Electronic copy available at: http://ssrn.

com/abstract=1706758
Stochastic Proportional Dividends
Working Paper
Hans Buehler, Anissa Dhouibi, Dimitri Sluys
JP Morgan Equity Derivatives Group
Quantitative Research
London
hans.buehler@jpmorgan.com
anissa.s.dhouibi@jpmorgan.com
dimitri.sluys@jpmorgan.com
First draft: January 2010 - This draft: December 2010,
Revision 1.013
Abstract
Motivated by recently increased interest in trading derivatives on div-
idends, we present a simple, yet ecient equity stock price model with
discrete stochastic proportional dividends. The model has a closed form
for European option pricing and can therefore be calibrated eciently to
vanilla options on the equity. It can also be simulated eciently with
Monte-Carlo and has fast analytics to aid the pricing of derivatives on
dividends. While its eciency makes the model very appealing, it has the
twin drawbacks that dividends in this model can become negative, and
that it does not price in any skew on either dividends or the stock price.
We present the model and also discuss various extensions to stochastic
interest rates, local volatility and jumps.
Keywords: Options on Dividends, Stochastic Dividends, Dividend Yield
1 Introduction
The recent years have seen an increased interest in trading more directly one
of the most basic features of equity stock prices, its dividends. Dividends are
comparable to coupons of a bond since they provide the equity holder with a
stream of income. The main dierence, of course, is that a company can decide
1
Electronic copy available at: http://ssrn.com/abstract=1706758
how much dividend it will pay, hence the dividend amounts are a function of
the companys performance.
In the derivatives worlds, dividends have traditionally been traded either di-
rectly via forms of over-the-counter (OTC) equity swaps or implicitly by trading
calendar spreads of forwards. The latter in particular allows to trade the real-
ized dividends against the assumed, or implied, dividends. Starting in 2001,
dividend swaps have become a popular way for dealers to hedge their dividend
exposure [11]. A dividend swap pays directly the aggregated value of all (gross)
dividends a company pays in exchange for the initially agreed strike price.
The next step in listed product development was to introduce dividend fu-
tures on an exchange. The rst such product, the Dow Jones EURO STOXX
50 Index Dividend Futures was introduced by Eurex in June 2008 [1]. Figure 1
shows the growth of the volume of dividend futures over the last two years.
Figure 1: Growth of dividend future volumes since beginning of 2009.
The culmination of this process so far was the introduction on May 25, 2010
of options on the S&P500 Annual Dividend Index (DIVD) and the S&P 500
Dividend Index (DVS) on the CBOE [12], and the introduction on the same day
of options on both Euro STOXX50 Index Dividend Futures (FEXD) and the
Euro STOXX50 Index Dividend Points (DVP) on Eurex [2]. The exchange
traded contracts have the potential to improve price discovery for the volatility
market on dividends substantially.
Aside from the listed market, there also the market of OTC derivatives.
Here, the potential for structuring ideas around dividends are wide: options on
dividends are one way to expresss a view on dividends, but other products such
as Dividend Yield Swaps and Knock-Out Dividend Swaps can give a much more
tailored exposure to the realized dividends of an index or a given equity.
In addition, there is also a natural interest in products which give exposure
to the yield gap between the yield on bond investments and the yield derived
from dividends.
2
Pricing with Dividends
In the light of these developments, it is surprising to nd that there is actually
not much literature on the modeling of dividends for derivatives pricing and risk
management, except on the question on whether these should be proportional
to the stock price, xed in cash or a mixture of the two.
1
Beyond this, the only
reference to our knowledge on modeling the dividends of an equity is implicitly
given by Gaspars theoretical work [7] on modeling the entire term-structure of
the forward curve as a diusion in an Hilbert-space. This HJM-approach is
related to a version of our model where a dividend yield is modeled instead of
a stream of discrete dividends.
With this article, we aim to contribute to the modeling discussion on div-
idends with a model where dividends are discrete and proportional; the (log-)
proportionality factor itself is given by a function of an Ornstein-Uhlenbeck
process, meaning that the resulting dividend yield is mean-reverting around a
long-term average yield level. While the drawback of using a normal process for
the proportionality factor is that it can lead to negative dividends, the upside is
that the model remains very tractable: pricing vanilla options on the equity for
the purpose of calibration, simulating the model with Monte-Carlo and evaluat-
ing future forwards are all ecient operations. This gives the model the avor
of a robust Black & Scholes-style workhorse for pricing various variants of
structures on dividends.
This article is structured as follows: rst, we will introduce our setup and
a set of interesting payos. We then present our main model and discuss its
properties as well as a calibration procedure which we apply to market data
for STOXX50E. In the next section we show how an ecient Monte-Carlo
scheme can be implemented and how we can compute dividend future prices
on the given Monte-Carlo paths. We also compute some sample option prices.
A nal section discusses extensions of the model to stochastic interest rates,
credit and jump risk, and local volatility (with numerical methods). The sec-
tion concludes with a comment on the connection to Gaspars HJM-framework
for the forward curve.
An appendix contains most of the actual calculations.
Acknowledgements
We are very grateful for the help of Christopher Jordinson, now at UBS, many
of whose ideas are incorporated in the model we present here. His input was
crucial to the development of the model.
1
For a thorough review of this setting, cf. Bermudez et al. [3] and, in more detail,
Buehler [6].
3
2 Modeling Dividends
We aim to model an equity stock price process S = (S
t
)
t
under an interest
rate process r = (r
t
)
t
and with eective borrowing costs of = (
t
)
t
. Both are
assumed to be deterministic. For notional convenience we write R
t
:= e

