cutting edge.
derivatives pricing
In the balance
Funding costs and counterparty credit risk adjustments
have become increasingly important contributions
to the total value of derivatives positions. Based on a
recently developed derivatives valuation framework
that incorporates these two effects in a unified way,
Christoph Burgard and Mats Kjaer discuss the
relationship of the funding cost adjustment to the
balance sheet. They also demonstrate two ways in
which the funding cost adjustment can be eliminated,
resulting in symmetric derivatives values
Funding costs OD
and the bilateral counterparty credit
risk of derivatives positions have
become increasingly hot topics since the beginning of the credit
crisis in 2008. It has become standard practice to adjust deriva-
tives prices for the counterparty risk. Similarly, funding costs are
PR
increasingly incorporated into derivatives prices one way or
another, but the conceptual foundations for such funding adjust-
ments are much less well understood.
Piterbarg (2010) showed how the funding costs of delta hedg-
ing affect derivatives pricing. Subsequently, Burgard & Kjaer
RE
(2011) developed a unified framework that combines funding
costs and bilateral counterparty credit risk. This framework speci-
fies how a positive cash account related to the hedging strategy of
an uncollateralised derivative can be used to fund the repurchase
of the issuer’s own bonds in order to hedge its own credit risk.
R
This demonstrates that the debit value adjustment (DVA) is
equivalent to a funding benefit adjustment and justifies its inclu-
FO
sion in a bilateral counterparty value adjustment. The framework
also includes the funding costs associated with a negative cash
account, and yields a corresponding additional adjustment to the
derivatives price, the funding cost adjustment (FCA).
In both Piterbarg (2010) and Burgard & Kjaer (2011), the size
T
of the FCA depends on the specific way the funding is achieved
and gives rise to prices that are dependent on the funding posi-
NO
tion of the issuer. In this context, ‘price’ does not mean clearing
or mark-to-market price but the economic value of the derivative
to the issuer including counterparty risk and funding cost. The
counterparty would clear the derivative with the issuer that has
the best funding position.
This article shows that the funding cost term is related to a
windfall to the issuer’s bondholders upon default of the issuer.
This leads us to examine the impact of the derivative asset and
related funding positions on the balance sheet within a simple
model. We show that the overall impact on the funding position
of the issuer reduces the effective marginal funding spread for the
new positions to zero. We then discuss two strategies of how the
balance-sheet impact can directly be neutralised, mitigating the
need for an FCA to the derivatives price. If such strategies can be
put into practice, they lead back to a state where symmetric prices
72 Risk November 2011
between issuer and counterparty are achieved, independent of
funding costs.
A unified framework for counterparty risk and funding
To set the stage for the following discussions, we review and inter-
pret the framework of Burgard & Kjaer (2011). In this framework,
we consider an uncollateralised derivative on an underlying asset S
between an issuer B and a counterparty C. The ‘risk-free’ value
V(S, t) is the counterparty credit risk-free and funding cost-free
ON
value, as seen from the issuer, of the derivative at time t. The ‘risky’
^
value V, in contrast, is defined to include these effects.
The process of the underlying asset S is assumed to be lognormal,
with the zero-recovery bonds PB and PC of the issuer and counter-
party assumed to follow independent jump-to-default processes
TI
dJB and dJC for the issuer and counterparty, respectively:
dS
= µdt + σdW (1)
UC
S
dPB
= rB dt − dJ B (2)
PB
dPC
PC
= rC dt − dJ C
For simplicity, the framework assumes deterministic rates and
(3)
default intensities. As such, it does not include any convexity
effects between funding rates and market factors, as discussed in
Piterbarg (2010), but could be extended to do so. Later, we will
consider a case that includes bonds of different seniority (and
therefore recovery) in order to neutralise the balance-sheet impact
of the derivatives.
^
The boundary conditions of the risky value of the derivative V(S,
t, JB, JC) upon default of the issuer or counterparty are defined as:
V̂ (t,S,1,0 ) = M + (t,S ) + RB M − (t,S ) B defaults first
(4)
V̂ (t,S,0,1) = RC M + (t,S ) + M − (t,S ) C defaults first
Contractual details will determine which values for the close-outs
M should be used upon default of the issuer or counterparty. In
this article, we consider the case of the risk-free close-out M =
V(t, S, 0, 0), as it more closely describes contracts that follow the
International Swaps and Derivatives Association master agree-
ments of 1992 and 2002.
Mirroring the classical Merton derivation of the Black-Scholes
equation, we use a replication strategy of holding d in the underly-
ing asset S, aC in the zero-recovery bonds PC of the counterparty,
and b(t) in cash. This cash, if positive, is used to fund the (re)pur-
chase of the required number aB of the issuer’s own zero-recovery
bonds PB, to hedge the issuer’s credit risk on the derivatives posi-
^
tion. Any remaining cash bF = –V – aBPB, if positive, is used to
purchase risk-free assets yielding rate r. If negative, it is financed
from an external funding provider at a funding rate rF. The cost of
this funding is included in the cost of the hedging strategy and
^
therefore the risky value V. While the risk of the issuer’s default on
the derivatives contract is hedged by the replication strategy, the
risk of the funding provider to the issuer’s credit is not, and the
funding provider correspondingly earns the spread sF = rF – r as
compensation for this risk. The hedge ratios found in Burgard &
Kjaer (2011) are:
δ = −∂SV − ∂SU (5)
 U + (1− RB ) V −
αB = − (6) 1 CVA, DVA, FCA and total adjustment U as a function of
PB
the counterparty credit spread (issuer spread is 150bp)
U + (1− RC ) V + 10
αC = − (7)
PC 5

