Credit Default Swaps

The financial crisis over the past year has called attention to CDS and developments in CDS prices. Prices send signals about the credit stand ings of the reference entities and have therefore been used to predict which financial institutions might be particularly exposed to the credit crisis.

However, CDS have not only attracted attention due to interesting pricing patterns. There have also been, and still are, extensive concerns about the risks linked to the very widespread use of CDS, and whether the CDS market could amplify the financial turmoil.

Many observers see these concerns as being one of the primary reasons why the Federal Reserve earlier this year, in an unconventional step, helped to bail out the investment bank Bear Stearns. Bear Stearns is a major player in the CDS market, acting as counterparty in many CDS contracts. If it were to default, the problems would spread to many enterprises that have used CDS to hedge credit risk. This could cause an avalanche of problems.

This article provides a description of CDS. Which factors influence CDS prices? What can these prices tell us?

CREDIT DEFAULT SWAPS – CDS

As stated above, a CDS is an agreement between two parties to trade credit risk on a reference asset issued by a third party, the reference entity. The reference asset is typically a bond, but could also be a loan. Chart 2 illustrates the underlying principle of a CDS contract.

One party to the agreement purchases, for an agreed period, protec tion against a credit event, i.e. the reference entity's default or failure to meet its payment obligations. Most contracts run for 5 years.

The other party to the agreement sells protection in return for regular payments from the buyer. Such payments are received throughout the term of the contract, or until a credit event occurs. The annual payment per nominal value of the reference asset is known as the CDS spread. This is determined at the outset and depends on factors such as the per ceived credit standing of the reference entity.

If the reference entity defaults or fails to meet its payment obliga tions, the contract must be settled, i.e. the buyer receives the difference between the value of the asset and its nominal value. A number of credit events can trigger the settlement of the contract. In the following, these events are collectively referred to as default. The contract can be settled physically or in cash. Physical settlement entails a swap, whereby the protection buyer delivers the reference asset to the seller against payment of its nominal value. In cash settlement, the seller of protection pays the buyer the difference between the nominal value of the refer ence asset and its market value after the credit event.

A CDS bears considerable resemblance to a guarantee, but a significant difference in relation to conventional guarantees is that a fairly stand ardised and liquid CDS market has emerged. This enables trading of credit risk without access to the reference assets.

Growth in the market has to a large extent been driven by a set of Credit Derivative Definitions issued by the International Swaps and De rivatives Association (ISDA) in 2003. These include definitions of credit events and settlement methods.

At more or less the same time, i.e. in late 2003, the first CDS index con tracts were introduced. The introduction of indices also supported the development of a liquid CDS market.

Index CDS
CDS contracts are traded both as contracts on single reference entities and as contracts on multiple reference assets. Multiple-entity contracts can be CDS on a small group of reference entities (basket CDS) or CDS on standardised indices of reference entities.

There are several standardised CDS indices, each comprising 25-125 reference entities. Reference entities are grouped according to fixed guidelines, and the contracts operate with standard maturities and pay ment dates. Such CDS are generally more liquid than single-entity CDS contracts.

The two most important groups of standardised CDS indices are CDX and iTraxx. The two CDS indices with the highest turnover are thus CDX.NA.IG, comprising the 125 most frequently traded North American CDS reference entities with high credit standings, and the corresponding index of European reference entities, iTraxx Europe.

Use
CDS allow banks to sell the credit risk associated with their lending activities so that they can increase lending more than they would be able to if they bore the risk themselves. For example, a bank that has granted a loan to a business enterprise may sell off the credit risk on the loan by purchasing a CDS contract with the loan in question, or an equivalent loan, as the reference asset.

Likewise, credit risk on a bond can be hedged by purchasing a CDS with an equivalent bond as the reference asset. However, the CDS con tract covers only credit risk, i.e. the risk of losses if the issuer of the bond (the reference entity) defaults.

Credit risk can be reduced by buying CDS, but it can also be increased by selling CDS. CDS thus provide an enhanced opportunity to diversify credit portfolios and reduce the overall credit risk on the portfolio. In this way, a CDS can be a useful tool for risk management.

CDS have been developed primarily with a view to hedging and man aging credit risk. However, as is often the case with financial products, the use of CDS has been expanded so that many contracts are now used to take positions in credit risk or to construct other credit products.

