JPM Bond CDS Basis Handb 2009-02-05 263815
JPM Bond CDS Basis Handb 2009-02-05 263815
JPM Bond CDS Basis Handb 2009-02-05 263815
AC
(44-20) 7742-7829
[email protected]
Saul Doctor
(44-20) 7325-3699
[email protected]
Yasemin Saltuk
(44-20) 7777-1261
[email protected]
J.P. Morgan Securities Ltd.
100
US
50
0
-50
-100
-150
-200
-250
-300
Jan-07
Jul-07
Jan-08
Jul-08
Jan-09
Bond-CDS Basis =
CDS Spread Bond Spread
CDS vs. Bond Spreads
Electricite de France
X-axis: Maturity. Y-axis: Spread (bp).
CDS Curv e
200
Bond Spread
150
Negativ e
100
basis
Positiv e
basis
50
0
09
10
12
13
14
16
17
19
Abel Elizalde
(44-20) 7742-7829
[email protected]
Table of Contents
1. Introduction...........................................................................3
Basis Definition ...........................................................................................................7
Main Drivers of the Bond-CDS Basis..........................................................................9
Why Do Investors Enter Negative Basis Trades? ......................................................12
Abel Elizalde
(44-20) 7742-7829
[email protected]
1. Introduction
Basis trades exploit the different pricing of bond and CDS on the same
underlying company.
They represent one of the closest trading techniques in the credit market to an
arbitrage free trade. If the bond is trading less expensive than the CDS: buy the
bond and buy CDS protection to lock in a risk-free profit. Although potentially
very attractive, basis trades are not usually straightforward and almost never
arbitrage free trades. Fortunately, the current levels of the basis are high enough to
make basis trades more attractive than ever.
Bond-CDS basis:
CDS spread minus Bond spread
The basis between a bond and a CDS measures the pricing differential between the
two, expressed on a running spread per year. In particular, it is computed as the CDS
spread minus the bond spread with a similar maturity. Throughout this report, we
will use the term basis to mean Bond-CDS basis, also referred to as basis-to-cash.
Investors frequently seek to exploit discrepancies in the Bond-CDS basis. The
rationale is that bond and CDS positions should offset each other in case of default,
allowing the investor to take a view on the relative pricing of bonds and CDS without
taking on credit risk.
Being different instruments, bond and CDS might have different exposures to
movements in the underlying companys credit risk. For example, higher liquidity in
the CDS market might cause CDS spreads to lead bond spreads after a negative piece
of information about a company reaches the market. Investors can set up basis trades
to profit from this, and related, phenomena.
Figure 1: Electricite de France CDS Spreads vs. Bond Spreads
X-axis: Maturity. Y-axis: Spread (in bp). Bond spread measured as the PECS.
200
CDS Curv e
Bond Spread
150
Negativ e basis
100
Positiv e basis
50
0
2009
2010
2012
2013
2014
2016
2017
2019
Figure 1 shows, for Electricite de France, the CDS spread curve together with the
spread on several bonds. The difference between the bond spread and the CDS
spread (with the same maturity) will give us the Bond-CDS basis.
Abel Elizalde
(44-20) 7742-7829
[email protected]
Basis trades aim to take advantage of two separate but related factors:
Hedging bonds with CDS should theoretically produce a risk-free trade. An
investor can buy a bond and matched CDS protection in order to be neutral to a
default event. If the income from the bond is larger than the cost of protection, an
investor should receive a risk-free income.
The basis should be mean reverting. The basis can be negative, positive or
zero. If it is close to zero, bonds and CDS trade in line and there might be no
relative value opportunities. However, if the basis is too positive or too negative,
the relative value between bond and CDS should attract investors to do trades that
could cause the absolute value of the basis to decrease. Thus, the basis is meanreverting, implying that investors with shorter trade horizons can also take
advantage of large positive or negative basis.
If the basis is negative, then the CDS spread is lower (tighter) than the bond spread.
To capture the pricing discrepancy when a negative basis arises, an investor could
buy the bond (long risk) and buy CDS protection (short risk) with the same maturity
as the bond. If the basis is positive, then the CDS spread is higher (wider) than the
bond spread. An investor could borrow and short the bond (if possible) and sell CDS
protection (long risk) with the same maturity (or as near as possible) as the bond.
Thus the investor is not exposed to default risk but still receive a spread equal to the
Bond-CDS basis.
Figure 2: Current Basis Distribution
in Europe
X-axis: Basis bucket; Y-axis: % of bonds with
basis within each bucket.
50%
We typically find that negative basis trades are more attractive than positive
basis trades. In a negative basis trade an investor will buy the bond and buy CDS
protection. For a positive basis trade however, an investor needs to repo the bond and
sell CDS protection. The difficulty with repo of corporate bonds and the cheapest-todeliver option that protection buyers own makes positive basis packages more
difficult to analyse and execute.
40%
30%
20%
10%
0%
-800 -600 -400 -200
Negative basis trades have been in the spotlight recently (and still are) due to the
historical levels reached in the Bond-CDS basis since October 2008. Figure 2 shows
the current distribution of basis in the European market. Figure 3 and Figure 4 show
the average historical basis for the European and US bond market. Since October
2008, the basis has reached historical (negative) levels as credit investors have
shunned cash bonds due to their lower liquidity and higher funding requirements
compared to CDS.
Abel Elizalde
(44-20) 7742-7829
[email protected]
In bp.
In bp.
IG
100
50
-200
-50
-400
-100
-600
-150
Jan-07
Source: J.P. Morgan.
IG
200
HY
-800
Jul-07
Jan-08
Jul-08
Jan-09
Jan-07
Jul-07
Jan-08
Jul-08
Jan-09
Negative basis trades have been a popular investment strategy during the last years.
Since the risk on these trades is partially hedged, compared to long risk positions in
bonds or CDS, they were used extensively. The sudden and extreme widening of the
basis to negative territory during the last months caused significant negative MtM
losses to negative basis investors. The combination of low liquidity in the cash
market and tightening funding conditions contributed to unwinds of previously
established negative basis trades.
The current market conditions are very attractive to set up negative basis
packages. However, investors should be fully aware of all the issues affecting
basis trades construction, risks and sensitivities. In this research note, we will
provide a comprehensive analysis of the most important aspects relating basis trades
such as their measurement, their trading, their risks as well as their historical patterns
and future outlook.
Abel Elizalde
(44-20) 7742-7829
[email protected]
The Bond-CDS basis measures the extra compensation CDS investors receive
relative to bond investors. It is expressed as a running spread measure.
Trading the Bond-CDS basis requires investors to:
1.
2.
Choose the notionals in the bond and CDS positions, which will determine
the trade sensitivity to spread movements and to a default, as well as the carry.
3.
The different sections of this report cover all the above considerations in detail. In
Section 1 we take a closer look at what exactly the Bond-CDS basis encapsulates,
how we define it, how we measure it, its main drivers and how it is generally traded.
Section 2 tackles the measurement of the basis by analyzing different spread
measures for bonds (e.g. asset swap spread, Z-spread) and how well suited they are
to be compared with CDS spreads. We present a bond spread measure (par
equivalent CDS spread or PECS) which we consider better suited for measuring the
basis.
Section 3 focuses on the trading aspects of basis trades. We start with a very simple
stylised basis trade example which we use to illustrate all the structural features of
bonds and CDS which affect the mechanics and economics of basis trades. The
section also reviews the most popular motivations to enter into basis trades: lock-in a
risk-free spread over the life of the trade, bet that the basis will revert back to zero
and profit from a default of the company.
In Section 4 we review the historical evolution of the basis and present our outlook
for 2009. Finally, Section 4 provides an overview of the current picture in the
European basis space. We analyse the basis breakdown by sector, rating and
maturity, as well as the most attractive basis trades. We also outline J.P. Morgans
new European Bond-CDS Basis Report, which will highlight on a daily basis the best
trading opportunities and the evolution of the basis per rating, maturity and sector. It
complements our existing US Corporate High Grade Basis Report and US Corporate
High Yield Basis Report for the US market. All of them are available on Morgan
Markets and provide an efficient way to track the Bond-CDS basis on a daily basis.
In the appendices, we include a brief remainder of CDS pricing as well as a
comprehensive review of bond spread measures.
Although we cover the most relevant aspects of basis trading extensively, we do not
analyse with all the detail that they might deserve some other important issues such
as the treatment of basis in the high yield space (where bonds generally have some
embedded optionality), the practical implementation of funding issues surrounding
basis trades, and the hedging of the interest rate and FX risks in basis trades. We plan
to tackle them in future research.
Abel Elizalde
(44-20) 7742-7829
[email protected]
Basis Definition
CDS and corporate bonds are both affected by the credit risk of a company. As the
perceived credit risk of a company increases, CDS spreads rise and corporate bond
prices fall. If CDS and bonds contain the same credit risk, do they price the same
credit risk?1 The Bond-CDS basis aims to be a measure of the discrepancy between
the risks priced into bonds and CDS.
Bond yield components:
In order to isolate the effect of the increased credit risk on a bond, we decompose the
bonds yield into four components:
Risk-Free rate: The bond holder could earn this yield in a default/risk-free
investment (for example, the US Treasury rate).
Swap Spread: The swap spread is the difference between the funding cost of a
AA rated company and the risk-free rate. The swap rate (swap spread + risk-free
rate) is the cost of capital for many investors since they can borrow or lend at this
rate rather than the risk-free rate.
Credit Spread: The spread to compensate for the risk that the company defaults
and investors lose future interest and principal payments. When investors refer to
bond spreads, they usually have some measure of this credit spread in mind.
Bond Liquidity Premium: The spread to compensate investors for the illiquidity
of the bond.2
Credit and liquidity risks are measured by traditional bond spread measures such as
Z-spread and asset swap spread.
CDS spreads are meant to be a clean measure of credit risk, although liquidity risks
can also be priced in.
Leaving aside liquidity considerations, CDS and bond spreads both compensate
investors for the same risk the risk that a company might default. As such, they
should be approximately equal. The Bond-CDS basis measures the extent to which
these spreads differ from each other. We thus define it as the difference, in basis
points, between the CDS spread and the bond spread with the same maturity dates:
In fact, do they actually contain the same risk? Throughout this report, we will dissect the
structural differences between bonds and CDS, which should contribute towards answering
this question.
Our credit strategists estimate that the current liquidity premium in investment
grade bond spreads could potentially range from 100to 200bp. See How big is the
liquidity or illiquidity premium?, P Malhotra et al, 30 January 2009.
7
Abel Elizalde
(44-20) 7742-7829
[email protected]
Our measure of the basis will still provide a measure of the different credit risk
priced in both instruments if we assume the liquidity premium is the same. If that
was not the case the basis will be a combination of credit and liquidity risks. This can
be particularly relevant in situations where the liquidity on each market is very
different.
Whereas CDS spreads are determined by market participants and are readily
observable in the market, bond spreads are a theoretical measure backed out from
bond prices. There are different bond spreads measures which can be used to
compute the Bond-CDS basis.
For the purposes of the Bond-CDS basis, we are concerned with the
comparability of different bond spread measures with CDS spreads. In later
sections we analyse in detail different bond spread measures (e.g. Z-spread and asset
swap spread) and judge their comparability with CDS spreads. In particular, we
highlight the problems of comparing traditional bond spread measures with CDS
spreads since, unlike CDS spreads, their calculation does not explicitly take into
account expected recovery rates and the term structure of default probabilities
(although their prices should do so).
We also propose a bond spread measure which tackles those problems and
represents, in our view, a more appropriate spread measure to be compared to CDS
spreads. We call such measure the bonds par equivalent CDS spread (PECS). Like
the CDS spread, the PECS is a function of the assumed bond recovery rate and the
term structure of default probabilities. J.P. Morgan introduced the PECS in 20053 and
we use it extensively in our research and analytics.
3
See Credit Derivatives: A Primer, E Beinstein et al, January 2005, and US High Yield Spread
Curve Report: A Guide, E Beinstein et al, 29 April 2005.
Abel Elizalde
(44-20) 7742-7829
[email protected]
Effect on Basis
Negative Basis
Negative Basis
Negative Basis
Negative Basis
Negative Basis
Negative Basis
Negative Basis
Positive Basis
Positive Basis
Positive Basis
Positive Basis
Positive Basis
Liquidity Premium
Credit instruments do price a liquidity risk premium on top of the credit risk
premium. The Bond-CDS basis captures all those premiums. Even though liquidity
premiums for bonds and CDS might be similar during normal times, periods of
financial stress can have very different consequences in the liquidity of bonds and
CDS respectively.
A significantly lower liquidity in bonds than in CDS will tend to make the basis
positive in spread widening scenarios, as investors find it easier to buy CDS
protection than to sell their bonds. In a spread tightening environment, the opposite
should be true. Therefore, the different liquidity of bonds and CDS can give basis
trades significant market directionality.
The grey box in the following page contains an extract of our CD Player published
on 8 October 2008, where we reviewed the situation of the Bond-CDS basis at the
time, highlighting the importance of liquidity on the basis.
Extract from CD Player - 8 October 2008: The Basis-to-Cash is at its historical negative levels of the last 2 years. For all practical purposes, CDS
continues to be the only way to short the credit market or to hedge cash positions. The lower liquidity and funding issues in bonds means investors are
asking for a substantial risk premium over CDS, taking the basis to the 53bp area (Figure 1 and Figure 2).
Liquidity has taken over new issuance concerns, which now look less relevant for the basis. It all points out that this situation should persist until the
bond market recovers in terms of liquidity. Some of the negative basis packages established during the last months run the risk of unwinding, which will
not help the basis to trend to positive territory. We expect investors with liquidity reserves to monitor the existing negative basis opportunities, but not to
deploy enough capital to meaningfully move the aggregate market basis as long as liquidity and funding in the bond market do not improve.
Abel Elizalde
(44-20) 7742-7829
[email protected]
Spread (bp)
80
Basis-to-Cash
60
200
60
150
40
20
100
0
-20
50
-40
40
20
0
-20
-40
-60
Jan-07 Apr-07
80
Jul-07
Jul-08
Oct-08
Current
-60
0
50
100
150
200
For an analysis of the impact of synthetic structured products on the Bond-CDS basis, see
Impact of Structured Product Activity on the Credit Markets, D Toublan et al, 23 January
2009.
10
Abel Elizalde
(44-20) 7742-7829
[email protected]
11
Abel Elizalde
(44-20) 7742-7829
[email protected]
Lock-In Risk-free Spread. If bond and CDS share the same credit risk but
they are pricing it differently, it might be possible to construct something akin to
an arbitrage-free trade to profit from it. As we show later, this is not generally
possible due to the trading conventions of bonds and CDS. However, the more
negative the basis the more attractive the basis trade.
2.
Trade the Basis. A negative basis trade (buy bond and buy CDS protection) can
be used to bet that an already negative basis will disappear, or to bet that the
basis will become positive.
For example, CDS spreads might react faster to negative news regarding
corporate events.9 In those cases, the basis can become positive until bond
spreads catch up. A negative basis trade established prior to the negative news
should profit from it.
3.
Profit from Default. If the bond and CDS legs of a basis trade are done in the
same notional, the investor can, after a default, deliver the bond to the CDS
counterparty and both legs of the trade will terminate with no further payment.
In that case, the investors gain will be the net cash flows the trade generated up
to that point. If the investor expects the default to happen soon, a short maturity
CDS can be more economical if the CDS spread curve is steep enough.
The sizing of the basis trades will be key in determining their performance.
In the next section we tackle the measurement of the basis by analyzing different
spread measures for bonds (e.g. asset swap spread, Z-spread, PECS) and how well
suited they are to be compared with CDS spreads.
The issue of whether CDS spreads react faster than bond spreads is not entirely clear cut, and
it depends on the relative liquidity of bonds and CDS on each particular market. For example,
the lower liquidity of the European bond market vs. the US one will affect such relationship.
There are some empirical academic papers which analyse this issue. See R Blanco, S Brennan
and W Marsh, 2005, An Empirical Analysis of the Dynamic Relation between Investment
Grade Bonds and Credit Default Swaps, Journal of Finance 60 (5). See also S Alvarez, 2004,
Credit default swaps versus corporate bonds: Do they measure credit risk equally?,
unpublished manuscript.
12
Abel Elizalde
(44-20) 7742-7829
[email protected]
Bond-CDS Basis
CDS Spread
minus
Comparable Cash Bond Spread
1.
Z-spread
2.
3.
Others include spread to benchmark, I-spread and True ASW. ASW and Z-spreads
have traditionally been the most widely used bond spread measures when dealing
with basis trades. However, these spread measures have features which make it
difficult to have a like-for-like comparison with CDS spreads. In particular, their
calculation does not explicitly account for expected recovery rates or the term
structure of default probabilities, which are key determinants of CDS spreads. PECS
can be thought of as a bond credit spread measure consistent with the recovery rate
and term structure of default probabilities priced into the CDS market.
Appendix I includes a reminder of CDS pricing, and Appendix II provides a
detailed analysis of each of the above bond spreads measures. Here, we
summarise the differences between CDS spread, Z-spread, ASW and PECS.
20%
10% PD
When trading and quoting CDS spreads, market participants do so under a recovery
rate assumption. As we outline in Appendix I, CDS spreads are a function of
recovery rates and default probabilities. In a simple one-step time period example,
Equation 1 shows that the CDS spread (S) equals the default probability (PD) times
the loss in case of default, given by one minus the expected recovery rate (R).10
Equation 1: CDS Spread as a Function of Default Probability and Recovery Rate
Simple one-step time period example.
S = PD (1 R )
20% PD
15%
10%
5%
0%
0%
Keeping default probabilities constant, changes in expected recovery rates will affect
CDS credit spreads. This effect is larger for high default probabilities (Figure 7).
When the likelihood of default is high enough, estimates of recovery rates are more
important and the cheapest-to-deliver optionality in CDS contracts is priced in more
accurately. Thus, expected recovery rates affect CDS spreads. For a better
comparison with CDS spreads, bond spread measures should explicitly take into
account assumed recovery rates.
10
Abel Elizalde
(44-20) 7742-7829
[email protected]
In an equal notional basis trade where the investor plans to deliver the bond into the
CDS contract in case of default, the assumed recovery rate should not play such a big
role (except for cheapest-to-deliver considerations and assuming the bond is
deliverable into the CDS contract). Movements on the assumed recovery rate should
therefore not affect our measure of the basis. Since recovery rate changes do affect
CDS spreads, in order to compute the basis we would prefer a bond spread measure
which is also sensitive to the assumed recovery rate.
Additionally, CDS spreads for a given tenor can not be considered in isolation from
the full CDS curve, especially when analyzing trades where there is a regular stream
of risky running payments; which is the case in basis trades. The shape of the term
structure of CDS spreads determines default probabilities over time, i.e. the
likelihood of those future running payments being realised. For example, Figure 9
shows the cumulative default probabilities implied from the three CDS curves in
Figure 8. The three CDS curves have a similar 1y spread but different shapes.
Figure 8: CDS Curves
20%
15%
Flat
Dow nw ard
Upw ard
10%
20%
15%
10%
5%
5%
De
c-0
8
M
ar
-0
9
Ju
n09
Se
p09
De
c-0
9
0%
0%
20-Dec-08
20-Mar-09
20-Jun-09
20-Sep-09
20-Dec-09
Therefore, looking at a CDS spread for a particular tenor in isolation will not give us
the full information regarding default probabilities and therefore regarding credit
risk.
As we review in the next section, the shape of the term structure of default
probabilities will impact the economics of the basis trade as long as the timing of
default has an impact on the trades performance.
The difficulty here lies in which term structure of default probabilities to use: the one
implied by bond spreads of different maturities or by CDS spreads. Either one would
add value versus not using it. The PECS measure we introduce later uses the term
structure of default probabilities implied by the CDS curve; this is generally more
convenient as CDS curves are more readily available.
14
Abel Elizalde
(44-20) 7742-7829
[email protected]
Full Running =
Upfront - AI
+ Fixed Coupon
RA
As we explain in the next section, the CDS trading format does have an impact on
the economics of basis trades. Thus, the trading format matters even if it implies the
same full running spread. This illustrates the fact that basis measures, expressed in a
full running format, are useful to judge the attractiveness of basis trades, but they are
not enough.
Next, we look at three alternative measures of bond credit risk to asses the best
comparison to CDS.
15
Abel Elizalde
(44-20) 7742-7829
[email protected]
Z-spread
The Z-spread is the parallel shift applied to the zero curve in order to equate the
bond price to the present value of the cash flows.11
We take the zero curve as an input and add a flat credit risk premium (the Zspread) for which the bonds discounted cash flows match its market price.
Effectively, the Z-spread can be thought of as the flat spread that can be added to the
risk-free curve to capture the risks of the bond apart from interest rate risk.
The Z-spread accounts for the term structure of interest rates, but assumes a flat term
structure of credit spreads (assuming credit is the only additional risk priced in by the
bond). Thus, it does not explicitly take into account the term structure of default
probabilities.
Figure 10: Z-spread as a Set of Risky Discount Factors
10%
Risk-Free Curv e
Risky Curv e
8%
6%
4%
2%
0%
0
When computing the Z-spread, we do not take into account the possibility of the
bond defaulting (i.e. we assume zero default probabilities) or, if we do take such
possibility into account, we assume a zero recovery rate. In any case, the expected
recovery rate is not explicitly taken into account in the Z-spread calculation.
Z-spreads are useful for comparing the relative value of bonds as they take into
account the full term structure of the risk-free rates. However, Z-spreads are not
traded in the market.
11
As the Z-spread is just a spread above a given risk-free rate, we can reference it either to
LIBOR/Swap zero rates or to government zero rates.
16
10
Abel Elizalde
(44-20) 7742-7829
[email protected]
Spread
+
Libo r
+
100
100 (Par)
Upfront
payments
100
(Dirty Price)
100
Coupons
+
Cpn
+
Cpn
100
(Dirty Price)
Source: J.P. Morgan.
Let us consider an investor who is entering into an asset swap package and buying
the bond at the same time. In the asset swap package, the investor pays 100 and
receives the bond price P, which is used to buy the bond. Therefore the investors
payment at inception is 100.
We assume the bond and asset swap package have the same maturity and payment
dates. The bond pays an annual coupon c on the 100 notional, which the investor
transfers through the asset swap package in exchange for Libor l plus the asset swap
spread s on a 100 notional. Therefore he receives a net amount of 100 x (l + s - c)
per year in the asset swap and receives 100 x c on the bond as long as there has
been no default. At maturity, the investor receives the notional from the bond.12
12
An asset swap package is an interest rate swap with 100 notional with an initial cost of
(100 P) where the investor pays fixed coupons (c) and receives floating coupons (l + s).
17
Abel Elizalde
(44-20) 7742-7829
[email protected]
On a net basis, the investor has paid 100 to buy the bond and the asset swap
package, and receives an annual coupon of 100 x ( l + s). Effectively, it has
transformed the fixed coupon bond into a floating one, hedging the interest rate risk.
The asset swap spread s is the extra-compensation above the (Libor) risk-free rate,
and therefore it can be interpreted as a measure of the bonds credit risk. However,
notice that such spread is being paid on the bonds notional value, not on its price P.
Par asset swap spreads are useful as they can be traded. An investor can find a dealer
who will pay him the annual par asset swap spread. In Appendix II, we show that the
asset swap spread can be computed as:
Equation 3: Par Asset Swap Spread Calculation
where the annuity used here is the risk-free annuity (present value of a 1bp annuity
stream) and PV represents the present value of the bonds future cash flows using the
risk-free discount curve.
The asset swap package does not go away in case of default. In the extreme case that
the bond defaults immediately after entering the asset swap package, the interest rate
swap part of the package remains in place with the same value, but the bonds value
goes down from P to R (its recovery price). Therefore, the loss for the investor
upon an instantaneous default is (P R). Although the loss is a function of the
bonds recovery rate, the asset swap spread does not explicitly take it into account.
The loss in case of default in a CDS position with a 100 notional would be (100
R), i.e. it is a function of the recovery rate. Unlike the ASW, the CDS spread is
affected by the assumed recovery rate.
The asset swap spread is an equivalent measure for the credit risk of a bond and,
unlike the Z-spread, it is a traded measure. However, like the Z-spread, its
computation does not explicitly take into account the expected recovery rate and the
term structure of default probabilities. As the Z-spread, the asset swap spread
represents a flat credit spread measure.
18
Abel Elizalde
(44-20) 7742-7829
[email protected]
In this section, we have argued that bond spread measures that explicitly take into
account the term structure of default probabilities and the assumed recovery rates are
more appropriate for measuring the Bond-CDS basis than other measures, such as Zspreads or asset swap spreads, which do not. In the next grey box, we expand our
argument regarding the recovery rate issue.
Recovery Rates in Bond and CDS Pricing
Both traded CDS spreads and bond prices do factor in some assumption not only about the expected recovery rate, but also about its distribution
function and its relationship with default and interest rates risk. However, when modeling bond and CDS prices, those recovery rate assumptions can not
generally be disentangled from the assumptions regarding default risk.
The market convention for CDS pricing is to assume an expected recovery rate (independent of default and interest rate risks) and derive,
from traded levels, information regarding default probabilities. Traders and investors will change their recovery rate assumptions to reflect changing
market conditions, especially in distressed environments like the current one where recovery rates become very important for pricing credit risk.
These changes in CDS recovery rates assumptions do not necessarily come with changes in spreads. If, for example, a trading desk or an investor
decides to mark down their recovery rate assumption on one particular credit from 40% to 20%, their pricing models will use this new assumption to
calibrate default probabilities to the same CDS spread. The new calibrated default probabilities (with a 20% recovery rate) will be lower than before.
When changing the recovery rate assumptions the credit risk priced into the CDS has not changed (it has the same spread), but the allocation
of that risk between default and recovery risks has changed. Using the simple one-step time period example Equation 1 and a spread level of
1000bp, changing the recovery rate from 40% to 20% moves the calibrated default probability from 17% to 13%.
Bond-CDS basis trades can be viewed as risky annuity trades. Let us take a simplistic example for illustration purposes. Imagine a negative basis
trade where the investor buys a 5% coupon bond and buys CDS protection trading at 2% running spread on the same notional and with the same
maturity. Let us not worry now about funding costs, interest rate riskThe trade involves an initial payment equal to the bond price plus a risky annuity of
3% per annum.
As we show later, this trade is not exposed to the realised recovery rate on default: the investor delivers the bond into the CDS contract and gets par,
irrespective of the realised recovery rate. However, the risky annuity is very sensitive to default probabilities and, as we argued above, default
probabilities are very sensitive to the assumed recovery rate.
Thus, changes in the assumed recovery rate will affect the value of the risky annuity and therefore the MtM of a basis trade. Since the assumed
recovery rate is an important element for basis trades, we would prefer to use spread measures (both for CDS and bonds) which do explicitly incorporate
the assumed recovery rate.
Bond measures like Z-spread or asset swap spread do not incorporate such assumption explicitly, even though they are indirectly affected by it (because
it will affect the bond price). But if the assumed recovery rate changes, the distribution of credit risk between default and recovery rates will change, and
those spread measures will not necessarily capture it.
The bond spread measure that does explicitly take into account both the term
structure of default probabilities and the assumed recovery rates is the PECS, which
we turn to next.
19
Abel Elizalde
(44-20) 7742-7829
[email protected]
Using the full CDS curve traded in the market and a recovery rate assumption,
we calculate the implied default probabilities for the company.
2.
Using an iterative process, we identify the parallel shift to the default probability
curve which will make the bond implied price equal to its market price.13
3.
Once we have matched the bond price, we convert these default probabilities
back into spreads.
The PECS is the CDS spread which would match the bond market price
respecting the recovery rate and term structure of default probabilities implied
by the CDS market.
13
In particular, we additively shift the hazard rates which characterise the term structure of
default probabilities.
20
Abel Elizalde
(44-20) 7742-7829
[email protected]
In bp.
In bp.
Z-Spread
300
1000
Z-Spread
900
250
800
200
700
150
600
100
500
50
PECS
0
0
50
100
150
200
250
Source: J.P. Morgan. Sample of over 500 European bonds. Data as of 6 Jan 09.
300
400
PECS
300
300
400
500
600
700
800
900
1000
Source: J.P. Morgan. Sample of over 500 European bonds. Data as of 6 Jan 09.
Figure 12 and Figure 13 compare Z-spreads and PECS for a sample of European
bonds. They show how the observations are spread around the 45% degree line (in
black), meaning that there is a clear correlation between both spread measures.
However, the deviations from the 45% degree line are widespread and can be
significant, independently of the spread levels. A similar picture emerges when
comparing ASW and PECS. Different spread measures aim at measuring the credit
risk of the bond, but the way they are constructed implies they do not measure
exactly the same thing. We will also see discrepancies between these different bond
spread measures in the next section.
As long as different bond spread measures exist, there will be different ways to
measure the Bond-CDS basis. We have reviewed the most common ones and their
advantages and disadvantages. Although the basis is a good indicator of the different
risks priced in bonds and CDS, we show in the next section that the structural
features of bonds and CDS are a relevant factor for the risk exposures of basis trades.
21
Abel Elizalde
(44-20) 7742-7829
[email protected]
2.
Full upfront CDS with the same maturity as the bond. Zero running spread and
no CDS margin requirements.
3.
Equal notional trade: the bond and CDS notional are the same.
4.
Flat 0% interest rate, which we assume remains fixed. This removes any interest
rate risk on the trade and allows the investor to borrow money to pay for the
bond, CDS upfront (and any CDS margin) without affecting the economics of
the trade.
5.
6.
Under the above assumptions, we can compare the bond price and the CDS upfront
directly without the need to transform them to spread levels. For example, if the bond
price is 94 and the CDS upfront premium is 5, an investor can: borrow 99 at no
cost, buy the bond and buy CDS protection to lock a 1 risk-free profit:
a.
In case of no default the investor receives, at maturity, par (100) on the bond,
pays back the borrowed 99 and keeps 1.
b.
If there is default before maturity, the investor delivers the bond into the CDS
contract and gets paid the contracted CDS notional (100). The investor pays
back the borrowed 99 and keeps 1.
In either circumstance the investor has locked 1 of risk-free profit; there are no
initial or interim payments in the transaction until maturity or default.
22
Abel Elizalde
(44-20) 7742-7829
[email protected]
Arbitrage?
In this simple example the focus has been on arbitrage, which would involve
trading the same notional on each product. In what follows, we continue with this
example and introduce different issues which increase the complexity of the trade
and eliminate the arbitrage opportunity.
Treatment upon default: while bond coupons are usually lost, the CDS protection buyer has to pay for the accrued
coupon since the last coupon date.
The investor benefits if the bond in the basis trade recovers more than the cheapest-to-deliver.
Maturity Mismatch
Very rarely investors will be able to exactly match bond and CDS maturities in a basis trade. This introduces a residual
naked long or short protection position for the last period of the basis trade, which will affect its economics.
Funding costs increase the exposure of a basis trade to the timing of default, and reduce the attractiveness of negative
basis trades.
A higher CDS running spread generates a negative basis trade with a worse carry profile (more negative or less
positive) but a better (less positive or more negative) JtD sensitivity during the first part of the trade.
Bonds with higher price and coupon generate a negative basis trade with a better carry profile (less negative or less
positive) but a worse (less positive or more negative) JtD sensitivity during the first part of the trade.
Basis trades are subject to interest rate risks in several dimensions such as funding or bond and CDS sensitivity to
interest rates. If bond and CDS do not refer to the same currency, the basis trade will be subject to movements in the
exchange rate.
Bond holders might have actual or potential rights, and bonds may have embedded options such as callable, puttable,
poison puts, etc. These features can affect the economics of basis trades.
If the bond in the basis trade is not deliverable into the CDS contract, the investor is exposed to the different recovery
rate of their bond and the cheapest-to-deliver (or to the result of CDS cash auction if they decide to cash settles their
CDS position).
If a restructuring event occurs, CDS contracts specify restrictions regarding which obligations can be delivered into the
contract by CDS protection buyers. These restrictions (e.g. on the maturity of the deliverable obligations) can affect the
bond in the basis trade.
Breaching a bond covenant does not necessarily trigger a CDS credit event.
Funding Costs
CDS Upfront vs. Running Premium: Carry vs.
Jump-to-Default (JtD)
Bond Price vs. Coupon: Carry vs. Jump-toDefault (JtD)
Interest Rate, FX Risks
Bondholders rights & Bond embedded options
Bond deliverability into CDS contract
CDS Restructuring Credit Events
Bond Covenant Breaching vs. CDS Credit Events
Source: J.P. Morgan.
1.
We assume that the above bond and CDS have a 5% annual coupon and running
spread respectively, maintaining the same price and upfront premiums above.
Moreover, we also assume that both bond and CDS have similar quarterly coupon
payment dates (20 March, 20 June, 20 September, 20 December) and that we enter
into the trade on 20 December 2008, i.e. there are no accrued components on the
bond coupon or CDS running spread.
The key difference between bond and CDS running coupons is their treatment upon
default: while bond coupons are lost, the CDS protection buyer has to pay for the
accrued coupon since the last coupon date. Other things equal, the lower and the
more frequent the bond coupons the better for the investor in a negative basis trade.
Figure 14 shows the total cash flows of our negative basis trade for different default
dates. We draw readers attention to the grey box on page 25, where we explain
carefully how to read figures which, like Figure 14, show the default exposure of
negative basis trades over time. We will use these figures throughout the rest of the
report.
23
Abel Elizalde
(44-20) 7742-7829
[email protected]
If the default happens right after a coupon payment date or after maturity, the
investor makes a similar amount of money, 1, as in the previous case (where there
were no coupons). However, if the default happens within coupon dates, the investor
losses the bond accrued coupon but has to pay the CDS accrued coupon. Thus, the
worst case is one where the default happens the day before a coupon payment date.
The 1 that the investor makes upon default is not enough to compensate for the full
1.25 quarterly bond coupon lost.
Treatment upon default: while bond coupons are usually lost, the CDS
protection buyer has to pay for the accrued coupon since the last coupon date.
Figure 14: Bond Coupons and CDS Running Spreads: Impact on Basis Trades (JtD Exposure)
Y-axis: Sum of total cash flows on the negative basis trade until default; X-axis: Assumed default dates.
Bond coupons are lost upon
default; however CDS protection
buyers pay for the accrued
coupon since the last coupon
date
1.50%
1.50%
1.25%
1.00%
1.25%
1.00%
0.75%
0.75%
0.50%
0.25%
0.50%
0.25%
0.00%
-0.25%
0.00%
-0.25%
-0.50%
-0.50%
20-Dec-08
20-Mar-09
20-Jun-09
20-Sep-09
20-Dec-09
Finally, two additional issues should be considered. The frequency and payment
dates on bond and CDS can be different. Bond coupons have different accrual day
conventions (e.g. 30/360, Act/Act), whereas CDS coupons (running spread) accrue
on an actual/360 convention.
24
Abel Elizalde
(44-20) 7742-7829
[email protected]
The total cash flow will include all of the cash flows from inception up to, and including, the default date: upfront
payments on bond and CDS, bond coupons and CDS running spread, funding costs, CDS margin, and payments on default.
When the date shown in the x-axis is past the trade maturity, the figure will effectively show the total net cash flows of the
trade if there is no default during the life of the trade.
In case of default, and obviating any risk-free discounting, the profit (or loss) for the investor in a basis trade is given by
Equation 4.
Equation 4: Basis Trade Profit on Default
CDS Notional x (100 Recovery CDS Upfront CDS Coupons Paid CDS Funding Costs Paid)
+ Bond Notional x (Recovery + Bond Coupons Received Bond Price Bond Funding Costs Paid)
Note: Bond Price refers to the dirty bond price.
If both legs are done on the same notional, the final profit is independent on the recovery rate of the bond.
Negative Basis Trades Total Cash Flow at Maturity
If there is no default during the life of the trade, and again ignoring any risk-free discounting, Equation 5 shows the total
cash flows of a negative basis trade at maturity (i.e. after both the bond and CDS legs have expired).
Equation 5: Basis Trade Profit on Maturity
Bond Notional x (100 + Bond Coupons Received Bond Price Bond Funding Costs Paid)
CDS Notional x ( CDS Upfront + CDS Coupons Paid + CDS Funding Costs Paid)
Note: Bond Price refers to the dirty bond price.
It can be shown that Equation 4 and Equation 5 generate the same results for equal notional basis trades. Notice that the
amount of bond coupons received, CDS coupons paid and funding costs paid will be different in both cases (since they
depend on the timing of default in Equation 4).
25
Abel Elizalde
(44-20) 7742-7829
[email protected]
2.
The assumption of a similar recovery rate in bonds and CDS can be challenged. As
we explained before, the cheapest-to-deliver option that CDS protection buyers
enjoy represents a potential extra-benefit in a negative basis trade in case of default.
Figure 15 shows the total cash flows of our negative basis trade for different default
dates. We compare two scenarios under the previous case (bond and CDS with 5%
coupon). These scenarios are: (i) bond and CDS recovery rates are similar, and (ii)
the cheapest-to-deliver bond has a recovery rate 5% lower than the bond in the
negative basis trade. As Figure 15 shows, the lower the recovery rate of the bond in
the basis trade compared to the cheapest-to-deliver bond, the better for the negative
basis investor.
Figure 15: Cheapest-to-delivery Option: Impact on Basis Trades (JtD Exposure)
Y-axis: Sum of total cash flows on the negative basis trade until default; X-axis: Assumed default dates.
Negative basis trades benefit if
the bond is not the cheapest-todeliver
Benefit: bond recovery cheapestto-deliver recovery
7%
7%
6%
6%
5%
4%
5%
4%
3%
Cheapest to deliv er bond recov ers 5% less than the basis trade bond
3%
2%
2%
1%
0%
1%
0%
-1%
20-Dec-08
-1%
20-Mar-09
20-Jun-09
20-Sep-09
20-Dec-09
14
Fannie and Freddie CDS Settlement Auctions, E Beinstein et al, 2 October 2008. Fannie
Mae and Freddie Mac CDS Settlement, E Beinstein et al, 6 October 2008.
26
Abel Elizalde
(44-20) 7742-7829
[email protected]
3.
Maturity Mismatch
Very rarely will investors be able to exactly match bond and CDS maturities in
a basis trade. If the bond maturity does not coincide with a standard CDS maturity,
investors will have to decide either to buy CDS protection in the previous or next
maturity around the bond maturity. This introduces a residual naked long or short
protection position for the last period of the basis trade, which will affect its
economics.
Investors might prefer to pair long dated bonds with short dated CDS in some cases
(e.g. to play a positive jump-to-default JtD negative basis trade with a short term
trading horizon). In that case, the investor is effectively entering a negative basis
trade and a curve flattener trade on the name. Whether the investor wants to profit
from an early default or just a correction of a negative basis dislocation will
determine the notional used in both legs (e.g. equal notional vs. duration neutral).
4.
Funding Costs
When interest rates are positive, funding (i) the bond purchase (or bond margin), (ii)
the CDS upfront cost (and/or running spreads) and (iii) the CDS margin will affect
the economics of a negative basis trade.
The margin that the investor posts for the CDS trade will depend on its credit quality
and the credit risk of the reference company. The margin is generally applied to the
CDS notional at risk: difference between notional and upfront premium (95 in our
example). The margin will differ if the investor is buying or selling protection.
In our stylised example, assuming a 5% CDS margin the investor would need to fund
103.75 (bond price 94 + CDS upfront 5 + CDS margin 4.75)15. Assuming a flat
and fixed 5% interest rate swap curve at which the investor can fund, the total cost of
funding the trade if there is no default, i.e. for one year, is 5.1875 (= 103.75 * 5%).
Such cost will completely outweigh the 1 risk-free profit of our original stylised
example.
Generally, the sooner the default happens, the lower the funding costs ultimately
paid. If the default occurs immediately after the trade is entered into, funding costs
will be pretty much zero and the investor will make 1, i.e. the same as in our
stylised example.
shows the total (not discounted) cash flows of our negative basis trade for different
default dates for our stylised example (i.e. no funding costs) and for the case where
the investor has funding costs. The sooner the default occurs, the better for the
investor.
15
CDS margin in this example (4.75) is calculated as 5% of the CDS notional at risk (95),
which is the difference between the CDS notional (100) and the upfront premium (5).
27
Abel Elizalde
(44-20) 7742-7829
[email protected]
2%
2%
1%
1%
0%
0%
-1%
-1%
-2%
-2%
-3%
-3%
-4%
-4%
-5%
-5%
20-Dec-08
20-Mar-09
20-Jun-09
20-Sep-09
20-Dec-09
The risk-free negative basis trade of our stylised example has effectively turned
into a timing of default trade: the investor makes money only if a default happens
soon enough. For a very high funding cost the negative basis trade might be
completely unattractive.
An alternative way of funding all or part of the bond purchase is through a bond repo
agreement.
The repo counterparty will take the bond as collateral and lend to the investor the
bond price minus a given haircut the investor has to pay. Effectively, the investor
will be funding part of the bond price at the repo rate and the rest at his standard
funding rate. The repo rate will be a function of the investors credit quality, the
credit quality of the bond and the length of the loan. Repo financing is generally done
on short terms (e.g. one week) and rolled over. If the investor can not secure funding
for a long enough period of time, he runs the risk of running out of funding and
having to sell the bond.
Funding costs increase the exposure of a basis trade to the timing of default.
For fully funded investors with capital to put at work, the cost of funding effectively
becomes an opportunity cost for the capital invested in the basis trade. This will
probably be lower than the cost of funding, but it can be significant anyway.
The following box contains an extract of CMOS (Credit Markets Outlook and
Strategy, E Beinstein et al) published on 31 October 2008, where the impact of
funding conditions on the basis is analysed.
28
Abel Elizalde
(44-20) 7742-7829
[email protected]
Extract from CMOS - 31 October 2008:16 The financing of the bond has changed significantly. Last year, investors could effectively get 20x leverage (a
5% haircut) with a financing cost of Libor. Now, investors may get 4x leverage (25% haircut) with a financing costs of Libor + 125bp. With 5Y Libor
rallying about 125bp over the year (effectively offsetting the increase in cost over Libor), financing costs have increased.
Exhibit 1: Repo is not available in most situations. When it is, it is more expensive than even one month ago
07-Jun
07-Dec
08-Jun
08-Sep
Current
Approx Haircut
5%
8%
10%
12-15%
20-25%
Approx Spread
LIBOR Flat
L+10bp
L+15-20bp
L+35-50bp
L+100-125bp
CDS trading requires more cash as well. Last year, an investor buying protection was often not required to post an initial margin. Now, short risk
positions may require collateral posting of 2-10% or more of the notional amount of the trade. Thus, costs have increased for CDS as well.
Given these assumptions, we calculate that basis must be about -155bp for investors to earn returns similar to those earned basis of -30bp when
financing conditions were easier. We calculate this by estimating the amount of capital required to establish the bond, interest rate swap, and CDS
position in January 2007 and currently. The amount of capital required is about 5x higher if it is available at all. In other words, basis must be about 155bp for investors to earn returns similar to those earned basis of -30bp when financing conditions were easier. This should be viewed as boundary for
basis (basis should be more negative), in our view, if the currently tight funding conditions persist.
Exhibit 2: The capital required to establish a negative basis trade (bond + interest rate swap + CDS) has increased almost 5x
Jan-07
Oct-08
Bond
Price of bond
Capital required
Leverage
$100
$5
20x
$100
$25
4x
$100
0%
$100
1%
CDS
notional amount
Capital required (%)
$100
1%
$100
5%
$6
$31
-30
-155
5.0
5.0
16
Updated with High Grade Bond and CDS 2009 Outlook, E Beinstein et al, 5 December
2008; pp. 21-23.
29
Abel Elizalde
(44-20) 7742-7829
[email protected]
5.
The cash flow structure of both legs in a negative basis trade (bond and CDS) will
have a significant impact on the attractiveness of the basis trade. Our initial stylised
example represented the extreme case where there are no running payments during
the trade, i.e. the trade carry is zero.
Here we look at the difference between the CDS trading on a full upfront (5%) or a
full running basis (e.g. with a 5.13% coupon). Both positions would be equivalent in
terms of their expected losses. Figure 17 shows the total cash flows of our negative
basis trade for both cases.
Figure 17: CDS Upfront vs. Running: Impact on Basis Trades (JtD Exposure)
Y-axis: Sum of total cash flows on the negative basis trade until default; X-axis: Assumed default dates.
CDS Running Spread: Higher
initial JtD and less attractive
carry profile
8%
8%
6%
6%
4%
4%
2%
2%
0%
0%
-2%
20-Dec-08
-2%
20-Mar-09
20-Jun-09
20-Sep-09
20-Dec-09
In the case of a full running CDS spread, an instantaneous default will make the
investor earn the difference between par and the bond price (6). If no default occurs,
the investor still receives 6 at maturity but he will have paid 5.13 for the CDS
protection, i.e. a 0.87 gain.
In the full running case, the sooner the default the better, as the investor stops paying
for the CDS protection. A basis trade with a full running CDS introduces an exposure
to the timing of defaults. In general, swapping upfront payments into running
payments (both in the bond and CDS leg) will expose the investor to default timing.
If there is no default, the investor will generally end up paying more for the CDS
protection in a full running contract than in the full upfront case. Therefore, there will
be a cut-off date such that: if the default happens before it, the full running CDS will
be more attractive, but if the default happens afterwards, the full upfront CDS will be
preferred. Such date will actually be given by the duration of the CDS, which in our
example was 0.97 (i.e. almost one year). The higher the CDS spread the lower its
risky duration.
Compared to a CDS trading on an upfront (or upfront plus running) format, an
equivalent full running CDS spread generates a negative basis trade with a
worse carry profile (more negative or less positive) but a better (less positive or
more negative) JtD sensitivity during the first part of the trade.
30
Abel Elizalde
(44-20) 7742-7829
[email protected]
6.
In the previous example, we analysed the trade-off between JtD and carry when CDS
trades on an upfront or running basis. There is a similar trade-off for bonds
depending on the relationship between price and coupon.
There are many variations of the price-coupon relationship which would involve the
same credit risk as measured, for example, by the bond Z-spread or PECS. In our
stylised example, a 94 zero-coupon bond corresponds to a Z-spread of 7% and a
PECS of 6.4% (assuming all CDS maturities trade at 5% full upfront). A par bond
with a 6.53% quarterly coupon would have a similar PECS of 6.4%.17 Figure 18
shows the total cash flows of our negative basis trade for the two cases (assuming a
full upfront 5% CDS).
Figure 18: Bond Price-Coupon: Impact on Basis Trades (JtD Exposure)
Y-axis: Sum of total cash flows on the negative basis trade until default; X-axis: Assumed default dates.
2%
0%
0%
-2%
-2%
-4%
-4%
-6%
-6%
20-Dec-08
20-Mar-09
20-Jun-09
20-Sep-09
20-Dec-09
In the case of a par bond with a 6.25% coupon, an instantaneous default will lose the
investor the 5 paid for the CDS protection. If no default occurs during the life of the
trade, the investors will pocket the 6.25 bond coupon minus the 5 CDS protection.
Bonds with higher price and coupon generate a negative basis trade with a
better carry profile (less negative or less positive) but a worse (less positive or
more negative) JtD sensitivity during the first part of the trade.
As before, there will be a cut-off date such that, if default happens before it the
investor would prefer the zero-coupon bond relative to the coupon bond, and vice
versa.
Basis trades tend to be measured, analysed and ranked in terms of the credit risk
priced in each leg (bond and CDS) through a credit risk spread measure. However,
this might not be enough: different combinations of price and coupon in bonds, and
upfront and running in CDS generate material differences in a basis trade, in
particular regarding carry and JtD exposure.
17
Abel Elizalde
(44-20) 7742-7829
[email protected]
7.
Basis trades are subject to interest rate risks in several dimensions such as funding or
bond and CDS sensitivity to interest rates (which in the case of the CDS will be
minimal and in the case of the bond will be a function of whether the bond has a
fixed or floating coupon). If bond and CDS are not denominated in the same
currency, the basis trade will be subject to movements in the exchange rate.
Through this report we assume these risks are being hedged. In case of default, the
investor will have to unwind the hedge, which will likely have a MtM impact.
8.
Bond holders have actual or potential rights, e.g. negotiating rights after bankruptcy,
contingent payments if the company changes the debt terms, tender offers, etc. CDS
protection sellers do not have any such rights.
Bonds may have embedded options such as callable, puttable, poison puts, etc. They
may also have step-up coupons if the company ratings are downgraded. CDS
contracts do not have such features.
In this report, we do not consider the impact of bond optionality in basis trades. The
standard bond spread measures that we consider to characterise Bond-CDS basis
(e.g. Z-spread, asset swap spread, PECS) do not take into account options embedded
in the bond.
9.
When analyzing basis trades, investors implicitly assume that in case of default they
will be able to deliver the bond into the CDS contract. This will be the case if the
bond is deliverable upon default. CDS contracts specify a list of characteristics that
bonds must satisfy to be deliverable into the CDS contract upon default.
Among others, CDS contracts generally specify the following deliverable obligation
characteristics: Not subordinated, standard specified currency, transferable, 30 years
maximum maturity, not bearer and not contingent.
Investors should check whether the bond they are buying satisfies these deliverability
conditions, otherwise the economics of the basis trade can change dramatically on
default. In particular, the investor is exposed to the different recovery rate of their
bond and the cheapest-to-deliver (or to the result of CDS cash auction if they decide
to cash settle their CDS position).
10. CDS Restructuring Credit Events
CDS contracts generally include restructuring within the different credit events that
trigger the contract. Without going in detail into the legal fine print relating a
restructuring event, it refers to a change (or changes) in the terms of one or more
obligations which negatively affect the holders of such obligation and it is not
expressly provided for under the terms of such obligation. These changes relate to,
for example, reduction in the obligation coupon payments, reduction in the obligation
principal payable at maturity, postponement or deferral of payments, changes in
subordination and currency of the obligation. Investors should check the complete
legal implications of the CDS contract clauses and definitions relating to
restructuring. For example, the type of restructuring specified in the contract might
affect the type of obligations which can trigger a restructuring credit event.
32
Abel Elizalde
(44-20) 7742-7829
[email protected]
33
Abel Elizalde
(44-20) 7742-7829
[email protected]
As we hinted in the examples included in this section, looking at basis trades through
a single basis number is a simplification which potentially leaves important
characteristics un-accounted for. Basis trades generate a distribution of future risky
cash flows.
In the rest of this section, we review the three main ways of trading and structuring
basis trades, each with a different rationale:
1.
2.
3.
34
1.
2.
3.
5% CDS margin.
4.
5% cost of funding available on a daily basis throughout the life of the trade.
The investor funds all upfront payments: bond price, CDS margin and upfront (if
any).
5.
6.
Both bond and CDS have the same quarterly coupon payment dates (20 March,
20 June, 20 September, 20 December) and the same maturity (20 December
2013). We enter into the trade on 20 December 2008, i.e. there is no bond
accrued coupon or CDS running spread.
7.
Abel Elizalde
(44-20) 7742-7829
[email protected]
Figure 19 and Figure 20 show the total cash flows of the trades assuming there is no
default during the life of the trade, and including carry, margin and funding costs.18
Since we assume the investor is funding the whole trade without committing any
money upfront, the payment due to the difference between the bond price and
notional occurs at maturity (or at default time if before maturity). 19
18
In Figure 19, for example, the trade involves a quarterly negative carry or around -1.6%,
which is the sum of the CDS quarterly coupon (-1.25% = 5% annual spread * year), the
bond quarterly coupon (+0.5% = 2% annual coupon * year), the CDS margin (-0.0625% =
5% margin * 5% funding cost * year) and the bond funding (-0.78% = 62.31% price * 5%
funding cost * year).
19
In Figure 19, for example, the trade involves a payment at maturity equal to 36.1%, which is
the sum of the CDS quarterly coupon (-1.25%), the bond notional plus the quarterly coupon
(+100.5%), the CDS margin (-0.0625%), the bond margin (-0.78%) and the bond price which
settles the bond funding (-62.31%).
35
Abel Elizalde
(44-20) 7742-7829
[email protected]
Figure 19: Low Price & Coupon: Basis Trade Cash Flows
Figure 20: High Price & Coupon: Basis Trade Cash Flows
40%
2%
0%
30%
-2%
20%
-4%
10%
-6%
0%
-8%
-10%
-10%
Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec08
09
09
10
10
11
11
12
12
13
13
Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec08
09
09
10
10
11
11
12
12
13
13
Bonds with high price and coupons will generate a better carry profile over the
life of the trade, but a worse exposure to default.
Figure 21 shows the total (not discounted) cash flows of our negative basis trade for
the two different bonds.20 If there is no default, both trades generate positive total
cash flows. However, the exposure of each trade to the timing of defaults is
completely different. Basis trades with low price-coupon bonds will benefit from an
early default; basis trades with high price-coupon bonds will suffer from it.
Figure 21: Bond Price & Coupon Relationship: Constant Z-spread (JtD Exposure)
Y-axis: Sum of total cash flows on the negative basis trade until default; X-axis: Assumed default dates.
50%
50%
40%
40%
30%
30%
20%
20%
10%
10%
0%
0%
-10%
-10%
-20%
-20%
Dec-08 Jun-09 Dec-09 Jun-10 Dec-10 Jun-11 Dec-11 Jun-12 Dec-12 Jun-13 Dec-13
Source: J.P. Morgan.
The cash flows and JtD exposure of both trades are significantly different, even
though they share the same Z-spread. Timing of default is paramount, and the shape
of the CDS curve gives us information about it.
20
Figure 18 and Figure 19 refer to the case where there is no default during the life of the
trade, and show the cash flows that are received / paid at each point in time. Figure 20 shows
the jump-to-default exposure of the trade. For dates higher than the trade maturity (20-Dec13), i.e. if there is no default during the life of the trade, the value of each line in Figure 21 is
just the sum of the cash flows in Figure 18 and Figure 19 respectively.
36
Abel Elizalde
(44-20) 7742-7829
[email protected]
Hazard (default intensity) rates for the two CDS spread curves in Figure 22.
Inv erted
25%
10.0%
20%
7.5%
15%
5.0%
10%
2.5%
5%
0%
0.0%
08 09
Source: J.P. Morgan.
10
11 12
13
Dec-08 Jun-09
Dec-09 Jun-10
Dec-10 Jun-11
Dec-11 Jun-12
Dec-12 Jun-13
The hazard rate at any point in time represents the instantaneous default probability
at the time assuming there has been no prior default. Inverted curves have higher
default hazard rates for short maturities: as time goes by, if the company does not
default, its credit quality should improve. Upward slopping (normal) curves
assume that the credit quality of the company will deteriorate as time goes by.
Table 3 shows the expected value of the basis trade for the four possible scenarios we
have considered. We take the trade cash flows assuming a default at different points
in time (every month) and weight them by the probability of defaulting in that period.
Table 3: Basis Trades: Expected Cash Flows
As a % of the bond notional.
Discounted Expected CF
8.77%
6.88%
-1.92%
-0.16%
Usually, inverted CDS curves coincide with low priced bonds, and upward slopping
curves with bonds trading above par. However, the four combinations in Table 3
represent a good sample of possibilities. Bonds with low price and coupon generally
offer a better alternative for negative basis trades since they involve less funding
costs and benefit from a default more than bonds with higher coupons (as these are
lost after a default).
Low price-coupon bonds benefit from front-loaded defaults and therefore from
inverted CDS curves. On the contrary, high price-coupon bonds benefit from backloaded defaults and therefore from upward slopping CDS curves.
37
Abel Elizalde
(44-20) 7742-7829
[email protected]
A measure to judge the attractiveness of the above basis trades should generate a
more negative basis for the case of a low price-coupon bond and inverted CDS curve
than for the case of a high price-coupon bond and upward slopping CDS curve. Table
4 shows the difference between the CDS spread (5% in all cases) and (i) PECS, (ii)
ASW and (iii) Z-spread.
Table 4: Basis
CDS Curve: Inverted
Bond: Low Price-Coupon
-4.78%
-0.56%
-2.02%
The PECS comes up as the best measure for the basis, as it is sensitive to both
the shape of the CDS curve and the cash flow structure of the bond.
Example: Telecom Italia 12s and 55s
A good, although probably extreme, example of the above considerations can be
found looking at two Telecom Italia (TITIM) bonds, one of them maturing in 2055.
Table 5 shows the coupon, price, ASW, Z-spread and PECS of both bonds on 18Dec-08. As Figure 24 shows, the Z-spread on both bonds was similar around that
date. Table 5 shows that the ASW was lower for the 55s but the PECS was higher.
Obviously, depending on which spread measure we use for the basis, the basis will
be different.
Table 5: TITIM Bonds
Maturity
01-Feb-12
17-Mar-55
Coupon
6.250%
5.250%
Clean Price
96.0
63.5
Par ASW
463
345
Z-spread
488
486
PECS
475
722
700
12s
55s
500
300
100
-100
Jan-08
Apr-08
Jul-08
Oct-08
Jan-09
Using these two TITIM bonds and buying CDS protection with Dec-11 maturity
(trading at around 552bp) and with a notional equal to the bond notional, Figure 26
shows the total cash flows of our two basis trades for different default dates. The
lower price of the 55s gives its basis trade a very positive default exposure
compared to the one for the 12s.
38
Abel Elizalde
(44-20) 7742-7829
[email protected]
Figure 26: TITIM Equal Notional Basis Trades: Buying Dec-11 CDS Protection (JtD Exposure)
In bp. As of 28-Dec-08.
Y-axis: Sum of total cash flows on the negative basis trade until default; X-axis: Assumed default dates.
650
60%
600
50%
40%
550
12s
55s
30%
500
20%
10%
450
400
350
0%
-10%
300
-20%
0y
2y
4y
6y
8y
10y
Dec-08
Jun-09
Dec-09
Jun-10
Dec-10
Jun-11
We next include an extract of a basis trade idea we published on 13 January 2009 using a Ford Motor Credit Company
(FMCC) bond maturing in 2010. It will serve to illustrate the concepts that we have dealt with so far, as well as to compare
the different ways of trading the basis.
Maturity
15-Jan-10
20-Mar-10
Currency
EUR
USD
Price / Upfront
75.5
13%
Coupon
4.875%
5.000%
Notional
10,000,000
10,000,000
Source: J.P. Morgan. Indicative mid prices shown. COB 12 Jan. Check with the trading desk for updated levels. CDS trades with a full upfront price plus a fixed 500bp running coupon.
According to the latest market pricing, we estimate a mid-to-ask bond cost of 1.5% and a mid-to-ask CDS cost of 2%,
both upfront. Please check with the trading desk for updated levels.
Figure 27 shows how FMCC 4.875% 2010 bonds and FMCC 5y CDS have evolved over the last year. Figure 28 shows
that the 10s Bond-CDS basis has ranged between +1000 and -2000bp during that time. Although the basis has retraced
from its historical (negative) wides, it still stands around -1100bp according to our calculations.
39
Abel Elizalde
(44-20) 7742-7829
[email protected]
LHS: Bond price (); RHS: 5y full running CDS (bp), inverted.
LHS: Bond price (); RHS: 5y full running CDS (bp), inverted.
3000
1000
90
1000
2000
2000
80
2000
100
3000
1000
4000
0
70
3000
4.875% Price
60
4000
50
5000
Feb-08
May -08
Aug-08
5000
-1000
7000
-3000
Jul-08
Nov -08
6000
Basis
-2000
8000
Figure 29 and Figure 18 show how the basis has, roughly, moved in line with FMCCs credit quality (as measured by the
bond price and PECS).
As we wrote in our 2009 CD Player Outlook, we think we have seen the widest levels on the CDS-cash basis. Basis trades
have attracted investors attention as one of the most attractive type of trade for 2009, and we think that as liquidity
conditions normalise so will the basis (on average).
Figure 29: FMCC 4.875% Basis vs. Bond Price
X-axis: Bond price (); Y-axis: Basis (bp). One year historical data.
X-axis: Bond PECS (bp); Y-axis: Basis (bp). One year historical data.
2000
Basis
2000
1000
1000
-1000
-1000
Current
-2000
Basis
Current
-2000
-3000
-3000
40
50
60
70
80
90
100
2000
4000
6000
8000
PECS
Bond Price
Source: J.P. Morgan.
With average high yield basis around -550bp for the US corporate high yield market, we think FMCC's 4.875% bonds
represent an attractive basis trade opportunity.
Table 7: FMCC 10s
Currency
Maturity
Coupon
Frequency
Mid Price
Dirty Price
Mod. Duration
Z-spread
PECS
Basis
Source: J.P. Morgan. Calculations using mid prices. Indicative mid prices shown. COB 12 Jan.
Check with the trading desk for updated levels.
40
Upfront
12.5%
13%
Source: J.P. Morgan. Indicative mid prices shown. COB 12 Jan. Check with the trading desk for
updated levels.
Abel Elizalde
(44-20) 7742-7829
[email protected]
If we construct the basis trade with the longer dated CDS (March-10), the negative basis trade has a positive jump-to-default
during all the trade life (especially if the default happens after the bond matures). If we construct the basis trade with the
Dec-09 CDS, we end up with a naked long bond exposure from the CDS maturity until the bond maturity (1 month
approximately).
We recommend use of the March-10 CDS contract rather than the Dec-09 one given the small difference in cost. It does
not expose the investor to a month of outright long credit exposure to FMCC (event risk remains high during all the trade
life) and it is slightly more bearish with respect to spread movements.
Sizing the Trade: Equal Notional
An equal notional basis trade would involve:
Buying the bond at a current mid-dirty price of around 80.43 (see Table 7). The bond pays an annual coupon on
January 15th (i.e. this Thursday).
Buying CDS protection, which requires an upfront payment plus a 500bp running premium. We assume the investor
buys the same USD CDS protection as the EUR notional bought on the bond, and make our calculations assuming the
EUR/USD exchange rate stays constant. Investors can hedge the FX risk by locking in a forward FX rate. With a current
FX rate of around 1.3375, the investor would have to buy USD 133.75 CDS protection for each EUR 100 of bond
notional bought.
Funding the bond purchase, the CDS upfront cost and the CDS margin. We assume a 5% funding cost and a 5%
margin on the CDS.
We recommend an equal notional trade. However, investors who want to trade movements in the basis but do not want to be
exposed to general movements in spreads (keeping the basis constant) would need to duration-hedge the trade.
Trade Default Scenarios
In Figure 31 we show what would be the total (undiscounted) cash flows of the basis trade if FMCC was to default anytime
from here to the end of the trade. Figure 32 contains the same information as Figure 31, but it only shows dates from today
to December 2009. We have assumed a 20% recovery rate for both bond and CDS in case of default.
In case of no default during the trade life, the bond coupon (4.875%) would be almost enough to cover for the 500bp
running payment on the CDS, and the investor would earn around EUR 5.21 for each EUR 100 of bond notional on the
trade, net of funding and margin costs: Around EUR 29.3 on the bond (coupons plus pull to par), minus EUR 19 from the
CDS protection (upfront plus coupons), minus EUR 5.1 on funding costs.
In case of a default after the bond maturity, but before the March-10 CDS maturity, the investor would earn a total EUR
86.2 (approximately): Around EUR 61.8 from the CDS (EUR 80 on default minus the upfront and running premiums paid),
plus EUR 29.3 on the bond (coupons plus pull to par), minus EUR 5 on funding costs.
41
Abel Elizalde
(44-20) 7742-7829
[email protected]
Y-axis: Sum of total cash flows on the negative basis trade until default; Xaxis: Assumed default dates.
Y-axis: Sum of total cash flows on the negative basis trade until default; Xaxis: Assumed default dates.
100%
12%
10%
80%
8%
60%
6%
40%
4%
20%
2%
0%
0%
Jan-09
Apr-09
Jul-09
Oct-09
Jan-10
Apr-10
Jan-09
Mar-09
May -09
Jul-09
Sep-09
Nov -09
Spread Scenarios
Using the March-10 CDS contract, Figure 33 and Figure 34 show the trade MtM for changes in the Bond-CDS basis and on
spread levels (with the same basis), respectively. The trade would make money as the basis goes down (less negative) and
also as spreads increase (since the duration of the CDS is slightly higher than the bond duration).
Figure 33: FMCC 10s Basis Trade: Basis Scenario Analysis
Y-axis: Trade MtM; X-axis: Changes in both CDS and bond PECS spreads.
6%
CDS Widens
1.5%
1.0%
0.5%
0.0%
-0.5%
-1.0%
-1.5%
-2.0%
4%
2%
0%
-2%
-4%
-6%
-5% -4% -3% -2%
-1%
0%
1%
2%
3%
4%
5%
0%
1%
2%
3%
4%
5%
Source: J.P. Morgan. Additive spread changes. MtM expressed as % of the bond notional. Using Source: J.P. Morgan. Additive spread changes. MtM expressed as % of the bond notional. Using
mids.
mids.
Recovery Rate Risk: The impact of the realised recovery rate on the trade PnL would be negligible (since we are using the
same notional in both bonds and CDS) unless the default happens between 15-Jan-10 and 20-March-10. In that case, the
investor would be benefit from a low recovery rate since the trade is outright short credit.
42
Abel Elizalde
(44-20) 7742-7829
[email protected]
Capital-at-Risk Hedging
In the previous examples, we assumed that the amount of CDS protection bought is similar to the bond notional. In that
case, the investor is not exposed to the realised recovery rate in the event of default. On default, the investor delivers the
bond into the CDS contract and receives par.
Equation 4 showed the profit for the investor in a basis trade upon default:
CDS Notional x (100 Recovery CDS Upfront CDS Coupons Paid CDS Funding Costs Paid)
+ Bond Notional x (Recovery + Bond Coupons Received Bond Price Bond Funding Costs Paid)
Assuming an instantaneous default (or not considering the coupons and funding costs on both legs), the loss on the bond is
given by Bond Notional x (Recovery Bond Price), and the profit on the CDS is given by CDS Notional x (100 Recovery
CDS Upfront). If both legs are done on the same notional, the final profit is independent on the recovery rate of the bond.
Equation 6 splits the trade cash flows into running and one-off payments.
Equation 6: Basis Trade Profit on Default
CDS Notional =
The first and most important problem with the above hedging alternative is that the recovery rate is not known in advance.
Thus, the hedging ratio will only achieve its purpose if the realised recovery is similar to the one the investor assumed in
the first place. Figure 35 shows the distribution of recovery rates for the period 1998-2007. The figure shows the high
variability of recovery rates, which makes challenging the success of the above hedging strategy. The mean recovery rate is
27%, and the mode (most common) recovery rate in our data sample since 1998 is 10-15%.
43
Abel Elizalde
(44-20) 7742-7829
[email protected]
120
101
100
84
80
60
61
46
45
40
57
34
30
35
37
25
19
20
13
15
23
11
95-
0
0-5% 5-10%
10-
15-
20-
25-
30-
35-
40-
45-
50-
55-
60-
65-
70-
75-
80-
85-
90-
15%
20%
25%
30%
35%
40%
45%
50%
55%
60%
65%
70%
75%
80%
85%
90%
95% 100%
Table 9 shows the CDS notional in a capital-at-risk basis trade for different bond prices as a % of the bond notional,
assuming the CDS trades on a full running basis (i.e. no upfront). When the bond trades at par, the CDS notional is similar
to the bond notional. For bonds trading below par, the CDS notional is lower than the bond notional, and vice versa.
Bond Price
Table 9: CDS Notional in a Capital-at-Risk Basis Trade (as a % of the bond notional)
60
70
80
90
100
110
120
10%
20%
30%
40%
Recovery Rate
50%
60%
70%
80 %
90%
56%
67%
78%
89%
100%
111%
122%
50%
63%
75%
88%
100%
113%
125%
43%
57%
71%
86%
100%
114%
129%
33%
50%
67%
83%
100%
117%
133%
20%
40%
60%
80%
100%
120%
140%
25%
50%
75%
100%
125%
150%
33%
67%
100%
133%
167%
50%
100%
150%
200%
100%
200%
300%
Compared to an equal notional basis trade, a capital-at-risk trade tries to eliminate the jump-to-default (JtD)
exposure of the trade at the expense of the carry profile. For discount bonds, where the JtD exposure is positive in an
equal notional trade,21 a capital-at-risk would imply buying a lower CDS notional, which reduces the JtD to zero and also
improves the carry of the trade. For bonds trading above par, the opposite holds.
Again, we should notice that a capital-at-risk trade leaves the investor with exposure to realised recovery rates.
Effectively, the JtD exposure will only zero if the realised recovery rate is similar to the one assumed to construct the
hedging notional.
21
44
Abel Elizalde
(44-20) 7742-7829
[email protected]
3Y CDS
4Y CDS
5Y CDS
144.6
114.9
96.6
3Y
4Y
5Y
0.24%
3.25%
5.17%
As Table 10 shows, the lower duration of short maturity CDS contracts involve a
higher notional to make the basis trade duration weighted.
Using a shorter maturity CDS in a duration weighted basis trade involves a
worse carry profile (less positive or more negative) but a better JtD exposure.
45
Abel Elizalde
(44-20) 7742-7829
[email protected]
Using a shorter maturity CDS will effectively overlay a curve flattener on top of
the basis trade (selling longer protection through the bond and buying shorter
protection through the CDS). This can be particularly attractive when the investor
has a bearish view on the credit since spread widening tends to flatten spread curves,
or when the investor wants to have a positive JtD exposure.
When the notional on both legs of a basis trade is not the same, the investor
takes on recovery rate risk.
This recovery rate risk affects the difference between the bond and the CDS notional.
If the bond notional is larger, the investor is long recovery rates, and vice versa.
FMCC Case Study (Cont.)
Coming back to the FMCC basis trade example we introduced previously, Figure 36 compares the trade sensitivity to
movements in spreads (keeping the basis constant) in an equal notional basis trade and in a duration weighted basis trade
(where we buy CDS protection on a notional equal to 75% of the bond notional bought). The lower CDS notional will also
change the default exposure of the trade, as Figure 37 shows (assuming a 20% recovery rate): the trade is not positive jumpto-default any more. Since we have bought less CDS protection than bond notional bought, the realised recovery rate in case
of default will also play an important role: the lower the recovery the higher the losses of the basis trade.
Figure 36: FMCC 10s Basis Trade: Spread Scenario Analysis
Y-axis: Trade MtM; X-axis: Changes in both CDS and bond PECS spreads.
Y-axis: Sum of total cash flows on the negative basis trade until default; Xaxis: Assumed default dates.
1.5%
1.0%
0.5%
0.0%
-0.5%
-1.0%
-1.5%
-2.0%
Equal Notional
100%
Duration Hedged
Equal Notional
Duration Hedged
80%
60%
40%
20%
0%
-5% -4% -3% -2% -1%
0%
1%
2%
3%
4%
5%
-20%
Jan-09
Apr-09
Jul-09
Oct-09
Apr-10
Source: J.P. Morgan. Expressed as % of the bond notional. Using Mids. Assuming 20%
recovery.
The above example illustrates clearly the impact of the trade sizing on the economics of a basis trade.
46
Jan-10
Abel Elizalde
(44-20) 7742-7829
[email protected]
Basis trades with different bond and CDS notional expose investors to recovery rate
risks. An equal notional basis trade will eliminate that risk and provide a cleaner
exposure to the pricing differential between bonds and CDS.
As we showed in the grey box at the beginning of this section (Jump-to-Default
Exposure), in case of default, and ignoring any risk-free discounting and funding
costs, the profit for the investor in an equal notional basis trade is given by Equation
8.
Equation 8: Equal Notional Basis Trade Profit on Default or Maturity (Ignoring risk-free
discounting and funding costs)
( 100 Bond Price + Bond Coupons Received CDS Upfront CDS Coupons Paid)
Note: Bond Price refers to the dirty bond price.
Other things being equal, it is clear from Equation 8 that in equal notional basis
trades that look to benefit from a default:
Bonds with low price-coupon are generally more attractive (see Figure 21) if the
default is soon enough.
Full running CDS are more attractive than full upfront CDS (again if the default
is soon enough).
The more negative the basis the better for a JtD trade; however, we can also find
positive JtD exposures in positive basis trades.
We can isolate two different aspects of negative basis trades with respect to their JtD
exposure:
The different cash flow structure of bond (price vs. coupon) and CDS
(upfront vs. running) generate a different JtD profile on each basis trade.
The lower the upfront payments on the bond (price) and CDS the better. Even in
situations with positive basis, a long position in a low price-coupon bond and a
short position in a full running CDS will provide a high initial positive JtD
exposure (which will turn negative at some point during the trade).
Default (like maturity) allows the investor to cash in the Bond-CDS basis.
Therefore, the more negative the basis the better for the investor in any negative
basis trade.
Figure 38 and Figure 39 illustrate the two points above. In Figure 38 we consider two
extreme examples of bond and CDS, both with a zero basis (measured as equivalent
full running CDS spread minus Z-spread). The figure shows the total (not
discounted) cash flows of our negative basis trade for the two following
combinations: (i) a low price-coupon bond paired with a full running CDS, in dark
blue, and (ii) a high price-coupon bond paired with a full upfront CDS, in grey. Both
trades generate similar basis, but the JtD exposure is completely different.
47
Abel Elizalde
(44-20) 7742-7829
[email protected]
Y-axis: Sum of total cash flows on the negative basis trade until default; Xaxis: Assumed default dates.
Y-axis: Sum of total cash flows on the negative basis trade until default; Xaxis: Assumed default dates.
4%
10%
2%
2%
0%
0%
0%
-2%
-2%
-4%
30%
20%
20%
10%
0%
-10%
-10%
-20%
-20%
-30%
-30%
-4%
-40%
-40%
-6%
09
10
10
11
11
12
12
13
13
4%
30%
09
6%
40%
08
6%
40%
Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec-
-6%
Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec08
09
09
10
10
11
11
12
12
13
13
Figure 39 shows two basis trades sharing the same bond (5% Z-spread) paired with a
full running CDS trading at 5% (in dark blue) and 4% (in grey) respectively. The
more negative the basis the better for the JtD exposure of the trade.
JtD Exposure: Basis Trade vs. Fixed Recovery CDS
Assuming the sum of bond price and the CDS upfront is below par, which is
necessary for an equal notional negative basis trade to have a positive payment from
an instantaneous default, we can compare the profit from two alternatives:
In a short CDS position, the profit from a default is given by (100 Recovery
CDS Upfront), and the carry of the trade is given by the CDS running spread.
Compared with a short CDS position and assuming similar notional, a negative basis
trade will earn more money from an instantaneous default as long as the bond price is
lower than the realised recovery rate. Unlike in a short risk CDS position, in a
negative basis trade the payment in case of default is known a priori and does not
depend on the realised recovery rate. Thus, an equal notional basis trade resembles a
fixed recovery CDS with the recovery struck at the bond price and with a spread
equal to the difference between the CDS spread and the bond coupon.22 However,
there are very clear differences which difficult the comparison between the basis
trade and the fixed recovery CDS: in the negative basis trade the investor receives the
payment of (100 Bond Price CDS Upfront) even if there is not a default (he gets
paid 100 and maturity and he has paid the bond price and CDS upfront at inception),
and in the basis trade the investor has to fund the bond price.
22
48
Notice the different treatment upon defaults for CDS spreads and bond coupons.
Abel Elizalde
(44-20) 7742-7829
[email protected]
Historical Basis
Both during the first nine and last two months of 2008, the CDS-cash basis
was remarkably market directional, i.e. driven by CDS spreads.
September and October saw the basis moving significantly into negative
territory due to the illiquidity in the cash market, at the same time spreads
reached record wide levels.
In December, jump-to-default hedging took the basis off its widest (negative)
levels. Investors bought CDS protection to hedge default risks, causing CDS
spreads to underperform cash spreads.
The basis has continued to be market directional during the first weeks of
2009. The rally during January has taken again basis levels closer to the
(negative) levels seen during the last quarter of 2008.
Our base case scenario for the evolution of the CDS-cash basis during 2009
is a gradual transition to a new normal regime where the CDS becomes,
again, the driver of the basis.
However, we can not rule out the possibility of a sudden transition to the
previous normal regime: the basis collapsing to zero due to jump-to-default
hedging by credit funds and correlation desks.
We think that taking a step back and looking at the drivers of the basis during
normal and non-normal times can be helpful to understand its recent movements
and to form an opinion on its future ones.
The Bond-CDS basis aims to be a measure of the discrepancy between the risks
priced in bonds and CDS. Leaving aside liquidity considerations, CDS and bond
spreads both compensate investors for the same risk the risk that a company might
default. As such, they should be approximately equal. The CDS-cash basis measures
the extent to which these spreads differ from each other.
49
Abel Elizalde
(44-20) 7742-7829
[email protected]
However, leaving aside liquidity considerations might be leaving aside too much.
Notice that our measure of the basis-to-cash will still provide a measure of the
different credit risk priced in both instruments if we assume the liquidity premium is
the same. If that was not the case the basis-to-cash will be a combination of credit
and liquidity risks. This can be particularly relevant in situations where the liquidity
on each market is very different.
Figure 40: Historical iTraxx Basis vs. Spread
Spread (bp).
300
250
150
Basis-to-Cash (RHS)
100
200
50
150
100
-50
50
-100
-150
Jan-08
Apr-08
Jul-08
Oct-08
Jan-09
80
60
40
20
0
-20
-40
-60
-80
-100
-120
Before Sept. 08
Sept.-Oct. 08
Since Nov . 08
0
50
100
150
200
250
300
350
CDS and bonds should price the same credit risk:23 on an aggregate level, the
basis between them represents largely a combination of different liquidity
factors affecting CDS and bond pricing. These liquidity factors refer mainly to
issuance/redemptions (synthetic and cash) and funding costs,24 which tend to
materialise in sudden shocks rather than gradual movements.
Liquidity is a very elusive concept which people tend to use for different purposes.
We do think that liquidity can be interpreted simply as measure of the balance
between supply and demand. When the demand for risk is high, any supply of risk is
easily absorbed and liquidity is good; when the demand for risk is low, the supply of
risk can not be absorbed and liquidity is poor. Generally, low demand for risk
coincides with high supply of risk and vice versa; otherwise spreads would stay
constant. Thus we call liquidity those factors that affect the demand and supply of
risk.
Normal and Sudden Basis Regimes
Going back to the basis, we believe there are two different regimes which explain
movements in the CDS-cash basis: a normal one and a sudden one.
23
Leaving aside soft credit events, the cheapest-to-deliver option in CDS, the different
treatment of bond and CDS coupons upon default, and any optionality embedded in bonds.
24
Funding conditions affect funded instruments (bonds) more than it does unfunded ones
(CDS).
50
Abel Elizalde
(44-20) 7742-7829
[email protected]
The normal CDS-cash basis regime is one where the relative liquidity between
bonds and CDS stays constant or changes very slowly. In this regime, movements
in the basis are driven by movements in CDS spreads since they tend to be a leading
indicator to bond spreads due to their flexibility and unfunded nature (see grey box in
the next page). We do realise that the issue of whether CDS spreads react faster than
bond spreads is not entirely clear cut, and it depends on the relative liquidity of bonds
and CDS on each particular market. However, we do believe that this is clearer in
Europe, where the bond market is less liquid than in the US.
This normal regime is characterised by the relative liquidity between bonds and
CDS staying more or less constant, but not necessarily zero. As we explained above,
liquidity represents a very vague concept where we would generally include the net
difference in protection selling pressure in the bond and CDS markets coming from
factors such as issuance/redemptions and funding.
As Figure 41 shows, this normal regime was in place before September 2008 and it
is also in place since November 2008. The relationship between the basis and
underlying spreads tends to be a positive one (i.e. as spreads widen the basis becomes
more positive/less negative).
Figure 42: Regimes in Cash-CDS Basis
CDS-Cash Basis
Normal Regime: the relative liquidity between
bonds and CDS stays constant and funding
conditions do not change. Movements in the
basis are driven by movements in CDS spreads
(leading indicator to bond spreads).
Spreads
51
Abel Elizalde
(44-20) 7742-7829
[email protected]
The sudden CDS-cash basis regime is one where the relative liquidity between
bonds and CDS suddenly changes due to a shock such as a significant change in
funding conditions, redemptions/inflows in cash portfolios or sudden
issuance/unwinds of synthetic products. These sudden movements in liquidity
significantly affect the basis (in either direction) and will likely be corrected
gradually.
This is, for example, what happened during September and October this year (see
Figure 41), where the illiquidity of cash bonds and funding pressures saw the basis
moving 100-200bp negative in investment grade. Our European credit strategists
estimate that the current liquidity premium in European investment grade bond
spreads could potentially range from 100to 200bp.25 Our US colleagues estimate that
funding alone has made the basis around 125bp wider (i.e. more negative).26
These sudden shocks to the CDS-cash basis should generally be corrected
gradually, taking the basis closer to zero and to a new normal regime. As we
argue next, we believe 2009 will bring a gradual transition to a new normal basis
regime (at wider spreads) driven by the easing of funding pressures and by the
pressure on CDS spreads coming from a slow burning CSO unwind process.
Price discovery takes place primarily in the CDS market Blanco, Brennan and Marsh (2005) 27
Blanco, Brennan and Marsh (2005) represented one of the first academic papers to analyse the empirical relationship between CDS and bond spreads.
The authors showed that price discovery takes place primarily in the CDS market. The main conclusion of their empirical study is that the pricing
discrepancy between CDS prices and credit spreads is closed primarily through changes in the credit spread, reflecting the CDS markets lead in price
discovery. It is through this error correction mechanism that both CDS and credit spreads price credit risk equally in the long run. The authors
speculate that price discovery occurs in the CDS market because of (micro)structural factors that make it the most convenient location for the trading
of credit risk, and because there are different participants in the cash and derivative market who trade for different reasons.
25
Our credit strategists estimate that the current liquidity premium in investment
grade bond spreads could potentially range from 100to 200bp. See How big is the
liquidity or illiquidity premium?, P Malhotra et al, 30 January 2009.
26
See High Grade Bond and CDS 2009 Outlook, E Beinstein et al, 5 December 2008.
27
R Blanco, S Brennan and W Marsh, 2005, An Empirical Analysis of the Dynamic Relation
between Investment Grade Bonds and Credit Default Swaps, Journal of Finance 60 (5). See
also S Alvarez, 2004, Credit default swaps versus corporate bonds: Do they measure credit
risk equally?, unpublished manuscript.
52
Abel Elizalde
(44-20) 7742-7829
[email protected]
2009 Outlook
Figure 42 sketched our base case scenario for the evolution of the CDS-cash basis
during 2009: a gradual transition to a new normal regime where the CDS becomes,
again, the driver of the basis. However, we do think the possibility of a sudden
transition to the previous normal regime can not be ruled out, i.e. the basis collapsing
to zero due to jump-to-default hedging by credit funds and correlation desks.
Figure 43: Cash-CDS Basis in 2009
CDS-Cash Basis
Spreads
Sept-Oct. 08.
We think the current negative basis is mainly a reflection of the deleveraging process
in cash portfolios.28 Basis packages are one of the alternatives to sell cash bonds in
the current market, and given the high funding costs, basis remain very negative. A
negative basis trade in which investors lock-in, say, 100bp default-free spread
during, say, 4 years, can be seen as a 400bp discount on cash bonds (before funding
costs and taxes). Cash bonds are being sold at discounted prices versus instruments
which provide similar risk profiles and require less funding.
In our view, this dynamic will persist going forward. The negative basis is slowly
attracting the interest of credit investors, who are mainly cherry-picking among the
existing opportunities. As this trend generalises the basis will gradually normalise,
helped by easier funding and unwinds of synthetic structures.
Unwinds of CSOs should help this trend during 2009, as dealers buy synthetic
protection to unwind their hedges. We have argued in recent reports that we think
this unwind will be a gradual process during 2009.
The risk for our base case scenario is jump-to-default hedging through CDS. We
have seen during December how spreads can pick-up very quickly in an environment
of low trading volumes and deteriorating credit conditions. The auto sector led a
wave of jump-to-default hedging by correlation desks and credit funds during the
last weeks of 2008, which has quickly expanded to the overall market.
28
Synthetic structures have suffered higher MtM losses than cash positions, and their
unwinding dynamics are different. Many investors view these synthetic products (mainly
CSOs) like options, due to their low prices, and do not have a lot of further downside from
holding them forward. Thus, the pressure to unwind these synthetic structures might be lower.
53
Abel Elizalde
(44-20) 7742-7829
[email protected]
In any case, we think we have seen the widest levels on the CDS-cash basis, but
we also think the normalization of the CDS-cash basis will take time and will be a
gradual process during the next months.
With bonds pricing at high discounts from par and wide spreads, negative basis
packages generally have a significant negative carry component (or initial cost if
the CDS is trading in upfront terms), but a positive jump-to-default exposure.
Thus, they will be seen as limited downside trades going into 2009, attracting
the interest and funds from investors.
The current market conditions are very attractive to set up negative basis packages.
The high illiquidity premium on cash bonds relative to CDS has taken the basis to
(negative) levels never seen before. Additionally, the high credit risk environment
has taken bond prices to very low prices, which give negative basis trades a positive
jump-to-default exposure.
Drawing from the data of the European Bond-CDS basis Report we present an
analysis of the current basis as well as the historical basis by rating, maturity and
sector. The European Bond-CDS basis Report complements our existing US
Corporate High Grade Basis Report and US Corporate High Yield Basis Report for
the US market. All of them are available on Morgan Markets and provide an efficient
way to track the Bond-CDS basis on a daily basis. Finally, we include a list of the
top 40 negative basis trades.
Current Basis Breakdown: Maturity, Rating and Sector
Table 11 shows the current average basis for the bonds included in the European
Bond-CDS basis Report. We should note that the universe of bonds used will imply
different average basis and that, as we review next, the dispersion around the average
basis can be significant.
It is clear from Table 11 that the basis is more negative the lower the rating. This is
also clear from Figure 44 and Figure 45, which show the relationship between the
basis and the bond spread. The lower the credit quality of the bond, the more
negative the basis-to-cash.
Table 11 also shows that the basis tends to be more negative for longer maturities.
This is reasonable in a distressed credit environment, where CDS curves tend to be
inverted. The CDS market has higher curve liquidity than the bond market:
investors use CDS, rather than bonds, to express curve views. Additionally, all the
jump-to-default hedging coming from structured product desks tends to focus on
short CDS maturities, which inverts CDS curves more than bond ones.
29
See First signs of a fade, or just a pause for thought?, S Dulake, A Elizalde, 8 January
2009.
54
Abel Elizalde
(44-20) 7742-7829
[email protected]
Basis
-68
-101
-131
113
-78
-49
-129
-891
15
-62
-112
-142
-160
-8
55
-64
-79
51
-96
-147
-144
10
-18
IG All
IG with maturity above 3y
IG with maturity above 5y
AAA
AA
A
BBB
Crossover
(1,3] years
(3,5] years
(5,7] years
(7,10] years
(10,100] years
Basic Industry
Capital Goods
Consumer Cyclicals
Consumer Non-Cyclicals
Energy
Financial Institutions
Technology
Telecom & Media
Transport
Utilities
Rating
Maturity IG
Sector IG
Count
478
341
191
35
97
198
148
6
137
150
82
82
27
51
30
38
43
9
146
2
87
1
71
PECS
295
324
359
150
244
228
452
1940
224
279
343
370
371
450
272
459
196
226
290
310
301
190
178
Source: J.P. Morgan. Basis = PECS Interpolated CDS Spread. Data as of 2 February 2009.
Basis
2000
Basis
1000
1000
PECS
500
PECS
0
-1000
-500
-2000
-3000
-1000
0
1000
2000
3000
4000
5000
200
400
600
800
1000
Figure 46 shows the distribution of the current basis as well as the basis one year
ago. It is clear how the current basis is very disperse compared to the one in
January 2009. Moreover, it is clearly skewed to negative levels.
Thus, there are currently attractive opportunities in basis below -300bp which
investors should look at very carefully. One of the sections of the European BondCDS basis Report highlights those opportunities daily (see also Table 12 in this
report for the current top 40 most negative basis).
55
Abel Elizalde
(44-20) 7742-7829
[email protected]
Jan 08
70%
Jan 09
60%
50%
40%
30%
20%
10%
0%
-800 -700 -600 -500 -400 -300 -200 -100
Figure 47 shows the basis distribution by rating category. Although the number of
crossover bonds the report covers is very small, it is clear how their basis tends to be
a lot more negative. In Figure 48 we compare the basis distribution of the entire
sample of bonds and the bonds within the consumer cyclical sector, which is one
with the highest number of very negative basis.
Figure 47: Basis Distribution By Rating
X-axis: Basis bucket; Y-axis: % of bonds with basis within each bucket.
X-axis: Basis bucket; Y-axis: % of bonds with basis within each bucket.
AA
80%
BBB
Crossov er
40%
50%
30%
40%
30%
20%
20%
10%
10%
0%
0%
-800
56
-700
-600
-500
-400
-300
-200
All
50%
70%
60%
-100
-800
-600
-400
-200
Consumer Cy clicals
200
400
600
800
Abel Elizalde
(44-20) 7742-7829
[email protected]
In 2008 before Lehmans default, the basis for investment grade credit was
relatively low and weakly directional. After the default it has traded in a
range from 200-275bps.
In the HY market, a similar transition took place, albeit slightly after the IG
transition. Since mid-October, the HY basis has been in a -450bps to 750bps range. It has also been increasing as CDS spreads widened, but the
relationship is relatively weak.
The increase in funding costs has been an important driver of the significant
widening of basis. IG basis has been somewhat contained by dealer
correlation desks hedging their synthetic CDO positions. This is less
important for HY basis.
The US IG basis has behaved similarly to the EU IG basis in 2008. As shown in
Figure 49 and Figure 50, the basis was mildly directional until September 2008, with
the basis increasing (i.e. more positive) as CDS spreads increased. However,
Lehmans default changed the trading pattern between CDS spreads and basis. After
a couple of weeks, the basis was reset at around -250bps. Furthermore, in this new
phase, the basis is not very directional.
Figure 49: US IG CDS spread and basis have diverged in mid Sep
300
150
0
1/8/8 - 9/10/8
-150
9/11/8 - 10/15/8
-150
10/16/8 - 1/29/9
CDS Spread
-300
Jan-08
Source: J.P. Morgan.
-300
Basis-to-Cash
Apr-08
Jul-08
Oct-08
Jan-09
1/29/2009
100
150
200
250
300
350
The HY basis drastically changed last Fall. Early in 2008, the basis was small and
changing almost independently of the CDS spreads (see Figure 52). Lehmans
default did not immediately affect this pattern. However, things started to unravel
near the end of September and, in a few days, a new relationship was established at
much more negative basis levels. Since then, the HY basis has traded as negative as
-720bps in mid December, but has somewhat recovered since then. Since this new
trading pattern as been established, the basis has become more negative as spreads
widened, which is the opposite of what has happened so far in either IG or HY.
57
Abel Elizalde
(44-20) 7742-7829
[email protected]
Figure 51: US HY CDS spread and basis have diverged in late Sep
1250
1000
750
-150
500
250
0
-300
-250
-500
-450
CDS Spread
-750
Basis-to-Cash
Jan-08
Source: J.P. Morgan.
Apr-08
-600
500
750
1000
1250
1/8/8 - 9/24/8
9/25/8 - 10/15/8
10/16/8 - 1/29/9
Jul-08
Oct-08
Jan-09
-750
1/29/2009
These resets are probably due to the abrupt increase in funding costs that took
place at the same time as the perception of counterparty risk significantly
increased. We believe that bond spreads and CDS will likely trade separately until
financing becomes more readily available and the arbitrage trade becomes
attractive. Finally, correlation desks activity to hedge their synthetic CDO positions
should help reduce the size of the IG basis, especially if there is an increase in
synthetic CDO unwinds. The HY basis is only slightly affected by this activity as
few HY names were used in structured products.
58
Abel Elizalde
(44-20) 7742-7829
[email protected]
We start with the list of all MAGGIE Euro Credit bond index components,
excluding bonds issued by entities belonging to the following categories: public
sector, asset backed, miscellaneous, Pfandbriefe and EMU governments.
We first eliminate bonds with low liquidity. The bonds with the following
characteristics are eliminated:
Bonds with more than 5 days of stale prices during the month prior to the
rebalancing date (which will generally correspond with month end).
Bonds which have not traded in the last 5 days prior to the rebalancing date.
Bonds with less than 3 dealers (as reported by ISDA).
Bonds with an average daily traded volume over the month lower than EUR
500,000.
Subordinated bonds.
We then match up the remaining bonds with a CDS curve. Bonds for which such
matching is not found are eliminated.
We further eliminate those bonds for which the corresponding CDS spread has
been stale more than 10 days during the month prior to the rebalancing date.
We plan to update the list of bonds in our report on a monthly basis.
59
Abel Elizalde
(44-20) 7742-7829
[email protected]
Table 12: Most Attractive Negative Basis Trades in the European Bond Universe
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Company
Maturity
Basis
GMAC
GMAC
FIAT FINANCE NTH AMER
FCE BANK
GLENCORE FINANCE EUROPE
FCE BANK
GMAC INTL FINANCE
GIE PSA TRESORERIE
CIT Group Inc
WPP GROUP
WPP FINANCE
RENTOKIL INITIAL
XSTRATA FINANCE (CANADA)
CIT GROUP
KINGFISHER
PORTUGAL TELECOM INTL
VOLVO TREASURY
CIT GROUP
CIT GROUP
XSTRATA FINANCE (CANADA)
TELECOM ITALIA
VALEO
MORGAN STANLEY & CO INTL
MORGAN STANLEY
WOLTERS KLUWER
KINGFISHER
MORGAN STANLEY
WPP GROUP
CLARIANT FINANCE
MORGAN STANLEY
MORGAN STANLEY
ALTADIS EMISIONES
PORTUGAL TELECOM INTL
SAINT GOBAIN
GOLDMAN SACHS
CASINO GUICHARD PERR
ALTADIS FINANCE
XSTRATA FINANCE (CANADA)
OTE
ERICSSON
Jun-11
Sep-10
Jun-17
Jan-12
Apr-15
Jan-13
May-10
Sep-33
May-14
May-16
Jan-15
Mar-14
Jun-17
Mar-15
Nov-12
Jun-25
May-17
Sep-16
Nov-12
May-11
Mar-55
Jun-13
Oct-16
Oct-17
Apr-18
Oct-10
Dec-18
Dec-13
Apr-13
Nov-15
May-19
Dec-15
Mar-17
Apr-17
Feb-15
Apr-14
Oct-13
Jun-12
May-16
Jun-17
(bp)
-1655
-1286
-1244
-931
-759
-716
-708
-698
-696
-670
-669
-589
-566
-563
-531
-510
-497
-494
-452
-432
-431
-423
-399
-389
-376
-368
-356
-344
-344
-339
-327
-326
-319
-302
-293
-291
-287
-284
-282
-279
1 week
change
(bp)
-248
+653
-257
+130
-114
+110
-201
-4
-77
-15
-12
+49
+304
-5
-5
-28
-31
-11
-21
-20
+17
-46
-5
-5
-15
+16
+11
-17
+1
+22
+5
-6
-6
-28
-2
-16
-9
+105
-12
-25
1 month
change
(bp)
429
+1,178
-342
-125
+805
-44
+704
+33
+116
-172
-37
-38
+1,502
+54
-38
-47
-62
-37
+867
+76
-2
-250
+30
+45
-28
+10
+56
-105
-150
+77
+13
-4
+19
-82
+4
-9
-51
+612
-37
-106
PECS
(bp)
2800
2565
2678
2052
1946
1820
2123
949
1428
1126
1130
1139
1469
1270
979
619
890
1171
1231
1582
756
1018
727
706
452
837
662
811
883
680
630
416
445
609
580
500
384
1394
409
551
Interp.
CDS*
(bp)
1145
1278
1434
1121
1188
1104
1416
252
732
456
461
549
903
707
449
110
393
677
779
1150
325
595
327
317
76
469
306
467
539
341
303
91
126
307
287
209
97
1110
126
272
Price
(Clean)
Accr.
Coupon
S&P
Rating
Moody's
Rating
63.8
72.5
48.6
69.8
58.7
67.8
80.5
60.5
62.7
67.9
66.9
67.6
57.0
60.6
75.5
62.1
66.8
60.3
69.3
76.8
60.7
71.6
70.1
73.5
88.4
89.7
77.4
75.6
77.4
72.0
71.3
81.9
79.2
74.9
78.1
86.3
92.7
71.8
84.2
78.7
3.5
2.0
3.6
0.3
5.5
0.4
4.0
2.2
3.6
4.8
0.0
3.9
3.3
3.7
0.8
2.8
3.3
1.7
0.8
4.0
4.6
2.3
1.3
1.8
5.2
1.2
4.4
0.7
3.6
0.8
3.7
0.5
3.7
3.8
4.0
3.9
1.7
3.1
3.2
3.2
CC
CC
BBBBBBB
BCC
BBB+
ABBB
BBB
BBBBBB+
ABBBBBB-
C
C
Baa3
B3
Baa2
B3
C
Baa2
Baa1
Baa2
Baa2
AABBB+
BBB
A+
A+
BBB+
BBBA+
BBB
BBBA+
A+
BBB
BBBBBB+
AABBBBBB
BBB+
BBB+
BBB+
Baa2
Baa1
Baa3
Baa2
A3
Baa1
Baa1
Baa2
Baa2
Baa3
A1
A1
Baa1
Baa3
A1
Baa2
Baa3
A1
A1
Baa3
Baa2
Baa1
Aa3
Baa3
Baa2
Baa2
Baa1
ASW
Par
(bp)
1717
1710
866
1462
1170
1242
1499
464
934
770
740
742
773
754
723
326
578
678
888
1499
349
1607
528
515
411
1127
531
655
754
502
440
379
344
451
454
449
400
1366
339
413
Z-spread
Industry
Coupon
Ticker
ISIN
(bp)
2320
2122
1371
1891
1771
1627
1751
709
1286
1003
959
948
1173
1037
871
458
769
951
1133
1931
504
6130
687
644
467
1308
650
805
926
633
567
451
410
559
545
508
437
1943
390
487
Consumer Cyclicals
Consumer Cyclicals
Consumer Cyclicals
Financial Institutions
Basic Industry
Consumer Cyclicals
Consumer Cyclicals
Consumer Cyclicals
Financial Institutions
Telecom & Media
Telecom & Media
Consumer Cyclicals
Basic Industry
Financial Institutions
Consumer Cyclicals
Telecom & Media
Capital Goods
Financial Institutions
Financial Institutions
Basic Industry
Telecom & Media
Consumer Cyclicals
Financial Institutions
Financial Institutions
Telecom & Media
Consumer Cyclicals
Financial Institutions
Telecom & Media
Basic Industry
Financial Institutions
Financial Institutions
Consumer Non-Cycl.
Telecom & Media
Basic Industry
Financial Institutions
Consumer Non-Cycl.
Consumer Non-Cycl.
Basic Industry
Telecom & Media
Technology
5.375%
5.750%
5.625%
7.125%
7.125%
7.125%
5.750%
6.000%
5.000%
6.625%
5.250%
4.625%
5.250%
4.250%
4.125%
4.500%
5.000%
4.650%
3.800%
5.875%
5.250%
3.750%
4.375%
5.500%
6.375%
4.500%
6.500%
4.375%
4.375%
4.000%
5.000%
4.000%
4.375%
4.750%
4.000%
4.875%
5.125%
4.875%
4.625%
5.375%
GMAC
GMAC
FIAT
F
GLEINT
F
GMAC
PEUGOT
CIT
WPPLN
WPPLN
RENTKL
XTALN
CIT
KINGFI
PORTEL
VLVY
CIT
CIT
XTALN
TITIM
VLOF
MS
MS
WKLNA
KINGFI
MS
WPPLN
CLAR
MS
MS
IMTLN
PORTEL
SGOFP
GS
COFP
IMTLN
XTALN
OTE
LMETEL
XS0187751150
XS0177329603
XS0305093311
XS0282593440
XS0359781191
XS0299967413
XS0301812557
FR0010014845
XS0192461837
XS0362329517
XS0329479728
XS0293496815
XS0305188533
XS0215269670
XS0235984340
XS0221854200
XS0302948319
XS0268133799
XS0234935434
XS0366203585
XS0214965963
FR0010206334
XS0270800815
XS0323657527
XS0357251726
XS0178322128
XS0366102555
XS0275759602
XS0249417014
XS0235620142
XS0298899534
XS0236951207
XS0215828913
XS0294547285
XS0211034540
FR0010455626
XS0176838372
XS0305189002
XS0275776283
XS0307504547
Source: J.P. Morgan, iBoxx, Standard and Poors, Modoys. Data as of 2 February 2009.. *: Interp. CDS = From the CDS curve, we interpolate the CDS spread with a maturity equal to the bond maturity.
60
Abel Elizalde
(44-20) 7742-7829
[email protected]
The key to CDS pricing is that at inception of a CDS trade, the fee leg and the
contingent leg of the contract are equal neither the buyer nor the seller has an
economic advantage.
With this in mind, we look at how the usual all running CDS contracts are valued
both at inception and at any subsequent time, leading to a mark-to-market on the
trade.
Finally we show that an upfront plus running contract can be seen as an all
running contract with a mark-to-market.
CDS Pricing in a Nutshell
The basic setup of a CDS contract is shown in Figure 53. The buyer of protection B
pays an annual fee or spread, S, to the seller of protection, C. Should a credit event
occur, the buyer of protection will deliver defaulted bonds with a notional equal to
that of the CDS contract and in return will receive 100% of the notional. We call the
price of these defaulted bonds the recovery rate R (usually a % of the notional).
The net result is that the buyer pays S annually and receives (1-R) in the event of a
credit event; these cash flows are shown in Figure 54. We remind readers of this
process because it highlights an important concept at inception of a CDS contract,
the present value of the spread payments is equal to the present value of the expected
default payments and the value of the contract to both buyer and the seller is equal.
Figure 53: Risk Transfer in a CDS Contract
Periodic Fee, S
Protection
Buyer, B.
Payment in Default of
(1-R)
Protection
Seller, C.
1-R
1-R
This is not to say that the buyer and seller are indifferent to buying or selling
protection. By entering the contract, they are clearly taking a view that one side of
the contract will be more valuable in the future. The buyer thinks he will receive
more from the default event than he pays in annual spreads and the seller thinks he
will receive more from the spreads than he will have to pay out on default. The
important point is that at inception of the contract, the economic value to each
party is equal the present value of the expected default payment (contingent leg)
must equal the present value of the expected annual fees paid (fee leg).
61
Abel Elizalde
(44-20) 7742-7829
[email protected]
S = PD (1 R )
(The discounting factors on each side cancel.) Where:
S = Spread
PD = Probability of Default
R = Recovery-rate in %
The important point to note is that a change to any one of these three components
must change at least one of the others for the equation to balance.
Simple Example
Suppose company XYZ is trading with a spread of 100bp and that recovery rate is
expected to be 40% of the notional. The implied probability of default in this case
can be backed out of Equation 10.
S = PD (1 R )
100bp = PD (100% 40% ) PD =
100bp
(100% 40%)
= 1.7%
SN
i =0
i PS i DF i + AI = (1 R )
PV ( Expected Spread)
PD DF
i= 0
62
Abel Elizalde
(44-20) 7742-7829
[email protected]
CDS spreads take into account expected bond recovery rates: Keeping
default probabilities constant, a lower expected bond recovery rate increases the
bond's credit risk.
2.
CDS spreads take into account the full term structure of default
probabilities: The shape of the default probability term structure conveys
information about conditional default probabilities. For example, an inverted
CDS curve means that if the company survives the next period of time (e.g.
year), the probability of defaulting the subsequent period will be lower.
1
.
10,000
Risky Annuities
We define the risky annuity as the present value of a 1bp risky cash flow or annuity
stream. Suppose we are due to receive 1bp each year for 5 years. In a world with zero
interest rates, the present value of this income stream is worth the sum of the 1bp
payments (5c). However, in a world with non-zero interest rates, we can find the
present value of this income stream by summing up the present value of each
payment of 1bp. This present value calculation is called the Risk-Free Annuity and is
given in Equation 12.
Equation 12: Risk-Free Annuity
N
RiskFreeAnnuity N =
DFi
i =0
where DFi =
(1 + RiskFreeRatei )i
Suppose now that this income stream of 1bp is dependent on company XYZ
surviving at each payment date. This is then a risky annuity and we add a probability
term to realise the likelihood of receiving each cash flow (Equation 13). As the
survival probabilities increase, the 1bp payments become more certain and the risky
annuity increases. The risky annuity is therefore just a simple way of calculating the
present value of a 1bp risky income stream.
Equation 13: Risky Annuity
N
RiskyAnnuity N =
PS i DFi
30
i =0
30
Abel Elizalde
(44-20) 7742-7829
[email protected]
If the risky annuity for a 5-year CDS contract is 4.5, 1bp per year conditional on no
default of company XYZ would be worth 4.5c paid today. If spreads rise, the
survival probabilities decline and the risky annuity also declines resulting in a present
value of less than 4.5c.
We can simplify Equation 11 by using the risky annuity. Since the risky annuity tells
us the present value of 1bp, we can also use it to tell us the present value of SN basis
points. Equation 11 therefore becomes:
Equation 14: Simplifying the CDS Pricing Equation
S N Annuity N = (1 R )
PD DF
i
= PV ( E[1 R ])
i =0
We have seen that pricing CDS relies on equating the expected spreads paid (fee leg)
to the probability weighted default payout (contingent leg). This ensures that at
inception neither the buyer nor the seller has an economic advantage. We now turn to
an example of finding the present value of a Par CDS contract.
Example: Present Value of a Par CDS Contract
Suppose we have a 5-year CDS contract on company XYZ with a maturity of 20
March 2013 and a spread of 100bp. Initially, if the market trades at 100bp, then the
MtM of the contract is zero. This is shown in the Market Val field of the Bloomberg
CDSW calculator. For a Par CDS contract, the risky annuity and risky duration,
Spread DV01, are very close. The PV of the fee leg is therefore equal to
PV ( FeeLeg ) = S 5,0 Annuity 5,0
= 100bp 4.55
= 4.55% of Notional
= $455,000 on a notional of 10m
At inception, this is also the PV of the contingent leg or the expected loss from
default.
We have seen how at inception of a CDS contract, the fee leg and contingent leg are
equal and the value to both buyer and seller is the same. As the market perception of
risk changes and spreads move away from their initial value, there will be a mark-tomarket on the contract.
Marking-to-Market a CDS contract
As we will show, the mark-to-market on a CDS contract is the difference spread
between the initial and final spread multiplied by the risky annuity. Lets
consider the value of the CDS contract to the buyer of protection. They own the
contingent leg and must pay the fee leg. Therefore the value of the contract is:
N
i =0
At inception, as we have said, this value must be equal to zero. As the market
perception of risk changes, the mark-to-market on the contract is the difference in the
value of the contract at the start and at time t.
64
Abel Elizalde
(44-20) 7742-7829
[email protected]
Annuity N ,t [PV0 ( E [1 R ])] S N ,0
Annuity N ,0
= [PVt ( E [1 R ])] S N ,0
i =0
i =0
Annuity N ,t 0
= [PVt ( E [1 R ])] S N ,0
i =0
N
N
Annuity N ,t S N ,0
Annuity N ,t
= S N ,t
i =0
i =0
Annuity N ,t
= (S N ,t S N ,0 )
i =0
So far we have seen how to value CDS contract that trade in all running format,
where the full CDS spread is paid annually, we now to other format of trading CDS.
Full Upfront or Upfront Plus Running CDS
Another way to trade a CDS is to agree on a fixed spread that differs from the market
spread and to pay the difference as an upfront amount. For example, suppose the fair
spread for a CDS contract is 1,500bp but the protection buyer only wants to pay
500bp annually. In order to do so, he will have to compensate the protection seller by
paying him an upfront amount equal to the risky present value of the 1,000bp that he
is underpaying annually. We call this method of trading CDS trading in points
upfront or trading upfront plus running, expressing that there is a points upfront
payment plus running spread payments.
Upfront + Running
i =0
PS i DFi + AI = (1 R )
PV ( Expected Spread)
PD DF
i
i= 0
where
Upfront = The upfront payment on the contract at inception
Running = The fixed Running spread for the contract
Combining Equation 14 and Equation 15 gives us Equation 16.
Equation 16: CDS Pricing Equation
Upfront = (S N Running )
Annuity
+ AI
i =0
This equation is the same as that used to Mark-to-Market a CDS contract. So trading
in points upfront is really no different to trading a CDS contract where the
market spread has moved away from the initial deal spread. For a par CDS
contract, there is no upfront amount paid. However, as spreads move away from the
initial spread, the contract will have a mark-to-market based on where the current
market spreads are relative to the initial deal spreads. A contract with an initial
spread of 500bp, with a market spread now at 1,500bp is simply an upfront plus
running contract.
65
Abel Elizalde
(44-20) 7742-7829
[email protected]
3.
Spread to benchmark
4.
I-spreads
5.
6.
Z-spreads
7.
Present Value =
C
C
C
C
1+ C
+
+
+
+
1
2
3
4
(1 + r1 )
(1 + r2 )
(1 + r3 )
(1 + r4 )
(1 + r5 ) 5
5
(1 + r )
i =1
1
(1 + r5 ) 5
For an n-year bond with coupons paid with a frequency of f per year we get
Equation 18: Bond Pricing
N =n f
Dirty price =
C/ f
[1 + r / f ]
i =1
ti
f ti
[1 + r
tN
1
/f
f t N
where ti is the time (in years from today) of each coupon payment and ri is the annual
zero-rate for a payment at time ti.
66
Abel Elizalde
(44-20) 7742-7829
[email protected]
Dirty price =
i =1
C/ f
[1 + y / f ]
f ti
[1 + y / f ] f t
67
Abel Elizalde
(44-20) 7742-7829
[email protected]
1. Spread to benchmark
Spread to benchmark (or benchmark spread) is the difference between the
yield-to-maturity (YTM) on a corporate bond and the YTM of a benchmark
bond. Investors buying a corporate rather than risk-free bond will receive an
additional yield as compensation for holding the corporate bond rather than the riskfree bond. We call this additional yield on the corporate bond the benchmark spread:
it is the compensation for the default risk and lower liquidity of the corporate bond.
We usually quote benchmark spreads with reference to a relevant government bond;
however we could also quote it with respect to a benchmark swap rate.
Figure 55: Example Benchmark Spread
x-axis: Years to Maturity; y-axis: Yield (%)
10%
Risk-Free Curv e
Benchmark Yield
8%
6%
4%
2%
0%
0
10
In order to calculate the benchmark spread, we calculate the yield on the risky bond
using Equation 19 and then calculate the yield on the reference government bond.
The difference between these tells us the benchmark yield.
Equation 20: Spread to Benchmark
Substituting into the equation for yield calculation, Equation 19, we get:
Equation 21: Bond Pricing
N = n f
Dirty price =
i =1
N = n f
i =1
C/ f
[1 + y / f ]
f ti
[1 + y / f ] f t
C/ f
[1 + ( y B + S B ) / f ]
where
y = Bond Yield to Maturity
yB = Benchmark Bond Yield to Maturity
SB = Benchmark Spread
68
f ti
[1 + ( y B + S B ) / f ] f t
Abel Elizalde
(44-20) 7742-7829
[email protected]
Investors often refer to the benchmark spread as the Govie spread as it is frequently
used for quotation and trading purposes. The reason for this is that it provides an
alternative to quoting bond prices. Since investors must ultimately pay the bond price
when buying a bond, we need an unambiguous method for calculating this price.
Given the relationship between bond prices and bond yields (Equation 19) quoting a
yield is equivalent to quoting a price. Since the yield for the benchmark bond is well
defined and easily available from a variety of pricing sources, we can add the 'Govie'
spread to this in order to get the corporate bond yield. Using Equation 19 we can
convert the corporate bond yield into a price.
The benchmark spread is a good alternative to quoting bond yields, it is however less
useful as an indicator of excess return or default probabilities of investing in
corporate rather than government bonds. While the maturity of the benchmark bond
is usually chosen to be close to, it is not exactly the same as that of the corporate
bond. The additional maturity of the corporate bond will result in additional yield,
which is not related to credit risk.
The benefit of using benchmark spreads are ease of calculation and of quotation for
trading purposes. The problems with using this spread measure are:
The mismatch in maturity will tend to overestimate the credit risk
As with all yields, the benchmark spread assumes that income is reinvested at the
bond yield
No account taken for the term structure of interest rates, since a single yield is
assumed
The I-spread, which we discuss next, corrects for the maturity mismatch, although it
also assumes that income is reinvested in the bond.
2. I-spread
The I-spreads is the difference in yield between a corporate bond and a
maturity matched benchmark bond; it is similar to benchmark spreads but more
accurately reflects the credit risk of a bond. Since the maturity of the benchmark
bond usually differs from the maturity of the corporate bond, the benchmark spread
will tend to overestimate the credit risk of a bond. Suppose we have a 7.5-year
corporate bond and a 7-year benchmark bond. If the benchmark yield curve is
upward sloping then the benchmark spread is the difference between the 7.5-year
yield on the corporate bond and the 7-year yield on the government bond. Since the
yield on a 7-year government bond is less than the yield on a 7.5-year government
bond, some of the benchmark spread is simply compensation for the additional 0.5years of maturity.
One way to adjust for this maturity mismatch is to calculate the yield of a theoretical
7.5-year benchmark bond. We do this by interpolating between the 7-year and the 8year bond. Once we have the yield on the theoretical 7.5-year benchmark bond, we
can compare it to the yield on the 7.5-year corporate bond. This will give a more
accurate measure of the additional yield on the corporate bond relative to the
benchmark bonds.
69
Abel Elizalde
(44-20) 7742-7829
[email protected]
The I-spread is then the difference between the yield on a corporate bond and
the yield on a theoretical maturity matched risk-free bond. As with benchmark
bonds, the I-spread can be seen as the additional yield above government rates or
swaps depending on what we use as the benchmark. The usual convention is to refer
to I-spreads as spreads above swap.
Figure 56: Example I-spread
x-axis: Years to Maturity; y-axis: Yield (%)
10%
Risk-Free Curv e
Benchmark Yield
8%
6%
4%
2%
0%
0
10
Substituting into the equation for yield calculation, Equation 19, we get something
that looks very similar to our calculation of the benchmark spread. The difference
here is that we use an interpolated benchmark bond yield, or the maturity matched
swap rate.
Equation 23: Bond Pricing
N =n f
Dirty price =
i =1
N =n f
i =1
C/ f
[1 + y / f ]
f ti
[1 + y / f ] f t
C/ f
[1 + ( y I + S I ) / f ]
f ti
[1 + ( y I + S I ) / f ] f t
where
y = Bond Yield to Maturity
yI = Interpolated Benchmark Bond Yield or Swap Rate to Maturity
SI =I-spread
While the I-spread adjusts for the maturity mismatch, and therefore better reflects the
credit risk of a corporate bond, it is less suited for quoting purposes. Since market
participants often construct the risk-free curve using different interpolation methods,
the yield for a theoretical risk-free bond it is not as universally defined as the yield
for a particular benchmark government bond. This could lead to different price
calculations which would hamper trading.
70
Abel Elizalde
(44-20) 7742-7829
[email protected]
Additionally, the I-spread looks at the spread above a single point, the yield. A
further improvement could be to calculate the credit spread above the risk-free curve.
The benefits of the using the I-spread are:
Account more precisely for the maturity of the bond
Ease of calculation as swap rates are easily available
The problems with using this spread measure are
As with all yields, the I-spread assumes that income is reinvested at the bond
yield
No account taken for the term structure of interest rates, since a single yield is
assumed
Different interpolation methods may lead to small differences in pricing
Comparing Benchmark Spread to I-spreads
We noted earlier that the spread to benchmark (or benchmark spread) is the
difference between the yield-to-maturity (YTM) on a corporate bond and the YTM of
a benchmark government bond:
Benchmark Spread = Bond Yield - Benchmark Bond Yield
And the I-spread is the difference in yield between a corporate bond and a maturity
matched, or interpolated benchmark bond:
The I-spread is therefore equal to the benchmark spread plus the difference between
the benchmark bond yield and the interpolated bond yield. For an upward sloping
yield curve, the interpolated bond spread will be higher than the benchmark bond
yield, which usually has a shorter maturity, so the difference between these will
always be negative. In an upward sloping yield environment, the I-spread is therefore
always less than the benchmark spread. For a downward sloping yield curve, the
reverse would hold.
Changing yield curves will impact the fixed difference between the I-spread
and benchmark spread. A parallel shift to the yield curve will preserve the
difference between the benchmark yield and the interpolated benchmark yield
and will therefore not affect the relationship between the I-spread and benchmark
spread. A flattening or steepening yield curve however will decrease this
difference (flatting) or increase the difference (steepening).
Recovery rates are not accounted for in either method and will therefore
have no effect on the benchmark spread and I-spread.
71
Abel Elizalde
(44-20) 7742-7829
[email protected]
3. Z-spread
The Z-spread is the parallel shift applied to the zero curve in order to equate the
bond price to the present value of the cash flows. We noted earlier that the price of
a bond is equal to the present value of the expected cash flows. For a risk-free bond,
each cash flow is discounted by the risk-free discount rate, rti, for that maturity ti,
(Equation 24).
Equation 24: Pricing a Risk-Free Bond
N =n f
C/ f
[1 + r / f ]
Dirty price =
i =1
ti
f ti
[1 + r
tN
/f
f t N
We also noted that the price of a risky corporate bond will be lower than the price of
a risk-free bond. The price of the bond will therefore be less than the present value of
the cash flows as we might not receive all these cash flows should the bond default.
In order to equate the price of a risky corporate bond to the present value of the
expected cash flows, we therefore increase each discount rate by an amount, z, the Zspread (Equation 25). This has the effect of lowering the present value of each cash
flow.
Equation 25: Pricing a Risky Bond
N = n f
Dirty price =
[1 + (r
i =1
ti
C/ f
) ]
+z / f
f ti
[1 + (r
tN
) ]f t
+z / f
Risk-Free Curv e
Risky Curv e
8%
6%
4%
2%
0%
0
Source: J.P. Morgan.
72
10
Abel Elizalde
(44-20) 7742-7829
[email protected]
The Z-spread therefore represents the credit risk in the bond. Z-spreads are
useful values of the credit risk of a corporate bond as they take into account both the
maturity and full term structure of interest rates. However, while Z-spreads indicate
credit risk, they cannot be traded. Asset swap spreads however can be traded. We
turn to these next.
Dirty price =
[1 + ( y
i =1
C
+ S I )]
[1 + ( y I
+ S I )]n
Dirty price =
[1 + (r + z )]
C
i =1
[1 + (ri + z )]i
[1 + (r + z )]
i =1
[1 + (ri + z )]
[1 + ( y
i =1
C
+ S I )]
[1 + ( y I
+ S I )]n
So if we have a flat discount curve equal to the interpolated yield, then the I-spread
and the Z-spread will be equal. If the curve is upward sloping then the Z-spread is
greater than the I-spread. The converse holds if the curve is downward sloping.
73
Abel Elizalde
(44-20) 7742-7829
[email protected]
Changing the zero curve will have an impact on the difference between the
Z-spread and the I-spread. With a flat curve, the I-spread and the Z-spread are
equal. As the curve pivots into an upward sloping curve, Z-spread will increase
above the I-spread. In a flattening scenario the reverse holds and the Z-spread
decreases relative to the I-spread.
Recovery rate is not accounted for in either method and will therefore have
no effect on the Z-spread and I-spread
Par asset swaps: the price of the asset swap and notional swapped is par
True asset swaps: the price of the asset swap and notional swapped is the
market (dirty) price of the bond
Our aim here is to look at the cash flows and calculations of each to better understand
what the asset swap spread represents.
74
Abel Elizalde
(44-20) 7742-7829
[email protected]
Figure 58: Asset Swap Package + Hedge: Cash flows from the Investors Perspective
Libor + Spread
Spread
+
Libo r
+
100
100 (Par)
Upfront
payments
100
(Dirty Price)
100
Coupons
+
Cpn
100
Coupons
+
Cpn
100
(Dirty Price)
Source: J.P. Morgan.
Although the bond and swap are traded as a package, the swap does not knock out in
the case of default. This means that if the bond defaults, the investor will be exposed
to interest rate risk, as well as any remaining MtM position resulting from the bond
trading away from par at inception.
Par Asset Swaps
The par asset swap spread is equal to the difference between the risk-free present
value of the bond coupons (plus principal) and the price paid for the bond. Since the
spread is paid on a running basis, we divide this amount by the annuity to find the
running spread amount. In the grey box below we will show that:
Equation 26: Par Asset Swap Spread Calculation
Spread =
An intuitive way to think about this is as follows. Suppose an investor has a choice of
two bonds, a risk-free bond and a risky bond, both with the same maturity and
coupon. The risky bond will be cheaper than the risk-free bond because there is a
chance that the risky bond defaults and the investor does not get back his money or
coupons. The difference in value between these two bonds is the compensation for
buying the risky bond over the corporate bond. Rather than paying this on an upfront
basis, we can pay it as an annual spread by dividing by the annuity.
75
Abel Elizalde
(44-20) 7742-7829
[email protected]
This spread is compensation because if the risky bond defaults, the investor will no
longer receive the coupons nor will they receive the notional at maturity; they would
still need to make the payments on the risk-free bond of both the coupons and
principal at maturity.
First lets consider a simple asset swap where an investor buys a bond and swaps the
fixed coupons into floating rate coupons through a swap (Figure 58). Ignoring the
bond for the moment, lets consider the swap contract first. The swap contract
essentially has two legs a floating leg which we will assume the investor is
receiving and a fixed leg, which we assume the investor is paying. The fixed leg is
like shorting a fixed rate bond, the investor receives the upfront price and will pay
the fixed coupon payments plus the notional at maturity. The floating leg is like
being long a floating rate bond priced at par; the investor will pay out the notional
amount and will receive a floating rate coupon plus the notional amount at expiry.
The floating rate is set to be equal to Libor + Spread.
Now the value of a contract to pay fixed coupons is:
Equation 27: Value of Fixed Contract
i.e. we receive the bond price but have to pay away the coupons plus the principal at
maturity.
And the value of a contract to receive floating coupons is:
Equation 28: Value of Floating Contract
i.e. we pay a 1 in order to receive the floating rate coupons plus the principal at
maturity. Since the initial value of the floating rate leg must be equal to the initial
value of the fixed rate leg, we get:
Equation 29: Equation the legs of the contract
Now in a par asset swap, the notional of the swap is set to be equal to the notional of
the bond, i.e. 1. We also set the Floating leg payments to be equal to Libor + Spread.
Equation 29 then becomes:
1 (PV[LIBOR + Spread + Principal] - 1) = 1 (PV[Coupon + Principal] - BondPrice )
PV[LIBOR + Principal] + PV[Spread ] - 1 = PV[Coupon + Principal] - BondPrice
Spread =
76
Abel Elizalde
(44-20) 7742-7829
[email protected]
Here we set the notional in the swap to be equal to the bond price.
BondPrice (PV [LIBOR + Spread + Principal ] - 1) = 1 (PV [Coupon + Principal ] - BondPrice )
BondPrice (PV [LIBOR + Principal ] + PV [Spread ] - 1) = PV [Coupon + Principal ] - BondPrice
PV [Coupon + Principal ] - BondPrice
PV [Spread ] =
BondPrice
77
Abel Elizalde
(44-20) 7742-7829
[email protected]
Spread =
TrueASW =
Bond Pr ice
When bonds are trading above par, then the true ASW spread will be lower than the
par ASW spread and vice-versa.
The benefits of using the asset swap spread are:
Asset swap spreads can be traded and investors can receive the fixed spread for
the life of the swap
The problems with using this spread measure are
Assumes that spreads are paid with certainty which causes over/underestimation
of credit risk for bonds trading above/under par
Calculations are not straightforward
Comparing Asset Swap Spreads to Z-spreads
We showed earlier that the par ASW spread is given by Equation 33.
Equation 33: Par Asset Swap Spread Calculation
ASW =
Alternatively, we can write this as Equation 34 where SwapRate is the swap rate for
the maturity of the bond with the same conventions. The Annuity here is the risk-free
annuity and is the present value of 1bp.
Equation 34: Par Asset Swap Spread Calculation
ASW =
Dirty price =
[1 + (r
i =1
ti
+z
)] [1 + (rt
ti
1
N
+z
)]t
If we let Y be the yield on the bond, then the present value of a bond with a coupon
of Y is 1. This gives:
78
Abel Elizalde
(44-20) 7742-7829
[email protected]
1=
[1 + (r
i =1
ti
+z
)] [1 + (rt
ti
1
N
+z
)]t
The next step is to assume a flat curve such that Y = r + z, where the yield is equal to
the risk-free rate plus the Z-spread. This gives us
n
1=
r+z
[1 + (r + z )]
ti
i =1
[1 + (r + z )]t
DP 1 =
[1 + (r + z )]
C
i =1
n
ti
[1 + (r + z )]
tN
r+z
[1 + (r + z )]
i =1
ti
[1 + (r + z )]t
r+z
[1 + (r + z )] [1 + (r + z )]
i =1
ti
i =1
tN
this says that if we present value the coupon minus the swap rate minus the Z-spread
we get the difference between the dirty price and 1. By substituting this into the asset
swap calculation, we find that
(Bond Pr ice 1)
ASW =
Annuity
(Bond Pr ice 1) + Zspread + 1-Bond Pr ice
=
RiskyAnnuity
Annuity
1
1
= Zspread + (1 Bond Pr ice)
Annuity RiskyAnnuity
Therefore, for a flat interest rate curve, if the risky annuity is equal to the Annuity
then both the Z-spread and the asset swap spread are zero. If the spreads are nonzero, then the risky annuity will be less than the risk-free annuity. We then get
If BondPrice >1 then ASW > Z-spread
If BondPrice <1 then ASW < Z-spread
When the riskless curve is upward slopping, the forward Libor rates will increase
over time; therefore, in case of a future default the interest rate swap on the asset
swap package will have a higher value than at inception. To compensate for this the
asset swap spread will decrease.
79
Abel Elizalde
(44-20) 7742-7829
[email protected]
a)
Z-spread is a movement of the discount (zero) curve, but true asset swap spread
represents a spread above the swap curve (from which the zero curve is
bootstrapped).
b) The true asset swap payments are paid on swap conventions, so are paid
quarterly on a 30/360 basis. The zero curve spread represents discount rates at
each bond payment.
The final spread measure we look at is the Par Equivalent CDS Spread (PECS). This
methodology explicitly accounts for recovery of a bond following default and is our
preferred method of calculating the bond CDS basis.
Calculate the implied survival probabilities from the market price of the
bond. The price of a risk-free bond is equal to the present value of the cash
flows, including the bond coupons plus and notional amount paid back at
maturity. For a corporate bond, where both the coupons and the notional amount
are risky (i.e. contingent on the survival of the company), the price of the bond
should be equal to the expected present value of the coupons plus notional. In
case of default, we assume the bond recovers a fraction of notional R, which is
the recovery rate assumed for pricing CDS contracts referencing the same
company.
SN
i =0
80
i PS i DFi + AI = (1 R)
PD DF
i
i =0
Abel Elizalde
(44-20) 7742-7829
[email protected]
P = PS (1 + C )
where
C = Coupon
PS = Probability of Survival
R = Recovery-rate in %
Once we have the implied probabilities, we can back out the spread using
S = (1 PS ) (1 R )
Simple Example
Suppose a bond issued by company XYZ is trading with a price of 100% and that the
bond coupon is 7% and the 1-year risk-free rate is 5%. Let us assume that the
expected recovery on the bond is 0%. Using Equation 34 we get:
1
1
+ R
1+ r
1+ r
1
1
1 = PS ((1 + 7%) 0 )
+ 0%
1 + 5%
1 + 5%
1.07
1 = PS
1.05
1.05
PS =
98%
1.07
P = PS ((1 + C ) R )
S = ( PD) (1 R ) = (1 PS ) (1 R )
S = (1 98%) (1 0% )
S = 2% = 200bp
81
Abel Elizalde
(44-20) 7742-7829
[email protected]
PV(E[Coupon Payments]) = C
PS i DFi
i =0
PV(E[Principal]) = 1 PS N DFNi + R
PD DF
i
i =0
Bond Price = C
i PS i DFi + 1 PS N DFN + R
i =0
PD DF
i
i =0
C = Bond Coupon
i = Length of time period i in years
PSi = Probability of survival to time i, as at time t0
PDi = Probability of default at time i,as at time t0 = (1 - PSi)
DFi = Risk-free discount factor to time i, as at time t0
R = Recovery rate on default
Readers familiar with CDS pricing will notice the similarity between Equation 40
and the standard CDS pricing equation (Equation 41).
Equation 41: CDS Pricing Equation
N
SN
i =0
PS i DFi + AI = (1 R )
PD DF
i
i =0
We now follow the two step procedure to calculate the PECS: (1) Using Equation 40
we calculate the implied survival probabilities from the bond, and (2) Use Equation
41, to back out the implied CDS spread.
In practice we reverse these steps. Using the full CDS curve traded in the market,
we calculate the implied survival probabilities for the company. By applying a
parallel shift to the survival curve, we use Equation 40 to match the price of the
bond to the expected value of the cash flows.31 In case of default, we assume the
bond recovers a fraction of notional R, which is the recovery rate assumed for pricing
CDS contracts referencing the same company. Once we have matched the bond
price, we convert these survival probabilities back into spreads.
31
In particular, we apply a parallel shift to all the hazard rates which characterise the term
structure of default probabilities. Generally, we use a piece-wise constant hazard rate model
calibrated to the available term structure of CDS.
82
Abel Elizalde
(44-20) 7742-7829
[email protected]
Since the survival probabilities implied by the CDS are explicitly dependent on the
assumed recovery rate R, using these probabilities to derive the PECS means it is
also explicitly dependent on R. Moreover, as we explained in the previous paragraph,
we first derive the default probabilities using the full CDS curve traded in the market,
and then use them to compute the PECS. Thus the PECS explicitly takes into account
the term structure of default probabilities.
The PECS therefore explicitly accounts for the recovery of the bond and is similar
to the pricing methodology used for CDS. Moreover, the PECS takes into account
the term structure of default probabilities implied by the CDS market in order to
derive the credit risk (spread) of the bond.
Comparing PECS to CDS
Intuitively, as we explained above, PECS can be thought of as a shift in the term
structure of CDS spreads in order to match the price of the bond. In particular, to
compute the PECS:
10. We take as inputs the term structure of default probabilities and recovery rate
derived from the CDS market.
Using these inputs and Equation 40, one could derive the CDS-implied bond
price. In general, such price does not coincide with the bond market (dirty)
price.
11. We find the shift we need to apply to those default probabilities in order to
match the bond price. We maintain the recovery rate assumption and shape of
the default probability term structure (we just shift it in parallel).
When the CDS-implied bond price is higher than the bond market price, the
PECS will be higher than the CDS spread: we need to add an extra component of
default risk to the CDS spreads in order to match the bond price.
The result is a PECS which shares with the CDS spread the recovery rate assumption
and the shape of the term structure of default probabilities.
Comparing PECS to Z-spread
PECS depends on the term structure of default probabilities and on the expected
recovery rate, whereas Z-spreads are not affected by them.
We saw earlier that we can calculate the survival probabilities from the bond price.
Equation 42: Bond Pricing Equation
N
i =0
i =0
83
Abel Elizalde
(44-20) 7742-7829
[email protected]
i =0
N
i =0
i =0
i =0
If we set the recovery to be zero, then this looks very similar to the calculation of the
Z-spread in continuous time.
Bond Price = C e (ri + s )ti + 1 e (rN + s )t N
N
i =0
Now since
ex = 1 x +
x2
+ O( x 3 )
2
1
= 1 x + x 2 + O( x 3 )
1+ x
then
ex <
1
1+ x
so in order to equate the cash prices, the zero recovery PECS (in continuous time)
needs to be lower than the Z-spread (in discrete time). In the case where the recovery
rate is non zero, the PECS increases relative to the zero-recovery PECS.
In the following tables we provide a summary of the main features of the various
spreads we have looked at.
84
Abel Elizalde
(44-20) 7742-7829
[email protected]
I-Spread
I-Spread = Bond Yield Interpolated Bond Yield
Z-Spread
The Z-spread is the parallel shift applied to the zero curve that equates
the bond price to the PV of the cash flows
Definition
Benchmark Spread
Benchmark Spread = Bond Yield Benchmark Yield
N = n f
i =1
C/ f
[1 + y / f ]
N = n f
i =1
f ti
[1 + y / f ]
C/ f
[1 + ( y B
N = n f
+ S B )/ f ] f t i
Dirty price =
f t N
[1 + ( y B
+ S B )/ f ] f t N
i =1
N = n f
i =1
C/ f
[1 + y / f ]
f ti
N = n f
[1 + y / f ]
C/ f
Dirty price =
f t N
i =1
ti
C/ f
) ]f t [1 + (rt
+z / f
) ]f t
+z / f
Recovery
[1 + ( y I + S I ) / f ] f t [1 + ( y I + S I ) / f ] f t
i
[1 + (r
Comparison
Downside
Calculation
Dirty price =
85
Calculation
Definition
Abel Elizalde
(44-20) 7742-7829
[email protected]
Spread =
Spread =
Bond Price = C
PS DF + 1 PS
i
DFN
i =0
N
+ R
PD DF
i
Comparison
Downside
Benefit
i =0
Asset swap spreads can be traded and investors can receive the
fixed spread for the life of the swap
Uses CDS pricing methodology and accounts for recovery rates and
credit curve
Recovery
86
TrueASW =
ParASW
Bond Pr ice
Explicitly accounts for the recovery of the bond and takes into account
the term structure of default probabilities implied by the CDS market.
Abel Elizalde
(44-20) 7742-7829
[email protected]
87
Abel Elizalde
(44-20) 7742-7829
[email protected]
88
Abel Elizalde
(44-20) 7742-7829
[email protected]
Analyst Certification:
The research analyst(s) denoted by an AC on the cover of this report certifies (or, where multiple research analysts are primarily
responsible for this report, the research analyst denoted by an AC on the cover or within the document individually certifies, with
respect to each security or issuer that the research analyst covers in this research) that: (1) all of the views expressed in this report
accurately reflect his or her personal views about any and all of the subject securities or issuers; and (2) no part of any of the research
analysts compensation was, is, or will be directly or indirectly related to the specific recommendations or views expressed by the
research analyst(s) in this report.
Important Disclosures
Important Disclosures for Credit Research Compendium Reports: Important disclosures, including price charts for all companies
under coverage for at least one year, are available through the search function on J.P. Morgans website
https://mm.jpmorgan.com/disclosures/company or by calling this U.S. toll-free number (1-800-477-0406)
Explanation of Credit Research Ratings:
Ratings System: J.P. Morgan uses the following sector/issuer portfolio weightings: Overweight (over the next three months, the
recommended risk position is expected to outperform the relevant index, sector, or benchmark), Neutral (over the next three months, the
recommended risk position is expected to perform in line with the relevant index, sector, or benchmark), and Underweight (over the next
three months, the recommended risk position is expected to underperform the relevant index, sector, or benchmark). J.P. Morgans
Emerging Market research uses a rating of Marketweight, which is equivalent to a Neutral rating.
Valuation & Methodology: In J.P. Morgans credit research, we assign a rating to each issuer (Overweight, Underweight or Neutral)
based on our credit view of the issuer and the relative value of its securities, taking into account the ratings assigned to the issuer by credit
rating agencies and the market prices for the issuers securities. Our credit view of an issuer is based upon our opinion as to whether the
issuer will be able service its debt obligations when they become due and payable. We assess this by analyzing, among other things, the
issuers credit position using standard credit ratios such as cash flow to debt and fixed charge coverage (including and excluding capital
investment). We also analyze the issuers ability to generate cash flow by reviewing standard operational measures for comparable
companies in the sector, such as revenue and earnings growth rates, margins, and the composition of the issuers balance sheet relative to
the operational leverage in its business.
J.P. Morgan Credit Research Ratings Distribution, as of December 31, 2008
EMEA Credit Research Universe
IB clients*
Overweight
Neutral
Underweight
21%
65%
52%
71%
27%
67%
Represents Ratings on the most liquid bond or 5-year CDS for all companies under coverage.
*Percentage of investment banking clients in each rating category.
Analysts Compensation: The research analysts responsible for the preparation of this report receive compensation based upon various
factors, including the quality and accuracy of research, client feedback, competitive factors and overall firm revenues. The firms overall
revenues include revenues from its investment banking and fixed income business units.
Other Disclosures
J.P. Morgan is the global brand name for J.P. Morgan Securities Inc. (JPMSI) and its non-US affiliates worldwide.
Options related research: If the information contained herein regards options related research, such information is available only to persons who
have received the proper option risk disclosure documents. For a copy of the Option Clearing Corporations Characteristics and Risks of
Standardized Options, please contact your J.P. Morgan Representative or visit the OCCs website at
http://www.optionsclearing.com/publications/risks/riskstoc.pdf.
Legal Entities Disclosures
U.S.: JPMSI is a member of NYSE, FINRA and SIPC. J.P. Morgan Futures Inc. is a member of the NFA. JPMorgan Chase Bank, N.A. is a
member of FDIC and is authorized and regulated in the UK by the Financial Services Authority. U.K.: J.P. Morgan Securities Ltd. (JPMSL) is a
member of the London Stock Exchange and is authorised and regulated by the Financial Services Authority. Registered in England & Wales No.
2711006. Registered Office 125 London Wall, London EC2Y 5AJ. South Africa: J.P. Morgan Equities Limited is a member of the Johannesburg
Securities Exchange and is regulated by the FSB. Hong Kong: J.P. Morgan Securities (Asia Pacific) Limited (CE number AAJ321) is regulated
by the Hong Kong Monetary Authority and the Securities and Futures Commission in Hong Kong. Korea: J.P. Morgan Securities (Far East) Ltd,
Seoul branch, is regulated by the Korea Financial Supervisory Service. Australia: J.P. Morgan Australia Limited (ABN 52 002 888 011/AFS
89
Abel Elizalde
(44-20) 7742-7829
[email protected]
Licence No: 238188) is regulated by ASIC and J.P. Morgan Securities Australia Limited (ABN 61 003 245 234/AFS Licence No: 238066) is a
Market Participant with the ASX and regulated by ASIC. Taiwan: J.P.Morgan Securities (Taiwan) Limited is a participant of the Taiwan Stock
Exchange (company-type) and regulated by the Taiwan Securities and Futures Bureau. India: J.P. Morgan India Private Limited is a member of
the National Stock Exchange of India Limited and Bombay Stock Exchange Limited and is regulated by the Securities and Exchange Board of
India. Thailand: JPMorgan Securities (Thailand) Limited is a member of the Stock Exchange of Thailand and is regulated by the Ministry of
Finance and the Securities and Exchange Commission. Indonesia: PT J.P. Morgan Securities Indonesia is a member of the Indonesia Stock
Exchange and is regulated by the BAPEPAM. Philippines: J.P. Morgan Securities Philippines Inc. is a member of the Philippine Stock Exchange
and is regulated by the Securities and Exchange Commission. Brazil: Banco J.P. Morgan S.A. is regulated by the Comissao de Valores
Mobiliarios (CVM) and by the Central Bank of Brazil. Mexico: J.P. Morgan Casa de Bolsa, S.A. de C.V., J.P. Morgan Grupo Financiero is a
member of the Mexican Stock Exchange and authorized to act as a broker dealer by the National Banking and Securities Exchange Commission.
Singapore: This material is issued and distributed in Singapore by J.P. Morgan Securities Singapore Private Limited (JPMSS) [MICA (P)
132/01/2009 and Co. Reg. No.: 199405335R] which is a member of the Singapore Exchange Securities Trading Limited and is regulated by the
Monetary Authority of Singapore (MAS) and/or JPMorgan Chase Bank, N.A., Singapore branch (JPMCB Singapore) which is regulated by the
MAS. Malaysia: This material is issued and distributed in Malaysia by JPMorgan Securities (Malaysia) Sdn Bhd (18146-X) which is a
Participating Organization of Bursa Malaysia Berhad and a holder of Capital Markets Services License issued by the Securities Commission in
Malaysia. Pakistan: J. P. Morgan Pakistan Broking (Pvt.) Ltd is a member of the Karachi Stock Exchange and regulated by the Securities and
Exchange Commission of Pakistan. Saudi Arabia: J.P. Morgan Saudi Arabia Ltd. is authorised by the Capital Market Authority of the Kingdom
of Saudi Arabia (CMA) to carry out dealing as an agent, arranging, advising and custody, with respect to securities business under licence number
35-07079 and its registered address is at 8th Floor, Al-Faisaliyah Tower, King Fahad Road, P.O. Box 51907, Riyadh 11553, Kingdom of Saudi
Arabia.
Country and Region Specific Disclosures
U.K. and European Economic Area (EEA): Unless specified to the contrary, issued and approved for distribution in the U.K. and the EEA by
JPMSL. Investment research issued by JPMSL has been prepared in accordance with JPMSL's policies for managing conflicts of interest arising
as a result of publication and distribution of investment research. Many European regulators require that a firm to establish, implement and
maintain such a policy. This report has been issued in the U.K. only to persons of a kind described in Article 19 (5), 38, 47 and 49 of the Financial
Services and Markets Act 2000 (Financial Promotion) Order 2005 (all such persons being referred to as "relevant persons"). This document must
not be acted on or relied on by persons who are not relevant persons. Any investment or investment activity to which this document relates is only
available to relevant persons and will be engaged in only with relevant persons. In other EEA countries, the report has been issued to persons
regarded as professional investors (or equivalent) in their home jurisdiction. Australia: This material is issued and distributed by JPMSAL in
Australia to wholesale clients only. JPMSAL does not issue or distribute this material to retail clients. The recipient of this material must not
distribute it to any third party or outside Australia without the prior written consent of JPMSAL. For the purposes of this paragraph the terms
wholesale client and retail client have the meanings given to them in section 761G of the Corporations Act 2001. Germany: This material is
distributed in Germany by J.P. Morgan Securities Ltd., Frankfurt Branch and J.P.Morgan Chase Bank, N.A., Frankfurt Branch which are
regulated by the Bundesanstalt fr Finanzdienstleistungsaufsicht. Hong Kong: The 1% ownership disclosure as of the previous month end
satisfies the requirements under Paragraph 16.5(a) of the Hong Kong Code of Conduct for persons licensed by or registered with the Securities
and Futures Commission. (For research published within the first ten days of the month, the disclosure may be based on the month end data from
two months prior.) J.P. Morgan Broking (Hong Kong) Limited is the liquidity provider for derivative warrants issued by J.P. Morgan
International Derivatives Ltd and listed on The Stock Exchange of Hong Kong Limited. An updated list can be found on HKEx website:
http://www.hkex.com.hk/prod/dw/Lp.htm. Japan: There is a risk that a loss may occur due to a change in the price of the shares in the case of
share trading, and that a loss may occur due to the exchange rate in the case of foreign share trading. In the case of share trading, JPMorgan
Securities Japan Co., Ltd., will be receiving a brokerage fee and consumption tax (shouhizei) calculated by multiplying the executed price by the
commission rate which was individually agreed between JPMorgan Securities Japan Co., Ltd., and the customer in advance. Financial Instruments
Firms: JPMorgan Securities Japan Co., Ltd., Kanto Local Finance Bureau (kinsho) No. 82 Participating Association / Japan Securities Dealers
Association, The Financial Futures Association of Japan. Korea: This report may have been edited or contributed to from time to time by
affiliates of J.P. Morgan Securities (Far East) Ltd, Seoul branch. Singapore: JPMSS and/or its affiliates may have a holding in any of the
securities discussed in this report; for securities where the holding is 1% or greater, the specific holding is disclosed in the Important Disclosures
section above. India: For private circulation only, not for sale. Pakistan: For private circulation only, not for sale. New Zealand: This
material is issued and distributed by JPMSAL in New Zealand only to persons whose principal business is the investment of money or who, in the
course of and for the purposes of their business, habitually invest money. JPMSAL does not issue or distribute this material to members of "the
public" as determined in accordance with section 3 of the Securities Act 1978. The recipient of this material must not distribute it to any third
party or outside New Zealand without the prior written consent of JPMSAL.
General: Additional information is available upon request. Information has been obtained from sources believed to be reliable but JPMorgan
Chase & Co. or its affiliates and/or subsidiaries (collectively J.P. Morgan) do not warrant its completeness or accuracy except with respect to any
disclosures relative to JPMSI and/or its affiliates and the analysts involvement with the issuer that is the subject of the research. All pricing is as
of the close of market for the securities discussed, unless otherwise stated. Opinions and estimates constitute our judgment as of the date of this
material and are subject to change without notice. Past performance is not indicative of future results. This material is not intended as an offer or
solicitation for the purchase or sale of any financial instrument. The opinions and recommendations herein do not take into account individual
client circumstances, objectives, or needs and are not intended as recommendations of particular securities, financial instruments or strategies to
particular clients. The recipient of this report must make its own independent decisions regarding any securities or financial instruments
mentioned herein. JPMSI distributes in the U.S. research published by non-U.S. affiliates and accepts responsibility for its contents. Periodic
updates may be provided on companies/industries based on company specific developments or announcements, market conditions or any other
publicly available information. Clients should contact analysts and execute transactions through a J.P. Morgan subsidiary or affiliate in their home
jurisdiction unless governing law permits otherwise.
90
Abel Elizalde
(44-20) 7742-7829
[email protected]
Copyright 2009 JPMorgan Chase & Co. All rights reserved. This report or any portion hereof may not be reprinted, sold or
redistributed without the written consent of J.P. Morgan.
91
Abel Elizalde
(44-20) 7742-7829
[email protected]
Team Assistant
Laura Hayes
(44 20) 7777-2280
[email protected]
Tatiana Tchembarova
(44-20) 7325-7278
[email protected]
www.morganmarkets.com/analyst/tatianatchembarova
Anh Dang
(44-20) 7777-4754
[email protected]
www.morganmarkets.com/analyst/anhdang
Andrew Webb
(44-20) 7777 0450
[email protected]
www.morganmarkets.com/analyst/andrewwebb
Credit Strategy
Stephen Dulake
(44-20) 7325-5454
[email protected]
www.morganmarkets.com/analyst/stephendulake
Daniel Lamy
(44-20) 7777-1875
[email protected]
www.morganmarkets.com/analyst/daniellamy
Priyanka Malhotra
(44-20) 7325-7043
[email protected]
www.morganmarkets.com/analyst/priyankamalhotra
Tina Zhang
(44-20) 7777-1260
[email protected]
www.morganmarkets.com/analyst/tinazhang
Research Distribution
To amend research distribution please contact Laura Hayes our Credit Research Administration, contact details above. JPMorgan research is available at http://www.morganmarkets.com
92