t
0
rss ds
for the drift factor of the equity and B
t
:= e

t
0
rs ds
for the cash account.
2
To model the dividend stream of the equity, we assume that the stock is
going to pay N random dividends = (
i
)
i=1,...,N
at dividend dates 0 <
1
<
<
N
.
3
(We will assume that the dividend dates are xed and known in
advance.) Modeling the dividends of the equity as stochastic means modeling
each paid dividend
i
as a random variable which is known by the dividend
time
i
.
4
Dividend Derivatives
A dividend swap on the equity between the dates T
1
and T
2
pays the accumu-
lated dividends between these two dates against a xed strike K, i.e.

i:T1iT2

i
K .
Similarly, an option on realized dividends, or in short OOD, is a vanilla payo
on the same quantity, i.e. in the case of a call on realized dividends
_
_

i:T1iT2

i
K
_
_
+
where we used the notation x
+
:= max(0, x). Note that dividend futures usually
pay the realized dividends of the respective underlying index. Hence, an option
on a dividend future is eectively also an option on realized dividends for the
period of the future.
More advanced OTC products on dividends are structures, where the payo
of the dividend is linked to the performance of the equity itself. One example is
a knock-out dividend swap, i.e. a dividend swap which knocks out if the equity S
trades below a given barrier. The payo for this product is
1
min
t:T
1
<tT
2
St<B
_
_

i:T1iT2

i
K
_
_
. (1)
2
We do not model the possibility of default here, but an extension to the case of a deter-
ministic hazard rate model is well within our framework, c.f. the discussion in [6].
3
For ease of exposure we assume that the payment dates are equal to the ex-div dates. An
extension to separate ex-div and payment dates is straight-forward.
4
An exception here is Japan, where dividends are usually announced after the ex-div date.
4
Another example is a dividend yield swap, which pays the sum of realized divi-
dends over the monthly average spot of the equity:

i:T1iT2

k:T1t
k
T2

St
k
(2)
where (t
i
)
i
are monthly xings (other variants scale the realized dividends by
the stock price at the end of the period, or divide each dividend by the stock
price of the previous trading day).
As mentioned before, there is a recent interest in rates-dividend hybrid prod-
ucts, for example those which allow to manage an exposure to the dierence in
dividend and bond yields. Figure 2 shows that the implied yield on equity is
recently higher than the yield which can be derived from investing in bonds.
A good example of a product which allows to hedge this yield gap would be a
Figure 2: Dividend yield is high in both absolute and relative terms compared to
bond yields.
Leveraged Div Yield Swap Certicate which basically pays coupons which are
a dierence between interest rates and dividends: We simplify the product for
the sake of simplicity. The simple version works as follows: dene the realized
dividend yield between two times T
1
and T
2
as
ryld(T
1
, T
2
) :=
1
S
T2

i:T1<tiT2

i
and let CMS
t
(1Y ) denote the 1Y libor rate observed at time t. Then, the
product pays
max
_
100% + 5

t=1y,2y,3
_
CMS
t
(1y) ryld(t, t + 1y)
_
, oor
_
where the oor is, for example, 30%. In this form, the product provides a
positive exposure to a drop in dividend yield vs. interest rates.
5
While we will mainly discuss our model in a deterministic interest-rate envi-
ronment, section 3.1 explains how it can be incorporate stochastic interest rates
in order to risk manage rates-dividend hybrids.
The Equity and its Dividends
To model an equity process with a stream of stochastic dividends, the challenge
is to nd a distribution for the dividends which still allows pricing and risk
management of options on the underlying eciently such that the model can be
calibrated easily to observed prices of equity vanilla options.
To approach this topic, we rst assume that we face an idealized liquid
market of dividend swaps, which allows us to forward-trade any dividend prior
to its dividend date. In particular, we assume that there is at any time t a
market for the dividends with dividend dates past t. We denote the forward-
price at time t for the ith dividend accordingly by
i
t
. As a consequence of our
assumptions, its value is given under any risk-neutral measure as