A CDS is a very flexible instrument for taking credit risk positions. An investor who believes that the credit risk on a reference entity is in creasing, and that the CDS spread on that entity will therefore widen, should purchase protection by way of a CDS contract.

If the CDS spread in the market for the reference entity in question subsequently widens, the investor profits. The gain can be realised by selling the contract. Alternatively, an opposite contract can be con cluded by selling equivalent protection to another counterparty. Since the spread has widened, the investor – who is now selling protection – will regularly receive payments exceeding those being paid for the pro tection previously purchased. Selling protection involves taking on a credit risk on the reference asset, but this risk is hedged by the pro tection initially purchased.

Likewise, it is possible to speculate in a reduced credit risk, and thus a narrowing of the credit spread, by selling protection.

Today, CDS are also extensively used as building blocks for "structured products". CDS make it possible to design products that include or sell off credit risk on one or more reference entities – without directly in vesting in or selling the bonds of the reference entity.

When taking positions in credit risk, and when constructing other credit products, buyers and sellers of CDS contracts are not necessarily exposed to the reference asset beforehand. As a result, the notional value of outstanding CDS on a reference entity may exceed the refer ence debt considerably. This is indeed often the case.

The market
The market for CDS contracts has been growing rapidly, particularly since 2003. Annual growth in the notional value of outstanding CDS was in excess of 100 per cent from 2004 to 2006. In 2007, the notional value of CDS increased by 81 per cent, from 34,423 billion dollars to 62,173 billion dollars.

Just over half of all outstanding CDS in 2007 were single-entity con tracts. The share of multiple-entity contracts is, however, increasing.

The largest part by far of CDS turnover takes place between financial enterprises. According to BIS (2008), non-financial enterprises were counterparties to less than 2 per cent of outstanding CDS in December 2007.

In a global perspective, banks are net purchasers of protection, while insurance companies and hedge funds are net sellers. According to Fitch Ratings, the 10 largest players in the CDS market are all banks, and in 2006 these 10 banks were counterparties to 89 per cent of all traded CDS. This means that a major share of the CDS market is concentrated on a small group of very large counterparties.

In Denmark, CDS are mainly traded on the following reference en tities: Carlsberg, Danske Bank, DONG Energy, ISS, TDC and FIH Erhvervs bank.¹

WHAT DRIVES CDS SPREADS – AND WHY ARE THEY SO INTERESTING?

The price paid by a buyer of protection, the CDS spread, is intended to compensate the seller for the risk of loss taken on by concluding the contract.

As this risk depends on the probability that the reference entity will default, the spread is significantly affected by how great the market perceives this probability to be. If the market believes that it has in creased, the CDS spread widens.

The probability that the reference entity will default depends on a number of factors, some of which are entity-specific, while others con cern the market in general. The former include e.g. conditions relating to the assets and loans of the reference entity. If the value of the entity's assets is believed to be falling, this may increase the perceived probabil ity of default and thus the CDS spread.

Market factors affecting the probability of default include business cycles and liquidity conditions. Recessions often increase the probability that the reference entity will default, thereby widening the CDS spread. Likewise, liquidity shortfalls in the market and strained credit conditions will also increase the probability of default.

The compensation payable to the seller also depends on this party's willingness to take on additional credit risk. The CDS spread is therefore also influenced by general risk appetite in the market.

Finally, the price of a CDS also reflects the liquidity of the product. If a CDS contract is difficult to resell, compensation is paid to the seller by way of a liquidity premium. The less liquid the contract is, the greater the liquidity premium and thus also the CDS spread.

What information do CDS spreads contain?
As stated, CDS spreads depend on many factors. Nevertheless, a liquid CDS market may provide insight into market perceptions of both credit risk on individual reference entities and systemic risk.

Changes in CDS spreads across the board reflect risk appetite among market participants, cyclical conditions and other general market factors. A downturn leads to widening of spreads in all segments of the CDS market.

Such general widening of CDS spreads has been observed in both Europe and the USA over the past year, cf. Chart 3. The indices shown comprise reference entities from a number of different sectors, e.g. the energy, industry and financial sectors, reflecting the general market trend. The widening of CDS spreads over the past year reflects, among other things, how the financial crisis has reduced the risk appetite of market participants, as well as less favourable macroeconomic conditions.

