Techniques Used On The Credit Derivatives Market: Credit Default Swaps

Techniques used on the credit
derivatives market:
credit default swaps
EMMANUELLE OLLON-ASSOUAN
Directorate General Operations
Market Operations Directorate
Foreign Exchange and Euro Market Operations Division
As a response to the need for credit risk protection, the credit derivatives market has seen substantial
growth over the past few years. While the deteriorating credit status of a large number of issuers in
2002 contributed to the expansion of this market, this growth can also be ascribed to the broadening
of the uses to which these credit derivatives are put, above and beyond their original function of
providing protection. This article describes how credit default swaps (CDSs) work and how they are
used, CDSs accounting for the lions share of the credit derivatives market.
One characteristic feature of CDSs is their premium, which is expressed in basis points and constitutes
the periodic cost of purchasing protection. As an initial approximation, this premium may be regarded
as being equivalent to the spread between the bonds issued by reference entities (the risk on which
the derivatives are used to hedge) and the level of the swap curve; however, for technical reasons this
is not in fact the case. This differential between the CDS premium and the bond yield spread
commonly known as the basis encourages the use of CDSs as a new instrument for arbitrage,
investment and position-taking on the credit market.
In addition to credit risk hedging, arbitrage strategies are put in place to take advantage of the
fluctuations in the basis and thus obtain maximum benefit from the now recognised blurring of
boundaries between the bond and credit derivatives markets. An illustration is given of their use in
the context of sovereign risk. As one of the original sources of demand for protection, the sovereign
CDS market constitutes an ideal testing ground for strategies regarding credit derivatives instruments.
94
Banque de France Financial Stability Review No. 4 June 2004
Techniques used on the credit derivatives market: credit default swaps
he credit derivatives market has seen

substantial growth in volume terms over the
past three years. The burgeoning of this
market, whose primary purpose is to give protection
to holders of credit risk, can partly be ascribed to
the deteriorating credit status of a large number of
issuers in 2002. Its continued expansion in 2003, in
spite of the improvement in the credit risk
environment (illustrated by the trend to upgrade
ratings by agencies and the narrowing of spreads
between corporate and government bonds) can be
attributed to the extension of the uses of these
instruments beyond the mere transfer of risk.
The aim of this article is to provide a technical
analysis of the functioning of the credit default swap
(CDS) market. It does not propose to re-examine
the issues raised by the development of these
derivative credit instruments from a financial
stability perspective, nor their probable impact on
the behaviour of banks and other types of players
on the financial markets. As far as data are
concerned, the financial centre survey carried out
in the second half of 2003 by the General Secretariat
of the Commission Bancaire, the Banque de France,
the Insurance Supervisory Commission (Commission
de contrle des assurances) and the Financial
Markets Authority among French market
participants, whose results are presented in another
article in this review, highlights the risks that the

use of these instruments can generate.
The extension of the use of credit derivatives such
as credit default swaps (CDSs) to a large number of
market participants has given rise to the emergence
of a new asset class and the standardisation of a
body of financial techniques and legal provisions
that are specific to these instruments.
This article1 describes these techniques and sets out
the terms governing the use of these instruments,
whether in the context of strategies of straightforward
speculation (trading), hedging of securities already
held in banks portfolios (so-called negative basis
strategies), or the creation of synthetic bonds
(fixed-rate or variable-rate credit linked notes CLNs).
All of these strategies are now widespread and are
utilised by different categories of market participants:
credit derivatives trading teams, managers of
traditional portfolios, asset and liability management
teams in large commercial banks, operators dealing
with emerging market debt, and teams specializing
in so-called high yield bonds.
The characteristics of the CDS market are
presented in the first section, while the second
reviews the arbitrage and hedging strategies
employed on this market.
1| Credit derivatives:
the characteristics of the CDS market
Credit default swaps are characterized by their flows,
their modes of settlement and the factors
determining their price.
1|1 Credit default swaps:

movement of flows
Initial CDS flows
CDSs make it possible to transfer credit risk between
two counterparties. The counterparty buying
protection seeks to rid himself of the credit risk on
a reference entity (the reference entity is the bond

issuer or the debtor whose debt is the subject of the
protection); the counterparty selling protection
assumes the credit risk on the reference entity for
investment or arbitrage purposes. The transfer of
credit risk between two counterparties takes place
via a swap in which counterparty A buys protection
and pays a regular premium to counterparty B, a
seller of protection, which thus takes on exposure
to the credit risk. This regular premium constitutes
the fixed part of the swap. The contingent part of
the swap is paid by the protection seller to the
protection buyer if, and only if, a predefined credit
event (see below) on the reference entity occurs.
This article has benefited from the input of several CDS practitioners, as well as the information compiled by many commercial banks on this
topic. In particular, we wish to thank Rayas Richa from CAI Indosuez, Xavier Brunet from CDC-IXIS and Catherine Lubochinsky, Professor at
Paris II University.
95
Payment of the regular premium

