Professional Documents
Culture Documents
Equity Variance Swaps With Dividends OpenGamma
Equity Variance Swaps With Dividends OpenGamma
Richard White
i=1
_
log
_
S
i
S
i1
__
2
K
2
_
(1)
where S
i
is the closing price of a stock or index on the i
th
observation date, the annualisation
factor A is set by the contract (and usually 252), N
var
is the variance notional and K
2
(the
variance strike) is usually chosen to make the initial value zero. The notional is often quoted in
units of volatility (i.e. $1M per vol), N
vol
with N
var
=
N
vol
2K
.
Since all the other terms are xed by the contract, we will focus on the realised variance part
RV (T) =
n
i=1
_
log
_
S
i
S
i1
__
2
(2)
1.1 The Black-Scholes World
If the stock follows the process
dS
t
S
t
= rdt +dW (3)
then the returns are given by
R
i
log
_
S
i
S
i1
_
= (r
2
/2)t +W (4)
Version 2.1
1
The name swap is a bit of a misnomer, as it is a forward contract on the realised variance
1
That is, they are normally distributed with mean (r
2
/2)t and variance
2
t. Ignoring the
mean (which will be negligible next standard deviation for a period of one day), we can write
the realised variance as
RV (T) =
2
t
n
i=1
Z
2
i
(5)
where Z
i
are standard normal random variables. Hence the realised variance has a chi-squared
distribution with n degrees of freedom scaled by
2
t - which is a Gamma distribution (n/2, 2
2
t).
The expected variance is simply
E[RV (T)] =
2
T (6)
but even in the Black-Scholes world, the realised variance can vary greatly from this.
2
1.2 Other diusions
By considering the Taylor series of log
_
S
i
S
i1
_
and
_
log
_
S
i
S
i1
__
2
about S
i1
, it is easy to show
that
_
log
_
S
i
S
i1
__
2
2 log
_
S
i
S
i1
_
+ 2
S
i
S
i1
S
i1
+O
_
(S
i
S
i1
)
3
S
3
i1
_
(7)
Summing the terms gives Neubergers formula, [Neu92].
RV (T) 2 log
_
S
T
S
0
_
+ 2
n
i=1
1
S
i1
(S
i
S
i1
) (8)
What this says is that you can replicate the realised variance up to the expiry, T, by holding
2 log(S
0
) zero coupon bonds with expiry at T, shorting two log-contracts
3
and daily delta hedging
with a portfolio whose delta is 2/S
i1
.
The expected value of the realised variance (or just expected variance) is
EV (T) = 2 log(F
T
) 2E[log(S
T
)] (9)
and therefore the fair value of the variance strike is
K =
_
2A
n
(log(F
T
) E[log(S
T
)]) (10)
where F
T
is the forward price. This is model independent provided the stock process is a diusion,
[Gat06].
1.3 Static Replication
Any twice dierentiable function, H(x) can be written exactly as [CM98].
H(x) = H(x
0
) + (x x
0
)H
(x
0
) +
_
x
0
0
H
(z)(z x)
+
dz +
_
x
0
H
(z)(x z)
+
dz (11)
2
The fractional error is approximately
2/n
3
A log-contact has the payo log(S
T
) at expiry - these are not liquid but can be statically replicated with
strips of Europeans puts and calls
2
for x 0. The (non-discounted) value of a derivative at time t with a payo H(S
T
) at T is
E[H(S
T
)|F
t
]. Expanding H(S
T
) using equation 11 we nd
E[H(S
T
)|F
t
] = H(s
) + (E[S
T
] s
)H
(s
)
+
_
s
0
H
(k)E[(k S
T
)
+
]dk +
_
s
(k)E[(S
T
k)
+
]dk
(12)
Choosing s
= F
T
and recognising that E[(S
T
k)
+
] and E[(k S
T
)
+
] are (non-discounted) call
and put prices at strike k, we can write
E[H(S
T
)|F
t
] = H(F
T
) +
_
F
T
0
H
(k)P(k)dk +
_
F
T
H
(k)C(k)dk (13)
So in principle, one can replicate any payo that is a function of the underlying at expiry, by an
innite strip of puts and calls on the underlying, with weights H
(k) = 1/k
2
, so the expected variance can be written as
EV (T) = 2
_
_
F
T
0
P(k)
k
2
dk +
_
F
T
C(k)
k
2
dk
_
(14)
In practise only a nite number of liquid call and put prices are available in the market
and the expiries may not match that of the variance swap. What is needed is an arbitrage free
interpolation of option prices. For a call with expiry T and strike k we have
A call with a zero strike equals the forward C(T, 0) = F
T
Positive calendar spreads,
C(T,k)
T
0
Call price is monotonically deceasing in strike,
C(T,k)
k
0
Positive butteries,
2
C(T,k)
k
2
0
High strike limit, c(T, ) = 0
If the underlying cannot reach zero (e.g. Exponential Brownian Motion (EBM)) then,
C(T,0)
k
= 1, otherwise (e.g. SABR)
C(T,0)
k
> 1
The corresponding conditions for puts follow from put-call parity. We carry out the interpo-
lation in implied volatility space, using arbitrage free smile models for interpolation in the strike
direction - this is detailed in [Whi12]. To extrapolate to strikes outside the range of market
quotes, we use a shifted log-normal model. This is just the Black option pricing formula where
the forward and the volatility are treated as free parameters which are calibrated so that either
the last two option prices are matched, or the price and dual delta
4
of the last option is matched.
