Rethinking OTC Credit Derivatives

The effects that OTC credit derivatives have on the broader financial markets are poorly understood. These effects can perhaps best be described as what economists call externalities or, as they are referred to in the context of financial markets, spillover effects. While the term externality is usually associated with a negative effect on a third party, such as environmental pollution, both externalities and spillover effects can be used more generally to describe effects that are not the express subject of a transaction or group of transactions, without any subsequent value judgment. So, for example, a spillover effect of a robust derivative market could be price information generated in that derivative market “spilling over” into the corresponding asset market, resulting in more accurate pricing in the latter.

Externalities of any kind tend to be controversial, since the existence of externalities could be used as a basis for regulatory intervention. For example, price volatility in spot commodities markets has been pegged to supposedly excessive speculation in futures markets, which has been the lynch pin of the recent push for position limits.[1] Similarly, credit default swaps (CDS) have been blamed for a wide range of effects observed outside the CDS market, from crashing individual equity prices to systemic failures in the broader financial markets. These arguments, which are usually unsupported by any cogent theory or evidence, have been used to justify a ban on naked CDS positions (un-hedged short positions)[2] and even banning CDS altogether.[3] And while the messenger may be merely posturing or aggrandizing, these arguments have a frighteningly large audience.

In a modest effort to contribute to a more sober dialogue, what follows is a discussion and analysis of the legitimately new products and relationships that CDS have created along with a largely theoretical discussion of how these new products and relationships could be expected to spillover into and alter behavior in the broader credit markets, using concepts rooted in basic economics and common sense. Specifically, we consider whether CDS could divert capital that would otherwise be allocated to the bond market, and thereby, in certain cases, mitigate damage to overall economic welfare.

What Is New About Credit Derivatives

No Need To Own the Underlying

In a world without credit derivatives, it would be necessary to first locate and purchase a bond (or loan, but for simplicity's sake we will refer only to bonds) issued by a particular issuer to take a position on that issuer's credit. Even taking a short position requires locating a bond to first borrow and then sell short into the market. This implies that, in the absence of credit derivatives, the aggregate dollar value at risk at any time that is tied to a particular issuer's default cannot be greater than the total face value of the issuer's bonds then outstanding.

Credit derivatives remove that ceiling by allowing parties to identify a particular issuer or issuance of bonds without either of them owning any of the underlying bonds. The vast majority of credit derivatives are CDS[4], and so are composed of two transactional legs: the short leg, which pays a fixed quarterly premium and possibly an upfront amount; and a long leg, which makes a payment upon the occurrence of certain predefined "default" events. Those that have incorrectly characterized CDS as insurance policies have focused only on these two payment legs. This characterization also ignores the CDS market’s structure, which has dealers that supply liquidity and end users that take both long and short positions on credit risk. This is in contrast to the financial guaranty insurance market, which has end users that take only short positions (the policy holders) and insurers that take only long positions on credit risk.

Because CDS are swaps, there is much more to the transaction than the premium and default payment. As a matter of general market practice, collateral flows back and forth between the parties based on daily mark-to-market valuations of each party's respective position, with the out-of-the-money party posting collateral to the other. The value of each party's position is a function of how CDS spreads have moved since the contract was executed. As a result, CDS have a dynamic life between execution and termination, and therefore, have more in common with other swaps and forward contracts than insurance policies.[5] Interestingly, this also implies that the value of a CDS position is determined by both the underlying bond (upon the occurrence of a default) and CDS spreads, which creates the potential for a disconnect between the two near default known as jump to default risk.

Some have argued that because CDS do not require ownership of the underlying, there is no limit to the amount of risk that can be created through their use and that as such one spillover effect generated by the CDS market is greater systemic risk.[6] This argument assumes the existence of a derivative vending machine that generates products upon payment and ignores the reality of contracts, which require human decision makers to mutually agree to assume risk. Embracing the latter view, we note that the two parties to a derivative contract necessarily assume diametrically opposing risks.[7] As such, while CDS eliminate the requirement of locating and purchasing a bond, they still require one party to identify another willing to assume the other side of the trade. While it could be that there are more fools than bonds, the argument that CDS facilitate the creation of unlimited credit risk ignores the simple fact that there are two parties to every derivative contract, with each assuming opposing risks. So while CDS remove the ceiling imposed by the supply of bonds, the CDS market is still subject to that market's willingness to assume risk.