i
t
:= E
t
_

. (3)
Note that the process (
i
t
)
t[0,i]
is a (local) martingale.
With the setup of the previous section, we can deduct a basic shape of any
equity model which is consistent with a given dividend stream: basically, the
assumptions that dividends can be traded means that the stock price cannot
drop below the properly discounted value of future dividends at any point.
Following the same arguments as laid out in Buehler [6], standard no-arbitrage
replication arguments imply that the stock price process has to have the form
S
t
= R
t
_

S
0
X
t
+

i:i>t

i
t
R
i
_
with

S
0
= S
0

i

i
/R
i
for a (local) martingale X. Accordingly, the forward
at time zero of the process is given as
F
t
= R
t
_
_
_
S
0

i:it

i
0
R
i
_
_
_
.
Black & Scholes
The standard extension to Black & Scholes model is the classic case where the
dividend is proportional to the equity, i.e.
5

i
= S
i
(1 e
Di
) with D
i
:= ln
_
1

i
0
F
i
_
. (4)
5
This follows since both F

e
D
i
= F
i
and F

i
= F
i
have to hold.
6
The corresponding stochastic model is then given by
S
t
= F
t
X
t
,
where X is a log-normal model with pure equity volatility term structure .
Its SDE is given as
dS
t
S
t
= (r
t

t
) dt +
t
dW
t

i:tit
(1 e
Di
)
i
(dt) (5)
(we use
x
() to denote the Dirac-measure in x). In this model, the dividend
stream is obviously stochastic it exhibits the same volatility as the underlying
equity.
Stochastic Proportional Dividends
The model we want to propose here is given as well in the form
dS
t
S
t
= (r
t

t
) dt +
t
dW
t

i
(1 e
di
)
i
(dt) (6)
but the dividend ratios d
i
are random: to model them, we use an Ornstein-
Uhlenbeck process
dy
t
= y
t
dt + dB
t
and set
d
i
:= (D
i
+ E
i
y
i
) + C
i
(7)
where D
i
is the Black & Scholes value for the proportional dividend as dened
in (4). The constant E
i
allows to blend between normal and log-normal volatility
for the dividend yield by using E
i
1 or E
i
D
i
, respectively. Finally, the
constant C
i
is determined by matching the forward such that E[S
t
] = F
t
for
all t using an analytical procedure; see appendix A.2 for details.
Remark 2.1 Our model can easily be extended to incorporate a xed cash-
dividend part for each dividend. For example, we could assume that a fraction
i
of each dividend
i
is xed in cash. Compared to (4), the stochastic dividend
in such a model becomes

i
= S
i
(1 e

Di
) + (1
i
)
i
0
.
for a suitable

D
i
which ensures that the forward of the process is correct. Ap-
pendix A.1 provides some more details.
To motivate our model choice, let us dene the forward yield between T
1
and T
2
, seen at a time t, as the sum of the expected dividends between the two
reference times, divided by the current spot level:
y
t
(T
1
, T
2
) :=

i:T1<iT2

i
t
S
t
.
7
It is easy to see that in the current model, each dividend has the property
that log
i
t
/S
t
is an ane function of y
t
, which implies that the log-yield
log y
t
(T
1
, T
2
) is also approximately proportional to y
t
,
y
t
(T
1
, T
2
) ay
t
+ b (8)
for some constants a and b.
Looking then at historical data, we indeed see that the log-forward yield
seems to mean-revert around some xed mean. During the great moderation
of the early 2000s in particular, we have seen a very stable pattern of mean-
reversion. Even though this pattern has unsurprisingly been severely disturbed
with the onset of the nancial crisis and, in particular, Lehmans default in
October 2008, the structure has since vaguely recovered, albeit to a regime with
a lower level of average implied yield. Figure 3 illustrates this point. In general,
Figure 3: The graph shows historical log-dividend yields for future periods. For
example, the 9Y/10Y point refers to the oating maturity forward dividend swap
between 9Y and 10Y, divided by the spot at the observation point.
the assumption of a mean-reverting dividend yield in a non-distressed market
is a very natural from an economic point of view, in particular if applied to
indices.
Model Properties
The rst important observation is that the models stock price is log-normal:
its explicit form is
S
t
= R
t
S
0
e

