There has been a lot of talk about what to do with the credit default swap (CDS) market. Given that most people had not actively followed the somewhat arcane fixed-income markets until the recent credit crisis, it is surprising how much airplay CDS have received. Are these really the root of all evil? Or do they actually fulfill a valid function in market processes?
Most people have probably heard the analogy that credit default swaps are similar to an insurance policy -- and that may not be a bad way to think about them. If I own a car, I buy insurance, paying regular premiums for the right to "put" my car to the insurance company should I get into an accident. Similarly, a CDS contract is an agreement between two parties where one pays a premium for "protection" on corporate or sovereign debt; in the event that entity should default, the protection buyer is made whole on a par claim.
But simple explanations never fully capture the complexities, and these are derivatives after all. For instance, what constitutes a default? What is the basket of securities that are deliverable? Should it matter if I own or don't own the underlying securities, given that there is a defined set of deliverables?
Before we can have an active discussion, we need to establish a better understanding of the market and its evolution, and then we can pick apart how to handle the shortcomings while maintaining the positives. For what it is worth, I believe that the positives outweigh the negatives in the case of CDS, so this market is worth delving into so we can better understand how to address it.
To start, we need to better appreciate the fixed-income markets more broadly. For the purposes of this discussion, when I refer to the fixed-income markets, I am referring to the corporate debt markets, but a lot of the concepts apply across the fixed-income spectrum. Capital structure aside, the debt markets differ from the equities markets in one key point: how the markets themselves function.
Fixed-income markets do not have exchanges -- there is no order flow to route and match. They are principal-based businesses, meaning that a dealer is on the other side of each transaction acting as principal, either buying bonds from or selling bonds to an end-user.
So if Pimco or Fidelity calls up and asks for a bid on 50 million of XYZ bond, the dealer places a bid as to where it will take down that risk (I'm talking about what happens in "normal" markets...). There is no time to accumulate an order book; dealers use their judgment as to where they can take down the risk and trade (or hedge) their way out of it.
The reason why this is the case is that the debt markets themselves are very fragmented. When we trade equity, everyone knows where to go and find the price. Stocks trade on the NYSE or Nasdaq, and anyone knows where to seek price discovery. There may be slight differences for an odd-lot or a block trade, but they all hover around that observable price.
But if you think of the debt markets, companies issue various forms of debt -- fixed rate, floating rate, secured bank loans, unsecured bonds, subordinated bonds, first mortgage bonds, callable bonds, convertible bonds, maturities that range anywhere from overnight commercial paper to 30 years and any maturity in between (and beyond), issued in various currencies and formats -- euro bonds, DTC-eligible bonds, Samurai bonds, and so on and so on. One company can have dozens, even hundreds, of individual debt issues, yet they have typically only one stock ticker. If I ask, "Where is IBM (IBM) equity trading?," you can log on to the Internet and tell me to the penny. If I ask, "Where is IBM debt trading?," you have to ask me several follow-up questions.
One other critical element is that each debt issue of a company is limited in size. Companies typically only borrow what they need at a time, so you will have debt issuance of $100 million, $200 million, maybe up to $500 million or $1 billion. Only on rare occasions do you get debt issues larger than $1 billion.
Furthermore, since the debt markets are dominated by institutions, many institutions do not lend out their portfolios. Both of these factors combine to connote that the "repo" market for corporate bonds is fragmented, meaning they are hard to borrow in order to deliver into a short sale (particularly in contrast to the equity market, where the float is the float -- it is all fungible, and much of it available to borrow).
What often happens is that if Pimco or Fidelity comes to look for a bond, it ends up in a game of "go fish." If I do not happen to be long the particular issue they are looking for, I have to pass, because I cannot guarantee I can even borrow that bond in order to short it to them. For a lot of issues, they have simply gone to "bond heaven" -- meaning no one will ever see them again, because they are locked up in an insurance company or pension fund portfolio to match a liability stream and will never trade again.
So while bonds may be quoted with a bid and an offer, the offered side generally is subject to having the bonds in inventory. Historically, only the "on-the-run" (usually larger bond deals that were recently issued) had actionable two-way activity. In short, the corporate bond market has historically been an inventory-driven business.