(no credit event)
Buyer of
credit risk
protection
Notional amount
(after credit event)
Seller of
credit risk
protection
Delivery of securities
(after credit event)
The regular premium (which is discussed in greater

detail in parts 1.2 and 1.3) is expressed in basis points
of the notional amount of the swap and is
conventionally known as the credit default swap
spread or credit default swap premium. It is
generally paid on a quarterly basis. Market
participants have adopted fixed dates for payment
(20 March, 20 June, 20 September and
20 December). More unusually, and particularly in
the context of imminent default, this premium may
be paid in the form of an upfront fee when the
contract is concluded. The latter is expressed as a
percentage of the notional amount rather than in
basis points. Thus, if a default is imminent, the
protection seller will ask for 20% of the notional
amount from the protection buyer in the form of a
single payment at the start of the contract.
Otherwise he could be forced to pay the notional
amount without having received prior compensation
for this risk.
Terms of settlement of a credit default swap

in the case of a credit event
If a credit event occurs, payment is made either in
cash or by delivery of the reference entitys
securities, with the protection seller paying the
nominal (par) value of the securities.
In the case of cash settlement, the reference entitys
debt is valued at market price. The protection
seller pays the protection buyer the difference
between the securitys nominal value and its
market price, or the difference between its par
value and the estimated recovery value. The
estimated recovery rate at the time of default is
determined on the basis of the reference
obligations market value. The parties to the
contract poll market makers to establish the
percentage of the par value constituting the
expected recovery rate. Thus, since Parmalat filed
for bankruptcy in late December 2003, its bonds
have traded at 20% of their par value, which
means an expected recovery rate of 20%.
96
In the case of physical settlement, the protection

buyer delivers a bond issued by the reference
entity or a bank claim that it holds on the entity in
exchange for payment of the par value. The
protection buyer may deliver any bond or bank
claim vis--vis the reference entity if they meet
certain criteria: they must be denominated in one
of the G 7 currencies (US dollar, pound sterling,
canadian dollar, yen or euro); their maturity may
not exceed 30 years; and they must be ranked pari
passu and therefore may not be subordinated.
In most cases, market participants opt for physical
settlement. This preference can be ascribed above
all to the fact that the protection seller, who receives
the securities in exchange for payment of the par
value, may anticipate a revival in the reference
entitys fortunes in the future. The use of physical
settlement means that losses are not irrevocably
fixed at the time of default by settlement in cash.
Example: an investor wishes to protect himself
against the credit risk on borrower X (see example
involving Worldcom below) the reference entity
over a five-year period for an amount of
EUR 10 million. The CDS premium is quoted at
100 basis points. The amount that the protection
buyer has to pay periodically is EUR 100,000 per
annum. If a credit event takes place, the protection
seller has to make the contingent payment. In the
case of physical settlement, he pays the notional
amount of the contract, i.e. EUR 10 million, and the
buyer delivers securities issued by the reference
entity. In the case of cash settlement, the protection
seller pays an upfront fee to the buyer, equivalent
to the par value less the market value of the
reference entitys securities; the protection buyer
retains the reference entitys securities. Worldcoms
bankruptcy in July 2002 can serve as an illustration
of these two modes of settlement.
Credit default swap involving Worldcom (a)
(cash settlement)
Protection
buyer
After Worldcoms bankruptcy

in July 2002,
payment of USD 8.3 million
(USD 8.3 million = 10 million - 1.7 million),
with a notional amount
of USD 10 million and securities
with a market value of
USD 1.7 million after bankruptcy
Protection
seller
a) The expected recovery rate at the time of Worldcoms failure was 17%.
Credit default swap involving Worldcom

(physical settlement)
deemed to constitute financial debt. As far as emerging

market sovereign issuers are concerned, the relevant
debt universe is confined to bonds. Protection selling
and buying can take place on reference entities that
have no bond debt, their debt being solely with credit
institutions.
Transfer of Worldcom securities

after the reference entitys bankruptcy
Protection
buyer
Protection
seller
Payment of
USD 10 million after
Worldcoms bankruptcy
in July 2002
Standard features of a single name

credit default swap contract
The standardisation of credit default swap contracts
arose out of a process of consultation within the
profession conducted by the International Swaps and
Derivatives Association (ISDA), which led to the
drawing up of the first convention in 1999, followed
by a second one in 2003. The amendments made to
the conventions resulted principally from the
lessons learned from the Russian and Argentine
sovereign debt crises, in 1998 and 2002 respectively.
In particular, they set out more exactly the debt
universe concerned and give a more precise
definition of what constitutes a credit event
(see Box 1).
The reference entitys debt universe on which a credit
event may be observed is strictly defined. Only
financial debt borrowed money to use the ISDAs
terminology should be considered, with operating
debt excluded. Bonds, loans and credit letters are
Credit events that trigger payment of the contingent

part of the swap are specified as being bankruptcy,
failure to pay (on a nominal amount of more than
USD 1 million), obligation acceleration, moratorium
and restructuring (nominal amount greater than
USD 10 million). In the case of sovereign debt,
bankruptcy is not listed as a credit event.
In practical terms, the date of the credit event is
taken to be that on which one of the counterparties
sends notification of default to the other. It is more
likely that this will be done by the protection buyer.
Nonetheless, the seller may have an interest in
triggering early payment of the contingent part if
he expects to obtain a higher recovery rate.
Notification of default is based on the cross-checking
of two different sources of public information
(Bloomberg, Reuters and daily financial papers).
If a cash settlement is made, the recovery rate, which
is used to determine the amount of the upfront fee,
is established by means of a valuation process
involving five independent bond market makers. In
the case of physical settlement, following agreement
as to the date of the credit event, the protection buyer
must notify the seller of the bonds or loans that he
intends to deliver to him. He has 30 calendar days
in which to give this notification. If he fails to do so
within this period, he loses his right to protection.
Box 1
Russias default in 1998

On 17 August 1998, the Russian government announced the restructuring of its sovereign bond debt and
suspended transactions on those securities. In response to this announcement, the foreign currency-denominated
external debt depreciated strongly, in spite of the fact that it had not been ascertained that Russia would also
default on this asset class.
The diverse nature of Russias debt highlighted the gaps in the existing documentation covering CDS contracts.
The complex structure of Russias debt and the litigation arising therefrom prompted the ISDA to tighten up
certain aspects of their documentation. More specific details were included regarding the issuers identity,
subordination clauses and creditors status.
.../...
97
Argentinas default in 2001

Following the moratorium on Argentinas external debt declared on 23 December 2001, the settlement of CDS
contracts did not encounter many difficulties given that at mid-February 2002, 95% of outstanding CDS contracts
had been settled. Nonetheless, the settlement of credit default swaps in this context provided an opportunity to
define contracts more precisely with regard to three main aspects:
Definition of a credit event: debt swaps executed on a voluntary basis may not be considered as credit
events. Protection buyers had wanted to include in the moratorium the exchange of debt that had taken place
in November 2001, arguing that it constituted a restructuring linked to the default.
Definition of deliverables: zero-coupon bonds that satisfy the criterion of non-contingence are accepted as
deliverables.
Option of partial cash settlement: partial cash settlement is permitted when the protection buyer is unable,
either for technical or legal reasons, to deliver securities, but is not allowed on the grounds of the high price
of the security on the market and therefore the difficulty of obtaining it. The background to this was that
physical settlement via delivery of Argentine debt securities had led to a rise in the price of deliverable
securities on the repo market with the effect that some protection buyers were unable to deliver securities.
1|2 Parallels between the bond

and CDS markets
The CDS premium expresses the general credit risk
on a reference entity. We can liken the CDS to the
additional yield relative to the Libor offered by the
reference entitys bond on an equivalent maturity
(asset swap spread), in that this additional yield
expresses the credit risk on the issuer. The CDS
premium differs from it, however, by virtue of
technical factors (distortions linked to the securities
lending market, use of the cheapest-to-deliver bond).
This difference is commonly known as the basis. The
use of mechanisms from the repo market, the
cheapest-to-deliver bond and the term basis are
derived from existing vocabulary and concepts on
the interest rate derivatives market.
The asset swap spread and CDS premium

There is a strong correlation between the yield
differential of a bond relative to the Libor rate (the
bond spread) and a CDS premium of the same
maturity for the same reference entity. In theory,
an investor can put a risk-free investment in place
by combining:
the purchase of protection on a reference entity
via a credit default swap;
98
and the purchase of a variable-rate security of

the same maturity issued by the same reference
entity. The bond coupon is made up of the Libor
rate plus a credit spread (S).
The purchase of a variable-rate bond is preferred as
it does not entail any interest rate risk. The
investment is risk-free in that, at maturity, with or
without a credit event, the investor receives the
notional amount of the bond, which he uses to repay
his financing. It is assumed that the investor finances
his strategy by borrowing at Libor plus or minus a
margin depending on his rating.
In the absence of a credit event, the investor receives
the Libor (L) plus the credit spread (S) on the
variable-rate bond. He pays the Libor plus or minus
a refinancing margin (RM) and the credit default
swap premium (CDS).
To eliminate any opportunity for arbitrage, the CDS
premium must be equal to the bond spread after
deduction of the investors refinancing margin.
Libor (L) + credit spread (S) =

Libor (L) +/- refinancing margin (RM)
+ CDS premium (CDS)
credit spread (S) =

CDS premium (CDS) +/- refinancing margin (RM)
If we assume that RM is negligible, we can regard

the credit spread as being equal to the CDS premium.
Technical determinants of the differential

between the asset swap spread
and the CDS premium
The parallel between the bond spread and the credit
default premium is however limited. A differential
between the CDS premium with a given maturity
and the additional yield relative to Libor offered by
a bond with an equivalent maturity can be observed.
This differential is known as the basis. The basis
may be negative or positive, and fluctuates
depending on the developments affecting credit risk
on the reference entity as well as technical factors.
A positive basis (premium CDS>bond spread) can be
ascribed to both structural and market-related factors
Among structural factors can be listed:
the repo market. By carrying out repo transactions,
the bondholder may borrow at a lower rate than
the uncollateralised rate. This reduction in the cost
of refinancing results in a narrowing of the credit
spread given that S = CDS + RM. This reduction
is even greater when the bond is in such demand
on the repo market that it is possible to obtain
refinancing at a much lower rate than the
uncollateralised rate (it is then traded as a specific
bond on the securities lending market);
the cheapest-to-deliver bond option. In the event of
default and physical settlement, the protection
buyer can choose, in return for payment of the
par value, to deliver any security of the reference
entity, denominated in any specified currency,
and with any maturity (provided, however, it
does not exceed 30 years). The protection buyer
delivers the bond with the lowest valuation on
the market. The uncertainty about which
securities will actually be delivered to the
protection seller justifies an increase in the CDS
premium that is paid to him;
management of bank portfolios. The hedging of
bank credit risk and the desire to release
regulatory capital by underwriting risks incurred
with respect to loans recorded on the banks
balance sheet generate demand for protection
instruments that cause CDS premiums to rise.
2
Among cyclical factors, we can note convertible bond

issues. The issuance of bonds convertible into equities
(consisting of a bond and an option on an equity
sometimes valued differently from the valuation used
on the equity option market) encourages arbitrage
funds to buy the convertible bonds to trade off against
the price of the equity option. In order to cover the
bonds flows and obtain an equity option, funds may
buy protection via credit default swaps. The strong
demand for protection leads temporarily to an
increase in the CDS premium.
A negative basis (CDS premium<bond spread)
is less common but can be caused
by a number of different factors
Credit default swaps make it possible to obtain a yield
without locking up capital. Market participants prefer
to receive payment for credit risk by selling protection
via credit default swaps. They may accept a lower
return than the credit spread on the spot market since
they have no costs in terms of locking up capital.
Counterparty risk. The protection buyer may be
exposed to a two fold default risk, that of the
reference entity and that of the protection seller.
If there is a strong correlation between the
counterpartys default risk and that of the reference
entity (for example between the reference entity
Parmalat and counterparties such as the Italian
banks that granted it credit lines), the premium paid
by the protection buyer will tend to be even lower.
Demand for protection linked to the structuring of
collateralised debt obligations synthetic CDOs.
The abundant supply of protection linked to the
structuring of a large number of synthetic CDOs2
leads to lower CDS premiums.
Variation in the basis constitutes a leading indicator
of increasing credit risk for investors.
Market developments over the last two years point to
a strong positive correlation between the levels of the
basis and of spreads. When yield spreads between
bonds issued by non-financial corporations and the
swap curve are wide, the basis is strongly positive,
and conversely, when bond spreads narrow, as was
the case last year, the basis tends to be zero or negative.
The main reason for this phenomenon is the
asymmetry of supply and demand under certain
market conditions.
The structuring of a synthetic CDO involves simultaneously investing capital at Libor and selling protection on about 100 reference entities.
99
It may also be noted that, in a context of deteriorating

credit risk and massive sell-offs of the reference
entitys bond debt, the refinancing rate obtained via
repo transactions involving the reference entitys
bonds tends to fall (on account of the need to hedge
short selling). As was mentioned above, rising prices
on the repo market leads to a widening of the basis.
While the basis tends to widen along with credit
spreads, the size of this movement varies according
to the credit status of the reference entities concerned.
For low-rated entities, the basis is generally very
positive (on account of substantial protection buying).
For issuers with very high ratings, the base is also
positive given that investors take speculative positions
on the CDS market (by buying protection on account
of the tendency of the credit situation to deteriorate).
For investors with an intermediate rating (type A-
and A), the basis tends to be negative since such
entities are more frequently subject to synthetic
CDO-type operations. This explains why there is a
greater supply of protection on these issuers, which
results in a certain tightening of the basis. Moreover,
investors do not feel the need to hedge against default
risk on these entities.
Chart 2
Turkey five-year EMBI spread,
Turkey five-year CDS and Turkey five-year Basis
1,500
1,400
1,300
1,200
1,100
1,000
900
800
700
600
500
400
300
200
100
0
-100
10/12 29/01 20/03 09/05 28/06 17/08 06/10 25/11 14/01 04/03
2002
2003
2004
Turkey EMBI spread

Turkey five-year CDS
Basis
The basis widened in early spring 2003, reflecting

concerns about the sustainability of Turkeys debt
and uncertainty as to the outcome of the war in
Iraq. The removal of these concerns resulted in a
substantial narrowing of the credit spread and a
tightening of the basis.
Chart 3
Correlation of the EMBI spread and the basis for Turkey
450
400
350
300
Basis
Thus, a deterioration in credit market conditions

prompts bondholders to buy protection to hedge
against possible default without getting rid of their
bond holdings in the first instance. This strong
demand for protection leads to an increase in CDS
premiums and as a result contributes to the widening
of the basis. Conversely, in a situation where credit
spreads are tending to narrow, demand for protection
is lower and investor appetite for the bond market
remains buoyant. As a result, the basis narrows.
250
y = 0.3349x 77.793
R2 = 0.7973
200
150
100
Chart 1
Basis according to rating
50
0
-50
0
Basis
50
200
400
600
800
Spread EMBI
1,000
1,200
40
30
20
10
Rating
0
-10 BBB- BBB BBB+ A-
AA
AAA
Variation in the basis may also be observed for a

single issuer according to changes in its credit status
over time. This phenomenon can be observed with
regard to Turkey during 2003.
100
These variations in the basis constitute a precious

indicator of the position of the market, making it
possible to answer the following questions: is the
market long or short of protection? Do structured
transactions weigh on the bases and do they
constitute a determining factor in the evolution of
the credit spread itself? etc.
1|3 Underlying pricing models:

their scope and limitations
Given that the CDS premium is an explicit
component of credit risk, it may be regarded as a
relatively pure expression of a reference entitys
credit risk. Credit default swaps in themselves are a
prime indicator for the measurement of implied
probabilities of default. The default rate may be
derived from historical or market data. The historical
default rate is calculated using statistical data on
actual corporate defaults; tables are published
showing this rate, notably by rating agencies.
Two categories of models are generally used to
estimate the default risk of a reference entity using
market data. The first category consists of structural
models derived from Mertons approach (1974), which
is based on modelling the value of a companys assets.
If the value of the reference entitys assets falls below
the face value of its debt, this constitutes a default.
In this case, the shareholder does not recover his
investment and the bondholder only receives the
residual value of the assets (which may be nil).
The probability of default depends on the one hand
on the reference entitys liabilities structure (a
reduction in the debt to equity ratio leads to a
reduction in default risk) and, on the other hand, on
the volatility of its asset value (the greater the
volatility, the greater the probability that the value
of its assets is lower than that of its debt). Share price
volatility is generally accepted as a measure of asset
value volatility, which is otherwise difficult to assess.
The structural model was discussed in articles in the
November 2002 and June 2003 issues of the FSR,
which the reader may refer to.3 The second category
comprises so-called reduced models, notably default
intensity models. These latter methods are in general
based on an approach that models different possible
occurrences of default over certain time horizons to
calculate implicit probabilities of default. Market
participants use above all default intensity models.
The default intensity model and

determination of default probabilities
on reference entities
The logic runs as follows.
For a bond of maturity T, it is considered that in each
period (one year) two events are possible: payment
3
4
of coupon C signalling the continuation of

repayments with probability 1-; or default with
probability leading to a recovery value RV.
The price therefore includes two variables: the
implied probabilities of default and the recovery
rates anticipated by the markets.
By way of hypothesis, it is assumed that the
conditions of the first period also hold good for the
subsequent periods, i.e. interest rates, implied
probability of default and recovery rates all
constant.
The theoretical value of the bond (P) may be
calculated in the following manner:
T
P=
t=1
(VR(1(1+ r) ) )+ [ ( C(1(1+ r)) )+ ( F(1(1+ r)) )]

t1
0 t
t=1
0 t
0 t
with 0rt the zero-coupon rate of government bonds

and F the face value of the security.
Assuming that we know the market price, we now
have an equation with two unknowns, RV and .
To solve this equation, we can fix one of the
unknowns and calculate the other. This method is
based on a very strong hypothesis of a constant
annual probability of default over time (i.e. a term
structure of probabilities of default is discounted,
although this hypothesis is invalidated on a daily
basis by observation of the term structure of CDS
premiums). More sophisticated models that use the
bootstrapping method make it possible to
reconstruct this term structure of the default
probability by iteration (the default probability at
one year making it possible through bootstrapping
to calculate that at a two-year horizon, and so on).
The only variable necessary for the valuation of a
credit default swap that cannot be observed on the
market is the default probability. Based on the
statistical tables available, the convention is to
consider that the average recovery rate is 40%.
Nevertheless, the latest statistical studies devoted
to this subject and the defaults that have been
observed seem to point to a drop in this recovery
rate. Following Worldcoms bankruptcy, the recovery
rate on its debt totalled only 17%. In the cases of
Enron and Parmalat, the recovery rate is thought to
stand at around 20%. According to a study conducted
by Fitch in 2001,4 the average recovery rate for all
types of bond assets was 22%.
See Haas (2003) and Lubochinsky (2002).

Bank loan and bond recovery study: 1997-2000, Fitch, March 2001. Study carried out on the basis of debt totalling USD 24 billion of concerning
35 companies.
101
Thus, by way of example, given a flat euro Libor

curve for all maturities at 3.3%, a CDS premium that
pays 30 basis points at five years and a recovery rate
of 40%, it appears that the default probability on the
reference entity at a horizon of 5 years is 2.4%. If the
recovery rate increases to 50% and the CDS
premium remains unchanged at 30 basis points, the
default probability rises to 2.9%, and conversely, if
the recovery rate falls to 30%, the default probability
drops to 2.1%.
A much simpler approximate formula may also be
used satisfactorily which does not take account
either of the present value of flows or of the term
structure in order to derive the default probability
from CDS premiums, i.e.:
Pcds x Tcds = x (1 RR)
with
Pcds = CDS premium
Tcds = CDS maturity
= default probability
RR = recovery rate
Thus, with a premium of 30 bp, a recovery rate of 50%,
and a five-year maturity, the implied default probability is 3%.
This relationship may also be expressed in another form,
i.e.:
with x (1 RR) = anticipated loss
Pcds x Tcds = 1 anticipated recovery value
The scope and limitations of this approach

