What’s the instability risk of CDS markets?
Kevin Drum has a couple of good questions about credit default swaps, and the final link in his post literally made me laugh out loud, so I’ll do my best to answer him.
If the bond issuer does default, and there are a hundred speculators who own CDS protection on one of its bond, you’ve gone from, say, a $10 million event to a $1 billion event. Basically, when things go bad — and eventually they always do — widespread CDS protection can cause things to spiral far more out of control than they would otherwise.
And what’s the upside of allowing this? The argument I hear most often is that broad market trading of CDS provides an efficient price discovery mechanism for the underlying securities. Moody’s may rate that bond AA, but the CDS market will tell you what traders really think.
I guess I have two questions about that. First, does it really work? Are CDS marks really reliable indicators of creditworthiness? That’s debatable. Second, even if they are, is this a big enough benefit to make the instability risk worth it?
The first thing to do here is to dispute the premise. Yes, things will always go bad. But when things go bad, is it true that “widespread CDS protection can cause things to spiral far more out of control than they would otherwise”? Kevin is honest enough to note that when things went bad in 2008, that didn’t happen. And conceptually, it’s not the existence of credit protection which is a problem. People who have bought protection make money, they’re happy. The problem is the people on the other side of the trade — the people who wrote the CDS, not the people who bought them.
Indeed, that’s exactly what we saw in 2008: the people who bought lots of credit protection made lots of money, while the big losers were triple-A-rated companies like AIG and MBIA which had written the protection and which, by dint of being triple-A, didn’t have to put up collateral against the CDS which they sold.
Today, there isn’t a company in the world — not even Berkshire Hathaway — which can write CDS protection without having to put up collateral. And even if credit default swaps aren’t moved to an exchange, they will likely move towards a much more centralized clearing mechanism, which will institutionalize collateral and margin calls and make sudden bankruptcies even less likely than they are now.
But the most important thing to note is that the true villain here is not CDS so much as it is leverage. If I buy a bond and it goes to zero, I lose money, but there are few systemic consequences. If I’m a leveraged institution, however, and I bought that bond with borrowed money, then the consequences can be dire. Writing CDS protection is essentially the same as buying a bond, and writing CDS protection while putting no money down — as AIG did — is just as dangerous as buying a bond on zero margin.
So now ask yourself exactly what’s going on in Kevin’s hypothetical, where there are a hundred speculators who own CDS protection on a single credit. As we’ve seen, the speculators who bought protection don’t pose a systemic risk, although they might alter the likelihood of a bankruptcy filing. The real risk comes from the people who are net sellers of protection on that credit. And who might those be?
Think about what’s going on here, for a second: we have a lot of speculators making a negative-carry trade, where they pay regular coupon-like insurance premiums and get paid a lump sum in the event of default. Negative-carry trades are almost exclusively the province of hedge funds: long-only investors almost never make them. Indeed, it’s the long-only investors who are on the other side of the trade — institutional investors who wrote CDS rather than buying bonds, because the yield was higher or because they simply couldn’t locate bonds to buy, or just because CDS are more liquid. For those investors, CDS are bond substitutes, and they’ll buy a CDS for the full cash value of the amount being protected, just as they would a bond. They’re not leveraged.
The middlemen in the picture, the broker-dealers, are leveraged, but they are nearly always pretty flat when it comes to their net CDS positions. Insofar as there’s a big net position in the market, with one group of people net buyers of protection and another net sellers, the net sellers of protection — the ones posing the systemic risk — are unlikely to be leveraged, and if they’re not leveraged, nothing spirals out of control.
Sensible CDS regulation can and should, then, put tight limits on how much leverage a net seller of credit protection can be allowed to have. And by all means, if you want to keep things symmetrical, apply the same limits on how much leverage a bond investor can be allowed to have. But the CDS market is more dangerous, just because it’s so much bigger.
As to Kevin’s two questions, the first asks what the upside of the CDS market is, and the answer is simply liquidity — which is much more than just price discovery. The CDS market worked, during the crisis, even as the market in corporate bonds failed: bonds weren’t pricing, and CDS were. Selling bonds is hard and opaque — that’s why bond traders make so much money. Buying protection on bonds you own is cheaper and easier and much more transparent. And don’t even get me started on loans. People will invest in bonds if they believe they can exit their positions when they want. And the existence of CDS makes exiting a bond position much easier than it ever used to be — which in turn makes the bond market much more efficient. If you banned CDS, the price of credit would surely rise.
And Kevin’s second question is whether this benefit outweighs what he calls “the instability risk”. I think it does, just because the instability risk is hard to quantify and easy to regulate. The bellyaching of the Greeks notwithstanding, there’s no real evidence that the existence of the CDS market makes credit markets more volatile. Indeed, I believe that it can keep primary credit markets open when otherwise they would be closed. So let’s consider the upside of the CDS market before banning it for no good reason.
Update: Kevin, clever chap that he is, sees a flaw in this argument:
To a large extent, the superior liquidity and price discovery of the CDS market is attributable to its greater size, which in turn is attributable to leverage. If you tightly limit leverage, don’t you also constrain the size of the CDS market? And if you do that, will it still provide any noticeable liquidity benefits compared to the market in the underlying securities themselves?
I have to admit I was wondering this myself, as I left the office this evening. And really, there’s only one way to find out. But I suspect that CDS will always be significantly more liquid than the bond market, just because most bonds are buy-and-hold investments which are very difficult to find or sell on the open market. You will always be able to find a broker-dealer willing to quote you a price on a credit default swap, because they can simply conjure one out of thin air by writing a contract with you. If you want to find a specific corporate bond, however, that can be all but impossible most of the time, and in any case it might well be trading well above par, or have some other attribute which makes it much less attractive than a CDS of the same duration. People who are used to the stock market are always shocked at the illiquidity and opacity of the bond market: insofar as credit default swaps are a positive financial innovation, it’s because they managed to bring equity-like liquidity and transparency to a market in desperate need of them.