The numbers that are thrown around are so mind-boggling that they are mind-numbing. The total amount of Credit Default Swap (CDS) obligations outstanding, according to the Bank of International Settlement, was 57 trillion US$ in December 2007 (pdf).
That is roughly Four Times the size of the US GDP.
These are the things that Warren Buffet called a "time bomb".
What are they? Well, suppose that we are watching a little old lady crossing a street, and want to take out a life insurance policy that pays if she gets clobbered by traffic. Unless she is close family, or she is a business partner, we can't do that ... we have no insurable interest.
But if we were watching a company, and wanted to buy a contract that pays off if the company can't pay on its bonds, we could. We'd buy a CDS.
You Too can be an Insurance Company
Well, no, not everybody can be a real insurance company. You need to get a pile of money together, and have to submit to regulation on the assets you hold against your insurance liabilities.
And not everybody can buy insurance ... so you have to check whether the people requesting coverage have an actual insurable interest in the thing they want to get covered. If "Matches" McGee comes in for fire insurance on the bank across the street, and he is not representing the bank across the street, or maybe if its a leased premises the landlord, then "Matches" McGee would have to be sent packing.
Ah, but that's real insurance, not derivative financial contracts. Under a Credit Default Swap, the buyer of the protection pays for protection against a "credit event", and the seller of the protection promises to pay off in the event of the credit event.
A common credit event is default on a bond.
Here's the "beauty" of it, though: the buyer of the protection does not have to hold the bond. They could be holding the bond ... or they could just be betting that the bond will fail. There's no requirement to have an insurable interest ... its just a contract that says this much money flows this way over this period of time, and if this event occurs during the period of the swap, then that much money flows that way.
So there can be $10b in bonds with $100b bet on whether the bonds will be paid off or not.
And to sell those Credit Default Swaps, you only have to persuade a CDS dealer to let you ... its a private, over the counter market. None of that nasty business of satisfying tedious, time consuming government regulations.
That's part how the CDS's get to a face value estimated at $45 TRILLION worldwide.
Laying off the bet
If a bookie takes a bet, sometimes they are getting all the money on one side ... and don't want to be betting all their own money on the other side. So they lay off part of the bet by covering the bet with other bookies.
Same thing with a CDS. A company that has sold protection thinks they are overexposed, they can take out CDS of their own. Anybody can take out a CDS after all ... well, anyone who can handle the big sums involved in each one.
Except, the devil is in the details. Take the case of Aon and Societe General:
Bear Stearns provided a loan of US$10 million to a Philippine entity and demanded the borrower obtain a surety bond from a Philippine government agency, the Government Service Insurance System (GSIS). Bear Stearns, to further hedge default risk on the US$10 million loan purchased protection contract from AON for US$0.425 million. AON, to hedge this risk purchased protection from Societle Generale for US$0.3 million believing it made a cool profit of US$0.1 million.
Easy Money! Except, there was a catch:
However, as the Philippines entity defaulted and the GSIS refused to pay on the surety bond, Bear Stearns sued AON based on the first CDS contract, which AON lost and had to eventually pay US$10 million to Bear Stearns. AON then went on to sue Societe Generale, and argued that the court's finding in the first action, that a "Credit Event" requiring payment had occurred under first CDS, mandated a similar result with respect to the second CDS.
... and Aon lost and Societe General did not have to pay, because there was slightly different wording in what credit events were covered. They sold protection against the GSIS refusing to pay, and bought protection
against a condition resulting from any act or failure to act by the Government of the Republic of the Philippines or any agency thereof that has the effect of causing a failure to honor any obligation issued by the government of the Republic of the Philippines.
... which, it seems the refusal of the GSIS to pay did not qualify for.
So what was supposed to be $100,000 easy money turned out to be a $10million loss.
And there, you have an example of the other thing amplifying the number of outstanding CDS ... laying off the risk of paying on the default by finding someone else willing to issue a CDS, who in turn may lay off the risk of some of their CDS liabilities, and so on.
And where it stops, in reality nobody really knows, because its a private over the counter dealership market.
What made CDS into Magic Money Machines are exactly what is Wrong With Them
This is from JPMorgan back in happier days (full disclosure: my checking account is at Chase), when they are peddling CDS to people engaged in international trade, who would otherwise be in the market for actual insurance:
The use of credit default swaps represented a new, less expensive, kind of risk mitigation for borrowers who could command very low spreads on borrowing, as well as pay low fees for the issuance of standby letters of credit issued for their account.