Criticism of this approach comes mainly from rating
agencies, whose assessments, which were based on
structural models, diverged from the evolution of
CDS premiums in 2002 and the default probabilities
that were derived from them. By way of illustration,
France Telecoms CDS premium widened to 730 basis
points in June 2002, whereas Standard & Poors

rating stressed the investment grade rating of the
issuer (BBB). The level of France Telecoms CDS
premium was thus very similar to that of Worldcom
three months before its failure.
Practitioners appear, in any case, to favour default
intensity models. First of all, they underline the ease
of use of default intensity models. They also point
out that the correlation between stock price volatility
(an important variable in the structural model) and
CDS premiums can be limited. While, during periods
of crisis, implied stock price volatility displays a
strong correlation with the CDS premium (see the
Adecco chart), the correlation seems to weaken
during phases of improvement in credit risk. For
example, the volatility derived from equity options
tends in part to ignore the dynamics at work on the
bond market, such as the search for higher returns
in a context of low interest rates, which can lead
investors to favour credit as a means of investment.
Chart 4
Adecco: implied stock price volatility (right-hand scale)
and CDS premium
350
100
300
85
250
70
200
55
150
40
100
25
50
10
0
17 Oct. 2003 17 Nov. 2003 17 Dec. 2003 17 Jan. 2004
-5
Implied volatility derived from equity option

CDS premiums
2| CDSs instruments for investment, speculation

and creation of synthetic securities: the strategies utilised
In addition to being credit risk hedging instruments,
CDSs have become instruments for investment and
arbitrage and are used to create structured products.
necessary to hedge the credit risk is calculated

depends on the hedging strategy adopted.
There are three types of hedging strategy.
2|1 Hedging credit risk via CDSs

This is the original function of CDSs and may involve
different procedures depending on the investors
objective. The way the notional amount of the CDS
102
Hedging the bonds face value. When the investment

horizon is long, the investor seeks to hedge the
nominal amount of the bond. The amount of the
CDS contract corresponds simply to the bonds
nominal amount. It is possible to adjust this
hedging ratio (one for one of the nominal

amount) using the ratio of anticipated recovery
values should these implied values be different
for the CDS and the bond.
This type of hedging is however unsuited to
short-term horizons if the bondholder wishes to
protect himself from spread movements or if, in
the context of imminent default on the part of
the reference entity, he wishes to safeguard his
portfolios market value. Sensitivity hedging
strategies are appropriate in the first case and the
so-called default neutral hedge ratio is suitable
in the second.
Spread hedging. This type of hedging is intended
to shield the value of the investors bond portfolio
from any variations in the credit spread. By way
of hypothesis, we assume that the variation is
uniform for all maturities. The aim is to ensure
that for a given spread variation the loss on the
bond is entirely offset by a gain on the CDS (and
vice versa).
Box 2
The sensitivity of the CDS price (discounted sum
of premiums multiplied by the notional amount)
to a 1-basis-point variation in the annual premium
is relatively close to that of a bond with the same
maturity.
S =
to be put in place. It takes into account a variation

of 1 basis point in the CDS premium and of 1 basis
point in the bond rate:
ratiosensitivity =
Sbond
Scds
with S bond being the bonds sensitivity.

Hedging the bonds market value (default neutral
hedge). When the default horizon is close, it is in
the investors interest to safeguard the market
rather than the face value of his portfolio. This
protection is especially necessary when the bond
has been acquired at above par market value, so
that the goodwill paid with respect to the nominal
amount is also hedged. The corresponding
hedging ratio is called the default neutral hedge
by market participants.
Default neutral hedge =
Pbond VRbond
100 VRcds
where Pbond= the bond market price including the

accrued interest; RVbond and RVcds are the implied
recovery values of the bond and the CDS
respectively.
The latter hedging strategy is effective (i.e. there is
no loss) if the implied recovery values in the ratio
match the recovery values actually obtained.
P P
R
This sensitivity is the price variation relative to

an interest rate variation. It is a function of the
maturity, the size of flows and the discount rates.
It is above all the size of flows, given that it is
different for the bond and the CDS, that leads to a
difference in sensitivity.
Thus, a CDS on France Telecom at horizon
December 2009 of 73 basis points displays
sensitivity of 5.19. The sensitivity of a bond of
similar maturity (France Telecom 7% bond due
December 2009) is 5.22.
As a consequence, we obtain a hedging ratio in

terms of sensitivity, which makes it possible to
calculate exactly the notional amount of the CDS
2|2 Directional strategies

The purchase or sale of protection may be part of a
purely directional strategy. Here it is a question of
anticipating a rise or fall in the CDS premium and
of taking the losses or profits on the position.
There are several different techniques for unwinding
a position on a credit default swap:
unwinding with the original counterparty to
whom is made or who makes a payment at
market value. The CDS is then terminated;
transfer of the CDS position to a third party at
market value thus making it possible to register
the gain or loss;
setting in place a reverse CDS.
103
In terms of final investors, the first two methods are

the most frequently used. They allow investors to
record losses or profits; in the second case, the future
market risks and legal risks are assumed by the new
counterparties. Market intermediaries tend to favour
the third option. Here the result is not immediate
but registered gradually. Thus, a market participant
may initially have bought protection at 100 basis
points; in order to unwind his position at a later date
he chooses to sell protection at 200 basis points on
the same reference entity. He gains a maximum of
100 basis points in annual premium until the maturity
of the contracts if there is no credit event. His profit
will be lower if a credit event occurs.
2|3 Arbitrage: basis

and relative value arbitrage
Market participants have formulated different
strategies that are derived from the basis. These are
commonly known as basis strategies and allow
arbitrage or credit positions to be taken.
A negative basis makes it possible to engage