This is done separately for low strikes (puts) and high strikes (calls).
The last practical detail is to nd an upper cut-o for the integral. Using the monotonically
decreasing property, we know that for a cut-o of A we have
_
A
C(k)
k
2
dk
_
A
C(A)
k
2
dk =
C(A)
A
(15)
So having integrated up to A we have an upper limit on the error.
4
dual delta is the rate of change of a option price with respect to the strike
3
Using LHopitals rule, we nd
lim
k0
P(k)
k
2
=
1
2
2
P(T, 0)
k
2
(16)
Which is half the probability density at zero, and is zero for our shifted log-normal extrapolation.
This limit on the integrand should be used in any numerical integration routine. Alternatively
we have for very small A
_
A
0
P(k)
k
2
dk
P(A)
2A
+ (17)
where is the probability mass at zero (zero for exponential Brownian motion but non-zero
for SABR etc). This can again be used to set a limit on the error.
2 Dividends
In this section we follow the approach of Buehler [OBB
+
06, Bue10]
5
.
Companies make regular dividend payments (e.g. every six months) of a xed amount per
share, with the amount announced well in advance
6
, so dividends over a 1-2 year horizon can be
treated as xed cash payments (per share) regardless of the share price. Further in the future, the
dividend payment is likely to depend on company performance, hence we can model payments
as being a xed proportion of the share price on the ex-dividend date. In general dividends are
paid
7
at times
1
,
2
, ... from today, with a cash amount
i
and a proportional about
i
. If the
stock price immediately before the dividend payment is S
i
= S
(1
i
)
i
(18)
If the stock price follows a diusion process, then this implies that it can go negative. One way to
avoid this is by making the dividend payment a (non-ane) function of the stock price, D
i
(S
)
such that D
i
(S
) S
1
+
1
) on the rst
dividend date after t. Rewriting this in terms of the post-dividend price S
1
we have
Dividend Payment
1
=
t
_
1
1
1
S
1
+
1
1
1
_
(19)
We use the part that is proportional to S
1
to buy new stock immediately after the payment,
then the stock holding will be
1
=
t
1
1
(20)
Our residual cash may be written as
1
.
If we repeat the procedure for each dividend payment, then at T we will own
T
=
t
/
j:t<
j
T
(1
j
)
shares. To have exactly one share at maturity, set
t
=
j:t<
j
T
(1
j
) (21)
5
It should be noted that both the book and the paper contain considerable typos in the mathematics
6
Although there may be no legal obligation to actually pay this amount if the companys performance is poor.
7
We assume the ex-dividend date and the payment date coincide.
8
They also consider credit risk and repo rates, which we ignore are present.