Scale and Composition Are Arbitrary

CDS are not bound to any multiple of the face value of a bond nor are the parties required to reference only one issuer. So long as the mutual interests of the parties are served, they can contract on whatever scale and reference any issuers they like. This feature facilitates the creation of a wide variety of CDS, such as basket CDS, n-th to default CDS, and the widely traded iTraxx credit indices. These products allow market participants to take highly structured positions, both long and short, on a basket of credits through a single trade. The creation of comparable products in the bond market would incur substantially greater transaction costs than that of their synthetic counterparts.

Reduced Funding

As a general matter, CDS are unfunded. That is, the long leg, the protection seller, does not post the face value of the contract, the notional amount, into an account for the benefit of the short leg, the protection buyer. While both parties must set aside some capital to cover potential collateral and payment obligations, the reduced funding requirement allows the protection seller to invest some portion of the notional amount elsewhere, earning a return. If that portion of the notional amount is posted or held in Treasuries, then the cash flows received by the protection seller will be roughly equivalent to the cash flows of the underlying bond. This is because the protection seller will receive both the risk free rate on the Treasuries and the premium from the short leg, which should be approximately the spread over the risk free rate that the underlying bond pays. But if the protection seller invests the portion of the notional amount available for investment in non-risk free instruments, he will be able to earn a return above the risk free rate on that portion in addition to earning the swap fee, which implies that the total return should be higher than that of the underlying bond. Thus, CDS allow for essentially unfunded exposure to credit risk, which can be used to facilitate a return that is higher than that of the underlying bond, albeit with a different risk profile.

Diverting Capital

CDS offer market participants an alternative to purchasing bonds. Assuming that the price of each comes to an equilibrium that excludes any opportunity for obvious arbitrage, we would expect at least some market participants to be indifferent between the two. Moreover, certain market participants might, and probably do, prefer the greater flexibility and liquidity of CDS to the control rights, security interests, and other benefits that traditional credit investments offer. This suggests capital that would otherwise be allocated to the bond market is being allocated to the CDS market. But unlike bonds, which are used to fund economic activity (e.g., purchasing equipment, hiring employees, etc.), CDS are for the most part unfunded, and in any case the funds allocated do not go to the issuer but to one of the counterparties. While the same is true of bond transactions in the secondary market, a portion of the capital that market participants are willing to allocate to credit risk is actually transferred to issuers in primary market transactions. Because this capital is then used by the issuer to purchase and use resources, it should be the case that transactions in the primary market have an impact on overall economic welfare. In contrast, CDS payments are made from one counterparty to the other where each counterparty is, in all likelihood, engaged in very similar activities. Payments of this sort should not affect overall economic welfare in the absence of a payment pushing a counterparty into insolvency. As such, CDS payments are similar to transactions in the secondary market in the sense that society should not care how payments and assets are distributed among a group of similar market participants.

Conclusion

By taking a long position through CDS a protection seller, who by definition desires credit exposure to the underlying bond, can achieve this exposure without increasing the amount of capital allocated to the issuer's economic activities. If it turns out that at least some market participants believe that a given issuer’s economic activities are overvalued by the bond market and that consequently the bond market has underpriced that issuer’s default risk, these market participants should be willing to take short positions on that issuer through CDS. Their willingness to enter into short positions creates liquidity for long positions, thereby offering an alternative to the issuer's bonds and potentially diverting capital away from the issuer's ostensibly overvalued economic activities and into a transaction that should have no effect on overall economic welfare. That said, if it turns out that the issuer’s activities are not overvalued, the same argument applies and so CDS could divert capital away from appropriately priced activities as well. However, as mentioned above, CDS are generally unfunded and so a protection seller that has taken a long position still has some portion of the notional amount available for investment. As a result, CDS absorb the demand for exposure to risk while requiring the use of only a fraction of the capital that would otherwise be required to satisfy that demand. As suggested above, this cuts two ways: if the issuer’s bonds are overvalued, then CDS satisfy the demand for exposure to that issuer’s credit risk without increasing the capital allocated to that issuer; but if the issuer’s bonds are priced appropriately, CDS absorb some fraction of the capital that would otherwise be allocated to that issuer. In the former case, the remaining capital is free to go to better use while the market’s demand for exposure to risk is satisfied in a manner that should have no effect on overall economic welfare, which is preferable to over allocating capital to the overvalued issuer. In the latter case, we have missed an opportunity to increase overall economic welfare by failing to allocate the capital that ended up being absorbed by the CDS market. In both of these cases, CDS have not decreased overall economic welfare, but either prevented it from decreasing needlessly or caused a failure to increase it .