the pure bets. But they also pose another kind of danger of injecting too much liquidity and leverage into the entire market. An investor can use credit to purchase a CDO bond. And a credit derivative counterparty may need to set aside only a fraction of its potential obligations (post collateral) – which is also a form of leverage. Higher leverage means higher potential returns in good times and higher losses in bad.
With increased leverage, more money is funneled back to borrowers. This can increase the effective supply of money and cause a boom in asset prices. Essentially, the network of securitization and credit derivatives – what is called the “shadow banking system” – effectively printed more Monopoly money that could be used to buy Park Place. In turn, booming asset prices can raise the value of collateral for loans and lead to more leverage. Margaret Blair and I wrote a short piece last summer that included a brief explanation of how this second type of credit derivative can increase system-wide leverage.
So hedges need some regulation too beyond just dealing with massive counterparty risk and the risk of falling dominoes. Controlling excessive leverage is the problem.
The CDO at the heart of the Goldman suit was a special kind of CDO called a synthetic CDO. That means instead of the investment vehicle purchasing bonds or securities, it entered into credit derivatives that would pay out as if the investment vehicle actually held bonds. It was a pure bet. So it didn’t necessarily increase overall leverage in the economy or help pump up housing prices. It didn’t have any social value though other than giving us all an opportunity to fight over fraud, which is something we desperately need.
Full employment for securities lawyers!
Could a party use a synthetic CDO to hedge actual risk? Theoretically yes, but that is not likely to happen. For one thing, there are much simpler ways to hedge risk – like entering into a simple credit default swap, buying bond insurance, or selling interests in the bonds or mortgages you own. According to media accounts, Paulson & Co. was looking to create new synthetic CDOs but because it couldn’t find enough ways to short mortgage-backed securities.
1. Per a comment from John Walker below, I should clarify that Professor Stout is focusing on an insurable interest requirement only for OTC derivatives.
2. Professor Bartlett provides an example of an early synthetic CDO product created by JP Morgan in which banks hedged at least some risk in their portfolio. It looks like this product was largely designed to help banks arbitrage regulatory capital requirements. This potential purpose should be a another big concern with using credit derivatives. This arbitrage can undermine the effectiveness of capital requirements and can lead us into a false sense of security that financial institution leverage is limited.
So to the extent credit risk was being hedged, this JP Morgan product would fall under my second category of credit derivative. And then we have to worry about its effect on injecting leverage in financial markets. It is also interesting to see another example of a product that had early, even arguably beneficial use, evolve into the device in the Goldman case.
TrackBack URL for this entry:
Links to weblogs that reference Two types of credit derivatives: two types of CDOS; two dangers: