Upload
sandeepagarwala
View
220
Download
0
Embed Size (px)
Citation preview
8/14/2019 11 September 2002
1/17
11 September 2002
IntroductionThe credit derivatives market has grown significantly in recent years, revolutionising
the trading of credit risk. Close to zero in the mid-nineties, the British Bankers
Association estimated the outstanding notional at just under US$1 trillion in 2000 and
expects it to top the US$1.5 trillion mark by the end of 2002. The (modified) 1999 ISDA
documentation now provides a legal framework and many of the problems of the early
days (such as the definition of credit events) have largely been ironed out. Recent
developments in the US indicate that the market is trending towards a two-credit-
event market there (thus more closely resembling the corporate bond market). This
should attract more participants and increase liquidity, helping the market to mature.
Credit derivatives facilitate the transfer and repackaging of credit risk. They have de-
linked credit risk from funding and separated the ownership of an asset from the
ownership of the risk. As such, they allow market participants to buy credit risk for names
not available in the cash market, and the number of actively traded names has increased
significantly. Why is this a very important development?
The risk attached to a credit includes the deterioration of spread and/or rating (not
necessarily going hand in hand, as witnessed over recent years), as well as default
and recovery (defining the ultimate loss). Within a portfolio, credits show varyingdegrees of correlation. Credit derivatives allow protection buyers to hedge these risks,
increase diversity and bring down concentrations in an efficient and flexible way. At the
same time, they offer protection sellers the opportunity to add credit exposure
according to their risk appetite. There may also be more liquidity than the cash market.
Risks include credit event definitions, spread volatility and counterparty risk.
Who would be interested in trading credit risk? Historically, banks have extended loans
to corporates, other banks and sovereigns, thus building their credit portfolios. In
addition, the bond markets have grown significantly over the past few years. The 1988
Basle Capital Accord requires banks to allocate a certain amount of capital for credit
exposure to different obligors, not taking into account credit ratings or correlation. As aresult, banks started to take advantage of the arbitrage possibilities regulatory capital
requirements created. The market has come a long way since. Banks now increasingly
manage their economic capital and the revised Accord (Basle 2), which is due to be
implemented by the end of 2006, will adopt a refined approach towards credit risk. In
addition, new protection sellers like insurers, reinsurers, fund managers and, as of late,
hedge funds are entering the credit markets.
Credit default swaps have dominated the market as the building blocks for most credit
derivative structures. A single-name credit default swap allows an investor to replicate a
cash position without using any funding. However, the scope of credit derivatives goes
far beyond this. There is a wide range of innovative applications, three of which the
credit-linked note, the synthetic CDO and the first-to-default basket we shall discuss.
Credit derivatives facilitate the
trading of credit risk
They have de-linked credit risk
from funding and separated the
ownership of an asset from its
risk
Credit risk encompasses
deterioration, default and
ultimate loss
Banks dominate the market
with increasing interest from
(re-)insurers, fund managers
and hedge funds
Credit default swaps dominate
as the building blocks for most
structures
8/14/2019 11 September 2002
2/17
The mechanics
What is a credit default swap?A credit default swap (CDS) is a bilateral contract in which one party (the protection
buyer or the seller of the credit risk) pays a periodic, fixed premium to another (the
protection seller or buyer of the credit risk) for protection related to credit events on
an underlying reference entity or obligation. Usually the premium is paid quarterly
and expressed in basis points per annum of the swaps notional. If a credit event
occurs, the protection seller is obliged to make a payment to the protection buyer in
order to compensate him for any losses that he might otherwise incur. Thus the credit
risk of the reference entity or obligation is transferred from the protection buyer to the
protection seller. A CDS is therefore similar to an insurance contract or a financial
guarantee i.e. of unfunded nature. Alternatively it can be thought of as an unfunded
FRN (which also largely excludes the interest rate risk), with the premium similar to the
credit spread over Libor.
A credit default swap
Source: DrKW
Determining the premium and the compensation payment both require a valuation of
credit risk. Credit risk is defined as the product of the likelihood of default (P d) and the
actual loss (L), which in turn depends on the recovery (R).
C = Pd* (1-R)*100 where L = 1-R
The credit spread is a function of P, R and maturity assuming a risk-free interest rate
(e.g. Libor). In addition, liquidity, regulatory capital requirements as well as the market
sentiment and perceived volatility and shape of curve are usually priced in. As a result,
in order to price a CDS, we need to know the credit, its default probability and recovery
rates, which depend on the level of seniority of the debt, plus market information. The
rating agencies, and in particular Moodys, provide a long history of statistical
information on one-year and cumulative default probabilities, as well as recovery rates
for different products.
Therefore, the counterparties in a CDS first need to agree on three broad issues; (i)
what the reference entity or obligation(s) is, (ii) what the credit events will be and (iii)
how and when the payment will be settled. Most CDS contracts1 are based on 1999
International Swaps and Derivatives Association (ISDA) standard documentation.
One party provides credit
protection to another according
to pre-defined credit events
Determining the premium and
protection payment require a
valuation of credit risk
CDS counterparties need to
specify three broad issues in
the documentation
Protection buyer
Premium (bp)
Reference entityor obligation
Protection seller
Contingent payment
Protection buyer
Premium (bp)
Reference entityor obligation
Protection seller
Contingent payment
1 It is estimated over 90%.
8/14/2019 11 September 2002
3/17
Reference entity or obligation
As no particular asset is sold, a reference entity upon which protection is being writtenneeds to be identified. It may be a corporate, financial institution or a sovereign. A CDS
may reference a specific obligation of this entity a reference obligation and if this is
the case credit events would be linked to this obligation only. But more commonly
protection will be written on the reference entity without specifying a particular
obligation. Therefore the seniority and type of obligation that will trigger a credit event
will be identified in the documentation. Typically a CDS will reference an issuers
senior unsecured debt, and credit events could be linked to bonds or loans or to the
broader category of borrowed money.
Definitions of obligations
Category Characteristics
- Payment - Pari passu ranking
- Borrowed money - Specified currency
- Bonds - Not sovereign lender
- Loans - Not domestic currency
- Bonds or loans - Not domestic law
- Reference obligation only - Listed
- Not contingent
- Not domestic issuanceSource: International Swaps and Derivatives Association
SettlementThere are two methods of settlement for a CDS physical or cash.
In a physically settled CDS, upon the occurrence of a credit event the protectionbuyer would deliver an obligation to the protection seller and receive par in return.
If credit events are linked to a reference obligation, the scope of deliverable
obligations usually extends to include all obligations that rank pari-passu with the
reference obligation. The protection seller will be delivered the obligation that is
cheapest to deliver.
In a cash settled transaction, the protection seller would pay the difference
between market value and par (instead of having an obligation delivered) after a
specified period of time (e.g. 30 or 60 days). The protection buyer would obtain the
market price either from a calculation agent or via dealer poll. Again, if the market
value of a reference obligation cannot be obtained for whatever reason, then a
pari-passu ranking obligation with a similar maturity would be valued instead.
Settlement of a credit default swap
Source: DrKW
Which obligation(s) credit eventswill be linked to need to be defined
Two methods of settlement
Cash settlement
Protection buyerObligation
Protection seller
Par
Protection buyer Protection sellerPar Market value
Physical settlement
Cash settlement
Protection buyerObligation
Protection seller
Par
Protection buyer Protection sellerPar Market value
Physical settlement
8/14/2019 11 September 2002
4/17
While the recovery value is usually the unknown variable, there are transactions in
which a fixed, pre-defined cash settlement amount (fixed recovery) is agreed.
The result is that the protection seller bears the loss on the obligation, but this loss
may differ according to the method of settlement. The protection seller may prefer
physical delivery if he thinks that by waiting and allowing the price of the obligation to
stabilise, he may realise a higher recovery. Intuitively one might think physical
settlement, or indeed a longer cash settlement period (i.e. 60 days rather than 30),
would result in a higher recovery as it gives the price of the credit more time to settle
down. However, recent history paints a cloudier picture, as this was not the case with
Enron and WorldCom, for example.
Credit events
The 1999 ISDA Credit Derivatives Definitions lists six credit events2.
Bankruptcy
Failure to pay
Restructuring
Obligation acceleration
Obligation default
Repudiation/moratorium
The first five relate to corporate obligations, whereas repudiation/moratorium is more
applicable to sovereigns. A CDS contract must define at least one credit event, but
usually it will include more. In recent months, obligation acceleration and default have
been dropped from standard CDS documentation, and most contracts are now based
on bankruptcy, failure to pay and restructuring. However, as we discuss recent
developments indicate that the market is moving towards even narrower credit event
definitions, with moves afoot, primarily in the US, to drop restructuring as a credit event
altogether. There are some obstacles to this, but at some point in the future it is
conceivable that protection may be written on bankruptcy and failure to pay only.
Bankruptcy
Generally this refers to a reference entity becoming insolvent or being unable to pay its
debts. The definition also includes a provision whereby any action taken by a reference
entity in furtherance3 of a bankruptcy would also constitute a credit event e.g. if a
company was known to be planning, or even just considering, filing for bankruptcy,
even if it had not yet done so.
Failure to pay
After the expiration of any applicable grace period, a reference entity fails to make
payments when due under one or more of its obligations. A minimum threshold (the
Payment Amount) needs to be specified and may vary depending on what obligations
the CDS is referencing.
The loss may differ according
to the method of settlement
Six credit events listed by ISDA
Bankruptcy, failure to pay andrestructuring have become the
standard credit events
but for how long?
2 The procedure for notifying the protection seller of a credit event is detailed in the documentation. Its occurrenceneeds to be confirmed by public sources e.g. the news wires.3 See 1999 ISDA Credit Derivatives Definitions, ISDA, for more details.
8/14/2019 11 September 2002
5/17
Restructuring
This has been the most controversial credit event, and it is most relevant in the contextof hedging loans rather than bonds. According to the 1999 ISDA Definitions,
restructuring is a change in the terms of an obligation due to a deterioration in credit
quality of a reference entity that leaves creditors materially worse-off. The five
restructuring events listed are a reduction in the rate or amount of principal payable, a
reduction in the amount of principal or premium payable at maturity or at scheduled
redemption dates, a postponement or deferral of interest or principal payments, a
change in the ranking of an obligation causing it to be subordinated, or a change in the
currency or composition of any interest or principal payments.
In August 2000, Conseco restructured its maturing bank loans in order to forestall an
impending liquidity crisis. The maturity of a portion of the loans was extended, and
even though creditors were compensated through an increased coupon, a new
corporate guarantee and additional covenants, this restructuring triggered a credit
event under the ISDA Definitions. This caused uproar and sparked considerable
discussion about the merits of restructuring as a credit event, and the market went
through an uncertain period during which liquidity declined.
There were two key points that were up for debate in the aftermath of Conseco-gate
and at which protection sellers felt particularly aggrieved. Firstly, protection sellers
argued that banks (protection buyers) were not materially worse off as a result of the
restructuring and were compensated for the maturity of Consecos debt being
extended. Secondly, as protection buyers were able to deliver any obligation ranking
pari-passu to or higher than the reference obligation, they delivered the cheapest-to-
deliver bond. Some of Consecos bonds were trading at distressed levels, therefore
protection sellers incurred losses that they arguably would not have done if they were
long the reference obligation (i.e. the bank loan) itself.
As a response to the discussion and in order to address the concerns of market
participants, ISDA produced some supplemental definitions in May 20014 that have
been termed modified restructuring. The new definitions include a Restructuring
Maturity Limitation, which limits the maturity of deliverable obligations to 30 months
and thereby prevents protection buyers from delivering long-dated obligations that
might be trading at a significant discount. To meet the Pari-Passu Ranking criterion,
an obligation must have the ranking in the priority of payment that the reference
obligation had as of the later of the trade date or its issue date. This is to ensure that if
a reference obligation is subordinated in a restructuring, protection sellers cannot
deliver a subordinated obligation.
These new definitions have largely addressed the uncertainty that arose from
Conseco, and market participants in the US and Asia have incorporated modified
restructuring into most CDS contracts. In Europe, the transition has not yet been made
and old (i.e. unmodified) restructuring persists for now. However, a European version
of modified restructuring is currently being negotiated and is expected to be ready by
the end of the year.
The most controversial andfrequently discussed credit
event
The Conseco restructuring in
August 2000 had a significant
impact on the market
It sparked considerable debate
about the ISDA 1999 definition
of restructuring
As a response, ISDA produced
some supplemental definitions
modified restructuring
Modified restructuring has
been incorporated in the US
and Asia, but not yet in Europe
4 See Restructuring Supplement to the 1999 ISDA Credit Derivatives Definitions, ISDA, 11 May 2001.
8/14/2019 11 September 2002
6/17
Despite the new definitions, controversy and uncertainty still surround restructuring as
a credit event. A credit event was called on Xerox in June this year when the companyextended the maturity of its bank loans, much to the dismay of protection sellers. There
is little doubt that Marconis restructuring, announced last month, in which it will swap
over 4bn of debt for equity, cash and some new debt should justifiably trigger a credit
event, however credit events have not yet been called. There is a debate over whether
CDS contracts have been legally triggered as the process has not yet reached court,
and there is also confusion over whether the relevant credit event is restructuring or
the companys effective bankruptcy. Even though the restructuring has not reached
court, the mere announcement of it should be sufficient to trigger payment. Credit
events could be called imminently, and the market is waiting for somebody to set a
precedent.
Reflecting the problems and uncertainty that still exist, J.P. Morgan, one of the leading
players in the credit derivatives market, has recently stated that it will drop
restructuring from the CDS contracts that it uses to hedge its own portfolio. The reason
cited is the growing concern amongst insurance companies and other protection
sellers over banks benefiting unduly from restructuring. We understand that other US
banks are moving in a similar direction, and the anti-restructuring movement is
gathering momentum5. Because the bond markets are more developed in the US, the
CDS market is based primarily on bonds and therefore dropping restructuring as a
credit event is a logical step in the US.
Outside the US, regulation is major obstacle. In Europe for instance, banks argue that
regulators would not grant them capital relief if contracts excluded restructuring,
although the Bank of International Settlements is currently being lobbied to accept
two-credit-event capital relief. The CDS market in Europe remains very much bank-
loan driven, therefore retaining restructuring makes more sense than in the US. But as
the bond markets continue to develop, moving towards two credit events would be a
natural progression. A two-credit-event CDS market i.e. where contracts would be
based on bankruptcy and failure to pay would certainly bode well for liquidity, as it should
attract more participants. That said, we may see two separate markets develop one for
bonds (without restructuring) and one for loans (with restructuring).
Obligation acceleration or default
The two are very similar, but they are no longer common in most contracts. Obligation
acceleration means that one or more obligations have become due and payable earlier
than they would otherwise have been due to the default (or similar condition) of a
reference entity. As with failure to pay, this is subject to a minimum threshold amount
(the Default Requirement) that needs to be specified in the documentation. The only
difference with obligation default is that one or more obligations have become capable
of being declared due and payable, even though they may have not yet been.
Repudiation/moratorium
A reference or government entity challenges or rejects the validity of one or more of its
obligations, or imposes a standstill or deferral on payment. Again this has to be for an
amount exceeding the Default Requirement.
Controversy surrounding
restructuring still exists
Could we see CDS contracts
exclude restructuring in the
future?
Outside the US, regulation is a
major obstacle
These are very similar, but have
recently been omitted from
most CDS transactions
Typically relates to sovereign
obligations
5 A tale of two restructurings, Creditflux, 5 September 2002
8/14/2019 11 September 2002
7/17
The credit derivatives market
Volume and compositionThe growth of the credit derivatives market has been exceptional, and in its 1999/2000
Credit Derivatives Survey the British Bankers Association (BBA) estimated that
outstanding volume would increase almost tenfold from 1997 to 2002. As this is an
over-the-counter market, it is difficult to gauge its exact size. ISDA puts the outstanding
volume of CDS transactions at US$918.9bn at end-20016. Their data includes single
name CDS, basket and portfolio trades, but assuming that the composition of the
market has not changed significantly, then the total outstanding volume of credit
derivatives should currently be well in excess of US$1 trillion. As the graph below
shows, the size of the credit derivatives market is also increasing vis--vis the
corporate bond market. In 1997, credit derivatives represented less than 10% of
outstanding US corporate bond issuance, but by the end of 2002 we expect this figure
to be somewhere around 40%.
The credit derivatives vs. the corporate bond market: outstanding volume (US$ bn)
* Excluding asset swaps
** The credit derivatives figure is a BBA estimate, while the corporate bonds figure is an estimate as of Q1 2002
Source: BBA 1999/2000 Credit Derivatives Survey, Bond Market Association
Single name CDS are the largest component of the market, but their share has fallen
from 52% in 1997 (2002 estimate: 37%) as the market has developed and become more
diverse. Increasingly, however, single name CDS are becoming the building blocks of
more complex structured products, particularly CDOs as banks and asset managers
seek to exploit arbitrage opportunities synthetically rather than via the cash market.
Portfolio/CDO products account for the second largest portion, and we have seen a
considerable number of deals referencing portfolios, both static and managed, of single
name CDS. This has been facilitated by the improved liquidity and deepening of the CDS
market, which itself is the result of the standardisation of documentation and greater use
of the market as understanding and technical expertise have increased.
In terms of underlying, the focus has shifted from sovereigns to corporates, which now
account for the largest share of the market as the graph on the following page shows.
The credit derivatives market
has shown exceptional growth
over the past five years
Single name CDS are the
largest component of the
market
Corporates dominate in terms
of underlying
6 See 2001 Year End Survey, www.isda.org
0
500
1,000
1,500
2,000
2,500
3,000
3,500
4,000
4,500
1997 1998 1999 2000 2002E**
Credit derivatives*
US corporate bonds
8/14/2019 11 September 2002
8/17
The global credit derivatives market
Source: BBA 1999/2000 Credit Derivatives Survey
Market participantsThe credit derivatives market remains bank-dominated. Banks are natural lenders and
should always have a demand to hedge credit risk. As protection buyers, they have
been able to hedge certain risk positions in their portfolios and thus reduce their
regulatory (or economic) capital requirements. This has also enabled them to reduce
concentration risk to particular industries or countries and fill unutilised credit lines. In
Europe in particular, banks have transferred the risk of large portfolios through
synthetic CDOs. This has become increasingly interesting in light of the need to
enhance shareholder value and return on equity. However, as the table below shows,
banks share of the market is estimated to have fallen since 1999 as insurance
companies and corporates in particular have become more active.
Given the similarity of a CDS to a financial guarantee, it is natural that insurance
companies (especially monolines) have been far more active sellers than buyers of
protection. They have played an important part in the CDO market by taking the high
quality super senior CDS positions in (largely unfunded) synthetic deals (see page 16),
as well as equity pieces which has allowed them to gain leveraged exposure to a
portfolio of credits. Similar to banks, they have also used credit derivatives to reduce
risk concentrations e.g. to hedge country risks. Securities firms maintain an important
role as market makers, reflected in an almost equal proportion of protection buying and
selling. Although their overall market share is still small, fund managers, particularly
hedge funds, are becoming increasingly important players.
Credit derivatives market participants
End-1999 actual End-2002 estimate
Protection Buyers Protection Sellers Protection Buyers Protection Sellers
Banks 63% 47% 51% 38%
Securities firms 18% 15% 15% 16%
Corporates 6% 3% 10% 5%
Insurance companies 5% 24% 11% 26%
Governments 2% 1% 3% 1%
Mutual funds 1% 2% 3% 4%
Pension funds 2% 3% 3% 5%
Hedge funds 3% 5% 4% 5%Source: BBA 1999/2000 Credit Derivatives Survey
The market remains bank-
dominated
but other players are
becoming more active
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1997 1999 2002E
Corporates Banks Sovereigns
Total return swaps
10%
Portfolio/CDOs
18%
Credit spread
products
6%
Basket trades
7%
Asset swaps
11%Credit default
swaps
37%
Credit-linked
notes
11%
Underlying Composition (end-2002 estimate)
8/14/2019 11 September 2002
9/17
Market characteristics
Why use credit derivatives?There are multiple reasons why banks, insurers, corporates and fund managers use
credit derivatives. Credit derivatives facilitate risk management by increasing diversity
or lowering concentrations, thereby potentially reducing regulatory (or economic)
capital requirements and increasing income. Theoretically, this could be achieved
through the cash market, but the credit derivatives market provides greater flexibility,
efficiency and liquidity. That said, there are some challenges including greater spread
volatility, counterparty risk as well as different default probabilities and, depending on
settlement, recovery values. The table below illustrates the motivation of protection
buyers and sellers.
The motivation of protection buyers and sellers
Protection sellers Protection buyers
Portfolio/credit risk management
Free regulatory capital
Manage economic capital
Exploit arbitrage
Increase return via fee income
Fill credit lines
Source: DrKW
The CDS vs. the cash marketFlexibilityCDS offer a more flexible way to both assume and short credit risk, as the restrictions
of the cash market can be overcome. In the cash market investors are limited to an
entitys outstanding bonds (and/or loans), whereas CDS contracts can be tailored in
terms of maturity, size and the type of obligation. By referencing borrowed money,
CDS also facilitate credit exposure to entities without public bond issues, creating a
more diverse credit universe and avoiding exposure to frequent issuers. For those
wishing to short credit, buying protection can circumvent the potential difficulties of
borrowing the cash instruments and the risks of the repo market (e.g. funding or short
squeezes on the bond). As credit derivatives are an over-the-counter product, more
complex structures can be tailor-made according to the investors risk appetite.
The cash market imposes a number of restrictions in terms of the volume and pattern
of issuance volume and number of issuers. If one takes the European corporate bond
market as an example, issuance rose tenfold between 1995 and 2000. Europes single
currency in 1999 set a milestone and corporate issuance reached 193bn7 in 2001.
Our credit strategists expect new issuance to fall to around 125bn7 in 2002, and for
outstanding volume to be around 600bn7 at the end of the year. The market in Europe
is still dwarfed by the US$ corporate bond market, where outstanding volume is
expected to reach US$4 trillion (including financials) by the end of 2002.
There are multiple reasons to
use credit derivatives
Restrictions of the cash market
can be overcome
These restrictions include the
volume and pattern of issuance
and number of issuers
7 Excluding financials
8/14/2019 11 September 2002
10/17
-denominated corporate bond supply excluding financials (bn)
Source: DrKW
In the past few years, corporate bond supply in Europe has been particularly strong
from industries in which companies were increasing leverage and/or expanding rapidly
via M&A e.g. telecoms, and more recently from companies with significant refinancing
needs. Still, issuance has been fairly irregular over time, making it difficult to plan credit
investments. A look at iBoxx, a credit index that includes issues exceeding 500m and
denominated in , shows that it currently includes 440 bonds from 210 different issuer
groups. By industry, banks and financial services dominate with over a 30% share,
followed by telecoms and autos as the chart below illustrates.
The CDS market in comparison does not rely on new bond issuance and therefore
offers better opportunities to select a more diversified portfolio. According to our credit
derivatives desk, there are around 400-500 credits that are traded (mostly investment
grade), which are represented in the chart below. We can see that the industry
distribution is far more even compared to iBoxx, as the market is not skewed in favour
of particular sectors. Industrials, energy and metals and consumer products account
for a larger proportion of the market, while the share of banks and financials, telecoms
and autos is much smaller.