0
β F = −V + − RBV − (8)

Euro (’000)
0 50 100 150 200 250 300

ON
–5
^
where U = V – V is the total adjustment. –10
The replication strategy with these hedge ratios is self-financing CVA
and results in the following partial differential equation (PDE) –15 DVA
for the risky value of the derivative: –20 FCA
Total
∂tV̂ + AtV̂ − ( r + λ B + λ C ) V̂ –25

TI
Counterparty credit spread (bp)
= − ( RB λ B + λ C ) V − − ( λ B + RC λ C ) V + + sFV + (9)
V̂ (T ,S ) = H ( S )

UC

2 CVA, DVA, FCA and total adjustment U as a function of
where:
the issuer credit spread (counterparty spread is 150bp)
1
AtV ≡ σ 2 S 2∂2SV + ( qS − γS ) S∂SV 10
2 (10)
sF ≡ rF − r λ B ≡ rB − r λ C ≡ rC − r 5

and qS is the net share position financing cost and gS is the divi-
dend income.
OD Euro (’000)

0
–5
0 50 100 150 200 250 300

This then yields the integral representation:

–10
U (t,S ) = CVA + DVA + FCA (11)
PR
–15 CVA FCA
with: –20 DVA Total
T Issuer credit spread (bp)
CVA = − (1− RC ) ∫ t λ C (u ) Dr+λ B +λC (t,u ) E t ⎡⎣V + (u,S (u ))⎤⎦ du
T
DVA = − (1− RB ) ∫ t λ B (u ) Dr+λ B +λC (t,u ) E t ⎡⎣V − (u,S (u ))⎤⎦ du (12)
RE

T
FCA = − ∫ t sF (u ) Dr+λ B +λC (t,u ) Et ⎡⎣V + (u,S (u ))⎤⎦ du
The first term is often referred to as the unilateral credit value
adjustment (CVA). It is clear from this representation that while
R
–
V of the derivatives value, it could instead invest the amount
–
–V in bonds with recovery R B.
The contributions of U to the hedge ratios, on the other hand,
come from the fact that upon default the close-out amount of V
differs from the (risky) value of the derivative just prior to default

both the DVA and the FCA are related to the credit position of by the amount U, because the credit and funding adjustments for
the issuer, they do not double count the issuer’s credit but capture the trade disappear upon default of the counterparty or the issuer.
FO

exposures of the mark-to-market value of the derivative of oppo-
site sign and as such are two sides of the same coin. The DVA
term itself can be seen as a funding benefit term, as it arises from
the issuer using a positive cash account to buy back its own bonds,
earning the spread while at the same time hedging out its own
T
Hence the credit risk on the full amount of U needs to be hedged,
and this is achieved by taking positions in zero-recovery bonds of
B and C corresponding to the full value of U.
^
Without the FCA term, the risky value V would be symmetric in
B and C – the counterparty and the issuer, following the same

credit risk on the derivatives position. methodology, would agree on the same value. If, however, the fund-
Figures 1 and 2 display the value of the CVA, DVA and FCA ing costs for the replication strategy are included and the funding
corrections as well as the total correction U as a function of the
NO

counterparty and issuer credit spreads for a five-year quarterly
payer swap of €10 million notional with present value V = 0.
Equation (11) deserves some further discussions. The hedging
strategy leading to this equation involves the issuer repurchasing its
own bonds, but does not involve any dealings in its own credit
default swaps (CDSs). The term lB is the spread of the yield of a
zero-recovery bond over the risk-free rate, not the hazard rate
derived from the CDS market. If there is a basis between bonds
and CDSs for the issuer B, it is the bond market that counts for
determining lB used in the DVA and FCA terms in equations (12).
For negative V, that is, positive cash account –V, the combina-
–
tion of zero-recovery zero-coupon bonds of value –(1 – R B)V and
–
–R BV amount of cash invested in risk-free assets replicates an
–
investment of value –V in a zero-coupon bond with recovery R B.
So for the issuer to hedge its own credit risk on the risk-free part
spread is non-zero, then the two parties both hedging their risks and
pricing in their funding costs would not agree on the price. A coun-
terparty that wants to buy this product would buy it from the issuer
with the smallest FCA term, that is, the lowest funding costs.
In the following sections, we discuss the origins of the FCA
term in more detail and show different ways the funding costs
can be mitigated. Doing so successfully can reduce the FCA term
to zero and produce prices that are independent of the funding
position of the issuer and are therefore symmetric.
Origin of the FCA term for unsecured funding of derivatives position
If the funding of the negative cash account is done unsecured,
then sF is the issuer’s unsecured funding spread. Assuming the
recovery upon issuer’s default on the unsecured funding is RB, that
is, the same as the recovery on the derivative’s close-out amount,
then the unsecured funding spread sF is related to the spread lB on

risk.net/risk-magazine 73
cutting edge. derivatives pricing
the zero-recovery bond via the relation sF = (1 – RB)lB and the
FCA term becomes:
T = r ( L0 + d ) dt + ( L1 − A1 ) λdt
FCA = − (1− RB ) ∫ t λ B Dr+λ B +λC (t,u ) E t ⎡⎣V + (u,S (u ))⎤⎦ du (13)
To understand the origin of this term, it is instructive to see what
^
happens when the issuer defaults while V is positive, that is, in-the-
money for the issuer. In this case, just prior to default, the cash
account bF is negative and its corresponding amount –V + is pro-
vided unsecured by the funding provider. If the issuer then defaults,
it will settle the derivative at the close-out amount with the coun-
terparty, that is, receive V +. The external, unsecured funding pro-
vider, on the other hand, will only receive RBV + on his funding, so
losing (1 – RB)V +. Prior to default, the funding provider is compen-
sated for this potential loss by receiving the spread sF. The position
of the issuer upon its own default, that is, receiving V + for the deriv-
ative and paying RBV + to the funding provider, is a positive wind-
UC
fall of (1 – RB)V + that goes towards the recovery of the bondholders
of the issuer. Inspection of the FCA term in equation (13) shows
that it is the expectation value of this windfall – it is the integral
over lB, representing the probability density of the issuer to default
implied by the bond market, times Dl +l , the probability that nei-
B C
times the expected loss –(1 – RB)Et[V +] to the funding provider
upon default of the issuer, all discounted to t.
So the following picture arises: the FCA term of equation (13) is
OD
ther the issuer nor the counterparty have defaulted up to that point,
the premium for the windfall to the bondholders of the issuer in the
PR
case where the issuer defaults and the derivative has positive value for
the issuer. This premium is a cost that arises while the issuer has not
^
defaulted. This cost is included in the derivative value V charged to
the counterparty by means of the FCA in equation (13).
RE
Impact on balance sheet and funding
As discussed above, derivatives and their funding positions con-
tribute to the asset and liability pools upon the issuer’s default.
Therefore they should affect the funding position of the issuer. In
R
this section, we quantify this feedback effect using a simple bal-
ance-sheet and funding model.
FO
Assume that, as in a reduced-form credit model, default of the
issuer is driven by an instantaneous default process with default
intensity l. Prior to entering into the derivative contract, let A 0 be
the expected assets upon default of the issuer and L 0 be the liabil-
ities, so that the expected recovery upon default is R0 = A 0/L 0.
Within this simple set-up, the funding spread sF of the issuer over
T
the risk-free rate that compensates for the expected loss upon its
default is sF = (1 – R0)l. Hence the instantaneous funding costs f0
NO
of the issuer over time dt for his total liability L 0 is:
f0 dt = ( r + (1− R0 ) λ ) L0 dt
Let the issuer now add a derivative with positive value d as an
asset, resulting in total assets of A1 = A 0 + d, and fund the corre-
sponding cash outflow by adding a corresponding liability, giving
new total liabilities of L1 = L 0 + d. Thus, the expected recovery
changes to:
A A +d
R1 = 1 = 0 (15)
L1 L0 + d
Adding the derivatives asset, assuming the issuer has hedged the
market and counterparty risk, does not change the default inten-
sity of the issuer. Thus, the instantaneous funding costs after add-
ing the derivatives are:
74 Risk November 2011
f1dt = ( r + (1− R1 ) λ ) L1dt
= rL0 dt + r ⋅ d ⋅ dt + ( L0 − A0 ) λdt (16)
= r ⋅ d ⋅ dt + r ⋅ L0 ⋅ dt + (1− R0 ) λL0 dt
= r ⋅ d ⋅ dt + f0 dt
ON
Hence the effective cost of funding for the additional liability d is
r ⋅ d ⋅ dt. While the new liability d draws the new funding spread (1
– R1)l, the change to the recovery and its effect on the funding of the
total liabilities results in an effective funding rate for d that is the risk-
free rate. Thus, within this balance-sheet model the spread sF is zero.
TI
While this balance-sheet model is somewhat simplistic, it shows
that proper accounting for the effects of the derivative assets on the
balance sheet can mitigate the funding costs and bring the FCA
term down to zero. With a vanishing FCA term, equation (11)
^
yields an adjustment U and risky value V that are symmetric
between issuer B and counterparty C. Practically, however, the
benefit of the balance-sheet impact is difficult to pin down at the
moment of trading and hedging the derivative contract, and there-
fore difficult to allocate as a benefit to the derivatives trading desk.
In the following, we will discuss two mechanisms that shield
the balance sheet from the impact of the derivative asset and
funding liability, and lower the funding costs directly.
Case where the derivative can be used as collateral
As mentioned in Burgard & Kjaer (2011), the derivative itself can be
used as collateral for the funding required for negative balances on
the cash account. The cash account is negative when the derivative
has positive value to the issuer. It could be ring-fenced from the
other assets and used as security for the required funding. To be
precise, the negative cash balance bF that needs funding is –V +, that
is, minus the risk-free value of the derivative, if V is positive. While
^
the value of the derivative to the issuer is the risky value V, its close-
out value is the risk-free value V, so it could indeed be used to secure
the required amount of funding of V if no haircut were applied.
In practice, there are a number of technical difficulties to be taken
into account when considering whether and how the derivative could
be used as collateral. One of them is that the value of the derivative
(and the amount of cash requiring funding) changes constantly.
Another is that, in general, one should expect a healthy haircut to be
applied between the value of the derivative collateral and the secured
funding amount. Both of these practical difficulties can be mitigated
to some extent by pooling derivatives assets together and obtaining
funding against this pool (or part of the pool). In the ideal case,
where the derivative can be used as collateral with zero haircut and
the secured funding rate is the risk-free rate, the spread sF and corre-
(14)
spondingly the FCA of equation (12) disappear and we are left with
the first two terms for the adjustment U in equation (11), represent-
ing the usual bilateral CVA as described in many papers (for exam-
ple, see Gregory, 2009), but its use being well justified in our frame-
work through the hedging strategy of repurchasing the issuer’s own
bonds from its positive cash account. At the same time, it does not
appear as an asset on the balance sheet of the issuer since the deriva-
tive has been pledged as collateral when the cash account is negative.
Balance-sheet management to mitigate funding costs
Another way to shield the balance sheet from the impact of the
derivative asset and funding liabilities is to actively manage the
balance sheet in such a way that the windfall from the derivative
asset and the funding position upon default of the issuer is bal-
anced out by a corresponding liability.
A. Value of the issuer’s P1 and P2 bond positions post
This can be achieved if the issuer is able to freely trade two of its own default
own bonds, P1 and P2 , with different recovery rates R1 and R2 ,
− R1R2V̂ −RR1V−R−R1RBV
+ −
P1 position value
that is, different seniority. 2 1

We thus change the earlier set-up such that we have the hedging

ON
instruments P1, P2, PC and S. All positive and negative cash in the − −R2 R1V̂+R2V +R2 RBV
+ −
P2 position value
R −R
2 1
cash account is invested and raised, respectively, by buying back and
issuing P1 and/or P2 bonds. The assets follow the dynamics:
dP1
Total position value (
− V + + RBV − )
= r1 (t ) dt − (1− R1 ) dJ B
P1

TI
dP2 ∂tV̂ + AtV̂ − ( r + λ B + λ C ) V̂
= r2 (t ) dt − (1− R2 ) dJ B
P2 (27)
(17) = − ( RB λ B + λ C )V − − ( λ B + RC λ C ) V +
dPC

UC
= rC (t ) dt − dJ C Comparing this with equation (9) implies financing of the nega-
PC
tive cash account at vanishing spread sF = 0. In particular, the
dS
= µ (t ) dt + σ (t ) dW strategy involves issuing the senior P2 bonds and using some of
S the proceeds to repurchase the junior and hence higher-yielding
where R1 ∈ [0, 1), R2 ∈ [0, 1) and R1 < R2 . Neither of the recovery P1 bonds. The excess return generated by this strategy exactly off-
rates R1 and R2 need to equal the derivative recovery rate R B.

As before, we set up a replicating hedge portfolio Π given by:
Π = α1P1 + α 2 P2 + αC PC + δS + βS + βC

OD
(18)
sets the additional funding costs, so the net-financing rate
becomes r. At the same time, the combined position of P1 and P2
bonds ensures that there is no windfall to the bondholders in the
case of default of the issuer while V is positive (and the cash
where bS = –dS is the funding account for the share position and account negative). This is shown in table A, which summarises
PR
bC = –aCPC is the funding position for the PC bonds. The fact that the total bond position at the issuer’s own default and which off-
Π is a self-financing replicating hedge portfolio implies that: sets perfectly the value of the derivative as defined in equation (4).
Π = α1P1 + α 2 P2 = −V̂ So if the issuer is able to offset the impact of the derivative and
(19) its funding on the balance sheet with a combination of going long
dΠ = −dV̂
RE

senior bonds and short junior bonds, then the windfall is effec-
so repeating the delta hedging arguments of Burgard & Kjaer tively monetised while the issuer has not defaulted and by doing
(2011) and defining s1 = r1 – r and s2 = r2 – r yields: so the funding cost term is reduced to zero.

δ = −∂SV̂ (20)
Conclusion
R

The inclusion of FCAs in derivatives prices leads to issuer-depend-

ΔV̂B = α1 (1− R1 ) P1 + α 2 (1− R2 ) p2 (21)
ent economic valuations. We have shown that the FCA is related
FO

to the impact of the derivative and corresponding funding posi-

ΔV̂C = αC PC (22)
tions on the balance sheet upon the issuer’s default and have dem-
α1s1P1 + α 2 s2 P2 + αC λ C PC onstrated ways this impact can be mitigated. Being able to follow
(23) any of these strategies will allow an issuer to value derivatives
1
= −∂tV̂ − σ 2 S 2∂2SV̂ + rV̂ − δ ( qS − γ S ) S∂SV̂ without an FCA and leads to prices that are symmetric between
2 issuer and counterparty. n
T

^ ^ ^
where DVB = V(t, S, 1, 0) – V(t, S, 0, 0) and DVC = V(t, S, 0, 1) –
^ ^ ^
V(t, S, 0, 0), with V(t, S, 1, 0) and V(t, S, 0, 1) given in equation (4) Christoph Burgard is global head of equity derivatives, securitisation derivatives
NO

with M = V. and counterparty-credit modelling and Mats Kjaer is a quantitative analyst for
From these equations, we can determine a1 and a2 to be: counterparty-credit modelling team in the quantitative analytics division at
R V̂ − V + − RBV − Barclays Capital in London. They would like to thank Tom Hulme and Vladimir
α1 = − 2 (24) Piterbarg for useful comments and suggestions. This article represents the views

( R2 − R1 ) P1 of the authors alone, and not the views of Barclays Capital or Barclays Bank.
−R V̂ + V + + RBV − Email: christoph.burgard@barclayscapital.com, mats.kjaer@barclayscapital.com
α2 = − 1 (25)

( R2 − R1 ) P2
References
which implies the following pricing PDE:
Burgard C and M Kjaer, 2011 Gregory J, 2009
R V̂ − V + − RBV − Partial differential equation Being two-faced over counterparty
∂tV̂ + AtV̂ − rV̂ = s1 2 representations of derivatives with credit risk
R2 − R1 (26) counterparty risk and funding costs Risk February, pages 86–90
Journal of Credit Risk 7(3), pages 1–19
−R V̂ + V + + RBV −
+s2 1
R − R
(
− λ C V − + RCV + ) Piterbarg V, 2010
Funding beyond discounting: collateral
2 1 agreements and derivatives pricing
If we furthermore assume zero basis between the bonds, that is, Risk February, pages 97–102
s1 = (1 – R1)lB and s2 = (1 – R2)lB, then (26) simplifies to:

risk.net/risk-magazine 75