Exotic interest-rate options

Marco Marchioro
www.marchioro.org

November 10th, 2012

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 1

Lecture Summary

• Exotic caps, floors, and swaptions

• Swaps with exotic floating-rate legs
• No-arbitrage methods (commodity example)
• Numeraires and pricing formulas
• stochastic differential equations (SDEs)
• Partial differential equations (PDEs)
• Feynman-Kac formula
• Numerical methods: analytical approximations

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 2

Common exotic interest-rate options

There are all sorts of exotic interest-rate options.

The least exotic, i.e. commonly traded, types are

• Caps and floors with a digital payoff

• Caps and floors with barriers

• Bermuda and American Swaptions

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 3

Caps and floors with a digital payoff

A vanilla Cap has caplets with payoff

CapletPayoff
V = N T (6m, 9m) max [(3m-Libor − K), 0] (1)

A digital Cap has caplets with payoff


 N T (6m, 9m) R for 3m-Libor ≥ K

CapletPayoff
D = (2)

 0 for 3m-Libor < K

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 4

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 5

Caps and floors with barriers

A barrier Cap has caplets with payoff

CapletPayoff
V = N T (6m, 9m) max [(3m-Libor − K), 0] (3)

• Knock-in: paid if 3m-Libor touches a barrier rknock-in

• Knock-out: paid if 3m-Libor does not reach rknock-out

Discrete barriers are checked at certain given barrier dates

Continuous barriers are checked on each daily Libor fixing
Digital payoffs are also available. A rebate is paid upon knock out.

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 6

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 7

Bermuda and American Swaptions

Very similar to standard Swaptions (known as European swaptions)

can be exercised at other dates

• Bermuda swaptions: exercised dates are discrete (typically with

the same tenor as one of the legs)

• American Swaption can be exercised at any time

Both are found in two different types of payoffs: co-terminal swap

and constant-maturity swap

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 8

Questions?

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 9

More exotic interest-rate options

• Caps and floors on Constant-Maturity-Swaps rates (CMS)

• Look-back options

• TARN (Target Accumulator Redemption Note) swap legs

• Spread options

Usually these options are gift-wrapped within a swaps paying a fixed

or a Libor rate

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 10

Exotic swap legs

• Constant-maturity swaps (CMS)

• Spread options

• Ratchet swap

• Range accruals

• Look-back options

• TARN (Target Accumulator Redemption Note) swap legs

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 11

Constant-maturity swaps (CMS)

A swap with a standard fixed-rate leg and a floating-rate leg that pays
at times Ti, with i = 1, . . . , n cash flows given by
Ci = N τi−1 CMSM (Ti−1) (4)
where

• τi is the year fraction between dates Ti and Ti+1

• CMSM (Ti) is the fair rate, observed at time Ti, of a vanilla swap
(fixed leg against Libor leg) with maturity date Ti + M

Constant-maturity swaps usually have Caps or Floors on the swap

rate
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 12

Swaps with spread options

A swap with a standard fixed-rate leg and a floating-rate leg that pays
a cash flow given by the difference of two swap rates.

For example the floating leg fixing at time Ti−1, and paying at time
Ti, payments are given by
 
Ci = N τi−1 max rfloor, CMS10Y (Ti−1) − CMS2Y (Ti−1)

• We are betting on the swap curve to steepen

• Usually a multi-factor model is needed to price this instrument

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 13

Swap with ratchet options

Also known as cliquet option.