On the other hand, fluctuations in the spread for a single reference entity, which do not mirror the general trend in the market, indicate that the market assesses the credit standing of the specific reference entity to have deteriorated (the spread widens) or improved (the spread narrows). In July 2007, the US investment bank Bear Stearns announced sizeable losses, and subsequently a pronounced widening of the Bear Stearns CDS spread was observed. As Chart 4 illustrates, the Bear Stearns spread did not follow the trend in the banking sector or the market in general. The market perception was that the credit stand ing of the bank had dropped as a consequence of the losses announced.

The correlation between spreads provides an indication of how the market perceives systemic risk. This is the risk that affects all reference entities in the market or within a given sector. Greater correlation between fluctuations in CDS spreads reflects the view that the probability of coinciding credit events, and thus the systemic risk, has increased.²

Market-based ratings
As an alternative to the traditional ratings used for credit assessment of counterparties, Fitch Ratings, among others, has begun to issue market-based ratings. These ratings are based on CDS spreads for the individual reference entities and measure the market assessment of the entities' credit standing.

Typically, traditional ratings are seldom revised. In contrast, CDS spreads on the most frequently traded reference entities can be observed on a daily basis. This means that the market-based ratings are in effect revised every day. Since CDS spreads rapidly respond to changes in the credit standings of reference entities, market-based ratings have often proved to be good at predicting subsequent changes in the traditional ratings. Chart 5 shows traditional and market-based ratings for a group of banks that have been officially downgraded within the last few months. In all cases, a decline in the market-based ratings has preceded the down grading.

It is also seen that the market-based ratings have been lower than the official ratings in the review period. The reason could be that CDS spreads, unlike traditional ratings, are cyclically sensitive. The unfavour able macroeconomic conditions at present entail wide CDS spreads and thus low market-based ratings. Traditional ratings, on the other hand, take resilience against cyclical fluctuations into account and therefore a general fall is not likely during a recession. This means that market-based ratings may very well be higher than traditional ratings during an upswing.

One of the reasons why declining market-based ratings are often followed by a fall in traditional ratings could, however, be that the pre dictions are to some extent self-fulfilling. If the market-based ratings go down, the reference entity may experience problems in the credit mar ket and in relation to customers, suppliers, etc., which could in turn con tribute to a downgrading of the traditional rating.

How useful are CDS spreads as measures of credit risk?
As seen, there can be advantages in using CDS spreads to assess credit risk, but the method also has its drawbacks.

In order to use CDS spreads for credit assessment of a counterparty, there must be a liquid market for CDS with the counterparty as the reference entity. That is not always the case. Although CDS contracts are often traded more frequently than the reference assets, the liquidity of CDS on a number of reference entities is nevertheless very limited. The lack of liquidity reduces the quality of the information derived from the CDS spreads. It is difficult to quantify the elements of credit risk and liquidity premium, respectively, in the spread.

Moreover, CDS are also traded for speculative purposes, which may in turn affect spreads. Speculation may thus lead to wider and more vola tile CDS spreads. Once again, the less liquid CDS are more vulnerable to speculation.

An advantage of CDS spreads is that they react very rapidly to news and market conditions that influence credit risk on the reference entity. However, this also means that CDS spreads can be sensitive to miscon ceptions of the credit standings of reference entities and to rumours that may be unfounded.

IMPLICATIONS FOR THE CREDIT MARKET

In principle, there is close a link between the CDS spread and the inter est on loans/yield on bonds issued by the reference entity.

For example, if a business enterprise issues a bond, the buyer of the bond may hedge the credit risk by purchasing a CDS with the bond issuer as the reference entity. The buyer of the bond then regularly receives the difference between the reference entity's borrowing rate and the CDS spread without having incurred any credit risk on the refer ence entity. However, the buyer does have a counterparty risk on the seller of the CDS. Appendix 1 includes a brief description of counterparty risk and other problems and risks related to the CDS market. Alterna tively, the buyer of the bond could buy a risk-free asset such as a govern ment bond and receive a risk-free return. As Chart 6 illustrates, the two strategies outlined ensure that the buyer of the bond receives regular payments without incurring the associated credit risk, and in principle the size of the two payment flows should therefore be the same.