Credit derivatives market vs. iBoxx - by industry sector
Source: iBoxx and DrKW
Corporate bond supply in
recent years has been skewed
towards particular industries
CDS market: a more diverse
universe to choose from
0
50
100
150
200
250
1998 1999 2000 2001 2002 (end-
Aug)
0
5
10
15
20
25
30
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
1999 2000 2001 2002
Banks
6%
Energy and metals
10%
Industrials
20%
Autos
4%
Financial services
9%
Healthcare
4%
Telecoms
5%
Consumer
products
11%
Media
6%
Retail
6%
Utilities
6%Transportation
4%
Technology
6%
Hotel and
leisure
1%Banks
20%
Energy and metals
2%
Industrials
8%Autos
14%
Hotel and
leisure
1%
Technology
2%
Transportation
1%
Utilities
10%
Retail
1%
Media
1%
Consumer
products
7%
Telecoms
20%
Healthcare
1%
Financial services
12%
CDS market iBoxx
Annual Monthly
8/14/2019 11 September 2002
11/17
Efficiency
As CDS are unfunded instruments, credit and funding decisions can be separated fromone another. Protection sellers can take pure credit risk, tailored to their own
specifications. They do not need to fund the position, and credit expertise can
therefore be leveraged more cost-effectively. In addition, banks can hedge credit
portfolios much more efficiently than if they had to physically transfer all the assets
(and fund such a transaction). These portfolios can be tranched into products with
different risk/return profiles and investors can chose risk positions according to their
risk appetite/return requirements. This increases the efficiency of the market.
Liquidity and market dynamicsCompared to the cash market, the credit derivatives market tends to be more volatile.
Liquidity, or the lack of it, is driven by news on a particular name or industry sector,
new issuance in the cash market, etc. Generally bid-offer spreads are greater in the
CDS market. In the more liquid names, the spread will be 10-20bp depending on
market conditions, but significantly greater in others. However, the CDS market has
tended to be more liquid than the cash market in off-the run names or in times of
distress e.g. Russia in 1998, Railtrack, as well as with large volumes.
As with most derivatives, CDS exaggerate movements in the underlying market,
meaning that protection sellers are subject to greater volatility than if they held an
equivalent cash position. Generally, in periods of deteriorating credit sentiment the
CDS widens more than the cash bond as market participants clamber to buy protection
in order to hedge their positions or go short. As we have already mentioned, actually
selling the bond or shorting it may be difficult when there is negative news surrounding
an issuer, but in the CDS market there is usually an offer for protection that can be hit.
Dealers mark out their offers as the demand for protection increases, and this creates
a more exaggerated widening compared to the cash market, although arguably it
represents a more accurate pricing of that issuers credit risk. Likewise a tightening in
the cash may be magnified in the CDS, as the premium for protection falls quickly as
credit concerns subside. These trends are more evident with weaker credits, but they
still exist in more stable names.
As CDS are unfundedinstruments, credit and funding
decisions can be separated
Greater liquidity in times of
stress, in off-the-run names
and large transactions
Spread volatility is typically
greater
8/14/2019 11 September 2002
12/17
Spread comparison of the CDS and cash market 1
1 All spreads are mid levels, and corporate bond spreads are asset swap spreads
Source: iBoxx and DrKW
Counterparty riskAs well as the credit risk of the reference entity, in a CDS there is also the default risk
of the counterparty to consider. For benchmark issuers, this typically results in theCDS market trading slightly wider than the cash (positive basis), as the graphs above
show. In off-the-run names where the cash is less liquid than the CDS, negative basis
may exist on occasions (i.e. the CDS is trading inside the cash). This creates a
possible arbitrage between the two markets by taking a cash position and
simultaneously buying protection. However, this trade can be difficult to identify in
practice, and because the cash position needs to be funded it does not represent the
most efficient use of capital.
Counterparty risk adds a few
basis points, so CDS usuallytrade outside the cash market
0
100
200
300
400
500
600
700
Jan-02 Mar-02 May-02 Jul-02 Sep-02
5yr CDS 6.75% '08
0
100
200
300
400
500
Jan-02 Mar-02 May-02 Jul-02 Sep-02
5yr CDS 5.875% '06
0
50
100
150
200
250
Jan-02 Mar-02 May-02 Jul-02 Sep-02
5yr CDS 6.125% '06
0
200
400
600
800
1000
1200
Jan-02 Mar-02 May-02 Jul-02 Sep-02
5yr CDS 5.75% '06
0
20
40
60
80
100
Jan-02 Mar-02 May-02 Jul-02 Sep-02
5yr CDS 5.375% '08
0
40
80
120
160
200
Jan-02 Mar-02 May-02 Jul-02 Sep-02
5yr CDS 5.25% '06
France Telecom Deutsche Telekom
DaimlerChrysler Fiat
RWE Endesa
8/14/2019 11 September 2002
13/17
Defaults and recoveries may differ from historical data
Default probability and recovery values for a given credit may differ from historical datafor the cash market, depending on the definitions of credit events as well as the terms
and conditions of settlement. In 2001, Moodys published a report7 in which they
discussed at length the notion that a CDS may, if certain (soft) credit events are
included in the contract, expose investors to a higher frequency and severity of losses
than if they held an equivalent cash position.
Moodys definition of a default encompasses any missed or delayed payment of
interest and/or principal, bankruptcy and receivership, and a distressed exchange.