t
0
s dWs
1
2

t
0

2
s
ds

i:
i
t
{Ci+Di+Ei[y0e

i
+

i
0
e
(
i
s)
dBs]}
,
8
with log-variance
Var(t) =
_
t
0

2
s
ds+

i:it
E
i
_
_
_

j:ji
E
j
_

2
1 e
2j
2
2
_
i
0

s
e
(is)
ds
_
_
_
_
.
(9)
This means that if we are given the dividend parameters y
0
, , and , then
we can bootstrap a piece-wise constant forward volatility (
k
)
k
to matching an
observed term structure of Black & Scholes (spot-) implied volatilities (
k
)
k
dened for maturities T
1
< T
2
< (see appendix A.4 for details). Since this is
an analytic procedure, this can be performed on-the-y in order to ensure that
the model always reprices a selected term structure of observed market implied
volatilities on the stock price.
The task remains to determine reasonable values for the dividend volatility
term structure via the two parameters and , and the correlation between
dividend yield and the spot price (the initial value y
0
can be left at zero). For
liquid indices, we can calibrate these parameters to ATM prices of options on
dividends. As an example, we have used internal JP Morgan price indications as
of 14 October 2010 for options on dividend prices for STOXX50E to calibrate
the model. The prices provided for Dec10-11 and Dec11-12 ATM options on
dividends are quoted in terms of Black & Scholes implied volatilities
6
as 7.8%
and 16.1%, respectively (the dividend swaps traded at 114.0 and 109.2).
To calibrate our model to just two maturities of options on dividend prices,
we have xed our correlation
7
at 95% and then run a simple minimizer which
yielded a dividend volatility of = 27% and a mean-reversion speed of = 1.4.
The resulting ATM implied BS volatilities for the two maturities in the model
are then 7.9% and 16.1%, respectively. These have been calibrated using a
Monte-Carlo simulation with 10,000 paths.
In gure 4 we show the calibrated model prices against market price indica-
tions for options on dividends. The data show that the market price indications
exhibit far more skew than the model is able to replicate. This is in line with
the models basic structure where the sum of dividends is roughly log-normal,
hence we do not expect much skew. Figure 5 shows the entire implied volatility
surface for options on dividends.
The Role of Dividend Correlation
Correlation plays a particular role in this model. Recall that the yield in the
model is approximatively an ane function of y
t
, cf. equation (8). Figure 6
illustrates the impact of a change of correlation on the relation between spot
and yield as expected, a positive correlation implies a positive yield with
raising spot while a negative correlation implies a drop in yield.
6
I.e. the forward in the Black & Scholes formula is set to the dividend swap fair strike and
then the respective implied volatility is calculated from observed market prices.
7
The very negative correlation is not only necessary to be able to provide a decent t to
the observed market prices; it is also economically sensible as we will explain below.
9
Figure 4: Prices for options on dividends; the model is calibrated against the ATM
strike of the market price indications. The data has been priced as of mid October 2010.
Between the two, the latter is a much more realistic assumption since an
increase in spot is not usually immediately followed by an increase in dividend
levels; in a sense, the negative correlation between spot and dividends sta-
bilizes the level of dividends and gives them some constantness. Figure 7
shows visually how realistic paths generated by a negative correlation regime
look until the crisis in September 2008. After the crisis, however, the two
were more correlated: dividends were cut along with the fall in stock prices.
More recently, the market has recovered and we see again a more anticorrelated
behavior.
This means that the correlation in our model should generally be signicantly
negative in order to produce the well-known eect that short term dividends
are much more certain than longer term dividends. This is also borne out of
the market prices for options on dividends: the above calibration only works
well for very negative correlations. If the correlation is set at, say, -50%, then
there is no combination of mean-reversion speed and dividend volatility which
ts the provided market.
Negative Dividends
While most of the properties of our model are appealing, the models structure
also reveals its main drawback: the fact that the dividends themselves can
become negative. This happens if d
i
becomes negative. The probability of this
happens at a dividend date
i
is
P
_
Y
_
1 e
2i
2