A swap with a standard fixed-rate leg and a floating-rate leg that pays
at times T2, . . . , Tn Caplet-like coupons,


Ci = N τi−1 max 0, Li−1 − Ki−1 (5)

where Li is the Libor rate observed at time Ti, and Ki is the strike
satisfying


Ki−1 = max Ki−2, Li−2 for i = 3, . . . (6)

Note: coupon rates are not decreasing. Usually K1 = F wd12

Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 14

Swaps with range accruals

A swap with a floating-rate leg paying coupons

Ci = N τi−1 hri (7)
where the average rate hri is given by
1 X
hri = f [L(tk )] (8)
K k∈[T ,T )
i−1 i
K business dates between Ti−1 and T1,
f being the range function of Libor rate

 rin
 for Lmin ≤ L ≤ Lmax
f (L) = (9)

 r
out otherwise
Lmin minimum Libor rate and Lmax maximum Libor rate
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 15

Look-back options

A swap with a floating-rate leg paying coupons on the maximum Libor

fixing over a certain past period

Ci = N τi−1 rmax(Ti−1, Ti) (10)

where
rmax = maxk∈[Ti−2,Ti−1)[L(tk )] (11)

• We always receive the best Libor fixing in a range

• Sometimes look-back periods extend further back than one coupon

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 16

TARN (Target Accumulator Redemption Note) swap legs

A swap with a floating-rate leg paying coupons up to a maximum

accumulated rate rmax
Ci = N τi−1 ci (12)
where

c1 = min [rmax, L0]

c2 = min [rmax − c1, L1]
c3 = min [rmax − c1 − c2, L2]
...
Coupon rates are subtracted to the maximum rate until a zero rate
is reached
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 17

Questions?

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 18

Crash course on commodity contracts

(Example of no-arbitrage methods)

• Introduction to commodities

• The forward-spot and the forward-forward relationship

• Convenience yields

• Example of commodity derivatives: The forward contract and the

futures spread

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 19

Exposure to commodity prices

If you are managing an hedge fund and want some exposure on the
price of live cattle what can you do?

A. Become a cowboy overnight and buy some live cattle

B. Enter in a derivative contract that gives you an exposure to the

live-cattle price (e.g. a futures contract)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 20

Commodity futures contracts

• Commodity futures are among the oldest financial instrument

traded on any trading floor

• Futures on the cotton price, for example, have traded on the

market for longer than a century

• Refer to specialized literature for more details

• In this talk we describe how to generate simulated spot prices and

convenience-yield curves (to be used in the computation of risk
figures)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 21

Copper, Corn, and WTI Oil

We consider three samples commodities

• Copper (a.k.a Dr. Copper): non-ferrous industrial metal quoted,

in $ per ton, on the London Metal Exchange (LME)

• Corn: an agriculture commodity quoted, in $-cents per bushel

(27,216 kg), on the Chicago Board of Trade (CBoT)

• West-Texas-Intermediate Oil: an energy commodity quoted, in $

per barrel (158.987 liters), on the New York Mercantile Exchange
(NYMEX)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 22

Commodity futures contracts

A contract that allows the delivery of a commodity at a certain future

date (however settled daily according to the close price).

E.g. on 2011-11-01 we observed the following market quotes

Commodity Maturity Date Price

copper 2012G 2012-02-24 7993.00
copper 2012H 2012-03-27 7995.00
copper 2012J 2012-04-25 7995.50

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 23

Commodity forward contracts

A commodity forward contract is struck between two parties that

agree to buy/sell a given commodity at a future date at a pre-
determined price (the strike price).

Settlement could be physical or cash. In case of cash settlement the

payoff P at the maturity date T (for the long side) is given by the
difference between the value of the underlying commodity spot price
S(T ) and a strike price K.

For W lots we have,

h i
P = W · S(T ) − K (13)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 24

Arbitrage-free strategies

An arbitrage-free strategy is ...

A series of physical or financial transactions that starting with

a portfolio with a zero value end up with a risk-less portfolio.

The no-arbitrage assumption states that the final portfolio value is

zero with 100% probability (otherwise we could buy the cheaper part
and sell the dearer one making a risk-free profit with some probability).

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 25

Spot-forward relationship (1/2)

For a storable commodity. Consider a forward contract on a cer-

tain storable commodity (no costs no benefits) asset and set up the
following strategy:

Today borrow an amount S of currency, exactly enough to buy the

spot asset (since T is a short maturity the borrowing can be made
at reasonable, i.e. risk-free like, interest rates) and enter into a short
forward contract to sell the asset at a future date T at a price f ; then
store the asset until the date T ; when the date T comes, enforce the
contract and sell the asset for a price fT , finally, payback the loan
with interests.

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 26

Spot-forward relationship (2/2)

The strategy can be summarized in the following table

Date Description Cash flows Asset exch.
borrow an amount S of cash +S
t=0 purchase the asset at spot price −S + asset
enter into a short forward at T 0
0<t<T store the asset
use the asset to service the forward fT - asset
t=T
repay debt with interests −S[1 + R(T ) T ]

Because of no arbitrage we have

h i
fT = S 1 + R(T ) T (14)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 27

Adding costs and benefits

In the case of a commodity, for a dividend-paying stock, for a bond,

and for some other asset types, in general there are costs and/or
benefits associated in holding the asset,

fT = S[1 + R(T ) T ] + Cos(T ) − Ben(T ) (15)

Assuming costs and benefits to be proportional to the asset price,

using simple compounding,
h i h i
fT = S 1 + R T + S YCos T − YBen T (16)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 28

Convenience yield
We can define the convenience yield as the difference between benefits
and costs,
y = YBen − YCos (17)
so that
h e−y T i
fT = S 1 + R T − Y T = S −r T (18)
e
where we used continuous compounding. Using the discount factor
we have,
S
fT = e−y T (19)
D(T )

Note: the convenience yield y usually depends on the maturity T .

Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 29

Forward-forward relationship

Strategy: at time t = 0 enter into a long forward contract to buy

the asset at T1 for f1 units of currency, at the same time enter into
a short contract to sell the asset at date T2 for a price f2; at date
t = T1 use the long forward contract to buy the asset for a price of
f1 financing the purchase by borrowing the money from the market;
at a later date T2 sell the commodity for f2

From arbitrage-free assumption we have

D(T1) −y12 (T2−T1)
f2 = f1 e (20)
D(T2)
where y12 is the forward convenience yield between T1 and T2.
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 30

General spot-forward relationship

The forward-forward relationship between T2 and T3 is

D(T2) −y23 (T3−T2)
f3 = f2 e (21)
D(T3)

Chain linking the spot-forward formula and multiple forward-forward

relationships we can write
S
fj = e−yj Tj (22)
D(Tj )
where

yj Tj = y1 T1 + y12 (T2 − T1) + y23 (T3 − T2) + . . . + yij (Tj − Ti)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 31

Forward-futures relationship

The main difference between a futures contract and a forward con-

tract is that the former must be settled daily while the latter is settled
at the contract maturity. It can be shown that we have,

F T = f T + D(T ) σS σr ρr−S (23)

• σs is the asset-price volatility

• σr is the interest-rate volatility

• ρS−r is the correlation between the asset price and the money-
market account.

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 32

Convenience yields from futures quotes

We assume |σS σr ρr−S |  1 so that
F T ' fT (24)

Consider n futures contracts with maturities T1, T2, . . . Tn, having, re-
spectively, quotes F1 F1, F2, . . . Fn at some reference date.

From equation (22) and (24) we have

S
Fi = e−yiTi , for i = 1, . . . , n (25)
D(Ti)
compute the convenience yields yi’s as
!
1 S
yi = log (26)
Ti D(Ti)Fi
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 33

Yield on a commodity (1/2)

What is exactly the meaning of the convenience yield? Recall that

1+rT
fT = S (27)
1+yT
and suppose we owe a commodity that is worth S today. Assume no
credit risk and follow the strategy:

• Today sell the commodity; invest the money in risk-free deposit;

enter in a forward contract to buy W = 1 + y T lots at T

• At T recover the money and interests from the depo, and use the
money to purchase the commodity at the forward-contract price

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 34

Yield on a commodity (2/2)

In summary, starting with 1 lot of the asset

Date Description Cash flows Asset exch.
sell the asset at spot price S +S -1 · asset
t=0 invest the money in a deposit -S
enter into W long fwd contracts 0
0<t<T do nothing
redeem the deposit +S[1 + r T ]
t=T
use money to service the contract −S[1 + r T ] +W · asset

with W = 1 + y T . The contract payoff at redemption is given by

!
1+rT
W · fT = (1 + y T ) · S = S (1 + r T )
1+yT
The commodity provided an interest y T !
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 35

Few convenience-yield curves for Copper

3

2.5
Copper convenience yield (%)

1.5

0.5

0
0 1 2 3 4 5 6 7 8 9 10 11
Maturity T (years)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 36

Recall the zero interest-rate curves

1.6

1.4

1.2
Zero interest rates (%)

0.8

0.6

0.4

0.2

0
0 1 2 3 4 5 6
Maturity T (years)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 37

Few convenience-yield curves for Corn

10

6
Corn convenience yield (%)

-2

-4

-6
0.5 1 1.5 2 2.5 3 3.5
Maturity T (years)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 38

Few convenience-yield curves for WTI Oil

5

3
WTI Oil convenience yield (%)

-1

-2
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5
Maturity T (years)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 39

NPV of a commodity forward contract

Given the convenience yield curve y(T ) define the (continuously-

compounded) convenience discount factor Dc(T ) as

Dc(T ) = e−y(T ) T (28)

so that the commodity forward price can be written as
Dc(T )
f (T ) = S r . (29)
D (T )
The commodity forward contract NPV is then given by
h i
s
NPV = W · D (T ) · f (T ) − K (30)
At inception NPV=0 implies K = f (T ).

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 40

Commodities as currencies

Recall that given two currencies e and $ and their exchange rate Xe$
1e
1$ =
Xe$
we have the arbitrage-free forward exchange rate
D$(T )
Xe$(T ) = Xe$ e
D (T )

To be compared with the commodity relationship

Dc(T )
f (T ) = S r
D (T )
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 41

Questions?

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 42

Generic pricing formula (1/2)

Remember the discount factor relationship with the money account?

 RT 
− 0 r(t)dt −r −...−r
h i
D(T ) = E e τ
=E e 1 1 τ
n n (31)

If one payment C is made at a future date T we have,

 RT 
P V = D(T ) C = E e− 0 r(t)dt
C (32)

Assume now n deterministic cash flows C1, C2, . . . , Cn, one each day,
i.e. at dates T1, T2, . . . , Tn

P V = D(T1) C1 + D(T2) C2 + D(T3) C3 + . . . (33)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 43

Generic pricing formula (2/2)

For example, in the simple case with only three dates we have
−r −r −r −r −r −r
h i
PV = τ τ τ
E e 1 1 C1 + e 1 1 2 2 C2 + e 1 1 2 2 3 3 C3 τ τ τ

−r −r −r
n h  io
= τ τ
e 1 1 C1 + E e 2 2 C2 + e 3 3 C3 τ
 
 i
−r −r −r
h 
τ τ τ
X
= e 1 1 C + P (r2) e 2 2 E C2 + e 3 3 C3|r2
 1
r2

   
 
−r τ −r τ −r τ
X X
= e 1 1 C1 + P (r2) e 2 2 C2 + P (r3|r2)e 3 3 C3
r2 r3
 

Wouldn’t it be great if we could do the same for random cash flows?

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 44

Pricing in absence of arbitrage

There is a fundamental result from Harrison and Kreps (1979) that

holds for stochastic cash flows

Theorem: in absence of arbitrage, the generic pricing formula for an

option with a payoff H(T, rt), depending on the interest rates rt, is
 RT 
P V = E e− 0 rt dt
H(T, rt) (34)
where E is the expected-value operator on the risk-neutral measure.

In terms of the money-market account M (t), recall M (0), we have

" #
H(T, rt)
P V = M (0) E (35)
M (T )
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 45

Numeraires

It turns out that formula (35) is a special case of a general theorem

that holds for generic numeraire assets.