One must also understand that a corporate bond carries risk beyond just the company's ability to repay -- there are other embedded risks (duration and optionality, to name two). Just because I may like Company XYZ, that does not mean that I necessarily want to take on 30 years' worth of interest rate exposure.
This also raises the point that investment-grade bonds typically trade on a yield spread to Treasuries, the risk-free rate. So the parties typically agree to the spread, then agree to the risk-free rate, to give the bond's yield in order to calculate the price. You need to recalculate a price based on that yield at any given time. This has added complexity if the bond is callable or prepayable.
And although I may be comfortable with Company XYZ for the next few years, that doesn't necessarily mean I want to take on XYZ credit exposure for 20 or 30 years. Supply does not always meet demand -- there may be demand for three-year XYZ paper, but only 20-year paper exists.
These issues just scrape the surface. Many factors go into a bond's price and trading, so price discovery on fixed-income instruments can be difficult.
Because of the myriad issues a company may have outstanding in the debt markets, and because of the myriad factors that go into pricing each piece of debt, it is virtually impossible to expect a market construct other than that of a principal market to develop. You need a market maker to step in and provide temporary liquidity and then redistribute that risk. That redistribution process can take weeks. That is the principal's risk; they hopefully make some bid-ask spread to compensate for this.
Enter the credit default swap market. The CDS market helps fill in some of the gaps in the fragmented debt markets. As mentioned at the beginning, the CDS market is focused on a company's credit risk. The premium exchanged strips out other ancillary variables, such as interest rate risk (to be sure, there is some interest rate risk in the pricing model, but nothing to the extent that you see in a fixed-rate corporate bond). CDS can be for terms unrelated to a company's actual debt maturity schedule. CDS are not dependant on being able to borrow a bond, although an active repo market helps establish arbitrage boundaries.
For example, XYZ could have a 10-year bond outstanding that trades at 220 basis points more than Treasuries. To price that bond, I need to know where the 10-year Treasury is, add the spread and then price the bond. But I may not want to take 10 years of XYZ exposure, or 10 years of interest rate risk. And a dealer might not have the bonds, even if I wanted them.
With CDS, I may be able to take on XYZ credit exposure for five years and get paid 150 basis points. This distills my trade to just focus on the credit risk of XYZ for five years. I have not taken on the concomitant interest rate exposure, and I have set the term of my exposure to only five years rather than 10. I get paid 150 basis points a year, and if XYZ defaults over that time frame, I need to pay the buyer par in exchange for the defaulted bonds. If I wanted to then take on interest rate exposure, I could always go buy Treasuries with the maturity of my choice.
When the first formal credit derivatives standards were put forth in 1998, CDS won out over the total return swap (TRS) market in terms of single-name risk transfer, because CDS were viewed as a "unifier" in these fragmented markets. As opposed to the TRS market, which is based on the total return of one bond, the CDS contract allows for multiple deliverables (similar to bond futures), so liquidity extends beyond any singular debt issue. One could deliver any bond or loan, any maturity, any G7 currency, of a given company to fulfill a contract. This created bridges across all of these markets and players, from the convert arb to the bank loan portfolio manager -- they could all meet up in the CDS market. Thus we could reach price discovery that truly had input from all players in the credit markets.
I can use CDS to quickly shed or take on exposure. I can customize my exposure. Say a company only has a 10-year and 30-year bond outstanding, but I only wish to take five years of exposure. With CDS I can do so. As a market maker, I will be more willing to take down a larger block of bonds if I know I can hedge the "jump to default" exposure, thus liquidity can be increased. I can better control my "spread duration" (duration is essentially the price reaction to a change in interest rates or in this case, spread). Furthermore, the advent of the CDS market has raised the level of sophistication in pricing corporate debt into components. Generally speaking, CDS help facilitate more complete corporate debt markets.
That isn't to say that the CDS market is free of issues, but I believe the issues can be managed in a way that does not call for draconian measures such as complete abolition or require their use for hedging purposes only. In the next section, we'll take a look at some of the issues surrounding CDS -- some valid, some not so valid -- and explore possible mitigants to the concerns.