Despite the issues listed above, the market for credit default swaps continues to grow for these reasons:
- Standardised documentation leading to improved process flow and efficiency
- Dependably timely pay-out in case of major default
- Investor recognition of similarities between CDS solutions and more standard insurance and cash products
- Better use of capital and increased optimisation of balance sheets
- Diversification through investment that not only provides better yields, but may offer access to an otherwise inaccessible transaction. According to ISDA, at the end of 2004, the notional amount for the interest rate and currency derivatives market was US$183.6 trillion Credit derivatives comprised 4.4% of the entire derivatives market at US$8.42 trillion The players include insurers, reinsurers, financial guarantors and hedge funds; but the largest players are commercial banks, who are net buyers of protection
Great! Except ... suppose that event occurs at the same time as a large number of other credit events ... and because of the massive exposure, the credit events drive a lot of the issuers of the CDS over the brink and they fold.
Oops. As Reuters noted last month:
"This was supposedly a way to hedge risk," says Ellen Brown, the author of the book Web of Debt.
"I'm sure their predictive models were right as far as the risk of the things they were insuring against. But what they didn't factor in was the risk that the sellers of this protection wouldn't pay ... That's what we're seeing now."
The Problem is Systemic Risk
The problem here is systemic risk. This is by a quite different route than How a Little House Threatens Pension Funds and Insurance Companies ... but the upshot is the same.
You can diversify against "stochastic" (random) risk. Its a straightforward proposition, and the random events that happen all the time provide a lot of data to develop quite good statistical models of what is going on. These are risks like rolling a hundred fair dice and losing big on each "1" but winning a bit on each on that is "4, 5 or 6".
The odds of all 100 dice coming up 1? Really, really low.
But as long as the economy is an interconnected system, there are events that drive other events that drive other events in either a virtuous or a vicious circle. And when its a vicious circle, that's a systemic risk.
Like the collapse of the housing bubble, where the housing bubble led to lending on the assumption that people could always sell out if they could not pay, which then led to people who could neither pay nor sell out going into foreclosure, which undermined the housing market, and made it more difficult to lend, which undermined the housing market, and so on and so forth.
And Fannie Mae and Freddie Mac collapsed, sparking the biggest settlement of CDS to date.
We Know How to Protect Against Systemic Risk
We know how to protect against systemic risks in insurance. Insurance is always exposed to systemic risks, so we don't allow insurance to be bought by those without an insurable interest, and we require the firms selling insurance to manage their assets in a financially prudent manner.
The question is, how did we forget this with CDS, and allow anybody with a large enough bankroll to act like an insurance company, and anybody at all to take out insurance, even as a purely speculative play?
Well, because systemic risks are about the birds coming home to roost. Systemic risk exposure is not realized on a steady, ongoing, basis like stochastic risk. Instead, very little systemic risk is realized from one year to the next to the next ... then there is a recession and a financial arrangement proves able to withstand that level of systemic risk exposure ... and then there is little systemic risk realized for another stretch of years ...
... and you look around and find that there the exposure itself has become part of the vicious circle.
What can we do about it?
The original incentive for scratching into CDS was thinking through different alternatives to a "Bail-Out 2.0" when the current bail-out runs out of gas, sometime later this year or early next year. But digging into it leads to thinking about what can be done to avoid this kind of massive systemic risk exposure.
Over the long term, one thing the US could do would be to explicitly rule out enforcing CDS contracts unless the recipient of the protection holds the asset that is being insured. If the court does not enforce CDS that are not equivalent to a "real" insurance contract, then that will substantially reduce the appeal of taking out new CDS contracts.
In the short term ... well, now you know why there has been so much focus on buying up shaky assets. If the government waits until firms go belly up, then that is a "credit event" and triggers a cascade of CDS payments ... or, if the firms issuing the CDS cannot pay, more firms going belly up, who themselves may have bonds for which there are outstanding CDS that have to be paid, and so on.
On the other hand, if the government prevents the firm from going belly up, so they continue to pay their bonds, then that holds the lid on an explosion of CDS obligations.
Mind you, it should not be free of charge ... that is why I favor a system of 50:50 shaky assets and equity stake, with limits on corporate pay, regulation of mergers and acquisition, and no handing out profits to shareholder dividends until and unless the Preferred Dividend to the Public is paid in full.