A numeraire is any tradable asset that

• is always positive
• does not pay any dividends (nor coupons)

Examples: risk-less money-market account, risk-less zero-coupon bond,

the price of a commodity (e.g. the gold price). See Hull book for more
examples.
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 46

Equivalent Martingale Measure

Theorem:
A continuous economy is arbitrage-free and every security is attain-
able if for every choice of numeraire there exists a unique equivalent
(martingale) measure. The security PV is then given by
" #
H(T, rt)
P V = N (0) E N (36)
N (T )

• A security is attainable if it can be replicated (recall stock options

replication with delta-hedging)

• The expectation E N is taken according to the equivalent measure

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 47

Change of Numeraire

How do we compute the expectation in (36)? E.g. using the following

theorem.

Given two numeraires M and N and a financial quantity G,

" #
M (T ) N (0)
E M [G] = E N G (37)
M (0) N (T )

• This result was first proved by Geman et al. in 1995

• It is incredibly useful in practice

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 48

Numeraire applications: payoff at maturity (1/2)

Consider the numeraire associated to a payoff of 1$ at time T (N (0) =

D(T ), N (T ) = 1) and E T the corresponding expectation, defining G
as
M (0)
G(T ) = H(T ) (38)
M (T )
the forward payoff corresponding to H, then
" #
M (T ) N (0)
P V = E M [G(T )] = E T G(T ) (39)
M (0) N (T )
so that ...

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 49

Numeraire applications: payoff at maturity (2/2)

.. we obtain
P V = D(T ) E T [H(T )] (40)
Notice how the stochastic rates disappeared.

Other examples (see Hull’s book)

• Choosing N as the period-compounded deposit yields the correct

measure for caplets and floorlets
• Choosing the fixed-rate annuity Af ixed as numeraire provides the
measure used to price swaptions

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 50

Questions?

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 51

Recall the Brownian motion

A continuous stochastic process Wt satisfying,

• W0=0
• Wt − Ws is independent form Ws (for 0 < s < t)
• Wt − Ws is normally distributed, precisely as N (0, t − s)

is a standard Brownian motion.

Complex stochastic processes can be built upon the Brownian motion

using stochastic differential equations.
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 52

Stochastic differential equations (1/3)

A stochastic differential equation for rt is defined as
drt = u(rt, t)dt + σ(rt, t)dWt (41)

• u is the generic drift (deterministic)

• σ is the generic volatility (deterministic)
• Wt is a Brownian motion (stochastic)

Every stochastic differential equation is really a shorthand for the

following stochastic (Ito’s) integral equation
Z t Z W
t
rt − r0 = u(rs, s)ds + σ(rs, s)dWs (42)
0 0
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 53

Stochastic differential equations (2/3)

We can integrate the stochastic integral equation over a short time

interval [t1, t2]. The drift term becomes

u(r1, t1)(t2 − t1) (43)

Since Wt is a Brownian motion the volatility term becomes

√
σ(r1, t1) (W1 − W2) = σ(r1, t1) t2 − t1 ε (44)
where ε is a Gaussian random number

ε ∼ N (0, 1) (45)

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 54

Stochastic differential equations (3/3)

In summary the stochastic differential equation for rt

drt = u(rt, t)dt + σ(rt, t)dWt (46)

can be interpreted, for a small interval [t, t + ∆t], as a simulation for

the future values of rt+∆t given the value rt

√
rt+∆t − rt = u(rt, t)∆t + σ(rt, t) ∆t εj (47)
with many samples εj ’s taken “Normally” randomly

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 55

Example: Martingales

In the particular case of processes where u = 0,

dXt = σ(Xt, t)dWt (48)

we have a Martingale.

Taking an expectation of any martingale yields the current value

E [Xt] = X0 (49)

Choosing Xt = Yt/Nt leads to the numeraire pricing equation

" #
Yt
Y0 = N 0 E (50)
Nt

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 56

Example: Hull–White SDE

Consider a Stochastic differential equation describing a financial model.