4
Then at some time s with t < s T we own
s
=
j:s<
j
T
(1
j
) (22)
shares. At each dividend date we also have residual cash equal to
i
. As these are xed, we
can realise their value now by selling
i
zero coupon bonds expiring at
i
. The total value of
the sale is
j:t<
j
T
j
P(t,
j
) (23)
If we dene the growth factor as
R(t, T) =
1
P(t, T)
j:t<
j
T
(1
j
) (24)
then the total initial investment that guarantees one share at time T is
P(t, T)R(t, T)
_
_
S
t
j:t<
j
T
j
R(t,
j
)
_
_
(25)
We can fund this by selling k zero coupon bonds with maturity T. The value of k that makes
the initial investment zero is the forward price, therefore
F(t, T) = R(t, T)
_
_
S
t
j:t<
j
T
j
R(t,
j
)
_
_
(26)
Since the share price cannot become negative, neither can the forward. This implies that
S
t
j:t<
j
T
j
R(t,
j
)
, and since this must be independent of the arbitrary maturity, T, of
a forward, we must have
S
t
D
t
j:
j
>t
j
R(t,
j
)
(27)
That is, the dividend assumption imposes a oor on the stock price equal to the growth factor
discounted value of all future cash dividends.
9
As a result of the oor on the stock price, the Black model, S
t
= F
t
X
t
,
10
where X
t
is a EBM
with E[X
t
] = 1, is no longer valid since P(S
t
< D
t
) > 0.
11
[OBB
+
06, Bue10] propose
S
t
= (F
t
D
t
)X
t
+D
t
(28)
with E[X
t
] = 1 and X
t
0.
12
The stock price experiences randomness just on the portion above
the oor D
t
. X
t
is known as the pure stock process.
9
We have overloaded the letter D - D
t
is the growth factor discounted value of all future cash dividends, while
D() is the dividend paid at time
10
F
t
F(0, t)
11
Any other process for X
t
is also invalid for the same reason.
12
Their model also includes credit risk, so they impose X
t
> 0 and have the stock price and all future dividends
fall to zero on a default. The setup with no credit risk and X
t
0 implied that the stock becomes a bond with
xed coupons if X
t
becomes zero.
5
2.1 European Options
Consider the price at time t of two european call options, one with expiry
, k
) & C
t
(, k). The rst payo is
(S
)
+
=
_
S
+
1
k
_
+
=
1
1
_
S
[(1 )k
]
_
+
(29)
If we let k = (1 )k
then the payo of the rst option is just 1/(1 ) times that of the
second. It then follows that
(1 )C
t
(
, k
) = C
t
(, (1 )k
) (30)
If the options are priced using Blacks formula with the same volatility, then the condition of
equation 30 does not hold. Conversely, if the condition does hold, and (Black) implied volatilities
are found, they will jump up across the dividend date. Once again, the Black model, or indeed
any model that has a continuous implied volatility surface, is not consistent with the dividend
assumption.
Another consideration is what happens to the price of an option with expiry, T, as the current
time, t, moves over a dividend date. We now write the call price as C(t, S
t
, T, k). Clearly
C(
, S
, T, k) = C(, S
, T, k) (31)
This is true for any European option (e.g. a log-payo) with T > . This condition must be
applied when solving a backwards PDE for option prices, as discussed later.
2.2 Options on the Pure Stock
Dene a call option on the pure stock as
C(T, x) = E[(X
T
x)
+
] (32)
This is related to a real call by
C(T, x) =
1
P(0, T)(F
T
D
T
)
C (T, (F
T
D
T
)x +D
T
)
C(T, k) = P(0, T)(F
T
D
T
)
C
_
T,
k D
T
F
T
D
T
_ (33)
Hence if the dividend structure is known, market prices for calls (and puts) can be converted
into pure call prices - which is turn can be converted to a pure implied volatility by inverting
the Black formula in the usual way.
Figure 1 shows the implied volatility surface if the underlying prices are generated from a at
pure implied volatility surface at 40%, and the pure implied volatility surface if the prices are
generated from a at implied volatility surface at 40%. These shapes are observed in [Bue10], but
without the jumps at the dividend dates. The condition leading to equation 30 means we expect
a jump in implied volatility if the pure implied volatility is at and vice versa. As Buehler does
not state the dividends schedule, we cannot make a direct comparison. However if we increase the
dividend frequency ten-fold and decrease each amount ten-fold (which may be more the case for
indices)
13
, then the implied volatility surface (for a at pure implied volatility surface) becomes
as shown in gure 2.
13
We have cash only dividends in year one, proportional only after 3 years and a linear switch over in year 2
6
0.01
0.2289
0.4478
0.6667
0.8856
1.1045
1.3234
1.5423
1.7612
1.9801
0.32
0.33
0.34
0.35
0.36
0.37
0.38
0.39
5
0
5
4
.5
5
9
6
3
.5
6
8
7
2
.5
7
7
8
1
.5
8
6
9
0
.5
9
5
9
9
.5
1
0
4
1
0
8
.5
1
1
3
1
1
7
.5
1
2
2
1
2
6
.5
1
3
1
1
3
5
.5
1
4
0
1
4
4
.5
1
4
9
1
5
3
.5
1
5
8
1
6
2
.5
1
6
7
1
7
1
.5
1
7
6
1
8
0
.5
1
8
5
1
8
9
.5
1
9
4
1
9
8
.5
Time to Expiry
I
m
p
l
i
e
d
V
o
l
a
t
i
l
i
t
y
Strike
0.01
0.2289
0.4478
0.6667
0.8856
1.1045
1.3234
1.5423
1.7612
1.9801
0.39
0.4
0.41
0.42
0.43
0.44
0.45
0.46
0
.5
0
.5
4
5
0
.5
9
0
.6
3
5
0
.6
8
0
.7
2
5
0
.7
7
0
.8
1
5
0
.8
6
0
.9
0
5
0
.9
5
0
.9
9
5
1
.0
4
1
.0
8
5
1
.1
3
1
.1
7
5
1
.2
2
1
.2
6
5
1
.3
1
1
.3
5
5
1
.4
1
.4
4
5
1
.4
9
1
.5
3
5
1
.5
8
1
.6
2
5
1
.6
7
1
.7
1
5
1
.7
6
1
.8
0
5
1
.8
5
1
.8
9
5
1
.9
4
1
.9
8
5
Time to Expiry
P
u
r
e
I
m
p
l
i
e
d
V
o
l
a
t
i
l
i
t
y
Strike
Figure 1: Implied Volatility (left) and Pure Implied volatility (right) for an initial stock price is
100, dividends every six months and the next is in one month, with a proportional part equal
to 1% of the stock price ( = 0.01) and a cash part of 1.0, and the risk free rate is 5%. In each
case the prices were generated from a at pure implied volatility surface (left), or a at implied
volatility surface (right) - both at 40%
0.01
0.2289
0.4478
0.6667
0.8856
1.1045
1.3234
1.5423
1.7612
1.9801
0.35
0.36
0.37
0.38
0.39
0.4
0.41
3
0
3
5
.
1
4
0
.
2
4
5
.
3
5
0
.
4
5
5
.
5
6
0
.
6
6
5
.
7
7
0
.
8
7
5
.
9
8
1
8
6
.
1
9
1
.
2
9
6
.
3
1
0
1
.
4
1
0
6
.
5
1
1
1
.
6
1
1
6
.
7
1
2
1
.
8
1
2
6
.
9
1
3
2
1
3
7
.
1
1
4
2
.
2
1
4
7
.
3
1
5
2
.
4
1
5
7
.
5
1
6
2
.
6
1
6
7
.
7
1
7
2
.
8
1
7
7
.
9
1
8
3
1
8
8
.
1
1
9
3
.
2
1
9
8
.
3
Time to Expiry
I
m
p
l
i
e
d
V
o
l
a
t
i
l
i
t
y
Strike
Figure 2: Implied Volatility surface for a pure implied volatility surface of 40%, and 20 dividend
payments per year
7
2.3 Dupire Local Volatility for Pure Stock Process
Following exactly the same argument as [Dup94], if a unique solution to the SDE
dX
t
X
t
=
X
(t, X
t
)dW
t
(34)
exists, then the pure stock local volatility is given by:
X
(t, X
t
)
2
= 2
C
T
x
2
2
C
x
2
(35)
Once we have the pure option prices (or equivalently the pure implied volatilities) from market
prices, we can proceed as in [Whi12] to construct a smooth Pure Local Volatility surface, and
numerically solve PDEs without having to impose jump conditions at the dividend dates. Con-
versely, if we solve a PDE expressed in the stock price, then at dividend dates the nodes must be
shifted such that for the i
th
stock price node s
i
= s
i
D(S
), where D(S
) is the dividend
amount, and the option price maps according to equation 31. Between dividend dates the local
volatility is
(t, S
t
) = 1
S
t
>D
t
S
t
D
t
S
t
X
_
t,
S
t
D
t
F
t
D
t
_
(36)
3 Expected Variance in the Presence of Dividends
3.1 Proportional Dividends Only
The stock price process can be written as
dS
t
S
t
= r
t
dt +
t
(S
t
)dW
t
j
(
j
) (37)
where the volatility is a function of the stock price immediately prior to any jump (and possibly
another stochastic factor), and (
j
) indicates that jumps only occur at one of the dividend dates
j
.
If we dene y
t
= log(S
t
) then
dy
t
= (r
t
2
t
/2)dt +
t
dW
t
+
j
log(1
j
)(
j
) (38)
and
y
s
= y
t
+
_
s
t
(r
t
2
t
/2)dt +
_
s
t
t
dW
t
+
j:t<
j
s
log(1
j
) (39)
Which is a formal solution since in general
t
is a function of y
t
. The expected squared return is
E
_
ln
2
_
S
i
S
i1
__
_
t
i
t
i1
2
t
dt +
j
log
2
(1
j
)
j
=t
i
(40)
which is the quadratic variance of the continuous part of the log-process, plus the contribution
(if any) from the dividend. The total expected variance is then
EV (T)
_
T
0
2
t
dt +
j:
j
T
log
2
(1
j
) (41)
8
Of course, if dividends are corrected for (as is usual for single stock), then the second term is
absent.
Again from equation 39 the expected value of a payo of log(S
T
) is
E[log(S
T
)] E[y
T
] = log(S
0
) +
_
T
0
r
t
dt +
j:
j
T
log(1
j
)
1
2
_
T
0
2
t
dt
= log(F
T
)
1
2
_
T
0
2
t
dt
(42)
where the replacement of the rst three terms by the forward comes from equation 26. It then
follows immediately that
EV (T) = 2E
_
log
S
T
F
T
_
+
j:
j
T
log
2
(1
j
) (43)
where the second term is dropped when an adjustment is made for the dividends - so in this
case, the expected variance is the same as the no dividend case,
14
.
3.2 With Cash Dividends
Following the argument [Bue06], the realised variance can be expressed as
RV (T) =
n
i=1
log
2
_
S
i
S
i1
_
+
j:t<
j
T
log
2
_
S
j
(1
j
)
S
j
+
j
_
(44)
S
i
is the stock price just before the dividend payment (if one occurs on that observation
date), hence the rst term is the realised variance of the continuous part, RV (T)
cont
. If dividends
are corrected for, then the second term disappears, and the realised variance is the same as the
realised variance of the continuous part, as above.
The SDE for the log of the stock price is given by
d(log S
t
) =
1
S
t
dS
t
1
2
dRV
cont
+
j
log
_
S
j
(1
j
)
S
j
+
j
_
j
(dt) (45)
The last term simply means that at dividend dates the log of the stock jumps by log
_
S
j
(1
j
)
S
j
+
j
_
,
but is otherwise continuous. Integrating equation 39 and combining with equation 44 gives
RV (T) =2 ln
S
T
S
0
+ 2
_
T
0
1
S
t
dS
t
+ 2
j:
j
T
log
_
S
j
(1
j
)
S
j
+
j
_
+
j:t<
j
T
log
2
_
S
j
(1
j
)
S
j
+
j
_
=2 ln
S
T
S
0
+ 2
_
T
0
r
t
dt + 2
_
T
0
t
dW
t
+ 2
j:
j
T
log
_
S
j
(1
j
)
S
j
+
j
_
+
j:t<
j
T
log
2
_
S
j
(1
j
)
S
j
+
j
_
(46)
14
The big caveat here is that option prices will not be the same with and without dividends, so the value will
be dierent.
9
Taking expectations we arrive at the result
EV (T) = 2E
_
log
S
T
S
0
_
2 log(P(0, T)) + 2
j:
j
T
E
_
G
j
(S
j
)
where G
j
(s) log
_
s(1
i
)
s +
i
_
+
1
2
log
2
_
s(1
i
)
s +
i
_
(47)
Unlike the pure proportion case, we cannot rewrite this in terms of the forward as
log(F
T
) = log(S
0
) log(P(0, T)) +
j:
j
T
E
_
log
_
S
j
(1
i
)
S
j
+
i
__
(48)
The correction term is a (twice dierentiable) function of the stock price just after a dividend
payment, so can be priced with a strip of call and puts with expiry
j
as in equation 13. When
dividends are corrected for (the usual case for single stock), then last term is removed. An
alternative derivation of this result is given in appendix A.
We now have a mechanical procedure to value variance swaps when dividends are paid on
the underlying stock and the stock price is a pure diusion between dividend dates.
Estimate the future dividend schedule. This can come for declared dividends (for short
time periods), equity dividend futures and calendar spreads.
Collect prices of all European options on the underlying stock or index
15
, and convert to
prices of pure put and call options
Construct the pure local volatility surface
solve the forward PDE up to the expiry of the variance swap
Convert the pure option prices on the PDE grid, to pure implied volatility to form a
continuous (interpolated) surface
Convert back to option prices at arbitrary expiry and strike
Compute the value of the log-contact and the correction at each dividend date using equa-
tion 13.
3.3 The Buehler Paper
In Buehler (2010), the realised variance which is not adjusted for dividends is given as
RV (T) = 2 log
S
T
S
0
+ 2
n
i=1
_
S
t
i
S
t
i1
1
_
+ 2
j:
j
T
E
j
(S
j
)
where E
j
(s) = log
s(1
j
)
s +
j
+
s
j
+
j
s +
j
_
s
j
+
j
s +
j
_
2
(49)
15
Most equity options are American style, but we will ignore this important detail for now.
10
Taking expectations, we nd the expected variance is
EV (T) = 2E
_
log
S
T
S
0
_
2 log(P(0, T)) + 2
n
i=1
E
_
E
i
(S
i
)
_
where
E
j
(s) = log
s(1
j
)
s +
j
_
s
j
+
j
s +
j
_
2
(50)
This is not the same as our equation 47, since the last term (which only applies if dividends are
not corrected for) is
_
s
j
+
j
s+
j
_
2
rather than
1
2
log
2
_
s(1
i
)
s+
i
_
. We believe this is a mistake in
the paper [Bue10], and have conrmed this using Monte Carlo methods described in the next
section.
4 Numerical Testing
We present four test cases with exaggerated parameters to highlight any errors. All cases have a
spot of 100, a pure local volatility of 30%, a xed interest rate of 10%, a single dividend payment
at 0.85 years and expiry of the variance swap at 1.5 years. The rst two cases have = 0 and
= 0.4, the former with and the latter without dividends corrected for in the realised variance.
The second two cases have = 30 and = 0.2, again with and without dividends correction.
Since the pure local volatility is at, we can nd call and put prices from the modied Black
formula of equation 33, which allows us to nd the expected variance using equations 47 with
the static replication of equation 13 (once at expiry and once at the dividend date). In the case
of pure proportion dividends (rst two cases), we have a simple analytical answer via equation
43
16
. All the numerical methods do not depend on a at pure local volatility surface.
Another method is to solve the backwards PDE with the log-payo as the initial condition.
This can be done using the pure stock or the actual stock as the spacial variable - in the latter
case the jump condition must be applied at the dividend. Since this method does not calculate
the terms G(S), it is only applicable to the rst test case ( = 0 with dividends corrected for in
the realised variance).
17
Solving the forward PDE for the pure call prices up to expiry, gives actual call and put prices
on a expiry-strike grid, which can be used as in the static replication case to price all four cases.
This in our principle numerical method.
Finally, it is simple to set up Monte Carlo to calculate expected variance in all four cases.
18
The results are shown in table 1 below, expressed in terms of K where K
2
is the variance
strike (i.e. K =
_
EV (T)/T). The agreement to 4 dp of the forward PDE to the Monte Carlo,
gives use condence in the method.
5 Conclusion
We have reviewed the pricing of variance swaps, and the ane dividend model of Buehler et al.
We believe that there is a mistake in the paper [Bue10], and that the expression presented here
is correct.
16
For a at pure local volatility,
x
we have E
log
S
T
F
T
=
1
2
(
x
)
2
T
17
The correction terms can also be priced using the backwards PDE by specifying their payo as an initial
condition. This was not implemented at the time of writing.
