David Einhorn, who enjoys his considerable reputation for hard-fought battles against firms with shaky finances and dubious accounting (Allied Capital and Lehman), has taken aim at a new and equally deserving target: credit default swaps.
In an interesting bit of synchronicity, Einhorn’s comments in a letter to investors overlap to a considerable degree with a post we wrote yesterday on why a clearinghouse for derivatives wasn’t a solution to the dangers posed by credit default swaps (and note the Orwellian branding, the reforms are about “derivatives” which include benign ones, names simple interest rate and currency swaps, yet the bill has loopholes that will let many, indeed probably most, credit default swaps escape).
Credit default swaps have no redeeming social value. They are a fee machine for Wall Street and their supposed value is considerably overstated (the world pre credit default swaps functioned perfectly well) and their costs, which are considerable, are not given the attention they warrant. And I don’t mean the failure of AIG, either.
Even though Einhorn gave a stinging, wide-ranging indictment, he missed one of the issues I find troubling, which is that credit default swaps result in information loss, which in turn lowers the quality of credit decisions. In other words, the product is inherently destructive.
In the world of old-fashioned fixed income investing, creditors would evaluate a borrower to make sure it had good odds of meeting its obligations. The lender could and usually did make inquiries about the borrower’s income, and its other commitments. If it was a business, the bank might also want to assess information that would help it evaluate the stability of the borrowers income (for instance, learning who its main customers were to determine how diverse and solid they were).
Just as with securitiztion, credit default swaps lower the incentive to do borrower due diligence. Why bother, when the CDS spreads on the reference entity tells you what the market thinks and you can use CDS to reduce or lay off the credit risk? But the original lender is in a privileged position; he is able to gather data from the borrower that it non-public and thus will not be incorporated in a market price. Thus giving creditors an incentive not to do that work systematically lower the quality of credit decisions.
But that reason is a bit abstract, although the costs are real. Einhorn focused on more tangible types of damage wrought by CDS, as summarized by the Financial Times. First, CDS are a means of extortion:
“I think that trying to make safer credit default swaps is like trying to make safer asbestos,” he writes in a recent letter to investors, adding that CDSs create “large, correlated and asymmetrical risks” having “scared the authorities into spending hundreds of billions of taxpayer money to prevent speculators who made bad bets from having to pay”.
Second, CDS speculators win if companies die. Given that the volume of CDS outstanding is a significant multiple of the amount of bonds outstanding, they are not used primarily for hedging, but for creating “synthetic” exposures. And those on the short side have compelling reasons to influence outcomes. When a company gets in trouble, the best outcome is often an out-of-court restructuring of debt before it gets even further in trouble. As much as the Chapter 11 process has certain advantages, it is also costly and risky. A CDS holder (one with a significant short position) can buy some bonds (now at a cheap price) of a struggling company to assure it has a seat at the table in negotiations so it can block a renegotiation of the debt and force a bankruptcy filing so it can assure its payoff on the CDS. From the Financial Times:
CDSs are “anti-social”, he goes on, because those who buy credit insurance often have an incentive to see companies fail. Rather than merely hedging their risks, they are actively hoping to profit from the demise of a target company. This strategy became prevalent in recent years and remains so, as holders of these so-called “basis packages” buy both the debt itself and protection on that debt through CDSs, meaning they receive compensation if the company defaults or restructures. These investors “have an incentive to use their position as bondholders to force bankruptcy, triggering payments on their CDS rather than negotiate out of court restructurings or covenant amendments with their creditors”
Einhorn also agrees with our contention, that a credit default swaps clearinghouse is not a viable solution. As we said yesterday in comments:
CDS are not economic if adequately margined. Adequate allowance for jump to default risk makes it very unattractive on a ROE basis. The way around that pre-crisis was making AIG and the monolines the bagholders. That game is over, but the Street is hooked on the revenues…..
….in invoking AIG, I am saying that an undercapitalized clearinghouse is a concentrated point of failure and a very big one too, a systemic risk all of its own.
“The reform proposal to create a CDS clearing house does nothing more than maintain private profits and socialised risk by moving the counterparty risk from the private sector to a newly created too big to fail entity,” he notes.