In other words, the yield on the reference entity's bond should be equal to the sum of the risk-free return and the CDS spread on the refer ence entity. The wider the CDS spread, the higher the borrowing costs for the reference entity.

All the same, this link does not quite hold true in practice due to a number of factors. Tax issues may play a role, or the liquidity of the CDS contract may differ from that of the reference entity's debt instruments. In many cases, CDS contracts are more liquid than the reference assets, which makes the CDS spread a purer measure of credit risk.

The difference in liquidity is indeed one of the reasons why the CDS spread, and not the price of the reference asset, is generally used for assessing the credit standing of e.g. a bank. In many cases, CDS spreads also react more rapidly to changes in the credit standing of the refer ence entity.

For a business enterprise that wishes to issue bonds or raise loans, it can thus be an advantage to be a reference entity on CDS traded with narrow spreads. Conversely, wide CDS spreads reduce the borrowing op portunities of the reference entity.

Widening of the CDS spread for a reference entity does not necessarily reflect actual deterioration of the reference entity's credit standing. Speculation, rumours or misconceptions may also be underlying factors. Nevertheless, the widening of the spread could make it more difficult and costly for the reference entity to borrow. Ultimately this could mean that faulty assessments or rumours may cause real problems for a reference entity.

CDS spreads may fluctuate considerably. Chart 7 shows the develop ment in the CDS spreads of the three large Icelandic banks Kaupthing Banki, Glitnir Banki and Landsbanki. The CDS spreads for these three banks have widened considerably.

So far, these banks have not had any great need to raise new loans, but the terms of the loans already raised have been less stringent in practice than the spreads would indicate. The banks have, for instance, raised loans via private placements, which could reflect that the CDS market is fairly thin in comparison with the actual lending market.

Besides the impact of CDS spreads on the borrowing costs of the individual reference entity, the CDS market may also influence the credit market in general. The CDS market has made it easier and less expensive to hedge credit risk and diversify credit portfolios. In principle, this should mean that credit risk on a loan is passed on to the party best able to assume the risk. In addition, CDS provide easily accessible information about the credit standing of reference entities. This could mean less ex pensive loans for references entities on relatively liquid CDS.³

However, reduced liquidity in the CDS market may have an impact on the lending market, which could in turn affect financial stability, cf. Appendix 2. If trading in credit risk becomes less efficient, it may be come more expensive and more difficult to borrow and raise liquidity.

LITERATURE

Ashcraft, Adam B. and João A. C. Santos (2007), Has the Credit Default Swap Market Lowered the Cost of Corporate Debt?, Federal Reserve Bank of New York, Staff Report, no. 290.

BIS (2007), New developments in clearing and settlement arrangements for OTC derivatives, March.

BIS (2008), OTC derivatives market activity in the second half of 2007, May.

Fitch Ratings (2007), CDx Survey – Markets Volumes Continue growing while New Concerns Emerge, July.

Mengle, David (2007), Credit Derivatives: An Overview, 2007 Financial Market Conference, Federal Reserve Bank of Atlanta, May.

Risk Magazine (2008), JP Morgan CDS may top $10trn notional, April.

Tarashev, Nikola and Haibin Zhu (2008), Market perception of systemic risk in the banking industry, BIS, Quarterly Review, March.

APPENDIX 1: PROBLEMS AND RISKS

For the individual market participants, benefits may be gained by trad ing credit risk in order to achieve the desired credit risk profile, but the CDS market also entails a number of problems and risks.

Counterparty risk
The buyer of protection passes on the credit risk on the reference asset, but instead incurs a credit risk on the seller. There is a risk that the latter is unable to pay if the reference entity defaults. The counterparty risk is also relevant to the buyer of protection if the CDS purchased has a posi tive market value, since the market value falls if the seller's credit stand ing deteriorates. Counterparty risk may also affect the CDS spread. The buyer of protection may be willing to accept a higher spread if the seller is assessed to have a high credit standing.

The seller of the guarantee assumes the credit risk traded, as well as a counterparty risk in relation to the regular payments from the buyer. Both parties would thus find themselves in a rather awkward position were the counterparty to default.

Another issue relating to counterparty risk is that the contract may be passed on to a third party by either the buyer or the seller. This can lead to uncertainty about the counterparty risk, as the identity of the actual counterparty is not clear.