18
Since the purpose of the Monte Carlo is to test the other methods, we are free to use a very large number of
paths.
11
Case1 Case2 Case3 Case4
Analytic 0.3 0.5138 N/A N/A
Static Replication 0.3000 0.5138 0.2463 0.6035
Backwards PDE (pure stock) 0.3000 N/A N/A N/A
Backwards PDE (actual stock) 0.3000 N/A N/A N/A
Forward PDE 0.3000 0.5138 0.2463 0.6035
Monte Carlo 0.3000 0.5136 0.2461 0.6035
Table 1: The expected variance expressed as K =
_
EV (T)/T for the 4 test cases and 5 numerical
methods discribled above.
A more general dividend model would require more numerical work to calibrate and price
variance swaps.
A Alternative Derivation
Assuming that S
i
is the closing price on date t
i
and that dividends are paid just before this, so
that
S
i
= S
i
D
i
(51)
where D
i
is the amount of the dividend (which is zero except for dividend dates). The realised
variance can be written as
RV (T) =
n
i=1
log
2
_
S
i
S
i1
_
=
n
i=1
log
2
_
S
i
D
i
S
i
S
i
S
i1
_
=
n
i=1
log
2
_
S
i
S
i1
_
+
j:t<
j
T
log
2
_
S
j
S
j
+D
j
_
+ 2
j:t<
j
T
log
_
S
j
S
j
+D
j
_
log
_
S
j
S
j
t
_
(52)
The last term (which is the return of continuous part multiplied by the return due to the dividend
payment) is negligible and can be ignored. This is the same as the result show in equation 44.
The rst term (which is the squared returns of the continuous part), can be approximated as in
equation 7.
log
2
_
S
i
S
i1
_
2 log
_
S
i
S
i1
_
+ 2
S
i
S
i1
S
i1
(53)
We rewrite the rst term as
log
_
S
i
S
i1
_
= log
_
S
i
S
i1
_
+ log
_
S
i
+D
i
S
i
_
(54)
Hence
RV (T) 2 log(S
T
/S
0
) + 2
n
i=1
S
i
S
i1
S
i1
+ 2
j:t<
j
T
log
_
S
j
S
j
+D
j
_
+
j:t<
j
T
log
2
_
S
j
S
j
+D
j
_ (55)
If we take the expectation and set D
j
=
j
S
j
+
j
1
j
we arrive at the result in equation 47.
12
References
[Bue06] Hans Buehler. Options on variance: Pricing and hedging. Technical report, IQPC
Volatility Trading Conference, 2006.
[Bue10] Hans Buehler. Volatility and dividends. Working paper, 2010.
[CM98] P. Carr and Dilip Madan. Volatility: New Estimation Techniques for Pricing Deriva-
tives, chapter Towards a Theory of Volatility Trading. Risk Books, 1998.
[DDKZ99] Demeter, Derman, Kamal, and Zou. More than you ever wanted to know about
volatility swaps. Technical report, Goldman Sachs Quantitative Strategies Research
Notes, 1999.
[Dup94] Bruno Dupire. Pricing with a smile. Risk, 7(1):1820, 1994. Reprinted in Derivative
Pricing:The Classical Collection, Risk Books (2004).
[Gat06] J. Gatheral. The Volatility Surface: a Practitioners Guide. Finance Series. Wiley,
2006.
[HHL03] Espen Haug, Jorgen Haug, and Alan Lewis. Back to basics: a new approach to the
discrete dividend problem. Technical report, 2003.
[Neu92] A Neuberger. Volatility trading. Technical report, London Business School working
paper, 1992.
[Nie06] M. H. Nieuwenhuis, J.W. Vellekoop. Ecient pricing of derivatives on assets with
discreate dividends. Applied Mathematical Finance, 2006.
[OBB
+
06] Marcus Overhaus, Ana Berm udez, Hans Buehler, Andrew Ferraris, Christopher
Jordinson, and Aziz Lamnouar. Equity Hybrid Derivatives. Wiley Finance, 2006.
[Whi12] Richard White. Local volatility. OG notes, OpenGamma, 2012. http://developers.
opengamma.com/quantitative-research/Local-Volatility-OpenGamma.pdf.
13