in credit risk arbitrage
The investor buys the bond and protection on the CDS
market. Owing to the negative basis
(premium<spread), he receives a credit spread on
the bond of a higher amount than the CDS premium
that he has to pay. If a credit event occurs, the
investor is protected in that, this strategy allows him
to deliver the bond purchased. He may, however,
incur a loss if he has not used the appropriate
hedging strategy for the bond via a CDS (see above).
The positive basis strategy amounts to the

creation of a synthetic security
CDSs are increasingly used as the cornerstone in the
synthetic structuring of spread products such as
credit-linked notes (CLN). The CLN is a structure that
allows an investor to buy a synthetic security that provides
all of the characteristics of a bond while benefiting from
a higher return by virtue of a positive basis.
The investor may, for example, combine the purchase
of a certificate of deposit bearing interest around Libor
with the sale of protection via a CDS. This strategy is
mainly used with respect to reference entities with
high ratings since investors judge that defaults by these
entities is improbable. This strategy is particularly
attractive for participants whose investment
management is compared to that of a benchmark
portfolio since it provides an additional yield.
The same type of product can be structured with a
fixed-rate CLN, used to replicate investment in a
bond. In this case, the buyer combines via the CLN
a variable-rate investment which is turned into a
fixed-rate one (by means of an interest rate swap)
with the sale of protection on a reference entity,
which allows him to obtain a higher return than
direct investment in the underlying bond.
Creation of a five-year fixed-rate CLN
Desk CDS
(1)
CLN
buyer
(4)
Bank in charge
of putting
the CLN in place
(2)
Interest rate
swap desk
Example
(3)
Purchase of a five-year France Telecom bond

with a spread of 78 basis points (2009)
Purchase of five-year protection on France
Telecom for which the premium is 64 basis
points
Negative basis of 64-78 = -14 basis points
(carry trade gain received)
104
Treasury desk
(1) The bank sells protection to its CDS

desk and receives a quarterly premium.
(2) The bank receives the fixed-rate of the five-year
swap and every quarter pays the three-month
Libor recorded at the beginning of the quarter for
a notional amount indentical to that of the CDS.
(3) The bank invests liquidity equivalent
to the notional amount of the CDS
on the market at three-month Libor.
(4) The bank sells a five-year fixed-rate CLN, with
the CDS premium allowing the buyer to receive
a higher return than the five-year fixed-rate.
The structuring of a fixed-rate CLN requires the

involvement of several sectors in the trading room
of a commercial bank: credit derivatives and interest
rate swap dealers and the treasury desk. The team
in charge of the structuring and trading of the
instrument are remunerated by the payment of part
of the spread received by the investor above Libor.
The CLN, however, differs from a traditional bond in
three fundamental ways. First, it is characterised by its
illiquidity and only meets an investors needs provided
it is not involved in secondary transactions. Second,
although it is a synthetic equivalent, it cannot give rise
to refinancing at favourable terms on the repo market.
Lastly, the nature of the credit risk borne by the CLN
buyer is different. In the case of a traditional security,
this risk is confined to a deterioration in the credit status
of the issuing entity. In the case of a CLN, this risk does
not only consist of that incurred on the underlying
entity, but also includes the credit risk borne on the
counterparty to the swap, i.e. the protection buyer who
has undertaken to pay the CDS premium at regular
intervals. Although it only concerns a portion of the
structures outstandings, this risk needs to be taken into
account especially when premiums are sold on
low-rated issuers (given that the premium to be paid is
higher). In the example shown in the diagram above,
the protection buyer is the CDS desk. This two fold risk
accounts in large part for the additional yield.
The strategy of relative value of two issuers

in the same sector
More riskily, the investor can buy and sell protection
on two issuers from the same sector of activity. In
this way, he can protect himself from deteriorating
credit risk in the sector. He can also benefit from
the positive carry trade on this position. He may,
moreover, hope for a potential gain in market value
when the position is unwound if the relative risk of
the two entities has evolved favourably. In the
example below, the investor anticipates a
deterioration in Volkswagens credit status and an
improvement in DaimlerChryslers credit status.
2|4 Investment strategies based

on sovereign CDSs
Historically, it was the sovereign credit market that
first gave rise to demand for protection. The defaults
of the 1980s, notably in Latin America, generated
Box 3
Carry trade
Purchase of protection on Volkswagen
(54 basis points)
Sale of protection on DaimlerChrysler
(84 basis points)
This results in a positive carry trade of
30 basis points per annum
Profit-taking
If the investors expectations have proved to be
correct:
the premium on Volkswagen rises from 54 bp

to 64 bp: the market value of the position
appreciates given that the premium bought is
resold 10 basis points higher;
the premium on DaimlerChrysler drops from
84 bp to 74 bp: the market value of the position
appreciates given that the premium is bought
10 basis points cheaper;
overall, if the investor has taken positions of
EUR 10 million on each of these two issuers,
he earns EUR 91,2985 if he unwinds his double
position on Volkswagen and DaimlerChrysler
at these levels.
1
EUR 45,561 on DaimlerChrysler and EUR 45,737

on Volkswagen, constituting the discounted sum of
additional future flows factoring in a rise of 10 bp.
the need for hedging on risk countries. The creation