Legal and operational risks
Legal risk arises from factors such as lack of documentation and stand ards. There have been problems with lack of clarity in relation to refer ence assets and the absence of definitions of the events that trigger settlement of the contract. This could mean that buyers of guarantees are in practice not obtaining the insurance that they believe.

These issues have to some extent been addressed by the ISDA Credit Derivative Definitions, as they reduce the uncertainties concerning the content of the contracts. However, not all CDS contracts are based on the ISDA Definitions, and the Definitions have not been seriously tested in court.

An operational issue in relation to the CDS market is the potentially large backlog of contract registrations. Again, extensive retrading ex acerbates the problem. Since late 2005, the Federal Reserve, among others, has had strong focus on eliminating this backlog – with consid erable success.

There are also operational issues concerning settlement of contracts, primarily because the notional value of outstanding CDS on a reference entity often by far exceeds the reference debt. Since physical settlement requires the buyer to deliver the reference asset to the seller, problems may arise because many buyers of protection will attempt to procure the asset if the reference entity defaults.

The first step towards an infrastructure to address the settlement issues has now been taken, in that the US Depository Trust & Clearing Corporation, DTCC, has entered into cooperation with CLS Bank⁴. In No vember 2007, CLS began to settle CDS registered at DTCC in US dollars, euro, Japanese yen, pounds sterling and Swiss francs.

In connection with cash settlement, it may be a problem to determine the market value of the reference asset. ISDA has, however, begun to conduct auctions to determine the market value of the reference entity's debt. No standards have as yet been developed for such auctions.

Incentive issues
The CDS market has contributed to a shift in the business models of some institutions. Previously, they took on most of the credit risk on lending themselves, but now they increasingly act as lending intermedi aries, i.e. they do not assume the credit risk.

This development can create incentive problems. It can be problematic that the credit assessment performed by the original issuer of the loan is less thorough than previously. In addition, the original lender does not suffer a loss if the reference entity defaults and therefore does not have the same incentive to find a solution to prevent this. On the contrary, the buyer of protection may have an interest in the default of the refer ence entity, since this could trigger contract settlement.

APPENDIX 2: SYSTEMIC RISK AND FINANCIAL STABILITY

CDS have become a frequently traded product. There are many players in the market, and extensive retrading creates long chains of counterparties. This tangle of counterparties in the CDS market can entail systemic risk.

Any difficulties experienced by any one active participant in the CDS market may spread to a wider range of counterparties. This can cause instability in the CDS market, which may in turn spill over into other financial markets. The tangle of counterparties could mean that it will take some time to realise how the default of one market participant will affect the rest. Consequently, the market may for a period become very illiquid, causing considerable widening of CDS spreads. As previously de scribed, this may affect the credit market, thereby causing the problems to spread.

The bailout of Bear Stearns indicates that the Federal Reserve is very much aware of the potential systemic risk in relation to CDS. Bear Stearns was one of the large players in the CDS market. According to an article in Risk Magazine, it is estimated that Bear Stearns had outstanding CDS contracts totalling around 2,000 billion dollars.⁵ Consequently many players would have experienced problems if Bear Stearns had defaulted.

The need to improve e.g. the registration of trades, handling of counterparty risk and the procedure for settlement of contracts was recently highlighted in a report by the Counterparty Risk Management Policy Group, which comprises representatives from a number of market participants and is chaired by Gerald Corrigan, former president of the Federal Reserve Bank of New York. The Policy Group emphasises the im portance of cooperation between market participants with a view to im plementing improvements. Regulatory authorities have also focused on containing systemic risk in the CDS market. For example, the estab lishment of a central counterparty – a clearing house – for all trades would reduce counterparty risk and facilitate the disentanglement of counterparties in the event that a reference entity or market participant defaults. Recently, players in the CDS market have themselves taken the initial steps to establish such a clearing house.

It seems like a paradox that CDS can cause so much concern. After all, a CDS allows efficient hedging of credit risk and diversification of credit portfolios, which should contribute to financial stability. However, finan cial innovations such as CDS are not only used to reduce risks. Further more, the rapid growth of the CDS market has led to operational and legal issues in the market. Consequently, the CDS market may have a negative impact on financial stability unless these issues are addressed in an efficient manner.