D
i
+ C
i
E
i
y
0
e
i
_
10
Figure 5: The implied volatility surface for options on dividends in our model.
Strikes are provided as oset to 100% ATM.
where Y is standard normal. For the calibrated model above, the annual prob-
ability of having negative dividends is around 1% in the model as is shown
in gure 8. The same gure also shows that the size of the negative dividends
in relation to the total forward is (predictably) small.
However, we think that the advantage of analytical tractability of the model
far outweighs the downside of having potentially negative dividends.
2.1 Using the Model
For most derivatives on dividends, we will need to revert to numerical methods
to compute the value of the payo and risk manage it. We will here discuss
briey how to implement an ecient Monte-Carlo and will show how the model
prices some options on dividends.
In order to simulate the model (6), we assume that we have some observa-
tion dates 0 < t
1
< < t
n
of interest at which points we wish to evaluate
our payo; for ease of exposure we assume w.l.g. that these dates include all
dividend dates in the respective period.
A convenient feature of our essentially normal model is that we do not need
to use Euler to simulate our SDE because the model is in fact normal between
two of the observation dates. Assuming constant equity vol of
k
between t
k1
and t
k
= dt
k
+ t
k1
, the variance of the increment dy is
y
k
=
2 1e
2dt
k
2
, the
variance for the equity increment is
S
k
:=
2
k
dt
k
and the covariance between
11
Figure 6: Impact of correlation on the relation between spot and yield; the graphs
show expected dividends between Dec 11 and Dec 12 divided by the spot at observation
time. Dividend volatility was 50%, correlation -80% and mean-reversion speed 1.
Figure 7: Visual comparison of a negative correlation case above visually with historic
data.
the two is
k
1e
dt
k

, which means that the correlation is

k
:=
1 e
dt
k

1 e
2dt
k
_
1
2
dt
k
.
Hence, if we chose a iid sequence of standard normal random variables (Y
k
,

Y
k
)
k
,
we can simulate on big step between t
k1
and t
k
using
y
t
k
= y
t
k1
e
dt
k
+
_

y
k
_

k
Y
k
+
_
1
2
k

Y
k
_
log S
t
k
= log S
t
k1
+
k
_
dt
k
Y
k

1
2

2
k
dt
k
+ log
F
k
F
t
k1
(E
k
y
t
k
+ C
k
) .
(We used a slightly lax notation here: the D
k
, E
k
and C
k
are either the relevant
coecients from a dividend payable at t
k
or zero.)
Another strength of the model is that it is possible to eciently calculate
12
Figure 8: Illustration of the risk of having negative dividends for the calibrated
model: the graph shows the annual probability of having negative dividends (around
one percent) and the ratio of the expected negative part over the full expected sum
over the same period, which hovers around 0.5%.
the forward price of the equity as a function of spot and the driving Ornstein-
Uhlenbeck process u at a given future time eciently since it has the form
F
t
(T) := E
t
[ S
T
] = S
t
R
T
R
t
e
At(T)yt+Bt(T)
for some deterministic functions A and B (see appendix A.3 below). This is
important because it allows us
8
computing future expected dividend values on
a given Monte-Carlo path eciently using

t
= F
t
(

) F
t
(

) = F
t
(t
1
)
R

R
t

1
F
t
(

)
(x y := max(x, y)). Hence, the model permits ecient access to pricing divi-
dend swaps analytically within a Monte-Carlo simulation. This means we can
use the model to price and risk manage not only options on realized dividends,
but also on more involved trades which depend in value on futures on dividends.
As an example, gure 9 shows a sample path of both stock price and a div-
idend future generated by the model.
To illustrate the pricing of derivatives with the model, we used it to price
two types of products with the parameters estimated before for STOXX50E.
The rst product is the aforementioned Knock-Out Equity Swap (1). Using the
same calibration as above, and a barrier of 90% of initial spot,
9
we get
8
We use the fact that by denition Ft(T) = R
T
/Rt(St

i:t
i
T
Rt/R
i
t(i)).
9
In practise, this model should not be used to price barriers due to its lack of skew for the
equity. Below is a brief comment on how to incorporate skew.
13
Figure 9: A Monte-Carlo path of spot and expected dividends divided by spot where
the dividends are computed from the observation time to a xed maturity (Dec 13).
Dec10-11 Dec11-12 Dec12-13 Dec13-14
35.1 46.9 45.9 47.6
Another interesting product is the Dividend Yield Swap (2). It pays the sum
of realized dividends over the monthly spot price average.
Our model gives us:
Dec10-11 Dec11-12 Dec12-13 Dec13-14
4.13% 4.09% 3.97% 3.97%
3 Extensions and Related Models
We briey want to give some overview over possible extensions to the current
model.
3.1 Stochastic Interest Rates
An interesting aspect of dividend modeling has to be the relation ship between
dividends and interest rates. From an economic point of view, both rates prod-
ucts and dividend-paying equity play a similar role in providing investors with
a stream of income. Consequently, we can assume that the markets dividend
yield is related to the level of interest an investor can earn by buying just a
coupon-bearing bond.
While we will not discuss the economic situation much further here, we would
like to point out that the current model can easily be adapted to the standard
Hull & White interest model [9], and that it retains its analytical tractability
since the distribution of the stock price remains log-normal.
To start with, let us assume that we can observe at time zero a market of
zero coupon bonds (P
t
)
t
for each maturity t. We are going to model the interest
14
short rate as an Ornstein-Uhlenbeck process by setting:
dr
t
= (
t
qr
t
) dt +
t
dW
r
t
.
Mean-reversion speed q, the interest volatility and the correlation structure
between the equity and the short rate are free parameters have to be calibrated
or estimated. The level of mean-reversion, , is implicitly given by tting the
model to the observed zero coupon bond term structure by stipulating
E
_
e

t
0
rs ds
_
!
= P
t
(in practical applications, one would never actually calculate but use it only
in integrated form; see [3]). Given our Hull & White model, we can now dene
the equity process via, once again,
dS
t
S
t
= (r
t

t
) dt +
t
dW
t

i
(1 e
di
)
i
(dt)
with d
i
given again as in (7) with adjusted Cs. Since r and therefore any
integral over it are normal, it follows that the stock price process in this joint
Hybrid model is still log-normal. Consequently, the analytic tractability of the
model is preserved and it is a very convenient candidate if we want to price
joint rates-dividend derivatives. We even maintain the ability to run an ecient
Monte-Carlo by using the methods discussed in Brockhaus et.al. [5], page 36.
Remark 3.1 Both our original and our rates/dividends hybrid model allow the
introduction of several Ornstein-Uhlenbeck factors to drive the respective curves;
see also Hulls description for the 2-factor rates model in [8].
3.2 Jump Risk, Crash Risk, Credit Risk
Handling plain credit risk within a deterministic intensity model is straight
forward following the recipe in [6]. If one is interested in adding less drastic
jumps to the equity process which also impact the dividend, then the classis
approach by Merton [10]. One point of consideration is the behavior of the
yield if the equity exhibits a sudden drop. We would assume that a sudden
drop in the spot price will also lead to a drop in yield. This can be seen very
well in gure 3 on page 8 which shows that the stock market drop in October
2008 lead to a substantial drop in dividend yield as well.
To model such a relationship, let us consider series of normal crash or jump
scenarios (
j
)
j=1,...,M
for the stock price process, each with mean m
j
and volatil-
ity s
j
, and driving Poisson processes (N
j
)
j
with crash intensities (
j
)
j
. We also
assume we have corresponding shock scenarios (g
j
)
j
for the dividend yield.
The diusion for the stock price is then
dS
t
S
t
= (r
t

t
) dt +
t
dW
t

i
(1 e
di
)
i
(dt)
_
_
M

j=1

j
dN
j
t
+

j
dt
_
_
.
15
with

j
:=
j
(e
mj+
1
2
s
2
j
1). In order to incorporate synchronous jumps
into the dividend yield as well, we additionally rewrite our driving Ornstein-
Uhlenbeck process as
dy
t
= y
t
dt + dB
t

j=1
g
j
dN
j
t
.
This model has the desired properties of synchronous crashes in equity and yields
while it maintains enough exibility to be adapted to small jumps which do
not aect the yield itself (they will aect the dividends being proportional to
the spot price level). Note that one drawback of incorporating negative jumps
into u is that the proportional dividend factors d
i
are more likely to become
negative.
In terms of analytical tractability this model is once again very convenient:
conditional on the number of jumps for each Poisson-process this model has a
log-normal distribution which means
that it can be used eciently using Fourier-based pricing for European op-
tions.
3.3 Fitting the Equity Smile
Since the proposed models stock is essentially log-normal, it does not have any
skew when pricing vanilla options on the equity. This obvious drawback can
be addressed at least in the case for deterministic interest rates by calibrating
numerically a local volatility function on top of the equity.
The corresponding SDE for the equity is
dS
t
S
t
= (r
t

t
) dt +
t
(S
t
) dW
t

i
(1 e
Di
)
i
(dt) .
The local volatility function itself can then be calibrated using forward-PDE
methods such as the one described in [3] for the calibration of a local volatility
function on top of a stochastic rates equity pricing model.
3.4 Stochastic Yield Dividends
One of the main features of our stochastic proportional dividend model is that
it keeps the dividend payments to exactly the payments found in the market
data. An alternative is to model a yield on top of the prevalent forward curve:
such a model is given by
dS
t
S
t
= (r
t

t
y
t

C
t
) dt +
t
dW
t

i
(1 e
Di
)
i
(dt)
where the D
i
are the proportional dividends from the Black & Scholes formu-
lation (5), and where the continuous function

C is chosen such that the model
16
ts the forward, i.e. E[S
t
] = F
t
.
10
While being numerically easier to handle
than the proportional dividend model above, it has the disadvantage that the
yield will often be negative over periods between the input market dividend
dates. As such, the model might be more useful if the equity forward itself is
given as a yield in fact, this model is related to Gaspars HJM-framework [7]
for forward curves which we mentioned in the introduction; it is basically the
normal 1F-version in her framework.
We will not discuss the details of this approach further, but the interested
reader will nd that most calculations are either similar or simpler compared to
our proportional dividend model.
4 Conclusions
We have discussed a stochastic dividend model with very tractable analytics for
calibration towards the vanilla market and for the pricing and risk management
of dividend derivatives. To the best of our knowledge, this is the rst attempt to
model the dividend stream of an equity in a consistent manner for the purpose
of derivatives pricing. The models simplicity qualies it as a standard tool in
a derivatives library, while the fact that its dividend are not always positive
and that it does not capture the implied volatility skew in either equity or
dividends means that there is clearly a further need for development in the
dividend modeling area.
A.1 Cash Dividends
In order to extend the model to support explicit cash dividends, we follow the
approach in [6]: rst, we dene for each dividend
i
are ratio
i
which denotes
how much of todays expected dividend value
i
0
will certainly be paid (an
extension to support simple credit risk is trivial). The model is then written as
follows: rst dene an adjusted spot

S
0
:= S
0

i
P(0,
i
)
i

i
0
and a proportional forward

F
t
:= R
t
_
_
_

S
0

i:it
(1
i
)
i
0
R
i
_
_
_
.
Accordingly, we dene new proportionality factors

D
i
:= ln
_
1
(1
i
)
i
0

F
i
_
.
10
I.e. we use the relation e

t
0

Cs ds
= E

t
0
s dWs
1
2

t
0

2
s
ds

t
0
ys ds

to nd

C.
17
The proportional dividend for a given

E
i
will then be modelled as

d
i
:=
_

D
i
+

E
i
y
i
_
+

C
i
which gives us the diusion
d

S
t

S
t
= (r
t

t
) dt +
t
dW
t

(1e e
di
)
i
(dt) .
The stock price is then given as
S
t
:=

S
t
+

i:i>t
P(t,
i
)
i

i
0
and will, with appropriately calculated

C
i
, reprice the forward. The calculations
for

C are the equivalent to the one for the model discussed in the text.
A.2 Matching the Forward
Fix some . We want to compute
() = E
t
_
e

0
t dWt
1
2

0

2
t
dt

j:j
dj
_
.
As a rst step, we change into the equity measure Q under which
dy
t
= (
t
y
t
) dt + dB
t
.
with
t
:=
t
. Under this measure, we have
() = E
Q
_
e

j:j
dj
_
.
Since we have de
t
y
t
= e
t

t
dt + e
t
dB
t
, we get
y
t
= y
u
e
(tu)
+
_
t
u
e
(ts)

s
ds +
_
t
u
e
(ts)
dB
s
.
That means that for any p,
E
Q
_
e
pyt

F
u

= e
pyuK(t,u)p(t,u)+p
2
(t,u)
. (10)
with
_

_
(t, u) :=
_
t
u
e
(ts)

s
ds
K(t, u) := e
(tu)
(t, u) :=
1
2

2 1 e
2(tu)
2
.
We also abbreviate

j
:= (
j
,
j1
) , K
j
:= K(
j
,
j1
) and
j
:= (
j
,
j1
) .
18
Iteration
The aim is to make sure with out choice of (C
j
)
j=1,...,N
that
1
!
=
E
Q
_
e

j:j
dj
_
e

j:j
Dj
= E
Q
_
e

j:j
(Ejy
j
+Cj)
_
.
Consider the formula
c

(p) := log E
Q
_
e

j:j<
(Ejy
j
+Cj)(E

+p)y

_
which obviously yields with C

:= c

(0) the desired correction terms. We will

derive these functions iteratively: we start with = 1. Using (10) yields readily
c
1
(p) = (E
1
+ p)K
1
y
0
(E
1
+ p)
1
+ (E
1
+ p)
2

1
.
The next step is > 1 to determine c

(p) assuming we know c

j
(q) and there-
fore C
j
for j < .
e
c

(p)
= E
Q
_
e

j:j<
(Ejy
j
+Cj)(E

+p)y

_
= E
Q
_
e

j:j<
(Ejy
j
+Cj)
E
Q

1
_
e
(E

+p)y

_ _
= E
Q
_
e

j:j<
(Ejy
j
+Cj)
e
(E

+p)K

y
1
(E

+p)

+(E

+p)
2

_
= E
Q
_
e

j:j<1
(Ejy
j
+Cj)(E
1
+(E

+p)K

)y
1
_
e
(E

+p)