of Brady bonds was the first form that it took. CDSs
grew out of investors interest in preventive hedging
on risk countries. A BIS study conducted in 2003
(see Quarterly Review, December) points out that
quotes on sovereign reference entities are still
confined to a small number of issuers. Brazil,
Mexico, Japan, South Africa and the Philippines
together account for 40% of quotes on sovereign
issuers. Moreover, Fitchs study of September 2003
corroborates this and underlines that the credit
status of sovereign reference entities is on average
lower than that of reference entities as a whole (only
60% of sovereigns on which CDSs are quoted are
investment grade, compared to 90% for reference
entities as a whole).
105
The credit derivatives market on emerging countries

appears to be particularly developed. According to
market participants, the outstanding notional
amount on emerging market reference entities is
around USD 300 billion. By way of comparison, the
JPMorgan EMBI index totals USD 220 billion
(USD 260 billion if the countries of Eastern Europe
are included). The high notional value on the credit
derivatives segment relative to outstandings on the
spot bond market makes it possible to devise
diversified strategies on this segment of activity and
makes CDSs on emerging market sovereigns an ideal
testing ground.
Strategy of private risk coverage at lower cost: the
investor can partially hedge against a deterioration
of risk on non-financial companies relative to
sovereign risk. For example, by selling protection
on Mexico (230 bp) and buying protection on the
company Pemex (270 bp), the investor pays only
40 basis points to protect himself against a serious
deterioration in Pemexs credit status, based on the
assumption that the Mexican government will not
default. A reduction in the cost of hedging on the
state-owned company (whose debt is not
underwritten by the government) is obtained via
the sale of protection on the government.
Strategy of differing risk on two sovereign issuers:
the investor takes advantage of the differences
in premium between two sovereign issuers,
which can be ascribed to macroeconomic
fundamentals. Thus, investors have recently sold
five-year protection on Brazil (premium received
1,375 bp) and simultaneously bought five-year
protection on Turkey (premium to be paid
920 bp). This strategy is based on the expectation
that there will be an improvement in Brazils risk
profile (due to structural reforms and fiscal
consolidation) and a deterioration of Turkeys
relative position.
Yield curve strategy: CDSs can be used to anticipate
developments (steepening or flattening) in the
slope of the CDS spread curve. By way of example,
an investor anticipates a steepening of the slope

of Russias credit curve, which is traditionally very
flat. This steepening could be prompted by a
slowdown in the pace of structural reform. In this
context, the investor sells one-year protection on
Russia at 135 bp and simultaneously buys
three-year protection at 280 bp. If the slope of the
CDS curve steepens, the transaction can be
unwound with profit-taking.
Strategy of buying and selling forward CDSs:
investors can create forwards by buying and
selling protection on different maturities for an
equivalent notional amount. Thus, an investor
buys one-year protection on Brazil at 725 basis
points and sells two-year protection on Brazil at
1,150 basis points. This amounts to selling
one-year protection in a years time, thereby
earning a premium of 1,5935 basis points. This
transaction makes it possible to garner a higher
return than on the Brazilian spot market;
moreover, the absence of one-year spot securities
means that such a strategy is not possible on the
bond market.
Although the share of sovereign issuers involved in
credit derivative transactions is on the decline (54%
of the notional amount in 1996, but less than 15%
in 2003 according to Fitch), sovereign reference
entities have become more diverse. A number of
CDS transactions on sovereigns such as France and
Germany have taken place since the end of 2001.
According to the Fitch survey published in
September 2003, Japan stands out as one of the
reference entities most quoted by the market
participants questioned. Non-emerging countries
accounted for 40% of transactions on sovereigns.
Information derived from quotes on OECD sovereign
CDSs must be interpreted with caution on account
of the uncertainty regarding transaction volumes
on this segment. Owing to the intrinsic quality of
CDS premiums, which make it possible to isolate
credit risk, these instruments nonetheless constitute
a valuable tool allowing for a hierarchy among
sovereign issuers to be established.
(1+7.25 %) (1+X %) = (1+11.5 %)2
106
Bibliography
Fage (P.) (2003): Hedging with Credit Default
Swaps, Credit Suisse First Boston, Emerging
Markets Sovereign Strategy Focus, 24 November
Lubochinsky (C.) (2002) : How much credit should

be given to credit spreads?, Banque de France,
Financial Stability Review, No. 1, November
FitchRatings (2003): Global Credit Derivatives:

A Qualified Success, September
Packer (F.) and Suthiphongchai (C.) (2003):

Sovereign credit default swaps, Bank of
International Settlements, Quarterly Review,
December
Haas (F.) (2003): Towards a market continuum?

Structural models and interaction between credit
and equity markets, Banque de France, Financial
Stability Review, No. 2, June
Kiff (J.), Michaud (F. L.) and Mitchell (J.) (2003): An
analytical review of credit risk transfer instruments,
Banque de France, Financial Stability Review, No. 2, June
Xu (D.) and Wilder (C.) 2003: Emerging markets

credit derivatives: market overview, products,
analyses and applications, Deutsche Bank, Global
Markets Research, May
107

Techniques Used On The Credit Derivatives Market: Credit Default Swaps

Uploaded by

Copyright:

Available Formats

Techniques Used On The Credit Derivatives Market: Credit Default Swaps

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Techniques Used On The Credit Derivatives Market: Credit Default Swaps

Uploaded by

Copyright:

Available Formats

Techniques used on the credit

Banque de France Financial Stability Review No. 4 June 2004