Some types of restructuring, Conseco being an example (see page 6), would fall
outside Moodys definition of default and would not be captured by a Moodys rating or
by its historical default statistics. Neither would obligation acceleration and default,
while some components of ISDAs definition of bankruptcy are not consistent with
Moodys. For example, the placing of Railtrack under administration in December 2001
constituted a credit event under the ISDA definition of bankruptcy, however all of its
debt service obligations were being met in a timely manner. There was no payment
default and Moodys retained the companys investment grade rating.
While Moodys has some justification, its argument is now less relevant given the
advent of modified restructuring and the fact that obligation acceleration and default
are now omitted from CDS documentation. Moreover, if restructuring is dropped
altogether, then default probabilities should begin to more closely reflect those implied
by historical data. But at present the argument does still hold in certain circumstances
so it is a risk that needs to be considered, albeit a relatively small one.
Default probability and recoveryvalues may differ from
historical data
Some credit events are not
consistent with Moodys
defintion of default
Moodys argument has some
justification, but it is now less
relevant
8 Understanding the Risks in Credit Default Swaps, Moodys Investors Service, 16 March 2001
8/14/2019 11 September 2002
14/17
CDS applicationsAs we have already mentioned, not only has the CDS market become very large in its
own right, but single name CDS are increasingly being used as the building blocks to
create more innovative structured products. Here we briefly explain three of the most
widely-used applications; the credit-linked note, the synthetic CDO and the first-to-
default basket.
Example 1: credit-linked noteA credit-linked note (CLN) can be thought of as a note with an embedded CDS, oralternatively as a collateralised CDS. The investor receives a pay-off that is dependent
on the performance of one or more reference credits. In its simplest form, a CLN pays
a coupon and redeems at par provided that no credit events have occurred on the
reference entity/obligation(s). If a credit event does occur, then the note is redeemed
for an amount equal to par minus the loss on the defaulted credit.
This is the basic design, but a CLN can be structured with additional features
according to investor requirements. It could be principal protected, meaning the
investor is repaid par at maturity regardless of whether a credit event has occurred or
not. A credit event would simply mean that the coupon stops being paid. Alternatively it
could be structured with coupon protection instead. Other variations include linking aCLN to credit spreads rather than credit events as such, or referencing it to an equity
or a commodity index.
There are two main types of CLN structure: an EMTN CLN and a SPV CLN. In the
former, the note is issued by a well-known entity (a bank or corporate), so the investor is
exposed to the credit risk of the issuer as well as the reference credit(s). In the latter, an
SPV issues the CLN and the note proceeds are invested in highly-rated collateral (e.g.
AAA ABS or government bonds), although this may also be done for an EMTN CLN.
EMTN vs. SPV CLN
Source: DrKW
CDS are increasingly be used
as the building blocks for more
innovative structured products
A CLN is effectively acollateralised CDS
It is a very flexible product and
can be tailored according to
investor requirements
There are two broad types: an
EMTN CLN and a SPV CLN
InvestorCDS
counterparty
CLN coupon
EMTN CLN
EMTN issuer
SPV CLN
InvestorCDS
counterpartySPV
Collateral
CLN coupon
Principal
Principal
Principal Bond coupon
CDS premium
CDS premium
Libor f unding
Bond coupon
InvestorCDS
counterparty
CLN coupon
EMTN CLN
EMTN issuer
SPV CLN
InvestorCDS
counterpartySPV
Collateral
CLN coupon
Principal
Principal
Principal Bond coupon
CDS premium
CDS premium
Libor f unding
Bond coupon
8/14/2019 11 September 2002
15/17
CLNs are on-balance sheet investments and therefore are attractive to investors such
as insurance companies who cannot use credit derivatives directly (note that there aresome exceptions). Typically they provide a pick-up over a reference entitys senior
unsecured bonds, and a huge advantage is their flexibility. They can offer investors
exposure to one or more reference entities in any currency, with any maturity and with
any coupon structure. If required, a CLN can be rated by the rating agencies.
Example 2: synthetic CDOBefore the emergence of synthetics, bank CDOs were cash flow structures where the
originating bank sold a pool of assets (usually loans) off-balance sheet and obtained
funding. As the credit derivatives market developed, synthetic CDOs began to emerge
towards the end of the 1990s and have allowed banks to decouple risk transfer and
funding. In this type of transaction, the originating bank transfers the risk of a
designated reference portfolio by buying protection, achieving regulatory (or economic)
capital relief, while the assets themselves remain on-balance sheet. This risk is then
tranched up. As many banks can achieve a lower cost of funding in the senior
unsecured debt market, synthetic CDOs tend to be largely unfunded. The majority
(around 90%) of the underlying credit risk is transferred via a super senior CDS. For
the remaining 10%, the bank buys protection from an SPV, which then issues notes.