For example the Hull-White short-rate process defined by

drt = [θ(t) − a rt] dt + σ dWt (51)

Consider an option V so that the underlying short rate satisfies the
Hull-White model. Proceeding with no-arbitrage arguments and using
Ito’s lemma it can be shown that V satisfies
∂V ∂V σ 2 ∂ 2V
+ [θ(t) − a r] + 2
−rV = 0 (52)
∂t ∂r 2 ∂r
which is the analogous of the Black-Scholes equation for the Hull-
White model

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 57

Partial differential equations

For every model described by a stochastic differential equation, the

corresponding option price V satisfies a partial differential equation
∂V ∂V σ(r, t)2 ∂ 2V
− r V + u(r, t) + 2
=0 (53)
∂t ∂r 2 ∂r

• r is the source term

• u is the drift (a velocity field)
• σ is the volatility (the diffusion coefficients)

Equation (53) is a parabolic partial differential equation, with the ap-

propriate boundary and final conditions can be solved to given V (r, T )
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 58

Feynman-Kac formula

Richard Feynman (physicist) and Mark Kac (mathematician) demon-

strated a result that applied to the equation
∂V ∂V σ(r, t)2 ∂ 2V
− r V + u(r, t) + 2
=0 (54)
∂t ∂r 2 ∂r
with the final condition, defined by the payoff H

V (T, r) = H(r) (55)

Has a solution
 RT 
V (t, r) = E e− t r(s)ds
H(r) (56)

The Feynman-Kac formula is another derivation of the generic price

formula. Most numerical methods use this equation.
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 59

Questions?

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 60

Numerical methods for interest-rate models

Even the most sophisticated interest-rate model is useless if it cannot

produce at least a single number

Definition 1:
An interest-rate model is a mathematical tool that describes interest
rates in the financial markets.

Definition 2:
A numerical method is an algorithm that is applied to a model in
order to compute numerical values for the financial variables

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 61

Analytical formulas

Analytical formulas are historically the first numerical method used

to compute actual numbers from financial models

In few rare and exceptional cases it is possible to find analytical solu-

tions to derivatives models. Important examples are

• Merton formula for Black-Scholes equations

• Formulas for barrier and digital options priced in Black-Scholes
model
• Discount-factor and bond-option formulas for Hull-White model

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 62

Analytical approximations (1/2)

Even the most basic analytical solutions needs numerical methods to

be evaluated. For example the cumulative normal distribution
t2 dt
Z x
1 −
N (x) = √ e 2 (57)
2 π −∞
needs to be approximated. A simple approximation is given by
1 −x/2  2 3

N (x) ∼ 1 − √ e a1 k + a2 k + a3 k (58)
2π
with
1
k= (59)
1 + 0.33267 x
and a1 = 0.4361836, a2 = −0.1201676, and a3 = 0.937298
Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 63

Analytical approximations (2/2)

Most analytical formulas are obtained from the mathematical theory

of analytical functions, i.e. functions that can be obtained from a
locally-convergent power series

A fairly exhaustive collection of analytical formulas for option pricing

has been compiled by the collector ∗ and can be found in

• The Complete Guide to Option Pricing Formulas, Espen Gaarder

Haug, Mc Graw Hill (from first edition)

∗ See also the interesting picture collection

Advanced Derivatives, Interest Rate Models 2010–2012
c Marco Marchioro
 Exotic interest-rate options 64

Questions?

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro
 Exotic interest-rate options 65

References

• Options, future, & other derivatives, John C. Hull, Prentice Hall

(from fourth edition)

• Efficient methods for valuing interest rate derivatives, Antoon

Pelsser, Springer Finance

• Interest rate models: theory and practice, D. Brigo and F. Mer-

curio, Springer Finance (from first edition)

• The collector web site: http://www.espenhaug.com

Advanced Derivatives, Interest Rate Models 2010–2012

c Marco Marchioro