+(E

+p)
2

C
1
= e
(E

+p)

+(E

+p)
2

+c
1
((E

+p)K

)C
1
which means that
c

(p) = (E

+ p)

+ (E

+ p)
2

+ c
1
((E

+ p)K

) C
1
(11)
is well-dened.
A.3 Future Forwards
A very similar calculation as the above allows us to compute the future forwards
of the model,
F
t
(T) := E
t
[ S
T
] .
We sketch the idea here: let n :
n
T <
n+1
and k :
k1
t <
k
. First of
all,
F
t
(T) = S
t
R
T
R
t
E
Q
t
_
e

j:kjn
(Ejy
j
+Cj+Dj)
_
.
The last term can be handled with the same method as above: dene for : n
c

(t; p) := E
Q
t
_
e

j:kj<
Ejy
j
(E

+p)y

_
.
19
The rst term we need to know is
c
k
(t; p) := (E
k
+ p)K(t,
k
)y
t
(E
k
+ p)(t,
k
) + (E
k
+ p)
2
(t,
k
) .
and all further terms have the same structure as (11), i.e.
c

(t; p) = (E

+ p)

+ (E

+ p)
2

+ c
1
(t; (E

+ p)K

) .
The forward is then given as
F
t
(T) = S
t
R
T
R
t
e
cn(t;0)

j:kjn
Cj+Dj
S
t
R
T
R
t
e
At(T)yt+Bt(T)
(12)
for some A and B (note that c
n
(t; 0) is a function of y
t
).
A.4 Fitting to Equity Vanillas
Assume we are given a term structure of reference maturities 0 < T
1
< T
2
<
for which we are given Black & Scholes-implied volatilities (
k
)
k
on the stock
price. We want to bootstrap a piecewise linear equity forward volatility term
structure (
k
)
k
which is constant over the intervals [T
k1
, T
k
].
To do so, we iteratively solve the following quadratic equation in
k
:
Var(T
k
) Var(T
k1
) =
2
k
(T
k
T
k1
)
2
k

i:T
k1
<iT
k
e
i
E
i
O
i
k
+

i:T
k1
<iT
k
E
i
_
E
i
e
i

r:r<i

r
K
r
_
with
K
r
:=
e
Tr
e
Tr1

, O
i
k
:=
e
i
e
T
k1

, E
i
:=

j:ji
E
j

2
1 e
2j
2
.
References
[1] B.Baldwin:
The World of Equity Derivatives - The Essential Toolbox for Investores,
Eurex, September 2008,
http://www.eurexchange.com/download/documents/publications/
TheWorldofEquityDerivatives.pdf
[2] B.Baldwin:
Dividend Derivatives: Introduction of Options on EURO STOXX50 Index
Dividend Futures, eurex circular 082/10, May 2010,
hhttp://www.eurexchange.com/download/documents/circulars/cf0822010e.pdf
20
[3] A.Bermudez, H.Buehler, A.Ferraris, C.Jordinson, A.Lamnouar,
M.Overhaus:
Equity Hybrid Derivatives, Wiley, 2006
[4] F.Black, P.Scholes:
The Pricing of Options and Corporate Liabilities, Journal of Political
Economy, 81, pp. 637-59, 1973
[5] O.Brockhaus, A.Ferraris, C.Gallus, D.Long, R.Martin, M.Overhaus:
Modelling And Hedging Equity Derivatives, Risk Books, 1999
[6] H.Buehler:
Volatility and Dividends - Volatility Modelling with Cash Dividends and
Simple Credit Risk, WP, June 7, 2008
http://ssrn.com/abstract=1141877
[7] R.Gaspar
Finite Dimensional Markovian Realizations for Forward Price Term Struc-
ture Models, Stochastic Finance, 2006, Part II, 265-320
[8] J.Hull: Interest Rate Derivatives: Models of the Short Rate, Options,
Futures, and Other Derivatives (6th ed), Upper Saddle River, N.J: Prentice
Hall. pp. 657658, 2006
[9] J.Hull, A.White:
One factor interest rate models and the valuation of interest rate derivative
securities, Journal of Financial and Quantitative Analysis, Vol 28, No 2,
(June 1993) pp 235254
[10] R.Merton: Option Pricing When Underlying Stock Returns are Discon-
tinuous, Journal of Financial Economics 3 (1976) pp. 125-144.
[11] D.Wood:
Uncertain Dividends, Risk Magazine, Oct 2007
[12] D.Wood:
Options on DIVD and DVS Indexes, CBOE Website, May 2010,
http://www.cboe.com/micro/dvs/introduction.aspx and
http://www.cboe.com/micro/dvs/DIVDFAQ.pdf
21