The note proceeds are usually invested in highly rated collateral, allowing a AAA rating
to be achieved on the senior notes. Deals have tended to issue AAA all the way down
to BB, with noteholders assuming any losses on the portfolio above the equity piece
(i.e. the first loss, which has usually been retained by the bank). Credit events aresettled by liquidating the collateral.
A synthetic CDO structure
Source: DrKW
By buying protection from an OECD bank, the originating bank reduces its regulatory
capital requirement on the super senior risk to 20% from 100%. Assuming that
government bonds are used as collateral, a 0% risk-weighting is achieved on the
portion of risk that is transferred via note issuance. If the equity is retained it attracts a
one-for-one capital charge.
CLNs offer investors a number
of benefits
Synthetic CDOs have been used
by banks to obtain regulatory
(and economic) capital relief
By freeing up capital, these
structures have allowed banks
to enhance shareholder value
Protection
xth loss 7% SPV
Originating bank
Noteholders
Collateral
Principal andinterest
CDS premium
Note proceeds
Note principaland interest
Referenceportfolio
Equity holder
Noteproceeds
Usually the
originating bank
First l oss 3%
Super senior 90%Super senior CDS
counterparty
Protection
CDS premium
Protection
xth loss 7% SPV
Originating bank
Noteholders
Collateral
Principal andinterest
CDS premium
Note proceeds
Note principaland interest
Referenceportfolio
Equity holder
Noteproceeds
Usually the
originating bank
First l oss 3%
Super senior 90%Super senior CDS
counterparty
Protection
CDS premium
8/14/2019 11 September 2002
16/17
Basle 2 has been delayed until 2006 at the earliest, but when it is finally implemented
we expect the motivation for bank balance sheet CDOs to shift from regulatory toeconomic capital relief, as the current regulatory arbitrage will largely disappear. But as
regulatory and economic capital will be more closely aligned, banks, primarily those
adopting the internal ratings-based approach, should be able to retain the super senior
pieces of CDOs and significantly reduce their regulatory capital requirements.
Synthetic CDOs have evolved considerably over the years, both in terms of their
structure and motivation. Structural modifications that we have seen include the
omission of an SPV, fully unfunded deals and the combination of synthetic and cash
flow techniques. The traditional synthetic CDO was driven by banks seeking to gain
capital relief on cash assets already on their balance sheet. But economic capital
models discourage banks from cherry-picking their good assets and currently render
regulatory balance sheet deals defunct. As the CDS market has become more liquid,
banks, asset managers and hedge funds are increasingly ramping up portfolios by
selling protection into the market and using CDOs for arbitrage rather than balance
sheet purposes. As CDS trade wider than their cash equivalents, there are
considerable opportunities. The increased flexibility also means that portfolios with
greater diversity and granularity can be constructed, potentially reducing event risk and
ratings volatility for investors. Pools may be static or managed, but the dynamics of
managing CDS are different to managing a pool of bonds, particularly in terms of
spread volatility. This is something investors need to be cognisant of.
Example 3: first-to-default basketThis is a leveraged product that can either be issued as a CLN or structured in
unfunded form. The first credit to default in a basket of (two or more) reference entities
would trigger a payment to the protection buyer, and the structure would subsequently
unwind. Transactions may be cash settled, but more commonly they will involve
physical delivery of the defaulted asset. Should more than one default occur
simultaneously in the basket, then the asset of one defaulted reference credit only is
delivered i.e. the investor only loses once. Therefore a first-to-default basket allows
investors to leverage their credit exposure without increasing their downside risk. They
can obtain the yield of a low quality asset by taking exposure to high quality assets that
they are familiar with.
Pricing is driven by the credit spreads and the correlation of the names in the basket.
For a perfectly correlated basket, the first-to-default spread will be equal to the name
with the widest spread in the basket. For a perfectly uncorrelated basket, it would be
equal to the sum of all the spreads in the basket. In reality though, correlation lies
somewhere between these two extremes, and baskets are priced accordingly.
Determining the correlation is thus crucial, and on the next page we illustrate the
pricing of a basket consisting of six German investment grade corporates from different
industries.
Basle 2 is likely to effect a shift
from regulatory to economiccapital CDOs
Synthetic CDOs have evolved in
terms of their structure and
their motivation
It investors to leverage their
credit exposure without
increasing downside risk
Pricing is driven by the
underlying credit spreads and
correlation
8/14/2019 11 September 2002
17/17
First-to-default pricing: an example (as of 01 September 2002)
Reference entity 5yr CDS Moody's S&P
Daimler Chrysler +145bp A3 BBB+
Metro + 70bp Baa1 BBB+
Deutsche Telekom +290bp Baa1 BBB+
Thyssenkrupp +150bp Baa1 BBB
Siemens + 60bp Aa3 AA-
Lufthansa +140bp Baa1 BBB+
First-to-default pricing Euribor +445bp
100% correlation Euribor +290bp
0% correlation Euribor +855bpSource: DrKW