That’s because it is almost impossible to adequately capitalise against such developments. “There is no way a clearing house could demand enough collateral,” he says. “The market can be so big and discontinuous that it is very hard to figure out the correct amount of collateral.”
I think you need more people recognizing that CDS serve the interests of the financial sector at the expense of the real economy, and calling for the product to be banned. Only then might you see radical enough action taken.
However, as much as I hate CDS, I have reluctantly concluded that they cannot be taken out overnight. They have become sufficiently enmeshed in our financial infrastructure that eliminating them is like disarming a web of nuclear weapons. If you make a mistake on any one, they all go boom. One (and this is far from the only) problem is that the big banks not only have large CDS exposures, but they have other hedges related to them (such as interest rate swaps). So simply putting CDS into runoff mode could lead to dislocations in other markets.
I prefer regulating them very intrusively (like insurance, to make sure the counterparties are adequately capitalized), limiting new CDS writing to hedging existing positions (that would need to be tightly defined and monitored) and limiting CDS writing to end users (which would include proprietary trading desks) to where the investor had an insurable interest, as in owned the bonds, and only up to his exposure. That plus increasing capital requirement over, say, a three year period, to reflect the true default risk of the product should shrink the market enough to allow regulators to then ascertain whether it could then be put in runoff mode. But the intent of policy should be loud and clear: to strangle CDS, with the hope of killing them.
And for those who hope netting might do the trick, reader Richard Smith disabuses us of that notion:
Another point is about the struggle to keep up with ‘financial innovation’ in the OTC market. A problem for clients and regulators alike. CDS are probably the nastiest of these. They are so polymorphous – part of a basis trade, or a directional bet, or a sort-of-legit hedge, or a synthetic, depending on context; and no cap on speculation a la Gambling Act; and then vaguely like derivatives, or insurance, or short bond positions, or a prediction market.
But you couldn’t rule out the possibility that equally nasty new products could be developed by some smart aleck. Maybe there should be a charge on the inventors to cover the cost of regulatory catch up. Or something equivalent to airworthiness regulations, which even libertarians accept without demur, as far as I understand. That would slow the innovators down a bit – proving the ‘wings’ aren’t going to come off their new financial products and kill all the passengers.
Another observation I’d been meaning to make on ‘CDS trade compression’: the 20-40% that some commentators are so pleased about. I worked on an app like this for a large IB (recently unpopular in the guise of an mollusc) at the turn of the millennium. They had half a million daily NASDAQ trades at that time and their settlement IT guy in NY was freaking out as his mighty mainframe began to wilt under the volumes. Even with quite a conservative approach to compression (there are choices about how aggressively you net the trades – we thought we could get it down to 25,000 trades per day if we really went for it) we got 80% compression straight away, so, 100,000 netted trades per day. Of course those are highly standardized trades. The aggregation was something like stock, side, settlement date, counterparty, trade flags. NASDAQ is often characterized as an OTC market so it is really the product standardization that matters, rather than the nature of the venue perhaps. I think it went to 90% within a month or two as we got bolder but I may be confabulating; it’s a while ago.
If they can only get 40% trade compression out of CDS, after a year, there must be an awful lot of detritus left over (especially when IIRC most of the counterparties are TBTFs). So things like contract clauses, reference entity, duration of cover must be all over the place in what remains. Difficult to hedge or lay off I should think. And some unconfirmed trades too no doubt. A total mess.
Ignoring all the other shortcomings of CDS the natural thing would be to standardize the product:: that’s happened so many times before, but IBs hate standardization of course for the margin erosion it brings, and anyway now we get this cartel-like protection of the margins, under the guise of support for ‘finanical innovation’.
The implication is that what is on the banks’ books now is a bit hairier to manage than they are ‘fessing up. As other experts who similarly hate the product, like Satyajit Das have observed, simply banning new protection writing would probably lead to hugely disfunctional behavior prior to the date and also lead to problems (as in big time losses, which in a worst case scenario could result in another bailout) as positions that were in runoff mode would be essentially frozen and could not be managed.
But if we can get agreement on aims, which is the product should be killed, then it becomes possible to debate the best (least painful and costly) means.